DIABETES/METABOLISM RESEARCH AND REVIEWS
RESEARCH
Diabetes Metab Res Rev 2014; 30: 232–240.
Published online in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/dmrr.2481
ARTICLE
Common quantitative trait locus downstream of RETN
gene identified by genome-wide association study is
associated with risk of type 2 diabetes mellitus in
Han Chinese: a Mendelian randomization effect
Chia-Min Chung1,2 Tsung-Hsien
Lin3,4 Jaw-Wen Chen5,6 Hsin-Bang
Leu5 Wei-Hsian Yin7 Hung-Yun Ho8
Sheng-Hsiung Sheu3,4 Wei-Chuan
Tsai9 Jyh-Hong Chen9 Shing-Jong
Lin5 Wen-Harn Pan1,2*
1
Institute of Biomedical Sciences, Academia
Sinica, Taipei, Taiwan
2
Division of Health Service Research and
Preventive Medicine, Institute of Population
Health Sciences, National Health Research
Institutes, Miaoli, Taiwan
3
Division of Cardiology, Department of
Internal Medicine, Kaohsiung Medical
University Hospital, Kaohsiung, Taiwan
4
Department of Internal Medicine, Faculty
of Medicine, College of Medicine, Kaohsiung
Medical University, Kaohsiung, Taiwan
5
Cardiovascular Research Center, National
Yang-Ming University, Taipei, Taiwan
6
Department of Medical Research and
Education, Taipei Veterans General
Hospital, Taipei, Taiwan
7
Division of Cardiology, Cheng-Hsin
General Hospital, and Faculty of Medicine,
School of Medicine, National Yang-Ming
University, Taipei, Taiwan
8
Taichung Veterans General Hospital,
Taichung, Taiwan
9
College of Medicine, National Cheng Kung
University, Tainan, Taiwan
*Correspondence to: Wen-Harn Pan,
Institute of Biomedical Sciences, Academia
Sinica, 128 Academia Rd. Sec.2, Taipei,
Taiwan and Division of Preventive Medicine
and Health Service Research, National
Health Research Institutes, Miaoli, Taiwan.
E-mail: [email protected],
[email protected]
Received: 1 March 2013
Revised: 16 September 2013
Accepted: 27 September 2013
Copyright © 2013 John Wiley & Sons, Ltd.
Abstract
Objective Plasma resistin level is a potential molecular link between obesity
and diabetes. Causal role of resistin, type 2 diabetes mellitus (T2DM) and genetic variants have not been thoroughly investigated. Therefore, we conducted
a genome-wide association study (GWAS) to identify quantitative trait loci
associated with resistin levels and investigated whether these variants were
prospectively associated with the development of metabolic syndrome (MetS)
and T2DM in an independent community-based cohort, the CardioVascular
Disease risk FACtors Two-township Study (CVDFACTS).
Research Design and Methods We genotyped 382 young-onset hypertensive (YOH) subjects with Illumina HumanHap550 chips and searched for quantitative trait loci (QTLs) of resistin in the 1st stage GWAS and confirmed the
finding in another 559 YOH subjects. Logistic regression was used to examine
the Mendelian randomization effects between genotypes of confirmed QTLs
and metabolic outcomes in 3400 subjects of CVDFACTS.
Results Two single nucleotide polymorphisms (SNP) (rs3745367 and
rs1423096) were significantly associated with resistin levels (p = 5.52 × 10 15
and p = 2.54 × 10 20) and replicated in another 559 YOH subjects
(p = 1.29 × 10 3 and p = 1.13 × 10 7), respectively. The SNP rs1423096
was further associated with the levels of HDL-C (p = 0.006), the risk of MetS
(OR = 2.21, p = 0.0034) and T2DM (OR = 1.62, p = 0.0063) in the CVDFACTS.
People with the haplotypes A-G and G-G determined by rs3745367 and rs1423096
showed a significantly increased T2DM risk (p = 0.0068 and p = 0.0035, respectively) compared with those with A-A haplotype.
Conclusion We have found that rs3745367 and rs1423096 on the RETN gene
were significantly associated with resistin levels. However, rs1423096, downstream of RETN, seems to be associated with MetS and T2DM risk more so than
rs3745367. The established genotype–disease association points to a causal
association of resistin and T2DM. Copyright © 2013 John Wiley & Sons, Ltd.
Keywords
GWAS; T2DM; Mendelian randomization
Introduction
The resistin level is potentially involved in the aetiology of insulin resistance
and metabolic traits, which has been implicated as a molecular link between
233
QTLs are associated with risk of T2DM
type 2 diabetes mellitus (T2DM) and obesity [1]. Some
population studies have shown that increased serum
resistin levels are associated with obesity, visceral fat
accumulation, insulin resistance [2], T2DM [3–5] and
inflammation [6]. Resistin plays important regulatory
roles in the cardiovascular system. Evidence suggests
that resistin is involved in pathological processes leading
to cardiovascular diseases (CVD) including endothelial
dysfunction [7], thrombosis [8], angiogenesis [9] and
smooth muscle cell dysfunction [10]. Despite some
controversial reports, accumulating clinical evidence
repeatedly demonstrates that elevated serum resistin
levels are predictive of risk of coronary artery disease
and peripheral artery disease [11–14]. These findings
support that the plasma levels of resistin may directly
lead to the development of insulin resistance, diabetes
and CVD, and are not merely a consequence of the
metabolic syndrome (MetS). Therefore, we would like to
examine the causal relationship between resistin and
cardiometabolic diseases by first discovering the genetic
factors involved in the modulation of plasma resistin
levels and second using a Mendelian randomization
approach to study the association between genotype and
cardiometabolic diseases. Mendelian randomization
takes advantage of the natural phenomenon of random
assortment of alleles at the time of gamete formation.
