Systemic and Intraplaque Mediators of Inflammation Are Increased in Patients Symptomatic for Ischemic Stroke
Background and Purpose— The concept of “vulnerable plaque” has been extended to the more recent definition of the “cardiovascular vulnerable patient,” in which “intraplaque” and “systemic” factors contribute to the cumulative risk of acute cardiovascular events. Thus, we investigated the possible role of systemic and intraplaque inflammation in patients asymptomatic versus symptomatic for ischemic stroke.
Methods— Regions upstream and downstream the blood flow were isolated from internal carotid plaques of patients asymptomatic (n=63) or symptomatic (n=18) for ischemic stroke. Specimens were analyzed for lipid, collagen, macrophage, lymphocyte, neutrophil, mast cell and smooth muscle cell content, and chemokine and cytokine mRNA expression. Chemokine receptors and adhesion molecules were assessed on circulating leukocytes by flow cytometry. Systemic inflammatory markers and biochemical parameters were measured on total blood, plasma, and serum.
Results— Tumor necrosis factor-α and CCL5 serum levels as well as intercellular adhesion molecule-1 expression on circulating neutrophils were increased in symptomatic as compared with asymptomatic patients. Collagen content and smooth muscle cell infiltration were decreased in symptomatic plaques. In upstream regions of symptomatic plaques, lipid content and lymphocyte infiltration were increased. In downstream regions of symptomatic plaques, macrophage, neutrophil, and mast cell infiltration were increased. Intraplaque collagen content was positively correlated with smooth muscle cell infiltration and inversely correlated with macrophages, neutrophils, or serum tumor necrosis factor-α. Collagen reduction in downstream regions and serum tumor necrosis factor-α were independently associated with the likelihood of being symptomatic.
Conclusions— Inflammatory mediators are increased in ischemic stroke. Despite statistically significant, the correlation between tumor necrosis factor-α serum level and intraplaque vulnerability was weak and probably of limited biological importance.
Variations in plaque composition, size, and severity of lumen stenosis have been identified as crucial aspects of plaque vulnerability.1,2 Inflammation, thin or fissured cap with large lipid core, and severe stenosis increase plaque vulnerability. In addition, superficial calcified nodules, hemorrhages, endothelial dysfunction, and expansive (positive) remodeling could also contribute to atherosclerotic plaque destabilization. More recently, clinical studies suggested that acute ischemic events could be also influenced by systemic factors and peripheral tissue resistance to hypoxia.3 Nonspecific serum markers of altered lipid profile, inflammation, and hypercoagulability have been identified.4 Thus, the concept of “vulnerable plaque” has been extended to the more recent definition of the “cardiovascular vulnerable patient,” in which “intraplaque” and systemic factors contribute to the cumulative risk of acute cardiovascular events.4 This composite risk score should be also considered as the most accurate method of risk stratification in both primary and secondary prevention. In patients symptomatic for stroke, inflammatory processes could favor recurrent stroke and carotid restenosis after vascular intervention.5 Thus, to identify some targets for possible selective anti-inflammatory treatments against stroke, we investigated the possible role of “intraplaque” and “systemic” inflammation in plaque remodeling and vulnerability after stroke. In particular, we compared asymptomatic patients with advanced carotid atherosclerosis with symptomatic subjects in early phases after stroke (within 30 days). To better assess the possible shear stress influence on intraplaque inflammation, internal carotid plaques were divided perpendicularly to the long axis through the point of maximum stenosis to obtain 2 regions (upstream and downstream the blood flow). Systemic inflammation was investigated through the detection of circulating inflammatory markers (previously shown as associated with cardiovascular diseases) and the assessment of adhesion molecule and chemokine receptor expression on circulating leukocytes.6
Patients and Study Design
We conducted an unmatched case-control study between March 2008 and March 2009 at a single hospital (San Martino Hospital) in Genoa, Italy. A case was defined as any symptomatic patient with a first episode of ipsilateral ischemic stroke, which had occurred in the period between 30 and 10 days before endarterectomy. Ischemic stroke was defined as ipsilateral focal neurological deficit of acute onset lasting >24 hours. A control subject was defined as an asymptomatic patient who had no personal history of ischemic symptoms, presenting a severe internal carotid stenosis incidentally diagnosed at ultrasound Doppler during the same time period as cases. Symptomatic (n=18) and asymptomatic (n=63) patients underwent carotid endarterectomy for extracranial high-grade internal carotid stenosis (>70% luminal narrowing). The day before endarterectomy, blood samples were obtained by peripheral venipuncture from these patients at fasting state to collect serum and plasma and to perform blood parameters and flow cytometry analysis of adhesion molecule and chemokine receptor expression on circulating leukocyte. The degree of luminal narrowing was determined by repeated Doppler echography and angiographic confirmation using the criteria of the North American Symptomatic Carotid Endarterectomy Trial (NASCET).7 The indication for carotid endarterectomy for asymptomatic patients was based on the recommendations published by the Asymptomatic Carotid Surgery Trial (ACST) and the indication for patients symptomatic was according to the recommendations of the European Carotid Surgery Trial (ECST) and the NASCET.8–10 Medications reported in Supplemetal Table I (available at http://stroke.ahajournals.org) were not modified in the 2 months before enrollment.
