Background

Myocarditis is best defined simply as inflammation of the heart muscle. Despite this simplistic definition, the diagnosis and treatment of myocarditis continue to present clinical problems. The more regular use of the endomyocardial biopsy has promoted greater understanding of the natural history of human myocarditis and prompted new efforts at classification. In 1991, Lieberman et al. proposed a classification based on analysis of the histologic findings of biopsies and clinical course of the patient [1]. They distinguished patients with fulminating myocarditis who become acutely ill after a distinct viral prodrome and have severe cardiovascular compromise, histologic evidence of active myocarditis, and left ventricular dysfunction. The disease in these patients either results in death or resolves spontaneously. Patients with acute myocarditis present with established ventricular dysfunction and show histologic evidence of active or borderline myocarditis. This form of disease may lead to complete resolution but sometimes progresses to dilated cardiomyopathy. Like active myocarditis, chronic active myocarditis also begins insidiously and results in left ventricular dysfunction but is most likely to become fibrotic and progress to dilated cardiomyopathy. Finally, chronic persistent myocarditis is characterized by continuing cardiovascular symptoms and a histologic infiltrate with foci of myocyte necrosis but without significant ventricular compromise. Other types of myocarditis can be recognized by their characteristic histologic features. They include giant cell myocarditis, a form with a poor outcome and high likelihood of death or cardiac transplantation [2]. Eosinophilic myocarditis also has a dire prognosis and is sometimes associated with prior exposure to drugs [3]. It may be part of a general eosinophilic syndrome. It is now possible to reproduce these forms of inflammatory heart muscle disease by experimental manipulations of mice or rats. These models are adding a great deal to our understanding of the pathogenesis of the disease.

Clinical Features

The clinical manifestations of myocarditis are highly variable, ranging from no symptoms to heart failure. Patients may present with a large variety of signs [4]. The major features include chest pain, arrythmias, embolic events, congestive heart failure, and sometimes cardiogenic shock. These clinical manifestations can be supported by electrocardiographic findings of ST-T wave abnormalities and atrial or ventricular arrythmias. Noninvasive methods like echocardiography or magnetic resonance imaging may be used to further the diagnosis. As the disease progresses, gallup rhythms and signs of congestive heart failure become increasingly evident. Sometimes evidence of pericarditis can be detected in the form of a pericardial rub.

More recently, efforts are underway to classify myocarditis and the subsequent dilated cardiomyopathy on the basis of etiology [5]. About 30% of cases of cardiomyopathy are inherited and are principally caused by genetic mutations in genes for critical cytoskeletal or contractile proteins of the cardiac myocyte [6].

Many microbial agents have been implicated as potential initiates of the myocarditic process [5]. Among the more common viral agents are Coxsackievirus, adenovirus, hepatitis C virus, influenza virus, herpes simplex virus, Epstein–Barr virus, cytomegalovirus, and parvovirus B19. Human immunodeficiency virus and vaccinia virus have also been cited in recent publications. In Central and South America, Trypanosoma cruzi, the causative agent of Chagas’ disease, is the most common cause of myocarditis.

Lessons about the etiology of human myocarditis can be gleaned from studies of its epidemiology and demography. The occurrence of myocarditis varies widely in different geographic areas with prevalence ranging from 1.06 to 10 cases per million [4]. The reason for these differences may reside in differing diagnostic criteria, but may also reflect differing types of infections, as well as genetic differences in susceptibility in the populations. Myocarditis and dilated cardiomyopathy are the major causes of heart failure and sudden unexpected death in young adults [7]. Athletes and others with similar heavy physical exertion appear to be particularly vulnerable to sudden unexpected death due to myocarditis For example, studies of military personnel in Israel, Finland, and the United States have all shown that this is an important problem [8–10]. Another group that appears to be particularly vulnerable are young women in the peripartum period.

Studies of the etiology of myocarditis and dilated cardiomyopathy have been facilitated by the availability of the endomyocardial biopsy [11]. These biopsies can be assessed for markers of genetic, immunologic or viral causes, are of great value in establishing the histopathology of lymphocytic, eosinophilic, or giant cell myocarditis and in predicting the long-term outcome [12]. An international team has recently reviewed in depth the advantages and risks of endomyocardial biopsy and provided carefully considered rationales for its use [13]. Looking to the future, it may well be that the use of cardiac biopsies will expand if they are shown to be the important tools in establishing the etiology and treatment of myocarditis and dilated cardiomyopathy.

