Pharmacogenetic Effect of the Stromelysin (MMP3)
Polymorphism on Stroke Risk in Relation to
Antihypertensive Treatment
The Genetics of Hypertension Associated Treatment Study
Richard Sherva, PhD; Charles E. Ford, PhD; John H. Eckfeldt, MD, PhD; Barry R. Davis, MD, PhD;
Eric Boerwinkle, PhD; Donna K. Arnett, PhD
Downloaded from http://stroke.ahajournals.org/ by guest on June 16, 2017
Background and Purpose—Atherothrombotic diseases including stroke share a common etiology of atherosclerosis, and
susceptibility to atherosclerosis has a genetic component. Stromelysin-1 (matrix metalloproteinase-3 [MMP3]) regulates
arterial matrix composition and is a candidate gene for atherothrombosis. A common polymorphism of MMP3 alters
expression levels and affects atherosclerotic progression and plaque stability. As part of the Genetics of Hypertension
Associated Treatment study, ancillary to the Antihypertensive and Lipid Lowering to Prevent Heart Attack Trial, we
evaluated the 5A/6A polymorphism in MMP3 to determine its association with stroke and determine whether it modifies
clinical outcome response to blood pressure–lowering drugs.
Methods—The effect of the MMP3 5A/6A polymorphism on stroke rates was examined by using multivariate-adjusted
Cox regression models, including a test for interactions between genotype and antihypertensive drug class.
Results—Compared with participants treated with chlorthalidone with the 6A/6A genotype, individuals with the 6A/6A
genotype randomized to lisinopril had higher stroke rates (hazard ratio⫽1.32; 95% CI, 1.08 to 1.61; P⫽0.007) and
5A/6A individuals taking lisinopril had lower stroke rates (hazard ratiointeraction⫽0.74; 95% CI, 0.53 to 1.04;
Pinteraction⫽0.08), whereas 5A/5A individuals taking lisinopril had the lowest stroke rate (hazard ratiointeraction⫽0.51;
95% CI, 0.31 to 0.85; Pinteraction⫽0.009). There were no pharmacogenetic differences in stroke rate by genotype in
patients taking amlodipine or doxazosin vs chlorthalidone.
Conclusions—The MMP3 6A/6A genotype is associated with an increased risk of stroke in hypertensive subjects taking
lisinopril compared with patients treated with chlorthalidone, whereas a protective effect was found for 5A/5A
individuals treated with lisinopril. Genetic screening for the MMP3 5A/6A genotype might be a useful tool to select
optimal antihypertensive therapy if this finding is replicated.
Clinical Trial Registration—URL: http://www.clinicaltrials.gov. Unique identifier: NCT00563901.
(Stroke. 2011;42:330-335.)
Key Words: stroke 䡲 pharmacogenetics 䡲 matrix metalloproteinase-3 䡲 antihypertensive agents
S
troke is the third leading cause of death in the United
States and the leading cause of long-term disability.
There are ⬇700 000 stroke cases per year, and estimates of
the annual burden placed on healthcare costs by stroke range
between $30 and $50 billion.1
In addition to numerous lifestyle risk factors, atherosclerosis2– 4 and hypertension5,6 are central to the pathogenesis of
stroke. Changes in the composition of the arterial matrix are
among the primary mechanisms underlying this progression. As
atherosclerosis and hypertension persist, there is a reduction in
the ratio of elastin to collagen in the arterial matrix, reducing
arterial compliance and increasing blood pressure (BP).7
Stromelysin-1 (matrix metalloproteinase-3 [MMP3]) is an important enzymatic regulator of extracellular matrix remodeling
secreted predominantly by connective-tissue cells.8 MMP3
cleaves several types of collagen at various sites in the helical
domain and degrades other constituents of the extracellular
matrix, including proteoglycan, fibronectin, and laminin.9 Because hypertension is a major cause of atherosclerotic remodeling of the large arteries and is also the most important stroke risk
factor, we chose stroke as the outcome through which to test for
an antihypertension pharmacogenetic effect of MMP3.
Received June 15, 2010; accepted September 2, 2010.
From the Department of Medicine, Genetics Program (R.S.), Boston University School of Medicine, Boston, MA; School of Public Health (C.E.F.,
B.R.D., E.B.), University of Texas, Houston, TX; Department of Laboratory Medicine and Pathology (J.H.E.), University of Minnesota, Minneapolis,
MN; and Department of Epidemiology (D.K.A.), University of Alabama, Birmingham, AL.
The online-only Data Supplement is available at http://stroke.ahajournals.org/content/full/stroke.110.593798/DC1.
Correspondence to Donna K. Arnett, PhD, Department of Epidemiology, University of Alabama at Birmingham, 1665 University Ave, Suite 220,
Birmingham, AL 32594. E-mail [email protected]
© 2011 American Heart Association, Inc.
