Matrix Metalloproteinases/Tissue Inhibitors of

Metalloproteinases
Relationship Between Changes in Proteolytic Determinants of Matrix
Composition and Structural, Functional, and Clinical Manifestations
of Hypertensive Heart Disease
S. Hinan Ahmed, MD; Leslie L. Clark, MS; Weems R. Pennington, MD; Carson S. Webb, MD;
D. Dirk Bonnema, MD; Amy H. Leonardi, BS; Catherine D. McClure, RDCS;
Francis G. Spinale, MD, PhD; Michael R. Zile, MD

Background—Chronic hypertension may cause left ventricular (LV) remodeling, alterations in cardiac function, and the
development of chronic heart failure (CHF). Changes in the composition of the extracellular matrix (ECM) known to
occur in hypertension are believed to be causally related to these structural, functional, and clinical outcomes. However,
Downloaded from http://circ.ahajournals.org/ by guest on February 4, 2018

whether the determinants of ECM composition, such as the balance between ECM proteases (matrix metalloproteinases
[MMPs]) and their tissue inhibitors [TIMPs]), are altered in hypertensive heart disease is unknown.
Methods and Results—Plasma MMP-2, -9, and -13 values, TIMP-1 and -2 values, and Doppler echocardiography images
were obtained for 103 subjects divided into 4 groups: (1) reference subjects (CTL) with no evidence of cardiovascular
disease, (2) hypertensive (HTN) subjects with controlled blood pressure and no LV hypertrophy, (3) hypertensive
subjects with controlled blood pressure and with LV hypertrophy (HTN⫹LVH) but no CHF, and (4) hypertensive
subjects with controlled blood pressure, LVH, and CHF (HTN⫹LVH⫹CHF). Compared with CTL, patients with HTN
had no significant changes in any MMP or TIMP. Patients with HTN⫹LVH had decreased MMP-2 and MMP-13 values
and increased MMP-9 values. Only patients with HTN⫹LVH⫹CHF had increased TIMP-1 values. A TIMP-1 level
⬎1200 ng/mL was predictive of CHF.
Conclusions—Patients with hypertension but normal LV structure and function had normal MMP/TIMP profiles. Changes
in MMP profiles that favor decreased ECM degradation were associated with LVH and diastolic dysfunction. An
increased TIMP-1 level predicted the presence of CHF. Although these findings should be confirmed in a larger
prospective study, these data do suggest that changes in the MMP/TIMP balance may play an important role in the
structural, functional, and clinical manifestations of hypertensive heart disease. (Circulation. 2006;113:2089-2096.)
Key Words: heart failure 䡲 hypertension 䡲 hypertrophy 䡲 matrix metalloproteinases

E pidemiological studies have shown that chronic arterial

hypertension causes left ventricular (LV) structural re-
modeling, significant alterations in cardiac function, and the
development of chronic heart failure (CHF).1– 8 The structural
and functional manifestations of hypertensive LV remodeling
have been associated with significant changes in the extra-
cellular matrix (ECM) composition.7,9 –12 For example, exper-
imental and clinical studies have shown that hypertensive
heart disease can result in increased fibrillar collagen content,
altered fibrillar collagen geometry, and an increased collagen
I to III isotype ratio.7,9 –12 However, the specific molecular
and biochemical determinants that contribute to this ECM
remodeling process in patients with hypertensive heart dis-
Clinical Perspective p 2096
ease have not been completely defined. Two of the major
determinants that alter collagen homeostasis are the rate and
extent of collagen degradation. Collagen degradation is under
the control of matrix metalloproteinases (MMPs), a family of
zinc-dependent interstitial enzymes, and their tissue inhibitors
(TIMPs).13,14 In hypertensive heart disease, if TIMPs are
increased and MMPs are decreased, this would favor de-
creased collagen degradation and increased collagen accumu-
lation. The present study focused on one pathophysiological
determinant of ECM composition, the regulatory role played
by changes in the ECM proteolytic system, ie, the balance
between MMPs and TIMPs. The purpose of this study was to

Received July 7, 2005; revision received January 30, 2006; accepted February 17, 2006.
From the Departments of Medicine (S.H.A., W.R.P., C.S.W., D.D.B., C.D.M., M.R.Z.); Biostatistics, Bioinformatics, and Epidemiology (L.L.C.); and
Surgery (L.L.C., A.H.L., F.G.S.), Medical University of South Carolina and RHJ Department of Veterans Affairs Medical Center, Charleston, SC.
Reprint requests to Michael R. Zile, MD, Medical University of South Carolina, Cardiology/Medicine, 135 Rutledge Ave, Ste 1201, PO Box 250592,
Charleston, SC 29425. E-mail zilem@musc.edu
© 2006 American Heart Association, Inc.
Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.105.573865

2089
 2090 Circulation May 2, 2006

TABLE 1. Demographic, LV Structure/Function, and increase in the TIMPs, a pattern known to impair ECM
MMP/TIMP Data turnover and favor ECM accumulation.
Reference Controls LVH Patients
(n⫽53) (n⫽50) Methods
Sex, male/female 20/33 24/26 Subjects
Age, y 59⫾1 60⫾2 Two groups of subjects were recruited into this study: reference
controls and patients with LV hypertrophy (LVH). Reference con-
Systolic blood pressure, mm Hg 127⫾2 137⫾3*
trols were identified from locally sponsored health fairs and volun-
Diastolic blood pressure, mm Hg 75⫾1 76⫾2 teers from the Medical University of South Carolina staff. Of the
EDV, mL/m2 51⫾2 52⫾2 reference controls screened, 35% were enrolled, 50% had 1 of the
exclusion criteria listed next, and 15% declined participation. LVH
Ejection fraction, % 66⫾1 72⫾2*
patients were identified from echocardiographic studies. Of the
2
LV mass, g/m 99⫾3 162⫾6* patient echocardiograms screened, 10% were enrolled, 75% had 1 of
Volume/mass, mL/g 0.54⫾0.02 0.32⫾0.01* the exclusion criteria listed below, and 15% declined participation.
There were some exclusion criteria common to both groups: (1)
Mitral E/A 0.95⫾0.04 0.91⫾0.05
history of myocardial infarction; (2) regional wall-motion abnormal-
Isovolumic relaxation time, ms 83⫾2 91⫾3* ity; (3) coronary revascularization surgery; (4) amyloidosis, sarcoid-
E-wave deceleration time, ms 208⫾8 234⫾10* osis, HIV, hypertrophic obstructive cardiomyopathy, or valvular
heart disease; (5) ejection fraction ⬍50%; (6) malignancy; (7)
Tissue doppler E⬘, cm/s 10.1⫾0.4 7.4⫾0.4*
significant renal or hepatic dysfunction; (8) rheumatological disease;
Downloaded from http://circ.ahajournals.org/ by guest on February 4, 2018

