Peroxisome proliferator-activated receptor (PPAR) g gene polymorphisms
and colorectal cancer risk among Chinese in Singapore
Sue A.Ingles2and Mimi C.Yu1
Department of Community, Occupational and Family Medicine, Yong Loo
Lin School of Medicine, National University of Singapore, Singapore,1The
Cancer Center, University of Minnesota, MN, USA and2USC/Norris
Comprehensive Cancer Center, Keck School of Medicine, University of
Southern California, Los Angeles, CA, USA
?To whom correspondence should be addressedat: Departmentof Community,
Occupational and Family Medicine, Yong Loo Lin School of Medicine,
National University of Singapore, MD3, 16 Medical Drive, Singapore 117597.
Tel: +65 6874 4975; Fax: +65 6779 1489;
Peroxisome proliferator-activated receptor (PPAR) g is a
ligand-activated nuclear receptor that plays a key role in
adipogenesis and adipocyte gene expression, and has
recently been linked with possible antineoplastic effects
in colonic carcinogenesis. PPARg2 and g3 are two tran-
scripts arising from the PPARg gene through differential
promoter usage and alternative splicing. We investigated
the associations between PPARg2 Pro12Ala and PPARg3
C-681G gene polymorphisms and colorectal cancer (CRC)
risk in a case–control study nested within the Singapore
Chinese Health Study. Genotypes for the PPARg2 and
PPARg3 polymorphisms were determined on 362 incident
CRC cases and 1164 cohort controls by direct sequencing
and by fluorogenic 50-nuclease assay. Unconditional logistic
regression models were used for statistical analyses. With
adjustment for CRC risk factors, subjects with one or two
copies of the G allele of the PPARg2 Pro12Ala polymorph-
ism showed a statistically significant reduction in risk com-
pared to those with the CC genotype [odds ratio (OR) ¼
0.53, 95% confidence interval (CI) ¼ 0.30–0.92]. For the
PPARg3 C-681G polymorphism, subjects with one or two
copies of the C allele showed a reduction in risk compared
to those with the GG genotype (OR ¼ 0.72, 95% CI ¼
0.51–1.04). When PPARg2 and PPARg3 genotypes were
considered simultaneously, thenumber of putative low-risk
genotypes was significantly associated with reduced risk of
CRC in a gene-dose-dependent manner; the OR (95% CI)
was 0.72 (0.49–1.07) among subjects possessing one
low-risk genotype (either PPARg2 or PPARg3), and the
comparable figure among subjects possessing both low-risk
genotypes was 0.19 (0.07–0.51).
Colorectal cancer (CRC) has one of the highest incidence rates
among cancers in developed countries. The geographical pat-
tern of high rates in the West and lower rates in Asia suggests
that in addition to genetic determinants, lifestyle and dietary
factors may be important contributors to colonic carcinogen-
esis (1,2). Among them, obesity, which is more prevalent
among Western populations, has been consistently associated
with higher risk of CRC among men and women in both case–
control and cohort studies (reviewed in ref. 3). Epidemiolo-
gical data also suggest that both Type 2 diabetes and impaired
glucose tolerance are risk factors for colon cancer in
Western populations. Together with a growing body of experi-
mental evidence, these epidemiological data suggest that
obesity-induced hyperinsulinemia, hyperlipidemia and insulin
resistance may play a role in colon carcinogenesis (reviewed
in ref. 4).
As a nuclear receptor which plays a pivotal role in regu-
lating adipocyte differentiation, glucose and lipid homeo-
stasis, and intracellular insulin-signaling events, peroxisome
proliferators-activated receptor (PPAR) g has received grow-
ing interest for its possible role in CRC. PPARg forms func-
tional heterodimers with members of the retinoid X-receptor
family of nuclear receptors and activates the transcription of
target genes by the release of corepressors and recruitment of
coactivators (5). Putative endogenous ligands for PPARg
include both polyunsaturated fatty acids (PUFAs) and arachi-
donic acid derivatives (6). Experimental evidence has sugges-
ted that activation of PPARg in the colon results in growth
inhibition and differentiation, and reduces the malignant
potential of CRC cells (7). In addition, administration of the
PPARg ligand troglitazone significantly inhibits chemically
induced colitis and formation of aberrant crypt foci in rats (8).
The PPARg gene produces four different PPARg mRNAs
by differential promoter usage and alternative splicing, giving
rise to two different protein isoforms. The PPARg1, PPARg3
and PPARg4 transcripts, although possessing different
upstream regulatory sequences, give rise to identical proteins
encoded by exons 1–6. A functional C-to-G polymorphism in
the promoter region for the PPARg3 transcript at position
?681 from the beginning of exon A2 has been associated
with increased body weight and circulating levels of choles-
terol (9,10). The PPARg2 transcript gives rise to a protein with
an additional 28 amino acids encoded by the PPARg2-specific
exon B (11,12). A polymorphism (proline-to-alanine substitu-
tion at codon 12) in exon B has been associated with reduced
risk of diabetes mellitus (13–15) and CRC (16). We investig-
ated the associations of these two polymorphisms in the
PPARg2 coding region and the PPARg3 regulatory region
with CRC risk in a nested case–control study within the Sin-
gapore Chinese Health Study, a prospective investigation of
diet and cancer in 63000 Chinese men and women.
