Download The haplotype of the growth-differentiation factor 15 gene is

Clinical Science (2010) 118, 137–145 (Printed in Great Britain) doi:10.1042/CS20080637
The haplotype of the growth-differentiation
factor 15 gene is associated with left ventricular
hypertrophy in human essential hypertension
Xiaojian WANG1 , Xu YANG 1 , Kai SUN, Jingzhou CHEN, Xiaodong SONG, Hu WANG,
Zhe LIU, Changxin WANG, Channa ZHANG and Rutai HUI
Sino-German Laboratory for Molecular Medicine, Key Laboratory for Clinical Cardiovascular Genetics of Ministry of Education,
FuWai Cardiovascular Hospital and Cardiovascular Institute, Chinese Academy of Medical Sciences and Peking Union Medical
College, Beijing 100037, China
A
B
S
T
R
A
C
T
GDF15 (growth-differentiation factor 15) is a novel antihypertrophic factor which is induced in
the heart in response to pressure overload and plays an important regulatory role in the process
of hypertrophy. In the present study, we have investigated the relationship between GDF15 gene
variants and left ventricular hypertrophy in human essential hypertension. A community-based
hypertensive population sample of 1527 individuals (506 men and 1021 women) was genotyped
for three GDF15 genetic variants, including one tag variant − 3148C > G (rs4808793) and
two exonic variants + 157A > T (rs1059369) and + 2438C > G (rs1058587). The effects of
those variants on gene expression were studied by use of luciferase reporter assays and the
determination of plasma GDF15 levels. Only the tag variant − 3148G was significantly associated
with a lower risk of left ventricular hypertrophy [odds ratio = 0.75 (95 % confidence interval,
0.63–0.89); P = 0.0009]. Multiple regression analyses confirmed that − 3148G predicted the
decrease in left ventricular end-diastolic diameter (β = −0.10, P = 0.0001), end-systolic diameter
(β = −0.09, P = 0.0007), mass (β = −0.11, P < 0.0001) and indexed mass (β = −0.12, P < 0.0001).
These effects were independent of conventional factors, including gender, age, body surface area,
blood pressure, diabetes, cigarette smoking and alcohol consumption. The transcription activity
of the − 3148G-containing construct was increased 1.45-fold (P = 0.015) at baseline and 1.73-fold
(P = 0.008) after stimulation with phenylephrine when compared with the − 3148C construct.
The − 3148G allele was also associated with a significant increase in the plasma GDF15 level
in hypertensive subjects (P = 0.04). In conclusion, the results show that a promoter haplotype
containing the − 3148G variant increases GDF15 transcription activity and is associated with
favourable left ventricular remodelling in human essential hypertension.
Key words: blood pressure, cardiac remodelling, growth-differentiation factor 15 (GDF15), hypertension, hypertrophy, pressure
overload.
Abbreviations: ACS, acute coronary syndrome; ANF, atrial natriuretic factor; BP, blood pressure; BSA, body surface area;
C/EBPγ , CCAAT/enhancer-binding protein; CHB, Chinese Han in Beijing; CI, confidence interval; DBP, diastolic BP; DMEM,
Dulbecco’s modified Eagle’s medium; GDF15, growth-differentiation factor 15; IVST, interventricular septal thickness; LD, linkage
disequilibrium; LV, left ventricular; LVH, LV hypertrophy; LVIDD, LV internal dimension at end-diastole; LVISD, LV internal
dimension at end-systole; LVM, LV mass; LVMBSA , LVM indexed to BSA; NCS, newborn calf serum; OR, odds ratio; PE,
phenylephrine; PWT, posterior wall thicknesses; RWT, relative wall thickness; SBP, systolic BP; TGFβ, transforming growth
factor β.
1
These two authors contributed equally to the work.
Correspondence: Dr Rutai Hui (email [email protected]).
C
The Authors Journal compilation
C
2010 Biochemical Society
137
138
X. Wang and others
INTRODUCTION
LVH [LV (left ventricular) hypertrophy] is a strong
independent risk factor for cardiovascular morbidity
and all-cause mortality [1,2]. The development of
LVH is dependent on genetic predisposition and the
accumulation of risk factors, including BP (blood
pressure), age, gender, lifestyle and diabetes [3]. BP can
only account for approx. 50 % of the increase in LVM
(LV mass) in hypertensive patients [4]. Much of the BPindependent variation in cardiac mass may be attributable
to inherited genetic causes [5,6].
GDF15 (growth-differentiation factor 15) is a distant
member of the TGFβ (transforming growth factor β)
superfamily. GDF15 was originally found to be involved
in tumour progression and invasiveness [7], but growing
evidence has shown that GDF15 is also a protective
factor in response to cardiovascular injury [8,9]. Although
baseline expression of GDF15 is negligible in the heart, it
is up-regulated rapidly in the myocardium following the
pathological or environmental stress [8,9]. After being
secreted as a disulfide-linked dimeric protein, GDF15
suppresses the development of maladaptive hypertrophy
and limits myocardial ischaemia/reperfusion injury [8,9].
Furthermore, the elevated concentration of GDF15 in
plasma has been associated with an increase in the risk
of developing future cardiovascular events in women
[10], and an increase in the risk of death in patients
with chronic heart failure and non-ST-elevation ACS
(acute coronary syndrome) [11–13]. The involvement
of GDF15 in cardiac regulation makes it a new target
for risk stratification and therapeutic decision making in
cardiovascular disease.
