Immunoglobulin G3 cardiac myosin autoantibodies correlate with left ventricular dysfunction in patients with dilated cardiomyopathy: Immunoglobulin G3 and clinical correlates

Immunoglobulin G3 cardiac myosin autoantibodies correlate with left ventricular dysfunction in patients with dilated cardiomyopathy: Immunoglobulin G3 and clinical correlates Rahat S. Warraich, PhD,a Michel Noutsias, MD,b Ilkay Kasac, MBBS,b Bettina Seeberg, MD,b Micheal J. Dunn, PhD,a Heinz-Peter Schultheiss, MD,b Magdi H. Yacoub, FRCS,a and Uwe Kuhl, MD, PhDb Middlesex, United Kingdom, and Berlin, Germany

Background Effector functions of an aberrant immune response have been implicated in the pathogenesis of idiopathic dilated cardiomyopathy (DCM). The immunologic determinants of myocardial dysfunction, however, remain poorly understood. This study sought to determine the relation of different immunologic responses to hemodynamic dysfunction in DCM. Methods Immunoglobulin (Ig) G class/subclass response ELISA (enzyme-linked immunosorbent assay) against cardiac myosin heavy chain, histologic characteristics (DALLAS criteria), immunohistochemistry, plasma interleukin-4 and plasma interferon gamma (IFN-␥) were determined in patients (n ⫽ 76) with clinically suspected myocarditis or DCM. Patients were prospectively evaluated, both clinically and hemodynamically, on admission (baseline) and at 6-month follow-up.

Results Indices of hemodynamic dysfunction (by cardiac catheterization and transthoracic echocardiography) correlated significantly with an Ig subclass response. IgG3 levels correlated with left ventricular ejection fraction (P ⫽ .02), pulmonary capillary wedge pressure (P ⬍ .0001), left ventricular end-systolic volume index (P ⫽ .002), left ventricular enddiastolic volume index (P ⫽ .033), left ventricular end-diastolic pressure (P ⫽ .04), right ventricular end-diastolic pressure (P ⫽ .039), and left ventricular end-systolic dimension and left ventricular end-diastolic dimension (P ⬍ .05). Patients positive for IgG3 (predominantly male, P ⫽ .01) had depressed left ventricular ejection fraction (ⱕ45%, relative risk 3.0, 95% CI 1.5-5.7, P ⫽ .005) at baseline and 6 months. Mitral-septal separation at follow-up improved in patients negative for IgG3 (P ⫽ .018), and the number of patients on conventional therapy in this group declined at 6-month follow-up (P ⬍ .05). Lymphocyte counts/high-power field; CD2, CD3, CD4, and CD8 (independent of IgG class/subclass response and left ventricular dysfunction) were significantly higher in patients positive for IFN-␥ (25%). A positive IFN-␥ response was higher in patients positive for IgG3. These patients, positive for IgG3 and IFN-␥ (10%), had significantly shorter duration of clinical symptoms: 0.17 years (0.12-2.36 y) versus 1.01 years (0.49-5.35 y, P ⫽ .04).

Conclusion IgG3 reactivity correlated with depressed myocardial dysfunction. This may render this subclass Ig a surrogate target for therapeutic intervention in DCM. With IFN-␥, IgG3 may reflect a more aggressive disease. (Am Heart J 2002;143:1076-84.)

Idiopathic dilated cardiomyopathy (DCM) presents as congestive heart failure and the current therapeutic

From the aDepartment of Cardiothoracic Surgery, National Heart and Lung Institute, Imperial College School of Medicine, Royal Brompton and Harefield Trust, Harefield Hospital, Middlesex, UK, and the bDepartment of Cardiology, University Hospital Benjamin Franklin, Free University of Berlin, Berlin, Germany. Presented in part at the 2000 Annual Meeting of the American Heart Association, New Orleans, La. Published in abstract form, Circulation 2000;102(2 Suppl):3772. Submitted December 28, 2001; accepted February 25, 2002. Reprint requests: Rahat Warraich, PhD, Department of Cardiothoracic Surgery, National Heart and Lung Institute, Imperial College School of Medicine, Royal Brompton and Harefield Trust, Harefield Hospital, UB9 6JH, Middlesex, United Kingdom. E-mail: [email protected] © 2002, Mosby, Inc. All rights reserved. 0002-8703/2002/$35.00 ⫹ 0 4/1/124406 doi:10.1067/mhj.2002.124406

modalities are largely directed toward ameliorating the secondary manifestations. Beneficial results from clinical trials of immunoadsorption,1,2 with improvement of left ventricular function in patients with myocarditis and DCM, support the role and contribution of humoral effector functions in disease. Conversely, the discouraging reports of immunosuppressive trials3,4 may in part reflect our inability to identify or characterize components of the immunologic repertoire that may respond better to therapy. A definitive measure of the immunologic correlates of progressive myocardial dysfunction may allow for better implementation of therapeutic measures in clinical practice. Functional blockade of Fc␥ receptors (target receptors for immunoglobulin [Ig]–Fc domains) with intrave-

