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
/
Epigenetic mechanisms driving bladder cancer
(USC Thesis Other)
Epigenetic mechanisms driving bladder cancer
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
EPIGENETIC MECHANISMS DRIVING BLADDER CANCER
by
Erika Michele Wolff
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)
August 2009
Copyright 2009 Erika Michele Wolff
ii
EPIGRAPH
“A man ceases to be a beginner in any given science and becomes a master in that
science when he has learned that he is going to be a beginner all his life.”
--Robin George Collingwood
iii
DEDICATION
This work I dedicate to my father who first inspired me to become a scientist, to both of
my grandfathers who lost their battles with cancer, and to the other members of my
family who have fought and won.
iv
ACKNOWLEDGEMENTS
I would like to thank the following people that made the work in this thesis possible:
Dr. Peter Jones, my mentor, who has shown me not just how to do science but,
more importantly, how to think like a scientist. You set an example for all of us by
coming in everyday with such passion, not to discover but to uncover our underlying
biology. Your words of wisdom will stay with me for the rest of my career. There is a
long-standing joke in the Ph.D. community that if you aren’t careful you tend to turn into
your thesis advisor. I can only hope to be so lucky.
Dr. Gangning Liang, my other mentor, without whom I could never have
completed the work in this thesis. Your patience and dedication to helping everyone
around you has truly been an inspiration. It has been my privilege to work with you
every day for the past five years. I can’t imagine working in a lab without you but you
have given me the tools and the confidence to do so.
Dr. Connie Cortez, my fellow graduate student, for her constant support and
friendship. Getting to spend the last five years with you was truly the greatest silver
lining earning a Ph.D. could ever have.
Dr. Terry Kelly, for being my senior postdoc even when you didn’t want to be! I
don’t know if I will ever be on such a similar wavelength, both scientifically and
personally, with someone else as I have been with you for the past year.
v
Dr. Gerda Egger, for being such an amazing role model. You always had a
levelheaded, yet passionate, approach to science. You were serious when you had to be
but also kept your sense of humor. You managed to balance both a career and a family. I
feel truly fortunate that I was able to learn so much from you.
Dr. Shinwu Jeong, for his amazing ability to have the answer to any experimental
question and an even more amazing ability to always be able to put a smile on my face.
Dr. Yvonne Tsai, for managing to run a lab full of people who didn’t always want
to be managed. I will be eternally grateful for how much you did for me.
My fellow graduate students Dr. Joy Lin, Dr. Christine Yoo, Flora Han, Shikhar
Sharma, Daniel Carvalho, and Sonia Escobar, who were all so much more than
colleagues. Postdocs Dr. Tina Miranda and Dr. Phillippa Oakford for all of your help and
for making the lab a truly amazing place to work everyday. My committee members, Drs.
Peter Laird, Gerry Coetzee, Judd Rice, and Darryl Shibata, for their support and feedback.
My scientific collaborators, Dr. Allen Yang, Dr. Hyang-Min Byun, and Dr. Yoshitomo
Chihara, my pathology collaborators, Dr. Peter Nichols and Moli Chen, and my
biostatistician collaborators, Dr. Susan Groshen and Dr. Kim Siegmund, for all of their
assistance.
vi
TABLE OF CONTENTS
Epigraph ii
Dedication iii
Acknowledgements iv
List of Tables viii
List of Figures ix
Abstract xii
Chapter 1: Mechanisms of Disease: Genetic and Epigenetic Alterations that
Drive Bladder Cancer 1
Introduction 1
Constitutive activation of the Ras-MAPK signal transduction pathway 4
Fibroblast growth factor receptor 3 6
Epidermal growth factor receptors 7
H-Ras 8
c-myc 9
E2F3 9
RASSF1A 10
DAPK 10
Inactivation of the p53 and Retinoblastoma pathways 11
p53 pathway 11
Retinoblastoma pathway 12
Chromosome 9 13
Inactivation of p53 and RB has a synergistic effect on
tumorigenesis 14
Epigenetic deregulation and genomic instability 16
Molecular pathways to bladder tumorigenesis 17
Chromosome 8 deletions 20
Conclusion 20
Overview of Thesis Research 22
Chapter 2: RUNX3 Methylation Reveals that Bladder Tumors are Older in
Patients with a History of Smoking 25
Introduction 25
Materials and Methods 28
Results 33
RUNX3 is frequently and specifically methylated in bladder
vii
tumors 33
RUNX3 methylation precedes methylation at the other eight
genes 38
RUNX3 methylation increases with age and a history of
smoking 38
Discussion 42
Conclusion 45
Chapter 3: An Epigenetic Field Defect is Present Across Bladders with
Cancer 47
Introduction 47
Materials and Methods 50
Results 54
Bladders with cancer have a widespread epigenetic field defect 54
The epigenetic field defect in bladders is not due to clonal
expansion 60
Non-invasive bladder tumors have a distinct pattern of
hypomethylation 65
Discussion 71
Conclusion 74
Chapter 4: Hypomethylation and Chromatin Remodeling of a LINE-1
Promoter Activates and Alternate Transcript of MET 76
Introduction 76
Materials and Methods 81
Results 82
Hypomethylation and expression of L1-MET is present across
bladders with cancer 82
L1-MET hypomethylation is a sensitive and specific marker for
premalignant bladder tissue 89
Fully unmethylated L1-MET promoters are found in
premalignant urothelium 91
L1-MET encodes truncated MET proteins 93
DNA methylation silences the L1-MET promoter 96
Chromatin remodeling accompanies transcriptional activation
of L1 promoters 98
A switch from a tetranucleosome to dinucleosome structure
accompanies transcriptional activation of the L1-MET promoter 98
The switch from a tetranucleosomal structure to a dinucleosomal
structure is a common event at L1 promoters 102
Discussion 108
Conclusion 112
viii
Chapter 5: Summary and Conclusions 114
Final Conclusion 118
References 121
ix
LIST OF TABLES
Table 1.1 Oncogenes and tumor-suppressor genes relevant to bladder 3
cancer.
Table 2.1 Clinicopathological characteristics of study population. 29
Table 2.2 Association of demographics with gene methylation (PMR) . 36
Table 2.3 Impact of sex, age, tumor stage, tumor grade, and smoking
history on RUNX3 methylation . 40
Table 4.1 Potentially active specific L1s and their associated alternate
transcripts. 78
x
LIST OF FIGURES
Figure 1.1 Integration of common genetic and epigenetic modifications in
bladder cancer into a common pathway. 5
Figure 1.2 Model for bladder cancer progression showing the molecular
pathways of tumorigenesis. 18
Figure 2.1 Methylation in bladder samples at nine different loci using
quantitative methylation-sensitive real-time PCR. 34
Figure 2.2 RUNX3 is methylated specifically in bladder tumors. 37
Figure 2.3 RUNX3 methylation increases with age and smoking history. 41
Figure 2.4 Model for bladder cancer progression showing the molecular
pathways of tumorigenesis. 46
Figure 3.1 DNA methylation assay scheme for GoldenGate. 51
Figure 3.2 Changes in DNA methylation occur in the corresponding normal
urothelium of bladders with cancer. 55
Figure 3.3 Methylation at MYOD in bladder tissue samples. 57
Figure 3.4 Methylation at CDH13 in bladder tissue samples. 58
Figure 3.5 Methylation at ZO2 in bladder tissue samples. 59
Figure 3.6 Methylation of ZO2 across the bladder. 61
Figure 3.7 ZO2 knockdown accelerates wound healing. 62
Figure 3.8 Silencing of ZO2 can be reversed by 5-aza-CdR treatment. 63
Figure 3.9 Two models of the development of bladder cancer. 64
Figure 3.10 X-inactivation patterns reveal that the urothelium of two female
patients with bladder cancer remains polyclonal. 66
Figure 3.11 Bladder tumors are hypermethylated and non-invasive tumors
have a unique hypomethylated pattern. 67
xi
Figure 3.12 Bladder tumors are hypermethylated mainly at CpG islands. 68
Figure 3.13 Non-invasive bladder tumors are hypomethylated mainly at
non CpG islands. 69
Figure 3.14 Illumina DNA methylation results were confirmed by
Pyrosequencing. 70
Figure 3.15 Model for bladder cancer progression showing the molecular
pathways of tumorigenesis. 75
Figure 4.1 Map of alternate transcripts from specific L1s. 83
Figure 4.2 Methylation and expression of L1-MET correlates in cell lines. 84
Figure 4.3 Methylation and expression of L1-ACVR1C correlates in cell
lines. 86
Figure 4.4 Methylation and expression of L1-RAB3IP correlates in cell
lines. 87
Figure 4.5 Methylation and expression status of specific L1s correlates in
bladder tissues. 88
Figure 4.6 ROC curves of specific L1s for bladder tissues. 90
Figure 4.7 Methylation of specific L1s across the bladder. 92
Figure 4.8 Bisulfite sequencing of L1-MET. 94
Figure 4.9 The truncated MET protein encoded by L1-MET. 95
Figure 4.10 DNA methylation silences the L1-MET promoter. 97
Figure 4.11 Chromatin remodeling occurs at active L1 promoters. 99
Figure 4.12 Nucleosome positioning in an active L1-MET promoter in
T24 cells. 101
Figure 4.13 Nucleosome positioning in an inactive L1-MET promoter in
LD419 cells. 103
Figure 4.14 Nucleosome positioning in an inactive L1-MET promoter in
HCT116 cells. 104
xii
Figure 4.15 Nucleosome positioning in an active L1-MET promoter in
HCT116 DKO cells. 105
Figure 4.16 Nucleosome eviction is a frequent occurrence at L1
promoters. 107
Figure 4.17 Model of the epigenetic alterations that occur between the
inactive L1-MET and active L1-MET. 111
Figure 4.18 Model for bladder cancer progression showing the molecular
pathways of tumorigenesis. 113
Figure 5.1 Model for bladder cancer progression showing the molecular
pathways of tumorigenesis. 120
xiii
ABSTRACT
Successes in the clinic have opened up the era of epigenetic therapy in which the
goal is to reactivate genes silenced inappropriately during carcinogenesis. The advantage
of using epigenetic therapy to target defects in DNA methylation is that, unlike mutations
in the DNA sequence, these alterations are reversible. The focus of this thesis was
twofold; to further elucidate the role of epigenetic alterations in bladder tumorigenesis
and determine whether they provide useful therapeutic targets for epigenetic therapy. An
ideal therapy for bladder cancer would address many of the unique aspects of bladder
cancer, such as preventing tumors in high-risk patients with a history of smoking, treating
both noninvasive and invasive tumors even though they develop via two separate
molecular pathways, and reducing the frequency of recurrences. Based on the findings in
this thesis epigenetic therapy, specifically DNA methylation inhibitors, has the potential
to address each of these concerns
Using aberrant DNA methylation at RUNX3 as a clock I have shown that bladder
tumors from smokers have undergone more cell divisions and may have initiated earlier
than tumors from nonsmokers. In addition, since RUNX3 methylation is present in early
lesions, epigenetic therapy may be useful in preventing high-risk patients from
developing tumors. Epigenetic therapy also has the potential for treating both
noninvasive and invasive bladder tumors since aberrant hypermethylation occurs at
numerous gene promoters in both tumor types. I have also revealed the presence of a
generalized epigenetic defect across bladders with cancer that is not due to clonal
xiv
expansion. These epigenetic defects involve hypermethylation of single copy genes and
also hypomethylation of specific LINE-1 elements. The hypomethylation of specific
LINE-1 elements activates alternate transcripts of genes across the bladder, including the
MET oncogene. The presence of so many epigenetic alterations in premalignant tissues
of the bladder indicates that treatment with epigenetic therapy may be beneficial not just
in the treatment of bladder tumors but also in the prevention of future recurrences.
.
1
CHAPTER 1
MECHANISMS OF DISEASE: GENETIC AND EPIGENETIC ALTERATIONS
THAT DRIVE BLADDER CANCER
INTRODUCTION
Bladder cancer was the fifth most commonly diagnosed cancer in the United
States in 2004 (after prostate, breast, lung, and colon cancers), with an estimated 60,240
new cases and 12,710 deaths (NCI). Risk factors for developing bladder cancer include
cigarette-smoking, exposure to arsenic, occupation in the rubber or fossil fuel industry,
and schistosomiasis, an inflammatory disease resulting from a parasitic infection
common in developing countries (Brandau and Bohle, 2001; Feng et al., 2002).
Approximately 80% of bladder cancers present as noninvasive papillary urothelial
carcinomas (UCs), 70% of which will recur, and 10-20% of which will progress and
invade the bladder muscle (Knowles, 2001). Of the patients initially presenting with
muscle-invasive UC, 50% will relapse with metastatic disease (Williams and Stein, 2004).
Around 5% of bladder tumors in the US are squamous-cell carcinomas (SCCs), but this
figure exceeds 75% in Egypt because of its association with chronic infections such as
schistosomiasis (Tsutsumi et al., 1998). Due to the low incidence of SCC in the US this
introductory chapter focuses on UC.
The difficulty in managing bladder cancer comes from the inability to predict
which tumors will recur or progress, necessitating frequent hospital visits for cystoscopy.
2
There is a need for noninvasive methods of diagnosis, prognosis, and monitoring for
recurrence and progression in bladder cancer. One such method includes analyzing DNA
isolated from the urine of bladder cancer patients. DNA sequencing for activating
mutations of the fibroblast growth factor receptor 3 (FGFR3) in urine complements
cystoscopy in monitoring patients with superficial and invasive bladder cancer (Rieger-
Christ et al., 2003). DNA from the tumors and urine of the same patient show similar
patterns of methylation, with hypermethylation of APC, p14, or RASSF1A detected in the
urine of almost all bladder cancer patients (Dulaimi et al., 2004). DNA methylation
levels of apoptosis-related genes, such as death-associated protein kinase (DAPK), are
increased in DNA from the urine of bladder cancer patients (Friedrich et al., 2004).
Most alterations found in noninvasive papillary UC can constitutively activate the
Ras mitogen-activated protein kinase (MAPK) pathway, an extracellular signal
transduction pathway involved in cellular proliferation. Additionally, there are two main
tumor suppressor pathways that appear to be inactivated in invasive bladder cancer, the
retinoblastoma (RB) and p53 pathways. Each of these three pathways can be altered in a
variety of genetic and epigenetic ways, and at a number of different sites, making it
difficult to draw conclusions based on monitoring only one type of alteration or one site
(Markl and Jones, 1998; Sarkar et al., 2000). The oncogenes and tumor-suppressor genes
relevant to bladder cancer, the way in which they are altered in this condition, and any
associations with tumor stage and grade are listed in Table 1.1. The main weakness in
bladder cancer studies stems from focusing on only one particular interest, be it an
oncogene, tumor-suppressor gene, single nucleotide polymorphism, or epigenetic
3
Table 1.1 Oncogenes and tumor-suppressor genes relevant to bladder cancer.
Genes Alteration Excludes Clinical association with
bladder tumors
Oncogenes
FGFR3 Activating mutation HRAS, TP53 Ta/low grade, low recurrence
EGFR/ERBB2 Overexpression None
HRAS Activating mutation,
overexpression
FGFR3 None
c-Myc Overexpression None
E2F Gene amplification,
overexpression
High stage/grade
Tumor-suppressor genes
RASSF1A Hypermethylation High stage
DAPK Hypermethylation High recurrence of Ta/T1
RB1 Deletion,
hyperphosphorylation
High stage/grade
TP53 Inactivating mutation,
deletion
FGFR3 T1 & high stage/grade
p14/p16 Homozygous deletion,
hypermethylation
High stage/grade
DBCCR1 Deletion,
hypermethylation
None
4
modification, and reporting its rate of occurrence. This introductory chapter synthesizes
the data gathered so far in the field of bladder cancer research into a model for bladder
carcinogenesis that involves the constitutive activation of the Ras-MAPK signal
transduction pathway and the inactivation of the p53 and pRb pathways. It illustrates the
necessity for all alterations in these pathways to be measured in the same tumor sample
populations in order to determine the optimal panel of markers for clinical use.
CONSTITUTIVE ACTIVATION OF THE RAS-MAPK SIGNAL
TRANSDUCTION PATHWAY
A remarkable convergence of anomalies in bladder cancers has led to the
realization that a single pathway may be disrupted at different levels. The Ras-MAPK
signal transduction pathway is constitutively activated in many different types of cancer
(Figure 1.1). The pathway tranduces extracellular growth signals into the nucleus of cells,
stimulating the transcription of certain genes, including c-Myc. Activation of the Ras-
MAPK pathway in tumor cells, through increased transcription of c-Myc, can deregulate
the cell cycle, leading to unchecked proliferation. The pathway is activated when growth
factors bind to their respective receptor tyrosine kinases, causing receptor dimerization
and autophosphorylation. The activated receptors recruit certain proteins that convert
Ras to its active state. Once active, Ras can transduce a signal through a variety of
different pathways including the MAPK pathway, which acts through MAPK/ERK
kinase and extracellular signal-regulated kinase (ERK). A downstream target of this
pathway is mitogen- activated and stress-activated protein kinase 1 (MSK1), a histone H3
5
Figure 1.1 Integration of common genetic and epigenetic modifications in bladder
cancer into a common pathway. Proteins encoded by oncogenes or with increased
expression are shown in green while proteins encoded by tumor-suppressor genes or with
decreased expression are shown in red. CDK4, cyclin-dependent kinase 4; c-Myc, Myc
proto-oncogene protein; DAPK, death-associated protein kinase; E2F, member of the
E2F family of transcription factors; EGF, epidermal growth factor; EGFR, epidermal
growth factor receptor; ERBB2, erythroblastic leukemia viral oncogene homolog 2; ERK,
extracellular signal-regulated kinase; FGF, fibroblast growth factor; FGFR3, fibroblast
growth factor receptor 3; Raf, RAF proto-oncogene serine/threonine-protein kinase; Ras,
GTPase H-Ras; MDM2, p53-binding protein Mdm2; MEK, MAPK/ERK kinase MSK1,
mitogen-activated and stress-activated protein kinase 1P, phosphorylated; RASSF1A, Ras
association (RalGDS/AF-6) domain family 1; RB, retinoblastoma-associated protein.
6
kinase involved in remodeling the structure of chromatin into a more relaxed and
transcriptionally accessible state (Dunn et al., 2005). The altered chromatin state induces
the c-myc gene (Hipfner and Cohen, 2004).
As listed in Table 1.1, the most common oncogenes and some of the tumor-
suppressor genes relevant to bladder cancer are components of this pathway, including
activating mutations of FGFR3 (Billerey et al., 2001), overexpression of two different
members of the epidermal growth factor receptor family (EGFR and ERBB2) (Nicholson
et al., 2001), overexpression or activating mutations of the Ras gene (Oxford and
Theodorescu, 2003), and overexpression of c-myc (Williams and Stein, 2004). A tumor-
suppressor gene, RASSF1A, is commonly silenced by methylation in bladder cancer and
is an inhibitor of the active form of Ras. Deletion of another tumor-suppressor gene,
TP53 (encoding p53), results in increased levels of H-Ras (Oxford and Theodorescu,
2003).
Fibroblast growth factor receptor 3
One of the most exciting discoveries in bladder tumorigenesis was the finding of
activating mutations in FGFR3 in the majority of papillary tumors (Cappellen et al.,
1999). FGFR3 is one of four known fibroblast growth factor (FGF) receptor tyrosine
kinases. Upon activation by one of 23 identified ligands, FGF receptors are involved in
signal pathways associated with angiogenesis, wound healing, and cellular differentiation.
When mutated in the germline, FGF receptor activation causes severe skeletal disorders.
Alterations in FGF receptors and the expression of their ligands occur in a number of
7
cancers, such as lung, prostate and breast cancer (Munro and Knowles, 2003). It has also
been discovered that FGFR3 and Ras gene family mutations are mutually exclusive,
suggesting that they affect the same pathways (Figure 1.1) (Jebar et al., 2005). FGFR3
mutations occur in approximately 68-88% of papillomas and noninvasive papillary UC
(stage Ta), possibly indicating that papillomas are a precursor of papillary urothelial
carcinomas (Billerey et al., 2001; van Rhijn et al., 2002). These mutations are also found
in 21% of UCs invading the lamina propria (stage T1), and 16% of muscle-invasive UCs
(stages T2-4), but are not found in carcinoma in situ (CIS) tumors, illustrating that
noninvasive papillary tumors do progress, although infrequently, and through a different
pathway than CIS (Billerey et al., 2001). Tumors with FGFR3 mutations have a lower
risk for recurrence than those without. This is the first marker linked to a positive
prognosis for bladder cancer (van Rhijn et al., 2002). FGFR3 mutations are not
associated with smoking and very few tumors have concurrent FGFR3 and TP53
mutations, with FGFR3 mutations found mostly in low stage (Ta and T1) and low grade
(1 and 2) tumors, and TP53 mutations in high stage (T2 to T4) and high grade (2 and 3)
tumors (Bakkar et al., 2003; Wallerand et al., 2005). FGFR3 mutations are detectable in
urine and are a good marker for monitoring recurrence of superficial papillary tumors
(Rieger-Christ et al., 2003).
Epidermal growth factor receptors
Overexpression of either of two members of the epidermal growth factor receptor
family, EGFR or ERBB2, has been associated with a poor prognosis in many cancers,
8
including bladder cancer (Nicholson et al., 2001). EGFR is a growth-factor-receptor
tyrosine kinase that activates several signal transduction pathways, including those that
contribute to cell growth and differentiation like Ras-MAPK (Figure 1.1). A high level
of EGFR expression in mice causes hyperplasia and does not have a synergistic effect
with Ras activation, indicating that they work through the same pathway. When EGFR
overexpression is combined with interruption of the RB and p53 pathways, CIS tumors
progress to high grade, but do not become invasive, suggesting that high levels of EGFR
promote tumor growth but not progression (Cheng et al., 2002).
H-Ras
H-Ras is a member of the Ras superfamily, which consists of more than 100
monomeric G proteins. The role of the Ras superfamily is to transduce signals from the
cellular membrane to the nucleus (Figure 1.1). Altered H-Ras contributes to progression
of bladder cancer either through activating mutations of HRAS or aberrantly high
expression of the protein. The frequency of activating mutations in HRAS is a matter of
some debate; however, it is generally accepted that they occur in an estimated 10 to 30%
of bladder tumors.
Increased gene expression of HRAS has been found in CIS and high grade (2 and
3) tumors, and allelic loss of p53, which is common in these tumor types, might
contribute to its upregulation (Oxford and Theodorescu, 2003). Mice overexpressing H-
Ras develop urothelial hyperplasia and eventually superficial papillary tumors (Zhang et
al., 2001). When further characterized, it was found that these Ras-induced hyperplasias
9
and superficial papillary tumors had lost expression of RB, indicating that Ras
overexpression may be compensated for by the RB pathway and its loss is necessary for
Ras-induced tumorigenesis (Garcia-Espana et al., 2005). These tumors, however, do not
progress unless there is also allelic TP53 loss (Gao et al., 2004).
c-myc
The Myc family is a group of transcription factors that control cell growth and
regulate the cell-cycle. c-myc is overexpressed in several types of cancer, including
bladder, although the exact mechanism of overexpression in bladder cancer is unknown.
Studies addressing the association of Myc overexpression with cancer stage or prognosis
have yielded conflicting results (Williams and Stein, 2004). Myc proteins promote
expression of cyclin D and cyclin-dependent kinase 4 (CDK4), which form a complex
that phosphorylates RB (Figure 1.1). Upon phosphorylation, RB releases the
transcription factor E2F and cellular proliferation is induced through cell-cycle
progression to S phase. Additionally, Myc downregulates inhibitors of CDK, including
p21, increasing the levels of phosphorylated RB (Hipfner and Cohen, 2004).
E2F3
The E2F transcription factor family consists of seven E2F genes, most of which
interact with RB and control cell division (Dimova and Dyson, 2005). Gene
amplification, and subsequently higher levels of E2F3 expression, has been found in
grade 2 and 3 bladder tumors. Approximately 14% of muscle-invasive (stage T2-4)
10
bladder cancers have gene amplification of E2F3, making this one of the most commonly
amplified bladder oncogenes, which induces cellular proliferation in tumor cells
(Oeggerli et al., 2004).
RASSF1A
RASSF1A is a tumor suppressor gene that encodes a protein with a Ras
association domain whose function may be to inhibit the active form of Ras, Ras-GTP
(Figure 1.1) (Oxford and Theodorescu, 2003). It is commonly methylated in bladder
cancer and can be detected in DNA from urine (Dulaimi et al., 2004), with increased
levels of methylation associated with higher tumor stage (Friedrich et al., 2004).
DAPK
Death-associated protein kinase (DAPK) is involved in promoting apoptosis. It
interacts with ERK in the Ras-MAPK pathway and prevents the translocation of ERK
from the cytoplasm to the nucleus (Figure 1.1) (Chen et al., 2005). Hypermethylation of
DAPK, which silences gene expression, is found in 29% of superficial bladder cancers,
and is associated with a high recurrence rate (Tada et al., 2002). An increased level of
methylation at the DAPK promoter region was detected in urine from bladder cancer
patients (Friedrich et al., 2004).