Given the rapid change in GDF15 expression levels
in response to pressure overload and the important role
of GDF15 in the regulation of cardiac remodelling,
we hypothesized that genetic variants which influence
the expression and function of GDF15 may confer the
susceptibility to LVH and cardiac size in human
hypertension. To test the hypothesis, we investigated
the potential relationship of genetic variants of GDF15
with LVH and LVM in a community-based hypertensive
population.
MATERIALS AND METHODS
Study population
The study subjects were recruited from a communitybased cross-sectional study, which investigated the prevalence and risk factors of LVH in a Chinese hypertensive
population. The present study was conducted in Xinyang
County, a central region in China from March 2005 to
August 2005. The study was approved by the Institutional
Review Boards of the participating hospitals and the
Ministry of Science and Technology of China. All subjects
C
The Authors Journal compilation
C
2010 Biochemical Society
who participated in the study provided written informed
consent.
BP was measured by trained professionals with a standardized mercury sphygmomanometer. Two readings were
recorded at least 30 s apart, after a minimum of 5 min of
rest, in the sitting position on the subject’s right upper
arm. The average was used for analysis. Hypertension was
defined as an SBP (systolic BP) 140 mmHg or a DBP
(diastolic BP) 90 mmHg, or the use of antihypertensive
medication. In total, 15835 subjects of both sexes
were invited to participate in the study from 60 rural
communities; of those, 13444 subjects (5270 men and
8174 women) completed the survey, and 5440 patients had
essential hypertension (40.5 %). Among the hypertensive
patients, 4869 underwent echocardiography (89.5 %).
A total of 25 communities were randomly selected
from the 60 communities for this association study. All
hypertensive patients with detailed echocardiography
determination were enrolled. Patients were excluded
if they took antihypertensive drugs, or had valvular
heart disease, pulmonary hypertension or coronary heart
disease. Eventually, 1527 subjects were recruited, 506 men
and 1021 women.
Echocardiographic measurement
M-dimensional and bi-dimensional echocardiography
was performed using commercially available instruments
[HP 5500 (Phillips Medical System); or HDI 3000 (ATL)].
Three physician-echocardiographers who were blinded
to the patients’ genotypes supervised the examination.
Measurements included LVIDD and LVISD (LV internal
dimension at end-diastole and end-systole respectively),
IVST (interventricular septal thickness), and PWT
(posterior wall thicknesses) in diastole up to three
cardiac cycles, in accordance with methods outlined by
the American Society of Echocardiography [14]. Two
cardiologists analysed each recording and the mean values
were used. LVM was calculated using the equation:
LVM(g) = 0.8 × [1.04 × (LVIDD + IVST + PWT)3
−(LVIDD)3 ] + 0.6 [15].
Intra-and inter-observer coefficients of variation for
LVM were 11–13% respectively. LVM was indexed to
BSA (body surface area) to obtain LVMBSA . LVH was
defined using the partition values of LVMBSA > 115 g/m2
for men and > 96 g/m2 for women [16]. RWT (relative
wall thickness) was calculated as 2 × the PWT/LVIDD
ratio [17].
Clinical data collection
All the participants completed a questionnaire on current
medication and lifestyle. A blood sample was taken
from all participants after overnight fasting. Plasma and
urine biochemical variables were determined by using
GDF15 variants and LVH in hypertension
Figure 1 Structure of the GDF15 gene, and the effect of the haplotype on the transcription activity of GDF15
(A) Structure of GDF15 gene. Boxes indicate exons, solid black lines indicate the intron and intergenic regions, and the vertical arrows indicate the genetic
variants. (B) Upper panels, schematic representation of the reporter gene constructs that contained either the C-C-C haplotype (pGL3-GDF15-wild) or T-T-G haplotype
(pGL3-GDF15-mutant) at the promoter region. Lower panel, assessment of the promoter activity in neonatal rat cardiomyocytes. Cardiomyocytes were transfected with
pGL3-GDF15-mutant and pGL3-GDF15-wild, stimulated with PE for 24 h and the relative transcription activity was measured. Values represent the mean +
− S.D. of
nine experiments. pGL3-basic was used as a negative control without any promoter sequence, and pGL3-control was used as a positive control. The P value was
determined by using a Student’s t test.
standard methods on an automatic analyser (Hitachi
7060). Diabetes mellitus was diagnosed when the subject
had a fasting glucose > 7.0 mmol/l or was taking
medication for diabetes [18]. Smoking was defined as
a history of smoking > 2 pack-years and/or smoking
in the last year. Alcohol consumption was defined as
continuously drinking for 1 year with the frequency of at
least once per week.
Selection of the GDF15 variants
The genetic architecture of the GDF15 gene and its
flanking regions (9.0 kb in size) were analysed by
using the pairwise method of the Haploview version
of the Tagger program in the HapMap CHB (Chinese
Han in Beijing) database [19]. There is a strong LD
(linkage disequilibrium) block in the promoter region,
including five variants: rs12459566 (located at position
− 3690T > C, relative to the transcription start site),
rs4808793 (−3148C > G), rs8101804 (−862C > T) and
rs12459782 (−474C > T) in the promoter region,
and rs1059519 (+ 56C > G) in exon 1 (Figure 1).