American Heart Journal Volume 143, Number 6

Warraich et al 1077

nous Ig (IVIg) is an adapted therapeutic avenue by which immunologic disorders reportedly responding to IVIg have expanded. However, the benefits of IVIg in myocarditis and DCM remain unclear. Ig-Fc domains are the regulatory elements of the humoral effector system that determine the Ig subclass specificity and define the functional capacity (primarily characteristic of IgG3 ⬎ IgG15,6) with which these Igs contribute to immune-mediated tissue injury. Their capacity to elicit inflammatory processes has been described in diverse clinical entities such as primary biliary cirrhosis,7 rheumatoid arthritis,8-10 type 1 diabetes,11 and membranoproliferative glomerulonephritis.12 Given their differential capacity to trigger effector functions, a diseasespecific Ig-subclass profile may be of significant clinical interest. We have previously identified an Ig subclass response against cardiac myosin heavy chain (MHC) in patients with heart failure with DCM.13 IgG3 reactivity was raised in patients with DCM compared with patients with ischemic heart disease and healthy blood donors. These patients, positive for IgG3, were associated with a relatively adverse clinical course after cardiac transplantation.14 Interestingly, our unpublished data show that IgG3 reactivity represents the predominant humoral immune response in patients with DCM and is very much independent of the autoantigens recognized in this clinical entity. In an effort to provide a better understanding of the role of IgG3 in DCM, the relationship of this Ig to different phenotypes of immunocompetent infiltrates (T lymphocytes CD2, CD3, CD4, and CD8; memory T cells CD45R0; early activated macrophages 27E10) and indices of cardiac dysfunction was sought in patients with clinically suspected myocarditis (subjects with relatively milder symptoms of disease) and DCM.

catheterization, coronary angiography, and hemodynamic measurements were performed before taking endomyocardial biopsy specimens from the right ventricular septum. All patients consented and the studies were approved by the ethical committee in accordance to the Helsinki Declaration of 1975.

Methods

Histologic and immunohistologic study

Patient selection criteria Patients with evidence of idiopathic myocardial dysfunction (by echocardiography and cardiac catheterization) and/or with asynergic segments (segmental wall motion defects in at least 2 wall segments) who were angiographically free of coronary artery disease were recruited for this study. Exclusion criteria for all patients were coronary artery or valvular heart disease (by selective coronary angiography), systemic blood pressure ⬎160/100 mm Hg, systemic or endocrine disease, or a history of chronic excessive alcohol consumption. DCM was diagnosed as described by Manolio et al.15 Diagnosis of a clinically suspected myocarditis and the indication for endomyocardial biopsy was made on the basis of a clinical presentation or a history of palpitations, dyspnea at rest or on exertion, reduced exercise tolerance, cardiac arrhythmias, atypical chest pain, or bundle-branch block, as previously reported.16 Invasive procedures, left and right heart

Study patients Only patients with adequate serum available for humoral autoimmune responses were enrolled in this study (n ⫽ 76). All patients were prospectively evaluated clinically (at first hospital visit [baseline] and at 6-month follow-up), histologically, and immunohistochemically. Serum from an additional group of patients (age and sex matched) was evaluated for humoral autoimmune responses. Serum from patients with end-stage heart failure with ischemic heart disease (n ⫽ 35, taken before transplantation at Harefield Hospital) was also evaluated for humoral-autoimmune responses. This was approved by the local ethics committee.

Echocardiography, cardiac catheterization, and ventriculography Mitral E-point septal separation17 (minimal separation between the anterior mitral valve leaflet at its E point and the interventricular septum) and left atrial and ventricular dimensions were estimated by transthoracic echocardiography (Vingmed-system 5, version 1.3.1, with a 2.5-MHz transducer, Solingen, Germany). Wall motion abnormality was determined by radionuclide ventriculography. Left ventricular enddiastolic volume index (LVEDVI) and left ventricular ejection fraction (LVEF) (by angiography) were determined by the method of Dodge et al18 (Cardio 500, Kontron GmbH, Munich, Germany). Left ventricular end-diastolic pressure (LVEDP) was determined with a left-ventricular pigtail catheter. Right ventricular pressure in systole, diastole, end-diastole, pulmonary artery mean pressure (systole, diastole), and pulmonary capillary wedge pressure (PCWP) were determined by use of a right-heart catheter. Cardiac index and stroke volume index were measured with a flow-directed thermodilution catheter.

Endomyocardial biopsy specimens (EMB) (n ⱖ6 per patient) were obtained by the standard percutaneous transvenous right femoral approach with a Cordis bioptome modified by Olsen.19 Two EMBs, formalin fixed and paraffin embedded, were assessed for inflammation and histologic characteristics by light microscopy (DALLAS classification20). Confirmation of active myocarditis was made in EMBs (eosin and hematoxylin stained) demonstrating lymphocytic infiltration in the vicinity of myocytolysis. Biopsy specimens with cellular infiltration without myocytolysis of adjacent cardiomyocytes, with or without interstitial or reparative fibrosis, were reported as borderline myocarditis. Immunohistochemical analysis of inflammation and quantification of infiltrates and adhesion molecules CD2, CD3, CD45RO, CD4, CD8, 27E10, HLA-I, HLA-DR, and ICAM1 was performed as previously described.21 Stained cells were counted per high-power field (HPF) with 400-fold magnification (1 HPF ⫽ 0.28 mm2 [Leica-MDRD microscope, Ben-

American Heart Journal June 2002

1078 Warraich et al

sheim, Germany]). Results were expressed as the mean of at least 10 counted fields per HPF.

wall motion defects were more prominent in the latter group.