The random assignment mimics that of a randomized
controlled trial [15]. Because Mendelian randomization
takes place in the beginning of the life cycle, it is able to
examine the causal association of a genetic variant as
well as the quantitative trait influenced by the variants
with disease. The method uses the genetic markers
related to the exposure of interest, resistin in this case,
as the instrumental variables. It provides an opportunity
to investigate whether resistin is causally associated with
T2D and MetS, with genotype (at eQTL for resistin) being
the instrumental variable.
The plasma resistin levels, a potential marker and
internal facet of metabolic diseases [16], have a reasonably high heritability with estimates ranging from 40%
to 70% [17,18]. Several studies have found the SNP
420 C/G (rs1862513) on promoter of resistin gene,
one of the most commonly studied polymorphisms, was
associated with serum resistin levels [16,19,20]. Another
large-scale fine-mapping analysis of the RETN gene has
found that the SNPs including rs1477341, rs4804765,
rs1423096 and rs10401670 in the gene 3′ untranslated
region (UTR) were associated with circulating resistin
levels in European population [21]. Although the polymorphisms of resistin structural gene was associated with
resistin level [2,22], it has not been carefully investigated
on whether other novel variants influence the changes in
resistin level and the roles of these genetic variants on
subsequent clinical outcomes, including MetS, T2DM
Copyright © 2013 John Wiley & Sons, Ltd.
and coronary artery disease. By using genetic variants
identified by genome-wide association study (GWAS) as
an instrument variable, it offers the potential to assess
causation of resistin for T2DM.
We performed a GWAS to identify the quantitative trait
loci (QTL) regulating the resistin levels by using phenotypic and genotypic information of 941 young-onset
hypertensive (YOH) subjects including the Illumina
HumanHap550 SNP data for the initial 382 subjects.
The variants responsible for variation of resistin levels
showed genome-wide significance in both the first-stage
(with 382 YOH cases) and the second-stage (with 559
YOH cases) studies; subsequently, we determined the
association of these SNP variants with the risk of MetS
and T2DM in an independent large-scale communitybased cohort study, the CardioVascular Disease risk
FACtors Two-township Study (CVDFACTS).
Methods
Two-stage GWAS to determine QTL
influencing the plasma resistin levels
We included 941 hypertensive subjects recruited by the
Academia Sinica Multi-centered Young-onset Hypertension
(AS-YOH) Genetic Study: 382 in the initial stage genomewide scan and 559 in the second stage confirmatory
study. The inclusion criteria for hypertensive subjects are
as follows: (1) systolic blood pressure (SBP) ≥ 140 mmHg
and/or diastolic blood pressure (DBP) ≥ 90 mmHg over
a 2-month period or SBP/DBP ≥ 120/80 mmHg in
patients on antihypertensive medications in two consecutive visits within 2 months; (2) age, between 20 and
51 years at the first diagnosis of hypertension; (3) no
secondary causes of hypertension, such as chronic renal
disease, renal arterial stenosis, primary aldosteronism,
coarctation of the aorta, thyroid disorders, Cushing syndrome and pheochromocytoma, confirmed by extensive
clinical examinations, including blood chemistry examination, renal function tests, endocrine examination and
abdominal sonogram; (4) fasting glucose (FG) level, less
than 126 mg/dl; (5) body mass index (BMI), less
than 35 kg/m2, where BMI was defined as body weight
in kilogramme (kg) divided by height in metres
squared (m2); and (6) self-reported Han Chinese ethnicity
in more than two generations.
Data were collected according to standardized protocols. Blood pressure was measured according to the
protocol established for the Nutrition and Health Survey
in Taiwan [23]. Serum levels of resistin for all samples
were assayed by Taipei Institute of Pathology (Taipei,
Taiwan), using the enzyme immunoassay kit [resistin
Diabetes Metab Res Rev 2014; 30: 232–240.
DOI: 10.1002/dmrr
C.-M. Chung et al.
234
(Human) ELISA kit; Phoenix Pharmaceuticals Inc,
Belmont, USA]. In addition, the data on sociodemographic factors, the smoking and drinking habits,
the past medical history and the current medications
were obtained by interviewing the subjects. Our multicentre study was approved by the Human Investigation
Committee of Academia Sinica. Informed consent was
obtained from each participant at his or her first visit to
the clinic.
Study population to determine the
association between QTL of resistin and
metabolic traits, and T2DM
The study subjects were those recruited to the CVDFACTS
[24,25]. Briefly, the CVDFACTS is a community-based
follow-up cohort study designed to study the risk
factors of cardiovascular and metabolic diseases in
Taiwan. This study was initiated in 1993, and all residents
older than 3 years in Chu-Dung (northwest Taiwan)
and Pu-Tzu (southwest Taiwan) were invited to participate in the baseline examination. The follow-up examinations were performed in 1994–1997, 1997–1999 and
2000–2002. Data on socio-demographic factors, anthropometric parameters, smoking and drinking habits, past
medical history and current medications were obtained
by interviewing the subjects, and fasting blood was drawn
for biochemical examination, including measurement of
the levels of serum glucose, triglyceride (TG) and insulin.