The natural haplotype block of the promoter region
was confirmed by sequencing the 3.9-kb 5 upstream
promoter region and 100 bp of exon 1 in 30 randomly
selected subjects. All five variants were in strong LD
(D > 0.9 and r2 > 0.9). Therefore any one of the five
variants could reflect the natural haplotype block of
GDF15 (T-C-C-C-C and C-G-T-T-G). The variant
− 3148C > G (rs4808793) was selected for the following
genotyping assay because there is a natural BsrI digestion
site within the sequence.
Two other potential functional genetic variants,
including the exonic variants rs1059369 (+ 157A > T, in
exon 1) and rs1058587 (+ 2438C > G, in exon 2), were
C
The Authors Journal compilation
C
2010 Biochemical Society
139
140
X. Wang and others
also selected for genotyping. These variations result in
non-synonymous amino acid changes at residue 48T > S
in the precursor protein and 6H > D in the mature
protein of GDF15 respectively. The variant rs1059369
was in complete LD with variant rs1804826 (D =1, r2 = 1)
according to the HapMap CHB data.
Genotyping
Genomic DNA was extracted from cellular buffy
coat. The three selected polymorphisms, − 3148C > G
(rs4808793), + 157A > T (rs1059369) and + 2438C > G
(rs1058587), were genotyped in all 1527 subjects by
PCR-RFLP (restriction-fragment-length polymorphism). The conditions for the PCR are shown in
Supplementary Table S1 (available at http://www.ClinSci.
org/cs/118/cs1180137add.htm). Reproducibility of genotyping was confirmed by bidirectional sequencing in 100
randomly selected samples and the reproducibility was
100 %.
Luciferase reporter constructions
The 1138 bp fragments (from − 1061 to + 77) were
amplified from genomic DNA with the forward primer
5 -CCGCCTCGAGGCATGTAATCCCACCACCAAG-3 (containing a XhoI restriction site underlined) and
the reverse primer 5 -CCGCAAGCTTAGCACCAGCAACACCAGG-3 (containing a HindIII restriction site
underlined). The resultant PCR products were doubledigested with XhoI/HindIII and cloned into the multiple
cloning sites of pGL3-basic vector (Promega). The vector
containing the genotype C-C-C (rs8101804-rs12459782rs1059519) was designated as pGL3-GDF15-wild and the
vector containing the T-T-G genotype as pGL3-GDF15mutant (Figure 1). The GDF15 promoter sequences in
both vectors were confirmed by direct sequencing.
DNA transfection and luciferase assay
Primary cultures of cardiac myocytes were prepared
from ventricles of 1–2-day-old neonatal Sprague–Dawley
rat hearts by a method reported previously [20].
After separation from non-myocyte fibroblasts through
pre-plating, myocytes were replated at a density of
0.5 × 105 in 24-well plates. Following 48 h of culture
in DMEM (Dulbecco’s modified Eagle’s medium; Life
Technologies) containing 10 % (v/v) NCS (newborn
calf serum; Hyclone), cardiomyocytes were grown in
fresh DMEM containing 1 % (v/v) NCS for 24 h.
GDF15 promoter–luciferase fusion plasmids (1 μg) and
0.05 μg of pRL-null plasmid (Promega) were cotransfected into cardiomyocytes with LipofectamineTM
2000 (Invitrogen). The cells were then stimulated with
1 × 10−4 mol/l PE (phenylephrine) in fresh DMEM containing 1 % (v/v) NCS for 24 h. Luciferase activity was
determined and normalized to the pRL-null luciferase
activity using the Dual Luciferase assay kit (Promega).
C
The Authors Journal compilation
C
2010 Biochemical Society
Three independent experiments were performed in
triplicate.
Determination of genotype-dependent
plasma levels of GDF15
The plasma level of GDF15 antigen was determined in
136 hypertensive patients (43 males) randomly selected
from the recruited subjects. To minimize any potential
confounding effects, the subjects were recruited only
when they met the following criteria: (i) 40–65 years
of age; (ii) SBP 160 mmHg and/or DBP 100
mmHg; (iii) using neither antihypertensive medication
nor NSAIDs (non-steroidal anti-inflammatory drugs);
and (iv) no history of diabetes mellitus. The variant
− 3148C > G (rs4808793) was genotyped in all patients.
Of the patients, 60 were homozygous for the C allele,
64 were heterozygous for the G allele and 12 were
homozygous for the G allele. The GDF15 concentrations
were determined twice by using an ELISA kit (R&D
Systems) as described previously [21].
Statistical analysis
Results are presented as percentages, medians (25–75th
percentiles) or means (S.E.M.), as indicated. A χ 2 test
was applied to compare qualitative variables and the
Hardy–Weinberg equilibrium. Differences in GDF15
plasma levels were analysed by the non-parametric
Mann–Whitney test between two groups, or the Kruskal–
Wallis test among distinct groups. Other quantitative
variables were tested with a Student’s t test or oneway ANOVA. The associations of the quantitative
echocardiography variables with genetic variants were
detected by multiple regression analyses adjusted for
age, gender, BSA, SBP, DBP, cigarette smoking, alcohol
consumption and diabetes status. The ‘gene–dosage
effect’ of the positively associated variants was tested by
linear regression, based on the number of polymorphic
alleles (0, 1 or 2). Stepwise multiple linear regression
analysis was used to examine whether the number of
variant alleles carried by each subject had statistical
influence on LV diameters and LVM, independent of the
conventional covariates mentioned above. The α level for
entry and removal of terms at each forward step was 0.10.