Stroke is available at http://stroke.ahajournals.org
DOI: 10.1161/STROKEAHA.110.593798
330
Sherva et al
A common polymorphism, MMP3 5A/6A, consisting of
either 5 or 6 adenosine residues, respectively, is located in the
interleukin-1 response element of the MMP3 promoter region. The 6A allele is associated with decreased MMP3
expression10,11 and more rapid progression of atherosclerosis,10,12–14 presumably due to collagen buildup in the arterial
matrix. Other studies have directly linked increased MMP3
expression and 5A alleles to plaque instability and myocardial infarction (MI),15–19 and heterozygotes are thought to
have the optimal balance of matrix accumulation and
degradation.11Although 6A alleles have been shown to predispose individuals to carotid artery stenosis, a known stroke
risk factor,20 –22 no studies have directly linked either allele to
stroke risk.
Materials and Methods
Study Design
Downloaded from http://stroke.ahajournals.org/ by guest on June 16, 2017
The Antihypertensive and Lipid Lowering to Prevent Heart Attack
Trial (ALLHAT) is a practice-based, double-blind, active-controlled,
randomized, clinical trial with 42 418 high-risk hypertensives randomized in 623 clinical centers. Participants are 55 years of age or
older, with systolic or diastolic hypertension and 1 or more risk
factors for cardiovascular disease. BP eligibility for ALLHAT was
based on current antihypertensive medication use and/or the average
of 2 seated measurements during 2 prerandomization visits. For
patients currently taking BP–lowering drugs, the systolic (SBP) and
diastolic (DBP) BP could not exceed 160/100 mm Hg at the first visit
or 180/110 mm Hg at the second visit. Patients were eligible if they
met any of the inclusion criteria (in addition to a diagnosis of
hypertension). Inclusion criteria were ST-T–wave ECG changes
indicative of ischemia, a history of coronary revascularization,
angina pectoris, other documented atherosclerotic disease, type 2
diabetes, HDL cholesterol ⬍35 mg/dL (any 2 determinations in the
past 5 years), current smoking, a combined ECG wall thickness of
0.25 mm measured in the past 2 years, and ECG evidence of left
ventricular hypertrophy in the past 2 years. Exclusion criteria were
symptomatic MI, stroke, or angina pectoris within the past 6 months;
symptomatic heart failure or ejection fraction 35%; nonhypertensive
indications for an aversion to the drugs being tested in ALLHAT;
requirement for ⬎2 medications to control BP; sensitivity/contraindications to any of the first-line study medications; serum creatinine
ⱖ2 mg/dL; low likelihood of compliance with the study protocol;
serious illness likely to lead to noncardiac death during the trial; or
current participation in another clinical trial.
Participants were randomized to 1 of 4 treatment arms, chlorthalidone, lisinopril, amlodipine, and doxazosin, in a ratio of 1.7:1:1:1,
respectively. They were initially started on the lowest possible dose
of each agent, and dosages were increased until BP control was
achieved (SBP ⬍140 and DBP ⬍90 mm Hg). If BP control was not
achieved at the maximum dose of the study medication, a secondline, open-label agent (reserpine, clonidine, or atenolol) and/or a
third-line, open-label agent (hydralazine) was added.
After BP goals were met and medication regimens stabilized,
clinic visits were scheduled at 3-month intervals during the first year
and at 4-month intervals thereafter. BP was measured at each visit.
Laboratory measurements, including potassium, glucose, creatinine,
total cholesterol, HDL cholesterol, triglycerides, and alanine aminotransferase, were measured after an 8-hour fast at years 1, 2, 4, and 6.
The primary outcome for ALLHAT was a composite measure
including coronary heart disease death and nonfatal MI. Major
secondary outcomes included all-cause mortality; fatal and nonfatal
stroke; combined coronary heart disease, defined by coronary revascularization or hospitalized angina; combined cardiovascular disease
defined by combined coronary heart disease, stroke, other treated
angina, heart failure, or peripheral arterial disease; and end-stage
renal disease defined by dialysis, renal transplantation, or death from
Pharmacogenetics of Antihypertensives and Stroke
331
kidney disease. For this analysis, we focused on 1 secondary
outcome, fatal and nonfatal stroke, which included ischemic and
hemorrhagic strokes. Outcomes were ascertained at clinic visits and
through national databases and death certificates, and hospitalization
records were used to support clinician-assigned outcomes.
In the Genetics of Hypertension Associated Treatment (GenHAT)
study, variants in candidate genes were typed in ALLHAT participants in an attempt to determine whether these genes interact with
antihypertensive medications to modify the risks of developing
adverse cardiovascular end points. Of the original ALLHAT sample,
39 114 had DNA available and were included in GenHAT. This
research was approved by local institutional review boards, and
informed consent was collected from all participants in ALLHAT.