Insulin sensitivity was estimated on the basis of the
homeostasis model assessment of insulin resistance
(HOMA-IR) formula (serum glucose levels × insulin / 22.5).
For the analysis of associations between SNPs and risk of
MetS (as defined by the NCEP ATP III criteria [26]) and
T2DM, we included 3400 subjects from the CVDFACTS. A
total of 748 subjects with MetS and 230 subjects with
T2DM were included in the current analysis. The more
detailed information of baseline characteristics of CVD
cohort study is provided in the Supplementary Table S1.
Our study was approved by the Human Investigation
Committee of Academia Sinica.
Genotyping methods
Genomic DNA was extracted from peripheral blood
samples of hypertensive subjects using the Puregene DNA
isolation kit (Gentra Systems, Minneapolis, MN, USA) for
YOH genetic study and using the phenol/chloroform
method for CVDFACTS. For the GWAS, genotyping
experiments were performed by deCODE genetics
(Reykjavik, Iceland) by using the Illumina Infinium II
Copyright © 2013 John Wiley & Sons, Ltd.
HumanHap550 SNP chips (Illumina, San Diego, CA,
USA) to analyze data from 382 leukocyte DNA samples.
Genotypic data were checked for quality as described in
detail elsewhere [27]. Two SNPs located at RETN with
log p ≥ 7 strongly correlated with serum resistin levels;
one at promoter (p = 6.97 × 10 6) and one at intron
1 (p = 6.94 × 10 4) of FN3KRP reaching statistical
threshold were also suggested as markers. These four were
genotyped in the second stage for further confirmation.
The genotyping for the samples in the second-stage confirmation study and for the CVDFACTS subjects was
performed using Sequenom MassARRAY System (San
Diego, CA, USA) by the Academia Sinica National
Genotyping Center (Taipei, Taiwan).
Statistical analysis
We performed a two-stage genome-wide QTL mapping
for resistin levels. General linear model (GLM) was
then used to associate the transformed resistin levels
with genotype adjusting for gender, age and BMI in both
the first-stage and the second-stage analysis, where the
distribution of resistin levels was normalized by taking a
square root transformation, because it is skewed to the
right. In the first stage, the genome-wide significance
level was set at 1 × 10 7 (≈ 0.05/509 174) according
to the Bonferroni multiple testing correction [28]. A multiple regression model was used to estimate the degree of
variation in plasma resistin levels explained by the
selected SNPs using the combined data from the first
and second stages. An examination of possible population
stratification was carried out using multidimensional
scaling analysis of PLINK software [29]. Quantile–
quantile (Q–Q) plot of the genome-wide QTL mapping
was also drawn to examine p-value distributions based
on 382 YOH patients.
The levels of serum glucose and TG in CVDFACTS
database were also normalized by square root transformation. The resistin QTL was examined against the aforementioned metabolic traits and outcomes. Analysis of
GLM was used to compare the mean levels of metabolic
parameters among genotype groups with adjustments
for age, gender, BMI and medication for T2DM as
covariates. The relation between the two SNPs on resistin
gene and the presence of MetS, T2DM and hypertension
was examined by logistic regression adjusting for age
and gender. The haplotypes were constructed with
rs3745367 and rs1423096. The logistic regression
analyses were used to examine associations between
haplotypes and risk of T2DM, adjusting for age and gender.
All statistical analyses were performed using SAS version 9.2
(Cary, North Carolina, USA), except that PLINK software
was used to evaluate population stratification.
Diabetes Metab Res Rev 2014; 30: 232–240.
DOI: 10.1002/dmrr
235
QTLs are associated with risk of T2DM
Results
Characteristics of subjects in the
AS-YOH Genetic Study
No significant differences were observed between subjects
in the first and second stages (Table 1) with respect to gender distribution, mean levels of HOMA-IR and resistin, and
proportion of smoking status. Significant differences were
noted for mean age and BMI of the subjects. The subjects
in the second stage were 4 years older than those in the first
stage, but the magnitude of the difference in BMI was small.
GWAS findings
The results of principal component analysis in stage 1 revealed no evidence for population stratification for
Table 1. Comparing characteristics of the study subjects included in the two stages of genome-wide association study
Characteristic
Initial stage
Second stage
p*
N
Male (%)
Age (years)
2
BMI (kg/m )
HOMA-IR
Resistin (μg/mL)
Smoking status (%)
382
68
38.4 ± 0.4
26.2 ± 0.2
3.04 ± 0.12
2.78 ± 0.01
21.9
559
69
42.9 ± 0.2
26.7 ± 0.1
3.43 ± 0.18
2.86 ± 0.11
24.7
0.67
<0.0001
0.03
0.06
0.56
0.36
Data are given either in percentage or in mean ± standard error.
BMI, body mass index; HOMA-IR, homeostasis model assessment
of insulin resistance
*Either t-test or chi-square was used to make comparisons between the first-stage and second-stage samples with a significance level set at p = 0.05.
Figure 2. Regional association plot of SNPs with circulating
resistin levels. It included 10 SNPs within the region from 7.6 to
7.67 Mb on chromosome 19
hypertensive subjects. Multidimensional scaling analysis
using PLINK also showed similar results (Figure S1).
To identify the QTLs influencing the resistin levels, we
performed GLM with adjustments for age, gender and
BMI to reduce the potential confounding effects. The major
results of the GWAS for resistin levels are shown in Figure 1.
