Article Text
PDF
Abstract
Background Congenital heart block (CHB) may develop in fetuses of women with anti-Ro/La autoantibodies following placental transfer of maternal autoantibodies and disruption of the fetal atrioventricular (AV) conduction system. Animal models of CHB currently rely on immunisation or transfer of anti-Ro/La antibodies purified from mothers of children with CHB, which does not allow precise identification of the disease-inducing antibody specificity.
Objective To determine the ability of different anti-Ro52 monoclonal antibodies to induce cardiac electrophysiological abnormalities in vivo and affect the calcium homoeostasis of cardiomyocytes in vitro.
Methods Monoclonal antibodies recognising different domains of Ro52 were generated and injected into pregnant rats, and ECG was recorded on newborn pups. Cultures of rat neonatal cardiomyocytes were established to assess the effect of the different anti-Ro52 monoclonal antibodies on calcium homoeostasis.
Results First-degree AV block and bradycardia developed after maternal transfer of antibodies specific for amino acids 200–239 of Ro52 (p200), while pups exposed to antibodies targeting N- or C-terminal epitopes of Ro52 did not show any electrocardiogram abnormalities. Addition of an anti-p200 antibody to cultured cardiomyocytes induced calcium dyshomoeostasis in a time- and dose-dependent manner, while addition of other Ro52 antibodies had no effect.
Conclusion These data for the first time show unambiguously that antibodies specific for amino acids 200–239 of Ro52 can induce cardiac conduction defects in the absence of other autoantibodies, and may therefore be the main initiators of cardiac pathology in the pool of anti-Ro52 antibodies in mothers of children with CHB.
Statistics from Altmetric.com
Congenital heart block (CHB) is a cardiac manifestation of the neonatal lupus syndrome, which may develop in fetuses of women with anti-Ro/SSA antibodies.1 During pregnancy, the maternal autoantibodies are transferred to the child and may initiate a series of events that will lead to cardiac inflammation, fibrosis and calcification, eventually blocking signal conduction at the atrioventricular (AV) node.2 3 Up to one-third of fetuses in anti-Ro52-positive pregnancies have been shown to develop signs of first-degree AV block and sinus bradycardia in utero, and these blocks may subsequently revert to normal sinus rhythm or progress to a more severe degree of AV block.4 This suggests that complete CHB may develop as a multistep process, where maternal antibodies induce fetal cardiac conduction disturbances, detected as a first-degree AV block, which may or may not subsequently progress to complete AV block depending on the presence of additional risk factors.5 Such risk factors may include environmental conditions or fetal genetic factors, of which fetal major histocompatibility (MHC) genes have been indicated in both patients and animal models.6 7
While the association of CHB with the presence of maternal anti-Ro antibodies is well established, recent studies indicate that the risk for fetal heart block may be higher in women in whom antibody reactivity targets the Ro52 rather than the Ro60 component of the Ro/SSA complex.8 The presence of maternal antibodies with specificity for amino acids 200–239 (p200) of the Ro52 protein has recently been suggested to increase the predictive value for fetal cardiac involvement.9
Immunisation of mice and rats with the Ro52 protein or the p200 peptide leads to AV block in the offspring10 11; however, whether only p200-specific antibodies carry the pathogenic potential of the maternal anti-Ro52 antibody pool or whether antibodies targeting other epitopes of the Ro52 protein also induce heart block remains to be determined. Interestingly, Ro52 expression is low in the heart12 and it has been suggested that the effect of the maternal autoantibodies stems from binding to a cross-reactive protein in the fetal heart.5 13 14
The generation of heart block-inducing antibodies has recently been shown to depend on maternal MHC,7 therefore limiting the value of active immunisation as a reliable method of inducing heart block in a genetically diverse population. By contrast, a passive antibody transfer model would bypass maternal genetic differences, allowing for a precise control of antibody specificity and for further investigation of maternal and fetal requirements for the development of heart block.
