β1-Adrenergic Receptor Autoantibodies Mediate Dilated Cardiomyopathy by Agonistically Inducing Cardiomyocyte Apoptosis
Background— Antibodies to the β1-adrenergic receptor (β1AR) are detected in a substantial number of patients with idiopathic dilated cardiomyopathy (DCM). The mechanism whereby these autoantibodies exert their pathogenic effect is unknown. Here, we define a causal mechanism whereby β1AR-specific autoantibodies mediate noninflammatory cardiomyocyte cell death during murine DCM.
Methods and Results— We used the β1AR protein as an immunogen in SWXJ mice and generated a polyclonal battery of autoantibodies that showed selective binding to the β1AR. After transfer into naive male hosts, β1AR antibodies elicited fulminant DCM at high frequency. DCM was attenuated after immunoadsorption of β1AR IgG before transfer and by selective pharmacological antagonism of host β1AR but not β2AR. We found that β1AR autoantibodies shifted the β1AR into the agonist-coupled high-affinity state and activated the canonical cAMP-dependent protein kinase A signaling pathway in cardiomyocytes. These events led to functional alterations in intracellular calcium handling and contractile function. Sustained agonism by β1AR autoantibodies elicited caspase-3 activation, cardiomyocyte apoptosis, and DCM in vivo, and these processes were prevented by in vivo treatment with the pan-caspase inhibitor Z-VAD-FMK.
Conclusions— Our data show how β1AR-specific autoantibodies elicit DCM by agonistically inducing cardiomyocyte apoptosis.
Received December 8, 2006; accepted May 11, 2007.
Idiopathic dilated cardiomyopathy (DCM) is a disease of putative autoimmune origin that frequently causes congestive heart failure. A sizable proportion of patients with DCM develop immune responses to cardiac self-antigens; thus, tissue-restricted proteins such as the β1-adrenergic receptor (β1AR) have been investigated as containing determinants for immune cell activation before or during idiopathic DCM.1–7
Clinical Perspective p 410
The β1AR is a 7-transmembrane G-protein–coupled receptor abundantly expressed on cardiomyocytes (reviewed elsewhere8–10). Cognate catecholamine binding to the β1AR transmits an intracellular signal through a canonical cAMP-dependent protein kinase A (PKA) pathway that drives functional cardiomyocyte contractile alterations as part of the sympathetic “fight-or-flight” response. Up to 33% of patients with DCM have been reported to produce detectable circulating autoantibodies directed against epitope regions of the β1AR protein.11 The pathogenic potential of β1AR-specific autoantibodies was affirmed by recent studies in which recipient rodents developed DCM after passive transfer of β1AR-specific antisera12,13 and by clinical trials showing hemodynamic benefit in patients with DCM after protein A immunoadsorption of serum.14,15
Although the qualitative effects of β1AR autoantibodies on cardiomyocyte function and survival have been described in vitro,16–19 a causal mechanism defining how these β1AR-specific antibodies exert their pathogenic effect in vivo is lacking. In this context, we investigated the pathogenesis of β1AR autoantibody–induced heart disease using a newly developed murine model of DCM. We found that sera obtained from SWXJ mice immunized with the β1AR protein are capable of inducing fulminant high-frequency DCM when transferred into naive SWXJ recipients. Development of DCM was dependent on selective binding of β1AR-specific IgG to host β1AR because immunoadsorption of β1AR-specific IgG or treatment with a β1AR but not a β2AR antagonist abrogated the development of DCM.
We found that β1AR antibodies act as classic functional agonists, causing an increase in cardiomyocyte contractility (positive inotropy) and an acceleration in cardiomyocyte relaxation (positive lusitropy) mediated by changes in the handling of intracellular calcium (Ca2+). Sustained agonism of cardiomyocytes by β1AR-specific autoantibodies elicits caspase-3 activation in vivo through the Ca2+- and cAMP-dependent PKA pathways. Taken together, our data define a novel mechanism whereby autoimmune responses directed against the β1AR elicit DCM by inducing noninflammatory cardiomyocyte cell death.
Passive Transfer of DCM
Six- to 8-week-old SWXJ (H-2q,s) male mice (Jackson Laboratories, Bar Harbor, Maine) were immunized subcutaneously in the abdominal flank on day 0 with either 100 μg ovalbumin (OVA; Sigma, St Louis, Mo) or 100 μg human recombinant β1AR (Perkin-Elmer, Boston, Mass) and 400 μg Mycobacteria tuberculosis H37RA (Difco, Detroit, Mich) in 200 μL of an emulsion of equal volumes of water and Freund’s adjuvant (Difco). Three weeks later, cell-free antisera were collected, pooled, and intravenously injected into naive recipient SWXJ male hosts at 200 μL per mouse. Recipient mice were euthanized at various times after transfer of antisera by CO2 asphyxiation followed by cervical dislocation.
For IgG depletion, antisera were passed through a PROSEP-G spin column (Millipore, Billerica, Mass) according to the manufacturer’s recommendations, and IgG depletion was confirmed by coating immunoassay plates (Nalge Nunc, Rochester, NY) with column eluate and performing a direct ELISA using goat anti-mouse IgG (Southern Biotechnology, Birmingham, Ala). For in vivo adrenergic antagonism, mice were treated orally in the drinking water with 40 μg/d of the β1AR antagonist CGP 20712 dichloride (Tocris, Ellisville, Mo) or the β2AR antagonist ICI 118551 (Tocris). All protocols were approved by the Cleveland Clinic Institutional Animal Care and Use Committee in compliance with the Public Health Service policy on humane care and use of laboratory animals.