The two SNPs (rs1423096 and rs3745367) most significantly associated with resistin levels were located in the 3′
UTR and intron 2 of the RETN gene (p = 2.54 × 10 20 and
5.52 × 10 15), respectively. Regional association plots of
SNPs in the RETN loci are displayed in Figure 2. In addition
to the most significant SNPs on the RETN gene, other
two suggestive significant SNPs located at promoter
Figure 1. Illumina HumanHap550 SNPs associated with resistin levels in the genome-wide association study. The general linear
model was used to examine associations between SNPs and resistin levels with adjustments for age, gender and body mass index with
the significance level set at log p = 7 (grey dashed horizontal line)
Copyright © 2013 John Wiley & Sons, Ltd.
Diabetes Metab Res Rev 2014; 30: 232–240.
DOI: 10.1002/dmrr
C.-M. Chung et al.
236
(rs1046875; p = 2.94 × 10 4) and intron 1 of FN3KRP
locus (rs2243523; p = 6.97 × 10 6) were also selected
for confirmation in the second stage by examining an additional 559 YOH patients. The two SNPs of RETN, but not
FN3KRP, showed significant associations with resistin levels
(Table 2): rs1423096 (p = 1.13 × 10 7) and rs3745367
(p = 1.29 × 10 3). When we entered these two SNPs simultaneously in the model, rs3745367 and rs1423096 are independently associated with resistin levels and accounted for
3.01% and 15.44% of the total variance of resistin levels in
the combined data analysis for 941 hypertensive subjects
(Table 2), respectively.
Association between rs1423096 and
rs3745367 in RETN and metabolic traits
We further examined the association of the two most significant SNPs, rs1423096 and rs3745367, with other metabolic
parameters, including hypertension, waist circumference
(WC) and the levels of high-density lipoprotein cholesterol
(HDL-C), TG and FG for the samples in CVDFACTS. The
SNP rs1423096 was significantly associated with HDL-C
levels (p = 0.003), not with TG, FG, WC and hypertension
(Table 3). The GG and GA genotypes of rs1423096 were
associated with a higher resistin level, lower HDL-C level
and an increased risk of MetS [OR = 2.21 (1.3–3.78),
p = 0.0034; Table 3]. Nevertheless, the SNP rs3745367
was not significantly associated with any metabolic traits
and MetS risk (Table 3).
Association between rs1423096 and
rs3745367 in RETN and the risk of T2DM
Results from the GLM with adjustments for sex, age and
BMI showed that G allele of rs1423096 was associated
with an increased risk of T2DM in a dominant fashion
[OR = 1.615 (1.145–2.277), p = 0.0063]. Similar phenomenon was also observed for G allele of rs3745367
[OR = 1.864 (1.148–3.026), p = 0.012; Table 4].
Associations of haplotypes defined by
rs3745367 and rs1423096 with risks of
T2DM
To determine which combination of these two SNPs is
most responsible for risk of T2DM, the haplotypes
constructed by rs3745367 and rs1423096 in this order
were then analyzed for their associations with T2DM risk
by logistic regression analyses using the covariates of gender, age and haplotypes. The results showed that subjects
with A-G and G-G haplotypes showed a significantly
increased T2DM risk with odds ratio (OR) of 1.362
(95% CI = 1.089–1.704; p = 0.0068) and 1.377 (95%
CI = 1.111–1.707; p = 0.0035) compared with those with
A-A haplotype, respectively, while subjects with G-A
1haplotype showed no significantly changed T2DM risks
compared with those with A-A haplotype (Figure 3.).
Therefore, these results indicate that rs1423096 is associated with the risk of T2DM development.
Discussion
We are the first to perform a GWAS for resistin levels and
discovered two QTLs (rs3745367 and rs1423096) on the
RETN gene associated with resistin levels in Chinese. Furthermore, rs1423096, downstream of RETN, is significantly
associated with MetS and T2DM. Based on basic assumptions of Mendelian randomization, we demonstrated the
evidence from the genotype–intermediate phenotype
Table 2. Association between resistin levels and the SNPs identified at the initial stage, confirmed at the second stage
SNP
rs1423096
rs3745367
rs1046875
rs2243523
Gene
Position
Allelea
MAF
β (SE)
β1 (SE)
RETN
RETN
FN3KRP
FN3KRP
3′ UTR
Intron 2
Promoter
Intron 1
Gb/A
Gb/A
G/Ab
Gb/T
0.22
0.36
0.49
0.44
0.366 (0.028)
0.238 (0.024)
0.046 (0.024)
0.018 (0.025)
0.322 (0.038)
0.199 (0.032)
0.036 (0.031)
0.001 (0.032)
Initial stage
(n = 382)
Second stage
(n = 559)
Combined
(n = 941)
p-valuec
p-valued
R (%)e
2.54 × 10
5.52 × 10
6.97 × 10
2.94 × 10
20
15
6
4
1.13 × 10
1.29 × 10
0.95
0.65
2
7
3
15.44
3.01
0.22
0.31
β, estimated effect size in initial stage; β1, estimated effect size in second stage; BMI, body mass index; GLM, general linear model; MAF,
minor allele frequency.
a
Major allele/minor allele.
b
Allele with higher resistin value.
c
Statistics corresponding to GLM testing association between genotypes and resistin levels after adjustments for age, gender and BMI in
the first-stage study.
d
Statistics corresponding to GLM testing association between genotypes and resistin levels after adjustment for age, gender and BMI in
the second-stage confirmatory study.
e 2
R was obtained by conducting multiple regression analysis with the combined samples obtained from both the first and second stages.
