Comprehensive profiling of DNA methylation in colorectal cancer reveals subgroups with distinct clinicopathological and molecular features.
ABSTRACT Most previous studies of the CpG island methylator phenotype (CIMP) in colorectal cancer (CRC) have been conducted on a relatively small numbers of CpG sites. In the present study we performed comprehensive DNA methylation profiling of CRC with the aim of characterizing CIMP subgroups.
DNA methylation at 1,505 CpG sites in 807 cancer-related genes was evaluated using the Illumina GoldenGate methylation array in 28 normal colonic mucosa and 91 consecutive CRC samples. Methylation data was analyzed using unsupervised hierarchical clustering. CIMP subgroups were compared for various clinicopathological and molecular features including patient age, tumor site, microsatellite instability (MSI), methylation at a consensus panel of CpG islands and mutations in BRAF and KRAS.
A total of 202 CpG sites were differentially methylated between tumor and normal tissue. Unsupervised hierarchical clustering of methylation data from these sites revealed the existence of three CRC subgroups referred to as CIMP-low (CIMP-L, 21% of cases), CIMP-mid (CIMP-M, 14%) and CIMP-high (CIMP-H, 65%). In comparison to CIMP-L tumors, CIMP-H tumors were more often located in the proximal colon and showed more frequent mutation of KRAS and BRAF (P<0.001).
Comprehensive DNA methylation profiling identified three CRC subgroups with distinctive clinicopathological and molecular features. This study suggests that both KRAS and BRAF mutations are involved with the CIMP-H pathway of CRC rather than with distinct CIMP subgroups.
- SourceAvailable from: PubMed Central[Show abstract] [Hide abstract]
ABSTRACT: A subset of colorectal cancers (CRCs) harbor the CpG island methylator phenotype (CIMP), with concurrent multiple promoter hypermethylation of tumor-related genes. A serrated pathway in which CIMP is developed from serrated polyps is proposed. The present study characterized CIMP and morphologically examined precursor lesions of CIMP. In total, 104 CRCs treated between January 1996 and December 2004 were examined. Aberrant promoter methylation of 15 cancer-related genes was analyzed. CIMP status was classified according to the number of methylated genes and was correlated with the clinicopathological features, including the concomitant polyps in and around the tumors. The frequency of aberrant methylation in each CRC showed a bimodal pattern, and the CRCs were classified as CIMP-high (CIMP-H), CIMP-low (CIMP-L) and CIMP-negative (CIMP-N). CIMP-H was associated with aberrant methylation of MLH1 (P=0.005) and with an improved recurrence-free survival (RFS) rate following curative resection compared with CIMP-L/N (five-year RFS rate, 93.8 vs. 67.1%; P=0.044), while CIMP-N tumors were associated with frequent distant metastases at diagnosis (P=0.023). No concomitant serrated lesions were present in the tumors, whereas conventional adenoma was contiguous with 11 (10.6%) of 104 CRCs, including four CIMP-H CRCs. CIMP-H was classified in CRCs by a novel CIMP marker panel and the presence of concomitant tumors revealed that certain CIMP-H CRCs may have arisen from conventional adenomas.Oncology letters 11/2014; 8(5):1937-1944. · 0.24 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: It is clear that colorectal cancer (CRC) develops through multiple genetic and epigenetic pathways. These pathways may be determined on the basis of three molecular features: (i) mutations in DNA mismatch repair genes, leading to a DNA microsatellite instability (MSI) phenotype, (ii) mutations in APC and other genes that activate Wnt pathway, characterized by chromosomal instability (CIN) phenotype, and (iii) global genome hypermethylation, resulting in switch off of tumor suppressor genes, indicated as CpG island methylator phenotype (CIMP). Each of these pathways is characterized by specific pathological features, mechanisms of carcinogenesis and process of tumor development. The molecular aspects of these pathways have been used clinically in the diagnosis, screening and management of patients with colorectal cancer. In this review we especially describe various aspects of CIMP, one of the important and rather recently discovered pathways that lead to colorectal cancer.Gastroenterology and hepatology from bed to bench. 01/2013; 6(3):120-128.
- [Show abstract] [Hide abstract]
ABSTRACT: High-throughput 'omic' data, such as gene expression, DNA methylation, DNA copy number, has played an instrumental role in furthering our understanding of the molecular basis in states of human health and disease. As cells with similar morphological characteristics can exhibit entirely different molecular profiles and because of the potential that these discrepancies might further our understanding of patient-level variability in clinical outcomes, there is significant interest in the use of high-throughput 'omic' data for the identification of novel molecular subtypes of a disease. While numerous clustering methods have been proposed for identifying of molecular subtypes, most were developed for single "omic' data types and may not be appropriate when more than one 'omic' data type are collected on study subjects. Given that complex diseases, such as cancer, arise as a result of genomic, epigenomic, transcriptomic, and proteomic alterations, integrative clustering methods for the simultaneous clustering of multiple 'omic' data types have great potential to aid in molecular subtype discovery. Traditionally, ad hoc manual data integration has been performed using the results obtained from the clustering of individual 'omic' data types on the same set of patient samples. However, such methods often result in inconsistent assignment of subjects to the molecular cancer subtypes. Recently, several methods have been proposed in the literature that offers a rigorous framework for the simultaneous integration of multiple 'omic' data types in a single comprehensive analysis. In this paper, we present a systematic review of existing integrative clustering methods.Translational Cancer Research 06/2014; 3(3):202-216.
Ang et al. BMC Cancer 2010, 10:227
Comprehensive profiling of DNA methylation in
colorectal cancer reveals subgroups with distinct
clinicopathological and molecular features
© 2010 Ang et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons At-
tribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any
medium, provided the original work is properly cited.