Genotyping
DNA was isolated and immobilized on FTA paper (Fitzco Inc,
Maple Plain, MN). Polymerase chain reaction amplification was
done on 96-well plates by robotic workstations and multichannel
pipettes with positive and negative controls. The MMP3 polymorphism was genotyped in the context of a larger panel of
cardiovascular disease candidate gene variants, so each sample was
amplified in a multiplex polymerase chain reaction and then hybridized to an array of immobilized, sequence-specific oligonucleotide
probes (Roche Molecular Systems, Pleasanton, CA) according to the
methods described by Cheng et al.23
Statistical Analysis
Hardy-Weinberg equilibrium tests were performed by ␹2 tests. The
MMP3 5A/6A polymorphism was analyzed in multivariate Cox
regression models with stroke as the outcome. The polymorphism
was initially tested as a 3-level categorical predictor (representing 2
homozygous and 1 heterozygous genotype). In addition, we compared 6A/6A individuals versus those with 6A/5A or 5A/5A, and
also 5A/5A individuals versus those with 6A/5A or 6A/6A. Finally,
as there is previous evidence suggesting that individuals who are
heterozygous for the MMP3 5A/6A polymorphism may have the
lowest cardiovascular risk profile,11 we tested a model comparing
heterozygotes with homozygotes for the polymorphism. For the fully
adjusted, model potential covariates including age, sex, race, smoking and diabetes status, BP (SBP and DBP), total cholesterol, body
mass index, aspirin use, and antihypertensive treatment group were
added. These covariates were chosen on the basis of previous
evidence suggesting that they might influence stroke risk. Those
found significant (Pⱕ0.05) were retained in the final model. We also
tested for interactions between the MMP3 gene and age, sex, race,
diabetes status, and smoking, as these were prespecified subgroups
within ALLHAT, and also for interactions between treatment group and
the 5A/6A polymorphism. The complete results for each genotypic
model and interaction test are shown in Table I in the online-only Data
Supplement, available at http://stroke.ahajournals.org.
In total, 78 variants were typed in the GenHAT study, which
would normally require probability values on the order of 0.001 to
meet Bonferroni-adjusted significance thresholds. Because only the
MMP3 variant was analyzed in the current study, less conservative
adjustment methods such as computing false-discovery rates were
not feasible. We therefore used the Bonferroni-adjusted significance
threshold as a general guideline, knowing that it is likely overly
conservative.
Because the focus of this analysis was primary stroke prevention,
we excluded 9076 participants with a history of stroke or MI at
baseline. This also allowed us to use a cumulative incidence analysis
approach and minimized potential bias due to changes in lifestyle
and risk factor profiles after 1 of these events. Because the doxazosin
arm of the trial was halted early, those data are not directly
comparable with those of the other 3 arms, and we therefore
excluded participants randomized to doxazosin. We tested for
pharmacogenetic effects of the MMP3 variant on that drug, however,
in a subset of the data that only contained information collected
332
Table 1.
Stroke
February 2011
Characteristics of Participants by MMP3 Genotype
Variable
5A/5A
5A/6A
6A/6A
Stroke, No. (%)
120 (3.6)
288 (3.2)
457 (4.2)
Age, mean⫾SD, y
66.9⫾7.5
66.6⫾7.5
66.1⫾7.8
1819 (54.9)
4548 (52.3)
5161 (46.4)
Male, No. (%)
Diabetes, No. (%)
Current smoking, No. (%)
1156 (34.9)
3288 (37.8)
4425 (39.8)
748 (22.6)
1979 (22.7)
2660 (23.9)
Baseline SBP, mean⫾SD,
mm Hg
144.6⫾13.8
6-month SBP, mean⫾SD,
mm Hg
139.4⫾16.0
Baseline DBP, mean⫾SD,
mm Hg
82.6⫾9.8
83.3⫾9.9
84.1⫾10.0
6-month DBP, mean⫾SD,
mm Hg
79.8⫾9.6
80.2⫾9.8
81.4⫾10.2
1538 (46.4)
3923 (45.1)
5093 (45.8)
Lisinopril, No. (%)
870 (26.3)
2381 (27.4)
3078 (27.7)
Amlodipine, No. (%)
906 (27.3)
2398 (27.6)
2949 (26.5)
Cholesterol, mean⫾SD,
mg/dL
215.7⫾42.4
216.4⫾42.6
217.3⫾44.2
BMI, mean⫾SD, kg/m2
29.7⫾6.2
29.7⫾5.9
30.2⫾6.6
Chlorthalidone, No. (%)
145.1⫾13.9
145.2⫾13.9
Table 2. MMP3 5A/6A Promoter Polymorphism Frequencies in
Participants With No History of Stroke or MI at Baseline by Race
MMP3
Genotype
EA
(n⫽13 840)
Black
(n⫽8533)
AI/AK
(n⫽286)
Asian/PI
(n⫽37)
Other
(n⫽925)
5A/5A
22.1%
1.7%
3.2%
13.5%
10.5%
5A/6A
47.2%
19.9%
27.3%
51.3%
41.6%
6A/6A
28.9%
76.0%
67.5%
67.5%
45.7%
1.8%
2.4%
2.1%
0%
2.2%
Missing
AI indicates American Indian; AK, Alaska native; and PI, Pacific Islander.
