key: cord-022439-8wy7rpqv
authors: DENMAN, A.M.
title: Viral Etiology of Polymyositis/Dermatomyositis
date: 2013-11-17
journal: Polymyositis and Dermatomyositis
DOI: 10.1016/b978-0-409-95191-2.50010-0
sha: 
doc_id: 22439
cord_uid: 8wy7rpqv

nan

A.M. DENMAN . . . for we all of us, grave or light, get our thoughts entangled in metaphors and act fatally on the strength of them.

George Eliot, Middlemarch

Any attempt to ascribe polymyositis and dermatomyositis (PM/DM) to a viral etiology must take into account what is already known about the pathogenesis of this group of disorders. First, it is a rare disease, and there are few pointers to a conventional infec tious etiology. Outbreaks of the disease have not been described, and the evidence for a seasonal onset is weak. Thus, theories based on a viral etiology must invoke a highly un usual host response to ubiquitous agents, the existence of uncommon viruses with the abil ity to provoke the disorder, or the possibility that virus infections operate synergistically with other factors. Second, the immunopathologic features of PM/DM strongly indicate the importance of immune mechanisms in the pathogenesis of most forms of the diseases [1] . Thus, advocates of a viral etiology are confronted with the same problems facing anyone seeking to implicate virus infections in autoimmune diseases in general. Finally, PM/DM often accompanies a variety of other inflammatory connective tissue diseases, and these associations must also be satisfactorily explained by any theory based on viral infec tion. Increasing understanding of the immunopathologic consequences of viral infections means that such points can be countered con-

The author would like to thank Mrs. Kathy Jameson for expert help in preparing the manuscript. Attempted isolation Electron-microscopic studies Viral probes Antiviral antibody titers Establishing clones from infiltrating T lymphocytes and screening their antigen specificities Exploring the antigen specificities of associated autoantibodies In vitro models of virus-infected muscle cells In vivo models of experimental polymyositis ceptually. However, one must concede that the evidence to date for a viral etiology is fragmentary. The available approaches for exploring this possibility are set out in Table  6 .1.

In recent years there have been several claims implicating specific virus infections in isolated patients with PM/DM. These reports have of ten been based on direct isolation of the agent or the ultrastructural appearances of the inflamed muscle. Coxsackieviruses have at tracted particular attention because of the relative ease with which these agents induce inflammatory disease in cardiac and striated muscle of mice. In addition, coxsackieviruses have been linked with myocarditis in human infants on epidemiologic grounds [2] . Struc tures resembling picornaviruses were de tected by electron microscopy in muscle biopsy from two patients who died with subacute dermatomyositis [3] . Similar particles were seen in the diaphragmatic and intercos tal muscles of an 11-year-old girl who suc cumbed to a chronic myopathy [4] . These authors also claimed that muscle suspensions induced cytopathic effects in culture of pri mary human amnion cells. Furthermore, the isolate reacted with antibody to coxsackievirus A9. Other viruses have been the subject of similar reports. McKinlay and Mitchell [5] described eight children with transient myositis associated with elevated serum levels of creatine (CK) following acute virus infec tions. Adenovirus type II was isolated from one patient and a parainfluenza virus from a second patient. PM complicated the clinical course of a patient with undoubted hepatitis caused by the hepatitis B agent since the diag nosis was confirmed by muscle biopsy and marked elevations of the serum CK level [6] . Hepatitis B surface antigen (HB s Ag) was de tected by immunofluorescence studies of the intranuclear and intracytoplasmic inclusions in the biopsy specimen of the affected muscle. Mikol et al. [7] have also described inclusion bodies in a muscle biopsy specimen obtained from a patient with long-standing PM; these workers also claimed to have cultured adeno virus type II from the biopsy. In addition to the foregoing reports there have been several reports based almost exclu sively on ultrastructural appearances (sum marized by Schiraldi and Iandolo [8] . Very few of these reports have been confirmed by isolating the virus suspected on ultrastruc-tural grounds. Nor indeed have there been many attempts to confirm the viral nature of the inclusion bodies by immuno-electron mi croscopic techniques or other specific stain ing methods.

Indeed, there are dangers in ascribing PM/DM to a viral etiology on the basis of ultrastructural changes in isolation. Katsuragi et al. [9] have emphasized this point by their description of histochemical studies of a mus cle biopsy specimen from an elderly man with PM. Some unidentified objects in the biopsy specimens were initially considered to be in clusion bodies resembling picornaviruses but were later shown to have the chemical charac teristics of glycogen. The most striking im pression from these reports is their scattered nature, implicating several different agents. There is a dearth of reports showing that agents can be consistently isolated and pas saged in vitro from patients with PM/DM. Nor have sequential studies of immune re sponses been appended indicating that pa tients with PM/DM attributed to a given agent mounted a cell-mediated or humoral immune response to that agent at the time of alleged infection.

There are diagnostic difficulties over some reports. Myalgia has often been reported in association with acute respiratory tract infec tions, and the symptoms may be accompa nied by electromyographic (EMG) abnormal ities [10] . An epidemic of acute myositis has also been described in children in association with influenza B virus infection [11] . Tran sient elevation of the CK level was noted in these children, and influenza B virus was iso lated from 11 of the 17 patients studied. How ever, the patients made a complete recovery in 4 to 5 days without specific treatment, and it is clear that postviral myositis of this nature is unrelated to persistent PM as generally en countered. Interestingly, the symptoms were almost exclusively confined to the gastrocnemius and soleus muscles. Acute myositis has also been described after rubella vaccination, although the accompanying clinical and patho logic features suggested that the muscles were involved as part of a more generalized Arthus reaction to vaccination [12] . The muscle dis orders most commonly provoked by viral in fections have usually taken the form of tran sient myalgia, in contrast to the more sustained character of idiopathic PM/DM. Furthermore, the muscle disorders provoked by most spe cific infections have also been focal rather than generalized and, therefore, clearly re lated to localized infection [13] . Muscle symptoms after influenza-like illnesses have also been described in older patients [14] , including objective muscle weakness and minimal EMG changes. However, the viral etiology of the antecedent illness was not con firmed, biopsy proof was not obtained, and there was no consistent rise in the serum CK level or the levels of other muscle enzymes.

Muscle problems in patients with immu nodeficiency deserve separate comment. Muscle inflammation attributable to echovirus infection presents a striking picture in patients with hypogammaglobulinemia [15] . The muscle inflammation is often unilateral, usually localized, and rarely involves the clas sic sites of muscle involvement in idiopathic PM/DM. The virus has usually been isolated from the cerebrospinal fluid rather than from involved muscle [16] . Local abnormalities are revealed by muscle biopsies and consist of perivascular mononuclear infiltrates and muscle fiber atrophy, occasionally accompa nied by extensive interstitial fibrosis. How ever, the disease picture in these unusual cir cumstances has not been shown to have a wider relevance.

There are additional reasons for doubting the relevance of most reports of this nature to the pathogenesis of PM/DM as commonly encountered. Ultrastructural changes of a kind likely to be caused by virus infection have not been reported in any systematic study of biopsy specimens from patients with PM/DM diagnosed by accepted criteria. Fur thermore, the epidemiologic evidence does not support the concept that exposure to common viruses is linked to the pathogenesis of PM/DM. Thus, influenza virus causes tran sient myalgia and has also been isolated from the muscle of a patient with myoglobulinuric PM [17] . Yet massive immunization with swine influenza vaccine during a 10-week period in 1976 was not followed by any in crease in the incidence of PM/DM [18] .