Multiple testing was adjusted by using the Bonferroni
correction. Two-tailed values of P < 0.05 were considered
statistically significant. Analyses were performed with
SPSS 13.0 for Windows.
RESULTS
The genotype frequencies for all the three GDF15
variants were in accordance with the Hardy–Weinberg
equilibrium. No LD was found between any pairwise
analysis among the three selected variants. As shown
in Table 1, only − 3148C > G was found to be
GDF15 variants and LVH in hypertension
Table 1 Association of the three variants of GDF15 studied with LVH in the study population
ORs were calculated with 2 × 3 or 2 × 2 cross-tabulation.
(A) − 3148C > G (rs4808793)
Genotype (n)
Allele (n)
Group
CC
GC
GG
P value
C
G
OR (95 % CI)
P value
With LVH
Without LVH
318 (58.9 %)
497 (50.4 %)
189 (35.0 %)
408 (41.3 %)
33 (6.1 %)
82 (8.3 %)
0.005
835 (76.6 %)
1402 (71.0 %)
255 (23.4 %)
572 (29.0 %)
0.75 (0.63–0.89)
0.0009
(B) + 157A > T (rs1059369)
Genotype (n)
Allele (n)
Group
AA
AT
TT
P value
A
T
OR (95 % CI)
P value
With LVH
Without LVH
253 (46.9 %)
412 (41.7 %)
227 (42.0 %)
469 (47.5 %)
60 (11.1 %)
106 (10.7 %)
0.109
733 (67.9 %)
1293 (65.5 %)
347 (32.1 %)
681 (34.5 %)
1.113 (0.95–1.30)
0.185
(C) + 2438C > G (rs1058587)
Genotype (n)
Allele (n)
Group
CC
GC
GG
P value
C
G
OR (95 % CI)
P value
With LVH
Without LVH
286 (53.0 %)
482 (48.8 %)
214 (39.6 %)
412 (41.7 %)
40 (7.4 %)
93 (9.4 %)
0.203
786 (72.8 %)
1376 (69.7 %)
294 (27.2 %)
598 (30.3 %)
1.162 (0.99–1.37)
0.074
significantly associated with LVH (P = 0.005). The C
allele of the − 3148 locus was more frequently identified
in individuals with LVH, whereas the G allele was
more often found in individuals without LVH {OR
(odds ratio), 0.75 [95 % CI (confidence interval) 0.63–
0.89]; P = 0.0009}. The difference was still statistically
significant after the Bonferroni correction (P < 0.017; i.e.
0.05/3), suggesting the minor allele − 3148G may be a
‘protective’ allele.
Table 2 shows the clinical characteristics according
to the three genotypes of − 3148C > G. Among
the anthropometric and biochemical variables, DBP
decreased significantly with an increase in G alleles
(P = 0.015). With respect to the echocardiographic
parameters, the − 3148C > G genotype was associated
with LVIDD, LVISD, LVM and LVMBSA in the
whole sample. The association remained, revealing
remarkably low P values, after correction for other
conventional factors that might influence cardiac size,
such as age, gender, BSA, SBP, DBP, cigarette smoking,
alcohol consumption and diabetes status (Table 3). The
coefficients of linear regression of − 3148 and each cardiac
variable on the number of G alleles were all statistically
significant (P < 0.0001), suggesting that − 3148G had a
‘gene–dosage’ or ‘addictive’ effect on these variables.
Stepwise multiple regression analyses showed that, of all
of the variables that might influence cardiac size, − 3148G
was one of the two (the other is BSA) most important
independent predictors of LV diameters, non-indexed and
indexed LVM (Table 3).
As the G allele of − 3148 reflects the C-G-TT-G (−3690/−3148/−862/−474/ + 56) haplotype of the
GDF15 gene, C-G-T-T-G might be a protective
haplotype against cardiac remodelling in hypertension.
To test whether the variants affected the transcriptional
activity of the GDF15 minimal promoter (−1061 to + 77,
relative to the transcript start site), the mutant haplotype
(T-T-G) and the wild-type haplotype (C-C-C) were
transfected into neonatal rat cardiomyocytes. As shown
in Figure 1, the promoter activity of mutant promoter
was 1.45-fold higher (P = 0.015) that of the wild-type
construct at baseline, and 1.73-fold higher (P = 0.008)
after the stimulation with PE. Moreover, PE significantly
increased luciferase activity of the GDF15 mutant
promoter in transfected cardiomyocytes (P = 0.02), but
failed to increase the activity of wild-type promoter.
These results suggest that the mutant haplotype promoter
was more sensitive to the hypertrophic stimulation and
had higher transcription activity than the wild-type
haplotype promoter.
To investigate further the expressional change induced
by the variant in the promoter, the plasma concentrations
of GDF15 were determined in hypertensive patients
(Figure 2). The − 3148G allele carriers had a significant
increase in GDF15 levels (P = 0.04, as determined using
a Kruskal–Wallis test). The plasma GDF15 levels were
significantly higher in patients carrying the GG genotype
[774 ng/l (25–75th percentiles, 555–1441 ng/l)] than in
those carrying the CC genotype [559 ng/l (25–75th
percentiles, 328–763 ng/l); P = 0.03]. These results were
C
The Authors Journal compilation
C
2010 Biochemical Society
141
142
X. Wang and others
Table 2 Clinical characteristics according to the three genotypes of − 3148C > G
Values are means (S.E.M) or numbers (percentage). HR, heart rate; Ccr, creatinine clearance rate; NS, not significant.