Cardiac sections were stained with hematoxylin and eosin and read in a blinded manner as described previously.20 Criteria for the presence of DCM included increased ventricular chamber widths and increased ratios of heart weight to body weight, calculated by dividing the heart weight in milligrams by the body weight in grams. Ratios of nuclei staining positively for terminal deoxynucleotidyl transferase–mediated dUTP nick-end labeling (TUNEL) fluorescein (Roche, Indianapolis, Ind) versus 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI) blue (Vector, Burlingame, Calif) were quantified in 6 high-powered fields per sample with ImagePro computer software (Media Cybernetics, Silver Spring, Md).
Histological DCM was confirmed by functional measurements assessed by echocardiography. Mice were imaged in the lateral decubitus position. Echocardiography was performed with a Vivid 7 machine (GE Medical, Milwaukee, Wis). Gray-scale 2-dimensional and M-mode data were collected with a 14-MHz epicardial linear transducer and analyzed with Echopac personal computer analysis software (GE Medical).
ELISA and Western Blot Analysis
For direct ELISA, immunoassay plates (Nalge Nunc) were coated overnight at 4°C with 20 μg/mL β1AR protein, and antisera diluted 1/1000 were overlaid at 4°C on coated plates. Serum antibody isotyping was performed according to the manufacturer’s instructions using the mouse MonoAB ID/SP ELISA kit (Zymed, South San Francisco, Calif). Serum brain natriuretic peptide (BNP) concentrations were measured with the BNP-32 Rat EIA kit (Phoenix Pharmaceuticals, Belmont, Calif). Western blots of cardiac protein extracts were performed with a β1AR-specific antibody (Abcam, Nottingham, UK), β1AR antisera, or an antibody (1/1000) recognizing PKA-specific phosphorylated consensus sequence (RRXS, Abcam).21
β1AR Competition Binding Experiments
Heart membrane proteins were prepared as previously described.22 Membrane protein concentrations were determined by the Bradford method.23 Competition binding experiments using the radioligand β1AR/β2AR antagonist [125I]cyanopindolol (CYP; Amersham, Piscataway, NJ) were carried out in a final volume of 0.5 mL containing HEM buffer (20 mmol/L Hepes, pH 7.5, 1.4 mmol/L EGTA, and 12.5 mmol/L MgCl), 150 pmol/L [125I] CYP (Perkin Elmer, Wellesley, Mass), cardiac cell membranes, and increasing concentrations of norepinephrine (0.1 nmol/L to 1 mmol/L; Sigma) in the presence or absence of β1AR-specific antibody or OVA-specific control antibody at 1/20 000 dilutions. Antisera and membranes were preincubated for 1 hour at 25°C before the addition of the other components. The reactions were then incubated for another 60 minutes at 25°C, terminated by addition of 3 mL binding buffer at 4°C, and rapidly filtered and washed through Whatman GF/C glass fiber filters (Fisher Scientific, Pittsburgh, Pa) using a model 290 PHD cell harvester (Brandel, Gaithersburg, Md). No significant difference existed in total binding in the presence or absence of β1AR antisera. Bound radioactivity was quantified with an auto-gamma 500 counter (Packard Instruments, Meriden, Conn). Binding data were analyzed with the iterative curve-fitting software program GraphPad Prism (GraphPad, San Diego, Calif). An F test using the sum of the squares determined whether the data fit statistically better to a 1- or 2-site model, and the percentages of high- and low-affinity sites were calculated.
Cardiomyocyte single-cell suspensions were obtained through collagenase perfusion as described24 and cultured in 24-well plates on glass coverslips precoated with 10 μg/mL laminin. Cardiomyocytes were suspended in serum-free Krebs-Henseleit buffer containing 118 mmol/L NaCl, 4.8 mmol/L KCl, 1.2 mmol/L MgCl2, 1.2 mmol/L KH2PO4, 1.2 mmol/L CaCl2, 37.5 mmol/L NaHCO3, and 16.5 mmol/L dextrose with a pH adjusted to 7.35. Pharmacological antagonists were added 1 hour before the addition of antisera. Before use in vitro, complement components of β1AR-specific and OVA-specific antisera were functionally depleted by 3 freeze-thaw cycles as confirmed by low cell surface detection of C5 on target cell membranes as previously described.25
Pharmacological agonists and antagonists used in our experiments include 100 nmol/L CGP 20712 dichloride (Tocris), 100 nmol/L ICI 118551 (Tocris), 5 μmol/L PKI 14-22 inhibitor (EMD Biosciences, San Diego, Calif), 100 μmol/L Rp-8-CPT-cAMPS (Biolog, San Diego, Calif), 20 μmol/L H-89 (EMD Biosciences), 40 μmol/L ApoBlock (BD Biosciences, San Jose, Calif), 1 μmol/L nifedipine (Sigma), 1 μmol/L isoproterenol (Sigma), 10 μmol/L KN62 (Calbiochem, San Diego, Calif), and 0.5 μmol/L KN93 (Calbiochem).
