Original Article
Common Variants of the Endothelial Nitric Oxide
Synthase Gene and the Risk of Coronary Heart Disease
Among U.S. Diabetic Men
Cuilin Zhang,1 Ruy Lopez-Ridaura,1 David J. Hunter,1,2,3 Nader Rifai,4 and Frank B. Hu1,2,3
Endothelial nitric oxide synthase (eNOS) gene represents
a promising candidate gene for coronary heart disease
(CHD) because of its impact on eNOS activity. We system-
atically examined the associations of eight variants of the
eNOS gene (two potentially functional variants [ⴚ786T>C
and Glu298Asp] and six tagging single nucleotide polymor-
phisms) with CHD risk in a large cohort of diabetic pa-
tients. Among 861 diabetic men (>97% Caucasian) from
the Health Professionals Follow-Up Study, 220 developed
CHD, and 641 men without cardiovascular disease were
used as control subjects. Genotype distributions of
ⴚ786T>C and Glu298Asp polymorphisms were not signifi-
cantly different between case and control subjects. CHD
risk was significantly higher among men with the variant
allele at the rs1541861 locus (intron 8 A/C) than men
without it (adjusted odds ratio 1.5 [95% confidence inter-
val 1.1–2.1]). Moreover, among control subjects, plasma
soluble vascular cell adhesion molecule concentrations
were significantly higher among carriers of this allele (P
0.019) and carriers of the variant allele of the ⴚ786T>C (P
0.010), or the Glu298Asp polymorphism (P 0.002), com-
pared with noncarriers. In conclusion, our data suggested
that ⴚ786T>C, Glu298Asp, and an intron 8 polymorphism
of the eNOS gene are potentially involved in the athero-
genic pathway among U.S. diabetic men. Diabetes 55:
2140 –2147, 2006
From the 1Department of Nutrition, Harvard School of Public Health, Boston,
Massachusetts; the 2Channing Laboratory, Department of Medicine, Brigham
and Women’s Hospital and Harvard Medical School, Boston, Massachusetts;
the 3Department of Epidemiology, Harvard School of Public Health, Boston,
Massachusetts; and the 4Department of Laboratory Medicine, Children’s
Hospital and Harvard Medical School, Boston, Massachusetts.
Address correspondence and reprint requests to Cuilin Zhang MD, PhD,
Department of Nutrition, Harvard School of Public Health, 665 Huntington
Ave., Boston, MA 02115. E-mail: nhcui@channing.harvard.edu or nhbfh@
channing.harvard.edu.
Received for publication 26 November 2005 and accepted in revised form 3
April 2006.
CABG, coronary artery bypass grafting; CHD, coronary heart disease; CRP,
C-reactive protein; eNOS, endothelial nitric oxide synthase; ICAM, intercellu-
lar cell adhesion molecule; NCBI, National Center for Biotechnology Informa-
tion; sICAM, soluble intercellular cell adhesion molecule; SNP, single
nucleotide polymorphism; sVCAM, soluble vascular cell adhesion molecule;
VCAM, vascular cell adhesion molecule.
DOI: 10.2337/db05-1535
© 2006 by the American Diabetes Association.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked “advertisement” in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
2140
E
ndothelial dysfunction and vascular inflamma-
tion contribute substantially to the accelerated
atherogenesis associated with type 2 diabetes.
Endothelial-derived nitric oxide (NO) is the pri-
mary compound responsible for vasodilation in arteries
(1) and maintenance of normal vascular homeostasis. It
can inhibit platelet aggregation and modulate leukocyte
endothelium interactions by inhibiting cell adhesion mol-
ecule expression (i.e., intercellular cell adhesion molecule
[ICAM] and vascular cell adhesion molecule [VCAM]),
reducing monocyte adherence (1), and inhibiting the pro-
liferation of smooth muscle cells (2). Endothelial-derived
NO is synthesized from L-arginine by NO synthase encoded
by endothelial NO synthase (eNOS or NOS3) gene,
mapped on chromosome 7q36 (3). The eNOS gene knock-
out mouse was demonstrated to have an increased sus-
ceptibility to develop atherosclerosis independent of
blood pressure changes (4). Both ex vivo and in vivo eNOS
gene transfer to atherosclerotic arteries can improve acety-
choline-induced endothelium-dependent vasodilation (5–10).
The association between several polymorphisms of the
eNOS gene and coronary heart disease (CHD) risk has
been studied recently. In particular, the ⫺786T⬎C poly-
morphism in the 5⬘-flanking region and the Glu298Asp
polymorphism in exon 7 have received the most attention;
these two polymorphisms have been associated with al-
terations in eNOS activity in experimental studies (11–14)
and with the existence, severity, and progression of CHD
in some, although not all, epidemiological studies (12,15–
17). The association of these potentially functional poly-
morphisms with CHD risk has not been well studied
among diabetic patients. Furthermore, we are unaware of
studies that have systematically investigated the associa-
tion of variants of the eNOS gene with CHD risk. In the
present study, we examined associations of the two po-
tentially functional single nucleotide polymorphisms
(SNPs) and tagging SNPs of the eNOS gene with CHD risk
among diabetic men. The associations of these eNOS SNPs
with circulating levels of soluble VCAM-1 (sVCAM-1) and
ICAM-1 were also examined.
RESEARCH DESIGN AND METHODS
The Health Professionals Follow-Up Study is a prospective cohort study of
51,529 American male health professionals aged 40 –75 years at study initia-
tion in 1986. This cohort has been and continues to be followed through
biennial mailed questionnaires focusing on various lifestyle factors and health
outcomes. In addition, between 1993 and 1999 (⬎95% of them between 1993
and 1995), 18,159 study participants provided blood samples by overnight
courier. Among them, 999 (⬎97% are Caucasian) had confirmed type 2
DIABETES, VOL. 55, JULY 2006

