Hum Genet (2007) 122:3340

DOI 10.1007/s00439-007-0370-y
1 3
ORI GI NAL I NVESTI GATI ON
Sex-speciWc eVects of ACE I/D and AGT-M235T on pulse pressure:
the HyperGEN Study
Amy I. Lynch Donna K. Arnett James S. Pankow
Michael B. Miller Kari E. North John H. Eckfeldt
Steven C. Hunt Dabeeru C. Rao Luc Djouss
Received: 9 February 2007 / Accepted: 18 April 2007 / Published online: 10 May 2007
Springer-Verlag 2007
Abstract Evidence shows that an elevated pulse pressure
(PP) may lead to an increased risk of cardiovascular morbidity
and mortality. There is also evidence that PP is a sexually
dimorphic trait, and that genetic factors inXuence inter-indi-
vidual variation in PP. The aim of this project was to assess
the genotype-by-sex interaction on PP in a sample of mostly
hypertensive African American and White participants
using candidate genes involved in the reninangiotensin
aldosterone system. Subjects were participants in the Hyper-
GEN Study, including men (43%) and women (57%) over
the age of 55 years (mean age = 65). Candidate gene
polymorphisms used were ACE insertion/deletion (1,789
subjects genotyped) and AGT-M235T (1,800 subjects geno-
typed). We employed linear regression methods to assess
the genotype-by-sex interaction. For ACE, genotype-by-sex
interaction on PP was detected (P = 0.04): the D/D geno-
type predicted a 2.2 mmHg higher pulse pressure among
women, but a 1.2 mmHg lower PP among men, compared to
those with an I allele, after adjusting for age, weight,
height, ethnicity, and antihypertension medication use. A
similar interaction was found for systolic blood pressure.
The genotype-by-sex interaction was consistent across eth-
nicity. The interaction was evident among those on antihy-
pertensive medications (P = 0.05), but not among those not
taking such medications (P = 0.55). In our analysis of AGT,
no evidence of a genotype-by-sex interaction aVecting PP,
SBP, or DBP was detected. This evidence for a genotype-
by-sex interaction helps our understanding of the complex
genetic underpinnings of blood pressure phenotypes.
Introduction
There is evidence that an elevated pulse pressure, which
indicates an increase in large artery stiVness, may lead to an
increased risk of cardiovascular morbidity and mortality
(Casiglia et al. 2002; Chae et al. 1999; Glynn et al. 2000;
Mitchell et al. 1997). Pulse pressure, deWned as the diVer-
ence between the arterial systolic blood pressure (SBP) and
diastolic blood pressure (DBP), is determined both by car-
diac and vascular factors. There is a tendency for SBP to
A. I. Lynch J. H. Eckfeldt
Department of Laboratory Medicine and Pathology,
University of Minnesota, Minneapolis, MN, USA
D. K. Arnett (&)
Department of Epidemiology,
University of Alabama at Birmingham,
RPHB 220E, 1530 3rd Avenue South,
Birmingham, AL 55294-0022, USA
e-mail: arnett@ms.soph.uab.edu
J. S. Pankow M. B. Miller
Division of Epidemiology and Community Health,
University of Minnesota, Minneapolis, MN, USA
K. E. North
Division of Epidemiology,
University of North Carolina at Chapel Hill,
Chapel Hill, NC, USA
S. C. Hunt
Department of Internal Medicine,
University of Utah School of Medicine,
Salt Lake City, UT, USA
D. C. Rao
Division of Biostatistics,
Washington University School of Medicine,
St Louis, MO, USA
L. Djouss
Division of Aging, Brigham and Womens Hospital
and Harvard Medical School, Boston, MA, USA
34 Hum Genet (2007) 122:3340
1 3
rise throughout the aging process while DBP tends to rise
until about the age of 50, when there is a leveling-oV, fol-
lowed by a decrease in DBP throughout the later years of
life, which creates the observed pattern of increasing pulse
pressure with age (Franklin et al. 1997).
Interestingly, pulse pressure appears to be a sexually
dimorphic trait, with women having lower mean levels of
pulse pressure than men until their mid-50s, at which time
women have increases in mean pulse pressure beyond the
level of men that continues throughout the rest of the aging
process (Franklin et al. 1997). Not only does the pulse pres-
sure measure diVer by sex, but outcomes associated with
pulse pressure and arterial stiVness also appear to diVer by
sex (Danchin et al. 2004; Darne et al. 1989; Mazza et al.
2001).
Pulse pressure has been found to be a heritable trait, and
there is evidence that genetic factors inXuence the inter-indi-
vidual variation in pulse pressure. The heritability has been
estimated to increase with age, suggesting that with
increased age, genetic factors play a greater role in deter-
mining pulse pressure (Brandon et al. 2003). Linkage analy-
ses and genotypic association studies have been performed
to locate genes that aVect pulse pressure, yielding some
signiWcant Wndings (Atwood et al. 2001; Bielinski et al.
2005; Camp et al. 2003; Mourad et al. 2002). However, the
genetic underpinnings of blood pressure traits have proven
diYcult to reveal, likely due to the complexities of gene
environment interactions, genegene interactions and multi-
ple genes aVecting traits through diVerent pathways.
Previous research Wndings on pulse pressure suggest
that: (1) pulse pressure is an important predictor of cardio-
vascular disease and mortality, (2) genes appear to
inXuence pulse pressure, and (3) gender may be an eVect
modiWer in the relation between pulse pressure and cardio-
vascular disease and mortality. This study attempts to
integrate these concepts by addressing whether renin
angiotensinaldosterone system (RAAS) candidate geno-
types, which have been shown to aVect blood pressure
(Castellano et al. 2003; Jeunemaitre et al. 1992; Lim et al.
2002; Siani et al. 2004; Tiago et al. 2003), are associated
with pulse pressure in a gender-speciWc manner, i.e., is
there an interaction between genes and gender in their eVect
on pulse pressure?
The candidate genes included were the angiotensin I-con-
verting enzyme (ACE) insertion/deletion (I/D) polymor-
phism on chromosome 17 and the AGT-M235T angiotensi-
nogen (AGT) polymorphism on chromosome 1. The ACE D/
D genotype has been associated with an increased risk of
high blood pressure and myocardial infarction in some stud-
