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
/
DNA hypermethylation: its role in colorectal tumorigenesis and potential clinical applications
(USC Thesis Other)
DNA hypermethylation: its role in colorectal tumorigenesis and potential clinical applications
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
DNA HYPERMETHYLATION: ITS ROLE IN COLORECTAL TUMORIGENESIS
AND POTENTIAL CLINICAL APPLICATIONS
by
Toshinori Hinoue
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)
May 2009
Copyright 2009 Toshinori Hinoue
ii
DEDICATION
This thesis is dedicated to my father Toshifumi Hinoue and my mother Fumiko
Hinoue, who have always believed in me and have offered me unconditional love and
support throughout the years.
iii
ACKNOWLEDGEMENTS
I would like to thank Dr. Peter W. Laird for providing me with the opportunity to
work under his supervision. I have learned tremendously from his insightful advice and
from the criticism given to me and to other members of the laboratory in our weekly lab
meeting. Carefully observing him over the years, I have also gained valuable skills in
presentation and writing. He also inspired me to start running, which has become an
integral part of my life for the last several years and will continue to do so for the rest of
my life.
Dr. Daniel J. Weisenberger and Dr. Mihaela Campan have always been available to
me and have given me great support and encouragement in tough times. I have benefited
greatly from their thoughtful discussions and advice. I also feel fortunate to share interest
in running with them. We have spent a lot time together outside of the laboratory as well.
I am very happy to have been able to share my first half- and full-marathon experience
with them in 2005 and 2006 respectively.
I am grateful to each member of my Ph.D. Guidance/Dissertation committee, Dr.
Gerhard A. Coetzee, Dr. Louis Dubeau, Dr. Judd C. Rice, and Dr. Robert H. Stellwagen,
who have been generous with their time and guidance.
I would also like to thank Dr. Peter A. Jones and each individual who has shared
their research projects with me over the years in the Jones lab meeting on Thursdays at
iv
8am. It has been a valuable experience to discuss my research projects in front of
distinguished professors and an experienced audience.
I would like to thank my fellow graduate students (Nicole M. Sodir, Shirley
Oghamian, Sahar Hooshdaran, and Kwangho Lee) for their friendships and supports. I
am very happy to have been able to share a very important part of my life with them. We
have helped each other, solved problems together, laughed together and cried together.
I would like to also acknowledge other past and present members of the Laird’s
laboratory and Dr. Ite A. Laird-Offringa and members of her laboratory: Daniele V.
Tauriello, Tiffany I. Long, Cindy Lin, Romulo Martin Brena, Michelle Davis, Fei Pan,
Benjamin Berman, Melissa Moffitt, Meleeneh Kazarian, Paul Anglim, Janice S Galler,
Suhaida Selamat, Nikhil Chopra, Charlene Lee, and Devon Pryor. I would like to thank
them for the support and the technical advice. I have had the pleasure of working with
them and sharing great memories throughout the course of my graduate school career.
Finally, I would especially thank Houtan Noushmehr at the Epigenome center,
who helped me to learn R programming by generously sharing his knowledge and
resources with me.
v
TABLE OF CONTENTS
DEDICATION ⅱ
ACKNOWLEDGEMENTS ⅲ
LIST OF FIGURES ⅷ
ABSTRACT ⅹ
CHAPTER1: Molecular genetics and epigenetics of colorectal tumorigenesis
Introduction 1
Types of genetic and epigenetic alterations in colorectal cancer 2
Hereditary and familial predisposition to colorectal cancer 2
High-penetrance rare genetic predisposition 3
Low-penetrance common genetic predisposition 4
Epimutations 5
Major molecular pathways to colorectal cancer 6
Genetic instability 6
Epigenetic regulatory defect 7
CIN vs. CIMP 8
Prognostic significance 9
Role of gene mutations in the development of colorectal cancer 9
Genetic landscape of CRC 9
APC 10
KRAS and BRAF 10
Role of epigenetic changes in the development of colorectal cancer 11
DNA methylation 11
Histone modifications 15
Thesis outline 16
CHAPTER2: Test of a causal role of BRAF
V600E
in the development of CIMP
in colorectal cancer
Introduction 17
Materials and Methods 19
Cell culture and genomic DNA isolation 19
MethyLight analysis of five CIMP-specific markers 19
Mutation analysis and MSI status of colorectal cancer cell lines 20
Western blotting analysis 20
Primary colorectal tissue samples 21
Illumina GoldenGate DNA methylation assay 21
Identification of CpG sites that undergo CIMP-associated DNA
methylation 21
vi
Results 22
Discussion 31
CHAPTER3: CIMP creates a permissible context for the acquisition of
BRAF
V600E
in colorectal cancer
Introduction 33
Materials and Methods 34
Primary colorectal tissue samples 34
Illumina GoldenGate DNA methylation assay 34
Identification of CpG sites specifically methylated in
BRAF
V600E
tumors 35
Cell culture and genomic DNA isolation 35
Quantitative Real-Time RT-PCR 36
Results 36
Table 3.1: The Illumina GoldenGate DNA Methylation targets
specifically methylated in BRAF
V600E
colorectal tumors 38
Discussion 45
CHAPTER4: Comprehensive DNA methylation profiling and identification
of novel CIMP targets in colorectal cancer
Introduction 49
Materials and Methods 50
Primary colorectal tissue samples 50
Genomic DNA and total RNA isolation 51
BRAF and KRAS mutation analysis 51
DNA methylation analyses
MethyLight 51
Illumina Infinium DNA methylation assay 52
Cluster analysis 53
Gene expression analysis
Assessment of RNA integrity 53
Illumina whole genome expression BeadChips 53
Results 55
Discussion 66
CHAPTER5: Identification of a panel of potential DNA methylation-based
biomarkers for colorectal caner
Introduction 68
Materials and Methods 70
Primary colorectal tissue samples 70
Control blood samples 71
DNA Methylation analyses
MethyLight 72
Illumina GoldenGate DNA methylation assay 72
Illumina Infinium DNA methylation assay 72
vii
Results 73
Discussion 84
CHAPTER6: Summary and conclusions 88
BIBLIOGRAPHY 91
Appendix A: MethyLight reaction probe and primer sequences 107
viii
LIST OF FIGURES
Figure 2.1: Characteristics of 21 colorectal cancer cell lines 24
Figure 2.2: Selection of BRAF
V600E
stably-transfected clones and their Illumina
GoldenGate DNA methylation profiles 26
Figure 2.3: Illumina GoldenGate DNA methylation profiles of CIMP-associated
loci in 58 primary colorectal tumors and COLO 320DM cells, empty
vector and BRAF
V600E
transfected COLO 320DM clones 29
Figure 2.4: Changes in DNA methylation levels over passages in BRAF
V600E
and EVC stably-transfected clones 30
Figure 3.1: IGFBP7 promoter DNA methylation in human colorectal cancers 43
Figure 3.2: Analysis of DNA methylation and mRNA expression of IGFBP7 in
colorectal cancer cell lines 44
Figure 4.1: CIMP and covariate analyses of 100 colorectal tumor samples 56
Figure 4.2: Unsupervised hierarchical clustering analysis of 100 colorectal
tumor samples 59
Figure 4.3: Identification of CIMP-associated CpG sites 60
Figure 4.4: Identification of cancer-specific non-CIMP-associated Infinium
probes 62
Figure 4.5: CIMP-specific DNA hypermethylation and embryonic stem cell
polycomb targets 64
Figure 4.6: Correlation heatmap based on the expression data 65
Figure 5.1: Unsupervised hierarchical clustering of DNA methylation data
obtained from the Infinium assay 74
Figure 5.2: Top 50 ranked markers from the Illumina Infinium analysis 77
Figure 5.3: Flow chart of the comprehensive DNA Methylation marker
discovery 78
Figure 5.4: Validation of the candidate markers by MethyLight 80
ix
Figure 5.5: Stringent filtering of the 31 DNA methylation markers against blood
samples 82
Figure 5.6: The stringent set of 8 candidate biomarkers 83
x
ABSTRACT
Aberrant DNA hypermethylation of CpG islands is a common and early event in
colorectal tumorigenesis. Promoter DNA hypermethylation is associated with
transcriptional gene silencing, and can contribute to tumorigenesis when it occurs at a
critical tumor suppressor gene. A distinct subset of colorectal cancers display the CpG
island methylator phenotype (CIMP), characterized by an exceptionally high frequency
of cancer-specific DNA hypermethylation. Recent studies have shown that an activating
mutation of BRAF (BRAF
V600E
) is specifically associated with CIMP. We used colorectal
cancer cell lines and primary tumors to elucidate the molecular mechanisms for the
association. We first examined whether expression of BRAF
V600E
causes CIMP-specific
DNA hypermethylation in the CIMP-negative, BRAF wild-type COLO 320DM colorectal
cancer cell line. We found that stable expression of BRAF
V600E
is not sufficient to induce
CIMP in our system. Secondly, considering the alternative possibility, we searched for
genes whose DNA hypermethylation was tightly linked to BRAF
V600E
and CIMP in
colorectal cancer. Intriguingly, we identified CIMP-dependent DNA hypermethylation
and transcriptional inactivation of IGFBP7, a mediator of BRAF
V600E
-induced cellular
senescence and apoptosis. Inactivation of IGFBP7 by DNA hypermethylation may
accommodate BRAF
V600E
by blocking the senescence pathway. Therefore, our work
provides a mechanistic rationale for the association between BRAF
V600E
and CIMP.
Moreover, in an attempt to further characterize CIMP-associated DNA hypermethylation,
we have quantitatively determined DNA methylation status of 27,578 CpG sites located
in 14,495 gene promoters in 100 colorectal tumors and 8 normal mucosae. Through
xi
stringent statistical methods, we have identified cancer-specific DNA hypermethylation
of 1,036 genes, of which 601 show CIMP-specific DNA hypermethylation. We have also
generated gene expression data. Our most comprehensive list of CIMP targets and their
expression data will help us understand the role of CIMP-associated DNA
hypermethylation in colorectal tumorigenesis, and provide the opportunity to study
structural and sequence characteristics of affected CpG islands. The latter might
ultimately give us insights into the underlying basis of CIMP-associated DNA
hypermethylation. Finally, we have developed a panel of novel cancer-specific DNA
methylation markers, which would potentially be useful for early diagnosis of colorectal
cancer using a noninvasive stool DNA test or a minimally invasive blood-based assay.
1
CHAPTER 1
Molecular genetics and epigenetics of colorectal tumorigenesis
Introduction
Colorectal cancer (CRC) is the third most common type of cancer in both men
and women and the second leading cause of cancer-related death in the U.S. In 2008,
there were an estimated 148,810 new cases and 49,960 deaths from this disease. (SEER,
2008; Jemal et al., 2008). Although, the incidence of CRC is declining in the U.S, parts of
the Asian countries including Japan, Taiwan, and Eastern Europe have continued to see
rising incidence and mortality (Sung et al., 2008). CRC represents a heterogeneous
disease (Jass, 2007a). Development of CRC typically takes place over a long period of
time through multiple distinct pathways (Jones et al., 2008; East et al., 2008). The basis
of the heterogeneity can be attributed to the accumulation of a distinct set of genetic and
epigenetic alterations during the course of its development. Both the genetic and
epigenetic alterations can occur as either inherited (germline) mutations or acquired
(somatic) mutations, which affect critical genes/pathways important for the tumor
initiation and progression (Hanahan and Weinberg, 2000; Vogelstein and Kinzler, 2004).
In this chapter, I will discuss proposed molecular pathways leading to CRC, different
types of molecular (genetic and epigenetic) alterations, and their role in colorectal
tumorigenesis. Finally, the outline of my thesis will be presented.
2
Types of genetic and epigenetic alterations in colorectal cancer
The genetic alterations commonly observed in colorectal cancers are largely
categorized into sequence and chromosomal copy number alterations. Sequence
alterations include single nucleotide changes involving point mutations (Wood et al.,
2007) and small insertions or deletions by slippage mutation due to mismatch repair
deficiency (Mori et al., 2001). Copy number alterations in colorectal cancer are mainly
represented as deletions and amplifications of large chromosomal regions (Leary et al.,
2008). The majority of colorectal cancer display aneuploidy. On the other hand, two
major types of epigenetic modifications that have been found in colorectal cancer are
DNA methylation and histone modifications (Jones and Baylin, 2007). Global DNA
hypomethylation and CpG island hypermethylation have been studied extensively and
play multiple roles in colorectal tumorigenesis. Specifically, transcriptional silencing of
critical genes by DNA hypermethylation and its contribution to the development of
colorectal cancer will be discussed in detail and will be the main focus of my thesis.
Hereditary and familial predisposition to colorectal cancer
Although, it is estimated that 20%-25% of all CRC show a familial clustering,
only 5-6% of CRC cases are currently attributable to a known genetic predisposition (de
la Chapelle, 2004; Rustgi, 2007). Lifetime risk of being diagnosed with CRC is
approximately 5% for an individual without a family history of CRC (SEER, 2008).
However, individuals with an inherited high-risk germline mutation face 40-100%
lifetime risk of developing CRC.
3
High-penetrance rare genetic predisposition. Two of the most well-documented
inherited forms of CRC occur in individuals with Lynch syndrome (or hereditary
nonpolyposis colorectal cancer (HNPCC)) and Familial adenomatous polyposis (FAP).
Both are inherited in an autosomal dominant manner. Lynch syndrome accounts for 3-4%
of CRC cases. Patients with Lynch syndrome who carry a germline mutation in genes
involved in the mismatch repair (in particular, MLH1, MSH2, MSH6, and PMS2) have an
up to 80% lifetime risk of developing CRC (Rustgi, 2007). Colorectal cancers in Lynch
syndrome exhibit microsatellite instability due to mismatch repair deficiency. FAP, on
the other hand, is found in nearly 1% of CRC patents and is caused by the inheritance of
a germline mutation in the APC gene. Patients with FAP typically develop numerous (up
to thousands) adenomatous polyps throughout the colon during their teenage years. For
patients with FAP, development of colorectal cancer is inevitable (100%) by the age of
40 (Rustgi, 2007). Two rather uncommon autosomal dominant hereditary syndromes,
familial juvenile polyposis syndrome (FJP) and Peutz–Jeghers syndrome (PJS), are
characterized by hamartomatous polyposis, which increases the risk of developing CRC
to 40-60% (Kaz and Brentnall, 2006). Disruption of the TGF-β signaling pathway
appears to be involved in the development of FJP. Germline mutations are reported in
either BMPR1A (bone morphogenic protein receptor 1A), or SMAD4, or ENG (endoglin,
an accessory receptor for TGF-β). By contrast, with PJS, germline mutation of tumor
suppressor gene STK11, a serine–threonine protein kinase, is present in 40–60% of
individuals. However, the exact function of STK11 has not been determined (Rustgi,
2007).
4
Familial CRC also occurs as an autosomal recessive or co-dominant condition. A
subset of patients with adenomatous polyposis syndrome (5 to 10%) without APC
mutations often carry biallelic germline mutations in MUTYH (designated as a MUTYH-
associated polyposis or MAP). MUTYH encodes for a protein that functions as an
adenine-to-guanine-specific DNA glycosylase, which is involved in base excision repair
(Kaz and Brentnall, 2006; Grady and Carethers, 2008). A co-dominant mode of
inheritance has been proposed for the development of hyperplastic polyposis syndrome
(HPS), which is estimated to affect about 1/2000 subjects (Young et al., 2007).
Importantly, patients with HPS develop multiple sessile serrated adenomas (SSA),
proposed precursor lesions of a distinct subset of CRC characterized by the CpG island
methylator phenotype (East et al., 2008). The risk of developing CRC in individuals with
HPS is estimated to be as high as 50% and multiple synchronous or metachronous
cancers may develop (Makinen, 2007). Search for the genetic variant associated with
HPS is underway (Young et al., 2007).
Low-penetrance common genetic predisposition. Genome-wide association
(GWA) studies have been conducted in order to understand the basis of the largely
uncharacterized familial predisposition to CRC (Broderick et al., 2007; Tomlinson et al.,
2007; Zanke et al., 2007; Houlston et al., 2008; Jaeger et al., 2008; Neklason et al., 2008;
Tenesa et al., 2008; Tomlinson et al., 2008). These studies corroborate the notion that a
part of the familial clustering of CRC is attributable to a combination of more common,
low-penetrance genetic predisposition and perhaps environmental influences (de la
Chapelle, 2004). Ten CRC susceptibility loci, most of which are not associated with
5
protein coding sequences, have been reported to date (Houlston et al., 2008) and await
further functional characterizations.
Recently, reduced germline allele-specific expression of TGFBR1 has been found
in a large proportion (~20%) of CRC patients with positive family history (Valle et al.,
2008). The odds ratio of CRC is estimated to be 8.7. Molecular mechanisms of altered
expression of TGFBR1 in colorectal tumorigenesis have also been investigated in mice
(Zeng et al., 2009). However, a mutation responsible for the reduced expression of
TGFBR1 has not been discovered.
Epimutations. Transcriptional silencing of tumor suppressor genes by aberrant
DNA hypermethylation contributes to the development of CRC (detailed discussion
below). Epigenetic information including DNA methylation is generally cleared between
generations through the reprogramming process that takes place in primordial germ cells
and early embryos (Santos and Dean, 2004). Intriguingly, however, there have been
reports suggesting that some cases of Lynch syndrome can be attributable to germline
epigenetic alterations. It has been found that the aberrant DNA hypermethylation and
inactivation of the MLH1 and MSH2 gene is heritable over generations (Suter et al., 2004;
Hitchins et al., 2007; Chan et al., 2006). However, the mechanisms of such germline
epimutation in human is contentious as to whether such epimutations can be directly
transferred across generations or indirectly generated through a cis-acting or trans-acting
genetic mutation (Chong et al., 2007; Horsthemke, 2007; Gosden and Feinberg, 2007;
Suter and Martin, 2007; Ligtenberg et al., 2008).
6
Major molecular pathways to colorectal cancer
Genetic instability. Approximately 70% of CRC are sporadic without evidence of
familial predisposition. Earlier investigations of two forms of hereditary colorectal
cancer, FAP and Lynch syndrome, initially provided insight into different molecular
mechanisms underlying colorectal tumorigenesis (Kinzler and Vogelstein, 1996). Two
major forms of genetic instability involve chromosome instability (CIN), found in FAP,
and microsatellite instability (MIN, MSI+), found in Lynch syndrome. The majority of
sporadic CRC (80-85%) exhibit CIN, characterized by aneuploidy and chromosome
rearrangements (Grady and Carethers, 2008). The molecular basis of CIN is largely
unknown, however several potential mechanisms have been suggested and genes
responsible for chromosome transmission fidelity are suspected to be involved (Grady
and Carethers, 2008). Recently, the DNA sequence of these candidate genes were
examined in order to elucidate the molecular mechanisms responsible for CIN in CRC
(Barber et al., 2008). Mutations were discovered specifically in a number of genes
involved in sister chromatid cohesion. Further studies of such genes in a cell cultures
system corroborate the hypothesis that a defect in sister chromatid cohesion may
represent a major mechanism for CIN in CRC (Barber et al., 2008).
MSI+ CRCs, found in 15-20% of sporadic cases, tend to be diploid and are
characterized by an abnormal expansion and contraction of simple nucleotide repeats due
to strand slippage during DNA replication. As in the case of tumor in Lynch syndrome,
mismatch repair gene inactivation underlies MSI+. MSI+ in sporadic CRC is mainly
caused by epigenetic silencing of MLH1 by promoter CpG island DNA
7
hypermethylation. Genes containing simple nucleotide repeats are specifically vulnerable
to the slippage mutations. Most frequent MSI target genes with suspected tumor
suppressor functions include TGFBR2 (containing A10 repeat, ~85% mutation
frequency), IGF2R (G8, 10%), BAX (G8, 50%), and other member of mismatch repair
genes MSH3 (A8, 50%) and MSH6 (C8, 33%) (Grady and Carethers, 2008). Other genes
have also been found to be altered by the slippage mutations (including nine genes in
>20% of MSI+ tumors) (Mori et al., 2001; Grady and Carethers, 2008). However, the
importance of the mutations of these genes in colorectal tumorigenesis have not been
elucidated.
Epigenetic regulatory defect. Aberrant DNA methylation at CpG islands has been
widely observed in cancer (Jones and Baylin, 2002). At least 15% of all CRC cases have
been described to have the CpG island methylator phenotype (CIMP), characterized by a
high frequency of DNA hypermethylation in a specific group of CpG islands (Toyota et
al., 1999; Issa, 2004; Samowitz et al., 2005; Weisenberger et al., 2006). Features
associated with CRC with CIMP include female sex, proximal location, and poorly
differentiated or mucinous histology (Kambara et al., 2004; Issa, 2004; Samowitz et al.,
2005; Weisenberger et al., 2006). Intriguingly, our study using a newly developed CIMP
marker panel in colorectal cancers found that MLH1 DNA hypermethylation occurs
almost exclusively in CIMP+ tumors (Weisenberger et al., 2006). Therefore, it appears
that sporadic MSI+ is caused by an implied epigenetic regulatory defect occurring in
CIMP that in certain circumstances affects MLH1 (Weisenberger et al., 2006).
Furthermore, our study found a strong association of CIMP+ with the presence of an
8
activated mutant form of BRAF (BRAF
V600E
) (Odds ratio = 203) (Weisenberger et al.,
2006). Importantly, this BRAF mutation is not simply caused by a mismatch repair
deficiency due to CIMP-specific inactivation of MLH1, since it is not found in Lynch
syndrome with germline mutations in mismatch repair genes (Domingo et al., 2004). The
molecular mechanism that accounts for CIMP is not known. In one study, cigarette
smoking was found to be associated with increased risk of developing CIMP+ CRC in a
dose dependent fashion (Samowitz et al., 2006)
CIN vs. CIMP. Colorectal cancers with CIN and CIMP have shown to be
inversely correlated (Goel et al., 2007; Cheng et al., 2008) and appear to develop in two
separate pathways (Jass, 2007b). Two major types of colorectal polyps exist based in
part on their morphological appearances, adenomatous polyps and serrated polyps.
