drug discovery today disease models€¦ · drug discovery today: disease models vols. 17–18,...

DRUG DISCOVERY

TODAY

DISEASEMODELS

In vivo and in vitro models of food allergyEdited by Michelle Epstein

Editorial and introduction by Kitty Verhoeckx, Liam O’Mahony and Michelle M. Epstein 1

In silico tools for exploring potential human allergy to proteinsM. Hayes, P. Rougé, A. Barre, C. Herouet-Guicheney and E.L. Roggen ........................................................................ 3

Applicability of epithelial models in protein permeability/transport studies and food allergyN. Cubells-Baeza, K.C.M. Verhoeckx, C. Larre, S. Denery-Papini, M. Gavrovic-Jankulovic and A. Diaz Perales ........ 13

Static and dynamic in vitro digestion models to study protein stability in the gastrointestinal tractD. Dupont and A.R. Mackie ........................................................................................................................................... 23

Epithelial models to study food allergen-induced barrier disruption and immune activationM. Gavrovic-Jankulovic and L.E.M. Willemsen ............................................................................................................. 29

IgE – the main player of food allergyH.C.H. Broekman, T. Eiwegger, J. Upton and K.L. Bøgh ............................................................................................... 37

Non-IgE mediated food allergyD. Lozano-Ojalvo, G. Lezmi, N. Cortes-Perez and K. Adel-Patient ............................................................................... 45

Experimental food allergy models to study the role of innate immune cells as initiators of allergen specifi c Th2 immune responses

M. Hussain, M.M. Epstein and M. Noti ......................................................................................................................... 55

The use of animal models to discover immunological mechanisms underpinning sensitization to food allergensJ.J. Smit, M. Noti and L. O’Mahony .............................................................................................................................. 63

Infl uence of microbiome and diet on immune responses in food allergy modelsW. Barcik, E. Untersmayr, I. Pali-Schöll, L. O’Mahony and R. Frei ............................................................................... 71

A review of animal models used to evaluate potential allergenicity of genetically modifi ed organisms (GMOs)N. Marsteller, K.L. Bøgh, R.E. Goodman and M.M. Epstein ......................................................................................... 81

Drug Discovery Today: Disease Models Vols. 17–18, 2015

Editors-in-ChiefJan Tornell – AstraZeneca, SwedenAndrew McCulloch – University of California, San Diego, USA

Contents

DOI: 10.1016/S1740-6757(16)30014-7 www.drugdiscoverytoday.com i

DRUG DISCOVERY

TODAY

DISEASEMODELS

DRUG DISCOVERY

TODAY

DISEASEMODELS

EDITORIAL

Editorial and introduction by KittyVerhoeckx, Liam O’Mahony andMichelle M. Epstein

Drug Discovery Today: Disease Models Vol. 17–18, 2015

Editors-in-Chief

Jan Tornell – AstraZeneca, Sweden

Andrew McCulloch – University of California, SanDiego, USA

In vivo and in vitro models of food allergy

Food allergies affect up to 6% of children and 4% of adults in

Western countries and are associated with a large impact on

health and well-being, the food sector and society. Food

allergy is a consequence of an inappropriate immune re-

sponse to ingested proteins. The most common foods pro-

voking allergy include peanuts, tree nuts (e.g., pecans,

walnuts, almonds), fish, shellfish, eggs, milk, wheat, soy

products, and fruit and vegetable pollen (oral allergy syn-

drome). These and other less common foods may trigger

shortness of breath, throat tightness and hoarseness, wheez-

ing, tongue swelling, difficulty swallowing, vomiting, ab-

dominal pain, dizziness, feeling faint, hives, eczema, and

anaphylaxis which may be fatal without immediate treat-

ment with epinephrine. These symptoms occur from minutes

to hours following ingestion. However, there are delayed

food allergies that cause gastrointestinal reactions, e.g., food

protein-induced enterocolitis syndrome (FPIES) in response

to foods such as milk and soy.

There is no cure for food allergies and there is no clear

approach for prevention. Food allergy is treated symptomati-

cally and prophylactically by avoiding the food containing the

allergen. In some cases, allergies, especially to eggs, milk,

wheat and soy may disappear over time in children. The

problem with avoidance is that sometimes there is cross-

reactivity between foods, which was unrecognized and there

may be hidden ingredients in processed foods and in restau-

rant meals that might elicit allergic responses. In the last

decade, however, required labeling of common food allergens

on food packaging and in restaurants such as gluten contain-

ing cereals, eggs, fish, milk, tree nuts, celery, mustard, sesame,

and sulphur dioxide has significantly aided food allergic indi-

viduals avoid offending allergens, but it is not always effective.

1740-6757/$ � 2016 Published by Elsevier Ltd. http://dx.doi.org/10.1016/j.ddmod.2016.11.0

It is not known why a person is susceptible to food allergy

and how they are caused. Although food allergy may be more

common in certain families, the genetic factors underlying

food allergy remain unknown. There are intrinsic character-

istics of food proteins that may facilitate their allergenicity in

susceptible individuals, such as nonspecific lipid transfer

proteins (LTPs), which are well known allergens. There are

additional factors that may associate with the risk of devel-

oping an allergic response, such as certain food processing

methods and lifestyle, e.g., obesity is considered a risk factor

for food allergy. Thus, research to understand the mecha-

nisms underlying food allergy is necessary especially as the

global population is growing and dietary habits are changing

(e.g., more meat and less fiber consumption), which may

consequently lead to protein containing food shortages.

Thus, there is a move towards introducing new foods from

sustainable and climate-resistant crops and other food pro-

teins like those derived from insects to meet the demand.

However, with the introduction of new proteins to the food

market, a comprehensive risk assessment is required to eval-

uate the nutritional, microbial, toxicological and allergenic

risks.

Current allergenicity (the ability of a food protein to

induce allergy) risk assessment strategies used by governmen-

tal agencies like European Food Safety Agency (EFSA) were

originally designed for GMOs (genetically modified organ-

isms), but are also used to evaluate novel non-GMO foods. For

instance, the EFSA ‘weight of evidence approach’ assesses the

impact of a food on individuals who have pre-existing aller-

gies via cross-reactivity and elicitation of allergic responses,

but is unable to assess the potential risk of causing new

allergies. Thus, there is a need for predictive tests to assess

02 1

http://crossmark.crossref.org/dialog/?doi=10.1016/j.ddmod.2016.11.002&domain=pdf

http://dx.doi.org/10.1016/j.ddmod.2016.11.002



Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015

allergenicity of novel and newly processed proteins to ensure

safety and facilitate the introduction of new protein sources

onto the market. However, this is challenging because of the

absence of accepted, predictive and validated methods for

allergy hazard and risk assessment and thus, the reason for

increased research efforts and the creation of a large interdis-

ciplinary European network, COST Action ImpARAS

(FA1402) which consists of more than 250 scientists from

29 countries with a broad range of expertise who are actively

using a variety of models to study food allergy.

Short introduction

This issue is a comprehensive overview of in silico, in vitro and

in vivo models currently used for studying food allergy and

predicting allergenicity. Included are (1) a review of in silico

tools, (2) a selection of in vitro models of food allergy such as

epithelial models in protein transport and immune modula-

tion, static and dynamic in vitro digestion models, and IgE

tests, and (3) in vivo animal models addressing adaptive

immune responses to sensitisation to food allergens, non-

IgE-mediated food allergy, interaction between the innate

immune system and Th2 immune-mediated food allergy, and

a review of animal models used to evaluate potential allerge-

nicity of genetically modified organisms (GMOs).

In ‘In silico tools for exploring potential human allergy to

proteins’ explores how bioinformatics assists our understand-

ing of food allergy. The review highlights bioinformatic tools

aiming to identify food allergens, cross-reactivity with exist-

ing allergens and identifying whether the allergic IgE anti-

body will bind new proteins in food that will potentially

cause allergy.

There are four reviews on in vitro models including (1)

Applicability of epithelial models in protein permeability/

transport studies and food allergy, (2) Static and dynamic in

vitro digestion models to study protein stability in the gas-

trointestinal tract, (3) Epithelial models to study food aller-

gen induced barrier disruption and immune activation, and

2 www.drugdiscoverytoday.com

(4) IgE – the main player of food allergy. These reviews address

how digestion of food might play a role in the generation of

food allergens, the role of epithelial models of protein per-

meability and transport in understanding food allergy, how

in vitro models of human intestinal epithelial cells and co-

culture models examine barrier disruption and immune acti-

vation induced by food allergens and, how protein stability in

the gastrointestinal tract using static and dynamic in vitro

digestion models focus on the role of proteins surviving

digestion in the induction of allergic reactions, and due

how novel approaches to measuring IgE, its allergen binding

sites and functionality aims to discriminate between asymp-

tomatic and symptomatic sensitisation and distinct allergic

phenotypes.

The use of animal models are highlighted in five

reviews on (1) non-IgE-mediated food allergy, (2) experi-

mental food allergy models to study the role of innate

immune cells as initiators of allergen specific Th2 immune

responses, (3) influence of microbiome and diet on immune

responses in food allergy models, (4) the use of animal

models to discover immunological mechanisms underpin-

ning sensitization to food allergens, and (5) a review of

animal models used to evaluate potential allergenicity of

genetically modified organisms (GMOs) to better under-

stand the complex immunological and pathophysiological

mechanisms of food allergies in a way that is not possible

with in vitro models. The main aim of in vivo models is to

determine the role of the immune system as a qualitative

readout for the sensitizing potential and risk assessment of

food proteins. The focus on the in vivo animal model reviews

is on many aspects of the allergic immune response to food,

such as the adaptive and innate immune system, IgE and

non-IgE-mediated food allergies, the influence of micro-

biome and diet, the use of in vivo models for testing new

approaches for prophylaxis and treatment of food allergy,

and the use of animal models for the evaluation of allerge-

nicity of GMOs.

DRUG DISCOVERY

TODAY

DISEASEMODELS

In silico tools for exploring potentialhuman allergy to proteinsMaria Hayes1,*, Pierre Rouge2, Annick Barre2,3,

Corinne Herouet-Guicheney4, Erwin L. Roggen5

1Teagasc, The Irish Agricultural and Food Development Authority, Food BioSciences Department, Ashtown, Dublin 15, Dublin, Ireland2Universite de Toulouse, UPS, IRD, UMR 152 PharmaDev, Universite Toulouse 3, Faculte des Sciences Pharmaceutiques, 31062 Toulouse

cedex 09, France3Paul Sabatier University – Toulouse II, Toulouse, France4Bayer SAS, Human and Animal Safety Assessment – Seeds, 355 rue Dostoievski, 06903 Sophia Antipolis, France53Rs Management and Consulting ApS, Denmark


Editors-in-Chief




Bioinformatics can help scientists to develop hypothe-

ses about proteins that may need to be tested further

for risks of causing allergy. In silico methodologies and

tools like databases and comparison software, play an

important role in the assessment of protein allergenic-

ity and allergenicity mechanisms. They can identify

whether a novel protein is an existing allergen and/or

has the potential to cross-react with an existing aller-

gen. They cannot identify whether a novel protein will

‘become’ an allergen. AllergenOnline is the tool cur-

rently used for the safety assessment of novel proteins,

but other tools are also available including the Struc-

tural Database of Allergenic Proteins (SDAP) and

AllerTOP. Information concerning PeptideRanker, as

well as the Hydrophobic Cluster Analysis (HCA) meth-

od used for identifying IgE-binding epitopes in food

allergens is discussed.

*Corresponding author: M. Hayes ([email protected])

1740-6757/$ � 2016 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ddmod.201

Section editors:Kitty Verhoeckx – K.C.M. Verhoeckx, TNO, Zeist,The NetherlandsLiam Mahony – Swiss Institute of Allergy and AsthmaResearch, University of ZurichMichelle Epstein – Medical University of Vienna,Department of Dermatology, DIAID, ExperimentalAllergy, Waehringer, Vienna, Austria

Introduction

Allergenicity is the potential of any material to cause sensiti-

zation and allergic reaction and is frequently associated with

the IgE antibody [1]. An existing allergy/allergen is a real and

immediate risk [2,3]. Allergens represent a small fraction of

the proteins that humans are routinely exposed to. The

reason why these proteins can cause T- and B-cell responses

remains largely unanswered. Furthermore, a sensitized indi-

vidual may respond to proteins that share certain structural

features with the protein that elicited the initial immune

reaction – a phenomenon known as cross-reactivity.

In silico methodologies can identify whether a novel protein

is an existing allergen or whether the novel protein has

potential to cross-react with an existing allergen. However,

they cannot identify whether a novel protein will ‘become’ an

allergen [2]. Data produced from the use of in silico methodol-

ogies may be used to make a decision about whether additional

6.06.001 3




mailto:[email protected]



in vitro and in vivo testing is required, by serum screening, as

recommended by Codex Alimentarius Commission (2009)

and Goodman, 2008 [2,4]. In practice, several in silico meth-

odologies for determination of protein allergenicity compare

amino acid sequences from a novel, trait protein to known

food, contact, and respiratory allergenic proteins found in

allergen databases [3].

State of the art – methods and tools currently used for

allergenicity assessment

According to the most recent guidelines on the allergenicity

evaluation of proteins, a novel protein should have a mini-

mum of 35% sequence identity over a window of 80 amino

acids when compared with known allergens to be considered

a potential allergen [4,5]. This is a very conservative approach

when we take into account the high degree of sequence

identity that is needed for actual cross-reactivity which is

often in excess of 50–60% sequence similarity, over signifi-

cant spans of the target protein [6].

AllergenOnline (www.allergenonline.org) focuses on se-

quence identity matches. It provides a detailed description

of accepted bioinformatics comparisons on the website. Pre-

viously, Siruguri et al. and Moran et al. have used AllergenOn-

line for regulatory comparisons [7,11,12]. AllergenOnline

provides access to a peer reviewed allergen list and a sequence

searchable database (FASTA) [7]. It is used for the identifica-

tion of proteins that may present a potential risk of allergenic

cross-reactivity. AllergenOnline is used currently by industry

for the risk assessment of genetically modified food including

proteins. The robust allergen database is updated annually by

a panel of independent scientists and clinicians.

Real health risks come from inclusion of proteins in a new

food, that are allergens from another source or highly likely to

be cross-reactive. A much lower risk is presented by the likeli-

hood that a protein will become an allergen de novo, or sensitize

de novo and lead to allergic sensitization [15]. This may be

indicated by stability in pepsin, abundance and thermal sta-

bility, but these factors could be important in elicitation not

sensitization. (Where does sensitization occur? Gut, skin,

mouth, airway). In using sequence comparisons, if the protein

is found to have been described previously as an allergen (100%

or nearly 100% identity), that is a significant risk (weight). If a

protein has high sequence identity (50–70+%, it suggests the

risk of probable cross-reactivity and would require serum IgE

tests with properly targeted allergic human sera. If >35%

identity over 80 or more amino acids between a novel and

existing protein is found, that is considered a potential allergy

risk by Codex [4] and should be evaluated further by serum IgE

testing if a proper set of serum donors can be identified (which

can be challenging for rarely reported allergenic sources).

In the past, several researchers also used step-wise contig-

uous identical amino acid segment searches (i.e. 6- and later

8-mer searches), as described in the FAO/WHO guidelines


[5,8] to predict human allergenicity to proteins, based on the

idea that these segments represented both a theoretical B-cell

epitope as well as a minimum size for a conserved T-cell

epitope. For instance, Stadler and Stadler [14] reported that

a 6-mer match resulted in more than two-thirds of all proteins

in SwissProt being predicted to be allergens, and >40% of the

human genome being predicted as such. This was confirmed

in other studies and as such, this approach was not seen as a

reliable criterion for predicting allergenic potential [10–12].

In the past, immunologists have tried to correlate ‘known’

and ‘predicted’ B cell and T cell epitopes with allergens,

compared to non-allergens or weak-allergens, and failed to

be able to develop solid predictions or clusters for allergy.

Unfortunately, the ideas outlined by Ladics [4], have not

come to fruition.

Overall, this comparison methodology of 35% identity over

at least 80 amino acids is considered to be useful for the

prediction of potential cross-reactivity with known allergens,

but also produces a number of false positive results. The predic-

tive value of sequence similarity searches for allergenicity po-

tential should be carefully deliberated using a weight of

evidence approach as no single method can be fully predictive

[18]. Moreover, a relatively high degree of identity at the amino

acid sequence level, as commonly seen between IgE cross-reac-

tive proteins, cannot guarantee that the protein is a cross-

reactive allergen [9,13]. In other words, no perfect correlation

exists between these in silico results and food allergenicity.

Protein families containing known allergens

The databases used in assessment of potential protein allerge-

nicity or cross reactivity should be composed of protein

sequences based on key criteria like the recognition of aller-

gens by IgE (food allergenicity marker), which involves bind-

ing to linear or conformational epitopes on allergen surfaces,

and should be proven by clinical data in humans. These

protein sequence databases should be updated regularly as

new allergens are discovered every year.

Ideally, the molecular basis of protein allergenicity should

also be studied through analysis of its sequence, structure and

B- or T-cell epitopes where they relate to allergenicity [4] but

these data are often missing for most of the known allergen

databases. Furthermore, B and T-cell epitope search tools may

not be able to distinguish between immunogenicity and

allergenicity.

Important allergenic protein families include the non-spe-

cific lipid transfer proteins (nsLTPs), the 2S albumins, and the

cupin superfamily containing the 11S and 7S globulins [19].

The nsLTP proteins account for severe allergic reactions and

are found in fruits from the Rosaceae family (peaches and

apples), pollen, tree nuts, vegetables and peanuts [20]. Pepsin

stability of proteins may be due to secondary and tertiary

structural features. For instance, the presence of disulfide

bridges is known to stabilize the protein structure. This is

http://www.allergenonline.org/

Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy

the case for the 2S albumin family, which has a 3D structure

containing four disulphide bridges. Furthermore, the abun-

dance should be taken into account as most of the plant

allergens are seed storage proteins, like the 7/8S and 11S

globulins that are major components in seeds from dicotyle-

donous species [20]; However no clear criteria exists to define

how much is too much. There is little sequence similarity

between cupins and globulins although they share a similar

fold, thus assessment of cross-reactivity is sometimes limited

[20]. Caseins, parvalbumins and tropomyosins are found in

dairy products, fish and crustaceans, molluscs and meat

respectively [20]. There are many proteins in these families

that have never been associated with allergy and this could be

due to: (1) the broadness of the family designations, (2) there

has been little or no exposure to these proteins and (3) the

overall structure and sequence similarities are not sufficiently

definitive in a biological sense [17].

New methodologies, new perspectives

Figure 1 illustrates these and links them to potential human

allergenicity and immunogenicity to protein prediction

steps. Assigning proteins as allergens may involve assessment

of their amino acid and dipeptide composition using support

vector machines (SVMs) [21,22]. Other methods could in-

clude motif-based techniques using the software MEME/

MAST and comparison algorithms with ‘Allergen Represen-

tative Proteins’ (ARPs) [23]. In silico methods for identification

of B-cell epitopes could include hydrophilicity scans, amino

acid property assessment and combinations of both methods.

Computational prediction methods for prediction of peptide

binding to human leucocyte antigen (HLA), which is a pre-

requisite for T-cell recognition, are based on binding motifs,

quantitative matrices or artificial intelligence methods and

can reduce the number of experiments required to identify

relevant T-cell epitopes [24,25]. However, to date, there has

not been any demonstration that these new models out-

perform a FASTA sequence comparison with a well-developed

allergen database using criteria of >35–40% identity over 80

amino acids. For the most part, the value of predictions made

using these databases depends upon the dataset.

Hydrophobic Cluster Analysis (HCA) as a relevant tool for predicting

the IgE-binding epitope regions in food allergens

The amino acid residues forming the IgE-binding epitopes

exposed on the surface of allergenic proteins usually share a

set of physico-chemical characteristics that can be used for

predicting the potential immunogenicity and allergenicity of

food proteins. These characteristics mainly consist of (1) the

hydrophilicity, due to the occurrence of polar residues (Asn/

N, Gln/Q, His/H, Ser/S, Thr/T, Tyr/Y residues), (2) the elec-

tronegative (Asp/D and Glu/E residues) and/or electropositive

characteristics (Arg/R and Lys/K residues) of residues and (3)

the flexibility of residues (Gly/G, Ser/S, Thr/T residues) [26].

Owing to the combination of the physico-chemical charac-

teristics of their building residues, most of these epitopes

coincide with loops, which often protrude from the surface of

the allergenic proteins. However, other secondary structural

features like strands of b-sheet or a-helix, can be readily

exposed on the surface and thus, participate in the IgE-

binding of food allergens.

Recently, researchers used hydropathic profiles based on

different scales of hydrophilicity/hydrophobicity, flexibility

and solvent exposure to predict the linear IgE-binding epi-

topes of allergenic proteins, either coupled with an epitope-

mapping approach or structural analysis. However, hydro-

pathic profiles suffer from inherent limitations with respect

to structural information which render them unsuitable for

the structural characterization of the predicted epitopes on

the surface of the allergens. In this respect, HCA offers an

efficient tool [26], allowing association of the predicted epi-

topes to structural features. The prediction of IgE-binding

epitopes with HCA was successfully applied to Pru p 3 and

Mal d 3, the nsLTPs from peach and apple fruits [26].

HCA was also recently applied to Sal s 1, the salmon (Salmo

salar) parvalbumin allergen, Jug r1, the English walnut

(Juglans regia) 2S albumin allergen, Pru p 3, the peach (Prunus

persica) lipid transfer protein and Pis v 1, the pistachio (Pis-

tacia vera) 2S albumin allergen. YASARA [27] was used to build

the three-dimensional models of the proteins. The three-

dimensional structure of Pru p 3 (PDB code 2ALG) was used.

The IgE-binding epitopes identified on Sal s 1, Jug r 1, Pru p 3,

and Pis s 1 were mapped on the molecular surface of the

corresponding allergens. Molecular surface cartoons were

drawn with Chimera. The HCA profiles of Sal s 1, Jug r 1,

Pru p 3, and Pis v 1, were drawn from the drawhca server

(http://bioserv.rpbs.univ-paris-diderot.fr/services/HCA/).

Segments of the HCA profiles were predicted as putative

continuous IgE-binding epitopes when they fulfilled at least

three out of the four following criteria: (1) exposure to the

solvent, (2) flexibility (Gly, Ser, Thr, His residues), (3) preva-

lence of hydrophilic residues (Asn, Gln, His, Ser, Thr, Tyr),

and (4) occurrence of electropositive (Arg, Lys) and/or elec-

tronegative (Asp, Glu) residues. As shown in Fig. 2 most of the

linear IgE-binding epitopes identified on Sal s 1, Jug r 1, Pru p

3 and Pis v 1, were correctly predicted on the HCA profiles of

the corresponding allergens. Both the predicted and identi-

fied epitopic stretches overlapped significantly. However,

some discrepancies were found, which related to (1) the

extent of the IgE-binding epitopic stretch, which is often

under-estimated, and (2) the prediction of extra-epitopes,

which have no counterparts among the IgE-binding epitopes

immunochemically identified on the molecular surface of the

allergens. This is the case for the HCA profiles of Sal s 1 and Pis

v 1 allergens, which exhibit an additional epitopic stretch at

the C-terminal end of the sequence. In spite of these dis-

crepancies, the critical analysis of the HCA profiles provides a

www.drugdiscoverytoday.com 5

http://bioserv.rpbs.univ-paris-diderot.fr/services/HCA/


Sensitization

Prediction of B-cell epitopes

DiscoTope http://www.cbs.dtu.dk/services/DiscoTope

ElliPro http://tools.immuneerpilcope.org/tools/ElliPro/iedb-input

PEPITO/BEPro http://pepto.proteomics.ics.uci,edu/

SEPPA http://lifecenter.sgst.cn/sepper/inex.php

Epitopia http://epitopia.tau.ac.il/

EPLES http://sysbio.unl.edu/services/EPCES/

EPSVR & EPMeta http://sysbio.unl.edu/services/

FASTA http://www.ebi.ac.uk/Tools/sss/fasta/

BLAST http://blast.ncbi.nlm.gov/Blast.cgi

ClustalW http://www.ebi.ac.uk/Tools/msa/clustalw2/

PIMA https://www.google.ie/?gws rd=ssl#g=PIMA+alignme

nt&safe=active&start=10

ExPASy http://www.expasy.ch

PredictProtein http://bioinf.cs.ucl.ac.uk/psipred/

Jpred2 http://www.compbio.dundee.ac.uk/ipred/

PAIRCOIL http://paircoil2.csail.mit.edu/

COILS http://www.ch.embnet.org/software/COILS form.html

PSORT http://psort.hgc.ip/

SYFPEITHI http://www.syfpeithi.de/

MULTIPRED http://antigen.i2r.a-star.edu.sg/multipred/

TEPITOPE http://www.bioinformation.net/ted/

VAGAT http://sdmc.i2r.a-star.edu/sg/vagat/

EpiDock http://bioinfo-pharma.u-strasbg.fr/cheminformatics-tools.php

PAProC http://www.paproc.de/

NetChop http://www.cbs.dtu.dk/services/NetChop/

PREDTAP http://antigen.i2r.a-star.edu.sg/predTAP/

IEDB http://www.immuneepitope.org/tools/do

EpiJen http://www.jenner.ac.uk/EpiJen/

Sequence structure andpattern analysis

Prediction of immunogenicity(coil structure and localisation in cells)

Prediction of T-cell epitopes

Gene expressionanalysis

Prediction of corss-reactivity

Drug Discovery Today: Disease Models

Figure 1. In silico tools that may help predict cross-reactivity potential.

rather accurate tool for the prediction of the IgE-binding

epitopes of the food allergenic proteins, since the regions

in which they occur have been rather correctly predicted.

Three-dimensional (3-D) structure of allergens

The allergenicity potential of proteins may also be identified

by using 3-D structure when conformational epitopes are

engaged in the allergenic reaction. Linear epitopes can be

identified with FASTA and BLAST. Sequence identity using

FASTA/BLAST is useful for predicting potential cross-reactivi-

ty (depending on the cut-off) as a 3-D structural prediction. In

fact, most structural predictions for proteins that have not

been tested by crystallography have had to have high FASTA

or BLAST alignments to ensure predictions that were accu-

rate. For risk assessment, the suggested program and link is

interesting http://scanmail.trustwave.com/?c=6600&d=

mcja11FmSSkm7qfkpOzDr5P9z6uTfrk8vKFVfMEU2w&s=61

&u=http%3a%2f%2fwww-bionet%2esscc%2eru%2fpsd%

2fcgi-bin%2fprograms%2fAllergen%2fallergen%2ecgi.


However, a general structural feature of allergens that causes

allergenicity has not been described up to now. Allergenicity

prediction methods require information about the 3-D struc-

ture of query protein; thereby considerably restricting analy-

sis to only those proteins whose 3-D structure is known. As a

consequence, many proteins with unknown structure could

be overlooked. A new method for allergenicity prediction was

developed using information on protein 3-D structure [28].

Three-dimensional structures of known allergenic proteins

were used for representing protein surface as patches desig-

nated as discontinuous peptides. Allergenicity was predicted

by searching for these peptides in query protein sequences. It

was demonstrated that the information on the discontinuous

peptides may help to predict more accurately potential hu-

man allergenicity to protein. The method is available at

http://www-bionet.sscc.ru/psd/cgi-bin/programs/Allergen/

allergen.cgi [28].

Many freely accessible websites offer comparison tools

associated with allergen databases (Table 1).

http://scanmail.trustwave.com/?c=6600%26d=mcja11FmSSkm7qfkpOzDr5P9z6uTfrk8vKFVfMEU2w%26s=61%26u=http%3a%2f%2fwww-bionet%2esscc%2eru%2fpsd%2fcgi-bin%2fprograms%2fAllergen%2fallergen%2ecgi




http://www-bionet.sscc.ru/psd/cgi-bin/programs/Allergen/allergen.cgi

http://www-bionet.sscc.ru/psd/cgi-bin/programs/Allergen/allergen.cgi


Sal s 1

Jug r 1

Pri p 3

Pis v 1

2

4

1 3


Figure 2. Identified linear IgE-binding epitopes identified on Sal s 1,

Jug r 1, Pru p 3 and Pis v 1 predicted using HCA profiles of the

corresponding allergens.

PREAL: prediction of allergenic protein by maximum relevance

minimum redundancy feature selection

PREAL (http://gmobl.sjtu.edu.cn/PREAL/index.php) predicts

potential human allergenicity to protein by integrating vari-

ous protein properties, including the physicochemical and

subcellular locations, using the Maximum Relevance Mini-

mum Redundancy (mRMR) and Incremental Feature Selec-

tion (IFS) procedures [29]. The mRMR method was developed

to rank each feature according to its relevance to the target

and redundancy with other features [30]. IFS procedures were

adopted to perform feature selection for analysing the key

properties of allergenicity.

Similarities are studied by using NCBI-BLAST software.

SSpro/ACCpro 4.03 [31] is used to predict secondary struc-

tures of proteins. Solubility is predicted by using the Protein

Structure and Structural Feature Prediction Server (SCRATCH;

http://download.igb.uci.edu/). The physicochemical proper-

ties based on (1) amino acid composition (2) molecular

weight (3) hydrophobicity (4) polarizability (5) normalized

van der Waals volume and (6) polarity are determined for

each protein. The molecular weight of each protein also is

also considered. The subcellular location description for pro-

teins also is also incorporated into the SVM.

The PREAL method uses 1176 distinct allergenic proteins

from the Swiss-Prot Allergen Index, IUIS Allergen Nomencla-

ture, SDAP and the Allergen Database for Food Safety (ADFS)

for building the positive allergen dataset. For building the

negative dataset, previously reported methods by Bjorklund

[32], Stadler [14] and Barrio and colleagues [33] are integrated

and sequence entries removed where identify similarities are

greater than 30% to known allergens [29]. In addition,

sequences less than 50 amino acids are removed. Using this

methodology, the subcellular locations (particularly extracel-

lular/cell surface and vacuole) and amino acid composition

were identified as the major markers for allergenicity for

specific wheat and soybean proteins previously [30].

AlgPred: prediction of allergenic potential of proteins and IgE

epitope mapping

AlgPred (http://www.imtech.res.in/raghava/algpred/) uses an

allergen representative peptide (ARP) strategy to try to predict

allergenic properties of allergens [23]. Allergens are predicted

by (1) MEME/MAST motif searches; (2) SVM-based classifica-

tion of allergens and non-allergens by single amino acid

composition and by dipeptide composition; and (3) BLAST

searches against allergen representative peptides. However, to

date, PREAL and Algpred have not been demonstrated to

outperform FAST or BLAST, depending on the criteria and

dataset used.

AllerTOP1.0: prediction of allergenic potential of proteins

AllerTOP (http://www.pharmfac.net/allertop) attempts to

predict allergenic potential of proteins by applying auto

cross-covariance (ACC) pre-processing to build a dataset of

known allergens, developing alignment-independent models

for allergen recognition based on the main physico-chemical

properties of proteins [34]. It uses five machine learning

methods for classification of proteins including discriminant

analysis by partial least square (DA-PLS), logistic regression

(LR), decision tree (DT), nai#ve Bayes (NB) and k nearest

neighbors (kNN). AllerTOP also try to identify the most

probable route of exposure. In comparison to other models

for allergen prediction, AllerTOP out-performs them with

94% sensitivity [35].

Allergen databases

On top of AllergenOnline, several databases exist for example

BIOPEP. Although not fully curated and regularly updated,

these databases can provide some insight on allergenicity

potential of allergens. They include the Allergome (http://

www.allergome.org/script/about.php), which has been

designed to supply information on IgE-mediated allergens

and associated clinical data. However, the use of Allergome is

limited as it does not have a searchable function.

BIOPEP (http://www.uwm.edu.pl/biochemia/index.php/

en/biopep) is a database of biologically active peptide


http://gmobl.sjtu.edu.cn/PREAL/index.php

http://download.igb.uci.edu/

http://www.imtech.res.in/raghava/algpred/

http://www.pharmfac.net/allertop

http://www.allergome.org/script/about.php

http://www.allergome.org/script/about.php

http://www.uwm.edu.pl/biochemia/index.php/en/biopep

http://www.uwm.edu.pl/biochemia/index.php/en/biopep


Table 1. In silico prediction tools for prediction of potential allergenicity of proteins or for supporting explanatory work.

Web tool Web tool access address Advantages of method Reference

AllergenOnline http://www.allergenonline.org/ � Methodology currently used for the allergenicity assessment of

novel proteins

[40]

� Peer reviewed allergen list (by independent scientists and

clinicians) and sequence searchable tool (FASTA, exact match

searches, yearly, curated and updated.

� Intended for the identification of proteins that may present a

potential risk of allergenic cross-reactivity

� Also celiac disease protein database risk assessment tool

� Hosted in the University of Nebraska, USA

AllerHunter http://tiger.dbs.nus.edu.sg/AllerHunter � Cross reactive allergen prediction program that uses a

combination of SVM and pairwise sequence similarity

[30]

� Hosted in the University of Singapore, Singapore

PREAL http://gmobl.sjtu.edu.cn/PREAL/index.php � Built on a combination of Support Vector Machine and protein

features

[6]

� Uses AllFam, UIS and Allergome allergen databases and ProAP

webtool

� Integrates protein biochemical and physicochemical properties

(molecular weight, secondary structure propensity, hydrophobicity,

polarizability, solvent accessibility, normalized van der Waals

volume, polarity, and length)

� Integrates sequential features and subcellular locations

� mRMR and IFS used to identify allergenicity features

� Hosted in the Shanghai Jiao Tong University, China

AllerTOP 1.0 http://www.pharmfac.net/allertop/ � Based on physicochemical protein properties [41,42]

� Uses a protein sequence mining method (autocross covariance

transformation of protein sequences into uniform equal-length

vectors). The proteins are classified by k-nearest neighbor

algorithm (kNN, k = 3) based on training set containing 2210 known

allergens from different species and 2210 non-allergens from the

same species.

� Hosted in the Sofia University, BulgariaBulgaria

SDAP http://fermi.utmb.edu/SDAP/ � Investigation of the cross-reactivity between known allergens and

in predicting the IgE-binding potential of food proteins

� 3-D searches

� Possibility to retrieve information related to an allergen from the

most common protein sequence and structure databases

(SwissProt, PIR, NCBI, PDB), to find sequence and structural

neighbors for an allergen, and to search for the presence of an

epitope other the whole collection of allergens

� Various computational tools that can assist structural biology

studies related to allergens

� Hosted in the University of Texas, USA

AlgPred http://www.imtech.res.in/raghava/algpred/ � Allows prediction of allergens (and its position) based on similarity

with known IgE epitopes

[23]

� Uses several tools (SVM, MEM/MAST, BLASTBLAST, 2890

allergen-representative peptides) and combined approaches

� Hosted in the Bioinformatics centre at CSIR-Institute of microbial

technology, India

BIOPEP http://www.uwm.edu.pl/biochemia � Contains data on allergenic proteins including names, sequence,

sequences of experimental/predicted epitopes

[39]

� Includes AllFam allergen family and epitopes

� Hosted in the University of Warmia and Mazury, Poland

Pole Bioinformatique

Lyonnais (PBIL)

http://pbil.univ-lyon1.fr/ � Presents information concerning peptide sequence bioactivities

on predicted and known allergenic proteins

[43]


http://www.allergenonline.org/

http://tiger.dbs.nus.edu.sg/AllerHunter

http://gmobl.sjtu.edu.cn/PREAL/index.php

http://www.pharmfac.net/allertop/

http://fermi.utmb.edu/SDAP/

http://www.imtech.res.in/raghava/algpred/

http://www.uwm.edu.pl/biochemia

http://pbil.univ-lyon1.fr/


Table 1 (Continued )

Web tool Web tool access address Advantages of method Reference

FLAPs [email protected] � Structure prediction of proteins [44]

� Filter-length adjusted allergen peptides (FLAPS) database

BcePred http://www.imtech.res.in/raghava/bcepred/ � Evaluates the performance of existing linear B-cell epitope

prediction methods. 1029 B-cell epitopes

[31]

� Based on physico-chemical properties (hydrophilicity, flexibility/

mobility, accessibility, polarity, exposed surface and/or turns) on a

non-redundant dataset from Swiss-Prot


technology, India

BepiPred 1.0 http://www.cbs.dtu.dk/services/BepiPred/ � Predicts the location of linear B-cell epitopes using both a hidden

Markov model and a propensity scale method

[45]

� Hosted in the Technical University of Denmark, Denmark

ABCpred http://omictools.com/abcpred-s6519.html � Predicts B cell epitopes in an antigen sequence, using artificial

neural network.

[23]

� IIs able to predict epitopes with 65.93% accuracy using recurrent

neural network


technology, India

Bpredictor https://code.google.com/p/my-project-bpredictor/ � Prediction of conformational B-cell epitopes from 3-D structures

by random forests with a distance-based feature.

[46]

� Limited update: last update in 2011

Epitopia http://epitopia.tau.ac.il/ � Detection of immunogenic regions in protein structures or

sequences (PDB and FASTA)

[47]

� Machine learning scheme (i.e. Naive Bayes classifier) to rank

individual amino acids in the protein, according to their potential of

eliciting a humoral immune response

� Identify B-cell epitopes (physico-chemical and structural-

geometrical properties)

� Hosted in Tel Aviv University, Israel

sequences associated with a program enabling the construc-

tion of profiles of the potential biological activity of protein

fragments, calculation of quantitative descriptors as measures

of the value of proteins as potential precursors of bioactive

peptides, and prediction of bonds susceptible to hydrolysis by

endopeptidases in a protein chain as well as allergenicity

potential. It contains a small number of proteins (i.e. 135)

but also allergenic epitopes [36]. Most of the epitopes used are

registered in the Immune Epitope Database (IEDB) [37]. Sec-

ondary peptide structures are predicted using GOR V program

[38]. BIOPEP is a database of peptides that contains recently

identified allergenic peptides. Recently, sixty sequences of

epitopes from the BIOPEP database attributed to tropomyosin

from the shrimp Farfantepenaeus aztecus (Pen a 1.0102) were

used as query sequences [39]. Vertebrate tropomyosins (e.g.

from vertebrates used as food resources) contain fragments

containing between 10 and 15 amino acid residues revealing

100% identity with epitopes from allergen Pen a 1.0102.

Fragments identical to epitopes from Pen a 1.0102 are com-

mon in sequences of invertebrate tropomyosins, including

those annotated in the Allergome database. Common epitopes

are a probable molecular basis for cross-reactivity between food

and non-food invertebrates. Some epitopes, especially rare

penta-peptides containing the DEERM sequence, are present

in sequences of proteins not sharing homology with tropo-

myosins. This fragment was found to be present in several

proteins, from edible plants and animals as well as pathogenic

microorganisms.

Conclusion

This paper reviews current in silico tools for assessing poten-

tial human allergenicity to proteins. These methods use a

number of physico-chemical features (mainly amino acid

searches) of proteins that can be predicted, but a strict,

structural correlation between these features and allergenici-

ty does not exist. Use of future innovative in silico methods for

the prediction of allergenicity will be largely influenced by

the choice of databases and algorithms that will be devel-

oped, standardized and most importantly empirically vali-

dated. Prediction of potential allergy is not proof of allergy.

Further biochemical testing (IgE blotting) and biological tests

including Basophil, skin prick tests, or in vivo challenge tests



http://www.imtech.res.in/raghava/bcepred/

http://www.cbs.dtu.dk/services/BepiPred/

http://omictools.com/abcpred-s6519.html

https://code.google.com/p/my-project-bpredictor/

http://epitopia.tau.ac.il/


with allergenic subjects are needed to validate allergy to

protein predictions.

Conflict of interest

The authors declare that there are no conflicts of interest.

Acknowledgements

This work was supported by the EU COST Action ImpARAS

FA1402. The opinions expressed herein and the conclusions

of this publication are those of the authors.

References[1] Verhoeckx KCM, Vissers YM, Baumert JL, Faludi R, Feys M, Flanagan S,

et al. Food processing and allergenicity. Food Chem Toxicol 2015;80:

223–40.

[2] Goodman RE. J Huazhong Agric Univ 2014;33(6):85–113.

[3] Goodman RE, Vieths S, Sampson HA, Hill D, Ebisawa M, Taylor SL, et al.

Allergenicity assessment of genetically modified crops—what makes

sense? Nat Biotechnol 2008;26(January (1)):73–81. http://dx.doi.org/

10.1038/nbt1343. Review. Erratum in: Nat Biotechnol. 2008

February;26(2):241. PMID: 18183024.

[4] Codex Alimentarius Commission guidelines. Foods derived from modern

biotechnology (2003 reprinted in the second edition in 2009). World

Health Organization and Agricultural Organization of the United Nations;

2009.

[5] Mirsky HP, Cressman RJ, Ladics GS. Comparative assessment of multiple

criteria for the in silico prediction of cross-reactivity of proteins to known

allergens. Regul Toxicol Pharmacol 2013;67(2):232–9.

[6] Wang J, Yu Y, Zhao Y, Zhang D, Li J. Evaluation and integration of existing

methods for computational prediction of allergens. BMC Bioinform

2013;14:S1. http://dx.doi.org/10.1186/1476-2105-14-S4-s1.

[7] Ladics GS. Current Codex guidelines for assessment of potential protein

allergenicity. Food Chem Toxicol 2008;46:S20–3.

[8] FAO/WHO. Evaluation of allergenicity of genetically modified foods. Report

of a Joint FAO/WHO Expert Consultation on Allergenicity of Foods derived

from biotechnology, 22nd–25th January, 2001. FAO/WHO; 2001. 15.

[9] Aalberese RC, Crameri R. IgE-binding epitopes: a reappraisal. Allergy

2011;66:1261–74.

[10] Goodman RE, Ebisawa M, Ferreira F, Sampson HA, van Ree R, Vieths S,

et al. AllergenOnline: a peer-reviewed, curated allergen database to assess

novel food proteins for potential cross-reactivity. Mol Nutr Food Res 2016.

http://dx.doi.org/10.1002/mnfr.201500769.

[11] Siruguri V, Bharatraj DK, Vankudavath RN, Mendu VV, Gupta V,

Goodman RE. Evaluation of Bar, Barnase and Barstar recombinant

proteins expressed in genetically engineered Brassica juncea (Indian

mustard) for potential risks of food allergy using bioinformatics and

literature searches. Food Chem Toxicol 2015;83:93–102.

[12] Moran DL, Tetteh AO, Goodman RE, Underwood MY. Safety assessment

of the calcium-binding protein, apoaequorin, expressed by Escherichia coli.

Regul Toxicol Pharmacol 2014;69(2):243–9.

[13] Puumalainen TJ, Poikonen S, Kotoyuori A, Vaali K, Kalkkinen N. Napins, 2

S albumins, are major allergens in oilseed rape and turnip rape. J Allergy

Clin Immunol 2006;117:426–32.

[14] Stadler MB, Stadler BM. Allergenicity prediction by protein sequence.

FASEB J 2003;17(9):1141–3.

[15] Ladics GS, Fry J, Goodman R, Herouet-Guicheney C, Hoffmann-

Sommergruber K, Madsen CB, et al. Allergic sensitization: screening

methods. Clin Transl Allergy 2014;4:13.

[17] Mittag D, Batori V, Neudecker P, Wiche R, Friis EP, Ballmer-Weber BK,

et al. A novel approach for investigation of specific and cross reactive IgE

epitopes on Bet v 1 and homologous food allergens in individual patients.

Mol Immunol 2006;43:268–78.

[18] Bragin AO, Demenkov PS, Kolshonov NA, Ivanisenko VA. Accuracy of

protein allergenicity prediction can be improved by taking into account


data on allergenic protein discontinuous peptides. J Biomol Struct Dyn

2013;1:59–64.

[19] Radauer C, Bublin M, Wagner S, Mari A, Breiteneder H. Allergens are

distributed into a few protein families and possess a restricted

number of biochemical functions. J Allergy Clin Immunol

2008;121(4):847–52.

[20] Shewry PR, Jenkins JA, Beaudoin F, Mills ENC. The classification,

functions and evolutionary relationships of plant proteins in relation to

food allergies. In: Mills CEN, Shewry PR, editors. Plant food allergens.

Blackwell Sciences; 2004. p. 24–86.

[21] Kumar KK, Sheloker PS. An SVM method using evolutionary information

for the identification of allergenic proteins. Bioinformation

2008;2(6):253–6.

[22] Muh HC, Tong JC, Tammi MT. AllerHunter: a SVM-pairwise system for

assessment of allergenicity and allergic cross-reactivity in proteins. PLoS

ONE 2009;4(6):e5861. http://dx.doi.org/10.1371/journal.pone.0005861.

[23] Saha S, Raghava SPS. AlgPred: prediction of allergenic proteins and

mapping of IgE epitopes. Nucleic Acid Res 2006;34. http://dx.doi.org/

10.1093/nar/gkL343.

[24] Tomar N, De RK. Immunoinformatics: an integrated scenario.

Immunology 2010;131(2):153–68.

[25] Lundegaard C, Lund O, Nielsen M. Predictions versus high-throughput

experiments in T-cell epitope discovery: competition and synergy? Expert

Rev Vaccines 2012;11(1):3–54.

[26] Borges J-P, Barre A, Culerrier R, Archimbaud N, Didier A, Rouge P. How

reliable is the structural prediction of IgE-binding epitopes of allergens? The

case study of plant lipid transfer proteins. Biochemie 2007;89(1):83–97.