ies, though this association is not consistently present (Cam-
bien et al. 1992; Pereira et al. 2003; Samani et al. 1996;
Uemura et al. 2000). Likewise, the T/T genotype at the AGT-
M235T locus has been shown in previous research to confer
an increased risk of hypertension and higher DBP, but again
the association is not consistent across all populations (Cai
et al. 2004; Fornage et al. 1995; Niu et al. 1999; Pereira et al.
2003; Say et al. 2005). One study which included only men
showed the M/M genotype to be associated with higher pulse
pressure (Ortlepp et al. 2003).
The presence of gene-by-sex interactions has been
examined for a number of diVerent traits, including blood
pressure, in animals and humans (Atwood et al. 2006; North
et al. 2003; ODonnell et al. 1998). Previous studies of the
ACE I/D and AGT-M235T polymorphisms in particular
have revealed some evidence of gene by sex interactions.
For example, in some studies the ACE D/D genotype has
been associated with an increased risk of hypertension, car-
diac hypertrophy, myocardial infarction and increased DBP
among men, but not among women (ODonnell et al. 1998;
Orlowska-Baranowska et al. 2004; Petrovic et al. 2004).
Similarly, The T/T AGT-M235T genotype has been shown
to be associated with increased risk of hypertension and
increased SBP among women, but not men (Sethi et al.
2001). When a gene-by-sex interaction is revealed, it can
uncover associations that are otherwise masked, as well as
provide a more biologically relevant model of the exposure/
disease relationship.
Methods
Study population
Subjects were participants in the hypertension genetic epi-
demiology network (HyperGEN), which is one of the four
networks of the NHLBI Family blood pressure program
(FBPP). The goal of HyperGEN is to identify and charac-
terize genes involved in the development of hypertension
(Williams et al. 2000). The recruitment criteria for Hyper-
GEN required that at least two siblings in a sibship be diag-
nosed with hypertension before the age of 60 years, with
hypertension being deWned as SBP equal to or greater than
140 mmHg and/or DBP equal to or greater than 90 mmHg
measured at two separate visits, or current treatment for
hypertension. In addition to recruiting sibships, HyperGEN
also recruited a random, unrelated sample from the source
population. Participants were recruited from Weld centers in
Salt Lake City, UT; Birmingham, AL; Forsyth County, NC;
Framingham, MA; and Minneapolis, MN.
The subjects for this study included men and women (43
and 57%, respectively) over the age of 55 years with pheno-
typic and genotypic data. We limited our analysis to those
over the age of 55 since this is the age at which population
DBP tends to change in trajectory from increasing to
decreasing, and the pulse pressure in women changes with
women having a lower pulse pressure than men at younger
Hum Genet (2007) 122:3340 35
1 3
age, and a higher pulse pressure than men at older age
(Franklin et al. 1997). For the ACE polymorphism, 1,789
individuals had both genotypic and phenotypic data. There
were 1,800 participants with AGT genotypic and phenotypic
data. Of the individuals included in these analyses, 65.8%
were members of hypertensive sibships, 26.7% were
randomly selected subjects who were not members of hyper-
tensive sibships, 6.8% were parents of the hypertensive sib-
lings, and 0.6% were oVspring of the hypertensive siblings.
Of the participants in hypertensive sibships, 93% were tak-
ing antihypertension medication, while 42% of the random
participants were taking antihypertension medication. None
of the parents or oVspring of the hypertensive siblings were
on antihypertensive medication, as per the inclusion criteria.
This research was approved by local Institutional
Review Boards, and informed consent was collected from
all participants.
Phenotyping
To ensure comparable measurements among the Wve Weld
centers of the HyperGEN network, all blood pressures were
obtained using automated Dinamap devices (model 1846
SX/P, Critikon, Tampa, FL). The HyperGEN Data Coordi-
nating Center employed quality control measures such as
monitoring for reproducibility of random repeat measures,
Weld center diVerences, and bias due to digit preference of
speciWc screening personnel (Williams et al. 2000). The
blood pressure measures used were taken while the partici-
pants were in a seated position. Three successive measure-
ments were taken, and the mean value of the second two
measurements was used in this research. Pulse pressure was
calculated by taking the diVerence of SBP and DBP
(SBP DBP). Height was measured while participants
were standing shoeless, heels together against a vertically
mounted ruler. Weight was measured while participants
were standing shoeless on a balance scale. Age, gender,
ethnicity and medication use were self-reported by partici-
pants.
Genotyping
The DNA extraction and puriWcation method used a salt
precipitation method for protein removal using commercial
Puregene reagents (Gentra System, Inc., Minneapolis, MN)
following sodium dodecylsulfate cell lysis. Approximately
300 g of DNA was obtained from each participant, which
was stored as a 20 g/ml stock solution in 1.0 mM EDTA
10 mM Tris, pH 7.3. The buVer type and concentration
requested by the genotyping facilities were made from this
dilution.
The ACE I/D polymorphism genotyping method was a
modiWcation of the procedure described by Kim et al.
(2001). This procedure was modiWed by substituting 0.11
g/l BSA for 8-methoxypsoralen and the following
cycling conditions: 95C for 7 min, 12 cycles of 95C for
1 min, 63C for 1 min, 72C for 1.5 min, followed by 30
cycles of 95C for 45 s, 65 C for 45 s, 72C for 45 s, and a
Wnal extension at 72C for 5 min. The AGT-M235T poly-
morphism, which leads to a substitution of a methionine
(M) or a threonine (T) at codon 235, was analyzed with a
standard polymerase chain reaction (PCR) and the M235T
variant was measured with mass spectrophotometry.
Statistical analysis
STATA version 9.2 was used for all analyses (STATA Cor-
poration, College Station, Texas). We employed linear