Serrated polyps are rather heterogeneous and comprise a number of histological subtypes
(Lauwers and Chung, 2006). Evidence is unequivocal that both types of polyps have
malignant potential (East et al., 2008). It is well accepted that adenomatous polyps are
precursors for the majority of CRCs characterized by CIN. Both inherited (FAP) and
sporadic CRC with CIN are believed to follow the adenomas carcinoma sequence
initiated by inactivation of the APC gene. This stepwise progression model involves
acquisition of distinct sets of genetic alterations (including mutations in APC, TP53, and
KRAS) at each step of the morphological transformation, from benign adenomatous
polyps to adenocarcinoma (Vogelstein et al., 1988). On the other hand, accumulating
evidence indicates that CIMP in CRC arises through a distinct pathway originating in
sessile serrated adenomas (SSAs), a subtype of serrated polyps. (Kambara et al., 2004;
9
O'Brien et al., 2006). The precise natural history and rate of progression in the serrated
pathway to CRC is not well characterized. However, strong evidence comes from the
recognition that SSAs show the histological and molecular features associated with
CIMP+ tumors. In one study, SSAs represented 9% of all colorectal polyps and were
more frequently found in the proximal colon (Spring et al., 2006). BRAF mutations and
DNA hypermethylation of CIMP-associated loci, including MLH1, are very common in
SSAs (O'Brien et al., 2004; Spring et al., 2006).
Prognostic significance. CRCs with CIN have been reported to have poor clinical
outcomes (Walther et al., 2008). By contrast, CIMP has been found to confer variable
prognosis in CRC depending on the molecular features associated with it. MSI+ tumors,
which are a subset of CIMP due to MLH1 DNA hypermethylation, have generally been
associated with significantly better prognosis compared to that of CRCs without
mismatch repair deficiency (Popat et al., 2005). The presence of BRAF mutation in
CIMP+ colorectal tumors has an adverse prognostic effect (Ogino et al., 2009). In one
study, the poor clinical outcome generally associated with the presence of the BRAF
V600E
did not supersede the good prognosis associated with MSI+ (Samowitz et al., 2005).
Role of gene mutations in the development of colorectal cancer
Genetic landscape of CRC. Recent sequencing analysis of coding regions of
18,191 unique genes in the Reference Sequence (RefSeq) database has provided us with a
comprehensive view of the genetic landscape of CRC (Wood et al., 2007). It has been
found that there are several commonly mutated genes including APC, TP53, and KRAS,
10
which have already been studied extensively in the past. However, in addition, there is a
much larger number of genes mutated in a small proportion (>5%) of tumors. Essentially,
each tumor has a unique sets of mutations (median = ~80), with only a handful of
mutated genes in common between any two tumors. Through statistical analysis, Wood
and colleagues concluded that most mutations are harmless passenger mutations and each
tumor carries <15 mutations that may contribute to CRC development (Wood et al.,
2007). Importantly, genetic diversity revealed by this study underscores the heterogeneity
of CRC (Velculescu, 2008).
APC. Germline mutation of the APC gene is responsible for the inherited CRC
syndrome FAP. APC mutations are the most common genetic alterations found in more
than 60% of sporadic CRCs (Powell et al., 1992; Samowitz et al., 2007) For the majority
of CRCs, APC mutation is considered to be a initiating event during tumorigenesis
(described as a gatekeeper mutation) (Powell et al., 1992; Kinzler and Vogelstein, 1996;
Samowitz et al., 2007). APC functions as a negative regulator of the Wnt signaling
pathway and APC inactivation is the most frequently employed mechanism for Wnt
signaling hyepractivation in CRC (Schneikert and Behrens, 2007). Importantly, APC
mutations are significantly less frequent in CRC with CIMP (and BRAF mutation)
(Samowitz et al., 2007), suggesting an alternative mechanisms may be utilized for Wnt
signaling hyepractivation or Wnt signaling hyepractivation is less important in the
development of CRC with CIMP.
KRAS and BRAF. The RAS-RAF-MEK-ERK signaling pathway is frequently
hyperactivated in colorectal cancer. KRAS mutations occur most frequently in 30-40% of
11
all colorectal cancer (Oliveira et al., 2004), and BRAF mutations
are present at a
frequency of 5-22%, in which the constitutively activated BRAF
V600E
variant accounts for
~90% of all the BRAF mutations (Garnett and Marais, 2004). Mutations in KRAS and
BRAF are generally mutually exclusive, implying equivalent downstream effects in
tumorigenesis (Davies et al., 2002). However, recent studies indicate that mutations of
these genes might play distinct roles in tumor development or maintenance (Haigis et al.,
2007; Rosenberg et al., 2007). Furthermore, a tight association between BRAF mutations
and CIMP has been documented in CRC (Weisenberger et al., 2006).
Role of epigenetic changes in the development of colorectal cancer
Epigenetic modifications play an important role during normal development and are
thought to be involved in various diseases (Feinberg, 2008). Two major types of
epigenetic modifications that have been found to be altered in cancer are DNA
methylation and covalent histone modifications (Jones and Baylin, 2007).
DNA methylation. DNA methylation occurs almost exclusively in the context of
CpG dinucleotides in mammals and is involved in genomic imprinting and X inactivation
(Bird, 2002). Mammalian genomes are globally methylated except for CpG-rich clusters
(~1,000 bp) known as CpG islands. Although CpG islands contain only a small fraction
of the CpGs present in the human genome (1-2% of total CpG), they are frequently found
in promoter regions of ~60% of human genes (Bird, 2002). Up to 4% of promoter
associated CpG islands are found to be differentially methylated in normal tissues,
indicating that CpG island DNA methylation can be utilized to regulate gene expression
12
during the normal development (Shen et al., 2007; Illingworth et al., 2008). Altered
genomic distribution of DNA methylation in cancer was originally discovered in 1980s
(Feinberg and Tycko, 2004). Global DNA hypomethylation and CpG island
hypermethylation have been studied extensively and certainly play multiple roles in
colorectal tumorigenesis (Feinberg et al., 2006).
DNA hypomethylation has been shown to result in genetic instability and
accelerates cancer development in mice (Eden et al., 2003; Gaudet et al., 2003). Notably,
in a mouse model of intestinal tumorigenesis, DNA hypomethylation in Apc
Min/+
mice
increases the numbers of intestinal microadenoma through loss of heterozygosity at Apc,
while suppressing the overall incidence of macroscopic intestinal tumors, suggesting a
role for DNA hypomethylation in the initiation of intestinal tumorigenesis (Yamada et al.,
2005). In addition, gene-specific DNA hypomethylation and associated gene activation
has also been described to contribute to the development of cancer (Feinberg, 2007). In
particular, loss of imprinting (LOI) of the insulin-like growth factor II (IGF2) gene
(activation of the normally silenced maternal allele) occurs through DNA
hypomethylation. LOI of IGF2 is found in normal appearing colonic mucosa of patients
with adenoma and CRC (Cui et al., 2003; Woodson et al., 2004). LOI of IGF2 has been
implicated as an important initiation event in colorectal tumorigenesis in mice and human
(Sakatani et al., 2005).
Aberrant DNA methylation at promoter CpG islands has been extensively studied
in cancer. Inactivation of selected tumor-suppressor genes in this pathway has now been
accepted as a critical contributor to the development of cancer (Jones and Baylin, 2007).
13
Cancer-specific DNA hypermethylation has been reported in aberrant crypt foci (ACF),
the earliest detectable lesion in the colonic mucosa (Chan et al., 2002). DNA
hypermethylation perhaps cooperates with other genetic mechanisms such as mutation or
deletion to alter key signaling pathways critical in colorectal tumorigenesis (Baylin and
Ohm, 2006). For instance, multiple members of the secreted frizzled-related proteins
(SFRPs) gene family are subject to inactivation by DNA hypermethylation in CRC
(Baylin and Ohm, 2006). Activation of the Wnt signaling pathway is a hallmark of CRC
(Schneikert and Behrens, 2007) and SFRPs function as a negative regulator of Wnt
signaling. Genetically, loss of function mutation or deletion of APC (germline or somatic)
or activation of CTNNB1 (β-catenin) are common mechanisms for Wnt signaling
activation in CRC (Baylin and Ohm, 2006; Schneikert and Behrens, 2007). In addition,
germline mutations of MLH1 is linked to Lynch syndrome, which leads to the
development of CRC with MSI+ at young age (discussed above). By contrast, biallelic
promoter DNA hypermethylation of MLH1 is the dominant mechanism for the
development of the sporadic MSI+ colorectal tumor (Veigl et al., 1998). Furthermore,
recent larger scale comparison between genes hypermethylated and mutated in CRC
revealed significant overlap between these two alterations. Importantly, DNA
hypemethylation appeared to be the preferred mechanism of gene inactivation when the
individual gene was examined (Schuebel et al., 2007; Chan et al., 2008).
MicroRNAs (miRNAs) control multiple gene targets by causing mRNA
degradation or by inhibiting translation. Altered expression (up- and down-regulation) of
various miRNAs has been implicated in different types of cancer (Esquela-Kerscher and
14
Slack, 2006). Intriguingly, miRNAs associated with CpG islands have shown to be
regulated by DNA methylation (Saito et al., 2006). Several CpG island-associated
miRNAs (miR-124a, miR-34a/b, and miR-342) that function as tumor suppressors have
been shown to acquire DNA hypermethylation and to be transcriptionally silenced in
CRC (Lujambio et al., 2007; Toyota et al., 2008; Grady et al., 2008).
New insights into the mechanisms and the role of CpG island hypermethylation in
cancer came out of recent studies using integrated analyses of different epigenetic
modifications. We and other groups have initially reported that genes that are targeted by
Polycomb group (PcG) proteins in embryonic stem cells are more susceptible to aberrant
DNA hypermethylation in cancer (Ohm et al., 2007; Schlesinger et al., 2007;
Widschwendter et al., 2007). PcG proteins control the expression of genes critical in
development including lineage-specific genes (Bernstein et al., 2007). PcG target genes
are characterized by H3K27me3 histone modification and are maintained at a low
expression state (Bernstein et al., 2007). Recently, it has been observed that genes
targeted by H3K27me3 in normal cells gain DNA hypermethylation and lose their
H3K27me3 mark in cancer (Gal-Yam et al., 2008; Rodriguez et al., 2008). Importantly,
epigenetic switching of H3K27me3 and DNA methylation occurs at genes that are not
expressed in normal cells (Gal-Yam et al., 2008; Rodriguez et al., 2008). The functional
significance of the replacement of these two epigenetic modifications in cancer is not
fully understood. However it has been proposed that the permanent and heritable
silencing of critical genes by DNA hypermethylation may affect cellular plasticity by
blocking proper differentiation (Widschwendter et al., 2007; Gal-Yam et al., 2008).
15
Histone modifications. Global loss of trimethylation of histone H4 Lys20
(H4K20me3) and acetylation of histone H4 Lys16 (H4K16Ac) have been reported in
various cancers including colorectal cancer (Fraga et al., 2005). Loss of these
modifications are observed along with DNA hypomethylation in repeat DNA sequences
such as juxtacentromeric Satellite 2 (Sat2) sequences and non-satellite subtelomeric
NBL2 and D4Z4 repeats (Fraga et al., 2005). In addition, enzymes responsible for
different types of histone modifications (methylation, acetylation, and deacetylation) as
well as proteins that read the specific covalent histone modifications (HP1 for H3K9,
BMI1 for K3K27, and ING proteins for K3K4) are also aberrantly expressed in many
cancer cells (Wang et al., 2007).
With notable exceptions, both di- and trimethylated H3K9, H3K27me3, and
H4K20me3 are generally enriched in heterochromatin and are associated with
transcriptional silencing (Bernstein et al., 2007). Notably, CpG islands that acquire DNA
hypermethylation during tumorigenesis are also found to be marked by H3K9me2
(Nguyen et al., 2002; Kondo et al., 2004). Moreover, it has been shown that the
H3K9me2 mark is removed following 5-aza-2’-deoxycytidine treatment, a DNA
methylation inhibitor, in colorectal cancer cell lines, indicating that the two modifications
might work together in the same gene silencing pathway (McGarvey et al., 2006).
Trimethylation of H3K27 is mediated by a component of the Polycomb repressive
complex (PRC), EZH2. Overexpression of EZH2 has been observed in many different
types of cancer (Wang et al., 2007). Genomic deletion of miRNA-101 has been proposed
as a mechanism for EZH2 overexpression (Varambally et al., 2008). Recently, EZH2-
16
mediated H3K27me3 gene silencing has been observed in cancer, which is largely
independent of DNA hypermethylation (Kondo et al., 2008; Gal-Yam et al., 2008).
H3K27me3 tend to be targeted to non- CpG island promoters and can effectively
inactivate tumor suppressor genes such as DACT3, a negative regulator of Wnt signaling
pathway, without DNA hypermethylation in CRC (Kondo et al., 2008; Jiang et al., 2008).
Thesis outline
This thesis focuses on one of the epigenetic alterations in CRC, DNA
hypermethylation. Chapter 2 and Chapter 3 examine mechanisms linking CIMP and
BRAF mutation in CRC. Two hypotheses in particular will be discussed in these chapters.
Chapter 4 employs the most comprehensive DNA methylation profiling of CRC for
identification of novel CIMP targets, which will be useful in the future in better
understanding the role of CIMP-associated DNA hypermethylation in colorectal
tumorigenesis, and studying structural and sequence characteristics of affected CpG
islands. Chapter 5 describes a development of a panel of potential DNA methylation
based biomarkers for CRC screening. Chapter 6 concludes the thesis with summary of the
results and discussion with future research proposals.
17
CHAPTER 2
Test of a causal role of BRAF
V600E
in the development
of CIMP in colorectal cancer
Introduction
Aberrant DNA methylation at CpG islands has been widely observed in cancer.
Promoter CpG island hypermethylation associated with inactivation of selected tumor
suppressor genes appears to be a critical event from the earliest stages of tumor
development to the maintenance of the tumor phenotype (Jones and Baylin, 2007).
Distinct subgroups of several types of human cancers have been proposed to have a CpG
island methylator phenotype (CIMP), in which an exceptionally high frequency of
cancer-specific DNA hypermethylation is found (Toyota et al., 1999; Issa, 2004).
Although this concept has been controversial (Yamashita et al., 2003), we recently
confirmed the existence of CIMP in colorectal cancer in a large-scale comprehensive
study (Weisenberger et al., 2006).
CIMP in colorectal cancer may arise through a distinct pathway originating in
certain subtypes of serrated polyps (Kambara et al., 2004; O'Brien et al., 2006) and
accounts for at least 15% of all colorectal cancer cases (Samowitz et al., 2005;
Weisenberger et al., 2006). Features associated with CIMP colorectal cancers include
female sex, proximal location, and poorly differentiated or mucinous histology (Kambara
et al., 2004; Issa, 2004; Samowitz et al., 2005; Weisenberger et al., 2006). Our study
using a newly developed CIMP marker panel in colorectal cancers demonstrated that
18
sporadic microsatellite instability (MSI+) occurs as a consequence of CIMP-associated
MLH1 methylation and we found an extremely strong association of CIMP with the
presence of an activated mutant form of BRAF (BRAF
V600E
) (Odds ratio = 203)
(Weisenberger et al., 2006). Both CIMP and BRAF mutation have been reported in the
earliest stages of colorectal neoplasia; CIMP in the apparently normal mucosa of patients
predisposed to multiple serrated polyps (Minoo et al., 2006) and BRAF mutation in
diminutive lesions called aberrant crypt foci (Rosenberg et al., 2007).
The RAS-RAF-MEK-ERK signaling pathway is frequently hyperactivated in
colorectal cancer. KRAS mutations occur most frequently in 30-40% of all colorectal
cancer (Oliveira et al., 2004), and BRAF mutations
are present at a frequency of 5-22%,
in which the constitutively activated BRAF
V600E
variant accounts for ~90% of all the
BRAF mutations (Garnett and Marais, 2004). Mutations in KRAS and BRAF are generally
mutually exclusive, implying equivalent downstream effects in tumorigenesis (Davies et
al., 2002). However, recent studies indicate that mutations of these genes might play
distinct roles in tumor development or maintenance (Haigis et al., 2007; Rosenberg et al.,
2007). The extremely tight associations between BRAF
V600E
and CIMP raise the
possibility that BRAF
V600E
might represent a specific molecular defect that results in
hypermethylation of multiple CpG islands, and therefore play a causal role in the
development of CIMP+ tumor. Interestingly, it has been reported that an aberrant RAS
signaling pathway may alter the expression of DNA methyltransferase 1 (Ordway et al.,
2004). Furthermore, inhibition of MEK-ERK by chemical inhibitors or siRNAs result in
decrease in DNA methylation at promoters of CDKN2A (p16
INK4A
) and CDKN1A
19
(p21
WAP1
) in SW1116 colorectal cancer cell line (Lu et al., 2007). In this chapter, we
sought to determine whether expression of BRAF
V600E
would induce DNA
hypermethylation at loci that are associated with CIMP in an in vitro cell culture system.
Materials and Methods
Cell culture and genomic DNA isolation
Colorectal cancer cell lines were obtained from American Type Culture Collection
(Manassas, VA). COLO 320DM cells were grown in DMEM supplemented with 10%
FBS, 1mM glutamine. An empty vector and an HA-tagged BRAF
V600E
cDNA clone
(pMEV-HA, pMEV-BRAF-V599E, Biomyx Technology San Diego, CA) were
transfected into COLO 320DM cells using Lipofectamine2000 (Invitrogen, Burlington,
ON). G418 (1 mg/ml) was added 48 hours after transfection, and resistant clones were
randomly isolated and expanded. Stable expressing clones were maintained in 500 µg/ml
of G418. Genomic DNA from each cell line was isolated as described previously (Laird
et al., 1991).
MethyLight analysis of five CIMP-specific markers
Genomic DNA was treated with sodium bisulfite and subsequently analyzed by
MethyLight as described (Eads et al., 2000; Weisenberger et al., 2005; Weisenberger et
al., 2006). The primer and probe sequences for the MethyLight reactions are listed in
Appendix A. The results of MethyLight analyses were scored as PMR (Percent of
20
Methylated Reference) values as previously defined (Weisenberger et al., 2006).
Mutation analysis and MSI status of colorectal cancer cell lines
Primer sequences and PCR conditions for direct sequencing of BRAF at codon 600 in
exon 15 and at codons 12 and 13 KRAS in exon 2 were reported previously (Davies et al.,
2002). The MSI status of each cell line was based on the Sanger Institute Cancer Genome
Project (http://www.sanger.ac.uk/) and based on a previous study (Suter et al., 2003).
Western blotting analysis
Whole cell extracts were prepared from each resistant clone at the first passage using
CelLytic M Cell Lysis Reagent (Sigma-Aldrich, St. Louis, MO). Equal amounts of
protein from whole cell extracts were separated on gradient (4-20%) polyacrylamide gels
(Invitrogen) and then transferred to polyvinylidene difluoride (PVDF) membranes (Bio-
Rad, Hercules, CA). Blots were probed with the anti-HA antibodies (Roche, Indianapolis,
IN) for HA-BRAF
V600E
, anti-phospho-ERK1/2 (Cell Signaling, Beverly, MA), and anti-
ERK1 (Santa Cruz Biotechnology, Santa Cruz, CA) followed by incubation with species
specific horseradish peroxidase-conjugated secondary antibodies (Santa Cruz). Proteins
were visualized using SuperSignal West Pico Chemiluminescent Substrate (Pierce,
Rockford, IL).
21
Primary colorectal tissue samples
Primary colorectal tissue samples were collected and DNA was extracted as previously
described (Weisenberger et al., 2006). The 58 sample set includes five CIMP+ tumors,
five CIMP– tumors and 48 randomly selected tumors as indicated previously
(Weisenberger et al., 2006). CIMP status and BRAF mutation status for each tumor
sample was previously determined (Weisenberger et al., 2006).
Illumina GoldenGate DNA methylation assay
Genomic DNA was sodium bisulfite converted using the EZ-96 DNA Methylation Kit
(ZYMO Research, Orange, CA) according to manufacturer’s protocol. Illumina
GoldenGate DNA methylation analyses were performed as described previously
(Bibikova et al., 2006) at the USC Genomics Core Facility. Target sequences for the
assay and detailed information on each interrogated CpG site and its associated gene on
the “GoldenGate Methylation Cancer Panel 1” are described at www.illumina.com.
Identification of CpG sites that undergo CIMP-associated DNA methylation
To determine CIMP-associated DNA methylation among 1,505 interrogated loci in the
GoldenGate Methylation Cancer Panel 1, we first screened 58 primary colon tumor
samples. We then performed hierarchical two-dimensional unsupervised clustering
analyses using the β-values (84 X-linked reactions were omitted, resulting in 1,421
reactions) and JMP 6.0 software (SAS Institute, Cary, NC). A distinct cluster of 11 tumor
samples, majority of which contains a BRAF mutation, show the frequent methylation of
22
known CIMP-associated markers, including CDKN2A, IGF2, and MLH1 (Weisenberger
et al., 2006). We defined this subgroup of tumor as CIMP-positive tumors. In order to
identify CIMP-associated loci, we then performed a t-test on the difference in the β-value
between the CIMP-positive group and CIMP-negative group. We selected 100
GoldenGate CpG sites with p < 0.001 after a correction for multiple-comparison
(Benjamini and Hochberg, 1995) and mean |Δβ| > 0.17, the estimated error in β (Bibikova
et al., 2006).
Results
Characterization of 21 human colorectal cancer cell lines.