[27] Krieger E, Koraimann G, Vriend G. Increasing the precision of

comparative models with YASARA NOVA – a self-parameterizing force

field. Proteins 2002;47(3):393–402.

[28] Bragin AO, Demenkov DS, Kolchanov NA, Ivanisenko VA. Accuracy of

protein allergenicity prediction can be improved by taking into account

data on allergenic protein discontinuous peptides. J Biomol Struct Dyn

2013;13(1):59–64.

[29] Wang J, Zhang D, Li J. PREAL: prediction of allergenicity protein by

maximum relevance minimum redundancy (mRMR) feature selection.

BMC Syst Biol 2013;7(5):59.

[30] Liu y, Gu W, Zhang W, Wang J. Predict and analyse protein glycation sites

with the mRMR and IFS methods. BioMed Res Int 2015. http://dx.doi.org/

10.1155/2015/561549.

[31] Magnan CN, Balidi P. Sspro/ACCpro 5: almost perfect prediction of

protein secondary structure and relative solvent accessibility using profile,

machine learning and structural similarity. Bioinformatics

2014;30(18):2592–7.

[32] Bjorklund AK, Soeria-Atmadja D, Zorzet A, Hammerling U, Gustafsson

MG. Supervised identification of allergen-representative peptides for in

silico detection of potentially allergenic proteins. Bioinformatics

2005;21(1):39–50.

[33] Barrio AM. EVALLER: a web server for in silico assessment of potential

protein allergenicity. Nucleic Acids Res 2007;35(Suppl 2):W694–700.

[34] Dimitrov I, Bangor I, Flower DR. AllerTop v 2 – a server for in silico

prediction of allergens. J Mol Model 2014;20(6):2278.

[35] Dimitrov I, Flower DR, Doytchinova I. AllerTop – a server for in silico

prediction of allergens. BMC Bioinform 2013;14(6):S4.

[36] Minkiewicz P, Dziuba J, Iwaniak A, Dziuba M, Darewicz M. BIOPEP

database and other programs for processing bioactive peptide sequences. J

AOAC Int 2008;91:965–80.

[37] Vita R, Zarebski L, Greenbaum JA, Emami H, Hoof I, Salimi N, et al.

The Immune Epitope Database 2.0. Nucleic Acids Res 2010;38:D854–62.

[38] Sen TZ, Jernigan RL, Garnier J, Kloczkowski A. GOR V server for protein

secondary structure prediction. Bioinformatics 2005;21(11):2787–8.

[39] Minkiewicz P, Sokolowska J, Darewicz M. The occurrence of sequence

identical with epitopes from the allergen pen a 1.0102 among food and

non-food proteins. Pol J Food Nutr Sci 2015;65(1):21–9.

[40] Goodman RE, Silvanovich A, Hileman RE, Bannon GA, Rice EA, Astwood

JD. Bioinformatic methods for identifying known or potential allergens in

the safety assessment of genetically modified crops. Comments Toxicol

2002;8(3):251–69.

http://refhub.elsevier.com/S1740-6757(16)30001-9/sbref0240




http://dx.doi.org/10.1038/nbt1343










http://dx.doi.org/10.1186/1476-2105-14-S4-s1

http://dx.doi.org/10.1186/1476-2105-14-S4-s1










http://dx.doi.org/10.1002/mnfr.201500769

http://dx.doi.org/10.1002/mnfr.201500769





































http://dx.doi.org/10.1371/journal.pone.0005861


http://dx.doi.org/10.1093/nar/gkL343





















http://dx.doi.org/10.1155/2015/561549

http://dx.doi.org/10.1155/2015/561549

http://dx.doi.org/10.1155/2015/561549






























[41] Wold S, Jonsson J, Sjostrom M, Sandberg M, Rannar S. DNA and peptide

sequences and chemical processes multi-variably modelled by principal

components analysis and partial least square projections to latent

structures. Anal Chim Acta 1993;277:239–53.

[42] Dimitrov I, Flower DR, Doytchinova I. AllerTOP – a server for in silico

prediction of allergens. BMC Bioinform 2013;14:S4.

[43] Pierriere G, Combet C, Penel S, Blanchet C, Thioulouse J. Integrated

databanks access and sequence/structure analysis services at PBIL. Nucleic

Acids Res 2003;31(13):3393–9.

[44] Soeria-Atmadja D, Lundell T, Gustafsson MG, Hammerling U.

Computational detection of allergenic proteins attains a new level of

accuracy with in silico variable-length peptide extraction and machine

learning. Nucleic Acids Res 2006;34(13):3779–93.

[45] Larsen JE, Lund O, Nielsen M. Improved method for predicting linear B-

cell epitopes. Immunome Res 2006;2:2. http://dx.doi.org/10.1186/1745-

7580-2-2.

[46] Zhang W, Xiang Y, Zhao M, Zou H, Ye X, Liu J. Prediction of

conformational B-cell epitopes from 3D structures by random forests with

a distance-based feature. BMC Bioinform 2011;17. http://dx.doi.org/

10.1186/1471-210512-341.

[47] Rubinstein ND, Mayrose I, Pupko T. A machine-learning approach for

predicting B cell epitopes. Mol Immunol 2008;46:840–7.















http://dx.doi.org/10.1186/1745-7580-2-2

http://dx.doi.org/10.1186/1745-7580-2-2

http://dx.doi.org/10.1186/1745-7580-2-2

http://dx.doi.org/10.1186/1471-210512-341

http://dx.doi.org/10.1186/1471-210512-341

http://dx.doi.org/10.1186/1471-210512-341



DRUG DISCOVERY

TODAY

DISEASEMODELS

Applicability of epithelial models inprotein permeability/transport studiesand food allergyN. Cubells-Baeza1, K.C.M. Verhoeckx2, C. Larre3, S. Denery-Papini3,

M. Gavrovic-Jankulovic4, A. Diaz Perales1,*1Center for Plant Biotechnology and Genomics (UPM-INIA), Pozuelo de Alarcon, Madrid, Spain2TNO, Zeist, The Netherlands3INRA, UR 1268 Biopolymeres Interactions Assemblages, Nantes, France4Department of Biochemistry, Faculty of Chemistry, University of Belgrade, Belgrade, Serbia


Editors-in-Chief




Measurement of protein transport across the intestinal

barrier might be a relevant approach in allergenicity

risk assessment. Traditionally, studies on protein trans-

port, were performed using stable cell lines cultured as

a monolayer. One of the major advantages of these

models is their relatively low price and easy handling.

However, monolayers lack a physiologically relevant

environment (presence of other cell-types and a mucus

layer), which may have an effect on transport charac-

teristics and thus correct prediction of protein allerge-

nicity. This paper summarizes the most widely used

epithelial models and discusses their benefits and lim-

itations for measuring protein transport and allergic

sensitization to food.

Introduction

Incorporation of new proteins into food crops and the intro-

duction of new protein sources onto the food market (e.g.

rapeseed) can lead to the introduction of new food allergens,

and consequently increasing the risk for the susceptible food-

allergic population. For that reason, allergenicity assessment

of these new proteins is needed.

*Corresponding author: A. Diaz Perales ([email protected])



In the past 10 years, we have made huge efforts to establish

the features that characterize an allergen [1]. However, despite

these efforts, we still do not know exactly what makes a

protein an allergen. Allergies are complex diseases in which

two phases can be distinguished: (1) the sensitization phase, in

which a protein is exposed to the mucosal immune system, is

recognized as an allergen and induces the production of

immunoglobulin E (IgE). (2) The symptomatic phase, in which

IgE is bound to the surface of effector cells via specific recep-

tors (FceRI) and binding of two IgE molecules with the aller-

gen, which induces the release of inflammatory mediators

responsible for allergic symptoms. For both phases, proteins

need to come into contact with the immune system and

therefore needs to be transported over the protective epithelial

layers of our body. Unfortunately, the role of allergen trans-

port and transport route (respiratory, skin or oral) in food

6.08.002 13







allergy is relatively unaddressed in the literature. For that

reason, it is necessary to expand our knowledge in this field

to enhance our understanding on the sensitization mecha-

nism.

Emerging evidence suggests that the skin may be a highly

relevant inductive site for allergic sensitization to food pro-

teins. Conversely, some routes of exposure have been proposed

to be inherently tolerogenic (e.g. oral and sublingual expo-

sure). Transport via the gut may lead to different immunologi-

cal responses as well. Transport of soluble milk allergens via

epithelial cells led to anaphylaxis, while transport of the

aggregated form of the same milk allergen after heating led

to sensitization in a cow’s milk mouse model [2]. These find-

ings suggest that the transport of soluble protein via villous

epithelial cells was the main pathway for anaphylactic

responses, while transport of the aggregated forms via Peyer’s

Patches (PP) was needed for sensitization. Furthermore, defects

in the integrity of the epithelial barrier have also been reported

in food allergy. Clinical studies in children with cow’s milk

allergy demonstrated that intestinal permeability increased

after, but not before the allergen challenge [2,3]. A recent study

based on small intestinal biopsy specimens exposed to food

allergens in vitro, showed decreased expression of tight junc-

tion (TJ) proteins (i.e. occludin, claudin-1, and ZO-1) in tissues

obtained from food allergic patients compared to healthy

subjects [3]. Both studies suggest, that in sensitized individuals

intestinal permeability and the passage of allergens is en-

hanced. For that reason, evaluation of protein transport across

the intestinal barrier and its effect on epithelial permeability

might be a relevant parameter in allergenicity risk assessment

[4]. The applicability of different epithelial cell models to study

these aspects is discussed in the present review.

The intestinal epithelium

The intestinal epithelium is the largest interface between the

host and the environment. It regulates fluxes of ions and

nutrients and limits host contact with luminal antigens [5].

Anatomically, the intestinal mucosa is divided into three

layers: (i) the first, which is closest to the intestinal lumen,

consists of a single layer of epithelial cells attached to a

basement membrane; (ii) the second layer, the lamina propria,

consists of subepithelial connective tissue, immune cells, and

lymph nodes; (iii) the third layer is known as the muscularis

mucosae and is composed of smooth muscle fibres [6].

The first layer of epithelial cells forms a biochemical and

physical barrier which separates microbiota in the lumen

from the underlying mucosa (Fig. 1a) [7]. The lymphoid tissue

in the mucosa is organized into inductive sites (Peyer’s patches

and mesenteric lymph nodes) and effector sites (normal intes-

tinal mucosa), which are responsible for the induction phase

of an immune response such as sensitization [8,9].

At the bottom of the crypts of the first layer, a pool of

pluripotent stem cells can differentiate into five epithelial cell


types: absorptive columnar cells (enterocytes), goblet cells,

endocrine cells, Paneth cells, and M (microfold) cells (Fig. 1a)

[10]. Goblet cells and Paneth cells secrete, respectively, mucus

and antimicrobial proteins (defensins, cathelicidins, and his-

tatins) to protect the epithelial surface from intruding bacte-

ria. M cells and enterocytes mediate transport of luminal

antigens and living bacteria across the epithelial monolayer

to the underlying lymphoid cells, such as antigen-presenting

cells (e.g. dendritic cells and intestinal macrophages) [11].

The permeability and polarity of the first epithelial layer are

maintained by the apical junctional complex, which is com-

posed of TJs, adherent junctions, and the subjacent desmo-

somes (Fig. 1b). Permeability depends mainly on the TJs,

which are composed of transmembrane proteins such as

occludin, claudin, junctional adhesion molecule A, and tri-

cellulin (Fig. 1b) [5,12].

Transport through the epithelium

Although most dietary proteins are degraded by digestive

enzymes and absorbed as amino acids and di/tripeptides,

some can resist the gastric environment (pH 1–2 and pepsin

hydrolysis). Large immunogenic peptides and intact proteins

are capable of reaching the lumen of the small intestine and

triggering immune cells in the mucosa [5]. Thus, resistance to

gastrointestinal digestion might contribute to allergenicity.

However, there are also examples of pepsin-sensitive aller-

gens whose resulting fragments still show IgE-binding activi-

ty [13]. It can be envisioned that the combination of

gastrointestinal digestion and protein transport is an impor-

tant factor for allergenicity.

Once the dietary proteins and peptides reach the small

intestine, they can be transported across the epithelial intes-

tinal barrier to the underlying basolateral side and distributed

throughout the body. Transport of proteins across the

intestinal mucosa depends on size (influenced by aggrega-

tion), polarity, and shape. Proteins can be transported via the

paracellular route or via transcellular routes (Fig. 1c). Paracellular

transport is the transfer of compounds through the intercel-

lular space and is regulated by the integrity of the TJs [14].

Normally, small hydrophilic compounds (up to 600 Da) are

transported by this route, although small proteins (less than

3.5 kDa) can also pass.

Transcellular transport comprises the absorption of com-

pounds via passive diffusion, vesicle endocytosis, and carrier-

mediated transport (Fig. 1c). The main route of transcellular

protein transport is endocytosis, which is known to occur in

different cell types. The transcellular transport of large par-

ticles has traditionally been ascribed to M cells overlying

Peyer’s patches, while soluble particles are transported via

the epithelial cells [15]. It has even been suggested that

transport via M cells will induce a local or systemic immune

response towards the antigen, while soluble antigens trans-

ported via enterocytes will lead to suppression of the immune


Undigested food

Commensal bacteria

Goblet cell

Paneth cells

Paracellular transport Transcellular transport

Desmosomes

Adherence junctions

Tight junctions

Gap junctions

1 2 3 4

Enterocytes

(a) (b)

(c)


Figure 1. Components of intestinal mucosa. (a) Simplified scheme of intestinal cells and structures in the small intestine. (b) Schematic drawing of

polarized epithelial cells with different types of intercellular contacts. (c) Different types of transport through the intestinal epithelium: 1, paracellular

transport; 2, passive diffusion; 3, vesicle-mediated transcytosis; 4, carrier-mediated uptake.

system and induction of tolerance to the antigen [16]. The

passage of aggregated antigens through M cells in combina-

tion of soluble antigens through IEC is thought to be critical

in the onset of milk allergy [Roth Walter].

Alternatively, intestinal DCs and macrophages have the

capacity to sample directly in the intestinal lumen by extend-

ing dendrites between epithelial cells. Antibodies such as IgA,

IgG, and IgE can also be involved in enterocytic protein

transport in the form of carrier-mediated transport [5]. IgE

binds to the antigen and the CD23 receptor, which transports

the IgE-antigen complex across the cell without lysosomal

degradation [17].

Intestinal epithelium models used in protein

permeability/transport studies

An epithelial model used to study protein transport should

preferentially include all the components of the intestinal

mucosa (e.g. mucus layer and epithelial and M cells).

However, in practice this is not always feasible, with the

result that various models have been developed. These mod-

els differ in complexity and applicability (Tables 1 and 2). It is

therefore important to define the study objective (e.g. trans-

port via M cells and/or enterocytes, allergenic activity, mucus

effect, permeability) in order to identify the best intestinal

model [18]. Irrespective of the cell model chosen, the main

readout parameters are the effects on epithelial barrier func-

tion, absorption, and trans-epithelial transport of the test

compound. The reliability of such studies depends on the

uniformity and integrity of the confluent and polarized cell

monolayer.

Tumor cell line models

Cell lines used to study the transport and absorption of

proteins include Caco-2, HT-29, T84, and IPEC-J2 (Table

1). In all cases, cells are grown in a Transwell system to form a

monolayer (Fig. 2).



Table 1. Comparison of the different cell lines used to study protein transport

Caco cell line HT29 IPEC-J2 Cell line

Growth Grows in culture as an adherent monolayer

of epithelial cells

Grows in culture as an

adherent monolayer of

epithelial cells

Grows in culture as an

adherent monolayer of

epithelial cells

Differentiation Takes 14–21 days after confluence under

standard culture conditions

Takes 21–28 days after

confluence under standard

culture conditions

Takes 4–9 days after

confluence under standard

culture conditions

Cell morphology Polarized cells with tight junctions and brush

border at the apical side

Polarized cells with tight

junctions, without ciliar border

at the apical side

Microvilli on their apical side

and tight junctions sealing

neighboring cells together

Mucus production In co-culture with HT29-MTX Yes Yes

Electrical parameters High electrical resistance Moderate electrical resistance High electrical resistance

Digestive enzymes Expresses typical digestive enzymes,

membrane peptidases and disaccharidases

of the small intestine (lactase,

aminopeptidase N, sucrase-isomaltase and

dipeptidylpeptidase IV)

Expresses sucrose-isomaltase,

aminopeptidase N,

dipeptidylpeptidase-IV and

alkaline phosphatase, but

lactase is absent.

Active transport Amino acids, sugars, vitamins, hormones Amino acids, sugars, vitamins,

hormones

Low active transport

Receptors Vitamin B12, vitamin D3, EGFR (epidermal

growth factor receptor), sugar transporters

(GLUT1, GLUT2, GLUT3, GLUT5, SGLT1)

Receptors usually expresses in

intestinal epithelium

Receptors usually expresses in

intestinal epithelium

Cytokine production IL-6, IL-8, TNFx, TGF-x1, thymic stromal

lymphopoietin (TSLP), IL-15

IL-6, IL-8, TNFx, TGF-B, TSLP,

IL-15, IL3, GM-CSF, VGF

IL-1a, -1b, -6, -7, -8, -12A, -

12B, -18

Applicability Studies about protein/drugs transportation Studies about proteins/drugs

transportation

Studies about proteins/drugs

transportation

Interaction with

immune system

Co-culture with different immune cells No described No described

Expenses Low Low Low

Monolayers of human colon carcinoma cell lines, the so-

called Caco-2, have been extensively used over the last 20

years to predict the permeability of the intestinal mucosa to

proteins [19]. The polarized monolayer of these well-differ-

entiated columnar absorptive cells expresses a brush border

on their apical surface with typical small intestinal enzymes

and transporters. The cells differentiate into a polarized apical

and basolateral membrane mimicking the luminal and mi-

crovilli side (apical) of the intestine and intercellular TJs.

Caco-2 exhibit features of enterocytes of the small intestine.

During differentiation, cells progressively express digestive

enzymes [20]. Conversely, the electrical properties and ionic

conductivity and permeability of the differentiated Caco-2

cells resemble those of the colonic crypt cells [19].

The Caco-2 model has many limitations, such as trans-

epithelial electrical resistance (TEER (used to address the

integrity of the monolayer)), which is higher in Caco-2

monolayers (up to 500 ohm/cm2) than in human

intestine (12–69 ohm/cm2) owing to over-expression of TJs


[21]. Moreover, no mucus layer is produced on the apical side

of Caco-2, thus limiting studies of the protein–mucosa inter-

action. Despite these limitations, the Caco-2 cell line has

proved to be the best model to date to study intestinal

absorption and toxicity.

HT-29 cells stem from a human colon adenocarcinoma

cell line that contains both absorptive and mucus-secreting

cells [19]. Under normal growth conditions, HT-29 cells grow

as a multilayer of non-polarized, high glucose-consuming,

and undifferentiated cells. When glucose is removed from the

growth medium and replaced with a different carbon source,

the cells differentiate after three to four weeks in culture,

leading to the appearance of both absorptive cells with char-

acteristics similar to those of the differentiated Caco-2 cells

and to goblet-like mucin cells. These differentiated cells

express brush border-associated hydrolases that are typical

for the small intestine. The cells have brush border microvilli

even though enzyme activity is much lower than for normal

intestinal epithelial cells, and they do not express lactase.


Table 2. Comparison of different types of culture

Caco cell line Organoids Primary culture

High-throughput High Low Low

Expenses Low High High

Redout Protein/drugs transportation Protein and drugs metabolism Transport and metabolism

Advantage Easy manage Maintain the integrity of the mucosa,

with all the specialized cells represented

Multi-cell system. All intestinal regions

can be used Less labor intensive than

Ussing chamber technology

Disadvantages No mucus, no specialized cell,

no 3D structure

The apical side is not accessible Biological variations

Limited viability

Applicability on food allergy Yes Yes, specially, interaction with

specialized cells

Yes, transport and interaction

with specialized cells

The T84 cell line is an epithelial model used to study

protein transport. The cell line was obtained from a pulmo-

nary metastasis of a human colon carcinoma. When the cells

are grown on microporous filter supports coated with colla-

gen, they form a cell monolayer with a highly polarized

morphology, few microvilli, and a very high TEER, thus

indicating the presence of well-differentiated TJs. Chloride

secretion in T84 cells is regulated, as is typical of colonic crypt

cells [22,23].

The porcine intestinal enterocyte cell line (IPEC-J2) is a

non-transformed, permanent intestinal cell line that was

originally isolated from the jejunal epithelium of a neonatal

unsuckled piglet. The cells have already proven to be a

valuable tool in the characterization of epithelial cell inter-

actions with enteric bacteria and viruses providing insight

into initial host–pathogen and host–non-pathogen (e.g. com-

mensal or probiotic) interactions [24]. The strength of the

IPEC-J2 cell line as an in vitro model originates from its

morphological and functional similarities with intestinal

epithelial cells in vivo. No brush border enzyme activity has

been described in IPEC-J2 cells. High TEER values and low

active transport rates are obtained when IPEC-J2 cells are

cultured in fetal bovine serum.

The Caco-2 model is mostly used to study protein/allergen

transport. For example Roth Walter et al. studied the trans-

port of native and aggregated b-lactoglobuline in a Caco-2

model and compared this with an in vivo mouse study [2]. The

transport of Ber e 1 and Ses i 1 was studied by Moreno et al.

[25]. The other epithelial cell models were also used to study

allergens. HT-29 was used to study endocytosis of Ara h2 [26],

T84 to study effect of Gly m 5 on TJ proteins [27] and IPEC J2

to study Gly m 1 uptake [28].

Co-culture of the Caco-2 cell line and HT29-MTX cells

Co-cultures of Caco-2 cells with HT29-MTX (HT29 cells trea-

ted with methotrexate) have been developed in an attempt to

overcome the lack of mucus in Caco-2 cultures [21]. Caco-2

and HT29-MTX are derived from intestinal absorptive and

goblet cells, respectively. The human intestinal cell line Caco-

2 differentiates into enterocytes, while HT-29MTX cells pro-

duce mucins, heavily glycosylated proteins that form a sur-

face protecting layer on epithelial cells (mucus). The

difficulty of this model is to culture the right Caco-2/HT-

29 MTX ratio.

Triple co-culture of Caco-2, HT29-MTX, and RajiB lym-

phoma was recently applied to complete the intestinal

mucosa model [29]. RajiB cells can differentiate into a M-cell

phenotype by co-culturing with Caco-2 enterocytes. This

relevant model is complex owing to the presence of three

non-adherent cell lines in RajiB, and the conversion of M-cell

has to be closely monitored [30]. Both co-culture models

were, to the best of our knowledge, not used for protein

transport studies. Rytkonen et al. used a co-culture of

Caco-2 cells with PP from mice to study the transport of

heated and native b-lactoglobulin [31]. However the co-cul-

ture of human cells with mouse cells seems a bit odd.

Intestinal organoids

Collectively, the intestinal organoid system (Fig. 2) enables

culture and expansion of intestinal stem cells ex vivo. Impor-

tantly, the resulting epithelial structures faithfully recapitu-

late the homeostasis and architecture of the functional

intestinal epithelium. To date, intestinal organoids have been

shown to have multiple applications, including analysis of

endogenous stem cell characteristics and gene function, as

well as disease modelling [32]. With the recent development

of efficient gene-editing tools, it is now possible to rapidly

engineer cultured intestinal organoids to generate highly

physiological models of human gastrointestinal disease for

use as research tools [33]. Despite the requirement for more

expensive technology than tumor cell lines, the genetic

profile of intestinal organoids seems closer to that of the

intestinal cell epithelium. Therefore, this system could

be an alternative for the study of protein transport and




Polarized

Everted sac

Intestinal organoids Ussing Chamber

Tumor cell line models(Caco 2, HT29, T84, IPEC-J2)

Confluence

Personalized organoid culture

Stem cells

mV

µA

Hum

an

col

onic

cry

pt is

olat

ion

Biopsy sample

Figure 2. Different types of epithelial models used to study protein transportation. (a) Tumor cell line models, epithelial cells are seeded on plastic

structures (Transwell) separating the bsasolateral and apical sides and allowing the polarization of the cells. (b) Intestinal epithelial organoids, three-

dimensional structures of cultured intestinal cells from isolated intestinal crypts including stem cells, allowing to remake a mini-intestine including all cell

types. (c) Everted sac, intestinal segments are cut and inverted so that the apical side of the mucosa is on the outside. (d) Ussing Chamber, an intestine

segment placed on an apparatus for measuring epithelial membrane properties in both side and the efficiency of transportation.

associated diseases. However, the inwards orientation of the

epithelial cells (directed to the lumen of the organoids),

makes the apical side relatively inaccessible for direct experi-

mental stimulation. This will impede protein transport stud-

ies. Furthermore, availability of human tissue is often a

bottleneck and hinders their possibilities as a standard ap-

proach. To the best of our knowledge no literature was found

on the use of organoids in allergen transport or permeability

studies.

Ex vivo models

The major drawbacks of the single cell models described

above can be overcome by using intestinal tissues. In these

models, the asymmetrical distribution of proteins and lipids

in the two plasma membrane domains facing the intestinal

lumen, the internal milieu, and the presence of highly orga-

nized structures (TJs) joining adjacent cells and separating

the two membrane domains, enable selective processes of

absorption, transport, and secretion to take place across the


intestinal mucosa [34]. The maintenance of these character-

istics ex vivo is particularly important for the study of absorp-

tion, metabolism, and toxicity.

Traditionally, ex vivo intestinal models for intestinal pro-

tein transport studies have been based on animal tissue. The

techniques used include the everted sac technique [35]

and the Ussing chamber [36] (Fig. 2). The everted sac is a

segment of animal intestine that is everted and used to assess

protein transport. In the case of the Ussing chamber, a fresh

intestinal segment is mounted into a complex apparatus for

measuring protein transport and epithelial membrane prop-

erties. Both techniques provide an accurate measurement of

intestinal permeability. The most relevant advantage for food

allergy is the possibility of studying the effect of sensitization

on intestinal protein absorption, using intestinal tissue from

sensitized animals [37,38].

However, both techniques have several limitations. First,

tissue viability is rapidly lost (2 h); second, the tissue can be

damaged during isolation, which may lead to overestimation


of protein transport. On the contrary, the presence of the

muscle layer in the everted sac method may lead to underes-

timation of protein transport [35]. Third, interspecies differ-

ences in anatomy, physiology, metabolism, diet, and

microbiota complicate extrapolation of data to humans

[4,39]. Pigs share more physiological and immunological

similarities with humans than rodents, and the use of mini-

pigs is becoming increasingly common in nutritional re-

search [40].

The recently developed InTESTineTM method is a medi-

um-throughput alternative to the Ussing chamber and is

based on intestinal tissue from pigs that is incubated on a

rocker platform in a high oxygen incubator [41]. The viability

of tissue could be retained for 2 h, paracellular absorption

transport resembles that of human intestinal tissue in the

Ussing chamber [42], and the transport of macromolecular

proteins is studied using radioactively labelled proteins [14].

This method seems to be a good alternative that should be

evaluated in future studies on food allergy.

Food allergy studies based on epithelial models

It can be envisioned that sensitization to food allergens

begins with the transport of these allergens across the intes-

tinal epithelium. Transport of luminal antigens occurs typi-

cally via M cells but intestinal epithelial cells also have the

capacity to transport luminal antigens across the intestinal

wall, but with a different capacity than M cells do [43].

Furthermore, the ability of allergens (intact or fragmented)

to cross the epithelial barrier could be based on the increased

permeability of TJs or on their immunogenic activity [2,31].

Most studies with food allergens using epithelial models

focus on the effect on TJs. For instance, Price et al. showed

that peanut allergens were able to alter the intestinal barrier

permeability and TJ localisation in a Caco-2 model. The

allergens passed through the epithelial monolayer by the

paracellular pathway [41]. Zhao et al. used a T84 porcine

model and reported that incubation with b conglycinin from

soy (Gly m 5) induced the downregulation of TJ proteins

(claudin-3, occludin, and ZO-1) [27]. The study from Cavic

et al. [44] demonstrated that Act c 1 (actinidin), which is a

kiwifruit allergen, exhibits persistent proteolytic activity dur-

ing digestion. Exposure of T84 cells to this allergen, resulted

in impairment of the epithelial barrier, which was related to

the degradation of occludin by the proteolytic action of

actinin. Furthermore, the alteration of this single TJ protein

led to nonselective paracellular transport of allergens. Not all

allergens affect epithelial permeability, for example Moreno

et al. [25] showed that the transcellular transport of purified

2s albumins Ber e 1 (Brazil nut) and Ses i 1 (Sesame seed)

within Caco-2 monolayers, did not affect permeability as

observed with no change to allergen absorption rate

and TEER. The same is true for wheat allergen v 5 and

lipid transfer protein (LTP) [45]. This paper also shows that

digestion of the wheat allergen v 5 protein enhanced their

transcellular transport capacity. Besides digestion also other

intrinsic properties and processing steps might influence

protein transport. Roth-Walter et al. showed, in vitro and in

vivo, that pasteurization of the soluble milk protein b-lacto-

globulin (which causes aggregation) shifted transport from

transcytosis through enterocytes to transport via Peyer’s

patches. The in vivo study also showed that aggregated b-

lactoglobulin induced IgE formation (Th2-associated anti-

body) and Th2 cytokine production (IL-5, IL-13, IFN-g, IL-

10) with respect to the soluble b-lactoglobulin. The findings

of this study suggested that the transport of soluble protein

via villous epithelial cells was the main pathway for anaphy-

lactic responses, while transport of aggregates via PP induced

sensitization [2]. So it can be hypothesized that parameters

such as transport route (M-cells or epithelial cells) and/or

transported protein size (intact or fragmented) could help us

to predict the allergenic potential of proteins, but more tests

are needed to confirm this.

Both increased the permeability of the epithelial T84

monolayer, and thus affected the apical-to-basal movement

of proteins, such as horseradish peroxidase, through both the

transcellular and paracellular pathway [46]. Moreover, there

is evidence that mediators, released from mast cells (e.g.

tryptase and tumor necrosis factor alpha) contribute to in-

creased epithelial paracellular permeability [47]. The latter

will take place in already sensitized individuals.

In summary, we can make the following assumptions: (1)

allergens may affect TJs (e.g. protease activity), (2) digestion

and processing influence protein transport, (3) allergens must

cross the gastric barrier in an immune reactive form, (4) size

and solubility determines transport route and immunological

response, and (5) immunological status (release compounds

mast cells and Th2 cytokines) may increase paracellular

transport (sensitized persons). However, further studies with

food allergens in the models described above are required in

order to clarify the precise scenario for proteins necessary to

induce food allergy.

Conclusions

Unfortunately, there is lack of data on the transport capabili-

ties of many food allergens and their route of exposure to the

mucosal immune system. It is highly likely that gut perme-

ability and allergen transport play a role in the development

of food allergy or tolerance. However, more data is needed on

the permeability and absorption of food allergens to draw any

conclusions regarding the influence of intestinal permeabili-

ty on the allergenic potential of proteins. Furthermore, the

combination of processing and digestion on permeability

should be explored, since they may have a huge effect on

the transport capacities of allergens and thus the immuno-

logical response thereafter. Understanding the role of protein

transport and gut permeability, will help us to develop better



in vitro models to measure these important parameters and to

predict allergenicity of new proteins, in the future.


The authors declare that they have no conflicts of interest.

Acknowledgements

This study was supported by the EU COST Action ImpARAS

FA1402. The opinions expressed herein and the conclusions

of this publication are those of the authors.

References[1] Ladics GS, Selgrade MK. Identifying food proteins with allergenic

potential: evolution of approaches to safety assessment and research to

provide additional tools. Regul Toxicol Pharmacol 2009;54(Suppl.):S2–6.

[2] Roth-Walter F, Berin MC, Arnaboldi P, Escalante CR, Dahan S, Rauch J,

et al. Pasteurization of milk proteins promotes allergic sensitization by

enhancing uptake through Peyer’s patches. Allergy 2008;63:882–90.

[3] Pizzuti D, Senzolo M, Buda A, Chiarelli S, Giacomelli L, Mazzon E, et al. In

vitro model for IgE mediated food allergy. Scand J Gastroenterol

2011;46:177–87.

[4] Sjogren E, Abrahamsson B, Augustijns P, Becker D, Bolger MB, Brewster M,

et al. In vivo methods for drug absorption – comparative physiologies,

model selection, correlations with in vitro methods (IVIVC), and

applications for formulation/API/excipient characterization including

food effects. Eur J Pharm Sci 2014;57:99–151.

[5] Menard S, Cerf-Bensussan N, Heyman M. Multiple facets of intestinal

permeability and epithelial handling of dietary antigens. Mucosal

Immunol 2010;3:247–59.

[6] Rao JN, Wang JY. Regulation of gastrointestinal mucosal growth.

Colloquium series on integrated systems physiology: from molecule to

function, 2011;p. 1–114.

[7] Lozupone CA, Stombaugh JI, Gordon JI, Jansson JK, Knight R. Diversity,

stability and resilience of the human gut microbiota. Nature

2012;489:220–30.

[8] Hamada H, Hiroi T, Nishiyama Y, Takahashi H, Masunaga Y, Hachimura S,

et al. Identification of multiple isolated lymphoid follicles on the

antimesenteric wall of the mouse small intestine. J Immunol

2002;168:57–64.

[9] Mowat AM. Anatomical basis of tolerance and immunity to intestinal

antigens. Nat Rev Immunol 2003;3:331–41.

[10] Barker N, van Es JH, Kuipers J, Kujala P, van den Born M, Cozijnsen M,

et al. Identification of stem cells in small intestine and colon by marker

gene Lgr5. Nature 2007;449:1003–7.

[11] Peterson LW, Artis D. Intestinal epithelial cells: regulators of barrier

function and immune homeostasis. Nat Rev Immunol 2014;14:141–53.

[12] Wells JM, Spence JR. How to make an intestine. Development

2014;141:752–60.

[13] Moreno FJ. Gastrointestinal digestion of food allergens: effect on their

allergenicity. Biomed Pharmacother 2007;61:50–60.

[14] Reitsma M, Westerhout J, Wichers HJ, Wortelboer HM, Verhoeckx KC.

Protein transport across the small intestine in food allergy. Mol Nutr Food

Res 2014;58:194–205.

[15] Jang MH, Kweon MN, Iwatani K, Yamamoto M, Terahara K, Sasakawa C,

et al. Intestinal villous M cells: an antigen entry site in the mucosal

epithelium. Proc Natl Acad Sci U S A 2004;101:6110–5.

[16] Brandtzaeg PE. Current understanding of gastrointestinal

immunoregulation and its relation to food allergy. Ann N Y Acad Sci

2002;964:13–45.

[17] Heyman M. Gut barrier dysfunction in food allergy. Eur J Gastroenterol

Hepatol 2005;17:1279–85.

[18] Chopra DP, Dombkowski AA, Stemmer PM, Parker GC. Intestinal

epithelial cells in vitro. Stem Cells Dev 2010;19:131–41.


[19] Navabi N, McGuckin MA, Linden SK. Gastrointestinal cell lines form

polarized epithelia with an adherent mucus layer when cultured in semi-

wet interfaces with mechanical stimulation. PLOS ONE 2013;8:e68761.

[20] Sambuy Y, De Angelis I, Ranaldi G, Scarino ML, Stammati A, Zucco F. The

Caco-2 cell line as a model of the intestinal barrier: influence of cell and

culture-related factors on Caco-2 cell functional characteristics. Cell Biol

Toxicol 2005;21:1–26.

[21] Beduneau A, Tempesta C, Fimbel S, Pellequer Y, Jannin V, Demarne F,

et al. A tunable Caco-2/HT29-MTX co-culture model mimicking variable

permeabilities of the human intestine obtained by an original seeding

procedure. Eur J Pharm Biopharm 2014;87:290–8.

[22] Lievin-Le Moal V, Servin AL. Pathogenesis of human enterovirulent

bacteria: lessons from cultured, fully differentiated human colon cancer

cell lines. Microbiol Mol Biol Rev 2013;77:380–439.

[23] Cencic A, Langerholc T. Functional cell models of the gut and their

applications in food microbiology – a review. Int J Food Microbiol

2010;141(Suppl 1):S4–14.

[24] Brosnahan AJ, Brown DR. Porcine IPEC-J2 intestinal epithelial cells in

microbiological investigations. Vet Microbiol 2012;156(3–4):229–37.

[25] Moreno FJ, Rubio LA, Olano A, Clemente A. Uptake of 2S albumin

allergens, Ber e 1 and Ses i 1, across human intestinal epithelial Caco-2 cell

monolayers. J Agric Food Chem 2006;54(22):8631–9.

[26] Song CH, Liu ZQ, Huang S, Zheng PY, Yang PC. Probiotics promote

endocytic allergen degradation in gut epithelial cells. Biochem Biophys

Res Commun 2012;426(1):135–40.

[27] Zhao Y, Liu D, Han R, Zhang X, et al. Soybean allergen glycinin induced

the destruction of the mechanical barrier function in IPEC-J2. Food Agric

Immunol 2015;26:601–9.

[28] Sewekow E, Bimczok D, Kahne T, Faber-Zuschratter H, Kessler LC, Seidel-

Morgenstern A, et al. The major soyabean allergen P34 resists proteolysis

in vitro and is transported through intestinal epithelial cells by a caveolae-

mediated mechanism. Br J Nutr 2012;108(9):1603–11.

[29] Demarne F, Jannin V, Lamprecht A, Pellequer Y, Beduneau A. Cell culture

method and system for establishing an in vitro model of the intestinal

barrier; 2012, WO/2012/110641.

[30] Araujo F, Sarmento B. Towards the characterization of an in vitro triple co-

culture intestine cell model for permeability studies. Int J Pharm

2013;458(1):128–34.

[31] Rytkonen J, Valkonen KH, Virtanen V, Foxwell RA, Kyd JM, Cripps AW,

et al. Enterocyte and M-cell transport of native and heat-denatured bovine

beta-lactoglobulin: significance of heat denaturation. J Agric Food Chem

2006;54(4):1500–7.

[32] Leushacke M, Barker N. Ex vivo culture of the intestinal epithelium:

strategies and applications. Gut 2014;63(8):1345–54.

[33] Schwank G, Koo BK, Sasselli V, Dekkers JF, Heo I, Demircan T, et al.

Functional repair of CFTR by CRISPR/Cas9 in intestinal stem cell

organoids of cystic fibrosis patients. Cell Stem Cell 2013;13(6):653–8.

[34] Sambruy Y, Ferruzza S, Ranaldi G, De Angelis I. Intestinal cell culture

models: applications in toxicology and pharmacology. Cell Biol Toxicol

2001;17(4–5):301–17.

[35] Alam MA, Al-Jenoobi FI, Al-Mohizea AM. Everted gut sac model as a tool in

pharmaceutical research: limitations and applications. J Pharm

Pharmacol 2012;64(3):326–36.

[36] Clarke LL. A guide to Ussing chamber studies of mouse intestine. Am J

Physiol Gastrointest Liver Physiol 2009;296(6):G1151–66.

[37] Lomasney KW, Hyland NP. The application of Ussing chambers for

determining the impact of microbes and probiotics on intestinal ion

transport. Can J Physiol Pharmacol 2013;91(9):663–70.

[38] He L, Yin Y, Li T, Huang R, Xie M, Wu Z, et al. Use of the Ussing chamber

technique to study nutrient transport by epithelial tissues. Front Biosci

(Landmark Ed) 2013;18:1266–74.

[39] Sjoberg A, Lutz M, Tannergren C, Wingolf C, Borde A, Ungell AL.

Comprehensive study on regional human intestinal permeability and

prediction of fraction absorbed of drugs using the Ussing chamber

technique. Eur J Pharm Sci 2013;48(1–2):166–80.

[40] Patterson JK, Lei XG, Miller DD. The pig as an experimental model for

elucidating the mechanisms governing dietary influence on mineral

absorption. Exp Biol Med (Maywood) 2008;233(6):651–64.


























































































































[41] Westerhout J, van de Steeg E, Grossouw D, Zeijdner EE, Krul CA, Verwei M,

et al. A new approach to predict human intestinal absorption using

porcine intestinal tissue and biorelevant matrices. Eur J Pharm Sci

2014;63:167–77.

[42] Van Mierlo LD, Meiland FJ, Van Hout HP, Droes RM. Toward an evidence-

based implementation model and checklist for personalized dementia

care in the community. Int Psychogeriatr 2015;1–13.

[43] Price D, Ackland L, Suphioglu C. Nuts ‘n’ guts: transport of food

allergens across the intestinal epithelium. Asia Pac Allergy 2013;3(4):

257–65.

[44] Cavic M, Grozdanovic MM, Bajic A, Jankovic R, Andjus PR, Gavrovic-

Jankulovic M. The effect of kiwifruit (Actinidia deliciosa) cysteine protease

actinidin on the occludin tight junction network in T84 intestinal

epithelial cells. Food Chem Toxicol 2014;72:61–8.

[45] Bodinier M, Legoux MA, Pineau F, Triballeau S, Segain JP, Brossard C, et al.

Intestinal translocation capabilities of wheat allergens using the Caco-2

cell line. J Agric Food Chem 2007;55(11):4576–83.

[46] Di Leo V, Yang PC, Berin MC, Perdue MH. Factors regulating the effect of

IL-4 on intestinal epithelial barrier function. Int Arch Allergy Immunol

2002;129(3):219–27.

[47] Jacob C, Yang PC, Darmoul D, Amadesi S, Saito T, Cottrell GS, et al. Mast

cell tryptase controls paracellular permeability of the intestine. Role of

protease-activated receptor 2 and beta-arrestins. J Biol Chem

2005;280(36):31936–48.


























DRUG DISCOVERY

TODAY

DISEASEMODELS

Static and dynamic in vitro digestionmodels to study protein stability in thegastrointestinal tractDidier Dupont1,*, Alan R. Mackie2

1STLO, Agrocampus Ouest, INRA, 35000 Rennes, France2Institute of Food Research, Norwich Research Park, Colney Lane, Norwich, Norfolk NR4 7UA, UK


Editors-in-Chief




Food protein allergenicity has been linked to the sur-

vival of the allergen in the gastrointestinal tract. There-

fore, in vitro digestion models have been widely used as

tools to help predicting allergenicity. A huge diversity

of static in vitro digestion models based on different

experimental conditions have been proposed in the

literature making the comparison between studies

impossible. For this reason, an internationally harmo-

nized static model has recently been developed. Dy-

namic in vitro digestion models are complex but more

physiologically relevant and could represent an excel-

lent alternative to study allergenic food digestion.

Overall, these models have shown that the ability of

a protein to survive in the gastrointestinal tract highly

depends on whether the protein is pure or embedded

into a complex food matrix.

Introduction

Introducing new protein sources to our daily diet is not easy

and requires making sure that these proteins will not generate

adverse reactions like allergy. However, there is a current lack

of methods that could allow prediction of the allergenic

properties of a food protein and the mechanisms that

make a protein an allergen are still under investigation.

*Corresponding author: D. Dupont ([email protected])



Nevertheless, it has been hypothesized that for eliciting an

allergenic reaction, a protein has to partly persist in the

gastrointestinal tract and pass through the epithelial barrier

to come into contact with immune cells. The present paper

aims to review the different types of in vitro digestion models

available and discuss their physiological relevance to investi-

gating food protein hydrolysis in the gastrointestinal tract.

Is there a link between digestibility and allergenicity?

A possible connection between the ability of a protein to

resist the digestive process and its ability to raise an allergic

reaction is still highly controversial. The protein does not

have to be intact when reaching the epithelial cells and

peptides generated by the digestion process and long

enough to contain at least 2 epitopes could be responsible

for sensitization [1]. The general opinion appears to be that

the lower limit for allergenicity of peptides is a Mw of

approximately 3.5 kDa [2]. Astwood et al. [3], using a rather

basic incubation test with pepsin, compared the resistance

to pepsin digestion of 16 known food allergens, that is,

6.06.002 23







ovalbumin, b-lactoglobulin, Ara h2, b-conglycinin, among

others and 9 common plant proteins considered to be non-

allergens like Rubisco LSU and SSU from spinach leaf,

lipoxygenase from soybean seed, sucrose synthetase from

wheat kernel, b-amylase from barney kernel or acid phos-

phatase and phosphofructokinase from potato tuber. They

showed that while major food allergens in general resisted

the digestion process, non-allergenic proteins (mainly

enzymes) were by contrast rapidly digested [3]. Using stur-

geon caviar and parvalbumin, the major fish allergen, as

examples, impairment of the digestion process was shown

to increase allergenicity of the proteins under investigation

in a Balb/c mouse model further supporting the hypothesis

of a link between resistance to digestion and allergenicity

[4]. These results were confirmed in human adults a few

years later by the same group [5]. When reviewing all the

literature available on digestibility studies of pure allergens,

Bøgh and Madsen did not find clear evidence of such a link

[6] but this could be due to the wide range of digestion

methods employed in the studies reviewed, many of which

were not physiologically relevant. Studies assessing the

allergenicity of digestion products, by either IgE-binding,

elicitation or sensitizing capacity shows that digestion may

abolish, decrease, have no effect, or even increase the

allergenicity of food allergens. For example, Fu et al. tested

several similar allergenic and nominally non-allergenic

proteins with similar cellular functions. They selected 23

allergens including 15 storage proteins (casein, b-lactoglob-

ulin, ovalbumin, conalbumin, Ara h1, Ara h2, among

others), 2 plant lectins from soybean and peanut, 5

enzymes (lysozyme, lactoperoxidase, papain, bromelain

and actinidin) and 1 contractile protein, that is, tropomy-

osin from shrimp. They compared the resistance of these

known allergens to 16 proteins with similar functions but

unproven allergenicity: 4 storage proteins (a-lactalbumin,

zein, and 2 trypsin inhibitors), 5 plant lectins from pea,

lentil, lima bean, jack bean and red kidney bean, 4 enzymes

(cytochrome c, rubisco, phosphofructokinase and sucrose

synthetase) and 3 contractile proteins, that is, tropomyosin

from bovine, chicken and pork. They found there was no

clear relationship between digestibility measured in vitro

and protein allergenicity [7]. The overall controversy can

certainly be explained by the different experimental con-

ditions (enzyme: substrate ratio, pH and duration of the

gastric phase, among others) that were used in those dif-

ferent studies and also by differences in analytical techni-

ques that were used to characterize the digested product.