All patients who developed spontaneous cerebral embolism during 30 minutes preoperatively and during the dissection phase of the operation (detected by transcranial Doppler insonation of the middle cerebral artery) were excluded from the study. Other exclusion criteria were malignant hypertension, acute coronary artery disease, any cardiac arrhythmias, congestive heart failure (New York Heart Association II, III, and IV classes), liver or renal disorders or function abnormalities, acute and chronic infectious disease, autoimmune and rheumatic diseases, cancer, endocrine diseases, inflammatory bowel diseases and anti-inflammatory (other than aspirin) medications, oral anticoagulant treatments, hormone, cytokine, or growth factor therapies.
The Medical Ethics Committee of San Martino Hospital approved the study, and participants provided written informed consent. The study was conducted in compliance with the Declaration of Helsinki.
Power calculation was based on prior published literature.11–13 Our sample size (18 cases and 63 control subjects) allowed us to detect large effect sizes (>0.80) for systemic inflammatory markers or intraplaque parameters between symptomatic and asymptomatic patients with a power of 80% taking a 2-sided Type I error of 5%.
Systemic Inflammatory Marker Detection
Serum C-reactive protein, soluble CD40 ligand, osteoprotegerin, tumor necrosis factor (TNF)-α, CCL2, CCL3, CCL4, CCL5, adiponectin, leptin, resistin, vascular cell adhesion molecule-1, intercellular adhesion molecule-1 (ICAM-1), P-selectin, and E-selectin levels were measured by colorimetric enzyme-linked immunosorbent assay (R&D Systems, Minneapolis, Minn) following the manufacturer’s instructions. Serum insulin levels were measured by colorimetric enzyme-linked immunosorbent assay (Millipore, St Charles, Mo) following the manufacturer’s instructions. Sodium citrate plasma d-dimer levels were measured by enzyme-linked fluorescent assay (Roche Diagnostics Systems, Basel, Switzerland) following the manufacturer’s instructions. The limits of detection were 31.25 pg/mL for C-reactive protein, 15.625 pg/mL for soluble CD40 ligand, 62.5 pg/mL for osteoprotegerin, 15.625 pg/mL for TNF-α, 15.6 pg/mL for CCL2, 7.80 pg/mL for CCL3, 15.625 pg/mL for CCL4, 15.60 pg/mL for CCL5, 62.50 pg/mL for adiponectin, 31.25 pg/mL for leptin, 31.25 pg/mL for resistin, 15.625 pg/mL for vascular cell adhesion molecule-1, 15.60 pg/mL for ICAM-1, 125 pg/mL for P-selectin, 93.75 pg/mL for E-selectin, 2 μU/mL for insulin, and 45 ng/mL for d-dimer. Mean intra- and interassay coefficients of variation were <8% for all markers. Glucose, triglycerides, total cholesterol, low-density lipoprotein cholesterol, and high-density lipoprotein cholesterol were routinely measured and expressed in milligrams per deciliter.
Flow Cytometry Analysis of Circulating Leukocytes
Peripheral blood samples (50 μL, obtained from 71 [n=55 asymptomatic and n=16 symptomatic] of total 81 patients of the study the day before endarterectomy) were incubated with antihuman CC chemokine receptor-1-phycoerythrin conjugated antibody (Ab; R&D Systems), antihuman CC chemokine receptor-2-PE conjugated Ab (R&D Systems), antihuman ICAM-phycoerythrin conjugated Ab (BD Pharmingen, Torreyana, Calif), antihuman CD11b-fluorescein isothiocyanate conjugated Ab (Invitrogen, Carlsbad, Calif), antihuman CD15-FITC Ab (selective marker of circulating neutrophils, from Invitrogen Corporation, Camarillo, Calif), antihuman CD14-PE Ab (selective marker of circulating monocytes, from BD Pharmingen™) as well as with corresponding isotype controls. After staining, blood samples were incubated 10 minutes with human erythrocyte lysis buffer (R&D Systems) at room temperature, washed with phosphate-buffered saline, and analyzed by flow cytometry. Cell Quest software (BD Biosciences, Heidelberg, Germany) was used for data analysis. The results of flow cytometry experiments were expressed as medians (interquartile range) of relative fluorescence intensity calculated as the ratio between the mean fluorescence intensity of cells stained with the selective Ab and the mean fluorescence intensity of cells stained with the corresponding isotype control.