Antigen preparation and enzyme-linked immunosorbent assay

Histologic study and immunocytochemistry

Antigen extraction, purification, isoform-specificity for MHC (␣-[atrial-specific] and ␤-[ventricular-specific]), and the enzyme-linked immunosorbent assay for Ig class IgG and subclasses (IgG1, IgG2, and IgG3 [Binding Site, Birmingham, United Kingdom]) have been previously described.13 Ageand sex-matched healthy blood donors (n ⫽ 54) served as controls. Ig reactivity was the optical density measured at 490 nm (Titertek Plus MS2 plate-reader, Dynatech, United Kingdom). The threshold for the Ig frequency was determined as an Ig response plus mean ⫹ 2 SD above the level of the relevant corresponding control Ig.

Cytokines Levels of plasma interferon gamma (IFN-␥) and interleukin-4 (IL-4) were determined by ELISA (Pelikine-compact, CLB, Amsterdam). Venous plasma was obtained before the biopsy procedures. Supernatant from EDTA-collected blood (after immediate centrifugation [800g for 15 min]) was aliquoted into 200-␮L samples and stored at ⫺80°C until use. Samples (in duplicate) were assayed in a blinded fashion. Sensitivity and assay range for IL-4 and IFN-␥ was 0.2 pg/mL (0.6-450 pg/mL) and 1 pg/mL ([mean blank ⫹ 3 SD] 2-1800 pg/mL), respectively.

Statistical analysis Correlation of continuous (Pearson correlation coefficient) and noncontinuous variables was computed with JMP statistical software version 3.2.2 (SAS Institute, Cary, NC). Data assessment by linear and multiple regression and correlation analysis was exclusively after normality of quantitative data (as appropriate). Comparison of Ig reactivity after categorical distribution was computed as median and interquartile range (Mann-Whitney U test). Comparison of noncontinuous, categorical, ranked, or qualitative data was determined by nonparametric methods. Ig frequency and qualitative data were determined by ␹2 test. Quantitative and qualitative data were compared with the Wilcoxon/KruskalWallis test on rank sums. Turkey-Kramer post hoc analysis was used for multiple comparisons. A P value ⬍.05 was considered statistically significant.

Results Study patients A larger number of the study patients had clinically suspected myocarditis16 or wall motion abnormalities with relatively preserved left ventricular function compared with DCM. Patients were therefore categorized by an arbitrary measure of hemodynamic severity: LVEF ⱕ45% (n ⫽ 21) and ⬎45% (n ⫽ 55) (Table I). By use of the criteria of Manolio et al,15 patients with LVEF ⱕ45% were categorized as having DCM. Diffuse hypokinesia was reported in 95% of the patients with LVEF ⱕ45% and in 30% with LVEF ⬎45%. Segmental

Histologic study by light microscopy (DALLAS criteria) was independent of cardiac function and patient demographics. Myocarditis was reported in 3 (4%) patients (LVEF 38%, 55%, and 64%) and borderline myocarditis in 4 (5%) patients (LVEF 54% and ⬎65%). With the exception of CD8 (ⱖ2 counts/HPF, n ⫽ 7), markers of T cells and macrophages (diffuse or as focal aggregates [ⱖ3 stained cells]) were independent of left ventricular dysfunction. By multiple regression (excluding CD45), CD8 activity correlated with LVEF (r ⫽ 0.36, P ⫽ .001). However, LVEF in a multiple regression model with other indices of function was not significant as an independent correlate.

Igs IgG class/subclass reactivity and frequency (Figure 1) versus controls (Table II, A) show that the predominant Ig response is primarily IgG3. IgG3 response discriminated LVEF status in the study group (Table II, B). Relative risk of IgG3 in patients with LVEF ⱕ45% was 3.0 (95% CI 1.5-5.7, P ⫽ .005). With logistic regression, IgG3 reactivity was independent of the designated cutoff value for LVEF as ⱕ45% (P ⫽ .01). IgG3 regressed with LVEF of ⱕ50% (P ⫽ .02) and ⱕ40% (P ⫽ .04). Compared with study patients (n ⫽ 76), IgG3 autoantibody response was significantly lower in patients with heart failure with ischemic heart disease, although levels of total IgG were raised (Table II, C and D), as previously reported.13 With linear, multiple, and stepwise regression followed by correlation, IgG3 associated significantly with indices of myocardial dysfunction (Table III). The correlation of IgG3 with interventricular septum (r ⫽ 0.234, P ⫽ .066), although not significant, was relatively higher than the colinearity of interventricular septum with echocardiographic variables: LVEDD (r ⫽ 0.02, P ⫽ .87), LVESD (r ⫽ 0.03, P ⫽ .81), and LVEF (r ⫽ 0.05, P ⫽ .68). Hence, despite the significant colinearity of pulmonary artery–mean pressure with indices of left and right heart catheterization, correlation with IgG3 was not significant (Table III). These results provide a descriptive measure of the order and magnitude in which IgG3 in the study patients related to hemodynamic dysfunction. Distribution of these variables with IgG3 status categorized as positive or negative reactivity (Table IV) coincided with the correlation analysis. IgG3 reactivity was significantly associated with the male sex; males (0.42 [0.22-0.72]) versus females (0.24 [0.11-0.43], P ⫽ .01) (median and interquartile-range values). Duration of symptoms was relatively (although

