Further, the role of T cells in the allergic reaction has been little explored, but T cells together with eosinophils have been regarded to be important for the late phase reaction [48]. Fenugreek anti-PD-1 monoclonal antibody exhibits properties that are inhibitory on all cytokine release, an effect evident both after in vivo and ex vivo exposure and in both models. The inhibitory effect on cytokines is in accordance with the suggested immunomodulatory properties of fenugreek [49, 50]. This makes it difficult to draw any conclusions based on the cytokine profile in the fenugreek model. In allergy testing

of humans the outcome may be that a number of IgE mediated serological reactions occur with no apparent clinical relevance [23, 51, 52]. In contrast, our mouse models showed clinical reactions sometimes without correlating serological responses, an event rarely seen in man. This could be related to the sensitivity or relevance of the laboratory tests, or it could be an expression of differences between man and mice, including a difference in allergen exposure. Mice live in a very controlled and sheltered environment and are essentially exposed to only one legume, the experimental one. Humans, on the other hand, are exposed to several different legumes from

early on in life, which makes co-sensitization Tyrosine Kinase Inhibitor high throughput screening a possible cause of apparent cross-allergic reactions [13, 52]. In the mouse models, we essentially have mono-sensitization and the observed reactions are thus true cross-allergic responses. In conclusion, we have in the mouse models shown clinically Celecoxib relevant cross-allergy between the four allergenic legumes, lupin, fenugreek, peanut and soy, reflected to some extent in serological and cellular tests. The effector immune mechanisms underlying cross-allergic reactions in mice and their relevance for man still remain to be fully elucidated.

Our models may prove valuable for the study of cross-allergy mechanisms and the role of individual allergen components. This study was financially supported by the Research Council of Norway, as part of the Strategic Institute Program (SIP) at the National Veterinary Institute lead by Eliann Egaas entitled “A coordinated research program into food allergen identification, quantification, modification and in vivo responses”. We thank Åse Eikeset, Else-Carin Groeng, Bodil Hasseltvedt, Berit A. Stensby and Astri Grestad for excellent technical assistance, and Lena Haugland Moen at the National Veterinary Institute for providing the food extracts. All authors declare no conflict of interest. Figure S1 IL-5, IL-10, IFN-γ and IL-2 responses in the two models are shown.