Since primary colonic epithelial cells were not readily available, we screened for
colorectal cancer cell lines that do not have substantial DNA methylation at CIMP-
defining loci and carry wild-type forms of both BRAF and KRAS. Such cell lines would
serve as suitable systems for the introduction of BRAF
V600E
. We selected 21 colorectal
cancer cell lines, characterized their DNA methylation profiles, and determined their
BRAF and KRAS mutation status (Figure 2.1). We used MethyLight to assess the DNA
methylation status of five CIMP-defining markers identified in our laboratory
(Weisenberger et al., 2006). Using a PMR (percent of methylated reference) of ≥10 as a
threshold for positive methylation as used previously (Weisenberger et al., 2006), we
identified six cell lines that lacked DNA methylation above this threshold for all five
CIMP-specific markers (Figure 2.1). Among the six colorectal cancer cell lines, three,
23
Caco-2, COLO 320DM, and LS1034, have neither BRAF nor KRAS mutations (Figure
2.1). To test our hypothesis, we chose COLO 320DM for its ease in culturing and
transfection.
24
Figure 2.1: Characteristics of 21 colorectal cancer cell lines. MethyLight was used
to assess the DNA methylation status of five CIMP-defining markers. A PMR of ≥10 is
used as a threshold for positive methylation. Dark gray boxes indicate PMR ≥ 10, and
white boxes indicate PMR < 10. The methylation frequencies of the five CIMP markers
increase from top to bottom. Microsatellite instability for each cell line is listed as
microsatellite stable (MSS) or harboring instability (MSI). The mutation status of BRAF
exon 15 and KRAS exon 2 are listed.
25
Stable transfection of BRAF
V600E
in COLO 320DM cells.
We transfected COLO 320DM cells with an HA-tagged BRAF
V600E
cDNA
and
picked G418-resistant clones. The expression level of BRAF
V600E
was determined by
western blotting using an antibody against the HA epitope (Figure 2.2A). The activity of
BRAF
V600E
was confirmed by examining the activation of ERK1/2 using an antibody
against phosphorylated ERK1/2 (Figure 2.2A). Eight stably-transfected clones exhibiting
high expression of BRAF
V600E
, as well as strong activation of ERK1/2, were grown in
culture, and genomic DNA was isolated at various passages (passages 2-27) from these
clones. Four empty-vector transfected clones (EVCs) were grown in the same conditions
and used as controls.
26
Figure 2.2: Selection of BRAF
V600E
stably-transfected clones and their Illumina
GoldenGate DNA methylation profiles. A. expression of BRAF
V600E
and ERK1/2
phosphorylation in stably-transfected COLO 320DM cells. Blots were probed with anti-
HA antibodies for HA-BRAF
V600E
, anti-phospho-ERK1/2, and anti-ERK1. Asterisks
indicate the eight BRAF
V600E
transfected clones that were subjected to DNA methylation
analysis at various cell passages. B. DNA methylation profiles of untransfected COLO
320DM cells, empty vector and BRAF
V600E
transfected COLO 320DM clones determined
by the Illumina GoldenGate DNA methylation assay. The DNA methylation data were
scored as β-values as previously defined (Bibikova et al., 2006). Each row corresponds to
an individual CpG locus and the data were sorted by average β-value across all samples.
Each clone is ordered from left to right in increasing number of passages. EVC, empty-
vector transfected clones.
HA-BRAF
V600E
p-ERK1/2
Total-ERK1/2
* * * * * * * *
Clones
A
B
Empty Vector BRAF
V600E
Transfected
1,505 CpG Loci
Parent
Clones:
Passages:
1
2
3
4
NA
11
15
85
90
93
96
111
115
!"
!#$"
%"
!#&"
!#'"
!#%"
!#("
!#)"
!#*"
!#+"
!#,"
!"-./012"
1
7
8
11
12
EVC 1
EVC 2
15
17
23
29
EVC 3
34
44
54
62
65
72
81
83
EVC 4
84
85
90
93
96
108
111
113
EVC 5
115
117
118
18
27
DNA methylation analysis of the BRAF
V600E
transfected COLO 320DM clones.
We next determined the DNA methylation levels in each of the eight BRAF
V600E
clones and four EVCs using the Illumina GoldenGate DNA Methylation Cancer Panel 1
platform (Figure 2.2B). The Illumina assay targets 1,505 CpG loci located at 807
different genes, most of which are within a promoter CpG island. The selection of these
CpG sites was heavily influenced by the use of human cell lines and previous reports
showing cancer-specific methylation. We found that the mean DNA methylation value
(β-value) across all 1,505 CpG loci in the BRAF
V600E
transfected clones (regardless of
their expression level) was very similar to those of empty-vector clones and relatively
stable over time, suggesting that there was no overall increase in DNA methylation in
BRAF
V600E
transfected clones (Figures 2.2B and 2.4A). We next determined whether the
stable expression of BRAF
V600E
specifically increased the DNA methylation of only the
CIMP-associated markers in the 1,505 interrogated loci. For this, we first screened 58
primary colorectal tumor samples using the Illumina GoldenGate DNA Methylation
assay to identify CIMP-associated DNA methylation markers. We selected 100 CpG loci
that have significantly higher levels of methylation in CIMP-positive (CIMP+) versus
CIMP-negative (CIMP–) tumors (The details are provided in the Materials and Methods
section) (Figure 2.3). Interestingly, we observed that the mean DNA methylation level of
the 100 CIMP-specific loci increased as a function of cell passage (Figure 2.4B).
However, this increase did not correlate with the levels of BRAF
V600E
expression, and
was also observed in cells transfected with an empty vector (Figure 2.4B). This general
increase in the mean β-value is specific for CIMP-associated loci, since the mean β-value
28
from several sets of 100 randomly selected loci does not show a similar trend (Figure
2.4C and D). Therefore, we conclude that although CIMP-associated CpG islands may be
prone to acquire DNA methylation in certain culture conditions, BRAF
V600E
does not
specifically induce CIMP in COLO 320DM cells.
29
Figure 2.3: Illumina GoldenGate DNA methylation profiles of CIMP-associated
loci in 58 primary colorectal tumors and COLO 320DM cells, empty vector and
BRAF
V600E
transfected COLO 320DM clones. CIMP+ tumors and CIMP-associated
loci in 58 primary tumor samples are defined as described in the Materials and Methods
section. Each row corresponds to an individual locus of the 100 locus panel, and the data
were sorted by the mean β-value of each locus over all 58 primary tumor samples.
Tumors with mutant BRAF and mutant KRAS are indicated by the circle and the triangle
symbols respectively. indicates that mutation status is not available. Each BRAF
V600E
transfected clone and EVC is ordered from left to right in increasing number of passages.
“P” indicates the DNA methylation profiles of parent untransfeted COLO320 DM cells.
30
Figure 2.4: Changes in DNA methylation levels over passages in BRAF
V600E
and
EVC stably-transfected clones. Black lines indicate BRAF
V600E
expressing clones and
gray lines represent empty vector transfected control clones. Each graphing point
represents mean β-values across indicated loci from the Illumina GoldenGate DNA
methylation assay at various cell passages for each clone: A. All 1,505 loci from the
Illumina GoldenGate assay. B. Only 100 CIMP-associated loci are profiled. C. One
hundred randomly chosen loci. D. One hundred non-CIMP loci which show mean β-
values similar to the 100 CIMP-associated loci.
!"#$%&'&(')*+,'-".%/&'0
!
! !
Clone 11
Clone 15
Clone 85
Clone 90
Clone 93
Clone 96
Clone 111
Clone 115
Parent cell
EVC1
EVC2
EVC3
EVC3
Cell passages Cell passages
31
Discussion
CIMP is characterized by an exceptionally high frequency of DNA
hypermethylation. Recent studies have shown that an activating mutation of BRAF
(BRAF
V600E
) is specifically associated with CIMP. In this study, we tested the causal
contribution of BRAF
V600E
to CIMP by stably expressing BRAF
V600E
in a colorectal
cancer cell line and showed that BRAF
V600E
is not sufficient to induce CIMP. However,
since this study was performed in transformed cancer cells, one caveat is that BRAF
V600E
could induce DNA methylation but only in the context of primary cells. It has been
proposed that CIMP tumors might arise through a distinct molecular pathway, in which
BRAF mutations have been described at the earliest stage of the tumor development
(Beach et al., 2005; O'Brien et al., 2006). Therefore, generating CIMP in a CIMP– cancer
cell line could be problematic due to fundamental differences between transformed cells,
which might have accumulated distinct genetic or epigenetic defects during later stages of
tumor progression, and the primary cells from which CIMP arises (Jass, 2007a; Laiho et
al., 2007).
Recently, Minoo et al. reported MLH1 DNA hypermethylation upon stable-
transfection of BRAF
V600E
into the NCM460 cell line (Minoo et al., 2007). Previously,
we described the CIMP-associated methylation of MLH1 as the underlying basis for
mismatch repair deficiency in sporadic colorectal cancer (Weisenberger et al., 2006).
Although we did not detect an increase in MLH1 methylation in our system, it might be
possible that in certain circumstances BRAF
V600E
can contribute to DNA
hypermethylation at a certain locus. Interestingly, in the proposed serrated pathway to
32
CIMP+ tumors, both the BRAF mutation and CIMP+ have been observed in early
precursor lesions, whereas MSI+ might represent a later stage event in the development
(O'Brien, 2007). However, Minoo and colleagues observed the DNA hypermethylation of
CDKN2A and 15 other CIMP-associated markers in parent NCM460 cells, which limited
their ability to study further the role of BRAF
V600E
in inducing CIMP in their
experimental system (Minoo et al., 2007).
Intriguingly, we observed in our stably-transfected cells that the overall DNA
methylation level of the CIMP-specific loci increases as a function of cell passage. It is
interesting to note that a selection drug in cultured cells has been described to result in
changes in the global chromatin structure (Muthuswami et al., 2000), and a similar
process may be associated with our observations here.
Finally, we found relatively large inter-clonal (among different clones) variation
in our methylation analyses in our transfection experiment (Figures 2.2B and 2.3), with
an average R
2
correlation calculated based on four EVCs of 0.88 ± 0.01 (± s.d.). Our
average intra-clonal (within clones at different passages) R
2
correlation is 0.97 ± 0.01 and
the R
2
correlation between technical replicates in Illumina GoldenGate DNA methylation
analysis is 0.98 ± 0.02 (Bibikova et al., 2006). Consequently, we found some large
differences in DNA methylation at several loci even among control clones (Figures 2.2B
and 2.3). This emphasizes the importance of using multiple clones for this type of studies.
33
CHAPTER 3
CIMP creates a permissible context for the acquisition of BRAF
V600E
in colorectal cancer
Introduction
Aberrant DNA methylation at promoter CpG islands has been extensively studied
in cancer. Inactivation of selected tumor-suppressor genes in this pathway has now been
accepted as a critical contributor to the development of cancer (Jones and Baylin, 2007).
A distinct subset of colorectal cancer displays CIMP, characterized by a high frequency
of DNA hypermethylation in a specific group of CpG islands. CIMP+ cancer may arise
through a distinct pathway originating in certain subtypes of serrated polyps (Kambara et
al., 2004; O'Brien et al., 2006) and accounts for at least 15% of all colorectal cancer cases
(Samowitz et al., 2005; Weisenberger et al., 2006). Features associated with CIMP+
colorectal cancers include female sex, proximal location, and poorly differentiated or
mucinous histology (Kambara et al., 2004; Issa, 2004; Samowitz et al., 2005;
Weisenberger et al., 2006). Our study using a newly developed CIMP marker panel in
colorectal cancers demonstrated that sporadic microsatellite instability (MSI+) occurs as
a consequence of CIMP-associated MLH1 methylation and we found a tight association
of CIMP with the presence of an activated mutant form of BRAF (BRAF
V600E
) (Odds
ratio = 203) (Weisenberger et al., 2006). The extremely strong association between
BRAF
V600E
and CIMP raises the question of whether BRAF
V600E
plays a causal role in the
development of CIMP or whether CIMP provides a favorable setting for the acquisition
34
of BRAF
V600E
. Previously, we examined the causal contribution of BRAF
V600E
to CIMP
by stably expressing BRAF
V600E
in the CIMP-negative, BRAF wild-type COLO 320DM
colorectal cancer cell line. We found that BRAF
V600E
does not specifically induce CIMP
in our system (Chapter 2). Here, we considered an alternative hypothesis that promoter
methylation of specific gene targets provides a favorable setting for the acquisition of
BRAF mutation in CIMP+ colorectal cancer.
Materials and Methods
Primary colorectal tissue samples
Primary colorectal tissue samples were collected and DNA was extracted as previously
described (Weisenberger et al., 2006). The ten pairs of normal and tumor colorectal tissue
samples (five CIMP+ tumors and five CIMP– tumors) were obtained as previously
described (Weisenberger et al., 2006). The 235 sample set includes the 48 randomly
selected samples together with an additional 187 randomly collected tumor samples
described previously (Weisenberger et al., 2006). CIMP status and BRAF mutation status
for each tumor sample was previously determined (Weisenberger et al., 2006).
Illumina GoldenGate DNA methylation assay
Genomic DNA was sodium bisulfite converted using the EZ-96 DNA Methylation Kit
(ZYMO Research, Orange, CA) according to manufacturer’s protocol. Illumina
GoldenGate DNA methylation analyses were performed as described previously
35
(Bibikova et al., 2006) at the USC Genomics Core Facility. Target sequences for the
assay and detailed information on each interrogated CpG site and its associated gene on
the “GoldenGate Methylation Cancer Panel 1” are described at www.illumina.com.
Identification of CpG sites specifically methylated in BRAF
V600E
colorectal tumors
We performed the Illumina GoldenGate DNA Methylation assay on 235 primary
colorectal tumor samples, whose CIMP status and BRAF mutation status have been
determined previously (Weisenberger et al., 2006). We dichotomized the DNA
methylation β-value (methylated or unmethylated) for each CpG site. We then calculated
odds ratios and P values (Fisher’s exact test) in order to identify CpG sites specifically
methylated in tumors harboring BRAF
V600E
. The dichotomization threshold was chosen
for each CpG site using the mean β-value + 3SD (standard deviations) from ten normal
mucosal samples. The 89 Illumina GoldenGate Methylation targets (of 60 genes) with p
< 0.001 (Fisher’s exact test), after Bonferroni correction for multiple comparisons were
selected.
Cell culture and genomic DNA isolation
Colorectal cancer cell lines were obtained from American Type Culture Collection
(Manassas, VA). Genomic DNA from each cell line was isolated as described previously
(Laird et al., 1991).
36
Quantitative Real-Time RT-PCR
Total RNA from colorectal cancer cell lines was isolated using RNeasy Mini Kit
(QIAGEN GmbH, Hilden, Germany). Reverse transcription reaction was performed
using SuperScript® III First-Strand Synthesis System for RT-PCR (Invitrogen,
Burlington, ON). The quantitative real-time PCR was performed with primers and a
probe purchased from Applied Biosystems (Assay ID Hs00266026_m1). The raw
expression values were normalized to those of PCNA.
Results
Identification of genes that are specifically methylated in colorectal tumors
harboring BRAF
V600E
.
We first determined the CIMP status and BRAF mutation status of 235 primary
colorectal tumor samples, and performed the Illumina GoldenGate DNA Methylation
assay on these samples. We found the BRAF
V600E
in 33 tumor samples (14.0%), and 31
of these tumors harboring BRAF
V600E
were classified as CIMP+ (93.7%). We calculated
odds ratios and p-values (Fisher’s exact test) in order to identify CpG sites specifically
methylated in the 33 BRAF
V600E
-positive tumors, (The details are provided in the
Materials and Methods section). The Illumina probes were then ranked based on the p-
values. We focused on 89 CpG sites (at 60 genes) selected with p < 0.001, after
Bonferroni correction for multiple comparisons (Table 3.1). These genes are candidates
for CIMP-specific inactivation, which may closely synergize with BRAF mutation to
37
promote tumorigenesis. We confirmed the recently observed associations between DNA
hypermethylation of BMP3 and MCC with CIMP+ and BRAF mutation in colorectal
cancer (Kohonen-Corish et al., 2007; Loh et al., 2008). We also found CIMP-specific
DNA hypermethylation of BMP6. The simultaneous epigenetic inactivation of BMP3 and
BMP6 was shown to be associated with the activation of the RAS-RAF-MEK-ERK
signaling pathway in non-small-cell lung cancer (Kraunz et al., 2005). Moreover, we
found an association of SLC5A8 and TIMP3 DNA methylation with BRAF mutation in
our colorectal tumor samples, as had been previously reported in papillary thyroid
carcinomas (Xing, 2007). The functional relevance of the inactivation of such tumor
suppressor genes linked with CIMP+/BRAF
V600E
remains speculative (Kraunz et al.,
2005; Kohonen-Corish et al., 2007; Xing, 2007; Loh et al., 2008). The most highly
ranked genes include the receptor tyrosine kinases (RTK), EPHA3, KIT, and FLT1.
Somatic mutations or overexpression of RTKs has been implicated in colorectal
tumorigenesis, which may involve the activation of the RAS-RAF-MEK-ERK signaling
pathway (Bardelli et al., 2003; Bates et al., 2003; Graells et al., 2004; Noren and
Pasquale, 2004; Bellone et al., 2006; Wood et al., 2007) (Table 3.1). Other highly ranked
genes such as SMO and HHIP are involved in the regulation of the Hedgehog (Hh)
signaling pathway (Table 3.1). Changes in the Hedgehog (Hh) signaling pathway in
relation to colorectal tumorigenesis has not been fully elucidated (Chatel et al., 2007).
38
Table 3.1: The Illumina GoldenGate DNA Methylation targets specifically methylated in BRAF
V600E
colorectal tumors.
NUMBER OF BRAF MUTANT
TUMORS
NUMBER OF BRAF WT
TUMORS
HGNC
SYMBOL
TARGET ID METHYLATED UNMETHYLATED METHYLATED UNMETHYLATED
ODDS
RATIO
P value
EPHA3 EPHA3_E156_R 32 1 38 160 134.7 1.35E-18
BMP6 BMP6_P163_F 33 0 51 151 N/A 1.28E-17
SMO SMO_E57_F 32 1 49 146 95.3 8.22E-16
KIT KIT_P405_F 31 2 46 156 52.6 2.12E-15
CALCA CALCA_P75_F 30 3 40 162 40.5 2.17E-15
MT1A MT1A_P49_R 19 14 5 196 53.2 6.28E-15
TMEFF1 TMEFF1_E180_R 31 2 49 153 48.4 9.09E-15
HHIP HHIP_E94_F 32 1 56 134 76.6 4.56E-14
FZD7 FZD7_E296_F 25 8 24 177 23.0 1.17E-13
TIMP3 TIMP3_seq_7_S38_F 25 8 24 177 23.0 1.17E-13
CALCA CALCA_E174_R 30 3 48 152 31.7 1.53E-13
PAX6 PAX6_E129_F 22 11 15 187 24.9 1.76E-13
FLT1 FLT1_P615_R 31 2 58 144 38.5 4.84E-13
IGF2 IGF2_E134_R 29 4 45 154 24.8 6.69E-13
BMP6 BMP6_P398_F 30 3 52 149 28.7 8.01E-13
NGFR NGFR_E328_F 29 4 48 153 23.1 2.04E-12
FLT3 FLT3_P302_F 32 1 68 129 60.7 2.12E-12
MT1A MT1A_E13_R 19 14 10 191 25.9 2.13E-12
FN1 FN1_E469_F 29 4 49 153 22.6 2.79E-12
WNT1 WNT1_P79_R 29 4 50 152 22.0 4.31E-12
WNT1 WNT1_E157_F 28 4 49 148 21.1 1.37E-11
ASCL1 ASCL1_E24_F 30 3 59 142 24.1 1.41E-11
MCC MCC_E23_R 26 7 37 165 16.6 1.54E-11
RAB32 RAB32_E314_R 21 12 18 184 17.9 1.63E-11
MCC MCC_P196_R 28 5 47 153 18.2 1.81E-11
MYLK MYLK_E132_R 19 14 13 189 19.7 2.79E-11
PLXDC2 PLXDC2_E337_F 19 14 13 189 19.7 2.79E-11
SLC5A8 SLC5A8_P38_R 26 7 39 163 15.5 4.00E-11
BMP3 BMP3_E147_F 33 0 90 112 N/A 4.91E-11
POMC POMC_P400_R 30 3 63 136 21.6 8.30E-11
39
Table 3.1 Continued
COL4A3 COL4A3_E205_R 32 1 81 121 47.8 9.50E-11
CDKN2A p16_seq_47_S188_R 28 5 51 148 16.3 1.08E-10
EPHA3 EPHA3_P106_R 28 5 52 149 16.0 1.27E-10
GDF10 GDF10_P95_R 31 2 72 127 27.3 1.38E-10
PAX6 PAX6_P50_R 30 3 65 135 20.8 1.50E-10
ABCB1 MDR1_seq_42_S300_R 26 7 42 160 14.1 1.55E-10
PROK2 PROK2_E0_F 30 3 66 136 20.6 1.65E-10
POMC POMC_P53_F 31 2 75 127 26.2 2.51E-10
WNT5A WNT5A_E43_F 29 4 61 141 16.8 3.23E-10
CDH11 CDH11_E102_R 30 3 69 132 19.1 5.27E-10
COL4A3 COL4A3_P545_F 25 8 41 161 12.3 8.27E-10
SIRPA PTPNS1_P301_R 30 3 71 131 18.5 9.01E-10
ROR2 ROR2_P317_R 23 10 33 168 11.7 1.42E-09
NOTCH3 NOTCH3_E403_F 25 8 43 157 11.4 2.34E-09
HLF HLF_E192_F 24 9 39 163 11.1 2.62E-09
BMP3 BMP3_P56_R 32 1 92 110 38.3 2.73E-09
ERBB4 ERBB4_P541_F 32 1 92 110 38.3 2.73E-09
IGFBP7 IGFBP7_P371_F 27 6 55 147 12.0 3.14E-09
THBS1 THBS1_E207_R 13 20 5 197 25.6 4.14E-09
ALPL ALPL_P278_F 30 3 75 124 16.5 4.78E-09
WNT5A WNT5A_P655_F 32 1 93 107 36.8 4.85E-09
KIT KIT_P367_R 30 3 76 125 16.4 5.04E-09
CALCA CALCA_P171_F 16 17 12 190 14.9 5.70E-09
CDH11 CDH11_P203_R 28 5 63 138 12.3 6.83E-09
SOX2 SOX2_P546_F 17 16 15 187 13.2 6.88E-09
PAX6 PAX6_P1121_F 28 5 64 137 12.0 9.49E-09
HTR2A HTR2A_P853_F 19 14 22 177 10.9 1.33E-08
CCKBR CCKBR_P361_R 31 2 88 114 20.1 1.33E-08
ERBB4 ERBB4_P255_F 32 1 97 103 34.0 1.49E-08
FGF8 FGF8_E183_F 33 0 111 89 N/A 2.56E-08
PLXDC1 PLXDC1_P236_F 29 4 75 127 12.3 3.11E-08
JAK3 JAK3_E64_F 21 12 32 170 9.3 3.51E-08
LOX LOX_P71_F 15 18 12 190 13.2 3.79E-08
DIO3 DIO3_P90_F 16 17 15 187 11.7 4.43E-08
40
Table 3.1 Continued
PROK2 PROK2_P390_F 32 1 100 98 31.4 4.46E-08
GJB2 GJB2_P791_R 19 14 25 176 9.6 5.32E-08
COL1A1 COL1A1_P5_F 27 6 64 137 9.6 7.03E-08
DIO3 DIO3_P674_F 11 22 4 198 24.8 7.58E-08
CADM1 IGSF4_P86_R 20 13 31 171 8.5 1.31E-07
NOTCH3 NOTCH3_P198_R 15 18 14 187 11.1 1.52E-07
SMO SMO_P455_R 32 1 107 95 28.4 1.54E-07
SEPT5 SEPT5_P441_F 18 15 24 178 8.9 1.73E-07
SLC5A8 SLC5A8_E60_R 25 8 56 146 8.1 2.02E-07
FLT1 FLT1_P302_F 24 9 47 141 8.0 2.31E-07
VIM VIM_P343_R 32 1 110 92 26.8 3.24E-07
FN1 FN1_P229_R 23 10 47 155 7.6 3.33E-07
PTCH2 PTCH2_P37_F 9 24 2 200 37.5 3.36E-07
TERT TERT_E20_F 29 4 84 118 10.2 3.84E-07
ROR2 ROR2_E112_F 26 7 64 138 8.0 3.96E-07
COL1A2 COL1A2_P48_R 26 7 63 136 8.0 4.02E-07
VCAN CSPG2_P82_R 33 0 124 78 N/A 4.68E-07
GDF10 GDF10_E39_F 32 1 112 90 25.7 5.26E-07
IGFBP7 IGFBP7_P297_F 21 12 39 163 7.3 5.59E-07
IGF2 IGF2_P1036_R 32 1 113 89 25.2 6.68E-07
HHIP HHIP_P307_R 31 2 103 99 14.9 6.88E-07
MATK MATK_P64_F 27 6 73 129 8.0 8.51E-07
LRP2 LRP2_E20_F 33 0 127 75 N/A 9.41E-07
CDKN2A CDKN2A_E121_R 25 8 60 138 7.2 1.01E-06
CDKN1C CDKN1C_P626_F 19 14 32 170 7.2 1.02E-06
Table 3.1. The Illumina GoldenGate DNA Methylation targets specifically methylated in BRAF
V600E
colorectal tumors.