11
INACTIVATION OF THE P53 AND RETINOBLASTOMA PATHWAYS
p53 pathway
The most common genetic change in human tumors is mutated TP53 (Williams
and Stein, 2004). The role of p53 in cells is to coordinate the detection of DNA damage
and cell-cycle arrest, stop the damage can be repaired. If the damage cannot be repaired,
then p53 signals the cell to undergo apoptosis. Activated p53 halts the cell cycle by
upregulating p21, an inhibitor of CDKs, resulting in hypophosphorylation of RB and
inhibition of cell-cycle progression (Figure 1.1) (Stein et al., 1998). p53 is targeted for
degradation by MDM2, and p14 prevents p53 degradation by sequestering MDM2
(Eymin et al., 2001). p53, MDM2, and p14 form an autoregulatory feedback loop, with
p53 downregulating p14 expression (Robertson and Jones, 1998). Mutations in p53 are
found at a high rate in dysplasia and CIS, 67% and 72%, respectively (Hartmann et al.,
2002). Allelic loss of TP53 in mice seems to be an early event in the development of CIS
(Cheng et al., 2003).
Mutations in TP53 are associated with smoking. Patients that smoke tend to have
tumors that are of a higher grade and stage than nonsmokers or ex-smokers (Wallerand et
al., 2005). The bladder carcinogen 4-aminobiphenyl (ABP) found in cigarette smoke is
metabolized and transported to the bladder where ABP-DNA adducts form in the
urothelial cells. Bladder epithelial cells with mutated TP53 are less able to repair these
adducts (Swaminathan et al., 2002). Additionally, the p53 mutational spectrum that is
specific to bladder cancer correlates with the 4-ABP binding spectrum in the p53 gene
12
(Feng et al., 2002). Many reports have shown a distinct mutational spectrum of p53 in
smokers compared with nonsmokers, although the exact nature of each spectrum is
controversial, with the role of 4-ABP remaining unclear (Spruck et al., 1993; Wallerand
et al., 2005).
Retinoblastoma pathway
The RB protein is involved in many different cellular functions. RB helps to
control the expression of genes involved in cellular proliferation, differentiation, and
apoptosis, through its interactions with chromatin and DNA-modifying enzymes and
transcription factors (Robertson et al., 2000). In normal, non-dividing urothelial cells,
RB is hypophosphorylated and can interact with members of the E2F transcription factor
family, thus preventing their binding to DNA to promote transcription of genes involved
in proliferation (Garcia-Espana et al., 2005). In dividing cells, RB is
hyperphosphorylated by cyclin-CDK complexes, releasing E2F and allowing cellular
growth (Chatterjee et al., 2004). Any perturbation in RB expression, be it loss of
expression or an increase in phosphorylation, can result in uncontrolled cell growth.
Increased levels of cyclin D or cyclin E, or decreased levels of CDK inhibitors including
p21, p27, and p16, can cause hyperphosphorylation of RB (Figure 1.1) (Chatterjee et al.,
2004).
Patients with tumors that have lost RB expression have an equally poor prognosis
as those whose tumors have aberrantly high expression of RB (Cote et al., 1998). In
bladder tumors expressing high amounts of RB, the protein is hyperphosphorylated,
13
through elevated cyclin D1 expression and/or loss of p16 expression, leading to
functional inactivation of RB (Chatterjee et al., 2004). RB, through its interaction with
transcription factor AP-2, is essential in maintaining the epithelial phenotype by
activating expression of E-cadherin. Loss of expression of RB is, therefore, linked to the
dedifferentiation process that occurs during tumorigenesis (Batsche et al., 1998). The
expression of p21 is also controlled by RB through Sp1 and Sp3 transcription factors;
loss of RB leads to reduced expression of p21 (Decesse et al., 2001). Loss of p21
correlates with a higher rate of recurrence and decreased survival (Stein et al., 1998).
Loss of heterozygosity (LOH) at RB1 increases with increasing stage, from 19% in
noninvasive papillary tumors to approximately 60% in muscle-invasive tumors (Wada et
al., 2000).
Chromosome 9
Deletions of part or all of chromosome 9 are a common event in bladder cancer
and occur early on in tumorigenesis (Hartmann et al., 2002; Orlow et al., 1999; Tsai et al.,
1990). Urothelial hyperplasias are a likely precursor to papillary urothelial carcinoma, as
they both have similar spectrums of deletions of chromosome 9 (Chow et al., 2000;
Hartmann et al., 1999). Finding a deletion in a specific region of a chromosome in tumor
cells usually indicates the location of a tumor-suppressor gene. One such locus on the 9q
arm, known as deleted in bladder cancer gene 1 (DBC1), is commonly deleted. A novel
gene in this region at 9q32-33, deleted in bladder cancer candidate region 1 (DBCCR1), is
postulated to be a tumor-suppressor gene, on the basis of the high rate of
14
hypermethylation at this locus in bladder cancer (Habuchi et al., 1998). When the
DBCCR1 protein is expressed in bladder cancer, growth is inhibited and cell death is
induced (Nishiyama et al., 2001; Wright et al., 2004). Homozygous deletion of the 9p21
region at the INK4A/ARF locus, which encodes both p16 and p14 respectively, occurs in
approximately 14% of superficial bladder tumors and deregulates both the RB and p53
pathways (Figure 1.1). In superficial papillary UCs, those with a homozygous deletion at
INK4A/ARF locus have a higher grade, are larger in size and have a higher rate of
recurrence than superficial papillary UCs that do not have the deletion (Orlow et al.,
1999). In addition to LOH at this locus, the promoters of p14 and p16 can be methylated
in bladder cancer (Dulaimi et al., 2004; Gonzalez-Zulueta et al., 1995). A similar
spectrum of chromosome 9 deletions occurs in dysplasia and CIS, indicating that
dysplasias are precursors to CIS (Hartmann et al., 2002).
Inactivation of p53 and RB has a synergistic effect on tumorigenesis
In superficial bladder tumors TP53 mutations are extremely rare (Bakkar et al.,
2003). Most alterations found in these tumors involve activation of the Ras-MAPK
pathway and inactivation of the Rb pathway. Homozygous deletion or silencing of the
INK4A locus that encodes for p14 and p16, and alters both the RB and p53 pathways, is
associated with a greater risk for recurrence (Figure 1.1) (Orlow et al., 1999). p14 also
has a role in the RB pathway, forming a complex with MDM2 to bind E2F and inhibit
cellular proliferation (Eymin et al., 2001).
15
Inactivation of both the p53 and RB pathways seems to be necessary for the
development of invasive bladder cancer. Inactivation of these pathways can result from
combinations of different events (Markl and Jones, 1998; Sarkar et al., 2000). A study of
19 bladder carcinoma cell lines revealed that they all had an altered p53 pathway, either
through a mutation of TP53 or a homozygous deletion of p14, and an altered RB pathway,
either through loss of the RB protein or hyperphosphorylation of the RB protein due to
hypermethylation or homozygous deletion of p16 (Markl and Jones, 1998). In another
study 12 bladder tumor samples examined had altered p53 and RB pathways, through
different genetic and epigenetic combinations, yielding four groups: -RB/-p53, -p16/-p53,
-p16/-p21, or –p16/-p14, where the loss of protein activity results from homozygous
deletion, gene mutation, or protein inactivation through an unidentified mechanism
(Sarkar et al., 2000). No single inactivating event has the same effect on the phenotype
of the cell, and, therefore, the prognosis for each patient will be different.
Cellular senescence can be bypassed either through the loss of p16 or RB, but
only the loss of RB eliminates the cell’s ability to respond to p53-mediated growth arrest,
and is associated with higher tumor stage and risk of progression. The p53 pathway is
mainly inactivated either through loss of p21, p14, or p53. Cells with loss of p14 and 16
can still respond to DNA damage by activating p53 and arresting the cell cycle (Sarkar et
al., 2000). Only 12% of patients with tumors characterized by the loss of p21 and altered
p53 expression remain recurrence-free after 5 years, compared with 64% of patients with
p21-positive tumors and either wild-type or altered p53 (Stein et al., 1998). The
expression of p21 is lost in the absence of RB, meaning that the p21-negative tumors
16
probably have no RB expression, while the positive tumors retain RB but have an altered
RB pathway through the loss of p16.
Epigenetic deregulation and genomic instability
In addition to the genetic abnormalities associated with cancer, alterations in
epigenetic patterns occur. Almost all cells in the human body contain the same DNA, i.e.
are genetically identical, but do not express the same genes and are, therefore,
epigenetically different. The differentiation of cells into various types is due to the
expression of a tissue-specific subset of genes. This expression is controlled by DNA
methylation, chromatin structure, and transcription factors. DNA can be methylated at
cytosine residues adjacent to guanine residues (CpG), and the distribution of CpG sites
throughout the genome is nonrandom, with islands of CpG sites occurring in the
promoter and exonic regions of genes. The pattern of methylation of cytosines in DNA is
significantly altered in cancer, with sites that are normally hypermethylated, such as
repetitive regions, becoming hypomethylated and inactivation of tumor-suppressor genes
by promoter hypermethylation (Jones and Baylin, 2002). Increased methylation of CpG
islands is associated with higher tumor stage (Salem et al., 2000a). CpG methylation
occurs through DNA methyltransferases (DNMTs), the most studied of which is DNMT1.
Expression of DNMT1 is increased in tumors. DNMT1 expression is repressed by
inactive p53, and when p53 is deleted or activated, DNMT1 expression increases by five-
fold to six-fold (Peterson et al., 2003). Also, DNMT1 has an E2F binding site linking it
to the RB pathway. When cells lose RB expression, DNMT1 becomes aberrantly
17
expressed during the cell cycle and hypermethylation of tumor-suppressor genes can
occur (McCabe et al., 2005). Transformation with certain oncogenes, such as HRAS,
alters the methylation of some genes (Ordway et al., 2004).
Changes in methylation status of genes involved in apoptosis can be detected
noninvasively through DNA in urine (Friedrich et al., 2004). Hypomethylation of
repetitive satellite regions of chromosome 9 is associated with LOH on chromosome 9
and increases with higher tumor stage and grade (Nakagawa et al., 2005). Loss of RB
causes genomic hypomethylation and altered pericentric and telomeric chromatin
structure, as it interacts with certain histone methyltransferases. RB helps to maintain
chromatin structure, and loss of functional activity may contribute to genomic instability
(Gonzalo et al., 2005). Mutations of TP53 in combination with cyclin E overexpression,
which inactivates RB, induce centrosome amplification and chromosome instability in
bladder cancer (Kawamura et al., 2004).
MOLECULAR PATHWAYS TO BLADDER TUMORIGENESIS
There is substantial evidence for the existence of mutually exclusive molecular
pathways of tumorigenesis in the formation of papillary and invasive carcinomas, a
model of which is shown in Figure 1.2 (Spruck et al., 1994). The most common genetic
alterations in low-grade (grade 1) papillary UC are LOH of part or all of chromosome 9
and activating mutations of FGFR3 (Billerey et al., 2001). Only one study has examined
LOH of chromosome 9 and FGFR3 mutations concurrently, revealing that out of 10
patients with LOH of chromosome 9, three also had FGFR3
18
Figure 1.2 Model for bladder cancer progression showing the molecular pathways of
tumorigenesis. One pathway involves the development of noninvasive papillary tumors
through the accumulation of activating mutations of FGFR3 (fibroblast growth factor
receptor 3) and/or loss of heterozygosity of either the 9p or 9q regions of chromosome 9
to give rise to low grade tumors, while homozygous deletion of the INK4A (p16) locus is
involved in high grade non-invasive tumor formation. Another pathway involves the
development of carcinoma in situ (CIS) through dysplasia, by the loss of heterozygosity
of either 9p or 9q preceded by mutations in TP53 (tumor protein p53). TP53 mutations
can occur in either exons 1-4 (orange) or exons 5-11 (pink). CIS progresses either
through loss of RB (retinoblastoma-associated protein) expression or homozygous
deletion of the INK4A locus depending on the region in which p53 is mutated (exons 5-11
or 1-4, respectively). The thickness of the arrow represents the frequency of occurrence.
Increasing genomic instability and aberrant methylation occur as a function of increasing
invasion and progression. Stage T2 and carcinoma in situ tumors express extracellular
matrix (ECM) remodeling genes and have deletion of the p region of chromosome 8.
CIS
Ta
Low Grade
Non-Invasive
Papillary
Carcinoma
in situ
Flat
Metastasis M
Muscle
Invasive
Dysplasia
Lamina Propria
Invasive
Lamina Propria
Invasive
T
1
8p-
p53
INK4a
INK4a
Rb
?
9-
High Grade
Non-Invasive
Papillary
Normal
Urothelium
T
1
T
2
-T
4
?
9-
9-
ECM
Remodeling
Genes
ECM
Remodeling
Genes
FGFR3
19
mutations (van Tilborg et al., 2000). In high-grade noninvasive papillary tumors
homozygous deletion of the INK4A locus on chromosome 9 correlates with an increased
risk of recurrence and worse prognosis compared to tumors without this deletion (Orlow
et al., 1999). The pathway of formation of invasive UC seems to start with the formation
of dysplasia, continues with progression to carcinoma in situ (CIS), and then invasion of
the lamina propria (Hartmann et al., 2002). The most frequent genetic alteration in
dysplasia and CIS is mutation of TP53 followed by LOH of chromosome 9 (Hopman et
al., 2002). Depending on the location of the TP53 mutation, progression from CIS to
stage T1 tumors may involve either loss of the RB protein or homozygous deletion of
INK4A (Markl and Jones, 1998). A marker for progression in UC is loss of chromosome
8p, which occurs in approximately 60% of tumors (Stoehr et al., 2004). Global trends of
increased genomic instability and aberrant methylation correlate with increased invasion
and progression of tumors (Salem et al., 2000a).
In addition to genetic studies, gene expression profiling with microarrays has
revealed significantly different patterns of expression in superficial UC compared to CIS,
with both stage T2 and high grade papillary UC with surrounding CIS expressing matrix
remodeling genes (Dyrskjot et al., 2004). The ability of microarrays to delineate between
stages of tumors indicates that there are distinct and reproducible changes that occur in
bladder tumor progression. Also, not only can superficial papillary (Ta) tumors, T1
tumors, and muscle invasive tumors (T2-T4) be distinguished from each other, but there
are subclasses of Ta tumors, and those that frequently recur have a unique gene-
expression pattern. Tumors with the worst prognosis, including high grade Ta tumors
20
with surrounding CIS and T2 tumors, expressed a group of genes involved in matrix
remodeling (Dyrskjot et al., 2003).
CHROMOSOME 8 DELETIONS
Deletions of chromosome 8p are associated with progression in papillary UC,
with several candidate tumor-suppressor genes proposed (Stoehr et al., 2004). The high
frequency of 8p deletions in tumors that have invaded the muscle, regardless of the
identity of any tumor-suppressor genes, may provide a valuable marker for progression
(Knowles, 1999).
CONCLUSION
In the general model for bladder tumorigenesis, normal urothelium is exposed to
endogenous or exogenous (i.e., carcinogens from cigarette smoke) factors that cause
DNA damage, inducing the overexpression of oncogenes. Once DNA replication
becomes deregulated, DNA damage continues to accumulate, until double stranded DNA
breaks occur at fragile sites of the chromosomes. The DNA damage response is induced
and cell growth is arrested until the damage can be repaired or, if the damage is too
extensive, the cells undergo apoptosis. In this state cells that can proliferate and evade
apoptosis are selected for and continue to grow, with increasing genomic instability and
continued selection for avoidance of senescence (Bartkova et al., 2005). Thus, any
acquisition of mutations that increase the ability of cells to proliferate increases the
likelihood that additional mutations will occur that can inactivate cellular growth
21
checkpoints and the apoptotic pathway. One of the dangers that arise from smoking
cigarettes is the high rate of mutations that occur in TP53, preventing the arrest of cell
growth and apoptosis.
It is surprising that most of the genetic and epigenetic changes common in bladder
cancer can activate a single pathway, that involving Ras-MAPK, p53, and RB (Figure
1.1). The model of the molecular mechanisms behind bladder cancer presented in this
introductory chapter (Figure 1.2) is a simplified version of the effect of specific
oncogenes and tumor-suppressor genes in bladder cancer, but also gives an insight into
the interrelated mechanisms behind bladder carcinogenesis. Using this model it is
possible to approach the development of a panel of markers for use in the diagnosis and
monitoring of bladder cancer in a logical way. Studies of a tumor population and
measurement of the many different genetic and epigenetic changes reviewed in this
introductory chapter would provide valuable information of clinical utility. This would
help to form a full picture of how these changes interact with each other and what their
affect on recurrence and prognosis might be.
22
OVERVIEW OF THESIS RESEARCH
As summarized already, much is known about the genetics of bladder
cancer. However the molecular basis behind the high rate of recurrence and what role
epigenetics may play remains unclear. The epigenome consists of several layers of
heritable transcriptional regulation imposed upon the genome, including DNA
methylation, histone modifications, and nucleosome positioning (Jones and Baylin, 2007).
DNA becomes methylated at CpG dinucleotides and is associated with transcriptional
repression when it occurs in CpG rich regions (CpG islands) located in gene promoters
(Miranda and Jones, 2007). During carcinogenesis, CpG islands become aberrantly
hypermethylated (Jones and Baylin, 2002).
Exposure to tobacco smoke is associated with increased DNA methylation at
certain genes in both lung and bladder tumors. In Chapter 2 I identified interactions in
bladder cancer between DNA methylation and a history of smoking, along with the effect
of aging. The prevalence of methylation at RUNX3 increased as a function of age at
diagnosis and a history of smoking. Using RUNX3 methylation as a molecular clock in
order to determine the age of a bladder tumor revealed that tumors from smokers are
“older” than tumors from nonsmokers either due to tumors in smokers initiating earlier or
undergoing more rapid cell divisions.
In Chapter 3 I conducted a large-scale analysis of DNA methylation in tissues
from cancer-free bladders, “normal” tissues from bladders with cancer, and bladder
tumors and found the presence of an epigenetic field defect in bladders with cancer. In
addition, I also demonstrated that the altered urothelium in bladders with cancer is not the
23
result of clonal expansion but instead occurs as a general epigenetic defect, likely due to
the fact that the urothelium of the bladder is uniformly exposed to any mitogens or
carcinogens present in the urine. These widespread epigenetic alterations may indicate
that by the time bladders have developed tumors no “normal” urothelium remains.
Bladder cancer is one of the most highly recurrent tumor types, necessitating
frequent and invasive monitoring. In Chapter 4 I revealed that hypomethylation of the
LINE-1 located within the MET oncogene activates an alternate MET transcript across the
urothelium of tumor-bearing bladders. Therefore surgical excision of the tumor would
leave behind large areas of the bladder that remain epigenetically altered and express a
potential oncogene. Global hypomethylation of repetitive elements may activate
numerous alternate transcripts of genes during tumorigenesis. In addition I also
demonstrated that in addition to hypomethylation, active L1 promoters undergo dramatic
chromatin remodeling resulting in a switch from a transcriptionally inactive
tetranucleosomal structure to a transcriptionally active dinucleosomal structure.
Taken together this work indicates that DNA methylation plays an important role
in the development of bladder cancer. At the conclusion of each chapter of this thesis I
added my findings into the model for bladder tumorigenesis presented in Figure 1.2. At
the end of Chapter 5 I have a final model incorporating all of my data. The epigenetic
defects I have uncovered involve hypermethylation of single copy genes and also
hypomethylation of specific LINE-1 elements. The hypomethylation of specific LINE-1
elements activates alternate transcripts of genes across the bladder, including the MET
oncogene. The presence of so many epigenetic alterations in premalignant tissues of the
24
bladder may indicate that treatment with epigenetic therapies would be beneficial not just
in the treatment of bladder tumors but also in the prevention of future recurrences.
25
CHAPTER 2
RUNX3 METHYLATION REVEALS THAT BLADDER TUMORS ARE OLDER
IN PATIENTS WITH A HISTORY OF SMOKING
INTRODUCTION
Cancer etiology involves interactions between the environment, the genome, and
the epigenome (Feinberg, 2004; Herceg, 2007). The epigenome consists of several layers
of heritable transcriptional regulation imposed upon the genome, including DNA
methylation, histone modifications, and nucleosome positioning (Jones and Baylin, 2007).
DNA becomes methylated at CpG dinucleotides and is associated with transcriptional
repression when it occurs in CpG rich regions (CpG islands) located in gene promoters
(Miranda and Jones, 2007). During carcinogenesis, CpG islands become aberrantly
hypermethylated and this hypermethylation increases with cell division (Jones and Baylin,
2002; Velicescu et al., 2002). Alterations in the epigenome accumulate during the aging
process (Fraga et al., 2005). For instance, normal epithelium in the colon acquires
methylation in an age-dependent manner at genes that become hypermethylated in colon
cancer (Ahuja et al., 1998). Therefore, it has been proposed that DNA methylation
patterns constitute a molecular clock and can be used to determine the “age” of normal
tissues, i.e., the number of times the cells have divided (Kim et al., 2005a). It may also
be possible to use methylation as a molecular clock to determine the age of tumor tissues
by using a locus that is specifically methylated in tumor cells and does not acquire
26
methylation in normal tissues as a function of aging. Information about the age of a
tumor may give insight into when tumorigenesis initiates, which has important
implications for early detection.
Numerous types of environmental exposures, such as tobacco smoke, arsenic,
cadmium, and nickel, are associated with aberrant DNA methylation (Herceg, 2007).
Specifically, tobacco-derived carcinogens are associated with DNA methylation at p16
(Kim et al., 2001), CYP1A1 (Anttila et al., 2003), and RASSF1A (Kim et al., 2003) in
lung tumors and p16 in bladder tumors (Marsit et al., 2006). Marsit et al. (Marsit et al.,
2007) recently measured promoter hypermethylation of 16 different genes in bladder
tumors and found a correlation between overall methylation and age, gender, and
smoking history. However, their analysis combined the methylation at all the genes into
one measure based on the premise that hypermethylation of promoters is not a targeted
event. While the precise mechanism resulting in de novo methylation of specific CpG
islands in cancer is currently unknown, recent work has demonstrated that genes targeted
by polycomb complexes in embryonic stem (ES) cells are more likely than other genes to
undergo promoter hypermethylation during carcinogenesis (Ohm et al., 2007; Schlesinger
et al., 2007; Widschwendter et al., 2007).
Bladder cancer is the fifth most commonly diagnosed cancer in the United States,
where the majority of tumors are urothelial carcinoma (UC) (Wolff et al., 2005). Long-
term cigarette smokers are 2.5 times more likely to develop bladder cancer than
nonsmokers (Castelao et al., 2001). In this study, I took advantage of a large number of
UC samples from patients with known age, gender, and exposure to tobacco smoke in
27
order to address the possibility that these factors are associated with epigenetic alterations.
I utilized the real-time based methylation sensitive PCR assay MethyLight, developed by
Eads et al. (Eads et al., 2000a), to specifically and sensitively detect DNA methylation at
nine different genes (RUNX3, BCL2, PTGS2 [COX2], DAPK, CDH1 [ECAD], EDNRB,
RASSF1A, TERT, and TIMP3) in bladder tumor DNA samples from 342 patients.
RUNX3 is a tumor suppressor gene involved in apoptosis and is both frequently silenced
by methylation in bladder cancer (Kim et al., 2005b; Suzuki et al., 2005) and a confirmed
polycomb target (Fujii et al., 2008; Ringrose, 2007; Schwartz et al., 2006). RUNX3
methylation is associated with bladder tumor grade, invasiveness (Suzuki et al., 2005),
stage, recurrence, and progression (Kim et al., 2005b). My results revealed interactions
between environmental exposure, aging, and the epigenome. In particular, I have shown
for the first time that DNA methylation, in addition to being used as a molecular clock to
determine the age of normal tissues, can determine the age of tumors and the RUNX3
methylation data suggests that bladder tumors from smokers are older than from
nonsmokers. Therefore, bladder tumorigenesis may initiate early in a smoker’s life,
while tumors from nonsmokers initiate at a later age.
28
MATERIALS AND METHODS
Study population. From 1987 to 1996 a population-based case-control study of bladder
cancer was conducted in Los Angeles County as described previously (Castelao et al.,
2001). Eligibility criteria for cases included histologically confirmed urothelial
carcinoma diagnosed between January 1, 1987 and April 30, 1996 among non-Asian
patients aged 25 to 65 years. In total 2,098 cases were identified through the Los Angeles
County Cancer Surveillance Program, and of these 1,582 patients were included as part
of a case-control pair.
Tissue collection. Hospitals and pathology laboratories provided tumor tissue blocks to
the Los Angeles County SEER Registry Slide Retrieval Program, a component of the
Tissue Procurement Core Resource of the USC/Norris Comprehensive Cancer Center.
