Localization of Apoptotic Macrophages at the Site of Plaque Rupture in Sudden Coronary Death

Selection of Cases

Cases of sudden coronary death were identified as part of a consultative service provided to the Office of the Chief Medical Examiner for the State of Maryland. 16 Sudden coronary death was based on unexpected death, witnessed within 6 hours of the onset of symptoms, or death of a person known to be in stable condition <24-hours antemortem. A complete autopsy, including a toxicology screen, ruled out noncardiac causes of death.

Twenty-five cases of plaque rupture were collected and of these, 21 were used for DNA fragmentation staining and immunohistochemistry and four were processed for immunoblot analysis. Of the 15 stable plaques, 11 were used for DNA fragmentation staining and four were processed for immunoblot analysis; four nonatherosclerotic arteries served as controls for immunoblot studies.

Definitions

Culprit plaques were identified as those with an acute luminal thrombus and plaque rupture or in the absence of a thrombus, the artery with the greatest degree of luminal narrowing relative to the internal elastic lamina. 4,15 Acute plaque ruptures consisted of a luminal platelet-fibrin thrombus continuous with an underlying lipid-rich core. The connection between the thrombus and the lipid-rich core was through a disrupted thin fibrous cap infiltrated by macrophages. Stable plaque was defined as cross-sectional luminal narrowing of ≥75% with a fibrous cap >65 μm in thickness in the absence of plaque rupture or thrombus.

Evaluation of the Hearts and Preparation of Tissue Specimens

Hearts were studied under the supervision of the medical examiner. Coronary arteries were serial sectioned at 3-mm intervals. Segments showing narrowing of >50% were submitted for histological analysis. Collected arterial specimens were immediately immersed in ice-cold medium 199. Vessels were then snap-frozen in optimal cutting temperature tissue processing medium (OCT; Miles Diagnostics, Elkhart, IN).

In Situ End-Labeling (ISEL) for DNA Fragmentation

Serial cryostat sections (6 μm) were cut and air-dried onto Superfrost microslides (Columbia Diagnostics, Inc., Springfield, VA) and stored at −70°C before use. In situ labeling of DNA fragmentation was performed using terminal deoxyribonucleotide transferase (TdT)-mediated nick end labeling based on an in situ apoptosis detection kit (TACS; Trevigen, Gaithersburg, MD). Sections were fixed for 10 minutes in neutral-buffered formalin, rinsed in phosphate-buffered saline (PBS), and incubated for 5 minutes in 0.1% Triton X-100 in 0.1% sodium citrate. Treating the slides with 0.3% hydrogen peroxide for 10 minutes inactivated endogenous peroxidase. DNA fragments were labeled with biotinylated nucleotides (dNTPs) and TdT for 1 hour at 37°C. The incorporation of biotinylated nucleotides into DNA was detected with a streptavidin-conjugated horseradish peroxidase. The chromogenic substance used to visualize the reaction was TACS Blue Label (Trevigen) which produces a blue reaction product. Sections of rat mammary glands (weaning animals) were used a positive controls for apoptosis (Oncor, Gaithersburg, MD).

Detection of Fragmented DNA by Electron Microscopy

The method of tissue processing for ultrastructural analysis and immunogold labeling was adapted from the procedure described by Berryman and Rodewald, 17 and all processing steps were performed at 4°C. Semithin sections were cut, stained with toluidine blue, and examined by light microscopy for the site of rupture. Artifact-free areas showing rupture sites were selected for ultrathin section cutting.

The TdT immunogold assay was performed as above with slight modifications. 18 Sections were etched for 10 minutes in absolute ethanol, rinsed in PBS, and permeabilized in CytoPore (Trevigen, Inc.). Endogenous peroxidase was quenched by immersion in 3% hydrogen peroxide in 40% methanol. Sections were then incubated in the TdT-labeling mixture in a humidified chamber at 37°C for 1 hour. The reaction was stopped and sections were then incubated overnight at 4°C with colloidal gold-streptavidin (Zymed, San Francisco, CA) for direct visualization of biotinylated dNTPs. Immunogold labeling was examined using an electron microscope (Zeiss EM 10, Germany). Omitting TdT during the procedure as the negative control checked the validity of the method.