Pei Woon Ang1,2, Marie Loh1,2, Natalia Liem3, Pei Li Lim3, Fabienne Grieu1, Aparna Vaithilingam2, Cameron Platell1,4,
Wei Peng Yong3, Barry Iacopetta1 and Richie Soong*2
Background: Most previous studies of the CpG island methylator phenotype (CIMP) in colorectal cancer (CRC) have
been conducted on a relatively small numbers of CpG sites. In the present study we performed comprehensive DNA
methylation profiling of CRC with the aim of characterizing CIMP subgroups.
Methods: DNA methylation at 1,505 CpG sites in 807 cancer-related genes was evaluated using the Illumina
GoldenGate® methylation array in 28 normal colonic mucosa and 91 consecutive CRC samples. Methylation data was
analyzed using unsupervised hierarchical clustering. CIMP subgroups were compared for various clinicopathological
and molecular features including patient age, tumor site, microsatellite instability (MSI), methylation at a consensus
panel of CpG islands and mutations in BRAF and KRAS.
Results: A total of 202 CpG sites were differentially methylated between tumor and normal tissue. Unsupervised
hierarchical clustering of methylation data from these sites revealed the existence of three CRC subgroups referred to
as CIMP-low (CIMP-L, 21% of cases), CIMP-mid (CIMP-M, 14%) and CIMP-high (CIMP-H, 65%). In comparison to CIMP-L
tumors, CIMP-H tumors were more often located in the proximal colon and showed more frequent mutation of KRAS
and BRAF (P < 0.001).
Conclusions: Comprehensive DNA methylation profiling identified three CRC subgroups with distinctive
clinicopathological and molecular features. This study suggests that both KRAS and BRAF mutations are involved with
the CIMP-H pathway of CRC rather than with distinct CIMP subgroups.
DNA hypermethylation-induced gene silencing is a com-
mon event in many malignancies and serves as an alter-
native mechanism to genetic mutation for the loss of
tumor suppressor functions [1,2]. Although the mecha-
nisms that underlie aberrant DNA methylation in cancer
cells remain to be elucidated, current evidence suggests
that it may be an early and possibly even an initiating
event in the development of colorectal cancer (CRC).
A subset of CRC has been shown to exhibit frequent
and concurrent hypermethylation at specific gene pro-
moters and is referred to as the CpG island methylator
phenotype (CIMP+) . CIMP+ CRC is associated with
distinct clinicopathological and molecular features
including proximal tumor location, preponderance in
elderly females, poorly differentiated and mucinous
tumor histology, microsatellite instability (MSI) and fre-
quent BRAF V600E mutation [3-10]. CIMP+ CRC often
lack the hallmark genetic alterations in APC, p53 and 18q
that characterize the classic
sequence. Instead, CIMP+ tumors are thought to develop
along an alternate serrated adenoma pathway in which
hypermethylation rather than mutation is used to inacti-
vate tumor suppressor genes .
* Correspondence: email@example.com
2 Cancer Science Institute of Singapore, National University of Singapore,
Full list of author information is available at the end of the article
Ang et al. BMC Cancer 2010, 10:227
Page 2 of 8
In an effort to establish CIMP+ CRC as a distinct sub-
group of CRC, Laird and colleagues analysed the methy-
lation of 195 individual gene promoter regions in 295
CRC using the quantitative MethyLight assay . From
their results, they proposed a panel of 5 CpG island meth-
ylation markers to standardize the classification of
CIMP+ CRC. However, different groups have continued
to use a variety of methylation markers to define CIMP+
CRC [7,12-14]. The lack of consensus markers has led to
reports of several CIMP subgroups according to the fre-
quency of CpG island methylation [13-15]. The investiga-
tors who originally proposed CIMP recently described
two subgroups of CIMP+, termed CIMP-1 and CIMP-2,
that displayed increased frequencies of BRAF and KRAS
mutations, respectively . Similarly, Nagasaka et al
described two distinct patterns of gene methylation in
CRC that also segregated with BRAF and KRAS muta-
tions [13,16]. Using a panel of 8 methylation markers,
Ogino et al identified a CRC subgroup which they termed
CIMP-low that was associated with frequent KRAS muta-
tion, MGMT methylation and occurrence in males .
Most previous studies of CIMP+ CRC have investigated
a relatively small number of CpG island markers for
methylation. The GoldenGate Methylation BeadArray
(Illumina, Inc.) technology provides the opportunity for
high-throughput methylation analysis of a large number
of CpG sites. In the present study the GoldenGate Methy-
lation Cancer Panel I containing 1,505 CpG loci within
807 cancer-related genes was used to study methylation
patterns in 91 unselected CRC. These genes were
selected based on their involvement in cell growth con-
trol, differentiation, migration, apoptosis, DNA damage
repair and oxidative metabolism. The GoldenGate tech-
nology allowed us to identify three distinct CRC sub-
groups according to their methylation pattern which
showed distinctive clinicopathological and molecular
characteristics and differed in their frequencies of BRAF
and KRAS mutation.
Unselected cases of CRC and adjacent normal colonic
mucosa were obtained from 91 patients undergoing sur-
gical resection at St John of God Hospital, Subiaco, West-
ern Australia. All samples were snap-frozen in liquid
nitrogen at the time of surgery and stored at -80°C until
use. This set of tumors contains well-annotated clinico-
pathological information including age, gender, tumor
location, staging, presence of lymphocytic infiltration and
careful pathological assessment of perineural (PNI), lym-
phovascular (LVI) and extramural invasion (EMVI).
Informed consent was obtained from all patients and the
project was approved by the Human Research Ethics
Committee of St John of God Hospital.
BRAF mutation, KRAS mutation and microsatellite
DNA was extracted from approximately 25 mg of tissue
using standard phenol-chloroform extraction. Hotspot
mutations in BRAF (V600E) and KRAS (codons 12 and
13) were identified using fluorescent single strand confor-
mation polymorphism (F-SSCP) as described previously
[18,19]. Deletions in the BAT-26 mononucleotide repeat
were detected using F-SSCP and this was used to estab-
lish MSI+ status .
