Autoantibodies Against Cardiac Troponin I: Friend or Foe?

This editorial refers to ‘Autoantibodies against cardiac troponin I in patients with congestive heart failure’ by H.-D. Dungen et al., published in this issue on page 668. For more than a decade, elevated levels of troponin T and I have been well-established biomarkers for the diagnosis and quantification of myocyte injury, irrespective of the underlying cause, e.g. ischaemic or inflammatory cardiac damage.1–3 Troponins are constituents of the regulatory complex of the myofibrillar thin filament critically involved in myocyte excitation–contraction coupling, and thus predominantly localized in the sarcoplasm. Under physiological conditions, such intracellular proteins are not accessible to the immune system. Within the aforementioned pathological conditions, however, the situation may change. Following ischaemic or inflammatory injury, the cellular integrity of cardiomyocytes may be altered causing a release of troponins into the circulation. The troponin efflux may then trigger auto-reactivity against previously concealed antigenic portions of this endogenous protein.3 Autoantibodies targeting essential cardiac structural proteins like troponins (anti-cTn-aabs) might to some extent have the potential to affect cardiomyocyte function and cardiac remodelling.3,4 The circumstances, however, under which such antibodies access their targets remain to be determined. In this context, a recent clinical pilot study in patients with acute myocardial infarction indicated that such autoantibodies may in fact negatively impact on recovery of left ventricular (LV) function.4 On the other hand, in chronic cardiac disease, the same investigators5 and several other groups6,7 reported that the outcome of anti-cTnI-aab-positive patients was not inferior to aab-negative patients with ischaemic or dilated cardiomyopathy (DCM), and in one recent study was even superior in anti-cTnI-aab-positive patients with DCM.5 Hence, whether the development of anti-cTn-aabs exerts positive effects on cardiac function and remodelling or not, and whether such aabs are beneficial for patient outcome or not, is still under debate. The first clue for a potential pathophysiological relevance of troponin-directed immune reactions was provided in mice at the beginning of this century.8 Programmed cell death-1 (PD-1) receptor deficient engineered mice developed severe inflammatory cardiomyopathy and antibodies directed against a distinct protein localized in the cytoplasm and in the membrane of cardiomyocytes which were subsequently identified as cardiac troponin I (cTnI).9 Because in these mice, cTnI was found abundantly on the outer surface of t-tubules known to be critically involved in excitation–contraction coupling, it was suggested that anti-cTnI antibodies could affect the electrophysiological function of cardiomyocytes. Subsequent in vitro experiments revealed indeed that anti-cTnI antibodies bound to the surface of isolated cardiomyocytes and significantly augmented the voltage-dependent calcium-current, indicating that this pathomechanism might contribute to altered cardiac function.9 In 2006, a similar inflammatory cardiac phenotype was induced in genetically susceptible mice either by immunization with cardiac troponin I (cTnI), but interestingly not with troponin T (cTnT), or by transfer of CD4+ T cells from cTnI-immunised into naïve animals.10 In these experiments, high anti-cTnI-antibody titres correlated well with the severity of cardiac inflammation and produced a cardiomyopathic phenotype characterized by myocardial fibrosis, LV dilatation, and impaired cardiac function.10 Subsequent transfer of anti-cTnI mouse (auto)antibodies into naïve BALB/c mice also induced a cardiomyopathic phenotype in recipients,10 thus providing direct evidence for a causative relationship between anti-cTnI-aabs and heart failure.11 In addition, experimental myocardial infarction produced more severe cardiac damage in cTnI-immunized mice than in non-immunised controls, which is suggestive for an unfavourable role of anti-cTnI-abs—at least in ischaemic heart disease.10 From the latter model, we also learned that the severity of ischaemic cardiac damage in anti-cTnI-ab-positive animals appears to positively correlate with the myocardial expression of selected pro-inflammatory chemokines, which underlines the importance of the individual genetic background for the phenotypic expression of immune-mediated (heart) diseases. It is commonly accepted that this assumption holds true for both mice and man.12 However, when trying to transfer knowledge from animal models into clinical practice, one has to keep in mind that the pathomechanisms in these species are often quite different. For example, in the intact human heart troponins are believed to be restricted to the