139.9⫾16.5
140.7⫾17.4
Downloaded from http://stroke.ahajournals.org/ by guest on June 16, 2017
BMI indicates body mass index.
before the time that the doxazosin arm was halted. All other results
presented do not include data from the doxazosin group.
Results
The results of the ALLHAT trial with regard to stroke have
been previously published.24 In brief, rates of stroke were
higher in patients randomized to lisinopril than chlorthalidone
(6.3% vs 5.6%; relative risk⫽1.15; 95% CI, 1.02 to 1.30).
This finding was primarily observed in black participants and
was probably mediated, at least in part, by higher a BP in
black subjects assigned to lisinopril. The number of individuals with complete genotype and phenotype information,
after excluding 9076 individuals with a history of stroke or
MI prior to baseline and those randomized to doxazosin, was
21 309. The average follow-up time was 4.6 years. Individuals with the 5A/6A genotype had the lowest unadjusted
stroke rates (6.8 per 1000 person-years [PYs]), followed by
5A/5A individuals (7.5 per 1000 PYs). Individuals with the
6A/6A genotype had the highest stroke rates (11.1 per 1000
PYs). Table 1 shows the means and standard deviations of
baseline levels of continuous traits and percentages for
categorical variables by MMP3 5A/6A genotype. The means
and percentages for potential confounding factors were
highly similar across genotypic classes. Age, sex, black race,
SBP, DBP, total cholesterol, current smoking, and diabetes
status were significant stroke predictors (Pⱕ0.05) and were
retained as covariates in final models.
MMP3 allele frequencies differed significantly by race,
with 5A alleles much more common in European Americans
(EAs). Table 2 shows the 5A/6A genotype frequencies by
race. The overall stroke rate was 8.2 per 1000 PYs in the
whole sample and 8.0 per 1000 PYs in people with no
previous stroke or MI. There was no significant association
between any MMP3 genotype and stroke rate after multivar-
iate adjustment. Also, black participants showed a small but
significantly greater response (2.2-mm Hg change in SBP
between baseline and 6 months, P⬍0.0001) to treatment
compared with a reduction in SBP in other race groups after
adjustment for multiple factors including baseline SBP. The
MMP3 polymorphism was not in Hardy-Weinberg equilibrium in blacks (P⫽0.008) or EAs (P⫽0.00004).
In light of the difference in allele frequency between EAs
and blacks and the apparent interaction between treatment
and genotype, we wanted to determine whether race could act
as a surrogate for genotype information to inform first-line
treatment options. To do so, we compared the fit between a
model containing genotype⫻treatment interaction terms and
a model containing a genotype⫻treatment term and a genotype⫻race term, but the model fit was not significantly
improved. Also, the race⫻treatment interaction term was
highly significant, with blacks taking lisinopril showing
increased stroke risk (hazard ratio [HR]⫽1.81; 95% CI, 1.33
to 2.45; P⫽0.0002). These analyses indicate that race may be
a valid surrogate for MMP genotype for informing treatment
choice. We also tested for the genotype⫻treatment interaction in the black and EA groups separately. The interaction
results were neither significant nor distinguishable from the
combined race results (as determined by the overlap of the
95% CIs from the HRs). These results are shown in Table 1.
There was a nominally significant interaction between
MMP3 genotype and antihypertensive treatment group in
predicting stroke rates, although the associated probability
values did not meet a Bonferroni-adjusted significance
threshold. The Figure shows a plot of stroke HR by treatment
group and genotype, illustrating the observed interaction.
Each combination of genotype and drug treatment was
compared with a common reference group, 6A/6A individuals randomized to chlorthalidone. In a multivariate-adjusted
model, the interaction term between 5A/6A genotype and
lisinopril treatment was not significant (HR⫽0.74; 95% CI,
0.53 to 1.04; P⫽0.08), but the interaction term for 5A/5A
individuals taking lisinopril was nominally significant
(HR⫽0.51; 95% CI, 0.31 to 0.85; P⫽0.009). There were no
pharmacogenetic differences in stroke rate by genotype in
patients taking amlodipine versus chlorthalidone, nor doxazosin versus chlorthalidone. Also, there were no main effects of
MMP3 genotype on stroke rate in any of the subgroups, nor
were there any significant genotype⫻subgroup interactions
after correcting for multiple tests.