MethyLight determination of CIMPW status
Sodium bisulfite modification was performed using the
EZ DNA methylation kit according to the manufacturer's
instructions (Zymo Research, Orange, CA) and eluted
into 20 μl of 10 mmol/L Tris-HCl (pH 8). The required
amount of genomic DNA to ensure reliable evaluation of
DNA methylation following bisulfite modification was
determined as described previously . DNA methyla-
tion levels for the panel of markers (RUNX3, CACNA1G,
IGF2, NEUROG1, SOCS1) described by Weisenberger et
al  were measured using MethyLight as described by
the authors. The percentage of methylated reference
(PMR) was calculated and normalised against β-actin to
account for variability in the amount of input bisulfite-
treated DNA. SssI methylase-treated DNA was used as
the methylated standard. A threshold PMR value of > 4
was used to classify loci as methylated or non-methy-
lated. In the present study, CIMPW refers to the classifica-
tion of CIMP using the panel of markers described by
Weisenberger et al., whereby CIMPW-high is defined as 3
or more methylated loci, CIMPW-low as 1 or 2 methy-
lated loci and CIMPW-negative as no methylated loci.
DNA methylation profiling using Illumina GoldenGate®
methylation bead array
Comprehensive DNA methylation profiling using the
Illumina Goldengate Methylation Arrays® (Illumina, San
Diego, CA) was carried out as described by Bibikova et al
 on 91 CRC and 28 randomly selected, matched nor-
mal colonic mucosa samples. Briefly, DNA was quantified
by real-time PCR and treated with bisulfite as for the
MethyLight assay. Human sperm DNA and Universal
methylated DNA (Chemicon, Temcula, CA) were
included in each run as unmethylated and methylated
controls, respectively. The bisulfite-converted DNA was
probed at 1,505 individual CpG loci contained within 807
genes in the GoldenGate Methylation Cancer Panel I
according to the manufacturer's instructions (Illumina).
Hybridised arrays were scanned using the BeadArray
Reader (Illumina). Extraction and normalization of inten-
sity data was performed using the Beadscan software. To
ensure adequate sample quality, only samples having >
Ang et al. BMC Cancer 2010, 10:227
Page 3 of 8
75% loci with a detection P-value of < 0.05 were included
The methylation level at each CpG site, or β-value, was
defined as the ratio of methylated allele to the sum of
methylated and unmethylated alleles and ranged from 0
(completely unmethylated) to 1 (completely methylated).
Normalisation of background intensity was estimated
from a set of built-in negative controls and subtracted
from each methylation data point as performed in other
studies [23,24]. All statistical analyses were carried out
using β-value as a continuous variable unless specified
otherwise. To compare the number of methylated genes
between different tumor subgroups, β-values were bina-
rized using a methylated threshold of 0.297. Using this
threshold, methylated controls in the array were classified
as unmethylated at a 5% false discovery rate. A total of 84
CpG sites contained within 39 X-chromosome genes
were excluded from the analysis in order to eliminate
Unsupervised and supervised hierarchical clustering
analyses were performed with the heatmap.2 function in
the gplots library. Unsupervised clustering was used to
characterize methylation patterns in an unbiased fashion,
as performed in other studies using methylation arrays
[14,25-27]. Supervised clustering analysis was used to
further investigate methylation differences observed in
unsupervised clustering. The optimal number of clusters
was determined using the Hubert & Levine internal clus-
ter quality index . The robustness of this number was
evaluated by bootstrap resampling analysis (n = 1000).
Additional evidence to support the delineation of clusters
was obtained through unsupervised principal component
analysis. The frequency and level of CpG methylation
across different clusters was compared using a two-sam-
ple proportion test based on both binarised and continu-
ous β-values. The association of clinicopathological and
molecular variables with each cluster was analysed using
continuous β-values and the two-sample proportion t-
test. All statistical analyses were performed in R version
2.7.1 (The R Foundation for Statistical Computing) at 5%
significance level unless otherwise stated. Where applica-
ble, Bonferroni correction was applied to adjust for mul-
DNA methylation patterns in normal and tumor tissue
Unsupervised hierarchical clustering of DNA methyla-
tion data from 1,505 CpG sites in 28 samples of normal
colonic mucosa revealed no distinct clusters [Additional
file 1]. As expected, the methylation status of 84 CpG
sites in 39 genes located in the X-chromosome was per-
fectly correlated with gender [Additional file 1]. These
genes were excluded from subsequent analyses. For the
91 tumor samples, three clusters were observed when
methylation data from all 1,505 loci were included in the
analysis [Additional file 2].
A total of 202 CpG sites, corresponding to 132 genes
(90 hypermethylated and 42 hypomethylated), were dif-
ferentially methylated between tumor and normal colonic
mucosa (P < 0.001, FDR 5%) [Additional file 3]. Unsuper-
vised hierarchical clustering of methylation data from
these 202 tumor-specific markers identified three major
tumor groups (Fig. 1), referred to here as CIMP-high
(CIMP-H; 59/91, 65%), CIMP-mid (CIMP-M; 13/91,
14%) and CIMP-low (CIMP-L; 19/91, 21%). The mean
methylation level (β-value) of the 202 CpG sites for these
groups was 0.617, 0.506 and 0.370, respectively (P <
0.001). Binarization of the methylation readings using a
β-value cut-off of 0.297 revealed a decreasing number of
methylated CpG sites for the three groups (167, 136 and
105 respectively; P < 0.001).
Although branching of the dendogram suggested the
existence of two subgroups within CIMP-H (Fig. 1), the
mean methylation level and the frequency of methylation
between these groups were not significantly different (P =
0.37 and P = 0.90 respectively). Additional evidence for
the validity of tumor segregation was obtained through
unsupervised principal component analysis (PCA) [Addi-
tional file 4]. CIMP-H could be clearly segregated from
CIMP-L. CIMP-H and CIMP-M could also be discrimi-
nated from each other, although less distinctly. This is
presumably because of a greater similarity between these
two groups [Additional file 4]. In further support, 3 was
the most frequent optimal number of clusters in boot-
strap resampling analysis.