cytoplasm which makes it difficult to explain how human anti-cTnI-aabs access their intracellularly localized target-proteins. Following the first reports showing that circulating human anti-cTnI-autoantibodies (anti-cTnI-aabs) interfere with diagnostic assays in patients with acute coronary syndromes (ACS),13–15 a smaller clinical study conducted by Shmilovich et al.6 analysed the prevalence of such aabs in sera from patients suffering from idiopathic DCM or ischaemic cardiomyopathy (ICM) because leakage of cTnI into the circulation has been postulated as an indicator of cardiomyocyte damage and degeneration also in chronic heart failure (CHF). About 16% (5/32) and 18% (6/33) of the DCM and ICM patients, respectively, but none of 42 healthy subjects, had circulating anti-cTnI-aabs.6 However, in contrast to the previously mentioned mouse models, cTnI-reactive IgG fractions isolated from anti-cTnI-aab-positive patients did not recognize cultured (neonatal) rat cardiomyocytes nor did they have a measurable effect on the Ca2+ transients of the same cells.6 This finding was indirectly confirmed by Miettinen et al.,7 who assessed the clinical significance of cTnI-efflux and anti-cTnI-aabs in 95 DCM patients. Whereas cTnI-positive (cTnI > 0.01 ng/mL, n = 19/95) compared with cTnI-negative patients had significantly more dilated hearts, greater intracardiac volumes (assessed by gated single photon emission tomography), higher levels of vasoreactive peptides, and more clinical endpoints during a median follow-up period of 4 years (combined endpoint: cardiovascular death/heart transplantation/need for ICD implantation), the presence of circulating anti-cTnI-aabs (n = 15/95, 16% of the DCM patients) was neither associated with a change in LV functional parameters nor with the patients' clinical status or outcome.7 In contrast, in a recent pilot study on patients with ACS,4 the absence or non-development of anti-cTnI-aabs predicted improvement of cardiac function after acute myocardial infarction. Roughly 10% of the ACS patients had anti-cTnI-aab titres above the cut-off value, and these patients also had impaired recovery of cardiac function after ischaemic injury (assessed by cardiac magnetic resonance imaging). In this context, it was suggested that the development of anti-cTnI-aabs might be promoted by higher initial concentrations and/or longer release of cTnI in ACS patients.16 However, neither the results by Shmilovich et al.6 nor the data presented by Düngen et al.17 in the current issue of this Journal appear to support this hypothesis. In this context, the limitation that the latter groups studied patients with chronic (i.e. ICM) and not acute (i.e. ACS) ischaemic cardiac injury, may be negligible: in a recent report from Leuschner et al.,4 all anti-cTnI-aab-positive ACS patients (n = 10/108) remained aab-positive during 6–9 months of follow-up, and none of the initially aab-negative patients developed anti-cTnI-aabs in the course of the study. This has given rise to the assumption that anti-cTnI-aabs might pre-exist in some patients as a consequence of ‘silent’ cardiac cTnI-releases prior to the first acute (transmural) ischaemic event.3 This hypothesis might perhaps also explain the relatively high proportion of anti-cTnI-aabs found in the (apparently) healthy population (12.7%).18 In this issue of the Journal, Düngen at al. re-addressed the potential clinical significance of anti-cTnI-aabs in a subgroup of well-characterized elderly CHF patients previously included into the CIBIS-ELD study (138 out of 883 study participants, mean age 73 ± 6 years).19 The authors analysed retrospectively the relationship between anti-cTnI-aabs and disease severity before (baseline) and after titration of beta-blocker therapy—which, however, was achieved only in one-third (34.8%) of this study population. In parallel, serum cTnI and NT-proBNP levels were also measured. Regarding CHF aetiology, the analysed study population comprised more than 50% patients with hypertensive and ∼30% with ischaemic heart disease, whereas only 3% of the patients had DCM.17 Results were compared with those from 300 healthy controls. Of 134 patients, less than 9% (n = 12/134) were anti-cTnI-aab-positive at baseline and remained so until the end of a median follow-up period of ∼7 years. During this period, eight patients initially considered anti-cTnI-aab-negative developed anti-cTnI-aabs above the cut-off level, which was accompanied by a small albeit significant increase in serum cTnI levels (0.5 vs. 4.0 µg/mL). However, taking all anti-cTnI-aab-positive patients together, there was no association between cTnI-efflux (i.e. cTnI serum levels) and the development of anti-cTnI-aabs. Further, and from a clinical point of view even more important, anti-cTnI-aab titres at baseline and at follow-up did not correlate