Of the specimens retrieved, 342 cases had sufficient tumor available to permit the
analysis of DNA methylation (Table 2.1). Frozen blocks of matched bladder tumor and
corresponding mucosa have been described previously (Byun et al., 2007) and urothelium
from cancer-free bladders was obtained from age-matched patients undergoing
prostatectomies. All study subjects had signed informed consent forms approved by the
Human Subjects Committee at the University of Southern California Keck School of
Medicine.
29
Table 2.1 Clinicopathological characteristics of study population
Variables N %
No. of patients 342
TNM Stage
Ta 154 45%
CIS 9 3%
I 93 27%
II 54 16%
III 18 5%
IV 13 4%
Grade
1 109 32%
2 134 39%
3 87 25%
4 11 3%
Sex
Male 259 76%
Female 83 24%
Age at diagnosis
≤40 22 6%
41-50 60 18%
51-60 175 51%
61-65 84 25%
Median (range) 58 (25-65)
Smoking History
Never/Irregular 45 13%
Former 137 40%
Current 159 46%
Duration of Smoking (years)
0 46 13%
1 to 15 36 11%
16 to 29 79 23%
30 to 39 95 28%
≥ 40 86 25%
Smoking in Pack-Years
0 46 13%
1 to 30 103 30%
31 to 60 105 31%
61 to 85 53 15%
≥ 86 35 10%
30
DNA isolation and sodium bisulfite modification. DNA extraction from frozen tissues
has been described previously (Byun et al., 2007). Microdissection was performed on
one to three consecutive 5-micron sections (H&E stained) composed mainly of tumor
tissue from 342 UC patients. DNA was extracted and bisulfite converted as previously
described (Friedrich et al., 2004; Widschwendter et al., 2004).
Quantitative methylation-sensitive real-time PCR. Methylation analysis was
performed as described previously (Friedrich et al., 2004). Briefly, the proportion of
bisulfite-converted DNA in each sample was controlled for by a collagen IIA (COL2A1)
reaction, which only amplifies bisulfite-converted DNA and is located in a region with no
CpG sites in order to be independent of the methylation status. TNFRSF25 (DR3) was
used as a positive control and FADD as a negative control for methylation. Genomic
DNA (Promega, Madison, WI) was treated with SssI DNA methyltransferase (New
England Biolabs) and used as a fully methylated reference to which all samples were
compared to yield the percentage of fully methylated DNA (PMR) (Friedrich et al., 2004).
Primer sequences and locations have been previously described for BCL2, DAPK,
EDNRB, FADD, RASSF1A, TERT, TNFRSF25 (DR3), COL2A1 (Friedrich et al., 2004);
CDH1 (ECAD) (Eads et al., 2000b); PTGS2 (COX2), and TIMP3 (Eads et al., 2001).
RUNX3 (NM_004350, -326/-258) forward primer, probe, and reverse primer sequences
were as follows: 5′-CGTTTTAGCGTTAGGGAGTTACG’3′, 6FAM5′-
TTTGAGAGAGGGCGGTAAGGGCG-3′BHQ1, 5′-AACGTCCGAATCCCACGA-3′.
31
Quantification of DNA methylation by methylation-sensitive single nucleotide
primer extension. Quantification of methylation at specific CpG sites using
methylation-sensitive single nucleotide primer extension, a method developed in our lab,
has been described previously (Gonzalgo and Liang, 2007). The promoter region of
RUNX3 was amplified with primers specific for bisulfite-converted DNA and with an
annealing temperature of 58ºC (forward 5′-GGGGTTGTAGAAGTTATAGGT-3′ and
reverse 5′-CCAATACCACAACCCAAAAC-3′) and two specific CpG sites were assayed
using SNuPE primers with an annealing temperature of 56ºC (5′-
GGGGTTGTAGAAGTTATAGGTT-3′ and 5′-TAGTAAGAGTTGGGGAAGTT-3′).
RUNX3 methylation was calculated as the average percent methylation of two CpG sites.
Statistical analysis. For each gene (BCL2, COX2, DAPK, ECAD, EDNRB, RASSF1A,
RUNX3, TERT, and TIMP3) the methylation, measured as the Percent of the fully
Methylated Reference (PMR) (Weisenberger et al., 2006), was scored two ways: as the
rank among study samples of all PMR values for that gene and as methylated (PMR ≥
10%) or unmethylated (PMR < 10%). Since the methylation at nine genes in 10
urothelium samples from cancer-free bladders were all less than 10% PMR, this value
was used as a biological cutoff for the presence or absence of methylation (Figure 2.1A).
For each patient, the following information was also included in the analysis: age at
diagnosis, sex, smoking history, tumor stage, and tumor grade (Table 2.1). To assess the
overall association between each of these characteristics and the PMR ranking for each
gene, patient and tumor characteristics were evaluated using either the Wilcoxon test for
32
the dichotomous characteristics of sex (male and female), grade (low grade and high
grade), and smoking history (nonsmokers, which included both never and irregular
smokers, and smokers, which included both former and current smokers) or the Kruskal-
Wallis test for the characteristics with 3 or more categories of tumor stage (Ta, CIS, T1,
and T2-4) and patient’s age at diagnosis (≤40, 41-50, 51-60, and 61-65) (Table 2.2).
A logistic regression model was used to evaluate the association between RUNX3
methylation, classified as positive (PMR ≥ 10%) or negative (PMR < 10%), and
clinicopathological characteristics (Table 2.3). The model used in the logistic regression
included the following variables: age at diagnosis (1-year increase), smoking history
(nonsmokers vs. smokers), sex, and tumor stage and grade (Ta low grade; Ta high grade,
CIS, T1; and T2-T4). The interaction term for smoking status and age was not significant
(p=0.23) and was omitted from the final model; similarly a quadratic term for age, to
capture a nonlinear relationship between age and the log odds of methylation, was also
not significant and was omitted in the final model. P-values were based on a likelihood
ratio test. The curves fit to this data display the association between age and probability
of methylation for smokers and non-smokers separately; the raw data were smoothed
using a LOESS procedure (SAS Institute Inc., 2000) and the estimated probability of
methylation was based on the logistic regression model with age as the only covariate.
All p-values were reported as two-sided.
33
RESULTS
RUNX3 is frequently and specifically methylated in bladder tumors
Since the methylation at nine genes in 10 urothelium samples from cancer-free
bladders were all less than 10% PMR, this value was used as a biological cutoff for the
presence or absence of methylation (Figure 2.1A). In matched sets of bladder tumors and
corresponding tissue EDNRB, RASSF1A, and BCL2 were frequently methylated in both
the tumors (65%, 60%, and 39%) and the corresponding tissues (15%, 11%, and 9%)
(Figure 2.1A). While TERT was methylated in 30% of tumors and only 4% of
corresponding tissues, RUNX3 was the most specific marker for bladder cancer, with
frequent methylation in the tumor samples (23/41, 56%) and infrequent methylation in
the corresponding tissues (2/46, 4%). The two corresponding tissues with RUNX3
methylation have nearly identical methylation patterns as their matched high grade and
invasive tumors, suggesting that these bladders either had epigenetic aberrancies
throughout or that the invasive tumor had spread across the bladder. The pathological
reports for these two cases indicated that the first case had multiple foci of carcinoma in
situ and severe urothelial atypia and the second case had diffuse lesions throughout the
bladder. Therefore the two corresponding tissue samples with RUNX3 methylation were
atypical and are not considered “normal”. Of the other 44 cases, two had areas of
urothelial hyperplasia and one had moderate chronic cystitis, although there was no
detectable methylation in the corresponding tissues, and one case had areas of squamous
metaplasia with methylation at TIMP3, COX2, and EDNRB in the corresponding tissue.
34
Figure 2.1 Methylation in bladder samples at nine different loci using quantitative
methylation-sensitive real-time PCR. Data is shown as percent of a fully methylated
reference (PMR), with PMR > 10% in grey, PMR ≤ 10% in black, and unavailable data
in white. A. Ten normal urothelium samples from cancer-free bladders were obtained
during prostatectomies from age-matched patients. Forty-six matched sets of bladder
tumor tissue and corresponding tissue were obtained during cystectomies. B.
Methylation data for 342 bladder cancer samples from paraffin-embedded tissues.
35
Since the matched tumor and corresponding tissue sets were taken from patients
undergoing cystectomies, most of the tumors were highly invasive. Therefore I extended
the study to include more noninvasive and low-grade tumor samples obtained by
transurethral resections. When I assayed the methylation status of nine genes in these
additional 342 bladder tumor samples, RUNX3 was still the most frequently methylated
gene (39%), with the next most frequently methylated genes being RASSF1A (37%),
EDNRB (30%), and BCL2 (28%) (Figure 2.1B, Table 2.2).
In order to confirm these methylation results I used another methylation assay,
Ms-SNuPE, which was developed in our lab (Gonzalgo and Liang, 2007) and is able to
quantify the methylation at specific CpG sites. When I examined two CpG sites just
downstream of the region assayed using MethyLight I found that 5 out of 7 bladder
tumors had high levels of RUNX3 methylation and only 1 out of 7 corresponding tissues
had a high level of RUNX3 methylation in an independent cohort of 7 matched sets
(Figure 2.2). Upon examination of the pathological reports for these patients it was noted
that in the patient with RUNX3 methylation in the corresponding tissue (Case 5) the entire
bladder showed signs of cystitis cystica, a type of proliferative cystitis, with acute and
chronic inflammation and had multifocal carcinoma in situ. Of the other six cases only
one had reported areas of urothelial atypia.
36
Table 2.2 Association of demographics with gene methylation (PMR)
Gene N
%
PMR
≥ 10%
Stage
1,3
Grade
2,4
Sex
2,5
Age
1,6
Smoked
2,7
RUNX3 304 39% 0.008 0.015 0.38 0.027 0.004
RASSF1A 323 37% <0.001 <0.001 0.25 0.49 0.81
EDNRB 316 30% 0.013 0.004 0.98 0.83 0.053
BCL2 324 28% 0.51 <0.001 0.10 0.50 0.12
COX2 239 13% 0.39 0.08 0.72 0.52 0.19
DAPK 253 4% 0.08 0.017 0.66 0.18 0.63
TERT 262 3% 0.10 0.42 0.43 0.35 0.13
TIMP3 222 1% 0.047 0.017 0.47 0.69 0.81
ECAD 241 < 1% 0.33 0.034 0.72 0.82 0.49
1
P values based on Kruskal-Wallis test, bold indicates p<0.05.
2
P values based on Wilcoxon test, bold indicates p<0.05.
3
Data grouped into the categories of Ta, CIS, T1, and T2-T4.
4
Data grouped into the categories of low grade (1-2) and high grade (3-4).
5
Data grouped into the categories of male and female.
6
Data grouped into the categories of ≤40, 41-50, 51-60, and 61-65 years at diagnosis.
7
Data grouped into the categories of nonsmokers (never/irregular) and smokers
(former/current).
37
Figure 2.2 RUNX3 is methylated specifically in bladder tumors. Average percent
methylation of two CpG sites located in the promoter of RUNX3 in 7 cases of matched
bladder tumors (black bars) and corresponding tissues (grey bars) measured by Ms-
SNuPE.
38
RUNX3 methylation precedes methylation at the other eight genes
Methylation at all genes except COX2 and TERT was significantly associated with
tumor grade and methylation at EDNRB, RASSF1A, RUNX3, and TIMP3 was associated
with tumor stage, revealing that methylation was more likely in tumors of higher grade or
stage (Table 2.2). Based on the assumption that tumors of the lowest stage and grade are
the precursors of tumors of higher stage and grade, I compared the methylation patterns
of the four most frequently methylated genes in Ta low grade tumors to tumors of higher
stage and grade in order to determine whether RUNX3 methylation precedes methylation
at the other genes (Figure 2.1B). An analysis of the discordance of RUNX3 methylation
revealed that of the 20 tumors that had only RUNX3 methylated, 17 cases were Ta low
stage tumors (85%), while only 36% of Ta low stage tumors had methylation at any of
the other genes besides RUNX3 (32/89, p<0.001, Fisher’s exact test). The observed
discordance is compatible with the hypothesis that RUNX3 methylation occurs early in
tumorigenesis before methylation of any of the other genes examined. In contrast,
methylation of EDNRB (35% compared to 31%, p=0.80) and BCL2 (33% compared to
28%, p=0.67) showed no discordance of methylation while RASSF1A showed the
opposite pattern of RUNX3, with significantly more methylation in tumors of higher
grade and stage (24% compared to 44%, p=0.042).
RUNX3 methylation increases with age and a history of smoking
Only methylation at RUNX3 was significantly associated with the age at which
diagnosis of UC was made (p=0.027) and a history of tobacco smoking (p=0.004) based
39
on univariable analyses (Table 2.2). Using a logistic regression model controlling for age,
sex, tumor stage, and grade, I found methylation of RUNX3 more frequently in tumors of
smokers versus nonsmokers (p=0.015, OR=2.76, 95% CI: 1.15, 6.66) (Table 2.3). There
was no effect of the duration of smoking when added to the above logistic regression
model (p=0.88). Methylation at RUNX3 was significantly associated with the age at
diagnosis (p=0.031) based on a logistic regression adjusting for sex, smoking history,
tumor stage and grade.
In order to examine the joint association between smoking, age, and RUNX3
methylation I used a logistic regression model with age, smoking history, tumor stage and
grade (Figure 2.3). I found a statistically significant association between RUNX3
methylation and both age (p=0.004) and smoking status (p=0.009). The probability of
RUNX3 methylation is higher in smokers and increases as a function of age in both
nonsmokers and smokers (Figure 2.3). Since RUNX3 methylation increases with age, is
not present in the normal urothelium, and occurs early in tumorigenesis, it can be used as
a molecular clock in order to determine the “age” of a bladder tumor and doing so reveals
that tumors from smokers are “older” than tumors nonsmokers.
40
Table 2.3 Impact of sex, age, tumor stage, tumor grade, and smoking history on
RUNX3 methylation
OR (95% CI) p-value
Age at Diagnosis 0.031
1-Year Increase 1.04 (1.00, 1.08)
Sex 0.49
Female 1.00 (reference)
Male 1.23 (0.68, 2.20)
Tumor Stage & Grade 0.010
Ta - Low 1.00 (reference)
Ta – High/CIS/T1 1.18 (0.68, 2.07)
T2+ 2.16 (1.18, 3.94)
Smoking 0.015
Nonsmokers 1.00 (reference)
Smokers 2.76 (1.15, 6.66)
p-values based on likelihood ratio test, bold indicates p<0.05.
(PMR ≥ 10%, N=304)
41
Figure 2.3 RUNX3 methylation increases with age and smoking history. Estimated
probability of RUNX3 methylation (PMR ≥ 10%) is plotted as a function of age at
diagnosis for smokers (N=265, dashed grey line) and nonsmokers (N=38, dashed black
line). These fitted curves are superimposed over the raw data that has been smoothed
using a LOESS procedure, with the smokers represented by the solid grey line and the
nonsmokers by the solid black line. The logistic regression analysis was performed on
tumors from 304 patients and yielded a statistically significant association between
RUNX3 methylation and smoking history (p=0.009, likelihood ratio test), adjusted by age
at diagnosis, tumor grade, and tumor stage.
42
DISCUSSION
In this study I evaluated the methylation of nine genes in urothelial carcinoma
samples from 342 patients about whom detailed demographic, clinicopathological, and
smoking information had been previously collected (Castelao et al., 2001). I found
several genes to be more frequently methylated in tumors of higher stage and grade. The
four most commonly methylated genes in this cohort of bladder tumors are involved in
apoptotic pathways, including RUNX3, and BCL2, EDNRB, RASSF1A (Figure 2.1),
consistent with a previous study (Friedrich et al., 2004). In addition, I found that RUNX3
is not methylated in normal-appearing tissue corresponding to tumors or tissue from
cancer-free bladders. However, since quantitative methylation-sensitive real-time PCR
detects individual strands of DNA that are simultaneously methylated at all sites a cutoff
value of 10% PMR cannot rule out the presence of sporadic sites of methylation.
Therefore, I measured methylation of two specific CpG sites in the RUNX3 promoter in a
small independent cohort of matched tumor and corresponding tissue. I found significant
levels of methylation in 5 of the 7 tumors and only one of the corresponding tissues
(Figure 2.2). However, this specific corresponding tissue is not considered to be non-
neoplastic, as it was taken from a bladder with cystitis cystica with acute and chronic
inflammation and multifocal carcinoma in situ. In addition, other studies have also found
that RUNX3 is not methylated in normal bladder (Kim et al., 2005b), prostate (Kang et al.,
2004; Kim et al., 2004b; Suzuki et al., 2005), lung (Suzuki et al., 2005), breast (Suzuki et
al., 2005), and colon (Kim et al., 2004b) tissues using a variety of methods and also never
or infrequently methylated in non-neoplastic tissue corresponding to bladder tumors
43
(Suzuki et al., 2005), gastric tumors (Kim et al., 2004b; Waki et al., 2003), breast tumors
(Lau et al., 2006), non-small cell lung tumors (Yanagawa et al., 2003), and hepatocellular
carcinomas (Park et al., 2005). Since RUNX3 is not methylated in normal urothelium or
normal-appearing tissue in bladders with cancer, we can assume that RUNX3 methylation
was not initially present in the normal urothelium of these patients, is specific to bladder
tumors, and is not acquired as a function of aging.
I have shown that RUNX3 methylation precedes methylation at the other 8 genes I
examined, indicating that it is an early event in bladder tumorigenesis. RUNX3
methylation appears to occur early in tumorigenesis in a variety of cancer types.
Methylation of RUNX3 has been found in invasive ductal breast carcinoma (IDC) and its
precursor lesion of ductal carcinoma in situ (DCIS) (Subramaniam et al., 2008), prostate
cancer and its precursor prostatic intraepithelial neoplasia (Kang et al., 2004), and gastric
cancer and its precursor lesions of chronic gastritis and intestinal metaplasia (Kim et al.,
2004b). RUNX3 methylation also occurs in Barrett’s esophagus (BE), a precursor of
esophageal adenocarcinoma (EAC), and is associated with progression to EAC
(Schulmann et al., 2005). In addition, I found methylation in the corresponding tissue
from three bladders, two of which had multifocal carcinoma in situ and acute and chronic
inflammation with cystitis cystica, a proliferative cystitis, or severe urothelial atypia, and
one of which with diffuse lesions throughout the bladder. Other groups have shown that
similar inflammatory lesions in the bladder express telomerase (Kavaler et al., 1998;
Lancelin et al., 2000), indicating that these lesions are likely to be undergoing
immortalization or malignant transformation. RUNX3 is a polycomb target (Fujii et al.,
44
2008) and several groups have found that polycomb targets identified in embryonic stem
cells become preferentially methylated during cancer (Ohm et al., 2007; Schlesinger et al.,
2007; Widschwendter et al., 2007). Several events occur in stem cells before they are
initiated and then clonally expand during promotion (Luebeck and Moolgavkar, 2002).
Initiating events are irreversible and if they occur in stem cells then that event will remain
in the asymmetrically dividing stem cell. Therefore, the results showing that
corresponding tissue from a bladder with widespread inflammatory lesions has elevated
RUNX3 methylation indicate that RUNX3 methylation occurs earlier than tumorigenesis
and accumulates in bladder stem cells that will eventually develop into a tumor.
According to a multivariable logistic regression model, the probability of RUNX3
methylation increases with age in both smokers and non-smokers. Biological age is a
surrogate marker for the number of times a cell has divided and the more times a cell
divides the more opportunities for aberrant methylation to accumulate (Preston-Martin et
al., 1990). Because the degree of hypermethylation in normal colonic epithelium is
related to age (Ahuja et al., 1998), it has been suggested that methylation can be used as a
molecular “clock” to predict the age of a tissue (Kim et al., 2005a). Therefore, I was able
to use RUNX3 methylation to show for the first time that methylation in tumor cells
might be useful as a molecular clock. I have shown that RUNX3 methylation is present
early in tumorigenesis and that the level increases with age (Figure 2.3). When I apply a
RUNX3 molecular clock to tumors from smokers versus nonsmokers, the data suggests
bladder tumors in smokers are “older” than in nonsmokers, i.e. have undergone more cell
divisions before diagnosis, since RUNX3 is more prevalent in tumors from smokers.
45
There are at least two possible explanations for such an observation. Either tumors in
smokers “age” or divide more quickly than in nonsmokers, resulting in faster
accumulation of methylation errors over a similar amount of time, or tumors in smokers
were initiated at an earlier age compared with tumors of nonsmokers, and therefore have
had more time to divide and accumulate RUNX3 methylation. More work is necessary to
distinguish between these two possibilities. Since RUNX3 methylation appears to be an
early event in bladder tumorigenesis and increases over time, then detection of RUNX3
methylation in biopsy or urine specimens could provide a marker to screen the at risk
population of cigarette smokers long before any symptoms are present.
CONCLUSION
Adding to the model for bladder cancer progression showing the molecular
pathways of tumorigenesis presented in Chapter 1, I have found that RUNX3 methylation
occurs before the development of both non-invasive and invasive tumors (Figure 2.4).
46
CIS
Ta
Low Grade
Non-Invasive
Papillary
Carcinoma
in situ
Flat
Metastasis M
Muscle
Invasive
Dysplasia
Lamina Propria
Invasive
Lamina Propria
Invasive
T
1
8p-
p53
INK4a
INK4a
Rb
?
9-
High Grade
Non-Invasive
Papillary
Normal
Urothelium
T
1
T
2
-T
4
?
9-
9-
ECM
Remodeling
Genes
ECM
Remodeling
Genes
FGFR3
RUNX3
Hypermethylation
Figure 2.4 Model for bladder cancer progression showing the molecular pathways of
tumorigenesis. RUNX3 hypermethylation occurs before the development of both non-
invasive and invasive tumors
47
CHAPTER 3
AN EPIGENETIC FIELD DEFECT IS PRESENT ACROSS
BLADDERS WITH CANCER
INTRODUCTION
Bladder cancer is characterized by a high rate of recurrence and necessitates
frequent and invasive monitoring. The biological basis for the high rate of recurrence
remains unknown, although the presence of a field defect has been postulated. A field
defect is an area of tissue that is predisposed to undergo transformation and has been
proposed as an underlying mechanism of tumor recurrences and multifocality. It remains
unclear whether such premalignant changes in an affected organ are due to expansion of a
single cell with a growth advantage or due to a generalized defect. DNA methylation has
been shown to be involved in field defects in other cancers, highlighting the possible role
of epigenetic alterations in the development of cancer. DNA methylation, nucleosome
positioning, and histone modifications are epigenetic processes that heritably change
gene expression without altering the DNA sequence (Egger et al., 2004). DNA
methylation of cytosine residues occurs in the context of CpGs. Most unmethylated
CpGs are found in GC-rich sequences referred to as CpG islands. Approximately half of
human genes have CpG islands at their promoters (Venter et al., 2001) and it has been
shown repeatedly that methylation of the CpG sites within these promoters can
effectively and heritably silence genes. Global hypomethylation of repetitive elements as
well as regional hypermethylation of CpG islands are common events during
48
tumorigenesis (Gaudet et al., 2003; Laird and Jaenisch, 1996). Methylation changes are
involved in both the initiation of carcinogenesis and progression (Goodman and Watson,
2002). Genome wide alterations occur in DNA methylation patterns in bladder cancer
(Gonzalgo et al., 1997; Liang et al., 2002; Liang et al., 1998; Markl et al., 2001). There
are progressive increases in the de novo methylation of CpG islands in bladder carcinoma
cells, suggesting that epigenetic gene silencing is involved in the development of bladder
cancer (Byun et al., 2007; Friedrich et al., 2005; Friedrich et al., 2004; Salem et al.,
2000a; Salem et al., 2000b; Wolff et al., 2008).
In order to determine whether there is an epigenetic field defect in bladder cancer
I conducted large scale DNA methylation profiling. The Illumina GoldenGate
methylation assay was used to measure DNA methylation in bladder urothelium samples
from three different sources: urothelium from 12 age-matched bladder cancer-free
patients (normal), 39 invasive bladder tumors and their corresponding apparently normal
tissues. Many CpG islands were hypermethylated not only in tumors but also in
apparently normal tissue when compared to normal urothelium from bladder cancer-free
patients, indicating the presence of an epigenetic field defect. These loci may provide
markers for the identification of individuals at risk for developing bladder cancer, as they
are only methylated in bladders with cancer.
Two different molecular pathways have been suggested in the development of
non-invasive and invasive urothelial carcinoma (UC). The non-invasive papillary tumors
tend to have mutations in FGFR3, while invasive cancers commonly have p53 mutations.
There may also be different epigenetic pathways between non-invasive and invasive UC.
49
In order to address this possibility I also conducted DNA methylation profiling on 52
non-invasive bladder tumors. I found a distinct pattern of hypomethylation at non-CpG
islands in the non-invasive tumors compared to invasive tumors, suggesting that there are
different epigenetic pathways involved in the development of these two tumor types.