Materials and methods
The study design and subject recruitment of the Singapore Chinese Health
Abbreviations: BMI, body mass index; CRC, colorectal cancer; EM
algorithm, expectation–maximization algorithm; LD, linkage disequilibrium;
PPAR, peroxisome proliferators-activated receptor.
Carcinogenesis vol.27 no.9 pp.1797–1802, 2006
Advance Access publication March 2, 2006
#The Author 2006. Published by Oxford University Press. All rights reserved. For Permissions, please email: firstname.lastname@example.org
by guest on June 1, 2013
men aged 45–74 years belonging to the Hokkien or Cantonese dialect group
were enrolled in the study between April 1993 and December 1998. At recruit-
ment,aface-to-face interviewwasconductedinthesubject’shomebya trained
interviewer, using a structured questionnaire, which requested information on
demographics, lifetime use of tobacco and alcohol, medical history, family
Boards at the University of Southern California and the National University of
Singapore had approved this study.
A smoker was defined as a subject who ever smoked at least one cigarette
per day for 1 year or longer. Ever smokers were further asked for current
smoking status at recruitment, age at which regular smoking was started, the
average number of cigarettes smoked per day and the number of years of
smoking. An alcohol drinker was defined as a subject who drank any alcoholic
beverages on a monthlybasis or more often. Each drinker was further asked for
the frequency and amount of each of the four types of alcoholic beverages
(beer, rice wine, grape wine and hard liquor) consumed. One drink was defined
as 375 ml of beer (13.6 g of ethanol), 30 ml of rice wine (10.9 g of ethanol),
118 ml of grape wine (11.7 g of ethanol) and 30 ml of hard liquor (10.9 g of
ethanol). Total number of drinks per day was computed based on the amount
and type of alcoholic beverages consumed.
BetweenApril1994and July1999, we attemptedto collect blood (orbuccal
cells if subject refused blood donation) and single-void urine specimens from a
random 3% sample of study enrollees. Details of the biospecimen collection,
processing and storage procedures have been described previously (18). We
collected blood/buccal cell samples from 1194 (63% of total eligible) subjects
during this period. After we excluded 18 subjects who had a history of CRC at
recruitment (n ¼ 5) or developed first CRC (n ¼ 13) by May 31, 2003, the
remaining 1176 subjects constituted the referent (i.e. control) group for the
present study. Compared with the rest of the cohort members, these subjects
had comparable distributions by age, sex, dialect group, level of education,
body mass index, alcohol drinking, history of diabetes, or familial history of
CRC or any cancer (all P values > 0.10). There was a slightly lower percentage
of ever smokers in the control group (27.8%) than that in the rest of the cohort
participants (31.6%, P ¼ 0.04).
We identified incident CRC cases through the population-based cancer
registry in Singapore (19). As of May 31, 2003, CRC occurred among 696
cohort participants. All cases were histologically confirmed except for 15 that
were ascertained by clinical evidence (2%) and six by death records (0.8%).
We also attempted to collect blood/buccal cell and urine samples from all
incident CRC cases. Blood (n ¼ 289) or buccal (n ¼ 85) specimens were
available on 374 (53.7%) incident CRC cases. Compared with CRC patients
who did not donate a blood or buccal sample, those who donated had a similar
mean of age at cancer diagnosis (66.2 versus 67.5 years). Male patients were
more likely to donate a biospecimen than female patients (57.1 versus 49.7%).