American Heart Journal Volume 143, Number 6

Warraich et al 1079

Table I. Patient demographic, clinical, and immunocytochemical characteristics Study patients (n ⴝ 76)

Age at diagnosis (y) Sex (M/F ratio) Duration of symptoms (y) Left ventriculographic data† CI (L/min/m2 BSA) SVI (mL/m2 BSA) PCWP (mm Hg) LVEDP (mm Hg) LVEDVI (mL/m2 BSA) LVESVI (mL/m2 BSA) LVEF (%) RVEDP (mm Hg) PA-mean (mm Hg) Echo-data LVEDD (mm) LVESD (mm) LA (mm) IVS (mm) PWT (mm) Immunohistochemistry‡ CD2 T-lymphocyte counts/HPF CD3 T-lymphocyte counts/HPF CD4 T-lymphocyte counts/HPF CD8 T-lymphocyte counts/HPF CD45 T-lymphocyte counts/HPF 102-27 Macrophage count/HPF HLA1 vessels/HPF HLA1 interstitial/HPF HLA2 vessels/HPF HLA2 interstitial/HPF ICAM1 vessels/HPF ICAM1 interstitial/HPF Cytokines IL-4 pg/mL IFN-␥ (no. of patients ⱖ1 pg/mL)

LVEF >45% (n ⴝ 55)

LVEF <45% (n ⴝ 21)

P-value*

47.9 (38.5-58.5) 55% (30/25) 0.85 (0.29-2.89)

55.5 (44.8-60.2) 76% (16/5) 1.91 (0.26-5.94)

.07 .14 .33

3.64 (3.17-4) 48.8 (41-58.8) 6.0 (4.3-10) 6.5 (5-9.3) 96.6 (90-110) 39.6 (33-48.5) 60.5 (53.8-66.3) 4.0 (2-5) 13.0 (10-15.75)

3.11 (2.2-3.5) 40.0 (26.8-54.3) 12.0 (8.8-19.7) 12.5 (7.3-20) 174.5 (122.7-235) 96.1 (77.8-145) 33.0 (28-37) 6.0 (4-8) 18.0 (15-27)

.015 .019 .0002 .0002 ⬍.0001 ⬍.0001 ⬍.0001 .0027 .003

54 (52-60) 38 (32-41) 37 (33.5-40.5) 10 (8-11) 9 (8-10)

69 (62.8-72.8) 55.5 (51-60) 45 (41.5-48) 9 (7-11) 9 (8-10)

⬍.0001 ⬍.0001 .0002 .467 .874

1.6 (1.1-2.3) 1.55 (1.07-2.5) 1.05 (0.52-1.5) 0.92 (0.92-1.3) 0.75 (0.3-1.22) 1.30 (0.7-2.0) 2.00 (1.0-2.0) 1.50 (1.0-1.5) 1.50 (1.0-2.0) 1.00 (1.0-1.5) 1.00 (1.0-1.5) 1.50 (1.0-2.0)

2.0 (1.0-4.37) 2.0 (1.3-4.20) 1.15 (0.8-2.50) 1.35 (0.8-1.90) 0.95 (0.5-1.67) 0.90 (0.7-1.65) 2.00 (1.5-2.0) 1.50 (1.0-1.9) 1.50 (1.5-2.0) 1.50 (1.0-1.5) 1.00 (1.0-1.5) 1.50 (1.13-2.0)

.15 .075 .046 .032 .13 .24 .16 .61 .2 .28 .93 .68

2.91 (0.70-5.34) 18% (10/55)

1.81 (0.46-5.18) 43% (9/21)

.61 .054

BSA, Adjusted to body surface area; LVEDD and LVESD, left ventricular end-diastolic and -systolic dimensions; LA, left atrial dimension; IVS, interventricular septum; PWT, posterior wall thickness. *Mann-Whitney U test. †P ⱕ .0035 (significance following Bonferroni correction for multiple comparison). ‡P ⱕ .004 (significance following Bonferroni correction for multiple comparison).

not significantly) shorter in patients positive for IgG3. Ig-class/subclass responses were independent of myocardial infiltrates (data not shown).

Igs and cytokines Plasma IFN-␥ or IL-4 was independent of demographic and hemodynamic variables. Patients positive for IFN-␥ (25% [19/76]; detectable levels ⱖ1 pg/mL) expressed significantly higher CD2, CD3, CD4, and CD8 lymphocyte counts per HPF (P ⬍ .05). IFN-␥ positive response, although independent of Ig class/subclass response, was relatively higher in patients positive for IgG3 (39%) compared with IgG1 (0%), IgG2 (12%), and IgG (21%). A finding of note-

worthy interest was the relation between IgG3 and IFN-␥. Patients with positive IgG3 and IFN-␥ response (10%) had significantly shorter duration of clinical symptoms: 0.17 years (0.120-2.36 y) versus 1.01 years (0.49-5.35 y, P ⫽ .04). These patients expressed increased lymphocyte counts per HPF (data not shown).

Patient 6-month follow-up IgG3 reactivity correlated with LVEF at 6-month follow-up (r ⫽ 0.28, P ⫽ .031). LVEF was significantly lower in patients with positive IgG3 reactivity versus patients with negative reactivity (42.5% [25%-61%] vs 55% [45.5%-62.5%], respectively, P ⫽ .03).