50
MATERIALS AND METHODS
Tissue Collection and DNA/RNA Isolation. Tumor tissue samples were collected from
the patients undergoing cystectomy or TURBT for bladder cancer. Normal bladder
epithelium was obtained from 12 patients undergoing radical prostatectomy for prostate
cancer (aged from 50 to 80) and 7 autopsy patients aged from 34 to 82, 5 of which were
from non-cancer related deaths and 2 from deaths due to cancers other than bladder). All
of these collections took place at Norris Cancer Hospital in IRB-approved protocols with
patients’ consent. Hematoxylin and eosin (H&E) sections marked with the location of
the adjacent urothelium or tumor were used to guide in microdissection. DNA was
bisulfite treated as previously described (Gonzalgo and Liang, 2007). RNA extraction
was done using a RNAeasy Micro Kit (Qiagen, Crawley, UK. RNA was reverse-
transcribed as previously described(Friedrich et al., 2004).
Illumina GoldenGate Methylation Assay. Bisulfite conversion of 2-4 ug of genomic
DNA was achieved through use of the EZ DNA Methylation-Gold kit (Zymo Research).
Detailed methods for the GoldenGate assay are described in Figure 3.1. The samples
were interrogated using the Illumina GoldenGate methylation cancer panel I. A β-value
of 0-1.0 was reported for each CpG site signifying percent methylation, from 0% to 100%
respectively. β-values were calculated by subtracting background using negative
controls on the array and taking the ratio of the methylated signal intensity the sum of
both methylated and unmethylated signals.
51
Figure 3.1 DNA methylation assay scheme for GoldenGate. A. Bisulfite conversion
of DNA. B. For each CpG site an allele-specific oligonucleotide (ASO) and a locus-
specific oligonucleotide (LSO) probe pair was designed for the methylated state and a
corresponding ASO-LSO pair for the unmethylated state. Each ASO consists of a 3'-
portion that hybridizes to the bisulfite-converted genomic DNA, with the 3'-base
complementary to either the C or T allele of the targeted CpG site, and a 5'-portion with a
universal PCR primer sequence P1 or P2. The LSOs consist of three parts: a CpG locus-
specific sequence, an address sequence complementary to a corresponding capture
sequence on the array, and a universal PCR priming site (P3). P1 and P2 were
fluorescently labeled, each with a different dye, and associated with the T (unmethylated)
allele or the C (methylated) allele, respectively.
52
Quantitation of DNA Methylation. Pyrosequencing was also performed as described
previously (Bollati et al., 2007). Briefly, PCR was performed on bisulfite converted
DNA using a biotin-labeled 3’ primer to enable purification and denaturation of the
product by Streptavidin Sepharose beads and was followed by annealing of a sequencing
primer to the single-stranded PCR product. Pyrosequencing
was performed using the
PSQ HS96 Pyrosequencing System and the degree of methylation was expressed for each
DNA locus as
percentage methylated cytosines over the sum of methylated and
unmethylated cytosines.
Bisulfite Sequencing and X-chromosome Analysis. To analyze the methylation status
of individual DNA molecules, I cloned bisulfite PCR fragments into the pCR2.1 vector
using the TOPO-TA cloning kit (Invitrogen, Carlsbad, CA). Individual colonies were
screened for the insert and the region of interest was sequenced using M13 primers. The
primers for amplification of bisulfite-converted DNA were the same used in the
Pyrosequencing or Ms-SNuPE assays. Bisulfite sequencing was performed on a region
of the X-chromosome with an A/T SNP (rs4826507) and six CpG sites that are
differentially methylated between the active and inactive X chromosomes (Hellman and
Chess, 2007), allowing for the simultaneous identification of which X-chromosome was
being interrogated and its activation status.
Cell Lines. The non-tumorigenic human urothelial cell lines UROtsa and NK2426 and
the normal fibroblast cell line LD419 have been described previously(Chapman et al.,
53
2006; Friedrich et al., 2004; Rossi et al., 2001). Human bladder carcinoma cell lines
were obtained commercially (T24, J82, HT1376, SCaBER, UM-UC-3, TCCSUP, and
RT4; American Type Culture Collection, Manassas, VA) or derived in our laboratory
(prefix LD). Cell culture, DNA and RNA purification were performed as previously
described(Friedrich et al., 2004).
ZO2 knockdown and wound healing assay. ZO2 was knocked down using siRNA from
Dharmafect in the immortalized UROtsa cell line according to the manufacturer’s
protocol. A wound healing assay was performed using 100 nm of siRNA for ZO2 or 100
nm of a scrambled siControl.
54
RESULTS
Bladders with cancer have a widespread epigenetic field defect
In Chapter 2 I used the semi-quantitative methylation assay MethyLight to
measure DNA methylation in bladder urothelium samples from three different sources:
urothelium from 10 age-matched bladder cancer-free patients, 46 bladder tumors, and
their corresponding normal tissues. I found that EDNRB, RASSF1A, and BCL2 were
frequently methylated in both tumors (65%, 60%, and 39%) and the corresponding
tissues (15%, 11%, and 9%) (Figure 2.1A) in contrast to RUNX3 methylation, which
appears to occur only in bladder tumors or transformed tissues. Since I saw some
evidence of aberrant DNA methylation being present in apparently normal urothelium
from patients with bladder cancer I expanded the study using the Illumina GoldenGate
methylation assay to simultaneously interrogate 1505 loci. I measured DNA methylation
in bladder urothelium from 12 age-matched bladder cancer-free patients, 34 invasive
bladder tumors, and their corresponding normal tissues. A comparison of methylation in
UC tissue versus bladder urothelium from bladder cancer-free patients yielded 158
hypermethylated loci and 356 hypomethylated loci in UC. In addition I found 275 loci
that were hypermethylated and 31 loci that were hypomethylated not only in tumors but
also in adjacent normal appearing tissue when compared to normal urothelium from
bladder cancer-free patients, suggesting the presence of an epigenetic field defect in
bladder cancer patients (Figure 3.2). These loci may provide markers for the
identification of individuals at risk for developing bladder cancer, as they are only altered
in bladders with cancer .
55
% Methylation
0-20%
21-50%
51-80%
81-100%!
Figure 3.2 Changes in DNA methylation occur in the corresponding normal
urothelium of bladders with cancer. DNA methylation measured by the Illumina
GoldenGate assay in tissue from cancer-free bladders, “normal” urothelium from
bladders with cancer, and bladder tumors. Bladder tumors show a high degree of
abnormal methylation at these loci compared to cancer-free bladders. However,
“normal” urothelium from bladders with cancer also shows significantly more abnormal
methylation at these loci than cancer-free bladders.
56
In addition to serving as biomarkers, the methylation changes found at these loci
may have functional significance. Increased methylation at one of these loci, MYOD, has
previously been shown to be associated with oncogenic transformation (Rideout et al.,
1994) (Figure 3.3), indicating that the normal appearing urothelium may actually be
premalignant tissue. H-cadherin, CDH13, is also hypermethylated in adjacent bladder
tissue (Figure 3.4). Previous studies have found CDH13 to be hypermethylated in
premalignant gastric tissue (Yamamoto et al., 2008) and loss of expression is associated
with increased invasion in melanocytes (Kuphal et al., 2009). Another gene of interest,
Zona occludens 2 (ZO2), is a tight junction binding protein. While the role of ZO2 in
cancer has not been studied in depth, expression of other components of tight junctions is
altered in cancer, with loss of expression leading to increased cell motility and
invasiveness. Examining the expression of ZO2 in the same clinical samples revealed
that while it was expressed in normal urothelium, it was less so in the adjacent tissue and
even less in tumors, indicating that hypermethylation of ZO2 is associated with reduced
expression (Figure 3.5A). Bisulfite sequencing confirmed the absence of methylation in
normal urothelium from a bladder cancer-free patient, a low level of methylation in
adjacent tissue, and a high level of methylation in a bladder tumor (Figure 3.5B).
In order to determine the full extent of the hypermethylation found in adjacent
normal tissue I collected a series of tissue samples taken from various distances away
from tumors in several bladders. I found hypermethylation of ZO2 across bladders with
cancer when compared to the average level of ZO2 methylation in normal urothelium
from bladder cancer-free patients, indicating the presence of a widespread epigenetic
57
Figure 3.3 Methylation at MYOD in bladder tissue samples. Methylation data from
the Illumina GoldenGate assay performed on normal urothelium from patients without
bladder cancer (N), normal appearing urothelium taken from bladders with cancer (CN),
and bladder tumors (T) at MYOD. Horizontal lines represent the median value. Red
arrows indicate the CpG sites queried. *** represents p<0.001, ** represents p<0.01, as
determined by unpaired t-tests (N&CN, N&T) or paired t-tests (CN&T).
N CN T
0.0
0.2
0.4
0.6
0.8
1.0
*** **
***
(n=12) (n=34) (n=42)
Methylation of MYOD at -50 (! value)
N CN T
0.0
0.2
0.4
0.6
0.8
1.0
*** ***
***
(n=12) (n=34) (n=42)
Methylation of MYOD at +156 (! value)
MYOD!
!""#$%
58
Figure 3.4 Methylation at CDH13 in bladder tissue samples. Methylation data from
the Illumina GoldenGate assay performed on normal urothelium from patients without
bladder cancer (N), normal appearing urothelium taken from bladders with cancer (CN),
and bladder tumors (T) at CDH13. Horizontal lines represent the median value. Red
arrows indicate the CpG sites queried. *** represents p<0.001, ** represents p<0.01, as
determined by unpaired t-tests (N&CN, N&T) or paired t-tests (CN&T).
N CN T
0.0
0.2
0.4
0.6
0.8
1.0
*** ***
***
(n=12) (n=34) (n=42)
Methylation of CDH13 at -88 (! value)
N CN T
0.0
0.2
0.4
0.6
0.8
1.0
*** ***
***
(n=12) (n=34) (n=42)
Methylation of CDH13 at +102 (! value)
!""#$%
CDH13!
59
Figure 3.5 Methylation at ZO2 in bladder tissue samples. A. Methylation data from
the Illumina GoldenGate assay performed on normal urothelium from patients without
bladder cancer (N), normal appearing urothelium taken from bladders with cancer (CN),
and bladder tumors (T) at ZO2. Expression of ZO2 and GAPDH were measured by RT-
PCR. Horizontal lines represent the median value. Red arrows indicate the CpG sites
queried. *** represents p<0.001, ** represents p<0.01, as determined by unpaired t-tests
(N&CN, N&T) or paired t-tests (CN&T). B. Bisulfite sequencing of ZO2 was performed
on one N (NN12), one CN (2730N), and one T (2730T) using primers indicated by the
black arrows in order to confirm the Illumina methylation results.
N CN T
0.0
0.2
0.4
0.6
0.8
1.0
Methylation of ZO2 at +518 (! value)
** ***
***
(n=12) (n=34) (n=42)
N CN T
0.0
0.2
0.4
0.6
0.8
1.0
Methylation of ZO2 at +330 (! value)
** ***
***
(n=12) (n=34) (n=42)
N CN T
0.00
0.02
0.04
0.06
Expression (ZO2/GAPDH)
**
(n=12) (n=16) (n=16)
A
B
!""#$%
&'(%
!!"#$
#%&'!$
#%&'($
60
field defect in bladders with cancer (Figure 3.6). Knockdown of ZO2 expression in an
immortalized urothelial cell line lead to an increase in cell motility, as indicated by a
wound-healing assay (Figure 3.7). Therefore, the hypermethylation of ZO2 in bladder
tissue may contribute to a more malignant and invasive phenotype. Moreover, I have
shown that ZO2 methylation can be reduced by treatment with the DNA methylation
inhibitor 5-aza-CdR in tumor, premalignant (Chapman et al., 2006), and normal cell lines
leading to re-expression (Figure 3.8). These results suggest that in addition to treating
bladder tumors, premalignant defects in bladders could be reversed by epigenetic therapy.
The epigenetic field defect in bladders is not due to clonal expansion
To address whether the widespread epigenetic field defect I have found in
bladders with cancer is due to clonal expansion across the bladder or a generalized
epigenetic alteration I first sequenced DNA from tumors of these same five patients for
FGFR3 mutations, one of the earliest genetic alterations in bladder cancer(Wolff et al.,
2005). I found FGFR3 to be mutated only in the tumor samples and not in any of the
adjacent samples (#6519 at S249C and #6671 at F384I). Since it is possible that any
clonal expansion could have occurred before the FGFR3 mutation was acquired, I also
examined the pattern of X-inactivation in the series of samples taken from the two female
patients, #6522 and #6671. X-inactivation patterns are permanent marks of clonality and
I found a random pattern of X-inactivation across these bladders, suggesting that the
widespread epigenetic alterations did not result in the clonal expansion of a cell with a
growth advantage (Figure 3.9). In addition, previous work from the lab has shown with a
61
C1
C2
a2
a1
aT
b3
b2
b1
bT
C1
C2
a3
a2
a1
b3
b2
b1
T
C1
C2
a4
a3
a1
T
a1
b3
b2
T
a5
a3
a1
aT
b4
b3
b2
bT
c4
c3
c2
c1
d4
d3
d1
0
20
40
60
80
100
6486 6515 6519 6522 6671
% DNA Methylation of ZO2
6486!
a!
b!
C1! C2!
6515!
a!
b!
C1! C2!
6519!
a! C1! C2!
C3!
6522!
a!
b!
C1!
b!
6671!
a!
c!
d!
C1! C2!
Corresponding Normal (> 5cm)!
Corresponding Normal (0.5-2.5cm)!
Tumor!
A
B
Figure 3.6 Methylation of ZO2 across the bladder. Tissue samples were taken from
five patients of their tumors (red, T) and at increasing distances from the tumor (0.5 to 2
cm) in the surrounding normal-appearing tissue in multiple directions (light blue, a to d).
Additionally, distant normal-appearing samples were taken at least 5 cm from the tumor
(dark blue, C). B. Methylation was measured pyrosequencing. The green line represents
the median methylation value of normal samples from cancer-free patients.
62
Figure 3.7 ZO2 knockdown accelerates wound healing. ZO2 was knocked down using
siRNA in the immortalized UROtsa cell line. A wound healing assay performed using
100 nm of siRNA for ZO2 revealed that after 48 hours cells with reduced ZO2 exhibited
greater cell motility.
ZO2 (160kDa)!
ZO2 siRNA!
50nm 48h!
100nm 48h!
100nm 72h!
siCONTROL!
Untreated!
PCNA (36kDa)!
B-Actin (42kDa)!
A!
B!
0h! 24h! 48h!
siControl!
siZO2!
(100nm)!
63
Figure 3.8 Silencing of ZO2 can be reversed by 5-aza-CdR treatment. A. Treatment
with 5-aza-CdR reduced levels of methylation at ZO2 not just in three bladder carcinoma
cell lines, but also in a fibroblast cell line as well as a premalignant urothelial cell line,
NK2462, which has been immortalized with hTERT but not transformed. B. ZO2 is
highly expressed in a fibroblast cell line and urothelial cell line (blue lines) but not in
most bladder carcinoma cell lines assayed (red lines). Treatment with 5-aza-CdR for 24
hours induced expression of ZO2 three days later (yellow lines).
64
Figure 3.9 Two models of the development of bladder cancer. Normal urothelium is
composed of clonal patches of approximately 1 cm in size. There are at least two
possible explanations for the presence of abnormal DNA methylation in apparently
normal urothelium from bladders with cancer. Either aberrant DNA methylation
accumulates in one cell, followed by the clonal expansion of that cell population across
the urothelium and subsequent transformation, or there is a generalized epigenetic defect
that occurs independently in many different cells, which does not result in a growth
advantage but does predispose cells to undergo transformation.
65
different set of bladder specimens that clonality is maintained across the urothelium of
female patients with bladder cancer (Tsai et al., 1995). These results suggest that the
generalized epigenetic defect does not give rise to a clonal dominance of the
histologically normal cells but may provide a permissive environment either for the
emergence of tumor clones containing genetic mutations or the seeding or migration of
tumor cells once they arise (3.10).
Non-invasive bladder tumors have a distinct pattern of hypomethylation
I also conducted methylation analysis on 52 non-invasive bladder tumors using
the Illumina GoldenGate assay. Both noninvasive and invasive tumors had numerous loci
that were hypermethylated compared to normal urothelium (Figure 3.11), most of which
were located in CpG islands (Figure 3.12). I found a distinct pattern of hypomethylation
in the non-invasive tumors suggesting that there are two separate molecular pathways to
non-invasive or invasive bladder tumors that differ epigenetically as well as genetically
(Figure 3.11). Most hypomethylated loci were in CpG poor regions (Figure 3.13). Dr.
Yoshitomo Chihara confirmed several of the hypermethylated and hypomethylated loci
using pyrosequencing (Figure 3.14).
66
Figure 3.10 X-inactivation patterns reveal that the urothelium of two female
patients with bladder cancer remains polyclonal. Bisulfite sequencing was performed
on a region located on the X-chromosome with an A/T SNP (rs4826507). This SNP is
associated with an exonic region that is methylated on the active X chromosome and
unmethylated on the inactive X chromosome(Hellman and Chess, 2007). Pink and blue
boxes represent the portion of sequences from each clonal population. The tumors are
monoclonal while the samples of corresponding normal-appearing urothelium taken at
various distances from the site of the tumors are polyclonal, indicating that a single clone
has not grown across the bladder.
Tumor
c1
(0.5 cm)
C1
(> 5 cm)
d4
(2.0 cm)
A T
a4
(2.0 cm)
a1
(0.5 cm)
b3
(1.5 cm)
Tumor
a1
(0.5 cm)
C1
(> 5 cm)
b1
(0.5 cm)
b2
(1.0 cm)
b3
(1.5 cm)
T A
Corresponding Urothelium Corresponding Urothelium
X-chromosome SNP
rs4826507
A/T
100bp
6522!
a!
b!
C1!
b!
6671!
a!
c!
d!
C1! C2!
67
Figure 3.11 Bladder tumors are hypermethylated and non-invasive tumors have a
unique hypomethylated pattern. A Supervised cluster analysis is shown of bladder
cancer samples at over 1300 loci that showed differential methylation using the Illumina
GoldenGate methylation assay. NN represents normal tissue from patients without
bladder cancer, AN represents corresponding normal tissue from bladder cancer patients,
T represents different stages of bladder tumors, ES represent human ES cell lines, the
number below represents the sample size.
68
Figure 3.12 Bladder tumors are hypermethylated mainly at CpG islands. A
supervised cluster analysis is shown of bladder cancer samples at CpG islands that
showed differential methylation using the Illumina GoldenGate methylation assay. NN
represents normal tissue from patients without bladder cancer, AN represents
corresponding normal tissue from bladder cancer patients, T represents different stages of
bladder tumors, ES represent human ES cell lines, the number below represents the
sample size.
NN AN
20
T1 T2 T3 T4 ES
11 34
Ta
32 13 20 6 7
CpG Islands
69
Figure 3.13 Non-invasive bladder tumors are hypomethylated mainly at non CpG
islands. A supervised cluster analysis of bladder cancer samples at non CpG islands that
showed differential methylation using the Illumina GoldenGate methylation assay. NN
represents normal tissue from patients without bladder cancer, AN represents
corresponding normal tissue from bladder cancer patients, T represents different stages of
bladder tumors, ES represent human ES cell lines, the number below represents the
sample size.
NN AN
20
T1 T2 T3 T4 ES
11 34
Ta
32 13 20 6 7
Non CpG Islands
70
Figure 3.14 Illumina DNA methylation results were confirmed by Pyrosequencing.
Bent arrows represent transcriptional start sites and each lower tick mark represents a
CpG site. The region between two black arrows represents the PCR amplified region for
Pyrosequencing and the red arrow indicates the CpG sites checked by Pyrosequencing.
The CpG sites with blue arrows represents those checked by the Illumina assay.
71
DISCUSSION
Bladder cancer is one of the most expensive cancers to treat due to the need for
numerous follow-up visits and additional treatments. Therefore, the ideal therapeutic for
bladder cancer would, in addition to targeting tumors themselves, help to prevent future
recurrences. One of the possible causes for the frequent recurrence of bladder tumors is
the presence of premalignant changes across the urothelium. These changes have not
been well defined molecularly. DNA mutations that are postulated to occur early in
bladder tumorigenesis are not typically found in surrounding tissues (van Rhijn et al.,
2002). These results indicate that the premalignant changes are epigenetic rather than
genetic, supporting the use of epigenetic therapy not just for targeting tumors but also to
help prevent future recurrences by reversing premalignant changes before a new tumor
forms. Our lab was the first to identify that azaucleosides were powerful inhibitors of
DNA methylation and inducers of gene expression in treated cells (Jones and Taylor,
1980). These compounds are integrated into both RNA and DNA as a cytosine analog
and can form a covalent bond with the DNA methyltransferase enzyme DNMT1,
resulting in a reduction of methylation levels and reversal of aberrant silencing of tumor
suppressor genes (Yoo and Jones, 2006).
I discovered numerous loci in CpG islands that were aberrantly methylated in the
“normal” tissues from bladders with cancer when compared to tissue from cancer-free
bladders (Figure 3.2). In order to elucidate the mechanism behind this observation, I
collected a series of tissue samples taken from various distances away from tumors in
several bladders. I analyzed methylation at ZO2, which was one of the genes identified
72
by the Illumina assay as being aberrantly methylated in both bladder tumors and the
corresponding “normal” tissue (Figure 3.5). These results showed that the aberrant DNA
methylation of this gene was present across the entire bladders of these patients (Figure
3.6). Moreover, I have shown that treatment with 5-aza-CdR can reduce ZO2
methylation and induce reexpression in tumor, premalignant (Chapman et al., 2006), and
normal cell lines (Figure 3.8 ). Reduction of ZO2 expression in normal cells resulted in
an increase in cell motility, suggesting that some of the hypermethylation found in
apparently normal urothelium may contribute to its eventual transformation (Figure 3.7).
These results provide evidence that in addition to treating bladder tumors, the
premalignant epigenetic defects in bladders could be reversed by treatment with a DNA
methylation inhibitor.
My data supports a field cancerization model where independent events occur
across the urothelium resulting in a field defect that is oligoclonal (Figure 3.10).
However, another possible explanation is that the abnormal methylation occurred during
early development before the bladder was fully formed. Another possibility, which
cannot be ruled out by this data, is that the presence of a tumor causes epigenetic changes
across the bladder. Whatever the underlying mechanism, I have demonstrated that
alterations in DNA methylation occur across the entire urothelium and may allow for a
more permissible environment for the growth of newly mutated cells or spread of
remaining tumor cells rather than directly conferring a growth advantage. Transurethral
resection of bladder tumors would leave behind large areas of epigenetically altered
urothelium, possibly contributing to the high level of recurrence of bladder cancer.
73
Fortunately, the hypermethylated loci may provide valuable biomarkers that have the
potential to significantly impact the diagnosis and treatment of bladder cancer.
In addition to identifying markers of premalignant tissue I also found widespread
alterations of DNA methylation in the bladder tumors of all stages and grades. Two
different molecular pathways have been suggested in the development of non-invasive
and invasive bladder cancer. The non-invasive papillary tumors rarely have p53
mutations but commonly have FGFR3 mutations, while invasive cancers commonly have
p53 mutations but rarely FGFR3 mutations. The pattern of hypomethylated loci in non-
CpG islands that I found suggests that there are also different epigenetic pathways
between non-invasive and invasive bladder cancer (Figure 3.11). The limitation of my
study is that I was unable to determine whether the widespread field defect I identified in
bladders with invasive tumors is also present in bladders with non-invasive tumors. All
of the adjacent normal tissues that we collected were from patients that were having their
entire bladders removed. Treatment of non-invasive bladders tumors usually involves
surgical excision of the tumor and does not result in the collection of any additional
normal-appearing tissue. Therefore, since I have found that non-invasive and invasive
bladder tumors appear to undergo two separate epigenetic pathways to tumorigenesis I
cannot rule out that the generalized epigenetic defect I have uncovered is specific to
bladders with invasive tumors.
74
CONCLUSION
Adding to the model for bladder cancer progression showing the molecular
pathways of tumorigenesis presented in Chapter 1, we have found that some CpG island
hypermethylation occurs in premalignant bladder tissue in addition to bladder tumors and
that non-invasive bladder tumors show a distinct pattern of hypomethylation at non-CpG
islands (Figure 3.15).
75
Figure 3.15 Model for bladder cancer progression showing the molecular
pathways of tumorigenesis. Some CpG island hypermethylation occurs in
premalignant bladder tissue in addition to bladder tumors and non-invasive bladder
tumors show a distinct pattern of hypomethylation at non-CpG islands.
CIS
Ta
Low Grade
Non-Invasive
Papillary
Carcinoma
in situ
Flat
Metastasis M
Muscle
Invasive
Dysplasia
Lamina Propria
Invasive
Lamina Propria
Invasive
T
1
8p-
p53
INK4a
INK4a
Rb
?
9-
FGFR3
High Grade
Non-Invasive
Papillary
Premalignant
Urothelium
T
1
T
2
-T
4
?