CIMP subgroups show distinctive clinicopathological and
The distribution of clinicopathological and molecular
features for 91 CRC in relation to the methylation pattern
obtained from analysis of all 202 differentially methylated
CpG sites is shown in Fig. 1. Calculation of associations
between these features and the three CIMP subgroups
are shown in Table 1. Similar to previous reports on
CIMP+, the CIMP-H tumors in this study were signifi-
cantly associated with older age, proximal tumor location
and BRAF mutation relative to CIMP-M and CIMP-L
tumors. CIMP-H was also significantly associated with
MSI+ when compared to CIMP-M, but not CIMP-L
tumors. Two of the 15 MSI+ tumors were observed in the
CIMP-L group and 13 in the CIMP-H group. Interest-
ingly, the two patients with CIMP-L MSI+ tumors were
aged 44 and 60 years, suggesting the underlying cause of
the MSI+ phenotype was germline or somatic mutation
of the mismatch repair genes rather than hMLH1 methy-
lation. Indeed, no hMLH1 methylation was detected in
Ang et al. BMC Cancer 2010, 10:227
Page 4 of 8
both these cases in the GoldenGate Methylation Array
All 16 tumors classified as CIMPW-high by Methylight
analysis using the panel of markers proposed by Weisen-
berger et al (>3/5 sites methylated) were contained within
the CIMP-H group, while all 18 tumors classified as
CIMPW-low (1/5 or 2/5 sites methylated) segregated into
the CIMP-H or CIMP-M groups. Unfortunately, CpG
sites for only 2 (RUNX3, IGF2) of the 5 genes in the
CIMPW panel were included in the Golden Gate arrays,
thus preventing comparison of CIMP status by array and
Methylight analysis. Nevertheless, there was good corre-
lation between the array and Methylight methods for
methylation levels of RUNX3 and IGF2 (all p < 0.05) using
All 15 tumors with BRAF mutation were CIMP-H. A
significantly higher frequency of KRAS mutation was
observed in CIMP-H compared to CIMP-L or CIMP-M
tumors. None of the 13 CIMP-M tumors contained a
KRAS mutation. The presence of extramural vascular
invasion (EMVI) was more frequent in CIMP-M com-
pared to CIMP-H or CIMP-L tumors. The presence of a
tumor-infiltrating lymphocytic response (TILS) was not
associated with any of the CIMP subgroups.
Differentially methylated genes in CIMP subgroups
Five clusters of CpG loci, termed A to E, were apparent
following unsupervised hierarchical clustering of methy-
lation data for the 202 CpG loci that showed tumor-spe-
cific methylation (Fig. 1). CpG sites in clusters A and C
were more highly methylated in CIMP-M and CIMP-H
Figure 1 Unsupervised hierarchical clustering of 202 tumor-specific probes (rows) in 91 CRC (columns). The 3 tumor clusters generated by this
analysis were termed CIMP-High (CIMP-H), CIMP-Mid (CIMP-M) and CIMP-Low (CIMP-L). Clinicopathological and molecular features are shown above
the heatmap. White rectangles are cases with missing data. Gender: female (red), male (blue); Age: ≥ 67 years (red), < 67 (blue); Tumor location: prox-
imal (red), distal (blue); Tumor stage (ACPS): A or B (blue), C or D (red); Lymphovascular invasion (LVI): present (red), absent (blue); Extramural vascular
invasion (EMVI): present (red), absent (blue); Perineural invasion (PNI): present (red), absent (blue); Tumor infiltrating lymphocytes (TILS): present (red),
absent (blue); CIMPW: CIMPW-high (red), CIMPW-low (yellow), CIMPW-negative (blue); BRAF: mutant (red), wildtype (blue); KRAS: mutant (red), wildtype
(blue); Microsatellite instability (MSI): positive (red), negative (blue). Five CpG clusters (A-E) were apparent from the analysis and showed differential
methylation amongst the 3 CIMP subgroups.
Ang et al. BMC Cancer 2010, 10:227
Page 5 of 8
tumors compared to CIMP-L tumors, while the converse
was true for the CpG sites in cluster D. CpG sites in clus-
ter B and cluster E showed uniformly high and low meth-
ylation, respectively, in each of the 3 CIMP subgroups.
Using published data from studies on human stem cells
, 50% (39/98) of the genes within clusters A and C
were found to be targets for binding by Polycomb repres-
sive complex 2 (PRC2) components and/or H3K27 trim-
ethylation. In contrast, only 12% (5/41) of the genes
within clusters B, D and E were targets (P < 0.001). These
observations support previous reports that hypermethy-
lated genes in cancer are frequent targets of PRC2-medi-
ated H3K27 trimethylation .
The current study is the first to use array-based technol-
ogy to enable comprehensive methylation profiling of
CRC. A total of 1,505 CpG sites contained within 807
genes were assessed in 91 consecutive cases of CRC. The
GoldenGate® arrays employed here were recently used to
Table 1: Clinicopathological and molecular characteristics of CIMP subgroups.