Endarterectomy Specimen Processing
Shortly after surgical excision, the internal carotid plaque specimens were taken from all patients and immediately transferred at 4°C to the laboratory for processing. Carotid plaques had the morphology of advanced lesions with a lipid-rich acellular core. All atherosclerotic plaques were cut perpendicular to the long axis through the point of maximum stenosis to obtain 2 portions (upstream and downstream the blood flow; Supplemental Figure I, available at http://stroke.ahajournals.org). Each portion was further divided perpendicular to the long axis in the middle in 2 subsegments. Half was snap-frozen in liquid nitrogen and stored at −80°C (for chemokine and cytokine mRNA isolation), and the other half was frozen in cryoembedding medium for histological analyses.
Frozen upstream and downstream carotid specimens were serially cut in 7-μm sections. Eight sections per each portion separated by 105 μm from each other were fixed in acetone at room temperature and immunostained with specific antibodies antihuman smooth muscle actin (dilution: 1:100; Dako Corporation, Glostrup, Denmark), antihuman CD68 (dilution: 1:100; Dako Corporation), antihuman CD3 (dilution: 1/25; Dako Corporation), antihuman CD66b (dilution: 1:50; Beckman Coulter), and antihuman mast cell tryptase (dilution: 1:100; Dako Corporation).14 Sections were counterstained with Mayer hemalune and rinsed in water. Quantifications of cell infiltration were performed with MetaMorph software. Data were calculated as percentages of stained area on total lesion area (for CD68+ cells) and as number of infiltrated cells on mm2 of total lesion area (for CD3+, CD66b+, or mast cell tryptase+ cells).
Sirius Red Staining for Collagen Content
Eight sections per each portion separated by 105 μm from each other (upstream and downstream the blood flow) were rinsed with water and incubated with 0.1% Sirius red (Sigma Chemical Co) in saturated picric acid for 90 minutes. Sections were rinsed twice with 0.01 N HCl for 1 minute and then immersed in water. After dehydration with ethanol for 30 seconds and coverslipping, the sections were photographed with identical exposure settings each section under ordinary polychromatic or polarized light microscopy. Total collagen content was evaluated under polychromatic light (Sirius red). Interstitial collagen subtypes were evaluated using polarized light illumination; under this condition, thicker Type I collagen fibers appeared orange or red, whereas thinner Type III collagen fibers were yellow or green.11 Quantifications were performed with MetaMorph software. Data were calculated as percentages of stained area on total lesion area.
Oil Red O Staining for Lipid Content
Eight sections separated by 105 μm from each other per each portion (upstream and downstream the blood flow) were incubated in 60% isopropanol for 2 minutes and then in Oil Red O solution for 20 minutes and rinsed in phosphate-buffered saline. Sections were then counterstained with Mayer hemalune and rinsed in distilled water. Quantifications were performed with MetaMorph software. Data were calculated as percentages of stained area on total lesion area.
Real-Time Reverse Transcription-Polymerase Chain Reaction
Total mRNA was isolated with Tri-reagent (MRC Inc) from upstream or downstream specimens of human carotid plaques. Reverse transcription was performed using the ImProm-II Reverse Transcription System (Promega, Madison, Wis) according to the manufacturer’s instructions. Real-time polymerase chain reaction (StepOne Plus; Applied Biosystems) was performed with the ABsolute QPCR Mix (ABgene).
Specific primers and probes (Supplemental Table II, available at http://stroke.ahajournals.org) were used to determine the mRNA expression of CCL2, CCL3, CCL4, CCL5, CX3CL1, CXCL8, TNF, and glyceraldehyde-3-phosphate dehydrogenase (housekeeping gene). The fold change of mRNA levels was calculated by the comparative Ct method. The resultant Ct values were first normalized to the internal control. This was achieved by calculating a delta Ct (ΔCt) by subtracting the internal control Ct values from the glyceraldehyde-3-phosphate dehydrogenase Ct value. A delta delta Ct (ΔΔCt) was calculated by subtracting the designated control ΔCt value from the other ΔCt values. The ΔΔCt was then plotted as a relative fold change with the following formula: 2−ΔΔCt.
Patient characteristics were described 1 day before endarterectomy. Cases (symptomatic patients) were compared with control subjects (asymptomatic patients) using Pearson χ2 test or Fisher exact test, when appropriate, for the comparison of qualitative variables and Mann-Whitney nonparametric test (the normality assumption of the variables’ distribution in both groups was violated) for comparisons of continuous variables. The comparisons between upstream and downstream portions of carotid plaques within asymptomatic and symptomatic groups were performed using the Wilcoxon signed rank test. For continuous variables, results were expressed as medians (interquartile range).
