Article Text
PDF
Abstract
Background: T-lymphocyte activation within atherosclerotic plaque, and widespread to the myocardium, has been shown in patients with acute coronary syndromes.
Objective: To investigate the presence of T-lymphocyte infiltrate at different stages of acute coronary syndromes by studying patients with sudden coronary death, acute myocardial infarction (AMI) and healed infarction, in comparison with patients with myocarditis and patients with non-ischaemic heart failure.
Methods: 72 cases were studied at autopsy: 12 dying of sudden coronary death (group 1), 12 dying <4 weeks (group 2) and 12 dying >4 months after AMI (group 3), 12 with active lymphocytic myocarditis (group 4), 12 with hypertensive heart disease (group 5), and 12 control subjects (group 6). Light microscopy was performed to measure the number of activated T-lymphocytes (CD3+/DR+) in the myocardium and coronary artery wall, and intercellular adhesion molecule-1 (ICAM-1) expression in the myocardium.
Results: Activated T-lymphocyte infiltrates and ICAM-1 myocardial expression in both remote and peri-infarction regions and activated T-lymphocytes within the epicardial coronary artery wall of both the infarct- and non-infarct-related arteries were found in groups 1, 2 and 3, whereas myocardial, but not coronary, infiltrates were found in groups 4 (p<0.001 vs groups 1, 2 and 3 for coronary infiltrates). Groups 5 and 6 had no evidence of myocardial or coronary inflammation (p<0.001 vs groups 1, 2 and 3).
Conclusions: The study shows the presence of a lymphocytic infiltrate in both coronary arteries and myocardium and a proinflammatory phenotype shift in the myocardium associated with acute coronary thrombosis in patients dying suddenly, shortly, or even late after coronary thrombosis.
Statistics from Altmetric.com
Coronary vascular inflammation has a pivotal role in the development of acute coronary syndromes (ACS).1 The theory of the “vulnerable plaque” has been expounded upon, and it is postulated that widespread coronary artery inflammation can lead to acute thrombosis in susceptible plaques and cause myocardial ischaemia and infarction in the underlying myocardium.1 The presence of an active myocardial infiltrate has also been demonstrated in regions remote from the infarcted myocardium,2 further supporting the notion of a diffuse inflammation. The pathophysiology of inflammation in ACS is, however, not well understood, and it is currently unclear whether inflammation precedes myocardial necrosis or is a consequence of it, and it is unknown whether postinfarction cardiomyopathy shares a common inflammatory activation with non-ischaemic cardiomyopathy and heart failure. Moreover, the characterisation of such a state may provide information leading to a better understanding of the disease process.
The aim of the study was therefore to investigate the presence of a widespread myocardial lymphocytic infiltrate in patients dying suddenly with acute coronary thrombosis (in which case an inflammatory response would be considered to exist before the ischaemic event), patients with recent or prior acute myocardial infarction (AMI; in which case the inflammatory response may be amplified by the occurrence of myocardial necrosis), and to compare these findings with those in subjects with myocarditis (in which case the inflammatory response would be unrelated to the coronary events), subjects with hypertensive heart failure (in which case the presence of an inflammatory response would most likely reflect an aspecifc response to myocardial damage), and control subjects without heart disease.
PATIENTS AND METHODS
Selection of cases
Seventy-two subjects were retrospectively selected from a series of over 3000 consecutive postmortem examinations at the University of Trieste starting from 31 December 2005 backwards to form six groups of 12 cases each. Twelve cases were selected consecutively according to the following inclusion and exclusion criteria: sudden death (within 6 hours of the onset of symptoms), acute coronary thrombosis, and corresponding early ischaemic changes in the myocardium (group 1, sudden coronary death). Twelve cases with AMI within 4 weeks (group 2, recent AMI), and 12 cases of myocardial infarction occurring at least 4 months before death and with an identifiable corresponding infarct lesion at pathology (group 3, healed AMI) were also included. Clinical or pathological evidence of continuing or very recent myocardial ischaemia or necrosis were exclusion criteria for groups 2 and 3. Twelve cases of active lymphocytic myocarditis diagnosed at autopsy were also selected and constituted group 4 (myocarditis). Twelve cases of hypertensive heart disease, cardiomegaly and heart failure without coronary artery disease or myocardial infarction (group 5, heart failure) were selected at autopsy as a model of non-ischaemic heart disease. Twelve consecutive control subjects were selected among the cases of death in the absence of any evident cardiac disease (group 6, controls). If more than 12 cases meeting the prespecified criteria for each group were found, only the first 12 cases in each group were studied. Delay between death and autopsy (>30 hours), advanced cancer, and the presence of chronic systemic inflammatory disease (requiring immunosuppressive treatment at any time before admission—with the exception of group 4 cases) were exclusion criteria for all groups.