Immunohistochemistry

Electrophoresis and Immunoblot Analysis

Culprit plaques were identified by gross histological analysis using a dissection microscope and further confirmed by histological examination of adjacent sections. Atherosclerotic segments (3 to 5 mm) with and without plaque rupture and thrombi were removed and immediately placed in ice-cold M199; nonatherosclerotic vessels were collected as controls. After removal of the adventitia, samples were snap-frozen in liquid nitrogen and stored at −70°C until protein extraction.

For protein extraction, the tissue was weighed, minced with scissors, and extracted in ice-cold lysis buffer (pH 7.6) containing 50 mmol/L Tris-HCl, 15 mmol/L NaCl, 1 mmol/L phenylmethylsulfonyl fluoride, 10 μg/ml aprotonin, 10 μg/ml leupeptin, 1% Nonidet P-40, and 1 mmol/L dithiothreitol.

Proteins (50 μg per lanes) from plaque homogenates were separated on 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis mini gels and transferred to nitrocellulose membranes (Bio-Rad, Richmond, CA) using a semidry blotting system. Blocking of nonspecific binding was achieved with incubation of the membrane in 5% milk PBS, pH 7.2, containing 0.05% Tween 20. Polyclonal antibodies against caspase-1 (Upstate Biotechnology, Lake Placid, NY) or caspase-3 (Santa Cruz Biotechnology, Inc.) were used for staining. Antigen detection was performed with a chemiluminescent detection system (ECL; Amersham, Arlington Heights, IL).

Statistical Analysis

All values are expressed as the mean ± SEM. Comparisons of apoptotic index among cell types in the different regions of the plaque were performed by factorial analysis of variance (StatView 4.5; Abacus Concepts Inc., Berkeley, CA) and analyzed simultaneously with post hoc testing by the Scheffé’s procedure; statistical significance was established at P < 0.05.

Evidence of Extensive Apoptosis at the Site of Plaque Rupture

Labeling of DNA fragmentation alone was undertaken for the detection of apoptosis by light microscopy in ruptured and stable plaques. Overall apoptosis within the fibrous cap was more prevalent in plaque ruptures than in stable plaques (P = <0.001). The bulk of this difference was accounted for by the high degree of apoptosis at plaque rupture sites (Figure 2) ▶ .

Ultrastructural examination was performed to confirm the occurrence of apoptosis in plaque rupture (Figure 3) ▶ . Many cells at the ruptured site demonstrated varying stages of apoptosis with characteristic changes of nuclear chromatin condensation and clumping, and loss of cell volume. DNA fragmentation was shown by combined ultrastructure and ISEL using a gold conjugate. The gold-labeled fragments were commonly found at the periphery of the nucleus at sites of chromatin condensation in cells undergoing apoptosis. Only occasional gold particles were noted in cells with normal ultrastructure and were found irrespective of nuclear chromatin material. Unlike ruptures, apoptotic cells were rarely observed in the fibrous caps of stable plaques.

Apoptosis at the Plaque Rupture Site Occurs Essentially in Macrophages

Immunohistochemical staining of serial sections for the identification of cell type showed that the majority of cells at the rupture site were macrophages. Of the total number of immunoreactive cells, 193 ± 10 per 0.1 mm 2 were macrophages; the number of smooth muscle and T cells were markedly less, 16 ± 2 and 15 ± 5 per 0.1 mm², respectively. In contrast, stable plaques showed a decreased prevalence of macrophages (30 ± 9 per 0.1 mm², P < 0.0001) and an increase in SMCs (80 ± 13, P = 0.0003); T cells comprised ∼3 ± 0.5 per 0.1 mm².

Double-staining for the immunohistochemical typing of cells in combination with in situ labeling for apoptosis showed that apoptotic nuclei at rupture sites were essentially those of macrophages (Figures 4 and 5) ▶ ▶ . Apoptosis in macrophages was more frequent at the rupture site (93 ± 8 per 0.1 mm²) as opposed to areas of intact fibrous cap (10 ± 2 per 0.1 mm², P = 0.03). Apoptotic smooth muscle and T lymphocytes at the rupture site were only 5 ± 0.9 per 0.1 mm 2 and 5 ± 1 per 0.1 mm², respectively, contributing little to the total apoptotic cell population. Conversely, stable plaques showed a much lower frequency of apoptotic cells compared with ruptures (Figures 2 and 4) ▶ ▶ . The number of apoptotic nuclei expressed per 0.1 mm 2 in stable plaques were 7 ± 1.2 for macrophages, 3 ± 0.8 for SMCs, and 0.8 ± 0.4 for T lymphocytes.