Variant of the − 3148C > G genotype
Characteristic
CC (n = 815)
CG (n = 597)
GG (n = 115)
P value
Age (years)
Male gender (n)
BSA (m2 )
SBP (mmHg)
DBP (mmHg)
HR (beats/min)
Ccr (ml/min)
Cigarette smoking (n)
Alcohol consumption (n)
Diabetes status (n)
IVST (mm)
PWT (mm)
LVIDD (mm)
LVISD (mm)
RWT (%)
LVM (g)
LVMBSA (g/m2 )
LVEF (%)
58.3 (0.3)
266 (32.6 %)
1.64 (0.01)
165.1 (0.8)
100.0 (0.4)
73.3 (0.5)
95.7 (1.2)
136 (16.7 %)
146 (17.9 %)
96 (11.8 %)
10.1 (0.1)
9.8 (0.1)
45.6 (0.2)
29.9 (0.2)
43.4 (0.3)
158.1 (1.5)
96.3 (0.9)
58.7 (0.4)
58.3 (0.4)
202 (33.8 %)
1.64 (0.01)
162.8 (1.0)
98.3 (0.5)
72.9 (0.5)
93.5 (1.4)
89 (14.9 %)
105 (17.6 %)
72 (12.1 %)
10.0 (0.1)
9.7 (0.1)
45.1 (0.2)
28.9 (0.2)
43.8 (0.3)
150.2 (1.7)
91.6 (1.0)
59.9 (0.4)
58.1 (0.8)
38 (33.0 %)
1.65 (0.01)
163.3 (2.3)
97.6 (1.1)
74.5 (1.3)
91.8 (2.8)
19 (16.5 %)
22 (19.1 %)
16 (13.9 %)
9.8 (0.1)
9.6 (0.1)
43.9 (0.4)
28.4 (0.5)
44.2 (0.8)
144.5 (3.5)
87.2 (1.9)
59.7 (0.9)
NS
NS
NS
NS
0.015
NS
NS
NS
NS
NS
NS
NS
<0.001
<0.001
NS
<0.001
<0.001
NS
Table 3 Echocardiographic variables according to the − 3148C > G genotype
Values are means (S.E.M.). In the univariate analysis, age, gender, BSA, SBP, DBP, cigarette smoking, alcohol consumption and diabetes status were used as covariates.
For the linear regression, a gene–dosage effect test (the independent variable was the number of G alleles, i.e. CC = 0, GC = 1 and GG = 2) was used. The
protective effects of − 3148G (CC = 0, GC = 1 and GG = 2) on cardiac variables were analysed in multiple regression models, adjusted for the factors that influence
cardiac remodelling, including age, gender, BSA, SBP, DBP, cigarette smoking, alcohol consumption and diabetes status. β, standardized regression coefficient; NS, not
significant.
Univariate analysis
Variant of the − 3148C > G genotype
Linear regression
Stepwise multiple regression
Variable
CC (n = 815)
CG (n = 597)
GG (n = 115)
P value
P value
β
P value
IVST (mm)
PWT (mm)
LVIDD (mm)
LVISD (mm)
RWT (%)
LVM (g)
LVMBSA (g/m2 )
LVEF (%)
10.1
9.8
45.5
29.8
43.3
157.1
95.5
59.8
10.0
9.6
44.7
28.9
43.6
150.4
91.4
60.6
9.8
9.6
43.8
28.4
44.3
142.7
86.8
60.2
0.131
0.266
0.0001
0.002
0.397
<0.0001
<0.0001
0.307
0.035
0.063
<0.0001
<0.0001
0.226
<0.0001
<0.0001
0.040
NS
NS
−0.10
−0.09
NS
−0.11
−0.12
NS
NS
NS
0.0001
0.0007
NS
<0.0001
<0.0001
NS
(0.1)
(0.1)
(0.2)
(0.2)
(0.3)
(1.4)
(0.8)
(0.6)
(0.1)
(0.1)
(0.2)
(0.2)
(0.3)
(1.6)
(1.0)
(0.6)
(0.1)
(0.1)
(0.4)
(0.5)
(0.7)
(3.5)
(2.2)
(1.0)
consistent with those from the luciferase activity
study.
DISCUSSION
The results of the present study show that a GDF15
haplotype, which significantly increases the promoter
C
The Authors Journal compilation
C
2010 Biochemical Society
activity and consequently the plasma level of GDF15,
was associated with a decrease in LVIDD, LVISD, LVM
and LVMBSA in hypertensive patients. The effects were
independent of BP levels, age, gender and other clinical
factors that influence cardiac size. To the best of our
knowledge, this is the first demonstration that genetic
variants of GDF15 are associated with favourable cardiac
hypertrophy in human essential hypertension.
GDF15 variants and LVH in hypertension
Figure 2 Correlation of serum levels of GDF15 with
genotypes of the variant − 3148C > G
The GDF15 levels are presented as box (25th percentile, median and 75th
percentile) and whisker (10 and 90th percentiles) plots. The number of patients
in each group is indicated. The correlation was significant (P = 0.041).
