The discussion of the possible pathophysiological role of plasma and/or serum nucleic acids from patients suffering from systemic lupus erythematosus is nearly identical with the still unsolved question regarding the antigen, i.e., the possible autoantigen-inducing antibodies against native double-strand DNA (dsDNA). This question is of special interest, since native dsDNA per se is not immunogenic. As repeatedly demonstrated, circulating antibodies of the isotype IgG against dsDNA can be correlated with the disease activity and, particularly, with renal involvement. Antibodies against native dsDNA were isolated from affected organs from SLE patients. The analysis of circulating immune complexes revealed dsDNA, as well as antibodies against dsDNA as complex components. DNA-anti-dsDNA serum immune complexes could also be demonstrated in patients with a so-called seronegative SLE, i.e., in patients not showing any free antibodies against native dsDNA in their sera. Furthermore, the involvement of antibodies against native dsDNA in pathogenic mechanisms of SLE patients becomes particularly evident from animal models. Regarding the induction of dsDNA antibodies, it is of special interest that in animal models such as the MLR/lpr mouse, anti-dsDNA antibodies are rather antigen-selected than they are a consequence of a "random" polyclonal B cell stimulation. Likewise, by the demonstration of somatic mutations of clonal human IgG-anti-dsDNA antibodies from SLE patients it has recently been possible to prove that these autoantibodies are also most probably antigen-selected.(ABSTRACT TRUNCATED AT 250 WORDS)