To define the specificity of heart block-inducing Ro52 antibodies we developed an animal model of heart block by transfer of a panel of anti-Ro52 monoclonal antibodies with specificity for different functional domains of Ro52 into female rats during gestation.
Methods
Experimental animals
Dark Agouti (DA) rats (NOVA-SCB, Sollentuna, Sweden)were kept and bred in the animal facility at the Center for Molecular Medicine at the Karolinska Institutet. All experimental protocols were approved by the Stockholm North Ethics Committee.
Antibody transfer and ECG recording
Female DA rats, 10–12 weeks old, were injected intraperitoneally with phosphate-buffered saline (PBS) or a total of 4 mg of monoclonal antibodies of IgG1 isotype (2 mg at days 6 and 9), unless specified otherwise. Three-lead electrocardiograms (ECGs) were recorded within 24 h of birth from all conscious pups using four microelectrodes attached to a body clip.11 ECGs were sampled for 5 s four times a minute, with a sampling rate of 1000 Hz. The ECG was digitalised and analysed with Pharmlab (AstraZeneca, Mölndal, Sweden).QRS complexes were averaged and used to calculate the PR interval. PR intervals were corrected for heart rate (HR) variation by expressing them as PR/√RR. First-degree AV block was defined as PR/√RR≥(mean PR/√RR)PBScontrol+2×SD. Bradycardia was defined as HR<(mean HR)PBScontrol−2×SD.
Monoclonal antibodies
Anti-Ro52 monoclonal antibodies were generated as previously described.15 Briefly, Balb/c mice were immunised with recombinant human Ro52 full-length protein. Spleen cells were fused with SP 2/0 myeloma cells and hybridomas supernatants were screened for Ro52 IgG antibodies in ELISA. Positive clones were subcloned at least twice by limiting dilution. All monoclonal antibodies used in this study were of the IgG1 isotype (Clonotyping System-HRP kit, Southern Biotechnology Associates, Birmingham, AL, USA).
Proteins and peptides
Full-length human Ro52 protein and deletion constructs were generated by enzymatic restriction, expressed as fusion proteins with the maltose-binding protein from the pMAL-vector and purified according to the manufacturer's instructions (New England Biolabs, Ipswich, MA, USA),as previously described.16 17 Ro52 peptides p200, pZIP, pOUT and A233 were purchased from Thermo BioSciences (Ulm, Germany).
ELISA
Ro52 and p200 ELISA were performed as previously described.15 Monoclonal antibodies were tested at a concentration of 1 μg/ml, and serum from female rats injected with a monoclonal antibody at a dilution of 1:100.
Cardiomyocyte culture
Cultures of cardiomyocytes were prepared using a kit (Worthington Biochemical, Lakewood, NJ, USA).Hearts from 1–2-day-old DA rats were dissected and prepared according to the manufacturer's instructions. The cardiomyocytes were cultured in Dulbecco's modified Eagle's medium/F12 supplemented with 10% fetal calf serum, penicillin (100 U/ml), streptomycin (100 μg/ml), 2.5 μg/ml insulin, 2.5 μg/ml transferrin, 2.5 ng/ml selenin, 30 μg/ml BrdU and 15 mM Hepes at 37°C with 5% CO2. Fresh medium was replenished every second day.