GDF15 has been identified as an antihypertrophic
endocrine factor [9]. It is barely detectable in the heart at
baseline, but is up-regulated rapidly following pressure
overload. The increase in GDF15 levels might be an
intrinsic protective mechanism that counter-regulates
the hypertrophic stimuli. In our present study, we
found that rs4808793 (−3148C > G) was associated
with an increase in GDF15 promoter activity and
circulating GDF15 protein levels. Furthermore, − 3148G
was negatively associated with LVM in both quantitative
and dichotomized phenotypes. It is noteworthy that
the significant decrease in LVM and LVMBSA with the
number of − 3148G alleles are contributed to by
decreased ventricular diameters, rather than differences
in wall thickness. This is consistent with the putative
role of GDF15 in modifying cardiac mass in animal
models. It has been shown that Gdf15-overexpressing
mice with a low- or medium-expression level have normal
LV thickness, but significantly a smaller LV chamber
compared with the wild-type mouse. Conversely, Gdf15
gene-targeted mice have larger LV dimensions after
being normalized to body weight [9]. Therefore one
can speculate that, in response to hypertension, the
haplotype of the GDF15 transcription regulatory region
might promote the GDF15 antihypertrophic effects by
increasing the protein level.
It has been reported that increased levels of circulating
GDF15 are associated with an increased risk of death
in patients with non-ST-elevation ACS and chronic heart
failure [11,12]. However, the high GDF15 level is unlikely
to be a cause of these cardiovascular events, because
overexpression of GDF15 by injection of recombinant
GDF15 protein has been shown to attenuate heart failure
in mlp (muscle lim protein) gene-targeted mice [9]. As
pathological cardiac hypertrophy and heart failure are
continuously progressive, the protective effect of the
GDF15 haplotype on LVH might only function within a
given stage in hypertensive patients. At an early stage, the
GDF15 plasma level is elevated primarily due to promoter
activation and plays an antihypertrophic role; however, if
the haemodynamic stress persists, this counter-regulatory
role would be limited and unable to provide protection.
Indeed, the protective role played by GDF15 in the
heart is reminiscent of ANF (atrial natriuretic factor).
Although the ANF level is strongly correlated with
the degree of ventricular dysfunction and eventual
mortality, hypertensive patients carrying the ANF
gene promoter variant, which is significantly associated
with lower plasma pro-ANF, have an increased LVM
[22].
The promoter haplotype is located in the main
5 region of GDF15 gene. This region contains the
binding sites for the important transcriptional factors
that regulate the expression of GDF15, such as p53,
Sp1 and Sp3 [23–26]. Indeed, the variants − 862 and
− 474 are both located immediately adjacent to the
potential binding sites for p53, which is essential
for the induction of GDF15 expression in tumours
[24,25]. Moreover, analysis of these sequences, using
the MAPPER database (http://bio.chip.org/mapper/),
predicts that the variant − 862T introduces a novel
consensus eukaryotic transcription factor C/EBPγ
(CCAAT/enhancer-binding protein γ )-binding site [27].
C/EBPγ has been reported to function as an activator
and modulate the expression of several genes [28,29].
Therefore it is possible that any of the five variants in
the LD block could regulate the expression of GDF15
in the response to hypertension via altering the interactive
effects of the transcription factors. These need to be
elucidated in future studies.
Neither exonic variants + 157A > T nor + 2438C > G
were associated with LVM or the concentration of
GDF15 (results not shown), consistent with a previous
study which reports that variant H6D (now called
+ 2438C > G) is not associated with the subsequent risk
of cardiovascular events and GDF15 plasma level in
women [10]. It is more likely that the cardiac protective
role of GDF15 is mainly restricted to the promoter
variants which regulate gene expression, rather than those
which alter the protein structure or function.
The strengths of our present study include that the
clinical relevance of the GDF15 promoter haplotype was
explored in a community-based hypertensive population, and the functional change was confirmed in
vitro and in vivo. As our hypertensive patients reside
in the same area and are of the same nationality,
the possibility of population stratification is low [30].
Nevertheless, our findings should be confirmed in other
independent hypertensive cohorts and explored in the
general population. As only the promoter region was
sequenced, the information of additional GDF15 genetic
C
The Authors Journal compilation
C
2010 Biochemical Society
143
144
X. Wang and others
variants is insufficient. Moreover, the plasma GDF15
level should be measured in the general population and
the triangular relationship between BP, the GDF15 level
and LVH should be assessed. Considering the powerful
ability of LVH to predict morbidity and mortality, it is
necessary to perform a longitudinal follow-up study in
the future to elucidate whether there are protective effects
of GDF15 polymorphisms on heart failure.
FUNDING
This work was supported by the National High
Technology Research and Development Program of
China [863 Program, grant number 2006AA02Z477]; and
the Ministry of Science and Technology of China [grant
number 2006DFA31500] to R.H.
REFERENCES
1 Vakili, B. A., Okin, P. M. and Devereux, R. B. (2001)
Prognostic implications of left ventricular hypertrophy.