The HRs for stroke and probability values associated with
the MMP3-lisinopril versus chlorthalidone interaction terms
were similar whether or not we adjusted for the change in BP
Sherva et al
Pharmacogenetics of Antihypertensives and Stroke
333
Table 3. Compliance With Randomized Study Drug and Use of
Second-Line Agents at 1 and 5 Years of Follow-Up
Drug
Downloaded from http://stroke.ahajournals.org/ by guest on June 16, 2017
Figure. Stroke HRs by treatment group and MMP3 genotype.
Results were based on the treatment⫻genotype interaction term
from a Cox regression model adjusted for age, total cholesterol,
black race, sex, current smoking, SBP, DBP, and diabetes status. The MMP3 polymorphism was modeled as a 3-level (genotypic) factor. All HRs and probability values are in reference to
individuals with the 6A/6A genotype who were randomized to
chlorthalidone.
between the first visit and after 6 months of treatment. The
consistency of the findings before and after controlling for BP
lowering suggests that the differential stroke rates were not
solely due to differential BP lowering.
Discussion
We found a novel pharmacogenetic effect of the MMP3
genotype for stroke outcomes, and our findings suggest that
angiotensin-converting enzyme inhibition is a more effective
treatment than chlorthalidone for hypertensives homozygous
for the MMP3 5A allele, whereas chlorthalidone is more
efficacious than lisinopril for hypertensives heterozygous for
the 6A allele. The mechanism for this pharmacogenetic effect
is not entirely clear. Previous data indicate that 6A alleles
correspond to lower expression levels, greater matrix deposition, and more rapid progression of atherosclerosis, whereas
homozygosity for the 5A allele leads to overexpression and
plaque instability. Although not replicated in this subset of
the data, ALLHAT results indicated that in the lisinopril
group, there was a significantly increased risk of stroke
compared with chlorthalidone. This effect may be partially
mediated by MMP3 genotype. Although there was no significant association between MMP3 genotype and stroke, we
found that the 5A/6A heterozygotes had the lowest stroke
rates, consistent with previous studies of the variant’s effect
on various cardiovascular traits.10,12–22
In addition to the expected difference in stroke rates
between black and EA participants, this large, multiethnic
sample showed marked differences in several other factors
across ethnic groups. The most striking was the difference in
MMP3 5A/6A allele frequencies across ethnic groups, with
5A alleles being much more common in EAs than in the other
ethnic groups. The greater proportion of heterozygotes in EA
participants may explain part of the increased stroke rates in
blacks.
Compliant
at Year 1
Compliant
at Year 5
Additional
Drugs at Year 1
Additional
Drugs at Year 5
Chlorthalidone
84%
72%
27%
41%
Lisinopril
78%
62%
33%
43%
Doxazosin
79%
27%
31%
46%
Amlodipine
84%
73%
26%
40%
There were differences in compliance rates between some
of the study drugs and also in the number of participants
requiring a second, open-label treatment. Doxazosin and
lisinopril had lower compliance rates than did the other drugs,
as well as higher rates of additional medications required.
Table 3 shows the compliance and second-line treatment rates
by randomized drug group. Also, there were differences in
the number of participants initially assigned to each treatment
group who required an additional, open-label antihypertensive drug. At 1 and 5 years, more participants assigned to
chlorthalidone and amlodipine continued the study medication than did participants who were assigned to lisinopril. In
addition, at both 1 and 5 years, more participants assigned to
lisinopril received open-label medication (32.6% at year 1
and 43.0% at year 5) than did those assigned to chlorthalidone
(26.7% at year 1 and 40.7% at year 5) or amlodipine (25.9%
at year 1 and 39.5% at year 5). In the ALLHAT cohort,
among the 2 treatment groups of interest for this analysis,
lisinopril and chlorthalidone, 5.8% of the chlorthalidone
group were also taking a calcium channel blocker and 9.3%,
an angiotensin-converting enzyme inhibitor in addition to the
randomized drug at 5 years. In the lisinopril group, 15.7%
were taking an angiotensin-converting enzyme inhibitor with
a diuretic at 5 years.13 The most common second-line agent
for all race and treatment subgroups was atenolol (21% to
29%), followed by hydralazine (8% to 22%). In the chlorthalidone group, 25% of black and 35% of nonblack participants
were prescribed atenolol. Among blacks, 25% of the lisinopril group received a diuretic and 17%, a calcium channel
blocker. Among nonblacks, a diuretic was prescribed for 21%
of lisinopril participants.25
Strengths and Limitations
These analyses represent the largest sample to date used to
test the MMP3 5A/6A promoter polymorphism for association with stroke and for interaction with antihypertensive
treatment. Because this was a randomized, prospective study,
we avoided the limitations of case-control designs while
retaining sufficient power to detect genetic associations and
interactions with treatment and lifestyle factors.