Time lapse Ca2+ imaging
Cardiomyocytes, prepared as described above, were cultured on poly-D-lysine-coated glass slips. Cells were loaded with the Ca2+-sensitive dye fluo-4-acetoxymethylester (fluo-4 AM, Molecular Probes, Life Technologies, Stockholm, Sweden),by incubation for 30 min at 37°C, 5% CO2 in conditioned medium containing fluo-4 AM (2 mM) mixed with pluronic acid (final concentration 0.2‰). Subsequently, dye-containing medium was removed and 10 min of de-esterification in fresh medium was performed before measurements began. The cover slips were mounted in a chamber with conditioned medium (37°C). Regions with cells showing contractions or calcium transient activity were identified using an HC Plan 10X, fluorescent lamp, Leica DM IRHC RF (Leica Microsystems, Wetzlar, Germany). Time-lapse imaging was performed using a Leica DM IRBE inverted confocal laser-scanning microscope equipped with a 40×/1.25 epifluorescence oil immersion objective (Leica Microsystems). Fluo-4 was excited at wavelength 488 nm. Images were collected every 1.7 s, corresponding to a Nyquist frequency of 294 mHz. Consecutive images collected every 1.7 s (and in some cases every 0.7 s) for 30–40 min were recorded for each experiment during two consecutive recordings. There was a time break (2–5 min) between the first and second recording. After 5 min control recording, antibody or vehicle containing medium was added (10–50–100 µg/ml). Warm fresh cardiomyocyte medium (500 µl) was added during recording to maintain pH and temperature.
Ca2+ data analysis
Fluorescence intensity of the Ca2+-sensitive dye was measured in four to six regions of oscillating cells in each experiment, using LCS Lite ver. 2.61 Build 1538 (Leica Microsystems Heidelberg GmbH) and subsequently analysed using Excel (MS Office 2007) and ORIGIN 8 (OriginLab). Median intensity of fluorescence between transients (Fm) was measured during control and every 5 min after application of antibody or PBS.
Statistical analyses
A Mann–Whitney U test or Kruskal–Wallis test followed by Dunn's post-tests (multiple comparisons) were used.
Results
Transfer of anti-Ro52 monoclonal antibodies during gestation induces AV block in rat pups in a dose- and time-dependent manner
To identify the specificity of heart block-inducing antibodies we developed an animal model of CHB using monoclonal antibodies generated against the human Ro52 protein.15 As antibodies to the Ro52 p200 peptide have been associated with the development of CHB both in humans and in a rat immunisation model,11 we selected a p200-specific monoclonal antibody (7.8C7) to establish our experimental system. Four milligrams of the antibody or vehicle were injected intraperitoneally into female DA rats (2 mg at days 6 and 9 of gestation, respectively) and the PR interval was measured on ECG performed on newborn pups. Pups exposed to the anti-p200 monoclonal antibody had a significantly longer PR interval than pups born to mothers receiving vehicle only (figure 1A). In addition, the mean heart rate of pups exposed to the anti-p200 antibody was significantly lower than that of control pups (figure 1B).
Induction of atrioventricular block in rat pups by passive transfer of an anti-Ro52 monoclonal antibody. Pregnant rats were injected with anti-Ro52 monoclonal antibody (7.8C7) specific for amino acids 200–239 (p200) or vehicle (PBS). An ECG was performed on pups within 24 h of birth. (A, B) Females were injected with 4 mg antibody or vehicle (2 mg at days 6 and 9 of gestation). PR interval (A) and heart rate (B) of 14 (vehicle) and 19 (7.8C7) pups (two litters per group). (C) Females were injected with a total of 6 mg of antibody or vehicle with the first injection at the indicated day of gestation and the second and third injections 2 and 4 days later. PR interval of nine (vehicle), eight (7.8C7/d6) and 14 (7.8C7/d9) pups (one or two litters per group). (A–C) PR values are shown as box plots (25th to 75th percentiles), line at median and whiskers showing min and max. Statistical difference was assessed by Mann–Whitney U test (A,B) and Kruskal–Wallis test followed by Dunn's post-tests (C). (D) Females were injected with the indicated amount of antibody (first injection at day 6 of gestation). PR interval (mean±SEM) of 9 (0), 18 (2 mg), 8 (4 mg) and 8 (6 mg) pups (one or two litters per group). PBS, phosphate-buffered saline.
In humans, the fetal AV block consistently develops during weeks 18–24 of gestation, and to understand the relevance of gestational age for developing AV block after antibody exposure in our model we varied the time point during pregnancy at which the antibody was administered. Pups born to mothers receiving the first antibody injection at day 6 of gestation had significantly longer PR intervals than pups of females injected with vehicle, while pups born to mothers who were given the antibody for the first time at day 9 of gestation had PR intervals comparable to those of the control group (figure 1C).