Most attempts to identify viruses in clini cal material obtained from patients with PM/ DM have depended on the presence of viruses that can either be recovered as fully infectious particles or recognized ultrastructurally. There is always the possibility that defective replication or latent infection might render the host cell the target for an immune re sponse and yet be undetectable by traditional methods. The powerful techniques now avail able for detecting viral nucleic acid sequences offer one obvious means of probing putatively infected muscle samples. This approach has been applied to the study of heart muscle from patients with myocarditis [19] . Coxsackievirus B has been implicated serologically in the etiology of myocarditis, but virus has not been convincingly detected in dis eased muscle by immunofluorescence or iso lated by conventional culture techniques. These authors used a complementary DNA (cDNA) copy of coxsackievirus B genomic RNA as a hybridization probe for identifying virus-specific sequences in myocardial biop sies. They claim to have obtained positive hybridization signals in nine of 17 samples from patients with myocarditis and related disorders. If confirmed and extended, these experiments will obviously be relevant to sim ilar attempts to identify viral sequences in biopsy specimens from patients with PM/DM. One can also speculate that latent infection might be more readily detectable if the in fected host cells could be cultured for periods of sufficient duration. Improving techniques for culturing differentiated muscle cells and for inducing the growth and expression of latent viruses makes this an important area for future research.

Antibody screening has been frequently used as indirect means of obtaining evidence for a viral etiology in PM/DM. Antibody titers to coxsackievirus B have received particular at tention because of the many experimental models in which this group of viruses induces myocarditis. This line of research has pro duced some suggestive but not conclusive evi dence for coxsackievirus infection in patients with cardiomyopathies. Cambridge et al. [20] encountered high neutralization titers to coxsackievirus B more commonly in patients with congestive cardiomyopathy than in con trols. Moreover, this occurred more com monly in patients with a short history than in patients with a febrile illness at the onset of their symptoms. Interestingly, endomyocardial biopsies did not produce evidence of direct viral infection. Although myositis and myo carditis occur most commonly as clinically distinguishable entities, cardiac involvement is often an important feature of idiopathic PM/DM. Thus, in one detailed study Haupt and Hutchins [21] found evidence of active myocarditis in 17 of 60 patients with PM or related diseases. Thus it is reasonable to sur vey antiviral titers to this group of viruses in patients with PM/DM in isolation. In a lim ited study of this problem Travers et al. [22] noted a marked rise of antibody titer to a specific serotype of coxsackievirus B in four patients with PM/DM without any rise in titers to other viruses. Detailed surveys of this kind have not been reported in PM/DM. Simi larly raised antibody titers to these viruses have been detected by complement fixation techniques in a preliminary study of children with juvenile PM/DM [23] .

The most informative studies of this kind have been those that have sought evidence for an IgM response to defined coxsackievirus strains, using enzyme-linked immunosorbent assay (ELISA) techniques, since this approach is more likely to detect serologic indication of recent infection. El-Hagrassy et al. [24] used ELISA techniques to detect coxsackievirus B-specific IgM responses in 37% of patients with acute myocarditis or pericarditis. In other areas of clinical investigation improved ELISA assays have also proved more discrimi nating. King et al. [25] detected IgM responses to a single coxsackievirus B strain in 39% of children aged 3 to 14 years with insulin-dependent diabetes mellitus; their findings also emphasize the important point that it may be easier to identify recent infec tions by defined viral strains in children than in adults since the latter can be expected to have had cumulative exposure to the entire range of these ubiquitous viruses.

Surveys of antiviral antibody titers of PM/DM have not provided any other useful clues to agents that might be implicated in the pathogenesis of the disorder. Occasionally, high antibody titers to specific agents have been described in isolated cases. Thus, John and Fink [26] noted a sustained rise in anti body titers to echovirus type 9 in one patient with PM. However, surveys of antiviral anti body titers to common viruses pose several problems in interpretation. First, antibodies If viruses induce PM/DM, the simplest mech anism by which they could induce this disease is by replicating in muscle cells, thereby serv ing as a direct target for an immune attack. Furthermore, ephemeral or latent infection of muscle cells could initiate an immune response subsequently sustained by other mechanisms. In a search for myotropic vi ruses, Klavinskis et al. [27] found that two strains of influenza A virus lyrically infected human synctial myotubes. In addition, one strain was found to infect unicellular precur sor myoblasts. These authors showed that viral proteins were synthesized in the infected cells and they also identified viral antigens on the surface of the infected cells by immunofluorescence and immuno-electron microscopy ( Fig. 6.1 ). There is little information about will be present in the majority of the popula tion. Rising antibody titers or the detection of IgM antibody against strains recently detected in the community might be helpful, but such surveys are difficult to organize, particularly given the rarity of the disease. Positive finding in cases drawn from widely disparate geo graphical areas would need extensive control studies in those same areas. Second, antibody titers to endemic viruses do not necessarily imply that those agents are pathogenetically related to the disease. This point is particu larly troublesome if the virus, such as mem bers of the herpes group, persists life-long in the human host; rising antibody titers may simply reflect nonspecific reactivation. Finally, atypical infections with unusual clinical con sequences may be encountered in patients with generalized immunodeficiency or impaired immunity specifically to the infecting agent. This immunodeficiency could be reflected in inappropriately low antibody titers; the com plications of echovirus infections in hypogammaglobulinaemic patients vividly illustrate this problem. the extent to which other viruses implicated in the pathogenesis of PM/DM grow in mus cle cells. Nor should interactions between viruses and muscle cells be analyzed exclu sively in terms of immune-mediated damage. There is also the possibility that PM/DM in duced by a putative infectious agent could result from a mixture of direct cytopathic changes and immune damage. For example, hemolytic paramyxoviruses induce perme ability changes in infected cells in addition to their ability to mediate cell fusion. Patel and Pasternak [28] have described membrane permeability changes in cell lines infected by influenza virus at low pH which persisted when the pH was shifted to the physiologic range. Persistent PM/DM is often character ized by degenerative changes in muscle fibers 

The suspected association between coxsackievirus infections and muscle disorders of man has prompted a particular interest in animal models of myocarditis and PM induced by this group of viruses. Moreover, it has proved relatively easy to induce inflammatory disorders of muscle with these agents. Coxsackievirus-induced myocarditis has been studied particularly intensively. The factors determining the outcome of such infections are set out in Table 6 .2. To some extent the outcome of infections by coxsackievirus B is governed by the genes peculiar to the infecting strain. Indeed, strains belonging to the same group differ in their ability to grow in host tissues and in their resulting virulence. Virulent and avirulent strains may appear identical by criteria such as antigens detected by conventional polyclonal antisera, biophysi cal properties, and genomic structure. Further more, even virulent coxsackievirus strains indistinguishable by conventional criteria induce a variety of clinical syndromes, including myocarditis. For example, adoles cent CD-I mice inoculated with one variant of coxsackievirus B3 developed readily observable myocarditis, whereas a second variant was virtually innocuous [29] . These variants were indistinguishable in terms of their ability to grow in cell lines in vitro or heart muscle cells in vivo, to stimulate defective particles that might interfere with viral replication, or to stimulate host immune factors such as interferon. However, more stringent analysis may reveal basic differences between virulent and avirulent strains. Some of the factors determining the outcome of coxsackievirus B infections are governed by the genes of the infecting strain. Prabhakar et al. [30] showed that monoclonal antibodies can detect differences between strains of virus considered antigenically homogeneous by conventional analysis. These authors used monoclonal antibodies to detect 13 antigeni cally distinct variants of coxsackievirus B4. They also found that the rate of antigenic mutation is extremely high in this strain, resembling that long recognized to occur in influenza viruses.