IGFBP7 is highlighted in gray and has been previously described as a mediator of the BRAF
V600E
induced senescence (Wajapeyee et
al., 2008).
41
Promoter DNA hypermethylation and transcriptional silencing of IGFBP7 in BRAF
mutant CIMP+ colorectal cancer.
Intriguingly, we also identified the IGFBP7 promoter CpG island as a target for
DNA methylation in BRAF mutant CIMP+ tumors. BRAF
V600E
has been shown to induce
cellular senescence (Zhu et al., 1998; Michaloglou et al., 2005; Dankort et al., 2007).
Oncogene-induced senescence (OIS) has been recognized as an important tumor
suppressor mechanism (Mooi and Peeper, 2006). The molecular basis of BRAF
V600E
-
induced senescence and apoptosis has been studied in detail. Wajapeyee et al. initially
identified IGFBP7 as a mediator of BRAF
V600E
-induced senescence and apoptosis in
human primary foreskin fibroblasts using a genome-wide shRNA screen (Wajapeyee et
al., 2008). Their subsequent findings suggest that IGFBP7 is both necessary and
sufficient to induce senescence and apoptosis in human primary melanocytes and
melanoma, respectively. Moreover, IGFBP7 was found to be epigenetically silenced by
CpG island promoter hypermethylation in melanoma samples carrying BRAF mutations
(Wajapeyee et al., 2008). Promoter-associated CpG island DNA hypermethylation of
IGFBP7 has been reported in some human colorectal cancer cell lines (Lin et al., 2007).
The recent study by the same group demonstrated that DNA methylation inhibitor 5-aza-
2'-deoxycytidine restores the expression of IGFBP7 in colorectal cancer cell lines,
indicating that the DNA hypermethylation plays a major role in silencing of IGFBP7
expression in colorectal cancer cells (Lin et al., 2008). However, its association with
BRAF mutation and CIMP+ status in human colorectal cancers has not been explored.
The Illumina GoldenGate DNA methylation Cancer Panel 1 platform contains two
42
different IGFBP7 targets (IGFBP7_P297_F and IGFBP7_P371_F) that interrogate DNA
methylation in the IGFBP7 promoter CpG island (Figure 3.1A). We found that these two
CpG sites of IGFBP7 are cancer specifically methylated (Figure 3.1B) and strongly
associated with both the BRAF mutation (Wilcoxon rank-sum test, p-value = 2.0 x 10
-10
)
and CIMP (p-value = 3.6 x 10
-9
) (Figure 3.1C). By contrast, IGFBP7 DNA
hypermethylation was not significantly associated with KRAS mutations compared with
tumors with wild-type BRAF (p-value = 0.85). Similarly, we observed that DNA
hypermethylation of IGFBP7 is mostly found in colorectal cancer cell lines with the
BRAF
V600E
and frequent DNA methylation of the five-gene CIMP–specific marker panel
(Figure 3.2). Real-time RT-PCR analyses of the colorectal cancer cell lines showed that
IGFBP7 mRNA expression was inversely related to DNA hypermethylation (Figure 3.2)
as previously reported (Lin et al., 2007). Among the CIMP− cell lines we examined, only
COLO 320DM shows DNA hypermethylation of IGFBP7 CpG island promoter with
minimal level of expression of the gene. In retrospect, this unique characteristic of COLO
320DM cells compared to the other CIMP− cell lines might have enabled these cells to
tolerate mutant BRAF overexpression in our previous study (Chapter 2), and may explain
our difficulties in obtaining BRAF
V600E
overexpression in other colorectal cancer cell
lines. Overall, these observations suggest that epigenetic inactivation of IGFBP7 in the
context of CIMP may accommodate BRAF mutation by blocking oncogene-induced
senescence, thus providing a mechanistic rationale for the association between BRAF
mutation and CIMP.
43
Figure 3.1: IGFBP7 promoter DNA methylation in human colorectal cancers. A.
Genomic map of IGFBP7 promoter-associated CpG island, transcription start site (TSS)
and exon 1 based on the UCSC genome browser (March 2006 assembly). The location of
CpG sites interrogated by the Illumina GoldenGate DNA methylation assay are indicated
by vertical arrows. B. DNA methylation levels of the two CpG dinucelotides in the
IGFBP7 promoter CpG island. β-values of each CpG site in 10 tumors (five CIMP–
tumors with wild type BRAF and five CIMP+ tumors with mutant BRAF, black bars) and
matched normal mucosa (gray bars) are listed. C. IGFBP7 promoter DNA methylation
box plots of 235 human colorectal tumors stratified by BRAF mutation status (left) and
CIMP+ status (right) at the IGFBP7 P371 locus. In the box plots, the ends of the box are
the 25th and 75th quartiles. The line within the box identifies the median β-value. The
bars above and below the box are at most 1.5 times the box width. The CIMP status of
each colorectal tumor sample is determined as described in the Materials and Methods
section.
!!"#$%&'()*+
!!"#$%&'()*+
P P
Matched normal mucosa
Tumor
–
B
C
A
!!"#$%&'()*+
!!"#$%&'(,-*
44
Figure 3.2: Analysis of DNA methylation and mRNA expression of IGFBP7 in
colorectal cancer cell lines. Quantitative real-time RT-PCR analysis of IGFBP7
expression. IGFBP7 expression levels are presented as relative to PCNA expression.
Each horizontal bar indicates the mean of triplicate RT-PCR. The error bars are the
standard deviation of the expression levels. The number of methylated loci among the
five CIMP markers and mutation status of BRAF and KRAS listed in Figure 2.1 are
provided.
β
Caco-2 0 WT WT
COLO320DM 0 WT WT
LS1034 0 WT WT
LS123 0 WT G12S
SW480 0 WT G12V
SW1463 1 WT G12C
SW1116 2 WT G12A
SW1417 2 V600E WT
SW948 2 WT WT
SW837 3 WT G12C
COLO205 4 V600E WT
HCT116 4 WT G13D
HT29 4 V600E WT
RKO 4 V600E WT
DLD1 5 WT G13D
HCT15 5 WT G13D
Relative level of
IGFBP7 mRNA
0 1 2 10 20 30 40 50
CELL LINES
# OF METHYLATED
CIMP MARKERS
BRAF EXON15
KRAS EXON2
0-0.1
0.1-0.3
0.3-0.6
>0.6
45
Discussion
CIMP in colorectal cancer provides a unique opportunity to study the molecular
mechanisms that lead to epigenetic changes in cancer and the contribution of these
changes in the development of cancer (Issa, 2004; Schuebel et al., 2006). The distinct
features found in CIMP are important clues in understanding this phenotype (Kambara et
al., 2004; Issa, 2004; Samowitz et al., 2005; Weisenberger et al., 2006). Particularly
striking is the extremely tight association between CIMP and BRAF mutation
(Weisenberger et al., 2006). Previously, we showed that stable expression of BRAF
V600E
is not sufficient to induce CIMP in an established colorectal cancer cell line (Chapter 2).
Here, we found a functional example supporting the hypothesis that inactivation of a
specific target gene in the context of CIMP creates a favorable context for the acquisition
of BRAF
V600E
in colorectal cancer. BRAF
V600E
has been shown to induce cellular
senescence in cultured cells (Zhu et al., 1998), a mouse model (Dankort et al., 2007), and
primary human cells (Michaloglou et al., 2005). Oncogene-induced senescence (OIS) has
been recognized as an important tumor suppressor mechanism (Mooi and Peeper, 2006).
In order for BRAF
V600E
to promote its oncogenic effects, additional cooperative events
are required to bypass senescence (Mooi and Peeper, 2006). A recent study showed that
IGFBP7 is both necessary and sufficient to induce senescence and apoptosis induced by
BRAF
V600E
(Wajapeyee et al., 2008). The underlying molecular mechanisms have also
been elucidated extensively in their study, which involves in part IGFBP7-mediated
induction of Raf kinase inhibitory protein (RKIP), a protein that interacts with BRAF and
antagonize the kinase activity of BRAF (Park et al., 2005) and up-regulation of BNIP3L,
46
a proapoptotic BCL2 family protein (Wajapeyee et al., 2008). They observed loss of
IGFBP7 in a BRAF
V600E
-positive melanoma and concluded that silencing of IGFBP7
expression is a critical step in the development of a melanoma harboring BRAF
V600E
.
Moreover, the recent study demonstrated that DNA methylation inhibitor 5-aza-2'-
deoxycytidine restores the expression of IGFBP7 in colorectal cancer cell lines,
indicating that the DNA hypermethylation plays a major role in silencing of IGFBP7
expression in colorectal cancer cells (Lin et al., 2008). In this study, we showed that
IGFBP7 DNA hypermethylation is tumor-specific and tightly associated with colorectal
tumors exhibiting CIMP, and that IGFBP7 expression was inversely related to DNA
hypermethylation. Collectively, we propose that CIMP-specific inactivation of
BRAF
V600E
-induced senescence and apoptosis pathway by IGFBP7 DNA
hypermethylation provides an explanation of the tight association between BRAF
V600E
and CIMP in colorectal cancers.
It is important to note that IGFBP7 DNA hypermethylation was not observed in
all of the BRAF mutant colorectal tumors. For instance, SW1417 cells exhibit strong
IGFBP7 expression despite having mutant BRAF (Figure 3.2). Evasion of BRAF
V600E
induced senescence, however, appears to be achieved by various ways in cancer cells.
Expression of CDKN2A (also known as p16
INK4a
) has also been described to be
important in the OIS (Mooi and Peeper, 2006), and IGFBP7 appears to be required for its
induction (Wajapeyee et al., 2008). Notably, in our study, frequent DNA
hypermethylation of CDKN2A (p16
INK4a
) has also been associated with CIMP+ colorectal
tumors (Table 3.1) (Toyota et al., 1999; Issa, 2004; Weisenberger et al., 2006). More
recently, activation of AKT3 has shown to cooperate with BRAF
V600E
in promoting
47
melanoma development (Cheung et al., 2008). In this case, BRAF
V600E
phosphorylation
by AKT3 results in reduced activity of BRAF
V600E
to the level that is compatible with cell
proliferation and cancer development. It is possible that CIMP-associated DNA
hypermethylation targets various pathways to impair the ability of normal cells to block
the oncogenic effects of BRAF
V600E
.
Disruption of BMP signaling has been proposed to play a role in colorectal
tumorgenesis (Loh et al., 2008). We found that both BMP3 and BMP6 are targeted for
CIMP-specific DNA hypermethylation and strongly linked with BRAF mutation.
Concurrent epigenetic inactivation of BMP3 and BMP6 was shown to be associated with
the hyperactivation of the RAS-RAF-MEK-ERK signaling pathway in non-small-cell
lung cancer (Kraunz et al., 2005). The functional relationship between epigenetic
inactivation of BMP signaling and activation of the RAS-RAF-MEK-ERK signaling
pathway in cancer has been postulated but remained to be investigated further (Kraunz et
al., 2005; Loh et al., 2008).
Furthermore, RTKs such as EPHA3, KIT, and FLT1 also showed CIMP-
associated DNA hypermethylation and ranked most significantly in our analysis (Table
3.1). Somatic mutations or overexpression of these genes has been implicated in
colorectal tumorigenesis, which may involve the activation of the RAS-RAF-MEK-ERK
signaling (Bardelli et al., 2003; Bates et al., 2003; Graells et al., 2004; Noren and
Pasquale, 2004; Bellone et al., 2006; Wood et al., 2007). The potential inactivation of
these genes in CIMP may lead to the development of tumors dependent on oncogenic
BRAF-driven hyperactivation of the RAS-RAF-MEK-ERK signaling pathway.
48
In this report, we have demonstrated a mechanistic basis for the association
between CIMP and BRAF mutation in human colorectal cancer. Our observations suggest
that inactivation of the senescence mechanism by CIMP-associated DNA
hypermethylation may facilitate the development of cancers by providing a permissive
environment for the acquisition of a BRAF mutation. Consistent with our findings, we
have previously shown that BRAF mutation rarely occurs outside the context of CIMP
(Weisenberger et al., 2006), in contrast to the occurrence of CIMP in the absence of
BRAF mutation. These observations highlight the importance of an implied epigenetic
regulatory defect in the development of this subset of colorectal cancer.
49
CHAPTER 4
Comprehensive DNA methylation profiling and identification of novel CIMP
targets in colorectal cancer
Introduction
Aberrant DNA hypermethylation of CpG islands is a common and early event in
colorectal tumorigenesis. Promoter DNA hypermethylation is associated with
transcriptional gene silencing and can contribute to tumorigenesis when it occurs at a
critical tumor suppressor gene (reviewed in Chapter 1). A distinct subset of colorectal
cancers display CIMP, characterized by an exceptionally high frequency of DNA
hypermethylation in a group of CpG islands. CIMP has been reported in the earliest
stages of colorectal cancer (Minoo et al., 2006). Clearly, a subset of CIMP-specific DNA
hypermethylation events play a role in the development of the tumor, and are responsible
for creating the distinct genetic and histopathological characteristics associated with
CIMP+ tumors. For instance, we reported that CIMP-associated DNA hypermethylation
of MLH1 generates sporadic MSI+ tumors due to mismatch repair deficiency
(Weisenberger et al., 2006). Furthermore, we found that the CIMP-specific inactivation
of IGFBP7-mediated senescence and apoptosis pathway may provide a permissive
environment for the acquisition of a BRAF mutation in CIMP+ tumors (Chapter 3).
Questions remain as to how widespread the aberrant CpG island hypermethylation
events are that take place in CIMP+ tumors, what is special about genes and associated
CpG islands targeted by CIMP, and most importantly, what the molecular mechanism is
50
that accounts for CIMP. In this chapter, we performed comprehensive DNA methylation
profiling of 100 colorectal tumors in an attempt to answer these questions. We employed
recently developed Illumina Infinium DNA methylation technology, which interrogates
27,578 CpG loci spanning 14,495 genes. We have also performed gene expression
analysis for the same 100 tumor samples. Our most comprehensive list of CIMP-
associated loci will provide us with the opportunity to study structural and sequence
characteristics of affected CpG islands, which might give us insights into the underlying
basis of CIMP-associated DNA hypermethylation. DNA methylation data combined with
expression data will help us better understand the role of CIMP-associated DNA
hypermethylation in colorectal tumorigenesis.
Materials and Methods
Primary colorectal tissue samples
A total of 100 well-characterized fresh-frozen colorectal adenocarcinoma samples with
clinical data were obtained from four hospitals within the Ontario Tumor Bank network
(The Ontario Institute for Cancer Research, Ontario, Canada). Accompanying clinical
information collected on each case include histology, patient history (including alcohol
consumption and smoking history), family history, treatment (surgery, chemotherapy,
and radiation), toxicity, and outcome data. Normal colonic biopsy samples were collected
at the time of screening or diagnostic colonoscopy at the University of Southern
California USC/Los Angeles County Medical Center and USC University Hospital. The
51
tissue collection and analyses were approved by the University of Southern California
Institutional Review Board.
Genomic DNA and total RNA isolation
Genomic DNA and total RNA were simultaneously extracted from the same sample by
using TRIZOL® Reagent (Invitrogen, Burlington, ON) according to the manufactures
protocol.
BRAF and KRAS mutation analysis
BRAF mutations at codon 600 in exon 15 and KRAS mutations at codons 12 and 13 in
exon 2 were determined by the pyrosequencing assay on a Pyrosequencing 96HS
(Biotage, Uppsala Sweden) per manufacturer’s protocol. A 224 bp fragment of the BRAF
gene containing exon 15 was amplified from the genomic DNA using the following
primers: 5’ TCA TAA TGC TTG CTC TGA TAG GA 3’ and 5’Biotin-GGC CAA AAA
TTT AAT CAG TGG A 3’and genotyped with the sequencing primer 5’ CCA CTC CAT
CGA GAT T 3’. Similarly, a 214 bp fragments of the KRAS gene containing exon 2 were
amplified from the genomic DNA using the following primers: 5’Biotin-GTG TGA CAT
GTT CTA ATA TAG TCA 3’ and 5’ GAA TGG TCC TGC ACC AGT AA 3’ and
genotyped with the sequencing primer 5’ GCA CTC TTG CCT ACG 3’.
DNA methylation analyses
MethyLight
Genomic DNA was treated with sodium bisulfite using the Zymo EZ DNA Methylation
52
Kit (Zymo Research, Orange, CA) and subsequently analyzed by MethyLight as
described recently (Campan et al., 2009). The primer and probe sequences for the
MethyLight reactions are listed in Appendix A. The results of the MethyLight assays
were scored as PMR (Percent of Methylated Reference) values as previously defined
(Weisenberger et al., 2006; Campan et al., 2009).
Illumina Infinium DNA methylation assay
Genomic DNA was bisulfite converted using the EZ-96 DNA Methylation Kit (ZYMO
Research) according to manufacturer's protocol. For improved bisulfite conversion
efficiency, a short denaturation step was included (16 cycles of 95ºC for 30 seconds
followed by 50ºC for 1 hour). After bisulfite conversion, the analyses were performed at
the USC Genomics Core Facility. The DNA methylation analysis technology utilizes the
Infinium technology previously described for SNP genotyping (Steemers et al., 2006). In
brief, bisulfite-converted DNA was whole genome amplified (WGA) and enzymatically
fragmented. The bisulfite-converted, fragmented WGA-DNA samples are then applied to
the BeadChips. During this hybridization, the WGA-DNA molecules anneal to allele-
specific DNA oligomers linked to individual bead types. The two bead types correspond
to each CpG locus, one to the methylated (C) and the other to the unmethylated (T) state.
A successfully annealed hybrid will allow for the incorporation of a labeled nucleotide
immediately adjacent to the interrogated CpG (or TpG) site. A measure of the level of
methylation at each CpG site is scored as continuous variables described as β-value: the
ratio of the intensity of the methylated bead type to the combined locus intensity. The β-
values are accessed for each locus in each sample via proprietary BeadStudio software
(Illumina).
53
Cluster analysis
For the clustering analysis shown in Figure 4.2, we only used autosomal genes. We also
eliminated a reaction if the probe sequence contains a single nucleotide polymorphism
(SNP) at the interrogated CpG locus or overlaps with a repeat sequence ≥10 base pair.