diabetes (at baseline or during follow up: 1986 –2000). From this group of
diabetic men, we excluded those with reported CHD or stroke at baseline,
those who developed stroke or other noncoronary cardiovascular disease
during follow-up, and those who developed CHD before diabetes was diag-
nosed (total exclusion n ⫽ 138). Among the remaining 861 diabetic men, 220
who developed CHD (i.e., nonfatal myocardial infarction n ⫽ 66, fatal CHD
n ⫽ 18, and coronary artery bypass grafting [CABG] n ⫽ 136) composed the
case group, and 641 diabetic men who remained free of CHD and stroke
composed the control group.
Diagnosis of type 2 diabetes. A diagnosis of diabetes was established when
at least one of the following criteria was reported on a supplementary
questionnaire sent to all men who reported a diagnosis of diabetes on any
biennial follow-up questionnaire: 1) one or more classic symptoms (excessive
thirst, polyuria, weight loss, hunger, or coma) plus a fasting plasma glucose
concentration of ⱖ140 mg/dl or a random plasma glucose concentration of
ⱖ200 mg/dl; 2) at least two elevated plasma glucose concentrations on
different occasions (fasting ⱖ140 mg/dl and/or random ⱖ200 mg/dl and/or
ⱖ200 mg/dl after 2 h or more on oral glucose tolerance testing) in the absence
of symptoms; or 3) treatment with hypoglycemic medication (insulin or oral
hypoglycemic agents). We used the National Diabetes Data Group criteria (18)
to define diabetes because the majority of our case subjects were diagnosed
before the release of the American Diabetes Association criteria (19). Men
with type 1 diabetes were excluded. A validation study in a sub-sample of the
Health Professionals Follow-Up Study demonstrated that our supplementary
questionnaire is reliable in confirming diabetes diagnosis (20).
Diagnosis of cardiovascular end points. The cardiovascular end points for
this analysis comprised fatal CHD, nonfatal myocardial infarction, CABG, and
percutaneous transluminal coronary angioplasty occurring between the return
of the 1986 questionnaire and 31 January 2000. Nonfatal myocardial infarction
was confirmed by a review of medical records based on World Health
Organization criteria that included characteristic symptoms with either typical
electrocardiographic changes or elevations of cardiac enzymes (21). Probable
cases of myocardial infarction (no available records but confirmed by
hospitalization and information from telephone interview/letter) were also
included in the analysis. Confirmation of CABG or percutaneous transluminal
coronary angioplasty was based on self-report only; hospital records obtained
for a sample of 102 men in the Health Professionals Follow-Up Study
confirmed the procedure for 96% (22). Deaths were reported by next of kin,
through the postal system, and through records of the National Death Index.
Using all sources combined, follow-up for deaths was ⬎98% (23). Fatal CHD
was confirmed by reviewing medical records or autopsy reports with the
permission of the next of kin. Physicians who reviewed the records had no
knowledge of genotype or the self-reported risk factor status of participants.
Sudden deaths (i.e., deaths within 1 h of symptom onset in men without known
disease that could explain death) were included in the fatal CHD category.
Ascertainment of characteristics of participants. At baseline, partici-
pants were asked to report their height and current body weight; weight was
then updated during biennial follow-up. Self-reports of body weight have been
shown to be highly correlated with technician-measured weights (r ⫽ 0.97) in
Health Professionals Follow-Up Study participants (24). Participants also pro-
vided information biennially on their cigarette smoking, aspirin use, and physical
activity. History of high blood pressure and family history of CHD were reported
in 1986. Alcohol intake was estimated with a dietary questionnaire in 1986.
Assessment of genotypes and biochemical variables. Resequencing infor-
mation from the Environmental Genome Project of the National Institute of
Environmental Health Sciences at the University of Washington was used to
generate haplotypes for the selection of tagging SNPs (25). There were 23
individuals reporting European descent with 119 genotyped SNPs in the
database. After excluding SNPs with a minor allele frequency of ⬍5%,
haplotypes were reconstructed with PHASE (26). We selected six tagging
SNPs based on both haplotype tagging SNPs algorithm with BEST (27) and
linkage disequilibrium pairwise algorithm (28). They are at chromosome
positions 150127045 (rs1800783), 150130092 (rs1800781), 150134981
(rs1541861), 150140689 (rs2566511), 150144031 (rs753482), and 150151284
(rs3800787) reported on the National Center for Biotechnology Information
(NCBI) website (geneID 4846) (Appendix 1). Two other SNPs (⫺786T⬎C,
rs2070744; and Glu298Asp, rs1799983), previously described as being associ-
ated with CHD risk, were also examined.
The collection and treatment of blood samples have been previously
described (29). DNA was extracted from the buffy coat fraction of centrifuged
blood using the QIAmp Blood kit (Qiagen, Chatsworth, CA). Case and control
subjects were genotyped using the TaqMan system (Applied Biosystems,
Foster City, CA). Primer and probe sequences are available from the authors
on request. Replicate quality control samples (10%) were included and
genotyped with ⱖ99% concordance. Genotype data were available for 828
(95% of case subjects and 97% of control subjects) of 861 study participants in
the present study.
DIABETES, VOL. 55, JULY 2006
C. ZHANG AND ASSOCIATES
Biochemical markers were measured among diabetic men who developed
CHD after their blood samples were collected in 1993–1994 (n ⫽ 120) and
control subjects (n ⫽ 641). All markers were assayed in the laboratory of Dr.
Nader Rifai (The Children’s Hospital, Boston, MA), which is certified by the
NHLBI/CDC Lipid Standardization program. Circulating levels of soluble
ICAMs (sICAMs) and sVCAMs were measured by enzyme-linked immunosor-
bent assay from R&D Systems (Minneapolis, MN). The day-to-day variability
of the assays at concentrations of 64.2, 117, 290, and 453 ng/ml sICAM and at
concentrations of 9.8, 24.9, and 49.6 ng/ml sVCAM was 10.1, 7.4, 6.0, and 6.1%
and 10.2, 8.5, and 8.9%, respectively. C-reactive protein (CRP) was measured
via an immunoturbidimetric assay using reagents and calibrators from Denka
Seiken (Niigata, Japan). The day-to-day variability of the assay at concentra-
tions of 0.91, 3.07, and 13.38 mg/l was 2.8, 1.6, and 1.1%, respectively.
Fibrinogen was measured with an immunoturbidimetric assay using reagents
and calibrators from Kamiya Biomedical (Seattle, WA). The day-to-day vari-
ability of the assay at concentrations of 4.92, 9.51, and 16.29 ␮g/l was 0.9, 1.1,
and 1.5%, respectively.
Statistical analysis. Frequency distributions of characteristics of study
participants were examined according to case-control status. Student’s t tests
were used for comparisons of means. ␹2 Tests were used to assess whether
genotype distributions were deviated from Hardy-Weinberg equilibrium and to
determine differences in genotype frequencies and proportions of categorical
variables between case and control subjects.
Unconditional logistic regression was used to calculate odds ratios (ORs)
and 95% CIs for the association between SNP and CHD risk adjusted for risk
factors for CHD. Adjustment for baseline variables (i.e., age, smoking status
[never, past, or current smoker], and alcohol consumption [nondrinker or
⬍5.0, ⬍10.0, or ⬎10.0 g/day]) and duration of diabetes changed the OR
slightly, but we kept these variables in the final model because of their
recognition as risk factors for CHD and interpretable variables that may
account for the heterogeneity of study participants. Adjustment for other
covariates such as family history of myocardial infarction, aspirin use, history
of hypertension, and hypercholesterolemia at baseline did not show a
significant effect on the ORs, and because they might be intermediates for the
association, they were not included in the final model. Generalized linear
models were used to compare geometric mean concentrations of log-trans-
formed biomarkers across the eNOS genotype groups among diabetic men
without CHD adjusting for multiple covariates. In addition, to account for
multiple statistical testing, we constructed permutation null-distribution data-
sets (10,000 resamples) to guide the interpretation of our results (30).
Two measures of linkage disequilibrium, squared correlation coefficient
(r2) and Lewontin’s standardized disequilibrium coefficient (D⬘), were com-
puted between pairs of SNPs. Haplotype frequencies were estimated with the
expectation-maximization algorithm, as implemented in SAS PROC Haplotype
(SAS Institute, Cary, NC). Global/Omnibus tests of haplotype association and
haplotype-specific ORs were calculated by haplotype replacement regression
(31), assuming an additive model using the probability of carrying each
haplotype provided by PROC Haplotype. Men who were homozygous for the
most common haplotype were used as the reference, and rare haplotypes
(combined frequency ⬍5%) were collapsed together. We assumed an additive
model, in which haplotype-specific parameters represent the per-haplotype
increase in log odds of disease. All reported P values are two-tailed, and
statistical significance was defined at the ␣ ⫽ 0.05 level. All analyses were
performed with SAS version 8.12 software (SAS Institute, Cary, NC).
RESULTS
Compared with those without CHD, diabetic men who de-
veloped CHD tended to be older and were more likely to
have a family history of myocardial infarction and to have
hypertension and hypercholesterolemia at baseline (Table 1).
They were also more likely to use aspirin at baseline, probably
because of their elevated cardiovascular risk factors.
We first examined the association of the two potentially
functional SNPs, eNOS ⫺786T⬎C and the Glu298Asp, with
CHD risk. Genotype frequency distributions of them did
not deviate from Hardy-Weinberg equilibrium among study
participants (combined case and control subjects, P ⫽ 0.47
for ⫺786T⬎C and 0.07 for Glu298Asp) (Table 2 and
Appendix 2). No statistically significant association be-
tween each of these two polymorphisms and CHD risk was
observed (minor allele frequency [95% CI] for case versus
control subjects: 38.4 [33.8 – 43.1] vs. 40.8 [38.1– 43.3] for
⫺786T⬎C; 30.5 [26.7–35.1] vs. 30.6 [27.5–32.8] for
2141
 eNOS GENE AND CHD RISK