Further analysis of antigenic variants of coxsackievirus B4 shows that clinical isolates display a mixture of highly conserved, moderately conserved, and poorly conserved epitopes [31] . Their studies suggested that a high rate of error in RNA replication is responsible for these frequent antigenic changes. Moreover, this appeared to be independent of any selection pressure. 

Receptor availability? Interferon induction in interstitial cells? Express la antigens?

Antibody-all relevant epitopes recognized?

Cell mediated immune responses: antiviral autoimmune access to myocardium interferon production See text for discussion and references.

Despite this marked antigenic diversity, all the isolates were neutralized by conventional polyclonal antiserum. Although most of the antigenic changes solely detected by mono clonal antibodies may be innocuous, there is also the possibility that these changes deter mine viral tropism for different target cells in vivo and hence their subsequent virulence. However, while there is considerable antigen variation between different strains of coxsackievirus B, repeated passage in intermediate hosts seems to stabilize their properties since there is preferential survival of strains that are tropic for some target cells such as pan creatic islet cells and other endocrine tar get cells. Thus, coxsackievirus B repeatedly passaged in mouse pancreata or beta cell cul tures acquired a specific tropism for those cells [32] . The nature of the host immune response also influences the outcome of experimental coxsackievirus infection. This important point is borne out by all the earlier work on this model (reviewed by Woodruff [2] ). Both humoral and cell-mediated immune responses contribute to host defense. In animal models, humoral mechanisms are decisive in limiting primary infection. Cell-mediated immunity also contributes to host defense but some times at the price of initiating immunopathologic reactions against virus-infected myocar dial cells. Thus cell-mediated responses are a mixed blessing for the infected host. There is good experimental evidence that cellmediated immunity may reduce the virulence of coxsackievirus infections. Khatib et al. [33] found that the virulence of one strain of coxsackievirus B4 was increased in a strain of Swiss mice given antithymocyte serum at the time of infection. In addition, immunosuppressive treatment with cyclophosphamide converted infection by an avirulent strain of B3 coxsackievirus into one produc ing myocarditis, probably because the virus was thereby allowed to replicate to abnor mally high titers [29] . However, the dual contribution of viral cytopathic effects and cell-mediated immunity is well illustrated by experiments in which athymic BALBc-nu/nu mice were infected with coxsackievirus B3 [34] . In the immunodeprived mice, the virus produced mild degenerative or necrotic changes only. However, intense inflammatory changes were induced after the adoptive transfer of spleen cells of immunocompetent nu/+ mice immunized with the same virus. In another athymic strain nu/nu mice could not eliminate infectious virus from their hearts, and this persistent infection was associated with myocardial abnormalities characteristic of cytopathology and also with increased mortality rates [35] . Euthymic nu/+ mice were infected for much shorter periods, yet developed persistent myocardial abnormali ties attributable to the cellular response. These observations emphasize that the myocardial changes are determined partly by viral factors and partly by the character of the T-cell response.

The immunopathologic consequences of T-cell immunity often offset any advantage to the host. This outcome is suggested by the frequent presence of mononuclear cell infil trates at sites of coxsackievirus replication in myocytes indicative of myocarditis. In vitro studies of sensitized lymphocytes in coxsackie virus infections emphasize the complexities of cell-mediated immunity to these agents and point to some pathways that may lead to immunopathologic catastrophes. In general, the briskness of the cytotoxic T-cell reaction to infected myocardial cells correlates with the severity of the myocarditis, which sug gests that hypersensitivity to viral antigens contributes to the immunopathologic conse quences of infection. However, there are also important observations that sensitized lym phocytes in coxsackievirus infections are able to kill uninfected myocardial cells and fibroblasts (Woodruff [2] ), and there is now good evidence that this autoreactivity contributes to the myocarditis. In addition, Paque et al. [36] have described experiments in which peritoneal exudate cells from mice infected with coxsackievirus B3 were tested for their reactivity with cardiac antigens using the mi gration inhibition assay. They found that the migration of these cells was specifically inhib ited by cardiac antigens isolated from mice infected with coxsackievirus B3 but not by antigens prepared from uninfected hearts or from hearts infected with an unrelated virus. Since the cardiac extracts failed to bind virusneutralizing antibodies, they concluded that virus-induced tissue antigens but not virusrelated antigens were responsible for this inhibition. Paque et al. have extended these observations to baboons infected with this virus [37] . In these experiments, sensitivity directed solely at antigens isolated from hearts infected with coxsackievirus was shown by a proliferative response of lympho cyte monitored by the incorporation of tritiated thymidine as well as by inhibition of migration.

Further evidence for autoimmunity to myocyte antigens is provided by experiments in which spleen cells of mice infected with coxsackievirus B3 were allowed to adsorb to uninfected or infected myocyte monolayers [38] . One population of T lymphocytes adsorbed to and lysed uninfected myocytes and were assumed to be autoreactive. In contrast, lymphocytes lysing infected cells were considered to be virus specific. The autoreactive lymphocytes were injected into recipient mice infected with the same virus but deprived of T lymphocytes by prior ex perimental manipulation. The transferred cells induced myocarditis in mice not other wise able to mount a T-cell response to infected myocytes. These observations raise important questions about the nature of the autoantigens, the manner in which auto immune reactions are induced, and the fac tors that dictate their immunopathologic importance.

So far, there is little information about the precise nature of the antigens in apparently uninfected heart cells which attract the T-cell response. These might be related to sites of viral transport across cell membranes during the initial infection or during the export of newly synthesized viral particles. Nor is it clear whether the response is directed at viral antigens or, as seems more likely, normal host cells whose production or expression is al tered by viral infection. Furthermore, while autoreactivity persists beyond the time in which viral persistence can be detected by conventional techniques, there is still the strong possibility that in reality it is sustained by viral persistence in latent or defective form. Appropriate experiments with viral probes or with techniques for reactivating vi rus have not yet been reported. The sequence of events by which transient or persistent coxsackievirus infection induces an autoreactive T-cell response also remains largely unexplored. Autoreactivity arises early in infection and is more vigorous in female than in male mice. The ability of T cells to gain access to infected myocardial fibers may also influence the outcome of the infection. Huber and Job [39] in 1983 described two strains of coxsackievirus B type 3 which are indistin guishable in terms of their antigenic proper ties and ability to grow to high titers in myo cardial cells. Judged by in vitro techniques these strains are equally efficient in inducing cytotoxic T lymphocytes. However, it is likely that in mice infected by the virulent strain, these cells are able to reach and damage myo cardial cells, whereas lymphocytes in mice infected by the nonvirulent strain lack this ability. However, other crucial pieces of in formation are lacking, such as the expression of viral antigens in infected cells, local Inter feron production, the expression of la anti gens by myocardial cells, and the kinetics of the infiltrating cell populations.

It is also possible that the extent to which coxsackievirus infection interferes with the re gulation of the immune response also deter mines the severity of the immunopathologic damage. As with other models of experimental virus infection, coxsackieviruses induce sup-It has long been recognized that some strains of echo virus induce PM in mice [43] . How ever, the PM results from acute necrosis of muscle fibers, and there is little evidence that the PM is mediated by immune mecha nisms. Indeed, paralysis only occurs in mice in which viral replication has reached very high titers. Thus, the disease is easily induced in newborn mice in which more than pressor T cells which affect immune responses in general and may also influence the character and efficiency of specific antiviral immune re sponses [40] . Deficient suppressor mechan isms could allow persistent autoreactive Tcell responses that damage heart fibers even after infectious virus has been eliminated and viral antigens are no longer expressed.

