Angiotensin II Receptor Blockade Reduces Tachycardia-Induced Atrial Adhesion Molecule Expression
Background— Increased levels of inflammatory markers are predictors of thromboembolic events during atrial fibrillation (AF). Increased endocardial expression of adhesion molecules (ie, vascular cell adhesion molecule [VCAM] and intercellular adhesion molecule [ICAM]) could be an important link between initiation of inflammatory and prothrombogenic mechanisms responsible for thrombus development at the atrial endocardium (endocardial remodeling).
Methods and Results— Tissue microarrays were used to screen right atrial tissue specimens obtained from 320 consecutive patients for differences in atrial expression of the prothrombogenic proteins VCAM-1, ICAM-1, thrombomodulin, plasminogen activator inhibitor-1, and von Willebrand factor. An in vitro organotypic human atrial tissue model and a pig model of rapid atrial pacing were used to determine the therapeutic impact of angiotensin II receptor blockade. Immunohistochemical analyses showed that all prothrombogenic proteins are expressed by endocardial cells. Using multivariable analysis, only the intensity of VCAM-1 expression was increased in patients with AF (P=0.03). Increased atrial VCAM-1 expression was confirmed by Western blotting in patients with persistent and paroxysmal AF (persistent AF 207±42% versus sinus rhythm 100±16%, P=0.028; paroxysmal AF 193±42%, P=0.024 versus sinus rhythm). In vitro pacing of ex vivo human atrial tissue slices confirmed that rapid activation causes VCAM-1 upregulation (mRNA and protein levels). Pacing-induced VCAM-1 expression was abolished by olmesartan. To confirm this finding in vivo, VCAM-1 expression was determined in 14 pigs after rapid atrial pacing (600 bpm). Atrial tachycardia caused an upregulation of VCAM-1 expression, which was prevented by irbesartan, consistent with the observed increase in plasma levels of angiotensin II. Alterations in the in vivo VCAM-1 expression were more pronounced in the left atrium (>5-fold compared with sham) than in the right atrium (3.5-fold compared with sham).
Conclusions— AF and rapid atrial pacing both increase endocardial VCAM-1 expression, which can be attenuated by angiotensin II receptor blockade. This provides evidence that angiotensin II plays a pathophysiological role in prothrombotic endocardial remodeling.
Received July 30, 2007; accepted November 14, 2007.
Atrial fibrillation (AF) is the most common cardiac arrhythmia in clinical practice. It induces significant electrophysiological and structural changes of atrial myocardium and may also be of relevance for the development of atrial thrombus formation.1–5 In general, thrombi develop in the presence of reduced blood velocities, when the activity of the clotting system is increased, and in the presence of endothelial alterations (Virchow’s triad). This also applies to atrial thrombus formation during AF. Although the loss of regular atrial contractions reduces blood flow velocities, particularly in the atria and most prominently in the atrial appendages,6,7 AF is also associated with an activation of the plasmatic clotting system and of thrombocytes.8–10 However, to date, the contribution of endothelial alterations to atrial thrombogenesis has not been fully explored.
Clinical Perspective p 742
Previously, it has been shown that AF is associated with an inflammatory response.11,12 In turn, increased inflammatory markers (C-reactive protein, interleukin-6) and increased leukocyte-platelet interactions are predictors of atrial thrombus formation and thromboembolic stroke. An increased endothelial expression of adhesion molecules (vascular cell adhesion molecule-1 [VCAM-1] and intercellular adhesion molecule-1 [ICAM-1]) could be an important link between the initiation of proinflammatory and prothrombogenic mechanisms responsible for atrial thrombogenesis13–15; however, atrial expression of adhesion molecules has not yet been studied systematically in patients with AF.
The purpose of the present study was to systematically investigate the impact of AF on the expression of various prothrombotic atrial proteins in a large cohort of patients by use of tissue microarrays. Positive array results, which were obtained for VCAM-1, were validated by Western blotting. Subsequently, we aimed to determine the putative value of VCAM-1 as a treatment target. To that end, the effect of angiotensin II type 1 receptor blockers (ARBs) on VCAM-1 expression was determined in an organotypic human atrial tissue culture model. Interestingly, the ex vivo pacing experiments showed that angiotensin II type 1 receptor blockade reduces pacing-induced VCAM-1 expression in human atria. The potential therapeutic effect of ARBs was then tested with a closed-chest rapid atrial pacing model, which confirmed the beneficial effects of ARBs in reducing tachycardia-induced atrial VCAM-1 expression in vivo.
Right atrial appendages were obtained from 320 consecutive patients with (n=61) and without (n=259) AF undergoing cardiac bypass surgery or valve replacement. The clinical characteristics are shown in Table 1. Paroxysmal AF was present in 23 of (7%) 320 patients, whereas chronic persistent/permanent AF was present in 38 (12%; Table 2). To exclude the impact of covariables on the results of the tissue microarrays, analyses were also performed with 49 matched patient pairs with and without AF (pairs of patients were matched with regard to age [±5 years], gender, concomitant cardiovascular diseases, left ventricular ejection fraction, and type of cardiac surgery; Data Supplement, Table I). The study was approved by the local ethics committee, and all patients gave written informed consent to participate.