Despite these strengths, some limitations exist. First, no
differentiation was made between stroke subtypes, and heterogeneity within the combined stroke phenotype could have
limited our power to detect an association. However, because
all stroke subtypes are caused in part by atherosclerosis, the
pathway through which MMP3 acts, any power loss would
likely be minimal. Second, although previous evidence shows
334
Stroke
February 2011
Downloaded from http://stroke.ahajournals.org/ by guest on June 16, 2017
that MMP3 genotype changes expression level, we had no
expression data to validate this in the GenHAT sample.
Although we tested for stroke risk associated with a single
genetic polymorphism with well-documented effects on gene
expression and multiple cardiovascular disease, we tested
several genetic models and for interactions with several risk
factors, and therefore, multiple comparisons become an issue.
Given that 4 modes of inheritance (3 with 1 df and 1 with 2
df) were tested, with the addition of a categorical (2 df)
genetic model for interactions with 5 variables, we conducted
15 tests in the initial analysis. A simple Bonferroni correction
for the tests conducted here would require a probability value
of 0.003 for significance, which was not met by any of the
tests. To adjust for all 78 tests conducted in the larger
GenHAT study would require a threshold of 0.0006. This
threshold is likely conservative, as the different genetic
models are not independent tests. Also, most of the variants
tested in GenHAT have established effects on protein function or expression levels. We therefore assumed more prior
evidence for true-positive associations than for the nonfunctional tag single-nucleotide polymorphisms that make up the
bulk of markers tested in many genome-wide association
studies. Still, the likelihood that these findings are falsepositives is not negligible, and independent replication is
necessary.
Confounding due to population stratification is also a
potential problem. Although analyses were adjusted for black
race, there may be additional population strata within race
groups. A recent simulation study showed that the magnitude
of this bias is small unless extreme differences exist in
genotype frequency, and the magnitude of bias decreases as
the number of admixed ethnicities increases. In addition, bias
due to population stratification may be small when baseline
risk differences are small within major categories of admixed
ethnicity.26 Finally, the MMP3 polymorphism was not in
Hardy-Weinberg equilibrium in the sample, which could
indicate a systematic genotyping error.
Summary
In summary, these data from the GenHAT study show that the
MMP3 5A/6A genotype may be a useful genetic marker for
determining who should avoid lisinopril in favor of chlorthalidone as first-line treatment for hypertension in terms of
stroke prevention. Specifically, our results suggest that lisinopril rather than chlorthalidone may not be as efficacious a
hypertension treatment option in patients with the MMP3
6A/6A genotype but is more efficacious in hypertensives with
the MMP3 5A/5A genotype. In heterozygotes, lisinopril
treatment appears identical to chlorthalidone for stroke prevention, and there was no statistical difference between stroke
rates between these drugs among MMP3 5A homozygotes.
The observed effects are quite small, however, and would
have to be replicated in an independent sample before any
changes in treatment are warranted. Our results also indicate
that race may be a potential surrogate for MMP3 genotype in
informing treatment choice.
Acknowledgments
We thank Victoria Brophy, Suzanne Cheng, Henry Erlich, Michael
Grow, and Calvin Mano at Roche Molecular Systems for kindly
providing genotyping materials and technical support.
Sources of Funding
This work was supported in part by HL63082 (GenHAT) from the
National Heart, Lung, and Blood Institute, Bethesda, MD, and
N01-HC-35130 (ALLHAT).
Disclosures
None.
References
1. Barone FC, Feuerstein GZ. Inflammatory mediators and stroke: new
opportunities for novel therapeutics. J Cereb Blood Flow Metab. 1999;
19:819 – 834.
2. Handa N, Matsumoto M, Maeda H, Hougaku H, Kamada T. Ischemic
stroke events and carotid atherosclerosis: results of the Osaka Follow-up
Study for Ultrasonographic Assessment of Carotid Atherosclerosis (the
OSACA Study). Stroke. 1995;26:1781–1786.
3. Roederer GO, Langlois YE, Jager KA, Primozich JF, Beach KW, Phillips
DJ, Strandness DE Jr. The natural history of carotid arterial disease in
asymptomatic patients with cervical bruits. Stroke. 1984;15:605– 613.
4. Gomez CR. Carotid plaque morphology and risk for stroke. Stroke.
1990;21:148 –151.
5. Seshadri S, Beiser A, Kelly-Hayes M, Kase CS, Au R, Kannel WB, Wolf
PA. The lifetime risk of stroke: estimates from the Framingham Study.
Stroke. 2006;37:345–350.
6. Strandgaard S. Hypertension and stroke. J Hypertens Suppl. 1996;14:
S23–S27.
7. Roman MJ, Saba PS, Pini R, Spitzer M, Pickering TG, Rosen S,
Alderman MH, Devereux RB. Parallel cardiac and vascular adaptation in
hypertension. Circulation. 1992;86:1909 –1918.
8. Kingwell BA, Medley TL, Waddell TK, Cole TJ, Dart AM, Jennings GL.
Large artery stiffness: structural and genetic aspects. Clin Exp Pharmacol
Physiol. 2001;28:1040 –1043.