Spearman rank correlation coefficients were used to assess correlations between serum markers levels, intraplaque infiltration of vascular and inflammatory cells, and intraplaque collagen content in both upstream and downstream regions of carotid atherosclerotic plaques. To assess the independent impact of several systemic inflammatory markers and intraplaque inflammatory mediators on the likelihood to be a symptomatic case, we used a parsimonious multivariable logistic regression model providing maximum likelihood estimates of the OR and its 95% CIs; we used a robust estimation of the variance due to the small number of events per independent variable leading to the risk of deviations from the distributional assumption.15 We verified the log linearity for all continuous variables and decided to use those reclassified in categorical variables (defined with quartiles’ cutoffs) when log linearity was not respected (ICAM-1 expression on circulating neutrophils, percentages of intraplaque macrophages, number of intraplaque lymphocytes, neutrophils, and mast cells). TNF-α was used transformed as 1/10th of initial value to assess the likelihood to be symptomatic for each increase of 10 pg/mL TNF-α. We controlled for the adequacy of our model using the Hosmer and Lemeshow goodness-of-fit test regrouping the predicted probabilities in 8 equal-sized groups due to a number of covariates equal to the number of observations. Values of P<0.05 (2-tailed) were considered significant. All analyses were done with Statistical Package for the Social Sciences, Version 11.0 (SPSS Inc, Chicago, Ill).
Clinical characteristics, biological parameters as well as medications in asymptomatic and symptomatic patients are described in Supplemental Table I. There was no significant difference between asymptomatic and symptomatic patients in terms of age, sex, classical risk factors for stroke, and medications. Antiplatelet treatment was more frequent in asymptomatic than symptomatic patients (90.5% versus 66.7%, P=0.021). Carotid lumen stenosis was significantly increased (80.0% versus 90%, P=0.020) in symptomatic as compared with asymptomatic patients.
Systemic Inflammatory Markers and Circulating Leukocyte Activation Are Increased in Symptomatic Patients
As shown in Table 1, serum TNF-α and CCL5 levels were significantly increased in symptomatic versus asymptomatic patients (TNF-α: P=0.001; CCL5: P=0.022, respectively). No difference was observed in the levels of other serum inflammatory markers (Table 1). As shown in Table 2, ICAM-1 expression on circulating neutrophil surface membrane was significantly (P=0.019) increased in symptomatic as compared with asymptomatic patients. Although a slight increase was observed in symptomatic patients, ICAM-1 and CC chemokine receptor-1 expression on circulating monocytes did not change significantly between the 2 groups (Table 2). No difference between the 2 groups was observed on leukocyte surface expression of other assessed molecules (Table 2).
Intraplaque Lipid Content Is Increased, Whereas Collagen and Smooth Muscle Cell Infiltration Are Decreased in Symptomatic Patients
Lipid content was significantly higher in upstream regions of symptomatic plaques in comparison with upstream portions of asymptomatic plaques and downstream portions of symptomatic plaques (Figure 1A–B). Conversely, total collagen, collagen I, and collagen III contents were increased in upstream regions of asymptomatic and symptomatic plaques as compared with their downstream portions (Figure 2A–D). Total collagen content was higher in both upstream and downstream regions of asymptomatic plaques as compared with symptomatic plaques (Figure 2A). Accordingly, smooth muscle cell (SMC) infiltration was increased in both upstream and downstream regions of asymptomatic as compared with symptomatic plaques (Figure 2E–F). Furthermore, upstream regions of asymptomatic plaques had increased SMC infiltration in comparison with their downstream portions (Figure 2E–F).
Intraplaque Inflammatory Cell Infiltration Is Increased in Symptomatic Patients
In both asymptomatic and symptomatic plaques, downstream regions had increased macrophage infiltration as compared with their upstream portions (Figure 3A–B). In symptomatic plaques, upstream regions had increased immunoreactivity for lymphocytes in comparison with downstream regions and upstream regions of asymptomatic plaques (Figure 3C-D). Neutrophil infiltration was higher in symptomatic as compared with asymptomatic plaques (Figure 3E–F). Furthermore, downstream regions of symptomatic plaques had increased neutrophil infiltration in comparison with upstream portions (Figure 3E–F). Mast cell infiltration was increased in downstream regions of symptomatic as compared with asymptomatic plaques (Figure 3G–H).
Intraplaque Collagen Content Is Positively Correlated With SMC and Inversely Correlated With TNF-α Serum Levels and Macrophage and Neutrophil Infiltration
TNF-α serum levels inversely correlated with total collagen content in upstream plaques (r=−0.251, P=0.031, Supplemental Figure II, available at http://stroke.ahajournals.org). No significant Spearman rank correlations were observed between intraplaque collagen content and other systemic inflammatory makers (data not shown).
In upstream regions of carotid plaques, collagen I content was positively correlated with intraplaque SMC infiltration (r=0.318, P=0.005; Supplemental Table III, available at http://stroke.ahajournals.org). Conversely, total collagen content was inversely correlated with macrophage infiltration (r=−0.250, P=0.029).