American Heart Journal June 2002

1080 Warraich et al

Figure 1

Ig subclass distribution in study patients (patients with DCM and clinically suspected myocarditis, n ⫽ 76) and healthy blood donors (controls, n ⫽ 54).

IgG3 correlated with baseline mitral E-point septal separation (EPSS)15,17 (r ⫽ 0.39, P ⫽ .016). At 6-month follow-up, EPSS significantly improved in patients with negative IgG3 reactivity (from 12 mm [7-17 mm] to 8 mm [6-11 mm], ⫺2, 95% CI ⫺4.0 to ⫺0.50, P ⫽ .018, Wilcoxon). Follow-up EPSS in the IgG3-positive group did not differ (P ⫽ .27). The number of patients on therapy in the IgG3-negative group decreased at follow-up. IgG3 positive reactivity may therefore identify patients with depressed left ventricular dysfunction. The number of patients on therapy (except diuretics) in the IgG3-positive group increased at 6-month follow-up and significantly differed compared with the IgG3-negative patients at follow-up (angiotensinconverting enzyme inhibitors, 91.6% vs 65%, P ⫽ .07; digitalis, 58.3% vs 27%, P ⫽ .047; anticoagulants, 75% vs 40.5%, P ⫽ .037; and diuretics, 58.3% vs 40.5%, P ⫽ not significant; respectively). At 6-month followup, IgG3 reactivity was significantly higher in patients on therapy than in patients not on therapy (angiotensinconverting enzyme inhibitors, 0.423 [0.23-0.72] vs 0.182 [0.062-0.34], P ⫽ .0087; digitalis, 0.435 [0.230.74] vs 0.25 [0.08-0.48], P ⫽ .047; and anticoagulants, 0.450 [0.276-0.999] vs 0.29 [0.11-0.53], P ⫽ .044).

Discussion This is the first study to demonstrate a correlation between a component of humoral autoimmunity and indices of systolic and diastolic dysfunction (deter-

mined by echocardiography and cardiac catheterization) in patients with DCM. IgG3 reactivity, after multiple comparison, correlated with increased filling pressures, ventricular volumes, and dimensions. Of the hemodynamic indices, PCWP was the most prominent independent correlate. CD4 and CD8 counts per HPF, although relatively higher in patients with LVEF ⱕ45%, were not significant after multivariate comparison with other markers of cellular infiltration, clinical or Igs, indicating the significant relevance of IgG3 autoantibodies in the clinical context of the disease process. The pathophysiologic characteristics of IgG3 in DCM are currently unknown. However, interaction of IgG3 with Ig receptors (Fc␥-Rs) expressed on immunocompetent cells, particularly Fc␥RI-A, a high-affinity receptor with a particularly high affinity and specificity for IgG3, may represent a potential mechanism for the observed immune-mediated correlates seen with hemodynamic indices. Hence, affinity of the Fc␥ domain of IgG2a (a high-affinity murine-Ig [analogue to human IgG3]) for the tumor necrosis factor–R1 has been described as one mechanism for disease relapse in an experimental model of collagen-induced arthritis (which generates IgG2a) after adenovirus-mediated blockade of tumor necrosis factor activity.22 Signal transduction of Fc␥RI orchestrates a host of functional events (complement-activation, antibody-dependent cell-mediated cytolysis, and release of perforin and cytolysins associated with the membrane attack complex). The magnitude of the effect is determined by the triggering (IgG3) stimuli while receptor proliferation is largely in response to IFN-␥. Therefore, a substantially high increase in the high-affinity Igs (IgG3), and in particular their correlation with myocardial dysfunction, may identify the likely candidates for immunomodulatory therapy. Furthermore, transmembrane motifs of Fc␥Rs also interact with the CD3/MHC complex (which establishes a cross-link between humoral and cellular responses). However, the observed discrepancy between the humoral and cellular activity in the study subjects may reflect differences in the time of onset, the amplitude, and possibly the duration of the different immunologic responses in the course of disease.

Igs and cytokines The relationship between IgG3 with IFN-␥ and shorter duration of symptoms, although intriguing, is currently a preliminary finding and warrants further work. Ig switching occurs either as a result of cognate interaction of T and B cells, preceding transcription of Ig-CH downstream of IgM (precommitted cells, determined by the antigenic stimuli), or in response to regulatory modalities of cytokines. Although IFN-␥ predominantly favors a Th1-type isotype switch IgG2a23

American Heart Journal Volume 143, Number 6

Warraich et al 1081

Table II A. Frequency and reactivity of Ig class and subclass response in study patients (n ⫽ 76) and healthy blood donors (n ⫽ 54) Ig frequency Ig class

Ig reactivity

Patients

Controls

P-value (␹2)

Patients

Controls

P-value

14/76 4/76 8/76 18/76

2/54 1/54 2/54 2/54

.02 .74 .257 .003†

0.673 (0.45-0.95) 0.254 (0.18-0.32) 0.125 (0.09-0.16) 0.379 (0.20-0.64)

0.568 (0.38-0.77) 0.253 (0.17-0.34) 0.09 (0.06-0.13) 0.104 (0.06-0.23)

.047 .701 .0035 ⬍.00001*

IgG IgG1 IgG2 IgG3

Ig reactivity, Ig response registered at an optical density (OD) of 490 nm; Ig frequency, Ig reactivity above mean OD ⫹2SD of control Ig. *Mann-Whitney U test. †␹2 analysis.