Myocarditis as a Viral Disease

The recognition that myocarditis often follows viral infection is well supported by both clinical and laboratory observations [14]. The virus may produce direct myocardial damage or it may be responsible for secondary, virus-initiated myocardial injury. The strongest support for the viral etiology of myocarditis has come from studies of the endomyocardial biopsies. In a pioneering investigation, Kandolf and colleagues applied in situ hybridization in order to detect viral genome in myocardial cells [15]. Hybridization and the polymerase chain reaction can be used to detect persisting viruses both in endomyocardial biopsy specimens from patients and from experimentally infected mice. This method has been applied clinically in many centers to study biopsies from patients with myocarditis or dilated cardiomyopathy. It can be used to detect both viral RNA in the instance of enteroviruses or DNA from adenoviruses and has clearly shown that the myocytes of patients with myocarditis and dilated cardiomyopathy can harbor a great variety of viral agents [16]. The viruses can persist for long periods of time in infected mice [17]. They may be associated with myocyte necrosis, fibrosis, calcification, and cardiac dilation [18]. These observations strengthen the concept that viral myocarditis can progress to dilated cardiomyopathy [19]. The viruses such as Coxsackieviruses can act directly to produce myocardial damage by disrupting, for example, the dystrophin–sarcoglycan complex, or indirectly by activating the innate immune response [20–22]. In an effort to establish that the Coxsackieviruses are not merely passengers on damaged myocardial cells, recent experiments have been reported by Shi et al., demonstrating that mice deficient in the Coxsackievirus–adenovirus receptor do not develop myocarditis [23].

Based on these observations in patients and experimental animals, antiviral treatment strategies have been tested in humans [24]. A number of patients with biopsy-proven enteroviral infection have been treated with interferon-α or interferon-β and have shown a significant decrease in lymphocytic infiltration of the heart. In some patients, viral genome was no longer detectable after interferon treatment. About two thirds of the patients showed clinical improvement that persisted for as long as 12 months. Clinical trials of patients with evidence of persisting viral genome using interferon treatment is underway, as will be described later.

Autoimmunity in Myocarditis

The presence of autoantibodies is a prominent feature of all forms of inflammatory cardiomyopathy. These autoantibodies can be directed to a number of different antigens, some of which are specific for heart cells. Some autoantibodies have functional consequences. The most widely used method for demonstrating heart-reactive antibodies in human sera is indirect immunofluorescence, a method pioneered by Caforio and colleagues [25, 26]. These antibodies can be found in many forms of cardiac inflammation such as rheumatic heart disease, Dressler and pericardiotomy syndrome, or cardiac transplant rejection. Using rat heart tissue as substrate, IgG antibodies were reported to be present in 59% of patients with myocarditis and 20% of those with dilated cardiomyopathy and less than 5% of normal controls. They may localize a variety of different patterns signifying reaction with a number of different cardiac antigens. To distinguish these different antigens, Neumann et al. used western immunoblots and found that the most prominent antigenic target in high-titer cardiomyopathy sera is myosin heavy chain [27]. Subsequent detailed studies showed that some of the antibodies were specific for the cardiac isoform of myosin heavy chain, whereas others cross-reacted with myosin from skeletal muscle or brain. The cardiac-specific antibodies of the IgG class were most closely associated with cardiomyopathy. IgM antibodies that are not specific for cardiac myosin are frequently found in other inflammatory processes.

An important question is whether these antiheart antibodies can signal the initiation of an autoimmune response to normal heart antigens or whether they are a response only to damaged heart cells. Caforio et al., in a prospective familial assessment of dilated cardiomyopathy, found that cardiac-specific antibodies precede disease development in asymptomatic relatives of cardiomyopathy patients [28]. Thus, antibodies to cardiac myosin may have a certain prognostic value.

It is not known whether antibodies to cardiac myosin exert functional effect in vivo. Lauer et al. found that antimyosin autoantibodies correlated with deteriorization of systolic and diastolic left ventricular function in patients with chronic myocarditis [29]. A number of other autoantibodies that are features of inflammatory cardiomyopathies have functional actions in vitro [30]. They include antibodies against several mitochondrial antigens such as the adenine nucleotide translocator (ANT) and branched-chain alpha-keto acid dehydrogenase. Whether these antibodies have functional effect in vivo is still controversial. In one study anti-ANT antibodies reacted with calcium channel complex proteins of myocytes. They increase energy demands within the myocyte and caused subsequent cell death in vivo [31]. A group of antibodies that may have definite functional importance are directed to cell surface receptors. Antibodies to the β1 adrenoreceptor were first described by Limas, Goldenberg, and Limas [32]. The antibodies bind preferentially to sequences in the second extracellular loop of the β1 adrenergic receptors and exert an inhibitory effect [33]. Fu and colleagues also identified a functional epitope on the muscarinic (M) acetylcholine II receptor and found that 39% of patients with dilated cardiomyopathy had anti-MII antibodies compared with 7% of normal controls. For comparison, significant inhibitory activity attributed to anti-β1 adrenoreceptor antibodies of the IgG class was found in as many as 75% of sera from patients with dilated cardiomyopathy and 18% of sera from normal subjects.