CIMP subgroup (n, %)
L vs MM vs HL vs H
19 (21)13 (14)59 (65)
Female 6 (32) 4 (31)30 (51)
Male13 (68) 9 (69)29 (49)0.9520.2110.175
Age ≥ 67 years6 (32)5 (38)37 (63)
Age < 67 years13 (68)8 (62)22 (37)0.5700.0030.012
5 (26)1 (8)29 (49)
Distal tumor site2
14 (74)12 (92) 29 (49)0.152 < 0.0010.001
ACPS Stage A or B 8 (42) 4 (31)36 (61)
ACPS Stage C or D 11 (58)9 (69) 23 (39)0.520 0.0250.105
LVI Negative15 (79) 6 (46)39 (66)
LVI Positive4 (21)7 (54) 20 (34)0.049 0.1880.126
EMVI Negative 19 (100)8 (62) 52 (88)
EMVI Positive0 (0) 5 (38)7 (12)0.0050.0310.024
PNI Negative17 (89) 11 (85)52 (88)
PNI Positive2 (11)2 (15) 7 (12)0.744 0.506 0.708
9 (47)5 (38) 26 (44)
TILS Positive7 (37) 8 (62)33 (56)0.326 0.5470.521
CIMPW - negative4
19 (100)10 (77)28 (47)
CIMPW - low 0 (0)3 (23)15 (25)
CIMPW - high0 (0)0 (0)16 (27)1.000< 0.001< 0.001
MSI+ 2 (11) 0 (0)13 (18)
MSI-17 (89)13 (100)46 (78)0.125 < 0.001 0.221
BRAF mutant0 (0)0 (0) 15 (25)
BRAF wildtype 19 (100)13 (100) 44 (75) 1.000< 0.001 < 0.001
KRAS mutant 3 (16)0 (0)26 (44)
KRAS wildtype16 (84) 13 (100)33 (56) 0.057 < 0.0010.014
L, CIMP-low; M, CIMP-mid; H, CIMP-high; LVI, lymphovascular invasion; EMVI, extramural vascular invasion; PNI, perineural invasion; TILS,
tumour infiltrating lymphocytes; CIMPW, classification of CIMP using the Weisenberger et al panel, whereby CIMPW-high is defined as 3 or
more methylated loci, CIMPW-low as 1 or 2 methylated loci and CIMPW-negative as no methylated loci. MSI, microsatellite instability; 1Tumor
location was unknown for 1 patient in CIMP-H, 2Number of distal CRC that were located at rectal site were 13, 1 and 11 in CIMP-L, CIMP-M and
CIMP-H respectively, 3TILS data unknown for 3 patients in CIMP-L, 4P value for CIMPW was generated from comparison between CIMPW-high
and CIMPW-low or CIMPW-negative.
Ang et al. BMC Cancer 2010, 10:227
Page 6 of 8
profile methylation in head and neck cancer , renal
cancer , glioblastoma  and hematological neo-
plasms [24,27]. The validity of these arrays for the quanti-
tative assessment of methylation was shown in several
previous studies by comparison with other quantitative
methods [23,26,34]. The finding that methylation of CpG
sites in X-linked genes correlated with gender provided
further validation [Additional file 1]. Many of the genes
found to be hypermethylated in this study were previ-
ously reported to be methylated in CRC [Additional file
5]. Finally, in agreement with earlier work on cancer ,
many of the genes showing de novo hypermethylation in
this study of CRC (cluster A and C genes, Fig. 1) are
known targets for PRC2 .
Similar to earlier studies in CRC that evaluated a lim-
ited number of methylation markers [3-10], comprehen-
sive methylation profiling in the present study revealed
the existence of distinct tumor subgroups (Fig. 1). The
three major subgroups identified by unsupervised hierar-
chical clustering were classified as CIMP-H, CIMP-M
and CIMP-L according to the level and frequency of
methylation. In agreement with previous studies, CIMP-
H tumors were associated with older patient age, proxi-
mal site and BRAF mutation (Table 1). All 16 tumors
identified as CIMPW-high using a proposed consensus
panel of 5 markers were contained within the CIMP-H
group, as well as all 15 tumors containing a BRAF muta-
tion. Using small numbers of methylation markers in
unselected CRC, the original studies by Toyota et al
reported CIMP+ frequencies of 62%  and 51% 
whereas subsequent studies reported lower frequencies
of 15-32% [5,7-10], . In contrast, by investigating a
large number of methylation sites and using unsuper-
vised hierarchical clustering to analyze the results, we
observed a relatively high proportion (65%) of CIMP-H
tumors in the present study.
Previous studies have reported inconsistent results for
the association between CpG island methylation and
KRAS mutation [8,13,17,35,36], probably because of the
different methylation markers used in each study. Analy-
sis of a large number of CpG sites in the present study
revealed that CIMP-H tumors showed a significantly
higher KRAS mutation frequency compared to both
CIMP-M and CIMP-L tumors (Table 1). This result
agrees with some studies [4,5,8,35] but not others that
found an inverse association between KRAS mutation
and CIMP+ [7,9,10].
Since BRAF mutations are strongly associated with
CIMP and mutually exclusive to KRAS mutations (;
Fig. 1), a point of interest is whether methylation patterns
differ between tumors with BRAF and KRAS mutations.
Supervised clustering analysis with Bonferroni correction
revealed that only 1 of the 202 tumor-specific CpG sites
was differentially methylated between these tumor
groups (HTR1B_P222_F, upregulated in BRAF mutant
tumors, p = 8.1 × 10-6). HTR1B (5-hydroxytryptamine
(serotonin) receptor 1B) is a G protein-coupled multi-
pass membrane protein involved in regulation of the
serotonin system . The gene is hypermethylated in
lung cancer and its chromosome locus (16q14.1) is fre-
quently deleted in a number of cancer types . How-
ever, no links with BRAF or RAS mutations or signaling
have been reported.
A novel finding of this array-based analysis was the
existence of an apparent CIMP-M group (Fig. 1). These
tumors displayed a higher frequency of EMVI compared
to both CIMP-L and CIMP-H, and a significantly higher
stage compared to CIMP-H (Table 1). CIMP-M tumors
were almost exclusively located in the distal colon or rec-
tum (12/13, 92%). MSI and KRAS and BRAF mutations
were notably absent in these tumors, although this may
be due to reportedly lower frequencies of these altera-
tions in distal tumors . Taken together, these results
suggest CIMP-M tumors could be a distinct clinical and
molecular entity, although confirmation in larger, inde-
pendent tumor series is required.