PCWP, mm Hg 10⫾1 16⫾1* or (9) blood pressure ⬎140/90 mm Hg.

MMP-2, ng/mL 1387⫾39 1205⫾44* One hundred three subjects were enrolled in this study: 53
reference control subjects and 50 subjects with evidence of LVH (LV
MMP-9, ng/mL 13⫾3 26⫾3* wall thickness ⬎1.2 cm and/or LV mass index ⱖ125 gm/m2; Table
TIMP-1, ng/mL 997⫾36 1291⫾70* 1). The reference control subjects were subdivided into 2 groups on
TIMP-2, ng/mL 44⫾4 58⫾7 the basis of the presence or absence of hypertension; 39 control
subjects (referred to as “reference controls without hypertension”)
Data are mean⫾SEM. had no history of hypertension, no evidence of cardiovascular
*P⬍0.05 compared with reference controls. disease, no symptoms or physical evidence of cardiovascular disease,
no cardiovascular medication use, and all echocardiographic mea-
test the hypotheses that (1) changes in the ECM proteolytic surements within the normal range (Table 2); and 14 patients
system are measurable by plasma assays of selected MMPs (referred to as “reference controls with hypertension”) had a history
and TIMPs; (2) each manifestation of hypertensive heart of arterial hypertension, controlled blood pressure (pharmacologi-
cally treated to meet JNC 7 criteria, ie, ⬍140/90 mm Hg), no LVH,1
disease is associated with a specific pattern of changes in the and all echocardiographic measurements within the normal range
ECM proteolytic system; and (3) hypertensive patients with (Table 2).
structural remodeling, diastolic dysfunction, and/or clinical LVH patients were subdivided into 2 groups on the basis of the
CHF are characterized by a decrease in the MMPs and an presence or absence of CHF. Twenty-three patients with hyperten-