We also filtered out probes if the difference between the highest methylation β-value and
the lowest methylation β-value in the 100 samples is less than 0.2. We used the average
linkage method and the correlation-based distance metric to perform the clustering as
implemented in the Cluster 3.0 software.
Gene expression analysis
Assessment of RNA integrity
The concentrations of RNA samples were measured using the NanoDrop 8000 (Thermo
Fisher Scientific Inc, Waltham, MA). The quality of the RNA samples was assessed
using the Experion RNA StdSens analysis kit, which is based on capillary electrophoresis
(Bio-Rad, Hercules, CA).
Illumina whole genome expression BeadChips
Expression analysis was performed using the Illumina whole genome expression
BeadChips (HumanRef-8 v3.0, 24,526 probes) (Illumina, San Diego, CA). Briefly, RNA
samples that passed the initial quality control checks were processed using the Illumina
TotalPrep RNA Amplification Kit (Illumina). 500ng of the RNA sample was subjected to
reverse transcription with an oligo(dT) primer bearing a T7 promoter. The cDNA then
underwent second strand synthesis and purification. Biotinylated cRNA was then
generated from the double-stranded cDNA template through in vitro transcription with
54
T7 RNA polymerase. 750ng of biotinylated cRNA was then hybridized to the
HumanRef-8 v3.0 BeadChips. The hybridized chips were stained and scanned using the
Illumina HD BeadArray scanner. Normalization and analysis of the samples across chips
were performed using BeadStudio 3.0.1 software (Illumina).
55
Results
CIMP and covariate analyses of 100 primary colorectal tumor samples
A total of 100 fresh-frozen colorectal tumor tissue samples were collected based
solely on the availability of accompanying clinical information from the Ontario Tumor
Bank network (Ontario, Canada). We first assessed the DNA methylation status of five
CIMP-defining markers and MLH1 promoter in these 100 tumor samples using
MethyLight. Using a PMR (percent of methylated reference) of ≥10 as a threshold for
positive methylation as used previously (Weisenberger et al., 2006), we identified 16
CIMP+ tumors (16%) that showed DNA methylation of at least three of five markers
(Figure 4.1A). We observed a clear bimodal distribution of tumors by number of
methylated markers (Figure 4.1B). We also determined BRAF and KRAS mutation status
in these 100 tumors by pyrosequencing. We confirmed the significant association of
CIMP+ tumors with female sex, mucinous histology, MLH1 DNA hypermethylation, and
BRAF mutation as previously found (Samowitz et al., 2005; Weisenberger et al., 2006)
(Figure 4.1A).
56
Figure 4.1: CIMP and covariate analyses of 100 colorectal tumor samples. A.
From left to right. Dichotomous heatmap of the DNA methylation data. Red bars indicate
PMR ≥ 10, and light blue bars indicate PMR < 10. Covariate status of each tumor.
Presence of BRAF and KRAS mutations are indicated as blue bars and wild-type are
indicated as green bars. MLH1 promoter DNA methylation status is indicated as either
methylated (blue) or unmethylated (green). Female, colon, and mucinous histology are
indicated in blue; male and rectum are indicated in green. P values: Fisher’s exact test. B.
Histogram showing the distribution of the numbers of tumors with different numbers of
methylated CIMP loci.
57
Comprehensive DNA methylation analysis and identification of novel CIMP-
associated genes
We next employed the Illumina Infinium DNA methylation technology in order to
perform comprehensive DNA methylation profiling in these 100 tumor samples. The
Illumina Infinium DNA methylation assay quantitatively measures DNA methylation for
27,578 CpG loci spanning 14,495 genes. These assays are designed in the promoters of
protein-coding genes curated in the consensus coding sequences (CCDS) database. The
majority of the reactions (72.5%) are located within a CpG island. We performed
hierarchical two-dimensional unsupervised clustering analysis using the DNA
methylation β-values (The details are provided in the Materials and Methods section).
Intriguingly, we found a large number of genes whose DNA methylation is strongly
associated with tumors with CIMP+ defined by the five-marker panel and with BRAF
mutation (Figure 4.2). Importantly, a number of the CIMP-defining markers including
CACNA1G, MLH1, RUNX3, and SOCS1 are found in this group of genes, indicating that
our five-marker CIMP panel appears to accurately identify CIMP+ tumors (Figure 4.2 on
the right).
In an attempt to obtain a list of CIMP-associated CpG sites, we used a Wilcoxon
rank-sum test to compare the DNA methylation β-value between the 16 CIMP-positive
tumors defined by the five-marker panel and 84 CIMP-negative tumors. We applied a
false-discovery rate (FDR) threshold to all Infinium loci using Benjamini and Hochberg
approach (Benjamini and Hochberg, 1995). At an FDR=0.001 cutoff, we identified 1,100
differentially methylated CpG sites. We then selected 786 CpG sites that showed a
58
minimum difference of mean β value > 0.2 between CIMP-positive group and CIMP-
negative group (Figure 4.3). Intriguingly, among these CpG sites, 776 sites (98.7%)
representing 601 genes were hypermethylated in CIMP+ tumors, and four of the five
CIMP-maker panel were included in the list; CACNA1G (FDR = 8.2×10
-11
), IGF2 (FDR
= 3.7×10
-9
), NEUROG1 (FDR = 2.9×10
-5
), and RUNX3 (FDR = 4.4×10
-9
). One of the
markers SOCS1 (FDR = 0.0038) was excluded from the list using the stringent cutoff of
FDR < 0.001. We also confirmed CIMP-specific hypermethylation of IGFBP7 as we
previously described (FDR = 6.8×10
-5
) (Chapter 3).
59
Figure 4.2: Unsupervised hierarchical clustering analysis of 100 colorectal tumor
samples. Right panel: DNA methylation profile of 100 colorectal tumors for 12,191 CpG
loci from the Illumina Infinium DNA methylation assay. Left panel: Expanded image
highlighting the CIMP-specific DNA methylation events. Tumors on the right, majority
of which is defined as CIMP+ by the five-marker panel and contains a BRAF mutation,
show high level of DNA methylation at these CpG sites. We used dark blue to yellow
gradient to represent the low and high DNA methylation β-value respectively. The details
of the clustering procedures are provided in the Materials and Methods section.
60
Figure 4.3: Identification of CIMP-associated CpG sites. The volcano plot
shows the 1 log
10
transformed FDR vs. the mean methylation difference between
CIMP+ and CIMP tumors. FDR = 0.001 and |Δβ| = 0.2 are used as a cutoff for
differential methylation. CIMP-associated DNA hypermethylation events are indicated in
red. In contrast, 10 CpG sites that are hypomethylated in CIMP+ tumors compared to
CIMP-negative tumors are indicated in green.
61
Comparison between CIMP-associated and non-CIMP-associated CpG islands
For the purpose of comparison, we also identified CpG sites that showed cancer-
specific but not CIMP-specific DNA hypermethylation. We compared the DNA
methylation β-values obtained from the Infinium assay between 100 tumors and 8 normal
colonic mucosa. Using a similar procedure as described above, we selected 594 Infinium
probes with FDR < 0.001 and mean |Δβ| > 0.2 between tumor and normal (Figure 4.4).
Among these probes, 5 probes were removed since they were also identified as CIMP-
associated probes. Therefore, we obtained 1,365 cancer-specific Infinium probes (1,036
genes), in which 776 were defined as CIMP-associated probes (601 genes) and 589 as
non-CIMP-associated probes (435 genes).
62
Figure 4.4: Identification of cancer-specific non-CIMP-associated Infinium
probes. The scatter plot on the left shows the mean normal DNA methylation vs. mean
tumor DNA methylation. FDR = 0.001 and |Δβ| = 0.2 between tumor and normal are used
as a cutoff for differential methylation. Only the cancer-specific 594 Infinium probes are
shown on the plot. The plot on the right compares mean methylation level between
CIMP-negative and CIMP-positive samples for the same 594 probes. These plots indicate
that majority of these probes are cancer-specific but not CIMP-specific.
63
We and other groups reported that genes that are targeted by Polycomb group
(PcG) proteins in embryonic stem (ES) cells are more susceptible to aberrant DNA
hypermethylation in cancer (Ohm et al., 2007; Schlesinger et al., 2007; Widschwendter et
al., 2007). Therefore, we decided to find out how the embryonic stem cell signature
differs between CIMP-associated loci and non-CIMP-associated loci. Interestingly, we
found that 23% of CIMP-associated genes contained all three ES cell repressive marks
(SUZ12, EED, H3K27Me3), whereas 30% of non-CIMP-associated genes also contained
all three ES cell polycomb marks (Figure 4.5). These observations indicate that there is a
specific group of ES cell polycomb target genes that are resistant to DNA
hypermethylation in the majority of colorectal cancers but succumb to aberrant
hypermethylation only in the context of CIMP. However, there is still a number of ES
cell polycomb target genes that appears to be protected from aberrant DNA
hypermethylation in colorectal cancer (Figure 4.5).
64
Figure 4.5: CIMP-specific DNA hypermethylation and embryonic stem cell
polycomb targets. The scatter plot on the left shows mean DNA methylation of
CIMP-negative tumors vs. the mean DNA methylation of CIMP-positive tumors.
Information on the polycomb occupancy for each Infinium locus was obtained from the
pervious study (Lee et al., 2006). Genes with all three polycomb marks (SUZ12, EED,
H3K27Me3) are indicated in green. Bar graph on the right compares percentage of the
embryonic stem cell polycomb targets in CIMP-specific genes and non-CIMP-specific
genes.
65
We also extracted total RNA simultaneously from the same tissue samples by
using TRIZOL® Reagent. We performed expression profiling on these samples using the
Illumina HumanRef-8 expression BeadChips, which measures the expression level of
18,626 different genes. Intriguingly, hierarchical unsupervised cluster analysis using the
log2-transformed expression data revealed similar expression profiles among tumors with
CIMP+, the majority of which carry a BRAF mutation (Figure 4.6).
Figure 4.6: Correlation heatmap based on the expression data. The heatmap
visualizes pearson correlation coefficients for each pairwise comparison. The samples are
arranged based on the order of unsupervised hierarchal clustering analysis using the
euclidean distance metrics. Only the variable genes with standard deviation of expression
across all samples > 1.0 was used for the clustering analysis. The black bar indicates
CIMP+ tumors. The blue and red bars indicate tumors with BRAF mutation and KRAS
mutation respectively.
66
Discussion
CIMP is defined as a high frequency of cancer-specific DNA hypermethylation in
a group of CpG islands in colorectal cancer. CIMP provides a unique opportunity to
study the molecular mechanisms that lead to aberrant DNA methylation in cancer and its
contribution in the development of cancer. In this study, we performed a comprehensive
DNA methylation profiling of colorectal tumor and normal colonic mucosa, and
identified cancer-specific DNA hypermethylation of 1,036 genes, of which 601 were
defined as CIMP-specific DNA hypermethylation. We have also generated gene
expression profiles from the same samples. Our most comprehensive list of CIMP target
genes and their expression data can be used to study the mechanisms underlying CIMP.
First, simple structural and sequence characteristics of the associated CpG islands such as
the GC content and the length might create differential susceptibility to DNA
hypermethylation in CIMP+ tumors. It is also possible that specific sequence motifs or
repeat sequences such as Alu and LINE surrounding the CpG islands might have a role in
attracting DNA hypermethylation specifically in CIMP+ tumors (Feltus et al., 2006).
Furthermore, analysis of differential gene expression between CIMP+ and CIMP– tumors
might reveal a specific trans-acting factor that causes CIMP. Intriguingly, CIMP+ tumors
showed similar expression profiles by hierarchical unsupervised cluster analysis. The
expression data may also provide us with insights into the role of CIMP-associated DNA
hypermethylation in colorectal tumorigenesis. The majority of CpG island
hypermethylation events may occur in gene promoters that are not expressed in normal
cells and are unlikely to be involved in tumor initiation or progression. Therefore, it is
67
important to determine which DNA hypermethylation events affect gene expression.
Finally, our statistical analyses indicate that it might be possible to develop a better
CIMP-marker panel that is sensitive and specific in defining CIMP+ tumors than the
currently utilized five-marker panel.
68
CHAPTER 5
Identification of a panel of potential DNA methylation-based biomarkers
for colorectal cancer
Introduction
Colorectal cancer (CRC) is the third most common type of cancer in both men
and women and the second leading cause of cancer-related death in the U.S, with
approximately 5% lifetime risk of being diagnosed with the disease. In 2008, in the U.S,
there were an estimated 148,810 new cases and 49,960 deaths (SEER, 2008; Jemal et al.,
2008). Although, the incidence of CRC are declining in the U.S, parts of the Asian
countries including Japan, Taiwan, and Eastern European countries have continued to see
rising incidence and mortality (Sung et al., 2008). Development of CRC typically takes
places over a long period of time. A stepwise progression model in CRC involves
acquisition of distinct sets of genetic and epigenetic alterations at each step of the
morphological transformation from benign adenomatous polyps to adenocarcinoma
(Jones et al., 2008). Therefore, in theory, CRC is largely preventable by the detection and
surgical removal of adenomatous polyps. The relative five-year survival rate of a
colorectal cancer patient differs dramatically depending on the stages at diagnosis. The
overall relative five-year survival rate has been reported to be 64% in the U.S. Although
the five-year survival rate is 90% when the cancer is diagnosed while still localized, only
39% of cases are detected at this early stage. As the cancer spreads to local lymph nodes,
the five-year survival rate of the patents decreases to 68%. Unfortunately, metastatic
69
colorectal cancer is present in 19% of patients at the time of initial diagnosis, and these
patient’s five-year survival rate is only 10% (SEER, 2008; Jemal et al., 2008). The
presence of a well-recognized pre-malignant lesion and the high survival rate associated
with early stages of the disease would make colorectal cancer an ideal candidate for
screening.
Current screening options developed for detection of early stages of colorectal
cancer or adenomatous polyps are divided into two general categories; stool tests and
structural exams. Noninvasive screening by testing for occult blood or the presence of
cancer-specific DNA (e.g. point mutations, microsatellite instability) in stool suffers
tremendously from its low sensitivity at detecting early stage cancer (Levin et al., 2008).
On the other hand, structural examinations such as colonoscopy and most recently
computed tomography (CT) colonography (also know as virtual colonoscopy) has been
demonstrated to be effective at detecting cancer and adenomas (Levin et al., 2008; Smith
et al., 2009). Colonoscopy is the only modality that can be used for biopsy or complete
polypectomy. There is a strong evidence indicating that removal of adenomatous polyps
results in reduced CRC incidence and morality (Winawer et al., 1993; Citarda et al.,
2001; Zauber et al., 2007; Sung et al., 2008). Despite the recent improvements in the rate
of CRC screening, particularly by colonoscopy, the prevalence of CRC screening among
average risk population remains low, and displays considerable socioeconomic disparity
(Meissner et al., 2006; CDC, 2008; Smith et al., 2009). The problem of the low
compliance with CRC screening is even worse in other parts of the world including some
of the Asian counties where the rate of CRC is raising (Sung et al., 2008).
70
Aberrant DNA methylation at CpG islands has been widely documented in
cancer. Aberrant DNA methylation is a prevalent and early event in cancer cells.
Promoter CpG island hypermethylation associated with inactivation of critical tumor
suppressor genes has been investigated extensively (Jones and Baylin, 2002).
Increasingly, this altered DNA methylation pattern has been recognized as a highly
promising cancer-specific biomarker (Laird, 2003; Laird, 2005; Cairns, 2007). It has been
demonstrated that aberrant DNA hypermethylation can be detected in stool (Chen et al.,
2005; Itzkowitz et al., 2008) and plasma samples (Lofton-Day et al., 2008; Grutzmann et
al., 2008) from CRC patients.
In this study, the goal was to identify novel markers to be used for colorectal
cancer screening. We performed the most comprehensive DNA methylation marker
screen to date in CRC using the newly developed Illumina Infinium and GoldenGate
DNA methylation platform, and validated the novel candidate markers identified using
highly sensitive MethyLight technology. Furthermore, I describe the rigorous filtering
steps we performed against control blood samples for potential use of these markers in
blood-based colorectal cancer screening. The panel of novel DNA methylation markers
that I describe here would potentially be beneficial for early diagnosis of colorectal
cancer using a noninvasive stool DNA test or a minimally invasive blood-based assay.
Materials and Methods
Primary colorectal tissue samples
Normal colonic biopsy samples were collected at the time of screening or diagnostic
colonoscopy at the University of Southern California USC/Los Angeles County Medical
71
Center and USC University Hospital. Primary colorectal tumor tissue samples for the
Illumina GoldenGate and Infinium analyses were collected and DNA was extracted as
previously described (Weisenberger et al., 2006). An independent set of colorectal tumor
samples for marker validation by MethyLight was obtained from four hospitals within the
Ontario Tumor Bank network (the Ontario Institute for Cancer Research, Ontario,
Canada). Genomic DNA and total RNA were extracted from those samples by using
TRIZOL® Reagent (Invitrogen, Burlington, ON) according to the manufacture’s
protocol. The tissue collection and analyses were approved by the University of Southern
California Institutional Review Board.
Control blood samples
Control DNA from peripheral blood mononuclear cells (referred to here as Buffy coat
DNA) and plasma samples was isolated from whole blood collected from 10 cancer-free
women (median age=59) (HemaCare Corporation, Van Nuys, CA). Plasma was separated
by spinning the blood at 1600g for 10 minutes. The isolated plasma was spun down for
another 10 minutes at 16,000g to remove any contaminating white blood cells. The white
blood cells located at the interphase between the plasma and the sedimented red blood
cells were collected in a separate tube. DNA from both the white blood cells and plasma
was isolated using the QIAamp DNA Blood Kit (QIAGEN GmbH, Hilden, Germany)
according to the manufacture’s instructions.
72
DNA methylation analyses
MethyLight
Genomic DNA was treated with sodium bisulfite using the Zymo EZ DNA Methylation
Kit (Zymo Research, Orange, CA) and subsequently analyzed by MethyLight as
described recently (Campan et al., 2009). The primer and probe sequences for the
MethyLight reactions are listed in Appendix A. The results of the MethyLight assays
were scored as PMR (Percent of Methylated Reference) values as previously defined
(Weisenberger et al., 2006; Campan et al., 2009).
Illumina GoldenGate DNA methylation assay
Genomic DNA was sodium bisulfite converted using the EZ-96 DNA Methylation Kit
(ZYMO Research) according to manufacturer’s protocol. Illumina GoldenGate DNA
methylation analysis was performed as described previously (Bibikova et al., 2006) at the
USC Genomics Core Facility. Target sequences for the assay and detailed information on
each interrogated CpG site and its associated gene on the “GoldenGate Methylation
Cancer Panel 1” are described at www.illumina.com
Illumina Infinium DNA methylation assay
Genomic DNA was bisulfite converted using the EZ-96 DNA Methylation Kit (ZYMO
Research) according to manufacturer's protocol. For improved bisulfite conversion
efficiency, a short denaturation step was included (16 cycles of 95ºC for 30 seconds
followed by 50ºC for 1 hour). After bisulfite conversion, the analyses were performed at
the USC Genomics Core Facility. The detailed assay procedures are described in
CHAPTER4, Materials and Methods section.
73
Results
Comprehensive DNA methylation marker discovery in CRC.
In an effort to discover novel candidate biomarkers for CRC, we employed the
recently developed Illumina Infinium DNA methylation platform. The Illumina Infinium
DNA methylation assay quantitatively measures DNA methylation for 27,578 CpG loci
spanning 14,495 genes. These assays are designed in the promoter of protein-coding
genes curated in the consensus coding sequences (CCDS) database. The Illumina
Infinium DNA methylation assay was performed on bisulfite converted DNA isolated
from 8 normal colonic mucosa and 20 primary colorectal tumors. We first performed
unsupervised hierarchical clustering analysis of the DNA methylation data obtained from
the Infinium assay, in order to compare the DNA methylation profiles (Figure 5.1). Two
separate classes were evident, each representing the normal colonic mucosa samples and
primary colorectal tumor samples (Figure 5.1). Notably, closer examination of the
clustered CpG sites showing cancer-specific DNA methylation events revealed extensive
DNA methylation in a group of CIMP+ tumors (Figure 5.1, lower panel).
74
Figure 5.1: Unsupervised hierarchical clustering of DNA methylation data
obtained from the Infinium assay We analyzed DNA methylation on 8 normal colonic
mucosa and 20 primary colorectal tumors by the Illumina Infinium DNA methylation
technology. Group of normal samples and tumor samples are clustered separately using
unsupervised clustering. Expanded image on the lower panel highlights the cancer-
specific DNA hypermethylation events. We used dark blue to yellow gradient to
represent low and high DNA methylation β-value respectively.
75
Marker ranking
We devised a marker ranking system intended to capture cancer-specific markers
represented even in tumors displaying a very low overall frequency of DNA
hypermethylation. We only used the markers on autosomal genes, since DNA
methylation level of X-linked loci naturally differs between male and female due to X
chromosome inactivation in female. We also filtered out markers with a fold difference in
average β-values between tumor and normal <1. For each of the remaining markers, we
first sorted tumor samples based on the DNA methylation β-values from the highest to
the lowest. We then calculated the average β-values of all the tumors below the 20
th
percentile. We also sorted the normal samples based on the DNA methylation β-values
and choose the highest β-value for each marker. We calculated the difference between the
average β-value of the lowest 20
th
percentile in tumors and the highest DNA methylation
β-values in normal for each marker, and then ranked markers by the calculated
difference. Figure 5.2 shows a heat map of the top 50 ranked markers for normal and
tumor as well as two buffy coat samples. DNA methylation information of the two buffy
coat samples will be useful for the development of potential blood-based DNA
methylation biomarkers (Figure 5.2). Among the top 50 ranked loci, which represent 47
unique genes, we designed 21 MethyLight reactions, including 10 newly designed
reactions and 11 reactions previously designed in our laboratory (Figures 5.2 and 5.3).