9-
9-
ECM
Remodeling
Genes
ECM
Remodeling
Genes
Normal
Urothelium
CpG Island
Hypermethylation
Non-CpG Island
Hypomethylation
CpG Island
Hypermethylation
76
CHAPTER 4
HYPOMETHYLATION AND CHROMATIN REMODELING OF A LINE-1
PROMOTER ACTIVATES AN ALTERNATE TRANSCRIPT OF MET
INTRODUCTION
Aberrant DNA methylation is involved in the initiation and progression of
carcinogenesis and includes both regional hypermethylation of CpG islands at gene
promoters and global hypomethylation of repetitive elements such as long interspersed
nuclear elements (LINE-1s or L1s) (Baylin et al., 1998). While a high frequency of
hypomethylation during tumorigenesis has been demonstrated, i.e., 70% of bladder
tumors display hypomethylation of L1s (Florl et al., 1999), the consequences remain
unclear. Genome-wide hypomethylation is thought to promote tumorigenesis through
chromosomal instability (Eden et al., 2003) and possibly overexpression of oncogenes
(Hanada et al., 1993; Lipsanen et al., 1988). Hypomethylation of transposable elements
in mouse models can lead to disruption of normal gene function (Waterland and Jirtle,
2003). Viable yellow agouti (A
vy
) mice have a retrotransposon inserted into one allele of
the agouti locus. When this retrotransposon is hypomethylated, which can occur in utero
by limiting the maternal intake of methyl donors, it acts as an alternate promoter and
ectopically induces expression of the agouti gene resulting in yellow coat color, obesity,
and increased incidence of tumors (Waterland and Jirtle, 2003). The existence of a
similar phenomenon in humans has yet to be proven. While there are nearly 500,000
77
copies of the L1 retrotranspositional element in the human genome (Ovchinnikov et al.,
2002) a relatively small portion have retained potentially functional promoters, and these
are located within CpG islands. A full length L1 sequence (6 Kb) has a sense promoter
driving transcription of its two open reading frames and an antisense promoter driving
transcription in the opposite direction that can act as an alternate promoter for
surrounding genes (Matlik et al., 2006; Nigumann et al., 2002; Speek, 2001). A recent
study has revealed that over 30% of transcription start sites in the human genome are
located within repetitive elements, with just over 7% in L1s (Faulkner et al., 2009). They
identified almost 600 retrotransposons capable of acting as alternate TSSs for nearby
genes, which were mainly active in embryonic and cancer cell lines. While repetitive
elements are hypomethylated in cancer, no one has ever directly demonstrated that
hypomethylation of a retrotransposon leads to ectopic gene expression in humans.
In order to provide such evidence we first used the functional promoter sequence
of L1s to identify specific L1 promoters capable of expressing alternate transcripts of
host genes (Table 4.1). One such L1 promoter is located within the MET oncogene (L1-
MET) and has been shown to function as an alternate promoter in a bladder carcinoma
cell line by analysis of the ESTs in GenBank (Matlik et al., 2006). This transcript may
encode a truncated protein, however its full length has not been determined. Several
truncated forms of the tyrosine kinase MET, which is the hepatocyte growth factor (HGF)
receptor, are constitutively active and promote invasion and migration through activation
of a variety of signal transduction pathways (Birchmeier et al., 2003; Wallenius et al.,
2000). Overexpression of MET has been correlated with hypomethylation of L1s in
78
Table 4.1 Potentially active specific L1s and their associated alternate transcripts
Black boxes represent exons of the host gene while red boxes represent a specific L1.
The black arrow represents the transcriptional start site of the host gene while the red
arrow represents the alternate transcriptional start site within the potentially active L1
promoter. Accession numbers for representative alternate transcripts are followed by the
number in parentheses of similar transcripts transcribed from the individual L1.
79
chronic myeloid leukemia (CML) (Roman-Gomez et al., 2005). However, until recently
it has not been possible to determine the methylation status of the promoters of individual
L1s since the sequences are too similar to design primers for one particular locus
(Chalitchagorn et al., 2004; Phokaew et al., 2008; Roman-Gomez et al., 2005). Therefore
a direct correlation between the methylation of a specific L1 and expression of its
associated transcript has not been possible.
In addition to DNA methylation there are other layers of epigenetic regulation
responsible for regulating expression of single copy genes, including histone
modifications and nucleosome occupancy. Until now the status of such modifications
has been largely ignored at repetitive elements, particularly active ones. H2A.Z has been
found at unmethylated copies of retrotransposons in Arabidopsis (Zilberman et al., 2008).
However, it remains unknown whether the chromatin structure of repeats is similar to that
of single-copy genes and what alterations occur between inactive and active copies.
For the first time, I have elucidated the role of methylation in the transcriptional
activity of L1s by utilizing assays capable of examining the methylation status of specific
L1s and expression of alternate transcripts originating from the L1 promoters. In addition
to L1s being hypomethylated and transcriptionally active in bladder tumors I also found
that a specific L1 located within the MET oncogene is active across entire bladders with
cancer. The clinical implication of my finding is that surgical excision of the tumor
would leave behind large areas of the bladder that remain epigenetically altered and
express a potential oncogene. I also revealed that in addition to hypomethylation, an
active L1 undergoes chromatin remodeling similar to that of active single copy gene,
80
resulting in a switch from an inactive tetranucleosomal structure to a dinucleosomal
structure.
81
MATERIALS AND METHODS
Chromatin immunoprecipitation. ChIP was performed as described previously (Lin et
al., 2007).
Luciferase assay. Flora Han performed these experiments as described previously (Klug
and Rehli, 2006).
Methylation-dependent single promoter analysis. M-SPA was performed as described
previously (Fatemi et al., 2005). Briefly, chromatin was isolated from 250,000 cells and
treated for 15 minutes with 50U of M. SssI. In order to increase the resolution of this
method and to examine endogenously methylated promoters I treated chromatin from
250,000 with the enzyme M. CviPI, which methylates GpC sites, for 15 minutes with
100U (Xu et al., 1998).
MNase digestion and Southern blot. Dr. Shinwu Jeong and Shikhar Sharma performed
these experiments as described in their current manuscript.
82
RESULTS
Hypomethylation and expression of L1-MET is present across bladders with cancer
We designed bisulfite-specific PCR primers with one located in the L1 promoter
and the other in the surrounding intronic region of the host gene (Figure 4.1). The L1-
MET promoter was highly methylated in normal cells and tissues, whereas 18 out of 20 of
the bladder carcinoma cell lines showed dramatic hypomethylation (Figure 4.2A). We
also measured methylation more globally of L1s using two primers that anneal within the
L1 promoter (Figure 4.1) and found that hypomethylation of L1s was not as dramatic as
L1-MET hypomethylation (Figure 4.2A). The transcript from the L1-MET anti-sense
promoter contains its own exons 1 and 2, referred to as L1-MET exon 1 and L1-MET
exon 2, and shares the same reading frame as MET (Figure 4.1). We designed RT-PCR
primers with one primer located in either the MET exon 2 or the L1-MET exon 1 and one
primer located in the shared exon 3 to examine the expression of the host gene MET and
the alternate transcript from L1-MET, respectively (Figure 4.1). We confirmed the
transcription start site of L1-MET by 5’RACE in the T24 bladder carcinoma cell line in
which the L1-MET promoter is completely unmethylated. Using the same strategy we
also designed bisulfite-specific PCR primers and RT-PCR primers for two other specific
L1s located within ACVR1C, a member of the TGF-Beta family able to induce
apoptosis(Kim et al., 2004a), and RAB3IP, and a protein whose exact function is
unknown (Figure 4.1).
MET and L1-MET were not expressed in normal tissues except for placenta (data
not shown). The L1-MET transcript was lowly expressed in one bladder fibroblast cell
83
Figure 4.1 Map of alternate transcripts from specific L1s. Exons are represented by
black boxes while the specific L1s are represented by red boxes. Horizontal arrows
indicate the primers for RT-PCR (RT-MET-5’, RT-L1-MET-5’, and RT-MET-3’) or
PCR primers for bisulfite converted DNA (Bi-L1-5’, Bi-L1-3’, and Bi-MET-3’). A
similar strategy was used for the other two specific L1s. The lower tick marks represent
each CpG site. Vertical arrows indicate the CpG sites analyzed by Ms-SNuPE assay for
L1-MET. Green arrows indicate the primers used to amplify the pyrosequencing product
and the black arrow in between indicates the location of the pyrosequencing primer for
L1-ACVR1C and L1-RAB3IP. The left bent arrow indicates transcriptional start sites and
ATGs indicate translational start sites.
L1-ACVR1C!
L1-RAB3IP!
L1-MET!
84
Figure 4.2 Methylation and expression of L1-MET correlates in cell lines. A. L1-
MET methylation (red bars) and L1 methylation (black bars) was analyzed by Ms-SNuPE
in 8 normal tissues, one normal bladder fibroblast cell line (LD419), two non-tumorigenic
urothelial cell lines (UROtsa and NK2426), and 20 bladder carcinoma cell lines. Values
are from one CpG site for L1 and an average of two CpG sites for L1-MET. B.
Expression of L1-MET was measured using real-time RT PCR in one normal bladder
fibroblast cell line, two normal urothelial cell lines and 10 bladder carcinoma cell lines.
Red bars indicate the methylation status of L1-MET, which is also represented in A, and
green bars represent the level of expression relative to GAPDH.
Sperm
Lung
Kidney
Liver
Brain
Muscle
Leukocyte
Placenta
LD419
UROTSA
NK2464
J82
HT29
HT1376
TCCSUP
LD700
LD137
LD627
LD660
LD692
LD600
LD630
T24
SCABER
UMUC3
LD605
LD71
LD583
LD611
LD679
RT4
0
20
40
60
80
100
L1-MET
L1
Normal tissues
& cell lines
Bladder cancer cell lines
% DNA Methylation
LD419
UROTSA
NK2464
UMUC3
LD600
HT1376
TCCSUP
SCABER
LD605
T24
LD583
LD71
LD137
0
20
40
60
80
100
0
1
2
3
4
5
6
7
% DNA Methylation of L1-MET
Expression of L1-MET/GAPDH x 10
-3
A
B
85
line (LD419) and two non-tumorigenic urothelial cell lines, UROtsa (Rossi et al., 2001)
and NK2426 (Chapman et al., 2006), and highly expressed in most bladder carcinoma
cell lines (Figure 4.2B). Therefore L1-MET hypomethylation correlated with the
expression of the alternate transcript (Figure 4.2). Similar results were found for the L1
located within ACVR1C (L1-ACVR1C) and the L1 located within RAB3IP (L1-RAB3IP)
(Figures 4.3 & 4.4), although the correlation between DNA methylation and expression
was not as strong as for L1-MET. Since we did not confirm the exact start sites of the
L1-ACVR1C and L1-RAB3IP by 5’RACE it is possible that the RT-PCR primers were
not located in the ideal region due to multiple possible transcription start sites.
The next step was to determine whether hypomethylation of the specific L1
promoters and their associated alternate transcripts were present in uncultured bladder
tumors and therefore could contribute to tumorigenesis. We found high levels of
methylation at L1-MET and low expression in normal bladder epithelium obtained from
12 age-matched cancer free bladders (Figure 4.5A). There was significant
hypomethylation of, and expression from, L1-MET in bladder tumors (Figure 4.5A).
Therefore, we have shown for the first time that hypomethylation of L1-MET is directly
correlated with activation of alternate transcripts of MET and this occurs during
tumorigenesis. As expected the expression of the host gene MET was not correlated with
hypomethylation of the L1-MET promoter, since the expression of MET is regulated by
its endogenous promoter (Figure 4.5A). MET has been shown to be strongly expressed
in bladder cancer but not in normal urothelium (Joseph et al., 1995; Li et al., 1998; Natali
et al., 1996), however these results suggest that it may be L1-MET that is over-expressed
86
Figure 4.3 Methylation and expression of L1-ACVR1C correlates in cell lines. A. L1-
ACVR1C methylation (red bars) and L1 methylation (black bars) was analyzed by
pyrosequencing in 6 normal tissues, one normal bladder fibroblast cell line (LD419), one
non-tumorigenic urothelial cell lines (UROtsa), and 10 bladder carcinoma cell lines.
Values are from one CpG site for L1 and an average of two CpG sites for L1-ACVR1C. B.
Expression of L1- ACVR1C was measured using real-time RT PCR in one normal
bladder fibroblast cell line, one normal urothelial cell line, and 10 bladder carcinoma cell
lines. Red bars indicate the methylation status of L1-ACVR1C, which is also represented
in A, and green bars represent the level of expression relative to GAPDH.
Sperm
Lung
Kidney
Liver
Brain
Placenta
LD419
UROTSA
HT1376
TCCSUP
LD137
T24
SCABER
UMUC3
LD605
LD71
LD611
RT4
0
20
40
60
80
100
L1
L1-ACVR1C
Normal tissues
& cell lines
Bladder cancer cell lines
% DNA Methylation
LD419
urotsa
LD137
RT4
UMUC3
T24
LD71
TCCSUP
SCABER
HT1376
LD611
LD605
0
20
40
60
80
100
0
1
2
3
4
5
6
7
% DNA Methylation of L1-ACVR1C
Expression of L1-ACVR1C/GAPDH x 10
-3
A
B
87
Figure 4.4 Methylation and expression of L1-RAB3IP correlates in cell lines. A. L1-
RAB3IP methylation (red bars) and L1 methylation (black bars) was analyzed by
pyrosequencing in 6 normal tissues, one normal bladder fibroblast cell line (LD419), one
non-tumorigenic urothelial cell lines (UROtsa), and 10 bladder carcinoma cell lines.
Values are from one CpG site for L1 and an average of two CpG sites for L1-RAB3IP. B.
Expression of L1-RAB3IP was measured using real-time RT PCR in one normal bladder
fibroblast cell line, one normal urothelial cell line, and 10 bladder carcinoma cell lines.
Red bars indicate the methylation status of L1-RAB3IP, which is also represented in A,
and green bars represent the level of expression relative to GAPDH.
Sperm
Lung
Kidney
Liver
Brain
Placenta
LD419
UROTSA
HT1376
TCCSUP
LD137
T24
SCABER
UMUC3
LD605
LD71
LD611
RT4
0
20
40
60
80
100
L1
L1-RAB3IP
Normal tissues
& cell lines
Bladder cancer cell lines
% DNA Methylation
LD419
urotsa
RT4
LD137
TCCSUP
UMUC3
HT1376
T24
LD71
LD611
SCABER
LD605
0
20
40
60
80
100
0
1
2
3
4
5
6
7
% DNA Methylation of L1-RAB3IP
Expression of L1-RAB3IP/GAPDH x 10
-3
A
B
88
Figure 4.5 Methylation and expression status of specific L1s correlates in bladder
tissues. Horizontal lines represent the median value. Methylation status (left column)
was analyzed by Ms-SNuPE or pyrosequencing in normal tissues (N), corresponding
normal tissues (CN), and bladder tumors (T). Values are an average of two CpG sites.
Expression of the alternate transcript from the specific L1s (middle column), the host
genes (right column), and the control gene GAPDH was measured by real-time RT-PCR.
*** represents p<0.001, ** represents p<0.01, and * represents p<0.05. The specific L1s
assayed are A. L1-MET; B. L1-ACVR1C; C. L1-RAB3IP.
0
20
40
60
80
100
*** ***
***
N CN T
(n=19) (n=94) (n=63)
% DNA Methylation of L1-MET
0
20
40
60
80
100
* ***
***
N CN T
(n=24) (n=97) (n=65)
% DNA Methylation of L1-ACVR1C
0
20
40
60
80
100
*** ***
***
N CN T
(n=25) (n=92) (n=62)
% DNA Methylation of L1-RAB3IP
0.00
0.01
0.02
0.03
0.04
***
***
N CN T
(n=12) (n=77) (n=38)
Expression of L1-MET/GAPDH
0.000
0.005
0.010
0.015
0.020
***
***
N CN T
(n=12) (n=62) (n=32)
Expression of L1-ACVR1C/GAPDH
0.00
0.02
0.04
0.06
0.08
0.10
***
***
N CN T
(n=12) (n=63) (n=32)
Expression of L1-RAB3IP/GAPDH
0.0
0.1
0.2
0.3
0.4
N CN T
(n=12) (n=77) (n=38)
Expression of MET/GAPDH
0.00
0.01
0.02
0.03
***
***
N CN T
(n=12) (n=62) (n=32)
Expression of ACVR1C/GAPDH
0.0
0.1
0.2
0.3
0.4
*
N CN T
(n=12) (n=62) (n=32)
Expression of RAB3IP/GAPDH
A
B
C
89
in bladder cancer. We also examined the methylation and expression of two additional
specific L1 promoters located within host genes (Figure 4.5B&C). Hypomethylation of
the L1-ACVR1C and L1-RAB3IP promoters occurred in bladder tumors (Figure 4.5B&C).
Therefore it appears that a general result of hypomethylation of functional L1s is the
alteration of gene expression by activation of alternate transcripts during tumorigenesis.
Surprisingly, we also found hypomethylation and associated alternate expression of L1-
MET in the corresponding histologically normal tissues from tumor-bearing bladders
taken at least 5 cm away from the tumor when compared to normal tissues from cancer-
free bladders (p<0.0001) (Figure 4.5A). Hypomethylation and expression of L1-MET
was more prevalent in the corresponding normal tissues than L1-ACVR1C, L1-RAB3IP
(Figure 4.5B&C), and L1s (Allen’s paper), indicating that L1-MET may be involved in
the premalignant changes in bladders.
L1-MET hypomethylation is a sensitive and specific marker for premalignant
bladder tissue
Given the prevalence of specific L1 hypomethylation in bladder tumors and
premalignant tissues, it may provide a valuable marker for tumor detection and prediction
of high-risk patients, respectively. Receiver operating characteristic (ROC) curves for
L1-MET (Figure 4.6A), L1-ACVR1C (Figure 4.6B), and L1-RAB3IP (Figure 4.6C) show
that hypomethylation of specific L1s have an extraordinary degree of both sensitivity and
specificity for detecting bladder tumors (N vs. T areas under the curve (AUC) of 0.97,
0.88, and 0.90, respectively) and premalignant tissue (N vs. CN AUCs of 0.89, 0.64, and
90
Figure 4.6 ROC curves of specific L1s for bladder tissues. ROC curves distinguish
between normal bladder tissue (N) and bladder tumors (T), normal bladder tissue (N) and
corresponding normal bladder tissues (CN), and corresponding normal (CN) and bladder
tumors (T) using A. L1-MET methylation, B. L1-ACVR1C methylation, and C. L1-
RAB3IP methylation.
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
AUC 0.9687
L1-MET N vs T
1 - Specificity
Sensitivity
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
AUC 0.8772
L1-ACVR1C N vs T
1 - Specificity
Sensitivity
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
AUC 0.8971
L1-RAB3IP N vs T
1 - Specificity
Sensitivity
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
AUC 0.8903
L1-MET N vs CN
1 - Specificity
Sensitivity
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
AUC 0.6424
L1-ACVR1C N vs CN
1 - Specificity
Sensitivity
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
AUC 0.7641
L1-RAB3IP N vs CN
1 - Specificity
Sensitivity
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
AUC 0.8446
L1-MET CN vs T
1 - Specificity
Sensitivity
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
AUC 0.8213
L1-ACVR1C CN vs T
1 - Specificity
Sensitivity
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
AUC 0.7746
L1-RAB3IP CN vs T
1 - Specificity
Sensitivity
A
B
C
91
0.76, respectively) (Figure 4.6A-C). While premalignant tissue would remain after
surgical resection of bladder tumors, the degree of hypomethylation of specific L1s may
be able detect the recurrence of a tumor, as demonstrated by the ROC curves for
corresponding normal tissue versus bladder tumors (CN vs. T AUCs of 0.84, 0.82, and
0.77, respectively) (Figure 4.6A-C). L1-MET hypomethylation is the most sensitive and
specific of the three L1s measured. Since aberrant methylation in bladder tumors can be
detected in urine sediments (Friedrich et al., 2004) and we are able to detect
hypomethylation of L1-MET in urine sediments of bladder cancer patients (data not
shown), the current method of highly invasive cystoscopy has the potential to be replaced
with a noninvasive urine test.
Fully unmethylated L1-MET promoters are present in premalignant urothelium
In order to study L1 hypomethylation in more detail, we collected histologically
normal tissue samples from five tumor-bearing bladders taken at various distances and
directions from the site of the tumors (Figure 4.7A). L1-MET showed dramatic
hypomethylation in the bladder tumors when compared to the average level of
methylation in normal tissues from cancer free bladders (Figure 4.7A). L1-MET was also
significantly hypomethylated in normal-appearing tissues across each of the tumor-
bearing bladders (green and blue bars, p<0.001) regardless of the distance from the site of
the tumor, (Figure 4.7A). Tumors were also hypomethylated at L1-ACVR1C, L1-
RAB3IP, and global L1 but normal-appearing tissues were not significantly
hypomethylated (Figure 4.6B-D). Bisulfite sequencing of L1-MET in the urothelium of
92
C1
C2
a2
a1
aT
b3
b2
b1
bT
C1
C2
a3
a2
a1
b3
b2
b1
T
C1
C2
C3
a4
a3
a2
a1
T
C1
a3
a1
b3
b2
b1
T
C1
a5
a4
a3
a2
a1
aT
b4
b3
b2
b1
bT
c4
c3
c2
c1
cT
d4
d3
d2
d1
dT
0
20
40
60
80
100
6486 6515 6519 6522 6671
% DNA Methylation of L1-MET
C1
C2
a2
a1
aT
b3
b2
b1
bT
C1
C2
a3
a2
a1
b3
b2
b1
T
C1
C2
C3
a4
a3
a2
a1
T
C1
a3
a1
b3
b2
b1
T
C1
a5
a4
a3
a2
a1
aT
b4
b3
b2
b1
bT
c4
c3
c2
c1
cT
d4
d3
d2
d1
dT
0
20
40
60
80
100
6486 6515 6519 6522 6671
% DNA Methylation of L1-ACVR1C
C1
C2
a2
a1
aT
b3
b2
b1
bT
C1
C2
a3
a2
a1
b3
b2
b1
T
C1
C2
C3
a4
a3
a2
a1
T
C1
a3
a1
b3
b2
b1
T
C1
a5
a4
a3
a2
a1
aT
b4
b3
b2
b1
bT
c4
c3
c2
c1
cT
d4
d3
d2
d1
dT
0
20
40
60
80
100
6486 6515 6519 6522 6671
% DNA Methylation of L1-RAB3IP
C1
C2
a2
a1
aT
b3
b2
b1
bT
C1
C2
a3
a2
a1
b3
b2
b1
T
C1
C2
C3
a4
a3
a2
a1
T
C1
a3
a1
b3
b2
b1
T
C1
C2
a5
a4
a3
a2
a1
aT
b4
b3
b2
b1
bT
c4
c3
c2
c1
cT
d4
d3
d2
d1
dT
0
20
40
60
80
100
6486 6515 6519 6522 6671
% DNA Methylation of L1
A
B
C
D
6486!
a!
b!
C1! C2!
6515!
a!
b!
C1! C2!
6519!
a! C1! C2!
C3!
6522!
a!
b!
C1!
b!
6671!
a!
c!
d!
C1! C2!
Corresponding Normal (> 5cm)!
Corresponding Normal (0.5-2.5cm)!
Tumor!
Figure 4.7 Methylation of specific L1s across the bladder. Tissue samples were taken
from five patients of their tumors (red, T) and at increasing distances from the tumor (0.5
to 2 cm) in the surrounding normal-appearing tissue in multiple directions (light blue, a to
d). Additionally, distant normal-appearing samples were taken at least 5 cm from the
tumor (dark blue, C). Methylation was measured by Ms-SNUPE or pyrosequencing. The
green line represents the median methylation value of normal samples from cancer-free
patients. A. L1-MET; B. L1-ACVR1C; C. L1-RAB3IP; D. L1.
93
patients without bladder cancer revealed the presence of fully methylated strands while
fully unmethylated strands were present not only in the tumor but also in corresponding
normal urothelium regardless of the distance from the tumor (Figure 4.8A). A plot of the
distribution of DNA strands versus the percent of methylated sites reveals a biphasic
distribution in the patient with bladder cancer, with the majority of strands either fully
methylated or fully unmethylated (Figure 4.8B). These fully unmethylated strands are
likely transcriptionally active and the presence of such strands in the corresponding
normal urothelium supports my finding that L1-MET is expressed in these tissues. These
results reveal an epigenetic alteration resulting in the complete loss of methylation of
individual L1-MET promoters that occurs across the urothelium of bladders with tumors.
To my knowledge this is the first time that any alteration, epigenetic or genetic, has been
found across an entire tumor-bearing organ.