Measurement of [Ca2+]i Transients and Cardiomyocyte Shortening
Intracellular Ca2+ concentration [Ca2+]i and cardiomyocyte shortening were simultaneously monitored as previously described.24 Briefly, ventricular cardiomyocytes at 5×105 cells/mL were incubated in Krebs-Henseleit buffer containing 2 μmol/L fura2/AM (Texas Fluorescence Labs, Austin, Tex). Fluorescence measurements were performed on single ventricular cardiomyocytes using a dual-wavelength spectrofluorometer at excitation wavelengths of 340 and 380 nm and an emission wavelength of 510 nm. The cells also were illuminated with red light at a wavelength >600 nm for simultaneous measurement of cardiomyocyte shortening using a video-edge detector (Crescent Electronics, Sandy, Utah). The video-edge detector was calibrated with a stage micrometer so that cell lengths during shortening and relengthening could be measured. Because calibration procedures rely on a number of assumptions, the ratio of the light intensities at the 2 wavelengths was used to measure qualitative changes in [Ca2+]i. Just before data acquisition, background fluorescence was measured and automatically subtracted from the subsequent experimental measurement. The fluorescence sampling frequency was 100 Hz, and data were collected using a software package from Photon Technology International (Lawrenceville, NJ). Fluorescence data for the [Ca2+]i transients and cardiomyocyte shortening were imported into Laboratory View (National Instruments, Austin, Tex), and both the [Ca2+]i transients and cardiomyocyte contractile responses were analyzed synchronously and simultaneously.
For analysis of [Ca2+]i and shortening data, the following variables were calculated for each individual contraction: resting [Ca2+]i and cell length; peak [Ca2+]i and cell length; change in [Ca2+]i (peak [Ca2+]i minus resting [Ca2+]i) and twitch amplitude; time to peak (Tp) for [Ca2+]i and shortening and time to 50% or 90% (T50relax, T90relax) resting [Ca2+]i; and relengthening. Variables from 10 contractions were averaged to obtain mean values at baseline and in response to the various interventions. Averaging the variables over time minimizes beat-to-beat variation.
Quantification of dead and apoptotic cells was determined in a blinded manner on the basis of manual counts of Trypan blue exclusion and fluorescent staining, respectively, of 3 to 5 separate high-power fields per sample containing >200 cells per field. Apoptosis was determined by fluorescein TUNEL staining (Roche) and by Annexin V PE staining (BD Biosciences) according to the manufacturers’ specifications with a modified 2-hour permeabilization period.
Caspase-3 and PKA Enzyme Activity
Caspase-3 activity was measured with the ApoAlert Caspase-3 Assay Plate (Clontech, Mountain View, Calif) according to the manufacturer’s instructions with a modified 60-minute reaction time at 37°C. PKA activity was measured with the PepTag Non-Radioactive cAMP Dependent Protein Kinase Assay (Promega, Madison, Wis) and calculated per the manufacturer’s specifications. Spectrophotometric measurements at 570 nm were performed on 40 μg/mL cardiac tissue after a 30-minute reaction time at room temperature.
Flow Cytometry Competition Assay
Cardiomyocyte single-cell suspensions were harvested from 24-well plates via mechanical disruption with a pipette and incubated with varying dilutions of human recombinant β1AR or β2AR protein for 20 minutes before staining with β1AR antisera at 1/50 dilution and cardiac myosin heavy-chain antibody (Abcam) at 1/100 dilution followed, respectively, by PE-conjugated goat anti-mouse IgG (BD Biosciences) and FITC conjugated goat anti-rabbit IgG (Vector). Results were analyzed with a BD FACSCalibur flow cytometer (BD Biosciences).
The nonpaired Student t test was used to compare differences in frequencies of apoptotic nuclei, caspase-3, and PKA enzyme activity; ratios of heart weight to body weight; echocardiography functional data; high-affinity receptor states; peak [Ca2+]i transients; peak cardiomyocyte shortening; and BNP serum levels in mice or cultures treated with β1AR versus OVA antisera.
The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.
Generation of β1AR-Specific Autoantibodies
To generate β1AR autoantibody, we immunized male SWXJ mice with recombinant human β1AR protein. At 2, 4, and 8 weeks after immunization, the presence of serum IgG binding to β1AR protein at a serum dilution of 1/1000 was determined by direct ELISA. Immunization with the β1AR protein caused a progressive increase over time in antigen-specific IgG (Figure 1A). We found that β1AR-specific autoantibodies consisted predominantly of IgG2a and IgG2b isotypes with relatively low levels of the IgG1 isotype, consistent with a Th1-like antibody phenotype26 (Figure 1B).
We next determined the specificity of autoantibody binding to the β1AR by Western blot analysis. We found that β1AR specific antisera at a 1/1000 dilution and a commercially purchased positive control β1AR-specific antibody showed similar Western blot staining patterns against cardiac protein extracts (Figure 1C, lanes 1 and 2, respectively). No signal was detected when β1AR antisera were blotted against β2AR recombinant protein (Figure 1C, lane 3). Specific immunostaining of endogenous β1AR by β1AR antisera was attenuated with increasing concentrations of soluble recombinant β1AR (Figure 1C, top right blot) but not β2AR (Figure 1C, bottom right blot) added as competition.