After adjustment for multiple testing, 170 CpG sites
were hypermethylated in CIMP-H compared to CIMP-L.
The 112 genes containing these CpG sites are ranked
according to significance in Additional file 5. Of these, 54
were previously reported as methylated in cancer, 38 as
methylated in gastrointestinal cancers and 30 in CRC
[Additional file 5]. Of the top 10 genes that were hyperm-
ethylated in CIMP-H compared to CIMP-L tumors, 5
have previously been implicated in the pathogenesis of
gastrointestinal tumors (NTRK3, HS3ST2, TWIST1,
CD40 and EYA4). Somatic mutation of NTRK3 has been
reported in human colon cancer , while methylation
of EYA4 has been documented previously in ulcerative
colitis-associated dysplasia  and CRC .
CIMP-M tumors were found to have a relatively high
incidence of EMVI (38%) compared to CIMP-H and
CIMP-L tumors (Table 1). Supervised analysis revealed
that HS3ST2, also known as 3-OST-2, was the only gene
to be differentially methylated between tumors showing
presence or absence of EMVI. Methylation-associated
silencing of HS3ST2 expression has been demonstrated in
breast, lung, pancreatic and colon cancers . This gene
encodes an enzyme that modifies heparin sulfate proteo-
glycans  involved in cell adhesion and migration ,
thus suggesting a possible mechanistic link between
HS3ST2 methylation and EMVI.
The use of Illumina GoldenGate® Beadarray technology
in this study allowed a large number of CpG sites to be
evaluated for methylation in an unbiased fashion. How-
ever, there are several limitations with this approach for
the characterization of CIMP subgroups in CRC. Firstly,
only a small fraction of all genes were investigated for
Ang et al. BMC Cancer 2010, 10:227
Page 7 of 8
methylation and in 70% of these just one CpG site per
gene was evaluated. Secondly, it is unclear whether the
methylation level at these sites relates to expression of the
genes. Thirdly, some of the probes used in this assay con-
tain single nucleotide polymorphisms (SNPs) or repeti-
tive elements that could influence methylation analysis
. The cost effectiveness of using arrays to characterize
CIMP-H is questionable, given the strong concordance
between CIMP-H from this study and CIMPw. Further
studies should clarify if the additional information pro-
vided by methylation arrays is worth the complexity and
Methylation profiling of 807 cancer-related genes
revealed the presence of three CRC subgroups with dis-
tinct clinicopathological and molecular features. Similar
to earlier studies that investigated fewer methylation
markers, CIMP-H CRC were associated with older
patient age, proximal location and mutations in BRAF
and KRAS. Further investigations in large and indepen-
dent population-based series are required to validate
these findings and to assess the clinical utility of CIMP
The authors declare that they have no competing interests.
PWA, FG, NL, PLL, AV carried out the experimental work. ML performed the sta-
tistical analysis. CP and WPY contributed gave critical clinical perspective to the
results, and BI and RS co-ordinated the study and compiled the manuscript. All
authors read and approved the final version of the manuscript.
This work is supported by the Singapore Cancer Syndicate (SCS#BU51), and the
Singapore National Research Foundation and the Ministry of Education under
the Research Center of Excellence Programme. We thank Patrick Tan and his
laboratory staff for assistance with data generation and technical support. PWA
is supported by the International Postgraduate Research Scholarship.
1School of Surgery, University of Western Australia, Perth, Australia, 2Cancer
Science Institute of Singapore, National University of Singapore, Singapore,
Singapore, 3Department of Haematology and Oncology, National University
Hospital, Singapore, Singapore and 4Department of Surgery, St John of God
Hospital, Subiaco, Australia
1.Esteller M: Epigenetics in cancer. N Engl J Med 2008, 358(11):1148-1159.
2. Issa JP: CpG island methylator phenotype in cancer. Nat Rev Cancer
3.Toyota M, Ahuja N, Ohe-Toyota M, Herman JG, Baylin SB, Issa JP: CpG
island methylator phenotype in colorectal cancer. Proc Natl Acad Sci
USA 1999, 96(15):8681-8686.
4. Barault L, Charon-Barra C, Jooste V, de la Vega MF, Martin L, Roignot P, Rat
P, Bouvier AM, Laurent-Puig P, Faivre J, et al.: Hypermethylator phenotype
in sporadic colon cancer: study on a population-based series of 582
cases. Cancer Res 2008, 68(20):8541-8546.
5. Hawkins N, Norrie M, Cheong K, Mokany E, Ku SL, Meagher A, O'Connor T,
Ward R: CpG island methylation in sporadic colorectal cancers and its
relationship to microsatellite instability. Gastroenterology 2002,
6. Ogino S, Cantor M, Kawasaki T, Brahmandam M, Kirkner GJ, Weisenberger
DJ, Campan M, Laird PW, Loda M, Fuchs CS: CpG island methylator
phenotype (CIMP) of colorectal cancer is best characterised by
quantitative DNA methylation analysis and prospective cohort studies.
Gut 2006, 55(7):1000-1006.
7. Ogino S, Kawasaki T, Kirkner GJ, Kraft P, Loda M, Fuchs CS: Evaluation of
markers for CpG island methylator phenotype (CIMP) in colorectal
cancer by a large population-based sample. J Mol Diagn 2007,
8.Samowitz WS, Albertsen H, Herrick J, Levin TR, Sweeney C, Murtaugh MA,
Wolff RK, Slattery ML: Evaluation of a large, population-based sample
supports a CpG island methylator phenotype in colon cancer.
Gastroenterology 2005, 129(3):837-845.
9.van Rijnsoever M, Grieu F, Elsaleh H, Joseph D, Iacopetta B:
Characterisation of colorectal cancers showing hypermethylation at
multiple CpG islands. Gut 2002, 51(6):797-802.