Tissue microarrays were generated from 320 donor paraffin blocks containing right atrial appendages. One tissue cylinder of 1.5-mm diameter and 3-mm depth was punched from the endocardial site of the donor block and placed into the recipient block by use of a precision instrument (Beecher Instruments, Silver Spring, Md). Paraffin sections (3 to 4 μm) were cut from each recipient block and stained with hematoxylin and eosin to confirm successful transfer of endocardium-containing atrial tissue (Figure 1).
For immunohistochemistry, deparaffinized serial sections were cut at 2 to 3 μm and placed on Superfrost Plus glass slides. Before immunostaining, sections were deparaffinized in xylene and rehydrated in an alcohol series. Different antigen retrieval procedures were performed. For the detection of VCAM-1 (monoclonal anti-VCAM-1 antibody) and plasminogen activator inhibitor-1 (PAI-1; polyclonal anti-PAI-1 antibody; both from Santa Cruz Biotechnology Inc, Santa Cruz, Calif), sections were pretreated with target retrieval buffer (pH 10; Dako Cytomation, Glostrup, Denmark) or 1 mmol/L EDTA (pH 8.0) for 20 minutes at 450 W in a microwave oven, respectively. Immunostaining with polyclonal anti-thrombomodulin antibody (Santa Cruz Biotechnology Inc) necessitated pretreatment with Protease 1 (Ventana, Strasbourg; France) for 20 minutes at 37°C. For the detection of von Willebrand factor (vWF; polyclonal anti-vWF antibody, Santa Cruz Biotechnology), sections were pretreated with 10 mmol/L sodium citrate (pH 6.0) for 3×10 minutes at 600 W in a microwave oven. Incubation with anti-ICAM antibody (monoclonal; Santa Cruz Biotechnology Inc) did not require antigen retrieval. Subsequently, the specimens were incubated with the primary antibodies for 1 hour at 37°C in a moist chamber, followed by incubation with biotinylated anti-mouse IgG/anti-rabbit IgG (1:200; Vector Laboratories; distributed by Camon, Wiesbaden, Germany) and ABC alkaline phosphatase reagent, each for 30 minutes at room temperature. Between steps, the sections were washed in Tris-buffered saline. FastRed (Fast Red Substrate Pack, Red 125, Zytomed GmbH, Berlin, Germany) served as chromogen. All specimens were counterstained with hematoxylin. Primary antibodies were omitted for negative controls. Owing to limitations in absolute amounts of tissue per patient, staining was performed in 249 samples for adhesion molecules, in 239 for PAI-1, in 226 for thrombomodulin, and in 240 for vWF.
The observed expression of VCAM-1, ICAM-1, PAI-1, thrombomodulin, and vWF in atrial tissue was assessed by use of 2 categories. Category A documented the number (quantity) of immunoreactive cells as 0 (no immunoreactive cells), 1 (<10%), 2 (10% to 50%), 3 (50% to 80%), or 4 (>80%). A positive case was defined as having a category A value of 1. Category B documented the intensity of the immunostaining as 0 (no immunostaining), 1 (weak), 2 (moderate), and 3 (strong). Lack of immunostaining of any cell was categorized as negative. For statistical analyses, each category was assessed separately (see below).
RNA Isolation and Reverse Transcription
Total RNA was prepared from ≈250 mg of atrial tissue or 100 mg of atrial tissue slices, respectively, by applying the method of Chomczynski and Sacchi, as described recently.16,17 One microgram of total RNA was transcribed into cDNA with AMV reverse transcriptase (Promega, Mannheim, Germany), as described previously.16
Quantitative Polymerase Chain Reaction
The iCycler (BioRad, Munich, Germany) was used for quantitative polymerase chain reaction. All samples were analyzed in triplicate. A 25-μL reaction mixture consisted of 1× SensiMix (Quantace, London, United Kingdom), 0.5 μL of SYBR-Green I (Quantace), 1 μL of cDNA, and 0.3 μmol/L of the specific primers for VCAM-1 (upstream: GTTTGGAAGTAACCTTTACTC; downstream: CCATCCTCATAGCAATTAAGG), and GAPDH (upstream: TCCAATCAAGTGGGGCGATGCT; downstream: ACCACCTGG-TGCTCAGTGTGACCC).
Initial denaturation at 95°C for 10 minutes was followed by 40 cycles with denaturation at 95°C for 15 seconds, annealing at 58°C for 30 seconds, and elongation at 72°C for 30 seconds. Quantities of GAPDH mRNA were used to normalize cDNA contents. The fluorescence intensity of the double-strand specific SYBR-Green I, which reflects the amount of the polymerase chain reaction product, was read in real time at the end of each elongation step. Then, amounts of specific initial template mRNA were calculated by determining the time point at which the linear increase of the sample polymerase chain reaction product began relative to the corresponding points of a standard curve; these are given as arbitrary units.