regression methods to assess the genotype by sex interac-
tion, designating pulse pressure as the dependent variable,
and the candidate genotype, sex, and the (genotype sex)
product as the independent variables, adjusting for age,
weight, height, ethnicity (African American or White), and
whether the participant was taking antihypertension medi-
cation (Y/N), all of which were potential confounders.
Because HyperGEN included family members, we used
mixed models (XTMIXED in STATA) to account for the
correlation of pulse pressure that may occur within fami-
lies. We used the family identiWcation number as the
group variable.
We modeled the genotypes additively as a categorical
variable with three levels (for example with the ACE I/D
polymorphism: D/D, I/D and I/I), and also collapsed geno-
typic groups. There was no assumption of linearity for the
main eVects or the interaction eVects of the genotype when
modeled in three categoriesthe test of signiWcance for the
interaction was a 2 of freedom test for the three genotypic
category analyses, and a 1 of freedom test for the two
genotypic category analyses.
To calculate the adjusted sex- and genotype-speciWc
means for the Wgures, after the estimation command
(XTMIXED in this case), the ADJUST command in
STATA provides adjusted predictions of xbeta (the mean of
the dependent variable in linear-regression) for the model.
We estimated the mean for each level of sex and genotype
by using the BY() statement. The adjustment variables
speciWed (for example, age, weight, height, ethnicity, anti-
hypertensive medication use) were set to their mean for the
estimation.
Results
Ethnic-speciWc genotypic frequencies for each of the candi-
date genes for all participants are shown in Table 1. The
observed genotypes do not diVer from those expected for
36 Hum Genet (2007) 122:3340
1 3
HardyWeinberg equilibrium for the ACE or AGT geno-
types when calculated in an ethnicity speciWc manner.
Table 2 provides summary statistics for participant charac-
teristics, stratiWed by candidate gene and gender, for those
individuals included in the analysis for ACE, and AGT,
respectively.
ACE I/D
We Wrst modeled the genotype additively as a categorical
variable with three levels: D/D, I/D and I/I. There was a
suggestion of a genotype by sex interaction, with D/D
women having an higher mean pulse pressure than I/I or I/
D women, while D/D men had a lower mean pulse pressure
compared with I/I or I/D men (P = 0.12). There was little
diVerence between the I/I and I/D groups after adjusting for
age, height, weight, ethnicity and hypertension medication
use. Therefore, we collapsed the I/I and I/D groups. When
the ACE polymorphism was modeled in two categories,
there was a more signiWcant genotype-by-sex interaction on
the pulse pressure trait (P = 0.04). Women with the D/D
genotype had a 2.2 mmHg higher mean pulse pressure than
women with an I allele. However, for men, the direction of
the eVect of the D/D genotype was reversed: men with the
D/D genotype had a 1.2 mmHg lower mean pulse pressure
than men with an I allele (Fig. 1). The interaction eVect and
its statistical signiWcance were robust to adjustment strat-
egy, with no appreciable diVerence when modeling with
diVerent combinations of the covariates. In an unadjusted
model with no covariates present, the P value for the inter-
action term was 0.05.
To assess whether this interaction was unique to pulse
pressure, separate from the SBP and DBP measures from
which pulse pressure was derived, we repeated the analysis
twice using SBP and DBP as dependent variables with the
same adjustment strategy. The genotype-by-sex interaction
on SBP was similar to that found for pulse pressure:
Women with the D/D genotype had a 2.0 mmHg higher
mean SBP than women carrying an I allele, while D/D men
had a 2.1 mmHg lower mean SBP than men carrying an I
allele (p = 0.07, Fig. 2). There was no signiWcant interac-
tion for DBP: both men and women carrying and I allele
had a slightly higher mean DBP (0.8 and 0.2 mmHg,
respectively) than those with the D/D genotype (P = 0.60).
Since our study population included African American
(n = 614) and white (n = 1,175) participants, we tested
whether the interaction was consistent among those partici-
pants of diVerent ethnicity. In the ethnicity-speciWc models,
the eVect of the genotype-by-sex interaction was consistent
in the two groups: Women with the D/D genotype had a
higher mean pulse pressure, while men with the D/D geno-
type had a lower mean pulse pressure when compared to
Table 1 Candidate genotypic frequencies, overall and by ethnicity
[frequency (column %)]
Genotype All African
American
White
ACE D/D 584 (33%) 228 (37%) 356 (30%)
ACE I/D 883 (49%) 291 (47%) 592 (50%)
ACE I/I 322 (18%) 95 (15%) 227 (20%)
Total 1,789 614 1,175
AGT M/M 407 (23%) 13 (2%) 394 (33%)
AGT T/M 766 (43%) 166 (27%) 600 (51%)
AGT T/T 627 (35%) 434 (71%) 193 (16%)
Total 1,800 613 1,187
Table 2 Participant summary statistics for blood pressure and adjustment variable measures [mean (SD), or percentage) by candidate gene and
gender
BMI body mass index
a
HT meds on antihypertensive medication
ACE I/D, n = 1,789 AGT m235t, n = 1,800
African American, n = 614 White, n = 1,175 African American, n = 613 White, n = 1,187
Women n = 406 Men n = 208 Women n = 613 Men n = 562 Women n = 407 Men n = 206 Women n = 623 Men n = 564
Pulse pressure
(mmHg)
64.8 (19.2) 59.9 (16.9) 62.1 (17.2) 55.7 (15.8) 64.7 (19.2) 59.7 (16.9) 62.0 (17.4) 55.7 (15.8)
SBP (mmHg) 136.1 (24.7) 136.4 (22.0) 126.8 (22.0) 126.8 (20.2) 136.0 (24.6) 136.3 (22.1) 126.7 (22.3) 126.8 (20.3)
DBP (mmHg) 71.3 (10.9) 76.6 (11.9) 64.7 (10.0) 71.0 (10.7) 71.3 (10.8) 76.6 (12.0) 64.7 (10.0) 71.1 (10.7)
Age (years) 64.3 (6.4) 64.4 (6.5) 65.4 (6.9) 65.1 (6.0) 64.3 (6.4) 64.3 (6.5) 65.4 (6.8) 65.1 (6.0)
Height (m) 1.62 (0.06) 1.74 (0.07) 1.61 (0.06) 1.74 (0.07) 1.62 (0.06) 1.74 (0.07) 1.61 (0.06) 1.75 (0.07)
Weight (kg) 85.0 (18.4) 89.8 (21.7) 75.7 (16.8) 90.1 (14.9) 85.0 (18.4) 89.7 (21.8) 75.7 (16.9) 90.1 (14.9)
BMI (kg/m
2
) 32.5 (6.8) 29.4 (6.4) 29.3 (6.2) 29.5 (4.4) 32.5 (6.8) 29.4 (6.5) 29.3 (6.2) 29.5 (4.4)
HT meds
a