We also analysed the amount of antibody required to induce AV block in the pups. Decreasing the dose of antibody injected into the mother by half led to an average PR interval in the offspring in the same range as that of pups born to mothers receiving vehicle only (figure 1D). Increasing the dose of antibody did not enhance disease phenotype—that is, it did not prolong the PR interval or affect development of a higher degree of AV block.
Antibodies to the p200 domain of Ro52 induce AV block in rat pups
To determine whether anti-Ro52 antibodies directed to epitopes outside the p200 peptide also have the ability to induce heart block we first screened the panel of our monoclonal antibodies for their binding specificity to various epitopes spanning the Ro52 protein (figure 2). In addition to antibodies specific for the p200 epitope (7.2H4, 7.8C7, 7.8H4, 7.9D3), we identified additional antibodies that bound to the zinc finger domain (7.13F2) or to the C-terminal domain of the protein (7.2F4). These antibodies were injected into pregnant female rats and the pups were analysed for development of AV block. Strikingly, 100% of pups born to females that had received anti-p200 antibodies developed first-degree AV block and had significantly longer PR intervals than pups exposed to antibodies targeting the N- or C-terminal domains of Ro52 (figure 3A). The average PR interval of the latter two groups did not differ significantly from the PR interval measured in pups of mothers injected with vehicle only. Serum analysis of mothers and pups confirmed a similar reactivity to Ro52 in all mothers (figure 3B) and a similar placental transfer of antibodies in all groups (figure 3C), confirming that the observed pathogenic differences relate to antibody specificity rather than level or differences in the antibody transfer. Additionally, pups exposed to anti-p200 antibodies had significantly lower heart rates than pups exposed to antibodies targeting the N- or C-terminal domains of Ro52, with 81% of pups developing bradycardia (figure 3D). The mean heart rate of pups exposed to non-p200 anti-Ro52 antibodies did not differ significantly from that of the control group.
Domain specificity of anti-Ro52 monoclonal antibodies. (A) Recombinant Ro52 protein constructs. (B) The specificity of the anti-Ro52 monoclonal antibodies for the different parts of Ro52 was determined by ELISA using the constructs described in (A). aa, amino acids.
Antibodies to amino acids 200–239 of Ro52 (p200) but not antibodies to other parts of the protein induce first-degree atrioventricular (AV) block in rat pups. (A) Female rats were injected with 4 mg of anti-Ro52 monoclonal antibodies, all of IgG1 isotype, targeting the indicated part of Ro52 or vehicle (PBS), and ECG was performed on pups within 24 h of birth. Dashed line represents the threshold for first-degree AV block. PR interval of 16 (PBS), 18 (RING/B-box), 68 (p200) and 14 (B30.2) pups (two or three litters per antibody group). (B, C) Reactivity to Ro52 of serum from female rats (B) before transfer of monoclonal antibodies (−) and after delivery (+) and from pups (C) was determined by ELISA. (D) Heart rate of pups analysed in (A). Statistical difference was assessed by Kruskal–Wallis test followed by Dunn's post-tests (A, D). PBS, phosphate-buffered saline.
All four anti-p200 monoclonal antibodies tested here induced heart block in the offspring in a similar manner, despite slightly different fine specificities, as shown by their different reactivities to mutated p200 peptides (figure 4). The p200-specific antibodies bind equally well to the A233 mutant as to p200, are heterogeneous in their reactivity to the pZIP peptide, but binding of all the heart block-inducing antibodies was abolished by mutations in the pOUT peptide. This indicates that the amino acids mutated in pOUT are critical to the epitope recognised by the pathogenic anti-p200 antibodies.
Fine specificity of anti-Ro52 p200 monoclonal antibodies. Reactivity to the p200 and mutated p200 peptides was determined by ELISA. Mutated amino acids in A233, pZIP and pOUT are indicated.