So did the patients with Cantonese dialect compared with those with Hokkien
dialect (59.0 versus 49.9%). Patients who did not donate a blood or buccal
(46.4%). There was no difference in the percentage of biopecimen availability
by level of body mass index, cigarette smoking, alcohol drinking or history of
DNA waspurifiedfrom buffycoatsof peripheral bloodand buccalcell samples
using a QIAamp 96 DNA Blood Kit (Qiagen, Valencia, CA). Alleles for the
PPARg2 Pro12Ala polymorphism (rs1801282) were identified using direct
sequencing of the polymorphic region. The region of the gene containing
the polymorphism was amplified by polymerase chain reaction (PCR) using
GC083rev (50-GCAGACAGTGTATCAGTGAAGG-30). PCR reaction mix
was prepared using HotStart Taq Polymerase (Qiagen, Valencia, CA) accord-
ing to manufacturer’s instructions using 20 ng of genomic DNA, 2 mM MgCl2
and 300 mM of each primer. PCR amplification was performed in a thermal
cycler (MWG Biotech, High Point, NC) using a touchdown protocol with an
initial step of 95?C for 15 min finishing with 35 cycles of 95?C per 25 s, 54?C
per min and 72?C per min. The PCR reactions are purified using a MAF-NOB
PCR purification plate (Millipore, Billerica, MA) to remove dNTPs and pri-
mers.Afractionofthe sampleswereanalyzedbyagarosegelelectrophoresis to
confirm the success of the PCR reactions and the absence of a product in
the negative controls. DNA sequencing was performed using primer
GC083S (50-ACTGATGTCTTGACTCATGGGTG-30) using ?10–20 ng of
purified PCR product using fluorescently labeled ddNTPs (ABI Dye Termin-
ator Sequencing Kit, Applied Biosystems) by cycle sequencing for 50 rounds
of 95?C per 15 s and 50?C for each 3.5 min. The sequencing reactions were
run on an ABI3730xl Capillary DNA Analyzer. The sequence files were run
through Phred/Phrap (University of Washington) to align the sequences and
mark possible polymorphic positions (20). The sequences were viewed in
Consed (University of Washington) and the polymorphic position scored (21).
The PPARg3 C-681G polymorphism (rs10865710) was genotyped using
the fluorogenic 50-nuclease assay (TaqMan Assay) (22). The TaqMan assays
were performed using a TaqMan PCR Core Reagent kit (Applied Biosystems,
FosterCity,CA) accordingto manufacturer’s instructions. The oligonucleotide
primers for amplification of the PPARg3 polymorphic region were GC028for
(50-CCTGATGATAAGGCTTTTGGCATT-30) and GC028rev (50-TATCT-
CTTATGAAAGGCTCAAGGATCCT-30). In addition, the fluorogenic oligo-
nucleotide probes (TaqMan MGB Probes; ABI) used to detect each of the
alleles were GC028F (50-TTTTCCATCAAGACAAAA-30) labeled with
6-FAM to detect the C allele and GC028V (50- TTTTCCATGAAGA-
CAAAA-30) labeled with VIC to detect the G allele. PCR amplification
using ?10 ng of genomic DNA was performed in a thermal cycler (MWG
Biotech, High Point, NC) with an initial step of 95?C for 10 min followed by
50 cycles of 95?C for 25 s and 60?C for 1 min. The fluorescence profile of each
well was measured in an ABI 7900HT Sequence Detection System and the
results analyzed with Sequence Detection Software (Applied Biosystems).
Experimental samples were compared with 12 controls to identify the three
genotypes at each locus (G/G, G/C, C/C). Samples that were outside the
parameters defined by the controls were identified as non-informative and
Twelve cases and twelve controls were non-informative in either the PPARg2
or PPARg3 genotypes. These subjects were excluded from the study. Thus, the
present analyses included 362 cases and 1164 controls. The cases patients and
control subjects were not genetically related.
The genotypes of control subjects were checked for Hardy–Weinberg equi-
librium using the exact test. The differences in genotype distribution between
cases and controls by demographic and selected lifestyle characteristics were
the two gene polymorphisms in control subjects using the Expectation–
Maximization (EM) algorithm (23) as implemented in Haploview version
3.2 (24). The measure of the statistical association for LD, R2, a better measure
for LD than the commonly used Lewontin’s D’ in the case of low allele
frequency of the gene, was used to assess the correlation of alleles at two
We used standard methods for unmatched case–control studies to examine
the effects of the PPARg gene polymorphisms on CRC risk (26). The strength
of a given gene–cancer association was measured by the odds ratios (ORs) and
their 95% confidence intervals (CIs) and P-values. Among control subjects, the
frequencies of the PPARg2 G allele were comparable between Cantonese and
Hokkiens (4.0 versus 3.7%, P ¼ 0.53). So were those of the PPARg3 G allele
(37.8 versus 35.7%, P ¼ 0.45). Therefore, we combined both dialect groups in
data analysis. We used a polytomous logistic regression model to estimate ORs
for age (year) at recruitment, year of recruitment, gender, dialect group (Can-
tonese, Hokkien), level of education (no formal schooling, primary school, and
secondary school or higher), body mass index (BMI) (<20, 20–<24, 24–<28,
and 28+ kg/m2), cigarette smoking (non-smokers, light and heavy smokers),
alcohol consumption (non-drinkers, <1 and 1+ drink/day), history of diabetes
mellitus (yes/no) and familial history of CRC (no/yes). Based on the entire
cohort, we classified subjects who started to smoke before 15 years of age and
smoked at least 13 cigarettes/day as ‘heavy’ smokers, whereas the remaining
smokers as ‘light’ smokers.
