Journal of Autoimmunity (1997) 10, 193–201 Characteristic Generated Alterations of Autoantibody Patterns in Idiopathic Thrombocytopenic Purpura Anne Sundblad 1,2 , C. Ferreira 1,3 , A. Nobrega 1 , M. Haury 1 , E. Ferreira 3 , F. Padua 3 and A. Coutinho 1 1 Unite ´ d’Immunobiologie, CNRS URA 1961, Institut Pasteur, Paris, France 2 Section of Hematology and Immunology, Department of Medicine, Karolinska Hospital, Stockholm, Sweden 3 Departamento de Medicina IV, Hospital de Santa Maria, Lisboa, Portugal Received 7 August 1996 Accepted 12 November 1996 Using a Western blot technique that allows quantitative detection of antibody reactivities to a large number of antigens, serum IgG and IgM antibody repertoires were compared in a group of 19 patients with a diagnosis of idiopathic thrombocytopenic purpura (ITP) and respective healthy controls. The results show that, irrespective of the duration of thrombocytopenia, age of the patients, and type of therapy, all ITP donors share characteristic alterations of serum antibody reactivity patterns on homologous erythrocyte and liver antigens. Multiparametric analyses of the immunoreactivity data readily segregated the groups of ITP and healthy donors. Similar analyses also distinguished ITP sera from those of a group of patients with systemic lupus erythematosus (SLE). We conclude that ITP is an autoimmune disease associated with generalized alterations of antibody repertoires, that may be characteristic enough to allow for diagnosis. © 1997 Academic Press Limited Key words: antibody repertoires, autoimmune disease, ITP, multiparametric analysis, natural antibodies Introduction Recent views of the immune system, arising from the observation that normal individuals manifest abundant and productive autoreactivities, propose that the physiology and pathology of autoreactivity correspond to different global states of repertoire selection [1]. While previous models proposed that autoimmune disease (AID) is associated with either clonally restricted specificities or polyclonal activation of B cells, these views instead suggest that AID is characterized by generalized and yet disease-specific alterations of antibody repertoires. These three classes of model predict, therefore, distinct autoantibody pro- files associated with a particular AID. Clarification of this issue could give insights into pathogenic mech- anisms and be relevant for diagnostic purposes. Thus far, however, no available methods could score a large enough number of antibody reactivities to address this question. The relationships between the presence, titre and characteristics of autoantibodies and AID are com- plex, and in only a few clinical conditions can signs or symptoms be directly ascribed to the activity of autoantibodies. Hence the difficulties in serological diagnosis in this group of diseases. Pathogenic autoantibodies have been demonstrated in the group of anti-receptor antibody AID, although a complete correlation between antibody titres and disease is often missing. In myasthenia gravis, for example, autoantibodies to the acetylcholine receptor (AChR) are not detectable in about 20% of patients [2], and healthy individuals, particularly among a patient’s relatives, may carry high titres of such antibodies in the serum. In other autoimmune conditions, some diagnostic autoantibodies are only present in a frac- tion of patients [3], and may also be detected in healthy individuals or in a variety of unrelated patho- logical conditions. The latter point is well illustrated by antinuclear antibodies, taken to be the serological hallmark of systemic lupus erythematosus (SLE), which can be detected in normal individuals and in various inflammatory diseases [4]. Idiopathic thrombocytopenic purpura (ITP) is a condition that is currently considered to be of auto- immune origin, but its diagnosis is normally reached following the exclusion of other causes of thrombo- cytopenia [5, 6], and there is no general agreement as to the diagnostic value of antiplatelet antibodies [6, 7]. Furthermore, autologous platelet antibody assays are difficult to perform, as they require isolation of pure platelet populations in thrombocytopenic patients and the detection of limited amounts of antibody associated with platelets [5, 8]. We have developed a method for quantitative analysis of a large number of serum antibody reactivi- ties [9], which permits screening of a large section of the antibody repertoires with no a priori limitation in Correspondence to: Anne Sundblad, Unite ´ d’Immunobiologie, Insti- tut Pasteur, 25 rue du Dr Roux, 75724 Paris CEDEX 15, France 193 0896-8411/97/020193+09 $25.00/0/au960116 © 1997 Academic Press Limited