There is little evidence to date that coxsackieviruses interfere directly with lym phocyte subpopulations. However, it has been shown that these viruses readily establish per sistent infections in human lymphoid cell lines and particularly in B-cell lines [41] . Only a minority of the cells are infected, hinting that in vivo infection might be difficult to identify if only a minority of circulating lym phocytes are infected. These observations also suggest that the infection may be maintained by mutants of the original infecting virus. The relevance of this experimental model of lym phocyte infection to putative defects in immunoregulation has not yet been established.

In addition to specific immune responses, other nonspecific factors influence the out come of experimental coxsackievirus infec tion. Thus, Reyes et al. [42] have shown that myocardial damage is accentuated in mice infected with coxsackievirus B3 that were given forced exercise in the form of daily swimming. Observations of this kind rein force the suspicion that excessive physical ac tivity may predispose to myocarditis and PM in clinical practice.

50% of skeletal muscle tissue is destroyed, older mice rapidly acquiring resistance to this process [44] . Similarly, acute myositis has long been recognized as a consequence of coxsackievirus B in neonatal mice, the dis order resulting from the cytopathic effects of viral growth [45] . Interestingly, even in this model, the myositis is selective for spe cific muscle groups, the hip extensors and

hindquarter knee flexors being particularly susceptible.

A more persuasive animal model of virusinduced PM is that described by Strongwater et al. [46] . In this model, the intraperitoneal inoculation of the Tucson strain of coxsackievirus Bl induced proximal hindquarter weak ness that persisted for more than 10 weeks. Moreover, the criteria for PM were satisfied by the characteristic EMG and histologic changes. Furthermore, these changes per sisted long after infectious virus could no

The extraordinary diversity of clinical syn dromes associated with the acquired immu nodeficiency syndrome (AIDS) has alerted clinical investigators to the possibility that demonstrable or covert infection by retroviruses could be etiologically related to PM, or indeed any chronic disorder whose cause has continued to baffle them. However, PM has not emerged as a frequent complication of AIDS despite the close monitoring of such patients for neurologic disorders. Thus, of 352 patients with AIDS or generalized lymphadenopathy studied at one center between 1979 and 1984, only five patients had muscle problems: two had persistent myalgias, two had myopathy, and only one had PM [47] . More recently, however, two additional pa tients with PM associated with AIDS were reported [47a] , as discussed in Chapter 1. Yet in a primate model of acquired immunodefi ciency, 50% of monkeys infected with the D retrovirus called SAIDS D showed muscle weakness and wasting, elevated serum con centrations of sarcoplasmic enzymes, and bi opsy features of PM [48] . The etiologic agent could be transmitted experimentally to healthy rhesus monkeys by tissue homogenates, blood, saliva, and urine. Reverse transcriptase assays revealed the presence of virus in muscle biopsies, but immunofluorescent studies with antibody to the agent longer be detected, suggesting that the disor der was mediated by immune mechanisms. Infectious virus could only be recovered from the affected muscles for 2 weeks after infec tion, during which period viral antigen could also be detected by immunofluorescence. No tably, electron microscopy did not reveal characteristic viral particle during this period of demonstrable infection. This model has many features in common with human PM and could also explain why it might be diffi cult to demonstrate virus infection in man.

showed that the agent was localized to the infiltrating inflammatory cells and interstitial fibroblasts rather than to muscle fibers. Inter estingly, the retrovirus replicated in myotubes and fibroblasts in culture but did not have any cytopathic effect on these cells. The predomi nantly immunopathologic features of this model make it an attractive analogue of the human disease. However, generalized immu nodeficiency is not a feature in the vast major ity of patients with PM/DM. Nevertheless, these observations emphasize the protean na ture of retroviral infection and indicate that this etiologic possibility should not be lightly discarded. In the past, retroviral infection has been invoked rather simplistically in the pathogenesis of inflammatory connective tis sue diseases, such as PM/DM. It is worth noting that Harbers et al. [49] microinjected cloned retroviral DNA into mouse zygotes and found that virus-specific RNA was subse quently expressed at different concentrations in the organs of infected mice. In one mouse, retrovirus specific information was expressed at far higher concentrations in muscle than in other tissues. While such viruses can be ac quired by horizontal infection, it is also possi ble that any involvement of retroviruses in the pathogenesis of PM might be more complicated.

The pathogenesis of autoimmune diseases in general continues to be keenly debated and frequently reviewed [50] . Viruses are com monly invoked in most schemata albeit in concert with genetic and immunologic factors [51] . It is instructive to apply these general hypotheses to PM/DM in order to see how viruses could invoke this autoimmune disease and the extent to which there is supporting data ( Table 6 .3).

It has long been a popular notion that virus infection might induce autoantigens and thereby make infected cells targets for auto immune attack. There is in fact little evidence that viruses induce novel antigens or alter the structure or expression of normal antigens. Nevertheless, there is good evidence that vi rus infection affects the recognition of surface antigens by the immune system, thereby in creasing the possibility that these will induce an immune response.

Classically, Lindenmann and Klein [52] in 1967 found that influenza virus infection of target cells increased the immunogenicity of host cell antigens expressed by the infected cells. The mechanisms of this phenomenon have been extensively studied. For example, it has clearly been shown that vaccinia virus infection of tumor cells increases cytotoxic T-cell and antibody responses against the in fected cells [53] . There are also several models of experimental viral infection that provoke autoimmune responses against the infected organs, with pathologic consequences. Mice infected with reovirus type I develop both transient diabetes and a runting syndrome [54] . The infected mice produce serum autoantibodies reactive with cytoplasmic an tigens in beta cells of the pancreatic islets of Langerhans, the anterior pituitary gland, and gastric mucosa. Additional autoantibodies are generated against both insulin and growth hormone, and serum concentrations of these hormones are abnormally low. Another perti nent example is provided by experimental in fection with the JHM strain of murine coronaviruses, which is neurotropic and which induces either acute or subacute encephalomyelitis, depending on viral and host genetic factors [55] . Subacute encephalomyelitis is determined not only by the antiviral immune response but also by an autoimmune reaction to brain antigens, including myelin.

An important issue in such models is the role of viral replication in the target cells. In the reovirus model of experimental autoim mune endocrine disease it seems likely that virus in the target organs is mandatory since reovirus type 3, which does not infect the anterior pituitary gland, does not induce autoantibodies to growth hormone. How ever, the period of virus infection may be transitory since the autoimmune process per sists long after local infection can readily be detected. Moreover, an immunopathologic basis for the syndrome seems certain, since it can be prevented by immunosuppressive treatment with antilymphocyte serum [56] , The extent to which persistent infection of brain cells by JHM virus is essential for the subacute demyelinating disease in susceptible hosts is still debatable.

One possible mechanism by which tran sient virus infection of target cells may induce a persistent T-cell response is the inappropri ate expression of la (class 2, or DR) antigens by the host cell. The expression of such anti gens has mainly been attributed to 7-interferon released by infiltrating T lymphocytes. In par ticular, this mechanism has been invoked to explain autoimmune disorders of the thyroid gland [57] . However, more recently in the JHM-induced model of demyelinating encephalomyelitis, it has been shown that viral particles directly induce la antigen expres sion by astrocytes [58] . Moreover, the extent of this interferon production is likely to be genetically controlled. Since both viral and host cell antigens are likely to be presented to the immune system in association with la an tigens, it is also evident that genetic factors control the generation of cytotoxic T cells directed at both sets of antigenic determinants.