The percentage of macrophage apoptosis was plotted against postmortem interval in 21 cases of plaque rupture to determine whether apoptosis of macrophages was dependent on the interval between death and autopsy. In our cases, the postmortem interval ranged between 10 and 53 hours with the majority of cases (n = 16) below 24 hours. Apoptosis of macrophages associated with plaque rupture sites showed no correlation with postmortem interval (R 2 = 6.2 × 10⁻⁵).

Apoptosis at the Rupture Site Is Associated with Interleukin-1β-Converting Enzyme Activation

Plaque rupture sites demonstrated strong immunoreactivity to caspase-1 corresponding to regions of apoptotic macrophages (Figure 6A) ▶ . In contrast, immunoreactivity for caspase-3 at rupture sites was weak (not shown). In stable plaques, areas positive for apoptotic cells, in particular regions bordering the necrotic core showed immunoreactivity to both caspases-1 and -3 (Figure 6B) ▶ ; staining in the fibrous cap was essentially negative for either protein.

Immunoblot analysis showed the presence of the 45-kd precursor form of caspase-1 in both ruptured and stable plaques (Figure 6C) ▶ . In contrast, the cleaved p20 band of caspase-1 was noted only in the ruptured plaques. Stable plaques did not show the evidence of caspase-1 activation. Nonatherosclerotic control arterial segments failed to show either the caspase-1 precursor or cleavage products. Unlike interleukin-1β-converting enzyme in plaque ruptures, caspase-3 cleavage to p17 protein was not detected in any of the cases; only the uncleaved 32-kd band was present (data not shown).

SMC Apoptosis and the Vulnerable Plaque

There is a well-established association of rupture with thin fibrous caps 19 and cap thickness constitutes a major part of the definition of the vulnerable plaque. 4 Although apoptosis of SMCs, in part may lead to fibrous cap thinning, this was not appreciated in the present study perhaps because the lesions were already ruptured. The paucity of SMCs at plaque rupture sites; however, suggests that SMC death may have occurred at an earlier time point. One preliminary study suggests that human monocytes/macrophages induce vascular SMC apoptosis; 20 although in fatty streaks, both appear to co-exist. The finding of decreased SMCs at plaque rupture sites agrees with earlier studies involving ulcerated lesions of the aorta. 21,22 The overall loss of SMCs within the fibrous cap, most likely represents a chronic process before the actual rupture event.

Apoptosis versus Oncosis

There is at least one paper on carotid plaques suggesting that oncosis contributes to cell death. 23 Recognition between apoptosis and oncosis is best appreciated at the ultrastructural level, because the terminal dUTP nick-end labeling assay by light microscopy lacks specificity. These pathways of cell death are best distinguished early because the phase of necrosis is similar for both. 24 In our study, early changes of cell death showed morphological features consistent with the nuclear changes associated with apoptosis. Because of the variable uptake of lipids and differences in cell size, cell shrinkage (another feature of apoptosis), was not always apparent. There was also some degree of uncertainty of cell type, because lipid-uptake was not exclusive to macrophages. As the morphological changes of cell death advanced the distinction between apoptosis and oncosis became difficult, thus, death by oncosis could not be excluded.

Dissolution of Matrix and Cell Apoptosis Anoikis

The precise etiology of macrophage apoptosis at the plaque rupture site is speculative. There is experimental evidence that the dissolution of extracellular matrix may threaten cell survival. In epithelial and vascular endothelium, inhibition of integrin function has been shown to produce a loss of cell-cell adhesion and apoptosis. 25,26 This type of detachment-induced cell death, a form of apoptosis, is referred to as “anoikis.” 27 In another study, basement membrane matrix was shown to suppress apoptosis of mammary epithelial cells in tissue culture and in vivo. In these studies, antibodies to α-integrins or overexpression of stromelysin-1 stimulated apoptosis, whereas the loss of extracellular matrix correlated with the expression of caspase-1 such that specific inhibitors of its activity rescued cells from apoptosis. Although data on cell attachment and death of human macrophages is lacking, an integrin dependence of cell survival has been shown in the murine macrophage-like cell line RAW264.7. 28