As further confirmation of the specificity of the β1AR antisera, cardiomyocytes were double stained with β1AR antisera and cardiac myosin heavy-chain antibody in the presence of varying dilutions of competing β1AR or β2AR protein. We observed diminishing frequencies of the double-staining β1AR-positive and cardiac myosin–positive cells with increasing concentration of added β1AR (Figure 1D, top row) but not β2AR protein (Figure 1D, bottom row). From these studies, we concluded that antisera generated against the β1AR protein contained substantial levels of Th1-like skewed autoantibodies that showed a pattern of β1AR-specific binding virtually identical to commercially available β1AR-specific antibodies.
β1AR-Specific Autoantibodies Cause DCM
To determine the pathogenic potential of β1AR autoantibodies, immune antisera were injected intravenously into naive male SWXJ hosts. At 3 weeks after passive transfer, we found that 7 of 12 mice receiving β1AR-specific antisera (Figure 2A, top) developed gross cardiac enlargement and biventricular dilatation consistent with DCM. None of 8 mice receiving control OVA-specific antisera developed DCM (Figure 2A, bottom). Evolution of the DCM phenotype in affected mice occurred without any observed increase in cardiac mononuclear cell infiltration. DCM occurred in mice (3 of 4) injected with the IgG-enriched fraction of β1AR antisera (Figure 2B, top) but was substantially abrogated in mice (1 of 8) injected with β1AR antisera depleted of IgG by immunoadsorption (Figure 2B, bottom). In addition, DCM occurred in mice (4 of 8) receiving 40 μg/d ICI 118551, a selective β2AR antagonist (Figure 2C, top), but was substantially abrogated in mice (1 of 8) receiving 40 μg/d CGP 20712, a selective β1AR antagonist (Figure 2C, bottom).
Correlate echocardiographic measurements were taken at multiple time points after transfer of β1AR antisera. Mice injected with β1AR- but not OVA-specific antisera showed functional changes consistent with DCM, with significantly reduced fractional shortening (Figure 2D, left) and ejection fractions (Figure 2D, middle) on days 17 and 28 (P<0.02) and significantly increased left ventricular end-systolic diameter (Figure 2D, right) on day 28 (P<0.03).
Cardiomegaly was assessed by measuring the ratios of heart weight (mg) to body weight (g) in all treatment groups. Our measurements showed significant differences in cardiomegaly between β1AR antisera– and OVA antisera–treated mice (P=0.001), between IgG fraction– and IgG-depleted fraction–treated mice (P=0.04), and between β1AR antisera–injected mice treated with the β2AR antagonist ICI 118551 and the β1AR antagonist CGP 20712 (P=0.03; the Table). Together, our data indicate that β1AR IgG autoantibodies are sufficient to mediate DCM in vivo and that the development of DCM is contingent on specific binding of autoantibody to endogenous cardiac β1AR.
β1AR-Specific Autoantibodies Act Agonistically
We hypothesized that β1AR autoantibodies may be exerting a direct toxic effect on host cardiomyocytes, possibly through alteration of cardiomyocyte signaling. To examine this possibility, cardiac surface membranes with abundant β1AR were isolated and exposed to β1AR antisera in the presence of the radioactive β1AR antagonist, CYP, and competing dilutions of norepinephrine. A sigmoid-shaped binding curve was constructed from 5 of these competition assays, and the percentage of receptors in the high-affinity state was calculated (Figure 3A, left). We found that β1AR antisera caused a leftward shift of the competition binding curve consistent with an increase in the proportion of receptors in the high-affinity state. Computer modeling indicated that the β1AR antisera caused a statistically significant increase (P<0.0001) in the frequency of high-affinity β1AR conformations (Figure 3A, right).
The observed high-affinity conformational shift implied that endogenous β1AR had become coupled to G protein, thereby increasing intracellular second messengers. To determine whether the intracellular signal derived from β1AR antibody binding is capable of altering cardiomyocyte contractile function, untreated male SWXJ cardiomyocytes were freshly isolated and loaded with fura2/AM for measurement of [Ca2+]i. Simultaneous measurements of [Ca2+]i (Δ340/380 ratio) and cardiomyocyte contractility (video-edge detection) were recorded in individual cells paced at 0.5 Hz before and after exposure to β1AR-specific antisera (1/1000), OVA-specific antisera (1/1000), or isoproterenol (1 μmol/L). Exposure of cardiomyocytes to β1AR antisera (Figure 3B, right top and bottom) but not OVA antisera (Figure 3B, left top and bottom) increased peak [Ca2+]i and peak shortening comparable to isoproterenol-treated cardiomyocytes (Figure 3B, middle top and bottom). An overlay depicting an average of 10 individual Ca2+ transients (Figure 3C, top left) and contractile events (Figure 3C, bottom left) clearly demonstrates the positive inotropic and lusitropic effects of β1AR antisera on cardiomyocyte function. Compared with OVA antisera–treated controls, exposure of cardiomyocytes to β1AR antisera resulted in an increase in cardiomyocyte peak [Ca2+]i (Δ340/380 ratio of 0.5 for OVA antisera versus 0.7 for β1AR antisera) and an increase in twitch amplitude (2.50 μm for OVA antisera versus 6.4 μm for β1AR antisera).