TABLE 2. Reference Controls With and Without Hypertension and LVH Patients With
and Without CHF
Reference Control Reference Control
Without Hypertension With Hypertension LVH Without CHF LVH With CHF
(n⫽39) (n⫽14) (n⫽23) (n⫽26)
Systolic blood pressure, mm Hg 126⫾3 131⫾4 138⫾3* 133⫾4
Diastolic blood pressure, mm Hg 74⫾2 77⫾2 82⫾2* 72⫾2‡
EDV, mL 97⫾3 94⫾5 98⫾6 104⫾5
Ejection fraction, % 65⫾1 66⫾1 70⫾2* 73⫾2*
2
LV mass, g/m 94⫾5 101⫾3 160⫾7*† 164⫾7*†
Mitral E/A 0.98⫾0.05 0.85⫾0.05 0.80⫾0.09* 0.97⫾0.07‡
Tissue Doppler E⬘, cm/s 10.0⫾0.4 9.8⫾0.5 8.4⫾0.4*† 7.2⫾0.5*†‡
PCWP, mm Hg 10⫾1 11⫾1 13⫾2 17⫾2*†‡
PCWP/EDV, mm Hg/mL 0.09⫾0.01 0.11⫾0.01 0.12⫾0.01 0.17⫾0.01*†‡
Ea, mm Hg/mL 1.50⫾0.05 1.61⫾0.09 1.67⫾0.10* 1.45⫾0.11‡
MMP-2, ng/mL 1383⫾44 1399⫾84 1119⫾48*† 1286⫾73
MMP-9, ng/mL 13⫾4 14⫾5 27⫾3*† 24⫾4*†
TIMP-1, ng/mL 1000⫾42 988⫾76 1092⫾77 1364⫾86*†‡
TIMP-2, ng/mL 42⫾4 48⫾7 58⫾11 59⫾9
Ea indicates effective arterial elastance. Data are mean⫾SEM.
Significant differences among all 4 groups were analyzed by ANOVA and Tukey’s multiple comparison tests:
*P⬍0.05 vs reference controls without hypertension; †P⬍0.05 vs reference controls with hypertension; ‡P⬍0.05 vs
LVH without CHF.
 Ahmed et al
TABLE 3. Plasma MMP/TIMP in Published Studies of Patients With Hypertension
Present Study or First
Author (Reference No.) MMP-2
Untreated hypertension
Yasmin (22) 1 1
Tayebjee (23-25)
Li-Saw-Hee (29)
Zervoudaki (30) 2 2
Timms (28)
Lindsay (27)
Laviades (26)
Treated hypertension c/w PreTx (c/w Ctl)
Tayebjee (24, 25)
Li-Saw-Hee (29)
Zervoudaki (30) 7 (2)
Laviades (26)
Downloaded from http://circ.ahajournals.org/ by guest on February 4, 2018
Present study (2)
c/w PreTx indicates compared with pretreatment value (ie, before institution of antihypertensive medications; c/w
Ctl, compared with normal age- and sex-matched reference control values.
sion, controlled blood pressure, and LVH but no CHF were referred
to as “LVH without CHF” (Table 2). The second subgroup consisted
of 26 patients with hypertension, controlled blood pressure, LVH,
and CHF and were referred to as “LVH with CHF.” All of these
patients had evidence of CHF as defined according to Framingham
criteria,2 evidence of abnormal relaxation (decreased E⬘), increased
stiffness (increased pulmonary capillary wedge pressure [PCWP]
and increased PCWP– end-diastolic volume [EDV] ratio), a mark-
edly reduced 6-minute walk distance (979⫾86 feet in the
LVH⫹CHF group compared with 1839⫾60 feet in the LVH⫺CHF
group; P⬍0.05), ejection fraction ⱖ50%, and therefore, had diastolic
heart failure.
The medications used to treat hypertension were chosen and
monitored by the patient’s primary physician and not the investiga-
tors. These included diuretics, renin-angiotensin-aldosterone antag-
onists (angiotensin-converting enzyme inhibitors, angiotensin II
receptor blockers, and aldosterone blockers), direct vasodilators
(nitrates and hydralazine), ␣-adrenergic blockers, central nervous
system blockers, aspirin, ␤-adrenergic receptor blockers, and cal-
cium channel blockers. The mean duration of antihypertensive
treatment was 6.4⫾1.5 years.
Echocardiographic Methods
Echocardiography was performed with a Sonos 5500 system (Agi-
lent Technologies, Andover, Mass) with a 4-MHz transducer. Mea-
surements were made according to American Society of Echocardi-
ography criteria.15,16 LV and left atrial volumes were calculated with
the method of discs.16 LV mass was calculated according to the
formula of Devereux et al.17 Doppler measurements of mitral inflow
E- and A-wave velocity, the E-A ratio, E-wave deceleration time,
and isovolumic relaxation time were made. Tissue Doppler (lateral
mitral annulus) measurements of mitral E⬘ -and A⬘-wave velocity
were made. PCWP was calculated from the formula 2⫹1.3(E/E⬘).18
Effective arterial elastance was calculated from the formula end-sys-
tolic pressure/stroke volume.
MMP/TIMP Plasma Measurements
Gelatinases (MMP-2 and -9), collagenase (MMP-13), and tissue inhib-
itors of MMPs (TIMP-1 and -2) were examined with the use of 2-site
ELISA kits (Amersham Pharmacia Biotech, Buckinghamshire, UK).
Plasma and the respective MMP standards were added to precoated
wells containing the antibody to the MMP or TIMP of interest and
washed. The resultant reaction was read at a wavelength of 450 nm
(Labsystems Multiskan MCC/340, Helsinki, Finland). Because
MMP/TIMP Profiles in Hypertensive Heart Disease 2091
MMP-9 TIMP-1 TIMP-2 MMP-13
7
1 1
2 2
1
1
1
c/w PreTx (c/w Ctl) c/w PreTx (c/w Ctl)
2 (?) 1 (1 1)
7 (2) 7 (2)
1 (?)
2 (1)
(1) (1) (7) (2)
MMP-13 was found in very low levels in plasma, the MMP-13 results
were categorized as either detectable or nondetectable.
Statistical Analysis
MMPs and TIMPs were measured every 2 hours for 6 hours to calculate
a coefficient of variance for MMP/TIMP measurements between and
within individual subjects in a subgroup of reference control subjects
(n⫽20) with a 1-way random-effects ANOVA. Then the coefficient was
calculated as the square root of the within-person mean square er-
ror⫻100. The intrapatient coefficient of variation for MMP-2 was
11.2⫾1.1%; for TIMP-1, 8.5⫾2.2%; and for TIMP-2, 14.3⫾1.7%. An
intra-assay coefficient of variation that quantified variation in the assay
technique itself was ⬍6% for all MMPs and TIMPs.
Initially, comparisons between reference controls versus LVH
subjects were made with a 2-tailed Student t test. Subsequently,
comparisons between all 4 groups (reference control with versus
without hypertension versus LVH with versus without CHF) were
analyzed by ANOVA and Tukey multiple-comparison tests. A value
of P⬍0.05 was considered significant. Simple linear regression was
used to examine the relation between MMP and TIMP levels and
measurements of LV structure and function. A Mantel Haenszel ␹2
and receiver operating curves were used to evaluate the association
between MMP-13 and TIMP-1 levels and the presence of LVH and
CHF. The potential effects of the medications on structure, function,
or plasma data were examined first by a univariate and then by a
multivariate regression analysis. The structure, function, MMP, or
TIMP measurement was the dependent variable, with the medication
entered as a dummy variable. A single drug was examined, and then
drugs in combination were examined.
The research protocol used in this study was reviewed and
approved by the institutional review board at the Medical University
of South Carolina. Written, informed consent was obtained from all
participants. The authors had full access to the data and take
responsibility for its integrity. All authors have read and agree to the
manuscript as written.
Results
Reference Control Versus LVH
Structure/Function Data
The reference control subjects had an age and sex distribution
similar to the LVH subjects (Table 1). Compared with
reference controls, LVH subjects had higher systolic blood
 2092 Circulation May 2, 2006
Figure 1. MMP-13 detectability in reference controls with and
without hypertension (HTN) and in LVH patients with and with-
out CHF. MMP-13 detectability decreased significantly in LVH
patients. *P⬍0.05 vs reference controls without hypertension.
Downloaded from http://circ.ahajournals.org/ by guest on February 4, 2018
#P⬍0.05 vs reference controls with hypertension.
pressures, significant concentric remodeling as evidenced by
a 60% greater LV mass index, no difference in EDV, and a
40% lower LV EDV–versus–mass ratio. Compared with
reference controls, LVH subjects had significant abnormali-
ties in the indices of LV diastolic relaxation and LV diastolic
stiffness: increased isovolumic relaxation time, increased
E-wave deceleration time, decreased E⬘, increased PCWP,
and an increased PCWP–versus–LV EDV ratio
(0.16⫾0.01 mm Hg/mL in LVH) compared with reference
controls (0.09⫾0.01 mm Hg/mL, P⬍0.05), suggesting that
there was an increase in the LV instantaneous end-diastolic
operating stiffness.
MMP and TIMP Plasma Profiles
Compared with reference controls, MMP-2 was decreased