Evidence of a functional role of circulating antibodies has come from clinical trials of ex vivo immune absorption [34, 35]. Immunoabsorption by anti-immunoglubulin G columns effectively eliminated total IgG and provided long-term therapeutic benefit in dilated cardiomyopathy. In contrast, absorption with protein A, which does not remove IgG3, does not induce improvement in the disease, showing that the functionally important antibodies are in the IgG3 subclass.

Recent studies have shown that patients with cardiomyopathy develop antibodies to troponin I, an early indicator of myocyte damage [36]. The potential for functional damage associated with this antibody is shown by experiments in mice. Immunization of susceptible strains of mice with recombinant troponin I induced severe inflammation followed by cardiomegaly, fibrosis, reduced fractional shortening, and 30% mortality. Immunization with troponin I also aggravated ischemia-induced cardiac injury in mice. There is, then, compelling evidence that autoimmunity plays an important and possibly critical role in the development of some forms of inflammatory heart muscle disease.

Lessons from Therapeutic Trials

The findings described above suggest that there may be distinct forms of inflammatory cardiomyopathy due to viral infection or to autoimmunity. Treatment of the disease based on etiology may be dramatically different. Frustaci et al., in a preliminary investigation, showed that it may be possible to distinguish these two etiologic forms of inflammatory cardiomyopathy [37]. He defined the virologic and immunologic profile of patients who responded to immunosuppressive treatment. He found that the nonresponders were characterized by a high prevalence of viral genomes in the myocardium and no detectable cardiac autoantibodies in their serum. Conversely, most of the responders were positive for antibodies, whereas, only a few of them presented viral particles in endomyocardial biopsies.

The present treatment of patients with acute myocarditis is designed to sustain life and restore cardiac output by reversing left ventricular failure. It usually depends on administration of angiotensin converting enzyme inhibitors, beta blockers, and diuretics. Most patients, especially those with fulminant or acute myocarditis, will improve and recover spontaneously. For patients with persisting impaired cardiac output or signs of critical arrythmias, further treatment is necessary. As pointed out by Liu and Mason, treatment of myocarditis may be approached as a viral infection, as an autoimmune disease, or both [38]. The use of type 1 interferons is a rational therapeutic approach to most cases of viral-mediated myocarditis. Other potentially effective antiviral agents include nucleoside analogs such as ribovirin. If regarded as an autoimmune disorder, myocarditis would be appropriately treated by immunosuppression or by removal of a pathogenic antibody.

The first large-scale, randomized, controlled clinical trial of immunosuppressive treatment for myocarditis was conducted by Mason et al. [39]. Using the strict Dallas criteria, patients received azathioprine or cyclosporin and prednisone or only standard immunosupportive therapy. The overall results did not show any advantage of treatment with immunosuppressive drugs; ventricular function improved, regardless of treatment. Nevertheless, long-term mortality remained high emphasizing the need for continued investigation.

Faced with these disappointing results for immunosuppressive treatment, Wojnicz et al. set out to distinguish patients with an immune-mediated disease from those whose disease may have an infectious etiology [40]. Using immunosuppressive treatment with steroids and azathioprine in patients with dilated cardiomyopathy, they were able to demonstrate a long-term benefit in patients with increased HLA expression in their cardiac biopsies. Frustaci et al. subsequently reported that patients with circulating antiheart antibodies, and no evidence of viral infection of the heart responded better to azathioprine and prednisone than patients with evidence of viral infection as described above [37, 41]. In a follow-up randomized placebo-controlled study, Fructaci and colleagues employed more extensive immunologic criteria including T cell infiltration in cardiac biopsies to define an immune-mediated disease [42]. Thirty-eight of 42 patients treated with immunosuppression improved when they fulfilled these selective criteria. The failure of response in 12% of cases suggest the presence of some viruses that were not screened for in the study. Based on these and other investigations, a large-scale multicenter clinical trial on treatment of cardiac inflammatory disease has been undertaken by European investigators. It will compare treatment with either prednisone and azathioprine for virus negative dilated cardiomyopathy, interferon (for enterovirus positive disease), or intravenous immunoglobulin (for cytomegalovirus, adenovirus, and parvovirus B19 dilated cardiomyopathy) [43]. Final results of this trial are not yet available.