10. Weisenberger DJ, Siegmund KD, Campan M, Young J, Long TI, Faasse MA,
Kang GH, Widschwendter M, Weener D, Buchanan D, et al.: CpG island
methylator phenotype underlies sporadic microsatellite instability and
is tightly associated with BRAF mutation in colorectal cancer. Nat
Genet 2006, 38(7):787-793.
11. Jass JR: Classification of colorectal cancer based on correlation of
clinical, morphological and molecular features. Histopathology 2007,
12. Ferracin M, Gafa R, Miotto E, Veronese A, Pultrone C, Sabbioni S, Lanza G,
Negrini M: The methylator phenotype in microsatellite stable colorectal
cancers is characterized by a distinct gene expression profile. J Pathol
13. Nagasaka T, Koi M, Kloor M, Gebert J, Vilkin A, Nishida N, Shin SK,
Sasamoto H, Tanaka N, Matsubara N, et al.: Mutations in both KRAS and
BRAF may contribute to the methylator phenotype in colon cancer.
Gastroenterology 2008, 134(7):1950-1960. 1960 e1951
14. Shen L, Toyota M, Kondo Y, Lin E, Zhang L, Guo Y, Hernandez NS, Chen X,
Ahmed S, Konishi K, et al.: Integrated genetic and epigenetic analysis
identifies three different subclasses of colon cancer. Proc Natl Acad Sci
USA 2007, 104(47):18654-18659.
15. Kawasaki T, Ohnishi M, Nosho K, Suemoto Y, Kirkner GJ, Meyerhardt JA,
Fuchs CS, Ogino S: CpG island methylator phenotype-low (CIMP-low)
colorectal cancer shows not only few methylated CIMP-high-specific
CpG islands, but also low-level methylation at individual loci. Mod
Pathol 2008, 21(3):245-255.
16. Nagasaka T, Sasamoto H, Notohara K, Cullings HM, Takeda M, Kimura K,
Kambara T, MacPhee DG, Young J, Leggett BA, et al.: Colorectal cancer
with mutation in BRAF, KRAS, and wild-type with respect to both
Additional file 1 Unsupervised hierarchical clustering of 1505 probes
(rows) in 28 normal colonic tissues (columns). Methylation of X-chromo-
some genes (enclosed within yellow rectangle) showed 100% correlation
to gender as indicated by the red (female) and blue (male) bar above the
Additional file 2 Unsupervised hierarchical clustering of 1505 probes
(rows) in 91 colorectal tumors (columns). Three tumor subgroups were
revealed when methylation data from all 1,505 loci were analysed.
Additional file 3 202 CpG sites differentially methylated between
normal and tumour tissues. Methylation status at 202 CpG loci differen-
tially methylated between normal and tumour tissues and presence of
repetitive element or single nucleotide polymorphisms within probes.
Additional file 4 Principal component analysis of 202 CpG loci that
were differentially methylated between tumor and normal colonic tis-
sue. This identified principal component 2 as the top ranking dimension
and which explained 20% of the variability in the dataset. CIMP-H tumors
are denoted in green, CIMP-M in black and CIMP-L in red.
Additional file 5 112 genes differentially methylated between CIMP-
H and CIMP-L. Known functions of 112 genes differentially methylated
between CIMP-H and CIMP-L and their reported methylation in cancer and
putative roles in gastrointestinal cancer.
Received: 4 November 2009 Accepted: 21 May 2010
Published: 21 May 2010
This article is available from: http://www.biomedcentral.com/1471-2407/10/227© 2010 Ang et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
BMC Cancer 2010, 10:227
Ang et al. BMC Cancer 2010, 10:227
Page 8 of 8
oncogenes showing different patterns of DNA methylation. J Clin
Oncol 2004, 22(22):4584-4594.
17. Ogino S, Kawasaki T, Kirkner GJ, Loda M, Fuchs CS: CpG island methylator
phenotype-low (CIMP-low) in colorectal cancer: possible associations
with male sex and KRAS mutations. J Mol Diagn 2006, 8(5):582-588.
18. Li WQ, Kawakami K, Ruszkiewicz A, Bennett G, Moore J, Iacopetta B: BRAF
mutations are associated with distinctive clinical, pathological and
molecular features of colorectal cancer independently of microsatellite
instability status. Mol Cancer 2006, 5:2.
19. Wang C, van Rijnsoever M, Grieu F, Bydder S, Elsaleh H, Joseph D, Harvey J,
Iacopetta B: Prognostic significance of microsatellite instability and Ki-
ras mutation type in stage II colorectal cancer. Oncology 2003,
20. Iacopetta B, Grieu F: Routine detection of the replication error
phenotype in clinical tumor specimens using fluorescence-SSCP.
Biotechniques 2000, 28(3):566-568. 570
21. Ang PW, Toh HB, Iacopetta B, Soong R: An improved quality control for
bisulfite-PCR-based DNA methylation analysis: cycle threshold value.
Clin Chem Lab Med 2008, 46(8):1117-1121.
22. Bibikova M, Fan JB: GoldenGate assay for DNA methylation profiling.
Methods Mol Biol 2009, 507:149-163.
23. Bibikova M, Lin Z, Zhou L, Chudin E, Garcia EW, Wu B, Doucet D, Thomas
NJ, Wang Y, Vollmer E, et al.: High-throughput DNA methylation profiling
using universal bead arrays. Genome Res 2006, 16(3):383-393.
24. Martin-Subero JI, Ammerpohl O, Bibikova M, Wickham-Garcia E, Agirre X,
Alvarez S, Bruggemann M, Bug S, Calasanz MJ, Deckert M, et al.: A
comprehensive microarray-based DNA methylation study of 367
hematological neoplasms. PLoS One 2009, 4(9):e6986.
25. Byun HM, Siegmund KD, Pan F, Weisenberger DJ, Kanel G, Laird PW, Yang
AS: Epigenetic profiling of somatic tissues from human autopsy
specimens identifies tissue- and individual-specific DNA methylation
patterns. Hum Mol Genet 2009, 18(24):4808-4817.