with changes in LV diameters and/or LV function (assessed by echocardiography) or with the severity of CHF (i.e. NYHA class), whereas cTnI and NT-proBNP serum levels correlated well with disease severity at baseline.17 After initiation and up-titration of beta-blocker therapy, NT-proBNP levels did not change substantially, whereas cTnI levels decreased significantly in the course of the study. The latter observation is not surprising and can be considered as a surrogate for cardiac stabilization in a CHF population comprising >50% patients with hypertensive and ∼30% patients with ischaemic heart disease after successful implementation of standard CHF therapy.20 Also, the finding that anti-cTnI-aabs do not appear useful for predicting disease severity and/or clinical outcome in this particular study population was to be expected: autoimmunity is thought to play a major role in idiopathic and/or inflammatory cardiomyopathy representing only about 3% of the patients analysed by Düngen et al.3,21 It is a generally accepted assumption that cardiac troponins are preferentially released after substantial (derived from extent and severity of) ischaemic or inflammatory cardiomyocyte damage,3 and then naturally decrease in the course of myocardial healing almost independent of initiation or up-titration of beta-blocking agents. However, in light of the recent clinical data from Dösch et al. suggesting a potential cardioprotective role of anti-cTnI-aabs in DCM,5 it is tempting to speculate that development of anti-cTnI-aabs might represent a surrogate for an appropriate (rather than inappropriate (!)) immune response to cardiomyocyte injury. Thus, the slight increase in anti-cTnI-aab titres observed by Düngen et al. in CHF patients after titration of beta-blocker therapy17 might indicate a kind of haemodynamic (and/or immunologic) stabilization of the underlying cardiac disorders.22 However, so far, a clinically relevant immuno-modulating effect of beta-blocking agents has never been substantiated in larger CHF trials,23 and a beta-blocker-induced increase in anti-cTnI-aab production and/or decrease in anti-cTnI-aab clearance appears rather speculative. Taken together, although the data by Düngen at al. presented here do not support the clinical relevance of anti-cTnI-aabs in human heart disease, further studies in larger and—with respect to CHF aetiology—better structured and/or more diseased study populations (again including a substantial proportion of elderly patients) are warranted in order to unravel the pathophysiological role of anti-cTnI-aabs in cardiac disorders.3 In this regard, a European multicentre diagnostic study with serial follow-up of patients after a first acute myocarditis or a first myocardial infarction has been launched just recently in order to assess Etiology, Titre-course, and effect on Survival (ETiCS) of distinct cardiac autoantibodies in human cardiac disease.12 Several European core laboratories will jointly study the hypothesis that both disorders may trigger autoimmune reactions leading to the formation of heart-directed aabs. Among other serum parameters, all study participants will be screened for aabs directed against cardiac membrane receptors (i.e. β1-adrenergic- and M2 acetylcholine-receptors),24,25 cardiac troponins,4 and a confined number of other potential cardiac antigens.26 The prospective study design will also enable an analysis of the components of the patients' immune system involved in the generation of cardiac aabs—including a search for pre-disposing genotypes.27 Once available, the results of ETiCS will contribute to a large range of diagnostic, pathophysiological, and prognostic issues in autoantibody-mediated heart disease. Insights are expected on cardiac events capable of triggering autoimmune processes and on the contribution of autoimmune processes to the initiation and/or progression of CHF.21,28 In particular, the results of the ETiCS study may help us to better understand whether the specific target of an autoantibody, or its titre and biological activity, or its conversion rate (persistence/clearance) relates to the complex processes of cardiac wounding and healing. Such knowledge might equally provide a solid basis for the development of novel and more specific therapeutic approaches targeting those cardiac autoantibodies that have been shown to be cardio-noxious.12 The ETiCS study is acknowledged by the German Competence Network Heart Failure (CNHF), and has equally been associated to the BMBF-funded Comprehensive Heart Failure Centre, University Hospital of Würzburg (CHFC). Research activities of the author are currently supported by the German Ministry of Research and Education (BMBF, MolDiag, project numbers 01ES0816, 01ES0901 and 01ES0902, and GoBio-1, project number 0315636). Conflict of interest: none declared.