and MMP-9 was increased in LVH subjects. Significant
differences were found in MMP-13 detectability (Figure 1).
Forty-seven percent of the reference control subjects had a
detectable level of MMP-13, whereas MMP-13 was detect-
able in only 15% of the LVH subjects (␹2⫽17.89, P⬍0.001;
odds ratio, 0.24). The plasma TIMP-1 value was significantly
increased in LVH compared with reference control subjects.
TIMP-2 and the MMP-9 –TIMP-1 and MMP-2–TIMP-2 ra-
tios were not different between reference control and LVH
subjects.
Reference Controls Without Hypertension Versus
Reference Controls With Hypertension
Structure/Function Data
Reference control subjects without hypertension served as the
age- and sex-matched reference control group for comparison
with the reference control with hypertension, the LVH
without CHF, and the LVH with CHF groups. There were no
significant differences in any demographic parameter or any
echocardiographic measurement of LV structure or function
between reference controls without hypertension versus ref-
erence controls with hypertension (Table 2). Left atrial
maximum volume (LAMV) and emptying fraction (LAEF)
were similar in reference controls without hypertension
(LAMV, 40⫾2 mL and LAEF, 42⫾3%) compared with
reference controls with hypertension (LAMV, 42⫾4 mL and
LAEF, 43⫾2%).
MMP and TIMP Plasma Profiles
There were no significant differences in any MMP or TIMP
plasma level between reference control subjects without
hypertension versus reference controls with hypertension.
LVH Without CHF Versus LVH With CHF
Structure/Function Data
There were no significant differences in systolic blood
pressure, LV volume, or mass between LVH without CHF
and LVH with CHF subjects (Table 2). However, diastolic
function was significantly more impaired in those with
LVH with CHF compared with LVH without CHF. Indices
of diastolic relaxation were slower, diastolic stiffness was
greater, and filling pressures were higher in LVH with
CHF compared with LVH without CHF. In particular, in
the LVH without CHF patients, tissue Doppler E⬘ was
decreased (8.4⫾0.4 cm/s with 95% confidence intervals
[CIs] of 7.4 to 9.3) compared with reference controls
without hypertension (10⫾0.4 cm/s; 95% CI, 9.3 to 11)
and reference controls with hypertension (9.8⫾0.5 cm/s;
95% CI, 8.1 to 11). E⬘ fell further in LVH with CHF
(7.2⫾0.5 cm/s; 95% CI, 6.2 to 8.3). In the LVH without
CHF patients, PCWP was unchanged (13⫾2 mm Hg; 95%
CI, 10.5 to 15.2) compared with reference controls without
hypertension (10⫾1 mm Hg; 95% CI, 9.3 to 10.6) and
reference controls with hypertension (11⫾1 mm Hg; 95%
CI, 9.1 to 12.2) but was increased in LVH with CHF
(17⫾2 mm Hg; 95% CI, 15.2 to 17.7). The PCWP–
versus–LV EDV ratio was unchanged in the LVH without
CHF patients but was significantly increased in the LVH
with CHF patients. Effective arterial elastance was in-
creased in LVH without CHF and was decreased in LVH
with CHF groups. LAMV was increased in LVH without
CHF subjects (LAMV, 53⫾4 mL; P⬍0.05, compared with
reference controls) and increased further in LVH with CHF
subjects (LAMV, 70⫾5 mL; P⬍0.05, compared with LVH
without CHF). LAEF was unchanged in the LVH with
CHF group (LAEF, 42⫾3%; P⬍0.05, compared with
reference controls) but was increased in LVH with CHF
(LAEF, 48⫾2% compared with LVH without CHF).
MMP and TIMP Plasma Profiles
There were no significant differences in the values of
MMP-2, -9, -13; TIMP-2; or MMP-TIMP ratios in LVH
without CHF compared with LVH with CHF (Figure 1).
However, TIMP-1 was significantly increased in LVH with
CHF (1364⫾86 ng/mL; 95% CI, 1185 to 1543) compared
with LVH without CHF (1092⫾77 ng/mL; 95% CI, 933 to
1252). In fact, TIMP-1 was elevated only in subjects with
CHF. TIMP-1 was unchanged in the LVH without CHF
patients compared with reference controls without hyperten-
sion (1000⫾42 ng/mL; 95% CI, 915 to 1085) and reference
controls with hypertension (988⫾76 ng/mL; 95% CI, 824 to
1152).
 Ahmed et al MMP/TIMP Profiles in Hypertensive Heart Disease
Downloaded from http://circ.ahajournals.org/ by guest on February 4, 2018
Figure 2. A, Relation between TIMP-1 and the LV volume-mass
ratio. Higher levels of TIMP-1 were associated with lower values
of LV volume-mass ratio, indicating more pronounce concentric
remodeling. r⫽⫺0.56, P⬍0.05. B, Relation between TIMP-1 and
TDI E⬘. Higher levels of TIMP-1 were associated with lower val-
ues of E⬘, indicating a slower LV diastolic relaxation rate.
r⫽⫺0.41, P⬍0.05. TDI indicates tissue Doppler imaging; E⬘,
peak early diastolic velocity of mitral annulus.
Relationship Between MMP and TIMP Plasma
Profiles and LV Structure and Function
There was a significant relation between TIMP-1 and the
extent of LV remodeling. As TIMP-1 increased, LV mass
increased (r⫽0.30, P⫽0.005), and the volume-mass ratio fell
(r⫽⫺0.56, P⫽0.001; Figure 2A). There was a significant
relation between TIMP-1 and the extent of diastolic dysfunc-
tion. As TIMP-1 increased, the mitral E-A ratio decreased
(r⫽⫺0.22, P⬍0.027), E⬘ fell (r⫽⫺0.62, P⫽0.001; Figure
2B), and PCWP increased (r⫽0.28, P⫽0.013). Finally, there
was a significant relation between the extent of CHF and
TIMP-1 levels. The mean value of TIMP-1 was higher in
LVH subjects with CHF who were in New York Heart
Association class III versus class II. Having a TIMP-1 level
⬎1200 ng/mL was predictive of having LVH with CHF
(␹2⫽4.6, P⫽0.03; specificity, 88%; positive predictive value,
94%, odds ratio, 3.54; 95% CI, 1.08 to 11.50). The area under
the receiver operator curve was 0.71.
There was no relation between the use of a specific
medication and differences in LV structure, function, or
plasma MMP/TIMP profiles between groups. Specifically,
there were no differences in any MMP or TIMP level between
patients grouped by any medication or combination of med-
ications. Nonetheless, it is recognized that this study was not
2093
powered sufficiently to completely address the effects of
drugs on LV structure, function, or plasma MMP/TIMP
profiles. Therefore, these data and analyses must be inter-
preted with appropriate caution.
Discussion
There are 3 unique findings in this study: (1) Patients with
hypertension but normal LV structure and function had a
normal MMP/TIMP profile, (2) changes in MMP and TIMP
profiles that favor decreased ECM degradation (decreased
MMP-2 and -13 and increased TIMP-1) were associated with
LVH and diastolic dysfunction, and (3) increased TIMP-1
predicted the presence of CHF.
Although pleiotropic in their substrates and actions,
changes in myocardial MMPs and TIMPs have predictable
effects on the ECM.13,14 For example, MMP-2 (a gelatinase)
degrades basement membrane proteins, fibrillar collagen
peptides, and newly synthesized collagen fibers. In the
present study, MMP-2 was significantly decreased in patients
with hypertensive LVH. MMP-9 (a gelatinase) has the same
structural protein substrates as MMP-2 but has a much lower
level of activity. However, MMP-9 has significant effects on
important biologically active proteins and peptides, such as
transforming growth factor-␤, and other “profibrotic” pro-
teins and pathways. Activation of profibrotic pathways by
increased MMP-9 would be expected to increase ECM
accumulation. Thus, the decreased MMP-2 and increased
MMP-9 levels found in LVH patients in the present study
may be one factor contributing to the observed structural and
functional changes seen in hypertensive heart disease.
MMP-13 is a collagenase that is found at very low levels in
plasma and is difficult to quantify accurately, even with a
high-sensitivity assay. Therefore, in the present study, rather
than reporting MMP-13 as a measured value, the results were
dichotomized. Detectable MMP-13 in the plasma of patients
with LVH was greatly reduced and was further reduced in
patients with LVH and CHF. The reduction in this collag-
enolytic enzyme would be expected to cause reduced fibrillar
collagen turnover, reduced degradation, and increased ECM
accumulation.
MMP activity is regulated at several levels that include not