Pathology
In groups 1–3, gross examination of the hearts was used to define the infarct area and infarct-related artery (IRA). Tissue specimens were obtained at peri-infarct regions and in myocardial regions remote from the infarcted area and from old scars (in cases of multiple AMI), which are apparently normal at gross pathology and associated with a non-occluded related coronary artery (in cases of multivessel disease). For the peri-infarct region, fields for the cell count were selected in the zone bordering the infarct only where viable myocardium was prevalent and reparative fibrosis only marginal, considering only those fields (×40) where more than 30 cardiomyocytes were present. Continuing or very recent cardiomyocyte necrosis in the peri-infarct or remote regions was excluded in groups 1, 2 and 3. Moreover, the presence of reparative fibrosis (expression of prior infarct)3 was excluded in remote regions. The anterior and lateral walls were sampled in groups 4–6. Two different sections of each major epicardial branch of the coronary arteries were taken in every case. Sections were taken in the proximal third and in correspondence to the highest degree of luminal stenosis.
The presence of activated cells in the myocardium was determined by immunohistochemical analysis. We used a mouse monoclonal antibody anti-CD3 (Ventana, Tucson, AZ, USA; 1:100 solution) to identify the presence of T-lymphocytes, and anti-HLA-DR (Ventana, 1:50 solution) to determine activation status. Expression of the intercellular adhesion molecule-1 (ICAM-1) was tested using a mouse monoclonal antibody (Diapath, CD54/ICAM-1 Ab-4, 1:10 dilution). Antibody binding was shown by the avidin–biotin–peroxidase complex technique (Ventana) after washing. The number of positive cells was counted on 50 random fields (×40) and expressed as the number of cells/mm2 for CD3+ cells. A semiquantitative four-grade score for ICAM-1 expression was used with 0 equal to no expression, 1 is mild/focal expression, and 4 is strong/diffuse expression. Two pathologists unaware of the clinical characteristics of the subjects independently performed a cell count.
Fluorescent microscopy was used to document CD3 and DR colocalisation in lymphocytes. A two-step process was used. After the primary anti-human CD3 antibody was incubated, a secondary rabbit anti-mouse IgG antibody was bound to Alexa Fluor 488 (Invitrogen, Carlsbad, CA, USA; 1:250 dilution). After multiple washings in phosphate-buffered saline (PBS), an additional primary mouse anti-human HLA- DR (Invitrogen, Carlsbad, CA, USA; 1:50 dilution) and a secondary goat anti-mouse IgG antibody was bound to Alexa Fluor 555 (Invitrogen; 1:250 dilution). Prolong Gold Antifade with 4,6-diamino-2-phenylindole (to stain for nuclear DNA)(Invitrogen) was applied, followed by mounting with a glass cover slip. The slides were visualised with a Nikon epifluorescent microscope with a ×60 oil objective and three different filters for DAPI, FITC, and rhodamine. Image acquisition was obtained with a MicroPublisher 3.3 CCD camera with Q-Capture Professional image analysis software (QImaging).
Negative control reactions, in which the primary antibody was omitted, were performed in each case to rule out non-specific binding of the secondary antibody.
Statistical analysis
SPSS 11.0 for Windows (SPSS, Chicago, IL, USA) was used. χ2 and Fisher exact tests were used to compare discrete variables, when appropriate. Quantitative results were expressed as median and interquartile range. Non-parametric tests were used to compare CD3/DR+ cells among different regions of each subject (Wilcoxon test for paired data) and among different subjects (Kruskal–Wallis test, with Mann–Whitney U test when one-to-one comparison was made). According to Bonferroni’s criteria for multiple comparisons, a p value of ⩽0.01 was considered significant.
RESULTS
Clinical features
Table 1 shows the clinical and demographic characteristics of the subjects in the six groups. All patients in group 1 had died suddenly. All patients in groups 2 and 3 had mild to severe symptoms secondary to ischaemic heart failure before death; class III/IV New York Heart Association (NYHA) symptoms were present in nine and seven cases, respectively. All group 4 subjects were symptomatic for NYHA class III–IV heart failure. Median time from the onset of symptoms to death was 2 hours in group 1 (range 0–6 hours), 18 days (10–25) in group 2, 135 days (120–210) in group 3, and 7 days (1–35) in group 4. None of the cases in groups 4, 5 or 6 had significant coronary artery disease, according to the selection criteria. None of the patients in group 1 had significant acute comorbidities, while the prevalence of infective, haemorrhagic, and traumatic disease was similar in the five remaining groups. All patients in groups 2 and 3 were treated with aspirin (81–325 mg/day).