TABLE 1
Comparison of cardiovascular risk factors between CHD case and control subjects at baseline (1986) among U.S. diabetic men
Characteristics CHD case subjects Control subjects P value
n 220 641
Age (years) 59.4 ⫾ 7.3 55.0 ⫾ 8.6 ⬍0.001
BMI (kg/m2) 27.9 ⫾ 4.3 27.6 ⫾ 4.1 0.52
Physical activity (MET/week) 14.4 ⫾ 20.7 14.0 ⫾ 17.7 0.81
Alcohol consumption (g/day) 8.9 ⫾ 16.4 11.0 ⫾ 16.5 0.10
Family history of myocardial infarction (%) 20.0 11.9 0.003
History of hypertension (%) 47.7 31.4 ⬍0.001
History of hypercholesterolemia (%) 25.9 12.2 ⬍0.001
Never smokers (%) 37.3 39.4 0.59
Aspirin users (%) 44.6 30.0 ⬍0.001
Race/ethnicity (% white) 99.0 97.4 0.79
ICAM-1 (ng/ml)* 379.1 ⫾ 80.0 356.7 ⫾ 88.6 0.10
VCAM-1 (ng/ml)* 1,425.4 ⫾ 384.5 1,387.8 ⫾ 367.4 ⬍0.001
CRP (mg/l) 3.4 ⫾ 4.5 2.9 ⫾ 4.8 0.04
Fibrinogen (mg/dl) 491.4 ⫾ 145.4 464.8 ⫾ 125.2 0.14
Data are means ⫾ SD unless otherwise indicated. *Only for subjects whose blood sample was available before CHD was identified (120 CHD
case subjects and 641 control subjects).

