The antigenicity of four soya-bean-based infant formulas (Prosobee powder, Prosobee liquid concentrate (Mead Johnson, Uxbridge, Middx), Wysoy (Wyeth, Maidenhead, Berks) and Formula S (Cow and Gate, Trowbridge, Wilts] was measured by enzyme-linked immunosorbent assays (ELISAs) specific for glycinin and beta-conglycinin. Results were compared with in vivo assessments of antigenicity using guinea-pigs, rabbits and calves. 2. The levels of antigenic glycinin and beta-conglycinin in Wysoy and Formula S were below the limits of detection of the ELISA. Both these proteins were detected in Prosobee powder and Prosobee liquid concentrate with the highest levels, especially for glycinin, being present in Prosobee powder. 3. Wysoy was sufficiently antigenic to evoke a soya-bean-specific serum antibody response in rabbits injected with this formula emulsified in complete Freunds adjuvant. A significantly greater response was obtained when rabbits were similarly injected with Prosobee powder. 4. The formulas varied in their ability to sensitize guinea-pigs for both anaphylaxis and antibody production when given orally, although the differences were not statistically significant. Prosobee powder appeared to be the most antigenic and Formula S the least, with Prosobee liquid concentrate and Wysoy being intermediate. 5. Similar variations in antigenicity were observed when Prosobee powder, Wysoy and Formula S were fed to soya-bean-sensitive calves. These formulas were all capable of provoking intestinal disturbances (seen as increased ileal flow-rate, decreased small intestinal transit time and decreased nitrogen absorption) but the most severe reactions were seen when Prosobee powder was fed and the least with Formula S. 6. Thus the four soya-bean-based infant formulas showed considerable differences in antigenicity. In vivo studies using guinea-pigs, rabbits and calves were in good agreement and broadly correlated with the immunochemical assessment of antigenicity. However, the in vitro and in vivo results did not correspond exactly and levels of glycinin and beta-conglycinin below the limit of detection by ELISA could evoke an immune response in the different animal species. We believe that these variations in antigenicity of different commercial products prepared from isolated soya-bean protein may be important when interpreting the results from studies of the development of allergy in infants given soya-bean-based formulas.