Copyright © 2013 John Wiley & Sons, Ltd.
Diabetes Metab Res Rev 2014; 30: 232–240.
DOI: 10.1002/dmrr
237
0.3360a
0.1808a
0.1122a
0.5955a
p-value
Table 4. Associations between genotypes of two SNPs on RETN
locus and the risk of type 2 diabetes mellitus
1
OR (95% CI)
1.30 (1.03–1.66)
p = 0.867*
1.18 (0.90–1.54)
p = 0.229*
1.14 (0.97–1.46)
p = 0.271*
1.02 (0.78–1.35)
p = 0.867*
Copyright © 2013 John Wiley & Sons, Ltd.
1
1.97 ± 0.23
80.5 ± 8.8
1.99 ± 0.07
43.18 ± 12.8
Mean ± SD
1.97 ± 0.24
81.6 ± 9.4
2.00 ± 0.08
42.68 ± 12.1
1.96 ± 0.24
80.4 ± 9.4
2.00 ± 0.09
42.60 ± 11.6
Genotypes
OR (95% CI)
p-value *
AG vs AA
GG vs AA
AG + GG vs AA
1.631 (1.053–2.527)
1.648 (1.164–2.334)
1.620 (1.148–2.287)
0.0283
0.0048
0.006
AG vs AA
GG vs AA
GA + GG vs AA
1.804 (1.09–2.982)
1.925 (1.159–3.199)
1.859 (1.145–3.018)
0.0214
0.0115
0.0122
SD, standard deviation; OR, odds ratio; CI, confidence interval.
A linear trend test is computed for continuous metabolic parameters across genotypes categories using generalized linear models.
*p-values were obtained from logistic regression analysis with adjustments for age, gender and body mass index.
OR, odds ratio; CI, confidence interval.
*p-values were obtained from logistic regression analysis with adjustments for age and gender.
a
1
Metabolic syndrome
1.28 (0.97–1.69)
p = 0.07*
1.51 (0.97–2.36)
p = 0.066*
OR (95% CI)
1.18 (0.90–1.55)
p = 0.23*
2.21 (1.3–3.78)
p = 0.0034*
1
rs3745367
Hypertension
0.9406a
0.1043a
0.2499a
0.006a
1.97 ± 0.24
79.6 ± 8.8
1.98 ± 0.08
46.6 ± 13.4
1.96 ± 0.24
81.0 ± 9.3
1.99 ± 0.08
42.49 ± 11.9
Triglyceride (mg/dl)
Waist circumference (cm)
Glucose (mg/dl)
HDL-C (mg/dl)
Mean ± SD
1.99 ± 0.26
80.9 ± 9.1
2.00 ± 0.08
42.35 ± 11.4
A/A (n = 468)
rs1423096
A/G (n = 1577)
G/G (n = 1272)
p-value
A/A (n = 169)
A/G (n = 214)
G/G (n = 2967)
Trait
rs3745367
SNP
rs1423096
Table 3. Associations between genotypes of rs1423096 and rs3745367 on RETN and metabolic parameters, levels of triglyceride, glucose and high-density lipoprotein cholesterol (HDL-C), and the waist circumference were averaged by genotypes of rs1423096 and rs3745367
QTLs are associated with risk of T2DM
Figure 3. The haplotypes in RETN gene and association with risk
of type 2 diabetes mellitus risks. The haplotypes were
constructed by rs3745367 and rs1423096 in this order. The logistic regression analyses were used to examine associations between haplotypes and risk of type 2 diabetes mellitus with
adjustments for age and gender. (*) indicates p < 0.05
(resistin levels) and the genotype–disease (T2DM) relations, which provided strong evidence supporting the
causal role of resistin level in the aetiology of T2DM.
The rs3745367 SNP located at intron 2 of resistin gene
found in our study was in moderate linkage disequilibrium
(LD) with rs1862513 (r2 = 0.45, D′ = 0.71 in Han Chinese
in Beijing (CHB) and Japanese in Tokyo (JPT) combined
samples). Our results are consistent with the findings of previous studies in that resistin concentration is significantly
influenced by SNPs on the promoter region. In addition, we
observed another stronger and separate signal for resistin
levels in the RETN locus at rs1423096 (2.54 × 10 20), which
resides in the 3′ UTR, about 887 kb downstream of the
reported variant rs1862513. Relatively weak LD exists
between rs1423096 and rs3745367 (r2 = 0.382, D′ = 0.64
in CHB and JPT combined samples). To test whether the
SNP identified in our study is a functioning SNP, we examined whether these SNPs are associated with certain cSNPs
or SNPs within 100 kb flanking the RETN gene. General
Diabetes Metab Res Rev 2014; 30: 232–240.
DOI: 10.1002/dmrr
C.-M. Chung et al.
238
linear model was used to examine the association between
imputed SNPs within 100 kb upstream/downstream of RETN
and resistin levels. The most significant SNPs were those located in RETN gene, but none of them are cSNPs (Supplementary Table S2). Further studies will be required to
identify the functional variant(s) responsible for resistin
levels or to understand how rs1423096 influences resistin
level and metabolic outcomes. In addition, two QTLs of
resistin, namely TYW3/CRYZ and NDST4 loci, were recently reported for Caucasian [30]. However, the SNPs in
these two QTLs were not available in the Illumina Infinium
II HumanHap550 SNP chips. We did not have the opportunity to study the association between these SNPs and
resistin levels in our population.