L1-MET encodes truncated MET proteins
We have confirmed the existence of 3.5 and 2.7 kb L1-MET transcripts by
sequencing of EST clones obtained from I.M.A.G.E. Both transcripts have start sites
located in the LINE-1 promoter. The 3.5 kb transcript extends to the 3’ end of the MET
gene and would encode a protein truncated at the N-terminus but otherwise identical to
the MET protein (truncated MET-1) while the 2.7 kb transcript would encode a protein
truncated at both termini (truncated MET-2) (Figure 4.9A). Therefore hypomethylation
of L1-MET could lead to expression of a transcript that encodes a truncated and
potentially constitutively active MET protein. In order to confirm that the two truncated
94
Figure 4.8 Bisulfite sequencing of L1-MET. A. Sequencing was performed on
samples from two bladder cancer-free patients (#4987 and #5240) and one bladder
cancer patient (#6519). White circles represent unmethylated CpGs and black circles
represent methylated CpGs. B. Biphasic distribution of L1-MET methylation status in
corresponding tissue from a patient with bladder cancer is revealed by plotting the
number of DNA strands by the percent of CpG sites methylated.
A
B
Number of DNA strands
Number of DNA strands
Patients without bladder cancer Patients with bladder cancer
% DNA methylation % DNA methylation
95
S1
1 2
Sema
Kinase 3 4
1 2
Kinase 3 4
PSI IPT TM
MET
Truncated MET - 1
!! !"#
SP
1 2 Kinase 3 4
1 2
Kinase Kinase
3 4
$%#
1 2 3 4
Truncated MET - 2
1 2 3 4
1 2
Sema
Kinase 3 4
1 2
Kinase Kinase 3 4
PSI IPT TM
MET
Truncated MET - 1
!! !"#
SP
1 2 Kinase 3 4
1 2
Kinase Kinase
3 4
$%#
1 2 3 4
Truncated MET - 2
1 2 3 4
pMEV
T - MET - 2
kDa
150
100
75
50
pMEV
2 2 2 4 4
pMEV
T - MET - 1
pMEV
T - MET - 2
kDa
150
100
75
50
pMEV
2 2 2 4 4
pMEV
T - MET - 1
kDa
150
100
75
50
pMEV
2 2 2 4 4
pMEV
T - MET - 1
A B
(µg)
Figure 4.9 The truncated MET protein encoded by L1-MET. A. The functional
domains of MET include the signal peptide (SP), sema domain at the N-terminus, the PSI
domain, IPT repeats, the transmembrane domain (TM), and the kinase domain at the C-
terminus. The truncated MET proteins 1 and 2 are shown. B. The two L1-MET
transcripts, T-MET-1 and T-MET-2, were cloned into a pMEV expression vector with 2
HA tags fused at the N-terminal. Hela cells were transfected with either the empty
pMEV vector, pMEV T-MET-1, or pMEV T-MET-2 and protein was extracted after 48
hours. The expression of truncated MET-1 (90kDa) and truncated MET-2 (60kDa) was
detected by western blot using an HA antibody.
96
transcripts actually encode proteins we cloned the coding regions into a pMEV
expression vector. Since most of the commercially available antibodies for detecting
MET are targeted to the N-terminus, which is missing in the truncated MET forms, we
used a pMEV expression vector with that has 2 HA tags fused to the N-terminus.
Transfection of Hela cells with the two constructs resulted in expression of two HA
tagged proteins of approximately 90 kDa and 60 kDa, indicating that the L1-MET
transcripts do encode truncated MET proteins (Figure 4.9B).
DNA methylation silences the L1-MET promoter
Thus far we have presented data revealing an association between
hypomethylation of the L1-MET promoter and expression of an alternate transcript.
Flora Han, a graduate student in our lab, utilized a luciferase promoter activity assay in
order to confirm the bidirectional L1-MET promoter activity and that DNA methylation
directly controls expression. A pCpGL luciferase reporter construct has been modified so
that it does not contain any CpG sites (Klug and Rehli, 2006). Therefore, after insertion
of the promoter sequence of interest the plasmid can be treated with the CpG
methyltransferase SssI, allowing the promoter to be methylated without affecting the
plasmid backbone. Flora inserted the L1-MET promoter sequence in both the sense
orientation to measure the L1 transcriptional activity and the antisense orientation to
measure the L1-MET activity. In both cases transcriptional activity was inhibited in the
methylated plasmid (Figure 4.10), confirming that DNA methylation suppresses
transcription from the L1-MET promoter.
97
Figure 4.10 DNA methylation silences the L1-MET promoter. The pCpGL luciferase
construct was modified to eliminate all CpG sites. The L1-MET promoter sequence was
ligated into the CpG less luciferase vector in either the antisense direction, measuring L1-
MET activity, or the sense direction, measuring L1 activity. Luciferase activity was high
in the untreated vector, SAM alone, and SssI alone. When SssI and SAM were added to
the luciferase vectors promoter activity was silenced. (Experiment conducted by Flora
Han)
CpG less Vector
L1-MET SAM SssI SssI+SAM
0
5.0!10
03
1.0!10
04
1.5!10
04
2.0!10
04
Luciferase Activity
L1 SAM SssI SssI+SAM
0
1.0!10
05
2.0!10
05
3.0!10
05
4.0!10
05
Luciferase Activity
L1-MET!
L1!
100bp!
CpG less Vector
98
Chromatin remodeling accompanies transcriptional activation of L1 promoters
In addition to DNA methylation, epigenetic regulation of gene transcription also
involves chromatin structure, specifically covalent modifications of histones,
incorporation of histone variants, and nucleosome occupancy. In order to better
understand the chromatin remodeling that occurs between inactive and active repetitive
elements we analyzed chromatin structure at single copies of L1s. We found that the
level of DNA methylation at each specific L1 is inversely proportional to the level of
enrichment of active histone marks (Figure 4.11). Specifically, comparing the T24
bladder carcinoma cell line in which MET-LINE is unmethylated to a non-tumorigenic
cell line with a high level of L1-MET methylation (Figure 4.11A), UROtsa, reveals a gain
of the active marks H3K4me3 and acetylated H3 and the histone variant H2A.Z (Figure
11B-D). Therefore transcriptional activation of a repetitive element results in a similar
pattern of chromatin remodeling found in single copy genes.
A switch from a tetranucleosome to dinucleosome structure accompanies
transcriptional activation of the L1-MET promoter
The Methylase-sensitive Single Promoter Analysis (M-SPA) has previously been
used to obtain single molecule resolution of nucleosome positioning at unmethylated
CpG island promoters (Fatemi et al., 2005). Briefly, chromatin was isolated and treated
with the CpG methyltransferase M. SssI, followed by DNA extraction, bisulfite
conversion, and genomic sequencing of individual clones. The resulting pattern of DNA
methylation reveals patches of protection, indicating the location of nucleosomes on
99
Figure 4.11 Chromatin remodeling occurs at active L1 promoters. A. DNA
methylation at specific and global L1s (with p16 as a control) was determined by
pyrosequencing in the immortalized urothelial cell line UROtsa and bladder carcinoma
cell line T24. The specific L1s had less methylation in the cancer cell line. Chromatin
immunoprecipitation was performed using antibodies for B. H3K4me3; C. acetylated H3;
and D. H2A.Z. The presence of active histone marks was associated with absence of
DNA methylation at the specific L1s in the cancer cell line.
P16
MET
RAB3IP
ACVR1C
GLOBAL
0
20
40
60
80
100
% DNA Methylation
0.0
0.2
0.4
0.6
0.8
H3K4me3 IP/Input (%)
0
2
4
6
8
Acetyl H3 IP/Input (%)
0.0
0.3
0.6
0.9
1.2
H2A.Z IP/Input (%)
P16
MET
RAB3IP
ACVR1C
GLOBAL
0
20
40
60
80
100
% DNA Methylation
0.0
0.2
0.4
0.6
0.8
H3K4me3 IP/Input (%)
0
2
4
6
8
Acetyl H3 IP/Input (%)
0.0
0.3
0.6
0.9
1.2
H2A.Z IP/Input (%)
L1s L1s
UROtsa T24
A
B
C
D
100
individual molecules. Previous work on the MLH1 bidirectional promoter demonstrated
that while each transcription start site lost the nucleosome directly upstream when active
(0 nucleosome), the nucleosome directly downstream was always maintained (+1
nucleosome) (Lin et al., 2007). The L1 promoter is a different type of bidirectional
promoter that generates partially overlapping sense and antisense transcripts, commonly
referred to as an antisense promoter (ASP). The L1 ASP has room for two nucleosomes
between the two transcription start sites, therefore each start site could have its own +1
nucleosome. We hypothesize that these two +1 nucleosomes are always maintained
while the active promoter loses the 0 nucleosome at both starts sites. Therefore an
inactive L1 promoter would exist in a tetranucleosomal state (two +1 and two 0
nucleosomes) while an active promoter would exist in a dinucleosomal state (two +1
nucleosomes). As expected, when we performed M-SPA on T24 cells in which MET-
LINE is unmethylated we found two nucleosomes occupying the region downstream of
the two transcription start sites and no nucleosomes upstream of either (Figure 4.12A).
However, the resolution of this assay is limited by the number and location of CpG sites
and the region upstream of the L1-MET start site contains only one CpG site. In order to
increase the resolution of this method we treated T24 chromatin with the enzyme M.
CviPI, which methylates GpC sites (Xu et al., 1998). There are numerous GpC sites
upstream of the L1-MET transcription start site and these were not protected from GpC
methylation, confirming the absence of the 0 nucleosome (Figure 4.12B).
The other limitation of the M-SPA method is that it cannot be used to assess
nucleosome positioning in a methylated region. However, the GpC methyltransferase
101
Figure 4.12 Nucleosome positioning in an active L1-MET promoter in T24 cells. T24
cells with a fully unmethylated L1-MET promoter have a dinucleosomal structure, as
determined by methylase dependent single promoter analysis (MSPA) with either A. SssI,
a CpG methyltransferase; or B. CviPI, a GpC methyltransferase.
102
enzyme can be used to avoid this problem provided any GpCpG sites are excluded from
analysis since it is not possible to distinguish between endogenous CpG methylation and
enzyme-induced GpC methylation at such loci. Therefore, by modifying our M-SPA
method by using a GpC methyltransferase I have conducted the first single molecule
analysis of nucleosome positioning at a methylated promoter. The endogenously
methylated L1-MET promoter in the LD419 fibroblast cell line (Figure 4.13A) was
completely occupied by nucleosomes, revealing that the methylated L1-MET promoter
exists in a tetranucleosomal structure (Figure 4.13B). The unmethylated MLH1 promoter
was used as a positive control for both M. SssI and GpC methyltransferase activity and
accessibility (Supplemental data). In addition, when DNA methylation levels are reduced
by knocking out expression of 2 of the 3 methyltransferases responsible for maintaining
DNA methylation, DNMT1 and DNMT3B, we see induction of expression of L1-MET
along with nucleosome eviction at the L1-MET promoter (Figures 4.14 & 4.15).
The switch from a tetranucleosomal structure to a dinucleosomal structure is a
common event at L1 promoters
While a single-molecule analysis of the nucleosome occupancy at the L1-MET
promoter confirmed my hypothesis that an active L1 promoter switches from a
tetranucleosomal structure to a dinucleosomal structure, we cannot generalize that such
remodeling occurs at other active L1s. In order to do so Dr. Shinwu Jeong took a cancer
cell line that has an inactive L1-MET promoter, HCT116, and performed chromatin
103
Figure 4.13 Nucleosome positioning in an inactive L1-MET promoter in LD419 cells.
A. LD419 cells with a fully methylated L1-MET promoter have a tetranucleosomal
structure, as determined by methylase dependent single promoter analysis (MSPA) with
B. CviPI, a GpC methyltransferase.
104
Figure 4.14 Nucleosome positioning in an inactive L1-MET promoter in HCT116
cells. A. HCT116 cells with a fully methylated L1-MET promoter have a
tetranucleosomal structure, as determined by methylase dependent single promoter
analysis (MSPA) with B. CviPI, a GpC methyltransferase.
105
Figure 4.15 Nucleosome positioning in an active L1-MET promoter in HCT116
DKO cells. HCT116 DKO (DNMT1 hypomorphs/DNMT3B knockout) cells with a fully
unmethylated L1-MET promoter have a dinucleosomal structure, as determined by
methylase dependent single promoter analysis (MSPA) with either A. SssI, a CpG
methyltransferase; or B. CviPI, a GpC methyltransferase.
106
fractionation using MNase digestion followed by sucrose gradient ultracentrifugation.
The fractions were run on an agarose gel and a genomic Southern using radioactively
labeled input DNA was performed. Most of the DNA was present in the
mononucleosome and dinucleosome fractions (Figure 4.16). The same blot was then
probed by Shikhar Sharma with the L1 promoter sequence and the enrichment of L1
promoters in both the dinucleosome and tetranucleosome fractions indicates that in
addition to L1-MET many other L1 promoters can be active during tumorigenesis (Figure
4.16).
107
Figure 4.16 Nucleosome eviction is a frequent occurrence at L1 promoters. Partial
MNase digestion of nucleosomes was followed by fractionation by a sucrose density
gradient. When a Southern for genomic DNA was performed on the DNA in each
fraction, enrichment in the mono- and dinucleosome fractions was revealed. When a
Southern for L1s was performed enrichment of L1s in the di- and tetranucleosome
fractions was found. According to our model the L1 promoters with a tetranucleosomal
structure should be inactive and methylated. (Experiment conducted by Dr. Shinwu Jeong
and Shikhar Sharma)
Genomic Southern! LINE-1 Southern !
6! 16! 6! 16!
108
DISCUSSION
The consequences of global hypomethylation at repetitive elements in cancer has
long been the subject of speculation regarding the generation of genomic instability and
potential activation of oncogenes. Ever since studies on viable yellow agouti (A
vy
) mice
revealed that hypomethylation of a retrotransposon could induce ectopic expression of a
gene and influence disease susceptibility it has been postulated that similar events may
occur in humans. We show, for the first time, a direct relationship between
hypomethylation of a retrotranspositional element and altered gene expression in humans
and its occurance in a diseased state. Specifically, hypomethylation of a LINE-1
promoter induces an alternate transcript of the MET oncogene in bladder tumors and
across the entire urothelium of tumor-bearing bladders (Figure 4.5). The robust
expression of the truncated MET does not lead to the clonal expansion of cells harboring
the defect, since the urothelium remains polyclonal (Figure 3.9). The widespread
epigenetic alteration is more likely to be permissive for the growth of newly mutated cells
rather than being directly responsible for the clonal evolution of the tumor. Receiver
operator characteristic (ROC) curves based on the hypomethylation of the specific LINE-
1 promoter within MET are extraordinarily specific and sensitive, providing a useful
marker for the diagnosis and treatment of bladder cancer (Figure 4.6).
The presence of L1-MET hypomethylation across the entire urothelium of tumor-
bearing bladders has several possible explanations. Epigenetic alterations such as
hypermethylation of tumor suppressor genes and hypomethylation of L1s have been
found in normal epithelia adjacent to several types of tumors, including breast (Yan et al.,
109
2006), esophageal (Eads et al., 2000b), and colon (Shen et al., 2005; Suter et al., 2004),
indicating the presence of a field defect. My data supports a field cancerization model
where independent events occur across the urothelium resulting in a field defect that is
oligoclonal (Jones et al., 2005). However, another possible explanation is that the loss of
L1-MET methylation occurred during early development before the bladder was fully
formed. Some evidence for such abnormal epigenetic programming exists, as a recent
study revealed that people who develop bladder cancer have slightly lower levels of
global DNA methylation in their blood than healthy control cases (Moore et al., 2008).
Another possibility, which cannot be ruled out by this data, is that the presence of a tumor
causes epigenetic changes across the bladder. Whatever the underlying mechanism, we
have demonstrated for the first time that the modulation of gene expression by
hypomethylation of a retrotransposon such as what occurs at the agouti locus in mice is
also found in humans. This leads to the activation of surrounding genes, including a
truncated form of MET that occurs across the entire urothelium and may allow for a more
permissible environment for the growth of newly mutated cells or spread of remaining
tumor cells rather than directly conferring a growth advantage. Transurethral resection of
bladder tumors would leave behind large areas of epigenetically altered urothelium,
possibly contributing to the high level of recurrence of bladder cancer. Fortunately, the
hypomethylation of L1-MET seems to provide a valuable biomarker that has the potential
to significantly impact the diagnosis and treatment of bladder cancer.
Besides revealing the relevance of specific L1 hypomethylation in bladder cancer
my study also elucidated the detailed chromatin structure of inactive and active L1s for
110
the first time (Figure 4.17). Hypomethylation and transcriptional activation of the L1-
MET promoter was accompanied by acquisition of the active histone marks H3K4me3
and acetylated H3 and the histone variant H2A.Z, similar to active single copy genes
(Figure 4.11). In addition, we found that the inactive and methylated L1-MET promoter
exists as a tetranucleosome structure (Figures 4.13&14) and when hypomethylated the
nucleosomes upstream of each transcription start site are lost, resulting in a dinucleosome
structure (Figures 4.12&15). We were able to apply a more genome wide approach to
show that the switch from a tetranucleosomal to dinucleosomal structure occurs at many
L1 promoters, indicating that global hypomethylation of L1 promoters results in
widespread chromatin remodeling and may have a large impact on the transcriptome in
cancer cells (Figure 4.16&17). Recently it was demonstrated that a large portion of the
human transcriptome can originate from within repetitive elements, highlighting the
widespread impact of such elements on gene expression (Faulkner et al., 2009). Active
L1s were mostly found in embryonic and cancer cell lines. We have provided direct
evidence that transcriptional activation of these elements is caused by hypomethylation
and chromatin remodeling at their promoters, occurs in a human diseased state, and may
be involved in disease predisposition.
111
ATG!
L1
1! 2! 3! 4 5!
MET!
L1-MET!
L1!
Hypomethylation!
L1-MET!
L1!
ATG!
Ac!
H2A.Z! H2A.Z!
K4! Ac!
K4!
L1
L1-MET!
Figure 4.17 Model of the epigenetic alterations that occur between the inactive L1-
MET and active L1-MET. The L1 located within the MET oncogene is usually silenced
by DNA methylation and has a compact chromatin structure with four nucleosomes
occupying the promoter. Upon hypomethylation the L1 promoter becomes
transcriptionally active. The active L1-MET promoter loses a nucleosome upstream of
each of the transcription start sites, resulting in a dinucleosome structure. The remaining
nucleosomes have acetylated H3, H3K4me3, and H2A.Z.
112
CONCLUSION
Adding to the model for bladder cancer progression showing the molecular
pathways of tumorigenesis presented in Chapter 1, we have found that specific L1s
become hypomethylated in premalignant bladder tissue while global L1 hypomethylation,
including at the specific L1s we assayed, occurs in bladder tumors (Figure 4.18).
113
CIS
Ta
Low Grade
Non-Invasive
Papillary
Carcinoma
in situ
Flat
Metastasis M
Muscle
Invasive
Dysplasia
Lamina Propria
Invasive
Lamina Propria
Invasive
T
1
8p-
p53
INK4a
INK4a
Rb
?
9-
FGFR3
High Grade
Non-Invasive
Papillary
Premalignant
Urothelium
T
1
T
2
-T
4
?
9-
9-
ECM
Remodeling
Genes
ECM
Remodeling
Genes
Normal
Urothelium
Specific L1
Hypomethylation
L1
Hypomethylation
Figure 4.18 Model for bladder cancer progression showing the molecular
pathways of tumorigenesis. Specific L1s become hypomethylated in premalignant
bladder tissue while global L1 hypomethylation, including at the specific L1s we
assayed, occurs in bladder tumors
114
CHAPTER 5
SUMMARY AND CONCLUSIONS
Successes in the clinic have opened up the era of epigenetic therapy in which the
goal is to reactivate genes silenced inappropriately during carcinogenesis (Yoo and Jones,
2006). The advantage of using epigenetic therapy to target defects in DNA methylation
is that, unlike mutations in the DNA sequence, these alterations are reversible. Both 5-
azacytidine (5-aza-CR) and 5-aza-2′-deoxycytidine (5-aza-CdR) have been approved by
the FDA to treat myelodysplastic syndrome (MDS). Several compounds that target
another epigenetic processes, such as the acetylation of histones, are currently undergoing
clinical trials and one (SAHA) has already been FDA approved for the treatment of
cutaneous T-cell lymphoma. The future of epigenetic therapy likely involves
administering combinations of epigenetic drugs and there are currently several clinical
trials doing so. In addition there is evidence that epigenetic therapy can reverse
resistance to chemotherapy, indicating that epigenetic therapy may help to enhance
traditional therapies (Cortez and Jones, 2008; Issa, 2007).
While epigenetic therapy is efficacious in treating certain types of leukemia it
remains controversial as a treatment method for solid tumors such as bladder cancer.
Therefore the focus of this thesis was twofold; to futher elucidate the role of epigenetic
alterations in bladder tumorigenesis and determine whether they provide useful
therapeutic targets for epigenetic therapy. An ideal therapy for bladder cancer would
115
address many of the unique aspects of bladder cancer, such as preventing tumors in high-
risk patients with a history of smoking, treating both noninvasive and invasive tumors
even though they develop via two separate molecular pathways, and reducing the
frequency of recurrences. Based on the findings in this thesis epigenetic therapy,
specifically DNA methylation inhibitors, has the potential to address each of these
concerns.
Markers of premalignant changes could be used to stratify high-risk patients with
a history of smoking in order to try to prevent formation of tumors. If tumors from
smokers initiate earlier than tumors from nonsmokers then RUNX3 methylation may be a
marker for detecting malignant changes in bladders many years before a tumor develops.
While smokers are at a much higher risk of developing bladder cancer than nonsmokers,
not all smokers develop bladder cancer. Therefore we would predict that smokers with
RUNX3 methylation present in urine sediments have precursor lesions in their bladders
and would be at an even higher risk for developing bladder cancer than smokers without
RUNX3 methylation and consequently should be monitored more closely. In addition,
since RUNX3 methylation is present in early lesions epigenetic therapy may be useful in
preventing high-risk patients from developing tumors.
Epigenetic therapy also has the potential for treating both noninvasive and
invasive bladder tumors since we found aberrant hypermethylation at numerous gene
promoters in both tumor types. In addition, noninvasive tumors also have a unique
pattern of hypomethylation at many gene promoters, indicating that a lower level of
methylation may be related to a less malignant phenotype and treatment of invasive
116
tumors with a DNA methylation inhibitor may result in a similar phenotype. We also
revealed that the urothelium in bladders with cancer is no longer “normal”. Instead the
urothelium in these diseased bladders has undergone widespread epigenetic alterations,
not due to a clonal expansion but as independent events. In contrast to RUNX3
methylation, which seems to be specific for transformed urothelial cells, these alterations
in premalignant tissue may predispose the tissue to undergo transformation or may
provide a more permissive environment for tumor cells to arise and spread, resulting in
future recurrences. Epigenetic therapy may help to reverse these alterations, preventing
tumors from recurring.
Cancer therapies must target cancer cells while not causing irreparable harm to
healthy cells. Therefore the obvious concern in using DNA methylation inhibitors to
treat or prevent bladder cancer is that normal urothelium may become globally
hypomethylated. Global hypomethylation is associated with chromosomal instability and
is commonly found in tumors. While systemic treatment of mice with DNA methylation
inhibitors causes some hypomethylation of repetitive elements in normal tissues, no
detrimental effects have been demonstrated thus far (Yoo et al). Even though global
hypomethylation in cancer was discovered before focal hypermethylation, most studies
have focused on the role of hypermethylation since it occurs at promoters and is known
to impact gene expression. When I conducted an in depth study on the role of
hypomethylation of repetitive elements in bladder tumorigenesis and I found that gene
expression is also impacted by hypomethylation at LINE-1 promoters.
117
In Chapter 4 we have revealed that hypomethylation at a promoter of a LINE-1
located within the MET oncogene activates an alternate MET transcript across the
urothelium of tumor-bearing bladders. Therefore surgical excision of the tumor would
leave behind large areas of the bladder that remain epigenetically altered and express a
potential oncogene. ROC curves based on the hypomethylation of the specific LINE-1
promoter within MET are extraordinarily specific and sensitive, providing a useful
marker for the diagnosis and treatment of bladder cancer that can be detected in urine
sediments. In addition to hypomethylation, we also show that the promoters of active
LINE-1 elements undergo chromatin remodeling similar to single copy genes.
Specifically, histone modifications associated with gene activity, including acetylated H3
and the histone variant H2A.Z, are acquired and nucleosomes upstream of the
transcription start sites are evicted, resulting in two nucleosome free regions.
The consequences of global hypomethylation at repetitive elements in cancer has
long been the subject of speculation. Ever since studies on agouti mice revealed that
hypomethylation of a retrotransposon could induce ectopic expression of a gene and
influence disease susceptibility it has been postulated that similar events may occur in
humans. We show, for the first time, that hypomethylation of a retrotranspositional
element leads to altered gene expression in humans and occurs in a diseased state. Global
hypomethylation of repetitive elements may activate numerous alternate transcripts of
genes during tumorigenesis. The potentially disturbing consequence of this finding is
that epigenetic therapy may activate alternate transcripts of genes in normal urothelium.