Anti-p200 antibodies disturb calcium homoeostasis of cultured cardiomyocytes in a dose-dependent manner
To investigate whether anti-Ro52 antibodies have a direct effect on cardiomyocyte function, we established primary neonatal cardiomyocyte cultures and analysed cellular calcium oscillations as an indicator of cardiomyocyte pacemaker activity and contractility.
Addition of an anti-p200 antibody (7.8C7) to spontaneously oscillating cardiomyocytes altered the frequency of calcium oscillations and induced progressive intracellular calcium accumulation (figure 5). These effects were dose-dependent as decreasing the amount of antibody added to the cultures by 10-fold reduced the effect on calcium homoeostasis and delayed its onset. Cardiomyocytes exposed to a non-p200 anti-Ro52 antibody (7.13F2) did not exhibit changes in calcium transients or calcium level at any of the doses tested (figure 5), indicating that a direct effect of anti-Ro52 antibodies on cardiomyocyte function may be limited to antibodies specific for the p200 peptide, thus accounting for their pathogenicity.
p200-specific antibodies dysregulate calcium homoeostasis in a dose-dependent manner. (A) Rat neonatal cardiomyocytes loaded with calcium-sensitive dye (Fluo-4) before addition of antibody or PBS. (B) Single cell tracing of [Ca2+]i fluorescence. Addition of the respective antibody (50 µg/ml) or PBS is indicated by a red line. (C) Quantification of change in baseline intracellular Ca2+ levels over time. (D) Same cardiomyocytes as seen in (A), 25 min after antibody or PBS addition. (A, D) Scale bar=50 µm. (A, B and D) Fluorescence intensity in arbitrary units (a.u.), 0 (dark red)–250 (blue) where intensity of fluorescent calcium signal is saturated. (C) Pooled results from 12, 10 and 3 independent experiments for 7.8C7, 7.13F2 and PBS respectively. PBS, phosphate-buffered saline.
Discussion
In this report we show that first-degree AV block can be induced in rat pups upon transfer of anti-Ro52 p200-specific monoclonal antibodies. In contrast to an immunisation-based model of heart block, where only 20% of pups born to rat mothers immunised with the Ro52-derived p200 peptide developed AV block,11 all of the pups exposed to specific anti-Ro52 p200 monoclonal antibodies in utero following intraperitoneal antibody injection in the mothers during pregnancy developed AV block. These results emphasise the importance of the fine specificity of the heart block-inducing autoantibodies, as immunisation with the 40 amino acid p200 peptide induces variable fine specificity in the maternal immune response,7 while the transfer model using Ro52 monoclonal antibodies presented herein ensures a uniform exposure to autoantibodies with a single specificity. A passive transfer of monoclonal antibodies therefore provides a simple and reliable model of heart block with 100% penetrance, and will be useful to investigate further the influence of the genetic make-up of the fetus in the development of more advanced stages of heart block.
While monoclonal antibodies targeting the p200 epitope of Ro52 proved highly pathogenic, antibodies directed towards epitopes in the N- or C-terminal part of Ro52 did not induce heart block upon transfer during gestation nor did they disturb the calcium homoeostasis of cultured cardiomyocytes. These data indicate that anti-Ro52 p200 antibodies are sufficient to induce AV block in vivo, whereas antibodies targeting other parts of the Ro52 protein are not, suggesting that the p200-specific antibodies may exert their pathogenic effects by cross-reacting with a molecule other than Ro52, expressed on fetal cardiac cells. Calcium channel subunits have been reported as cross-reactive targets for maternal anti-Ro/La antibodies,13 18 and IgG purified from mothers of children with CHB have been shown to inhibit L-type and T-type calcium currents in ventricular myocytes as well as in sinoatrial node cells and exogenous expression systems.13 18,–,21 Cross-reactivity of anti-Ro52 antibodies to the 5-hydroxytryptamine cardiac receptor has also been reported,14 though mouse pups born to females immunised with Ro52 peptides that had been selected on the basis of recognition by anti-5-hydroxytryptamine receptor 4 antibodies did not develop any sign of AV block or other cardiac dysfunction.22 In view of the effect of anti-p200 monoclonal antibodies on the calcium homoeostasis of cultured cardiomyocytes seen in this study, it is likely that a cross-reactive target would be a cardiac membrane protein involved in calcium homoeostasis and control of electric signal generation and/or conduction.