Resistin gene (RETN), a 12.5-kDa polypeptide, is
located on chromosome 19p13.3, which is also known as
FIZZ-3 (found in inflammatory zone 3) [6]. The polypeptide hormone belongs to the resistin-like molecule family
of cysteine-rich protein [31]. In humans, adipocytes contribute to only a small fraction of the resistin production
[32], and macrophages are thought to be the predominant source of circulating resistin [33,34]. Resistin is considered as a biomarker or mediator of insulin resistance,
inflammation and atherosclerosis [35,36]. Resistin via
crosstalk with other adipokines and compounds may be
detrimental to cardiovascular physiology. Resistin and
adiponectin have direct reciprocal effects on vascular
endothelial cells. Resistin induced the expression o
VCAM-1 and ICAM-1, but the induction of VCAM-1 and
ICAM-1 was inhibited by adiponectin [37]. A functional
link between leptin and resistin is suggested by suppression of resistin mRNA expression and protein levels in
ob/ob mice by leptin treatment [38]. Some population
studies [21,39–41] have shown that circulating concentrations of resistin are indeed associated with metabolic
traits and insulin resistance, suggesting that resistin might
play a role in the pathophysiology of cardiometabolic diseases. However, genetic associations between previously
studies RETN polymorphism and T2DM or related metabolic traits remain controversial, because of population
genetic heterogeneity and limited sample sizes.
Our study showed that two RETN QTLs (rs3745367
close to promoter and rs1423096 near 3′ UTR) are independently associated with increased risk of T2DM in a
large scale of cohort study. In several Japanese’s study,
the SNPs close to the promoter have been shown to associate with resistin level and risk of T2DM [4,19,20,40,42].
Although the association between rs3745367 and
increased risk of T2DM was recently reported for
Caucasians [21], this association was not found in the
Framingham study and in another cohort from Italy
[18,21]. The minor allele frequencies (MAF) of
rs3745367 in our study were similar to the Japanese
population (MAF 0.36 vs 0.41) but much higher than
Copyright © 2013 John Wiley & Sons, Ltd.
Caucasians (MAF 0.36 vs 0.24). Maybe this is why
significant associations can be seen more often in Asian
populations. Although several SNPs including rs1423096
in the 3′ UTR of RETN had been associated with resistin
levels in Caucasians and Japanese population [20,21],
we further demonstrated that this SNP is also associated
with risk of MetS and T2DM. The MAF of rs1423096 is
less than 10% in Caucasians but much higher (22%) in
Taiwanese. This discrepancy may make its differential
contributions to risk of MetS and T2DM in Caucasian
and Asians.
We performed haplotype analysis to determine which
combination of rs3745367 and rs1423096 is most responsible for risk of T2DM. We confirmed through this
approach that G allele at rs1423096 is the risk factor of
T2DM rather than G allele of rs3745367 (Figure 3). The
rs1423096 was located at 3′ UTR, which may play a vital
role in the formation of the 3′ end of mRNAs as well as in
stabilizing the transcripts.
A few limitations are discussed as follows. We did not
measure resistin levels in the CVDFACTS. Therefore, we
could not examine the magnitude of RETN polymorphisms on resistin levels and compare the observed and
the predicted effect on T2DM in this study. Second, only
two SNPs on RETN were included in the Illumina
HumanHap550 chip; further fine-mapping and functional
studies are needed to identify the most critical region.
Third, this study was conducted in hypertensive subjects.
There might still be concerns to apply the results of our
study to the general population.
In conclusion, our study provides evidence that GWAS
has the potential to reveal critical genetic variant influencing resistin levels. We have found that both rs3745367
and rs1423096 on the RETN gene were associated with
resistin levels. However, rs1423096, downstream of RETN,
seems to be associated with MetS and T2DM risk more so
than rs3745367. Our study provides evidence for the causal
link between resistin and T2DM development.
Author contributions
Conceived and designed the experiments: WHP CMC THL.
Performed the experiments: THC CMC JWC HBL WHY
HYH CTT SHS WCT JHC SJL WHP. Analyzed the data:
CMC. Contributed reagents/materials/analysis tools:
THL JWC WHP HCY. Wrote the article: CMC WHP.
Acknowledgements
We thank Academia Sinica National Genotyping Center (Taipei,
Taiwan) for performing genotyping in the second-stage study.
Diabetes Metab Res Rev 2014; 30: 232–240.
DOI: 10.1002/dmrr
239
QTLs are associated with risk of T2DM
This research has been supported by the Department of Health
(DOH90-TD-1037), the National Science Council (NSC91-3112-P001-025-YGPCP91-25), Academia Sinica (AS91IBMS2PP-A) and
Academia Sinica Genomics and Proteomics Program (2003–2006)
in Taiwan.
Conflict of interest
The authors have declared that there is no conflict of
interest.
References
1. Steppan CM, Bailey ST, Bhat S, Brown
EJ, Banerjee RR, Wright CM, et al. The
hormone resistin links obesity to diabetes. Nature 2001; 409(6818): 307–312.
2. Smith SR, Bai F, Charbonneau C,
Janderova L, Argyropoulos G. A promoter
genotype and oxidative stress potentially
link resistin to human insulin resistance.