[freq, (%)]
310 (76.4) 152 (73.1) 424 (69.2) 408 (72.6) 313 (76.9) 152 (73.8) 430 (69.0) 410 (72.7)
Hum Genet (2007) 122:3340 37
1 3
those with an I allele (P = 0.22 for African Americans,
P = 0.09 for Whites).
While we adjusted for antihypertensive medication use
in the multivariate models, we also assessed whether the
interaction was consistent among those on antihypertensive
medication (n = 1,294) and those not on antihypertensive
medication (n = 495). The stratiWed analyses revealed that
there was evidence of the ACE genotype-by-sex interaction
present among those on medication (P = 0.05) (Fig. 3), but
not in the unmedicated group (P = 0.55). It appears that
SBP is largely contributing to the pulse pressure interaction
among those taking antihypertensive medications: when we
evaluate the SBP measures separately among those on anti-
hypertensive medications versus those not on antihyperten-
sive medications, for those on medications, men with the
D/D genotype had a lower mean SBP than men with an I
allele (129.6 mmHg vs. 132.8 mmHg, respectively), while
women with the D/D genotype had higher mean SBP than
women with an I allele (131.5 mmHg vs. 129.8 mmHg,
respectively) (P = 0.06, Fig. 3). There was no similar inter-
action aVecting SBP among those participants not on medi-
cation (P = 0.70). For DBP, neither the medicated nor the
unmedicated subgroup showed an ACE genotype-by-sex
interaction (P = 0.38 and P = 0.73, respectively).
AGT M235T
The AGT variant showed no evidence of a genotype-by-sex
interaction on pulse pressure in our analyses. For the addi-
tive interaction model, men had a mean pulse pressure of
58.6, 58.4 and 58.1 mmHg for the T/T, M/T and M/M
genotypes, respectively. For women, the means were 62.2,
61.2 and 62.6 mmHg, respectively. The P value for the test
of interaction was 0.70. We also looked at SBP and DBP as
Fig. 1 Pulse pressure, ACE I/D-by-sex interaction, 3 and 2 genotypic
categories, adjusted for age, height, weight, ethnicity and antihyperten-
sive medication use
All Participants (n=1789)
Pulse Pressure: ACE by sex interaction (p=0.120)
59.8
58.6
57.7
61.7
61.0
63.4
57
61
65
I/I I/D D/D
ACE genotype
r
u
s
s
e
r
P