The Illumina GoldenGate assay was also performed on 10 non-tumor tissues from CRC
patients and 235 colorectal tumor tissues. The Illumina GoldenGate assay targets 1,505
CpG loci spanning 807 different genes, most of which are within promoter CpG islands.
76
A similar ranking method to the one described above was applied on the DNA
methylation data obtained from the Illumina GoldenGate assay. Based on the top 50
markers (33 genes), we designed 15 unique MethyLight reactions (Figure 5.3). Together,
we designed 36 MethyLight reaction based on the Infinium and GoldenGate assays
(Figure 5.3). After omitting one MethyLight reaction because it failed to amplify SSS1
treated control DNA, we tested the remaining 35 MethyLight reactions in an independent
tumor sample set (Figure 5.3).
77
Figure 5.2: Top 50 ranked markers from the Illumina Infinium analysis. Each DNA methylation β-value is visualized
by dark blue to yellow gradient to represent the low to high DNA methylation level. Top 50 ranked markets are sorted by the
mean of the β-value of the two buffy coat samples. Pink boxes beneath the gene name indicate that we have designed
MethyLight reactions for that gene.
78
Figure 5.3: Flow chart of the comprehensive DNA Methylation marker discovery.
1 All MethyLight markers were validated on an independent set of tumor samples
2 Candidate for noninvasive stool test
3 Candidate for blood-based screening
79
A highly sensitive and specific DNA methylation assay is required to detect
cancer-specific DNA hypermethylation present in very low-quantities in blood or stool of
CRC patients (Diehl et al., 2005; Diehl et al., 2008). MethyLight is a TaqMan-based real-
time PCR method for DNA methylation analysis and is well-suited for this purpose
(Weisenberger et al., 2008; Campan et al., 2009). We tested the 35 MethyLight reactions
on an independent set of 20 colorectal samples including 7 non-tumor tissues from CRC
patients and 13 colorectal tumor tissues (Figure 5.4). We validated cancer-specific DNA
methylation for most of the MethyLight reactions except a few reactions amplifying
DNA from the non-tumor tissue samples (Figures 5.3 and 5.4). We found that both
Illumina DNA methylation platforms, the Infinium assay and the GoldenGate assay,
performed equally well in identifying novel cancer-specific markers (Figure 5.4 bottom).
80
Figure 5.4: Validation of the candidate markers by MethyLight. DNA methylation of 20 colorectal samples including 7
normal and 13 colorectal tumor samples were analyzed by MethyLight. Samples are obtained either from Australia (Aus) or
Canada (OTB). PMR=10 was used to dichotomize the DNA methylation data. Blue and yellow indicates PMR <10 and PMR ≥
10 respectively. At the bottom, information on the platforms from which these markers are originally discovered are included.
GG: GoldenGata, Inf: Infinium.
81
Stringent filtering of the DNA methylation markers against blood samples
For the purpose of developing blood-based colorectal cancer screening, we
applied two additional filters on these markers. The main objective is to minimize the
false positive signals generated from WBC-derived DNA or other sources of DNA
present in blood. Our filtering processes involve two steps. For the first filtering, we
tested the MethyLight reactions on 50ng buffy coat DNA isolated from 2 healthy
individuals (> 60 years old). We eliminated fifteen markers with C(t) < 35 in any one of
the two samples (Figures 5.3 and 5.5). Next, we run remaining 16 MethyLight reactions
on 100µl plasma DNA from 10 healthy individuals (> 60 years old). We eliminated 8
markers with detectable methylation (PMR>0) in any of the 10 controls. We therefore
identified 8 markers that are not present in our control blood (Figures 5.3 and 5.5).
In summary, we identified 31 candidate markers that are cancer-specific, of which
8 markers are not present in blood from healthy individuals (Figures 5.3)
82
Figure 5.5: Stringent filtering of the 31 DNA methylation markers against blood
samples
83
Figure 5.6: The Stringent set of 8 candidate biomarkers Green and Red indicates
PMR <10 and PMR ≥ 10 respectively. Blue indicates PMR=0
84
Discussion
Despite the national guidelines endorsing regular CRC screening in average risk
individuals over 50 years old and strong evidence indicating its benefit in reducing CRC-
related mortality, the prevalence of CRC screening remains low. Colonoscopy has been
demonstrated to be effective at detecting cancer and adenomas and can be used for biopsy
or complete polypectomy (Levin et al., 2008; Smith et al., 2009). However, the major
obstacles for the widespread use of colonoscopy include a rigorous bowel preparation and
fear of discomfort or embarrassment (especially among women) associated with the
invasive procedure (Inadomi, 2008). Ideally, a robust noninvasive screening would be
available to select individuals who are likely to benefit from a follow-up colonoscopy. As
a noninvasive screening for early and sensitive detection of CRC, a newer DNA-based
stool test has shown the great potential (Chung, 2008). Traditionally, DNA-based stool
testing has been based on cancer-specific genetic changes such as point mutations and
microsatellite instability makers (Levin et al., 2008). Recently, aberrant DNA
hypermethylation at CpG islands has been recognized as a highly promising cancer-
specific biomarker (Laird, 2003; Laird, 2005; Cairns, 2007). DNA hypermethylation is a
frequent and early event in CRC and occurs in discrete regions (Jones and Baylin, 2002).
In this study, we performed the most comprehensive DNA methylation profiling to date
on primary colorectal samples and identified a panel of novel cancer-specific DNA
methylation markers in CRC. We also validated these markers in an independent set of
tumors using highly sensitive MethyLight technology. We believe that our makers are
superior to the existing DNA methylation-based marker under commercial development
85
(Ahlquist et al., 2008) for the following reasons. Previously, stool DNA test based on
DNA hypermethylation of Vimentin was evaluated (Chen et al., 2005). Chen and
colleagues used the traditional gel-based methylation-specific PCR (MSP) method to
examine DNA methylation. They reported 43% sensitivity for detecting stage I and II
colorectal cancers. However, they also found that 10% of healthy control individuals also
tested positive (Chen et al., 2005). In our Infinium analysis, we confirmed the cancer-
specific DNA hypermethylation of Vimentin. Furthermore, consistent with the previous
report, we found that the frequency of Vimentin DNA hypermethylation in our colorectal
tumor samples was only 55% (Chen et al., 2005). Our markers might be better at least for
two reasons. First, our marker detection is based on MethyLight technology, which is
more sensitive and specific in detecting DNA methylation than traditional MSP. Second,
in our system, our markers show more frequent DNA hypermethylation in CRC than
Vimentin.
Eight makers passed our rigorous tests on a panel of buffy coat DNA and plasma
samples from healthy individuals over 60 years old. Therefore, these markers would
potentially be used in blood-based CRC screening as well. It has been demonstrated that
circulating mutant DNA can be found in blood from >60% of patients with stage I or II
colorectal cancers (Diehl et al., 2005). However, detection of adenoma-derived mutant
DNA in blood was extremely problematic (Diehl et al., 2005). Recently, DNA
hypermethylation of SEPT9 was tested in plasma samples from CRC patients (Lofton-
Day et al., 2008). Lofton-Day and colleagues detected SEPT9 DNA hypermethylation in
69% of CRC plasma samples and 14% of plasma samples from healthy individuals
86
(Lofton-Day et al., 2008). In our Infinium DNA methylation analysis, we failed to
confirm cancer-specific DNA hypermethylation of SEPT9 at two locations in its
promoter CpG island. We found that SEPT9 promoter is methylated in our normal
samples and stay methylated at the same level in tumors. One possible explanation for
this discrepancy is the CpG island region examined by Grutzmann and colleagues shows
a different DNA methylation pattern than the two locations we analyzed by Infinium
(Grutzmann et al., 2008).
Fecal-based DNA analysis might be more promising for detecting early lesions of
CRC including adenoma, despite the inconvenience of obtaining specimens and difficulty
in preserving and purifying DNA from these types of specimens (Oberwalder et al., 2008;
Diehl et al., 2008). Nevertheless, insufficient quantity of DNA present in test specimens
is one of the biggest hurdles facing DNA methylation-based early detection of colorectal
cancer. Emerging technologies in highly sensitive DNA methylation detection will likely
help address this issue. We have developed digital MethyLight technology, which is
superior to the traditional MethyLight in sensitive detection for small amount of
methylated DNA in large background of unmethylated DNA (Weisenberger et al., 2008).
Using this method, we were able to detect methylated markers in 100µl plasma samples
obtained from one stage II and several stage IV breast cancer patients (Weisenberger et
al., 2008). Recently, MSP has been coupled to the single-quantum-dot-based DNA
nanosensor detection. It has been reported that this novel DNA methylation technology
can interrogate multiple DNA methylation markers in 800nl/reaction volume and detect
87
as little as 15 pg of methylated DNA in the presence of a 10,000-fold excess of
unmethylated alleles (Vasudev et al., 2008).
Future directions
We have collected 65 colorectal polyp samples and matched plasma samples from
individuals who underwent colonoscopy. Our DNA methylation makers need be tested
on these polyp samples to determine whether they are also useful in detecting precursor
lesions of CRC. The ultimate aim is to detect polyp-derived methylated DNA in plasma
using the most sensitive technology available.
88
CHAPTER 6
Summary and conclusions
The main goal of my thesis has been to understand the molecular mechanisms that
lead to epigenetic changes in colorectal cancer (CRC) and the contribution of these
changes in the development of CRC. Aberrant DNA hypermethylation at CpG islands is a
frequent and early event in CRC. Despite the large number of studies conducted over
many years, the molecular mechanism that accounts for the aberrant DNA
hypermethylation in cancer is not well understood. However, as I discussed in chapter 1,
it is clear that DNA hypermethylation plays a role in colorectal tumorigenesis when it
occurs at selected tumor suppressor genes. Recent advances in technology for DNA
methylation analysis started to reveal a large number of genes methylated in CRC.
Determining which DNA hypermethylation events actually contribute to initiation and
progression of colorectal cancer requires integrated analyses involving not only other
types of epigenetic alterations but also genetic alterations as well as transcriptome
changes.
The CpG island methylator phenotype (CIMP) exhibits an exceptionally high
frequency of cancer-specific DNA hypermethylation and is tightly linked to an activating
mutation of BRAF (BRAF
V600E
) in human colorectal cancer. Mechanisms linking
epigenetic (CIMP) and genetic (BRAF mutation) events and the temporal sequence in
which these two events take place have been widely discussed and cited in the literature.
However, no one has provided a satisfying explanation to date. In chapter 2 and 3, we
89
used colorectal cancer cell lines and primary tumors to elucidate the molecular
mechanisms for the association between BRAF mutation and CIMP in human colorectal
cancer. We considered two scenarios. First, that BRAF
V600E
plays a causal role in the
development of CIMP, or second, that CIMP provides a specific cellular context for
BRAF
V600E
to be favorably selected. In chapter 2, we examined whether expression of
BRAF
V600E
causes CIMP-specific DNA hypermethylation in cell culture system. We
found that stable expression of BRAF
V600E
is not sufficient to induce CIMP in an
established cancer cell line. In chapter 3, considering the alternative possibility, we
searched for genes whose DNA hypermethylation was tightly linked to BRAF
V600E
and
CIMP in colorectal cancer. Intriguingly, we identified CIMP-dependent DNA
hypermethylation and transcriptional inactivation of IGFBP7, which has been shown to
mediate BRAF
V600E
-induced cellular senescence and apoptosis. Inactivation of IGFBP7
by DNA hypermethylation may accommodate BRAF
V600E
by blocking the senescence
pathway. Therefore, our work provides a mechanistic rationale for the association
between BRAF
V600E
and CIMP and makes an important contribution to understanding the
role of aberrant DNA methylation in colorectal tumorigenesis.
In chapter 4, in an attempt to further characterize CIMP-associated DNA
hypermethylation, we have quantitatively determined DNA methylation status of 27,578
CpG sites located in 14,495 gene promoters in 100 colorectal tumors and 8 normal
colonic mucosae, using recently developed Illumina Infinium DNA methylation platform.
Through stringent statistical analyses, we have identified cancer-specific DNA
hypermethylation of 1,036 genes, of which 601 were defined as CIMP-specific DNA
90
hypermethylation. Our most comprehensive list of CIMP targets will provide the
opportunity to study structural and sequence characteristics of affected CpG islands in the
future, which might ultimately give us insights into the underlying basis of CIMP-
associated DNA hypermethylation. We have also determined expression of 18,626 genes
in the same 100 colorectal tumor samples. Unsupervised hierarchical cluster analyses
using the expression data revealed similar expression profiles among CIMP+ tumors,
majority of which carry BRAF mutation. Comprehensive integrated analysis of DNA
methylation and transcriptome profiling will help us better understand the role of CIMP-
associated DNA hypermethylation in colorectal tumorigenesis and find any trans-acting
factors that might cause CIMP.
In chapter 5, we have developed a panel of novel cancer-specific DNA
methylation markers that would potentially be useful for early diagnosis of colorectal
cancer using a noninvasive stool DNA test or a minimally invasive blood-based assay.
These markers will be tested on a panel of colorectal polyp samples we have collected.
Our ultimate goal is to develop a panel of maker that detects precancerous polyps using a
noninvasive stool DNA test or a minimally invasive blood-based assay.
91
BIBLIOGRAPHY
Ahlquist, D. A., Sargent, D. J., Loprinzi, C. L., Levin, T. R., Rex, D. K., Ahnen,
D. J. et al. (2008). Stool DNA and occult blood testing for screen detection of
colorectal neoplasia. Ann Intern Med, 149(7), 441-50, W81.
Barber, T. D., McManus, K., Yuen, K. W., Reis, M., Parmigiani, G., Shen, D. et
al. (2008). Chromatid cohesion defects may underlie chromosome instability in
human colorectal cancers. Proc Natl Acad Sci U S A, 105(9), 3443-3448.
Bardelli, A., Parsons, D. W., Silliman, N., Ptak, J., Szabo, S., Saha, S. et al.
(2003). Mutational analysis of the tyrosine kinome in colorectal cancers. Science,
300(5621), 949.
Bates, R. C., Goldsmith, J. D., Bachelder, R. E., Brown, C., Shibuya, M., Oettgen,
P. et al. (2003). Flt-1-dependent survival characterizes the epithelial-mesenchymal
transition of colonic organoids. Curr Biol, 13(19), 1721-1727.
Baylin, S. B., & Ohm, J. E. (2006). Epigenetic gene silencing in cancer - a
mechanism for early oncogenic pathway addiction? Nat Rev Cancer, 6(2), 107-
116.
Beach, R., Chan, A. O., Wu, T. T., White, J. A., Morris, J. S., Lunagomez, S. et al.
(2005). BRAF mutations in aberrant crypt foci and hyperplastic polyposis. Am J
Pathol, 166(4), 1069-1075.
Bellone, G., Smirne, C., Carbone, A., Buffolino, A., Scirelli, T., Prati, A. et al.
(2006). KIT/stem cell factor expression in premalignant and malignant lesions of
the colon mucosa in relationship to disease progression and outcomes. Int J Oncol,
29(4), 851-859.
Benjamini,Y., & Hochberg, Y. (1995). Controlling the false discovery rate - a
practical and powerful approach to multiple testing. J Roy Stat Soc B Met, 57(1),
289-300.
Bernstein, B. E., Meissner, A., & Lander, E. S. (2007). The mammalian
epigenome. Cell, 128(4), 669-681.
92
Bibikova, M., Lin, Z., Zhou, L., Chudin, E., Garcia, E. W., Wu, B. et al. (2006).
High-throughput DNA methylation profiling using universal bead arrays. Genome
Res, 16(3), 383-393.
Bird, A. (2002). DNA methylation patterns and epigenetic memory. Genes Dev,
16(1), 6-21.
Broderick, P., Carvajal-Carmona, L., Pittman, A. M., Webb, E., Howarth, K.,
Rowan, A. et al. (2007). A genome-wide association study shows that common
alleles of SMAD7 influence colorectal cancer risk. Nat Genet, 39(11), 1315-1317.
Cairns, P. (2007). Gene methylation and early detection of genitourinary cancer:
the road ahead. Nat Rev Cancer, 7(7), 531-543.
Campan, M., Weisenberger, D. J., Trinh, B., & Laird, P. W. (2009). MethyLight.
Methods Mol Biol, 507, 325-337.
CDC. (2008). Use of colorectal cancer tests--United States, 2002, 2004, and 2006.
MMWR Morb Mortal Wkly Rep, 57(10), 253-258.
Chan, A. O., Broaddus, R. R., Houlihan, P. S., Issa, J. P., Hamilton, S. R., &
Rashid, A. (2002). CpG island methylation in aberrant crypt foci of the
colorectum. Am J Pathol, 160(5), 1823-1830.
Chan, T. A., Glockner, S., Yi, J. M., Chen, W., Van Neste, L., Cope, L. et al.
(2008). Convergence of mutation and epigenetic alterations identifies common
genes in cancer that predict for poor prognosis. PLoS Med, 5(5), e114.
Chan, T. L., Yuen, S. T., Kong, C. K., Chan, Y. W., Chan, A. S., Ng, W. F. et al.
(2006). Heritable germline epimutation of MSH2 in a family with hereditary
nonpolyposis colorectal cancer. Nat Genet, 38(10), 1178-1183.
Chatel, G., Ganeff, C., Boussif, N., Delacroix, L., Briquet, A., Nolens, G. et al.
(2007). Hedgehog signaling pathway is inactive in colorectal cancer cell lines. Int
J Cancer, 121(12), 2622-2627.
Chen, W. D., Han, Z. J., Skoletsky, J., Olson, J., Sah, J., Myeroff, L. et al. (2005).
Detection in fecal DNA of colon cancer-specific methylation of the nonexpressed
vimentin gene. J Natl Cancer Inst, 97(15), 1124-1132.
93
Cheng, Y. W., Pincas, H., Bacolod, M. D., Schemmann, G., Giardina, S. F.,
Huang, J. et al. (2008). CpG island methylator phenotype associates with low-
degree chromosomal abnormalities in colorectal cancer. Clin Cancer Res, 14(19),
6005-6013.
Cheung, M., Sharma, A., Madhunapantula, S. V., & Robertson, G. P. (2008). Akt3
and mutant V600E B-Raf cooperate to promote early melanoma development.
Cancer Res, 68(9), 3429-3439.
Chong, S., Youngson, N. A., & Whitelaw, E. (2007). Heritable germline
epimutation is not the same as transgenerational epigenetic inheritance. Nat Genet,
39(5), 574-5; author reply 575-6.
Chung, D. C. (2008). Stool DNA testing and colon cancer prevention: another step
forward. Ann Intern Med, 149(7), 509-510.
Citarda, F., Tomaselli, G., Capocaccia, R., Barcherini, S., & Crespi, M. (2001).
Efficacy in standard clinical practice of colonoscopic polypectomy in reducing
colorectal cancer incidence. Gut, 48(6), 812-815.
Cui, H., Cruz-Correa, M., Giardiello, F. M., Hutcheon, D. F., Kafonek, D. R.,
Brandenburg, S. et al. (2003). Loss of IGF2 imprinting: a potential marker of
colorectal cancer risk. Science, 299(5613), 1753-1755.
Dankort, D., Filenova, E., Collado, M., Serrano, M., Jones, K., & McMahon, M.
(2007). A new mouse model to explore the initiation, progression, and therapy of
BRAFV600E-induced lung tumors. Genes Dev, 21(4), 379-384.
Davies, H., Bignell, G. R., Cox, C., Stephens, P., Edkins, S., Clegg, S. et al.
(2002). Mutations of the BRAF gene in human cancer. Nature, 417(6892), 949-
954.
de la Chapelle, A. (2004). Genetic predisposition to colorectal cancer. Nat Rev
Cancer, 4(10), 769-780.
Diehl, F., Li, M., Dressman, D., He, Y., Shen, D., Szabo, S. et al. (2005).
Detection and quantification of mutations in the plasma of patients with colorectal
tumors. Proc Natl Acad Sci U S A, 102(45), 16368-16373.
94
Diehl, F., Schmidt, K., Durkee, K. H., Moore, K. J., Goodman, S. N., Shuber, A.
P. et al. (2008). Analysis of Mutations in DNA Isolated From Plasma and Stool of
Colorectal Cancer Patients. Gastroenterology.
Domingo, E., Laiho, P., Ollikainen, M., Pinto, M., Wang, L., French, A. J. et al.
(2004). BRAF screening as a low-cost effective strategy for simplifying HNPCC
genetic testing. J Med Genet, 41(9), 664-668.
Eads, C. A., Danenberg, K. D., Kawakami, K., Saltz, L. B., Blake, C., Shibata, D.
et al. (2000). MethyLight: a high-throughput assay to measure DNA methylation.
Nucleic Acids Res, 28(8), E32.
East, J. E., Saunders, B. P., & Jass, J. R. (2008). Sporadic and syndromic
hyperplastic polyps and serrated adenomas of the colon: classification, molecular
genetics, natural history, and clinical management. Gastroenterol Clin North Am,
37(1), 25-46.
Eden, A., Gaudet, F., Waghmare, A., & Jaenisch, R. (2003). Chromosomal
instability and tumors promoted by DNA hypomethylation. Science, 300(5618),
455.
Esquela-Kerscher, A., & Slack, F. J. (2006). Oncomirs - microRNAs with a role in
cancer. Nat Rev Cancer, 6(4), 259-269.
Feinberg, A. P. (2007). Phenotypic plasticity and the epigenetics of human
disease. Nature, 447(7143), 433-440.
Feinberg, A. P. (2008). Epigenetics at the epicenter of modern medicine. JAMA,
299(11), 1345-1350.
Feinberg, A. P., Ohlsson, R., & Henikoff, S. (2006). The epigenetic progenitor
origin of human cancer. Nat Rev Genet, 7(1), 21-33.