There are several structural families of allergens that are

more resistant to proteolysis than others. For example, so

called lipid transfer proteins have been shown to be a pan-

allergen with a degree of cross-reactivity comparable to

profilin. It shows significant resistance to pepsin digestion

[8]. Similarly, the IgE binding capacity of thaumatin-like


protein Act d 2 from kiwi was found to be largely unaffect-

ed by low pH and simulated digestion [9]. By contrast,

protein families such as patatin, zein, chlorophyll binding

or flavodoxin contain few or no known allergens [10].

Another important aspect to consider is that allergens are

not consumed as pure proteins but are embedded into com-

plex food matrices. Interactions with other food constituents

or differences in the propensity of proteases to interact with

different proteins might dramatically modify the hydrolysis

of an allergen in the gastrointestinal tract. Furthermore, the

pH of a food is usually between 4 and 7 and its buffering

capacity will significantly increase the pH of the stomach

during the first stages of digestion consequently limiting the

activity of the main gastric protease, that is, pepsin whose

optimal activity is around pH 2 [11]. This will strongly reduce

the proteolysis and intact proteins have been shown to reach

the small intestine even after 20 min of gastric digestion [12].

Finally, protein structure can be significantly affected by the

physico-chemical conditions found in the gastrointestinal

tract, affecting the rate of proteolysis. One of the best exam-

ples to emphasize the importance of these structural mod-

ifications is the case of milk caseins. Caseins consist of 4

individual proteins (as1, as2, b and k) that are organized in

milk into a supramolecular structure called the casein mi-

celle. Submitted individually as pure proteins to an in vitro

digestion model, caseins will be cleaved and reduced into

short peptides within a few minutes [13]. However, when

ingested in the form of milk, caseins will clot in the stomach

due to the acid conditions and form a curd that will be retain

in the stomach and slowly released as curd particles in the

small intestine. For this reason, caseins have been called ‘slow

proteins’ [14] and it is therefore not surprising that caseins are

considered as major allergens for the pediatric population. By

contrast, the whey protein, b-lactoglobulin, is generally high-

ly resistant to gastric proteolysis. However, when it becomes

adsorbed to the surface of oil droplets its digestibility is

altered radically and a significant proportion, most probably

the population of molecules directly adsorbed to the oil

droplet surface, becomes highly digested, probably as a con-

sequence of denaturation [15]. In addition, the whey portion

of milk remains in solution under gastric conditions and so is

emptied from the gastric compartment relatively quickly and

is subsequently hydrolyzed by duodenal proteases and has

thus been designated as a ‘fast protein’ [16].

Finally other routes for generating allergic reactions to food

have been described like the respiratory mucosa [17] or the

skin [18]. For example, inhalant allergens are able to sensitize

subjects that will exhibit an allergic reaction when cross-

reacting food allergens are ingested [19–21].

The pepsin resistance test

In vitro testing has a central place in the risk assessment

process for allergenicity evaluation. In vitro digestion tests,


cell-based assays and IgE-binding tests are among the tools

that can be combined to have a rough idea of the allergenic

potential of a protein source. One of the first tests to assess

protein digestibility as a way to predict allergenicity was the

pepsin resistance test formerly proposed by Astwood et al. [3].

It consists of hydrolyzing food proteins with 0.32% pepsin at

pH 1.2. Three patterns of stability of the allergens included in

the study were observed:

1. Complete stability resisting pepsinolysis for 60 min

2. Intermediate stability, proteins resisting digestion for at

least 30 s but being digested within 60 min

3. Protein completely susceptible to proteolysis with no in-

tact protein remaining after the first time point sampled

(15 s), with stable fragments being observed for at least

8 min.

This study concluded that resistance to pepsinolysis was

indicative of allergenic potential, and as a consequence it was

proposed to include the pepsin resistance test in the decision

tree approach to allergenicity risk assessment by Metcalfe

et al. [22] which was then taken up by FAO/WHO Codex

Alimentarius Commission [23].

In vitro gastrointestinal digestion models for predicting

allergenicity

The pepsin resistance test is based on drastic conditions that

exacerbate the hydrolytic action of pepsin. The pH is ex-

tremely low (1.2) and the enzyme: substrate ratio is high, far

from the physiological reality [24]. Furthermore, this test

takes only the gastric phase into account whereas it has been

shown that a protein can be highly resistant to gastric

digestion but be completely hydrolyzed within a few min-

utes when entering the small intestine [25]. Therefore, other

groups have developed gastroduodenal or gastrointestinal

models taking intestinal proteolysis into account and dozens

of in vitro digestion models have been developed and pub-

lished. Among these models, some have been specifically

used for assessing protein allergenicity. For example, a sim-

ulated gastrointestinal digestion has been carried out on

purified peach lipid transfer protein, one of the main aller-

gens among the population of the Mediterranean area and

the major allergen of peach allergic patients [26]. About two

thirds of the proteins were hydrolyzed during digestion and

the peptides formed essentially derived from trypsin action,

whereas the protein appeared to be resistant to pepsin and

chymotrypsin. The intact protein and some high Mw pep-

tides were found to be recognized by patients’ sera. More

recently, three edible mealworm species (Tenebrio molitor,

Zophobas atratus and Alphitobius diaperinus) subjected to

processing and in vitro digestion were analyzed for IgE cross-

reactivity [27]. IgE from crustaceans or house dust

mite allergic patients showed cross-reactivity to mealworm

tropomyosin or alpha-amylase, hexamerin 1B precursor and

muscle myosin, respectively. Heat processing as well as in

vitro digestion did diminish, but not eliminate, house dust

mite or tropomyosin IgE cross-reactivity. These two exam-

ples selected among many others show the interest of in vitro

digestion protocols as first screening tools to assess the

allergenicity of food proteins or new protein sources. How-

ever, whereas the outcome of digestion studies is sometimes

clear and easy to interpret for proteins that are either highly

resistant to digestion or rapidly and fully hydrolyzed, it is

more difficult for proteins that show an intermediate behav-

ior. How to should a protein that needs a long time to be fully

digested be assessed? More data are needed for a better

guidance to interpret digestion outcomes. Another difficulty

is that all these models differ in their physicochemical con-

ditions (pH, enzyme: substrate ratio, ionic strength of the

medium) and their duration making a comparison of data

between different studies impossible.

The Infogest consensus in vitro digestion protocol

Infogest was a COST Action (http://www.cost-infogest.eu) that

took place between May 2011 and May 2015. The objective of

this international network was to gather scientists from differ-

ent disciplines (food science, nutrition, gastroenterology,

among others) to improve health properties of food by sharing

our knowledge on the digestive process. It involved 340 scien-

tists from 130 institutes in 37 countries (Europe but also New

Zealand, Australia, USA, Argentina, among others). One aim of

the network was to consolidate conditions for simulated di-

gestion of food and find a consensus, if possible, for a digestion

model. A frameset of parameters including the oral, gastric and

small intestinal digestion were outlined and their relevance

discussed in relation to available in vivo data and enzymes. A

consensus paper was released [24] giving a detailed protocol

and line-by-line guidance, recommendations and justifica-

tions but also limitation of the proposed simple static model.

A YouTube channel was created with videos showing how to

run the model, calibrate the digestive enzymes and quantify

the bile salts allowing the new comers to conduct experiments

in the proper way (https://www.youtube.com/channel/

UCdc-NPx9kTDGyH_kZCgpQWg). To validate this protocol,

an inter-laboratory trial on the in vitro digestion of skimmed

milk was conducted within the INFOGEST network [28]. The

degree of consistency in protein hydrolysis was investigated.

Analysis of the hydrolyzed proteins, after the gastric and

intestinal phases, showed that caseins were mainly hydrolyzed

during the gastric phase, whereas b-lactoglobulin was, as

previously shown, resistant to pepsin. Moreover, generation

of free amino acids occurred mainly during the intestinal

phase. The study also showed that a few crucial steps were

responsible for the remaining inter-laboratory variability. The

largest deviations arose from the determination of pepsin

activity. Therefore, this step was further clarified, standardized,


http://www.cost-infogest.eu/

https://w**w**w.youtube.com/channel/UCdc-NPx9kTDGyH_kZCgpQWg

https://w**w**w.youtube.com/channel/UCdc-NPx9kTDGyH_kZCgpQWg


and implemented in a third inter-laboratory study. The ‘har-

monized’ static, in vitro digestion method for food which will

aid the production of more comparable data in the future and

has started to be used all around the world. It has been used to

study the digestion of major allergens of egg [29], milk [30] and

pasta [31]. It has been recently compared with in vivo data

obtained in pigs for the digestion of skimmed milk showing an

excellent correlation with the extent of proteolysis observed

with the animal model used (manuscript in preparation). Since

the model has been detailed in an open access publication and

media, challenged in inter-laboratory trials, validated toward

in vivo data and is currently widely used, it represents an

excellent tool for assessing the resistance of new protein

sources to digestion including processed foods containing

these proteins.

Would dynamic in vitro digestion models be relevant?

Digestion is a dynamic process. Food entering the gastroin-

testinal tract will be transferred from one compartment to

another at variable rates depending on its structure, caloric

content, osmolarity and rheological properties. Physico-

chemical conditions (pH, ionic strength, digestive enzyme

concentrations, among others) occurring in the different

compartments will evolve with time. Static in vitro digestion

models do not take these evolutions with time into account.

By contrast, several dynamic multi-compartmental models

have been developed during the past decades and recently

reviewed [32]. One of the most well-known is the TIM model

that was developed at TNO (the Netherlands) in the nineties

[33] and is commercially available. The model has been used

to study the fate of gluten [34] and milk allergens [35] in the

digestive tract. Another multi-compartmental dynamic mod-

el is the SHIME1 that was developed at Ghent University

(Belgium), representing the gastrointestinal tract (GIT) of the

adult human, as described by Molly et al. [36]. It consists of a

succession of five reactors (stomach, small intestine, ascend-

ing, transverse and descending colon) simulating the differ-

ent parts of the gastrointestinal tract. More recently, new

dynamic models have been developed like the DIDGI1 at

INRA (France) [37] and the SIMGI1 at CSIC (Spain) [38] and

mainly used for studying the digestion of milk and dairy

products [39]. When relevant physiological parameters are

available for setting up these systems, they have been shown

to be able to closely mimic the fate of food in the gastroin-

testinal tract and have been validated against in vivo data

[37,40,41].

To be physiologically relevant, in vitro dynamic models

need to be properly programmed. For most of the existing

systems, key information needs to be entered in the software.

For instance, the gastric emptying half-time is one of these

key parameters and will be highly dependent on the proper-

ties of the food (caloric charge, viscosity, structure, osmolari-

ty) that contains the allergens. Also the evolution of pH in the


stomach is of crucial importance and will also highly depend

on the buffering capacity of the food itself. For these reasons,

it is rather difficult to use dynamic models to study the

digestion of pure allergens in aqueous solution but these

models are extremely relevant to study the digestion of

allergens in real foods. Harmonizing at the international level

the physiological parameters that would be relevant to digest

different families of foods in dynamic conditions is one of the

future objectives of the Infogest network.

Conclusion and perspectives

Resistance to digestion is one of the criteria to distinguish

allergenic from non-allergenic proteins/foods. This criteria

will be properly assessed only if physiologically relevant in

vitro digestion models are used. The Infogest consortium have

developed a simple static model that can be used in a consis-

tent manner and gives results that appear to mimic the

situation in vivo. Nevertheless, interpretation of digestion

data is sometimes difficult especially for allergens not show-

ing a strong resistance or a rapid hydrolysis and more guid-

ance on digestion output is needed. Recently, the model has

been applied to food allergens but more evidence is needed to

make sure that, for allergens, it would correlate with in vivo

data. Dynamic models are more complex but much more

physiologically relevant than static ones. They would be of

great interest in the future to study the persistence in the GI

tract of allergens embedded in their foods. Research effort is

urgently needed to validate these models for their ability to

predict allergenicity. Microsystems are currently being devel-

oped [42] and would help in limiting the quantity of pure

allergens to digest. In silico models [43] could also be of

interest for simulating food digestion, but have not been

applied so far to food allergens to our knowledge. Finally

more models simulating the digestive process of specific

populations like infant [44] or the elderly [45] will need to

be tested for their ability to predict protein allergenicity.


The authors have no conflict of interest to declare.

References[1] Mills ENC, Jenkins JA, Alcocer MJC, Shewry PR. Structural, biological, and

evolutionary relationships of plant food allergens sensitizing via the

gastrointestinal tract. Crit Rev Food Sci Nutr 2004;44(5):379–407.

[2] Lack G, Chapman M, Kalsheker N, King V, Robinson C, Venables K. Report

on the potential allergenicity of genetically modified organisms and their

products. Clin Exp Allergy 2002;32(8):1131–43.

[3] Astwood JD, Leach JN, Fuchs RL. Stability of food allergens to digestion in

vitro. Nat Biotechnol 1996;14(10):1269–73.

[4] Untersmayr E, Scholl I, Swoboda I, Beil WJ, Forster-Waldl E, Walter F, et al.

Antacid medication inhibits digestion of dietary proteins and causes food

allergy: a fish allergy model in Balb/c mice. J Allergy Clin Immunol

2003;112(3):616–23.

[5] Untersmayr E, Bakos N, Scholl I, Kundi M, Roth-Walter F, Szalai K, et al.

Anti-ulcer drugs promote IgE formation toward dietary antigens in adult

patients. FASEB J 2005;19(1):656–8.

















[6] Bøgh KL, Madsen CB. Food allergens: is there a correlation between

stability to digestion and allergenicity? Crit Rev Food Sci Nutr

2015;56(5):1545–1567.

[7] Fu TT, Abbott UR, Hatzos C. Digestibility of food allergens and

nonallergenic proteins in simulated gastric fluid and simulated intestinal

fluid – a comparative study. J Agric Food Chem 2002;50(24):7154–60.

[8] Asero R, Mistrello G, Roncarolo D, de Vries SC, Gautier MF, Ciurana LF,

et al. Lipid transfer protein: a pan-allergen in plant-derived foods that is

highly resistant to pepsin digestion. Int Arch Allergy Immunol

2000;122(1):20–32.

[9] Bublin M, Radauer C, Knulst A, Wagner S, Scheiner O, Mackie AR, et al.

Effects of gastrointestinal digestion and heating on the allergenicity of the

kiwi allergens Act d 1, actinidin, and Act d 2, a thaumatin-like protein. Mol

Nutr Food Res 2008;52(10):1130–9.

[10] Jenkins JA, Griffiths-Jones S, Shewry PR, Breiteneder H, Mills ENC.

Structural relatedness of plant food allergens with specific reference to

cross-reactive allergens: an in silico analysis. J Allergy Clin Immunol

2005;115(1):163–70.

[11] Piper DW, Fenton BH. pH stability and activity curves of pepsin with

special reference to their clinical importance. Gut 1965;6(5):506–8.

[12] Barbe F, Menard O, Le Gouar Y, Buffiere C, Famelart MH, Laroche B, et al.

The heat treatment and the gelation are strong determinants of the

kinetics of milk proteins digestion and of the peripheral availability of

amino acids. Food Chem 2013;136(3–4):1203–12.

[13] Dupont D, Mandalari G, Molle D, Jardin J, Leonil J, Faulks RM, et al.

Comparative resistance of food proteins to adult and infant in vitro

digestion models. Mol Nutr Food Res 2010;54(6):767–80.

[14] Boirie Y, Dangin M, Gachon P, Vasson MP, Maubois JL, Beaufrere B. Slow

and fast dietary proteins differently modulate postprandial protein

accretion. Proc Natl Acad Sci U S A 1997;94(26):14930–35.

[15] Macierzanka A, Sancho AI, Mills ENC, Rigby NM, Mackie AR.

Emulsification alters simulated gastrointestinal proteolysis of beta-casein

and beta-lactoglobulin. Soft Matter 2009;5(3):538–50.

[16] Veldhorst MAB, Nieuwenhuizen AG, Hochstenbach-Waelen A, van Vught

AJAH, Westerterp KR, Engelen MPKJ, et al. Dose-dependent satiating effect

of whey relative to casein or soy. Physiol Behav 2009;96(4–5):675–82.

[17] Jahn-Schmid B, Radakovics A, Luttkopf D, Scheurer S, Vieths S, Ebner C,

et al. Bet v 1(142–156) is the dominant T-cell epitope of the major birch

pollen allergen and important for cross-reactivity with Bet v 1-related food

allergens. J Allergy Clin Immunol 2005;116(1):213–9.

[18] Strid J, Hourihane J, Kimber I, Callard R, Strobel S. Epicutaneous exposure

to peanut protein prevents oral tolerance and enhances allergic

sensitization. Clin Exp Allergy 2005;35(6):757–66.

[19] Crespo JF, Rodriguez J, James JM, Daroca P, Reano M, Vives R. Reactivity to

potential cross-reactive foods in fruit-allergic patients: implications for

prescribing food avoidance. Allergy 2002;57(10):946–9.

[20] Fernandez-Rivas M, Bolhaar S, Gonzalez-Mancebo E, Asero R, van LA,

Bohle B, et al. Apple allergy across Europe: how allergen sensitization

profiles determine the clinical expression of allergies to plant foods. J

Allergy Clin Immunol 2006/8;118(2):481–8.

[21] Zuidmeer L, Goldhahn K, Rona RJ, Gislason D, Madsen C, Summers C,

et al. The prevalence of plant food allergies: a systematic review. J Allergy

Clin Immunol 2008;121(5):1210–8.

[22] Metcalfe DD, Astwood JD, Townsend R, Sampson HA, Taylor SL, Fuchs RL.

Assessment of allergenic potential of foods derived from genetically

engineered crop plants. Crit Rev Food Sci Nutr 1996;36:S165–86.

[23] Commission CA.In: Alinorm 03/34: Joint FAO/WHO Food Standard

Programme, Codex Alimentarius Commission, 25th Session; 2003.p. 47–60.

[24] Minekus M, Alminger M, Alvito P, Ballance S, Bohn T, Bourlieu C, et al. A

standardised static in vitro digestion method suitable for food – an

international consensus. Food Funct 2014;5(6):1113–24.

[25] Macierzanka A, Bottger F, Lansonneur L, Groizard R, Jean A-S, Rigby NM,

et al. The effect of gel structure on the kinetics of simulated

gastrointestinal digestion of bovine b-lactoglobulin. Food Chem

2012;134(4):2156–63.

[26] Cavatorta V, Sforza S, Aquino G, Galaverna G, Dossena A, Pastorello EA,

et al. In vitro gastrointestinal digestion of the major peach allergen Pru p

3, a lipid transfer protein: molecular characterization of the products and

assessment of their IgE binding abilities. Mol Nutr Food Res

2010;54(10):1452–7.

[27] van Broekhoven S, Bastiaan-Net S, de Jong NW, Wichers HJ. Influence of

processing and in vitro digestion on the allergic cross-reactivity of three

mealworm species. Food Chem 2016;196:1075–83.

[28] Egger L, Menard O, Delgado-Andrade C, Alvito P, Assuncao R, Balance S,

et al. The harmonized INFOGEST in vitro digestion method: from

knowledge to action. Food Res Int 2016;217–225.

[29] Benede S, Lopez-Exposito I, Molina E, Lopez-Fandino R. Egg proteins as

allergens and the effects of the food matrix and processing. Food Funct

2015;6(3):694–713.

[30] Picariello G, Miralles B, Mamone G, Sanchez-Rivera L, Recio I, Addeo F,

et al. Role of intestinal brush border peptidases in the simulated digestion

of milk proteins. Mol Nutr Food Res 2015;59(5):948–56.

[31] Manione G, Nitride C, Picariello G, Addeo F, Ferranti P, Mackie A.

Tracking the fate of pasta (T. Durum Semolina) immunogenic proteins by

in vitro simulated digestion. J Agric Food Chem 2015;63(10):2660–7.

[32] Guerra A, Etienne-Mesmin L, Livrelli V, Denis S, Blanquet-Diot S, Alric M.

Relevance and challenges in modeling human gastric and small intestinal

digestion. Trends Biotechnol 2012;30(11):591–600.

[33] Minekus M, Marteau P, Havenaar R, Huisintveld JHJ. A

multicompartmental dynamic computer-controlled model simulating

the stomach and small-intestine. ATLA – Altern Lab Animals

1995;23(2):197–209.

[34] Mitea C, Havenaar R, Drijfhout JW, Edens L, Dekking L, Koning F. Efficient

degradation of gluten by a prolyl endoprotease in a gastrointestinal

model: implications for coeliac disease. Gut 2008;57(1):25–32.

[35] Havenaar R, de Jong A, Koenen ME, van Bilsen J, Janssen AM, Labij E, et al.

Digestibility of transglutaminase cross-linked caseinate versus native

caseinate in an in vitro multicompartmental model simulating young

child and adult gastrointestinal conditions. J Agric Food Chem

2013;61(31):7636–44.

[36] Molly K, Woestyne MV, Verstraete W. Development of a 5-step

multichamber reactor as a simulation of the human intestinal microbial

ecosystem. Appl Microbiol Biotechnol 1993;39(2):254–8.

[37] Menard O, Cattenoz T, Guillemin H, Souchon I, Deglaire A, Dupont D,

et al. Validation of a new in vitro dynamic system to simulate infant

digestion. Food Chem 2014;145:1039–45.

[38] Barroso E, Cueva C, Pelaez C, Martinez-Cuesta MC, Requena T.

Development of human colonic microbiota in the computer-controlled

dynamic SIMulator of the GastroIntestinal tract SIMGI. LWT – Food Sci

Technol 2015;61(2):283–9.

[39] Sanchez-Rivera L, Menard O, Recio I, Dupont D. Peptide mapping during

dynamic gastric digestion of heated and unheated skimmed milk powder.

Food Res Int 2015;77:132–139.

[40] Etcheverry P, Grusak MA, Fleige LE. Application of in vitro bioaccessibility

and bioavailability methods for calcium, carotenoids, folate, iron,

magnesium polyphenols, zinc, and vitamins B-6, B-12, D, and E. Front

Physiol 2012;3:1–22.

[41] Molly K, Vandewoestyne M, Desmet I, Verstraete W. Validation of the

simulator of the human intestinal microbial ecosystem (SHIME) reactor

using microorganism-associated activities. Microb Ecol Health Dis

1994;7(4):191–200.

[42] Marze S, Algaba H, Marquis M. A microfluidic device to study the digestion

of trapped lipid droplets. Food Funct 2014;5(7):1481–8.

[43] Le Feunteun S, Barbe F, Remond D, Menard O, Le gouar Y, Dupont D, et al.

Impact of the dairy matrix structure on milk protein digestion kinetics:

mechanistic modelling based on mini-pig in vivo data. Food Bioprocess

Technol 2013;7:1099–1113.

[44] de Oliveira SC, Bourlieu C, Menard O, Bellanger A, Henry G, Rousseau F,

et al. Impact of pasteurization of human milk on preterm newborn in vitro

digestion: gastrointestinal disintegration, lipolysis and proteolysis. Food

Chem 2016;211:171–9.

[45] Levi CS, Lesmes U. Bi-compartmental elderly or adult dynamic digestion

models applied to interrogate protein digestibility. Food Funct

2014;5(10):2402–9.










































































































































DRUG DISCOVERY

TODAY

DISEASEMODELS

Epithelial models to study foodallergen-induced barrier disruptionand immune activationMarija Gavrovic-Jankulovic1, Linette E.M. Willemsen2,*1Department of Biochemistry, Faculty of Chemistry University of Belgrade, Belgrade, Serbia2Division of Pharmacology, Utrecht Institute for Pharmaceutical Sciences, Faculty of Science, Utrecht University, Utrecht, The

Netherlands


Editors-in-Chief




Changes in lifestyle, diet and environmental factors in

westernized countries correspond with the rise in non-

communicable diseases affecting metabolic and im-

mune disorders, such as allergies. Therefore the mech-

anisms by which environmental factors and allergens

are capable of elicitating allergic sensitization need to

be further unraveled. In vitro models using human

epithelial cells, with or without immune cells, are

needed to achieve this purpose. Epithelial cells cover

mucosal surfaces and provide a barrier between the

external and internal environment. In mucosal tissues

such as the respiratory and gastro-intestinal tract,

epithelial cells not only contribute to barrier integrity

but also actively regulate dendritic cell function and

adaptive immune responses and can support tolerance

induction or allergic sensitization. Certain allergens

contain protease activity which may facilitate them

to cross the barrier, others are transported via trans-

cytosis. In addition, certain allergens may provoke

epithelial activation resulting in production of TH2

driving immune mediators. Preserving epithelial

homeostasis is important to suppress allergic sensiti-

*Corresponding author: Linette E.M. Willemsen ([email protected])


zation. This review describes in vitro models of human

intestinal epithelial cells and co-culture models that are

currently available to determine barrier disruption or

immune activation induced by food allergens. These

can be used for future development of in vitro models to

study the contribution of intestinal epithelial cells in

allergic sensitization and to identify sensitizing proper-

ties of novel proteins.


Introduction

Changing living conditions in industrialized countries, in-

cluding dietary alterations, increased exposure to environ-

mental pollutants, microbiome alterations and a sedentary

lifestyle, have been linked to the increase in non-communi-

cable diseases including allergies [1–3]. In the western world

depending on the country 5–30% of young people are affect-

ed with asthma and/or rhinitis and 6% of children and 3–4%

of adults with food allergy [4–6]. Allergic sensitization occurs

6.09.002 29







for a large range of food allergens such as cow’s milk, hen’s

egg and peanut proteins and/or inhalant allergens like house

dust mite or pollen. Mucosal tissues covering the lung and

gastro-intestinal tract provide a barrier against environmen-

tal antigens, and support immunological tolerance for harm-

less agents while immunity is raised against pathogenic

intruders [7,8]. However in case of allergic sensitization a T

helper cell 2 (TH2) driven IgE mediated immune response is

raised against relatively harmless proteins (allergens). Epithe-

lial cells protect underlying mucosal lymphoid tissues from

excessive exposure to allergenic proteins. They express pat-

tern recognition receptors (e.g. Toll like receptors), glycan

binding receptors (e.g. galectins), cytokine and chemokine

receptors and produce cytokines, chemokines, galectins and

growth factors that drive immune polarization by affecting

dendritic cell (DC) function and the adaptive immune re-

sponse [9,10]. This review describes the current knowledge on

the contribution of intestinal epithelial cells (IEC) to allergic

sensitization with regard to barrier properties and production

of immune mediators and human in vitro models that can be

used and/or further developed to study these processes.

Epithelial barrier and defects related to allergic

sensitization

In the intestine a monolayer of epithelial cells exhibits nu-

merous physical adaptations to separate the mucosal im-

mune system from the external environment. A brush

border on the apical surface of the epithelium produces

digestive enzymes and allows uptake of nutrients, while

intercellular tight junctions between neighboring epithelial

cells prevent paracellular transport of immunogenic macro-

molecules. This physical barrier is reinforced by a glycocalyx

formed by secretion and apical attachment of a heavily

glycosylated mucin-rich layer further protecting the epithe-

lial lining from microbial attachment and pathogen invasion

[11]. In addition, IgA and digestive enzymes prevent the

uptake of antigenic macromolecules into the body. The gut

epithelium is created from a pool of pluripotent stem cells,

which give rise to five types of IEC: absorptive columnar cells

(enterocytes), goblet, endocrine, Paneth, and M (microfold)

cells. Enterocytes form the vast majority and 10–25% of IEC

consist of mucus producing goblet cells [8]. Cohesion and

Table 1. Examples of (food) allergens with proteolytic activity

activate mediator release in vitro

Allergen source Enzyme Mode of action

House dust mite [24] Der p 1 Cleavage of tight-junction molecules (o

cysteine protease activity

Kiwifruit [25,26] Act d 1 Cleavage of tight-junction molecules (o

Pineapple [27] Ana c2 Widening intercellular junctions, strong

Papaya [27] Car p 1 Loosening of tight junctions


polarity of the epithelial layer are maintained by the apical

tight and adherens junctions, and by the subjacent desmo-

somes [12]. Numerous aeroallergens (house dust mite (Der

p1, Der p9), cockroach, pollen, Penicillium sp., Aspergillus sp.,)

[11,13] and food derived allergens reveal protease activity (see

Table 1). These allergens are involved in the pathogenesis of

allergic diseases through (i) inducing the release of pro-in-

flammatory cytokines via activation of protease-activated

receptors (PARs), which are widely expressed on leukocytes,

endothelium, epithelium, and many airway cells; (ii) the

cleavage of CD23 from activated B cells and CD25 from T

cells to favor the development of TH2-type responses [14,15];

(iii) the degradation of junctional proteins, thus increasing

the permeability of the epithelium in vitro. Also non-proteo-

lytic food allergens can cross the epithelial barrier for exam-

ple via transcytosis (Table 2). Aeroallergens such as house

dust mite allergen Der p2 or Timothy grass allergen Phl p1

[11,13,16–18], have recently been shown to induce airway

epithelial activation resulting in the release of IL-1a, IL-33, IL-

25, TSLP and/or GM-CSF which may contribute to recruit-

ment and activation of DC and innate lymphoid group 2 cells

(ILC2) and consequent TH2 polarization. Similar aspects may

apply for certain food derived allergens such as Peach LTP and

peanut allergens (Table 2). In addition, IL-4 and IL-13 pro-

duced by TH2 cells and/or ILC2, and tryptase secreted by mast

cells, can enhance epithelial permeability via the IL-4/IL-13

receptor or PAR2 receptor respectively [19–22]. Beyond aller-

gens increasing paracellular permeability and crossing the

epithelial barrier via the transcellular route, IgE-allergen

complexes can be transported over IEC via the low affinity

IgE receptor CD23b [23].

Epithelial cells contribute to tolerance induction or

allergic sensitization

The intestinal epithelium is in close contact with dendritic

cells (DC) that sample luminal antigens. M-cells that cover

Peyer’s, caecal and colonic patches, are specialized in the

uptake of particulate antigens and transfer these to DC

in the subepithelial dome that can instruct naıve T-cells

and B-cells [8]. The lamina propria is the effector site of

the intestinal mucosa and contains DC, macrophages, ILC,

T-cells, B-cells, intra epithelial T-cells, eosinophils and mast

known to affect intestinal epithelial barrier integrity and/or

Effect

ccludin, claudin) via Increase in epithelial permeability of intestinal human

biopsy

ccludin) Increase in epithelial permeability of Caco-2 and T84

mucolytic activity Increase in epithelial permeability of Caco-2

Increase in epithelial permeability of Caco-2


Table 2. Examples of food allergens with non-proteolytic activity that traffic the intestinal epithelial barrier and/or activatemediator release in vitro

Allergen source Allergen Mode of action Effect

Peach LTP [28] Pru p3 Lipid raft mediated uptake

and increased epithelial TSLP,

IL-33, IL-25 mRNA

Crosses epithelial barrier and activates Caco-2 epithelial cells

Cow’s milk [29] aLac

bLac

Transcytosis Crosses epithelial barrier of Caco-2

Peanut [30,31] Ara h1/h2 Transcytosis Crosses epithelial barrier of Caco-2

Ara h2 Cellular activation Stimulates a pro-inflammatory response in Caco-2/TC7 cells

Wheat [32] v5-gliadin LTP Transcytosis Crosses epithelial barrier of Caco-2

Egg white [33,34] Gal d1 Transcytosis Crosses epithelial barrier of human breast

Gal d 2 Transcytosis Crosses epithelial barrier of human gastro-intestinal tract

Brazil nuts [35] Ber e 1 Transcytosis Crosses epithelial barrier of Caco-2

Sesame seeds [35] Ses i 1 Transcytosis Crosses epithelial barrier of Caco-2

cells [8]. Intestinal CD103+ DC are crucial in determining the

adaptive immune response to oral antigens, and they traffic

to the mesenteric lymph nodes (MLN) in a CCR7 dependent

manner where they promote tolerance or immunity [36,37].

The CX3CR1hi resident macrophages directly underlie the

epithelium and under homeostatic conditions produce high

amounts of IL-10. They can extend transepithelial dendrites

through the epithelium via the paracellular space to sample

luminal antigen and transfer this to CD103+ DC via connex-

ion 43. Similarly goblet cells transfer antigen via channels to

CD103+ DC [38]. Also CD103+ DC themselves are in close

contact with the epithelium and sample from the lumen. Oral

tolerance is abolished in absence of MLN or CCR7 expressing

DC, while the Peyer’s patches are dispensable. This suggests

that CD103+ migratory DC from the LP are key in oral

tolerance induction [39]. If these cells are instructed to pro-

duce retinoic acid (RA) (high expression of vitamin A con-

verting enzyme aldehyde dehydrogenase) and TGFb and/or

indoleamine 2,3-dioxygenase (IDO) they can induce gut

trophic a4b7+CCR9+FoxP3+regulatory T cells (Treg) that

home back to the lamina propria where they are further

differentiated and expanded by IL-10 producing CX3CR1+

macrophages [8,36,37,40]. Local intestinal factors that gen-

erate these tolerogenic CD103+ DC include the microbiome,

dietary components, leukocytes, stromal cells and neuroen-

docrine mediators as well as IEC derived factors (including

TGFb, TSLP, RA and mucin MUC2) [8,37,40–44] (Fig. 1).

Epithelial cells can instruct TH2 driving OX40L expressing

DC that secrete CCL17 and CCL22 and activate ILC2

[16,45,46]. This was convincingly shown for aero-allergens

like HDM which contains specific allergens (Derp2) and LPS

that activate NFkB signaling in airway epithelial cells. In

response they release IL-1a which via a positive feedback

loop induces IL-33, IL-25, TSLP and endogenous danger

factors such uric acid and airway epithelial cells also can

release DC chemo-attractants CCL2 and CCL20 upon aller-

gen exposure [16,47]. TSLP, IL25, IL33 and uric acid are also

increased in the intestine of mice affected with food allergy,

and in particular IL-33 and uric acid contribute to allergic

sensitization not only for inhalant allergen HDM but also for

food allergen peanut in mice (Fig. 1) [46,48–50].

Human in vitro models of intestinal epithelial cells

The use of in vitro IEC models for transport studies and

allergen uptake focusses on absorptive cells. Because of the

difficulties in culturing isolated primary human IEC and

limited viability, monolayers of human colorectal adenocar-

cinoma cell lines Caco-2, HT-29 and T84 are most often used.

Caco-2 cells are the most popular for use and serve as model

for human intestinal enterocytes. They differentiate sponta-

neously into polarized intestinal cells possessing an apical

brush border and tight junctions between adjacent cells, and

they express hydrolases and typical microvillar transporters

[32]. In the context of food allergy the Caco-2 cell line is the

most often used for allergen uptake [32,35]. However it

remains to be revealed if the permeability data obtained from

the Caco-2 model are predictive for human gastro-intestinal

tract absorption since it is very difficult to measure absorption

of proteins in vivo. HT-29 is another often used human cell

line, and although essentially undifferentiated, HT29 cells in

culture are heterogeneous and contain a small proportion

(i.e. <5%) of mucus-secreting cells and columnar absorptive

cells [51]. HT29-MTX, a stable homogenous subpopulation

obtained from methotrexate treated HT29, exhibit an entire-

ly differentiated goblet cell-like phenotype secreting low

amounts of intestinal type MUC2 mucins [52]. The T84 cell

line has been used as a model of intestinal cells which

produces high molecular weight mucus [53,54]. A very high



Allergic sensitization Tolerance

TSLP

IL-25 CX3CR1+CCR7–

mesentericlymph node

Entrocyte Goblet cell Myofibroblast

RA

CD103+CCR7+CD86+OX40L+CCL17, CCL22, <IL12

CD103+CCR7+TGF-β, RA

homing to lamina propria

CD103+CCR7+ CX3CR1+CCR7–intestinallamina propria

IL-33

Uric acidIL-13IL-5

IL-10RA

TGFβ

TGF-β

TGF-βIL-4

CCR9+ CCR9+

IL-10

TSLP


Figure 1. Antigen uptake in the intestine and contribution of IEC in tolerance induction or allergic sensitization. Allergen exposed IEC are in close contact

with DC that sample luminal antigens. Allergens can enter the mucosa via the transcellular or paracellular route or be transferred via M-cells (not shown),

goblet cells or sampled by resident macrophages and carried to migratory CD103+ DC, which can also directly sample from the lumen. Allergen loaded

CD103+ DC traffic to the MLN in a CCR7 dependent manner where they instruct tolerance or allergic sensitization. If migratory DC are instructed by IEC

derived factors (including TGFb, TSLP, retinoic acid (RA) and mucin MUC2) to produce RA (via retinaldehyde dehydrogenase (RALDH)) and TGFb they

can generate CCR9+ regulatory T cells (Treg). The CX3CR1hi resident macrophages directly underlie the epithelium and under homeostatic conditions

produce high amounts of IL-10 to expand these gut homing Treg. On the other hand allergens and environmental triggers can induce IL-33, IL-25, TSLP and

uric acid release by IEC which can activate ILC2 and instruct TH2 driving CD86 and OX40L expressing migratory DC that secrete CCL17 and CCL22. In

particular IL-33 and uric acid contribute to allergic sensitization for food allergen in mice (peanut allergy model).

trans-epithelial electrical resistance (TEER) is an indication of

the enterocyte phenotype with well differentiated tight junc-

tions. When grown on microporous filter supports coated

with collagen cultures T84 cells maintain the polarity of

goblet-like cells.

M-cells have a reduced glycocalix, irregular brush border

with reduced microvilli and lack apical digestive enzymes.

They are highly specialized for the phagocytosis and trans-

cytosis of particulate antigens and pathogenic or commensal

microorganisms [55]. An in vitro model system composed of a

monolayer of Caco-2 cultivated with the human B-lympho-

ma cell line Raij has been widely used to study M cells [56].

Although these cells display efficient transcytosis activity, it is

uncertain whether they accurately represent the character-

istics of M-cells in vivo. They highly express CCL20, but lack

expression of mature M-cell marker genes, such as glycopro-

tein 2. A novel potentially physiologically relevant in vitro


M-cell-model system was reported in which RANKL (Receptor

Activator of Nuclear Factor-kB Ligand) stimulation induces

M-cell differentiation in gut organoid cultures established

from intestinal crypts or single LGR5+ (Leu-rich repeat-con-

taining G protein-coupled receptor 5-expressing) crypt stem

cells [57]. Besides exhibiting high transcytosis activity, the

range of genes expressed by these organoid cultures closely

resembles those of M-cells in vivo.

The studies on primary murine or human stem cell derived

intestinal epithelium are expanding. Embryonic stem cells

(ESCs) are grown under specific conditions to self-organize

into organoids or ‘mini guts’ [58]. They form three-dimen-

sional structures that incorporate many key features of the in

vivo intestinal epithelium, including a crypt-villus structure

that surrounds a functional central lumen. Intestinal orga-

noids incorporate all of the known cell types found in the

adult intestinal epithelium, and provide a physiologically


relevant model. Several methods have been used to grow

‘organoids’ from the small intestine [59], but the most suc-

cessful method is a Matrigel-based three-dimensional culture

system that supports the growth of self-renewing, near-native

intestinal epithelia in the absence of stromal niche compo-

nents [60].

The limit of the above in vitro intestinal models is that they

do not recapitulate the mechanically active microenviron-

ment of living intestine (peristaltic motions and intraluminal

fluid flow) and cannot be colonized by microbes over a

prolonged period [61,62]. Although not using primary cells

a human gut-on-a-chip microfluidic device enables Caco-2

cells to be cultured in the presence of physiologically relevant

luminal flow and peristalsis-like mechanical deformations,

which promotes formation of intestinal villi lined by all

epithelial cell types of the small intestine [63]. They could

be co-cultured with a probiotic gut microbe (Lactobacillus

rhamnosus GG) for more than two weeks.

Hence several cell lines can be used to study epithelial

function and gut-on-a-chip and primary epithelial cell cul-

tures using organoids are being developed. Polarized Caco-2

cells are successfully used when studying barrier crossing

properties of allergens via the paracellular or transcellular

route in vitro. Alternatively T84 cells can be used since they

also contain highly functional tight junction structures. In

addition, these cells are sensitive for environmental triggers

such as TH2 driving IL-4 and IL-13 and PAR ligands [21,22].

Beyond studying the barrier crossing capacities of (potential)

allergens, allergen induced epithelial activation may be in-

dicative for its allergenicity. This phenomenon has only

recently been revealed for airway sensitization and similar

mechanisms may underlie food protein sensitization when

occurring in the intestine [16,46,50]. Sensitive epithelial

models enabling to measure this for food proteins are cur-

rently lacking and need to be developed. When developing

these tools one should take into account that IEC are in close

contact with the underlying mucosal cells such as DC (see

Fig. 1) and effector immune cells which also may have impact

on the epithelial interaction with allergens and environmen-

tal factors. Co-culture models combining IEC with mixed

immune cells or DC may provide a better reflection of the

mucosal tissue organization and allow cross talk between

certain cell types in their reaction on allergens either or

not in presence of other environmental factors. 2D and 3D

co-culture models may be used to study these interactions.

Human 2D and 3D co-culture models of (intestinal) epithelial and

immune cells

In a recent study colonic biopsies of healthy adults mounted

in Ussing Chambers kept under high oxygen pressure were

used to determine HDM induced intestinal barrier disruption

and effects on IL-10 and TNF-a levels [24]. Hence it may be

possible to maintain human intestinal biopsies for prolonged

time. However the availability of fresh human intestinal

biopsies for research purposes is limited and requires ethical

approval. Co-culture models allowing cross-talk between

structural cells and immune cells are being developed. Trans-

well 2D co-cultures in which T84 cells were grown on inserts

and exposed to anti-CD2/CD28 activated lamina propria

mononuclear cells (LPMC) in the basolateral compartment

can be used to study the epithelial cell immune cell cross talk

and barrier dysfunction [64]. Based on this model a 2D co-

culture model using HT-29 and more easily accessible periph-

eral blood mononuclear cells (PBMC) instead of LPMC was

developed. In this model the epithelial cells modified the

cytokine secretion of underlying anti-CD3/CD28 activated

PBMC when exposed to TLR ligands [65,66]. Epithelial de-

rived galectin-9 (in HT-29 as well as T84) contributed to Treg

and TH1 polarization of PBMC and epithelial derived super-

natant instructed Treg and TH1 inducing monocyte derived

DC (moDC). Epithelial galectin-9 expression was confirmed

in the murine intestine and increased intestinal and systemic

galectin-9 levels in association with enhanced intestinal Treg

and TH1 markers and suppression of food allergy symptoms,

indicating the translational value of this 2D co-culture model

[67,68]. Although allergens were not studied, in the co-cul-

ture LPS exposed HT-29 released TSLP and CCL22 (MDC) was

increased [65]. Also Caco-2 may be able to produce TH2

polarizing mediators. In a 2D Caco-2/PBMC co-culture Pru

p3 transport and enhanced TSLP, IL-25 and IL-33 mRNA

expression was measured while IL-1b, IL-6, IL-10 and TNFa

mRNA in underlying PBMC was increased [28]. Hence, this

type of model may not only indicate whether a food allergen

induces epithelial activation, it may also determine the con-

sequence of this effect on the underlying immune cells. In

most cases human 2D co-cultures combine epithelial cells

with DC. Caco-2 cells are grown on filters and moDC are

seeded at the basolateral side and inflammatory mediator

release and DC activation and migration is studied [69].

Supernatants of epithelial cells from healthy donors or

Caco-2 enhanced CD103 expression on moDC or CD1c+

DC from human PBMC and instructed CD103+CCR7+ DC

from human MLN to induce Treg [70]. RA, TGFb and TSLP in

the supernatant of Caco-2 cells were responsible for the

induction of these Treg driving tolerogenic moDC [70].

When Caco-2 were cultured with moDC in the basolateral

compartment and apically exposed to bacteria, epithelial

derived TGFb suppressed pro-inflammatory cytokine produc-

tion by the moDC [71]. Caco-2 can also be grown inverted on

the basolateral side of the filter while moDC are added to the

apical compartment (contact model) [72]. In both models

MHCII, CD86 and CD80 expression on moDC was reduced in

the presence of IEC. However, only in the contact model also

TGFb concentrations increased while IL-8 decreased and

moDC were less responsive to LPS maturation [72]. Future

studies are warranted to determine whether this model would



be suitable to study intestinal epithelial cell/DC cross talk

upon allergen exposure. Such an approach is already being

developed using lung epithelial cells. For example, in a 2D

bronchial epithelial cell (16HBe140) (inverted)/moDC con-

tact model the effect of allergen exposure was studied.

