Human enteroviruses are members of the Picornavirus family, which includes polioviruses, group A and group B Coxsackieviruses, echoviruses and new enteroviruses types 68–71. The virus particles are 28–30 nm in diameter with icosahedral symmetry, and consist of a single-stranded positive-sense RNA of 7,400–7,500 bases and a protein capsid. Viral proteins are synthesized initially as a polyprotein, cleaved subsequently to produce four capsid proteins VP1, 2, 3, 4 and nonstructural proteins including viral proteases. The outer surface of the virion consists of VP1, 2 and 3, whereas VP4 is located in the interior of the capsid. The epitopes for neutralizing antibodies are located predominantly in VP1 and are serotype specific. This protein is potentially the basis of future vaccines [18]. There are enterovirus group-common epitopes in VP1, not displayed on the surface of an infectious virion: antibodies against these determinants are heterotypic and non-neutralizing [20].

Enteroviruses are common and important pathogens of humans. In the United States alone, enteroviruses are estimated to cause 5–10 million symptomatic infections annually [28] and infection is associated with over 20 clinically recognized diseases and syndromes. This group of virus has a particular role in myocardial diseases, accounting for half of myocarditis cases [18,32]. Persistence of these viruses in myocardium is associated with the progression from acute myocarditis to dilated cardiomyopathy (DCM), a chronic heart muscle disease with impaired or compensated cardiac function. The association between enterovirus infection and myocardial disease has been based on virus isolation, serological studies, nucleic acid hybridization and reverse transcription-polymerase chain reaction (RT-PCR) [17]. Virus isolation from myocardial tissue is rarely achieved; retrospective serological evidence of enterovirus infection does not prove cardiac involvement; in situ hybridization can localize viral RNA in the involved heart [7,11,12] but requires particular technical skills and, unlike RT-PCR, is not routinely performed in many laboratories. However, RT-PCR has generated some disparity in the detection of enteroviral RNA in heart [4,6,15]. Previously, demonstration of enteroviral antigens in fixed tissue samples was limited [32], but an enterovirus group-specific monoclonal antibody (mAb) 5-D8/1 is now commercially available. This antibody reacts specifically with a conserved, non-neutralizing epitope in the capsid protein VP1 of a panel of enterovirus serotypes tested [30,33]. It does not cross-react with other RNA viruses, e.g., rhinoviruses, measles virus, or with human tissue. We have recently utilized this mAb to optimize an immunohistochemical technique for detection of VP1 in formalin-fixed and paraffin-embedded tissue samples from patients with myocarditis or DCM [13,14,34]. VP1 was readily seen and localized in cardiomyofibers adjacent to necrotic tissue and inflammatory infiltrates. VP1-positive myocytes were distributed focally or over a large area of myocardial sections. Other organs, e.g., the liver and lungs, were affected in cases of fatal myocarditis. VP1 was also detected in six of eight end-stage DCM cases requiring cardiac transplantation [13]. The enteroviral antigen detected by immunostaining in these DCM cases was present in scattered myocytes/myofibers: in some cases, only a few myocytes were positive in an entire section. VP1 was not present in myocardial tissue from the comparison group. In the present study we employed this technique to investigate a new and larger patient series of histologically proven myocarditis, healing/borderline myocarditis or DCM to provide new data on prevalence of enterovirus persistence in heart muscle disease. None of these cases had been investigated in our previous studies.

Patients and tissue samples

Formalin-fixed, paraffin-embedded endomyocardial biopsy, explanted or autopsy tissue samples from 136 subjects were collected randomly from King’s College Hospital, Harefield Hospital, Royal Brompton Hospital and entered into this retrospective study. These comprised 10 histologically proven cases of acute fatal myocarditis, a DCM group consisting of 89 cases, and a comparison group of samples from 37 unused donor hearts and cases with other conditions, including sarcoid, hypertension or malignancy. Of the 89 DCM cases, endomyocardial biopsy of the right ventricle was performed in 48 patients (including 10 patients with healing/borderline myocarditis) attending Cardiology Clinic in King’s College Hospital. Three samples from different areas of the right ventricular myocardium were fixed in the neutral buffered-formalin immediately for histological and immunohistochemical analysis. Explanted ventricular tissue samples were obtained from 41 patients who underwent cardiac transplantation in Harefield Hospital. Autopsy ventricular tissue samples were obtained from 10 cases of acute myocarditis. Samples were fixed similarly. Consent was obtained from patients or their relatives and approved by the Local Research Ethic Committee. Specimens were not studied chronologically but were removed from storage and analyzed randomly by a single operator, unaware of the patients’ clinical details. These specimens had not been investigated for enterovirus involvement using other detection methods.