In downstream regions of carotid plaques, total collagen, collagen I, and collagen III contents were positively correlated with SMC infiltration (respectively: r=0.407, P<0.001; r=0.718, P<0.001; r=0.702, P<0.001; Supplemental Table III). Conversely, collagen I and III contents were inversely correlated with macrophage infiltration (r=−0.284, P=0.011; r=−0.224, P=0.047, respectively). Total collagen was inversely correlated with intraplaque infiltrated neutrophils (r=−0.472, P<0.001; Supplemental Table III). No significant Spearman rank correlations were observed between collagen content and other inflammatory cell infiltration in upstream and down regions of carotid plaques (Supplemental Table III).
Collagen Content in Downstream Regions Is Independently Associated With the Likelihood of Being Symptomatic for Ischemic Stroke
We also assessed the independent association of several systemic markers and intraplaque parameters of vulnerability with the likelihood to be symptomatic. Several variables were significantly associated with the risk of symptomatic stroke at univariate (antiplatelet use, percentage of carotid stenosis, serum TNF-α, percentages of lipid and total collagen in upstream regions, and total collagen and number of infiltrated in downstream regions), but due to collinearity and interdependence between several of them, we made a selection of the covariates to be included in the multivariate model. The final model presented in Table 3⇓ (multivariate model) fits reasonably well (χ2 test, P=0.92) and the area under the curve is high (0.969). Total collagen content in downstream regions of carotid plaques reduced the likelihood to be symptomatic with an OR of 0.59 (95% CI: 0.42 to 0.82, P=0.002). The other significant predictor was serum TNF-α, which slightly increased the likelihood to be symptomatic for each increase of 10 pg/mL of TNF-α (OR: 1.12: 95% CI: 1.01 to 1.24, P=0.04). We also adjusted the model on the percentage of carotid stenosis, antiplatelet use, and the proportion of lipid in upstream regions of carotid plaques. These parameters were near the significance level (between 0.05 and 0.1), except for the percentage of carotid stenosis (P=0.31).
Inflammatory Chemokine and Cytokine mRNA Expression Is Increased in Downstream Regions of Carotid Plaques
As shown in Supplemental Figure III (available at http://stroke.ahajournals.org), CCL2, CCL3, CCL4, TNF-α, and CX3CL1 mRNA expression was increased in downstream regions of asymptomatic plaques as compared with their upstream regions. A marked nonsignificant increase in CXCL8 and TNF-α mRNA expression was found in downstream regions of symptomatic plaques as compared with the correspondent portions in asymptomatic patients (downstream regions asymptomatic versus symptomatic: CXCL8, P=0.052; TNF-α, P=0.078).
We investigated the potential influence of both systemic and local inflammation on global cardiovascular vulnerability in patients asymptomatic and symptomatic for ischemic stroke undergoing endarterectomy. Systemic inflammatory markers, circulating leukocytes, and intraplaque inflammation were assessed in the first period (within 30 days) after stroke. This short period avoided the potential interference of time.12 Several exclusion criteria were introduced to reduce the possible influence of inflammatory and infectious diseases and cancer on atherosclerotic inflammation. To assess the possible influence of shear stress, we also divided carotid plaques in upstream and downstream regions considering the blood flow direction. In carotid plaque portions, we assessed markers of vulnerability such as lipid and collagen content as well as inflammatory and vascular cell infiltration. This research approach to assess different cell distribution within atherosclerotic plaques was already performed.13 However, authors only focused on plaque shoulders, whereas in our work, we considered the entire plaque (also including lipid core). Thus, we designed the present study considering both systemic and intraplaque vulnerability to better assess the global risk of stroke. No differences in clinical characteristics between asymptomatic and symptomatic groups were observed, except for antiplatelet therapy (significantly more frequent in asymptomatic patients) and carotid lumen stenosis (significantly increased in symptomatic group). These differences might be explained by clinical investigations indicating that antiplatelet therapy induces a small reduction of cardiovascular risk and that carotid stenosis increases incidence of stroke.16 However, no correlation was found between carotid stenosis and local or systemic inflammation (data not shown). On the other hand, the regression logistic model (corrected for difference between symptomatic and asymptomatic patients due to carotid stenosis or antiplatelet use) assessed the association between a symptomatic stroke and the value of independent biological inflammatory markers and intraplaque components (Table 3⇑). Concerning systemic inflammation, we showed that only TNF-α and CCL5 levels were increased in symptomatic as compared with asymptomatic patients. TNF-α serum levels inversely correlated with upstream total collagen content, suggesting that, among several systemic markers, only TNF-α is an inflammatory mediator associated with an intraplaque parameter of vulnerability. TNF-α serum levels were also significantly associated with the likelihood of being symptomatic for stroke in a univariate model. This association was also confirmed by a multivariate model, indicating that TNF-α level was an independent risk factor for stroke. However, although these correlations were statistically significant, they were very weak. Thus, we doubt that TNF-α could be considered a biologically relevant systemic marker of intraplaque vulnerability. Previous clinical studies have already shown that TNF-α levels are increased in the first week after ischemic stroke and correlates with infarct volume and severity of neurological impairment.17 Elevated blood concentrations of TNF-α could be due to its intracerebral