9. Sellers A, Murphy G. Collagenolytic enzymes and their naturally
occurring inhibitors. Int Rev Connect Tissue Res. 1981;9:151–190.
10. Ye S, Eriksson P, Hamsten A, Kurkinen M, Humphries SE, Henney AM.
Progression of coronary atherosclerosis is associated with a common
genetic variant of the human stromelysin-1 promoter which results in
reduced gene expression. J Biol Chem. 1996;271:13055–13060.
11. Medley TL, Kingwell BA, Gatzka CD, Pillay P, Cole TJ. Matrix
metalloproteinase-3 genotype contributes to age-related aortic stiffening
through modulation of gene and protein expression. Circ Res. 2003;92:
1254 –1261.
12. Humphries SE, Luong LA, Talmud PJ, Frick MH, Kesaniemi YA, Pasternack A, Taskinen MR, Syvanne M. The 5A/6A polymorphism in the
promoter of the stromelysin-1 (MMP-3) gene predicts progression of
angiographically determined coronary artery disease in men in the
LOCAT gemfibrozil study. Lopid Coronary Angiography Trial. Atherosclerosis. 1998;139:49 –56.
13. Ye S, Watts GF, Mandalia S, Humphries SE, Henney AM. Preliminary
report: genetic variation in the human stromelysin promoter is associated
with progression of coronary atherosclerosis. Br Heart J. 1995;73:
209 –215.
14. Zhang B, Ye S, Herrmann SM, Eriksson P, de Maat M, Evans A, Arveiler
D, Luc G, Cambien F, Hamsten A, Watkins H, Henney AM. Functional
polymorphism in the regulatory region of gelatinase B gene in relation to
severity of coronary atherosclerosis. Circulation. 1999;99:1788 –1794.
15. Henney AM, Wakeley PR, Davies MJ, Foster K, Hembry R, Murphy G,
Humphries S. Localization of stromelysin gene expression in atherosclerotic plaques by in situ hybridization. Proc Natl Acad Sci U S A.
1991;88:8154 – 8158.
16. Galis ZS, Sukhova GK, Lark MW, Libby P. Increased expression of
matrix metalloproteinases and matrix degrading activity in vulnerable
regions of human atherosclerotic plaques. J Clin Invest. 1994;94:
2493–2503.
17. Beyzade S, Zhang S, Wong YK, Day IN, Eriksson P, Ye S. Influences of
matrix metalloproteinase-3 gene variation on extent of coronary athero-
Sherva et al
18.
19.
20.
21.
22.
sclerosis and risk of myocardial infarction. J Am Coll Cardiol. 2003;41:
2130 –2137.
Terashima M, Akita H, Kanazawa K, Inoue N, Yamada S, Ito K, Matsuda
Y, Takai E, Iwai C, Kurogane H, Yoshida Y, Yokoyama M. Stromelysin
promoter 5A/6A polymorphism is associated with acute myocardial
infarction. Circulation. 1999;99:2717–2719.
Nojiri T, Morita H, Imai Y, Maemura K, Ohno M, Ogasawara K, Aizawa
T, Saito A, Hayashi D, Hirata Y, Sugiyama T, Yamazaki T, Nagai R.
Genetic variations of matrix metalloproteinase-1 and -3 promoter regions
and their associations with susceptibility to myocardial infarction in
Japanese. Int J Cardiol. 2003;92:181–186.
Ghilardi G, Biondi ML, DeMonti M, Turri O, Guagnellini E, Scorza R.
Matrix metalloproteinase-1 and matrix metalloproteinase-3 gene
promoter polymorphisms are associated with carotid artery stenosis.
Stroke. 2002;33:2408 –2412.
Rundek T, Elkind MS, Pittman J, Boden-Albala B, Martin S, Humphries
SE, Juo SH, Sacco RL. Carotid intima-media thickness is associated with
allelic variants of stromelysin-1, interleukin-6, and hepatic lipase genes:
the Northern Manhattan Prospective Cohort Study. Stroke. 2002;33:
1420 –1423.
Rauramaa R, Vaisanen SB, Luong LA, Schmidt-Trucksass A, Penttila
IM, Bouchard C, Toyry J, Humphries SE. Stromelysin-1 and
interleukin-6 gene promoter polymorphisms are determinants of asymp-
Pharmacogenetics of Antihypertensives and Stroke
23.
24.
25.
26.
335
tomatic carotid artery atherosclerosis. Arterioscler Thromb Vasc Biol.
2000;20:2657–2662.
Cheng S, Grow MA, Pallaud C, Klitz W, Erlich HA, Visvikis S, Chen JJ,
Pullinger CR, Malloy MJ, Siest G, Kane JP. A multilocus genotyping
assay for candidate markers of cardiovascular disease risk. Genome Res.
