Circumferential Stress and Matrix Metalloproteinase 1 in Human Coronary Atherosclerosis
Implications for Plaque Rupture
Atherosclerotic plaque rupture may occur when regions of weakened extracellular matrix are subjected to increased mechanical stresses. Since collagen is a major determinant of extracellular matrix strength, enzymes that degrade collagen may play an important role in destabilizing the atherosclerotic lesion. To test the hypothesis that matrix metalloproteinase 1 (interstitial collagenase, or MMP-1), which initiates degradation of fibrillar collagens, colocalizes with increased stress in the fibrous cap of the atherosclerotic lesion, 12 unruptured human coronary lesions were studied. Finite-element analysis was used to determine the distribution of stress in the lesion, with estimates of material properties from previous measurements of human tissues. A computerized image analysis system was used to determine the distribution of immunoreactive MMP-1 within the fibrous tissue of the lesion. There was a significant correlation between immunoreactive MMP-1 and circumferential tensile stress in the fibrous cap within a given lesion (median Spearman rank correlation coefficient, .36; interquartile range, −.02 to .81; P<.02). Within a given lesion, the highest-stress region had twofold greater MMP-1 expression than the lowest-stress regions. In unruptured human atherosclerotic coronary lesions, overexpression of MMP-1 is associated with increased circumferential stress in the fibrous plaque. Degradation and weakening of the collagenous extracellular matrix at these critical high-stress regions may play a role in the pathogenesis of plaque rupture and acute ischemic syndromes.
- Received November 11, 1995.
- Revision received March 5, 1996.
Studies of thrombosis and occlusion in native coronary arteries have identified acute plaque rupture as an early event in most transmural myocardial infarctions.1 2 In addition, plaque rupture may play a critical role in the pathogenesis of unstable angina and in the asymptomatic progression of coronary lesions.3 Analyses of postmortem lesions strongly suggest that the distribution of mechanical stress based on the complex structure of lesions is an important determinant of the site of plaque disruption.4 5 However, the ability of the fibrous cap of the atheroma to withstand forces imposed on its tissue also depends on the mechanical integrity of the tissue. Fibrillar collagens are increased in the atherosclerotic lesion and can bear enormous stresses.6 Degradation of these fibrils could therefore compromise the ability of the matrix to withstand the increased mechanical stress. An important enzyme of the metalloproteinase family that initiates degradation of fibrillar collagen is interstitial collagenase, or matrix metalloproteinase 1 (MMP-1). Human atherosclerotic lesions have regions of focal overexpression of MMP-1, frequently at areas that would be anticipated to have increased mechanical stress.7 Therefore, this study was performed to test the hypothesis that expression of MMP-1 in the fibrous portion of the plaque is associated with regions of high stress in the atherosclerotic coronary lesion.
Twelve unruptured atherosclerotic lesions were removed by dissection of total artery specimens after fixation of whole hearts from eight patients with ischemic cardiomyopathy undergoing orthotopic heart transplantation. No lesion had evidence of plaque rupture, luminal thrombosis, or subintimal thrombosis. Specimens were fixed in the zero-stress state and stained with hematoxylin-eosin as well as Masson's trichrome stain. Specimens were not decalcified, and regions of lipid pool, normal artery, calcified plaque, and fibrous plaque highlighted by the routine morphological stains were traced by a cardiac pathologist as previously described.8 Paraffin sections were also stained with rabbit antiserum to human MMP-1 as previously described.7 The specificity of the antibodies was previously confirmed in our laboratory by immunoprecipitation and Western blotting of antigens secreted in vitro by cultured human vascular smooth muscle cells. Endogenous peroxidase activity was reduced by preincubation of sections with 0.3% hydrogen peroxide in Dulbecco's PBS. The sections were then incubated with primary antibodies diluted in PBS supplemented with 10% horse serum at room temperature for 60 minutes. After the sections were washed in PBS and then in 100 mmol/L Tris-HCl/150 mmol/L NaCl containing 2% horse serum, biotinylated secondary antibodies were applied, followed by avidin-peroxidase complexes, and the reaction was visualized with 3-amino-9-ethyl carbazole as substrate. In addition, adjacent sections were stained with the monoclonal anti–human macrophage antibody HAM-56. For each specimen, an approximate lumen center was identified, as well as the origin of an r-θ coordinate system. This coordinate system was used to divide the cross section into eight equal sectors of 45° and to allow comparison of the quantitative immunohistochemistry results with the independently performed structural analysis.