and/or peripheral increased production. A strong release of TNF-α has been found in blood cells collected the early days after stroke. This elevated production could play a dual (either cytoprotective or toxic) role in stroke depending on the concentrations and receptor activation.17 The increase of TNF-α levels after stroke could represent an important limitation in our study design. A more rational approach could be represented by a prospective study looking at the ability of systemic markers to predict cerebrovascular events. However, because it was clearly not possible to have symptomatic carotid plaques before an ischemic stroke, we planned to collect blood samples from symptomatic patients exclusively in the period between 10 and 30 days after the acute ischemic event. Furthermore, blood samples were collected also the day before endarterectomy to avoid the possible interference of the surgical procedure. Importantly, the low values of other inflammatory markers (including C-reactive protein that has been shown elevated after an ischemic stroke) further support the quality of this scientific approach to reduce possible bias due to stroke itself.18
Little is also known about the possible role of CCL5 in ischemic stroke. Zaremba and coworkers reported no elevation of CCL5 serum levels in the first 3 days after ischemic stroke.19 Although a significant increase of CCL5 levels in symptomatic patients has been observed, a univariate model did not confirm that circulating CCL5 is associated with the risk of symptomatic ischemic stroke. This lack of significance could be explained by the power of our sample size, whereas for CCL5, we only show an effect size <0.6 between symptomatic and asymptomatic patients (data not shown). The possible effects of systemic inflammation were also investigated in circulating neutrophils and monocytes. Although blood leukocyte count did not significantly differ between groups, ICAM-1 was upregulated on circulating neutrophils in symptomatic patients. Several controversies could be raised by these surprising data. ICAM-1 expression on human neutrophil surface membrane is controversial and partially influenced by leukocyte isolation methods.20 Indeed, some authors did not detect ICAM-1 on human neutrophils.21 The classical paradigm of intracerebral neutrophil recruitment during reoxygenation shows CD11b upregulation (a selective ligand of ICAM-1) on neutrophil membrane and ICAM-1 upregulation on human brain microvascular endothelial cells.22,23 In the present article, fluorescence-activated cell sorter analysis showed that neutrophils were strongly positive for ICAM-1 (average percentage of ICAM-1+ neutrophils was, respectively, 56.9% in asymptomatic patients and 67.7% in symptomatic patients). We also detected ICAM-1 expression on human neutrophils from 9 healthy volunteers (average percentage of ICAM-1+ neutrophils: 55.5%, average median relative fluorescence intensity: 6.2), confirming that this adhesion molecule is expressed on human neutrophil surface membrane. Thus, the role of ICAM-1 upregulation on human neutrophils in symptomatic patients remains unclear. We recently provided in vitro evidence that ICAM-1 upregulation might modulate leukocyte activation.24 Thus, the enhanced expression of ICAM-1 on neutrophils in symptomatic patients might be involved in the increased neutrophil infiltration in their carotid plaques. Despite these limitations in understanding the molecular mechanisms of neutrophil recruitment, we observed a significant increase of infiltrated neutrophils and mast cells in downstream regions of symptomatic plaques as compared with the corresponding regions in asymptomatic plaques. Although highly speculative, differences in neutrophil infiltration between carotid plaque regions might be explained by the expression of neutrophil chemoattractants (such as CXCL8 and TNF-α) that are increased in downstream regions in symptomatic patients. Importantly, the number of neutrophils within downstream regions of atherosclerotic plaques was inversely correlated with collagen content. Neutrophil infiltration has been already shown in culprit lesions from coronary arteries obtained at autopsy.14 Our data suggest for the first time that the increased neutrophil infiltration in symptomatic plaques might be directly involved in the degradation of collagen, mediating plaque vulnerability. Vulnerability in downstream regions of symptomatic plaques could be increased by neutrophil-mediated inflammation. These data further support the pivotal role of neutrophils in atheroprogression and plaque vulnerability, as recently shown by Zernecke and coworkers in atherosclerotic mice.25 The significant reduction of collagen content in symptomatic plaques was also associated with the decreased infiltration of SMC (crucial cells in plaque stabilization). Thus, 2 possible complementary mechanisms (increase of inflammatory cell infiltration and decrease of SMC recruitment) might reduce collagen content in downstream regions of human carotid plaques. Collagen content plays a crucial protective role within atherosclerotic plaques.2,4 A multivariable logistic regression model showed that the reduction of collagen content in downstream regions of carotid plaques is an independent risk factor for symptomatic ischemic stroke. On the other hand, vulnerability in upstream regions could be increased by different mechanisms. We showed that T lymphocyte infiltration and lipid content were increased in upstream regions of symptomatic plaques as compared with the corresponding regions in asymptomatic plaques. However, the number of T lymphocytes did not significantly correlate with collagen content, indicating that mainly plaque structure (increased lipid content) might increase vulnerability in upstream regions. The predictive value of lipid content in upstream regions of carotid plaques has been also independently assessed by a multivariable logistic regression model but was not significantly associated with the symptomatic presentation, reflecting again a lack of power of our sample size.