Table II B. Distribution of Ig class and subclass response (frequency and reactivity) with LVEF status (LVEF ⬎45% and ⱕ45%) LVEF >45% (n ⴝ 55) Ig class

LVEF <45% (n ⴝ 21)

P-value

Frequency (ⴙ/ⴚ)

Ig reactivity

Frequency (ⴙ/ⴚ)

Ig reactivity

␹2

Reactivity

11/44 4/51 5/50 8/47

0.51 (0.33-0.77) 0.23 (0.18-0.31) 0.12 (0.07-0.15) 0.25 (0.16-0.44)

3/18 0/21 3/18 10/11

0.61 (0.45-0.80) 0.26 (0.19-0.34) 0.13 (0.10-0.17) 0.52 (0.20-1.00)

.45 .36 .35 .006

.34 .59 .12 .003*

IgG IgG1 IgG2 IgG3

⫹/⫺, Number of patients with a positive and negative Ig response. *Mann-Whitney U test.

Table II C. Comparative analysis of Ig class and subclass response in IHD patients (n ⫽ 35) versus study patients (n ⫽ 76) and healthy blood donors (n ⫽ 54) Ig frequency (%)

IHD vs

Controls P-value (␹2)

Study patients P-value (␹2)

(28) 10/35 (3) 1/35 (6) 2/35 (6) 2/35

(3.7) ⬍.0001 (2) .7 (3.7) .95 (3.7) .7

(18.4) .341 (5.4) .944 (10.5) .636 (24) .04

Ig class IgG IgG1 IgG2 IgG3

Ig reactivity

IHD vs

Controls P-value

Study patients P-value

0.740 (0.54-1.13) 0.241 (0.17-0.43) 0.113 (0.09-0.18) 0.105 (0.042-0.25)

.011 .403 .049 .804

.02 .674 .601 ⬍.00001*

*Mann-Whitney U test.

Table II D. Comparative analysis of IgG3 frequency between different groups (␹2) P-value

Patients Study patients (LVEF ⬎45%) IHD Controls

Study patients (LVEF <45%)

Study patients (LVEF >45%)

.006 ⬍.0001 ⬍.0001

– .339 .103

American Heart Journal June 2002

1082 Warraich et al

Table III. Stepwise regression and correlation of Igs with hemodynamic variables Independent predictors (Igs [IgG, IgG1, IgG2, IgG3])

Correlation IgG*

Hemodynamic variables (n ⴝ 76) Echocardiographic data (mm) LVESD LVEDD LA IVS PWT Ventriculographic indices LVEF (%) PCWP (mm Hg) LVESVI (mL/m2 BSA) LVEDVI (mL/m2 BSA) LVEDP (mm Hg) RVEDP (mm Hg) RV-sys (mm Hg) RV-dia (mm Hg) PA-mean (mm Hg) PA-sys (mm Hg) PA-dia (mm Hg)

1st step

2nd step

IgG3 IgG3 – – –

– – – – –

IgG3 IgG3 IgG3 IgG3 IgG3 IgG3 IgG3 IgG3 – – –

– – IgG1 ⫺IgG ⫺IgG – – – – –

Correlation r

IgG3† P-value

Correlation r

0.033 0.183 0.103 0.196 0.136

.806 .419 .147 .126 .293

0.272 0.263 0.243 0.234 0.103

⫺0.152 0.037 0.158 0.118 ⫺0.151 ⫺0.147 0.013 0.018 0.011 0.028 0.031

.189 .760 .317 .446 .212 .222 .913 .884 .951 .815 .719

⫺0.252 0.411 0.463 0.323 0.238 0.246 0.295 0.263 0.193 0.208 0.208

P-value

.038‡ .036‡ .050 .066 .426 .028‡ ⬍.0001‡ .002‡ .033‡ .047‡ .039‡ .014‡ .027‡ .149 .084 .085

sys, Systole; dia, diastole. *IgG correlation (Pearson’s correlation coefficient) with hemodynamic indices. †IgG3 correlation (Pearson’s correlation coefficient) with hemodynamic indices. ‡P ⬍ .05.

([murine] complement-fixing antibodies), switching in human cell lines is less well defined. Therefore, it is not possible to ascribe the exerting effects of IFN-␥ to IgG3 switching (complementary to IgG2a) from the current results. There is, however, an IFN-␥–responsive region within the human Fc␥RI-A gene promoter.24 It may be that activation of the IFN-␥–responsive region by IFN-␥ mediates functional effects of the Fc␥Rs, which may further be fortified by the coupling of IgG to the ␥subunit of the receptor. The original antigenic stimuli, which predict the fate of the Ig switch, may also dictate the immunologic milieu that complements it. Therefore, the production of IFN-␥ in conjunction with IgG3 may possibly reflect earlier phases of disease in particular individuals. The shorter duration of symptoms may therefore reflect the time course in which these patients clinically presented.

