A pitfall in two-site sandwich 'one-step' immunoassay with monoclonal antibodies for the determination of human alpha-fetoprotein.

Utilizing monoclonal antibodies directed to 2 distinct antigenic determinants of the human alpha-fetoprotein (AFP), Uotila, Ruoslahti and Engvall developed the 2-site sandwich immunoassay. They found that due to the different specificities of monoclonal antibodies, 2 antigen-antibody reactions, AFP with immobilized antibody and AFP with labeled antibody, could be accomplished in a single step. We have found, however, that above a certain concentration, AFP detectable by their method decreased because the labeled antibody tended to bind with AFP that failed to react with the immobilized antibody. Consequently, 2 different AFP concentrations produced the same result, and an extremely high, still clinically expectable concentration gave a false negative result. Such an inhibition in high AFP concentrations was not observed in the conventional '2-step' immunoassay within the range of concentrations tested (0.1 ng-3 mg/ml). On the basis of these observations, the 2-site '1-step' immunoassay for AFP would have to be applied on multiple dilutions of the serum to avoid an erroneous interpretation of the results.