There is so far little evidence that similar mechanisms operate in PM. In particular, there is no information about the steps that initiate the infiltration of muscles by lympho cytes. Analysis of the infiltrating cells by immunocytochemical techniques has shown that the majority of the infiltrating cells are T lymphocytes bearing antigen determinants characteristic of activated lymphocytes [59] . Furthermore, these cells predominate in un treated acute rather than chronic disease [60] .

Interferons have also been detected in muscle biopsies [61] . However, the antigens against which these cells react have not been charac-Viral Etiology of Polymyositis 109

terized. It is possible that after isolation and expansion by cloning techniques the infiltrat ing T cells may prove specifically reactive to muscle cells, using the kind of assays that Cambridge and Stern [62] have shown are true of circulating blood lymphocytes in PM. Cellular infiltrates resembling those en countered in PM/DM are also provoked by exercise, emphasizing that these do not pro vide specific clues to the nature of the local initiating event, infectious or otherwise [62a] . Antimyosin antibodies have been detected by radioimmunoassay in 90% of sera from pa tients with PM [63] , but the nature of the stimuli provoking these autoantibodies is un known. There have been claims that the stain ing pattern of autoantibodies in patients with acute myocarditis can distinguish between patients with preceding coxsackie B infection and those with other etiologies, but these an tibodies were present in low titer, and their reactivity showed considerable overlap with antigens in other tissues [64] . It is purely conjectural that transient infec tion of muscle cells initiates the immunopathologic events of PM/DM. However, many vi ruses are able to replicate in vascular endothelial cells [65] . These cells could serve as sites for the initial growth of viruses implicated in the pathogenesis of PM/DM. Indeed, it has been shown that 7-interferon induces la ex pression on cultured human vascular endothelial cells and dermal fibroblasts [66] , Fur thermore, Ashida, Johnson, and Lipsky [67] have shown that these cells function as antigen-presenting cells in model systems. It is feasible, therefore, that muscle fibers become secondarily inflamed as the result of changes in the microenvironment of local blood ves sels and that persistent local immune responses are thereby established. Indeed, fully developed lymphoid follicles have been noted as a histologic feature of PM/DM [68] , It is customary to invoke genetic factors to account for the rarity of such events. How ever, in contrast to the strong HLA associa tions with most organ-specific autoimmune diseases, there is only weak evidence for such associations in PM/DM [69] . Claimed associ ations with HLA-B8 and HLA-DR3 in both adult and juvenile dermatomyositis are not compelling and have been observed in rela tively small numbers of patients.

The specificities of the autoantibodies characteristic of PM/DM have been invoked as evidence for a possible viral etiology for this group of diseases. Several authors, nota bly Plotz [70] and Bernstein et al. [71] , have stressed that the autoantibodies associated with inflammatory connective tissue diseases, including PM/DM, are directed at a relatively small number of cellular components. These autoantigens have been fairly well character ized as proteins, often present in complexes with additional proteins or nucleic acid mole cules. The antinucleoprotein autoantibody peculiar to many connective tissue diseases reacts with small ribonucleoproteins (RNPs) in which the antigens are commonly the small nuclear RNA species, Ul, U2, U4, U5, or U6 [72] . In addition, the antibodies characteristic of PM/DM commonly react with cytoplasmic antigens, notably aminoacyl-tRNA synthetases. These antigens are associated with mechanisms of protein synthesis [73] . In the first stage of protein synthesis, RNApolymerases catalyze the formation of messenger RNA complementary to the DNA sequence. This messenger specifies the sequence in which amino acids are assembled in the na scent polypeptide chain. The amino acids are initially attached to tRNA molecules before their polymerization into polypeptides. A specific set of enzymes termed aminoacyl-tRNA synthetases couple each amino acid to its matching tRNA. Anti-La autoantibody reacts with precursor forms of ribosomal RNA and tRNAs for five different amino acids [72] . The transcription of these RNA sequences is controlled by RNA polymerase III. Similarly, anti-Ro autoantibodies precipi tate RNPs containing RNAs whose transcrip tion is also dependent on polymerase III [74] . The Jo-1 antigen is a protein-RNA complex incorporating the synthetase needed to com plex histidine with the corresponding tRNA [75] , Moreover, anti-Jo autoantibody blocks the incorporation of histidine into nascent proteins, a reaction that depends upon this enzyme. Since these protein synthetic steps are also utilized in the synthesis and assembly of some virus particles, it is conceivable that these autoantibodies might represent an im mune response by the host aimed at disrupt ing viral synthesis. Anti-La antibody precipi tates not only cellular RNPs but also RNP complexes containing RNAs involved in the synthesis of adenovirus and Epstein-Barr vi rus (EBV) [72] . Many plant viral RNAs and a few animal viral RNAs contain a tRNA-like structure at their free end which may utilize tRNA synthetases to regulate viral assembly [76] , Equally, these autoantibodies might be directed at enzymes and proteins involved in protein synthesis which have been rendered autoantigenic by the transcription of viral se quences [71] . This mechanism is analogous to that likely to explain lupus-like syndromes provoked by drugs such as hydralazine.

However, it is only an assumption that the antigen specificities of the autoantibodies hint at a viral etiology. These autoantibodies may simply reflect an autoimmune response to a spectrum of antigens common to many tissues regardless of the provoking cause. Similarly, there is an obvious analogy with experimental reovirus infection in that the autoantibodies might determine patterns of tissue damage but be of little help in determin ing the nature of the initial insult.

A similar dilemma arises from the convinc ing observations that the sera of patients with autoimmune connective tissue diseases con tain antibodies to antigens coded by retroviruses. Rucheton et al. [77] have shown that sera from patients with anti-RNP and anti-La autoantibodies react with several viral polypep tides coded by the viral gag gene. How ever, as the authors emphasize, these antibod ies could be of secondary importance, simply reflecting the nonspecific activation of endo genous viral antigens rather than a response to exogenous retroviral infection.