Widespread coronary and myocardial inflammation
T-lymphocytes in the myocardium were found in all group 1, 2, 3 and 4 patients, but in none of groups 5 and 6 patients (p<0.001). Virtually all T-lymphocytes were in an activated state coexpressing the CD3 and DR antigens (fig 1), as previously reported.1 4
The infiltrate was widespread (present in both peri-infarct and remote regions) in all group 1 and 2 patients, in 67% of group 3 patients, and in 100% of group 4 patients (p = NS). The number of T-lymphocytes in the peri-infarct myocardium was different in the different groups, with the infiltrate being significantly greater in the peri-infarct area (vs remote) in group 2 (p<0.001), but not in group 1 and group 3. Table 2 and figs 2 and 3 show median and interquartile values for each group. The number of T-lymphocytes in group 4 was extremely variable between different fields, regions, and cases, ranging from few activated lymphocytes (23–32 cells/mm2, grade I) to many (>73 cells/mm2, grade IV).
In group 2 and group 3 patients, the intensity of the T-lymphocyte infiltrate was independent of the presence or absence of symptoms/signs of heart failure, sepsis, or other comorbidities. Moreover, the inflammatory infiltrate intensity was independent of the infarct size at pathology in groups 2 and 3, or the number of vessels with significant disease in groups 1, 2 and 3.
No T-lymphocytes were found in group 5 (p<0.001 vs remote and peri-infarct regions in groups 1, 2 and 3 and vs group 4) or in group 6 (p<0.001 vs remote and peri-infarct regions in groups 1, 2 and 3 and vs group 4).
ICAM-1 expression in cardiomyocytes was found in 92% of cases of group 1 and 100% of cases in groups 2 and 3 (both in peri-infarct and remote regions), and in 92% of cases in group 4 (p = NS)(fig 1). The intensity of ICAM-1 expression was significantly higher in the peri-infarct area in group 2 cases vs groups 1 and 3. ICAM-1 expression intensity in the remote areas was similar among the three groups. Group 4 subjects had significantly higher ICAM-1 intensity than the remote regions in groups 1–3 and peri-infarct regions in groups 1 and 3 (figs 2 and 3).
Activated T-lymphocytes within the epicardial coronary artery wall were found in both the infarct- and non-infarct-related arteries in 83–92% of cases in groups 1–3, and in none of the cases in groups 4–6 (p<0.001; figs 1 and 2). The infiltrate was present in the intima, media and adventitia of the vessels in all cases. In 83% of cases it was present also in the epicardial adipose tissue. In 25% of cases an adventitial granuloma was seen (fig 1G).
No correlation between the cause of death or comorbidities and inflammatory variables was found in groups 2–5 and no correlation between clinical variables (such as multiple MI and reperfused MI) in groups 2 and 3 and inflammatory variables was found (data not shown).
DISCUSSION
This study shows for the first time the presence of a lymphocytic infiltrate in the myocardium and coronary arteries, and a proinflammatory phenotype shift in the myocardium that accompanies acute coronary thrombosis in patients dying suddenly, shortly, or even late after coronary thrombosis. Indeed, we describe, for the first time, a significantly higher prevalence of activated T-lymphocytic infiltrate both in the peri-infarct and remote myocardial regions in patients who died suddenly with acute coronary thrombosis and also in patients who died 4–7 months after AMI, in comparison with patients with hypertensive heart failure or controls. Thus, the infiltrate may not be the mere consequence of acute necrosis as it is present as early as a few hours after coronary thrombosis and it persists chronically up to 7 months later. In the same cases we show myocardial expression of adhesion molecules demonstrating a tissue response to inflammation. Considering that the expression of ICAM-1 is dependent upon new protein synthesis and unrelated to the ischaemic insult,5 the presence of ICAM-1 in subjects who died within 2 hours of symptoms (group 1, sudden coronary death) suggests that the stimuli for myocardial inflammation were present several hours before death. Moreover, the lack of activated lymphocytes in patients with symptomatic hypertensive heart failure makes the hypothesis of a non-specific response to cell damage/death unlikely.
Although, the pivotal role of inflammation in ACS is generally accepted, the concept that inflammation is diffuse throughout the entire coronary tree, myocardium, and even to remote arterial districts is relatively new.[1, 2, 4, 6 to 8] Neri-Serneri et al6 and Abbate et al2 independently described intramyocardial inflammatory infiltrates in patients with ACS (unstable angina and AMI, respectively). However, the causes, course and destiny of such infiltrates over time is undefined. The new findings presented in this paper may help to rule out some possibilities and focus on others.