Frozen tissue samples from 42 human right atrial appendages (Table 3), as well as right and left atrial appendages from all animals, were pulverized in liquid nitrogen and subsequently homogenized in lysis buffer (50 mmol/L Tris/HCl, pH 7.5, 5 mmol/L EDTA, 100 mmol/L NaCl, 0.5% Triton X-100, 10% glycerol, 10 mmol/L K2HPO4, 0.5% NP-40), which contained a protease inhibitor cocktail (Boehringer Mannheim, Mannheim, Germany), 1 mmol/L sodium vanadate, 0.5% deoxycholate, 0.1 mmol/L PMSF, 20 mmol/L NaF, and 20 mmol/L glycerol 2-phosphate (all from Sigma, Heidelberg, Germany). Tissue homogenates were centrifuged at 15 000 rpm for 15 minutes, and the resulting supernatant was stored at −20°C until further use.
Extracted proteins (30 μg per lane) were separated by 10% PAGE, followed by transfer to nitrocellulose membranes (Schleicher and Schuell BioScience, Dassel, Germany). Membranes were blocked with Roti-Block (Roth, Karlsruhe, Germany) and then incubated with primary antibodies against VCAM-1 (CBL 507 mouse monoclonal IgG1, diluted in PBS 1:200; Chemicon Europe, Hofheim, Germany). Anti-mouse horseradish peroxidase–conjugated antibodies (diluted 1:2000 in PBS; Cell Signaling, Frankfurt/Main, Germany) were applied after the blots were washed 3 times in PBS. For chemiluminescence detection, the SuperSignal West Dura Extended Duration substrate (Pierce, Rockford, Ill) was used. To compare the different groups, densitometric quantification was performed only on equally processed blots and exposed on the same radiographic film.
Culturing of Tissue Slices and Ex Vivo Pacing
For ex vivo culture experiments, atrial tissue samples were taken from 20 patients (mean age 68±9 years; 10 men, 10 women) with sinus rhythm (SR) and without history of AF. Atrial tissue was obtained during open heart bypass surgery and processed directly for culturing. The tissue was sliced (350 μm) immediately, and 12 slices were placed on top of a 0.02-μm Anopore membrane tissue culture insert (25 mm, Nunc, Wiesbaden, Germany), which was placed in a Petri dish (8-cm diameter; Nunc) filled with culture medium as described previously.16 To make an intraindividual comparison of the experiments, tissue from each individual patient was cut in slices to serve both as a control specimen and for functional experiments, which were performed simultaneously.
To electrically stimulate adult human atrial tissue, a pair of custom-built carbon electrodes (12.5×6×32 mm) were submersed at opposite ends of the Petri dish. The distance between the electrodes was ≈8 cm. Copper wires, which were electrically isolated with silicon rubber, were inserted into holes drilled into the carbon electrodes and connected to a stimulation unit (GRASS stimulator, Astro-Med, West Warwick, RI). In accordance with previous in vitro pacing experiments, a biphasic square-wave impulse (150 V) of 5-ms duration was used for electrical field stimulation.16,18 A biphasic impulse was used to minimize electrolysis at the electrodes. The minimum distance between the tissue samples and the electrodes was ≈1.5 cm. Pacing of the tissue slices was performed within the cell culture incubator up to 24 hours at 37°C at a pacing rate of 0.6 and 2 Hz. Previously, it had been shown that slow in vitro pacing (0.6 Hz) does not alter tissue response compared with unpaced tissue samples.16,18 Thus, unpaced tissue slices also served as controls in the present set of experiments. To determine the effect of olmesartan medoxomil (provided by Daiichi Sankyo Pharma, Tokyo, Japan), including any potential concentration-dependent effects, 1 and 10 μg/mL of olmesartan medoxomil were added to the media throughout the pacing period. Thereafter, VCAM-1 expression was determined at the mRNA level. For Western blotting analysis, a dose of 10 μg/mL olmesartan medoxomil was chosen.
Angiotensin II Plasma Concentrations
Trunk blood was collected into chilled Vacutainer collection tubes (9NC, Becton-Dickinson, Heidelberg, Germany). Protease inhibitor cocktail (Roche Diagnostics, Mannheim, Germany) was added immediately, and the samples were centrifuged at 3000 rpm at 4°C for 15 minutes. Plasma was drawn without disturbing the packed cells. Aliquots were stored in liquid nitrogen until analysis.
Plasma concentrations of angiotensin II were measured by ELISA (BioCat, Heidelberg, Germany). Absorbance was recorded at 450 nm, and the concentration was calculated from the standard curve generated for angiotensin II.
Rapid Atrial Pacing Model
A total of 14 pigs (weight 24±3 kg) were studied to confirm the ex vivo finding that angiotensin II receptor blockade reduces tachycardia-induced atrial VCAM-1 expression. Animal experiments were approved by the Institutional Animal Care and Use Committee of the University of Magdeburg.