Diabetes 2003; 52(7): 1611–1618.
3. Kunnari A, Ukkola O, Kesaniemi YA.
Resistin polymorphisms are associated
with cerebrovascular disease in Finnish
type 2 diabetic patients. Diabet Med
2005; 22(5): 583–589.
4. Osawa H, Onuma H, Murakami A, Ochi
M, Nishimiya T, Kato K, et al. Systematic
search for single nucleotide polymorphisms in the resistin gene: the absence
of evidence for the association of three
identified single nucleotide polymorphisms with Japanese type 2 diabetes.
Diabetes 2002; 51(3): 863–866.
5. Sentinelli F, Romeo S, Arca M, Filippi E,
Leonetti F, Banchieri M, et al. Human
resistin gene, obesity, and type 2 diabetes: mutation analysis and population
study. Diabetes 2002; 51(3): 860–862.
6. Holcomb IN, Kabakoff RC, Chan B,
Baker TW, Gurney A, Henzel W, et al.
FIZZ1, a novel cysteine-rich secreted
protein associated with pulmonary inflammation, defines a new gene family.
EMBO J 2000; 19(15): 4046–4055.
7. Chen C, Jiang J, Lu JM, Chai H, Wang X,
Lin PH, et al. Resistin decreases expression of endothelial nitric oxide synthase
through oxidative stress in human
coronary artery endothelial cells. Am J
Physiol Heart Circ Physiol 2010; 299(1):
H193–201.
8. Calabro P, Cirillo P, Limongelli G,
Maddaloni V, Riegler L, Palmieri R,
et al. Tissue factor is induced by resistin
in human coronary artery endothelial
cells by the NF-kB-dependent pathway.
J Vasc Res 2011; 48(1): 59–66.
9. Mu H, Ohashi R, Yan S, Chai H, Yang H,
Lin P, et al. Adipokine resistin promotes
in vitro angiogenesis of human endothelial cells. Cardiovasc Res 2006; 70(1):
146–157.
10. Calabro P, Samudio I, Willerson JT, Yeh
ET. Resistin promotes smooth muscle
cell proliferation through activation of
extracellular signal-regulated kinase
1/2 and phosphatidylinositol 3-kinase
pathways. Circulation 2004; 110(21):
3335–3340.
11. Frankel DS, Vasan RS, D’Agostino RB Sr,
Benjamin EJ, Levy D, Wang TJ, et al.
Copyright © 2013 John Wiley & Sons, Ltd.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
Resistin, adiponectin, and risk of heart
failure the Framingham Offspring Study.
J Am Coll Cardiol 2009; 53(9): 754–762.
Weikert C, Westphal S, Berger K, Dierkes
J, Mohlig M, Spranger J, et al. Plasma
resistin levels and risk of myocardial
infarction and ischemic stroke. J Clin
Endocrinol Metab 2008; 93(7):
2647–2653.
Chen BH, Song Y, Ding EL, Roberts CK,
Manson JE, Rifai N, et al. Circulating
levels of resistin and risk of type 2 diabetes in men and women: results from two
prospective cohorts. Diabetes Care 2009;
32(2): 329–334.
Ding Q, White SP, Ling C, Zhou W.
Resistin and cardiovascular disease.
Trends Cardiovasc Med 2011; 21(1):
20–27.
Davey Smith G, Ebrahim S. What can
Mendelian randomisation tell us about
modifiable behavioural and environmental exposures? BMJ 2005; 330(7499):
1076–1079.
Norata GD, Ongari M, Garlaschelli K,
Tibolla G, Grigore L, Raselli S, et al.
Effect of the 420C/G variant of the
resistin gene promoter on metabolic syndrome, obesity, myocardial infarction
and kidney dysfunction. J Intern Med
2007; 262(1): 104–112.
Hussain S, Bibi S, Javed Q. Heritability of
genetic variants of resistin gene in
patients with coronary artery disease a
family-based study. Clin Biochem 2011;
44(8–9): 618–622.
Menzaghi C, Coco A, Salvemini L,
Thompson R, De Cosmo S, Doria A, et al.
Heritability of serum resistin and its
genetic
correlation
with
insulin
resistance-related features in nondiabetic
Caucasians. J Clin Endocrinol Metab
2006; 91(7): 2792–2795.
Osawa H, Yamada K, Onuma H,
Murakami A, Ochi M, Kawata H, et al.
The G/G genotype of a resistin singlenucleotide polymorphism at
420
increases type 2 diabetes mellitus susceptibility by inducing promoter activity
through specific binding of Sp1/3. Am J
Hum Genet 2004; 75(4): 678–686.
Asano H, Izawa H, Nagata K, Nakatochi
M, Kobayashi M, Hirashiki A, et al.
Plasma resistin concentration determined
by common variants in the resistin gene
and associated with metabolic traits in an
aged Japanese population. Diabetologia
2010; 53(2): 234–246.
Hivert MF, Manning AK, McAteer JB,
Dupuis J, Fox CS, Cupples LA, et al.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
Association of variants in RETN with
plasma resistin levels and diabetesrelated traits in the Framingham Offspring Study. Diabetes 2009; 58(3): 750–
756.
Lau CH, Muniandy S. Adiponectin and
resistin gene polymorphisms in association with their respective adipokine
levels. Ann Hum Genet 2011; 75(3):
370–382.