Compared with treatment of cardiomyocytes with OVA antisera, substantially increased [Ca2+]i (Figure 3C, top left) and significantly increased (P=0.001) twitch amplitude (Figure 3C, bottom left) occurred after exposure of 10 individual cardiomyocytes to β1AR antisera. Moreover, an acceleration in the decline of the Ca2+ transient to 50% (T50relax) or 90% (T90relax) of the baseline value, accompanied by an acceleration in cardiomyocyte relengthening, also was observed (data not shown). We were unable to assess the potential for changes in cardiac chronotropy occurring in response to any intervention because the cardiomyocytes were electrically paced. Taken together, our data indicate that β1AR-specific antisera trigger a positive inotropic and lusitropic effect in cardiomyocytes that is typical of classic β1AR activation of the heart.10
β1AR-Specific Autoantibodies Induce Cardiomyocyte Apoptosis In Vitro
Sustained agonism induces cell death in numerous cell types.27–29 Additionally, persistent isoproterenol-induced agonism of β1AR/β2AR leads to cardiomyocyte cell death.30 To determine whether β1AR autoantibodies affect cardiomyocyte survival, single-cell suspensions of cardiomyocytes were cultured with varying dilutions of complement-depleted β1AR-specific antisera and pharmacological inhibitors. After 24 hours, cardiomyocytes were stained with a 1/50 dilution of Trypan blue, and the cumulative percentage of cell death was measured by Trypan blue exclusion in 3 to 5 high-power fields. Compared with cultures treated with OVA antisera (Figure 4A, top left, and Figure 4B), prolonged exposure to β1AR-specific antisera caused a dose-dependent increase in cell death (Figure 4A, right column, and Figure 4B) similar to that seen in cultures treated with the nonselective β1AR/β2AR agonist isoproterenol (Figure 4B). Cumulative cell death induced by β1AR antisera (1/200) was attenuated by pretreatment with 100 nmol/L CGP 20712, a selective β1AR antagonist, but not with 100 nmol/L ICI 118551, a selective β2AR antagonist (Figure 4B). Most notably, total cell death in response to β1AR antisera was prevented by pretreatment of cardiomyocytes for 1 hour with the pan-caspase inhibitor ApoBlock (Figure 4B), suggesting that apoptosis was the predominant mode of cell death.
To directly assess apoptotic frequencies, cardiomyocytes were examined for TUNEL and Annexin V staining. Compared with a 1/200 dilution of OVA antisera–treated cultures (Figure 4C, top left, and Figure 4D), cardiomyocytes treated with a 1/200 dilution of β1AR antisera showed a higher frequency of apoptosis as shown by TUNEL staining (Figure 4C, bottom left, and Figure 4D). Frequencies of apoptotic cells were substantially decreased by addition of the β1AR antagonist CGP 20712 (Figure 4C, top right, and Figure 4D) but not by addition of the β2AR antagonist ICI 118551 as shown by Annexin V staining (Figure 4C, middle right, and Figure 4D). β1AR antisera–induced apoptosis as measured by TUNEL staining was prevented by pretreatment of cardiomyocytes for 1 hour with the pan-caspase inhibitor ApoBlock (Figure 4C, bottom right, and Figure 4D). Taken together, our data indicate that sustained agonism by β1AR autoantibodies caused cardiomyocyte apoptosis.
β1AR-Specific Autoantibodies Induce Cardiomyocyte Apoptosis In Vivo
We next determined whether apoptosis of endogenous cardiomyocytes occurs in vivo in mice treated with β1AR antisera. One week after transfer of β1AR- or OVA-specific antisera into naive SWXJ male mice, hearts were examined by TUNEL staining for the presence of apoptotic nuclei. Compared with hearts from OVA antisera–treated mice, hearts from β1AR antisera–treated mice showed an increased frequency of apoptotic TUNEL-positive green nuclei versus nonapoptotic DAPI-positive blue nuclei (Figure 5A through 5C). Digital quantification of TUNEL-stained nuclei showed a significantly increased (P<0.01) mean frequency of apoptotic TUNEL-positive nuclei 1, 3, and 7 days after injection of β1AR antisera compared with mice injected with OVA antisera (Figure 5D). The frequency of apoptotic nuclei in recipients of β1AR antisera is consistent with those observed in prior studies in transgenic mice showing that relatively low frequencies of cardiac apoptosis over time are sufficient to elicit severe DCM.31
Concurrently, we also determined levels of caspase-3 activity in recipient hearts at various times after transfer of β1AR- and OVA-specific antisera and found that hearts from mice treated with β1AR antisera showed significantly increased caspase-3 enzymatic activity (P=0.001) compared with hearts from mice treated with OVA antisera (Figure 5E). The caspase-3 activity in hearts from mice treated with β1AR antisera was significantly increased at the 12-hour, 24-hour, 4-day, and 7-day time points after transfer of antisera (P<0.01).