The advent of monoclonal antibodies has provided a powerful means of testing the long-held belief that autoimmune diseases may arise from cross-reactions between mi croorganisms and tissue antigens. Molecular mimicry of this kind can now be sought by direct experiment and by computer tech niques comparing known microbial and tis sue antigenic sequences. These techniques have been used to test the idea that molecular mimicry between viruses and muscle antigens might account for the continued immune re sponses characteristic of PM/DM. Srinivasappa et al. [78] screened 600 monoclonal antibodies to 11 different viruses for their reactivity with 14 different organs from normal mice. They found that 3.5% of the antiviral antibodies reacted with specific cells in these organs and that several of the antibodies reacted with antigens in more than one organ. However, no cross reactions with muscle-specific anti gens were uncovered in this survey or referred to in the accompanying review of similar searches. Immunofluorescence studies have revealed a monoclonal antibody to coxsackievirus B4 that reacts exclusively with A bands in the myofibrils of myocardial muscle, al though only one of 66 antibodies tested showed this pattern of reactivity [79] . Walker and Jeffrey [80] adopted a different strategy by using a computer alignment procedure to seek microbial sequences comparable with amino acid sequences of histidyl-tRNA synthetase and alanyl-tRNA synthetase since these are known to be the target molecules for autoantibody responses in PM/DM. Close matches were discovered between histidyl-tRNA synthetase and protein sequences of EBV. Similarly, of the 30 closest matches with alanyl-tRNA synthetase, more than half were with viral proteins, including those coded by EBV, some influenza virus hemagagglutinins, and a protein coded by adenovirus 2. These studies also uncovered sequence similarities with the myosin light chain tropomyosin and the skin component keratin. However, as em phasized by the authors, the tRNA synthetases are universally distributed, and it is diffi cult therefore to understand why skeletal muscle fibers should be a specific target organ in this disease. Furthermore, sequence homologies may be entirely fortuitous or result when part of a cellular protein is incorporated into viral protein during the course of viral replication. Furthermore, similarities in se quence do not take into account conformational differences which may in reality deter mine the ability of an antibody to fit a given epitope [81] . Moreover, regions of high mo bility in a protein also determine antibody binding so that comparison of sequences can never replace the need for experimentally studying the ability of an antibody which reacts with one structure to react with an other that appears to mimic it [82] . It is also important to test experimentally the idea that any observed cross-reactivity accounts for persistent autoimmune reactions. Thus Fujinami and Oldstone [83] in 1985 reported sequence homology between amino acid se quences in myelin basic protein and hepatitis B virus polymerase; lymphocytes from rab bits immunized with the viral sequence pro duced a proliferative response in vitro to chal lenge with basic myelin protein and also developed histologic changes reminiscent of experimental allergic encephalomyelitis.

It is now accepted that antibodies to novel sequences in the variable portion of the anti body molecule (idiotypes) contribute to the regulation of antibody production. Further more, since there are sequence similarities be tween the provoking antigen, receptors for those antigens on cell surfaces, and the idiotype, anti-idiotype antibodies might be expected to recognize antigen receptors. This pathway for inducing autoantibodies reactive with cell membrane antigens has been recog nized as a possible source for autoimmune diseases and, indeed, anti-idiotype autoanti bodies have been shown to have immunopathologic consequences in some experimen tal models [70, 84] .

The complexities of possible derange ments of the idiotype network have been in creased by the recent demonstration that a human monoclonal antibody reactive with a membrane protein of multiple organs con tains an idiotype which in turn stimulates an anti-idiotype antibody, and that an antibody raised against this anti-idiotypic antibody it self recognizes the protein determinant that initiated the whole chain of events [85] . Ago nized readers are referred to ancient chain songs dealing with kids brought to market.

Hybridoma techniques have revealed monoclonal antibodies that react with epitopes shared by multiple organs and tis sues. One such epitope in SLE is the phosphodiester bond [86] , while monoclonal auto antibodies reactive with pituitary, thyroid, pancreas, and stomach have also been de scribed [87] . Antibodies to these idiotypes have been produced experimentally and used to screen the sera of patients with autoim mune diseases for antibodies bearing these idiotypes. If autoimmune diseases are pro voked by a persistent autoantibody response to idiotypes, one might predict that the sera of patients with autoimmune diseases would contain a higher concentration of antibodies with this idiotype than do sera of normal individuals. However, Essani et al. [88] found that idiotypic markers thought to be related to autoimmune diseases are expressed in normal individuals and that antibodies bearing these markers do not necessarily bind to autoantigens. Similarly, the number of circulating B lymphocytes reactive with anti-idiotype antibody was identical in patients with autoimmune diseases and in controls. Similar observations showing the public nature of autoantibody idiotypes have been made by other authors [84] . Thus far theories about the pathogenesis of autoim mune diseases based on generation of antiidiotype antibodies have not provided any etiologic clues in autoimmune diseases in gen eral. With respect to PM/DM, those autoreactive antibodies whose idiotypes have been analyzed do not react with muscle. However, the idiotypes of muscle-reactive antibodies in PM/DM have not been examined. There are also caveats concerning the idea that analyz ing autoantibody idiotypes might further im plicate viruses that are putatively involved in the etiology of PM. McClintock et al. [89] in 1986 reported that anti-idiotypic antibodies raised to mouse monoclonal antibodies reac tive with independent epitopes on coxsackievirus B4 did not block the binding of this virus to cells bearing appropriate receptors.

The original studies of mice with an autoim mune polyendocrine disorder provoked by infection with reovirus type 1 revealed that spleen cells fused with a myeloma cell line produced monoclonal antibodies reactive with cells in the anterior pituitary gland, pan creas, small intestine, and stomach [90] . Ad sorption studies on antigen affinity columns confirmed that the antibody was indeed re acting with antigens common to the different target organs. It subsequently became evident that cells producing monoclonal antibodies with the same repertoire of multiple organ reactivity can be isolated from the spleens of normal mice [91] . Similarly, Dighiero et al. [92] have established hybridomas with spleen cells of normal mice synthesizing monoclonal autoantibody reactive with myosin, doublestranded DNA, actin, tubulin, and spectrin. In clinical studies, EBV transformed blood lymphocytes from both normal individuals and patients with autoimmune diseases pro duced monoclonal autoantibodies reactive with antigens in multiple organs [93] . The target antigens were detected in thyroid, pan creas, stomach, stratified squamous epithe lium, and nerve axons. The reactions could only be detected by the highly sensitive avidin-biotin-immunoperoxidase assay. Fur thermore, this group [94] used the same tech nique of screening EBV-transformed B cells to reveal that B cells from patients with autoim mune diseases produce monoclonal antibod ies reactive with both pancreas and thyroid. Evidently B cells in patients with autoimmune disease synthesizing serologically recogniz able autoantibodies may also synthesize autoantibodies reactive with different organs which are undetectable by standard tech niques. Thus the association of PM/DM with other autoimmune diseases could result from the simultaneous production of muscle-re active autoantibodies that have not yet been characterized. It also follows that PM/DM, like other autoimmune diseases, could result from the inappropriate activation of B cells already programmed to react with muscle an tigens. The multiple ways in which this acti vation could be achieved include the bypass ing of suppressor T-cell mechanisms and the inappropriate presentation of muscle-specific antigens to B cells by T cells reacting with surface antigens expressed by muscle fibers [95] . Viral infections could provide the ini tiating signal; experimentally Bromberg et al. [96] have shown that herpes simplex, New castle disease, and vaccinia viruses can induce immune responses against membrane anti gens that are normally weakly immunogenic. There is only meager evidence to support the alternative possibility that viruses putatively inducing PM/DM latently infect suppressor cells, thereby leading to their functional abla tion. Hollingworth et al. [97] showed that herpes simplex virus failed to replicate in PHA-stimulated blood lymphocytes from pa tients with PM/DM, suggesting that some cells might harbor virus particles that inter fered with the growth of the challenge virus. Since viral infection of a minority of circulat ing lymphocytes has marked immunodepressive effects [98] , this possibility can not be lightly discarded. It has recently been shown that measles virus has a strong tropism for lymphocytes and that continued infection of these cells may contribute to the pathogenesis of the persistent viral disease subacute sclerosing panencephalitis [99] .

