Antibodies to microbial hsp 60 and/or hsp 70 have been demonstrated in patients with various rheumatic and autoimmune diseases, including rheumatoid arthritis, SLE, Sjogren's syndrome, systemic sclerosis, mixed connective tissue disease, and ankylosing spondylitis, to name a few (19-26). In many of these reports, however, it is unclear whether detection of antibodies to microbial stress proteins in patient sera represents a true increase in specific antibodies to agents of putative etiological significance or whether their detection simply reflects a polyclonal increase in immunoglobulin levels. Indeed, a careful analysis using ELISA techniques found some IgG and/or IgM antibodies to mycobacterial hsp 60 in essentially all sera from patients with rheumatoid arthritis and various other rheumatic diseases and in all normal sera; moreover, patients and controls did not differ significantly in the levels of antibodies to this antigen (27).
Only limited data are available concerning the development of true autoantibodies to stress proteins; that is, antibodies that react with human antigens. Unconfirmed reports suggest that patients with SLE exhibit autoantibodies to ubiquitin and a charged synthetic octapeptide of ubiquinated histone H2A (28,29). Using immunoblotting techniques, our group described autoantibodies to hsp 90 and to the constitutively expressed 73-kDa member of the hsp 70 family in a minority of patients with SLE (2,3). Others also have reported autoantibodies to human hsp 70 in rheumatoid arthritis, SLE, and ankylosing spondylitis (24). Our initial studies have been extended in a more comprehensive survey of 268 patients with SLE or various other rheumatic diseases, inflammatory bowel disease, and several autoimmune skin diseases (30). Autoantibodies to human hsp 60, hsp 73, and hsp 90 were absent or rare in Sjogren's syndrome, rheumatoid arthritis, ankylosing spondylitis, Reiter's syndrome, and in control sera and infrequent (<20% of sera positive) in SLE and progressive systemic sclerosis. On the other hand, >20% of patient sera exhibited IgM and/or IgG autoantibodies to hsp 60 in mixed connective tissue disease, polymyositis/dermato-myositis, psoriatic arthritis, inflammatory bowel disease, epidermolysis bullosa acquisita, and bullous pemphigoid and to hsp 73 in Lyme disease and ulcerative colitis. In a similar study that used ELISA techniques rather than immunoblotting, some IgG and IgM autoantibodies to human hsp 60 (recombinant PI protein) were detected in nearly all sera from patients with SLE, various other rheumatic disease, and normal subjects (27). Here the majority of sera from patients with SLE or Sjogren's syndrome had significantly higher levels of IgG antibodies to human hsp 60 than to mycobacterial hsp 60. KindAs-Miigge and colleagues (31) also observed a similar (~ 20-30% frequency of IgG and IgM antibodies to hsp 72 and hsp 73 in patients with SLE and control subjects, with 2 of 47 patients with SLE exhibiting rather higher levels of IgM antibodies. Both of these latter studies probably are consistent with our own if one considers that the immunoblotting conditions that we used excluded the very low levels of anti-stress protein autoantibodies, that, like rheumatoid factor, may be present in essentially all human sera. In any case, it is clear that patients with autoimmune disease do not develop high-affmity/high-titer autoantibodies to stress proteins of the kind that are commonly seen to DNA and certain other nuclear antigens in SLE.
The ultimate significance of autoantibodies to stress proteins remains to be determined. Precise information concerning their diagnostic specificity, linkage with other autoantibody systems, relationship to disease activity status, and reactive epitopes is not available. Although anti-stress protein autoantibodies clearly do not appear to be disease-specific "markers," and their relatively low concentration and prevalence argues against a role in disease pathogenesis, these issues should remain open until examined more directly in acutely ill or early-onset patients. For example, certain stress proteins are spontaneously shed or selectively released (16,17,32) and theoretically could form immune complexes of potential significance in tissue injury. Perhaps of more interest is the possibility that stress proteins, such as hsp 70 or ubiquitin, may be ingredients in the "immunogenic particle" concept (33) of the origin of antinuclear and other autoantibodies. Thus, hsp 70 associates with proteins in the nucleus and nucleolus in a cell cycle-dependent fashion (34), migrates among different cell compartments, and binds to a variety of nuclear and cytoplasmic proteins in virus-infected cells and in cells exposed to other stressful stimuli (35-37). Consistent with this idea are recent data from Furakawa and colleagues demonstrating that induction of hsp 72 synthesis in cultured human keratinocytes by ultraviolet light or a cytotoxic prostaglandin is accompanied by increased binding of anti-UlRNP and anti-SS-A/Ro to cultured keratinocytes (38).
Was this article helpful?