e
s
l
u
P

n
a
e
M

d
e
t
s
u
j
d
A
e
)
g
H
m
m
(
Women
Men
All Participants (n=1789)
Pulse Pressure: ACE by sex interaction (p=0.040)
58.9
57.7
61.2
63.4
57
61
65
I/I or I/D D/D
ACE genotype
r
u
s
s
e
r
P

e
s
l
u
P

n
a
e
M

d
e
t
s
u
j
d
A
e
)
g
H
m
m
(
Women
Men
Fig. 2 SBP, ACE I/D-by-sex interaction, adjusted for age, height,
weight, ethnicity and antihypertensive medication use
All Participants (n=1789)
SBP: ACE by sex interaction (p=0.068)
131.8
129.7 128.4
130.4
128
131
134
I/I or I/D D/D
ACE genotype
g
H
m
m
(

P
B
S

n
a
e
M

d
e
t
s
u
j
d
A
)
Women
Men
Fig. 3 Participants on antihypertensive medication, pulse pressure
and SBP, ACE I/D-by-sex interaction, adjusted for age, height, weight
and ethnicity
Participants on anti-hypertensive medication (n=1294)
Pulse Pressure: ACE by sex interaction (p=0.054)
59.7
57.8
62.6
64.5
57
61
65
I/I or I/D D/D
ACE genotype
A
d
j
u
s
t
e
d