Statistical analysis was carried out using the SAS software Version 9.1
(SAS Institute, Cary NC), unless otherwise indicated. All P-values quoted
are two-sided. The two-sided P-values < 0.05 were considered statistically
Of 362 cases, 206 (57%) had cancers of the colon, and the
remaining156(43%) had either rectal or rectosigmoid cancers.
The mean age of cases at the time of diagnosis was 66.1 (SD
7.9) years, with a range of 47–81 years. The median time
interval between the baseline interview and cancer diagnosis
was 4.7 years (range, 1 month–9.9 years). Table I shows the
distributions by selected demographic characteristics and
potential risk factors for CRC in the study population. Relative
to control subjects, case patients were older at recruitment
[mean age at recruitment for cases was 61.0 (SD ¼ 7.6)
years and for controls was 56.4 (SD ¼ 8.1) years], had a greater
proportion of men, had lower level of education, had higher
W.-P.Koh et al.
by guest on June 1, 2013
BMI, were more likely to be cigarette smokers and daily alco-
hol drinkers, more likely to have a history of diabetes mellitus
and more likely to have a first-degree relative with CRC.
Among the control subjects, the frequencies of the C and G
alleles of the PPARg2 polymorphism were 0.961 and 0.039,
respectively, whereas the frequencies of the C and G alleles of
the PPARg3 polymorphism were 0.634 and 0.366, respect-
ively. All genotypic distributions were in Hardy–Weinberg
equilibrium (P-values > 0.7). The R2for the correlation of
alleles at two sites (i.e., for LD) was 0.07 between the
PPARg2 Pro12Ala and PPARg3 C-681G polymorphisms,
indicating a lack of LD. Table II shows PPARg2 and
PPARg3 genotypes in relation to CRC risk. For the
PPARg2 polymorphism, the CG and GG genotypes were
grouped owing to the low frequency of the GG genotype.
Subjects possessing at least one copy of the G allele had an
?50% reduction in risk of CRC (OR ¼ 0.53, 95% CI ¼
0.30–0.92). For the PPARg3 polymorphism, subjects with at
least onecopy of the Callele showeda >25% reductionin CRC
risk compared with those homozygous for the G allele. Having
two copies of the C allele was not associated with any addi-
tional reduction in risk. Hence the PPARg2 CG/GG genotypes
and the PPARg3 GC/CC genotypes were considered as
putative low-risk genotypes for CRC. When both PPARg2
and PPARg3 gene polymorphisms were examined simultan-
eously, the ORs (95% CIs) for CRC were 0.72 (0.49–1.07) for
subjects possessing only one low-risk genotype and 0.19
(0.07–0.51) for those possessing two low-risk genotypes
when compared to subjects without any low-risk genotype
(P for trend ¼ 0.002). There is no material difference in
any of the gene–risk associations across subsites (i.e. colon
versus rectal cancer).
We also examined the association between the PPARg2 and
PPARg3 genotypes and risk of CRC among non-diabetic sub-
jects. After exclusion of subjects who reported a history of
CRC for subjects possessing only one and both low-risk geno-
types were 0.66 (0.43–1.01) and 0.16 (0.05–0.41), respect-
ively, compared with those without any low-risk genotype
of the two PPARg polymorphisms studied (P for trend ¼
0.0001). We also performed subgroup analyses stratified by
age (<60 and 60+ years), gender, dialect group (Cantonese,
Hokkien), body mass index (<24 and 24+ kg/m2), cigarette
smoking (no/yes), alcohol drinking (no/yes) and familial his-
tory of CRC (no/yes). The inverse associations between the
combined genotypes of PPARg2 and PPARg3 and risk of CRC
risk were comparable between subgroups of any stratifying
variable (data not shown).
In this cohort of Singapore Chinese, we reported a significant
effect of the PPARg2 Pro12Ala and PPARg3 C-681G gene
polymorphisms on CRC risk. Although the risk reduction asso-
ciated with the PPARg2 variant is similar to that reported for
CRC in a Spanish population (16) and colorectal adenoma in a
US population (27), this is the first study relating a PPARg3
gene polymorphism with CRC risk.
The PPARg2 gene polymorphism in this study is a C-to-G
missense mutation, which results in a proline (Pro) to alanine
(Ala) substitution at codon 12 in the PPARg2-specific exon B.
This substitution is within a domain of the protein that
enhances ligand-independent activation (28,29). The reported
allele frequency of this variant varies in different populations,
from 0.12 among Caucasian Americans, 0.10 among Mexican
Americans and 0.03 among African Americans to 0.01 among
Chinese from mainland China (29). Our Ala allelic frequency
of 0.039 among the controls in our Chinese population is
consistent with the frequency reported previously for Singa-
pore Chinese (30). Our finding of a lack of strong LD between
the two polymorphic sites is consistent with the HapMap data
(31), which show an R2value of 0.08 for Han Chinese and 0.31
for Caucasians. In contrast to R2, Lewontin’s D0values were
high in our dataset (1.0, 95% CI ¼ 0.92–1.0), and in the
HapMap dataset (1.0, 95% CI ¼ 0.11–0.99 for Han Chinese;
1.0, 95% CI ¼ 0.69–1.0 for European whites), indicating high
LD. However, in the case of low allele frequency, such as the
case of PPARg2 (0.039 in our study population; 0.075 in Hap-
Map whites), D0may not be valid since it can be greatly
inflated (25). Therefore, in our study, we think it is more
appropriate to use R2instead of D0, and the low value of R2
in our study indicated a lack of LD.