However, since these alterations are already present in bladders with tumors and since
118
DNA methylation inhibitors are preferentially taken up by tumors cells, it is possible that
the benefits from epigenetic therapy would outweigh any detriments. Future studies will
reveal whether inhibiting DNA methylation in the bladder will help to prevent tumors.
In summary, we have clearly shown that altered DNA methylation is a frequent
occurrence in both bladder tumors and premalignant tissues. We have used DNA
methylation as a clock to show that bladder tumors from smokers have undergone more
cell divisions and may have initiated earlier than tumors from nonsmokers. We have
revealed the presence of a generalized epigenetic defect across bladders with cancer that
is not due to clonal expansion. These epigenetic defects involve hypermethylation of
single copy genes and also hypomethylation of specific LINE-1 elements. The
hypomethylation of specific LINE-1 elements activates alternate transcripts of genes
across the bladder, including the MET oncogene. The presence of so many epigenetic
alterations in premalignant tissues of the bladder may indicate that treatment with
epigenetic therapy would be beneficial not just in the treatment of bladder tumors but also
in the prevention of future recurrences.
FINAL CONCLUSION
Adding to the model for bladder cancer progression showing the molecular
pathways of tumorigenesis presented in Chapter 1, we have found that: RUNX3
methylation occurs before the development of both non-invasive and invasive tumors but
after premalignant alterations, indicating that it occurs specifically in cells that have
already transformed. In addition, some CpG island hypermethylation occurs in
119
premalignant bladder tissue in addition to bladder tumors and non-invasive bladder
tumors show a distinct pattern of hypomethylation at non-CpG islands, which may be a
hallmark of a less malignant phenotype. Also, specific L1s become hypomethylated in
premalignant bladder tissue while global L1 hypomethylation, including at the specific
L1s we assayed, occurs in bladder tumors (Figure 18).
120
Figure 5.1 Model for bladder cancer progression showing the molecular pathways of
tumorigenesis. A final model incorporating the major findings in this thesis. Distinct
epigenetic alterations occur in premalignant tissues and noninvasive and invasive tumors
differ epigenetically in addition to genetically.
CIS
Ta
Low Grade
Non-Invasive
Papillary
Carcinoma
in situ
Flat
Metastasis M
Muscle
Invasive
Dysplasia
Lamina Propria
Invasive
Lamina Propria
Invasive
T
1
8p-
p53
INK4a
INK4a
Rb
?
9-
FGFR3
High Grade
Non-Invasive
Papillary
Premalignant
Urothelium
T
1
T
2
-T
4
?
9-
9-
ECM
Remodeling
Genes
ECM
Remodeling
Genes
Normal
Urothelium
Specific L1
Hypomethylation
CpG Island
Hypermethylation
Non-CpG Island
Hypomethylation
CpG Island
Hypermethylation
(i.e. RUNX3)
L1
Hypomethylation
121
REFERENCES
Ahuja, N., Li, Q., Mohan, A.L., Baylin, S.B., and Issa, J.P. (1998). Aging and DNA
methylation in colorectal mucosa and cancer. Cancer Res 58, 5489-5494.
Anttila, S., Hakkola, J., Tuominen, P., Elovaara, E., Husgafvel-Pursiainen, K.,
Karjalainen, A., Hirvonen, A., and Nurminen, T. (2003). Methylation of cytochrome
P4501A1 promoter in the lung is associated with tobacco smoking. Cancer Res 63, 8623-
8628.
Bakkar, A.A., Wallerand, H., Radvanyi, F., Lahaye, J.B., Pissard, S., Lecerf, L.,
Kouyoumdjian, J.C., Abbou, C.C., Pairon, J.C., Jaurand, M.C., et al. (2003). FGFR3 and
TP53 gene mutations define two distinct pathways in urothelial cell carcinoma of the
bladder. Cancer Res 63, 8108-8112.
Bartkova, J., Horejsi, Z., Koed, K., Kramer, A., Tort, F., Zieger, K., Guldberg, P.,
Sehested, M., Nesland, J.M., Lukas, C., et al. (2005). DNA damage response as a
candidate anti-cancer barrier in early human tumorigenesis. Nature 434, 864-870.
Batsche, E., Muchardt, C., Behrens, J., Hurst, H.C., and Cremisi, C. (1998). RB and c-
Myc activate expression of the E-cadherin gene in epithelial cells through interaction
with transcription factor AP-2. Mol Cell Biol 18, 3647-3658.
Baylin, S.B., Herman, J.G., Graff, J.R., Vertino, P.M., and Issa, J.P. (1998). Alterations
in DNA methylation: a fundamental aspect of neoplasia. Adv Cancer Res 72, 141-196.
Billerey, C., Chopin, D., Aubriot-Lorton, M.H., Ricol, D., Gil Diez de Medina, S., Van
Rhijn, B., Bralet, M.P., Lefrere-Belda, M.A., Lahaye, J.B., Abbou, C.C., et al. (2001).
Frequent FGFR3 mutations in papillary non-invasive bladder (pTa) tumors. Am J Pathol
158, 1955-1959.
Birchmeier, C., Birchmeier, W., Gherardi, E., and Vande Woude, G.F. (2003). Met,
metastasis, motility and more. Nat Rev Mol Cell Biol 4, 915-925.
Bollati, V., Baccarelli, A., Hou, L., Bonzini, M., Fustinoni, S., Cavallo, D., Byun, H.M.,
Jiang, J., Marinelli, B., Pesatori, A.C., et al. (2007). Changes in DNA methylation
patterns in subjects exposed to low-dose benzene. Cancer Res 67, 876-880.
Brandau, S., and Bohle, A. (2001). Bladder cancer. I. Molecular and genetic basis of
carcinogenesis. Eur Urol 39, 491-497.
Byun, H.M., Wong, H.L., Birnstein, E.A., Wolff, E.M., Liang, G., and Yang, A.S. (2007).
Examination of IGF2 and H19 loss of imprinting in bladder cancer. Cancer Res 67,
10753-10758.
122
Cappellen, D., De Oliveira, C., Ricol, D., de Medina, S., Bourdin, J., Sastre-Garau, X.,
Chopin, D., Thiery, J.P., and Radvanyi, F. (1999). Frequent activating mutations of
FGFR3 in human bladder and cervix carcinomas. Nat Genet 23, 18-20.
Castelao, J.E., Yuan, J.M., Skipper, P.L., Tannenbaum, S.R., Gago-Dominguez, M.,
Crowder, J.S., Ross, R.K., and Yu, M.C. (2001). Gender- and smoking-related bladder
cancer risk. J Natl Cancer Inst 93, 538-545.
Chalitchagorn, K., Shuangshoti, S., Hourpai, N., Kongruttanachok, N., Tangkijvanich, P.,
Thong-ngam, D., Voravud, N., Sriuranpong, V., and Mutirangura, A. (2004). Distinctive
pattern of LINE-1 methylation level in normal tissues and the association with
carcinogenesis. Oncogene 23, 8841-8846.
Chapman, E.J., Hurst, C.D., Pitt, E., Chambers, P., Aveyard, J.S., and Knowles, M.A.
(2006). Expression of hTERT immortalises normal human urothelial cells without
inactivation of the p16/Rb pathway. Oncogene 25, 5037-5045.
Chatterjee, S.J., George, B., Goebell, P.J., Alavi-Tafreshi, M., Shi, S.R., Fung, Y.K.,
Jones, P.A., Cordon-Cardo, C., Datar, R.H., and Cote, R.J. (2004). Hyperphosphorylation
of pRb: a mechanism for RB tumour suppressor pathway inactivation in bladder cancer. J
Pathol 203, 762-770.
Chen, C.H., Wang, W.J., Kuo, J.C., Tsai, H.C., Lin, J.R., Chang, Z.F., and Chen, R.H.
(2005). Bidirectional signals transduced by DAPK-ERK interaction promote the
apoptotic effect of DAPK. Embo J 24, 294-304.
Cheng, J., Huang, H., Pak, J., Shapiro, E., Sun, T.T., Cordon-Cardo, C., Waldman, F.M.,
and Wu, X.R. (2003). Allelic loss of p53 gene is associated with genesis and maintenance,
but not invasion, of mouse carcinoma in situ of the bladder. Cancer Res 63, 179-185.
Cheng, J., Huang, H., Zhang, Z.T., Shapiro, E., Pellicer, A., Sun, T.T., and Wu, X.R.
(2002). Overexpression of epidermal growth factor receptor in urothelium elicits
urothelial hyperplasia and promotes bladder tumor growth. Cancer Res 62, 4157-4163.
Chow, N.H., Cairns, P., Eisenberger, C.F., Schoenberg, M.P., Taylor, D.C., Epstein, J.I.,
and Sidransky, D. (2000). Papillary urothelial hyperplasia is a clonal precursor to
papillary transitional cell bladder cancer. Int J Cancer 89, 514-518.
Cortez, C.C., and Jones, P.A. (2008). Chromatin, Cancer and Drug Therapies. Mutat Res
In press.
Cote, R.J., Dunn, M.D., Chatterjee, S.J., Stein, J.P., Shi, S.R., Tran, Q.C., Hu, S.X., Xu,
H.J., Groshen, S., Taylor, C.R., et al. (1998). Elevated and absent pRb expression is
associated with bladder cancer progression and has cooperative effects with p53. Cancer
Res 58, 1090-1094.
123
Decesse, J.T., Medjkane, S., Datto, M.B., and Cremisi, C.E. (2001). RB regulates
transcription of the p21/WAF1/CIP1 gene. Oncogene 20, 962-971.
Dimova, D.K., and Dyson, N.J. (2005). The E2F transcriptional network: old
acquaintances with new faces. Oncogene 24, 2810-2826.
Dulaimi, E., Uzzo, R.G., Greenberg, R.E., Al-Saleem, T., and Cairns, P. (2004).
Detection of bladder cancer in urine by a tumor suppressor gene hypermethylation panel.
Clin Cancer Res 10, 1887-1893.
Dunn, K.L., Espino, P.S., Drobic, B., He, S., and Davie, J.R. (2005). The Ras-MAPK
signal transduction pathway, cancer and chromatin remodeling. Biochem Cell Biol 83, 1-
14.
Dyrskjot, L., Kruhoffer, M., Thykjaer, T., Marcussen, N., Jensen, J.L., Moller, K., and
Orntoft, T.F. (2004). Gene expression in the urinary bladder: a common carcinoma in situ
gene expression signature exists disregarding histopathological classification. Cancer Res
64, 4040-4048.
Dyrskjot, L., Thykjaer, T., Kruhoffer, M., Jensen, J.L., Marcussen, N., Hamilton-Dutoit,
S., Wolf, H., and Orntoft, T.F. (2003). Identifying distinct classes of bladder carcinoma
using microarrays. Nat Genet 33, 90-96.
Eads, C.A., Danenberg, K.D., Kawakami, K., Saltz, L.B., Blake, C., Shibata, D.,
Danenberg, P.V., and Laird, P.W. (2000a). MethyLight: a high-throughput assay to
measure DNA methylation. Nucleic Acids Res 28, E32.
Eads, C.A., Lord, R.V., Kurumboor, S.K., Wickramasinghe, K., Skinner, M.L., Long,
T.I., Peters, J.H., DeMeester, T.R., Danenberg, K.D., Danenberg, P.V., et al. (2000b).
Fields of aberrant CpG island hypermethylation in Barrett's esophagus and associated
adenocarcinoma. Cancer Res 60, 5021-5026.
Eads, C.A., Lord, R.V., Wickramasinghe, K., Long, T.I., Kurumboor, S.K., Bernstein, L.,
Peters, J.H., DeMeester, S.R., DeMeester, T.R., Skinner, K.A., et al. (2001). Epigenetic
patterns in the progression of esophageal adenocarcinoma. Cancer Res 61, 3410-3418.
Eden, A., Gaudet, F., Waghmare, A., and Jaenisch, R. (2003). Chromosomal instability
and tumors promoted by DNA hypomethylation. Science 300, 455.
Egger, G., Liang, G., Aparicio, A., and Jones, P.A. (2004). Epigenetics in human disease
and prospects for epigenetic therapy. Nature 429, 457-463.
Eymin, B., Karayan, L., Seite, P., Brambilla, C., Brambilla, E., Larsen, C.J., and Gazzeri,
S. (2001). Human ARF binds E2F1 and inhibits its transcriptional activity. Oncogene 20,
1033-1041.
124
Fatemi, M., Pao, M.M., Jeong, S., Gal-Yam, E.N., Egger, G., Weisenberger, D.J., and
Jones, P.A. (2005). Footprinting of mammalian promoters: use of a CpG DNA
methyltransferase revealing nucleosome positions at a single molecule level. Nucleic
Acids Res 33, e176.
Faulkner, G.J., Kimura, Y., Daub, C.O., Wani, S., Plessy, C., Irvine, K.M., Schroder, K.,
Cloonan, N., Steptoe, A.L., Lassmann, T., et al. (2009). The regulated retrotransposon
transcriptome of mammalian cells. Nat Genet 41, 563-571.
Feinberg, A.P. (2004). The epigenetics of cancer etiology. Semin Cancer Biol 14, 427-
432.
Feng, Z., Hu, W., Rom, W.N., Beland, F.A., and Tang, M.S. (2002). 4-aminobiphenyl is
a major etiological agent of human bladder cancer: evidence from its DNA binding
spectrum in human p53 gene. Carcinogenesis 23, 1721-1727.
Florl, A.R., Lower, R., Schmitz-Drager, B.J., and Schulz, W.A. (1999). DNA methylation
and expression of LINE-1 and HERV-K provirus sequences in urothelial and renal cell
carcinomas. Br J Cancer 80, 1312-1321.
Fraga, M.F., Ballestar, E., Paz, M.F., Ropero, S., Setien, F., Ballestar, M.L., Heine-Suner,
D., Cigudosa, J.C., Urioste, M., Benitez, J., et al. (2005). Epigenetic differences arise
during the lifetime of monozygotic twins. Proc Natl Acad Sci U S A 102, 10604-10609.
Friedrich, M.G., Chandrasoma, S., Siegmund, K.D., Weisenberger, D.J., Cheng, J.C.,
Toma, M.I., Huland, H., Jones, P.A., and Liang, G. (2005). Prognostic relevance of
methylation markers in patients with non-muscle invasive bladder carcinoma. Eur J
Cancer 41, 2769-2778.
Friedrich, M.G., Weisenberger, D.J., Cheng, J.C., Chandrasoma, S., Siegmund, K.D.,
Gonzalgo, M.L., Toma, M.I., Huland, H., Yoo, C., Tsai, Y.C., et al. (2004). Detection of
methylated apoptosis-associated genes in urine sediments of bladder cancer patients. Clin
Cancer Res 10, 7457-7465.
Fujii, S., Ito, K., Ito, Y., and Ochiai, A. (2008). Enhancer of zeste homologue 2 (EZH2)
down-regulates RUNX3 by increasing histone H3 methylation. J Biol Chem.
Gao, J., Huang, H.Y., Pak, J., Cheng, J., Zhang, Z.T., Shapiro, E., Pellicer, A., Sun, T.T.,
and Wu, X.R. (2004). p53 deficiency provokes urothelial proliferation and synergizes
with activated Ha-ras in promoting urothelial tumorigenesis. Oncogene 23, 687-696.
Garcia-Espana, A., Salazar, E., Sun, T.T., Wu, X.R., and Pellicer, A. (2005). Differential
expression of cell cycle regulators in phenotypic variants of transgenically induced
bladder tumors: implications for tumor behavior. Cancer Res 65, 1150-1157.
125
Gaudet, F., Hodgson, J.G., Eden, A., Jackson-Grusby, L., Dausman, J., Gray, J.W.,
Leonhardt, H., and Jaenisch, R. (2003). Induction of tumors in mice by genomic
hypomethylation. Science 300, 489-492.
Gonzalez-Zulueta, M., Bender, C.M., Yang, A.S., Nguyen, T., Beart, R.W., Van Tornout,
J.M., and Jones, P.A. (1995). Methylation of the 5' CpG island of the p16/CDKN2 tumor
suppressor gene in normal and transformed human tissues correlates with gene silencing.
Cancer Res 55, 4531-4535.
Gonzalgo, M.L., and Liang, G. (2007). Methylation-sensitive single-nucleotide primer
extension (Ms-SNuPE) for quantitative measurement of DNA methylation. Nat Protoc 2,
1931-1936.
Gonzalgo, M.L., Liang, G., Spruck, C.H., 3rd, Zingg, J.M., Rideout, W.M., 3rd, and
Jones, P.A. (1997). Identification and characterization of differentially methylated
regions of genomic DNA by methylation-sensitive arbitrarily primed PCR. Cancer Res
57, 594-599.
Gonzalo, S., Garcia-Cao, M., Fraga, M.F., Schotta, G., Peters, A.H., Cotter, S.E., Eguia,
R., Dean, D.C., Esteller, M., Jenuwein, T., et al. (2005). Role of the RB1 family in
stabilizing histone methylation at constitutive heterochromatin. Nat Cell Biol 7, 420-428.
Goodman, J.I., and Watson, R.E. (2002). Altered DNA methylation: a secondary
mechanism involved in carcinogenesis. Annu Rev Pharmacol Toxicol 42, 501-525.
Habuchi, T., Luscombe, M., Elder, P.A., and Knowles, M.A. (1998). Structure and
methylation-based silencing of a gene (DBCCR1) within a candidate bladder cancer
tumor suppressor region at 9q32-q33. Genomics 48, 277-288.
Hanada, M., Delia, D., Aiello, A., Stadtmauer, E., and Reed, J.C. (1993). bcl-2 gene
hypomethylation and high-level expression in B-cell chronic lymphocytic leukemia.
Blood 82, 1820-1828.
Hartmann, A., Moser, K., Kriegmair, M., Hofstetter, A., Hofstaedter, F., and Knuechel, R.
(1999). Frequent genetic alterations in simple urothelial hyperplasias of the bladder in
patients with papillary urothelial carcinoma. Am J Pathol 154, 721-727.
Hartmann, A., Schlake, G., Zaak, D., Hungerhuber, E., Hofstetter, A., Hofstaedter, F.,
and Knuechel, R. (2002). Occurrence of chromosome 9 and p53 alterations in multifocal
dysplasia and carcinoma in situ of human urinary bladder. Cancer Res 62, 809-818.
Hellman, A., and Chess, A. (2007). Gene body-specific methylation on the active X
chromosome. Science 315, 1141-1143.
Herceg, Z. (2007). Epigenetics and cancer: towards an evaluation of the impact of
environmental and dietary factors. Mutagenesis 22, 91-103.
126
Hipfner, D.R., and Cohen, S.M. (2004). Connecting proliferation and apoptosis in
development and disease. Nat Rev Mol Cell Biol 5, 805-815.
Hopman, A.H., Kamps, M.A., Speel, E.J., Schapers, R.F., Sauter, G., and Ramaekers,
F.C. (2002). Identification of chromosome 9 alterations and p53 accumulation in isolated
carcinoma in situ of the urinary bladder versus carcinoma in situ associated with
carcinoma. Am J Pathol 161, 1119-1125.
Issa, J.P. (2007). DNA methylation as a therapeutic target in cancer. Clin Cancer Res 13,
1634-1637.
Jebar, A.H., Hurst, C.D., Tomlinson, D.C., Johnston, C., Taylor, C.F., and Knowles, M.A.
(2005). FGFR3 and Ras gene mutations are mutually exclusive genetic events in
urothelial cell carcinoma. Oncogene.
Jones, P.A., and Baylin, S.B. (2002). The fundamental role of epigenetic events in cancer.
Nat Rev Genet 3, 415-428.
Jones, P.A., and Baylin, S.B. (2007). The epigenomics of cancer. Cell 128, 683-692.
Jones, P.A., and Taylor, S.M. (1980). Cellular differentiation, cytidine analogs and DNA
methylation. Cell 20, 85-93.
Jones, T.D., Wang, M., Eble, J.N., MacLennan, G.T., Lopez-Beltran, A., Zhang, S.,
Cocco, A., and Cheng, L. (2005). Molecular evidence supporting field effect in urothelial
carcinogenesis. Clin Cancer Res 11, 6512-6519.
Joseph, A., Weiss, G.H., Jin, L., Fuchs, A., Chowdhury, S., O'Shaugnessy, P., Goldberg,
I.D., and Rosen, E.M. (1995). Expression of scatter factor in human bladder carcinoma. J
Natl Cancer Inst 87, 372-377.
Kang, G.H., Lee, S., Lee, H.J., and Hwang, K.S. (2004). Aberrant CpG island
hypermethylation of multiple genes in prostate cancer and prostatic intraepithelial
neoplasia. J Pathol 202, 233-240.
Kavaler, E., Landman, J., Chang, Y., Droller, M.J., and Liu, B.C. (1998). Detecting
human bladder carcinoma cells in voided urine samples by assaying for the presence of
telomerase activity. Cancer 82, 708-714.
Kawamura, K., Izumi, H., Ma, Z., Ikeda, R., Moriyama, M., Tanaka, T., Nojima, T.,
Levin, L.S., Fujikawa-Yamamoto, K., Suzuki, K., et al. (2004). Induction of centrosome
amplification and chromosome instability in human bladder cancer cells by p53 mutation
and cyclin E overexpression. Cancer Res 64, 4800-4809.
Kim, B.C., van Gelder, H., Kim, T.A., Lee, H.J., Baik, K.G., Chun, H.H., Lee, D.A.,
Choi, K.S., and Kim, S.J. (2004a). Activin receptor-like kinase-7 induces apoptosis
127
through activation of MAPKs in a Smad3-dependent mechanism in hepatoma cells. J
Biol Chem 279, 28458-28465.
Kim, D.H., Kim, J.S., Ji, Y.I., Shim, Y.M., Kim, H., Han, J., and Park, J. (2003).
Hypermethylation of RASSF1A promoter is associated with the age at starting smoking
and a poor prognosis in primary non-small cell lung cancer. Cancer Res 63, 3743-3746.
Kim, D.H., Nelson, H.H., Wiencke, J.K., Zheng, S., Christiani, D.C., Wain, J.C., Mark,
E.J., and Kelsey, K.T. (2001). p16(INK4a) and histology-specific methylation of CpG
islands by exposure to tobacco smoke in non-small cell lung cancer. Cancer Res 61,
3419-3424.
Kim, J.Y., Siegmund, K.D., Tavare, S., and Shibata, D. (2005a). Age-related human
small intestine methylation: evidence for stem cell niches. BMC Med 3, 10.
Kim, T.Y., Lee, H.J., Hwang, K.S., Lee, M., Kim, J.W., Bang, Y.J., and Kang, G.H.
(2004b). Methylation of RUNX3 in various types of human cancers and premalignant
stages of gastric carcinoma. Lab Invest 84, 479-484.
Kim, W.J., Kim, E.J., Jeong, P., Quan, C., Kim, J., Li, Q.L., Yang, J.O., Ito, Y., and Bae,
S.C. (2005b). RUNX3 inactivation by point mutations and aberrant DNA methylation in
bladder tumors. Cancer Res 65, 9347-9354.
Klug, M., and Rehli, M. (2006). Functional analysis of promoter CpG methylation using
a CpG-free luciferase reporter vector. Epigenetics 1, 127-130.
Knowles, M.A. (1999). The genetics of transitional cell carcinoma: progress and potential
clinical application. BJU Int 84, 412-427.
Knowles, M.A. (2001). What we could do now: molecular pathology of bladder cancer.
Mol Pathol 54, 215-221.
Kuphal, S., Martyn, A.C., Pedley, J., Crowther, L.M., Bonazzi, V.F., Parsons, P.G.,
Bosserhoff, A.K., Hayward, N.K., and Boyle, G.M. (2009). H-cadherin expression
reduces invasion of malignant melanoma. Pigment Cell Melanoma Res 22, 296-306.
Laird, P.W., and Jaenisch, R. (1996). The role of DNA methylation in cancer genetic and
epigenetics. Annu Rev Genet 30, 441-464.
Lancelin, F., Anidjar, M., Villette, J.M., Soliman, A., Teillac, P., Le Duc, A., Fiet, J., and
Cussenot, O. (2000). Telomerase activity as a potential marker in preneoplastic bladder
lesions. BJU Int 85, 526-531.
Lau, Q.C., Raja, E., Salto-Tellez, M., Liu, Q., Ito, K., Inoue, M., Putti, T.C., Loh, M., Ko,
T.K., Huang, C., et al. (2006). RUNX3 is frequently inactivated by dual mechanisms of
128
protein mislocalization and promoter hypermethylation in breast cancer. Cancer Res 66,
6512-6520.
Li, B., Kanamaru, H., Noriki, S., Fukuda, M., and Okada, K. (1998). Differential
expression of hepatocyte growth factor in papillary and nodular tumors of the bladder. Int
J Urol 5, 436-440.