only transcriptional regulation but also posttranslational mod-
ification, such as TIMP binding. The TIMPs bind to active
MMPs in a 1:1 relationship, inhibit MMP enzymatic activity,
and thereby form an important control point with respect to
net ECM proteolytic activity.13,14,19 The present study showed
that plasma levels of TIMP-1 increased in patients with LVH
and CHF. As a result, the balance between MMPs and TIMPs
was altered in favor of reduced ECM proteolytic activity,
which would therefore facilitate ECM accumulation. There
are 4 known TIMPs, and the transcriptional regulation of
these molecules is not homogeneous.19 Discordant levels of
TIMPs have been observed both in animal models of heart
failure and in patients with cardiomyopathic disease.20,21 In
the present study, a robust increase in TIMP-1 was observed
in LVH patients with CHF. In contrast, only a small increase
in TIMP-2 was observed in LVH patients with or without
CHF. These observations likely underscore the different
functions and regulatory pathways for TIMPs in the LV
 2094 Circulation May 2, 2006
remodeling process. A unique finding of the present study
was that a specific type of TIMP, TIMP-1, was strongly
associated with the development of CHF. In patients with
LVH and CHF, it is not clear whether the increased TIMP-1
levels contributed to the development of CHF or were the
result of its development. What is clear, however, is that
increased TIMP-1 was uniquely present in patients with LVH
and CHF, and a plasma TIMP-1 value ⬎1200 ng/mL was
predictive for CHF. Therefore, this plasma analyte should be
considered in the development of diagnostic criteria for heart
failure with a normal ejection fraction (diastolic heart failure)
and for the design of novel therapeutic management strategies
for diastolic heart failure. However, it is recognized that the
partition value of TIMP-1 at 1200 ng/mL was chosen in a post
hoc rather then a prospective fashion. Therefore, the validity
of its predictive value must be interpreted with appropriate
caution and confirmed in additional studies that use a large,
prospective, serial study design.
Downloaded from http://circ.ahajournals.org/ by guest on February 4, 2018
The changes in MMP/TIMPs that occur in patients with
hypertensive heart disease may effect growth regulation in
both the extracellular and cardiomyocyte compartments,
which together result in concentric LVH and increased
collagen content. Collagen homeostasis is determined by the
balance between synthesis, posttranslational modification,
and degradation. In hypertensive heart disease, Diez et al10
and Lopez et al11,12 have shown that increased collagen
content was associated with increased plasma markers of
collagen synthesis, decreased collagen degradation, and de-
creased collagen turnover. Changes in the MMP/TIMP pro-
files found in the present study suggest potential mechanisms
by which changes in synthesis, degradation, and turnover
may take place.
Although there are many determinants of LV structural
remodeling, blood pressure is one of the most important.
However, data from the present study suggest that even after
blood pressure has been adequately controlled, ongoing
changes in MMPs and TIMPs may predict, probably deter-
mine, and are certainly associated with persistent concentric
remodeling, LVH, and diastolic heart failure. Regression of
LVH requires appropriate remodeling of the ECM, including
degradation and turnover of ECM components (particularly
the basement membrane proteins) and alterations in cardio-
myocyte–matrix interactions. The present study has shown
that patients with hypertensive LVH had persistent abnormal-
ities in specific MMP (decreased MMP-2) and TIMP (in-
creased TIMP-1) profiles, which would be expected to favor
continued cardiomyocyte– basement membrane–matrix con-
nections and not the ECM turnover necessary to accommo-
date LV mass regression. It seems likely, therefore, that the
ongoing changes in MMPs and TIMPs seen in the present
study contribute to the phenotypic and structural changes
present in hypertensive heart disease.
Table 3 summarizes some of the previous clinical studies
that have examined changes in plasma MMP and TIMP
profiles in patients with hypertension.22–30 There does not
appear to be any consistent pattern of changes in MMP/
TIMPs across these studies. Some of this variability may be
based on the methods used to make the plasma measure-
ments. For example, some studies used ELISA methods,
whereas other studies used zymography and reverse zymog-
raphy to measure MMP/TIMP plasma levels. Zymography
has been successfully used in tissue extracts, but there are
some limitations to this approach, particularly for plasma
samples. First, a number of binding proteins exist within the
plasma, which covalently bind MMPs and make their extrac-
tion from plasma samples difficult. Second, plasma protein
binding of MMPs and TIMPs will yield multiple bands on
zymography, which make quantification and identification of
specific MMP and TIMP types problematic. Third, the
zymographic approach does not readily yield an absolute
MMP or TIMP value but rather yields relative comparative
values. In the present study, a validated and calibrated
high-sensitivity ELISA provided quantitative assessment of
specific MMPs and TIMPs. In addition to these technical
considerations, MMP/TIMP levels may change, dependent on
ambient blood pressure, the antihypertensive medications
used to treat hypertension, the development of LVH, changes
in LV afterload and preload, and the development of clinical
heart failure.31,32 Because changes in each of these factors
may alter MMPs and TIMPs, each was assessed in the present
study. Most of the previous studies, however, did not specif-
ically measure the changes in LV structure, LV function, LV
load, or clinical heart failure status in their patients. As a
consequence of these differences in experimental design,
direct comparisons between the present and previous studies
with regard to MMP and TIMP levels must be done with
caution. Nevertheless, Table 3 attempts to place previous
studies in context with the present study.
The present study used plasma levels of MMPs and TIMPs
as surrogate markers to reflect changes in myocardial levels
of these enzymes and peptides. However, there are several
limitations to this approach that must be acknowledged. First,
MMP activation and TIMP binding are compartmentalized
processes that occur within the myocardial interstitium.13,14
Thus, plasma levels do not necessarily reflect the net ECM
proteolytic activity that occurs within the myocardium. For-
tunately, data from previous clinical and experimental studies
suggest that the differences in plasma MMP and TIMP levels
observed between reference controls and patients with hyper-
tensive heart disease in the present study are likely to reflect
differences at the myocardial level.33–35 A second limitation
to plasma sampling of MMPs and TIMPs is that it is possible
that the myocardium is not the only source of MMPs and
TIMPs in LVH patients. Therefore, measurements of plasma
MMP and TIMP levels represent the summation of MMPs
and TIMPs released from both cardiac as well as noncardiac
sources. However, the specific exclusion criteria in the
present study helped eliminate significant changes in the
major noncardiac sources of MMPs and TIMPs. Neverthe-
less, it must be recognized that patients with hypertension and
LVH, with or without CHF, may have changes in other
noncardiac tissues, such as the kidneys and vasculature, that
may contribute to MMP and TIMP release into the plasma.
We clearly recognize, however, that the findings of the
present study that demonstrate differences in plasma MMP
and TIMP levels between reference controls and LVH pa-
tients are associative findings. Whether these changes are
 Ahmed et al MMP/TIMP Profiles in Hypertensive Heart Disease
mechanistically related to myocardial ECM accumulation
warrants further study.