A case–control study among Spanish participants found that
subjects with the G allele had an ?45% reduction in CRC risk
(16), which is similar to the magnitude of risk reduction noted
Table I. Distributions by selected demographic characteristics and potential
risk factors for CRC in the study population at baseline, the Singapore
Chinese Health Study
(n ¼ 362)
(n ¼ 1164)
Level of education
No formal schooling
Secondary school or higher
History of diabetes
First-degree relative had CRC
?Light smokers were those who either started smoking at 15+ years of
age or smoked <13 cigarettes/day whereas heavy smokers were
those who started smoking before 15 years of age and smoked
PPARg gene polymorphism and CRC risk
by guest on June 1, 2013
in our study. These empirical data suggest that the putative
low-risk G allele may be associated with biologically higher
PPARg2 activity. On the other hand, Jiang et al. (32) examined
the PPARg2 Pro12Ala polymorphism among CRC cases and
controls in an Indian population in Chennai, India, which had a
higher G allelic frequency of 0.11 compared with our study
population, and observed no significant association between
this gene polymorphism and CRC risk (OR ¼ 1.06; 95% CI ¼
0.70–1.61). Since putative ligands for PPARg include both
dietary polyunsaturated fatty acids (PUFAs) and arachidonic
acid derivatives (6), it is plausible that dietary factors can exert
an influence on the effect of PPARg on colorectal carcinogen-
esis. In support of this hypothesis, Jiang et al. (32) did find a
reduced risk of CRC associated with the G allele among sub-
jects with higher fish intake (OR ¼ 0.51), although this did not
reach statistical significance. Thus, differences in the preval-
ence of dietary cofactors may explain, at least in part, the
disparate findings among different populations.
C-to-G substitution ?681 bp upstream from the PPARg3 tran-
scription startsiteandislocatedinaputative DNA-bindingsite
for transcription factors of the signal transducer and activator
of transcription (STAT) family (9). The G allele frequency has
been describedas0.25intheFrenchpopulation,which islower
than that found among our Chinese population. The G variant
completely abolishes the binding of STAT5B to the cognate
promoterelementanddecreases the transactivation of PPARg3
promoter by the growth hormone/STAT5B pathway and has
been associated with increased plasma apolipoprotein B and
LDL-cholesterol levels (10). Alternatively, the C allele is asso-
ciated with higher PPARg3 receptor activity in vitro (9).
Adipose PPARg has been identified as an important medi-
ator for the maintenance of insulin sensitivity in the body.
of PPARg in adipose tissues improves obesity-associated insu-
lin resistance by regulating expression of adipocyte-secreted
hormones that regulate glucose homeostasis (reviewed in ref.
33). Obesity-induced insulin resistance results in hyperinsu-
linemia that can in turn lead to a decrease in synthesis of
insulin-like growth factor binding protein and an increase in
levels of bioavailable insulin-like growth factor-1 (IGF-1).
Both insulin and IGF-1 can signal through their receptors to
promote cellular proliferation and angiogenesis, or inhibit
apoptosis in CRC cells (34–36). Such experimental evidence
is consistent with epidemiological data associating clinical
conditions of high levels of insulin and IGF-1 with increased
risk of colon cancer (reviewed in ref. 37). Hence better insulin
sensitivity and lower insulin levels mediated by adipose
PPARg can potentially reduce CRC risk. Consistent with
this hypothesis, the Pro12Ala polymorphism of the PPARg2
gene that is associated with lower CRC risk in our study has
indeed been associated with lower fasting insulin concentra-
tions, improved insulin sensitivity and reduced risk of Type II
diabetes mellitus (13–15).