Pigs were premedicated and intubated. Anesthesia was maintained with 2% isoflurane and supplemental oxygen given via the endotracheal tube. Oxygen saturation was monitored continuously and kept above 90% throughout the procedure. An arterial sheath (6F) was introduced surgically into the left carotid artery. Sheaths (7F and 8F) were also introduced into the left jugular vein and the left femoral vein. After complete instrumentation, 1000 U of heparin was given intravenously and repeated every hour. A Swan-Ganz catheter (6F) was used to determine right atrial and pulmonary artery wedge pressures. Two quadripolar diagnostic electrode catheters were used to determine the effective refractory periods (ERPs) in the high right atrium. ERPs were measured after a train of 8 stimuli (S1) at a basic cycle length of 400 ms (0.5 ms, twice diastolic threshold). Inducibility of AF was tested twice on each occasion by rapid atrial burst stimulation for 10 seconds. Rapid atrial pacing was performed in 5 animals (pacing group) at a rate of 600 bpm (twice diastolic threshold, 2-ms pulse duration) for 7 hours. In 5 additional animals, rapid atrial pacing was performed in the presence of irbesartan infusion (1 mg/kg bolus followed by 0.3 mg · kg−1 · h−1 IV; irbesartan group), and 4 pigs were instrumented without further interventions (sham). After the 7-hour pacing period, the chest and pericardial sac were opened, and the heart was exposed. The right and left atrial appendages were cross-clamped and excised. Tissue samples were immediately frozen in liquid nitrogen and used for Western blotting as described above. Thereafter, animals were euthanized by clamping the LAD, which resulted in ventricular fibrillation, and by lethal injection of pentobarbital.
The description of the clinical characteristics has the form of mean±SD for continuous variables or n (%) for variables with discrete levels. Univariate comparisons of the 2 groups with AF and without AF (SR) were done with the t test (Welch test, Satterthwaite’s approximation to compute the degrees of freedom) for metric parameters. The χ2 test was used to assess differences between patient groups in categorical clinical parameters and immunohistochemistry results. For contingency tables with small sample sizes, in which the minimum expected value of any cell was <5, Fisher’s exact test was performed. The subgroup analysis to compare paroxysmal AF versus persistent AF was also done with the above-mentioned tests. The number of parameters analyzed is demonstrated in Tables 1 through 3⇑⇑. Five criteria were used for the evaluation of VCAM expression, as given in Tables 4 and 5⇓. To focus interest on the influence of AF on electrophysiological and structural changes, we repeated these analyses with pairs of patients. The pairs were matched with regard to age (±5 years), gender, concomitant cardiovascular diseases, left ventricular ejection fraction, and type of cardiac surgery. In addition, an ANOVA was performed with VCAM intensity and VCAM quality, respectively (dependent variable) and 8 independent variables (main factors: patient group [AF/SR], gender, Canadian Cardiovascular Society class, type of cardiac surgery, and diabetes status; covariables: age, C-reactive protein, and left ventricular ejection fraction). Furthermore, cross-effects between variables are expected, and therefore, in a second step, we used a factor model with 7 main effects and 6 two-factorial cross-effects between patient groups and the remaining independent variables. Baseline clinical characteristics and VCAM-1 levels were comparable in patients with paroxysmal and persistent AF. Therefore, patients with paroxysmal and persistent AF were summarized in a common group for multivariable analyses.
The in vitro and in vivo experiments were also analyzed with ANOVA. Values of these experiments are expressed as mean±SEM. All statistical decisions were 2-tailed with a critical probability of α=5% without α-adjustment, and probability values supported the interpretation. The results should be interpreted in an exploratory manner. Statistical analyses were performed with SAS software (SAS Institute, version 9.1; Cary, NC).
The authors had full access to the data and take full responsibility for its integrity. All authors have read and agree to the manuscript as written.
Atrial Expression of Prothrombotic Proteins in Human AF Tissue Microarray Analyses and Western Blotting
Tissue microarray analysis was used to screen a large cohort of patients for differences in expression of the prothrombogenic proteins VCAM-1, ICAM-1, thrombomodulin, PAI-1, and vWF (Table 4). Immunohistochemical analyses showed that all proteins analyzed were expressed by endothelial cells of right atrial appendages; however, only expression of VCAM-1 was increased in patients with AF compared with patients in SR (Figure 2). This difference remained after comparison of 49 matched patient pairs with AF and SR (P=0.06; Data Supplement Table II). Because cross-effects between variables were expected, a factor model with main effects and 2 factorial cross-effects between patient groups and the remaining independent variables was used for multivariable analyses. We found a significant association between the intensity of endocardial VCAM expression with AF (P=0.030) and left ventricular ejection fraction (P=0.044). Other clinical variables (age, gender, New York Heart Association class, history of myocardial infarction, diabetes mellitus, type of surgery, and medical therapy) were not related to endocardial VCAM-1 expression; however, analysis of cross-effects revealed that the combination of diabetes mellitus and AF had a significant impact on both VCAM-1 intensity (P=0.02) and quantity (P=0.016). Combinations of AF and ejection fraction (P=0.052), as well as AF and gender (P=0.005), affected VCAM-1 quantity.