Analysis of immunohistochemistry images was performed with a personal computer–based quantitative 24-bit (16.2 million unique combinations) color image analysis system. First, photographs of images were scanned into a 1K×1K image buffer of the Optimas 4.0 image analysis system (Optimas Co). A color threshold mask was defined to detect the red-brown color of immunoperoxidase activity by sampling, and this threshold was applied to the image. Within a given lesion, the same mask was used, although between lesions, minor adjustments were sometimes made on the threshold mask to eliminate detection of obvious artifacts. For each cross section, the fibrous component in each of the eight sectors was traced, excluding the necrotic core (as outlined by the pathologist's tracing from routine morphology stains), which was consistently immunoreactive with both MMP-1 and HAM-56 antibodies. In addition, the endothelial layer was excluded from the analysis area; this eliminated the influence of edge artifacts as well as positive staining of endothelial cells with HAM-56. The analysis was limited to the fibrous cap, because collagen in the fibrous cap is a primary determinant of the integrity of the lesion and the necrotic core bears little stress. For each sector, the percentage of the total area with the positive color was recorded.
Calculation of Circumferential Stress
Structural analysis was performed without knowledge of the immunohistochemistry analysis. The finite-element method was used to determine the distribution of stress in the lesion as previously described.8 The two-dimensional tracing of each section was digitized with a personal computer. The design program I-DEAS Master Series 1.3c (Structural Dynamics Research Co) and a Hewlett Packard model 730 workstation were used to create finite-element meshes based on the digitized contours. Quadrilateral, eight-noded, plane-strain, mixed-formulation elements were created using a free-meshing algorithm with a characteristic element dimension of 0.1 to 0.2 mm. Mesh resolution was increased near material interfaces to increase the accuracy of the solution. Assignment of material properties was identical to our previously published isotropic incompressible model with the following moduli (E): normal artery E=100 kPa; fibrous plaque E=1000 kPa; calcified plaque E=10 000 kPa; and lipid E=0.5 kPa.8 9 A static pressure of 14.6 kPa (110 mm Hg) was applied to the luminal wall of each model. The finite-element models were solved by use of ABAQUS/Standard 5.3 (Hibbitt, Karlsson & Sorenson Co) using a large-strain, hyperelastic, nonlinear strain energy potential–based solution algorithm; the solver was nonlinear to account for nontrivial strains. For each of the eight sectors of a lesion, the maximum circumferential stress within the fibrous cap tissue was recorded. The Cartesian normal stresses were transformed into cylindrical coordinates, and circumferential stresses were calculated with respect to the center of the lumen.
The primary data were not normally distributed according to the Wilk-Shapiro test either before or after correction for individual differences. Statistical analysis was performed only after completion of image and structural analysis of all 12 lesions. The primary hypothesis of this study was tested by comparison of the distribution of immunohistochemical staining of interstitial collagenase with circumferential stress by calculation of Spearman's rank correlation coefficients for each lesion. Because the nonparametric correlation coefficients were not normally distributed, they were transformed with Fisher's z transformation. These values were regarded as a random sample from a normal distribution with unknown mean and variance and tested for zero mean with a one-sample Student's t test. Because four patients provided two lesions each, the significance level from Student's t test was adjusted for correlated data by use of a generalized estimating equation approach to correlated data. Secondary analyses (correlation between macrophage density and stress or interstitial collagenase) were performed in an identical manner. A value of P<.05 was considered statistically significant. To estimate how much higher interstitial collagenase was in regions of high stress, the eight sectors within a given lesion were divided into four pairs by magnitude of stress; these groups were designated low-, moderate-, high-, and highest-stress regions. Collagenase expression in the three higher-stress groups was compared with collagenase expression in the lowest-stress group.