Generation of IgG3 autoreactivity The restrictive pattern of an Ig repertoire represents a natural process of affinity maturation and the effective elimination of antigen, whereas autoreactivity may occur as a result of inadequate regulation of these protective immune responses. Naive B-cell progeny possess little or no reactivity against self-determinants; this

predominantly arises through the process of somatic mutation,25 which generates high-affinity antibodies (IgG subclass switching). Elimination of such clones (positive selection) is how tolerance is maintained. Such B cells, which survive positive selection, have been shown to correlate with antiapoptotic activity (bcl-2 expression).26 Increased distribution of bcl-2 activity, within the zonal regions of memory B cells but not in the areas of primary immune responses of secondary germinal centers,27 indicates that it is the secondary wave of antibody generation that is likely to contribute to pathologic characteristics as a result of somatic mutation. Disorders in which predominantly protective antibodies acquire pathogenicity through affinity maturation have been reported.12 Hence, autoreactivity of IgG3 (a 4-fold increase) after affinity maturation may reflect inadequate regulation of a predominantly protective immune response in select individuals. Molecular mimicry may, in particular, facilitate acquisition for antiself specificity. Sequence homologic characteristics between the MHC and a host of infectious agents (chlamydia, enteroviruses [coxsackie B3], and streptococcal proteins)28-30 have been described in myocarditis and DCM. Furthermore, IgG3 and IgG1 are elicited predominantly by viral infections (reported Igs

American Heart Journal Volume 143, Number 6

Warraich et al 1083

Table IV. Association of Ig class and subclass status with hemodynamic characteristics Study patients (n ⴝ 76) Ig class/subclass frequency (%) IgG (18) PCWP LVEDP LVEF LVESVI LVEDVI IgG1 (5) PCWP LVEDP LVEF LVESVI LVEDVI IgG2 (10.5) PCWP LVEDP LVEF LVESVI LVEDVI IgG3 (24) PCWP LVEDP LVEF LVESVI LVEDVI

Antibody-positive patients

Antibody-negative patients

P-value*

10 (5-11.0) 6.5 (4-12.3) 53.5 (45.5-62.0) 48.5 (33.8-128) 105.1 (92.5-190)

7.0 (5-10.0) 8.0 (6-12.8) 54.5 (39.8-64.5) 45.1 (34.4-76.0) 105.4 (92.0-130)

.54 .4 .64 .78 .96

8.5 (4.5-17.7) 12 (5-17.0) 59 (52-66) 46.5 99.9

7.0 (5-10.0) 7.0 (5-12.0) 54 (39.3-64) 45.1 (34.0-79.5) 106.9 (91.8-134)

.72 .54 .44 .92 .67

6.5 (4.5-13.7) 5.0 (2-10.0) 53 (38-62) 65.3 (27-144) 115.7 (79-216)

7.5 (5-10.8) 8.0 (6-13) 55 (41.3-65) 46.3 (34.5-72.2) 105.4 (92.5-124)

.86 .1 .7 .98 .84

11 (.9-16.0) 11 (6.3-19.0) 45.5 (30-62) 81.3 (43-145) 122.7 (97.6-197.7)

6.0 (5-10.0) 7.0 (5-10.5) 55.0 (48-64) 41.4 (33.6-56.5) 102.5 (90.2-123.4)

.005† .05 .049 .009† .02

*Mann-Whitney U test. †P ⱕ .01 (significance following Bonferroni correction for multiple comparison).

against human enteroviruses31), often in the early and subacute phases of disease (IgG2a,32 corresponding murine-Ig). Because clonal restriction of an Ig repertoire determines, in part, the pathophysiologic requisite for the response, generation of IgG3 may, in part, represent the course of viruses in human disease.16 In summary, acquisition of abnormally raised IgG3 response in select patients may reflect dysregulation in mechanisms whereby tolerance to self is abrogated. The high-affinity characteristics of Igs are likely to generate antiself specificity and therefore are likely to represent humoral autoimmunity. The clinical and hemodynamic correlates of IgG3 conform to the characteristic capacity of these Igs to trigger effector functions. These responses may contribute to immunemediated pathogenicity and therefore represent potential therapeutic targets for immunomodulatory therapy. However, further studies are warranted to better evaluate the time course in which these responses evolve together with the underlying immunoregulatory mechanisms responsible for Ig switching.

References 1. Muller J, Wallukat G, Dandel M, et al. Immunoglobulin adsorption in patients with idiopathic dilated cardiomyopathy. Circulation 2000;101:385-91.

2. Felix SB, Staudt A, Dorffel WV, et al. Hemodynamic effects of immunoadsorption and subsequent immunoglobulin substitution in dilated cardiomyopathy. J Am Coll Cardiol 2000;35:1590-8. 3. Parrillo JE, Cunnion RE, Epstein SE, et al. A prospective randomized, controlled trial of prednisone for dilated cardiomyopathy. N Engl J Med 1989;321:1061-8. 4. Mason JW, O’Connell JB, Herskowitz A, et al. A clinical trial of immunosuppressive therapy for myocarditis. N Engl J Med 1995; 333:269-75. 5. Bruggemann M, Williams GT, Bindon CI, et al. Comparison of the effector functions of human immunoglobulins using a matched set of chimeric-antibodies. J Exp Med 1987;166:1351-60. 6. Bindon CI, Hale G, Bruggeman M, et al. Human monoclonal IgG isotypes differ in complement activating function at the level of C4 as well as C1q. J Exp Med 1988;168:127-42. 7. Torgano G, Vecchi M, Podda M, et al. Primary biliary cirrhosis is associated with specific changes in liver IgG-bearing cell subpopulations. J Hepatol 1995;22:545-50. 8. Hoffman WL, Goldberg MS, Smiley JD. Immunoglobulin IgG3 subclass production by rheumatoid synovial tissue cultures. J Clin Invest 1982;69:136-44. 9. Wong A, Kenny TP, Ermel R, et al. IgG3 reactive rheumatiod factor in rheumatoid arthritis: etiologic and pathogenic considerations. Autoimmunity 1994;19:199-210. 10. Bonagura VR, Agostino N, Borretzen M, et al. Mapping IgG epitopes bound by rheumatoid factors from immunised controls identifies disease-specific rheumatoid factors produced by pa-