Conclusions
A specific pattern of changes in the ECM proteolytic system
was associated with each structural, functional, and/or clini-
cal manifestation of hypertensive heart disease. Subjects with
adequately controlled blood pressure with no structural or
functional changes in the LV did not have any changes in the
MMP/TIMP signature. However, patients with LVH despite
adequate blood pressure control had decreased MMP-2 and
-13 values. Increases in TIMP-1 were found in patients with
LVH and CHF. In particular, the transition from hypertensive
LVH to the development of CHF may be heralded by changes
in MMPs and TIMPs, such as an increase in TIMP-1 ⬎1200
ng/mL or the absence of MMP-13. However, the present
study had a limited sample size, used a cross-sectional design,
and did not perform serial studies over time. These limitations
Downloaded from http://circ.ahajournals.org/ by guest on February 4, 2018
mandate that our observations be further tested and confirmed
in a large, prospective, serial study design. Nonetheless, the
data from the present study suggest that the observed stochas-
tic changes in MMP/TIMPs may play an important role in the
manifestations of hypertensive heart disease. Understanding
this ECM-dependent pathophysiology may lead to improved
diagnosis and treatment of patients with hypertensive heart
disease.
Acknowledgments
This study was supported by a Heart Failure Society of America
Fellowship Award (S.H. Ahmed) and grants from the Research
Service of the Department of Veterans Affairs (M.R. Zile, F.G.
Spinale) and the National Heart, Lung, and Blood Institute (PO1-
HL-48788: M.R. Zile, F.G. Spinale; RO1-HL-59165: F.G. Spinale;
MO1-RR-01070-251: M.R. Zile, F.G. Spinale).
Disclosures
None.
References
1. Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL
Jr, Jones DW, Materson BJ, Oparil S, Wright JT Jr, Roccella EJ; National
Heart, Lung, and Blood Institute Joint National Committee on Prevention,
Detection, Evaluation, and Treatment of High Blood Pressure; National
High Blood Pressure Education Program Coordinating Committee. The
seventh report of the Joint National Committee on Prevention, Detection,
Evaluation, and Treatment of High Blood Pressure: the JNC 7 report.
JAMA. 2003;289:2560 –2572.
2. Levy D, Larson MG, Vasan RS, Kannel WB, Ho KKL. The progression
from hypertension to congestive heart failure. JAMA. 1996;275:
1557–1562.
3. Lloyd-Jones DM, Larson MG, Leip EP, Beiser A, D’Agostino RB,
Kannel WB, Murabito JM, Vasan RS, Benjamin EJ, Levy D. Lifetime
risk for developing congestive heart failure: the Framingham Study.
Circulation. 2002;106:3068 –3072.
4. Schillaci G. Pasqualini L, Verdecchia P, Vaudo G, Marchesi S, Porcellati
C, De Simone G, Mannarion E. Prognostic significance of left ventricular
diastolic dysfunction in essential hypertension. J Am Coll Cardiol. 2002;
39:2005–2011.
5. Kenchaiah S, Pfeffer MA. Cardiac remodeling in systemic hypertension.
Med Clin North Am. 2004;88:115–130.
6. Zile MR, Brutsaert DL. New concepts in diastolic dysfunction and dia-
stolic heart failure, part I: diagnosis, prognosis, measurements of diastolic
function. Circulation. 2002;105:1387–1393.
7. Zile MR, Brutsaert DL. New concepts in diastolic dysfunction and dia-
stolic heart failure, part II: causal mechanisms and treatment. Circulation.
2002;105:1503–1508.
2095
8. Wachtell K, Smith G, Gerdts E, Dahlof B, Nieminen MS, Papademetriou
V, Bella JN, Ibsen H, Rokkedal J, Devereux RB. Left ventricular filling
patterns in patients with systemic hypertension and left ventricular hy-
pertrophy (the Life Study). Am J Cardiol. 2000;85:466 – 472.
9. Weber KT, Sun Y, Guarda E. Structural remodeling in hypertensive heart
disease and the role of hormones. Hypertension. 1994;23:869 – 877.
10. Diez J, Querejeta R, Lopez B, Gonzalez A, Larman M, Martinez Ubago
JL. Losartan-dependent regression of myocardial fibrosis is associated
with reduction of left ventricular chamber stiffness in hypertensive
patients. Circulation. 2002;105:2512–2517.
11. Lopez B, Querejeta R, Varo N, Gonzalez A, Larman M, Ubago JLM,
Diez J. Usefulness of serum carboxy-terminal propeptide of procollagen
type I in assessment of the cardioreparative ability in antihypertensive
treatment in hypertensive patients. Circulation. 2001;104:286 –291.
12. Lopez B, Gonzalez A, Varo N, Laviades C, Querejeta R, Diez J. Bio-
chemical assessment of myocardial fibrosis in hypertensive heart disease.
Hypertension. 2001;38:1222–1226.
13. Spinale FG. Matrix metalloproteinases: regulation and dysregulation in
the failing heart. Circ Res. 2002;90:520 –530.
14. Chapman RE, Spinale FG. Extracellular protease activation and
unraveling of the myocardial interstitium: critical steps toward clinical
applications. Am J Physiol. 2004;286:H1–H10.
15. Sahn DJ, DeMaria A, Kisslo J, Weyman A. Recommendations regarding
quantitation in M-mode echocardiography: results of a survey of echo-
cardiographic measurements. Circulation. 1978;58:1072–1083.
16. Schiller NB, Shah PM, Crawford M, DeMaria A, Devereux R,
Feigenbaum H, Gutgesell H, Reichek N, Sahn D, Schnittger I. Recom-
mendations for quantitation of the left ventricle by two-dimensional
echocardiography. American Society of Echocardiography Committee on
Standards, Subcommittee on Quantitation of Two-Dimensional Echocar-
diograms. J Am Soc Echocardiography. 1989;2:358 –367.
17. Devereux RB, Alonso DR, Lutas EM, Gottlieb GJ, Campo E, Sachs I,
Reichek N. Echocardiographic assessment of left ventricular hypertrophy:
comparison to necropsy findings. Am J Cardiol. 1986;57:450 – 458.
18. Nagueh SF, Mikati I, Kopelen HA, Middleton KJ, Quinones MA, Zoghbi
WA. Doppler estimation of left ventricular filling pressure in sinus
tachycardia: a new application of tissue Doppler imaging. Circulation.
1998;98:1644 –1650.
19. Brew K, Dinakarpandian D, Nagase H. Tissue inhibitors of metallopro-
teinases: evolution, structure and function. Biochim Biophys Acta. 2000;
1477:267–283.
20. Wilson EM, Gunasinghe HR, Coker ML, Sprunger P, Lee-Jackson D,
Bozkurt B, Deswal A, Mann DL, Spinale FG. Plasma matrix metallopro-
teinase and inhibitor profiles in patients with heart failure. J Card Failure.
2002;8:390 –398.
21. Stroud RE, Deschamps AM, Lowry AS, Hardin AE, Mukherjee R,
Lindsey ML, Ramamoorthy S, Zile MR, Spencer WH, Spinale FG.
Plasma monitoring of the myocardial specific tissue inhibitor of
metalloproteinase-4 after alcohol septal ablation in hypertrophic
obstructive cardiomyopathy. J Card Fail. 2005;11:124 –130.
22. Yasmin, Wallace S, McEniery CM, Dakham Z, Pusalkar P, Maki-Petaja
K, Ashby MJ, Cockcroft JR, Wilkinson IB. Matrix metalloproteinase-9
(MMP-9), MMP-2, and serum elastase activity are associated with sys-
tolic hypertension and arterial stiffness. Arterioscler Thromb Vasc Biol.
2005;25:372–378.
23. Tayebjee MH, Nadar SK, MacFadyen RJ, Lip GY. Tissue inhibitor of
metalloproteinase-1 and matrix metalloproteinase-9 levels in patients
with hypertension: relationship to tissue Doppler indices of diastolic
relaxation. Am J Hypertens. 2004;17:770 –774.
24. Tayebjee MH, Nadar S, Blann AD, Gareth Beevers D, MacFadyen RJ,
Lip GY. Matrix metalloproteinase-9 and tissue inhibitor of
metalloproteinase-1 in hypertension and their relationship to cardiovas-
cular risk and treatment: a substudy of the Anglo-Scandinavian Cardiac
Outcomes Trial (ASCOT). Am J Hypertens. 2004;17:764 –769.
25. Tayebjee MH, Lim HS, Nadar S, MacFadyen RJ, Lip GY. Tissue inhib-
itor of metalloproteinse-1 is a marker of diastolic dysfunction using tissue
Doppler in patients with type 2 diabetes and hypertension. Eur J Clin
Invest. 2005;35:8 –12.
26. Laviades C, Varo N, Fernandes J, Mayor G, Gil MJ, Monreal I, Diez J.
Abnormalities of the extracellular degradation of collagen type I in
essential hypertension. Circulation. 1998;98:535–540.
27. Lindsay MM, Maxwell P, Dunn FG. TIMP-1, a marker of left ventricular
diastolic dysfunction and fibrosis in hypertension. Hypertension. 2002;
40:136 –141.
 2096 Circulation May 2, 2006