The overall number of VCAM-1–immunoreactive endothelial cells was not different in 23 patients with paroxysmal AF and 38 patients with persistent AF (paroxysmal AF 1.4±1.2 versus persistent AF 1.5±1.3, P=0.814; Table 5). The distribution of VCAM-1 expression in paroxysmal AF revealed VCAM-1 staining in subendocardial tissue (27.6%) and endocardial cells (71.4%), whereas VCAM-1 was found exclusively in the endocardium in patients with persistent AF. VCAM-1 expression was not different in patients with (192/320) and without (128/320) angiotensin-converting enzyme (ACE) inhibitor/ARB therapy; however, medical therapy with ACE inhibitors and ARBs was not standardized with regard to pharmacological substance, dosage, and treatment duration.
Atrial VCAM-1 expression was also determined with unfixed atrial tissue homogenates. Western blotting confirmed an approximately 2-fold increased expression of VCAM-1 during both persistent and paroxysmal AF compared with SR (persistent AF 207±42% versus SR 100±16%, P=0.028; paroxysmal AF 193±42%, P=0.024 versus SR; Figures 3A and 3B), which was also associated with increased VCAM-1 mRNA levels in patients with persistent AF (AF 207±44% versus SR 100±11%, P=0.024; data not shown).
Ex Vivo Pacing of Human Atrial Tissue Slices
Angiotensin II is known to induce expression of adhesion molecule in the vascular endothelium. Therefore, our primary hypothesis was that ARBs reduce atrial expression of VCAM-1. To test this hypothesis, we used a previously described organotypic human atrial tissue culture model.16 Rapid pacing of human atrial tissue slices induced a significant upregulation of VCAM-1 expression at the mRNA and protein levels (Figure 4). Pretreatment with olmesartan medoxomil attenuated atrial VCAM-1 mRNA expression in a concentration-dependent manner (Figure 4B). Furthermore, the increase in VCAM-1 protein levels was abolished by pretreatment with 10 μg/mL olmesartan medoxomil, which highlights the involvement of angiotensin II type 1 receptor activation in the regulation of atrial VCAM-1 expression during ex vivo pacing.
Effect of Angiotensin II Receptor Blockade on Adhesion Molecule Expression In Vivo
An in vivo model of rapid atrial pacing was used to confirm and validate the potential beneficial effects of angiotensin II receptor blockade on atrial adhesion molecule expression as suggested by the ex vivo results. Hemodynamic data of the 3 different animal groups are provided in Table 6. Irbesartan infusion had no direct effect on atrial ERP at baseline (200±8 versus 203±8 ms, P=0.79). Atrial ERP decreased significantly during 7 hours of rapid atrial pacing (200±10 versus 132±8 ms, P=0.013), and this was attenuated by irbesartan (baseline 203±8 versus 163±10 ms, P=0.082). After 7-hour pacing, atrial ERP was significantly longer in the irbesartan group than in the pacing group (163±10 versus 132±8 ms, P=0.047).
Western blotting showed increased atrial VCAM-1 protein expression after 7 hours of rapid atrial pacing (Figure 5). Alterations in VCAM-1 expression were more pronounced in the left atria (5.3-fold compared with sham) than in the right atria (3.5-fold compared with sham). Importantly, irbesartan therapy abolished the pacing-induced increase in atrial VCAM-1 expression (Figure 5).
Effect of Acute Rapid Pacing In Vivo on Plasma Angiotensin II Levels
Seven-hour pacing of pigs led to an increase in plasma levels of angiotensin II compared with sham-operated animals (0.27±0.01 versus 0.46±0.07 pmol/L, sham group versus pacing group, *P=0.048). Pacing in the presence of irbesartan led to an additional increase in angiotensin II concentrations in plasma (irbesartan group 0.61±0.11 pmol/L, P=0.021 versus sham group; Figure 6).
To the best of our knowledge, the present study is the first to demonstrate increased expression of atrial VCAM-1 during AF in a large cohort of patients. We also show that ARBs reduce tachycardia-induced atrial VCAM-1 expression ex vivo and in vivo, providing evidence that angiotensin II has a significant pathophysiological role in the prothrombogenic process of endocardial remodeling.