Anti-p200 antibodies led to AV block in the offspring only when transferred to the mother before day 9 of gestation, suggesting that the first 8 days of fetal development in the rat constitute a critical window where the heart is vulnerable to the pathogenic effects of maternal antibodies and indicating that expression of a cross-reactive target may be confined to, or at a peak, during this time. These results mirror the situation in humans, where the fetal but not the maternal heart is affected by anti-Ro52 autoantibodies.
Decreasing the dose of transferred antibody by half abolished development of AV block, while increasing it by twofold and even fourfold (data not shown) did not induce further PR prolongation or a higher-degree block. These data suggest that the observed dose effect may be a ‘threshold effect’ where a certain level of the antibody in the fetal circulation has to be reached in order to induce first-degree heart block, but that development of complete AV block requires additional factors. Alternatively, it is possible that in our model the last injection of antibody made at day 11, although effectively increasing antibody levels in the fetus, did not induce any further damage in the heart because occurring too late in the developmental process of the fetal conduction system.23
Despite a 100% penetrance of first-degree AV block in newborn pups following in utero exposure to anti-p200 antibodies, no second- or third-degree AV block was seen. Progression to a complete third-degree AV block has been suggested to depend on fetal genetic susceptibility factors, as the recurrence rate in pregnancies after a CHB pregnancy is only 12–20% despite persisting maternal autoantibodies, and differing susceptibility to CHB depending on fetal MHC genes has been identified in a rat model.5 7 It will therefore be interesting to investigate whether it is possible to achieve induction of higher-degree blocks by passive transfer of anti-p200 antibodies in other rat and mouse strains or genetically modified mice that probably differ in their susceptibility to autoantibody-induced CHB. It is also possible that the single specificity of the transferred antibody, although sufficient to initiate first-degree AV block, cannot support the full-blown inflammation in the heart that eventually leads to fibrosis and complete heart block.3 Autoantibodies to Ro52 are more commonly found than anti-Ro60 and anti-La antibodies in mothers of children with CHB,8 24 suggesting that anti-Ro60 and anti-La antibodies may not play a major role in the initiation of CHB. The involvement of anti-Ro60 and anti-La antibodies in CHB pathogenesis will, however, be worth investigating in further studies, using the monoclonal antibody transfer model described here. In particular, it will be interesting to assess whether concomitant administration of anti-Ro52 p200 antibodies and anti-Ro60 antibodies will induce higher-degree blocks, as anti-Ro60 antibodies may contribute to amplifying the inflammatory reaction taking place in the fetal heart and leading to complete AV block.25
In summary, our results suggest that anti-Ro52 p200-specific antibodies are the main initiators of heart block development, while non-Ro52 p200 autoantibodies may subsequently amplify the inflammatory processes, leading to the establishment of heart block. In addition, the model that we have developed provides a simple and reproducible technique of inducing first-degree AV block by passive transfer of monoclonal antibodies specific for Ro52, which will enable understanding of genetic and pathogenic mechanisms involved in the development of CHB.
References
Footnotes
AA and VD contributed equally.
-
Competing interests None.
-
Funding Supported by grants from the Swedish Research Council, the Heart-Lung Foundation, the Stockholm County Council, Karolinska Institutet, the Swedish Rheumatism association, the King Gustaf the 5th 80-year foundation, the Göran Gustafsson Foundation and the Torsten and Ragnar Söderberg Foundation.
-
Provenance and peer review Not commissioned; externally peer reviewed.