1084 Warraich et al

11.

12.

13.

14.

15.

16.

17.

18.

19. 20. 21.

tients with rheumatoid arthritis. J Immunol 1998;160:2496505. Bonifacio E, Scirpoli M, Kredel K, et al. Early autoantibody responses in prediabetes are IgG1 dominated and suggest antigenspecific regulation. J Immunol 1999;63:525-32. Imai H, Hamai K, Komatsuda A, et al. IgG subclasses in patients with membrano-proliferative glomerulonephritis, membranous nephropathy, and lupus nephritis. Kidney Int 1997;51:270-6. Warraich RS, Dunn MJ, Yacoub MH. Subclass specificity of autoantibodies against myosin in patients with idiopathic dilated cardiomyopathy: proinflammatory antibodies in DCM patients. Biochem Biophysiol Res Commun 1999;259:255-61. Warraich RS, Pomerance A, Stanley A, et al. Cardiac myosin autoantibodies and acute rejection in patients with idiopathic dilated cardiomyopathy. Transplantation 2000;69:1609-17. Manolio TA, Baughman KL, Rodeheffer R, et al. Prevalence and etiology of idiopathic dilated cardiomyopathy: summary of a National Heart, Lung and Blood Institute workshop. Am J Cardiol 1992;69:1458-66. Pauschinger M, Doerner A, Kuhl U, et al. Enteroviral RNA replication in the myocardium of patients with left ventricular dysfunction and clinically suspected myocarditis. Circulation 1999;99:889-5. Massie BM, Schiller NB, Ratshin RA, et al. Mitral-septal separation: new echocardiographic index of left ventricular function. J Am Coll Cardiol 1977;39:1008-16. Dodge HT, Sheehan FH. Quantitative contrast angiography for assessment of ventricular performance in heart disease. J Am Coll Cardiol 1983;1:73-81. Olsen EG. Endomyocardial biopsies. Int J Cardiol 1983;3:240-3. Aretz HT. Myocarditis: the DALLAS criteria. Hum Pathol 1987;18: 619-24. Kuhl U, Noutsias M, Seeberg B, et al. Immunohistological evidence for a chronic intramyocardial inflammatory process in dilated cardiomyopathy. Heart 1996;75:295-300.

American Heart Journal June 2002

22. Quattrocchi E, Walmsley M, Browne K, et al. Paradoxical effects of adenovirus-mediated blockade of TNF activity in murine collagen-induced arthritis. J Immunol 1999;163:1000-9. 23. Stevens TL, Bossie A, Sanders VM, et al. Regulation of antibody isotype secretion by subsets of antigen-specific helper T cells. Nature 1988;334:255-8. 24. Pearse RN, Feinmann R, Ravetch JV. Characterization of the promoter of the human gene encoding the high affinity IgG receptor: transciptional induction by ␥-interferone is mediated through common DNA response elements. Proc Natl Acad Sci USA 1991;88: 11305-9. 25. French DL, Laskow R, Scharff MD. The role of somatic hypermutation in the generation of antibody diversity. Science 1989;244: 1152-7. 26. Hockenbery DM, Zutter M, Hickey W, et al. Bcl2 protein is topographically restricted in tissues characterised by apoptotic cell death. Proc Natl Acad Sci USA 1991;88:6961-5. 27. Nunez G, Hockenbery D, McDonnell TJ, et al. Bcl-2 maintains B cell memory. Nature 1991;353:71-3. 28. Bachmaier K, Neu N, de-la-Maza LM, et al. Chlamydia infections and heart disease linked through antigenic mimicry. Science 1999;283:1335-9. 29. Cunningham MW, Antone SM, Gulizia JM, et al. Cytotoxic and viral neutralizing antibodies crossreact with streptococcal Mprotein, enteroviruses and human cardiac myosin. Proc Natl Acad Sci USA 1992;89:1320-4. 30. Gauntt CJ, Arizpe HM, Hingdon AL, et al. Molecular mimicry, anti-coxsackievirus B3 neutralizing monoclonal antibodies and myocarditis. J Immunol 1995;154:2983-95. 31. Torfason EG, Reimer CB, Keyserling HL. Subclass restriction of human enterovirus antibodies. J Clin Microb 1987;25:1376-9. 32. Coutelier JP, van-der Logt JT, Heessen FW, et al. Virally induced modulation of murine IgG antibody subclasses. J Exp Med 1988; 168:2373-8.