28. Timms PM, Wright A, Maxwell P, Campbell S, Dawnay AB, Srikanthan
V. Plasma tissue inhibitor of metalloproteinase-1 levels are elevated in
essential hypertension and related to left ventricular hypertrophy.
Am J Hypertens. 2002;15:269 –272.
29. Li-Saw-Hee FL, Edmunds E, Blann AD, Beevers DG, Lip GYH. Matrix
metalloproteinase-9 and tissue inhibitor metalloproteinase-1 levels in
essential hypertension: relationship to left ventricular mass and anti-
hypertensive therapy. Int J Cardiol. 2000;75:43– 47.
30. Zervoudaki A, Economou E, Stefanadis C, Pitsavos C, Tsioufis K, Aggeli
C, Vasiliadou K, Toutouza M, Toutouzas P. Plasma levels of active
extracellular matrix metalloproteinases 2 and 9 in patients with essential
hypertension before and after antihypertensive treatment. J Hum
Hypertens. 2003;17:119 –124.
31. Deschamps AM, Apple KA, Leonardi AH, McLean JE, Yarbrough WM,
Stroud RE, Clark LL, Sample JA, Spinale FG. Myocardial interstitial
matrix metalloproteinase activity is altered by mechanical changes in LV
load: interaction with the angiotensin type 1 receptor. Circ Res. 2005;27:
96:1110 –1118.
32. Tsuruda T, Costello-Boerrigter LC, Burnett JC Jr. Matrix metallopro-
teinases: pathways of induction by bioactive molecules. Heart Fail Rev.
2004;9:53– 61.
33. Joffs C, Himali R, Gunasinghe RS, Multani MM, Dorman BH, Kratz JM,
Crumbley AJ III, Crawford FA Jr, Spinale FG. Cardiopulmonary bypass
induces the synthesis and release of matrix metalloproteinases. Ann
Thorac Surg. 2001;71:1518 –1523.
34. Yarbrough WM, Mukherjee R, Escobar P, Mingoia JT, Sample JA,
Hendrick JW, Dowdy KB, McLean JE, Lowry AS, O’Neil TP, Spinale
FG. Selective targeting and timing of matrix metalloproteinase inhibition
in post-myocardial infarction remodeling. Circulation. 2003;108:
1753–1759.
35. Lindsey ML, Mann DL, Entman ML, Spinale FG. Extracellular matrix
remodeling following myocardial injury. Ann Med. 2003;35:316 –326.