1999;9:936 –949.
ALLHAT Officers and Coordinators for the ALLHAT Collaborative
Research Group. The Antihypertensive and Lipid-Lowering Treatment to
Prevent Heart Attack Trial. Major outcomes in high-risk hypertensive
patients randomized to angiotensin-converting enzyme inhibitor or
calcium channel blocker vs diuretic: the Antihypertensive and LipidLowering Treatment to Prevent Heart Attack Trial (ALLHAT). J Am Med
Assoc. 2002;288:2981–2997.
Cushman WC, Ford CE, Einhorn PT, Wright JT Jr, Preston RA, Davis
BR, Basile JN, Whelton PK, Weiss RJ, Bastien A, Courtney DL,
Hamilton BP, Kirchner K, Louis GT, Retta TM, Vidt DG, ALLHAT
Collaborative Research Group. Blood pressure control by drug group in
the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart
Attack Trial (ALLHAT). J Clin Hypertens (Greenwich). 2008;10:
751–760.
Wang Y, Localio R, Rebbeck TR. Evaluating bias due to population
stratification in case-control association studies of admixed populations.
Genet Epidemiol. 2004;27:14 –20.
Downloaded from http://stroke.ahajournals.org/ by guest on June 16, 2017
Pharmacogenetic Effect of the Stromelysin (MMP3) Polymorphism on Stroke Risk in
Relation to Antihypertensive Treatment: The Genetics of Hypertension Associated
Treatment Study
Richard Sherva, Charles E. Ford, John H. Eckfeldt, Barry R. Davis, Eric Boerwinkle and Donna
K. Arnett
Downloaded from http://stroke.ahajournals.org/ by guest on June 16, 2017
Stroke. 2011;42:330-335; originally published online December 23, 2010;
doi: 10.1161/STROKEAHA.110.593798
Stroke is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2010 American Heart Association, Inc. All rights reserved.
Print ISSN: 0039-2499. Online ISSN: 1524-4628
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://stroke.ahajournals.org/content/42/2/330
Data Supplement (unedited) at:
http://stroke.ahajournals.org/content/suppl/2010/12/23/STROKEAHA.110.593798.DC1
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Stroke can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.
Once the online version of the published article for which permission is being requested is located, click
Request Permissions in the middle column of the Web page under Services. Further information about this
process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Stroke is online at:
http://stroke.ahajournals.org//subscriptions/
Supplementary Data
Pharmacogenetic effect of the stromelysin (MMP3) polymorphism
on stroke risk in relation to antihypertensive treatment: The
GenHAT Study
R Sherva, CE Ford, JH Eckfeldt, BR Davis, E Boerwinkle, DK Arnett
Table S1. Complete results for all genotypic, interaction, and race-stratified models.
All models were adjusted for age, gender, AA race, total cholesterol, current smoking,
SBP, DBP, and diabetes status except for the race-stratified models which were not
adjusted for AA race.
Variable
Genotypic Models
5A any vs. none
6A any vs. none
5A/6A vs. other
Interaction Tests
5A/5A x age
5A/6A x age
5A/5A x sex
5A/6A x sex
5A/6A x SBP
5A/5A x SBP
5A/5A x diabetes
5A/6A x diabetes
5A/5A x amlodipine
5A/6A x amlodipine
5A/6A x smoking
5A/5A x smoking
5A/5A x EA
5A/6A x EA
5A/5A x AA
5A/6A x AA
Race Stratified Models
EA
5A/5A x lisinopril
5A/6A x lisinopril
AA
5A/5A x lisinopril
5A/6A x lisinopril
Beta
SE
ChiSquare
Hazard
Ratio
P-value
-0.13
-0.01
-0.15
0.08
0.10
0.08
2.43
0.00
3.38
0.88
0.99
0.87
0.12
0.95
0.07
0.03
0.00
0.11
0.02
0.00
0.00
-0.09
0.10
0.18
-0.09
-0.24
-0.62
0.29
-0.05
0.06
0.16
0.01
0.01
0.22
0.16
0.00
0.01
0.22
0.16
0.24
0.18
0.19
0.28
0.44
0.18
0.47
0.18
5.87
0.20
0.24
0.02
0.26
0.08
0.16
0.42
0.56
0.22
1.73
4.92
0.44
0.07
0.01
0.79
1.03
1.01
1.11
1.02
1.00
1.00
0.92
1.11
1.19
0.92
0.78
0.54
1.34
0.95
1.06
1.17
0.02
0.65
0.63
0.88
0.61
0.78
0.69
0.52
0.45
0.64
0.19
0.03
0.51
0.79
0.91
0.37
-0.36
-0.03
0.31
0.26
1.32
0.02
0.70
0.97
0.25
0.90
0.37
-0.20
0.92
0.30
0.16
0.44
1.45
0.82
0.69
0.51