Glu298Asp). Because of the relatively small number of ho-
mozygous carriers of the variant allele in each polymorphism
among case subjects, we combined data of heterozygotes
with homozygotes in the following analysis. Adjustment for
smoking status, alcohol consumption, duration of diabetes,
and BMI did not change the results materially.
Among cardiovascular disease–negative control sub-
jects, plasma concentrations of both sVCAM-1 and
sICAM-1 were significantly higher for participants with the
298T variant allele (genotype GT/TT) than those without
this allele (genotype GG) (multivariate adjusted mean
1,385 vs. 1,299 ng/ml, P ⫽ 0.002 for sVCAM-1; 354 vs. 339
ng/ml, P ⫽ 0.020 for sICAM-1) (Fig. 1). Similarly, plasma
concentrations of sVCAM-1 and ICAM-1 were higher
among carriers of the ⫺786 C variant allele (genotype
CT/CC), although the difference in ICAM levels was not
statistically significant. No statistically significant differ-
ence in CRP and fibrinogen levels according to genotypes
of these two SNPs was observed.
We next examined the hypothesis that there might be
common noncoding variants that influence the risk of
CHD. We conducted haplotype analyses according to the
linkage disequilibrium pattern. The linkage disequilibrium
analyses (Appendix 3) revealed two regions characterized
by relatively strong linkage disequilibrium: block 1, includ-
ing SNP1 and SNP2 in promoter as indicate by D⬘ and r2,
and block 2, including SNPs 4 –7, which meets criteria for
haplotype “block” as defined by Gabriel et al. (32) (haplo-
view 4.0, 2004). SNPs 3 and 8 appeared not to belong to
these two blocks. We examined their associations with
CHD risk separately; no significant association was ob-
served (Table 3). Within block 2, haplotype 4 was signifi-
cantly associated with CHD risk, but the global test was
not significant. Because of the current controversies in
defining haplotype blocks, we also examined the associa-
tion of haplotypes with CHD risk using a sliding window
approach. The haplotypes containing the variant allele of
SNPs 4 and/or 5 were associated with a significantly
increased risk for CHD.
Because SNP4 (Glu298Asp) alone was not significantly
associated with CHD risk in our analyses, we examined
whether SNP5 was associated with CHD risk. We observed
that the frequency of heterozygous carriers of the variant
allele was significantly higher among case subjects than
control subjects (52 vs. 39%, P ⫽ 0.005) (Table 4). After
adjustment for age, BMI, duration of diabetes, smoking,
and alcohol consumptions, heterozygous carriers had 1.8-
fold increased risk of CHD compared with participants
without the variant allele (95% CI 1.3–2.6, P ⫽ 0.004).
Frequency distributions of homozygous carriers of the
variant allele were more similar among control subjects
than case subjects. Because of the small number of the
homozygous carriers among case subjects, we combined
heterozygous and homozygous data in the following anal-

TABLE 2
Associations of functional eNOS genotypes with the risk for CHD among U.S. diabetic men
Genotypes CHD case subjects Control subjects OR (95% CI)* OR (95% CI)†
⫺786T⬎C
TT 101 (47) 319 (51) 1.00 1.00
Carrier CT/CC 113 (52) 311 (49) 0.9 (0.6–1.3) 0.9 (0.7–1.3)
CT 95 (44) 240 (38) 0.9 (0.6–1.3) 1.0 (0.7–1.4)
CC 18 (8) 71 (11) 0.9 (0.6–1.5) 0.9 (0.5–1.7)
Glu298Asp (G894T)
GG 81 (38) 218 (34) 1.00 1.00
Carrier GT/TT 134 (62) 413 (66) 1.2 (0.8–1.6) 1.2 (0.9–1.8)
GT 103 (48) 313 (50) 1.2 (0.9–1.7) 1.3 (0.9–1.9)
TT 31 (14) 100 (16) 0.9 (0.5–1.5) 0.9 (0.5–1.7)
Data are n (%) and OR (95% CI). *Adjusted for age and duration of diabetes. †Adjusted for age, duration of diabetes, BMI, smoking status,
race/ethnicity, and alcohol consumption.

2142 DIABETES, VOL. 55, JULY 2006

 C. ZHANG AND ASSOCIATES

1, major allele; 0, minor allele. *Adjusted for age, BMI, duration of diabetes, smoking status, race/ethnicity, and alcohol consumption.

SNP8

Block 2

SNP3

Block 1

rs1800783
1

Association of haplotypes of the eNOS gene with the risk of CHD among U.S. diabetic men
TABLE 3
P for global test*
Others (Frq ⬍5%)

P for global test*

Others (Frq ⬍5%)
1
0

⫺786T⬎C
rs2070744
0.16

0.33

1
0

2
rs1800781
1
0

3
GLU298ASP
rs1799983

SNPs
1
1
0
0

4
rs1541861
1
1
0
0

5
FIG. 1. The associations of ⴚ786T>C, Glu298Asp, and the intron 8

rs2566511
(rs1541861) polymorphisms with plasma concentrations of sVCAM-1
and sICAM-1 among diabetic men without CHD at blood drawing. The
0
0
1
0

6
data are presented as means. P values for test of difference, using
log-transformed sVCAM-1 and sICAM-1, were adjusted for age, dura-
tion of diabetes, BMI, smoking status, and alcohol consumption. 䡺,
homozygous for common allele; f, carriers of rare allele.

ysis. Carriers of the variant allele (including heterozygous

and homozygous) had higher risk of CHD than noncarriers rs753482
0
1
0
0

(multivariate adjusted OR 1.5 [95% CI 1.1–2.1]; P ⫽ 0.03). 7

Moreover, plasma concentrations of sVCAM-1 and
sICAM-1 were higher among carriers than those without
this allele in control subjects (multivariate adjusted mean
rs380078

1,366 vs. 1,300 ng/ml, P ⫽ 0.019 for sVCAM-1; 350 vs. 340
ng/ml, P ⫽ 0.12 for sICAM-1) (Fig. 1). No statistically
1
0

significant difference in CRP and fibrinogen levels accord-

ing to genotypes of this SNP was observed. After account-
ing for multiple testing, we observed that empirical P
subjects

values exceeded 0.05 for the association between the

Case

intron 8 SNP and CHD risk (corrected P value ⫽ 0.17;

42.8
57.2

14.7
23.0
26.4
31.0
17.8
82.2

37.0
61.4

(%)
4.9

1.6

Appendix 2) and sVCAM concentrations (corrected P

value ⫽ 0.07), whereas empirical P values remained ⬍0.05
for the associations of the Glu298Asp and ⫺786T⬎C SNPs
with sVCAM-1 concentrations.
subjects
Control
39.2
60.8

12.8
19.9
28.0
33.1
17.3
82.7

40.0
58.3

(%)
6.2

1.7

DISCUSSION
In the present study of diabetic men, we did not observe
significant associations between the two potentially
functional eNOS polymorphisms (i.e., ⫺786T⬎C and
1.4 (0.9–1.9)

1.6 (1.1–2.4)
1.1 (0.5–9.2)
1.3 (0.9–1.7)

1.0 (0.8–1.3)

1.1 (0.7–1.5)