A widespread immune infiltrate in the myocardium after AMI may be related to the causes of AMI or may be secondary to an autoimmune response elicited by the occurrence of myocardial necrosis,9 as suggested by some experimental studies. The extrapolation of these experimental results to the clinical situation, however, is difficult.10 11 Our findings differ from those reported in the animal models10 11 for several reasons: (a) the smaller extent of the lymphoid infiltration, which, although significant, is represented by a low-grade infiltrate with scattered T-cells; (b) the finding of significant infiltrate also in cases of sudden coronary death (within 6 hours), and the lack of a time-related pattern of T-cell infiltration in remote regions; and (c) the lack of cardiomyocyte necrosis associated with the infiltrate. These findings argue against the hypothesis of a non-specific response to necrosis and/or an autoimmune myocarditis that implies CD8-mediated tissue destruction.11 12 Indeed, when compared with subjects with myocarditis (group 4), those with sudden coronary death, recent AMI or healing AMI (groups 1–3) not only had a smaller amount of infiltrating T-lymphocytes but also had no evidence of cytotoxicity of the infiltrate.
Notably, in sudden coronary death and in healed AMI the intensity of the infiltrate was similar in the remote compared with peri-infarct regions, while being significantly higher in the peri-infarct area (vs remote) in recent AMI, thus confirming that ischaemic necrosis appears to be only one of several possible stimuli responsible for myocardial inflammation associated with AMI. Possibly, subjects dying suddenly had had brief episodes of ischaemia in the days before death eliciting an inflammatory response in the myocardium, although this is not supported by the review of the clinical chart in any case.
The lack of activated T-lymphocytes in cases of symptomatic hypertensive heart failure makes the hypothesis of a non-specific response to cell damage/death even more unlikely. Conversely, the presence of inflammation in ACS as early as 2 hours after coronary thrombosis, and with or without evidence of necrosis seems to confirm that inflammation is a pathogenetic event related to the cause of coronary instability, rather than a consequence of myocardial necrosis.2 4 6 Also, the observation that the infiltrate is smaller in cases of sudden coronary death compared with cases of evolving AMI, lends further support to the hypothesis that myocardial necrosis might simply amplify a pre-existing inflammatory process (favouring an increase in CD4+ lymphocytes over CD8+ cells). Accordingly, the activated lymphocyte infiltrate is approximately 10 times higher in AMI2 than in unstable angina.6
The notion that activated T-lymphocytes persist in the myocardium after AMI may lead to a more focused search for the candidate triggers of inflammation associated with ACS. Characterisation of the myocardial and coronary infiltrates in cases with ACS (sudden coronary death, recent and healing AMI (groups 1–3)) and in cases with myocarditis (group 4) shows a lymphocytic infiltrate in the myocardium in all the above-mentioned conditions (with evidence of cytotoxicity in the myocarditis group only) and a lymphocytic infiltrate also in the coronary arteries of the patients with ACS and not in those with myocarditis. This may suggest different culprit antigens in the two processes. Of note, myocardial activated T-lymphocyte infiltration has recently been reported in 51 patients with “idiopathic” dilated cardiomyopathy symptomatic for angina and heart failure and with evidence of endothelial dysfunction.13 Notably, myocardial inflammation was associated with endothelial cell activation and dysfunction but not correlated with the presence of virus DNA/RNA.13 The similarities encountered in ACS and dilated cardiomyopathy (where an autoimmune mechanism may have an important pathogenic role) may suggest similar deregulated autoimmunity in the two diseases with different target antigens, common to coronary vessels and to myocardium in the former, confined to the myocardium in the latter.
The search for such antigens, therefore, remains challenging while stimulating the speculative, yet untested, hypothesis of possible immune therapy for ACS. Chlamydia pneumoniae is a potential candidate but a pathogenetic link is not supported.14 The possibility of antigen-independent, ischaemia-driven T-cell activation needs also to be considered,15 although a link between regional ischaemia and widespread myocardial inflammation seems unlikely.4 6
In conclusion, the present study, although limited in its observational postmortem design, small sample size and lack of bio-humoral characterisation of the inflammatory status before death, describes the presence of a lymphocytic infiltrate in the myocardium and coronary arteries and a proinflammatory shift in the myocardium associated with acute coronary thrombosis in patients dying suddenly, shortly, or even late after coronary thrombosis. The finding of diffuse coronary and myocardial inflammation has been consistently reported,2 4 6 16 17 and the presence of active inflammation seems to be the major determinant for clinical instability, more so than the presence of the so-called “vulnerable” plaques.16 Mauriello et al have indeed shown that, while “vulnerable” plaques are seen also in stable patients and controls, the presence of diffuse and active inflammation is specific for acute myocardial infarction.16 These findings may open new avenues for the study of the triggers of inflammation in ACS.
Acknowledgments
We thank Dr Vera Di Trocchio (Richmond, VA, USA) for her editorial support.
REFERENCES
Footnotes
This paper is dedicated to the memory of Agostino Abbate who died suddenly on the Great Wall of China in February 2005.
Competing interests: None.