M
e
a
n

P
u
l
s
e

P
r
e
s
s
u
r
e
A
d
j
u
s
t
e
d

M
e
a
n

S
B
P

(
m
m
H
g
)
(
m
m
H
g
)
Women
Men
Participants on anti-hypertensive medication (n=1294)
SBP: ACE by sex interaction (p=0.056)
132.8
129.6
129.8
131.5
128
131
134
I/I or I/D D/D
ACE genotype
Women
Men
38 Hum Genet (2007) 122:3340
1 3
dependent variables, and again there was no evidence of a
genotype-by-sex interaction (P = 0.53 for SBP, P = 0.41
for DBP). Because the allele frequencies between African
American and White participants were dramatically diVer-
ent at the AGT-M235T locus (see Table 1), we performed
ethnicity speciWc analyses for this candidate genotype as
well. There was no evidence of a genotype-by-sex inter-
action on either pulse pressure, SBP or DBP in African
American or White participants.
Discussion
In our analysis of the AGT-M235T polymorphism, no
evidence of a genotype-by-sex interaction aVecting pulse
pressure, SBP, or DBP was detected, whether all partici-
pants were included or when the data were stratiWed by
ethnicity. However, for the ACE I/D candidate polymor-
phism, an interaction on pulse pressure was detected,
which was consistent across ethnicity, with the D/D geno-
type predicting a 2.2 mmHg higher pulse pressure among
women, but a 1.2 mmHg lower pulse pressure among men,
compared to the I/D or I/I genotype (P = 0.04). We
observed a similar interaction for the SBP measure.
Although, pulse pressure is considered to be a marker of
arterial stiVness, and may provide information about car-
diovascular disease risk above the SBP and DBP measures
taken without regard to their diVerence, in this study the
interaction is captured by the SBP measure alone, and does
not seem to be a function of a dynamic interplay between
SBP and DBP. We are not aware of any previously
published research showing a gene-by-sex interaction on
pulse pressure for either the AGT-M235T or the ACE I/D
polymorphisms.
The interaction was evident among those on antihyper-
tensive medications (P = 0.05), but not among those not
taking such medications (P = 0.55) in stratiWed analyses.
What appears to be driving the interaction among the medi-
cated participants is that the D/D men have a substantially
lower mean SBP than men with an I allele (129.6 mmHg
vs. 132.8 mmHg), while the D/D women have a higher
mean SBP than women with an I allele (131.5 mmHg vs.
129.8 mmHg). It is possible that the SBP of men with a D/
D genotype is more responsive to blood pressure lowering
medication than for men with the I/I or I/D genotype, while
the opposite may be true for women. This is an interesting
Wnding, since it corroborates a report by Schwartz et al.
(2002) who found a statistically signiWcant gender-speciWc
eVect of the ACE I/D polymorphism on blood pressure
response to a thiazide diuretic: women with the D/D geno-
type experienced the smallest overall blood pressure lower-
ing response for both SBP and DBP to the medication
(hydrochlorothiazide, 25 mg per day) when compared with
I/I or I/D women, while among men, the D/D genotype
resulted in the greatest response.
It is possible that sex hormones work to alter the activity
of the reninangiotensin system, which may account for the
possible genesex interaction on blood pressure traits. In a
review article, Sandberg and Ji (2003) conclude that while
there is much evidence that sex steroids act to alter the
activity of the reninangiotensin system, much is still
unknown about the molecular mechanisms through which
estrogen and androgen alter the system. It is also possible
that gender is a marker for some unmeasured variable that
accounts for any observed interaction.
These results must be interpreted with caution for several
reasons: With limited sample sizes, the power to detect inter-
actions is also limited, as this study was not a priori powered
to detect such interactions. The majority of the participants
were recruited from hypertensive sibships, and the applica-
bility to a normotensive population is uncertain. Another
concern is the eVect of anti-hypertensive medications on the
blood pressure measures, which may have caused a distor-
tion of the true pulse pressure phenotype, blurring real
inter-individual diVerences. Data regarding the eVect of
antihypertensive treatment on pulse pressure are conXicting.
While most studies show a decrease in pulse pressure with
antihypertensive use, this is not always true for all age
groups, for all medications, or for all ethnic groups
(Cushman et al. 2001; Safar et al. 2000). For this reason, in
our medication-adjusted models we used a dichotomous
variable. Because of concern over the eVect of antihypertensive
medication on pulse pressure, we also reported results strati-
Wed by antihypertensive use. Among the participants in the
present study, 73% overall were taking at least one antihy-
pertensive medication. These Wndings may therefore be
suggestive of a genotype-by-sex interaction on post-medica-
tion pulse pressure level, rather than general pulse pressure
level. We must also note that the ACE and AGT genes con-
tain 26 and 5 exons, respectively, and therefore the polymor-
phisms used here represent only a small amount of the
variation present in these candidate genes. Finally, we recog-
nize that our study is exploratory in nature; we performed
multiple statistical tests and did not adjust our P values for
multiple testing. Therefore, we cannot rule out the possibility
of a false positive Wnding of a genotype-by-sex interaction
on pulse pressure. Since we performed a total of 21 statistical
tests, if one prefers to be conservative, a Bonferroni correc-
tion for multiple testing would yield a threshold for signiW-