CD80 and PD-L1 expression on moDC was increased and

the DC started to produce eotaxin and IL-10, which did not

occur when DC were cultured with the epithelial cell

supernatant. Furthermore, upon exposure to birch, grass

or HDM extracts the DC from the co-culture model had

reduced capacity to enhance autologous T-cell proliferation

and T cell cytokine release [73]. In another 2D airway co-

culture model BEAS-2B cells or primary bronchial epithelial

cells from allergic donors that were cultured inverted on

collagen coated transwell filters were basolaterally exposed

to Der p1 and moDC precursors were added to the apical

compartment. Der p1 increased the epithelial chemokine

release and enhanced moDC migration [47]. Hence, in

analogue to these 2D models studying the crosstalk be-

tween airway epithelial cells and DC upon aeroallergen

exposure, this could be studied for food proteins using

IEC. Beyond 2D also 3D co-cultures are being developed

which include connective tissue cells that produce immune

mediators as well as extracellular matrix components. In a

3D co-culture model T84 cells were grown on inserts on top

of primary human CCD-18Co intestinal myofibroblasts and

exposed to activated LPMC in the basolateral compart-

ment. These studies revealed myofibroblasts to protect

against inflammatory induced barrier disruption [64]. For

lung disease such types of models have been further devel-

oped and combine epithelial cells, DC and fibroblasts. In a

model in which human Calu-3 lung epithelial cells, moDC

and human MRC-5 lung fibroblasts are grown on separate

polyethylene terephthalate (PET) filters, papain induced

barrier disruption was less pronounced when the fibroblasts

were present. DC were found to migrate to the apical

epithelial compartment upon exposure to HDM or LPS

[74]. In an air exposed model in which MRC-5 cells, moDC

and 16HBE bronchial epithelial cells were grown directly

on top of each other on filters containing a collagen

matrix, CCL17 and CCL22 release by DC was silenced,

while CCL18 concentrations were high [75,76]. These 2D

and 3D cultures show that several cell types present in

mucosal tissues functionally interact and may impact on

whether or not an allergen, in absence or presence of

additional environmental triggers, can induce allergic sen-

sitization. Hence, future development of in vitro IEC models

that can identify the potential sensitizing capacity of aller-

gens or novel proteins may not only make use of epithelial

cells alone but also bring them in context with local tissue

cells such as fibroblasts known to affect epithelial function

and/or DC or mixed immune cells to reflect the impact on

the immune response.


Conclusion

IEC models to study intestinal allergen uptake are widely

used. Novel developments include the more physiological

‘gut-on-a-chip’ and stem cell derived primary organoids or

‘mini guts’ which in the future may be exploited for allergen

testing as well. In addition, epithelial models suitable to

measure TH2 driving mediators such as IL-33, IL-25 and TSLP

and relevant chemokines should be developed taking into

account not only the exposure of the allergens but also

environmental factors (such as inflammatory mediators, bac-

terial components or mycotoxins [77]) that can act as a

secondary trigger to activate the sensitization cascade. Fur-

thermore, taking into account the complexity of the mucosal

tissue, in vitro models to study the sensitizing potency of

allergens should also combine relevant mucosal cell types

since their interaction may affect the functional response of

IEC and therefore be more representative for the in vivo

setting.


LW is employed at the Utrecht University and collaborates

with Danone/Nutricia Research B.V. within a strategic alli-

ance between the Utrecht Institute for Pharmaceutical

Sciences of the Utrecht University and Danone/Nutricia

Research B.V, Utrecht, The Netherlands.

Acknowledgements

MGJ and LW are members of the Cost Action Working Group

FA1402, Improving allergy risk assessment strategy for new

food proteins (Imparas).

References[1] Haahtela T, Laatikainen T, Alenius H, Auvinen P, Fyhrquist N, Hanski I,

et al. Hunt for the origin of allergy – comparing the Finnish and Russian

Karelia. Clin Exp Allergy 2015;45:891–901.

[2] Prescott SL. Early-life environmental determinants of allergic diseases and

the wider pandemic of inflammatory noncommunicable diseases. J

Allergy Clin Immunol 2013;131:23–30.

[3] Thorburn AN, Macia L, Mackay CR. Diet, metabolites, and ‘western-

lifestyle’ inflammatory diseases. Immunity 2014;40:833–42.

[4] Platts-Mills TA. The allergy epidemics: 1870–2010. J Allergy Clin Immunol

2015;136:3–13.

[5] Pawankar R, Canonica GW, Holgate ST, Lockey RF. Allergic diseases and

asthma: a major global health concern. Curr Opin Allergy Clin Immunol

2012;12:39–41.

[6] Beasley R, Semprini A, Mitchell EA. Risk factors for asthma: is prevention

possible? Lancet 2015;386:1075–85.

[7] Mowat AM. Anatomical basis of tolerance and immunity to intestinal

antigens. Nat Rev Immunol 2003;3:331–41.

[8] Mowat AM, Agace WW. Regional specialization within the intestinal

immune system. Nat Rev Immunol 2014;14:667–85.

[9] Artis D. Epithelial-cell recognition of commensal bacteria and

maintenance of immune homeostasis in the gut. Nat Rev Immunol

2008;8:411–20.

[10] de Kivit S, Kraneveld AD, Garssen J. Glycan recognition at the interface of

the intestinal immune system: target for immune modulation via dietary

components. Eur J Pharmacol 2011;668(Suppl. 1):S124–32.

[11] Hammad H, Lambrecht BN. Dendritic cells and epithelial cells: linking innate

and adaptive immunity in asthma. Nat Rev Immunol 2008;8:193–204.





























[12] Menard S, Cerf-Bensussan N, Heyman M. Multiple facets of intestinal

permeability and epithelial handling of dietary antigens. Mucosal

Immunol 2010;3:247–59.

[13] Gandhi VD, Vliagoftis H. Airway epithelium interactions with

aeroallergens: role of secreted cytokines and chemokines in innate

immunity. Front Immunol 2015;6:147.

[14] Hewitt CR, Brown AP, Hart BJ, Pritchard DI. A major house dust

mite allergen disrupts the immunoglobulin E network by selectively

cleaving CD23: innate protection by antiproteases. J Exp Med

1995;182:1537–44.

[15] Schulz O, Sewell HF, Shakib F. Proteolytic cleavage of CD25, the alpha

subunit of the human T cell interleukin 2 receptor, by Der p 1, a major

mite allergen with cysteine protease activity. J Exp Med 1998;187:271–5.

[16] Lambrecht BN, Hammad H. Allergens and the airway epithelium

response: gateway to allergic sensitization. J Allergy Clin Immunol

2014;134:499–507.

[17] Roschmann K, Farhat K, Konig P, Suck R, Ulmer AJ, Petersen A. Timothy

grass pollen major allergen Phl p 1 activates respiratory epithelial cells by a

non-protease mechanism. Clin Exp Allergy 2009;39:1358–69.

[18] Roschmann KI, van Kuijen AM, Luiten S, Jonker MJ, Breit TM, Fokkens WJ,

et al. Purified Timothy grass pollen major allergen Phl p 1 may contribute

to the modulation of allergic responses through a pleiotropic induction of

cytokines and chemokines from airway epithelial cells. Clin Exp Immunol

2012;167:413–21.

[19] Mochizuki T, Satsu H, Totsuka M, Shimizu M. Transepithelial transport of

macromolecular substances in IL-4 treated human intestinal T84 cell

monolayers. Biosci Biotechnol Biochem 2009;73:2422–6.

[20] Heller F, Florian P, Bojarski C, Richter J, Christ M, Hillenbrand B, et al.

Interleukin-13 is the key effector Th2 cytokine in ulcerative colitis that

affects epithelial tight junctions, apoptosis, and cell restitution.

Gastroenterology 2005;129:550–64.

[21] Jacob C, Yang PC, Darmoul D, Amadesi S, Saito T, Cottrell GS, et al. Mast

cell tryptase controls paracellular permeability of the intestine. Role of

protease-activated receptor 2 and beta-arrestins. J Biol Chem

2005;280:31936–48.

[22] Ceponis PJ, Botelho F, Richards CD, McKay DM. Interleukins 4 and 13

increase intestinal epithelial permeability by a phosphatidylinositol 3-

kinase pathway. Lack of evidence for STAT 6 involvement. J Biol Chem

2000;275:29132–37.

[23] Tu Y, Salim S, Bourgeois J, Di Leo V, Irvine EJ, Marshall JK, et al. CD23-

mediated IgE transport across human intestinal epithelium: inhibition by

blocking sites of translation or binding. Gastroenterology 2005;129:928–40.

[24] Tulic MK, Vivinus-Nebot M, Rekima A, Rabelo Medeiros S, Bonnart C, Shi

H, et al. Presence of commensal house dust mite allergen in human

gastrointestinal tract: a potential contributor to intestinal barrier

dysfunction. Gut 2016;65:757–66.

[25] Grozdanovic MM, Cavic M, Nesic A, Andjelkovic U, Akbari P, Smit JJ,

et al. Kiwifruit cysteine protease actinidin compromises the intestinal

barrier by disrupting tight junctions. Biochim Biophys Acta

2016;1860:516–26.

[26] Cavic M, Grozdanovic MM, Bajic A, Jankovic R, Andjus PR, Gavrovic-

Jankulovic M. The effect of kiwifruit (Actinidia deliciosa) cysteine protease

actinidin on the occludin tight junction network in T84 intestinal

epithelial cells. Food Chem Toxicol 2014;72:61–8.

[27] Lorkowski G. Gastrointestinal absorption and biological activities of

serine and cysteine proteases of animal and plant origin: review on

absorption of serine and cysteine proteases. Int J Physiol Pathophysiol

Pharmacol 2012;4:10–27.

[28] Tordesillas L, Gomez-Casado C, Garrido-Arandia M, Murua-Garcia A,

Palacin A, Varela J, et al. Transport of Pru p 3 across gastrointestinal

epithelium – an essential step towards the induction of food allergy? Clin

Exp Allergy 2013;43:1374–83.

[29] Caillard I, Tome D. Transport of beta-lactoglobulin and alpha-

lactalbumin in enterocyte-like Caco-2 cells. Reprod Nutr Dev

1995;35:179–88.

[30] Price DB, Ackland ML, Burks W, Knight MI, Suphioglu C. Peanut allergens

alter intestinal barrier permeability and tight junction localisation in

Caco-2 cell cultures. Cell Physiol Biochem 2014;33:1758–77.

[31] Starkl P, Krishnamurthy D, Szalai K, Felix F, Lukschal A, Oberthuer D, et al.

Heating affects structure, enterocyte adsorption and signalling, as well as

immunogenicity of the peanut allergen Ara h 2. Open Allergy J 2011;4:24–34.

[32] Bodinier M, Legoux MA, Pineau F, Triballeau S, Segain JP, Brossard C, et al.

Intestinal translocation capabilities of wheat allergens using the Caco-2

cell line. J Agric Food Chem 2007;55:4576–83.

[33] Hirose J, Ito S, Hirata N, Kido S, Kitabatake N, Narita H. Occurrence of the

major food allergen, ovomucoid, in human breast milk as an immune

complex. Biosci Biotechnol Biochem 2001;65:1438–40.

[34] Husby S, Jensenius JC, Svehag SE. Passage of undegraded dietary antigen

into the blood of healthy adults. Further characterization of the kinetics of

uptake and the size distribution of the antigen. Scand J Immunol

1986;24:447–55.

[35] Moreno FJ, Rubio LA, Olano A, Clemente A. Uptake of 2S albumin

allergens, Ber e 1 and Ses i 1, across human intestinal epithelial Caco-2 cell

monolayers. J Agric Food Chem 2006;54:8631–9.

[36] Laffont S, Siddiqui KR, Powrie F. Intestinal inflammation abrogates the

tolerogenic properties of MLN CD103+ dendritic cells. Eur J Immunol

2010;40:1877–83.

[37] Scott CL, Aumeunier AM, Mowat AM. Intestinal CD103+ dendritic cells:

master regulators of tolerance? Trends Immunol 2011;32:412–9.

[38] McDole JR, Wheeler LW, McDonald KG, Wang B, Konjufca V, Knoop KA,

et al. Goblet cells deliver luminal antigen to CD103+ dendritic cells in the

small intestine. Nature 2012;483:345–9.

[39] Worbs T, Bode U, Yan S, Hoffmann MW, Hintzen G, Bernhardt G, et al.

Oral tolerance originates in the intestinal immune system and relies on

antigen carriage by dendritic cells. J Exp Med 2006;203:519–27.

[40] Shan M, Gentile M, Yeiser JR, Walland AC, Bornstein VU, Chen K, et al.

Mucus enhances gut homeostasis and oral tolerance by delivering

immunoregulatory signals. Science 2013;342:447–53.

[41] Cerovic V, Bain CC, Mowat AM, Milling SW. Intestinal macrophages and

dendritic cells: what’s the difference? Trends Immunol 2014;35:270–7.

[42] Iliev ID, Spadoni I, Mileti E, Matteoli G, Sonzogni A, Sampietro GM, et al.

Human intestinal epithelial cells promote the differentiation of

tolerogenic dendritic cells. Gut 2009;58:1481–9.

[43] Coombes JL, Siddiqui KR, Arancibia-Carcamo CV, Hall J, Sun CM, Belkaid

Y, et al. A functionally specialized population of mucosal CD103+ DCs

induces Foxp3+ regulatory T cells via a TGF-beta and retinoic acid-

dependent mechanism. J Exp Med 2007;204:1757–64.

[44] McDonald KG, Leach MR, Brooke KW, Wang C, Wheeler LW, Hanly EK,

et al. Epithelial expression of the cytosolic retinoid chaperone cellular

retinol binding protein II is essential for in vivo imprinting of local gut

dendritic cells by lumenal retinoids. Am J Pathol 2012;180:984–97.

[45] Halim TY, Steer CA, Matha L, Gold MJ, Martinez-Gonzalez I, McNagny

KM, et al. Group 2 innate lymphoid cells are critical for the initiation of

adaptive T helper 2 cell-mediated allergic lung inflammation. Immunity

2014;40:425–35.

[46] Chu DK, Llop-Guevara A, Walker TD, Flader K, Goncharova S, Boudreau

JE, et al. IL-33, but not thymic stromal lymphopoietin or IL-25, is central

to mite and peanut allergic sensitization. J Allergy Clin Immunol

2013;131. 187–200 e1–8.

[47] Pichavant M, Charbonnier AS, Taront S, Brichet A, Wallaert B, Pestel J,

et al. Asthmatic bronchial epithelium activated by the proteolytic allergen

Der p 1 increases selective dendritic cell recruitment. J Allergy Clin

Immunol 2005;115:771–8.

[48] Kong J, Chalcraft K, Mandur TS, Jimenez-Saiz R, Walker TD, Goncharova S,

et al. Comprehensive metabolomics identifies the alarmin uric acid as a

critical signal for the induction of peanut allergy. Allergy 2015;70:495–505.

[49] Chu DK, Jimenez-Saiz R, Verschoor CP, Walker TD, Goncharova S, Llop-

Guevara A, et al. Indigenous enteric eosinophils control DCs to initiate a

primary Th2 immune response in vivo. J Exp Med 2014;211:1657–72.

[50] Chu DK, Mohammed-Ali Z, Jimenez-Saiz R, Walker TD, Goncharova S,

Llop-Guevara A, et al. T helper cell IL-4 drives intestinal Th2 priming to

oral peanut antigen, under the control of OX40L and independent of

innate-like lymphocytes. Mucosal Immunol 2014;7:1395–404.

[51] Huet G, Kim I, de Bolos C, Lo-Guidice JM, Moreau O, Hemon B, et al.

Characterization of mucins and proteoglycans synthesized by a mucin-

secreting HT-29 cell subpopulation. J Cell Sci 1995;108(Pt 3):1275–85.











































































































































[52] Gagnon M, Zihler Berner A, Chervet N, Chassard C, Lacroix C.

Comparison of the Caco-2, HT-29 and the mucus-secreting HT29-MTX

intestinal cell models to investigate Salmonella adhesion and invasion. J

Microbiol Methods 2013;94:274–9.

[53] McCool DJ, Marcon MA, Forstner JF, Forstner GG. The T84 human colonic

adenocarcinoma cell line produces mucin in culture and releases it in

response to various secretagogues. Biochem J 1990;267:491–500.

[54] Willemsen LE, Koetsier MA, van Deventer SJ, van Tol EA. Short chain fatty

acids stimulate epithelial mucin 2 expression through differential effects

on prostaglandin E(1) and E(2) production by intestinal myofibroblasts.

Gut 2003;52:1442–7.

[55] Mabbott NA, Donaldson DS, Ohno H, Williams IR, Mahajan A. Microfold

(M) cells: important immunosurveillance posts in the intestinal

epithelium. Mucosal Immunol 2013;6:666–77.

[56] Kerneis S, Bogdanova A, Kraehenbuhl JP, Pringault E. Conversion by

Peyer’s patch lymphocytes of human enterocytes into M cells that

transport bacteria. Science 1997;277:949–52.

[57] de Lau W, Kujala P, Schneeberger K, Middendorp S, Li VS, Barker N, et al.

Peyer’s patch M cells derived from Lgr5(+) stem cells require SpiB and are

induced by RankL in cultured ‘miniguts’. Mol Cell Biol 2012;32:3639–47.

[58] Huch M, Koo BK. Modeling mouse and human development using

organoid cultures. Development 2015;142:3113–25.

[59] Sala FG, Matthews JA, Speer AL, Torashima Y, Barthel ER, Grikscheit TC. A

multicellular approach forms a significant amount of tissue-engineered

small intestine in the mouse. Tissue Eng Part A 2011;17:1841–50.

[60] Sato T, Vries RG, Snippert HJ, van de Wetering M, Barker N, Stange DE,

et al. Single Lgr5 stem cells build crypt-villus structures in vitro without a

mesenchymal niche. Nature 2009;459:262–5.

[61] Sanderson IR. The physicochemical environment of the neonatal

intestine. Am J Clin Nutr 1999;69:1028S–34S.

[62] Kim HJ, Huh D, Hamilton G, Ingber DE. Human gut-on-a-chip inhabited

by microbial flora that experiences intestinal peristalsis-like motions and

flow. Lab Chip 2012;12:2165–74.

[63] Kim HJ, Ingber DE. Gut-on-a-Chip microenvironment induces human

intestinal cells to undergo villus differentiation. Integr Biol (Camb)

2013;5:1130–40.

[64] Willemsen LE, Schreurs CC, Kroes H, Spillenaar Bilgen EJ, Van Deventer

SJ, Van Tol EA. A coculture model mimicking the intestinal mucosa

reveals a regulatory role for myofibroblasts in immune-mediated barrier

disruption. Dig Dis Sci 2002;47:2316–24.

[65] de Kivit S, van Hoffen E, Korthagen N, Garssen J, Willemsen LE. Apical TLR

ligation of intestinal epithelial cells drives a Th1-polarized regulatory or

inflammatory type effector response in vitro. Immunobiology

2011;216:518–27.


[66] de Kivit S, Tobin MC, DeMeo MT, Fox S, Garssen J, Forsyth CB, et al. In

vitro evaluation of intestinal epithelial TLR activation in preventing food

allergic responses. Clin Immunol 2014;154:91–9.

[67] de Kivit S, Kraneveld AD, Knippels LM, van Kooyk Y, Garssen J, Willemsen

LE. Intestinal epithelium-derived galectin-9 is involved in the

immunomodulating effects of nondigestible oligosaccharides. J Innate

Immun 2013;5:625–38.

[68] de Kivit S, Saeland E, Kraneveld AD, van de Kant HJ, Schouten B, van Esch

BC, et al. Galectin-9 induced by dietary synbiotics is involved in

suppression of allergic symptoms in mice and humans. Allergy

2012;67:343–52.

[69] Rimoldi M, Chieppa M, Vulcano M, Allavena P, Rescigno M. Intestinal

epithelial cells control dendritic cell function. Ann N Y Acad Sci

2004;1029:66–74.

[70] Iliev ID, Mileti E, Matteoli G, Chieppa M, Rescigno M. Intestinal epithelial

cells promote colitis-protective regulatory T-cell differentiation through

dendritic cell conditioning. Mucosal Immunol 2009;2:340–50.

[71] Zeuthen LH, Fink LN, Frokiaer H. Epithelial cells prime the immune

response to an array of gut-derived commensals towards a

tolerogenic phenotype through distinct actions of thymic stromal

lymphopoietin and transforming growth factor-beta. Immunology

2008;123:197–208.

[72] Butler M, Ng CY, van Heel DA, Lombardi G, Lechler R, Playford RJ, et al.

Modulation of dendritic cell phenotype and function in an in vitro model

of the intestinal epithelium. Eur J Immunol 2006;36:864–74.

[73] Papazian D, Wagtmann VR, Hansen S, Wurtzen PA. Direct contact

between dendritic cells and bronchial epithelial cells inhibits T cell recall

responses towards mite and pollen allergen extracts in vitro. Clin Exp

Immunol 2015;181:207–18.

[74] Harrington H, Cato P, Salazar F, Wilkinson M, Knox A, Haycock JW,

et al. Immunocompetent 3D model of human upper airway for

disease modeling and in vitro drug evaluation. Mol Pharm

2014;11:2082–91.

[75] Nguyen Hoang AT, Chen P, Bjornfot S, Hogstrand K, Lock JG, Grandien A,

et al. Technical advance: live-imaging analysis of human dendritic cell

migrating behavior under the influence of immune-stimulating reagents

in an organotypic model of lung. J Leukoc Biol 2014;96:481–9.

[76] Nguyen Hoang AT, Chen P, Juarez J, Sachamitr P, Billing B, Bosnjak L,

et al. Dendritic cell functional properties in a three-dimensional tissue

model of human lung mucosa. Am J Physiol Lung Cell Mol Physiol

2012;302:L226–37.

[77] Bol-Schoenmakers M, Braber S, Akbari P, de Graaff P, van Roest M,

Kruijssen L, et al. The mycotoxin deoxynivalenol facilitates allergic

sensitization to whey in mice. Mucosal Immunol 2016.

























































































DRUG DISCOVERY

TODAY

DISEASEMODELS

IgE – the main player of food allergyHenrike C.H. Broekman1, Thomas Eiwegger2,3, Julia Upton2,

Katrine L. Bøgh4,*1Department of Dermatology/Allergology, University Medical Centre Utrecht (UMCU), Utrecht, The Netherlands2Division of Immunology and Allergy, Food Allergy and Anaphylaxis Program, The Department of Paediatrics, Hospital for Sick Children,

Toronto, Canada3Research Institute, Physiology and Experimental Medicine, The University of Toronto, Toronto, Canada4National Food Institute, Technical University of Denmark, Søborg, Denmark


Editors-in-Chief




Food allergy is a growing problem worldwide, presently

affecting 2–4% of adults and 5–8% of young children. IgE

is a key player in food allergy. Consequently huge

efforts have been made to develop tests to detect

either the presence of IgE molecules, their allergen

binding sites or their functionality, in order to provide

information regarding the patient’s food allergy. The

ultimate goal is to develop tools that are capable of

discriminating between asymptomatic sensitization and

a clinically relevant food allergy, and between different

allergic phenotypes in an accurate and trustworthy

manner. This may generate better diagnostic, prognos-

tic and therapeutic monitoring tools for the future.

Introduction

Immunoglobulin E (IgE)-mediated food allergy is an immu-

nologic, non-toxic adverse reaction to otherwise harmless

antigens in food. The mechanisms underlying IgE-mediated

food allergy consist of a sensitization and an elicitation phase

(Fig. 1). Sensitization may occur upon the first contact with

the food allergen, and results in generation of allergen-spe-

cific IgE (sIgE). Elicitation of symptoms occurs upon subse-

quent contact with the respective allergen leading to

symptoms. Symptoms occur within minutes to hours of

allergen ingestion [1], and involve one or more of the follow-

ing systems; the skin (pruritus, urticaria, or angioedema), the

*Corresponding author at: National Food Institute, Technical University of Denmark, MørkhøjBygade 19, DK-2860 Søborg, Denmark.: K.L. Bøgh ([email protected])



gastro-intestinal tract (diarrhea, vomiting, contractions, in-

creased bowel movement), the respiratory tract (asthma at-

tack, hoarseness, stridor/laryngeal angioedema) or the

cardiovascular system (dizziness, drop in blood pressure, loss

of consciousness) [2,3].

Food allergy appears to be a rising problem worldwide, and

currently affects 2–4% of adults and 5–8% of young children

[4,5].

Although there is some evidence that the first year of life is

decisive to develop allergies or asthma later on, the time

point an allergic sensitization occurs is very individual. De-

spite of crude patterns of sensitization (food allergy in early

childhood vs sensitization to inhalant allergens later on),

sensitization may already occur in utero or at any time point

after birth [6]. Most likely a combination of genetic predis-

position, pro-allergenic, environmental factors and allergen

exposure is required to induce sensitization and overcome

natural mechanisms of tolerance induction to innocuous

environmental antigens. Mechanisms that are responsible

for tolerance maintenance in sensitized individuals and for

the re-induction of tolerance in allergic individuals are not

well understood. Proposed mechanisms include regulatory

6.07.001 37







Allergen

Allergen specific epitope

Dendritic cell

Naïve T cellMHC calss II

MHC class II

TCR

TCR

B cell

Plasmacell

BasophilSpecifi IgE

CD300a

CD63

CD203c



Specific IgE

CD300a

CD63 expression on membrane

BAT & Histaminerelease test mimicElicitation

Elicitation

Allergic symptoms

Allergen

Allergen

Mastcell Specific IgESpecific IgE

CD203cFCεRI

FCεRIFCεRI

CD40LTh 2 cell

CD40

IL-4&IL-13

Humoral cell free testmeasuresensitization

Sensitization


Figure 1. The role of IgE in the context of IgE-mediated allergy. The pathogenesis of IgE-mediated food allergy is divided into two phases; a sensitization

phase and an elicitation phase. In the sensitization phase the allergen is taken up by antigen-presenting cells, leading to activation of Th2 cells, which again

contributes to the activation of B-cells that differentiate and proliferate into IgE secreting plasma cells. This forms the basis for the cell-free IgE-based in vitro

test methods as well as the IgE epitope mapping based tests, based on IgE from the food allergic patients. In the elicitation phase allergens may, upon

reexposure, cross-link FceRI bound IgE on mast cells and basophils leading to mediator release and the symptoms characteristic of food allergy. This forms

the basis for the cell-based in vitro test methods, based on the functionality of the sIgE repertoire rather than just the presence of sIgE.

T-cells, blocking antibodies, tolerogenic dentritic cell popu-

lations, lack of epitope diversity and clonal deletion due to

constant exposure [6].

The gold-standard for food allergy diagnosis is the oral

food challenge (OFC), but it is expensive, time-consuming

and carries a risk of severe reactions [4,5]. Hence, there is

great interest in developing diagnostic in vitro methods.

After the discovery of IgE, allergen-sIgE-based tests were

developed for diagnosis and have resulted in the standard

we use today. Despite of good clinical applicability, limita-

tions of these tests have led to considerable efforts in

investigating the role and clinical value of IgE binding to

specific allergens as well as IgE binding to specific sites on

the allergen. Detecting sIgE binding patterns could be a

promising approach to predict food allergy and the associ-

ated clinical manifestations [7]. This review discusses the

applicability and value of IgE, its binding specificity and

functionality in the context of food allergy, in order to

predict patient’s individual clinical history and to assess

treatment efficacy.

IgE based approaches

Immunoglobulins, also designated antibodies, are produced

by B cells and consist of two heavy and two light chains. The

Fc-region (consisting of the heavy chains) of IgE binds

through the high affinity Fc-receptor (FceRI) to other cells


of the immune system, while the Fab region (part heavy and

variable light chains) binds to the antigen [8,9]. The binding

site of the Fab region (the paratope) binds to a specific part of

the antigen, in case of allergy an allergen, which is called the

epitope. When an allergen cross-links two FceRI-bound IgE

antibodies on either mast cells or basophils, these effector

cells degranulate and release mediators such as histamine,

prostaglandins, and leukotrienes, causing the allergic symp-

toms of food allergy [1].

Various IgE-based tests have been developed in order to

provide information about food allergy. These methods can

either be cell-free or cell-based (Table 1).

Cell-free IgE-based in vitro test methods

Total IgE

Total IgE can be measured by multiple methods and is

measured in international units (IU)/mL. Competitive dis-

placement radioimmunoassay (RIA), two-sided immunora-

diometric assays (IRMA), two-sided enzyme immunoassay

(EIA), and kinetic nephelometry are the currently favoured

methods [10].

The clinical applicability of total IgE is limited. IgE is not

necessarily specific to food allergens and can be elevated in

other atopic diseases, infections and primary immunodefi-

ciencies. Additionally, a low total IgE does not exclude a food


Table 1. Pros and cons of methods used for detection of sIgE and its functionality.

Humoral and cell based IgE test Pros Cons References

Total IgE Easy method Limited clinical applicability [10,11]

Specific IgE

RAST/FEIA

Performed in both commercial and

research laboratories

Relatively quick assay

Levels positively correlate with

likelihood of clinical allergy for

many foods

Need for clinical validation

False negative and false positive results may occur

Results not interchangeable with other sIgE tests

Recent approved CAP assay has better quality

[12,13]

Specific IgE

ELISA

Performed in both commercial and

research laboratories

Need for clinical validation False negative and false

positive results may occur

[12,13]

Results not interchangeable with other sIgE tests

Specific IgE

Immunoblot

Allows for identification of both

linear and conformational peptides

Specific protein recognition

Laborious method [12,13]

Components ISAC microarray Easy method

Large data output

Able to distinguish a clinical relevant

allergy and severity for some foods

High cost

Not sensitive for all foods

Not available or applicable for all foods

[19,22–24]

Basophil Activation Test Highly specific and sensitive for

several foods

Laborious method

False negative results may occur

No established extracts

Not enough clinical data available

Not suitable for screening approaches

[36–40]

Humanized RBL Easily standardized Laborious method

Need for validation

No established extracts

Low stability

[42,43]

Histamine release assay Mimics mast cell activation at a

larger scale

Laborious method

High cost

[46,47]

allergy. An expert panel has advised against using total IgE in

diagnosing food allergy [11].

Specific IgE

Allergen specific IgE (sIgE) can be measured by multiple

methods and is measured in units of allergen (UA)/mL.

Enzyme-linked immunosorbent assay (ELISA), enzyme aller-

gosorbent test (EAST), fluorescence enzyme immunoassay

(FEIA), radioallergosorbent test (RAST) and immunoblotting

are methods currently applied for measurement of sIgE [12].

Measurement of sIgE typically involves using allergens bound

to a solid phase to capture IgE and is quantified by the use of

labeled anti-IgE antibodies. These tests are performed both by

commercial and research laboratories as well as in many

hospital settings. The sIgE levels obtained for a particular

protein by different commercial tests are not inter-change-

able. There are no international standards for specific IgE

assays but rather they are calibrated with the WHO reference

preparation for total serum IgE [10].

sIgE levels usually positively correlate with the likelihood

of having a clinically relevant food allergy – the higher the

sIgE to a given food, the higher the likelihood of clinical

reactions upon ingestion. However, the ability to rule out

allergy (sensitivity, percentage of allergic individuals with a

negative test) and to diagnose allergy (specificity, percentage

of individuals with positive test that are allergic) is limited

and there is significant variability across populations [13].

sIgE to the respective food may be observed in subjects

without a clinical relevant food allergy and it may not be

detected in those with a confirmed food allergy [14,15]. This

is illustrated by a population-based birth cohort study from

the UK where 12 percent of children were sensitized to

peanut, but only 2 percent were peanut allergic [16]. Both,

the indication to perform sIgE measurement and the assess-

ment of the clinical relevance of a given sIgE value require

individual assessment by a clinician. Neither the test nor the

interpretation should be done without knowledge of the

patient’s history.

The sIgE/total IgE ratio has been examined regarding addi-

tional diagnostic utility with mixed results. It has been

reported that it did not contribute to the diagnosis beyond

the sIgE [17]. However, a recent study suggested superiority of



the sIgE/total IgE ratio as compared to sIgE alone to appro-

priately assign patients to a food challenge [18]. Overall the

clinical applicability of this ratio requires more data before

being integrated into clinical decision making.

Component resolved diagnosis

As clinically available test methods utilizing sIgE binding in

vitro are generally not as efficient as a food challenge in

diagnosing food allergy, attempts to further improve the

diagnostic accuracy of sIgE testing have been made by intro-

ducing the terminus component resolved diagnostics (CRD).

It defines reactivity to individual allergens and not to an

allergen extract. The binding patterns to homologous aller-

gens from different species may be explained by cross reac-

tivity amongst proteins within the same protein family [19].

For some food allergies the usage of component resolved

diagnosis has been proven to increase the ability to predict

clinical reactivity [20,21]. This is reflected in superior receiver

operator curves integrating sensitivity (proportion of allergic

patients correctly identified) and specificity (proportion of

non-allergic individuals identified) as compared to extract

testing. Currently there is good evidence for the usage of

CRD in diagnosing peanut and hazelnut allergy. Further, de-

tection of IgE against specific allergens within these foods, such

as the peanut allergen Ara h 2 and the hazelnut allergens Cor a

8, Cor a 9 and Cor a 14, have been shown to predict a clinically

relevant food allergy, as well as to help in distinguishing

between cross-reactive and ‘true’ sensitizing allergens [22–24].

CRD has improved our knowledge on the sensitization

patterns of some of the more prevalent allergen sources,

including peanut, tree nut, egg and milk, but also some less

prevalent allergies such as wheat dependent exercise induced

allergy and soy allergy [25,26]. In addition to its diagnostic

values, CRD may confer therapeutic importance for the de-

velopment of allergen-specific immunotherapy, as it may

enable us to use only the clinically relevant allergens [27].

IgE avidity/affinity

Affinity (the attractive force between substances or particles

that causes them to enter into and remaining a chemical

combination) of an antibody for its antigen has been shown

to be an important determinant of the biological efficacy of

the antibody [28]. Measuring the affinity of a single clone of

IgE antibodies or the avidity (the additive strength of multi-

ple affinities of non-covalent binding interactions) of a poly-

clonal IgE antibody response in serum is difficult because of

the low serum concentrations of IgE (�150 ng/mL [29]), and

sIgE levels are only a fraction of the total serum IgE. In

contrast to vaccine research, affinity and avidity measure-

ments to allergens are not commonly used. Nevertheless, El-

Khouly et al. [30] showed in a study investigating the anti-

body avidity characteristics of peanut allergic patients that

the peanut allergen Ara h 2-specific avidity correlated with


the severity as measured by a food challenge score. Shortly

afterwards, Wang et al. [31] reported that IgE affinity corre-

lated with severity of milk allergy. Recently, Surface Plasmon

Resonance imaging, has led to satisfactory measures of the

affinity of human IgE antibodies [32]. Despite only being

scarcely described affinity/avidity measures could be a prom-

ising future tool providing information on the food allergic

disease.

In vitro functional assays

Various cell-based methods for an indirect analysis of the

performance of sIgE have been developed using surrogate

biomarkers of effector cell activation such as surface markers

or released mediators [33]. Mediators which have been in-

vestigated include histamine, heparin, tryptase, chymase,

carboxypeptidase A3, prostaglandin D2 and cysteinyl leuko-

trienes. However, none of these biomarkers have yet proven

to be of more value than existing allergy tests [33].

Basophil activation test

Human basophils can be stimulated with allergens in vitro and

the ability to activate them can be linked to food allergy. In

the basophil activation test (BAT), activation of basophils via

allergens is reflected in an up-regulation of the cell-surface

molecules CD63 or CD203c [34]. BATs have been used in the

diagnosis, management and as a tool to decide the perfor-

mance of OFC in milk, egg and peanut allergy, and also in the

diagnosis of pollen food syndromes, as reviewed elsewhere

[35]. BAT has in some instances shown higher specificity and

negative predictive value than sIgE measurement, without

losing sensitivity or positive predictive value [36]. In particu-

lar, in young children with peanut allergy the BAT proved to

be superior to other diagnostic tests in discriminating be-

tween peanut allergy and tolerance and the results are en-

couraging that BAT may significantly reduce the need for

OFCs in the future [37]. In the context of ascertaining degrees

of baked milk product tolerance the BAT results reached a

statistically significant trend [38]. For discriminating between

peanut tolerance and reactivity in adult peanut sensitized

individuals [39] the BAT showed some utility. Recently,

passive sensitization of basophils with sera from allergic

donors after stripping of membrane bound IgE has provided

promising results in peanut allergic individuals which await

confirmation [40].

Humanized RBL assay

Humanized rat basophilic leukemia (RBL) cell-lines trans-

fected with human FceRI have been developed for the use

in functional allergen–IgE interaction research [41]. Human-

ized RBL cells can be cultured permanently, providing im-

proved standardization. However, this test has not found

widespread acceptance among clinicians [42], likely because

of the overall low stability of the humanized RBL assay due to


loss of the humanized receptor. These assays have a lower

degree of sensitivity as compared to human basophil tests.

The most recent degranulation assay developed is based on

the huFceRI-RBL-2H3 cells, which was tested for sensitivity

and specificity for food allergens [43]. Nevertheless, the non-

humanized version of the RBL assay, has shown to be efficient

for studying the IgE functionality [44,45].

Histamine release assay

Histamine release in vitro by stem cell derived mast cells

loaded with serum derived IgE may possess the ability to test

food-induced mast cell activation at a larger scale due to

recent improvements [46]. However, currently this test is

not applicable in a real life diagnostic setting because of its

costs and the complexity of the method [47].

IgE epitopes based approaches

IgE binding epitopes, defined as the allergenic regions of the

allergen recognized by IgE molecules, are generally catego-

rized as either linear or conformational based on the vicinity

of the amino acids in the primary structure being involved in

the IgE binding [48]. Whereas the linear epitope consists of a

contiguous stretch of amino acids juxtaposed in the primary

structure, the conformational epitope consists of amino acids

distant from one another in the primary structure but

brought together by the structural folding of the protein

[49–51].

There is no clear boundary at the amino acid level for those

amino acids which comprise the epitope [50,52]. Antibody

binding epitopes have been suggested to consist of approxi-

mately 15 amino acids [53], but there is no evidence that each

amino acid in the epitope necessarily interacts with the

antibody, and energy calculations have indicated that as

few as five to six amino acids are the actual contributors in

the binding between epitope and antibody [50,54,55].

Various methods can be applied for identification of IgE

binding epitopes, however, for experimental reasons some

approaches only allow for identification of the linear type

(Table 2).

Methods for identification of linear epitopes

Several IgE epitope mapping methods are based on binding of

IgE molecules to peptides derived from the primary structure

of the allergen [56,57], thereby allowing for the identification

of only linear epitopes. The epitope mapping technology of

such peptide arrays, by means of immobilized peptides on a

surface, have been subjected to rapid and substantial devel-

opment over the last decades [58,50]. Typically, overlapping

peptides of 10–20 amino acid residues are synthetized in

parallel, for example, on a glass slide or a nitrocellulose

membrane [60]. Just a few years ago standard peptide synthe-

sis could only synthetize a few hundred peptides, but with the

more recent improvements in the field synthesis of up to

2 100 000 peptides in parallel is now a possibility [61]. These

advances in peptide arrays have recently allowed for the

identification of epitopes on the amino acid level [62]. By

substituting each amino acid in the synthetic peptides with

an alanine (alanine scan) Hansen et al. [62] were able to

identify the amino acids within an epitope contributing to

the binding to IgE of peanut allergic patients.

Methods for identification of conformational epitopes

Identification of conformational IgE binding epitopes

requires more sophisticated techniques, such as X-ray crys-

tallography, nuclear magnetic resonance (NMR), site-direct-

ed mutagenesis or phage display technology [50,56,60]. The

only complete method for identification of an IgE binding

epitope is co-crystallization of an allergen:IgE antibody com-

plex by X-ray crystallography [56,63,64], and thus this tech-

nique is considered the gold-standard. However, X-ray

crystallography is a very laborious procedure that only allows

for the identification of a monoclonal response, and conse-

quently has a very low output. Another sophisticated tech-

nique is based on NMR that allows for a dynamic picture of

the allergen:IgE antibody complex [63,65]. IgE epitope map-

ping by site-directed mutagenesis is based on systematic

introduction of residue substitutions along the allergen,

and a subsequent determination of the effect of each muta-

tion on allergen recognition by IgE [56]. However, like X-ray

crystallography, NMR and site-directed mutagenesis techni-

ques only allow for identification of a monoclonal response

[63,66]. Another approach is phage display technology which

is based on the screening of a random peptide library, for

affinity selection of peptides mimicking structures of epitopes

bound by specific IgE antibodies, followed by competitive

immune-screening with the specific allergen for elution of

peptides of interest [67,68]. In order to predict the location of

IgE binding epitopes on allergens in a structural context

different in silico based methods are available [69–71]. In

contrast to other approaches allowing for identification of

conformational epitopes, this technique allows for the iden-

tification of IgE binding epitopes of a polyclonal antibody

response as well as for patient-specific identification of amino

acids contributing to the IgE binding [67,68]. Recent

advances in coupling the phage display technique with

high-throughput sequencing, has allowed a tremendous in-

crease in the data output [68]. However, a massive challenge

with the phage display technology is the notorious selection

of unspecific allergen unrelated peptides, which necessitate

the use of control subjects [68].

Clinical applicability of IgE and IgE binding epitope

based approaches

As sIgE is the main player in food allergic diseases, great effort

has been made in order to find biomarkers that discriminate

between asymptomatic sensitization and a clinical relevant



Table 2. Pros and cons of methods for IgE binding epitope identification.

Epitope mapping technique Pros Cons References

Peptide-array Allows for identification of a polyclonal response

Easy method

Large data output

Identification of only linear epitopes [60–62]

X-ray crystallography Allows for identification of both linear and

conformational epitopes

All types of interactions are realized

Most accurate structural information

Difficult to obtain diffracting crystals

Laborious method

Low data output

Identification of only monoclonal responses

[56,63,64]

NMR Allows for identification of both linear and


A rather fast technique

Dynamic behavior of the allergen–antibody

complex can be investigated

Generally identification of only monoclonal responses

Limited to allergens:antibody complexes of small sizes

Low data output

[63,65]

Site-directed mutagenesis Allows for identification of both linear and


Identification of only monoclonal responses

Laborious method

Low data output

[63,66]

Phage display technology Allows for identification of both linear and


Allows for identification of a polyclonal response

Large data output

Laborious method

Selection of unspecific allergen unrelated peptides

[67,68]

food allergy and between allergic phenotypes. Such biomark-

ers could be useful in predicting the course of the disease or

the efficacy of therapeutic interventions.

Food allergy is a very heterogeneous disease according to

clinical manifestations (severity and persistency). Conse-

quently it would tremendously increase the diagnostic and

therapeutic value of the available IgE and IgE binding epitope

based approaches if the IgE binding characteristics, both at

the allergen as well as the epitope level, could be correlated

with the clinical phenotype. CRD facilitated the detection of

patient-specific patterns at an allergen level. It revealed a

broad heterogeneity in the allergen-specific responses be-

tween patients [72,73]. In some conditions this is helpful

to understand the food allergic phenotype. In particular in

peanut allergy both, the diversity as well as the recognition of

specific allergens, such as Ara h 2, have been associated with

severe peanut allergy [74,75]. In peanut allergy also cell-based

approaches, such as BAT, have provided clinically meaning-

ful data to predict food allergy [37]. Even though the appli-

cability of cell-based assays in identifying a clinical relevant

food allergy and the associated phenotype are only scarcely

described, these could be promising future diagnostic and

monitoring tools as the assays are based on the functionality

of the raised IgE response rather than just the presence of sIgE.

Investigating the role of IgE binding epitopes in food

allergy has involved the attempt to correlate patterns of

IgE binding epitope recognition as well as the attempt to

correlate individual epitope biomarkers with a clinically rel-

evant food allergy and the associated allergic phenotype [7].

On the epitope level a great heterogeneity exists between


individual patients, with each having their own unique

pattern of IgE binding epitopes [62,67]. Further, IgE epitope

mapping performed with the inclusion of alanine scan has

revealed that patients reacting towards the same epitope may

indeed react with heterogeneity at the amino acid level,

revealing different patterns of amino acids contributing to

the antibody binding [62]. However, the clinical relevance of

the binding pattern at the amino acid level needs to be

elucidated. Several studies suggest an association between

IgE epitope diversity and persistency [31,76,77] as well as

severity [31,78,79] of the food allergy. In milk and egg allergy

the recognition of specific IgE binding epitopes has been

suggested as biomarkers of persistency and/or severity

[76,77]. On equal terms epitope mapping may be utilized

in the monitoring of therapeutic efficacy [80]. Additionally,

there has been an interest in the therapeutic utilization of

epitope mapping [66,81], by means of modifying specific

allergenic areas of the allergen or identifying new therapeutic

targets [81]. Although all methods allowing for identification

of IgE binding epitopes have limitations, epitope mapping

could be a promising future tool for diagnosis and treatment

of food allergic individuals.

Conclusion

New or improved approaches based on allergen sIgE, their

binding sites or functionality have the potential to become

accurate and trustworthy tools for diagnosis, prognosis and

monitoring of therapeutic efficacy in food allergy and will

add to our understanding of the etiology and pathology of

this disease. However, more research is needed in order to


invent tools providing accurate information on the course of

the food allergic disease.



Acknowledgement

The work was supported by the COST Action FA1402 entitled:

Improving Allergy Risk Assessment Strategy for New Food

Proteins (ImpARAS).

