Operationally defined single‐ and double‐stranded DNA antigens in the Farr assay: diagnostic value

The specificity of the Farr assay for the detection of antibodies to double‐stranded (ds) DNA is highly dependent on the properties of the radiolabelled antigen. Since dsDNA is notoriously difficult to prepare free of single‐stranded (ss) regions, we have examined the feasibility of using in vitro 125 I‐labelled calf thymus DNA separated in predominantly ss and ds fractions upon hydroxyapatite chromatography as antigens in the ammonium sulphate assay. Characterization of these defined antigens showed for the ‘ss fraction’: a sedimentation coefficient of 7S, molecular weights between 0.05 and 0.25 times 10 6 daltons and the presence of 34% ss‐regions. The corresponding values for the ‘ds fraction’ were: 9.5S, 0.2‐1.5 times 106 daltons and 10% ss regions. Using each of these 125I‐labelled DNA fractions, the ammonium sulphate test was performed on a variety of sera from patients with rheumatic and immunological diseases. The results show that 80% of the sera from systemic lupus erythematosus (SLE) sera have dsDNA binding over 2 standard deviations of the normal mean and that this binding is always higher than that to the ss‐antigen. On the other hand, a minority of non‐SLE sera gave a slightly abnormal binding with the ds‐antigen but in contrast to SLE sera, these always had a higher binding value with the ss‐antigen. Furthermore, when the binding of SLE and of these non‐SLE sera was tested against a perfectly double‐stranded DNA preparation of Adenovirus serotype 2, SLE sera still gave high values, whereas the binding of non‐SLE sera was abolished. Taken altogether, these results suggest that these two operationally defined DNA antigens can, when used simultaneously, provide a specific serological diagnosis. These data also support the idea that the presence of true anti‐dsDNA binding in a human sera is an exclusive SLE feature.