Pan WH, Hung YT, Shaw NS, Lin W, Lee
SD, Chiu CF, et al. Elderly Nutrition and
Health Survey in Taiwan (1999–2000):
research design, methodology and content. Asia Pac J Clin Nutr 2005; 14(3):
203–210.
Chen HJ, Bai CH, Yeh WT, Chiu HC, Pan
WH. Influence of metabolic syndrome and
general obesity on the risk of ischemic
stroke. Stroke 2006; 37(4): 1060–1064.
Chuang SY, Bai CH, Chen WH, Lien LM,
Pan WH. Fibrinogen independently predicts the development of ischemic stroke
in a Taiwanese population: CVDFACTS
study. Stroke 2009; 40(5): 1578–1584.
Executive Summary of the Third Report
of the National Cholesterol Education
Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High
Blood Cholesterol in Adults (Adult Treatment Panel III). JAMA 2001; 285(19):
2486–2497.
Chung CM, Lin TH, Chen JW, Leu HB,
Yang HC, Ho HY, et al. A genome-wide
association study reveals a quantitative
trait locus of adiponectin on CDH13 that
predicts cardiometabolic outcomes. Diabetes 2011; 60(9): 2417–2423.
Genome-wide association study of
14,000 cases of seven common diseases
and 3,000 shared controls. Nature
2007; 447(7145): 661–678.
Purcell S, Neale B, Todd-Brown K,
Thomas L, Ferreira MA, Bender D, et al.
PLINK: a tool set for whole-genome association and population-based linkage
analyses. Am J Hum Genet 2007; 81(3):
559–575.
Qi Q, Menzaghi C, Smith S, Liang L, de
Rekeneire N, Garcia ME, et al. Genomewide association analysis identifies
TYW3/CRYZ and NDST4 loci associated
with circulating resistin levels. Hum Mol
Genet 2012; 21(21): 4774–4780.
McTernan PG, Kusminski CM, Kumar S.
Resistin. Curr Opin Lipidol 2006; 17(2):
170–175.
Fain JN, Cheema PS, Bahouth SW, Lloyd
HM. Resistin release by human adipose
tissue explants in primary culture.
Diabetes Metab Res Rev 2014; 30: 232–240.
DOI: 10.1002/dmrr
C.-M. Chung et al.
240
33.
34.
35.
36.
Biochem Biophys Res Commun 2003;
300(3): 674–678.
Kaser S, Kaser A, Sandhofer A,
Ebenbichler CF, Tilg H, Patsch JR.
Resistin messenger-RNA expression is
increased by proinflammatory cytokines
in vitro. Biochem Biophys Res Commun
2003; 309(2): 286–290.
Patel L, Buckels AC, Kinghorn IJ,
Murdock PR, Holbrook JD, Plumpton C,
et al. Resistin is expressed in human
macrophages and directly regulated by
PPAR gamma activators. Biochem Biophys
Res Commun 2003; 300(2): 472–476.
Lehrke M, Reilly MP, Millington SC, Iqbal
N, Rader DJ, Lazar MA. An inflammatory
cascade leading to hyperresistinemia in
humans. PLoS Med 2004; 1(2): e45.
Reilly MP, Lehrke M, Wolfe ML, Rohatgi
A, Lazar MA, Rader DJ. Resistin is an
inflammatory marker of atherosclerosis
in humans. Circulation 2005; 111(7):
932–939.
37. Kawanami D, Maemura K, Takeda N,
Harada T, Nojiri T, Imai Y, et al. Direct reciprocal effects of resistin and adiponectin
on vascular endothelial cells: a new
insight into adipocytokine–endothelial
cell interactions. Biochem Biophys Res
Commun 2004; 314(2): 415–419.
38. Rajala MW, Qi Y, Patel HR, Takahashi N,
Banerjee R, Pajvani UB, et al. Regulation
of resistin expression and circulating
levels in obesity, diabetes, and fasting.
Diabetes 2004; 53(7): 1671–1679.
39. Norata GD, Ongari M, Garlaschelli K,
Raselli S, Grigore L, Catapano AL. Plasma
resistin levels correlate with determinants of the metabolic syndrome. Eur J
Endocrinol 2007; 156(2): 279–284.
40. Osawa H, Tabara Y, Kawamoto R, Ohashi
J, Ochi M, Onuma H, et al. Plasma
resistin, associated with single nucleotide
polymorphism 420, is correlated with
insulin resistance, lower HDL cholesterol,
and high-sensitivity C-reactive protein in
the Japanese general population. Diabetes
Care 2007; 30(6): 1501–1506.
41. Perepelov AV, L’Vov VL, Liu B,
Senchenkova SN, Shekht ME, Shashkov
AS, et al. A similarity in the O-acetylation
pattern of the O-antigens of Shigella
flexneri types 1a, 1b, and 2a. Carbohydr
Res 2009; 344(5): 687–692.
42. Onuma H, Tabara Y, Kawamura R, Tanaka
T, Ohashi J, Nishida W, et al. A at single
nucleotide polymorphism-358 is required
for G at 420 to confer the highest plasma
resistin in the general Japanese population. PLoS One 2010; 5(3): e9718.
Supporting information
Additional supporting information may be found in the online version of this article at the publisher’s web site.
Copyright © 2013 John Wiley & Sons, Ltd.
Diabetes Metab Res Rev 2014; 30: 232–240.
DOI: 10.1002/dmrr