References[1] Bischoff S, Crowe SE. Gastrointestinal food allergy: new insights into

pathophysiology and clinical perspectives. Gastroenterology

2005;128:1089–113.

[2] Ortolani C, Ispano M, Pastorello E, Bigi A, Ansaloni R. The oral allergy

syndrome. Ann Allergy 1988;61:47–52.

[3] Sicherer SH, Sampson HA. Food allergy. J Allergy Clin Immunol

2006;117:S470–5.

[4] Sicherer SH, Sampson HA. Food allergy: epidemiology, pathogenesis,

diagnosis, and treatment. J Allergy Clin Immunol 2014;133:291–307.

[5] Longo G, Berti I, Burks AW, Krauss B, Barbi E. IgE-mediated food allergy in

children. Lancet 2013;382:1656–64.

[6] Ponce M, Diesner SC, Szepfalusi Z, Eiwegger T. Markers of tolerance

development to food allergens. Allergy 2016. in press.

[7] Lin J, Sampson HA. The role of immunoglobulin E-binding epitopes in the

characterization of food allergy. Curr Opin Allergy Clin Immunol

2009;9:357–63.

[8] von Bahr-Lindstrom H, Bennich H. Human immunoglobulin E. The

primary structure of the third constant (C epsilon 3) domain of the epsilon

chain. FEBS Lett 1974;40:57–61.

[9] Ravetch JV, Kinet JP. Fc receptors. Annu Rev Immunol 1991;9:457–92.

[10] Eiwegger T, Szepfalusi Z. IgE testing. In: Encyclopedia of medical

immunology. Springer; 2014. p. 389–92.

[11] Boyce JA, Assa’ad A, Burks AW, Jones SM, Sampson HA, Wood RA, et al.

Guidelines for the diagnosis and management of food allergy in the

United States: report of the NIAID-sponsored expert panel. J Allergy Clin

Immunol 2010;126:S1–58.

[12] van Ree R, Vieths S, Poulsen LK. Allergen-specific IgE testing in the

diagnosis of food allergy and the event of a positive match in the

bioinformatics search. Mol Nutr Food Res 2006;50:645–54.

[13] Kattan JD, Sicherer SH. Optimizing the diagnosis of food allergy. Immunol

Allergy Clin North Am 2015;35:61–76.

[14] Soares-Weiser K, Takwoingi Y, Panesar SS, Muraro A, Werfel T, Hoffmann-

Sommergruber K, et al. The diagnosis of food allergy: a systematic review

and meta-analysis. Allergy 2014;69:76–86.

[15] Lin XP, Magnusson J, Ahlstedt S, Hlman-Hoglund A, Hanson LL,

Magnusson O, et al. Local allergic reaction in food-hypersensitive adults

despite a lack of systemic food-specific IgE. J Allergy Clin Immunol

2002;109:879–87.

[16] Nicolaou N, Poorafshar M, Murray C, Simpson A, Winell H, Kerry G, et al.

Allergy or tolerance in children sensitized to peanut: prevalence and

differentiation using component-resolved diagnostics. J Allergy Clin

Immunol 2010;125:191–7.

[17] Mehl A, Verstege A, Staden U, Kulig M, Nocon M, Beyer K, et al. Utility of

the ratio of food-specific IgE/total IgE in predicting symptomatic food

allergy in children. Allergy 2005;60:1034–9.

[18] Machinena-Spera A, Giner-Munoz MT, Varo-Lozano M, Iniesta-Benedicto

R, Lozano-Blasco J, Piquer-Gibert M, et al. Can total IgE/specific IgE ratio

predict tolerance in cow’s milk allergic children. Pediatr Allergy Immunol

2014;25:823–6.

[19] Nettis E, Bonifazi F, Bonini S, Di LE, Maggi E, Melioli G, et al. Molecular

diagnosis and the Italian Board for ISAC. Eur Ann Allergy Clin Immunol

2014;46:68–73.

[20] Suratannon N, Ngamphaiboon J, Wongpiyabovorn J, Puripokai P,

Chatchatee P. Component-resolved diagnostics for the evaluation of

peanut allergy in a low-prevalence area. Pediatr Allergy Immunol

2013;24:665–70.

[21] Prosperi MC, Belgrave D, Buchan I, Simpson A, Custovic A. Challenges in

interpreting allergen microarrays in relation to clinical symptoms: a

machine learning approach. Pediatr Allergy Immunol 2014;25:71–9.

[22] Werfel T, Asero R, Ballmer-Weber BK, Beyer K, Enrique E, Knulst AC, et al.

Position paper of the EAACI: food allergy due to immunological cross-

reactions with common inhalant allergens. Allergy 2015;70:1079–90.

[23] Masthoff LJ, Mattsson L, Zuidmeer-Jongejan L, Lidholm J, Andersson K,

Akkerdaas JH, et al. Sensitization to Cor a 9 and Cor a 14 is highly specific

for a hazelnut allergy with objective symptoms in Dutch children and

adults. J Allergy Clin Immunol 2013;132:393–9.

[24] Klemans RJ, Liu X, Knulst AC, Knol MJ, Gmelig-Meyling F, Borst E, et al.

IgE binding to peanut components by four different techniques: Ara h 2 is

the most relevant in peanut allergic children and adults. Clin Exp Allergy

2013;43:967–74.

[25] Hofmann SC, Fischer J, Eriksson C, Bengtsson GO, Biedermann T, Jakob T.

IgE detection to alpha/beta/gamma-gliadin and its clinical relevance in

wheat-dependent exercise-induced anaphylaxis. Allergy 2012;67:1457–60.

[26] Klemans RJ, Knol EF, Michelsen-Huisman A, Pasmans SG, de Kruijf-

Broekman W, Bruijnzeel-Koomen CA, et al. Components in soy allergy

diagnostics: Gly m 2S albumin has the best diagnostic value in adults.

Allergy 2013;68:1396–402.

[27] Larche M, Akdis CA, Valenta R. Immunological mechanisms of allergen-

specific immunotherapy. Nat Rev Immunol 2006;6:761–71.

[28] Mita H, Yasueda H, Akiyama K. Affinity of IgE antibody to antigen

influences allergen-induced histamine release. Clin Exp Allergy

2000;30:1583–9.

[29] Gould HJ, Sutton BJ, Beavil AJ, Beavil RL, McCloskey N, Coker HA, et al.

The biology of IgE and the basis of allergic disease. Annu Rev Immunol

2003;21:579–628.

[30] El-Khouly F, Lewis SA, Pons L, Burks AW, Hourihane JO. IgG and IgE

avidity characteristics of peanut allergic individuals. Pediatr Allergy

Immunol 2007;18:607–13.

[31] Wang J, Lin J, Bardina L, Goldis M, Nowak-Wegrzyn A, Shreffler WG, et al.

Correlation of IgE/IgG4 milk epitopes and affinity of milk-specific IgE

antibodies with different phenotypes of clinical milk allergy. J Allergy

Clin Immunol 2010;125:695–702.

[32] Chardin H, Mercier K, Frydman C, Vollmer N. Surface Plasmon Resonance

imaging: a method to measure the affinity of the antibodies in allergy

diagnosis. J Immunol Methods 2014;405:23–8.

[33] Metcalfe DD, Pawankar R, Ackerman SJ, Akin C, Clayton F, Falcone FH,

et al. Biomarkers of the involvement of mast cells, basophils and

eosinophils in asthma and allergic diseases. World Allergy Organ J

2016;9:7.

[34] Caubet JC, Sampson HA. Beyond skin testing: state of the art and new

horizons in food allergy diagnostic testing. Immunol Allergy Clin North

Am 2012;32:97–109.

[35] Hoffmann HJ, Santos AF, Mayorga C, Nopp A, Eberlein B, Ferrer M, et al.

The clinical utility of basophil activation testing in diagnosis and

monitoring of allergic disease. Allergy 2015;70:1393–405.

[36] Muraro A, Werfel T, Hoffmann-Sommergruber K, Roberts G, Beyer K,

Bindslev-Jensen C, et al. EAACI food allergy and anaphylaxis guidelines:

diagnosis and management of food allergy. Allergy 2014;69:1008–25.

[37] Santos AF, Douiri A, Becares N, Wu SY, Stephens A, Radulovic S, et al.

Basophil activation test discriminates between allergy and tolerance in

peanut-sensitized children. J Allergy Clin Immunol 2014;134:645–52.

[38] Ford LS, Bloom KA, Nowak-Wegrzyn AH, Shreffler WG, Masilamani M,

Sampson HA. Basophil reactivity, wheal size, and immunoglobulin levels

distinguish degrees of cow’s milk tolerance. J Allergy Clin Immunol

2013;131:180–6.

[39] McGowan EC, Savage JH, Courneya JP, Sterba PM, Parihar S, Lin J, et al.

Relationship of IgE to basophil phenotypes in peanut-sensitized adults. J


[40] Santos AF, James LK, Bahnson HT, Shamji MH, Couto-Francisco NC, Islam

S, et al. IgG4 inhibits peanut-induced basophil and mast cell activation in













http://dx.doi.org/10.1111/all.12953















































































































peanut-tolerant children sensitized to peanut major allergens. J Allergy

Clin Immunol 2015;135:1249–56.

[41] Wan D, Wang X, Nakamura R, Alcocer MJ, Falcone FH. Use of humanized

rat basophil leukemia (RBL) reporter systems for detection of allergen-

specific IgE sensitization in human serum. Methods Mol Biol

2014;1192:177–84.

[42] Falcone FH, Alcocer MJ, Okamoto-Uchida Y, Nakamura R. Use of

humanized rat basophilic leukemia reporter cell lines as a diagnostic tool

for detection of allergen-specific IgE in allergic patients: time for a

reappraisal? Curr Allergy Asthma Rep 2015;15:67.

[43] Knipping K, van Roest M, Kruijssen L, Smits M, Teunis M, Cox L, et al.

Intra- and inter-laboratory validation of an innovative

huFcepsilonRIalpha-RBL-2H3 degranulation assay for in vitro

allergenicity assessment of whey hydrolysates. Toxicol In Vitro

2016;33:29–34.

[44] Bøgh KL, Kroghsbo S, Dahl L, Rigby NM, Barkholt V, Mills ENC, et al.

Digested Ara h 1 has sensitizing capacity in Brown Norway rats. Clin Exp

Allergy 2009;39:1611–21.

[45] Kroghsbo S, Bøgh KL, Rigby NM, Mills ENC, Rogers A, Madsen CB.

Sensitization with 7S globulins from peanut, hazelnut, soy or pea induces

IgE with different biological activities which are modified by soy

tolerance. Int Arch Allergy Immunol 2011;155:212–24.

[46] Schmetzer O, Valentin P, Smorodchenko A, Domenis R, Gri G, Siebenhaar

F, et al. A novel method to generate and culture human mast cells:

peripheral CD34+ stem cell-derived mast cells (PSCMCs). J Immunol

Methods 2014;413:62–8.

[47] Siebenhaar F, Falcone FH, Tiligada E, Hammel I, Maurer M, Sagi-Eisenberg

R, et al. The search for mast cell and basophil models – are we getting

closer to pathophysiological relevance. Allergy 2015;70:1–5.

[48] Van Regenmortel MHV. Mimotopes, continuous paratopes and

hydropathic complementarity: novel approximations in the description

of immunochemical specificity. J Dispers Sci Technol 1998;19:1199–219.

[49] Aalberse RC. Structural biology of allergens. J Allergy Clin Immunol

2000;106:228–38.

[50] Van Regenmortel MHV. Mapping epitope structure and activity: from

one-dimensional prediction to four-dimensional description of antigenic

specificity. Methods 1996;9:465–72.

[51] Barlow DJ, Edwards MS, Thornton JM. Continuous and discontinuous

protein antigenic determinants. Nature 1986;322:747–8.

[52] Van Regenmortel MH. What is a B-cell epitope. Methods Mol Biol

2009;524:3–20.

[53] Kringelum JV, Nielsen M, Padkjaer SB, Lund O. Structural analysis of B-cell

epitopes in antibody:protein complexes. Mol Immunol 2013;53:24–34.

[54] Laver WG, Air GM, Webster RG, Smith-Gill SJ. Epitopes on protein

antigens: misconceptions and realities. Cell 1990;61:553–6.

[55] Bannon GA, Ogawa T. Evaluation of available IgE-binding epitope data

and its utility in bioinformatics. Mol Nutr Food Res 2006;50:638–44.

[56] Pomes A. Relevant B cell epitopes in allergic disease. Int Arch Allergy

Immunol 2010;152:1–11.

[57] Uttamchandani M, Yao SQ. Peptide microarrays: next generation biochips

for detection, diagnostics and high-throughput screening. Curr Pharm

Des 2008;14:2428–38.

[58] Berrade L, Garcia AE, Camarero JA. Protein microarrays: novel

developments and applications. Pharm Res 2011;28:1480–99.

[59] Weinrich D, Jonkheijm P, Niemeyer CM, Waldmann H. Applications of

protein biochips in biomedical and biotechnological research. Angew

Chem Int Ed Engl 2009;48:7744–51.

[60] Matsuo H, Yokooji T, Taogoshi T. Common food allergens and their IgE-

binding epitopes. Allergol Int 2015;64:332–43.

[61] Forsstrom B, Axnas BB, Stengele KP, Buhler J, Albert TJ, Richmond TA,

et al. Proteome-wide epitope mapping of antibodies using ultra-dense

peptide arrays. Mol Cell Proteomics 2014;13:1585–97.


[62] Hansen CS, Dufva M, Bøgh KL, Sullivan E, Patel J, Eiwegger T, et al. Patient

specific and persistent epitopes in peanut allergens demonstrated by

linear epitope mapping. J Allergy Clin Immunol 2016. pii: S0091-

6749(16)30610-8.

[63] Dall’antonia F, Pavkov-Keller T, Zangger K, Keller W. Structure of

allergens and structure based epitope predictions. Methods 2014;66:

3–21.

[64] Malito E, Carfi A, Bottomley MJ. Protein crystallography in vaccine

research and development. Int J Mol Sci 2015;16:13106–40.

[65] Takeuchi K, Wagner G. NMR studies of protein interactions. Curr Opin

Struct Biol 2006;16:109–17.

[66] Gershoni JM, Roitburd-Berman A, Siman-Tov DD, Tarnovitski FN, Weiss

Y. Epitope mapping: the first step in developing epitope-based vaccines.

BioDrugs 2007;21:145–56.

[67] Bogh KL, Nielsen H, Eiwegger T, Madsen CB, Mills EN, Rigby NM, et al. IgE

versus IgG4 epitopes of the peanut allergen Ara h 1 in patients with severe

allergy. Mol Immunol 2014;58:169–76.

[68] Christiansen A, Kringelum JV, Hansen CS, Bøgh KL, Sullivan E, Patel J,

et al. High-throughput sequencing enhanced phage display enables the

identification of patient-specific epitope motifs in serum. Sci Rep

2015;5:12913.

[69] Batori V, Friis EP, Nielsen H, Roggen EL. An in silico method using an

epitope motif database for predicting the location of antigenic

determinants on proteins in a structural context. J Mol Recognit

2005;18:1–9.

[70] Mayrose I, Shlomi T, Rubinstein ND, Gershoni JM, Ruppin E, Sharan R,

et al. Epitope mapping using combinatorial phage-display libraries: a

graph-based algorithm. Nucleic Acids Res 2007;35:69–78.

[71] Huang J, Ru B, Dai P. Bioinformatics resources and tools for phage display.

Molecules 2011;16:694–709.

[72] Lieberman JA, Sicherer SH. Diagnosis of food allergy: epicutaneous skin

tests, in vitro tests, and oral food challenge. Curr Allergy Asthma Rep

2011;11:58–64.

[73] Hong X, Caruso D, Kumar R, Liu R, Liu X, Wang G, et al. IgE, but not IgG4,

antibodies to Ara h 2 distinguish peanut allergy from asymptomatic

peanut sensitization. Allergy 2012;67:1538–46.

[74] Lewis SA, Grimshaw KE, Warner JO, Hourihane JO. The promiscuity of

immunoglobulin E binding to peanut allergens, as determined by

Western blotting, correlates with the severity of clinical symptoms. Clin

Exp Allergy 2005;35:767–73.

[75] Kukkonen AK, Pelkonen AS, Makinen-Kiljunen S, Voutilainen H, Makela

MJ. Ara h 2 and Ara 6 are the best predictors of severe peanut allergy: a

double-blind placebo-controlled study. Allergy 2015;70:1239–45.

[76] Jarvinen KM, Beyer K, Vila L, Bardina L, Mishoe M, Sampson HA.

Specificity of IgE antibodies to sequential epitopes of hen’s egg

ovomucoid as a marker for persistence of egg allergy. Allergy 2007;62:

758–65.

[77] Jarvinen KM, Beyer K, Vila L, Chatchatee P, Busse PJ, Sampson HA. B-cell

epitopes as a screening instrument for persistent cow’s milk allergy. J


[78] Flinterman AE, Knol EF, Lencer DA, Bardina L, den Hartog Jager CF, Lin J,

et al. Peanut epitopes for IgE and IgG4 in peanut-sensitized children in

relation to severity of peanut allergy. J Allergy Clin Immunol

2008;121:737–43.

[79] Battais F, Mothes T, Moneret-Vautrin DA, Pineau F, Kanny G, Popineau Y,

et al. Identification of IgE-binding epitopes on gliadins for patients with

food allergy to wheat. Allergy 2005;60:815–21.

[80] Vickery BP, Lin J, Kulis M, Fu Z, Steele PH, Jones SM, et al. Peanut oral

immunotherapy modifies IgE and IgG4 responses to major peanut

allergens. J Allergy Clin Immunol 2013;131:128–34.

[81] De Groot AS. Immunomics: discovering new targets for vaccines and

therapeutics. Drug Discov Today 2006;11:203–9.





























































































































DRUG DISCOVERY

TODAY

DISEASEMODELS

Non-IgE mediated food allergyDaniel Lozano-Ojalvo1, Guillaume Lezmi2, Naima Cortes-Perez3,

Karine Adel-Patient3,*1Instituto de Investigacion en Ciencias de la Alimentacion (CIAL, CSIC-UAM), 28049 Madrid, Spain2Pediatric Gastroenterology Service, Hopital Necker Enfants Malades, F-75015 Paris, France3Service de Pharmacologie et Immunoanalyse (SPI), CEA, INRA, Universite Paris-Saclay, F-91191 Gif-sur-Yvette, France


Editors-in-Chief




Non-IgE-mediated food allergies (FA) are highly prev-

alent within food allergic patients, notably in the first

years of life. The most prevalent non-IgE FA are mainly

induced by cow’s milk and soya, but many other foods

can be involved. Non-IgE FA encompass a wide range of

immune-related disorders that differ in prevalence,

clinical manifestation, and pathophysiology. Although

some in vivo models have been developed for their

study, further investigations are needed to fully delin-

eate the pathogenic mechanisms involved.

Introduction

Food allergies (FA) correspond to ‘an adverse health effect

arising from a specific immune response that occurs repro-

ducibly on exposure to a given food’ [1]. FA can be either IgE-

or non-IgE-mediated, both resulting from barrier dysfunction

and immune dysregulation. Although prospective cohort

studies demonstrated that almost 50% of allergic infants

endure non-IgE FA [2,3], they are often misdiagnosed and

less well studied than IgE-mediated FA.

The most prevalent non-IgE FA are eosinophilic esophagi-

tis (EoE), food-protein induced enterocolitis syndrome

(FPIES), proctocolitis (FPIAP) and entheropathy (FPE). In

the present review, we will mainly focus on EoE for which

clinical data and animal models are the more abundant.

Celiac disease, a prevalent adverse immune reaction triggered

by gluten, has been already largely described and then will

not be considered here.

*Corresponding author: K. Adel-Patient ([email protected])

1740-6757/$ � 2016 Published by Elsevier Ltd. http://dx.doi.org/10.1016/j.ddmod.2016.09.0


Eosinophilic esophagitis (EoE)

Prevalence, clinical manifestations, and trigger foods

EoE is considered a non-IgE FA based on immunological findings

and clinical evidences such as the inefficiency of anti-IgE therapy

or the increased frequency of EoE observed after oral immuno-

therapy for IgE FA [4]. EoE prevalence increased in the past years,

reaching 0.05–0.1% of the general population in the US [5].

EoE is a chronic disease characterized by esophageal dys-

function and eosinophilic inflammation of the esophagus [6].

In symptomatic patients, EoE is diagnosed after an esophage-

al biopsy showing at least 15 intraepithelial eosinophils per

high-power field after an 8–12 weeks course of proton pump

inhibitor to rule out a gastro-esophageal reflux disease-related

eosinophilia [6].

Patients with EoE are highly atopic, with elevated rates of

allergic rhinitis, asthma, eczema, or even IgE FA [6,7]. In infants

and children, symptoms include feeding difficulty, nausea,

vomiting, and failure to thrive. School-aged children often suffer

abdominal pain and frequently vomiting, whereas adolescents

and adults usually have dysphagia or food impactions [6,8,9].

Although previous results suggest the role of aeroallergens

in EoE, a recent meta-analysis study failed to show seasonal

03 45







recrudescence of EoE [10]. Conversely, marked improve-

ments of clinical and histologic symptoms upon strict ami-

no-acid based diet, and recurrence of symptoms upon food

challenge, evidenced the role of foods as triggers of EoE [11],

the most commonly involved being cow’s milk (CM), wheat,

soy, and egg [12]. Amino acid-based elemental diet, six food

elimination diet (based on the avoidance of the 6 most

common triggering foods), and allergy-test result-directed

food eliminations diets are effective in inducing clinical

and histologic remission [13]. Swallowed topical corticoste-

roids demonstrated clinical and histologic improvements,

but clinical symptoms recurred upon discontinuation and

response varied between patients [14].

Mechanisms

EoE may result from an alteration of the oesophageal epithe-

lial barrier induced by intrinsic dysfunction or environmen-

tal aggressions, such as gastric acid, microbiota changes and/

or xenobiotic. This may lead to an increased oesophageal

permeability, and activation of innate immune response.

Together with exposure to food antigens, these events could

initiate (local) food-specific Th2 cell activation. Recent

reviews give new insights into the pathogenesis of EoE

[15–17] and the more recent findings in humans are

highlighted below and summarized in Fig. 1.

Intrinsic barrier dysfunction(genetic)

E(gastric acid, xen

Oesophageal epithelium alterat(altered desmoglein/keratin expression)

Innate cell activation(ILC2, mast cells, basophils,

eosinophils, iNKT?)

Increase permeability

Dysregulatedregulatoryresponses

EosinophiliaMast cell recruitment & activat

DC activation and induction of Th2 adaptive immune response

(ILC2?, iNKT?)

Food allergenexposure

IL-33, TSLP

IL-4, IL-5, IL-13IL-9

IL-4

Amphiregulin(repair)

Figure 1. Pathogenesis of EoE. EoE may initially result from an alteration of t

environmental aggressions. This may lead to an increased oesophageal permeabil

food antigens, these events could initiate (local) food-specific Th2 cell activation

dysfunction.


Alteration of epithelial integrity and activation of innate

immunity

Oesophageal biopsies from patients with EoE showed a de-

creased expression of molecules essential for epithelial integ-

rity, such as desmoglein-1 (DSG-1) [18]. Downstream, both

thymic stromal lymphopoietin (TSLP) and IL-33, two epithe-

lial cell-derived cytokines that may promote Th2 type re-

sponse after an epithelial aggression through the activation

of innate lymphoid cells 2 (ILC2) and/or basophils, are over-

expressed in oesophageal biopsies from EoE patients [19,20].

Accordingly, the percentage of ILC2 and basophil responses

are higher in the oesophageal mucosa of patients with active

EoE than in those with inactive EoE or control subjects [19,21].

Upstream from TSLP/IL-33, TNF-related apoptosis-inducing

ligand (TRAIL), which has been shown to induce TSLP pro-

duction in vitro, is upregulated in the oesophageal mucosa of

patients with EoE [22]. In addition, the eosinophil chemoat-

tractant eotaxin-3 (CCL26) is overexpressed in oesophageal

epithelial cells from patients with EoE as a result from single-

nucleotide polymorphism mutation or epigenetic modifica-

tion [23,24]. The numbers of invariant natural killer T cells

(iNKTs) and mast cells are also increased in the oesophageal

mucosa of patients with EoE compared with healthy controls

[25,26]. Likewise, the expression of some histamine receptors

is increased in epithelial eosinophils of patients with EoE [27].

xtrinsic factorsobiotics, microbiota dysbiose/infection)

ion

Th2 chronic inflammationFibrotic modifications

Oesophagealdysfunction

ionspecific

IL-13

Eotaxin

LT, PGDTh2 cytokinesEotaxins


he oesophageal epithelial barrier induced by intrinsic dysfunction or

ity and activation of innate immune response. Together with exposure to

leading to chronic inflammation, fibrotic modifications and oesophageal


Evidence for a Th2 adaptive response

EoE is characterized by an increased expression of Th2 cyto-

kines in oesophageal biopsies and blood; these cytokines may

not only enhance activation and survival of eosinophils but

also increase/maintain inflammation and induce fibrotic

modifications [28–30]. EoE patients are more likely to present

polymorphisms in gene encoding IL-13 [31] and epithelial

cells from biopsies have shown an overexpression of IL-13

receptor (IL13R) and a high production of IL-13. In response

to IL-13, cultured epithelial cells produce eotaxin-3 and

reduce their expression of DSG-1 [18,32]. Polymorphisms

in the IL-5 gene have also been associated with EoE, as well

as increased circulating IL-5+ T cells and esophageal levels of

the receptor IL-5R [28,31,33]. Thus, a high secretion of IL-5,

potentiated by eotaxin-3, may increase the number of eosin-

ophils in the esophagus. However, IL-5 blockage reduced only

partially the recruitment and activation of eosinophils in

patients with EoE [34], suggesting that other pathways may

be involved in EoE pathogenesis. In that connection, an

immune dysregulation may be present in EoE, with regulato-

ry T cells (Treg) being increased in children and decreased in

adults with EoE [35,36]. Other cytokines such as IL-15, IL-18

or TGF-b1 are also overexpressed in patients with EoE [37–39].

In vivo models

The complex interplay between epithelial barrier, innate

and adaptive cells, and mediators within the esophageal

mucosa is difficult to reproduce in animal models. However,

various models of EoE have been proposed, mainly in mice

(Table 1). Although most of these models could not repro-

duce EoE as observed in humans (i.e. eosinophilic inflam-

mation restricted to the esophagus independently of IgE

production, improvement of experimentally-induced EoE

by local anti-inflammatory therapy or after removal of the

stimuli, food impaction), they recreate some of the observa-

tions in humans and give further insights in the putative

pathogenesis of EoE. These models also provide interesting

results which lead to the development of potential new

therapies.

Airway exposure to aeroallergen/cytokine

Initial studies showed that repeated intranasal (i.n.) admin-

istrations of aeroallergens such as Aspergillus fumigatus in

mice induced eosinophilia in the lungs, esophagus and

blood, but not in the stomach or small intestine [40]. Eosin-

ophils (�25–35 eosinophils/mm2) were predominant in the

lamina propria and the submucosa of the esophagus, and

about 50% of eosinophils were undergoing cell death. Esoph-

ageal eosinophilia (�6 eosinophils/mm2) was also induced

in mice sensitized to ovalbumin (Ova) through the intra-

peritoneal (i.p.) route and then repeatedly i.n. exposed.

Conversely, esophageal eosinophilia was neither induced

by intra-gastric (i.g.) or oral administration of A. fumigatus,

nor by i.n. inoculation in non-anesthetized mice (i.e. in mice

able to swallow). The critical role of IL-5 and the partial role

of eotaxin in the induction of esophageal eosinophilia were

then demonstrated using this model. As IL-5 has been related

to lung eosinophilia, whereas eotaxin is critical for basal

homing of eosinophils in the gastrointestinal tract (stomach

and small intestine), the authors proposed a causal link

between respiratory and esophageal hypersensitivity, which

is debatable when considering human data [10,11,34].

The same group reproduced A. fumigatus-induced-eosino-

phil influx in esophagus (and lung), and epithelial cell

hyperplasia after intratracheal delivery of IL-13, following

a protocol known to induce experimental asthma [41]. The

esophageal eosinophilia was dose-dependent and involved

STAT5, IL-5, and partially eotaxin-1, implicating Th2 cells in

EoE pathogenesis, although systemic Th2 response was not

observed.

CD4+, CD8+ and B cells influx was further evidenced in

this A. fumigatus model [42]. Interestingly, B cells or antigen-

specific antibodies were not involved in the recruitment of

eosinophils in the esophagus, whereas a critical role for

adaptive T-cell immunity was demonstrated. However,

CD4+ T cells dependency was less important in the esophagus

than in the lung, and CD8+ T cells were dispensable. This

study also showed that intra-tracheal (i.t.) administration was

as efficient as i.n. application for inducing esophageal (and

lung) eosinophilia (50–66 eosinophils/mm2 counted after

nine administrations in both cases), thus questioning the

importance of topical exposure to esophagus initially stated

by these authors in [40]. For both, i.n. and i.t. exposure, at

least six doses were needed to induce significant eosinophilia

in esophagus, whereas lung eosinophilia was induced after 4–

5 applications. Unfortunately, in most of these studies, both

the gender and the strain of mice were not specified. A

comparison of the results obtained in BALB/c (Th2 biased)

versus C57BL/6 (Th1 biased) strains would be very informa-

tive.

Recently, the roles of TRAIL and TRAIL-induced TSLP

production were evidenced in the initial phase of A. fumiga-

tus-induced airway and esophageal eosinophilia, in relation

to EoE clinical data [22]. Esophageal eosinophilia (�40 eo-

eosinophils/mm2) was associated with increased mast cell

number, but not CD4+ T cells. This study evidenced the

complex spatial and temporal expression pattern of various

factors (such as TRAIL, TSLP, eotaxin, IL-5, IL-13 and TGF-b).

A. fumigatus-induced eosinophilia in the esophagus, but not

in the respiratory tract, was further shown to be dependent

on IL-15 produced by macrophages/dendritic cells, suggest-

ing that different mechanisms may be involved in esophageal

and airway eosinophilia [37]. In vitro, IL-15 primes CD4+ T

cells for Th2 cytokines production via STAT5 activation, and

induces eotaxin production by esophageal epithelial cells

[37].



Table 1. Main animal models for EoE: experimental procedures and main outcomes. i.n.: intra-nasal, i.p.: intra-peritoneal, i.g.:intra-gastric, i.t.: intra-tracheal, NP: not provided

Model Mouse strain Genetic

modifications

Age/gender/

breeding

conditions

EoE induction Main outcome Ref

Airway

exposure to

aeroallergen

Aspergillus

fumigatus

BALB/c

C57BL/6

Wild-type

Eotaxin or IL-5

deficient

8–10 week-old

Males/females

SPF

Repeated i.n., oral or

i.g. administrations of

A. fumigatus (3

treatments/week, for

3 weeks)

I.p. sensitization to

ovalbumin (alum) and

repeated i.n. dosing

with ova (150 mg, 7

exposures over10

days)

Esophageal (and lung) eosinophilic

inflammation, extracellular granule

deposition and epithelial cell

hyperplasia after i.n.

administration (100 mg in 50 ml) –

importance of anesthesia

Critical role of IL-5/partial role of

eotaxin in the pathophysiological

changes

[40]

BALB/c

C57BL/6

Wild-type

Lymphocytes (RAG1)-,

B cell (IgH6)-,

T cell (Foxn1)-,

CD4- or

CD8a-deficient

6–8 week-old

Males/females

SPF

Repeated i.n. or i.t.

administrations of

(endotoxin-free) A.

fumigatus in

anesthetized mice

Validation of i.n. and i.t. routes for

experimental airway-EoE and

determination of the number of

doses required

Evidence of the role of adaptive T

cell immunity

[42]

BALB/c Wild type

TRAIL-deficient

8–12 week-old

Male

NP

Repeated i.n.

administrations of A.

fumigatus in

anesthetized mice

TRAIL expression is detected as

soon as 24 hours after the first

administration of A. fumigatus.

TRAIL then induced TSLP, which is

sufficient to induced esophagus

eosinophilia and remodeling

[22]

BALB/c Wild type

IL-15Ra deficient

6–8 week old

Males/females

SPF

IL-15 produced by esophageal

macrophages and dendritic cells

activates CD4+ cells to produce

IL-5 and IL-13 and epithelial cells

to produce eotaxin, thus

participates to experimental EoE

but not lung eosinophilia

[37]

Intratracheal

cytokine

administration

BALB/c Wild type

STATS-6,

eotaxin-1 or

IL-5 deficient

8–12 week-old

Males/females

NP

Repeated i.t.

administrations of

various doses of IL-

13, IL-4, IL-10,or IL-9

IL13 reproduce A. fumigatus-

induced EoE and is dependent on

IL-5, eotaxin-1, and STAT6

[41]

Systemic

sensitization

and local

exposure with

food antigens

BALB/c mice Wild type

Smad3 deficient

8-week old

Females

NP

2 i.p. sensitizations

with Ova (50 mg,

Alum) and repeated

intra-esophageal

administrations of

Ova(10 mg, 3 times/

week for 4 weeks)

Development of an Ova-induced

model of EoE (eosinophilia,

esophagus remodeling,

angiogenesis), but no info at other

sites that esophagus (stomach,

small intestine)

TGF-b signaling is critical in

esophageal remodeling

[44]

Wild type Role of eosinophils in

inflammation but also in

angiogenesis, deposition of

fibronectin and basal zone

hyperplasia.

[43]

BALB/c Wild type

Lta and CD1d

deficient

6–8 week-old

Males and

females (mix)

SPF

2 i.p. sensitizations

with corn or peanut

extract (200 mg,

alum) and repeated

i.n., oral or i.g.

exposures (every

other days, 100 mg)

Development of a peanut/corn

induced EoE (associated with lung

eosinophilia) after i.n. challenges

Role of iNKTs and eotaxin 1/2 in

the initiation of EoE

[45]




Model Mouse strain Genetic

modifications

Age/gender/

breeding

conditions

EoE induction Main outcome Ref

Oral

sensitization

and exposure to

food antigen

BALB/c Wild type 4 week-old

Female

Conventional

breeding

Sensitization by

repeated i.g.

administrations of

crude peanut protein

extract (1 mg) with

Cholera Toxin and

sustained oral and i.g.

exposures over 10

days

Induction of eosinophilia in

esophagus (and jejunum),

associated with a strong local and

systemic Th2 specific immune

response

Pre-clinical model for testing the

efficiency of epicutaneous

immunotherapy, through Treg

induction

[46]

Epicutaneous

sensitization

and i.n. or oral

exposure to

food antigen

BALB/c Wild-type

STAT6, IL-5, IL-13,

IL-4/IL-13

deficient

4–8 week old

Gender not

specified

SPF

Epicutaneous

exposures to Ova or

A. fumigatus using

occlusive patch (2 to

3 one-week

applications) and a

single i.n. challenge

(25 mg) on

anesthetized mice

Development of esophageal (and

lung) eosinophilia

Essential role of IL-5 and partial

role of IL-4, IL-13 and STAT6 in

pathogenesis

[48]

BALB/c

C57BL/6

Wild type

Igh-7-/TSLP

receptor-deficient

BALB/c and C57BL/6

Baso–DTR mice

8–12 week-old

Males and females

SPF

Epicutaneous

sensitization (daily,

for 14 days) to

ovalbumin or crude

peanut extract with

TSLP-inducing agents

and further sustained

intra-gastric and oral

exposure

Development of eosinophilic

inflammation in esophagus (and

stomach and small intestine),

structural changes and food

impactation in 30% of mice

TSLP and basophil contribute to

the pathogenesis of EoE

No role for IgE

[19,49]

Although very informative, the A. fumigatus model has two

main limitations. First, it could be argued that esophageal

eosinophilia may be secondary to lung eosinophilia. Second,

the esophageal eosinophilia observed after exposure through

different routes in mice may be mechanistically distinct from

that observed in humans with EoE.

Systemic sensitization combined with local delivery of food

antigens

Other studies evidenced the induction of EoE in mice

sensitized by the i.p. route and then repeatedly challenged

by the esophageal route using food allergens such as Ova

[43,44], peanut or corn [45]. I.p. sensitization and repeated

esophageal exposure to a high dose of Ova led to a dramatic

esophageal eosinophilia in the lamina propria (�120 eo-

eosinophils/mm2) [44], suggesting that previous work by

Mishra and coworkers [40] used too low doses during

challenges. TGF-b expression and esophagus remodeling

(angiogenesis, fibronectin deposit, epithelial basal

zone hyperplasia) were also evidenced, as the same as blood

eosinophilia. However, although the analysis procedure

and reagents were similar in both studies (anti-major

basic protein (MBP) staining), the number of esophageal

eosinophils in control mice highly differed (�12 eosino-

eosinophils/mm2 vs < 1). Moreover, no information was

provided on the impact of Ova treatment at other (gastro-

intestinal) sites. This model was used to demonstrate the

role of TGF-b-Smad3 axis in the late phase of EoE, that is,

fibrosis and angiogenesis [44], and the efficiency of target-

ing eosinophils using anti-Siglec (sialic acid-binding immu-

noglobulin-like lectin) antibodies as a new therapeutic

approach [43].

Experimental EoE was also successfully induced in mice

after i.p. sensitizations and repeated i.n. or i.g. exposures with

peanut extract [45]. In this study, several features of human

EoE were evidenced such as the esophageal eosinophilic

influx in lamina propria and within the epithelium, eosino-

philic micro abscesses, extracellular MBP+ granules, eosino-

phil-related cytokines mRNA expression, but also mast cell

accumulation and altered epithelial mucosa. I.n. challenge

induced a more severe esophageal eosinophilia than i.g.

challenge, and an eosinophilia in the airways, whereas i.g.

challenge induced eosinophilia also in the small intestine.

Conversely, oral challenge did not induce esophageal eosin-

ophilia. However, i.n. challenged mice were apparently not

anaesthetized, and then allergen could have been partially



swallowed. Moreover, kinetic of eosinophilia in lung and

esophagus was not provided, thus, the esophageal eosino-

philia as a consequence of pulmonary eosinophilia cannot be

ruled out – all the more para-esophagal lymph nodes, allow-

ing trafficking of eosinophils from airway to esophagus, have

a critical role in this model. Interestingly, basal influx of

eosinophils was induced after systemic sensitization and

was further highly increased both after i.n. and i.g. chal-

lenges, with mean values of 167–186 and 36–39 eosino-

phils/mm2, respectively. This influx was far higher than

that induced by i.n. exposure to A. fumigatus, highlighting

the impact of systemic sensitization. Peanut was a more

potent inducer of esophageal eosinophilia than corn, suggest-

ing the possibility of ranking food for their potency to induce

EoE. Additionally, mechanistic studies evidenced the critical

role of iNKTs and eotaxin 1 and 2 in the initiation of the

pathophysiology in this model, in accordance with clinical

human data.

Oral exposure to food antigens

Esophago-gastro eosinophilia was observed in mice sensitized

with peanut protein extract via the oral route, using cholera

toxin as an adjuvant, maintained on an elimination diet for 8

weeks and then submitted to a sustained oral exposure to

peanut [46]. Eotaxin and IL-13 mRNA expressions were in-

creased as early as the second day of oral exposure, but high

eosinophilic infiltration and IL-5 mRNA were only detected

on day 10, that is, after combined and intense oral plus i.g.

peanut protein exposures. The eosinophil influx was high

(�130 eosinophils/mm2), although this level was not reached

in another study using the same experimental procedure

(�36–70 eosinophils/mm2) [47]. Esophageal inflammation

was accompanied with jejunal lesions (necrosis, eosinophilic

inflammation, villous sub-atrophy) and high systemic Th2

response (specific IgE, spleen cells specific secretion of Th2

cytokines). This model was used as a pre-clinical model for

testing the efficiency of epicutaneous immunotherapy,

through Treg induction.

Epicutaneous sensitization and i.n. or oral exposure to food

antigen

Some groups tested the impact of cutaneous exposure on EoE

development in mice. Epicutaneous exposure to Ova or A.

fumigatus using occlusive patches induced eosinophils and

mast cells influx in skin, blood eosinophilia and systemic

sensitization when a further unique i.n. challenge with anti-

gen induced esophageal (�28–35 eosinophils/mm2) and lung

eosinophilia [48]. Esophageal eosinophilia appeared 4 hours

after i.n. challenge, was totally dependent on IL-5, and par-

tially dependent on IL-4, IL-13 and STAT6. Conversely, IL-5

deficient and wild type mice demonstrated the same levels of

total IgE and specific IgG1, thus dissociating antibody re-

sponse and the development of esophageal eosinophilia.


Eosinophil influx (evidenced by histology and flow cyto-

metry), edema, inflammation, food impaction and Th2-re-

lated cytokines expression in the esophagus were also

demonstrated in mice daily exposed to Ova or peanut via

the cutaneous route and then submitted to sustained and

combined i.g. and oral exposures to high doses of antigens

[19]. The associated structural changes of esophagus were

assessed using optical coherence tomography [49]. Interest-

ingly, sensitization was achieved by exposing Ova after tape

stripping or with vitamin D analog to increase TSLP produc-

tion in the skin, as a model of atopic dermatitis. In this

model, experimental EoE is shown to be dependent on

TLSP-elicited basophils and independent on IgE. Notably,

basophil depletion during cutaneous sensitization reduced

eosinophilia and implication of TSLP and basophils corre-

lated with data obtained on pediatric EoE population. Eo-

sinophilia in stomach and small intestine and systemic Th2

cytokine responses was also evidenced, and further studies

demonstrated that intestinal immediate FA is also induced in

this model, which is TSLP-elicited skin basophils and IgE-

dependent [50].

All these models mimicking more and more closely human

EoE improved our understanding of the initial events of EoE

pathogenesis. Additional studies using, for example, altered

esophageal epithelium (physical/chemical/microbial stress,

intrinsic dysfunction/defaults in tight junction or filaggrin

mutation, among others) combined with studies in humans

identifying homing and chemokine receptors, both on eo-

sinophils and T cells, will be useful to finalize and validate EoE

experimental models.

Food protein-induced enterocolitis syndrome (FPIES),

proctocolitis (FPIAP) and enteropathy (FPE): clinically

relevant pathologies without animal models

The main data concerning clinical manifestations, triggered

foods, biological and histologic features and supposed mech-

anisms for FPIES, FPIAP and FPE are gathered in Table 2.

Although these pathologies are clinically relevant, patho-

physiological mechanisms involved are deeply unknown.

To the best of our knowledge, no animal models have been

developed for FPIES, FPIAP or FPE. Human data allowing to

determine the most relevant models to be used (genetic

background, i.e. Th2 biased/Treg dysfunction, tight junction

dysfunction or basal cytokines/chemokines secretion), the

environmental factors potentially involved (factors leading

to alteration of the epithelium, role of microbiota) and the

key players of the immune system in initiating events and in

maintaining inflammation are still lacking, which renders

the development of relevant models still difficult. Moreover,

some symptoms can be difficult to analyse or even to induce

in ‘classical’ models: for example mice/rats do not vomit

which renders the development of models for acute FPIES

challenging.


Table 2. FPIES, FPIAP and FPIE clinical and biological features and putative mechanisms SPT: skin prick test. CM: cow’s milk

Diseases Prevalencea Clinical manifestations Foods Biological and histologic

features

Mechanisms References

FPIES 0.34% Acute form (1–3 hours

after exposure)

Intermittent exposure to the

offending food

Repetitive vomiting,

dehydration, pallor, lethargy,

hypotension

Delayed diarrhea

No skin or respiratory

involvement (6¼ IgE-mediated

anaphylaxis)

90% resolution at 3–5 years

of age

CM, soy

Single food

Variable villous blunting, colitis

Intestinal inflammation

(eosinophils, neutrophils,

lymphocytes)

Blood neutrophilia,

thrombocytosis, acidosis

Methemoglobinemia

Positive SPT or IgE: rare

Increased duodenal mucosal

expression of TNF-a/

decreased duodenal

mucosal expression of TGF-

b

Local production of food

specific IgA and IgM

Immune-neuroendocrine

interplay?

[51–54]

Chronic form

Regular intake of the offending

food

Chronic diarrhea, vomiting,

failure to thrive,

90% resolution at 3–5 years of

age

Hypoalbuminemia, anemia

FPIAP 0.16% Rectal bleeding in otherwise

well children

Possible mucus in stools,

diarrhea, abdominal pain

Resolution of most cases at 1

year of age

CM, soy, egg Eosinophils and lymphocytes

infiltration in the rectal mucosa

Eosinophilia, anemia

Delayed maturation of the

gastrointestinal immune

system/Delayed microbiota

establishment Increased

mucosal expression of

eotaxin-1, CXCL13

[51,53,55–57]

FPE Protracted diarrhea, vomiting,

malabsorption

Abdominal distension, failure

to thrive

Resolution of most cases at 2–3

years of age

CM, soy Duodenal and colic

lymphonodular hyperplasia

with increased intraepithelial

lymphocytes

Hypoprotidemia

Increased IFN-g and IL4

expression in jejunum

Increased intraepithelial

cytotoxic CD8+ T cells

[51,55]

a Prevalence among CM-allergic infants (1%), in a prospective population-based birth-cohort study in Israel [2].