Although both the PPARg2 and PPARg3 isoforms are
expressed in adipose tissues, the PPARg protein expressed
in normal and neoplastic colonic epithelial cells is predomin-
antly, if not solely, the g3 isoform. The distribution of PPARg3
mRNA in human colon cancer cell lines is 100-fold more
abundant than the g2 isoform under basal condition and
increased by >600-fold over g2 with induced differentiation
(38). Exposure of cultured human CRC cell lines to PPARg
agonists induces growth inhibition that is associated with a G1
cell cycle arrest (7), an increase in several markers of differ-
entiation (39) and in apoptosis (40). In addition, PPARg lig-
ands have been shown to dramatically attenuate cytokine gene
expression in colon cancer cell lines by inhibiting the activa-
tion of nuclear factor-kappa B, suggesting that PPARg3 may
also be important in modulating the intestinal inflammatory
response (41). Examination of human CRC cells has revealed
loss-of-function mutations in the PPARg gene in 7% of tumors
from 55 unrelated individuals (42). Collectively, these experi-
mental studies have provided evidence for a putative role of
PPARg3 as a tumor suppressor in CRC. The association of the
putative high-activity C allele with a lower risk of CRC in our
study supports the hypothesis that a higher PPARg3 activity in
the colon also leads to enhanced antineoplastic and prodiffer-
entiation effects in colonic epithelial cells.
Table II. PPARg2 and PPARg3 genotype in relation to risk of CRC, the Singapore Chinese Health Study
Controls CRCColon cancer Rectal cancer
CasesOR (95% CI)a
CasesOR (95% CI)a
Cases OR (95% CI)a
Sum of putative low-risk genotypesb
P for trend
aORs were adjusted for age (year) at recruitment, year of recruitment, gender, dialect group (Cantonese and Hokkien), level of education (no formal
schooling, primary school and secondary school and higher), BMI (<20, 20–<24, 24–<28 and 28+ kg/m2), cigarette smoking (non-smokers, light and
heavy smokers), alcohol consumption (none, <1 and 1+ drink/day), history of diabetes (no/yes) and familial history of CRC (no/yes).
bThe putative low-risk genotype was defined as the presence of at least one G allele of the PPARg2 genotype or the presence of at least one C allele of
the PPARg3 genotype.
W.-P.Koh et al.
by guest on June 1, 2013
A chief limitation of our study is a lack of information on the
use of PPARg agonists such as thiazolidinediones, a class of
recently available drugs for the treatment of diabetes mellitus.
These drugs may potentially intensify the effects of PPARg.
We have shown that our results on PPARg2 and risk were
independent of the subjects’ self-reported history of diabetes
mellitus, and remained unchanged after exclusion of subjects
with a history of diabetes, thus suggesting that the use of these
drugs was not a significant confounder. Another limitation of
the present study was lower statistical power due to small sam-
ple size, especially when we examined the genotype–CRC risk
association in stratified analyses. The present study had a 73%
statistical power to detect a 50% risk reduction in CRC among
subjects with the PPARg2 CG and GG genotypes, and a 59%
statistical power to detect a 30% risk reduction associated with
genotypes were combined, the present study had sufficient
power (>80%) to detect a statistically significant trend in the
risk of developing colon (minimal detectable OR ¼ 0.36) or
rectal cancer (minimal detectable OR ¼ 0.32) alone or CRC
combined (minimal detectable OR ¼ 0.43).
The current study has several strengths. Singapore is a small
city-state where there is good access to specialized medical
care. The nationwide cancer registry has been in place since
1968 and has been shown to be comprehensive in its recording
of cancer cases (43). Thus, CRC case ascertainment can be
assumed to be complete. Our study subjects originated from
two contiguous regions in South China, leading to a high
degree of genetic homogeneity. All dietary factors, measure-
ment of BMI and history of diabetes mellitus were assessed
prior to cancer diagnosis and, thus, can be presumed to be free
of recall bias (17).
In summary, the present study implicates a role for PPARg
in CRC. These findings may have clinical implications. One
may target the PPARg activation pathway in drug development
efforts as part of prevention and treatment strategies for CRC.
We thank Ms Siew-Hong Low of the National University of Singapore for
supervising the field work of the Singapore Chinese Health Study and Ms
Kazuko Arakawa of the University of Southern California for the development
and management of the cohort study database. We also thank the Singapore
Cancer Registry for assistance with the identification of cancer outcomes.
This work was supported by grants R01 CA55069, R35 CA53890, R01
CA80205 and R01 CA98497 from the United States National Cancer Institute,
Conflict of Interest Statement: None declared.
1.Armstrong,B. and Doll,R. (1975) Environmental factors and cancer
incidence and mortality in different countries, with special reference to
dietary practices. Int. J. Cancer, 15, 617–631.
2.Doll,R. and Peto,R. (1981) The causes of cancer: quantitative estimates of
3.IARC Working Group on the Evaluation of Preventive Agents. (2002),
IARC Handbook on Cancer Prevention. Vol 6: Physical Activity and
Weight Control. IARC Scientific Publication, Lyon.
4.Calle,E.E. and Kaaks,R. (2004) Overweight, obesity and cancer:
epidemiological evidence and proposed mechanisms. Nat. Rev. Cancer,
5.Glass,C.K. and Rosenfeld,M.G. (2000) The coregulator exchange in
transcriptional functions of nuclear receptors. Genes Dev., 14, 121–141.