Immunoreactive MMP-1 was detected in 11 of the 12 coronary lesions. The median Spearman correlation between fibrous cap circumferential tensile stress and percentage of area with immunoreactive MMP-1 was .36 (interquartile range, −.02 to .81; P<.02; Figure).⇓ Adjustment for correlation between specimens from the same individual improved the statistical significance for this relation. Within a given lesion, the three greater-stress regions had increased MMP-1 expression compared with the lowest-stress group (median increase, 33% for moderate stress; 43% for high stress; and 100% for highest stress). These data indicate that regions of high mechanical stress in the coronary atherosclerotic lesion have increased expression of MMP-1.
Evidence for immunoreactive clusters of macrophages (HAM-56 staining) was present in all lesions. There was no correlation between peak fibrous cap circumferential stress and macrophage staining. The median Spearman correlation coefficient between fibrous cap circumferential stress and percentage of area staining positively for macrophages was .23 (interquartile range, −.28 to .48; P=.55). Similarly, there was no significant correlation between macrophage staining and MMP-1 staining (median Spearman correlation coefficient, .15; interquartile range, −.56 to .63; P=.4).
In this study of stable coronary artery atherosclerotic fibrous caps, MMP-1 was preferentially expressed in regions of higher mechanical circumferential stress. This association may have important implications for the pathogenesis of acute ischemic syndromes as well as the natural history of the atheroma. Compelling evidence has emerged indicating that particular geometric features of the plaque, such as a thin fibrous cap and large lipid pool, cause regions of high circumferential stress in the fibrous cap and that fatal plaque rupture commonly occurs in these high-stress regions.9 The finding that MMP-1 is overexpressed at high-stress regions suggests that matrix degradation at these critical locations also plays a role. It is important to note that stress is not the only determinant of MMP-1 expression, because the lipid core was consistently immunoreactive despite bearing little mechanical stress.
The source of MMP-1 and other metalloproteinases such as stromelysin in the stable atheroma may be vascular smooth muscle cells, the primary cellular component of the fibrous cap. Production of MMP-1 by vascular smooth muscle cells is associated with the shift of these cells to the modulated phenotype typical of smooth muscle cells in the atheroma.10 Signals that induce metalloproteinase secretion by cultured vascular smooth muscle cells include cytokines, such as interleukin-1β and tumor necrosis factor-α, that are found in the human atherosclerotic lesion.11 Mechanical events can also modulate metalloproteinase secretion. Werb et al12 found that expression of collagenase and stromelysin by rabbit synovial fibroblasts depends on alteration of cellular morphology and of the actin cytoskeleton. In addition, metalloproteinases can be induced in monolayer cultures of vascular smooth muscle cells by mechanical injury of the monolayer.13 One potential explanation of our results is that expression of MMP-1 is an adaptive response to mechanical stress, allowing the fibrous cap to remodel itself to a more structurally sound configuration. In this case, expression of MMP-1 could be beneficial, delaying or preventing plaque rupture. Our study did not address the hypothesis that MMP-1 expression could be beneficial in some circumstances.