Conclusions

Elucidation of the immune mechanisms and environmental

factors involved in the pathogenesis of non-IgE FA needs

further investigations, in both patients (biopsies, blood sam-

ples, feces) and animal models. Animal models development

using relevant mouse strains and age, clinically relevant food

allergens (mainly CM and soy) and realistic route of exposure,

are needed to better understand the complex mechanisms

involved in epithelial barrier alteration and downstream

dysregulation of the immune system. In addition, such mod-

els will be useful for the development of new preventive and

therapeutic strategies, but also for the assessment of allerge-

nicity of new foods that will soon integrate the human

nutrition due to predicted shortage of proteins for human

consumption.

Acknowledgements

DLO and KAP are part of the COST Action FA1402 entitled:

Improving Allergy Risk Assessment Strategy for New Food

Proteins (ImpARAS). DLO acknowledges his FPU Grant

(MECD) and financial support through AGL2014-59771R

project.

References[1] Sampson HA, Aceves S, Bock SA, James J, Jones S, Lang D, et al. Food

allergy: a practice parameter update – 2014. J Allergy Clin Immunol

2014;134:1016–25.

[2] Katz Y, Goldberg MR, Rajuan N, Cohen A, Leshno M. The prevalence and

natural course of food protein-induced enterocolitis syndrome to cow’s

milk: a large-scale, prospective population-based study. J Allergy Clin

Immunol 2011;127:647–53.

[3] Grimshaw KE, Bryant T, Oliver EM, Martin J, Maskell J, Kemp T, et al.

Incidence and risk factors for food hypersensitivity in UK infants: results

from a birth cohort study. Clin Transl Allergy 2015;6:1.

[4] Simon D, Cianferoni A, Spergel JM, Aceves S, Holbreich M, Venter C, et al.

Eosinophilic esophagitis is characterized by a non-IgE-mediated food

hypersensitivity. Allergy 2016;71:611–20.

[5] Dellon ES. Epidemiology of eosinophilic esophagitis. Gastroenterol Clin

North Am 2014;43:201–18.

[6] Liacouras CA, Furuta GT, Hirano I, Atkins D, Attwood SE, Bonis PA, et al.

Eosinophilic esophagitis: updated consensus recommendations for

children and adults. J Allergy Clin Immunol 2011;128:3–20.


http://refhub.elsevier.com/S1740-6757(16)30007-X/sbref0290



















[7] Spergel JM. An allergist’s perspective to the evaluation of eosinophilic

esophagitis. Best Pract Res Clin Gastroenterol 2015;29:771–81.

[8] Furuta GT, Katzka DA. Eosinophilic esophagitis. N Engl J Med

2015;373:1640–8.

[9] Cianferoni A, Spergel J. Eosinophilic esophagitis: a comprehensive review.

Clin Rev Allergy Immunol 2016;50:159–74.

[10] Lucendo AJ, Arias A, Redondo-Gonzalez O, Gonzalez-Cervera J. Seasonal

distribution of initial diagnosis and clinical recrudescence of eosinophilic

esophagitis: a systematic review and meta-analysis. Allergy 2015;70:

1640–50.

[11] Kelly KJ, Lazenby AJ, Rowe PC, Yardley JH, Perman JA, Sampson HA.

Eosinophilic esophagitis attributed to gastroesophageal reflux:

improvement with an amino acid-based formula. Gastroenterology

1995;109:1503–12.

[12] Spergel JM, Brown-Whitehorn TF, Cianferoni A, Shuker M, Wang ML,

Verma R, et al. Identification of causative foods in children with

eosinophilic esophagitis treated with an elimination diet. J Allergy Clin

Immunol 2012;130:461–7.

[13] Greenhawt M, Aceves SS, Spergel JM, Rothenberg ME. The management of

eosinophilic esophagitis. J Allergy Clin Immunol Pract 2013;1:332–40.

[14] Rajan J, Newbury RO, Anilkumar A, Dohil R, Broide DH, Aceves SS. Long-

term assessment of esophageal remodeling in patients with pediatric

eosinophilic esophagitis treated with topical corticosteroids. J Allergy Clin

Immunol 2016;137:147–56.

[15] Hill DA, Spergel JM. The immunologic mechanisms of eosinophilic

esophagitis. Curr Allergy Asthma Rep 2016;16:9.

[16] Sherrill JD, Rothenberg ME. Genetic and epigenetic underpinnings of

eosinophilic esophagitis. Gastroenterol Clin North Am 2014;43:269–80.

[17] Leung J, Beukema KR, Shen AH. Allergic mechanisms of eosinophilic

oesophagitis. Best Pract Res Clin Gastroenterol 2015;29:709–20.

[18] Sherrill JD, Kc K, Wu D, Djukic Z, Caldwell JM, Stucke EM, et al.

Desmoglein-1 regulates esophageal epithelial barrier function and

immune responses in eosinophilic esophagitis. Mucosal Immunol

2014;7:718–29.

[19] Noti M, Wojno ED, Kim BS, Siracusa MC, Giacomin PR, Nair MG, et al.

Thymic stromal lymphopoietin-elicited basophil responses promote

eosinophilic esophagitis. Nat Med 2013;19:1005–13.

[20] Judd LM, Heine RG, Menheniott TR, Buzzelli J, O’Brien-Simpson N, Pavlic

D, et al. Elevated IL-33 expression is associated with pediatric eosinophilic

esophagitis, and exogenous IL-33 promotes eosinophilic esophagitis

development in mice. Am J Physiol Gastrointest Liver Physiol

2016;310:G13–25.

[21] Doherty TA, Baum R, Newbury RO, Yang T, Dohil R, Aquino M, et al.

Group 2 innate lymphocytes (ILC2) are enriched in active eosinophilic

esophagitis. J Allergy Clin Immunol 2015;136:792–4.

[22] Collison AM, Sokulsky LA, Sherrill JD, Nightingale S, Hatchwell L, Talley

NJ, et al. TNF-related apoptosis-inducing ligand (TRAIL) regulates

midline-1, thymic stromal lymphopoietin, inflammation, and

remodeling in experimental eosinophilic esophagitis. J Allergy Clin

Immunol 2015;136:971–82.

[23] Lim E, Rothenberg ME. Demethylation of the human eotaxin-3 gene

promoter leads to the elevated expression of eotaxin-3. J Immunol

2014;192:466–74.

[24] Krishnamurthy P, Sherrill JD, Parashette K, Goenka S, Rothenberg ME,

Gupta S, et al. Correlation of increased PARP14 and CCL26 expression in

biopsies from children with eosinophilic esophagitis. J Allergy Clin

Immunol 2014;133:577–80.

[25] Jyonouchi S, Smith CL, Saretta F, Abraham V, Ruymann KR, Modayur-

Chandramouleeswaran P, et al. Invariant natural killer T cells in children

with eosinophilic esophagitis. Clin Exp Allergy 2014;44:58–68.

[26] Arias A, Lucendo AJ, Martinez-Fernandez P, Gonzalez-Castro AM, Fortea

M, Gonzalez-Cervera J, et al. Dietary treatment modulates mast cell

phenotype, density, and activity in adult eosinophilic oesophagitis. Clin

Exp Allergy 2016;46:78–91.

[27] Merves J, Chandramouleeswaran PM, Benitez AJ, Muir AB, Lee AJ, Lim

DM, et al. Altered esophageal histamine receptor expression in

eosinophilic esophagitis (EoE): implications on disease pathogenesis.

PLOS ONE 2015;10:e0114831–114840.


[28] Blanchard C, Stucke EM, Rodriguez-Jimenez B, Burwinkel K, Collins MH,

Ahrens A, et al. A striking local esophageal cytokine expression profile in

eosinophilic esophagitis. J Allergy Clin Immunol 2011;127:208–17.

[29] Cheng E, Zhang X, Huo X, Yu C, Zhang Q, Wang DH, et al. Omeprazole

blocks eotaxin-3 expression by oesophageal squamous cells from patients

with eosinophilic oesophagitis and GORD. Gut 2013;62:824–32.

[30] Rieder F, Nonevski I, Ma J, Ouyang Z, West G, Protheroe C, et al. T-helper 2

cytokines, transforming growth factor beta1, and eosinophil products

induce fibrogenesis and alter muscle motility in patients with

eosinophilic esophagitis. Gastroenterology 2014;146:1266–77.

[31] Namjou B, Marsolo K, Caroll RJ, Denny JC, Ritchie MD, Verma SS, et al.

Phenome-wide association study (PheWAS) in EMR-linked pediatric

cohorts, genetically links PLCL1 to speech language development and

IL5-IL13 to Eosinophilic Esophagitis. Front Genet 2014;5:401–10.

[32] Blanchard C, Mingler MK, Vicario M, Abonia JP, Wu YY, Lu TX, et al. IL-13

involvement in eosinophilic esophagitis: transcriptome analysis and

reversibility with glucocorticoids. J Allergy Clin Immunol 2007;120:

1292–300.

[33] Kinoshita Y, Furuta K, Ishimura N, Ishihara S. Elevated plasma cytokines

in Japanese patients with eosinophilic esophagitis and gastroenteritis.

Digestion 2012;86:238–43.

[34] Assa’ad AH, Gupta SK, Collins MH, Thomson M, Heath AT, Smith DA,

et al. An antibody against IL-5 reduces numbers of esophageal

intraepithelial eosinophils in children with eosinophilic esophagitis.


[35] Stuck MC, Straumann A, Simon HU. Relative lack of T regulatory cells in

adult eosinophilic esophagitis - no normalization after corticosteroid

therapy. Allergy 2011;66:705–7.

[36] Fuentebella J, Patel A, Nguyen T, Sanjanwala B, Berquist W, Kerner JA,

et al. Increased number of regulatory T cells in children with eosinophilic

esophagitis. J Pediatr Gastroenterol Nutr 2010;51:283–9.

[37] Zhu X, Wang M, Mavi P, Rayapudi M, Pandey AK, Kaul A, et al.

Interleukin-15 expression is increased in human eosinophilic esophagitis

and mediates pathogenesis in mice. Gastroenterology 2010;139:182–93.

[38] Niranjan R, Rajavelu P, Ventateshaiah SU, Shukla JS, Zaidi A, Mariswamy

SJ, et al. Involvement of interleukin-18 in the pathogenesis of human

eosinophilic esophagitis. Clin Immunol 2015;157:103–13.

[39] Frischmeyer-Guerrerio PA, Guerrerio AL, Oswald G, Chichester K, Myers

L, Halushka MK, et al. TGFbeta receptor mutations impose a strong

predisposition for human allergic disease. Sci Transl Med 2013;5:

195ra94.

[40] Mishra A, Hogan SP, Brandt EB, Rothenberg ME. An etiological role for

aeroallergens and eosinophils in experimental esophagitis. J Clin Invest

2001;107:83–90.

[41] Mishra A, Rothenberg ME. Intratracheal IL-13 induces eosinophilic

esophagitis by an IL-5, eotaxin-1, and STAT6-dependent mechanism.


[42] Mishra A, Schlotman J, Wang M, Rothenberg ME. Critical role for adaptive

T cell immunity in experimental eosinophilic esophagitis in mice. J

Leukoc Biol 2007;81:916–24.

[43] Rubinstein E, Cho JY, Rosenthal P, Chao J, Miller M, Pham A, et al. Siglec-F

inhibition reduces esophageal eosinophilia and angiogenesis in a mouse

model of eosinophilic esophagitis. J Pediatr Gastroenterol Nutr

2011;53:409–16.

[44] Cho JY, Doshi A, Rosenthal P, Beppu A, Miller M, Aceves S, et al. Smad3-

deficient mice have reduced esophageal fibrosis and angiogenesis in a

model of egg-induced eosinophilic esophagitis. J Pediatr Gastroenterol

Nutr 2014;59:10–6.

[45] Rajavelu P, Rayapudi M, Moffitt M, Mishra A, Mishra A. Significance of

para-esophageal lymph nodes in food or aeroallergen-induced iNKT cell-

mediated experimental eosinophilic esophagitis. Am J Physiol

Gastrointest Liver Physiol 2012;302:G645–54.

[46] Mondoulet L, Dioszeghy V, Larcher T, Ligouis M, Dhelft V, Puteaux E,

et al. Epicutaneous immunotherapy (EPIT) blocks the allergic esophago-

gastro-enteropathy induced by sustained oral exposure to peanuts in

sensitized mice. PLoS ONE 2012;7:e31967.

[47] Dioszeghy V, Mondoulet L, Dhelft V, Ligouis M, Puteaux E, Dupont C,

et al. The regulatory T cells induction by epicutaneous immunotherapy is










































































































































sustained and mediates long-term protection from eosinophilic disorders

in peanut-sensitized mice. Clin Exp Allergy 2014;44:867–81.

[48] Akei HS, Mishra A, Blanchard C, Rothenberg ME. Epicutaneous antigen

exposure primes for experimental eosinophilic esophagitis in mice.


[49] Alex A, Noti M, Wojno ED, Artis D, Zhou C. Characterization of

eosinophilic esophagitis murine models using optical coherence

tomography. Biomed Opt Express 2014;5:609–20.

[50] Noti M, Kim BS, Siracusa MC, Rak GD, Kubo M, Moghaddam AE, et al.

Exposure to food allergens through inflamed skin promotes intestinal

food allergy through the thymic stromal lymphopoietin-basophil axis. J


[51] Nowak-Wegrzyn A, Katz Y, Mehr SS, Koletzko S. Non-IgE-mediated

gastrointestinal food allergy. J Allergy Clin Immunol 2015;135:

1114–24.

[52] Berin MC. Immunopathophysiology of food protein-induced

enterocolitis syndrome. J Allergy Clin Immunol 2015;135:1108–13.

[53] Nowak-Wegrzyn A. Food protein-induced enterocolitis syndrome and

allergic proctocolitis. Allergy Asthma Proc 2015;36:172–84.

[54] Chung HL, Hwang JB, Park JJ, Kim SG. Expression of transforming growth

factor beta1, transforming growth factor type I and II receptors, and TNF-

alpha in the mucosa of the small intestine in infants with food protein-

induced enterocolitis syndrome. J Allergy Clin Immunol 2002;109:150–4.

[55] Feuille E, Nowak-Wegrzyn A. Food protein-induced enterocolitis

syndrome, allergic proctocolitis, and enteropathy. Curr Allergy Asthma

Rep 2015;15:50–60.

[56] Ohtsuka Y, Jimbo K, Inage E, Mori M, Yamakawa Y, Aoyagi Y, et al.

Microarray analysis of mucosal biopsy specimens in neonates with rectal

bleeding: is it really an allergic disease? J Allergy Clin Immunol

2012;129:1676–8.

[57] Kumagai H, Maisawa S, Tanaka M, Takahashi M, Takasago Y, Nishijima A,

et al. Intestinal microbiota and secretory immunoglobulin A in feces of

exclusively breast-fed infants with blood-streaked stools. Microbiol

Immunol 2012;56:657–63.




































DRUG DISCOVERY

TODAY

DISEASEMODELS

Experimental food allergy models tostudy the role of innate immune cells asinitiators of allergen specific Th2immune responsesMaryam Hussain1, Michelle M. Epstein2, Mario Noti1,*1Institute of Pathology, Division of Experimental Pathology, University of Bern, Bern 3010, Switzerland2Medical University of Vienna, Department of Dermatology, Division of Immunology, Waehringer Guertel 18-20, A-1090 Vienna, Austria


Editors-in-Chief




Although our knowledge of the pathophysiology of food

allergies has significantly improved over the last years,

a more comprehensive understanding of basic immune

mechanisms driving disease pathogenesis is important

to develop new intervention strategies. The recent

development of animal model systems recapitulating

features of clinical food allergy provides an essential

tool to study the immunology of IgE-mediated food

allergies. While immunological effector responses have

been well documented, how food allergic immune

responses are initiated is not well defined. In this short

review, we discuss the use of experimental mouse

models both to study the role of innate immune cell

populations in promoting disease and to test new

treatment regimens that may prevent the onset of

IgE-mediated food allergies.

Introduction

Food allergy is an adverse type-2-immune cell driven allergic

response that occurs reproducibly on exposure to a given

food. As the public health and economic impact of food

*Corresponding author: M. Noti ([email protected])



allergies continues to grow, there is an urgent need to develop

new intervention strategies to prevent and treat this disease.

Despite their often criticized limitation to accurately mimic

human pathophysiology and to predict treatment efficacy,

experimental animal models have significantly contributed

to a better understanding of the immunology of food allergy.

The purpose of this review is to summarize the latest devel-

opments in the field of innate immune cells as initiators of

food allergic responses. Furthermore, we will discuss poten-

tial new therapeutic modalities targeting innate immune cell

populations which have emerged from experimental food

allergy models and hold promise for future clinical studies.

Immunology of food allergy

Food allergies are characterized as adverse immune reactions to

food proteins that affect up to 6% of children and 3–4% of

adults [1]. Despite food allergies represent a growing clinical

6.08.001 55







problem, disease etiology remains largely unknown. While

genetic predisposition is a significant risk factor for the devel-

opment of food allergies, the large increase in food allergies

over the last two decades suggests that genetic predisposition

alone cannot account for the observed phenomenon. Emerg-

ing evidence suggests that changes in lifestyle (e.g. diet, in-

creased vaccination rate, antibiotics, changes in microbiome)

modify innate and adaptive immunity which results in sus-

ceptibility to allergic sensitization to foods (reviewed in [2]).

Allergic responses to foods encompass a range of disorders

from IgE-mediated food allergies to delayed cell-mediated

reactions (also referred to as non-IgE food allergies) affecting

the gastrointestinal tract, airways or skin. Throughout this

review, we will focus on IgE-mediated food allergies. Food

allergies are characterized by uncontrolled type-2 mediated

immune responses that occur reproducibly on exposure to a

given food. Both arms of the host’s immune system, the

innate- and the adaptive immune system contribute to disease

pathogenesis. Cells of the innate immune system (mast cells,

granulocytes, mononuclear phagocytes, innate lymphoid

cells) are located at the interface between the external envi-

ronment and the internal adaptive immune system. In con-

trast to the innate immune system, adaptive immune cells are

able to generate allergen-specific receptor molecules and im-

munological memory. Thus, in response to allergen re-expo-

sure, cells of the adaptive immune system are able to mount a

memory immune response against the same allergen.

A key feature of the allergic cascade is the polarization of

allergen specific T helper (Th)2 cells. Th2 cells represent impor-

tant sources of pro-allergic cytokines and regulate B cell class

switching to IgE through production of IL-4, recruit eosino-

phils through IL-5 or mast cells through IL-4/IL-9 signaling

resulting in tissue eosinophilia and mast cell hyperplasia. B cell

derived allergen-specific IgE is dispersed systemically and binds

to its high affinity receptor FceRI on tissue resident mast cells

and circulating basophils resulting in allergen sensitization.

Upon allergen re-exposure, IgE cross-linking initiates degran-

ulation of mast cells and basophils which release a number of

pro-allergic factors including pre-formed- or newly generated

granule mediators, chemokines or cytokines causing smooth

muscle contraction, vascular permeabilization and further re-

cruitment of immune cells to sites of inflammation [3].

To better understand the immunological processes under-

lying the pathogenesis of food allergy, it is important to

understand how cells of the innate immune system regulate

adaptive immune responses to food allergens. IL-4 plays a

critical role in the polarization of Th2 cells by regulating

STAT6-mediated expression of GATA3, the master regulator

of Th2 differentiation. Given the importance of IL-4 on the

polarization of Th2 cells, IgE synthesis and mucosal mast cell

expansion in the development of experimental IgE-mediated

food allergy, identifying the initial source(s) of IL-4 is key for

a better understanding on how food allergen-specific Th2


immune responses are initiated. As naıve T cells are poor

producers of IL-4 and IL-4 is important for optimal Th2

polarization in most experimental settings, this raises the

chicken-and-egg question of the cellular origin of IL-4. Recent

studies using IL-4 reporter mice in models of helminth infec-

tion or allergic inflammation have highlighted numerous IL-

4 competent innate immune cells that actively contribute to

optimal Th2 polarization [4]. Emerging literature further

suggests that epithelial cells play a fundamental role in the

recruitment of IL-4 competent innate immune cells to sites of

epithelial stress through secretion of IL-25, IL-33 and thymic

stromal lymphopoietin (TSLP) [5,6].

Here, we address how the use of different sensitization

protocols in experimental murine food allergy models can

work in synergy with human studies to investigate the role

of innate immune cell populations as initiators of food allergic

responses. Further, we discuss potential new treatment proto-

cols that may interfere with the recruitment and activation of

innate immune cell populations in the context of food allergy.

Experimental models of food allergy

Epicutaneous sensitization protocols

Studies in humans and mice have demonstrated that epicu-

taneous food allergen sensitization represents a significant

risk factor for the development of food allergy, likely by

bypassing the induction of oral tolerance [7]. Epicutaneous

food allergen sensitization models often rely on physical

impairment of the skin barrier induced by physical damage,

chemical-induced damage or genetic manipulation resulting

in local tissue inflammation. Physical damage of the skin

epithelium can be induced by repeated tape-stripping of skin

allowing for food allergen sensitization on a compromised

skin barrier causing food allergy upon oral or systemic aller-

gen challenge [8]. Chemicals used to promote epicutaneous

food allergen sensitization include topical treatment with

trinitrobenzene sulfonic acid (TNBS) [9], sodium dodecyl

sulfate (SDS) [10] or calcipotriol (MC903), a vitamin D ana-

logue that is widely used to induce atopic dermatitis (AD)-like

skin lesions in mice. Because AD is a risk factor for the onset of

food allergy and asthma in humans, recent studies made use

of the MC903-mediated food allergen sensitization protocol

to induce food allergy or allergic airway inflammation in mice

[11,12]. Alongside these methods many genetic approaches

have been made to develop models with skin barrier defects,

such as mice with defects in skin matrix proteins, for example,

fillaggrin. Importantly, mutations in the filaggrin gene have

been strongly associated with the pathogenesis of AD and food

allergy in humans and mice [13]. Other genetic models allow-

ing for epicutaneous food allergen sensitization include skin-

specific over-expression of TSLP in keratinocytes resulting in

gastrointestinal food allergy or allergic lung inflammation

upon re-exposure of food allergens via the gastrointestinal

tract or the airways, respectively [14]. Together, epicutaneous


food allergen sensitization protocols represent valuable tools

to study immune cell functions at distinct physical sites; the

skin as site of allergen sensitization and the gastrointestinal

tract or the airways as site of allergen challenge.

Recruitment of innate immune cell populations to sites of

epicutaneous food allergen sensitization

Disruption of skin barrier function induced by mechanical,

chemical or genetic manipulation is often required for opti-

mal epicutaneous food allergen sensitization. In response to

stress, skin epithelial cells secrete a number of cytokines,

including IL-25, IL-33 and TSLP all of which have been

implicated in promoting Th2 cytokine responses in vivo

through attraction or stimulation of innate immune cells

[15]. Signaling between skin epithelial cells and innate im-

mune cells via TSLP and IL-33 has been implicated in the

pathogenesis of AD in humans and mice [16]. Recent studies

highlighted that epicutaneous sensitization to food aller-

gens on a developing AD-like skin lesion resulted in rapid

infiltration of TSLP-elicited basophils that were both neces-

sary and sufficient for the development of allergic responses

to food in the gastrointestinal tract. Antibody-mediated

depletion or genetic manipulation of TSLP-elicited basophils

led to a significant reduction in Th2 polarization and aller-

gen-specific IgE synthesis [12]. As basophils are potent pro-

ducers of IL-4 in response to FceRI cross-linking or TSLP

signaling, basophil-mediated Th2 polarization and associat-

ed development of experimental IgE-mediated food allergy

may be regulated through basophil intrinsic IL-4 production

[17]. Further studies revealed that basophils and group 2

innate lymphoid cells (ILC2) accumulate in close proximity

to each other in AD lesional skin of humans and mice. In

these settings, basophil-derived IL-4 was shown to promote

proliferation of ILC2s promoting AD-like skin inflammation

[18]. Despite their disease promoting role in AD, whether

skin ILC2s contribute to food allergen sensitization remains

to be determined.

While basophils are known to act as Th2-inducing antigen

presenting cells (APCs) and are required for Th2 polarization

in vitro and in vivo [19], TSLP-primed dendritic cells (DCs) play

a critical role in the differentiation of Th2 cells [20]. Given

their strategic resident location in the skin, immature DCs

may be important for internalization of food allergens and

the presentation of processed food allergens to naıve T cells.

Recent studies by Leyva-Castillo and colleagues demonstrat-

ed that optimal Th2 polarization is dependent on an orches-

trated immune cascade in a model of AD. TSLP-activated DCs,

through OX40L signaling, prime naıve CD4T cells to produce

IL-3 resulting in basophil recruitment and Th2 differentiation

[21]. In addition to the above described cross-talk between

basophils-DCs and T cells, basophils influence localized eo-

sinophil recruitment, another IL-4 competent innate cell

population, in a model of IgE-dependent eosinophilic skin

inflammation [22]. Furthermore, targeting basophil responses

in food-induced allergic inflammation resulted in a significant

reduction of eosinophils to sites of allergen sensitization and

challenge [23]. As eosinophils are capable of producing IL-4

and IL-13, it is likely that eosinophils contribute not only to

local tissue inflammation, but also to Th2 polarization. Fur-

ther studies are necessary to determine a potential role for

eosinophils in epicutaneous food allergen sensitization. In-

nate immune cell pathways contributing to Th2 polarization

in response to epicutaneous food allergen sensitization are

illustrated in Fig. 1.

Innate immune cells as initiators of food allergic responses in oral

sensitization protocols

In contrast to allergen sensitization via epicutaneous routes,

ingested food antigens are subject to denaturation and deg-

radation in the digestive tract resulting in either immuno-

logical ignorance or induction of oral tolerance. Failure to

induce tolerance to food proteins can result in the develop-

ment of celiac disease or food allergies. The cellular and

molecular events involved in the breakdown of oral toler-

ance, food allergen sensitization and the development of

food allergies are incompletely understood (reviewed in

[15]). Studies in mice demonstrated that oral administration

of potent mucosal adjuvants, for example, cholera toxin (CT)

or staphylococcus aureus enterotoxin B (SEB) together with

food antigens is sufficient to overcome oral tolerance, pro-

mote food allergen sensitization and the development of IgE-

mediated food allergy. Co-administration of allergens togeth-

er with CT induces the production of antigen-specific IgE

promoting anaphylactic responses in response to intra-gastric

food allergen challenge of sensitized mice [24]. Mechanisms

underlying food allergen sensitization in this model system

rely on the up-regulation of co-stimulatory molecules OX40L

[25] and TIM-4 [26] on intestinal DCs resulting in enhanced

migration of DCs from the lamina propria to mesenteric

lymph nodes where matured DCs present captured food

antigens to naıve T cells to induce Th2 polarization [27].

Importantly, experimental manipulation of these Th2 polar-

izing co-stimulatory molecules has been shown to reduce

Th2-associated food allergic responses in mice, suggesting

their importance in food allergen sensitization [28]. A recent

study using CT-mediated allergen sensitization to peanut

highlighted that IL-33, but not TSLP or IL-25 promotes up-

regulation of OX40L on DCs. IL-33, which is predominantly

produced by epithelial cells, is known to increase mucosal

permeability and promote Th2 skewing by attracting IL-4

competent innate immune cells [29,30]. Among these, ILC2s

– although poor producers of IL-4 – are major targets of IL-33

in the gastrointestinal tract. Despite their pathological role in

models of allergic inflammation, Th2 priming under the

control of OX40L-OX40 interactions in the CT food allergy

model was independent of ILC2s [31]. These data suggest that



Stressor: - Mechanic- Chemical- Genetic

food allergens

immatureDC

Basophil

compromisedskin barrier

Eosinophil

IL-4

IL-5

IL-4 IL-4

IL-4?

IL-4IL-13

B cell

Th2

Th0

OX40Lmature

DC

Bone marrow Skin draining lymph node

IgE

IL-5IL-13

? ?

ILC2

Mast cell

IL-33IL-33

TSLP


Figure 1. Contribution of innate immune cells to Th2 polarization in epicutaneous food allergen sensitization protocols. Skin epithelial cells secrete TSLP

and IL-33 following skin barrier impairment resulting in the recruitment and/or activation of innate immune cell populations. In response to TSLP, basophils

are recruited rapidly from the bone marrow to sites of epithelial damage or stress. Basophil-derived IL-4 production promotes the accumulation of ILC2 to

the skin that release significant amounts of IL-5 and IL-13 further attracting other immune cells such as eosinophils. Allergen uptake by DCs induces their

maturation and migration to skin draining lymph nodes, where they present processed allergen epitopes to cognate T cells. Interaction of basophils with

DCs induces the expression of the Th2 priming co-stimulatory molecules OX40L and TIM-4 on DCs, likely in an IL-4 dependent manner. In the presence of

IL-4 or IL-13 derived from basophils, eosinophils, ILC2s or tissue resident mast cells, naıve T cells differentiate into effector Th2 cells. Th2 cells promote IgE

isotype switching in B cells resulting in the secretion of allergen-specific IgE. Allergen-induced cross-linking of IgE on FceRI expressed on basophils and mast

cells induces cell degranulation and the release of pro-inflammatory mediators further amplifying the allergic cascade.

other IL-4 competent cells such as tissue resident mast cells or

eosinophils may be involved in OX40L-mediated polariza-

tion of Th2 cells. Studies by Chu et al. using intra-gastric

immunization to the common food allergen peanut with the

classical oral Th2-inducing adjuvant CT demonstrated that

indigenous enteric eosinophils control OX40L expression on

CD103+ DCs by means of secretion of the eosinophil-specific

granule protein eosinophil peroxidase (EPO). In this model,

eosinophil deficient mice were protected from Th2-mediated

food allergy and anaphylaxis while Th2 polarization was

restored by transfer of IL-4 sufficient or deficient eosinophils

into eosinophil deficient hosts [32].


Other mouse models relying on oral allergen sensitization

make use of enterotoxin B from Staphylococcus aureus (SEB)

[33]. S. aureus is a common organism colonising the airways,

and therefore, exposure to S. aureus derived super-antigens

may represent a physiologically relevant factor for allergic

sensitization to food proteins. SEB co-administered with food

proteins applied via the oral route resulted in Th2-mediated

IgE-dependent food allergy upon oral re-exposure of sensi-

tized animals [34]. Similar to the CT model, SEB treatment

resulted in the up-regulation of TIM-4 on intestinal DCs, and

blockade of TIM-4 inhibited food-induced allergic responses

[35]. Together, while DCs and indigenous enteric eosinophils


can contribute to mucosal food allergen sensitization in the

CT food allergy model, the role of indigenous enteric eosin-

ophils in allergen sensitization in the SEB model remains to

be determined (summarized in Fig. 2). In contrast, ILC2

responses may not contribute to IgE-mediated food allergy

in animals sensitized via the oral route [31], but have recently

been shown to be critical mediators of IgE-mediated food

allergy in mice systemically sensitized with alum before oral

allergen feeding [36]. Thus, depending on the route of food

Mucosal adjuvants: - CT- SEB

immatureCD103+ DC

maturationmigration

indigenous entericeosinophil

IL-4?

Th2

Th0

OX40LTIM4

EPO

matureCD103+ DC

Mesenteric

Figure 2. Contribution of innate immune cells to Th2 polarization in oral food

Staphylococcus aureus (SEB) are potent mucosal adjuvants and their detoxified der

models, oral administration of CT or SEB together with food proteins induces OX

indigenous enteric eosinophils leads to degranulation of eosinophil peroxidase (E

to mesenteric lymph nodes where they present allergen epitopes to naıve T ce

regulation of Th2 priming co-stimulatory receptors on CD103+ DCs likely thr

immune cells such as eosinophils or mast cells. In response to systemic allergen

resulting in secretion of prodigious amounts of IL-9 and pro-inflammatory media

allergen sensitization protocols.

allergen sensitization, different innate immune cells are re-

quired across various experimental systems to generate food

allergen specific adaptive immune responses.

Genetic food allergy models to study innate immune cell functions

In addition to epicutaneous, oral or systemic food allergen

sensitization protocols, recently established models of food

allergy include genetic manipulation of key type-2 cytokines

or their corresponding receptors. Enteral exposure to food

food allergens

IL-4?

IL-33

IL-4?

IL-9

IgE

IL-4IL-13

B Cell

Mast cell

lymph nodeDrug Discovery Today: Disease Models

allergen exposure protocols. Cholera toxin (CT) or enterotoxin B from

ivatives are important for the development of mucosal vaccines. In animal

40L- and IL-4-dependent Th2 priming in the small intestine. Activation of

PO) that promotes the activation of CD103+ DCs and their mobilization

lls resulting in Th2 differentiation. Epithelial-derived IL-33 promotes up-

ough recruitment or activation of IL-4 competent tissue resident innate

sensitization, IL-33 has been demonstrated to activate mucosal mast cells

tors upon oral allergen challenge, a scenario that is also likely in oral food



allergens in mice harboring activating mutations of the IL-

4 receptor a-chain (Il4raF709) induced strong allergen-spe-

cific IgE responses and intestinal mastocytosis resulting in

systemic anaphylactic responses upon allergen challenge

[36–38]. Other models use transgenic overexpression of IL-

9 predisposing mice for oral food allergen sensitization and

the development of IgE-mediated intestinal anaphylaxis.

The induction of food allergic responses in this model

system required intestinal mast cells promoting increased

intestinal permeability. Furthermore, overexpression of

IL-9 in the intestine promoted local allergen-specific Th2

responses upon intra-gastric allergen feeding [39]. Another

recently established genetic model of food allergy relies on

the overexpression of IL-25 in the small intestinal epithe-

lium (iL25Tg). Repeated oral sensitization of iL25Tg mice

with the model food allergen ovalbumin was sufficient

to promote symptomatic features of IgE-mediated food

allergy including diarrhea, hypothermia, intestinal masto-

cytosis and increased serum allergen-specific IgE levels. In

this model system, IL-25 responsive ILC2 promoted sus-

ceptibility to experimental food allergy [36]. Collectively,

genetic food allergy models represent valuable tools to

assess multifactorial functions of innate immune cell popu-

lations in the pathogenesis IgE-mediated food allergic

responses.

Potential new therapeutic targets arising from studies in

experimental food allergy models

Currently, there is no cure for food allergies and available

strategies to prevent or block food allergic responses include

strict allergen avoidance or injection of epinephrine in

emergency situations. Therefore, developing novel or im-

proved therapeutic strategies is an active area of food allergy

research. The recent use of pre-clinical experimental food

allergy models led to the discovery of several innate immune

cell pathways that promote the pathogenesis of food allergy.

As a result prevention and treatment strategies have been

successfully tested in animal food allergy models with prom-

ising results in clinical trials, including allergen-nonspecific-

and allergen-specific therapeutic approaches (reviewed in

[2]). For example, TSLP – a cytokine that is predominantly

produced by innate immune cells – represents a promising

new therapeutic target for preventing the onset of food

allergies [40]. Given the importance of TSLP for AD patho-

genesis in animal models [11] and AD representing a signifi-

cant risk factor for the development of food allergies [41],

targeting TSLP-TSLP-receptor interactions in AD patients

may limit food allergen sensitization on impaired barrier

skin and thus, prevent the progression to food allergy

later in life. Importantly, blocking TSLP signaling in asth-

matic patients attenuated allergen-induced asthmatic

responses, highlighting its potential clinical value for the

treatment of allergic inflammatory disorders [40]. Other


epithelial-derived cytokines including IL-25 or IL-33 may

represent potential new therapeutic targets for prevention or

treatment of food allergies. In addition to interrupting the

secretion of epithelial derived type-2 cytokines, targeting IL-

4 signaling may represent a promising new therapeutic

approach in the treatment of food allergies given its impor-

tance in Th2 polarization and the pathogenesis of IgE-medi-

ated experimental food allergy. Recent clinical trials using

dupilomab, an IL-4 receptor alpha blocking antibody

revealed significant efficacy and safety for the treatment of

AD [42]. Targeting IL-4 receptor signaling may not only show

efficacy in patients with moderate to severe AD but may also

limit food allergen sensitization on a compromised skin

barrier.

While the suppression of immune responses is a common

therapeutic strategy applied to various inflammatory disor-

ders including allergic inflammation, there is rarely a benefit

without potential harm. All biological targets discussed above

actively interact with cellular and molecular innate immune

cell functions that are important to maintain tissue homeo-

stasis or promote tissue repair in the healthy host. A future

challenge will be to determine the optimal therapeutic strat-

egy (e.g. dosage, single or combinatorial treatment protocols)

for the individual patient.

Conclusions

Experimental mouse models of food allergy significantly

contributed to a better understanding of disease pathogene-

sis, validation of existing therapeutics and development of

new treatment strategies. The use of these model systems

highlighted a central role for various innate immune cell

populations including basophils, mast cells, eosinophils,

ILC2, or dendritic cells as initiators and amplifiers of patho-

logic allergen specific Th2 responses. While targeting the

above discussed innate pathways of allergic inflammation

show promising results in preventing or ameliorating disease

in animal models, ongoing and future clinical trials will have

to demonstrate the efficacy of such intervention strategies in

food allergic patients. Together, despite experimental food

allergy models do not completely mimic human pathophysi-

ology, they have repeatedly demonstrated their utility in

translational discoveries. A future challenge using animal

models of food allergy will be the establishment of validated

and predictive preclinical models to translate findings from

bench to bedside [43].


The authors declare no conflict of interest.

Acknowledgments

We apologize to our colleagues whose work could not be

cited due to space restrictions. M.N. and M.E. are members of

the COST Action FA1402 entitled: Improving Allergy Risk


Assessment Strategy for New Food Proteins (ImpARAS). This

work was further supported by funding from the Swiss Na-

tional Science Foundation (PZ00P3-136486 to M.N.) and the

Olga Mayenfisch Foundation to M.N.

References[1] Burks AW, Tang M, Sicherer S, Muraro A, Eigenmann PA, Ebisawa M, et al.

ICON: food allergy. J Allergy Clin Immunol 2012;129(April (4)):

906–20.

[2] Sicherer SH, Sampson HA. Food allergy. J Allergy Clin Immunol

2010;125(February (Suppl. 2)):S116–25.

[3] Chinthrajah RS, Hernandez JD, Boyd SD, Galli SJ, Nadeau KC. Molecular

and cellular mechanisms of food allergy and food tolerance. J Allergy Clin

Immunol 2016;137(April (4)):984–97.

[4] Paul WE, Zhu J. How are T(H)2-type immune responses initiated and

amplified? Nat Rev Immunol 2010;10(April (4)):225–35.

[5] Saenz SA, Noti M, Artis D. Innate immune cell populations function as

initiators and effectors in Th2 cytokine responses. Trends Immunol

2010;31(November (11)):407–13.

[6] Hammad H, Lambrecht BN. Barrier epithelial cells and the control of type

2 immunity. Immunity 2015;43(July (1)):29–40.

[7] Lack G. Update on risk factors for food allergy. J Allergy Clin Immunol

2012;129(May (5)):1187–97.

[8] Strid J, Hourihane J, Kimber I, Callard R, Strobel S. Disruption of the

stratum corneum allows potent epicutaneous immunization with protein

antigens resulting in a dominant systemic Th2 response. Eur J Immunol

2004;34(August (8)):2100–9.

[9] Liu Z-Q, Yang P-C. Hapten may play an important role in food allergen-

related intestinal immune inflammation. N Am J Med Sci 2011;3(March

(3)):103–6.

[10] Terada M, Tsutsui H, Imai Y, Yasuda K, Mizutani H, Yamanishi K, et al.

Contribution of IL-18 to atopic-dermatitis-like skin inflammation

induced by Staphylococcus aureus product in mice. Proc Natl Acad Sci U S A

2006;103(June (23)):8816–21.

[11] Li M, Hener P, Zhang Z, Kato S, Metzger D, Chambon P. Topical vitamin

D3 and low-calcemic analogs induce thymic stromal lymphopoietin in

mouse keratinocytes and trigger an atopic dermatitis. Proc Natl Acad Sci U

S A 2006;103(August (31)):11736–41.

[12] Noti M, Kim BS, Siracusa MC, Rak GD, Kubo M, Moghaddam AE, et al.

Exposure to food allergens through inflamed skin promotes intestinal

food allergy through the thymic stromal lymphopoietin-basophil axis. J

Allergy Clin Immunol 2014;133(May (5)):1390–9. 1399.e1–6.

[13] Fallon PG, Sasaki T, Sandilands A, Campbell LE, Saunders SP, Mangan NE,

et al. A homozygous frameshift mutation in the mouse Flg gene facilitates

enhanced percutaneous allergen priming. Nat Genet 2009;41(May

(5)):602–8.

[14] Han H, Thelen TD, Comeau MR, Ziegler SF. Thymic stromal

lymphopoietin-mediated epicutaneous inflammation promotes acute

diarrhea and anaphylaxis. J Clin Invest 2014;124(December (12)):

5442–52.

[15] Pabst O, Mowat AM. Oral tolerance to food protein. Mucosal Immunol

2012;5(May (3)):232–9.

[16] Brandt EB, Sivaprasad U. Th2 cytokines and atopic dermatitis. J Clin Cell

Immunol 2011;2(August (3)).

[17] Oetjen LK, Noti M, Kim BS. New insights into basophil heterogeneity.

Semin Immunopathol 2016.

[18] Kim BS, Wang K, Siracusa MC, Saenz SA, Brestoff JR, Monticelli LA, et al.

Basophils promote innate lymphoid cell responses in inflamed skin. J

Immunol (Baltim Md 1950) 2014;193(October (7)):3717–25.

[19] Perrigoue JG, Saenz SA, Siracusa MC, Allenspach EJ, Taylor BC, Giacomin

PR, et al. MHC class II-dependent basophil-CD4+ T cell interactions

promote T(H)2 cytokine-dependent immunity. Nat Immunol

2009;10(July (7)):697–705.

[20] Yoshimoto T, Yasuda K, Tanaka H, Nakahira M, Imai Y, Fujimori Y, et al.

Basophils contribute to T(H)2-IgE responses in vivo via IL-4 production

and presentation of peptide-MHC class II complexes to CD4+ T cells. Nat

Immunol 2009;10(July (7)):706–12.

[21] Leyva-Castillo JM, Hener P, Michea P, Karasuyama H, Chan S, Soumelis V,

et al. Skin thymic stromal lymphopoietin initiates Th2 responses

through an orchestrated immune cascade. Nat Commun 2013;

4(November):2847.

[22] Cheng LE, Sullivan BM, Retana LE, Allen CDC, Liang H-E, Locksley RM.

IgE-activated basophils regulate eosinophil tissue entry by modulating

endothelial function. J Exp Med 2015;212(April (4)):513–24.

[23] Noti M, Wojno EDT, Kim BS, Siracusa MC, Giacomin PR, Nair MG, et al.

Thymic stromal lymphopoietin-elicited basophil responses promote

eosinophilic esophagitis. Nat Med 2013;19(August (8)):1005–13.

[24] Yamamoto M, Kiyono H, Kweon MN, Yamamoto S, Fujihashi K, Kurazono

H, et al. Enterotoxin adjuvants have direct effects on T cells and antigen-

presenting cells that result in either interleukin-4-dependent or -

independent immune responses. J Infect Dis 2000;182(July (1)):180–90.

[25] Blazquez AB, Berin MC. Gastrointestinal dendritic cells promote Th2

skewing via OX40L. J Immunol (Baltim Md 1950) 2008;180(April

(7)):4441–50.

[26] Liu T, He S-H, Zheng P-Y, Zhang T-Y, Wang B-Q, Yang P-C. Staphylococcal

enterotoxin B increases TIM4 expression in human dendritic cells that

drives naıve CD4T cells to differentiate into Th2 cells. Mol Immunol

2007;44(July (14)):3580–7.

[27] Gagliardi MC, Sallusto F, Marinaro M, Langenkamp A, Lanzavecchia A, De

Magistris MT. Cholera toxin induces maturation of human dendritic cells

and licences them for Th2 priming. Eur J Immunol 2000;30(August

(8)):2394–403.

[28] Wang Y-H, Liu Y-J. OX40–OX40L interactions: a promising therapeutic

target for allergic diseases? J Clin Invest 2007;117(December (12)):3655–7.

[29] Yang Z, Sun R, Grinchuk V, Fernandez-Blanco JA, Blanco JAF, Notari L,

et al. IL-33-induced alterations in murine intestinal function and cytokine

responses are MyD88, STAT6, and IL-13 dependent. Am J Physiol

Gastrointest Liver Physiol 2013;304(February (4)):G381–9.

[30] Turnquist HR, Thomson AW. IL-33 broadens its repertoire to affect DC.

Eur J Immunol 2009;39(December (12)):3292–5.




innate-like lymphocytes. Mucosal Immunol 2014;7(November (6)):

1395–404.

[32] Chu DK, Jimenez-Saiz R, Verschoor CP, Walker TD, Goncharova S, Llop-

Guevara A, et al. Indigenous enteric eosinophils control DCs to initiate a

primary Th2 immune response in vivo. J Exp Med 2014;211(July

(8)):1657–72.

[33] Miyata A, Natsuaki M, Yamanishi K. Staphylococcal enterotoxin B

enhances a flare-up reaction of murine contact hypersensitivity through

up-regulation of interferon-gamma. Exp Dermatol 2008;17(October

(10)):843–8.

[34] Ganeshan K, Neilsen CV, Hadsaitong A, Schleimer RP, Luo X, Bryce PJ.

Impairing oral tolerance promotes allergy and anaphylaxis: a new murine

food allergy model. J Allergy Clin Immunol 2009;123(January (1)).

231–8.e4.