AF and Adhesion Molecules
Atrial systolic dysfunction increases the stasis of blood, particularly in the atrial appendages. Therefore, atrial contractile dysfunction is important to explain the high risk of thromboembolic stroke in patients with AF.6,7 Furthermore, the plasmatic clotting system is activated in patients with AF.8–10 Despite this, it is well known that local thrombus formation also depends on endothelial changes. Endothelial damage or prothrombotic endothelial alterations (“endocardial remodeling”) must be present in fibrillating atria to explain the development of thrombi at the atrial endocardium.19–21 However, the process of endocardial remodeling in AF is still ill-defined. Cai et al19 have shown in an in vivo animal model that the left atrial endocardium produces large amounts of nitric oxide (NO), which is impaired during AF by reduced expression of endothelial NO synthase. Loss of NO production in vascular endothelium affects the expression of adhesion molecules, thereby allowing the development of vascular thrombi. In the atrial myocardium, AF is associated with increased activity of NADPH oxidase, which causes the generation of reactive oxygen species and oxidative stress.21,22 Oxidative stress is another potent stimulus to increase the expression of adhesion molecules.23–25 Angiotensin II increases via angiotensin II type 1 receptor NADPH oxidase activity, thereby increasing the expression of adhesion molecules. This, in turn, promotes the adhesion of lymphocytes and monocytes to the endothelial surface, which causes endothelial dysfunction and further contributes to the initiation and progression of thrombus formation. The pathophysiological significance of adhesion molecules in AF is supported by the study by Kamiyama,26 who showed that rapid atrial pacing (8 hours) increases adhesion molecule expression in the left atrial endocardium in rabbits, in association with leukocyte adherence to the atrial endocardium. Thus, endothelial expression of adhesion molecules might be a very early pathophysiological step in thrombogenesis by promoting adherence of leukocytes, monocytes, and platelets to the endocardial surface.23–26
Recent analyses of the Perindopril Protection Against Recurrent Stroke Study (PROGRESS), which included >6000 patients, showed that in addition to traditional risk factors, systemic levels of VCAM-1 provide prognostic information about recurrent ischemic stroke.27 Interestingly, VCAM-1 levels were found to be elevated even in patients with silent cerebral infarction.28 However, the role of adhesion molecules has not yet been clarified in patients with AF. Preliminary data from our group showed that systemic plasma levels of ICAM-1 and VCAM-1 are increased in patients with AF, reaching the highest levels with overt atrial thrombi. In that study, increased systemic VCAM-1 levels were an independent predictor of atrial thrombi in multivariate analysis.13
In the present study, we found increased atrial expression of VCAM-1 in patients with paroxysmal and persistent AF. Importantly, we can demonstrate that the increased expression of atrial adhesion molecules occurs within hours during rapid atrial pacing. This may explain why there was no difference in atrial VCAM-1 expression in patients with persistent and paroxysmal AF. One important factor that may influence endocardial protein expression is left ventricular failure that causes cardiac pressure and volume overload.29 In the present study, we also demonstrated an association between left ventricular function (left ventricular ejection fraction) and atrial VCAM expression using multivariable analysis; however, hemodynamic alterations were absent in the in vitro system or were constant in the in vivo experiments. This demonstrates that AF itself can induce alterations in VCAM expression. Nevertheless, heart failure is an important risk factor for stroke in patients with AF. One explanation for this finding might be the impact of left ventricular dysfunction on atrial expression of VCAM-1. Accordingly, concomitant diseases (eg, heart failure and diabetes mellitus) and gender appear to influence atrial VCAM-1 levels during AF, which may also explain persistently elevated VCAM-1 levels in the systemic circulation after successful cardioversion of AF.13 Nevertheless, the contribution of diabetes mellitus and heart failure per se to atrial VCAM expression needs further evaluation.
Importantly, we found quantitative differences in VCAM-1 expression in the right and left atria as a response to AF. During rapid atrial pacing, adhesion molecule levels increased more significantly in the left atrial tissue than in the right atrium. This may help to explain why thrombus formation is typically observed in the left atrial appendage, whereas right atrial thrombi with consecutive pulmonary embolism is a rare event in patients with AF. For the present large series of tissue microarray analyses, only right atrial tissue samples were available; however, the results of our animal experiments suggest that analyses of left atrial tissue samples might have revealed even more pronounced differences in endocardial VCAM expression. Nevertheless, the impact of other prothrombogenic proteins in the left atrium may merit further investigation, because some experimental data suggest that AF may lead to downregulation of thrombomodulin or upregulation of PAI-119,30; however, an increased expression of these proteins could not be shown in the present analysis of human tissue samples.
ARBs and Adhesion Molecules
The impact of ARB therapy on systemic adhesion molecule levels has been shown in patients with atherosclerosis and heart failure23,24,29; however, to the best of our knowledge, the present study is the first to demonstrate the direct impact of ARB treatment on endocardial expression of adhesion molecules. Both the presented ex vivo experiments using organotypic human atrial tissue cultures and the in vivo experiments showed that ARB treatment reduces adhesion molecule expression in atrial tissue during rapid pacing. Thus, rapid atrial stimulation influences prothrombotic endocardial remodeling via the angiotensin II type 1 receptor. In support of this view, we demonstrated elevated plasma concentrations of angiotensin II in response to rapid pacing in vivo. These findings are in line with recent studies that have already shown a significant impact of the atrial angiotensin II system on the occurrence of AF.3,31,32 Previous experiments demonstrated that ACE inhibitors and ARBs influence short-term electrical atrial alterations (shortening of the atrial ERP), which was confirmed in the present study.33 This effect, however, appears to diminish in the long term.34 In addition to their variable electrophysiological effects, ACE inhibitors and ARBs have been clearly demonstrated to reduce the amount of proarrhythmic structural atrial alterations, such as atrial fibrosis. The experimentally observed antiarrhythmic effects of ACE inhibitors and ARBs are supported by various retrospective analyses of clinical trials (TRACE [TRAndolapril Cardiac Evaluation], ValHeFT [Valsartan in Heart Failure Trial], SOLVD [Studies Of Left Ventricular Dysfunction], CHARM [Candesartan in Heart failure Assessment of Reduction in Mortality and morbidity], and LIFE [Losartan Intervention For End point reduction in hypertension]).35,36
In addition to antiarrhythmic effects, the present study also leads to the assumption that ARBs may effect atrial thrombus formation. Nevertheless, it has yet to be proven whether the effect of ARBs observed in the short-term in vitro and in vivo models in the present study corresponds to a clinically relevant reduction of thromboembolic events in patients with AF. Interestingly, the first clinical results from the LIFE study already point in that direction. Wachtell et al36 showed that losartan treatment substantially reduces the incidence of stroke in patients with new-onset AF (hazard ratio 0.49, 95% CI 0.29 to 0.86, P=0.01). Prospective clinical trials (ANTIPAF, ACTIVE-I) are being carried out to determine the impact of ARB therapy on the occurrence of thromboembolic stroke during AF. Furthermore, experiments in chronic AF models are warranted to elucidate the impact of ARB in the long term on prothrombotic atrial alterations at the molecular level.