The present study has further limitations. First, this observational study does not control for chronological time, which does not permit assessing the causative link between the increase of systemic inflammatory markers or circulating leukocyte activation and the risk of stroke recurrence. Second, our sample size was powered to show large differences (effect size >0.80) for the inflammatory markers or the intraplaque components between symptomatic and asymptomatic patients, thus failing to identify further independent biological predictors to be symptomatic (such as serum CCL5 or lipid content in upstream regions of carotid plaques), which were near the significance level.
In conclusion, systemic and local inflammation is increased in patients symptomatic for stroke. Although high levels of circulating CCL5 failed to be associated with the likelihood to be symptomatic, they might contribute to the increase of “systemic” vulnerability in symptomatic patients. Infiltrated neutrophils and macrophages were inversely associated with intraplaque collagen content and might influence plaque vulnerability. TNF-α serum levels were inversely correlated with upstream total collagen content. Downstream intraplaque collagen content and serum TNF-α were also independently associated with the likelihood to be symptomatic risk factors for ischemic stroke. However, despite being statistically significant, correlations using serum TNF-α were very weak. Thus, our data could hardly support serum TNF-α as a potential biological marker of intraplaque vulnerability for ischemic stroke.
Systemic and intraplaque inflammation could influence global patient vulnerability for ischemic stroke. We investigated the role of different soluble inflammatory markers and cells in the bloodstream and in upstream and downstream portions of carotid plaques in symptomatic and asymptomatic patients. Systemic and local inflammation is increased in patients symptomatic for stroke. Although high levels of circulating CCL5 failed to be associated with the likelihood to be symptomatic, they might contribute to the increase of “systemic” vulnerability in symptomatic patients. Infiltrated neutrophils and macrophages were inversely associated with intraplaque collagen content and might influence plaque vulnerability. TNF-α serum levels were inversely correlated with upstream total collagen content. Downstream intraplaque collagen content and serum TNF-α were also independently associated with the likelihood to be symptomatic risk factors for ischemic stroke. However, despite being statistically significant, correlations using serum TNF-α were very weak. Thus, our data could hardly support serum TNF-α as a potential biological marker of intraplaque vulnerability for ischemic stroke.
We thank Dr M.-L. Bochaton-Piallat from the Department of Pathology and Immunology of the University of Geneva for her technical advice.
Sources of Funding
This research was funded by EU FP7, grant number 201668, AtheroRemo, supported by a grant from the Swiss National Science Foundation (#310030-118245) and De Reuter Foundation to F.M. This work was funded by the “Sir Jules Thorn Trust Reg” fund to F.M.. This work was also funded by a fund by Carige Foundation to F.D.. L.R. is a recipient of a fellowship from Fondazione Italiana per la Lotta al Neuroblastoma.
- Received January 12, 2010.
- Revision received February 28, 2010.
- Accepted March 22, 2010.
Naghavi M, Libby P, Falk E, Casscells SW, Litovsky S, Rumberger J, Badimon JJ, Stefanadis C, Moreno P, Pasterkamp G, Fayad Z, Stone PH, Waxman S, Raggi P, Madjid M, Zarrabi A, Burke A, Yuan C, Fitzgerald PJ, Siscovick DS, de Korte CL, Aikawa M, Juhani Airaksinen KE, Assmann G, Becker CR, Chesebro JH, Farb A, Galis ZS, Jackson C, Jang IK, Koenig W, Lodder RA, March K, Demirovic J, Navab M, Priori SG, Rekhter MD, Bahr R, Grundy SM, Mehran R, Colombo A, Boerwinkle E, Ballantyne C, Insull W Jr, Schwartz RS, Vogel R, Serruys PW, Hansson GK, Faxon DP, Kaul S, Drexler H, Greenland P, Muller JE, Virmani R, Ridker PM, Zipes DP, Shah PK, Willerson JT. From vulnerable plaque to vulnerable patient: a call for new definitions and risk assessment strategies: part I. Circulation. 2003; 108: 1664–1672.