[35] Yang P-C, Xing Z, Berin CM, Soderholm JD, Feng B-S, Wu L, et al. TIM-4

expressed by mucosal dendritic cells plays a critical role in food antigen-

specific Th2 differentiation and intestinal allergy. Gastroenterology

2007;133(November (5)):1522–33.

[36] Lee J-B, Chen C-Y, Liu B, Mugge L, Angkasekwinai P, Facchinetti V, et al.

IL-25 and CD4+ TH2 cells enhance type 2 innate lymphoid cell-derived IL-

13 production, which promotes IgE-mediated experimental food allergy. J

Allergy Clin Immunol 2016;137(April (4)). 1216–25.e5.

[37] Mathias CB, Hobson SA, Garcia-Lloret M, Lawson G, Poddighe D,

Freyschmidt E-J, et al. IgE-mediated systemic anaphylaxis and impaired

tolerance to food antigens in mice with enhanced IL-4 receptor signaling.

J Allergy Clin Immunol 2011;127(May (3)). 795–805.e1–6.

[38] Burton OT, Darling AR, Zhou JS, Noval-Rivas M, Jones TG, Gurish MF,

et al. Direct effects of IL-4 on mast cells drive their intestinal expansion

and increase susceptibility to anaphylaxis in a murine model of food

allergy. Mucosal Immunol 2013;6(July (4)):740–50.

[39] Forbes EE, Groschwitz K, Abonia JP, Brandt EB, Cohen E, Blanchard C,

et al. IL-9- and mast cell-mediated intestinal permeability predisposes




































































































































to oral antigen hypersensitivity. J Exp Med 2008;205(April (4)):

897–913.

[40] Gauvreau GM, OByrne PM, Boulet L-P, Wang Y, Cockcroft D, Bigler J, et al.

Effects of an anti-TSLP antibody on allergen-induced asthmatic responses.

N Engl J Med 2014;370(May (22)):2102–10.

[41] Rance F. Food allergy in children suffering from atopic eczema. Pediatr

Allergy Immunol Off Publ Eur Soc Pediatr Allergy Immunol 2008;19(May

(3)):279–84 [quiz 285].


[42] Hamilton JD, Suarez-Farinas M, Dhingra N, Cardinale I, Li X, Kostic A,

et al. Dupilumab improves the molecular signature in skin of patients with

moderate-to-severe atopic dermatitis. J Allergy Clin Immunol

2014;134(December (6)):1293–300.

[43] Bøgh KL, van Bilsen J, Głogowski R, Lopez-Exposito I, Bouchaud G,

Blanchard C, et al. Current challenges facing the assessment of the

allergenic capacity of food allergens in animal models. Clin Transl Allergy

2016;6:21.

















DRUG DISCOVERY

TODAY

DISEASEMODELS

The use of animal models to discoverimmunological mechanismsunderpinning sensitization to foodallergensJoost J. Smit1,*, Mario Noti2, Liam O’Mahony3

1Institute for Risk Assessment Sciences, University Utrecht, Utrecht, The Netherlands2Institute of Pathology, Department of Experimental Pathology, University of Bern, Bern, Switzerland3Swiss Institute of Allergy and Asthma Research, University of Zurich, Davos, Switzerland


Editors-in-Chief




In almost all countries, food allergy is of growing con-

cern affecting all age groups. Given the increased prev-

alence of food allergies, current research focuses on

developing new treatment strategies and to predict

allergenicity of novel and modified food proteins. The

recent use of animal models has significantly contrib-

uted to a better understanding of the complex immu-

nological and pathophysiological mechanisms of food

allergies. Central to the development of food allergy is

the allergic cascade driven by cells of the innate and

adaptive immune system. These models can now be

integrated into the risk assessment of possible aller-

genic proteins. In this review, we discuss the role of the

immune system as a qualitative readout for the sensi-

tizing potential and risk assessment of food proteins.

Introduction

Previously, a lack of suitable animal models and immunologi-

cal techniques has made it difficult to grasp the role of the

immune system in the sensitization to food allergens. How-

ever, since the early 90s, there have been several food allergy

*Corresponding author: i.J. Smit ([email protected])


Section editors:Kitty Verhoeckx – K.C.M. Verhoeckx, TNO, Zeist, TheNetherlandsLiam Mahony – Swiss Institute of Allergy and AsthmaResearch, University of ZurichMichelle Epstein – Medical University of Vienna,Department of Dermatology, DIAID, ExperimentalAllergy, Waehringer, Vienna, Austria

models investigating the role of the immune system in the

pathogenesis of food allergy [1,2]. One key finding was that

oral administration of a protein to an animal might result in

sensitization or may induce oral tolerance [3,4]. A better

understanding of the mechanisms leading to sensitization

versus tolerance indicates that oral tolerance is probably the

normal physiological response and that a breakdown of this

process results in sensitization to food allergens. One possibil-

ity is that sensitization to food allergens actually occurs via

other sites like airways or skin in contrast to the intestine,

where oral tolerance is considered the default response. For

example, alterations in skin barrier integrity due to filaggrin

gene mutations were associated with increased rates of food

sensitization [4]. However, studies on IgE responses and di-

gestibility of food protein suggest that exposure via the oral

route is also important for sensitization to food allergens [5].

Oral tolerance to food antigens requires the robust induc-

tion of regulatory T (Treg) cells within the mucosa. The gut

6.09.001 63







micro-environment promotes expansion of Tregs through

multiple mechanisms including the presence of retinoic acid

and bacterial-derived metabolites such as short chain fatty

acids [6]. The main mechanisms underpinning Treg cell

function include production of inhibitory cytokines (IL-10,

TGF-b and IL-35), effector cell cytolysis (via secretion of

granzymes A and B), direct targeting of DCs via inhibitory

PD-1 and CTLR4 cell surface molecules and metabolic dis-

ruption of effector cells (CD25, cAMP, adenosine, CD39, and

CD73) [7]. However, why Tregs fail to suppress the sensitiza-

tion and effector phases of allergic reactions remains incom-

pletely understood. A recent study by Noval Rivas et al.

demonstrated that uncontrolled IL-4 signaling blocks the

generation of allergen-specific Treg cells and thus favors

the pathogenesis of food allergy [8].

Most food proteins are largely digested by gastric acids

in the stomach and intestinal enzymes after ingestion.

The remaining intact food proteins and peptides are then

Dendritic cell

Allergen-specificT cell

Th

LSINVD

Allergen

Epithelium

Food allergy

Exposure

Systemic symptomsAirway obstructionHivesShock

OX40L

GM-CSF

IL-25

IL-33

TSLP

ILC2

IEL

Eosinophil Basophil

EPO IL-4

Epithelial stress

Figure 1. Humoral and cellular mechanisms of food allergy. Allergens pass the e

promotes the transfer of allergen and the release of GM-CSF, IL-25, IL-33 or TSLP

followed by the differentiation of naıve allergen specific T cells into Th2 type resp

basophils (via IL-4). In addition, IL-33 induces ILC2s. ILC2 and Th2 cells promote c

IL-9, involved in the propagation of mast cells. Mast cell bound IgE induces mast ce

systemic food allergic responses.


transferred from the lumen to the mucosa via gut epithelial

cells (IECs) by specialized M cells lining the Peyer’s Patches or

by direct sampling of mucosal dendritic cells (DCs) [9]. Acti-

vated epithelial cells can secrete type 2 promoting cytokines

including TSLP, IL-25 and IL-33 to attract IL-4 competent

innate immune cells such as eosinophils, basophils or group 2

innate lymphoid cells (ILC2) [10] that promote surface ex-

pression of Th2 permissive co-stimulatory molecules (e.g.

OX40L) on DCs [11]. The activation of distinct DC subsets

and expression of co-stimulatory molecules are important for

determining the resulting immune response [12]. Activated

DCs process proteins and peptides, move to T cell areas and

present them on major histocompatibility complex (MHC) II

where they can interact with naıve T cells to induce T helper

(Th) 2 cell polarization [13] (Fig. 1). Migration and activation

of IELs, including gd T cells, also occurs in response to allergic

sensitization in mice [14]. While innate immune cells con-

tribute to the initiation of allergen specific Th2 responses [15]

2

IL-4IL-13

B cell Allergen-specificIgE

Mast cell

Mast celldegranulation

HistaminesLeukotrienes

CytokinesProstaglandins

PAF

ocal symptomswelling

tchingauseaomitingiarrhea

Sensitization

IL-9


pithelium and are captured by DCs. Epithelial stress, under control of IELs,

. These mediators upregulate the co-stimulatory molecule OX40L on DC,

onses under the influence of eosinophils (via eosinophilic peroxidase) or

lass switching of B-cells into IgE via IL-4 and IL-13. These cells also secrete

ll degranulation and release of mast cell mediators which induce local and


Th2 cells are the main intermediate effector cells of disease.

Studies in experimental food allergy models have demon-

strated the importance of Th2 cells as depletion of CD4 T cells

protects mice from food allergic responses while transfer of

allergen specific CD4 T cells into naıve mice can transfer

disease upon exposure to allergen [16].

B cells have an essential role in humoral immune responses

via their secretion of antigen-specific antibodies. B cell secre-

tion of immunoglobulin E (IgE) is a fundamental mediator in

atopic diseases and a hallmark of allergic sensitization [17].

Following help from Th2 cells, B cells proliferate, undergo

immunoglobulin isotype class switch recombination (CSR)

toward IgE and differentiate into antibody-secreting plasma

cells. IgE mediates immediate phase reactions like mast cell

and basophil degranulation. In addition to antibody secre-

tion, B cells can limit aggressive immune reactivity. B cells

regulate immune responses mainly via IL-10, which has been

shown in experimental models of infection, allergic inflam-

mation and tolerance [18].

Together, despite the significant advances that have been

made to understand cellular and molecular pathways associ-

ated with the pathogenesis of food allergy we do not fully

understand why default immune responses to food proteins

deviate from induction of tolerance to Th2-biased immune

responses that promote food allergy. One prominent hypoth-

esis is that the observed increase in the prevalence of food

allergy in the Western world strongly correlates with changes

in our lifestyle [19].

Food allergy models

To study immune mechanisms driving the pathogenesis of

food allergy and the sensitizing potency of food allergens,

researchers have established numerous in vivo rodent models

(Table 1). Some large animal models in pig, dogs or sheep

have been used, and might be more relevant for modeling

human responses [20]. However, the availability, ethical

concerns, high costs and extensive practical considerations

have limited the use of these models. Notably, the Brown

Norway rat model, which was first established in the 90s [21],

develop food-specific IgE in the absence of an adjuvant, after

a high frequency of intragastric dosing. However, this model

is hampered by the variable number of IgE responders and

practical disadvantages including the daily dosing for a long

period with a relatively high amount of protein. Thus, the

mouse as a food allergy model system has gained momentum

due to the need for less protein allergen and the immunolog-

ical tools available.

Adjuvants in food allergy models

In most mouse models, feeding the protein alone induces oral

tolerance. Therefore, adjuvants such as alum or cholera toxin

(CT) are frequently used to induce allergic sensitization to co-

administered proteins. Alum is administered systemically, by

intraperitoneal injection and boosts adaptive immunity by

inflammatory mediators and activating inflammatory DCs

[22]. CT is administered by intragastric administration and

induces innate immune changes that trigger allergen-specific

T- and B-cell responses, leading to an allergic phenotype.

These innate immune changes induced by CT involve acti-

vation of epithelial cells (ECs), IELs, DCs and induction of co-

stimulatory molecules, such as OX40L [14,23–26]. Under-

standing the mechanisms involved in the disruption of tol-

erance by mucosal adjuvants is highly relevant because the

same pathways may be operative in the pathogenesis of

human disease. For instance, molecular stress imposed on

gut epithelial cells by CT or other mucosal adjuvants are a

principal trigger for IEC and IEL to subsequently activate DCs,

T- and B-cells during allergic sensitization [27]. In addition,

the IEC-mediated intestinal barrier function also plays a

fundamental role in mucosal allergic responses, which is

illustrated by studies in mice following oral administration

of alcohol during allergen sensitization [28] that increases

small and large intestinal permeability thus facilitating sen-

sitization and allergic effector responses.

Sensitization to food proteins is a prerequisite for induc-

tion of effector immune responses upon allergen re-exposure

(Table 1). During the sensitization phase, an increase in

serum allergen-specific IgE and Th2-type responses in lym-

phoid organs is evident. Subsequent allergen challenges lead

to manifestations of food allergy. For example, intradermal or

intragastric sensitization leads to local manifestations includ-

ing itching, redness and swelling of skin or diarrhea, respec-

tively. Systemic exposure to allergens via intraperitoneal or

intravenous routes can result in anaphylactic reactions mea-

sured by reduced body temperature. The milk allergen beta-

lactoglobulin causes anaphylaxis after intragastric exposure

in relatively low amounts (JJ Smit, unpublished data), while

for peanut allergens doses over 200 mg are necessary to

induce anaphylaxis [28]. For peanut and other allergens,

multiple dosing of the allergen and additional treatments,

such as alcohol administration, may be necessary to induce

allergic responses.

Mouse food allergy models

There are many different types of mouse and rat food allergy

models, which influence disease outcomes and make com-

parisons between different models difficult. For instance, the

sensitizing material may be a protein extract or an isolated

individual protein [20,29]. The dose of allergen and the

frequency of allergen administration during sensitization

and challenge and the type and dose of adjuvant used (CT,

SEB) may influence the response to the specific allergen

[27,30,31]. In addition, factors that damage the epithelial

barrier (e.g. alcohol or toxins), route of exposure

[20,21,27,28,30], matrix effects (e.g. lipids, sugars, aggregated

proteins in protein extracts) [32], microbial contamination of


Dru

g D

isco

very

To

day:

Dise

ase

Mo

dels

| In

viv

o a

nd

in v

itro m

od

els

of

foo

d a

llerg

y

Vo

l. 1

7–18,

2015

Table 1. Summary of adaptive rodent models for food allergy.

Reference Species; strain Allergen Sensitization Challenge Parameters

Route Frequency Dose Adjuvant Route Frequency Dose

[29] (review) Rat; BN OVA

PN

CM

HEW

Ara h1, Sol t1,

Pen a1, Ber e1

IG 42�(daily for 6 weeks)

1–10 mg None IG 1� 10–100 mg Allergen-specific IgG, IgE

Gut permeability

[42] Mouse; BALB/c OVA

CM

WPE

IP 2� (2-weekly) 10–100 mg Alum IG 6–10�(every 3 days)

10–50 mg Allergen-specific IgG, IgE

Anaphylaxis (score + temperature)

Diarrhea

MMCP-1, histamine

Intestinal histology

Cell population counts

Cytokines

[43,44] Mouse; C3H/HeJ,

BALB/c

OVA

CPE

Soy

ALA

Ara h1

Ara h2

EP 6� (weekly) 0.1–1 mg None IG

IP

1�1�

50 mg

100 mg

Allergen-specific IgG, IgE


Cytokines

[45] Mouse; BALB/c WPE

Cashew

TD 4–6� (weekly) 1 mg None IG 1� 15 mg Allergen-specific IgG, IgE


Cytokines

[28,46,47] Mouse; C3H/HeJ CPE

CM

BLG

IG 4–6� (weekly) 0.2–10 mg CT or

CT + Vodka

IG 1� 10–200 mg Allergen-specific IgG, IgE


MMCP-1, histamine

Cytokines


[27,35,38,48,49] Mouse; C3H/HeJ,

C3H/HeOuJ, BALB/c,

C57BL/6

CPE

HEW

WPE

OVA

Ara h 1-6

Spinach

Turkey Brazil Nut

among others

IG 4–8� (weekly) 0.25–20 mg CT

SEB

DON

IP

ID

1�

1�

0.1–5 mg

50 mg

Allergen-specific IgG, IgE


MMCP-1, histamine

Cytokines


Ear swelling

BN: Brown Norway, OVA: ovalbumin, PE: peanut extract, CM: cow’s milk, WPE: whey protein extract, HEW: hen’s egg white, ALA: alfa-lactalbumin, BLG: beta-lactoglobulin, IG: intragastric, IP: intraperitoneal, EP: epicutaneous, TD: transdermal, CT:

cholera toxin, SEB: Staphylococcal enterotoxin B, DON: deoxynivalenol.

66

w

ww

.dru

gdisco

verytoday.co

m


protein extracts [33], composition of the microbiota [34] and

the animal strain used in the experiments [30,35] can all

influence the outcome of the model.

Allergic responses can be assessed by measuring changes

in cellular subsets, release of mediators and the appearance

of disease symptoms. Cytokine production upon re-stimu-

lation of mesenteric lymph nodes (MLN) and/or spleen

cells determine whether Th2-type allergic mediators are

upregulated. Flow cytometry of isolated cell populations

from intestinal tissues including MLN, lamina propria and

Peyer’s Patches can be used to identify allergen-specific

or Th2-associated cell subsets. Measurement of allergen-

specific IgG1, IgG2a, IgE or IgA antibody levels can help

establish the class of the immune response that was eli-

cited. Histological analysis of the intestine or lymphoid

organs is useful to quantify the severity of the inflamma-

tory response. In vivo, mast cell degranulation can be

assessed by measuring histamine or mouse mast cell prote-

ase-1 release. Functional parameters include gut permeabil-

ity and airway reactivity in response to ingested or inhaled

allergen challenges. Anaphylaxis is measured by monitor-

ing changes in body temperature and using an anaphylaxis

scoring system [36]. Ear swelling after intra-dermal chal-

lenge and passive cutaneous anaphylaxis can be assessed by

extravasation of Evans Blue dye [37]. Together, numerous

quantitative and semi-quantitative measurements can be

assessed in murine food allergy models to assess severity of

allergic manifestations. However, these manifestations de-

pend upon multiple exogenous and endogenous factors

(reviewed in [20,29,30]. The pros and cons of each animal

model is extensively reviewed recently elsewhere [29,31].

Predictability of food allergy models

There is extensive information on the chemical and physical

characteristics of food allergens, which belong to only 2% of

protein families. It remains unknown why certain proteins are

allergenic, compared to the large majority of food proteins,

which are not allergenic. Animal models that discriminate

between low or non-allergenic proteins from high-allergenic

proteins would be ideal for understanding food allergy mech-

anisms and for allergenicity risk assessment of novel proteins.

Dearman et al. showed that known allergenic proteins in-

duced protein-specific IgE upon systemic exposure in mice,

while non-allergenic proteins produced low IgE titers [33]. By

contrast, using the same route of administration, it was not

possible to differentiate between known allergens and puta-

tive non-allergens, for example, rubisco and soy lipoxygenase

[29]. However, oral protein administration allowed research-

ers to distinguish allergenic from non-allergenic food extracts.

Peanut, egg white and Brazil nut allergens were distinguished

from low-allergenic spinach and turkey after 2 weeks of feed-

ing a dose of 2 mg [38]. Additionally, using an ex vivo/in vitro

DC-T cell assay and an in vivo mouse model, it was possible

to distinguish known allergenic food proteins (Ara h1, beta-

lactoglobulin, shrimp tropomyosin, bovine serum albumin,

whey protein isolate) from low/non allergenic food proteins

(soy lipoxygenase, gelatin, beef tropomyosin, rubisco, patatin)

[39]. However, in this model prolonged exposure (>28 days)

may elicit responses to both allergen and non-allergen pro-

teins. Importantly, in these models, there is the possibility

that allergens are contaminated with endotoxin, which will

enhance allergen-stimulated proliferation and reduce the

threshold for T cell activation [40].

The use of highly purified proteins versus raw food extracts

is another important factor in food allergy models. The food

matrix influences responses to individual proteins dependent

on the route of administration [32]. For example, proteins

from the same source display different allergenic properties

while being ingested in the same matrix, that is, not all

proteins in peanut are allergenic and allergenic peanut pro-

teins induce significantly different allergic responses [41],

thus suggesting that protein-specific factors are possible.

Summary

Animal models have allowed us to uncover many of the

cellular responses and molecular mediators involved in the

induction of oral tolerance or allergic sensitization to food

antigens. Determination of regulatory T cell activity and

induction of Th2 lymphocyte polarization and B cell IgE class

switching are important parameters. In addition, in vitro or ex

vivo model systems should be complementary to animal

models and may provide information on cellular processing

and presentation of food proteins. Many different known and

unknown factors can influence the outcome of a food allergy

model and the development of a reference protein toolbox

(with validated high- and low-allergenic proteins) is essential

and would help standardize animal model responses across

different laboratories.



References[1] Eigenmann PA. T lymphocytes in food allergy: overview of an intricate

network of circulating and organ-resident cells. Pediatr Allergy Immunol

2002;13:162–71.

[2] Gould HJ, Ramadani F. IgE responses in mouse and man and the

persistence of IgE memory. Trends Immunol 2015;36:40–8. http://

dx.doi.org/10.1016/j.it.2014.11.002.

[3] Strobel S, Ferguson A. Immune responses to fed protein antigens in mice.

3. Systemic tolerance or priming is related to age at which antigen is first

encountered. Pediatr Res 1984;18:588–94. http://dx.doi.org/10.1203/

00006450-198407000-00004.

[4] Verhasselt V. Oral tolerance in neonates: from basics to potential

prevention of allergic disease. Mucosal Immunol 2010;3:326–33. http://

dx.doi.org/10.1038/mi.2010.25.

[5] Dupont D, Mackie AR. Static and dynamic in vitro digestion models to

study protein stability in the gastrointestinal tract and help predict

allergenicity. Drug Discov Today Dis Models 2016 [in press].





http://dx.doi.org/10.1016/j.it.2014.11.002



http://dx.doi.org/10.1203/00006450-198407000-00004

http://dx.doi.org/10.1203/00006450-198407000-00004

http://dx.doi.org/10.1203/00006450-198407000-00004

http://dx.doi.org/10.1038/mi.2010.25

http://dx.doi.org/10.1038/mi.2010.25

http://dx.doi.org/10.1038/mi.2010.25





[6] Frei R, Lauener RP, Crameri R, O’Mahony L. Microbiota and dietary

interactions: an update to the hygiene hypothesis? Allergy 2012;67:451–

61. http://dx.doi.org/10.1111/j.1398-9995.2011.02783.x.

[7] Palomares O, Martın-Fontecha M, Lauener R, Traidl-Hoffmann C,

Cavkaytar O, Akdis M, et al. Regulatory T cells and immune regulation of

allergic diseases: roles of IL-10 and TGF-b. Genes Immun 2014;15:511–20.

http://dx.doi.org/10.1038/gene.2014.45.

[8] Noval Rivas M, Burton OT, Oettgen HC, Chatila T. IL-4 production by

group 2 innate lymphoid cells promotes food allergy by blocking

regulatory T-cell function. J Allergy Clin Immunol 2016. http://

dx.doi.org/10.1016/j.jaci.2016.02.030.

[9] Schulz O, Jaensson E, Persson EK, Liu X, Worbs T, Agace WW, et al.

Intestinal CD103+, but not CX3CR1+, antigen sampling cells migrate in

lymph and serve classical dendritic cell functions. J Exp Med

2009;206:3101–14. http://dx.doi.org/10.1084/jem.20091925.

[10] Saenz SA, Noti M, Artis D. Innate immune cell populations function as

initiators and effectors in Th2 cytokine responses. Trends Immunol

2010;31:407–13. http://dx.doi.org/10.1016/j.it.2010.09.001.

[11] Chu DK, Llop-Guevara A, Walker TD, Flader K, Goncharova S, Boudreau

JE, et al. IL-33, but not thymic stromal lymphopoietin or IL-25, is central

to mite and peanut allergic sensitization. J Allergy Clin Immunol

2013;131. http://dx.doi.org/10.1016/j.jaci.2012.08.002. 187–200.e1–8.

[12] Schiavi E, Smolinska S, O’Mahony L. Intestinal dendritic cells. Curr Opin

Gastroenterol 2015;31:98–103. http://dx.doi.org/10.1097/

MOG.0000000000000155.

[13] Pabst O, Mowat AM. Oral tolerance to food protein. Mucosal Immunol

2012;5:232–9. http://dx.doi.org/10.1038/mi.2012.4.

[14] Bol-Schoenmakers M, Marcondes Rezende M, Bleumink R, Boon L, Man S,

Hassing I, et al. Regulation by intestinal gd T cells during establishment of

food allergic sensitization in mice. Allergy 2011;66:331–40. http://

dx.doi.org/10.1111/j.1398-9995.2010.02479.x.

[15] Hussain M, Epstein MM, Noti M, Noti. Experimental food allergy models

to study the role of innate immune cells as initiators of allergen specific

Th2 immune responses. Drug Discov Today Dis Models 2016 [in press].

[16] Knight AK, Blazquez AB, Zhang S, Mayer L, Sampson HA, Berin MC. CD4 T

cells activated in the mesenteric lymph node mediate gastrointestinal

food allergy in mice. Am J Physiol Gastrointest Liver Physiol

2007;293:G1234–43. http://dx.doi.org/10.1152/ajpgi.00323.2007.

[17] Oettgen HC. Fifty years later: emerging functions of IgE antibodies in host

defense, immune regulation, and allergic diseases. J Allergy Clin Immunol

2016;137:1631–45. http://dx.doi.org/10.1016/j.jaci.2016.04.009.

[18] Stanic B, van de Veen W, Wirz OF, Ruckert B, Morita H, Sollner S, et al. IL-

10-overexpressing B cells regulate innate and adaptive immune responses.

J Allergy Clin Immunol 2015;135:771–8. http://dx.doi.org/10.1016/

j.jaci.2014.07.041.

[19] Bloomfield SF, Stanwell-Smith R, Crevel RWR, Pickup J. Too clean, or not

too clean: the hygiene hypothesis and home hygiene. Clin Exp Allergy

2006;36:402–25. http://dx.doi.org/10.1111/j.1365-2222.2006.02463.x.

[20] Van Gramberg JL, de Veer MJ, O’Hehir RE, Meeusen ENT, Bischof RJ. Use

of animal models to investigate major allergens associated with food

allergy. J Allergy (Cairo) 2013;2013. http://dx.doi.org/10.1155/2013/

635695. 635695–10.

[21] Atkinson HA, Miller K. Assessment [correction of Assessment] of the

brown Norway rat as a suitable model for the investigation of food allergy.

Toxicology 1994;91:281–8.

[22] Kool M, Soullie T, van Nimwegen M, Willart MAM, Muskens F, Jung S,

et al. Alum adjuvant boosts adaptive immunity by inducing uric acid and

activating inflammatory dendritic cells. J Exp Med 2008;205:869–82.

http://dx.doi.org/10.1084/jem.20071087.




innate-like lymphocytes. Mucosal Immunol 2014;7:1395–404. http://

dx.doi.org/10.1038/mi.2014.29.

[24] Smit JJ, Bol-Schoenmakers M, Hassing I, Fiechter D, Boon L, Bleumink R,

et al. The role of intestinal dendritic cells subsets in the establishment of

food allergy. Clin Exp Allergy 2011;41:890–8. http://dx.doi.org/10.1111/

j.1365-2222.2011.03738.x.


[25] Berin MC, Mayer L. Immunophysiology of experimental food allergy.

Mucosal Immunol 2009;2:24–32. http://dx.doi.org/10.1038/

mi.2008.72.

[26] Arias K, Chu DK, Flader K, Botelho F, Walker T, Arias N, et al. Distinct

immune effector pathways contribute to the full expression of peanut-

induced anaphylactic reactions in mice. J Allergy Clin Immunol

2011;127. http://dx.doi.org/10.1016/j.jaci.2011.03.044. 1552–61.e1.

[27] Bol-Schoenmakers M, Braber S, Akbari P, de Graaff P, van Roest M,

Kruijssen L, et al. The mycotoxin deoxynivalenol facilitates allergic

sensitization to whey in mice. Mucosal Immunol 2016. http://dx.doi.org/

10.1038/mi.2016.13.

[28] Srivastava KD, Siefert A, Fahmy TM, Caplan MJ, Li X-M, Sampson HA.

Investigation of peanut oral immunotherapy with CpG/peanut

nanoparticles in a murine model of peanut allergy. J Allergy Clin

Immunol 2016. http://dx.doi.org/10.1016/j.jaci.2016.01.047.

[29] Ladics GS, Fry J, Goodman R, Herouet-Guicheney C, Hoffmann-

Sommergruber K, Madsen CB, et al. Allergic sensitization: screening

methods. Clin Transl Allergy 2014;4:13. http://dx.doi.org/10.1186/2045-

7022-4-13.

[30] Liu T, Navarro S, Lopata AL. Current advances of murine models for food

allergy. Mol Immunol 2016;70:104–17. http://dx.doi.org/10.1016/

j.molimm.2015.11.011.

[31] Bøgh KL, van Bilsen J, Głogowski R, Lopez-Exposito I, Bouchaud G,

Blanchard C, et al. Current challenges facing the assessment of the


2016;6:21. http://dx.doi.org/10.1186/s13601-016-0110-2.

[32] Wavrin S, Bernard H, Wal J-M, Adel-Patient K. Influence of the route of

exposure and the matrix on the sensitisation potency of a major cows’

milk allergen. Clin Transl Allergy 2015;5:3. http://dx.doi.org/10.1186/

s13601-015-0047-x.

[33] Dearman RJ, Kimber I. Determination of protein allergenicity: studies in

mice. Toxicol Lett 2001;120:181–6.

[34] Barcik W, Untersmayr E, Pali-Scholl I, O’Mahony L, Frei R. Influence of

microbiome and diet on immune responses in food allergy models. Drug

Discov Today Dis Models 2016.

[35] Smit JJ, Willemsen K, Hassing I, Fiechter D, Storm G, van Bloois L, et al.

Contribution of classic and alternative effector pathways in peanut-

induced anaphylactic responses. PLoS ONE 2011;6:e28917. http://

dx.doi.org/10.1371/journal.pone.0028917.

[36] Li XM, Serebrisky D, Lee SY, Huang CK, Bardina L, Schofield BH, et al. A

murine model of peanut anaphylaxis: T- and B-cell responses to a major

peanut allergen mimic human responses. J Allergy Clin Immunol

2000;106:150–8. http://dx.doi.org/10.1067/mai.2000.107395.

[37] Evans H, Killoran KE, Mitre E. Measuring local anaphylaxis in mice. J Vis

Exp 2014;e52005. http://dx.doi.org/10.3791/52005.

[38] Bowman CC, Selgrade MK. Differences in allergenic potential of food

extracts following oral exposure in mice reflect differences in digestibility:

potential approaches to safety assessment. Toxicol Sci 2008;102:100–9.

http://dx.doi.org/10.1093/toxsci/kfm288.

[39] Smit J, Zeeuw-Brouwer ML, van Roest M, de Jong G, van Bilsen J.

Evaluation of the sensitizing potential of food proteins using two mouse

models. Toxicol Lett 2016;262:62–9. http://dx.doi.org/10.1016/

j.toxlet.2016.09.005.

[40] Velickovic TC, Jankov RM. Design and modifications of allergens for

improving specific immunotherapy. Inflamm Allergy Drug Targets

2008;7:270–8.

[41] Smit JJ, Pennings MT, Willemsen K, van Roest M, van Hoffen E, Pieters RH.

Heterogeneous responses and cross reactivity between the major peanut

allergens Ara h 1, 2,3 and 6 in a mouse model for peanut allergy. Clin

Transl Allergy 2015;5:13. http://dx.doi.org/10.1186/s13601-015-0056-9.

[42] Kucuk ZY, Strait R, Khodoun MV, Mahler A, Hogan S, Finkelman FD.

Induction and suppression of allergic diarrhea and systemic anaphylaxis

in a murine model of food allergy. J Allergy Clin Immunol

2012;129:1343–8. http://dx.doi.org/10.1016/j.jaci.2012.03.004.

[43] Tordesillas L, Goswami R, Benede S, Grishina G, Dunkin D, Jarvinen KM,

et al. Skin exposure promotes a Th2-dependent sensitization to peanut

allergens. J Clin Invest 2014;124:4965–75. http://dx.doi.org/10.1172/

JCI75660.

http://dx.doi.org/10.1111/j.1398-9995.2011.02783.x

http://dx.doi.org/10.1111/j.1398-9995.2011.02783.x

http://dx.doi.org/10.1038/gene.2014.45

http://dx.doi.org/10.1038/gene.2014.45

http://dx.doi.org/10.1016/j.jaci.2016.02.030



http://dx.doi.org/10.1084/jem.20091925






http://dx.doi.org/10.1097/MOG.0000000000000155



http://dx.doi.org/10.1038/mi.2012.4

http://dx.doi.org/10.1038/mi.2012.4

http://dx.doi.org/10.1111/j.1398-9995.2010.02479.x

http://dx.doi.org/10.1111/j.1398-9995.2010.02479.x

http://dx.doi.org/10.1111/j.1398-9995.2010.02479.x




http://dx.doi.org/10.1152/ajpgi.00323.2007

http://dx.doi.org/10.1152/ajpgi.00323.2007






http://dx.doi.org/10.1111/j.1365-2222.2006.02463.x

http://dx.doi.org/10.1111/j.1365-2222.2006.02463.x

http://dx.doi.org/10.1155/2013/635695

http://dx.doi.org/10.1155/2013/635695

http://dx.doi.org/10.1155/2013/635695






http://dx.doi.org/10.1038/mi.2014.29

http://dx.doi.org/10.1038/mi.2014.29

http://dx.doi.org/10.1038/mi.2014.29

http://dx.doi.org/10.1111/j.1365-2222.2011.03738.x

http://dx.doi.org/10.1111/j.1365-2222.2011.03738.x

http://dx.doi.org/10.1111/j.1365-2222.2011.03738.x

http://dx.doi.org/10.1038/mi.2008.72

http://dx.doi.org/10.1038/mi.2008.72

http://dx.doi.org/10.1038/mi.2008.72



http://dx.doi.org/10.1038/mi.2016.13

http://dx.doi.org/10.1038/mi.2016.13

http://dx.doi.org/10.1038/mi.2016.13



http://dx.doi.org/10.1186/2045-7022-4-13

http://dx.doi.org/10.1186/2045-7022-4-13

http://dx.doi.org/10.1186/2045-7022-4-13

http://dx.doi.org/10.1016/j.molimm.2015.11.011



http://dx.doi.org/10.1186/s13601-016-0110-2

http://dx.doi.org/10.1186/s13601-016-0110-2

http://dx.doi.org/10.1186/s13601-015-0047-x

http://dx.doi.org/10.1186/s13601-015-0047-x

http://dx.doi.org/10.1186/s13601-015-0047-x









http://dx.doi.org/10.1067/mai.2000.107395

http://dx.doi.org/10.1067/mai.2000.107395

http://dx.doi.org/10.3791/52005

http://dx.doi.org/10.3791/52005

http://dx.doi.org/10.1093/toxsci/kfm288

http://dx.doi.org/10.1093/toxsci/kfm288

http://dx.doi.org/10.1016/j.toxlet.2016.09.005






http://dx.doi.org/10.1186/s13601-015-0056-9

http://dx.doi.org/10.1186/s13601-015-0056-9



http://dx.doi.org/10.1172/JCI75660




[44] Wavrin S, Bernard H, Wal J-M, Adel-Patient K. Influence of the route of

exposure and the matrix on the sensitisation potency of a major cows’

milk allergen. Clin Transl Allergy 2015;5:3. http://dx.doi.org/10.1186/

s13601-015-0047-x.

[45] Gonipeta B, Parvataneni S, Tempelman RJ, Gangur V. An adjuvant-free

mouse model to evaluate the allergenicity of milk whey protein. J Dairy

Sci 2009;92:4738–44. http://dx.doi.org/10.3168/jds.2008-1927.

[46] Li X-M, Srivastava K, Huleatt JW, Bottomly K, Burks AW, Sampson HA.

Engineered recombinant peanut protein and heat-killed Listeria

monocytogenes coadministration protects against peanut-induced

anaphylaxis in a murine model. J Immunol 2003;170:3289–95.

[47] Frossard CP, Zimmerli SC, Rincon Garriz JM, Eigenmann PA. Food allergy

in mice is modulated through the thymic stromal lymphopoietin

pathway. Clin Transl Allergy 2015;6:2. http://dx.doi.org/10.1186/s13601-

016-0090-2.

[48] van Esch BCAM, van Bilsen JHM, Jeurink PV, Garssen J, Penninks AH, Smit

JJ, et al. Interlaboratory evaluation of a cow’s milk allergy mouse model to

assess the allergenicity of hydrolysed cow’s milk based infant formulas.

Toxicol Lett 2013;220:95–102. http://dx.doi.org/10.1016/

j.toxlet.2013.04.008.

[49] Ganeshan K, Neilsen CV, Hadsaitong A, Schleimer RP, Luo X, Bryce PJ.

Impairing oral tolerance promotes allergy and anaphylaxis: a new murine

food allergy model. J Allergy Clin Immunol 2009;123:231–238.e4. http://

dx.doi.org/10.1016/j.jaci.2008.10.011.


http://dx.doi.org/10.1186/s13601-015-0047-x

http://dx.doi.org/10.1186/s13601-015-0047-x

http://dx.doi.org/10.1186/s13601-015-0047-x

http://dx.doi.org/10.3168/jds.2008-1927

http://dx.doi.org/10.3168/jds.2008-1927





http://dx.doi.org/10.1186/s13601-016-0090-2

http://dx.doi.org/10.1186/s13601-016-0090-2

http://dx.doi.org/10.1186/s13601-016-0090-2







DRUG DISCOVERY

TODAY

DISEASEMODELS

Influence of microbiome and diet onimmune responses in food allergymodelsWeronika Barcik1, Eva Untersmayr2, Isabella Pali-Scholl2,3,

Liam O’Mahony1, Remo Frei1,4,*1Swiss Institute of Allergy and Asthma Research, University of Zurich, Davos, Switzerland2Department of Pathophysiology and Allergy Research, Center for Pathophysiology, Infectiology and Immunology, Medical University of

Vienna, Vienna, Austria3Comparative Medicine, Messerli Research Institute of the University of Veterinary Medicine Vienna, Medical University Vienna and

University Vienna, Austria4Christine Kuhne – Center for Allergy Research and Education (CK-CARE), Davos, Switzerland


Editors-in-Chief




The intestinal immune system is intimately connected

with the vast array of microbes present within the gut

and the diversity of food components that are con-

sumed daily. The discovery of novel molecular mecha-

nisms, which mediate host–microbe–nutrient

communication, have highlighted the important roles

played by microbes and dietary factors in influencing

mucosal inflammatory and allergic responses. In this

review, we summarize the recent important findings in

this field, which are important for food allergy and

particularly relevant to animal models of food allergy.

Introduction

Food allergies are a growing health problem affecting a

significant proportion of the population, associated with a

substantial impact on quality of life and economic burden

[1,2]. Why some individuals develop allergic reactions to

specific foods, while the majority tolerates these food anti-

gens, is largely unknown. However, it is likely that the

interplay between genetic factors, microbial composition

*Corresponding author: R. Frei ([email protected])


Section editor:Kitty Verhoeckx – K.C.M. Verhoeckx, TNO, Zeist,The NetherlandsLiam Mahony – Swiss Institute of Allergy and AsthmaResearch, University of ZurichMichelle Epstein – Medical University of Vienna,Department of Dermatology, DIAID, ExperimentalAllergy, Waehringer, Vienna, Austria

and metabolic activity, dietary factors, or timing of antigen

exposure may play a crucial role. Animal models of food

allergy have allowed investigators to individually modulate

and test specific factors, which influence sensitization and

severity of disease. This review is focused on the recent

knowledge gained from animal models investigating the

influence of the microbiome and diet on the development

of food allergies. Table 1 shows an overview about the models

discussed in this review.

Food allergy models overview

In animal models, adjuvants are usually required to induce

sensitization to food allergens and are typically applied in

parallel with the allergen. Experimental adjuvants include

cholera toxin, staphylococcal enterotoxin B (also known to

6.06.003 71







Table 1. Overview of food allergy models.

Animal

model

Strain/mutation Antigen Adjuvant Treatment Reference

Mouse C57BL/6 Peanut Cholera toxin Kanamycin (4 mg/mL), gentamicin (0.35 mg/

mL), colistin (8500 U/mL), metronidazole

(2.15 mg/mL), and vancomycin (0.45 mg/

mL) After weaning, the Abx were

administered at 50-fold dilution except for

vancomycin, which was maintained at

0.5 mg/mL

Colonization of Clostridia

[15]

Mouse BALB/c

WT and Il4raY709

Egg protein OVA Staphylococcal

enterotoxin B

WT mice reconstitution with flora derived

from OVA-sensitized WT or Il4raF709 mice

[16]

Mouse BALB/c na na Bifidobacterium breve AH1205,

Bifidobacterium longum AH1206 and

Lactobacillus salivarius AH102 of human

origin

[20]

Mouse C3H/HeN b-Lactoglobulin

Whey protein

Cholera toxin Colonization with the infant microbiota

(dominance of Bifidobacterium and

Bacteroides species)

[18]

Mouse BALB/c Egg protein OVA Al(OH)3 Control diet containing 15% casein as a

protein source or an experimental diet

containing 15% of a mixture of amino acids

[22]

Mouse BALB/c Egg protein OVA Al(OH)3 Raw bovine milk, raw bovine milk heated to

878C, raw bovine milk gamma irradiated

[24]

Mouse BALB/c Egg protein OVA Aluminum

potassium

sulfate

Ag-free diet, amino acid diet (AAD), L-

amino-acid defined AIN-93G diet, irradiated

and vacuum-packed AAD

Ampicillin (1 g/L), neomycin (1 g/L),

metronidazole (1 g/L), and 0.5 g/L of

vancomycin (1 g/L)

[23]

Mouse BALB/c

C57BL/6

WT and CD1d�/� and Ja18�/�

Ber e 1 na Different lipid fractions (600 mg) from Brazil

nut seeds

[26]

Mouse BALB/cAnNCrl mice Ovomucoid

b-Lactoglobulin

Al(OH)3 Untreated antigen, sham-nitrated antigen or

nitrated antigen

[25]

Mouse BALB/c Egg protein OVA Freund’s adjuvant Oil diet [28]

Mouse C3H/HeOuJ Whey protein Cholera toxin Cows’ milk protein free AIN-93G diet

(containing 7% soyabean oil) or a 10%

soyabean oil diet (59.1% PUFA, of which

53.1% was LA (n-6 PUFA), 5.6% a-linolenic

acid (n-3 PUFA), 24.9% MUFA (oleic acid)

and 15.1% SFA (palmitic acid and stearic

acid))

[29]

Mouse WT and Sphk1�/�,

Sphk2�/�, S1pr2�/�,

and S1pr3�/�,

S1pr4�/�S1pr1loxp/loxp-Mx,

WSh/WSh

DNP36-HSA

(+ DNP-specific IgE)

na Sphingosine-1-phosphate

Polyinosinic-polycytidylic acid, histamine,

albumin

[30]




Animal

model


Mouse C57BL/6

WT and

SphK1�/�

SphK2�/� l

Egg protein OVA Al(OH)3 Proton-pump-inhibitor omeprazole,

sucralfate (i.e. anti-acid medication)

[4]

Mouse BALB/c OlaHsd Egg protein OVA Cholera toxin Different polyphenol-enriched apple

extracts, polyphenol-enriched cocoa

extract or purified epicatechin

[32]

Rat Brown Norway Egg protein OVA Al(OH)3, toxin from

Bordetella pertussis

Diet with no polyphenols, two cocoa-

enriched diets either including conventional

cocoa (CC) or cocoa flavonoids from

nonfermented cocoa (NFC), both

containing 0.4% of polyphenols

[33]

Mouse BALB/c Celery proteins Al(OH)3 Acid-suppression by proton pump inhibitor,

followed by application of the celery extract

mixed with 2 mg sucralfate; control groups:

celery extract alone.

[35]

Mouse BALB/c Egg protein OVA Al(OH)3 Diets containing either 0.08, 0.25, or

2.7 ppm Se.