To determine the significance of apoptosis in the development of DCM, male hosts were treated with the pan-caspase inhibitor Z-VAD-FMK or a vehicle control solution after transfer of β1AR antisera. At 3 weeks, recipient hearts and sera were analyzed. Mice treated with vehicle developed a high incidence (4 of 6) of DCM (Figure 5F, top left), whereas DCM incidence was substantially decreased (1 of 6) in mice treated with Z-VAD-FMK (Figure 5F, bottom left). The decreased incidence of DCM in Z-VAD-FMK–treated mice was accompanied by a significant reduction (P<0.0001) in the number of TUNEL-positive apoptotic nuclei 1 day after transfer of β1AR antisera (Figure 5F, top right), by a significantly reduced ratio of heart weight to body weight (P<0.01; Figure 5F, middle right), and by significantly decreased serum levels of BNP, a marker for heart failure32 (P<0.01; Figure 5E, bottom right). Taken together, our data indicate that β1AR autoantibodies induce cardiomyocyte apoptosis in vivo and that apoptosis is required for β1AR autoantibody–mediated DCM.
β1AR Autoantibody–Induced Apoptosis and DCM Are PKA Dependent
A recent study has shown that agonism of β1AR with the nonselective β1AR/β2AR agonist isoproterenol induces cardiomyocyte apoptosis by a Ca2+/calmodulin-dependent protein kinase II (CaMKII) –associated pathway independently of canonical cAMP-dependent PKA signaling.30 Therefore, we investigated whether the cAMP-dependent PKA pathway was responsible for providing the proapoptotic signal in β1AR antisera–treated cardiomyocytes.
Single-cell suspensions of male SWXJ cardiomyocytes were pretreated for 1 hour with 40 μmol/L H-89, a nonselective inhibitor of β1AR/β2AR pathways. Twenty-four hours later, cardiomyocyte cell death as measured by Trypan blue exclusion and apoptosis as measured by TUNEL and Annexin V immunostaining were reduced to levels observed in untreated control cultures (Figure 6A), indicating that H-89 completely rescued cardiomyocytes from β1AR antisera–induced apoptosis (P<0.001).
To determine whether the cAMP-dependent PKA pathway was responsible for mediating this apoptotic rescue, selective inhibitors of either the PKA/cAMP-dependent or CaMKII-dependent pathway were added to cardiomyocyte cultures before the addition of β1AR antisera. A PKA-specific inhibitor, PKI 14-22 (5 μmol/L), and an inhibitory cAMP analogue, RP-8-CPT-cAMPS (100 μmol/L), were added to inhibit the PKA/cAMP-dependent pathway. The selective CaMKII inhibitors KN-62 (10 μmol/L) and KN-93 (0.5 μmol/L) were added to inhibit the CaMKII-dependent pathway. At 20 hours, we found that specific inhibition of cAMP-dependent PKA signaling completely rescued cardiomyocytes from β1AR antisera–induced apoptotic cell death as measured by frequencies of TUNEL- or Annexin V–positive cells (P<0.001; Figure 6A). However, selective CaMKII inhibition resulted in only a mild nonsignificant rescue from cell death and apoptosis, indicating that β1AR autoantibodies induce cardiomyocyte apoptosis predominantly by activating the canonical cAMP-dependent PKA pathway.
Activation of the cAMP-dependent PKA pathway results in phosphorylation of multiple protein substrates that functionally increase [Ca2+]i.8–10 Because we observed an increase in peak [Ca2+]i after exposure to β1AR antisera (Figure 3B) and because disruption in intracellular Ca2+ homeostasis elicits apoptosis,33 we hypothesized that inhibiting peak [Ca2+]i would similarly prevent cardiomyocyte cell death. To test this hypothesis, cardiomyocytes were pretreated with the L-type Ca2+ channel antagonist nifedipine 1 hour before the addition of β1AR antisera. At 24 hours, we found that nifedipine inhibition of Ca2+ entry into cardiomyocytes abrogated total cell death and apoptosis to levels similar to those observed in PKI 14-22– and RP-8-CPT-cAMPS–pretreated cultures (Figure 6A). A similar level of rescue from apoptotic cell death was observed when Ca2+ was buffered in the media with 5 mmol/L EGTA before the addition of β1AR antisera (data not shown). Taken together, our results indicate that β1AR antisera–induced cardiomyocyte apoptotic cell death is mediated through the cAMP-dependent PKA pathway, likely through alterations in Ca2+ homeostasis.
We next determined whether PKA activation occurred in vivo during the development of β1AR antisera–induced DCM. To this end, cardiac proteins were extracted for analysis of PKA enzyme activity at various time points after injection of naive recipient mice with either β1AR-specific or OVA-specific antisera. We found that PKA-specific enzyme activity was increased in hearts from mice treated with β1AR antisera at all the time points examined compared with hearts from OVA antisera–treated mice (Figure 6B). Moreover, Western blot analysis showed that hearts from mice treated with β1AR antisera had increased levels of phosphorylated PKA substrates at 1-, 2-, and 3-hour time points after injection with antisera (Figure 6C). The increase in phosphorylated substrates was dependent on β1AR-specific binding because levels of protein phosphorylation decreased with concurrent β1AR-selective pharmacological blockade with the β1AR antagonist CGP 20712 (data not shown). Taken together, our data indicate that β1AR autoantibodies initiate intracellular cAMP-dependent PKA signaling in cardiomyocytes in vivo.