TIMP-3 and Cell Survival in the Fibrous Cap

There is emerging evidence that inhibitors of metalloproteinases, specifically TIMP-3, directly affect cell survival. 29 For example, adenoviral-induced overexpression of TIMP-3 promotes apoptosis of vascular SMCs both in vivo and in vitro. Although little TIMP-3 protein is found in normal arteries, it is markedly elevated in advanced plaques and is primarily associated with intimal macrophages at rupture-prone sites. 30 It remains to be demonstrated whether macrophages themselves commit cellular suicide through the expression of TIMP-3.

Apoptosis of Macrophages and Thrombogenicity

Macrophage apoptosis may also facilitate the acute thrombotic event arising from the rupture itself. In a recent study, Mallat et al 31 examined the role of apoptotic cell death in relation to plaque thrombogenicity. Shed membrane microparticles, products of apoptotic cells from carotid plaques were captured by annexin V and analyzed for their procoagulant activity. These microparticles were primarily monocytic and lymphocytic in origin, and very rich in tissue factor activity. Thus, it is conceivable that microparticles from macrophage apoptosis at the rupture site may increase the procoagulant potential of the plaque.

Caspases and Cell Death in Plaques

Apoptosis in mammalian cells is essentially regulated by caspases that are homologous to the C. elegans (worm) death genes. 32,33 The prototypic enzyme caspase-1 is synthesized as a 45-kd proenzyme that is autocatalytically cleaved to generate an active homodimeric enzyme of 20- and 10-kd subunits. 34 Geng and Libby 10 first reported on the co-localization of caspase-1 with apoptosis in advanced human atherosclerotic plaques. This data is corroborated by our study in which extensive expression of caspase-1 was detected at rupture sites rich in apoptotic macrophages. In comparison, caspase-1 immunolocalization in stable plaques was mostly confined to areas surrounding the necrotic core with minimal staining within the fibrous cap.

Although caspase-3 was also detected in ruptured plaques, its expression was not localized to plaque rupture sites and was more associated with cells surrounding the necrotic core. In the apoptotic cascade, caspase-3 has been described as an important mediator, especially with DNA damage and non-DNA damage stress. Cytokine-mediated stress, however, routes through activation of caspase-8 and caspase-1. 35 This raises the possibility of differential caspase activation depending on the environment within the plaque.

The present study demonstrates significant caspase-1 cleavage with plaque rupture when compared with stable plaque. Although cell types are indistinguishable in plaque homogenates, the strong expression of the caspase-1 precursor and its cleavage fragment complements the immunohistochemical localization of caspase-1 in macrophages at plaque rupture sites. The strong evidence of caspase-1 cleavage in the ruptured plaque may stem from the focal and concentrated presence of macrophages undergoing apoptosis. However, the absence of activation of caspase-1 in stable plaques does not rule out the possibility of apoptosis in these lesions but only suggests that the degree of apoptosis is too infrequent to be detected in whole plaque homogenates.

Study Limitations

Plaque rupture is a dynamic process and it is conceivable that an active cell such as macrophage plays an important role in the process. It must be emphasized that acute fatal ruptures are at their morphological endpoint and many of the mechanistic processes leading to the rupture event likely have occurred. Because autopsy material only provides an end-stage specimen, the physiological relevance of the present observations cannot be identified. Furthermore, limitations of autopsy material, including postmortem interval, must also be taken into account. In our study, specimens were often collected more than 8 hours after death, possibly allowing sufficient time for macrophages to undergo apoptosis after rupture was triggered rather than before. The lack of correlation of macrophage apoptosis at rupture sites with postmortem interval, however, argues against apoptosis occurring after the patient’s death. Other published studies in which simulated postmortem delay (up to 15 hours) failed to show increased apoptosis in myocardial and brain tissue corroborate our findings. 36,37 Finally, extensive macrophage apoptosis was primarily localized to the rupture site, whereas DNA fragmentation in macrophages in the fibrous cap away from the rupture site and in stable plaques with similar postmortem intervals was significantly less.