PM/DM commonly occurs in association with other inflammatory connective tissue diseases in which autoimmune reactions are promi nent. Thus, theories attributing autoimmune diseases to the uncontrolled proliferation of autoreactive lymphocyte clones are also ger mane to the pathogenesis of PM/DM. Classic ally, the emergence of such clones has been considered to result from the failure of regula tory mechanisms that normally lead to their elimination. Alternatively, mutations in the progenitor cells for autoimmune reactive lymphocytes might lead to their autonomous proliferation in a manner independent of con ventional antigenic stimulation. Indeed, the proliferation of autoreactive cells has often been considered intermediate between a con-ventional immune response and overt malig nant proliferation. Theoretically it is now possible to test the strength of such hypothe ses by comparing the characteristics of autoantibody secreting cells with frankly malig nant lymphoma cells (Table 6 .4). There is so far no evidence that autoreactive cells origi nate from a single progenitor cell in a manner analogous with the B cells of patients with chronic myeloid leukemia; the latter point has been clearly established in black female patients with the disease who are heterozy gous for the enzyme glucose-6-phosphate hydrogenase [100] . Nor have gene rearrange ments of the kind observed in neoplastic lymphoid cells [101] been detected in the genes coding for the antigen receptors of autoreac tive T cells or the immunoglobulin coding genes of autoreactive B cells. Gene probes for T-cell receptors have been used to document the maturation of mutant cells lacking hypoxanthine-guanine phosphoribosyltransferase (HPRT) activity [102] . Furthermore, assays for rearrangements of genes coding for T-cell antigen receptors have provided insight into a clinically benign but biologically lymphoproliferative disorder, lymphomatoid papulosis [103] .

Immunoglobulin synthesis involves a se ries of differentiation steps in which the prod- These mutations usually involve a sin gle base substitution, but it is not known whether these are entirely spontaneous or de pendent on antigenic stimulation. Similarly, there is a high rate of mutation in the genes coding for the variable portions of the beta chain of the antigen receptor of T cell al though at a lower rate than that occurring in genes coding for the variable portion of the immunoglobulin molecule. Thus, classic the ories postulating that autoimmune diseases arise from mutation in antibody synthesizing clones must be reconciled with observations showing that mutation is a physiologic step in the generation of antibody diversity. Further more, such theories must explain the circum stances that confer continuous and appar ently autonomous proliferation on such clones. Certainly a single somatic mutation can result in the production of an antibody with autoantibody specificity [104] , but anal ogous mutations in human autoantibodyproducing cells have not yet been described.

In malignant lymphomas, substantial evi dence has been adduced implicating the inap- propriate activation, mutation, or translocation of oncogenes in abnormal B-cell proliferation; moreover, many of these mod els postulate that these transforming events are initiated or enhanced by exogenous viral infection. There is as yet no evidence that similar events govern the proliferation of autoreactive B cells in connective tissue dis eases. The activation of the c-myc gene in SLE, for example, seems simply to reflect the extent of B-cell proliferation rather than an unusual event initiating this proliferation [105] . Nevertheless, the wealth of interaction observed between exogenous infectious agents and the deregulation of myc-genes in B cell tumours [106] means that such ideas should not be lightly discarded in other lymphoproliferative disorders. B-cell differentiation and maturation involves multiple steps, each of which is governed by interactions with ex ternal ligands, such as B cells and their prod ucts, activated components of the comple ment sequence, and other as yet unknown factors. It is noteworthy that the B-cell hyperplasia characteristic of mouse strains prone to spontaneous autoimmune disease is deregu-

There is so far little evidence that PM/DM is caused directly or indirectly by viral infec tions. Conventional methods have failed to isolate viruses in the vast majority of cases and serologic surveys have proved uncon vincing. The case for continuing to seek a viral etiology rests mainly on animal models. Myocarditis and PM induced by coxsackievirus infection is mediated by immunopathologic mechanisms that may continue to dam age muscle fibers long after infectious virus and even viral antigens have been eliminated. Clearly, if analogous mechanisms operate in human disease, it will prove extremely diffi cult to identify the provoking agent by the time the disease is clinically apparent. Fur thermore, the combination of viral and host genetic factors necessary to induce muscle in flammation may be extremely rare and im-lated at one or more points in their matura tion [106] . There is little information about the detailed regulation of human B cells syn thesizing conventional or autoantibodies. However, the introduction of improved tech niques for propagating human B cells in cul ture should remedy this situation. There is evidence, for example, that lupus B cells spon taneously synthesize B-cell growth factors and are refractory to exogenous growth fac tors [107] . The failure of normal repair processes may increase the possibility for recombination events between immunoglobulin-coding genes and genes related to controlled proliferation. This mechanism has also been invoked as a factor in mouse strains susceptible to plasmacytoma induction [106] , and it is note worthy that lymphocytes from patients with PM/DM are abnormally susceptible in vitro to irradiation, ultraviolet irradiation, and the mutagen methylnitrosourea [108] . In general, the relevance of such concepts to the pathogenesis of PM/DM has not been experimen tally tested. possible to define by orthodox biologic and immunologic techniques. Indeed, the succes sive steps leading to autoimmune inflamma tion of muscle in PM/DM or inflammation of organs in other autoimmune diseases are tortuous to unravel even when the disease is deliberately incited. The hope of progress lies in improved techniques for isolating, expand ing, and characterizing the effector cells from the blood and local lesions of patients with this group of disorders.

The association of PM/DM with other auto immune diseases encourages the belief that genetically determined mutations in T or B lymphocytes may be responsible for the dis order. If PM/DM results from the quasi-malig nant proliferation of autoreactive lymphocyte clones, such hypotheses can only be tested by detailed analysis of the multiple steps leading to the emergence of such clones. Such analo gies imply that viruses might not act by simple cause and effect. Instead, it is possible that they might act at several stages in the emer gence of such clones. Viruses could promote recombinational events between transform ing genes and genes coding for immunoglobu-lin or T-cell receptor sequences. Alternative ly, viruses could act by overcoming restric tions to lymphocyte proliferation normally imposed by a variety of regulatory ligands. However, overall, the stimulus for further in vestigating a viral etiology for P M / D M still rests on faith and metaphors.

Immunological features of polymyositis/dermatomyositis

Viral myocarditis

Picornavirus-like crystals in subacute polymyositis

Chronic myopathy associ ated with coxsackievirus type A9: a com bined electron microscopical and viral iso lation study

Transient acute myositis in childhood

Hepatitis B antigen and polymyositis

Inclusion-body myositis: clinicopathological studies and isolation of an adenovirus Type 2 from muscle biopsy speci men

Polymyositis ac companying Coxsackie virus B2 infection

Piconavirus-like inclusions in polymyosi tis: aggregation of glycogen particles of the same size

Myo sitis and acute respiratory infection

Acute myositis associated with influ enza B infection

Vasculitis and myositis sec ondary to rubella vaccination

Infectious myositis and related syndromes

Benign postinfection polymyositis

Echovirus disease in hypogammaglobulinemic patients

Persistent and fatal central nervous system echovirus infections in patients with agammaglobulinemia

Isolation of influenza virus from muscle in myoglobinuric polymyositis

Lack of asso ciation of A/NJ/76 (swine flu) vaccine and polymyositis

Detection of coxsackie-Bvirus-specific RNA sequences in myocardial biopsy samples from patients with myocarditis and dilated cardiomyopathy

Antibodies to Coxsackie B viruses in congestive car diomyopathy

The heart and cardiac conduction system in polymyositis-dermatomyositis: a clinicopathologic study of 16 autopsied patients

Coxsackie B neutralisation titres in polymyositis/dermatomyositis

Antibody to Coxsackie-B virus increased incidence in sera from children with recently diagnosed juve nile dermatomyositis

Coxsackie B virus specific IgM re sponses in patients with cardiac and other diseases

Coxsackie-B-virus-specific IgM responses in children with insulindependent (juvenile-onset: type I) diabetes mellitus

Myositis, myoglobinemia and myoglobinuria associated with enterovirus Echo 9 infection