In the present study, we generated a polyclonal pool of β1AR-specific antibodies that elicited high-frequency DCM in vivo through endogenous agonist-induced apoptosis of native cardiomyocytes. Our studies demonstrate the central role of apoptosis in the development of β1AR antibody–induced DCM and reveal a mechanism whereby autoimmune responses to a cardiac self-antigen, the β1AR protein, incite end-organ cell death through a noninflammatory mechanism.
Autoreactive responses directed against cardiac receptor signaling elements or contractile apparatus proteins play a significant role in the pathogenesis of DCM. Our studies have demonstrated a definitive pathogenic mechanism of β1AR-specific antibodies in vivo that was lacking in prior studies using cardiomyocyte cultures.16–19 On the basis of this work, idiopathic DCM can be considered in the same light as other autoimmune diseases such as myasthenia gravis and Grave’s disease in which autoantibody recognition of protein receptors pathologically modulates parenchymal organ function. Indeed, autoantibodies directed against a multiplicity of cardiac antigens, including cardiac myosin,5 the β1AR,6,11,12 the m2-muscarinic acetylcholine receptor,34 cardiac laminin,35 and the F1 ATP synthase,36 have been documented in patients with idiopathic DCM, and because these target autoantigens physiologically modulate cardiac function, antibody self-association to 1 or more of these proteins may represent a significant mechanism toward the overall pathogenesis of idiopathic DCM.
Most notably, we observed that autoantibody-induced cardiomyocyte apoptosis and pathological dilatory remodeling were completely reversed by in vivo inhibition of caspase activity with Z-VAD-FMK. This observation highlights the critical role of apoptosis in the pathogenesis of β1AR antibody–induced DCM and suggests that antibody-mediated proinflammatory mechanisms such as antibody-dependent cell-mediated cytolysis or complement-dependent cytolysis likely play relatively minor roles in the development of disease. In this regard, agonist-induced cardiomyocyte apoptosis may represent one of several noninflammatory modes of antibody-mediated pathology in DCM. Natural antibodies, composed of an evolutionarily conserved repertoire of broadly self-reactive IgM, for example, may promote tissue destruction through noninflammatory means, including direct target antigen proteolysis37 and target antigen opsonization.38 Thus, inflammatory destruction of cardiomyocytes may not necessarily constitute an obligate pathogenic component of DCM in which, especially in chronic disease, antibody-mediated end-organ destruction may persist long after immune self-tolerance has been established and cellular inflammatory processes have been attenuated.
Although species-defined differences almost always exist in immune recognition of identical proteins, it is important to note that our SWXJ mice developed autoimmune heart failure in response to human recombinant β1AR. Thus, in our study, mouse antibody recognition of human β1AR epitopes resulted in DCM, implying that identification of such epitopes may likely have relevant implications for human disease mediated by antibodies to the β1AR. Within the polyclonal pool of β1AR autoantibodies used in this study, a high degree of antibody heterogeneity is likely with regard to antigen binding sites on the β1AR protein and functional outcomes associated with such differential binding. Multiple pathogenic epitopes may be disseminated throughout the β1AR protein, and antibody binding to multiple sites may be required to allosterically convert endogenous β1AR to the high-affinity G-protein–coupled state. Alternatively, immunodominant determinants may be selectively enriched within a small region of the β1AR protein, analogous to the main immunogenic region on the nicotinic acetylcholine receptor in myasthenia gravis.39 Because antibodies are structurally large molecules relative to native β1AR catecholamine ligands, it is likely that binding sites for β1AR-specific autoantibodies are confined to portions of the extracellular loop domains of the β1AR protein. These loop domains have previously been shown to initiate G-protein–coupled signaling19 and represent putative autoantibody binding sites, in addition to the canonical catecholamine signaling domain buried within the 7-transmembrane pore of the β1AR protein, which is likely inaccessible to autoantibody binding because of molecular size constraints.
The low frequency of apoptotic nuclei observed in our study after β1AR antisera transfer is in close alignment with a previous study showing that apoptosis in as few as ≈230 per 1×106 cardiomyocytes is sufficient to cause DCM with complete penetrance. Previous studies in humans with heart failure40–42 indicated much higher frequencies of apoptotic nuclei. However, given the lack of functional redundancy of the heart as an organ and with limited regenerative capacity of its constituent cardiomyocytes, it is unlikely that the heart can functionally support high frequencies of apoptotic events at any given point in time, especially during the often chronic course of heart failure. Our data instead support the notion that chronic, low-level cardiomyocyte apoptosis incites DCM and heart failure.
Sources of Funding
This work was supported by an American Heart Association fellowship 0215114B (Dr Jane-wit) and National Institutes of Health grants HL-65661 (Dr Wang), HL-065701 (Dr Damron), HL-74400 (Dr Penn), HL-61438 (Dr Perez), AI-51837 (Dr Tuohy), and DC-006422 (Dr Tuohy).
Neumann DA, Burek CL, Baughman KL, Rose NR, Herskowitz A. Circulating heart-reactive antibodies in patients with myocarditis or cardiomyopathy. J Am Coll Cardiol. 1989; 16: 839–846.
Limas CJ. Cardiac autoantibodies in dilated cardiomyopathy: a pathogenetic role? Circulation. 1997; 95: 1979–1980.