Feinberg, A. P., & Tycko, B. (2004). The history of cancer epigenetics. Nat Rev
Cancer, 4(2), 143-153.
Feltus, F. A., Lee, E. K., Costello, J. F., Plass, C., & Vertino, P. M. (2006). DNA
motifs associated with aberrant CpG island methylation. Genomics, 87(5), 572-
579.
95
Fraga, M. F., Ballestar, E., Villar-Garea, A., Boix-Chornet, M., Espada, J.,
Schotta, G. et al. (2005). Loss of acetylation at Lys16 and trimethylation at Lys20
of histone H4 is a common hallmark of human cancer. Nat Genet, 37(4), 391-400.
Gal-Yam, E. N., Egger, G., Iniguez, L., Holster, H., Einarsson, S., Zhang, X. et al.
(2008). Frequent switching of Polycomb repressive marks and DNA
hypermethylation in the PC3 prostate cancer cell line. Proc Natl Acad Sci U S A.
Garnett, M. J., & Marais, R. (2004). Guilty as charged: B-RAF is a human
oncogene. Cancer Cell, 6(4), 313-319.
Gaudet, F., Hodgson, J. G., Eden, A., Jackson-Grusby, L., Dausman, J., Gray, J.
W. et al. (2003). Induction of tumors in mice by genomic hypomethylation.
Science, 300(5618), 489-492.
Goel, A., Nagasaka, T., Arnold, C. N., Inoue, T., Hamilton, C., Niedzwiecki, D. et
al. (2007). The CpG island methylator phenotype and chromosomal instability are
inversely correlated in sporadic colorectal cancer. Gastroenterology, 132(1), 127-
138.
Gosden, R. G., & Feinberg, A. P. (2007). Genetics and epigenetics--nature's pen-
and-pencil set. N Engl J Med, 356(7), 731-733.
Grady, W. M., & Carethers, J. M. (2008). Genomic and Epigenetic Instability in
Colorectal Cancer Pathogenesis. Gastroenterology.
Grady, W. M., Parkin, R. K., Mitchell, P. S., Lee, J. H., Kim, Y. H., Tsuchiya, K.
D. et al. (2008). Epigenetic silencing of the intronic microRNA hsa-miR-342 and
its host gene EVL in colorectal cancer. Oncogene, 27(27), 3880-3888.
Graells, J., Vinyals, A., Figueras, A., Llorens, A., Moreno, A., Marcoval, J. et al.
(2004). Overproduction of VEGF concomitantly expressed with its receptors
promotes growth and survival of melanoma cells through MAPK and PI3K
signaling. J Invest Dermatol, 123(6), 1151-1161.
Grutzmann, R., Molnar, B., Pilarsky, C., Habermann, J. K., Schlag, P. M., Saeger,
H. D. et al. (2008). Sensitive Detection of Colorectal Cancer in Peripheral Blood
by Septin 9 DNA Methylation Assay. PLoS ONE, 3(11), e3759.
96
Haigis, K. M., Wistuba, I. I., & Kurie, J. M. (2007). Lung premalignancy induced
by mutant B-Raf, what is thy fate? To senesce or not to senesce, that is the
question. Genes Dev, 21(4), 361-366.
Hanahan, D., & Weinberg, R. A. (2000). The hallmarks of cancer. Cell, 100(1),
57-70.
Hitchins, M. P., Wong, J. J., Suthers, G., Suter, C. M., Martin, D. I., Hawkins, N.
J. et al. (2007). Inheritance of a cancer-associated MLH1 germ-line epimutation. N
Engl J Med, 356(7), 697-705.
Horsthemke, B. (2007). Heritable germline epimutations in humans. Nat Genet,
39(5), 573-4; author reply 575-6.
Houlston, R. S., Webb, E., Broderick, P., Pittman, A. M., Di Bernardo, M. C.,
Lubbe, S. et al. (2008). Meta-analysis of genome-wide association data identifies
four new susceptibility loci for colorectal cancer. Nat Genet.
Illingworth, R., Kerr, A., Desousa, D., Jorgensen, H., Ellis, P., Stalker, J. et al.
(2008). A novel CpG island set identifies tissue-specific methylation at
developmental gene loci. PLoS Biol, 6(1), e22.
Inadomi, J. M. (2008). Taishotoyama Symposium Barriers to colorectal cancer
screening: economics, capacity and adherence. J Gastroenterol Hepatol, 23 Suppl
2, S198-204.
Issa, J. P. (2004). CpG island methylator phenotype in cancer. Nat Rev Cancer,
4(12), 988-993.
Itzkowitz, S., Brand, R., Jandorf, L., Durkee, K., Millholland, J., Rabeneck, L. et
al. (2008). A Simplified, Noninvasive Stool DNA Test for Colorectal Cancer
Detection. Am J Gastroenterol.
Jaeger, E., Webb, E., Howarth, K., Carvajal-Carmona, L., Rowan, A., Broderick,
P. et al. (2008). Common genetic variants at the CRAC1 (HMPS) locus on
chromosome 15q13.3 influence colorectal cancer risk. Nat Genet, 40(1), 26-28.
Jass, J. R. (2007a). Classification of colorectal cancer based on correlation of
clinical, morphological and molecular features. Histopathology, 50(1), 113-130.
Jass, J. R. (2007b). Molecular heterogeneity of colorectal cancer: Implications for
cancer control. Surg Oncol, 16 Suppl 1, S7-9.
97
Jemal, A., Siegel, R., Ward, E., Hao, Y., Xu, J., Murray, T. et al. (2008). Cancer
statistics, 2008. CA Cancer J Clin, 58(2), 71-96.
Jiang, X., Tan, J., Li, J., Kivimae, S., Yang, X., Zhuang, L. et al. (2008). DACT3
is an epigenetic regulator of Wnt/beta-catenin signaling in colorectal cancer and is
a therapeutic target of histone modifications. Cancer Cell, 13(6), 529-541.
Jones, P. A., & Baylin, S. B. (2007). The epigenomics of cancer. Cell, 128(4),
683-692.
Jones, P. A., & Baylin, S. B. (2002). The fundamental role of epigenetic events in
cancer. Nat Rev Genet, 3(6), 415-428.
Jones, S., Chen, W. D., Parmigiani, G., Diehl, F., Beerenwinkel, N., Antal, T. et
al. (2008). Comparative lesion sequencing provides insights into tumor evolution.
Proc Natl Acad Sci U S A, 105(11), 4283-4288.
Kambara, T., Simms, L. A., Whitehall, V. L., Spring, K. J., Wynter, C. V., Walsh,
M. D. et al. (2004). BRAF mutation is associated with DNA methylation in
serrated polyps and cancers of the colorectum. Gut, 53(8), 1137-1144.
Kaz, A. M., & Brentnall, T. A. (2006). Genetic testing for colon cancer. Nat Clin
Pract Gastroenterol Hepatol, 3(12), 670-679.
Kinzler, K. W., & Vogelstein, B. (1996). Lessons from hereditary colorectal
cancer. Cell, 87(2), 159-170.
Kohonen-Corish, M. R., Sigglekow, N. D., Susanto, J., Chapuis, P. H., Bokey, E.
L., Dent, O. F. et al. (2007). Promoter methylation of the mutated in colorectal
cancer gene is a frequent early event in colorectal cancer. Oncogene, 26(30),
4435-4441.
Kondo, Y., Shen, L., Cheng, A. S., Ahmed, S., Boumber, Y., Charo, C. et al.
(2008). Gene silencing in cancer by histone H3 lysine 27 trimethylation
independent of promoter DNA methylation. Nat Genet.
Kondo, Y., Shen, L., Yan, P. S., Huang, T. H., & Issa, J. P. (2004). Chromatin
immunoprecipitation microarrays for identification of genes silenced by histone
H3 lysine 9 methylation. Proc Natl Acad Sci U S A, 101(19), 7398-7403.
98
Kraunz, K. S., Nelson, H. H., Liu, M., Wiencke, J. K., & Kelsey, K. T. (2005).
Interaction between the bone morphogenetic proteins and Ras/MAP-kinase
signalling pathways in lung cancer. Br J Cancer, 93(8), 949-952.
Laiho, P., Kokko, A., Vanharanta, S., Salovaara, R., Sammalkorpi, H., Jarvinen,
H. et al. (2007). Serrated carcinomas form a subclass of colorectal cancer with
distinct molecular basis. Oncogene, 26(2), 312-320.
Laird, P. W., Zijderveld, A., Linders, K., Rudnicki, M. A., Jaenisch, R., & Berns,
A. (1991). Simplified mammalian DNA isolation procedure. Nucleic Acids Res,
19(15), 4293.
Laird, P. W. (2003). The power and the promise of DNA methylation markers. Nat
Rev Cancer, 3(4), 253-266.
Laird, P. W. (2005). Cancer epigenetics. Hum Mol Genet, 14 Spec No 1, R65-76.
Lauwers, G. Y., & Chung, D. C. (2006). The serrated polyp comes of age.
Gastroenterology, 131(5), 1631-1634.
Leary, R. J., Lin, J. C., Cummins, J., Boca, S., Wood, L. D., Parsons, D. W. et al.
(2008). Integrated analysis of homozygous deletions, focal amplifications, and
sequence alterations in breast and colorectal cancers. Proc Natl Acad Sci U S A,
105(42), 16224-16229.
Lee, T. I., Jenner, R. G., Boyer, L. A., Guenther, M. G., Levine, S. S., Kumar, R.
M. et al. (2006). Control of developmental regulators by Polycomb in human
embryonic stem cells. Cell, 125(2), 301-313.
Levin, B., Lieberman, D. A., McFarland, B., Smith, R. A., Brooks, D., Andrews,
K. S. et al. (2008). Screening and surveillance for the early detection of colorectal
cancer and adenomatous polyps, 2008: a joint guideline from the American Cancer
Society, the US Multi-Society Task Force on Colorectal Cancer, and the American
College of Radiology. CA Cancer J Clin, 58(3), 130-160.
Ligtenberg, M. J., Kuiper, R. P., Chan, T. L., Goossens, M., Hebeda, K. M.,
Voorendt, M. et al. (2008). Heritable somatic methylation and inactivation of
MSH2 in families with Lynch syndrome due to deletion of the 3' exons of
TACSTD1. Nat Genet.
99
Lin, J., Lai, M., Huang, Q., Ma, Y., Cui, J., & Ruan, W. (2007). Methylation
patterns of IGFBP7 in colon cancer cell lines are associated with levels of gene
expression. J Pathol, 212(1), 83-90.
Lin, J., Lai, M., Huang, Q., Ruan, W., Ma, Y., & Cui, J. (2008). Reactivation of
IGFBP7 by DNA demethylation inhibits human colon cancer cell growth in vitro.
Cancer Biol Ther, 7(12).
Lofton-Day, C., Model, F., Devos, T., Tetzner, R., Distler, J., Schuster, M. et al.
(2008). DNA methylation biomarkers for blood-based colorectal cancer screening.
Clin Chem, 54(2), 414-423.
Loh, K., Chia, J. A., Greco, S., Cozzi, S. J., Buttenshaw, R. L., Bond, C. E. et al.
(2008). Bone morphogenic protein 3 inactivation is an early and frequent event in
colorectal cancer development. Genes Chromosomes Cancer, 47(6), 449-460.
Lu, R., Wang, X., Chen, Z. F., Sun, D. F., Tian, X. Q., & Fang, J. Y. (2007).
Inhibition of the extracellular signal-regulated kinase/mitogen-activated protein
kinase pathway decreases DNA methylation in colon cancer cells. J Biol Chem,
282(16), 12249-12259.
Lujambio, A., Ropero, S., Ballestar, E., Fraga, M. F., Cerrato, C., Setien, F. et al.
(2007). Genetic unmasking of an epigenetically silenced microRNA in human
cancer cells. Cancer Res, 67(4), 1424-1429.
Makinen, M. J. (2007). Colorectal serrated adenocarcinoma. Histopathology,
50(1), 131-150.
McGarvey, K. M., Fahrner, J. A., Greene, E., Martens, J., Jenuwein, T., & Baylin,
S. B. (2006). Silenced tumor suppressor genes reactivated by DNA demethylation
do not return to a fully euchromatic chromatin state. Cancer Res, 66(7), 3541-
3549.
Meissner, H. I., Breen, N., Klabunde, C. N., & Vernon, S. W. (2006). Patterns of
colorectal cancer screening uptake among men and women in the United States.
Cancer Epidemiol Biomarkers Prev, 15(2), 389-394.
Michaloglou, C., Vredeveld, L. C., Soengas, M. S., Denoyelle, C., Kuilman, T.,
van der Horst, C. M. et al. (2005). BRAFE600-associated senescence-like cell
cycle arrest of human naevi. Nature, 436(7051), 720-724.
100
Minoo, P., Baker, K., Goswami, R., Chong, G., Foulkes, W. D., Ruszkiewicz, A.
R. et al. (2006). Extensive DNA methylation in normal colorectal mucosa in
hyperplastic polyposis. Gut, 55(10), 1467-1474.
Minoo, P., Moyer, M., & Jass, J. (2007). Role of BRAF-V600E in the serrated
pathway of colorectal tumourigenesis. J Pathol, 212(2), 124-133.
Mooi, W. J., & Peeper, D. S. (2006). Oncogene-induced cell senescence--halting
on the road to cancer. N Engl J Med, 355(10), 1037-1046.
Mori, Y., Yin, J., Rashid, A., Leggett, B. A., Young, J., Simms, L. et al. (2001).
Instabilotyping: comprehensive identification of frameshift mutations caused by
coding region microsatellite instability. Cancer Res, 61(16), 6046-6049.
Muthuswami, R., Mesner, L. D., Wang, D., Hill, D. A., Imbalzano, A. N., &
Hockensmith, J. W. (2000). Phosphoaminoglycosides inhibit SWI2/SNF2 family
DNA-dependent molecular motor domains. Biochemistry, 39(15), 4358-4365.
Neklason, D. W., Kerber, R. A., Nilson, D. B., Anton-Culver, H., Schwartz, A. G.,
Griffin, C. A. et al. (2008). Common familial colorectal cancer linked to
chromosome 7q31: a genome-wide analysis. Cancer Res, 68(21), 8993-8997.
Nguyen, C. T., Weisenberger, D. J., Velicescu, M., Gonzales, F. A., Lin, J. C.,
Liang, G. et al. (2002). Histone H3-lysine 9 methylation is associated with
aberrant gene silencing in cancer cells and is rapidly reversed by 5-aza-2'-
deoxycytidine. Cancer Res, 62(22), 6456-6461.
Noren, N. K., & Pasquale, E. B. (2004). Eph receptor-ephrin bidirectional signals
that target Ras and Rho proteins. Cell Signal, 16(6), 655-666.
O'Brien, M. J., Yang, S., Mack, C., Xu, H., Huang, C. S., Mulcahy, E. et al.
(2006). Comparison of microsatellite instability, CpG island methylation
phenotype, BRAF and KRAS status in serrated polyps and traditional adenomas
indicates separate pathways to distinct colorectal carcinoma end points. Am J Surg
Pathol, 30(12), 1491-1501.
O'Brien, M. J. (2007). Hyperplastic and serrated polyps of the colorectum.
Gastroenterol Clin North Am, 36(4), 947-68, viii.
101
O'Brien, M. J., Yang, S., Clebanoff, J. L., Mulcahy, E., Farraye, F. A., Amorosino,
M. et al. (2004). Hyperplastic (serrated) polyps of the colorectum: relationship of
CpG island methylator phenotype and K-ras mutation to location and histologic
subtype. Am J Surg Pathol, 28(4), 423-434.
Oberwalder, M., Zitt, M., Wontner, C., Fiegl, H., Goebel, G., Zitt, M. et al. (2008).
SFRP2 methylation in fecal DNA--a marker for colorectal polyps. Int J Colorectal
Dis, 23(1), 15-19.
Ogino, S., Nosho, K., Kirkner, G. J., Kawasaki, T., Meyerhardt, J. A., Loda, M. et
al. (2009). CpG island methylator phenotype, microsatellite instability, BRAF
mutation and clinical outcome in colon cancer. Gut, 58(1), 90-96.
Ohm, J. E., McGarvey, K. M., Yu, X., Cheng, L., Schuebel, K. E., Cope, L. et al.
(2007). A stem cell-like chromatin pattern may predispose tumor suppressor genes
to DNA hypermethylation and heritable silencing. Nat Genet, 39(2), 237-242.
Oliveira, C., Westra, J. L., Arango, D., Ollikainen, M., Domingo, E., Ferreira, A.
et al. (2004). Distinct patterns of KRAS mutations in colorectal carcinomas
according to germline mismatch repair defects and hMLH1 methylation status.
Hum Mol Genet, 13(19), 2303-2311.
Ordway, J. M., Williams, K., & Curran, T. (2004). Transcription repression in
oncogenic transformation: common targets of epigenetic repression in cells
transformed by Fos, Ras or Dnmt1. Oncogene, 23(21), 3737-3748.
Park, S., Yeung, M. L., Beach, S., Shields, J. M., & Yeung, K. C. (2005). RKIP
downregulates B-Raf kinase activity in melanoma cancer cells. Oncogene, 24(21),
3535-3540.
Popat, S., Hubner, R., & Houlston, R. S. (2005). Systematic review of
microsatellite instability and colorectal cancer prognosis. J Clin Oncol, 23(3), 609-
618.
Powell, S. M., Zilz, N., Beazer-Barclay, Y., Bryan, T. M., Hamilton, S. R.,
Thibodeau, S. N. et al. (1992). APC mutations occur early during colorectal
tumorigenesis. Nature, 359(6392), 235-237.
Rodriguez, J., Munoz, M., Vives, L., Frangou, C. G., Groudine, M., & Peinado,
M. A. (2008). Bivalent domains enforce transcriptional memory of DNA
methylated genes in cancer cells. Proc Natl Acad Sci U S A.
102
Rosenberg, D. W., Yang, S., Pleau, D. C., Greenspan, E. J., Stevens, R. G., Rajan,
T. V. et al. (2007). Mutations in BRAF and KRAS differentially distinguish
serrated versus non-serrated hyperplastic aberrant crypt foci in humans. Cancer
Res, 67(8), 3551-3554.
Rustgi, A. K. (2007). The genetics of hereditary colon cancer. Genes Dev, 21(20),
2525-2538.
Saito, Y., Liang, G., Egger, G., Friedman, J. M., Chuang, J. C., Coetzee, G. A. et
al. (2006). Specific activation of microRNA-127 with downregulation of the
proto-oncogene BCL6 by chromatin-modifying drugs in human cancer cells.
Cancer Cell, 9(6), 435-443.
Sakatani, T., Kaneda, A., Iacobuzio-Donahue, C. A., Carter, M. G., de Boom
Witzel, S., Okano, H. et al. (2005). Loss of imprinting of Igf2 alters intestinal
maturation and tumorigenesis in mice. Science, 307(5717), 1976-1978.
Samowitz, W. S., Albertsen, H., Herrick, J., Levin, T. R., Sweeney, C., Murtaugh,
M. A. et al. (2005). Evaluation of a large, population-based sample supports a
CpG island methylator phenotype in colon cancer. Gastroenterology, 129(3), 837-
845.
Samowitz, W. S., Albertsen, H., Sweeney, C., Herrick, J., Caan, B. J., Anderson,
K. E. et al. (2006). Association of smoking, CpG island methylator phenotype, and
V600E BRAF mutations in colon cancer. J Natl Cancer Inst, 98(23), 1731-1738.
Samowitz, W. S., Slattery, M. L., Sweeney, C., Herrick, J., Wolff, R. K., &
Albertsen, H. (2007). APC mutations and other genetic and epigenetic changes in
colon cancer. Mol Cancer Res, 5(2), 165-170.
Samowitz, W. S., Sweeney, C., Herrick, J., Albertsen, H., Levin, T. R., Murtaugh,
M. A. et al. (2005). Poor survival associated with the BRAF V600E mutation in
microsatellite-stable colon cancers. Cancer Res, 65(14), 6063-6069.
Santos, F., & Dean, W. (2004). Epigenetic reprogramming during early
development in mammals. Reproduction, 127(6), 643-651.
Schlesinger, Y., Straussman, R., Keshet, I., Farkash, S., Hecht, M., Zimmerman, J.
et al. (2007). Polycomb-mediated methylation on Lys27 of histone H3 pre-marks
genes for de novo methylation in cancer. Nat Genet, 39(2), 232-236.
103
Schneikert, J., & Behrens, J. (2007). The canonical Wnt signalling pathway and its
APC partner in colon cancer development. Gut, 56(3), 417-425.
Schuebel, K., Chen, W., & Baylin, S. B. (2006). CIMPle origin for promoter
hypermethylation in colorectal cancer? Nat Genet, 38(7), 738-740.
Schuebel, K. E., Chen, W., Cope, L., Glockner, S. C., Suzuki, H., Yi, J. M. et al.
(2007). Comparing the DNA hypermethylome with gene mutations in human
colorectal cancer. PLoS Genet, 3(9), 1709-1723.
SEER. (2008). SEER Cancer Statistics Review, 1975-2005.
http://seer.cancer.gov/csr/1975_2005.
Shen, L., Kondo, Y., Guo, Y., Zhang, J., Zhang, L., Ahmed, S. et al. (2007).
Genome-wide profiling of DNA methylation reveals a class of normally
methylated CpG island promoters. PLoS Genet, 3(10), 2023-2036.
Smith, R. A., Cokkinides, V., & Brawley, O. W. (2009). Cancer screening in the
United States, 2009: A review of current American Cancer Society guidelines and
issues in cancer screening. CA Cancer J Clin, 59(1), 27-41.
Spring, K. J., Zhao, Z. Z., Karamatic, R., Walsh, M. D., Whitehall, V. L., Pike, T.
et al. (2006). High prevalence of sessile serrated adenomas with BRAF mutations:
a prospective study of patients undergoing colonoscopy. Gastroenterology,
131(5), 1400-1407.