6.Kliewer,S.A., Sundseth,S.S., Jones,S.A. et al. (1997) Fatty acids and
eicosanoids regulate gene expression through direct interactions with
peroxisome proliferator-activated receptors alpha and gamma. Proc.
Natl Acad. Sci. USA, 94, 4318–4323.
7.Brockman,J.A., Gupta,R.A. and Dubois,R.N. (1998) Activation of
PPARgamma leads to inhibition of anchorage-independent growth of
human colorectal cancer cells. Gastroenterology, 115, 1049–1055.
proliferator-activated receptors alpha and gamma inhibit chemically
induced colitis and formation of aberrant crypt foci in rats. Cancer Res.,
9.Meirhaeghe,A., Fajas,L., Gouilleux,F., Cottel,D., Helbecque,N., Auwerx,J.
and Amouyel,P. (2003) A functional polymorphism in a STAT5B site of
the human PPAR gamma 3 gene promoter affects height and lipid
metabolism in a French population. Arterioscler. Thromb. Vasc. Biol.,
10.Meirhaeghe,A., Tanck,M.W., Fajas,L., Janot,C., Helbecque,N., Cottel,D.,
Auwerx,J., Amouyel,P. and Dallongeville,J. (2005) Study of a new
PPARgamma2 promoter polymorphism and haplotype analysis in a
French population. Mol. Genet. Metab., 85, 140–148.
11.Fajas,L., Auboeuf,D., Raspe,E. et al. (1997) The organization, promoter
analysis, and expression of the human PPARgamma gene. J. Biol. Chem.,
12.Fajas,L., Fruchart,J.C. and Auwerx,J. (1998) PPARgamma3 mRNA: a
distinct PPARgamma mRNA subtype transcribed from an independent
promoter. FEBS Lett., 438, 55–60.
13.Deeb,S.S., Fajas,L., Nemoto,M., Pihlajamaki,J., Mykkanen,L., Kuusisto,J.,
Laakso,M., Fujimoto,W. and Auwerx,J. (1998) A Pro12Ala substitution in
PPARgamma2 associated with decreased receptor activity, lower body
mass index and improved insulin sensitivity. Nat. Genet., 20, 284–287.
14.Altshuler,D., Hirschhorn,J.N., Klannemark,M. et al. (2000) The common
PPARgamma Pro12Ala polymorphism is associated with decreased risk of
type 2 diabetes. Nat. Genet., 26, 76–80.
Altshuler,D.M., Hunter,D.J. and Manson,J.E. (2003) Prospective study
of theassociation betweenthe
polymorphism in the PPARgamma gene and type 2 diabetes. Diabetes
Care, 26, 2915–2917.
16.Landi,S., Moreno,V., Gioia-Patricola,L., Guino,E., Navarro,M., de Oca,J.,
Capella,G. and Canzian,F. (2003) Association of common polymorphisms
in inflammatory genes interleukin (IL)6, IL8, tumor necrosis factor alpha,
NFKB1, and peroxisome proliferator-activated receptor gamma with
colorectal cancer. Cancer Res., 63, 3560–3566.
17.Hankin,J.H., Stram,D.O., Arakawa,K., Park,S., Low,S.H., Lee,H.P. and
Yu,M.C.(2001) Singapore Chinese
validation,and calibration of
questionnaire. Nutr. Cancer, 39, 187–195.
18.Koh,W.P., Yuan,J.M., Sun,C.L., van den Berg,D., Seow,A., Lee,H.P. and
Yu,M.C. (2003)Angiotensin I-converting
polymorphism and breast cancer risk among Chinese women in
Singapore. Cancer Res., 63, 573–578.
19.Seow,A., Koh,W.P., Chia,K.S.,
Shanmugaratnam,K.S. (2004) Trends in cancer incidence in Singapore,
1968–2002. The Singapore Cancer Registry Report No. 6. Singapore.
20.Nickerson,D.A., Tobe,V.O. and Taylor,S.L. (1997) PolyPhred: automating
the detection and genotyping of single nucleotide substitutions using
fluorescence-based resequencing. Nucleic Acids Res., 25, 2745–2751.
21.Gordon,D., Abajian,C. and Green,P. (1998) Consed: a graphical tool for
sequence finishing. Genome Res., 8, 195–202.
22.Lee,L.G., Connell,C.R. and Bloch,W. (1993) Allelic discrimination by
nick-translation PCR with fluorogenic probes. Nucleic Acids Res., 21,
23.Slatkin,M. and Excoffier,L. (1996) Testing for linkage disequilibrium in
genotypic data using the expectation–maximization algorithm. Heredity,
24.Barrett,J.C., Fry,B., Maller,J. and Daly,M.J. (2005) Haploview: analysis
and visualization of LD and haplotype maps. Bioinformatics, 21,
25.Ardlie,K.G., Kruglyak,L. and Seielstad,M. (2002) Patterns of linkage
disequilibrium in the human genome. Nat. Rev. Genet., 3, 299–309.