Neu N, Rose NR, Beisel K, Herskowitz A, Gurri-Galss G, Craig SW. Cardiac myosin induces myocarditis in genetically predisposed mice. J Immunol. 1987; 139: 3630–3636.
Liao L, Sindhwani R, Rojkind M, Factor S, Leinwand L, Diamond B. Antibody-mediated autoimmune myocarditis depends on genetically determined target organ sensitivity. J Exp Med. 1995; 181: 1123–1131.
Limas CJ, Goldenberg IF, Limas C. Autoantibodies against beta-adrenoceptors in human idiopathic dilated cardiomyopathy. Circ Res. 1989; 64: 97–103.
Dorffel WV, Felix SB, Wallukat G, Brehme S, Bestvater K, Hofmann T, Kleber FX, Baumann G, Reinke P. Short-term hemodynamic effects of immunoadsorption in dilated cardiomyopathy. Circulation. 1997; 95: 1994–1997.
Magnusson Y, Wallukat G, Waagstein F, Hjalmarson A, Hoebeke J. Autoimmunity in idiopathic dilated cardiomyopathy: characterization of antibodies against the β1adrenoreceptor with positive chronotropic effect. Circulation. 1994; 89: 2760–2767.
Jane-wit D, Yu M, Edling AE, Kataoka S, Johnson JM, Stull LB, Moravec CS, Tuohy VK. A novel class II-binding motif selects peptides that mediate organ-specific autoimmune disease in SWXJ, SJL/J, and SWR/J mice. J Immunol. 2002; 169: 6507–6514.
Ross SA, Rorabaugh B, Chalothorn D, Yun J, Gonzalez-Cabrera PJ, McCune DF, Piascik MT, Perez DM. The alpha(1B)-adrenergic receptor decreases the inotropic response in the mouse Langendorff heart model. Cardiovasc Res. 2003; 60: 598–607.
Kanaya N, Murray PA, Damron DS. Propofol and ketamine only inhibit intracellular Ca++ transients and contraction in rat ventricular myocytes at supraclinical concentrations. Anesthesiology. 1998; 84: 397–403.
Brodbeck WG, Liu D, Sperry J, Mold C, Medof EM. Localization of classical and alternative pathway regulatory activity within the decay-accelerating factor. J Immunol. 1996; 156: 2528–2533.
Chen M, Wang YH, Wang Y, Huang L, Sandoval H, Liu YJ. Wang J. Dendritic cell apoptosis in the maintenance of immune tolerance. Science. 2006; 311: 1160–1164.
Goodyear CS, Silverman GJ. Death by a B cell superantigen: in vivo VH-targeted apoptotic supraclonal B cell deletion by a Staphylococcal toxin. J Exp Med. 2003; 197: 1125–1139.
Zhu WZ, Wang SQ, Chakir K, Yang D, Zhang T, Brown JH, Devic E, Kobilka BK, Cheng H, Xiao RP. Linkage of beta1-adrenergic stimulation to apoptotic heart cell death through protein kinase A-independent activation of Ca2+/calmodulin kinase II. J Clin Invest. 2003; 11: 617–625.
McConkey DJ, Hartzell P, Amador-Perez JF, Orrenius S, Jondal M. Calcium-dependent killing of immature thymocytes by stimulation via the CD3/T cell receptor complex. J Immunol. 1989; 143: 1801–1806.
Zhang M, Alicot EM, Chiu I, Li J, Verna N, Vorup-Jensen T, Kessler B, Shimaoka M, Chan R, Friend D, Mahmood U, Weissleder R, Moore FD, Carroll MC. Identification of the target self-antigens in reperfusion injury. J Exp Med. 2006; 203: 141–152.
Tzartos SJ, Barkas T, Cung MT, Mamalaki A, Marraud M, Orlewski P, Papanastasiou D, Sakarellos C, Sakarellos-Daitsiotis M, Tsantili P, Tsikaris V. Anatomy of the antigenic structure of a large membrane autoantigen, the muscle-type nicotinic acetylcholine receptor. Immunol Rev. 1998; 163: 89–120.
Up to 33% of patients with idiopathic dilated cardiomyopathy have circulating antibodies directed against the β1-adrenergic receptor (β1AR). In this report, we show that β1AR-specific autoantibodies are pathogenic, and we elucidate the mechanism whereby these antibodies mediate dilated cardiomyopathy. We show that β1AR-specific antibodies cause dilated cardiomyopathy in vivo and that these autoantibodies act as agonists, shifting endogenous β1ARs into the agonist-coupled high-affinity state, thereby triggering an intracellular protein kinase A–dependent signaling cascade in cardiomyocytes. Sustained cardiomyocyte activation by β1AR-specific antibodies causes intracellular calcium mishandling, cardiomyocyte apoptosis, and ultimately congestive heart failure with echocardiographically defined dilated cardiomyopathy. These in vivo effects are prevented by inhibition of caspase activation and subsequent apoptosis. Our findings are significant in defining how β1AR-specific autoantibodies exert their pathogenic effect in vivo. Furthermore, our findings support the view that cardiac-specific autoantibodies are pathogenic and that therapies aimed at clearance of circulating autoantibodies may be beneficial in patients with idiopathic dilated cardiomyopathy.