Liang, G., Gonzalgo, M.L., Salem, C., and Jones, P.A. (2002). Identification of DNA
methylation differences during tumorigenesis by methylation-sensitive arbitrarily primed
polymerase chain reaction. Methods 27, 150-155.
Liang, G., Salem, C.E., Yu, M.C., Nguyen, H.D., Gonzales, F.A., Nguyen, T.T., Nichols,
P.W., and Jones, P.A. (1998). DNA methylation differences associated with tumor tissues
identified by genome scanning analysis. Genomics 53, 260-268.
Lin, J.C., Jeong, S., Liang, G., Takai, D., Fatemi, M., Tsai, Y.C., Egger, G., Gal-Yam,
E.N., and Jones, P.A. (2007). Role of nucleosomal occupancy in the epigenetic silencing
of the MLH1 CpG island. Cancer Cell 12, 432-444.
Lipsanen, V., Leinonen, P., Alhonen, L., and Janne, J. (1988). Hypomethylation of
ornithine decarboxylase gene and erb-A1 oncogene in human chronic lymphatic leukemia.
Blood 72, 2042-2044.
Luebeck, E.G., and Moolgavkar, S.H. (2002). Multistage carcinogenesis and the
incidence of colorectal cancer. Proc Natl Acad Sci U S A 99, 15095-15100.
Markl, I.D., Cheng, J., Liang, G., Shibata, D., Laird, P.W., and Jones, P.A. (2001).
Global and gene-specific epigenetic patterns in human bladder cancer genomes are
relatively stable in vivo and in vitro over time. Cancer Res 61, 5875-5884.
Markl, I.D., and Jones, P.A. (1998). Presence and location of TP53 mutation determines
pattern of CDKN2A/ARF pathway inactivation in bladder cancer. Cancer Res 58, 5348-
5353.
Marsit, C.J., Houseman, E.A., Schned, A.R., Karagas, M.R., and Kelsey, K.T. (2007).
Promoter Hypermethylation is Associated with Current Smoking, Age, Gender, and
Survival in Bladder Cancer. Carcinogenesis.
Marsit, C.J., Karagas, M.R., Danaee, H., Liu, M., Andrew, A., Schned, A., Nelson, H.H.,
and Kelsey, K.T. (2006). Carcinogen exposure and gene promoter hypermethylation in
bladder cancer. Carcinogenesis 27, 112-116.
Matlik, K., Redik, K., and Speek, M. (2006). L1 antisense promoter drives tissue-specific
transcription of human genes. J Biomed Biotechnol 2006, 71753.
129
McCabe, M.T., Davis, J.N., and Day, M.L. (2005). Regulation of DNA methyltransferase
1 by the pRb/E2F1 pathway. Cancer Res 65, 3624-3632.
Miranda, T.B., and Jones, P.A. (2007). DNA methylation: The nuts and bolts of
repression. J Cell Physiol 213, 384-390.
Moore, L.E., Pfeiffer, R.M., Poscablo, C., Real, F.X., Kogevinas, M., Silverman, D.,
Garcia-Closas, R., Chanock, S., Tardon, A., Serra, C., et al. (2008). Genomic DNA
hypomethylation as a biomarker for bladder cancer susceptibility in the Spanish Bladder
Cancer Study: a case-control study. Lancet Oncol 9, 359-366.
Munro, N.P., and Knowles, M.A. (2003). Fibroblast growth factors and their receptors in
transitional cell carcinoma. J Urol 169, 675-682.
Nakagawa, T., Kanai, Y., Ushijima, S., Kitamura, T., Kakizoe, T., and Hirohashi, S.
(2005). DNA hypomethylation on pericentromeric satellite regions significantly
correlates with loss of heterozygosity on chromosome 9 in urothelial carcinomas. J Urol
173, 243-246.
Natali, P.G., Prat, M., Nicotra, M.R., Bigotti, A., Olivero, M., Comoglio, P.M., and Di
Renzo, M.F. (1996). Overexpression of the met/HGF receptor in renal cell carcinomas.
Int J Cancer 69, 212-217.
Nicholson, R.I., Gee, J.M., and Harper, M.E. (2001). EGFR and cancer prognosis. Eur J
Cancer 37 Suppl 4, S9-15.
Nigumann, P., Redik, K., Matlik, K., and Speek, M. (2002). Many human genes are
transcribed from the antisense promoter of L1 retrotransposon. Genomics 79, 628-634.
Nishiyama, H., Gill, J.H., Pitt, E., Kennedy, W., and Knowles, M.A. (2001). Negative
regulation of G(1)/S transition by the candidate bladder tumour suppressor gene
DBCCR1. Oncogene 20, 2956-2964.
Oeggerli, M., Tomovska, S., Schraml, P., Calvano-Forte, D., Schafroth, S., Simon, R.,
Gasser, T., Mihatsch, M.J., and Sauter, G. (2004). E2F3 amplification and overexpression
is associated with invasive tumor growth and rapid tumor cell proliferation in urinary
bladder cancer. Oncogene 23, 5616-5623.
Ohm, J.E., McGarvey, K.M., Yu, X., Cheng, L., Schuebel, K.E., Cope, L., Mohammad,
H.P., Chen, W., Daniel, V.C., Yu, W., et al. (2007). A stem cell-like chromatin pattern
may predispose tumor suppressor genes to DNA hypermethylation and heritable silencing.
Nat Genet 39, 237-242.
Ordway, J.M., Williams, K., and Curran, T. (2004). Transcription repression in
oncogenic transformation: common targets of epigenetic repression in cells transformed
by Fos, Ras or Dnmt1. Oncogene 23, 3737-3748.
130
Orlow, I., LaRue, H., Osman, I., Lacombe, L., Moore, L., Rabbani, F., Meyer, F., Fradet,
Y., and Cordon-Cardo, C. (1999). Deletions of the INK4A gene in superficial bladder
tumors. Association with recurrence. Am J Pathol 155, 105-113.
Ovchinnikov, I., Rubin, A., and Swergold, G.D. (2002). Tracing the LINEs of human
evolution. Proc Natl Acad Sci U S A 99, 10522-10527.
Oxford, G., and Theodorescu, D. (2003). The role of Ras superfamily proteins in bladder
cancer progression. J Urol 170, 1987-1993.
Park, W.S., Cho, Y.G., Kim, C.J., Song, J.H., Lee, Y.S., Kim, S.Y., Nam, S.W., Lee, S.H.,
Yoo, N.J., and Lee, J.Y. (2005). Hypermethylation of the RUNX3 gene in hepatocellular
carcinoma. Exp Mol Med 37, 276-281.
Peterson, E.J., Bogler, O., and Taylor, S.M. (2003). p53-mediated repression of DNA
methyltransferase 1 expression by specific DNA binding. Cancer Res 63, 6579-6582.
Phokaew, C., Kowudtitham, S., Subbalekha, K., Shuangshoti, S., and Mutirangura, A.
(2008). LINE-1 methylation patterns of different loci in normal and cancerous cells.
Nucleic Acids Res 36, 5704-5712.
Preston-Martin, S., Pike, M.C., Ross, R.K., Jones, P.A., and Henderson, B.E. (1990).
Increased cell division as a cause of human cancer. Cancer Res 50, 7415-7421.
Rideout, W.M., 3rd, Eversole-Cire, P., Spruck, C.H., 3rd, Hustad, C.M., Coetzee, G.A.,
Gonzales, F.A., and Jones, P.A. (1994). Progressive increases in the methylation status
and heterochromatinization of the myoD CpG island during oncogenic transformation.
Mol Cell Biol 14, 6143-6152.
Rieger-Christ, K.M., Mourtzinos, A., Lee, P.J., Zagha, R.M., Cain, J., Silverman, M.,
Libertino, J.A., and Summerhayes, I.C. (2003). Identification of fibroblast growth factor
receptor 3 mutations in urine sediment DNA samples complements cytology in bladder
tumor detection. Cancer 98, 737-744.
Ringrose, L. (2007). Polycomb comes of age: genome-wide profiling of target sites. Curr
Opin Cell Biol 19, 290-297.
Robertson, K.D., Ait-Si-Ali, S., Yokochi, T., Wade, P.A., Jones, P.L., and Wolffe, A.P.
(2000). DNMT1 forms a complex with Rb, E2F1 and HDAC1 and represses transcription
from E2F-responsive promoters. Nat Genet 25, 338-342.
Robertson, K.D., and Jones, P.A. (1998). The human ARF cell cycle regulatory gene
promoter is a CpG island which can be silenced by DNA methylation and down-
regulated by wild-type p53. Mol Cell Biol 18, 6457-6473.
131
Roman-Gomez, J., Jimenez-Velasco, A., Agirre, X., Cervantes, F., Sanchez, J., Garate, L.,
Barrios, M., Castillejo, J.A., Navarro, G., Colomer, D., et al. (2005). Promoter
hypomethylation of the LINE-1 retrotransposable elements activates sense/antisense
transcription and marks the progression of chronic myeloid leukemia. Oncogene 24,
7213-7223.
Rossi, M.R., Masters, J.R., Park, S., Todd, J.H., Garrett, S.H., Sens, M.A., Somji, S.,
Nath, J., and Sens, D.A. (2001). The immortalized UROtsa cell line as a potential cell
culture model of human urothelium. Environ Health Perspect 109, 801-808.
Salem, C., Liang, G., Tsai, Y.C., Coulter, J., Knowles, M.A., Feng, A.C., Groshen, S.,
Nichols, P.W., and Jones, P.A. (2000a). Progressive increases in de novo methylation of
CpG islands in bladder cancer. Cancer Res 60, 2473-2476.
Salem, C.E., Markl, I.D., Bender, C.M., Gonzales, F.A., Jones, P.A., and Liang, G.
(2000b). PAX6 methylation and ectopic expression in human tumor cells. Int J Cancer 87,
179-185.
Sarkar, S., Julicher, K.P., Burger, M.S., Della Valle, V., Larsen, C.J., Yeager, T.R.,
Grossman, T.B., Nickells, R.W., Protzel, C., Jarrard, D.F., et al. (2000). Different
combinations of genetic/epigenetic alterations inactivate the p53 and pRb pathways in
invasive human bladder cancers. Cancer Res 60, 3862-3871.
SAS Institute Inc. (2000). The Loess Procedure. In SAS/STAT User's Guide, Version 8
(Cary, NC, SAS Publishing), pp. 1853-1900.
Schlesinger, Y., Straussman, R., Keshet, I., Farkash, S., Hecht, M., Zimmerman, J., Eden,
E., Yakhini, Z., Ben-Shushan, E., Reubinoff, B.E., et al. (2007). Polycomb-mediated
methylation on Lys27 of histone H3 pre-marks genes for de novo methylation in cancer.
Nat Genet 39, 232-236.
Schulmann, K., Sterian, A., Berki, A., Yin, J., Sato, F., Xu, Y., Olaru, A., Wang, S., Mori,
Y., Deacu, E., et al. (2005). Inactivation of p16, RUNX3, and HPP1 occurs early in
Barrett's-associated neoplastic progression and predicts progression risk. Oncogene 24,
4138-4148.
Schwartz, Y.B., Kahn, T.G., Nix, D.A., Li, X.Y., Bourgon, R., Biggin, M., and Pirrotta,
V. (2006). Genome-wide analysis of Polycomb targets in Drosophila melanogaster. Nat
Genet 38, 700-705.
Shen, L., Kondo, Y., Rosner, G.L., Xiao, L., Hernandez, N.S., Vilaythong, J., Houlihan,
P.S., Krouse, R.S., Prasad, A.R., Einspahr, J.G., et al. (2005). MGMT promoter
methylation and field defect in sporadic colorectal cancer. J Natl Cancer Inst 97, 1330-
1338.
132
Speek, M. (2001). Antisense promoter of human L1 retrotransposon drives transcription
of adjacent cellular genes. Mol Cell Biol 21, 1973-1985.
Spruck, C.H., 3rd, Ohneseit, P.F., Gonzalez-Zulueta, M., Esrig, D., Miyao, N., Tsai, Y.C.,
Lerner, S.P., Schmutte, C., Yang, A.S., Cote, R., et al. (1994). Two molecular pathways
to transitional cell carcinoma of the bladder. Cancer Res 54, 784-788.
Spruck, C.H., 3rd, Rideout, W.M., 3rd, Olumi, A.F., Ohneseit, P.F., Yang, A.S., Tsai,
Y.C., Nichols, P.W., Horn, T., Hermann, G.G., Steven, K., et al. (1993). Distinct pattern
of p53 mutations in bladder cancer: relationship to tobacco usage. Cancer Res 53, 1162-
1166.
Stein, J.P., Ginsberg, D.A., Grossfeld, G.D., Chatterjee, S.J., Esrig, D., Dickinson, M.G.,
Groshen, S., Taylor, C.R., Jones, P.A., Skinner, D.G., et al. (1998). Effect of
p21WAF1/CIP1 expression on tumor progression in bladder cancer. J Natl Cancer Inst 90,
1072-1079.
Stoehr, R., Wissmann, C., Suzuki, H., Knuechel, R., Krieg, R.C., Klopocki, E., Dahl, E.,
Wild, P., Blaszyk, H., Sauter, G., et al. (2004). Deletions of chromosome 8p and loss of
sFRP1 expression are progression markers of papillary bladder cancer. Lab Invest 84,
465-478.
Subramaniam, M.M., Chan, J.Y., Soong, R., Ito, K., Ito, Y., Yeoh, K.G., Salto-Tellez, M.,
and Putti, T.C. (2008). RUNX3 inactivation by frequent promoter hypermethylation and
protein mislocalization constitute an early event in breast cancer progression. Breast
Cancer Res Treat.
Suter, C.M., Martin, D.I., and Ward, R.L. (2004). Hypomethylation of L1
retrotransposons in colorectal cancer and adjacent normal tissue. Int J Colorectal Dis 19,
95-101.
Suzuki, M., Shigematsu, H., Shames, D.S., Sunaga, N., Takahashi, T., Shivapurkar, N.,
Iizasa, T., Frenkel, E.P., Minna, J.D., Fujisawa, T., et al. (2005). DNA methylation-
associated inactivation of TGFbeta-related genes DRM/Gremlin, RUNX3, and HPP1 in
human cancers. Br J Cancer 93, 1029-1037.
Swaminathan, S., Torino, J.L., and Burger, M.S. (2002). Human urinary bladder
epithelial cells lacking wild-type p53 function are deficient in the repair of 4-
aminobiphenyl-DNA adducts in genomic DNA. Mutat Res 499, 103-117.
Tada, Y., Wada, M., Taguchi, K., Mochida, Y., Kinugawa, N., Tsuneyoshi, M., Naito, S.,
and Kuwano, M. (2002). The association of death-associated protein kinase
hypermethylation with early recurrence in superficial bladder cancers. Cancer Res 62,
4048-4053.
133
Tsai, Y.C., Nichols, P.W., Hiti, A.L., Williams, Z., Skinner, D.G., and Jones, P.A. (1990).
Allelic losses of chromosomes 9, 11, and 17 in human bladder cancer. Cancer Res 50, 44-
47.
Tsai, Y.C., Simoneau, A.R., Spruck, C.H., 3rd, Nichols, P.W., Steven, K., Buckley, J.D.,
and Jones, P.A. (1995). Mosaicism in human epithelium: macroscopic monoclonal
patches cover the urothelium. J Urol 153, 1697-1700.
Tsutsumi, M., Tsai, Y.C., Gonzalgo, M.L., Nichols, P.W., and Jones, P.A. (1998). Early
acquisition of homozygous deletions of p16/p19 during squamous cell carcinogenesis and
genetic mosaicism in bladder cancer. Oncogene 17, 3021-3027.
van Rhijn, B.W., Montironi, R., Zwarthoff, E.C., Jobsis, A.C., and van der Kwast, T.H.
(2002). Frequent FGFR3 mutations in urothelial papilloma. J Pathol 198, 245-251.
van Tilborg, A.A., de Vries, A., de Bont, M., Groenfeld, L.E., van der Kwast, T.H., and
Zwarthoff, E.C. (2000). Molecular evolution of multiple recurrent cancers of the bladder.
Hum Mol Genet 9, 2973-2980.
Velicescu, M., Weisenberger, D.J., Gonzales, F.A., Tsai, Y.C., Nguyen, C.T., and Jones,
P.A. (2002). Cell division is required for de novo methylation of CpG islands in bladder
cancer cells. Cancer Res 62, 2378-2384.
Venter, J.C., Adams, M.D., Myers, E.W., Li, P.W., Mural, R.J., Sutton, G.G., Smith,
H.O., Yandell, M., Evans, C.A., Holt, R.A., et al. (2001). The sequence of the human
genome. Science 291, 1304-1351.
Wada, T., Louhelainen, J., Hemminki, K., Adolfsson, J., Wijkstrom, H., Norming, U.,
Borgstrom, E., Hansson, J., Sandstedt, B., and Steineck, G. (2000). Bladder cancer:
allelic deletions at and around the retinoblastoma tumor suppressor gene in relation to
stage and grade. Clin Cancer Res 6, 610-615.
Waki, T., Tamura, G., Sato, M., Terashima, M., Nishizuka, S., and Motoyama, T. (2003).
Promoter methylation status of DAP-kinase and RUNX3 genes in neoplastic and non-
neoplastic gastric epithelia. Cancer Sci 94, 360-364.
Wallenius, V., Hisaoka, M., Helou, K., Levan, G., Mandahl, N., Meis-Kindblom, J.M.,
Kindblom, L.G., and Jansson, J.O. (2000). Overexpression of the hepatocyte growth
factor (HGF) receptor (Met) and presence of a truncated and activated intracellular HGF
receptor fragment in locally aggressive/malignant human musculoskeletal tumors. Am J
Pathol 156, 821-829.
Wallerand, H., Bakkar, A.A., de Medina, S.G., Pairon, J.C., Yang, Y.C., Vordos, D.,
Bittard, H., Fauconnet, S., Kouyoumdjian, J.C., Jaurand, M.C., et al. (2005). Mutations in
TP53, but not FGFR3, in urothelial cell carcinoma of the bladder are influenced by
134
smoking: contribution of exogenous versus endogenous carcinogens. Carcinogenesis 26,
177-184.
Waterland, R.A., and Jirtle, R.L. (2003). Transposable elements: targets for early
nutritional effects on epigenetic gene regulation. Mol Cell Biol 23, 5293-5300.
Weisenberger, D.J., Siegmund, K.D., Campan, M., Young, J., Long, T.I., Faasse, M.A.,
Kang, G.H., Widschwendter, M., Weener, D., Buchanan, D., et al. (2006). CpG island
methylator phenotype underlies sporadic microsatellite instability and is tightly
associated with BRAF mutation in colorectal cancer. Nat Genet 38, 787-793.
Widschwendter, M., Fiegl, H., Egle, D., Mueller-Holzner, E., Spizzo, G., Marth, C.,
Weisenberger, D.J., Campan, M., Young, J., Jacobs, I., et al. (2007). Epigenetic stem cell
signature in cancer. Nat Genet 39, 157-158.
Widschwendter, M., Siegmund, K.D., Muller, H.M., Fiegl, H., Marth, C., Muller-Holzner,
E., Jones, P.A., and Laird, P.W. (2004). Association of breast cancer DNA methylation
profiles with hormone receptor status and response to tamoxifen. Cancer Res 64, 3807-
3813.
Williams, S.G., and Stein, J.P. (2004). Molecular pathways in bladder cancer. Urol Res
32, 373-385.
Wolff, E.M., Liang, G., Cortez, C.C., Tsai, Y.C., Castelao, J.E., Cortessis, V.K., Tsao-
Wei, D.D., Groshen, S., and Jones, P.A. (2008). RUNX3 Methylation reveals that
Bladder Tumors are Older in Patients with a History of Smoking. Cancer Res 68, In press.
Wolff, E.M., Liang, G., and Jones, P.A. (2005). Mechanisms of Disease: genetic and
epigenetic alterations that drive bladder cancer. Nat Clin Pract Urol 2, 502-510.
Wright, K.O., Messing, E.M., and Reeder, J.E. (2004). DBCCR1 mediates death in
cultured bladder tumor cells. Oncogene 23, 82-90.
Xu, M., Kladde, M.P., Van Etten, J.L., and Simpson, R.T. (1998). Cloning,
characterization and expression of the gene coding for a cytosine-5-DNA
methyltransferase recognizing GpC. Nucleic Acids Res 26, 3961-3966.
Yamamoto, E., Toyota, M., Suzuki, H., Kondo, Y., Sanomura, T., Murayama, Y., Ohe-
Toyota, M., Maruyama, R., Nojima, M., Ashida, M., et al. (2008). LINE-1
hypomethylation is associated with increased CpG island methylation in Helicobacter
pylori-related enlarged-fold gastritis. Cancer Epidemiol Biomarkers Prev 17, 2555-2564.
Yan, P.S., Venkataramu, C., Ibrahim, A., Liu, J.C., Shen, R.Z., Diaz, N.M., Centeno, B.,
Weber, F., Leu, Y.W., Shapiro, C.L., et al. (2006). Mapping geographic zones of cancer
risk with epigenetic biomarkers in normal breast tissue. Clin Cancer Res 12, 6626-6636.
135
Yanagawa, N., Tamura, G., Oizumi, H., Takahashi, N., Shimazaki, Y., and Motoyama, T.
(2003). Promoter hypermethylation of tumor suppressor and tumor-related genes in non-
small cell lung cancers. Cancer Sci 94, 589-592.
Yoo, C.B., and Jones, P.A. (2006). Epigenetic therapy of cancer: past, present and future.
Nat Rev Drug Discov 5, 37-50.
Zhang, Z.T., Pak, J., Huang, H.Y., Shapiro, E., Sun, T.T., Pellicer, A., and Wu, X.R.
(2001). Role of Ha-ras activation in superficial papillary pathway of urothelial tumor
formation. Oncogene 20, 1973-1980.
Zilberman, D., Coleman-Derr, D., Ballinger, T., and Henikoff, S. (2008). Histone H2A.Z
and DNA methylation are mutually antagonistic chromatin marks. Nature 456, 125-129.
Abstract (if available)
Abstract
Successes in the clinic have opened up the era of epigenetic therapy in which the goal is to reactivate genes silenced inappropriately during carcinogenesis. The advantage of using epigenetic therapy to target defects in DNA methylation is that, unlike mutations in the DNA sequence, these alterations are reversible. The focus of this thesis was twofold
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
DNA methylation as a biomarker in human reproductive health and disease
PDF
DNA methylation inhibitors and epigenetic regulation of microRNA expression
PDF
Identification of novel epigenetic biomarkers and microRNAs for cancer therapeutics
PDF
Use of DNA methylation inhibitors for chemotherapy and chemoprevention of cancer
PDF
DNA hypermethylation: its role in colorectal tumorigenesis and potential clinical applications
PDF
Effects of chromatin regulators during carcinogenesis
PDF
Transcriptional regulation by epigenetic mechanisms
PDF
Identification and characterization of PR-Set7 and histone H4 lysine 20 methylation-associated proteins
PDF
Epigenetic regulation of non CPG island gene promoters
PDF
CpG methylation profiling in lung cancer cell lines, tumors and non-tumors
PDF
Functional DNA methylation changes in normal and cancer cells
PDF
The role of microRNAs in cancer
PDF
Cancer epigenetics: linking basic mechanisms to therapy
PDF
The relationship between DNA methylation and transcription factor binding in colon cancer cells
PDF
Identification of DNA methylation markers in diffuse large B-cell lymphoma
PDF
The role of DNA methylation in early detection and progression of pancreatic cancer
PDF
Using epigenetic toggle switches to repress tumor-promoting gene expression
PDF
Development of DNA methylation based biomarkers for the early detection of squamous cell lung cancer
PDF
DNA methylation changes in the development of lung adenocarcinoma
PDF
DNA methylation markers for blood-based detection of small cell lung cancer in mouse models
Asset Metadata
Creator
Wolff, Erika Michele
(author)
Core Title
Epigenetic mechanisms driving bladder cancer
School
Keck School of Medicine
Degree
Doctor of Philosophy
Degree Program
Biochemistry and Molecular Biology
Degree Conferral Date
2009-08
Publication Date
08/05/2009
Defense Date
06/13/2009
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
bladder cancer,epigenetics,methylation,OAI-PMH Harvest
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Jones, Peter A. (
committee chair
), Coetzee, Gerhard A. (
committee member
), Laird, Peter W. (
committee member
), Rice, Judd C. (
committee member
), Shibata, Darryl K. (
committee member
)
Creator Email
erikawol@usc.edu,ewolff@gmail.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-m2487
Unique identifier
UC1155484
Identifier
etd-Wolff-3070 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-181217 (legacy record id),usctheses-m2487 (legacy record id)
Legacy Identifier
etd-Wolff-3070.pdf
Dmrecord
181217
Document Type
Dissertation
Rights
Wolff, Erika Michele
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Repository Name
Libraries, University of Southern California
Repository Location
Los Angeles, California
Repository Email
cisadmin@lib.usc.edu
Tags
bladder cancer
epigenetics
methylation