Macrophages are another potentially important source of matrix-degrading enzymes in the atherosclerotic lesion. Human monocytes and macrophages are a source of not only cytokines but also metalloproteinases.14 In addition, macrophages commonly localize at sites of fatal human coronary plaque rupture.15 16 The present study did not find a significant correlation between macrophage density and MMP-1 expression in stable coronary lesions. This suggests that there may be two stages that determine the “vulnerability'' of the atheroma to rupture. In the first stage, which may last many years, the lesion accumulates a necrotic core beneath the fibrous cap, and smooth muscle cell–derived MMP-1 weakens the high-stress regions. In the second stage, which is presumably more acute, the lesion accumulates macrophages, which accelerate the matrix degradation process at the rupture location. The recent description of a systemic inflammatory state in some unstable coronary syndromes is consistent with this hypothesis.17 18
The present study used a quantitative image processing system to evaluate immunohistochemistry specimens. The analysis process required photography, scanning into a digital 24-bit format, and digital color analysis; each of these steps could induce color imbalances or distortions that could affect the analysis. However, immunohistochemistry specimens were analyzed independently from structural analyses, and within a given lesion, no bias favoring distortion of a specific sector would be anticipated. In addition, the resolution of this system is insufficient to detect staining at the single-cell level or to detect focal expression associated with microvasculature. We cannot exclude the possibility that more focal expression of MMP-1 could play a significant role in plaque rupture. It is also important to note that matrix degradation is determined not only by expression of metalloproteinases but also by activation of the enzymes in the extracellular space, the degree of inhibition by endogenous inhibitors, and the presence and activity of enzymes that do not degrade metalloproteinase matrix. Further understanding of the balance of matrix degradation in the fibrous plaque at high-stress regions may provide insight into why some plaques rupture and others remain stable.
This study was supported in part by a Grant-in-Aid from the American Heart Association, Massachusetts Affiliate, and a grant from the National Heart, Lung, and Blood Institute (HL-48743). Dr Loree is a recipient of a National Research Service Award from the National Heart, Lung, and Blood Institute (2-T32-GM-07753-12). The authors thank E. John Orav, PhD, of the Harvard School of Public Health for his expert assistance in the statistical analysis. The authors also thank Scott D. Solomon, MD, for his advice regarding the image analysis system and Elena Rabkin for her technical assistance in immunohistochemical studies.
- ↵Davies MJ. A macro and micro view of coronary vascular insult in ischemic heart disease. Circulation. 1990;82(suppl II):II-38-II-46.
- ↵Fuster V. Mechanisms leading to myocardial infarction: insights from studies of vascular biology. Circulation. 1994;90:2126-2146.
- ↵Cheng GC, Loree HM, Kamm RD, Fishbein MC, Lee RT. Distribution of circumferential stress in ruptured and stable atherosclerotic lesions. Circulation. 1993;87:1179-1187.
- ↵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. 1995;94:2493-2503.
- ↵Cheng GC, Loree HM, Kamm RD, Fishbein MC, Lee RT. Distribution of circumferential stress in ruptured and stable atherosclerotic lesions: a structural analysis with histopathological correlation. Circulation. 1993;87:1179-1187.
- ↵Galis ZS, Muszynski M, Sukhova GK, Simon-Morrissey E, Unemori EN, Lark MW, Amento E, Libby P. Cytokine-stimulated human vascular smooth muscle cells synthesize a complement of enzymes required for extracellular matrix digestion. Circ Res. 1994;75:181-189.
- ↵Werb Z, Hembry RM, Murphy G, Aggeler J. Commitment to expression of the metalloendopeptidases, collagenase and stromelysin: relationship of inducing events to changes in cytoskeletal architecture. J Cell Biol. 1986;102:697-702.
- ↵Moreno PR, Falk E, Palacios IF, Newell JB, Fuster V, Fallon JT. Macrophage infiltration in acute coronary syndromes: implications for plaque rupture. Circulation. 1994;90:775-778.
- ↵van der Wal AC, Becker AE, van der Loos CM, Das PK. Site of intimal rupture and erosion of thrombosed coronary atherosclerotic plaques is characterized by an inflammatory process irrespective of the dominant plaque morphology. Circulation. 1994;89:36-44.
- ↵Jude B, Agraou B, McFadden EP, Susen S, Bauters C, Lepelley P, Vanhaesbroucke C, Devos P, Cosson A, Asseman P. Evidence for time-dependent activation of monocytes in the systemic circulation in unstable angina but not in acute myocardial infarction or in stable angina. Circulation. 1994;90:1662-1668.