26.Breslow,N.E. and Day,N.E. (1980) Statistical methods in cancer research.
Vol. I—The Analysis of Case–Control Studies. IARC, Lyon.
27.Gong,Z., Xie,D., Deng,Z., Bostick,R.M., Muga,S.J., Hurley,T.G. and
Hebert,J.R. (2005) The PPARg Pro12Ala polymorphism and risk for
incident sporadic colorectal adenomas. Carcinogenesis, 26, 579–585.
PPARg gene polymorphism and CRC risk
by guest on June 1, 2013
28.Werman,A., Hollenberg,A., Solanes,G., Bjorbaek,C., Vidal-Puig,A.J. and
Flier,J.S. (1997) Ligand-independent activation domain in the N terminus
of peroxisome proliferator-activated receptor gamma (PPARgamma).
Differential activity of PPARgamma1 and -2 isoforms and influence of
insulin. J. Biol. Chem., 272, 20230–20235.
29.Yen,C.J., Beamer,B.A., Negri,C., Silver,K., Brown,K.A., Yarnall,D.P.,
Burns,D.K., Roth,J. and Shuldiner,A.R. (1997) Molecular scanning of
the human peroxisome proliferator activated receptor gamma (hPPAR
gamma) gene in diabetic Caucasians: identification of a Pro12Ala PPAR
gamma 2 missense mutation. Biochem Biophys Res. Commun., 241,
Tan,C.E. and Ordovas,J.M. (2004) Differential effects of the C1431T
and Pro12Ala PPARgamma gene variants on plasma lipids and diabetes
risk in an Asian population. J. Lipid Res., 45, 674–685.
31.The International HapMap Consortium. (2003) The International HapMap.
Project. Nature, 426, 789–796.
32.Jiang,J., Gajalakshmi,V., Wang,J., Kuriki,K., Suzuki,S., Nakamura,S.,
Akasaka,S., Ishikawa,H. and Tokudome,S. (2005) Influence of the
C161Tbut notPro12Ala polymorphism
proliferator-activated receptor-gamma on colorectal cancer in an Indian
population. Cancer Sci., 96, 507–512.
33.Rangwala,S.M. and Lazar,M.A. (2004) Peroxisome proliferator-activated
receptor gamma in diabetes and metabolism. Trends Pharmacol. Sci., 25,
34.Bustin,S.A.,Dorudi,S., Phillips,S.M., Feakins,R.M. and Jenkins,P.J. (2002)
Local expression of insulin-like growth factor-I affects angiogenesis in
colorectal cancer. Tumour Biol., 23, 130–138.
35.Reinmuth,N., Fan,F., Liu,W., Parikh,A.A., Stoeltzing,O., Jung,Y.D.,
Bucana,C.D., Radinsky,R., Gallick,G.E. and Ellis,L.M. (2002) Impact of
insulin-like growth factor receptor-I function on angiogenesis, growth, and
metastasis of colon cancer. Lab Invest., 82, 1377–1389.
37.Giovannucci,E. (2001) Insulin, insulin-like growth factors and colon
cancer: a review of the evidence. J. Nutr., 131, 3109S–3120S.
38.Bull,A.W., Steffensen,K.R., Leers,J. and Rafter,J.J. (2003) Activation of
PPAR gamma in colon tumor cell lines by oxidized metabolites of
linoleic acid, endogenous ligands for PPAR gamma. Carcinogenesis, 24,
39.Gupta,R.A., Brockman,J.A., Sarraf,P., Willson,T.M. and DuBois,R.N.
(2001) Target genesof
receptor gamma in colorectal cancer cells.
40.Shimada,T., Kojima,K., Yoshiura,K., Hiraishi,H. and Terano,A. (2002)
Characteristics of the peroxisome proliferator activated receptor gamma
(PPARgamma) ligand induced apoptosis in colon cancer cells. Gut, 50,
41.Su,C.G., Wen,X., Bailey,S.T., Jiang,W., Rangwala,S.M., Keilbaugh,S.A.,
Flanigan,A., Murthy,S., Lazar,M.A. and Wu,G.D. (1999) A novel therapy
for colitis utilizing PPAR-gamma ligands to inhibit the epithelial
inflammatory response. J. Clin. Invest., 104, 383–389.
de la Chapelle,A., Spiegelman,B.M. and Eng,C. (1999) Loss-of-function
mutationsin PPAR gamma associatedwith human colon cancer. Mol. Cell,
43.Parkin,D.M., Whelan,S.L., Ferlay,J., Teppo,L. and Thomas,D. (2002)
Cancer Incidence in Five Continents. IARC, Lyon.
J. Biol. Chem., 276,
Received November 23, 2005; revised January 31, 2006;
accepted February 21, 2006
W.-P.Koh et al.
by guest on June 1, 2013