cance of P = 0.002 (0.05/21), in which case none of the
Wndings reported here would reach statistical signiWcance. It
may be argued, however, that since the 21 tests were not all
independent tests (for example, the outcomes of pulse pres-
sure, SBP and DBP are correlated), the Bonferroni correction
may be too restrictive. The limitations of this research not-
withstanding, this Wnding of a genotype-by-sex interaction
Hum Genet (2007) 122:3340 39
1 3
helps our understanding of the genetic underpinnings of a
complex disease such as arterial stiVness. In future research,
attempts to account for gene-by-sex, and more broadly gene-
by-environment and gene-by-gene interactions should be
encouraged. This will guide us to a better description of the
etiologic disease model.
Acknowledgments Supported in part by grants HL55673, HL54471,
HL 54472, HL54473, HL54495, HL54496, HL54497, HL54509 HL
54515 and HL007972 from the National Heart, Lung and Blood
Institute, Bethesda, MD, and grant M10RR0047-34 (GCRC) from the
National Institutes of Health, Bethesda, MD.
References
Atwood LD, Samollow PB, Hixson JE, Stern MP, MacCluer JW
(2001) Genome-wide linkage analysis of pulse pressure in Mexi-
can Americans. Hypertension 37:425428
Atwood LD, Heard-Costa NL, Fox CS, Jaquish CE, Cupples LA
(2006) Sex and age-speciWc eVects of chromosomal regions
linked to body mass index in the Framingham Study. BMC Genet
7:7
Bielinski SJ, Lynch AI, Miller MB, Weder A, Cooper R, Oberman A,
Chen YD, Turner ST, Fornage M, Province M, Arnett DK (2005)
Genome-wide linkage analysis for loci aVecting pulse pressure:
the family blood pressure program. Hypertension 46:12861293
Brandon DT, WhitWeld KE, Sollers JJ 3rd, Wiggins SA, West SG, Vo-
gler GP, McClearn GE, Thayer JF (2003) Genetic and environ-
mental inXuences on blood pressure and pulse pressure among
adult African Americans. Ethn Dis 13:193199
Cai SY, Yu F, Shi YP (2004) [Association of angiotensinogen gene
M235T variant with essential hypertension]. Zhejiang Da Xue
Xue Bao Yi Xue Ban 33:151154
Cambien F, Poirier O, Lecerf L, Evans A, Cambou JP, Arveiler D, Luc
G, Bard JM, Bara L, Ricard S et al. (1992) Deletion
polymorphism in the gene for angiotensin-converting enzyme is a
potent risk factor for myocardial infarction. Nature 359:641644
Camp NJ, Hopkins PN, Hasstedt SJ, Coon H, Malhotra A, Cawthon
RM, Hunt SC (2003) Genome-wide multipoint parametric link-
age analysis of pulse pressure in large, extended utah pedigrees.
Hypertension 42:322328
Casiglia E, TikhonoV V, Mazza A, Piccoli A, Pessina AC (2002) Pulse
pressure and coronary mortality in elderly men and women from
general population. J Hum Hypertens 16:611620
Castellano M, Glorioso N, Cusi D, Sarzani R, Fabris B, Opocher G,
Zoccali C, Golin R, Veglio F, Volpe M, Mantero F, Fallo F, Rossi
GP, Barlassina C, Tizzoni L, Filigheddu F, Giacche M, Rossi F
(2003) Genetic polymorphism of the renin-angiotensin-aldoste-
rone system and arterial hypertension in the Italian population:
the GENIPER Project. J Hypertens 21:18531860
Chae CU, PfeVer MA, Glynn RJ, Mitchell GF, Taylor JO, Hennekens
CH (1999) Increased pulse pressure and risk of heart failure in the
elderly. Jama 281:634639
Cushman WC, Materson BJ, Williams DW, Reda DJ (2001) Pulse
pressure changes with six classes of antihypertensive agents in a
randomized, controlled trial. Hypertension 38:953957
Danchin N, Benetos A, Lopez-Sublet M, Demicheli T, Safar M, Mou-
rad JJ (2004) Aortic pulse pressure is related to the presence and
extent of coronary artery disease in men undergoing diagnostic
coronary angiography: a multicenter study. Am J Hypertens
17:129133
Darne B, Girerd X, Safar M, Cambien F, Guize L (1989) Pulsatile ver-
sus steady component of blood pressure: a cross-sectional analy-
sis and a prospective analysis on cardiovascular mortality.
Hypertension 13:392400
Fornage M, Turner ST, Sing CF, Boerwinkle E (1995) Variation at the
M235T locus of the angiotensinogen gene and essential hyperten-
sion: a population-based case-control study from Rochester, Min-
nesota. Hum Genet 96:295300
Franklin SS, Gustin Wt, Wong ND, Larson MG, Weber MA, Kannel
WB, Levy D (1997) Hemodynamic patterns of age-related chang-
es in blood pressure. The Framingham Heart Study. Circulation
96:308315
List of HyperGEN participating institutions and principal staV
Network Center/University of Utah Field Center Steven C. Hunt, Roger R. Williams (deceased), Hilary Coon,
Paul N. Hopkins, Janet Hood, Lily Wu, Jan Skuppin
University of Alabama at Birmingham Field Center Albert Oberman, Cora E. Lewis, Michael T. Weaver,
Phillip Johnson, Susan Walker, Christie Oden
Boston University/Framingham Field Center R. Curtis Ellison, Richard H. Myers, Yuqing Zhang,
Luc Djouss, Jemma B. Wilk, Greta Lee Splansky
University of Minnesota Field Center Donna Arnett, Aaron R. Folsom, Mike Miller, Jim Pankow,
Gregory Feitl, Barb Lux
University of North Carolina Field Center Gerardo Heiss, Barry I. Freedman, Kari North,
Kathryn Rose, Amy Haire
Data Coordinating Center, Washington University D. C. Rao, Michael A. Province, Treva Rice, Charles Gu,
Ingrid B. Borecki, Derek Morgan, Karen Schwander,
Aldi Kraja, Stephen Mandel, Matthew Brown
Central Biochemistry Lab, University of Minnesota John H. Eckfeldt, Catherine Leiendecker-Foster,
Ronald C. McGlennen, Greg Rynders, Michael Y. Tsai, Jean Bucksa
Molecular Genetics Laboratory, University of Utah Mark Leppert, Steven C. Hunt, Jean-Marc Lalouel, Robert Weiss
Echocardiography Reading Center, Cornell Richard B. Devereux, Jonathan N. Bella
National Heart, Lung & Blood Institute Susan E. Old, Millicent Higgins (retired), Cashell Jaquish,