CLINICAL PERSPECTIVE
Chronic arterial hypertension is a common cause of left ventricular (LV) concentric hypertrophy, decreased relaxation rate,
and increased stiffness. The structural and functional changes caused by hypertension result from changes to both of the
Downloaded from http://circ.ahajournals.org/ by guest on February 4, 2018

principle constituents of the myocardium, the cardiomyocyte and particularly the extracellular matrix (ECM). These LV
structural and functional changes create the substrate necessary for the development of diastolic heart failure. However,
what controls these changes in the ECM, whether blood pressure control alone can prevent or reverse these changes, and
whether knowledge of the ECM control mechanisms would aid the diagnosis or treatment of hypertensive heart disease are
unknown. The present study showed that changes in the pattern of specific ECM proteolytic proteins/peptides (matrix
metalloproteinases [MMPs]) and their tissue inhibitors [TIMPs]) were associated with each structural, functional, and
clinical manifestation of hypertensive heart disease. Subjects with adequately controlled blood pressure and no LV
structural or functional changes did not have any changes in the MMP/TIMP signature. Therefore, treatment of
hypertension can prevent changes in the ECM and its proteolytic system. However, patients with residual or resistant LV
hypertrophy, despite adequate blood pressure control, had abnormal MMPs. The development of diastolic heart failure was
heralded by an increase in TIMP-1 to ⬎1200 ng/mL. These data suggest that regression of LV hypertrophy and prevention
of diastolic heart failure are dependent on more than just changes in blood pressure alone, and the need to target and
normalize changes in MMP/TIMPs may be warranted. Understanding this ECM-dependent pathophysiology may lead to
improved diagnosis and treatment of patients with hypertensive heart disease.
 Matrix Metalloproteinases/Tissue Inhibitors of Metalloproteinases: Relationship Between
Changes in Proteolytic Determinants of Matrix Composition and Structural, Functional,
and Clinical Manifestations of Hypertensive Heart Disease
S. Hinan Ahmed, Leslie L. Clark, Weems R. Pennington, Carson S. Webb, D. Dirk Bonnema,
Amy H. Leonardi, Catherine D. McClure, Francis G. Spinale and Michael R. Zile
Downloaded from http://circ.ahajournals.org/ by guest on February 4, 2018

Circulation. 2006;113:2089-2096; originally published online April 24, 2006;

doi: 10.1161/CIRCULATIONAHA.105.573865
Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2006 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7322. Online ISSN: 1524-4539

The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circ.ahajournals.org/content/113/17/2089

Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Circulation 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 Circulation is online at:

http://circ.ahajournals.org//subscriptions/