Glu298Asp) and CHD risk. However, we did find a signif-

OR*
1.0

1.0

1.0

1.0

icant elevation in the endothelial dysfunction markers

sVCAM-1 and sICAM-1 associated with these two polymor-
phisms. Another main finding was that carriers of the
DIABETES, VOL. 55, JULY 2006 2143
 eNOS GENE AND CHD RISK
TABLE 4
Associations of genotypes of the intron 8 locus of the eNOS Gene (rs1541861, SNP5) with the risk for CHD among U.S. diabetic men
Genotypes CHD case subjects
AA 69 (33)
AC/CC (carriers of the minor allele) 141 (67)
AC 109 (52)
CC 32 (15)
Data are n (%) and OR (95% CI). *Adjusted for age and duration of diabetes. †Adjusted for age, duration of diabetes, BMI, smoking status,
race, and alcohol consumption.
variant allele of a SNP located at intron 8 were associated
with increased risk of CHD among diabetic men. The
haplotype analyses confirmed this association. Moreover,
plasma concentrations of sVCAM-1 were significantly
higher among carriers of this allele.
The Glu298Asp polymorphism is the only common
variant identified so far in the coding region of eNOS gene.
Mechanistic studies indicated that this substitution alters
function. For instance, associations have been described
between this polymorphism and eNOS activity (11,33) or
endothelial function (34,35). A mechanism by which eNOS
Glu298Asp might reduce NO bioavailability has also been
reported. eNOS Asp298 is subject to selective proteolytic
cleavage in endothelial cells and vascular tissues, which
could lead to reduced vascular NO generation (11,33)
because the cleaved fragments would be expected to lack
NO synthase activity. A functional effect for the ⫺786 T⬎C
variant was observed as assessed by in vitro reporter gene
assays; the T⫺786C variant suppressed eNOS gene trans-
mission and resulted in a significant reduction in the eNOS
gene promoter activity (12). Lower eNOS mRNA and
serum nitrite/nitrate levels have been found in individuals
with the ⫺786C variant in some studies as well (36 –38).
Despite the substantial evidence that endothelial dys-
function and vascular inflammation contribute to the
accelerated atherogenesis associated with type 2 diabetes,
very few studies have been conducted among diabetic
patients. Findings from epidemiological studies on the
association between these two functional polymorphisms
and CHD risk are conflicting, and studies of the associa-
tion of eNOS polymorphism with CHD risk among Cauca-
sian diabetic patients are sparse. In a recent meta-analysis
of 26 studies, Casas et al. (15) found that the 298 polymor-
phism was associated with higher risk of ischemic heart
disease (OR 1.34 [95% CI 1.03–1.75]), whereas no signifi-
cant association of the ⫺786 polymorphism with CHD risk
was observed. Insufficient sample size may have limited
our power to detect a significant association between the
Glu298Asp polymorphism and CHD risk. We estimated a
priori detectable ORs on the basis of available data. We set
type 1 error at ␣ ⫽ 0.05, sample size as 220 case subjects
and 641 control subjects, power at ␤ ⫽ 0.80, and minor
allele frequency as reported previously in NCBI database
(minimum 0.33). The estimated detectable OR of the
Glu298Asp variant (1.54) fell into the range of reported
ORs (1.0 – 4.2) for the association between this SNP and
CHD risk among Caucasians (15). However, if true effect
of this SNP on CHD risk was smaller, we would have had
lower power and may have failed to detect an association.
Identifying significant associations of genetic variants
with complex qualitative trait as the CHD may demand a
larger sample size than that for quantitative traits. When
examining the association of these two polymorphisms
with quantitative trait, we observed that these two variants
2144
Control subjects OR (95% CI)* OR (95% CI)†
267 (44) 1.0 1.0
353 (56) 1.4 (1.1–1.9) 1.5 (1.1–2.1)
244 (39) 1.5 (1.2–2.1) 1.8 (1.3–2.6)
109 (17) 1.1 (0.7–1.7) 1.1 (0.7–1.9)
were associated with higher concentrations of cellular
adhesion molecules (i.e., soluble ICAM and/or VCAM).
Cellular adhesion molecules mediate the adhesion and
migration of leukocytes, steps proposed to play a critical
role in early atherogenesis (39,40). Consistent with find-
ings from the present study, circulating levels of VCAM-1
have been positively associated with clinical atherosclero-
sis (41– 43). Although we are unaware of studies that have
directly examined the association of eNOS polymorphisms
with CAM in diabetic patients, our results are consistent
with findings from a recent study, which demonstrated
that gene transfer of eNOS suppressed arterial VCAM-1
expression in an animal model (44).
There are a few plausible explanations for the observed
association between the intron 8 variant and CHD risk. It
is plausible that the intron 8 locus could act as a marker
for another functional, yet unidentified polymorphism, in
the eNOS gene or functional SNPs in other genes. In
addition to the eNOS gene, chromosome 7q36 contains a
few other genes that are plausibly implicated in the
atherogenic pathway because of their functional signifi-
cance, for example, insulin-induced gene 1 (154.48 –154.49
Mb; MIM 602055) (45), protein kinase, AMP-activated, 2
noncatalytic subunit (MIM 602743), and fatty acid binding
protein 5-like 3 (46). It is possible that the association we
found for the intron 8 locus is the result of linkage
disequilibrium with certain SNPs in genes within this area.
Another possibility is that the intron 8 allele has intrinsic
functional significance. Although this variant lies within an
intron, it could affect mRNA stability and enzyme levels by
affecting splicing. In the present study, the associations of
CHD risk with the intron 8 SNP was only due to the
heterozygous genotype. This could indicate only a weak
dominant effect or could be due to the lack of power to
detect a significant effect for the homozygotes of the rare
allele because of the small number of study subjects falling
into this category.
Lastly, our observed positive association may be due to
chance, in particular considering that multiple variants at
a locus may increase type I error (47,48). Stringent statis-
tical criteria to correct for multiple testing and replication
were recommended to determine whether the observed
associations are false-positive (49). We constructed 10,000
permutation null datasets and found that empirical P
values exceeded 0.05 for the association between the
intron 8 SNP and CHD risk, whereas empirical probability
values remained ⬍0.05 for the associations of the
Glu298Asp and ⫺786T⬎C SNPs with sVCAM-1 concentra-
tions. Thus, despite observing nominal significant associ-
ations of the intron 8 SNP with phenotypes (i.e., CHD risk
and sVCAM concentrations) among diabetic patients, dec-
laration of a genuine association between this polymor-
phism and phenotypes would be premature. Instead, we
regarded our observed associations of the intron 8 SNP
DIABETES, VOL. 55, JULY 2006
 C. ZHANG AND ASSOCIATES