Am. Heart J. 141, 334–341
2 Manyari, D. E. (1990) Prognostic implications of
echocardiographically determined left ventricular mass in
the Framingham Heart Study. N. Engl. J. Med. 323,
1706–1707
3 Galderisi, M., Anderson, K. M., Wilson, P. W. and Levy, D.
(1991) Echocardiographic evidence for the existence of a
distinct diabetic cardiomyopathy (the Framingham Heart
Study). Am. J. Cardiol. 68, 85–89
4 Ganau, A., Devereux, R. B., Pickering, T. G., Roman, M. J.,
Schnall, P. L., Santucci, S., Spitzer, M. C. and Laragh, J. H.
(1990) Relation of left ventricular hemodynamic load and
contractile performance to left ventricular mass in
hypertension. Circulation 81, 25–36
5 Post, W. S., Larson, M. G., Myers, R. H., Galderisi, M. and
Levy, D. (1997) Heritability of left ventricular mass: the
Framingham Heart Study. Hypertension 30, 1025–1028
6 Tanase, H., Yamori, Y., Hansen, C. T. and Lovenberg, W.
(1982) Heart size in inbred strains of rats. Part 1. Genetic
determination of the development of cardiovascular
enlargement in rats. Hypertension 4, 864–872
7 Bauskin, A. R., Brown, D. A., Kuffner, T., Johnen, H.,
Luo, X. W., Hunter, M. and Breit, S. N. (2006) Role of
macrophage inhibitory cytokine-1 in tumorigenesis and
diagnosis of cancer. Cancer Res. 66, 4983–4986
8 Kempf, T., Eden, M., Strelau, J., Naguib, M., Willenbockel,
C., Tongers, J., Heineke, J., Kotlarz, D., Xu, J., Molkentin,
J. D. et al. (2006) The transforming growth factor-β
superfamily member growth-differentiation factor-15
protects the heart from ischemia/reperfusion injury.
Circ. Res. 98, 351–360
9 Xu, J., Kimball, T. R., Lorenz, J. N., Brown, D. A.,
Bauskin, A. R., Klevitsky, R., Hewett, T. E., Breit, S. N.
and Molkentin, J. D. (2006) GDF15/MIC-1 functions as a
protective and antihypertrophic factor released from the
myocardium in association with SMAD protein activation.
Circ. Res. 98, 342–350
10 Brown, D. A., Breit, S. N., Buring, J., Fairlie, W. D.,
Bauskin, A. R., Liu, T. and Ridker, P. M. (2002)
Concentration in plasma of macrophage inhibitory
cytokine-1 and risk of cardiovascular events in women: a
nested case-control study. Lancet 359, 2159–2163
11 Wollert, K. C., Kempf, T., Peter, T., Olofsson, S., James, S.,
Johnston, N., Lindahl, B., Horn-Wichmann, R., Brabant,
G., Simoons, M. L. et al. (2007) Prognostic value of
growth-differentiation factor-15 in patients with non-STelevation acute coronary syndrome. Circulation 115,
962–971
C
The Authors Journal compilation
C
2010 Biochemical Society
12 Kempf, T., von Haehling, S., Peter, T., Allhoff, T., Cicoira,
M., Doehner, W., Ponikowski, P., Filippatos, G. S.,
Rozentryt, P., Drexler, H. et al. (2007) Prognostic utility of
growth differentiation factor-15 in patients with chronic
heart failure. J. Am. Coll. Cardiol. 50, 1054–1060
13 Wollert, K. C., Kempf, T., Lagerqvist, B., Lindahl, B.,
Olofsson, S., Allhoff, T., Peter, T., Siegbahn, A., Venge, P.,
Drexler, H. and Wallentin, L. (2007) Growth
differentiation factor 15 for risk stratification and selection
of an invasive treatment strategy in non ST-elevation acute
coronary syndrome. Circulation 116, 1540–1548
14 Sahn, D. J., DeMaria, A., Kisslo, J. and Weyman, A. (1978)
Recommendations regarding quantitation in M-mode
echocardiography: results of a survey of echocardiographic
measurements. Circulation 58, 1072–1083
15 Devereux, R. B., Alonso, D. R., Lutas, E. M., Gottlieb,
G. J., Campo, E., Sachs, I. and Reichek, N. (1986)
Echocardiographic assessment of left ventricular
hypertrophy: comparison to necropsy findings.
Am. J. Cardiol. 57, 450–458
16 Lang, R. M., Bierig, M., Devereux, R. B., Flachskampf,
F. A., Foster, E., Pellikka, P. A., Picard, M. H., Roman,
M. J., Seward, J., Shanewise, J. et al. (2006)
Recommendations for chamber quantification.
Eur. J. Echocardiogr. 7, 79–108
17 Ganau, A., Devereux, R. B., Roman, M. J., de Simone, G.,
Pickering, T. G., Saba, P. S., Vargiu, P., Simongini, I. and
Laragh, J. H. (1992) Patterns of left ventricular
hypertrophy and geometric remodeling in essential
hypertension. J. Am. Coll. Cardiol. 19, 1550–1558
18 Expert Committee on the Diagnosis,
Classification of Diabetes Mellitus (1997) Report
of the Expert Committee on the Diagnosis and
Classification of Diabetes Mellitus. Diabetes Care 20,
1183–1197
19 International HapMap Consortium (2005) A
haplotype map of the human genome. Nature 437,
1299–1320
20 De Windt, L. J., Lim, H. W., Haq, S., Force, T. and
Molkentin, J. D. (2000) Calcineurin promotes protein
kinase C and c-Jun NH2 -terminal kinase activation in the
heart. Cross-talk between cardiac hypertrophic signaling
pathways. J. Biol. Chem. 275, 13571–13579
21 Kempf, T., Horn-Wichmann, R., Brabant, G., Peter, T.,
Allhoff, T., Klein, G., Drexler, H., Johnston, N., Wallentin,
L. and Wollert, K. C. (2007) Circulating concentrations of
growth-differentiation factor 15 in apparently healthy
elderly individuals and patients with chronic heart failure
as assessed by a new immunoradiometric sandwich assay.