Infection of cultured human mus cle cells by influenza virus

Permeability changes elicited by influenza and Sendai viruses: separation of fusion and leakage by pH-jump experiments

Properties of Coxsackievirus B3 variants which are amyocarditic or myocarditic for mice

High frequency of antigenic variants among naturally occurring Cox sackie B4 virus isolates identified by monoclonal antibodies

Detection of conserved and nonconserved epitopes on coxsackievirus B: frequency of antigenic change

Virus-induced diabetes mellitus: glucose abnormalities produced in mice by the six members of the Coxsackie B virus group

Alterations in coxsackievirus B4 heart muscle disease in ICR Swiss mice by antithymocyte serum

The role of T lymphocytes in hepathogenesis of coxsackievirus B3 heart disease

Cocksackievirus B3 persistence and myocarditis in N:NIH(S) II nu/nu and +/nu mice

Assessment of cell-medi ated hypersensitivity against coxsackie virus B3 viral-induced myocarditis utiliz ing hypertonic salt extracts of cardiac tis sue

Assess ment of cell-mediated immunity against coxsackievirus B3-induced myocarditis in a primate model {Papio papioic)

Coxsackievirus B-3 myocarditis in Balb/c mice: evidence for autoimmunity to myocyte antigens

Differences in cytolytic T cell response of BALB/c mice infected with myocarditic and non-myocarditic strains of coxsackievirus group B, Type 3

Impairment of immunocompetent mouse spleen cell func tions by infection with Coxsackie virus B3

Group B coxsackieviruses readily establish persistent infections in human lymphoid cell lines

A mouse model of dilated-type cardiomyopathy due to Coxsackievirus B3

Factors determining pathogenicity of variants of ECHO 9 virus for newborn mice

Age dependence of paralysis in duced by Echovirus type 9 in infant mice

Selec tive polymyositis induced by coxsackie virus Bl in mice

A murine model of polymyo sitis induced by coxsackievirus Bl (Tucson strain)

Neurological manifestations of the acquired immunodeficiency syndrome (AIDS): ex perience of UCSF and review of the litera ture

Polymyositis in patients with AIDS

Polymyositis in an immunodefi ciency disease in monkeys induced by a type D retrovirus

Microinjection of cloned retroviral genomes into mouse zygotes: integration and expression in the animal

The autoimmune diseases

Virus-induced autoimmunity

Viral oncolysis: increased immunogenicity of host cell anti gen associated with influenza virus

The augmentation of tumour-specific immu nity by virus help. I. Demonstration of vac cinia virus-reactive helper T cell activity involved in enhanced induction of cytotoxic T lymphocyte and antibody responses

Virusinduced diabetes mellitus XX polyendocrinopathy and autoimmunity

Adoptive transfer of ERE-like lesions from rats with coronavirus-induced demyelinating encephalomyelitis

Virus-induced dia betes mellitus: autoimmunity and polyendocrine disease prevented by immunosuppression

Epithelial cells expressing aber rant MHC class II determinants can pres ent antigen to cloned human T cells

Viral particles induce la antigen expression on astrocytes

Characterization of polymyosi tis infiltrates using monoclonal antibodies to human leukocyte antigens

Monoclonal antibodies to human leuko cyte antigens in polymyositis and muscular dystrophy

Localization of Interfer ons and interleukin w in polymyositis and muscular dystrophy

The uptake of tritium-labelled carnitine by monolayer cultures of human fetal muscle and its po tential as a label in cytotoxicity studies

Cellu lar infiltrates in human skeletal muscle: Exercise-induced damage as a model for inflammatory muscle disease

Radioimmunoassay for antibodies to human skeletal muscle myosin in serum from patients with polymyositis

Diagnostic rele vance of humoral and cell-mediated im mune reactions in patients with acute viral myocarditis

Virus infection of endothelial cells

Lymphocytes recognize human vascular endothelial and dermal fibroblast la anti gens induced by recombinant immune In terferon

Hu man endothelial cell-lymphocyte interac tion

Polymyositis with infiltration by lymphoid follicles

Autoimmune diseases of muscle

Hypothesis: viruses and autoantibodies. Autoantibodies and anti-idiotype antibodies to anti-viral antibodies

Cellular protein and RNA antigens in auto immune disease

Precursor molecules of both human 5S ribosomal RNA and trans fer RNAs bound by a cellular protein reac tive with anti-La lupus antibodies

Basic genetic mechanism

Genes for two small cytoplasmic Ro RNAs are adjacent and ap pear to be single-copy in the human ge nome

Myositis autoantibody inhibits histidyl-tRNA synthetase: a model for autoimmunity

Possible functional role of viral tRNA-like struc tures

Presence of circulating antibodies against gag-gene MuLV proteins in patients with autoim mune connective tissue disorders

Molecular mimicry: frequency of reac tivity of monoclonal antiviral antibodies with normal tissues

Monoclonal antibody to coxsackievirus B reacts with myocardium

Polymyositis and molecular mimicry: a mechanism of auto immunity

On the use of se quence homologies to predict protein structure: Identical pentapeptides can have completely different conformations

Which structural features determine protein antigenicity?

Amino acid homology between the encephalitogenic site of myelin basic protein and virus: mechanism for autoimmunity

Idiotypes and autoimmunity

Multiple organ-reactive IgG antibody induced by an antiidiotypic antibody to a human mono clonal IgM autoantibody

Polyspecificity of monoclonal lupus auto antibodies produced by human-human hybridomas

Human multiple-organ reactive mono clonal autoantibody recognizes growth hormone and a 35,000 molecular weight protein

Anti-idiotypic antibodies against a human multiple organreactive autoantibody: detection of idiotopes in normal individuals and patients with autoimmune diseases

Anti-idiotypic antibodies to mono clonal antibodies that neutralize coxsackievirus B, do not recognize viral receptors

Multiple organ-reactive monoclonal autoantibodies

Lymphocytes capable of mak ing monoclonal autoantibodies that react with multiple organs are a common feature of the normal B cell repertoire

Murine hybridomas secreting natural monoclonal antibodies reacting with selfantigens

Epstein-Barr virus-transformed lympho cytes produce monoclonal autoantibodies that react with antigens in multiple organs

Hu man monoclonal autoantibodies that react with both pancreatic islets and thyroid

Mecha nisms of autoimmunity: a role for crossreactive idiotypes

Viral antigen act as helper determinants for anti body responses to cell surface antigens

Effects of immunosuppression on herpes simplex virus growth in lymphocytes of patients with connective tissue diseases

Interactions of viruses with the immune system

Detection of measles virus RNA in lym phocytes from peripheral-blood and brain perivascular infiltrates of patients with subacute sclerosing panencephalitis

Chronic myelocytic leuke mia origin of some lymphocytes from leukemic stem cells

Rearrangement of genes for the antigen receptor on T cells as mark ers of lineage and clonality in human lymphoid neoplasms

Use of T-cell receptor gene probes to quantify the in vivo hprt mutations of human T-lymphocytes

Clonal T-cell populations in lymphomatoid papulosis: evidence of a lympho-proliferative origin for a clinically benign disease

Somatic muta tion of the T-15 heavy chain gives rise to an antibody with autoantibody specificity

Induction of c-myc expression early in the course of B-cell activation: studies in normal humans and patients with systemic lupus erythematosus

Mechanisms in B-cell neoplasia

Spontaneous production of B cell growth factors by SLE lymphocytes

Defective repair of O 6 -methylguanine in autoimmune diseases