Histological analysis

Tissue samples were examined and diagnosed as acute myocarditis, healing/borderline myocarditis, DCM or specific heart muscle disease according to Dallas criteria [3,23]. The changes in DCM are nonspecific, consisting of hypertrophy of myocardial fibers, thickening of the endocardium, myocardial dilatation and a variable degree of interstitial fibrous replacement.

Improved immunohistochemistry

Enterovirus group-specific mAb 5-D8/1 (IgG2a), normal mouse IgG2a, blocking reagent, antibody diluent and detection system, EnVision/HRP, were purchased from Dako Ltd (Cambridge, UK). VP1 immunohistochemistry was performed as described previously [34]. Briefly, paraffin tissue sections were deparaffined and rehydrated. Heat-mediated antigen retrieval was achieved by boiling tissue sections in citrate buffer (pH 6.0) in a microwave oven. Primary antibody mAb 5-D8/1 was diluted according to manufacturer’s instructions or titrated by the user. Application of polymer/peroxidase-conjugated secondary antibody (EnVision) was performed according to published protocols [26]. As a control, the primary antibody was omitted or replaced with diluent only or subclass- and concentration-matched mouse IgG2a. Additionally, known enterovirus-positive or -negative myocardial sections, measured by RT-PCR and immunohistochemistry, from a case of fatal myocarditis and an accidental death were processed similarly as positive and negative controls respectively.

Statistics

Statistical analysis was carried out by an independent Medical Statician. Yates corrected χ² was used to calculate the significance of the difference between two groups.

Presence of enteroviral VP1 in myocardial tissue samples

Histological sections from 7 of 10 fatal myocarditis cases, 47 of 89 patients (52.8%) with DCM (including healing/borderline myocarditis) were positive for the viral capsid protein VP1 by immunohistochemical staining (Table 1). Consecutive sections of positive samples were negative when the primary mAb was omitted or replaced with diluent only or subclass and concentration-matched normal mouse IgG2a. In contrast, only 3 of 37 samples (8.1%) in the comparison group were positive (Yates corrected χ²=19.99, P<0.001: odds ratio =12.68). Positive control sections from proven enteroviral myocarditis and negative control sections of an enterovirus-negative heart from an accidental death gave the expected results. Endomyocardial biopsy samples from 5 of 10 patients (50%) with healing/borderline myocarditis and from 18 of 38 DCM cases (47.4%) were positive for VP1, making an overall positive rate of 47.9% in biopsy samples. Explanted myocardial samples from 24 of 41 DCM patients (58.5%) were positive for VP1. There was no statistically significant difference (P>0.05) in the VP1-positive rate between patients attending the cardiology clinic (biopsy samples) and patients with end-stage DCM who underwent cardiac transplantation (explanted samples).

Localization and distribution of enteroviral capsid protein VP1

Signals of VP1 immunostaining were localized in the cytoplasm of myocytes, but not seen in the nuclei (Figs. 1,2). This is consistent with the established knowledge and observation made by electron microscopic approach [12]. The VP1 staining was usually distributed in individual or grouped myofibers in myocardial tissue sections (Figs. 1B,2B, D). In some cases VP1 was detected in only a few myofibers within the entire section of a biopsy (Fig. 2C) or explanted tissue (Fig. 2E) specimen. This is in contrast to the intense extensive immunostaining of VP1 in myocardium from acute fatal myocarditis (Fig. 1) [13,14].

Detection of enterovirus capsid protein VP1 in fatal myocarditis by immunohistochemistry. Serial sections of an autopsy sample from a case of acute fatal myocarditis were stained with hematoxylin and eosin (A) or immunostained with either mAb 5-D8/1 (B) or normal mouse IgG2a (C). Yellow brown deposits represent specific detection of VP1 seen in (B), and (C) shows a negative result. B, C Counterstaining with hematoxylin. A–C ×100)

