Sera of patients suffering from birch pollinosis were studied in the radio-allergo-sorbent test (RAST) for the presence of IgE antibodies to various allergens of vegetable origin. The sera selected were positive in the RAST for both birch pollen and fruits. IgE antibodies directed against at least three different cross-reacting determinants in birch pollen were detected. In addition to periodate-susceptible cross-reacting determinants, which are found on a number of glycoproteins, two non-related periodate-resistant determinants were found in birch pollen, with molecular weights of 20 and 18 kD, respectively. The 20-kD component appears to be responsible for the co-occurrence of the binding of IgE to allergens of fresh fruits, whereas the 18-kD component appears to cause the cross-reactivity among grass pollen, potato and fruits.

IgG4 antibodies to banana were found to occur far more frequently than expected. The most important antigen involved proved to be a lectin, BanLec-I. Because of the lectin nature of the antigen, it was important to establish the antibody nature of the lectin-IgG4 interaction and to exclude an interaction between the sugar-binding site of the lectin and glycosidic chains on IgG4. Three arguments in support of immune binding are: (1) the binding of BanLec-I to IgG4 is mannoside resistant, whereas the binding to all other glycoproteins tested is mannoside inhibitable; (2) only a minor fraction of the IgG4 in serum and none of five IgG4 myelomas tested was bound, and (3) the lectin binds to the Fab fragment of the IgG4 molecule. A curious finding was that in the presence of high-molecular-weight glycoproteins the interaction between IgG4 and BanLec-I was enhanced by α-methyl mannoside. The probable explanation of this phenomenon is that complexes of the lectin with high-molecular-weight glycoproteins by sterical interference inhibit the interaction with human IgG4 antibodies (or with rabbit antibodies to the lectin). This inhibition is prevented in the presence of α-methyl mannoside. These results support the earlier suggestion that some lectins are particularly prone to induce an immune response upon oral feeding. This banana lectin might be a potentially useful carrier protein for oral antihapten immunization in humans.

In the present investigation we have tested the hypothesis that children with a high IgG antibody response to foods have an increased risk of developing IgE antibodies to inhalant allergens. Sera from 106 children with an increased risk of developing IgE-mediated allergy were analysed. During the follow-up, in 54 of these children IgE antibodies to inhalant allergens appeared. A positive/negative IgG1 and IgG4 anti-food score was determined as described previously: sera from age-clustered unselected children were tested for the levels of IgG1 and IgG4 antibodies to common foods. For each IgG RAST and each age group, the 75-percentile was chosen as cut-off value. Each antibody level was thus converted into a positive (higher than the 75-percentile of the age group) or negative value. The number of positive tests was used as the score. High-risk children with a high IgG1 anti-food score more often developed inhalant-specific IgE antibodies than high-risk children with low IgG1 titres: 50% of the children with a high IgG1 anti-food score developed IgE antibodies to grass pollen. Fifty per cent of the children with a high and 14% of the children with a low IgG1 anti-food score developed IgE antibodies to cat dander. For the prediction of the development of IgE anti-mite (house dust mite), the IgG4 anti-food scores appeared less useful than the IgG1 anti-food scores; 46% of the IgG4 high responders versus 22% of the IgG4 low responders acquired IgE anti-mite, whereas for IgG1 these percentages were 73 and 19, respectively.