Martha Lundberg, Mariana Gerschenson
40 Hum Genet (2007) 122:3340
1 3
Glynn RJ, Chae CU, Guralnik JM, Taylor JO, Hennekens CH (2000)
Pulse pressure and mortality in older people. Arch Intern Med
160:27652772
Jeunemaitre X, Soubrier F, Kotelevtsev YV, Lifton RP, Williams CS,
Charru A, Hunt SC, Hopkins PN, Williams RR, Lalouel JM et al.
(1992) Molecular basis of human hypertension: role of angioten-
sinogen. Cell 71:169180
Kim DS, Choi SI, Lee HS, Park JK, Yi HK (2001) Determination of
human angiotensin converting enzyme (ACE) gene polymor-
phisms in erectile dysfunction: frequency diVerences of ACE
gene polymorphisms according to the method of analysis. Clin
Chem Lab Med 39:1114
Lim PO, Macdonald TM, Holloway C, Friel E, Anderson NH, Dow E,
Jung RT, Davies E, Fraser R, Connell JM (2002) Variation at the
aldosterone synthase (CYP11B2) locus contributes to hyperten-
sion in subjects with a raised aldosterone-to-renin ratio. J Clin
Endocrinol Metab 87:43984402
Mazza A, Pessina AC, Gianluca P, TikhonoV V, Pavei A, Casiglia E
(2001) Pulse pressure: an independent predictor of coronary and
stroke mortality in elderly females from the general population.
Blood Press 10:205211
Mitchell GF, Moye LA, Braunwald E, Rouleau JL, Bernstein V, Gelt-
man EM, Flaker GC, PfeVer MA (1997) Sphygmomanometrical-
ly determined pulse pressure is a powerful independent predictor
of recurrent events after myocardial infarction in patients with im-
paired left ventricular function. SAVE investigators. Survival and
ventricular enlargement. Circulation 96:42544260
Mourad JJ, Ducailar G, Rudnicki A, Lajemi M, Mimran A, Safar ME
(2002) Age-related increase of pulse pressure and gene polymor-
phisms in essential hypertension: a preliminary study. J Renin
Angiotensin Aldosterone Syst 3:109115
Niu T, Chen C, Yang J, Wang B, Wang Z, Schork N, Fang Z, Xu X
(1999) Blood pressure and the T174M and M235T polymor-
phisms of the angiotensinogen gene. Ann Epidemiol 9:245253
North KE, Martin LJ, Dyer T, Comuzzie AG, Williams JT (2003) HDL
cholesterol in females in the Framingham Heart Study is linked to
a region of chromosome 2q. BMC Genet 4(Suppl 1):S98
ODonnell CJ, Lindpaintner K, Larson MG, Rao VS, Ordovas JM,
Schaefer EJ, Myers RH, Levy D (1998) Evidence for association
and genetic linkage of the angiotensin-converting enzyme locus
with hypertension and blood pressure in men but not women in
the Framingham Heart Study. Circulation 97:17661772
Orlowska-Baranowska E, Placha G, Gaciong Z, Baranowski R, Zak-
rzewski D, Michalek P, HoVman P, Rawczynska-Englert I (2004)
InXuence of ACE I/D genotypes on left ventricular hypertrophy in
aortic stenosis: gender-related diVerences. J Heart Valve Dis
13:574581
Ortlepp JR, Metrikat J, Mevissen V, Schmitz F, Albrecht M, Maya-
Pelzer P, Hanrath P, Zerres K, HoVmann R (2003) Relation
between the angiotensinogen (AGT) M235T gene polymorphism
and blood pressure in a large, homogeneous study population. J
Hum Hypertens 17:555559
Pereira AC, Mota GF, Cunha RS, HerbenhoV FL, Mill JG, Krieger JE
(2003) Angiotensinogen 235T allele dosage is associated with
blood pressure phenotypes. Hypertension 41:2530
Petrovic D, Bregar D, Guzic-Salobir B, Skof E, Span M, Terzic R, Pet-
rovic MG, Keber I, Letonja M, Zorc M, Podbregar M, Peterlin B
(2004) Sex diVerence in the eVect of ACE-DD genotype on the
risk of premature myocardial infarction. Angiology 55:155158
Safar ME, Rudnichi A, Asmar R (2000) Drug treatment of hyperten-
sion: the reduction of pulse pressure does not necessarily parallel
that of systolic and diastolic blood pressure. J Hypertens
18:11591163
Samani NJ, Thompson JR, OToole L, Channer K, Woods KL (1996)
A meta-analysis of the association of the deletion allele of the
angiotensin-converting enzyme gene with myocardial infarction.
Circulation 94:708712
Sandberg K, Ji H (2003) Sex and the renin angiotensin system: impli-
cations for gender diVerences in the progression of kidney dis-
ease. Adv Ren Replace Ther 10:1523
Say YH, Ling KH, Duraisamy G, Isaac S, Rosli R (2005) Angiotensi-
nogen M235T gene variants and its association with essential
hypertension and plasma renin activity in Malaysian subjects: a
case control study. BMC Cardiovasc Disord 5:7
Schwartz GL, Turner ST, Chapman AB, Boerwinkle E (2002) Interact-
ing eVects of gender and genotype on blood pressure response to
hydrochlorothiazide. Kidney Int 62:17181723
Sethi AA, Nordestgaard BG, Agerholm-Larsen B, Frandsen E, Jensen
G, Tybjaerg-Hansen A (2001) Angiotensinogen polymorphisms
and elevated blood pressure in the general population: the Copen-
hagen City Heart Study. Hypertension 37:875881
Siani A, Russo P, Paolo Cappuccio F, Iacone R, Venezia A, Russo O,
Barba G, Iacoviello L, Strazzullo P (2004) Combination of renin-
angiotensin system polymorphisms is associated with altered re-
nal sodium handling and hypertension. Hypertension 43:598602
Tiago AD, Badenhorst D, Nkeh B, Candy GP, Brooksbank R, Sareli P,
Libhaber E, Samani NJ, Woodiwiss AJ, Norton GR (2003) Im-
pact of renin-angiotensin-aldosterone system gene variants on the
severity of hypertension in patients with newly diagnosed hyper-
tension. Am J Hypertens 16:10061010
Uemura K, Nakura J, Kohara K, Miki T (2000) Association of ACE I/
D polymorphism with cardiovascular risk factors. Hum Genet
107:239242
Williams RR, Rao DC, Ellison RC, Arnett DK, Heiss G, Oberman A,
Eckfeldt JH, Leppert MF, Province MA, Mockrin SC, Hunt SC
(2000) NHLBI family blood pressure program: methodology and
recruitment in the HyperGEN network. Hypertension genetic epi-
demiology network. Ann Epidemiol 10:389400