Cells were seeded in 24-well plates and stimulated with CPE or soy extract. Cells were recovered at several time points for RNA isolation with the RNeasy Mini Kit (Qiagen).Bone marrow-derived DCs. Bone marrow cells were collected from femurs and cultured for 10 days in RPMI 1640 with l-glutamine (Gibco, Life Technologies), supplemented with 10% fetal bovine serum (Gemini Bio-products), penicillin/streptomycin (Gibco, Life Technologies), 50 μM 2-Mercaptoethanol (Sigma-Aldrich), and 20 ng/ml GM-CSF (Peprotech). At day 10, bone marrow-derived DCs were collected and resuspended in complete RPMI at a concentration of 5 × 10 5 cells per ml. Cells were stimulated for 3 hours with CPE (50 μg/ml), followed by RNA isolation with the RNeasy Mini Kit (Qiagen).Real-time PCR. RT-PCR was performed, starting from 1 μg total RNA, using SuperScript II Reverse Transcriptase (Invitrogen, Life Technologies). cDNA was amplified using the Power SYBR Green PCR Master Mix (Applied Biosystems, Life Technologies) and run on an Applied Biosystems 7300 Real-Time Detection System, using the following primers: mouse Il6, F, TTCCATCCAGTTGCCTTCTTG; R, GGGAGTGGTATCCTCTGTGAAGTC; mouse Il1b (65), F, TTGACGGACCCCAAAAGAT; R, GAGCGCTCACGAACAGTTG; mouse Il10 (25), F, TGCTATGCTGCCTGCTCTTA; R, TCATTTC-CGATAAGGCTTGG; mouse Ox40l (25), F, CCCTCCAATCCAAA-GACTCA; R, ATCCTTCGACCATCGTTCAG; mouse Il33, F, ATC-GGGTACCAAGCATGAAG; R, GACTTGCAGGACAGGGAGAC; mouse Tslp (48), F, GAGAGAAATGACGGTACTCA; R, CTACAGT-TAGGTTTGCCCTA; mouse Il25, F, TGTTGCATTCTTGGCAAT-GATC; R, GACTGCAGCCCTCCTGGAT; human IL6, F, AAA-GAGGCACTGGCAGAAAA; R, CAGGGGTGGTTATTGCATCT; human IL33, F, GAGCTAAGGCCACTGAGGAA; R, TGGGCCTTT-GAAGTTCCATA.GADPH and β-actin were used as the housekeeping genes. Relative quantification was performed using the comparative threshold cycle method (2 -ΔΔCt ). The changes in gene expression were calculated with respect to the untreated cells. All amplifications were carried out in duplicates. Antigen presentation assay. Antigen presentation assay was performed as previously described (19). BALB/c mice were exposed to 10 mg OVA in the presence or absence of 1 mg CPE. After 24 hours, DCs were purified from inguinal lymph nodes by using CD11c microbeads (Miltenyi Biotec). DCs were cultured at a ratio of 1:5 with DO11.10 CD4 + T cells. After 72 hours, cells were restimulated with anti-CD3/CD28, supernatants were harvested, and cytokines were measured by ELISA according to manufacturer's instructions (all from eBioscience).In neutralization experiments, anti-ST2 was injected prior to exposure with OVA and CPE, as described above. For the OX40L neu- The Journal of Clinical InvestigationR e s e a R c h a R t i c l e 4 9 7 4

This review searched for published evidence that could explain how different physicochemical properties impact on the allergenicity of food proteins and if their effects would follow specific patterns among distinct protein families. Owing to the amount and complexity of the collected information, this literature overview was divided in two articles, the current one dedicated to protein families of plant allergens and a second one focused on animal allergens. Our extensive analysis of the available literature revealed that physicochemical characteristics had consistent effects on protein allergenicity for allergens belonging to the same protein family. For example, protein aggregation contributes to increased allergenicity of 2S albumins, while for legumins and cereal prolamins, the same phenomenon leads to a reduction. Molecular stability, related to structural resistance to heat and proteolysis, was identified as the most common feature promoting plant protein allergenicity, although it fails to explain the potency of some unstable allergens (e.g. pollen-related food allergens). Furthermore, data on physicochemical characteristics translating into clinical effects are limited, mainly because most studies are focused on in vitro IgE-binding. Clinical data assessing how these parameters affect the development and clinical manifestation of allergies is minimal, with only few reports evaluating the sensitising capacity of modified proteins (addressing different physicochemical properties) in murine allergy models. In vivo testing of modified pure proteins by SPT or DBPCFC is scarce. At this stage, a systematic approach to link the physicochemical properties with clinical plant allergenicity in real life scenarios is still missing.

Key determinants for the development of an allergic response to an otherwise 'harmless' food protein involve different factors like the predisposition of the individual, the timing, the dose, the route of exposure, the intrinsic properties of the allergen, the food matrix (e.g. lipids) and the allergen modification by food processing. Various physicochemical parameters can have an impact on the allergenicity of animal proteins. Following our previous review on how physicochemical parameters shape plant protein allergenicity, the same analysis was proceeded here for animal allergens.We found that each parameter can have variable effects, ranging on an axis from allergenicity enhancement to resolution, depending on its nature and the allergen. While glycosylation and phosphorylation are common, both are not universal traits of animal allergens. High molecular structures can favour allergenicity, but structural loss and uncovering hidden epitopes, can also have a similar impact. We discovered that there are important knowledge gaps in regard to physicochemical parameters shaping protein allergenicity both from animal and plant origin, mainly because the comparability of the data is poor. Future biomolecular studies of exhaustive, standardized design together with strong validation part in the clinical context, together with data integration model systems will be needed to unravel causal relationships between physicochemical properties and the basis of protein allergenicity.