The present study provides evidence that AF and pacing-induced atrial tachycardia increase the expression of adhesion molecules in the atrial endocardium within hours. In addition, AF-induced expression of VCAM-1 is higher in the left atrium than in the right atrium. The prothrombogenic process of endocardial remodeling can be attenuated substantially by ARB treatment, which suggests that angiotensin II has a significant pathophysiological role in endocardial remodeling; however, experiments in chronic AF models are warranted to determine the long-term efficacy of ARBs in this process.
Sources of Funding
This work was supported by a grant from the Bundesministerium für Bildung und Forschung, Germany (Kompetenznetz Vorhofflimmern, grant 01GI0204). The tissue microarray instrument was supported by grants from the Rudolf Bartling Foundation. We thank Daiichi Sankyo for providing olmesartan medoxomil and Sanofi-Aventis for providing irbesartan.
Dr Goette received speaker honoraria from Daiichi Sankyo and Sanofi-Aventis. The other authors report no conflicts.
Goette A, Honeycutt C, Langberg JJ. Electrical remodeling in atrial fibrillation: time course and mechanisms. Circulation. 1996; 94: 2968–2974.
Ausma J, Wijffels M, Thone F, Wouters L, Allessie M, Borgers M. Structural changes of atrial myocardium due to sustained atrial fibrillation in the goat. Circulation. 1997; 96: 3157–3163.
Chung MK, Martin DO, Sprecher D, Wazni O, Kanderian A, Carnes CA, Bauer JA, Tchou PJ, Niebauer MJ, Natale A, Van Wagoner DR. C-reactive protein elevation in patients with atrial arrhythmias: inflammatory mechanisms and persistence of atrial fibrillation. Circulation. 2001; 104: 2886–2891.
Hammwöhner M, Ittenson A, Dierkes J, Bukowska A, Lendeckel U, Klein HU, Goette A. Platelet expression of CD40/CD40 ligand and its relation to inflammatory markers and adhesion molecules in patients with atrial fibrillation. Exp Biol Med. 2007; 232: 581–589.
Gawaz M, Neumann FJ, Dickfeld T, Koch W, Laugwitz KL, Adelsberger H, Langenbrink K, Page S, Neumeier D, Schomig A, Brand K. Activated platelets induce monocyte chemotactic protein-1 secretion and surface expression of intercellular adhesion molecule-1 on endothelial cells. Circulation. 1998; 98: 1164–1171.
Schild L, Bukowska A, Gardemann A, Polczyk P, Keilhoff G, Tager M, Dudley SC, Klein HU, Goette A, Lendeckel U. Rapid pacing of embryoid bodies impairs mitochondrial ATP synthesis by a calcium-dependent mechanism: a model of in vitro differentiated cardiomyocytes to study molecular effects of tachycardia. Biochim Biophys Acta. 2006; 1762: 608–615.
Cai H, Li Z, Goette A, Mera F, Honeycutt C, Feterik K, Wilcox JN, Dudley SC Jr, Harrison DG, Langberg JJ. Downregulation of endocardial nitric oxide synthase expression and nitric oxide production in atrial fibrillation: potential mechanisms for atrial thrombosis and stroke. Circulation. 2002; 106: 2854–2858.
Minamino T, Kitakaze M, Sato H, Asanuma H, Funaya H, Koretsune Y, Hori M. Plasma levels of nitrite/nitrate and platelet cGMP levels are decreased in patients with atrial fibrillation. Arterioscler Thromb Vasc Biol. 1997; 17: 3191–3195.
Dudley SC Jr, Hoch NE, McCann LA, Honeycutt C, Diamandopoulos L, Fukai T, Harrison DG, Dikalov SI, Langberg J. Atrial fibrillation increases production of superoxide by the left atrium and left atrial appendage: role of the NADPH and xanthine oxidases. Circulation. 2005; 112: 1266–1273.