Steemers, F. J., Chang, W., Lee, G., Barker, D. L., Shen, R., & Gunderson, K. L.
(2006). Whole-genome genotyping with the single-base extension assay. Nat
Methods, 3(1), 31-33.
Sung, J. J., Lau, J. Y., Young, G. P., Sano, Y., Chiu, H. M., Byeon, J. S. et al.
(2008). Asia Pacific consensus recommendations for colorectal cancer screening.
Gut, 57(8), 1166-1176.
Suter, C. M., Norrie, M., Ku, S. L., Cheong, K. F., Tomlinson, I., & Ward, R. L.
(2003). CpG island methylation is a common finding in colorectal cancer cell
lines. Br J Cancer, 88(3), 413-419.
Suter, C. M., & Martin, D. I. (2007). Inherited epimutation or a haplotypic basis
for the propensity to silence? Nat Genet, 39(5), 573; author reply 576.
104
Suter, C. M., Martin, D. I., & Ward, R. L. (2004). Germline epimutation of MLH1
in individuals with multiple cancers. Nat Genet, 36(5), 497-501.
Tenesa, A., Farrington, S. M., Prendergast, J. G., Porteous, M. E., Walker, M.,
Haq, N. et al. (2008). Genome-wide association scan identifies a colorectal cancer
susceptibility locus on 11q23 and replicates risk loci at 8q24 and 18q21. Nat
Genet, 40(5), 631-637.
Tomlinson, I., Webb, E., Carvajal-Carmona, L., Broderick, P., Kemp, Z., Spain, S.
et al. (2007). A genome-wide association scan of tag SNPs identifies a
susceptibility variant for colorectal cancer at 8q24.21. Nat Genet, 39(8), 984-988.
Tomlinson, I. P., Webb, E., Carvajal-Carmona, L., Broderick, P., Howarth, K.,
Pittman, A. M. et al. (2008). A genome-wide association study identifies
colorectal cancer susceptibility loci on chromosomes 10p14 and 8q23.3. Nat
Genet, 40(5), 623-630.
Toyota, M., Suzuki, H., Sasaki, Y., Maruyama, R., Imai, K., Shinomura, Y. et al.
(2008). Epigenetic silencing of microRNA-34b/c and B-cell translocation gene 4
is associated with CpG island methylation in colorectal cancer. Cancer Res,
68(11), 4123-4132.
Toyota, M., Ahuja, N., Ohe-Toyota, M., Herman, J. G., Baylin, S. B., & Issa, J. P.
(1999). CpG island methylator phenotype in colorectal cancer. Proc Natl Acad Sci
U S A, 96(15), 8681-8686.
Valle, L., Serena-Acedo, T., Liyanarachchi, S., Hampel, H., Comeras, I., Li, Z. et
al. (2008). Germline allele-specific expression of TGFBR1 confers an increased
risk of colorectal cancer. Science, 321(5894), 1361-1365.
Varambally, S., Cao, Q., Mani, R. S., Shankar, S., Wang, X., Ateeq, B. et al.
(2008). Genomic loss of microRNA-101 leads to overexpression of histone
methyltransferase EZH2 in cancer. Science, 322(5908), 1695-1699.
Vasudev, B., Zhang, Y., Easwaran, H., Griffiths, E., Herman, J., Baylin, S. et al.
(2008). High throughput, quantitative DNA methylation screening using a
quantum dot based nanotechnology assay. AACR Meeting
Abstracts/3_Molecular_Diagnostics_Meeting, PL01, PL01.
105
Veigl, M. L., Kasturi, L., Olechnowicz, J., Ma, A. H., Lutterbaugh, J. D.,
Periyasamy, S. et al. (1998). Biallelic inactivation of hMLH1 by epigenetic gene
silencing, a novel mechanism causing human MSI cancers. Proc Natl Acad Sci U
S A, 95(15), 8698-8702.
Velculescu, V. E. (2008). Defining the blueprint of the cancer genome.
Carcinogenesis, 29(6), 1087-1091.
Vogelstein, B., Fearon, E. R., Hamilton, S. R., Kern, S. E., Preisinger, A. C.,
Leppert, M. et al. (1988). Genetic alterations during colorectal-tumor
development. N Engl J Med, 319(9), 525-532.
Vogelstein, B., & Kinzler, K. W. (2004). Cancer genes and the pathways they
control. Nat Med, 10(8), 789-799.
Wajapeyee, N., Serra, R. W., Zhu, X., Mahalingam, M., & Green, M. R. (2008).
Oncogenic BRAF induces senescence and apoptosis through pathways mediated
by the secreted protein IGFBP7. Cell, 132(3), 363-374.
Walther, A., Houlston, R., & Tomlinson, I. (2008). Association between
chromosomal instability and prognosis in colorectal cancer: a meta-analysis. Gut,
57(7), 941-950.
Wang, G. G., Allis, C. D., & Chi, P. (2007). Chromatin remodeling and cancer,
Part I: Covalent histone modifications. Trends Mol Med, 13(9), 363-372.
Weisenberger, D. J., Campan, M., Long, T. I., Kim, M., Woods, C., Fiala, E. et al.
(2005). Analysis of repetitive element DNA methylation by MethyLight. Nucleic
Acids Res, 33(21), 6823-6836.
Weisenberger, D. J., Siegmund, K. D., Campan, M., Young, J., Long, T. I., Faasse,
M. A. et al. (2006). CpG island methylator phenotype underlies sporadic
microsatellite instability and is tightly associated with BRAF mutation in
colorectal cancer. Nat Genet, 38(7), 787-793.
Weisenberger, D. J., Trinh, B. N., Campan, M., Sharma, S., Long, T. I.,
Ananthnarayan, S. et al. (2008). DNA methylation analysis by digital bisulfite
genomic sequencing and digital MethyLight. Nucleic Acids Res.
106
Widschwendter, M., Fiegl, H., Egle, D., Mueller-Holzner, E., Spizzo, G., Marth,
C. et al. (2007). Epigenetic stem cell signature in cancer. Nat Genet, 39(2), 157-
158.
Winawer, S. J., Zauber, A. G., Ho, M. N., O'Brien, M. J., Gottlieb, L. S.,
Sternberg, S. S. et al. (1993). Prevention of colorectal cancer by colonoscopic
polypectomy. The National Polyp Study Workgroup. N Engl J Med, 329(27),
1977-1981.
Wood, L. D., Parsons, D. W., Jones, S., Lin, J., Sjoblom, T., Leary, R. J. et al.
(2007). The genomic landscapes of human breast and colorectal cancers. Science,
318(5853), 1108-1113.
Woodson, K., Flood, A., Green, L., Tangrea, J. A., Hanson, J., Cash, B. et al.
(2004). Loss of insulin-like growth factor-II imprinting and the presence of screen-
detected colorectal adenomas in women. J Natl Cancer Inst, 96(5), 407-410.
Xing, M. (2007). BRAF mutation in papillary thyroid cancer: pathogenic role,
molecular bases, and clinical implications. Endocr Rev, 28(7), 742-762.
Yamada, Y., Jackson-Grusby, L., Linhart, H., Meissner, A., Eden, A., Lin, H. et
al. (2005). Opposing effects of DNA hypomethylation on intestinal and liver
carcinogenesis. Proc Natl Acad Sci U S A, 102(38), 13580-13585.
Yamashita, K., Dai, T., Dai, Y., Yamamoto, F., & Perucho, M. (2003). Genetics
supersedes epigenetics in colon cancer phenotype. Cancer Cell, 4(2), 121-131.
Young, J., Jenkins, M., Parry, S., Young, B., Nancarrow, D., English, D. et al.
(2007). Serrated pathway colorectal cancer in the population: genetic
consideration. Gut, 56(10), 1453-1459.
Zanke, B. W., Greenwood, C. M., Rangrej, J., Kustra, R., Tenesa, A., Farrington,
S. M. et al. (2007). Genome-wide association scan identifies a colorectal cancer
susceptibility locus on chromosome 8q24. Nat Genet, 39(8), 989-994.
Zauber, A. G., Winawer, S. J., O'brien, M. J., Shi, W. J., & Bayuga, S. (2007).
Significant long term reduction in colorectal cancer mortality with colonoscopic
polypectomy: Findings of the National Polyp Study. Gastroenterology, 132(4),
A50-A50.
107
Zeng, Q., Phukan, S., Xu, Y., Sadim, M., Rosman, D. S., Pennison, M. et al.
(2009). Tgfbr1 haploinsufficiency is a potent modifier of colorectal cancer
development. Cancer Res, 69(2), 678-686.
Zhu, J., Woods, D., McMahon, M., & Bishop, J. M. (1998). Senescence of human
fibroblasts induced by oncogenic Raf. Genes Dev, 12(19), 2997-3007.
108
Appendix A: MethyLight reaction probe and primer sequences
HUGO Gene
Name
Reaction
Number
Previously
Published?
Forward Primer Sequence Reverse Primer Sequence Probe Oligo Sequence
BNIP3 HB-363 No TGCGTACGCGTAGGTTTTAAGTC CAAAACGAAACCGAACGAATCT CGACCGCGTCGCCCATTAACC
CACNA1G HB-158 Yes (a) TTTTTTCGTTTCGCGTTTAGGT CTCGAAACGACTTCGCCG AAATAACGCCGAATCCGACAACCGA
CDO1 HB-684 No CGGGTTTTTTTTAAGCGTTTG CCACCTTTCTTAAATAACTCTCCG CTCGTCCTCCCTCCGAAAACCGTT
CMTM2 HB-724 No TTTTCGTTGTTTTGTGTGAAAGTC AACTACGAACACCGAATCACGA CCGCGAACGCAACGCCAAA
DBC1 HB-552 No CGAATATACGTATTAAACGTAAACGACG CTAATTACGAATTTAAATAAATAACGTATATATCCC CCGCGCGCTCCCGACGTAT
EPHA5 HB-549 No GTTCGTCGAGGTGTAGAGTAGTAGTCG GAAACCCCGAAATACGAAACG ACCCCCAAACGACGACGACGACA
EYA4 HB-316 No GGAAAGAGTTGCGGGAAAAGT ACCAAAACTCCGAACTACGACAAA AACGCGCCCAACCGCCG
EYA4 HB-317 No TGGATAGGATGGAAGTTTTGCG AACTACCGACAACGCGACG CGCTCCGACCGTTCCCGACTT
FAM19A4 HB-728 No GCACCCCTACTCAAACGACG GAGTGGTGGTTATAGCGAGGC CCACTACCGACCTCCAAACGCCTCT
FGF8 HB-532 No TTCGGAGATTATTGCGTAGTCG CGAAAAACCCCGAACCTAAACTT TCTAAATCTCCGCCTCCGAACCCAACC
FLT1 HB-548 No CGGGTGGTTGAGCGTAGC CGCTTAAATCCCCAAACCG CGCTCGACGAACGCCCGC
FOXE1 HB-417 No GGGTTAGTTCGCGACGATTTT CGAACCTAACGTCCCCGA CGAACGCTCGACCCTTCTACGAAAAACT
GAS7 HB-705 No GCGTCGGGAAGTAGAGATTCG CTAAAATCCGCGAACTACCCG CGAAAACGTCGCCGCTCGCTCT
HS3ST2 HB-517 No AGTATTTTTCGGTTTTAGGCGATG CGAACCCGACGCACGTA CGAACCGCCCGACTACCCGAAAC
IGF2 HB-319 Yes (a) GAGCGGTTTCGGTGTCGTTA CCAACTCGATTTAAACCGACG CCCTCTACCGTCGCGAACCCGA
ITPKB HB-726 No CGAAACTCAACACTACCCTCCTC AGTTTGGTGATTATGAATAGCGTTAAC AAACGAAAACGTCTCGCTACCACTAAACCC
KCNQ5 HB-725 No CTCTACAACCCCTCCTTTCCC CGTTCGTTAGTTAGAGATTTTTGAGTC CCCCGAAACCCGAACGATTTACTAACTACC
LRRC4 HB-721 No GAACGAACCGAAATAAACGTCA TTTCGATCGGTTTAAAATAGTTGC AAACCGCGCGCAAATCGCAA
MLH1 HB-150 Yes (b) AGGAAGAGCGGATAGCGATTT TCTTCGTCCCTCCCTAAAACG CCCGCTACCTAAAAAAATATACGCTTACGCG
NEFL HB-520 No TGGAGACGTTTCGGGTGTATATTT AAACACCGACGCCGAATAAC AAACGTACGATACTATAACCGCTACGCACGCT
NEFL HB-528 No GTACGGATAGCGAGGAAGATATCG TACGTAAAAACGCCCCG AATAC ACGCAACGACTACAACACCGCACG
NEUROG1 HB-261 Yes (a) CGTGTAGCGTTCGGGTATTTGTA CGATAATTACGAACACACTCCGAAT CGATAACGACCTCCCGCGAACATAAA
NGF HB-730 No CTAACGCGCCAACTACCGA AAGTTTACGGTTATGTTAGGGTTTAGTTC CTCTCCGCGCCCATTCGCTCT
NPY HB-529 No TATTCGTGCGGTTGCGGT AAAACCTCGAACAATCACGCTT CGCCACTTCCCGCCCCTAAATAACG
NTRK3 HB-550 No GGTTGCGTATTTTAGGTGATTTCG GACGACCGAAATCCTCTCCG CCGAATCGCGCCTCGCCAA
POU3F1 HB-682 No GTTCGAGTTGGGCGTATTTTC CACCTACACCCGAACGC ACGACGACGACTCATCGATAAACGAACA
PRKAR1B HB-651 no ATTTTGAGTCGTTTATTTTAGTCG GCGATAATCGTAAACGAAACG CATCAAAAACGACCCCGCGACCT
RASGRF1 HB-555 No AGGGATTGGCGATTTGCG GCGCTACTAAAACGAACTCTTCG CGTCACATAACTCCGCCTACCCCTCG
RUNX3 HB-181 Yes (a) CGTTCGATGGTGGACGTGT GACGAACAACGTCTTATTACAACGC CGCACGAACTCGCCTACGTAATCCG
109
Appendix A: continued
SCGB3A1 HB-194 Yes (a) GGCGTAGCGGGCGTC CTACGTAACCCTATCCTACAACTCCG CGAACTCCTAACGCGCACGATAAAACCTAA
SFRP1 HB-201 Yes (a) GAATTCGTTCGCGAGGGA AAACGAACCGCACTCGTTACC CCGTCACCGACGCGAAAACCAAT
SFRP2 HB-279 No TTTATAATTTTGATTTTTTTACGGTATTGG GAAACCGCCTCGACGAACT CTCGAATCTCCAACCACCGTTCAACAA
SFRP2 HB-280 Yes (a) GCGTTTTAGTCGTCGGTTGTTAGT AAACGACCGAAATTCGAACTTATC CGAACCCGCTCTCTTCGCTAAATACGA
SLC18A3 HB-632 No TCGTTTCGTCGACGTCGTA TCGAAACTATAACGCGACGAAT CCCGAATAATACCCGAACCACTTACGTCG
SFRP1 HB-201 Yes (a) GAATTCGTTCGCGAGGGA AAACGAACCGCACTCGTTACC CCGTCACCGACGCGAAAACCAAT
SFRP2 HB-279 No TTTATAATTTTGATTTTTTTACGGTATTGG GAAACCGCCTCGACGAACT CTCGAATCTCCAACCACCGTTCAACAA
SFRP2 HB-280 Yes (a) GCGTTTTAGTCGTCGGTTGTTAGT AAACGACCGAAATTCGAACTTATC CGAACCCGCTCTCTTCGCTAAATACGA
SLC18A3 HB-632 No TCGTTTCGTCGACGTCGTA TCGAAACTATAACGCGACGAAT CCCGAATAATACCCGAACCACTTACGTCG
SOCS1 HB-042 Yes (b) GCGTCGAGTTCGTGGGTATTT CCGAAACCATCTTCACGCTAA ACAATTCCGCTAACGACTATCGCGCA
TFPI2 HB-361 No TTGGAGTAGAAAGTCGCGTATTTTT AAAAATCGAACGACCCGCTA ACAAAACGTCCGAAAAAACGCCTAACGAA
TMEFF2 HB-274 Yes (a) CGACGAGGAGGTGTAAGGATG CAACGCCTAACGAACGAACC TATAACTTCCGCGACCGCCTCCTCCT
TMEFF2 HB-628 No GTTAAATTCGCGTATGATTTCGAGA TTCCCGCGTCTCCGAC AACGAACGACCCTCTCGCTCCGAA
TWIST1 HB-047 Yes (c) GTAGCGCGGCGAACGT AAACGCAACGAATCATAACCAAC CCAACGCACCCAATCGCTAAACGA
WNT2 HB-551 No AAGGTTAGTTCGGATCGTTTCG AAAAACGCGCAAAACTTTCG ACTTTCGAACTACAAACGCTCGCTACGCCT
All primer and probe sequences are listed 5' to 3'.
All probes have a 5' 6FAM fluorophore, and a Black Hole Quencher (BHQ-1) at the 3' end.
(a) Weisenberger, D.J. et al Nature Genet 38, 787-793 (2006).
(b) Fiegl, H. et al Cancer Epidemiol BioMark A. Faasseers Prev 13,882-888 (2004)
(c) Muller, H.M. et al. Cancer Lett 209, 231-236 (2004)
Abstract (if available)
Abstract
Aberrant DNA hypermethylation of CpG islands is a common and early event in colorectal tumorigenesis. Promoter DNA hypermethylation is associated with transcriptional gene silencing, and can contribute to tumorigenesis when it occurs at a critical tumor suppressor gene. A distinct subset of colorectal cancers display the CpG island methylator phenotype (CIMP), characterized by an exceptionally high frequency of cancer-specific DNA hypermethylation. Recent studies have shown that an activating mutation of BRAF (BRAFV600E) is specifically associated with CIMP. We used colorectal cancer cell lines and primary tumors to elucidate the molecular mechanisms for the association. We first examined whether expression of BRAFV600E causes CIMP-specific DNA hypermethylation in the CIMP-negative, BRAF wild-type COLO 320DM colorectal cancer cell line. We found that stable expression of BRAFV600E is not sufficient to induce CIMP in our system. Secondly, considering the alternative possibility, we searched for genes whose DNA hypermethylation was tightly linked to BRAFV600E and CIMP in colorectal cancer. Intriguingly, we identified CIMP-dependent DNA hypermethylation and transcriptional inactivation of IGFBP7, a mediator of BRAFV600E-induced cellular senescence and apoptosis. Inactivation of IGFBP7 by DNA hypermethylation may accommodate BRAFV600E by blocking the senescence pathway. Therefore, our work provides a mechanistic rationale for the association between BRAFV600E and CIMP. Moreover, in an attempt to further characterize CIMP-associated DNA hypermethylation, we have quantitatively determined DNA methylation status of 27,578 CpG sites located in 14,495 gene promoters in 100 colorectal tumors and 8 normal mucosae. Through stringent statistical methods, we have identified cancer-specific DNA hypermethylation of 1,036 genes, of which 601 show CIMP-specific DNA hypermethylation. We have also generated gene expression data.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Integrated genomic & epigenomic analyses of glioblastoma multiforme: Methods development and application
PDF
DNA methylation as a biomarker in human reproductive health and disease
PDF
The kinetic study of engineered MBD domain interactions with methylated DNA: insight into binding of methylated DNA by MBD2b
PDF
DNA methylation inhibitors and epigenetic regulation of microRNA expression
PDF
DNA methylation markers for blood-based detection of small cell lung cancer in mouse models
PDF
Effects of chromatin regulators during carcinogenesis
PDF
Epigenetic mechanisms driving bladder cancer
PDF
Epigenetic regulation of non CPG island gene promoters
PDF
Functional DNA methylation changes in normal and cancer cells
PDF
Investigating the role of SASH1 gene located on chromosome 6 in ovarian cancer
PDF
Ancestral inference and cancer stem cell dynamics in colorectal tumors
PDF
Identification of DNA methylation markers in diffuse large B-cell lymphoma
PDF
Gene expression and angiogenesis pathway across DNA methylation subtypes in colon adenocarcinoma
PDF
The role of microRNAs in cancer
PDF
Differential role of two coactivators, CCAR1 and CARM1, for dysregulated beta-catenin activity in colorectal cancer cell growth and gene expression
PDF
Role of DNA methyltransferases 3A and 3B in inheritance of DNA methylation patterns
PDF
DNA methylation and gene expression profiles in Vidaza treated cultured cancer cells
PDF
Differential methylation analysis of colon tissues
PDF
The expression of human carboxylesterases in normal tissues and cancer cell-lines
PDF
The role of histone H4 lysine 20 monomethylation in gene expression and differentiation
Asset Metadata
Creator
Hinoue, Toshinori
(author)
Core Title
DNA hypermethylation: its role in colorectal tumorigenesis and potential clinical applications
School
Keck School of Medicine
Degree
Doctor of Philosophy
Degree Program
Biochemistry and Molecular Biology
Degree Conferral Date
2009-05
Publication Date
05/13/2011
Defense Date
03/27/2009
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
CIMP,colorectal cancer,DNA methylation,OAI-PMH Harvest
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Laird, Peter W. (
committee chair
), Coetzee, Gerhard A. (
committee member
), Dubeau, Louis (
committee member
), Rice, Judd C. (
committee member
)
Creator Email
thinoue@usc.edu,tnori22@gmail.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-m2247
Unique identifier
UC1208159
Identifier
etd-Hinoue-2850 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-238424 (legacy record id),usctheses-m2247 (legacy record id)
Legacy Identifier
etd-Hinoue-2850.pdf
Dmrecord
238424
Document Type
Dissertation
Rights
Hinoue, Toshinori
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
CIMP
colorectal cancer
DNA methylation