The presence of antibodies to β1 adrenergic receptor in many patients with dilated cardiomyopathy as alluded to earlier. Functional studies have suggested that they might play a critical role in causing heart failure [44]. Based on these and similar considerations, Felix and Staudt tested nonspecific immunoabsorption as a potential therapy for dilated cardiomyopathy [45, 46]. In some experiments, immunoadsorption columns were used to remove all immunoglobulins, whereas, in other IgG subclass 3 was preserved. Not only did one course of treatment induce clinical benefits, but there was still hemodynamic evidence of improvement after 6 months. Since the treatment may result in reduction or removal of any circulating antibodies, it is not possible at this time to discern its precise antibody target.

It is possible that the long-term improvement, which was seen following immunoadsorption, may be related to the administration of intravenous immunoglobulins (IVIG) as replacement in some patients in the course of the trial [47]. High-dose immunoglobulin therapy potentially has both antiviral and immunmodulatory effects. It has been widely used for the treatment of dilated cardiomyopathy and myocarditis with many case reports of success [48, 49]. Yet, a perspective placebo-controlled trial of IVIG was carried out by MacNamara et al. [50]. The treatment did not effect improvement in left ventricular ejection fraction or functional capability in the short-term or even after 12 months.

Giant cell myocarditis presents a particularly urgent problem with respect to improved treatments. The disease effects primarily young, predominantly healthy adults and results in death or cardiac transplantation in 89% of the subjects. A clinical trial described by Cooper et al. [13] suggested that immunosuppressive therapies including cyclosporin, azathioprine, or both, but not corticosteroids alone, may prolong the time to transplantation or death in patients with giant cell myocarditis [51].

Future improvements in treatment for myocarditis and dilated cardiomyopathy will require a better understanding of etiology and pathogenesis. At the first level, it would seem important to distinguish infectious from autoimmune disease since the same methods of treatment will not be optimal for both forms of heart muscle disease. If the disease is primarily a viral infection, either actively underway or persistent, more specific antiviral drugs would likely be the drugs of choice. In an autoimmune disease, the optimal treatment would be specific for the disease-inducing, “pathogenic” antigen. The presence and specificity of autoantibodies is, generally, the first clue in identifying the most relevant antigens. Antibody, however, may serve only as a signal for cell-mediated disease. These issues are notoriously difficult to sort out in human subjects. The use of experimental models of disease will provide the guideposts that finally take us to a rational therapy.

Lesions from Experimental Models

Continuing studies of the pathogenesis and kinetic control of myocarditis in experimental animals has led to fresh insights that can guide future approaches to therapy. Our own investigations have recently been reviewed and will not be repeated in detail [5]. They were initiated by studying the most common of the virus agents of myocarditis, Coxsackievirus B3, in mice. We found that there were great differences in susceptibility to Coxsackievirus-induced myocarditis among different strains of mice [52]. Most strains of mice developed a severe acute myocarditis, but completely recovered. In a few strains of mice, however, particularly those on “A” background, the disease continued to be chronic myocarditis. In these strains of mice, antibodies to heart, which mimicked those described in human cases of myocarditis and dilated cardiomyopathy could be found [53]. The major cardiac target antigen recognized by the virus-induced antibodies was identified as cardiac myosin heavy chain, the same antigen targeted by human autoantibodies.

Purified cardiac myosin made from mice is capable of inducing heart-specific antibodies, at the same time an inflammatory disease of the heart muscle closely resembling chronic myocarditis in humans followed active immunization [54, 55]. Significantly, only those strains of mice genetically susceptible to the chronic myocarditis following Coxsackievirus infection responded to immunization with murine cardiac myosin. Based on these findings, we proposed that chronic myocarditis can be induced by cardiotropic strains of Coxsackievirus, as well as other viruses that infect the heart. The infection can initiate an autoimmune response producing ongoing myocarditis sometimes leading to dilated cardiomyopathy [56, 57]. Furthermore, this ongoing disease can be reproduced in the absence of virus by immunization with cardiac myosin. We suggested that the virus itself is responsible for mobilizing intracellular myosin from infected and injured myocytes, thereby providing an indigenous source of antigen and also the necessary bystander or “adjuvant” effect [58].