with CHD risk as a hypothesis generating observations and
meriting testing in other cohorts of diabetic patients.
However, it should also be noted that, by controlling the
false-positive rate through empirical methods, we may
have increased our false-negative rate (48). Rigid adher-
ence to an empirical P ⬍ 0.05 significance threshold could
be overly conservative and obscure some true positive
associations (50). Finally, the Health Professionals Fol-
low-Up Study does not represent a random sample of U.S.
diabetic men. The relative socioeconomic homogeneity of
this population, on the other hand, tends to reduce the
impact of unknown confounders.
In conclusion, our data suggested that two potentially
functional polymorphisms (i.e., 786T⬎C and Glu298Asp)
and an intron 8 polymorphism of the eNOS gene are
potentially involved in the atherogenic pathway among
diabetic men. It will be important to confirm these findings
with additional investigations among larger samples of
diabetic patients. The biological importance of the intron
SNP is currently unclear, and further functional testing of this
SNP is required if this finding is confirmed by further studies.
ACKNOWLEDGMENTS
This research has received National Institutes of Health
Grants HL-65582 and HL-35464. F.B.H. is partly supported
by an American Heart Association Established Investiga-
tor Award.
APPENDIXES

APPENDIX 1
Descriptions of selected SNPS in the eNOS gene
Minor allele frequency
Chromosome Current Detectable
SNPs position DbSNP Location NCBI* sample† OR‡
1 150127045 rs1800783 5⬘ (A/T) 40.0–56.5 40.8 1.51
2 150127727 rs2070744 5⬘ (C/T) 36.0–41.0 41.3 1.52
3 150130092 rs1800781 Intron 2 (A/G) 15.0–19.6 17.3 1.65
4 150133759 rs1799983 Exon 7 (G/T) 33.3–50.0 30.6 1.54
5 150134981 rs1541861 Intron 8 (A/C) 45.8–47.8 37.2 1.50
6 150140689 rs2566511 Intron 13 (C/T) 22.7–28.3 28.4 1.56
7 150144031 rs753482 Intron 18 (G/T) 30.4 22.0 1.54
8 150151284 rs3800787 3⬘-UTR (C/G) 40.9 40.2 1.51
Data are percent and OR. *Data are from UW-FHRCRC-PGA European ancestry and Hapmap European ancestry in the NCBI SNP website.
†Case and control subjects combined. ‡Detectable OR estimated under the condition ␣ ⫽ 0.05, power ⫽ 0.80, and minor allele frequency
reported from NCBI. UTR, untranslated region.
APPENDIX 2
Comparison of minor allele frequency distributions between CHD case and control subjects among diabetic men
Minor allele frequency (95% CI)
SNPs CHD case subjects (n ⫽ 220) Control subjects (n ⫽ 641) P value* Corrected P value*†
1 37.2 (32.8–41.7) 40.8 (38.3–43.3) 0.19 0.57
2 39.4 (33.8–43.1) 41.7 (38.1–43.3) 0.46 0.80
3 17.9 (14.2–21.4) 17.3 (15.1–19.5) 0.61 0.80
4 30.6 (26.7–35.1) 30.3 (27.5–32.8) 0.34 0.75
5 38.3 (33.7–42.8) 37.2 (34.0–40.1) 0.03 0.17
6 28.9 (24.8–33.3) 27.6 (25.1–30.2) 0.31 0.75
7 19.8 (16.3–23.8) 22.5 (20.0–24.8) 0.52 0.80
8 42.8 (38.3–47.7) 39.2 (36.4–42.2) 0.09 0.31
*P value from Fisher’s exact tests of difference based on dominant model. †Corrected P value based on permutation test of 10,000 resampling
under null distribution.
APPENDIX 3
Pairwise measure of linkage disequilibrium (D⬘) between eNOS SNPs
SNP
SNP rs1800783 rs2070744 rs1800781 rs1799983 rs1541861 rs2566511 rs753482 rs3800787
rs1800783 — 0.923 0.192 0.177 0.078 0.028 0.293 0.0002
rs2070744 0.969 — 0.193 0.200 0.093 0.022 0.289 0.0005
rs1800781 0.797 0.791 — 00.086 0.115 0.470 0.051 0.234
rs1799983 0.538 0.559 ⫺0.976 — 0.682 0.162 0.477 0.104
rs1541861 0.302 0.326 ⫺0.963 0.972 — 0.228 0.372 0.114
rs2566511 0.221 0.199 0.934 ⫺0.986 ⫺0.990 — 0.094 0.430
rs753482 0.834 0.823 ⫺0.918 0.838 0.871 ⫺0.916 — 0.117
rs3800787 0.014 0.022 ⫺0.858 ⫺0.607 ⫺0.540 0.856 ⫺0.786 —
r2 values are given above the diagonal, and D⬘ values are given below the diagonal.