Full size image

Detection of enterovirus capsid protein VP1 in myocardial tissue sections by immunohistochemistry. A, B Serial sections of an endomyocardial biopsy sample from a patient with healing myocarditis immunostained with either normal mouse IgG (A) or mAb 5-D8/1 (B). Yellow brown deposits represent specific detection of VP1 seen in B. C1, C2 Viral antigen is present in only a few myofibers in the entire section of a biopsy sample from a patient with DCM. D Scattered immunostaining of VP1 in individual or grouped myofibers in an explanted heart of DCM. E Viral antigen is localized in cytoplasm of myocytes (arrows) in an explanted heart sample from a DCM case (DCM dilated cardiomyopathy). A–E Counterstaining with hematoxylin. A, B, C1, D, E ×100; C2, boxed areas ×600

Full size image

The main advantages of this improved immunohistochemical technique include its capacity of detecting a wide range of enterovirus serotypes, its high sensitivity to detect persistent infection and that it can be performed on fixed tissue samples allowing investigation of archival materials. It offers a useful addition to investigation and laboratory diagnosis of enterovirus infection, especially enteroviral heart disease. The success of this technique depends on the availability of mAb 5-D8/1 and the nature of the VP1 epitope recognized by this antibody. The 5-D8/1 epitope has been mapped to residues 40–48 at the N terminus of VP1 [27]. This region of the protein does not form an exposed domain on the virus capsid [19]. Sequence alignment shows that this is a highly conserved region among serotypes of enterovirus. Experimentally, 5-D8/1 was immunoreactive with 36 enterovirus serotypes tested [30] and reacted with VP1 of CVB3 expressed in either eukaryotic or insect cells in Western blotting. The epitope was recognized by mAb 5-D8/1 in formalin-fixed, paraffin-embedded sections after heating, but not after proteinase digestion [34]. These results suggest that this is a linear or non-conformational epitope and its exposure is heat mediated, explaining why detection of enterovirus capsid protein antigens in the fixed myocardial tissue was rarely achieved previously. The EnVision detection system was employed to increase the signal/noise ratio and enhance the sensitivity of VP1 detection [26]. In this system, the second antibody is coupled to a dextran backbone to which many peroxidase enzyme molecules are bound, amplifying the signal. This is particularly useful in investigation of persistent enterovirus infection in chronic myocardial disease, e.g., DCM where VP1-containing cells can be sparse.

Enterovirus VP1 was detected in a high proportion of cases of acute or fatal myocarditis or heart muscle disease with a history of enteroviral infection in several previous reports involving relatively small numbers of cases [8,13,14,34]. The current study provided an opportunity to investigate the prevalence of enteroviral VP1 in samples from an unbiased patient group with DCM. The results show that VP1 was detected in 7 of 10 cases of acute fatal myocarditis and in a high proportion of DCM cases. VP1-positive rate in the biopsy samples (47.9%) is not significantly lower than that in explanted tissue samples (58.5%), although there are clinical differences in disease severity between these two groups. These data might suggest a trend that the VP1-positive rate in end-stage DCM cases requiring cardiac transplantation could be higher, implying an association between the presence of enterovirus and poor prognosis of DCM. The presence of enteroviral RNA was shown previously to be an independent predictor of poor prognosis in DCM [31], although another study did not support this [16].

Previous studies have demonstrated the persistence of enterovirus-specific IgM or VP1 antigen in the serum and of enterovirus genomic or template strand RNA in chronic heart or skeletal muscle disease [2,22,24,31], suggesting the possibility of synthesis of viral proteins during virus persistence. In the present study, detection of viral capsid protein in different disease stages from acute myocarditis, healing/borderline myocarditis to DCM or end-stage DCM requiring cardiac transplantation provides direct evidence of expression of enteroviral proteins during persistent infection, which may facilitate the progression of viral myocarditis to DCM. They might also assemble limited number of progeny or more likely defective virus particles. Antibodies directed against enterovirus group-common epitopes in capsid protein VP1 can react with defective or inactivated virus [21]. Recent studies indicate that expression of the nonstructural proteins, proteases 2A and 3CD of CVB3, may impair heart functions by cleaving cardiac dystrophin and the p21^ras GTPase-activating protein RasGAP respectively, possible molecular mechanisms of pathogenesis of DCM associated with enterovirus infection [5,10]. Presence of VP1 in acute and chronic myocardial diseases also has implications in the immune response and antiviral therapy. There are dominant T cell and neutralizing epitopes in VP1, important for vaccine development [9,29]. A novel promising anti-enterovirus and other picornavirus drug, pleconaril blocks viral receptor binding and uncoating by binding to viral capsid in a hydrophobic pocket, formed by VP1, at the bottom of the ‘canyon’ [25].