Clin. Chem. 53, 284–291
22 Rubattu, S., Bigatti, G., Evangelista, A., Lanzani, C.,
Stanzione, R., Zagato, L., Manunta, P., Marchitti, S.,
Venturelli, V., Bianchi, G. et al. (2006) Association of atrial
natriuretic peptide and type a natriuretic peptide receptor
gene polymorphisms with left ventricular mass in human
essential hypertension. J. Am. Coll. Cardiol. 48,
499–505
23 Li, P. X., Wong, J., Ayed, A., Ngo, D., Brade, A. M.,
Arrowsmith, C., Austin, R. C. and Klamut, H. J. (2000)
Placental transforming growth factor-β is a downstream
mediator of the growth arrest and apoptotic response of
tumor cells to DNA damage and p53 overexpression.
J. Biol. Chem. 275, 20127–20135
24 Tan, M., Wang, Y., Guan, K. and Sun, Y. (2000) PTGF-β, a
type β transforming growth factor (TGF-β) superfamily
member, is a p53 target gene that inhibits tumor cell
growth via TGF-β signaling pathway. Proc. Natl. Acad.
Sci. U.S.A. 97, 109–114
25 Baek, S. J., Wilson, L. C. and Eling, T. E. (2002)
Resveratrol enhances the expression of non-steroidal
anti-inflammatory drug-activated gene (NAG-1) by
increasing the expression of p53. Carcinogenesis 23,
425–434
26 Baek, S. J., Horowitz, J. M. and Eling, T. E. (2001)
Molecular cloning and characterization of human
nonsteroidal anti-inflammatory drug-activated gene
promoter. Basal transcription is mediated by Sp1 and Sp3.
J. Biol. Chem. 276, 33384–33392
GDF15 variants and LVH in hypertension
27 Marinescu, V. D., Kohane, I. S. and Riva, A. (2005)
MAPPER: a search engine for the computational
identification of putative transcription factor binding sites
in multiple genomes. BMC Bioinformatics 6, 79
28 Omori, S. A., Smale, S., O’Shea-Greenfield, A. and Wall, R.
(1997) Differential interaction of nuclear factors with the
leukocyte-specific pp52 promoter in B and T cells.
J. Immunol. 159, 1800–1808
29 Wall, L., Destroismaisons, N., Delvoye, N. and Guy, L.-G.
(1996) CAAT/enhancer-binding proteins are involved in
β-globin gene expression and are differentially expressed in
murine erythroleukemia and K562 cells. J. Biol. Chem.
271, 16477–16484
30 Cardon, L. R. and Palmer, L. J. (2003) Population
stratification and spurious allelic association. Lancet 361,
598–604
Received 2 December 2008/29 May 2009; accepted 8 June 2009
Published as Immediate Publication 8 June 2009, doi:10.1042/CS20080637
C
The Authors Journal compilation
C
2010 Biochemical Society
145
Clinical Science (2010) 118, 137–145 (Printed in Great Britain) doi:10.1042/CS20080637
SUPPLEMENTARY ONLINE DATA
The haplotype of the growth-differentiation
factor 15 gene is associated with left ventricular
hypertrophy in human essential hypertension
Xiaojian WANG1 , Xu YANG1 , Kai SUN, Jingzhou CHEN, Xiaodong SONG,
Hu WANG, Zhe LIU, Changxin WANG, Channa ZHANG and Rutai HUI
Sino-German Laboratory for Molecular Medicine, Key Laboratory for Clinical Cardiovascular Genetics of Ministry of Education,
FuWai Cardiovascular Hospital and Cardiovascular Institute, Chinese Academy of Medical Sciences and Peking Union Medical
College, Beijing 100037, China
Table S1 Description of the GDF15 variants, primers and restriction enzymes
The mismatched base that was introduced to create restriction sites or facilitated the PCR is underlined. T m, annealing temperature.
Variant
Primer
PCR product (bp)
T m (◦ C)
Restriction enzyme
− 3148C > G
Forward, 5 -AGTGAGTCCTTGTGTCTCTTAC-3
Reverse, 5 -GCAGGCTGGTGTAGAGTC-3
Forward, 5 -GGCTGCTTGGGGGGTGGGAG-3
Reverse, 5 -GCAAGTTTCTCGGGACCCTCAGAGTTGTAC-3
Forward, 5 -GCAGAATCTTCGTCCGCAC-3
Reverse, 5 -CCAGCCCAGGTCTTCCAG-3
251
58
BsrI
209
61
BsrGI
165
62
AvaII
+ 157A > T
+ 2438C > G
Received 2 December 2008/29 May 2009; accepted 8 June 2009
Published as Immediate Publication 8 June 2009, doi:10.1042/CS20080637
1
These two authors contributed equally to the work.
Correspondence: Dr Rutai Hui (email [email protected]).
C
The Authors Journal compilation
C
2010 Biochemical Society