The limits of sensitivity of three gel diffusion methods are compared and their utilization in the detection of small amounts of antibody to antigens present in traces in a preparation is illustrated with the diphtheria toxin-human antitoxin system. The Preer modification of the Oakley-Fulthorpe technique and the Oudin tube method were found more sensitive than the Ouchteriony plate method, and permitted the detection of as little as 3 µg. antibody N/ml. of serum.
Antisera from eight Schick-negative individuals immunized with purified diphtheria toxoid have all been shown to contain, in addition to antitoxin, antibodies to substances present as impurities in the purified toxoid injected. The amounts of these antibodies in a serum and a partial characterization of their antigen-antibody curves have been determined through the combined use of quantitative precipitin and gel diffusion methods.
Different amounts of antibody have been precipitated by toxin and toxoid from individual sera. Evidence is presented that this may have been due to slight differences in antigenic specificity.
A serum, Hu, which had been held to contain no precipitating antibody, has now been shown, by the Preer and Oudin techniques, to contain at least 12 µg. of precipitating antibody N (per ml. serum) against an impurity in the toxoid preparation. This estimate has been confirmed by quantitative precipitin determinations. The presence of antibodies to impurities in all human antitoxins examined in the present work brings into question the assumption that human antitoxin as such has a skin-sensitizing capacity.