DIABETES, VOL. 55, JULY 2006 2145

 eNOS GENE AND CHD RISK
REFERENCES
1. Moncada S, Higgs A: The L-arginine-nitric oxide pathway. N Engl J Med
329:2002–2012, 1993
2. Tsao PS, Buitrago R, Chan JR, Cooke JP: Fluid flow inhibits endothelial
adhesiveness: nitric oxide and transcriptional regulation of VCAM-1.
Circulation 94:1682–1689, 1996
3. Marsden PA, Heng HH, Scherer SW, Stewart RJ, Hall AV, Shi XM, Tsui LC,
Schappert KT: Structure and chromosomal localization of the human
constitutive endothelial nitric oxide synthase gene. J Biol Chem 268:
17478 –17488, 1993
4. Kuhlencordt PJ, Gyurko R, Han F, Scherrer-Crosbie M, Aretz TH, Hajjar R,
Picard MH, Huang PL: Accelerated atherosclerosis, aortic aneurysm for-
mation, and ischemic heart disease in apolipoprotein E/endothelial nitric
oxide synthase double-knockout mice. Circulation 104:448 – 454, 2001
5. Chen AF, Ren J, Miao CY: Nitric oxide synthase gene therapy for
cardiovascular disease. Jpn J Pharmacol 89:327–336, 2002
6. Ooboshi H, Toyoda K, Faraci FM, Lang MG, Heistad DD: Improvement of
relaxation in an atherosclerotic artery by gene transfer of endothelial nitric
oxide synthase. Arterioscler Thromb Vasc Biol 18:1752–1758, 1998
7. Tsutsui M, Chen AF, O’Brien T, Crotty TB, Katusic ZS: Adventitial
expression of recombinant eNOS gene restores NO production in arter-
ies without endothelium. Arterioscler Thromb Vasc Biol 18:1231–1241,
1998
8. Channon KM, Qian H, Neplioueva V, Blazing MA, Olmez E, Shetty GA,
Youngblood SA, Pawloski J, McMahon T, Stamler JS, George SE: In vivo
gene transfer of nitric oxide synthase enhances vasomotor function in
carotid arteries from normal and cholesterol-Fed rabbits. Circulation
98:1905–1911, 1998
9. Sato J, Mohacsi T, Noel A, Jost C, Gloviczki P, Mozes G, Katusic ZS,
O’Brien T, Mayhan WG: In vivo gene transfer of endothelial nitric oxide
synthase to carotid arteries from hypercholesterolemic rabbits enhances
endothelium-dependent relaxations. Stroke 31:968 –975, 2000
10. Qian H, Neplioueva V, Shetty GA, Channon KM, George SE: Nitric oxide
synthase gene therapy rapidly reduces adhesion molecule expression and
inflammatory cell infiltration in carotid arteries of cholesterol-fed rabbits.
Circulation 99:2979 –2982, 1999
11. Tesauro M, Thompson WC, Rogliani P, Qi L, Chaudhary PP, Moss J:
Intracellular processing of endothelial nitric oxide synthase isoforms
associated with differences in severity of cardiopulmonary diseases:
cleavage of proteins with aspartate vs. glutamate at position 298. Proc Natl
Acad Sci U S A 97:2832–2835, 2000
12. Nakayama M, Yasue H, Yoshimura M, Shimasaki Y, Kugiyama K, Ogawa H,
Motoyama T, Saito Y, Ogawa Y, Miyamoto Y, Nakao K: T-7863 C mutation
in the 5⬘-flanking region of the endothelial nitric oxide synthase gene is
associated with coronary spasm. Circulation 99:2864 –2870, 1999
13. Wang XL: Cigarette smoking, DNA variants in endothelial nitric oxide
synthase gene and vascular disease. Contrib Nephrol 130:53– 67, 2000
14. Wang XL, Sim AS, Wang MX, Murrell GA, Trudinger B, Wang J: Genotype
dependent and cigarette specific effects on endothelial nitric oxide syn-
thase gene expression and enzyme activity. FEBS Lett 471:45–50, 2000
15. Casas JP, Bautista LE, Humphries SE, Hingorani AD: Endothelial nitric
oxide synthase genotype and ischemic heart disease: meta-analysis of 26
studies involving 23028 subjects. Circulation 109:1359 –1365, 2004
16. Antoniades C, Tousoulis D, Vasiliadou C, Pitsavos C, Chrysochoou C,
Panagiotakos D, Tentolouris C, Marinou K, Koumallos N, Stefanadis C:
Genetic polymorphism on endothelial nitric oxide synthase affects endo-
thelial activation and inflammatory response during the acute phase of
myocardial infarction. J Am Coll Cardiol 46:1101–1109, 2005
17. Cam SF, Sekuri C, Tengiz I, Ercan E, Sagcan A, Akin M, Berdeli A: The
GLU298ASP polymorphism on endothelial nitric oxide synthase gene is
associated with premature coronary artery disease in a Turkish popula-
tion. Thromb Res 116:287–292, 2005
18. National Diabetes Data Group: Classification and diagnosis of diabetes
mellitus and other categories of glucose intolerance. Diabetes 1979;28:
1039 –1057
19. Expert Committee on the Diagnosis and Classification of Diabetes Melli-
tus: Report of the Expert Committee on the Diagnosis and Classification of
Diabetes Mellitus. Diabetes Care 1997;20:1183–1197
20. Hu FB, Leitzmann MF, Stampfer MJ, Colditz GA, Willett WC, Rimm EB:
Physical activity and television watching in relation to risk for type 2
diabetes mellitus in men. Arch Intern Med 161:1542–1548, 2001
21. Rose G, Blackburn H: WHO Monograph Series: Cardiovascular Survey
Methods. Geneva, World Health Org., 1982
22. Rimm EB, Giovannucci EL, Willett WC, Colditz GA, Ascherio A, Rosner B,
Stampfer MJ: Prospective study of alcohol consumption and risk of
coronary disease in men. Lancet 338:464 – 468, 1991
2146
23. Stampfer MJ, Willett WC, Speizer FE, Dysert DC, Lipnick R, Rosner B,
Hennekens CH: Test of the National Death Index. Am J Epidemiol
119:837– 839, 1984
24. Rimm EB, Stampfer MJ, Colditz GA, Chute CG, Litin LB, Willett WC:
Validity of self-reported waist and hip circumferences in men and women.
Epidemiology 1:466 – 473, 1990
25. NIEHS SNPs [article online]. NIEHS Environmental Genome Project,
University of Washington, Seattle, WA, USA. Available from http://egp.gs.
washington.edu. Accessed 8 April 2004
26. Stephens M, Donnelly P: A comparison of bayesian methods for haplotype
reconstruction from population genotype data. Am J Hum Genet 73:1162–
1169, 2003
27. Sebastiani P, Lazarus R, Weiss ST, Kunkel LM, Kohane IS, Ramoni MF:
Minimal haplotype tagging. Proc Natl Acad Sci U S A 100:9900 –9905, 2003
28. Carlson CS, Eberle MA, Rieder MJ, Yi Q, Kruglyak L, Nickerson DA:
Selecting a maximally informative set of single-nucleotide polymorphisms
for association analyses using linkage disequilibrium. Am J Hum Genet
74:106 –120, 2004
29. Schulze MB, Rimm EB, Li T, Rifai N, Stampfer MJ, Hu FB: C-reactive
protein and incident cardiovascular events among men with diabetes.
Diabetes Care 27:889 – 894, 2004
30. Westfall PH, Soper KA: Nonstandard uses of PROC MULTTEST: Permuta-
tional peto tests, permutational and unconditional t and binomial tests. In
Proceedings of the 19th Annual SAS User’s Group International Confer-
ence, 1994. Cary, NC, SAS Institute, 1994, p. 986 –989
31. Kraft P, Cox DG, Paynter RA, Hunter D, De Vivo I: Accounting for
haplotype uncertainty in matched association studies: a comparison of
simple and flexible techniques. Genet Epidemiol 28:261–272, 2005
32. Gabriel SB, Schaffner SF, Nguyen H, Moore JM, Roy J, Blumenstiel B,
Higgins J, DeFelice M, Lochner A, Faggart M, Liu-Cordero SN, Rotimi C,
Adeyemo A, Cooper R, Ward R, Lander ES, Daly MJ, Altshuler D: The
structure of haplotype blocks in the human genome. Science 296:2225–
2229, 2002
33. Persu A, Stoenoiu MS, Messiaen T, Davila S, Robino C, El-Khattabi O,
Mourad M, Horie S, Feron O, Balligand JL, Wattiez R, Pirson Y, Chauveau
D, Lens XM, Devuyst O: Modifier effect of ENOS in autosomal dominant
polycystic kidney disease. Hum Mol Genet 11:229 –241, 2002
34. Savvidou MD, Vallance PJ, Nicolaides KH, Hingorani AD: Endothelial nitric
oxide synthase gene polymorphism and maternal vascular adaptation to
pregnancy. Hypertension 38:1289 –1293, 2001
35. Leeson CP, Hingorani AD, Mullen MJ, Jeerooburkhan N, Kattenhorn M,
Cole TJ, Muller DP, Lucas A, Humphries SE, Deanfield JE: Glu298Asp
endothelial nitric oxide synthase gene polymorphism interacts with envi-
ronmental and dietary factors to influence endothelial function. Circ Res
90:1153–1158, 2002
36. Miyamoto Y, Saito Y, Nakayama M, Shimasaki Y, Yoshimura T, Yoshimura
M, Harada M, Kajiyama N, Kishimoto I, Kuwahara K, Hino J, Ogawa E,
Hamanaka I, Kamitani S, Takahashi N, Kawakami R, Kangawa K, Yasue H,
Nakao K: Replication protein A1 reduces transcription of the endothelial
nitric oxide synthase gene containing a ⫺786T3 C mutation associated
with coronary spastic angina. Hum Mol Genet 9:2629 –2637, 2000
37. Jeerooburkhan N, Jones LC, Bujac S, Cooper JA, Miller GJ, Vallance P,
Humphries SE, Hingorani AD: Genetic and environmental determinants of
plasma nitrogen oxides and risk of ischemic heart disease. Hypertension
38:1054 –1061, 2001
38. Ohtoshi K, Yamasaki Y, Gorogawa S, Hayaishi-Okano R, Node K, Matsuhisa
M, Kajimoto Y, Hori M: Association of (⫺)786T-C mutation of endothelial
nitric oxide synthase gene with insulin resistance. Diabetologia 45:1594 –
1601, 2002
39. Ross R: The pathogenesis of atherosclerosis: a perspective for the 1990s.
Nature 362:801– 809, 1993
40. Osborn L, Hession C, Tizard R, Vassallo C, Luhowskyj S, Chi-Rosso G,
Lobb R: Direct expression cloning of vascular cell adhesion molecule 1, a
cytokine-induced endothelial protein that binds to lymphocytes. Cell
59:1203–1211, 1989
41. Rohde LE, Lee RT, Rivero J, Jamacochian M, Arroyo LH, Briggs W, Rifai N,
Libby P, Creager MA, Ridker PM: Circulating cell adhesion molecules are
correlated with ultrasound-based assessment of carotid atherosclerosis.
Arterioscler Thromb Vasc Biol 18:1765–1770, 1998
42. Ridker PM, Hennekens CH, Roitman-Johnson B, Stampfer MJ, Allen J:
Plasma concentration of soluble intercellular adhesion molecule 1 and
risks of future myocardial infarction in apparently healthy men. Lancet
351:88 –92, 1998
43. Ridker PM: Intercellular adhesion molecule (ICAM-1) and the risks of
developing atherosclerotic disease. Eur Heart J 19:1119 –1121, 1998
44. Li L, Crockett E, Wang DH, Galligan JJ, Fink GD, Chen AF: Gene transfer
of endothelial NO synthase and manganese superoxide dismutase on
DIABETES, VOL. 55, JULY 2006