Mihm MJ, Yu F, Carnes CA, Reiser PJ, McCarthy PM, Van Wagoner DR, Bauer JA. Impaired myofibrillar energetics and oxidative injury during human atrial fibrillation. Circulation. 2001; 104: 174–180.
Wang HD, Xu S, Johns DG, Du Y, Quinn MT, Cayatte AJ, Cohen RA. Role of NADPH oxidase in the vascular hypertrophic and oxidative stress response to angiotensin II in mice. Circ Res. 2001; 88: 947–953.
Campbell DJ, Woodward M, Chalmers JP, Colman SA, Jenkins AJ, Kemp BE, Neal BC, Patel A, MacMahon SW. Soluble vascular cell adhesion molecule 1 and N-terminal pro-B-type natriuretic peptide in predicting ischemic stroke in patients with cerebrovascular disease. Arch Neurol. 2006; 63: 60–65.
Kawamura T, Umemura T, Kanai A, Nagashima M, Nakamura N, Uno T, Nakayama M, Sano T, Hamada Y, Nakamura J, Hotta N. Soluble adhesion molecules and C-reactive protein in the progression of silent cerebral infarction in patients with type 2 diabetes mellitus. Metabolism. 2006; 55: 461–466.
Tsutamoto T, Wada A, Maeda K, Mabuchi N, Hayashi M, Tsutsui T, Ohnishi M, Sawaki M, Fujii M, Matsumoto T, Kinoshita M. Angiotensin II type 1 receptor antagonist decreases plasma levels of tumor necrosis factor alpha, interleukin-6 and soluble adhesion molecules in patients with chronic heart failure. J Am Coll Cardiol. 2000; 35: 714–721.
Yamashita T, Sekiguchi A, Iwasaki YK, Sagara K, Hatano S, Iinuma H, Aizawa T, Fu LT. Thrombomodulin and tissue factor pathway inhibitor in endocardium of rapidly paced rat atria. Circulation. 2003; 108: 2450–2452.
Li D, Shinagawa K, Pang L, Leung TK, Cardin S, Wang Z, Nattel S. Effects of angiotensin-converting enzyme inhibition on the development of the atrial fibrillation substrate in dogs with ventricular tachypacing-induced congestive heart failure. Circulation. 2001; 104: 2608–2614.
Nakashima H, Kumagai K, Urata H, Gondo N, Ideishi M, Arakawa K. Angiotensin II antagonist prevents electrical remodeling in atrial fibrillation. Circulation. 2000; 101: 2612–2617.
Shinagawa K, Mitamura H, Ogawa S, Nattel S. Effects of inhibiting Na(+)/H(+)-exchange or angiotensin converting enzyme on atrial tachycardia-induced remodeling. Cardiovasc Res. 2002; 54: 438–446.
Wachtell K, Lehto M, Gerdts E, Olsen MH, Hornestam B, Dahlof B, Ibsen H, Julius S, Kjeldsen SE, Lindholm LH, Nieminen MS, Devereux RB. Angiotensin II receptor blockade reduces new-onset atrial fibrillation and subsequent stroke compared to atenolol: the Losartan Intervention For End Point Reduction in Hypertension (LIFE) study. J Am Coll Cardiol. 2005; 45: 712–719.
Vascular thrombi develop in the presence of reduced blood velocities, when the activity of the clotting system is increased, and in the presence of endothelial alterations (Virchow’s triad). Virchow’s triad also applies to atrial thrombus formation during atrial fibrillation (AF). It is well known that AF causes loss of regular atrial contractions, and AF thereby reduces atrial blood flow velocities. Furthermore, AF is associated with activation of the plasmatic clotting system and of thrombocytes. However, the contribution of endothelial alterations to atrial thrombogenesis has not been fully explored. Increased endocardial expression of adhesion molecules (such as vascular cell adhesion molecule) could be an important link between initiation of inflammatory and prothrombogenic mechanisms responsible for thrombus development at the atrial endocardium (“endocardial remodeling”). The present study investigates the impact of AF on the expression of various prothrombotic atrial proteins in a large cohort of patients. The main finding of the study is an increased endocardial vascular cell adhesion molecule-1 expression in patients with AF. Interestingly, endocardial vascular cell adhesion molecule levels during AF are further increased by well-known risk factors for thromboembolic events such as diabetes mellitus and heart failure. We can show that angiotensin II receptor blockers reduce tachycardia-induced atrial vascular cell adhesion molecule-1 expression ex vivo and in vivo, providing evidence that angiotensin II has a significant pathophysiological role in the prothrombogenic process of endocardial remodeling. Thus, the present study provides an important piece of information to explain the impact of angiotensin II receptor blockers on inflammatory and prothrombotic alterations at the atrial endocardium during AF.
↵*The first 3 authors contributed equally to this work.
The online-only Data Supplement, which contains Tables I and II, is available with this article at http://circ.ahajournals.org/cgi/content/full/ CIRCULATIONAHA.107.730101/DC1.