26. Christensen BC, Houseman EA, Marsit CJ, Zheng S, Wrensch MR, Wiemels
JL, Nelson HH, Karagas MR, Padbury JF, Bueno R, et al.: Aging and
environmental exposures alter tissue-specific DNA methylation
dependent upon CpG island context. PLoS Genet 2009, 5(8):e1000602.
27. O'Riain C, O'Shea DM, Yang Y, Le Dieu R, Gribben JG, Summers K, Yeboah-
Afari J, Bhaw-Rosun L, Fleischmann C, Mein CA, et al.: Array-based DNA
methylation profiling in follicular lymphoma. Leukemia 2009,
28. Hubert LJ, Levin JR: A general statistical framework for assessing
categorical clustering in free recall. Psychological Bulletin 1976,
29. Lee TI, Jenner RG, Boyer LA, Guenther MG, Levine SS, Kumar RM, Chevalier
B, Johnstone SE, Cole MF, Isono K, et al.: Control of developmental
regulators by Polycomb in human embryonic stem cells. Cell 2006,
30. Widschwendter M, Fiegl H, Egle D, Mueller-Holzner E, Spizzo G, Marth C,
Weisenberger DJ, Campan M, Young J, Jacobs I, et al.: Epigenetic stem cell
signature in cancer. Nat Genet 2007, 39(2):157-158.
31. Marsit CJ, Christensen BC, Houseman EA, Karagas MR, Wrensch MR, Yeh
RF, Nelson HH, Wiemels JL, Zheng S, Posner MR, et al.: Epigenetic profiling
reveals etiologically distinct patterns of DNA methylation in head and
neck squamous cell carcinoma. Carcinogenesis 2009, 30(3):416-422.
32. McRonald FE, Morris MR, Gentle D, Winchester L, Baban D, Ragoussis J,
Clarke NW, Brown MD, Kishida T, Yao M, et al.: CpG methylation profiling
in VHL related and VHL unrelated renal cell carcinoma. Mol Cancer
33. Martinez R, Martin-Subero JI, Rohde V, Kirsch M, Alaminos M, Fernandez
AF, Ropero S, Schackert G, Esteller M: A microarray-based DNA
methylation study of glioblastoma multiforme. Epigenetics 2009,
34. Ladd-Acosta C, Pevsner J, Sabunciyan S, Yolken RH, Webster MJ, Dinkins T,
Callinan PA, Fan JB, Potash JB, Feinberg AP: DNA methylation signatures
within the human brain. Am J Hum Genet 2007, 81(6):1304-1315.
35. Toyota M, Ohe-Toyota M, Ahuja N, Issa JP: Distinct genetic profiles in
colorectal tumors with or without the CpG island methylator
phenotype. Proc Natl Acad Sci USA 2000, 97(2):710-715.
36. Nosho K, Irahara N, Shima K, Kure S, Kirkner GJ, Schernhammer ES, Hazra A,
Hunter DJ, Quackenbush J, Spiegelman D, et al.: Comprehensive
biostatistical analysis of CpG island methylator phenotype in
colorectal cancer using a large population-based sample. PLoS One
37. Drago A, Alboni S, Nicoletta B, De Ronchi D, Serretti A: HTR1B as a risk
profile maker in psychiatric disorders: a review through motivation and
memory. Eur J Clin Pharmacol 2010, 66(1):5-27.
38. Takai D, Yagi Y, Wakazono K, Ohishi N, Morita Y, Sugimura T, Ushijima T:
Silencing of HTR1B and reduced expression of EDN1 in human lung
cancers, revealed by methylation-sensitive representational difference
analysis. Oncogene 2001, 20(51):7505-7513.
39. Slattery ML, Curtin K, Wolff RK, Boucher KM, Sweeney C, Edwards S, Caan
BJ, Samowitz W: A comparison of colon and rectal somatic DNA
alterations. Dis Colon Rectum 2009, 52(7):1304-1311.
40. Wood LD, Calhoun ES, Silliman N, Ptak J, Szabo S, Powell SM, Riggins GJ,
Wang TL, Yan H, Gazdar A, et al.: Somatic mutations of GUCY2F, EPHA3,
and NTRK3 in human cancers. Hum Mutat 2006, 27(10):1060-1061.
41. Osborn NK, Zou H, Molina JR, Lesche R, Lewin J, Lofton-Day C, Klatt KK,
Harrington JJ, Burgart LJ, Ahlquist DA: Aberrant methylation of the eyes
absent 4 gene in ulcerative colitis-associated dysplasia. Clin
Gastroenterol Hepatol 2006, 4(2):212-218.
42. Schatz P, Distler J, Berlin K, Schuster M: Novel method for high
throughput DNA methylation marker evaluation using PNA-probe
library hybridization and MALDI-TOF detection. Nucleic Acids Res 2006,
43. Miyamoto K, Asada K, Fukutomi T, Okochi E, Yagi Y, Hasegawa T, Asahara T,
Sugimura T, Ushijima T: Methylation-associated silencing of heparan
sulfate D-glucosaminyl 3-O-sulfotransferase-2 (3-OST-2) in human
breast, colon, lung and pancreatic cancers. Oncogene 2003,
44. Shworak NW, Liu J, Petros LM, Zhang L, Kobayashi M, Copeland NG,
Jenkins NA, Rosenberg RD: Multiple isoforms of heparan sulfate D-
glucosaminyl 3-O-sulfotransferase. Isolation, characterization, and
expression of human cdnas and identification of distinct genomic loci.
J Biol Chem 1999, 274(8):5170-5184.
45. Perrimon N, Bernfield M: Specificities of heparan sulphate
proteoglycans in developmental processes. Nature 2000,
The pre-publication history for this paper can be accessed here:
Cite this article as: Ang et al., Comprehensive profiling of DNA methylation
in colorectal cancer reveals subgroups with distinct clinicopathological and
molecular features BMC Cancer 2010, 10:227