A noteworthy outcome of these studies is that the genetic susceptibility to postviral myocarditis in mice is mirrored by susceptibility to myocarditis induced by cardiac myosin immunization [59]. Although the genetics is quite complex, the major histocompatibility complex (H-2) plays an important modifying role in the severity of disease [60]. A whole genome search revealed that susceptibility is determined by an aggregation of several genetic traits acting in concert. Thus, there is a full spectrum of susceptibilities from marked resistance to high susceptibility, depending upon the number of these genes involved [61]. Some of the genes involved have been identified as controlling the normal regulation of the immune response, suggesting that autoimmune disease in humans may also be the result of a chance aggregation of certain alleles of normal immunoregulatory genes.

In most strains of mice, the pathogenesis of autoimmune myocarditis depends upon CD4 T cells [62]. Depending upon experimental conditions, CD8 T cells may either enhance or suppress the severity of disease [63]. In several mouse strains, the short peptide recognized by disease-inducing T cells has been identified [64]. In one strain of mice, the disease can actually be transferred by passive immunization with antimyosin antibody [65]. Perhaps in these mice, myosin is found not only within the confines of a myocyte, but also on the cell surface where it can be attacked by antibody. Complement contributes to the severity of disease, although it acts more with the T cell rather than in conjunction with antibody [66].

Accumulating evidence shows that there is a step-wise process leading from the initiation of an innate immune response by the viral infection to the final pathogenesis of dilated cardiomyopathy and heart failure. Within the first day or even hours after infection with Coxsackievirus, there is a shift in the cytokine profile of mice susceptible to subsequent development of autoimmune myocarditis. An early marker of heightened susceptibility is the production of elevated levels of IL-1β and TNF-α [67, 68]. These two cytokines are not only necessary but, when given to nonsusceptible mice, sufficient to induce the autoimmune form of myocarditis. When acting early in the immune response to virus or following immunization with myosin, IL-10 provides a downregulatory signal [69]. Other cytokines assume importance in directing the immune response. Afanasyeva et al. found that IL-4 actually promotes a severe form of eosinophilic myocarditis in highly susceptible A/J mice [70]. On the other hand, depleting interferon-γ leads to an especially severe myocarditis frequently followed by dilated cardiomyopathy and heart failure. IL-13 also exerts a protective effect in autoimmune myocarditis, whereas, IL-17 appears to promote disease [71]. Thus, it is the balance of cytokines during the entire course of a virus infection that determines whether a subject will recover completely following acute viral myocarditis or progress to an autoimmune myocarditis and eventually, to a life-threatening dilated cardiomyopathy.

Lessons learned from the studies of experimental autoimmune myocarditis in the laboratory can help to guide thinking towards future therapies. A treatment directed solely to the “pathogenic” antigen or even the precise antigenic determinant responsible for initialing the disease would be the ideal goal. As an autoimmune response evolves, however, additional epitopes represented on the initiating antigen are recruited. More important, other antigens, specific for the same organ, join the ongoing autoimmune response. Therefore, the target of antigen-specific immunotherapy expands during the course of disease. It seems likely, therefore, that a mode of treatment directed towards inducing tolerance to a single antigenic determinant or even a single molecule may not be effective when disease reaches its clinical manifestations. These problems will only be relieved when we can recognize disease in its earliest stages and intervene at that time with an effective but benign treatment [72].

Studies of the experimental disease in mice suggest a complementary approach. A limited number of cytokines is required for a virus infection to progress to pathological autoimmune disease. Inhibiting the critical cytokines, singly or collectively, may well arrest disease progress. Such an approach may be applicable to a disease like myocarditis in which there is continuing uncertainty over whether the pathology is due mainly to infection, to autoimmunity, or to a combination of the two.

Final Comments

Autoimmune myocarditis presents a remarkable opportunity to study the complex interaction between infection and autoimmunity. It is a unique example where the primary antigen responsible for progression from a self-limited viral infection to subsequent autoimmune disease has been identified. Virtually the same disease, chronic inflammatory myocarditis, can be induced by Coxsackievirus infection or by immunization with cardiac myosin or troponin I in the absence of virus. This model offers the possibility of finding markers that can distinguish these two pathogenetic processes in humans. In the animal model, the interplay of cytokine mediators of inflammation can be dissected from the beginning to the end of the process. The lessons learned from investigation of myocarditis in humans and animals will surely be applicable to many other autoimmune disorders.