arterial vascular cell adhesion molecule-1 expression and superoxide
production in deoxycorticosterone acetate-salt hypertension. Arterioscler
Thromb Vasc Biol 22:249 –255, 2002
45. Li J, Takaishi K, Cook W, McCorkle SK, Unger RH: Insig-1 “brakes”
lipogenesis in adipocytes and inhibits differentiation of preadipocytes.
Proc Natl Acad Sci U S A 100:9476 –9481, 2003
46. Li WD, Dong C, Li D, Garrigan C, Price RA: A genome scan for serum
triglyceride in obese nuclear families. J Lipid Res 46:432– 438, 2005
47. Neale BM, Sham PC: The future of association studies: gene-based analysis
and replication. Am J Hum Genet 75:353–362, 2004
DIABETES, VOL. 55, JULY 2006
C. ZHANG AND ASSOCIATES
48. Wacholder S, Chanock S, Garcia-Closas M, El Ghormli L, Rothman
N: Assessing the probability that a positive report is false: an approach
for molecular epidemiology studies. J Natl Cancer Inst 96:434 – 442,
2004
49. Lohmueller KE, Pearce CL, Pike M, Lander ES, Hirschhorn JN: Meta-
analysis of genetic association studies supports a contribution of common
variants to susceptibility to common disease. Nat Genet 33:177–182,
2003
50. Thomas DC, Clayton DG: Betting odds and genetic associations. J Natl
Cancer Inst 96:421– 423, 2004
2147