Naghavi M, Libby P, Falk E, Casscells SW, Litovsky S, Rumberger J, Badimon JJ, Stefanadis C, Moreno P, Pasterkamp G, Fayad Z, Stone PH, Waxman S, Raggi P, Madjid M, Zarrabi A, Burke A, Yuan C, Fitzgerald PJ, Siscovick DS, de Korte CL, Aikawa M, Airaksinen KE, Assmann G, Becker CR, Chesebro JH, Farb A, Galis ZS, Jackson C, Jang IK, Koenig W, Lodder RA, March K, Demirovic J, Navab M, Priori SG, Rekhter MD, Bahr R, Grundy SM, Mehran R, Colombo A, Boerwinkle E, Ballantyne C, Insull W Jr, Schwartz RS, Vogel R, Serruys PW, Hansson GK, Faxon DP, Kaul S, Drexler H, Greenland P, Muller JE, Virmani R, Ridker PM, Zipes DP, Shah PK, Willerson JT. From vulnerable plaque to vulnerable patient: a call for new definitions and risk assessment strategies: part II. Circulation. 2003; 108: 1772–1778.
Montecucco F, Mach F. Common inflammatory mediators orchestrate pathophysiological processes in rheumatoid arthritis and atherosclerosis. Rheumatology (Oxford). 2009; 48: 11–22.
Barnett HJ, Taylor DW, Eliasziw M, Fox AJ, Ferguson GG, Haynes RB, Rankin RN, Clagett GP, Hachinski VC, Sackett DL, Thorpe KE, Meldrum HE, Spence JD. Benefit of carotid endarterectomy in patients with symptomatic moderate or severe stenosis. North American Symptomatic Carotid Endarterectomy Trial Collaborators. N Engl J Med. 1998; 339: 1415–1425.
Halliday A, Mansfield A, Marro J, Peto C, Peto R, Potter J, Thomas D; MRC Asymptomatic Carotid Surgery Trial (ACST) Collaborative Group. Prevention of disabling and fatal strokes by successful carotid endarterectomy in patients without recent neurological symptoms: randomised controlled trial. Lancet. 2004; 363: 1491–1502.
Crisby M, Nordin-Fredriksson G, Shah PK, Yano J, Zhu J, Nilsson J. Pravastatin treatment increases collagen content and decreases lipid content, inflammation, metalloproteinases, and cell death in human carotid plaques: implications for plaque stabilization. Circulation. 2001; 103: 926–933.
Peeters W, Hellings WE, de Kleijn DP, de Vries JP, Moll FL, Vink A, Pasterkamp G. Carotid atherosclerotic plaques stabilize after stroke: insights into the natural process of atherosclerotic plaque stabilization. Arterioscler Thromb Vasc Biol. 2009; 29: 128–133.
Dirksen MT, van der Wal AC, van den Berg FM, van der Loos CM, Becker AE. Distribution of inflammatory cells in atherosclerotic plaques relates to the direction of flow. Circulation. 1998; 98: 2000–2003.
Naruko T, Ueda M, Haze K, van der Wal AC, van der Loos CM, Itoh A, Komatsu R, Ikura Y, Ogami M, Shimada Y, Ehara S, Yoshiyama M, Takeuchi K, Yoshikawa J, Becker AE. Neutrophil infiltration of culprit lesions in acute coronary syndromes. Circulation. 2002; 106: 2894–2900.
Kent JT. Robust properties of likelihood ratio tests. Biometrika. 1982; 69: 19–27.
Corso G, Bottacchi E, Brusa A, Benedetto MD, Giardini G, Lia C, Reggiani M, Veronese Morosini M. Is there a prognostic role for C-reactive protein in ischemic stroke? Acta Neurol Scand. 2009 Nov 19 [Epub ahead of print].
Meisel SR, Shapiro H, Radnay J, Neuman Y, Khaskia AR, Gruener N, Pauzner H, David D. Increased expression of neutrophil and monocyte adhesion molecules LFA-1 and Mac-1 and their ligand ICAM-1 and VLA-4 throughout the acute phase of myocardial infarction: possible implications for leukocyte aggregation and microvascular plugging. J Am Coll Cardiol. 1998; 31: 120–125.
Soriano SG, Coxon A, Wang YF, Frosch MP, Lipton SA, Hickey PR, Mayadas TN. Mice deficient in Mac-1 (CD11b/CD18) are less susceptible to cerebral ischemia/reperfusion injury. Stroke. 1999; 30: 134–139.
Hess DC, Zhao W, Carroll J, McEachin M, Buchanan K. Increased expression of ICAM-1 during reoxygenation in brain endothelial cells. Stroke. 1994; 25: 1463–1467.
Montecucco F, Steffens S, Burger F, Pelli G, Monaco C, Mach F. C-reactive protein (CRP) induces chemokine secretion via CD11b/ICAM-1 interaction in human adherent monocytes. J Leukoc Biol. 2008; 84: 1109–1119.
Zernecke A, Bot I, Djalali-Talab Y, Shagdarsuren E, Bidzhekov K, Meiler S, Krohn R, Schober A, Sperandio M, Soehnlein O, Bornemann J, Tacke F, Biessen EA, Weber C. Protective role of CXC receptor 4/CXC ligand 12 unveils the importance of neutrophils in atherosclerosis. Circ Res. 2008; 102: 209–217.