[38]

Mouse BALB/c

WT and

H2R�/�

na na L. saerimneri 30a, famotidine [41]

Mouse BALB/c

WT and

H2R�/�

na na L. rhamnosus [42]

Mouse C57Bl/6

WT and

Gpr43�/�

Acute DSS colitis, chronic DSS colitis,

TNBS-induced colitis, K/BxN inflammatory

arthritis model, allergic airway disease

(OVA/alum)

[43]

Mouse C57BL/6J na na Diet-induced obesity (high fat diet) [46]

Mouse C57BL/6 na na S. flexneri 5a (M90T), IpaB4 deletion mutant

S. flexneri 5a (M90TDIpaB4), wild-type

Salmonella typhimurium (UK-1), and non-

invasive Shigella strains (BS176)

Sphingosine-1-phosphate

[47]

Mouse C57BL/6

WT and

CD1d�/�

na na KRN7000 (1 mg/ml), bacterial lipid GSL-

Bf717,

PE-Cers

[48]

Mouse C57BL/6 Dextran sodium sulfate (DSS) colitis model

B. infantis

[49]

Mouse BALB/c Wheat-deaminated

gliadins (pups)

Al(OH)3 (pups) Diet supplemented with 4% galacto-

oligosaccharides and inulin in a 9:1 ratio,

during pregnancy and breastfeeding

[57]

Mouse C57BL/6

GPR41�/� and

GPR43�/�

HDM na Low-fiber diet, high-fiber diet (normal chow

supplemented with 30% cellulose or 30%

pectin)

Sodium propionate, sodium acetate

[52]




Animal

model


Mouse BALB/c Egg protein OVA Al(OH)3 Standard-fiber chow (4% content) or a low-

fiber chow (1.75% content), an extra fiber

supplementation of soluble pectin or

insoluble cellulose

[53]

Mouse BALB/c and

DO.11.10

transgenic mice

na na L. rhamnosus (JB-1), L. salivarius UCC118

heme oxygenase inhibitor (Chromium(III)

Mesoporphyrin IX chloride)

[55]

Mouse C3H/HeJ Shrimp tropomyosin Cholera toxin Probiotic VSL#3 (lyophilized mixture of

Lactobacillus acidophilus, L. delbrueckii subsp.

bulgaricus, L. casei, L. plantarum,

Bifidobacterium longum, B. infantis, B. breve,

Streptococcus salivarius subsp. thermophilus)

[56]

Mouse BALB/cByJ Cow’s milk Cholera toxin GF mice were orally inoculated with a 1:100

dilution of fecal homogenate freshly

prepared from CV mice

[61]

Mouse BALB/c,

Swiss-Webster, C57BL/6,

Rag1�/�

BaS-TRECK

Csf2rb�/�

Csf2rb�/�

Igh-7�/�

IL-4/eGFP reporter

Il4�/�

Myd88�/�

Nod1�/�

Tslp�/�

HDM Ampicillin (0.5 mg ml�1), gentamicin

(0.5 mg ml�1), metronidazole

(0.5 mg ml�1), neomycin (0.5 mg ml�1), and

vancomycin (0.25 mg ml�1)

Papain, antibodies, CpG, diphtheria toxin

[12]

play a role in human allergic diseases), or aluminum hydrox-

ide. They generally induce a strong T helper cell type-2

response via their influence on dendritic cell or macrophage

phenotypes or can also inhibit regulatory T cells [1,3]. While

sensitization can be induced via different routes such as oral,

intranasal, sublingual, or cutaneous, the presence of adju-

vants or danger signals (e.g. tape stripping the skin prior to

cutaneous exposure), or the impairment of physiological

gastric digestion [4,5] is crucial. In addition to measuring

sensitization (e.g. IgE induction), allergen challenge can re-

sult in anaphylaxis, which is assessed by symptoms (scratch-

ing, diarrhea, piloerection, labored respiration, cyanosis

around mouth and tail, reduced activity, tremors, convulsion

or death) and drop in body temperature [6]. Besides murine

and rat food allergy models, there are also food allergy models

in pigs, dogs or sheep. The advantages and disadvantages of

different food allergy animal model design parameters have

been reviewed extensively elsewhere [7].

Influence of microbiota on food allergy

There has been an increase in the number of individuals

suffering from allergic and inflammatory diseases over the


last decades, particularly in Western, developed countries [8].

The hygiene hypothesis suggests that altered exposure to

environmental factors may play a part in this phenomenon.

It suggests that excessive hygiene practices and limited con-

tact with microorganisms may contribute to allergic sensiti-

zation, including sensitization to food allergens [9]. Other

factors have also been linked with alterations in the gut

microbiome and increased risk of food allergy, such as exces-

sive antibiotic use (especially during infancy), high fat diet

and mode of delivery [10,11].

The importance of the host microbiota on immune

responses in mice models was highlighted with germ-free

mice and broad-spectrum antibiotic treated mice. In both

cases, serum IgE levels as well as basophil numbers were

increased and mice displayed exaggerated allergic

responses. Moreover, signals derived from commensal bac-

teria regulated bone marrow basophil development, which

shows that the microbiota can influence hematopoietic

programs in addition to regulation of immune cells in the

mucosa. These studies are relevant to human diseases, espe-

cially in the context of children’s exposure to antibiotics

early in life [12].


DIET MICROBIOTA

Undigestedfood

components

DC

Foodcomponents

Bacteriacomponents

IgGIgG IgE IgE

TH1

Treg

Treg

e.g.LPS

e.g.SCFA

e.g.vit. D

B BB

B B

IgGIgE

TH2

TH2TH2

TH1

TH1

TregTreg

Treg Treg


Figure 1. Microbiome and diet influences mucosal immune responses. Bacterial components, dietary components and metabolites released by bacterial

metabolism of undigested dietary components impact on epithelial cells, innate immune cells and adaptive immune cell polarisation. Exposure of the immune

system to bacterial components or dietary factors can either promote allergic responses (e.g. LPS) or dampen allergic responses (e.g. SCFA). Allergic

responses are characterized by an increase in IgE, IgG and Th2 cell numbers and a decrease of Treg cell numbers or activity.

The human gut is colonized with approximately 1014

bacteria, which represents approximately 1500 different spe-

cies, [13] typically dominated by two phyla, the Bacteroidetes

and the Firmicutes [14]. Many human studies and animal

models have now demonstrated that appropriate host–micro-

biota interactions are essential for immunological develop-

ment and oral tolerance. An overview of host–microbiota

interactions is illustrated in Fig. 1.

Neonatal mice treated with broad-spectrum antibiotics

became more prone to peanut allergy, as evidenced by in-

creased levels of circulating peanut specific IgE and IgG1

antibodies [15]. In addition, another study examining the

influence of a dysbiotic microbiota on food allergy was

investigated in Il4raF709 mice [16]. Il4raF709 mice carry a

mutation in the IL-4 receptor chain. This gain-of-function

mutation results in augmented signal transducer and activa-

tor of transcription 6 activation by IL-4 and IL-13, which

promotes allergy responses by increasing IgE and mast cell

levels, after antigen sensitization [17]. The study clearly

demonstrated that the microbiome composition differed

between allergic (Il4raF709 mice) OVA-sensitized mice and

wild type food allergic mice. Moreover, when the microbiome

from Il4raF709 mice was transplanted into germ-free wild-

type mice, these mice developed higher ova-specific IgE

antibody titres and more severe allergic reactions, suggesting

that allergic sensitization may be influenced by microbiome

composition. In this animal model, allergic responses were

associated with a decreased abundance of Firmicutes and

increased abundance of Proteobacteria [16].

Colonization of gnotobiotic mice with Clostridia (Firmi-

cutes phylum) suggested a food allergy-protective effect.

Moreover, when Clostridia were reintroduced to antibiotic

treated mice, the sensitization to food allergen was decreased.

The reason might be connected with the fact that Clostridia

induce IL-22 production which reduces uptake of antigens

from food into the systemic circulation [15]. Germ free mice



colonized with microbiota from healthy human infants

exhibited milder allergic symptoms after sensitization with

whey protein in comparison to mice that remained germ free.

The healthy infant gut microbiome is typically dominated by

Bifidobacteria and Bacteroides, which are known to have anti-

inflammatory properties [18]. Decreased diversity of the Bac-

teroidetes phylum was also reported in a separate study

examining infants with atopic eczema [19]. However, not

all Bifidobacteria species are equally effective in the generation

of mucosal regulatory T cells or in their protective effects in

food allergy models [20]. Interestingly, germ free mice colo-

nized with microbes from healthy humans were character-

ized by lower plasma level of antigen specific IgG1, with no

significant changes in IgE levels, suggesting that the protec-

tive effects on allergic responses are IgE-independent in this

model [18].

Influence of dietary components on food allergy

The accurate assessment of food allergy in animal models

requires careful control of dietary factors, in addition to the

microbiome itself. It is important to establish a stabilized

animal model, as both diet and microbiota play an important

role in food allergy and the response of immune system may

be influenced by microbiome-diet interactions. In addition,

when attempting to translate food allergy model across dif-

ferent laboratories, it is important to consider the effect of a

different microbiota and diet on the study results. When

establishing a food allergy animal model, the composition

of the diet has to be considered specifically, as dietary anti-

gens are the triggering factor for eliciting an immune re-

sponse in these model systems. Dietary pre-exposure to the

test antigen must be avoided, not only in experimental

animals but also from parental generations due to the poten-

tial antigen transfer in utero [21]. Moreover, selection of

suitable, relevant antigens for immunization and deciding

on using whole foods (including food matrixes and related

component) versus single purified allergens are crucial con-

siderations for sensitization and challenge outcomes.

Timing of exposure and the nature of proteins themselves

can contribute to immune activation and maturation, as the

absence of dietary proteins until adulthood was associated

with milder allergen-specific immune responses in mice fol-

lowing sensitization and paradoxically also hampered oral

tolerance induction [22]. A recent study reported that deple-

tion of dietary antigens by feeding an elemental diet was

associated with decreased numbers of regulatory T cells de-

veloping extra-thymically in the intestine from conventional

T cells and was associated with less severe mucosal inflam-

matory and allergic responses [23]. In addition, protein de-

naturation, for example, milk heat treatment or milk gamma

sterilization, or modification of the amino acids, for example,

tyrosine nitration, have a substantial influence on the im-

mune outcome in food allergy models [24,25].


Lipid-containing food matrixes influence the allergic re-

sponse, as specific allergen bound lipid fractions were revealed

to be essential for induction of Brazil-nut specific IgE and IgG1

antibodies in mice after intraperitoneal administration [26].

Lipids as matrix components not only influence food allergy

development by interaction with allergenic proteins, but also

have intrinsic immunomodulating properties. Polyunsaturat-

ed fatty acids (PUFA) and short chain fatty acids (SCFA) have

been extensively investigated in this regard. Intake of n-6 PUFA

rich soyabean oil increased the allergic response towards whey

proteins in a concentration dependent manner and hindered

tolerance induction when feeding partial whey hydrolysate

before sensitization [27]. In contrast, the anti-inflammatory

effects of n-3 PUFA containing linseed oil were mediated by

conversion of dietary n-3 a-linolenic acid to 17,18-epoxyeico-

satetraenoic acid in the gut [28]. Feeding of fish oil rich in n-3

PUFA was associated with prevention of cow’s milk sensitiza-

tion and protection was transferred by injection of CD25+ T

regulatory cells into naive recipient animals [29]. Other lipid

components are also important contributors to food allergy.

Sphingolipids are essential constituents of the outer cellular

membranes but also have bioactive functions, for example,

activation of immune cells. Sphingosine-1-phospate (S1P) is

produced by mast cells and signals back to these cells in an

autocrine manner. Even though the S1P converting enzyme

Sphingosine Kinase (SphK) 1 as well as the S1P receptor 2 were

reported to be essential for recovery from severe anaphylactic

reactions [30], it was demonstrated that intrinsic S1P produc-

tion via both SphK 1 and 2 was essential for food allergen

sensitization and effector cell activation in a oral mouse food

allergy model, potentially via impaired intestinal epithelial

barrier function [4].

In addition to the food compounds mentioned above,

there is a large number of micronutrients that influence food

allergy outcome by direct immune modulation, such as

vitamins, trace elements and plants polyphenols. For exam-

ple, vitamin D can be taken up via the diet, even though the

major part is produced in the skin upon UV exposure. The

inverse correlation of vitamin D levels with food allergy

development has been extensively studied in human as well

as animal studies underlining its modulatory effects on the

innate as well the adaptive immune system [31]. Polyphenols

also show immunomodulating properties. In mouse as well as

rat food allergy models, polyphenols from plant sources such

as cocoa were associated with reduced Th2 antibody response

and an overall anti-allergic protective effect [32,33]. Trace

elements support physical barriers (skin/mucosa), cellular

immunity and antibody production, and modulate immune

cell function by regulating redox-sensitive transcription fac-

tors, thus affecting production of cytokines and prostaglan-

dins [34]. Both iron and zinc serum levels were significantly

reduced in aged animals when compared to younger adult

mice, however food allergy could be induced equally in both


groups under acid-suppressing conditions [35]. Novel in vitro

data suggest an effect of low iron levels on allergy induction.

The milk allergen Bosd 5 as well as the aeroallergen Bet v1

from birch induced higher CD4+ T cell numbers and Th2-

cytokine responses in addition to IFN-g in human PBMC from

healthy as well as from allergic patients only when not-loaded

with iron [36,37]. Selenium has important functions in lym-

phocyte activity, and protects immune cells against oxidative

damage as component of selenoproteins. Selenium deficien-

cies reduce antibody production, lymphocyte proliferation

and cytotoxicity of immune-competent cells, whereas syn-

thesis of proinflammatory eicosanoids increases. In a murine

study, high dietary selenium prevented the induction of

asthma in OVA-sensitized mice [38].

Apart from these beneficial constituents of the diet, also

detrimental components such as toxins can influence the

immune response. Mediators such as histamine, which is

produced by bacterial metabolism of the amino acid histidine

and is associated with ageing of certain foods, have direct

effects on the immune system via interaction with receptors

on immune cells. These receptors are expressed by mucosal

cells and receptor expression is altered during mucosal in-

flammatory responses [39,40]. Binding of histamine to its

receptor 2 on immune cells such as T cells, B cells and

dendritic cells (DC) was reported to influence mucosal im-

munity and response to microbial ligands [41,42].

Interaction between diet, microbiota, metabolism and

the immune system

Dietary factors not only directly influence immune signaling,

but also indirectly affect the microbiome composition and

metabolic activity of the host. SCFA derived from intestinal

microbes are important for mucosal homeostasis. The SCFA

butyrate is an important energy source for colonocytes, and

regulates the assembly and organization of tight junctions. In

addition, SCFA bind G-protein coupled receptors (GPCRs),

such as GPR41 and 43, thereby suppressing inflammation.

Similarly to germ-free mice, mice deficient in GPR43 showed

increased inflammatory responses in models of colitis, arthri-

tis and asthma [43].

While a number of studies have shown SCFA-protective

effects in murine asthma models, similar effects in murine

food allergy models are less well described. However, the

beneficial effect of a high fiber diet and SCFA production

on gut inflammation has been demonstrated [44]. A high

fiber diet or oral administration of SCFA increase regulatory T

cells in the lamina propria of GF or antibiotic-treated mice.

Moreover, tolerance to cow’s milk was improved in cow’s

milk allergic infants following treatment with a probiotic-

formula that expanded butyrate-producing bacteria within

the gut [45].

In a mouse model, feeding a high-fat diet (HFD) resulted in

a (reversible) altered microbiota composition and bacterial

diversity significantly declined in the HFD group after only 2

weeks of feeding. Furthermore, a gradual and significant

increase of the relative abundance of Firmicutes and Proteo-

bacteria, paralleled by a decrease in Bacteroidetes was ob-

served [46]. Moreover, intestinal microbes release lipid

mediators such as glycosphingolipids or modulate S1P-relat-

ed genes of the host intestinal tissue resulting in attenuated

intestinal inflammation and regulated natural killer T cell

homeostasis [47,48].

In addition to diet influencing microbiome activities, spe-

cific microbes can alter the host metabolism of dietary com-

ponents. For example, vitamin A is metabolized by gut

dendritic cells resulting in the secretion of retinoic acid,

and retinoic acid is important for modulating mucosal in-

flammatory and tolerogenic responses. Specific bifidobacter-

ial strains can upregulate expression of the enzyme that

converts vitamin A into retinoic acid, thereby maximizing

the anti-inflammatory effects of this vitamin [49].

Despite significant interest in this topic, a limited number

of studies have been published linking the influence of diet

on allergic responses via alterations of the microbiome

(reviewed in [50]). Studies by Bouchaud et al. demonstrated

that mice fed with the prebiotics galacto-oligosaccharides

and inulin during pregnancy and breastfeeding, and their

offspring were sensitized to wheat-gliadin after weaning [51].

Young animals showed a significantly reduced clinical and

cellular Th2 response while T-regulatory responses increased

and the intestinal barrier was preserved. Importantly, in this

study the intestinal microbiota in feces were investigated in

parallel in the offspring before sensitization, and showed that

the supplemented maternal diet was associated with a higher

total bacterial load, higher proportions of Lactobacillus and

Clostridium leptum, and lower abundance of Clostridium coc-

coides in the offspring. However, similar changes in micro-

biota composition were observed after allergy induction in

both offspring groups [51]. In another study, where diet and

microbiome and allergy induction were evaluated in parallel,

mice were fed a low-fiber diet before nasal sensitization with

house dust mite extract. These animals developed higher

local Th2 responses associated with increased mucus and

goblet cell hyperplasia. In parallel the composition of the

microbiome changed, with increased Erysipelotrichaceae in

the low-fiber group, while a high-fiber diet promoted Bater-

oidaceae and Bifidobacteriaceae [52]. The latter diet increased

circulating levels of SCFA and administration of the SCFA

propionate enhanced generation of macrophage and DC

precursors from bone marrow and subsequent presence of

dendritic cells with high phagocytic capacity in lung tissue,

associated with an impaired ability to induce Th2 effector cell

functions. These effects were shown to depend on GPR41, but

not GPR43 [52]. In another allergic OVA asthma mouse

model, dietary fiber intake significantly prevented clinical

symptoms, lowered eosinophil infiltration and goblet cell



metaplasia in nasal and lung mucosa, reduced serum OVA-

specific IgE levels as well as Th2 cytokines in NALF and BALF,

which was paralleled by increased Bacteroidetes and Actino-

bacteria, whereas Firmicutes and Proteobacteria were reduced

in fecal samples [53].

Development of novel dietary and microbiome

approaches to protect against food allergy

Clearly dysbiosis of the gut microbiome can negatively influ-

ence intestinal homeostasis. Thus, novel immunotherapeutic

strategies, mostly prebiotic and probiotic, but also fecal trans-

plantation approaches are being examined to modify bacterial

composition and metabolic activity and consequently improve

tolerance and regulatory responses within the mucosa [9].

Probiotics are defined as live microorganisms which when

administered in adequate amounts confer a health benefit on the

host [54]. This is a relatively recent definition, however hypoth-

esis relating to the beneficial effects associated with the con-

sumption of live microbes was initially proposed at the

beginning of 20th century by Metchnikoff. He observed a

connection between health and longevity of Bulgarian with

their daily diet, which contained fermented milk products [44].

T regulatory cells play a crucial role in blocking allergic

reactions. In vitro, it was found that probiotics such as Lacto-

bacillus rhamnosus can influence Foxp3 expression by Treg

cells [55]. In murine models, oral treatment with a probiotic

mixture VSL#3 protected against shrimp tropomyosin-in-

duced anaphylaxis [56]. Decreased IL-4, IL-5 and IL-13 secre-

tion was observed in parallel with increased levels of IFN-g,

IL-10, TGF-b, and IL-27 [56]. In humans, probiotic studies

have given mixed results, although oral administration of

Bifidobacterium longum 35624 resulted in increased circulating

levels of Foxp3+ lymphocytes, elevated ex vivo IL-10 secretion

and reduced serum proinflammatory biomarkers such as CRP

in patients with psoriasis, irritable bowel syndrome and

ulcerative colitis [57,58]. A double-blind, randomized, place-

bo controlled trial in 119 infants with cow milk allergy whose

diet was supplemented with combination of Lactobacillus

casei CRL431 and Bifidobacterium lactis Bb-12 did show signif-

icant beneficial effects. The percentage of tolerance to cow

milk at 6 and 12 months was 77% in the probiotics group

versus 81% in the placebo group [59]. Another study showed

that maternal consumption of Lactobacillus rhamnosus or

Bifidobacterium lactis probiotics can influence fetal immune

parameters and increase protective factors in breast milk [60].

This and many other findings (including mouse models,

where colonization of germ free offsprings resulted in re-

duced production of specific antibodies, compared to germ

free controls [61]) suggest that food supplementation with

probiotics may be most effective during pregnancy or during

the first months of life.

Prebiotics are food components which have beneficial

influence on composition and activity of human gut


microbiota, such as fiber. Fiber metabolism by colonic

bacteria results in the production of metabolites, such as

short chain fatty acids [62], the beneficial effects of which

are described above.

A novel approach as a potential microbiome therapy

against food allergy is microbiota fecal transplantation. Fecal

material from a healthy non-allergic donor is administrated

to the upper gastrointestinal tract and proximal colon of a

patient with dysbiosis [63]. There are many studies ongoing

examining fecal transplantation in the treatment of IBD, IBS,

obesity and Clostridium difficile infection [64], however no

data is currently available on the usefulness of fecal micro-

biome transplantation therapy in humans with food allergy.

Conclusions

Animal models of food allergy are an important tool in deci-

phering the complex in vivo molecular and cellular interac-

tions between the diet and microbiome, which protect against,

or promote, mucosal allergic responses. In addition, investi-

gators should carefully control for dietary and microbiome

parameters in their experimental models. These parameters

should be taken into account when attempting to translate

food allergy model results across different laboratories.

When establishing a food allergy animal model, the

composition of the diet and the microbiome has to

be considered specifically, as dietary antigens are

the triggering factor for eliciting an immune re-

sponse in this model system. It is important to

establish a stabilized animal model, as both diet

and microbiota play an important role in food

allergy and the response of immune system may

be influenced by microbiome–diet interactions.

Funding

During research for this article, partial support was obtained

by the Austrian Science Fund Grants KLI284, WKP039 and

SFB F4606-B28.



References[1] Pelz BJ, Bryce PJ. Pathophysiology of food allergy. Pediatr Clin North Am

2015;62(6):1363–75.

[2] Van Gramberg JL, O’Hehir RE, de Veer MJ, Meeusen EN, Bischof RJ. Use of

animal models to investigate major allergens associated with food allergy.

J Allergy 2013.

[3] Johnston LK, Chien KB, Bryce PJ. The immunology of food allergy. J

Immunol 2014;192(6):2529–34.

[4] Diesner SC, Olivera A, Dillahunt S, Schultz C, Watzlawek T, Forster-Waldl

E, et al. Sphingosine-kinase 1 and 2 contribute to oral sensitization and











effector phase in a mouse model of food allergy. Immunol Lett

2012;141(2):210–9.

[5] Pali-Scholl I, Jensen-Jarolim E. Anti-acid medication as a risk factor for

food allergy. Allergy 2011;66(4).

[6] Dunkin D, Berin MC, Mayer L. Allergic sensitization can be induced via

multiple physiologic routes in an adjuvant-dependent manner. J Allergy

Clin Immunol 2011;128(6):1251–8.

[7] Lindholm Bøgh K, van Bilsen J, Głogowski R, Lopez-Exposito I, Bouchaud

G, Blanchard C, et al. Current challenges facing the assessment of the


2016;6:21.

[8] Thorburn AN, Macia L, Mackay CR. Diet, metabolites, and ‘‘western-

lifestyle’’ inflammatory diseases. Immunity 2014;40(6):833–42.

[9] Frei R, Lauener RP, Crameri R, O’Mahony L. Microbiota and dietary

interactions: an update to the hygiene hypothesis? Allergy

2012;67(4):451–61.

[10] Blaser M. Antibiotic overuse: stop the killing of beneficial bacteria. Nature

2011;476(7361):393–4.

[11] Koplin J, Allen K, Gurrin L, Osborne N, Tang ML, Dharmage S. Is caesarean

delivery associated with sensitization to food allergens and IgE-mediated

food allergy: a systematic review. Pediatr Allergy Immunol 2008;8:682–7.

[12] Hill DA, Siracusa MC, Abt MC, Kim BS, Kobuley D, Kubo M, et al.

Commensal bacteria-derived signals regulate basophil hematopoiesis and

allergic inflammation. Nat Med 2012;18(4):538–46.

[13] Lee YK, Mazmanian SK. Has the microbiota played a critical role in the

evolution of the adaptive immune system? Science 2010;330(6012):1768–

73.

[14] Ley RE, Lozupone CA, Hamady M, Knight R, Gordon JI. Worlds within

worlds: evolution of the vertebrate gut microbiota. Nat Rev Microbiol

2008;6(10):776–88.

[15] Stefka AT, Feehley T, Tripathi P, Qiu J, McCoy K, Mazmanian SK, et al.

Commensal bacteria protect against food allergen sensitization. Proc Natl

Acad Sci U S A 2014;111(36):13145–50.

[16] Noval Rivas M, Burton OT, Wise P, Zhang YQ, Hobson SA, Garcia Lloret M,

et al. A microbiota signature associated with experimental food allergy

promotes allergic sensitization and anaphylaxis. J Allergy Clin Immunol

2013;131(1):201–12.

[17] Noval Rivas M, Burton OT, Oettgen HC, Chatila T. IL-4 production by

group 2 innate lymphoid cells promotes food allergy by blocking

regulatory T-cell function. J Allergy Clin Immunol 2016.

[18] Rodriguez B, Prioult G, Hacini-Rachinel F, Moine D, Bruttin A, Ngom-Bru

C, et al. Infant gut microbiota is protective against cow’s milk allergy in

mice despite immature ileal T-cell response. FEMS Microbiol Ecol

2012;79(1):192–202.

[19] Abrahamsson TR, Jakobsson HE, Andersson AF, Bjorksten B, Engstrand L,

Jenmalm MC. Low diversity of the gut microbiota in infants with atopic

eczema. J Allergy Clin Immunol 2012;129(2):434–40. e1–2.

[20] Lyons A, O’Mahony D, O’Brien F, MacSharry J, Sheil B, Ceddia M, et al.

Bacterial strain-specific induction of Foxp3+ T regulatory cells is

protective in murine allergy models. Clin Exp Allergy 2010;40(5):811–9.

[21] Bernard H, Ah-Leung S, Drumare MF, Feraudet-Tarisse C, Verhasselt V,

Wal JM, et al. Peanut allergens are rapidly transferred in human breast

milk and can prevent sensitization in mice. Allergy 2014;69(7):888–97.

[22] Paula-Silva J, Santiago AF, Oliveira RP, Rosa ML, Carvalho CR, Amaral JF,

et al. Effect of a protein-free diet in the development of food allergy and

oral tolerance in BALB/c mice. Br J Nutr 2015;113(6):935–43.

[23] Kim KS, Hong SW SW, Han D D, Yi J J, Jung J, Yang BG, et al. Dietary

antigens limit mucosal immunity by inducing regulatory T cells in the

small intestine. Science 2016;351(6275):858–63.

[24] Hodgkinson AJ, McDonald NA, Hine B. Effect of raw milk on allergic

responses in a murine model of gastrointestinal allergy. Br J Nutr

2014;112(3):390–7.

[25] Diesner SC, Schultz C, Ackaert C, Oostingh GJ, Ondracek A, Stremnitzer C,

et al. Nitration of beta-lactoglobulin but not of ovomucoid enhances

anaphylactic responses in food allergic mice. PLOS ONE

2015;10(5):e0126279.

[26] Mirotti L, Florsheim E, Rundqvist L, Larsson G, Spinozzi F, Leite-de-

Moraes M, et al. Lipids are required for the development of Brazil

nut allergy: the role of mouse and human iNKT cells. Allergy

2013;68(1):74–83.

[27] van den Elsen LW, van Esch BC, Dingjan GM, Hofman GA, Garssen J,

Willemsen LE. Increased intake of vegetable oil rich in n-6 PUFA enhances

allergic symptoms and prevents oral tolerance induction in whey-allergic

mice. Br J Nutr 2015;114(4):577–85.

[28] Kunisawa J, Arita M, Hayasaka T, Harada T, Iwamoto R, Nagasawa R, et al.

Dietary omega3 fatty acid exerts anti-allergic effect through the

conversion to 17,18-epoxyeicosatetraenoic acid in the gut. Sci Rep

2015;5:9750.

[29] van den Elsen LW, Meulenbroek LA, van Esch BC, Hofman GA, Boon L,

Garssen J, et al. CD25+ regulatory T cells transfer n-3 long chain

polyunsaturated fatty acids-induced tolerance in mice allergic to cow’s

milk protein. Allergy 2013;68(12):1562–70.

[30] Olivera A, Eisner C, Kitamura Y, Dillahunt S, Allende L, Tuymetova G,

et al. Sphingosine kinase 1 and sphingosine-1-phosphate receptor 2 are

vital to recovery from anaphylactic shock in mice. J Clin Invest

2010;120(5):1429–40.

[31] Suaini NH, Zhang Y, Vuillermin PJ, Allen KJ, Harrison LC. Immune

modulation by vitamin D and its relevance to food allergy. Nutrients

2015;7(8):6088–108.

[32] Singh A, Demont A, Actis-Goretta L, Holvoet S, Leveques A, Lepage M,

et al. Identification of epicatechin as one of the key bioactive

constituents of polyphenol-enriched extracts that demonstrate an anti-

allergic effect in a murine model of food allergy. Br J Nutr

2014;112(3):358–68.

[33] Abril-Gil M, Perez-Cano FJ, Franch A, Castell M. Effect of a cocoa-enriched

diet on immune response and anaphylaxis in a food allergy model in

Brown Norway rats. J Nutr Biochem 2016;27:317–26.

[34] Maggini S, Wintergerst ES, Beveridge S, Hornig DH. Selected vitamins and

trace elements support immune function by strengthening epithelial

barriers and cellular and humoral immune responses. Br J Nutr

2007;98(Suppl 1):S29–35.

[35] Untersmayr E, Diesner SC, Bramswig KH, Knittelfelder R, Bakos N,

Gundacker C, et al. Characterization of intrinsic and extrinsic risk factors

for celery allergy in immunosenescence. Mech Ageing Dev

2008;129(3):120–8.

[36] Roth-Walter F, Pacios LF, Gomez-Casado C C, Hofstetter G G, Roth GA

GA, Singer J, et al. The major cow milk allergen Bos d 5 manipulates T-

helper cells depending on its load with siderophore-bound iron. PLOS

ONE 2014;9(8):e104803.

[37] Roth-Walter F, Gomez-Casado C, Pacios LF, Mothes-Luksch N, Roth GA,

Singer J, et al. Bet v 1 from birch pollen is a lipocalin-like protein acting as

allergen only when devoid of iron by promoting Th2 lymphocytes. J Biol

Chem 2014;289(25):17416–21.

[38] Hoffmann PR, Jourdan-Le Saux C, Hoffmann FW, Chang PS, Bollt O, He

Q, et al. A role for dietary selenium and selenoproteins in allergic airway

inflammation. J Immunol 2007;179(5):3258–67.

[39] Smolinska S, Groeger D, Perez NR, Schiavi E, Ferstl R, Frei R, et al.

Histamine receptor 2 is required to suppress innate immune responses to

bacterial ligands in inflammatory bowel disease patients. Inflamm Bowel

Dis 2016;22(7):1575–1586.

[40] Smolinska S, Jutel M, Crameri R, O’Mahony L. Histamine and gut mucosal

immune regulation. Allergy 2014;69(4):273–81.

[41] Ferstl R, Frei R, Schiavi E, Konieczna P, Barcik W, Ziegler M, et al.

Histamine receptor 2 is a key influence in immune responses to intestinal

histamine-secreting microbes. J Allergy Clin Immunol 2014;134(3).

[42] Frei R, Ferstl R, Konieczna P, Ziegler M, Simon T, Rugeles TM, et al.

Histamine receptor 2 modifies dendritic cell responses to microbial

ligands. J Allergy Clin Immunol 2013;132(1):194–204.

[43] Maslowski KM, Vieira AT, Ng A, Kranich J, Sierro F, Yu D, et al. Regulation

of inflammatory responses by gut microbiota and chemoattractant

receptor GPR43. Nature 2009;461(7268):1282–6.

[44] Feehley T, Stefka AT, Cao S, Nagler CR. Microbial regulation of allergic

responses to food. Semin Immunopathol 2012;34(5):671–88.

[45] Feehley T, Nagler CR. Cellular and molecular pathways through which

commensal bacteria modulate sensitization to dietary antigens. Curr

Opin Immunol 2014;31:79–86.











































































































































[46] Zhang C, Zhang M, Pang X, Zhao Y, Wang L, Zhao L. Structural resilience

of the gut microbiota in adult mice under high-fat dietary perturbations.

ISME J 2012;6(10):1848–57.

[47] Kim YI, Yang JY, Ko HJ, Kweon MN MN, Chang SY. Shigella flexneri inhibits

intestinal inflammation by modulation of host sphingosine-1-phosphate

in mice. Immune Netw 2014;14(2):100–6.

[48] An D, Oh SF, Olszak T, Neves JF, Avci FY, Erturk-Hasdemir D, et al.

Sphingolipids from a symbiotic microbe regulate homeostasis of host

intestinal natural killer T cells. Cell 2014;156(1–2):123–33.

[49] Konieczna P, Ferstl R, Ziegler M, Frei R, Nehrbass D, Lauener RP, et al.

Immunomodulation by Bifidobacterium infantis 35624 in the murine lamina

propria requires retinoic acid-dependent and independent mechanisms.

PLoS ONE 2013;8(5):e62617.

[50] Maslowski KM, Mackay CR. Diet, gut microbiota and immune responses.

Nat Immunol 2011;12(1):5–9.

[51] Bouchaud G, Castan L, Chesne J, Braza F, Aubert P, Neunlist M, et al.

Maternal exposure to GOS/inulin mixture prevents food allergies and

promotes tolerance in offspring in mice. Allergy 2016;71(1):68–76.

[52] Trompette A, Gollwitzer ES, Yadava K, Sichelstiel AK, Sprenger N, Ngom-

Bru C, et al. Gut microbiota metabolism of dietary fiber influences allergic

airway disease and hematopoiesis. Nat Med 2014;20(2):159–66.

[53] Zhang Z, Shi L, Pang W, Liu W, Li J, Wang H, et al. Dietary fiber intake

regulates intestinal microflora and inhibits ovalbumin-induced allergic

airway inflammation in a mouse model. PLOS ONE 2016;11(2):e0147778.

[54] Guidelines for the evaluation of probiotics in food; 2002, Report of a Joint

FAO/WHO Working Group on Drafting Guidelines for the Evaluation of

Probiotics in Food.

[55] Karimi K, Kandiah N, Chau J, Bienenstock J, Forsythe P. A Lactobacillus

rhamnosus strain induces a heme oxygenase dependent increase in Foxp3+

regulatory T cells. PLoS ONE 2012;7(10):e47556.


[56] Schiavi E, Barletta B, Butteroni C, Corinti S, Boirivant M, Di Felice G. Oral

therapeutic administration of a probiotic mixture suppresses established

Th2 responses and systemic anaphylaxis in a murine model of food

allergy. Allergy 2011;66(4):499–508.

[57] Konieczna P, Groeger D, Ziegler M, Frei R, Ferstl R, Shanahan F, et al.

Bifidobacterium infantis 35624 administration induces Foxp3 T regulatory

cells in human peripheral blood: potential role for myeloid and

plasmacytoid dendritic cells. Gut 2012;61(3):354–66.

[58] Groeger D, O’Mahony L, Murphy EF, Bourke JF, Dinan TG, Kiely B, et al.

Bifidobacterium infantis 35624 modulates host inflammatory processes

beyond the gut. Gut Microbes 2013;4(4):325–39.

[59] Hol J, Elink Schuurman vLE, de Ruiter BE, Samsom LF, Hop JN, Neijens W,

et al., Cow’s Milk Allergy Modified by Elimination and Lactobacilli Study

Group. The acquisition of tolerance toward cow’s milk through probiotic

supplementation: a randomized, controlled trial. J Allergy Clin Immunol

2008;121(6):1448–54.

[60] Romano-Keeler J, Weitkamp JH. Maternal influences on fetal microbial

colonization and immune development. Pediatr Res 2015;77(1–2):

189–95.

[61] Morin S, Fischer R, Przybylski-Nicaise L, Bernard H, Corthier G, Rabot S,

et al. Delayed bacterial colonization of the gut alters the host immune

response to oral sensitization against cow’s milk proteins. Mol Nutr Food

Res 2012;56(12):1838–47.

[62] Frei R, Akdis M, O’Mahony L. Prebiotics, probiotics, synbiotics, and the

immune system: experimental data and clinical evidence. Curr Opin

Gastroenterol 2015;31(2):153–8.

[63] Kelly CR, Kahn S, Kashyap P, Laine L, Rubin D, Atreja A, et al. Update on

fecal microbiota transplantation 2015: indications, methodologies,

mechanisms, and outlook. Gastroenterology 2015;149(1):223–37.

[64] ClinicalTrials.gov. Clinical trials for fecal transplant.




























































DRUG DISCOVERY

TODAY

DISEASEMODELS

A review of animal models used toevaluate potential allergenicity ofgenetically modified organisms(GMOs)Nathan Marsteller1, Katrine L. Bøgh2, Richard E. Goodman1,

Michelle M. Epstein3,*1Food Allergy Research and Resource Program, University of Nebraska-Lincoln, 1901 North 21st Street, PO Box 886207, Lincoln, NE

68588-6207, USA2National Food Institute, Div. for Diet, Disease Prevention and Toxicology, Technical University of Denmark, Mørkhøj Bygade 19, DK-

2860 Søborg, Denmark3Medical University of Vienna, Department of Dermatology, Division of Immunology, Allergy and Infectious Diseases, Experimental

Allergy, Wahringer Gurtel 18-20, A-1090 Vienna, Austria


Editors-in-Chief




Food safety regulators request prediction of allergenic-

ity for newly expressed proteins in genetically modified

(GM) crops and in novel foods. Some have suggested

using animal models to assess potential allergenicity. A

variety of animal models have been used in research to

evaluate sensitisation or elicitation of allergic

responses. However, protocols for sensitisation and

challenge, animal species and strains, diets and other

environmental factors differ widely. We present a com-

prehensive review of published, peer-reviewed experi-

mental animal models used for the evaluation of

allergenicity of genetically modified organisms (GMOs).

Introduction

The prevalence of allergy and, in particular, food allergy with

potentially life threatening reactions has increased in the last

decades [1], without the identification of obvious environ-

*Corresponding author: M.M. Epstein ([email protected])



mental or genetic causal factors [2]. Food allergy is a complex

disease resulting from primary (sensitisation) and secondary

(elicitation) responses against food proteins. During sensiti-

sation in susceptible individuals, food proteins may induce

specific Th2-type allergic responses. T cells stimulate immu-

noglobulin class-switching in B cells to produce allergen-

specific IgE, which binds to mast cells and basophils, and

upon re-exposure to the allergen induces the release of med-

iators that elicit allergic symptoms. Although it is not clear

why food allergies are more prevalent now, some authors

suggest that this increase is due to widespread use of GM crops

for food production ever since their introduction in 1996

([3], http://www.globalresearch.ca/genetically-modified-

6.11.001 81


http://www.globalresearch.ca/genetically-modified-foods-unsafe-evidence-that-links-gm-foods-to-allergic-responses-mounts/7277






foods-unsafe-evidence-that-links-gm-foods-to-allergic-

responses-mounts/7277).

Is there a need for animal models in GMO allergenicity

assessment/evaluation?

An incomplete understanding of factors that affect allergic

sensitisation has driven the search for predictive strategies in

allergy risk assessment. International guidelines and regula-

tions from various countries state that GMOs should be

assessed for potential allergy risks based on the source of

the gene, amino acid sequence identity matches to known

allergens and stability to in vitro pepsin digestion (Box 1) [4,5].

In the European Union, however, the European Food Safety

Authority (EFSA) has recommended using animal models to

evaluate sensitising potential of novel proteins on a case-by-

case basis [6], even though there are no validated animal

models that are broadly predictive for allergenicity in

humans [4,7]. Nevertheless, if a highly predictive animal

model was developed, it would be useful for answering several

critical questions about the basic mechanisms underlying

food allergy (Box 2). Such models would likely improve the

risk assessment process for GMOs that do not have a clear

history of safe consumption by humans.

GMOs evaluated in animal models

To date, Cry1Ab and Cry1Ac proteins of Bacillus thuringiensis

(Bt) and grain from the transformed maize host have been the

most frequently tested materials in experimental animal

Box 2. Questions potentially addressed using experi-mental animal models

� Is a protein without a history of safe human dietary exposure likely

to sensitise and cause allergic reactions?

� Does the food matrix alter potential sensitisation, tolerance or

elicitation?

� Are there ‘threshold’ doses for sensitisation or elicitation using

various routes of exposure?

� How does proteolysis or heat processing alter the sensitising and

eliciting properties of an allergen?

Box 1. Primary risks and overall focus for evaluatingpotential risks of allergy from GMOs [4]

� Is the protein from the transferred gene an existing allergen (food,

airway or contact) as suggested by the allergenicity of the gene

source or sequence comparison to known allergens? If indicated,

perform serum IgE tests using samples from appropriately allergic

donors.

� Is the protein encoded by the transgene likely to cause cross-

reactions as suggested by even modest amino acid sequence identity

matches to known allergens? If so perform serum IgE tests.

� Are the characteristics of the protein similar to known common

food allergens; stable in pepsin at acidic pH and abundant in food

grade materials suggesting potential risk?


models. Yet these proteins have not been found to induce

allergy in animal models or in humans who have consumed

food produced from the GM crops (see Bt-related references in

Table 1). These crystal proteins are encoded by a non-aller-

genic source. They are relatively large proteins that are rapidly

digested in pepsin and have a low abundance in the GM crop.

In addition to Bt and its cloned Cry1 proteins, other GMOs

have been tested in vivo including alpha amylase inhibitor

(aAI) peas [8,9], PHA-E lectin in rice [10], sunflower seed

albumin in narrow leaf lupin [11] and lactoferrin [12].

Animal models for testing potential allergenicity of

GMOs

Several animal species have been fed material from GM plant

varieties, near isogenic or non-GM (nGM) varieties. The evalu-

ated animal responses included weight gain and overall health,

toxic effects or development of allergy. In particular, Bt-maize

expressing Cry1Ab has been studied in pigs [13,14], salmon

[15], sheep [16], cattle [17], zebrafish (cross generational feed-

ing) [18], rats [19,20], and mice [21]. Although a scientific

rationale could be argued for using one or more of these species

to evaluate nutritional or ecological impacts of agricultural

plant varieties, each species differs markedly from humans in

some physiological and immunological responses. Rodents are

the most frequently used model for food allergy, even though

there is little evidence that their responses are highly predictive

for ranking the allergenicity of diverse proteins in humans [22].

Thus, we sought to review data from studies in rodents and

other animal models looking for evidence that they are useful

for the evaluation of allergic sensitisation, elicitation and

adjuvant activity (Table 1). We have excluded studies designed

to evaluate nutritional or toxic properties without evaluating

potential immunogenicity or allergenicity.

Allergic sensitisation

Most of the studies identified in this review used rats or mice

of a single genetic strain. Sensitisation was accomplished with

either purified protein or extracts or feed containing whole

GM crop materials. The antigens were administered orally by

feeding or gavage, by dermal application, intraperitoneal (i.p.)

injection or intranasal (i.n.) dosing. In most studies, the

materials were provided repeatedly over time and in some

cases with added adjuvants, such as alum or cholera toxin. For

oral sensitisation, the length of daily exposure varied from 30

days to more than 90 days [21], and in a few cases multigen-

erational exposure was evaluated [23,24]. Finamore et al. [21]

fed mice with diets incorporating MON810 maize or control

maize for 30 or 90 days, evaluated CD4+ T cell counts in

peripheral blood and measured differences in levels of IL-6,

and IL-13 in the serum of both newly weaned and old mice.

They reported that the parameters varied little between groups

[21]. Although differences were found for non-antigen-specif-

ic immunological markers, clarification for these findings is



Vo

l. 1

7–18,

2015

D

rug

Disc

overy

To

day:

Dise

ase

Mo

dels

| In

viv

o a

nd

in v

itro m

od

els

of

foo

d a

llerg

y

Table 1. Summary of animal models used for assessing allergenicity and/or immunogenicity of GMOs.

Protein Animal strain

and sex

Sensitisation

route

# of immunisations Conventional

adjuvant

Challenge Measured

parameters

Notes Year/citation

Soluble Cry1Ac

protoxin from

E. coli JM103,

crystalline from

BTHD-73, or BSA

Female BALB/c

mice

i.g. or i.p. with

Mg-Al

hydroxide and

i.p. in PBS as

well as

3 times on weekly

intervals

Either no adjuvant,

cholera toxin, or Alum

No Specific IgM, IgG, and IgA Cry 1Ab produced from E. coli JM103

(pOS9300) obtained from D. Dean of Ohio

State University. Immunogenicity (IgA, IgG and

IgM) measured by i.g. or i.p. crystalline or

soluble Cry1 administration

1999 Vazquez-

Padron [47]

Cry1Ac protoxin Female BALB/c

mice

i.p., rectally, or

i.n.

3 times on weekly

intervals

No No Specific IgA, IgM, IgG in

sera, BAL, vaginal, small

intestine, and large

intestine wash fluids

Observed mucosal immunogenicity of Cry1Ac,

but no control protein used. The negative

control was PBS

2000 Moreno-

Fierros [34]

Cry1A protoxins

Cry1A (Cry1Aa,

Cry1Ab, and Cry1Ac)

(130–133 kDa),

Cry3A protoxin,

devoid of C-terminal

half

Male BALB/c

mice

i.p. or i.n. 3 times on weekly

interval

No No Specific IgG, IgM, and IgA Aimed to define immunogenic regions of Cry

proteins using i.p. and i.n. route with n = 5 mice

per treatment measuring IgG, IgM and IgA to

suggest N-terminal region more immunogenic,

no treatment replicates

2004 Guerrero

[36]

Seed meal from GM

aAI peas, nGM peas;

aAI purified from

beans; OVA;

SSA-Lupin; GM aAI

chickpeas or nGM

pinto beans

BALB/c mice (sex

was not

indicated)

i.g., or i.p. Twice per week for 4

weeks

No (positive control

mice received alum, but

treatment mice did not)

By airway (asthma) and

in foot pad

(subcutaneously for

delayed type

hypersensitivity

responses)

Footpad thickness

measured, specific IgG1;

mucus secreting cells,

eosinophils, and Th2 cell

number

Multiple sensitisation and challenge schemes.

Indicates i.g. aAI peas increased OVA

responses. Compare to Lee et al. [8]

2005 Prescott [9]

Cry1A toxins

(Cry1Aa, Cry1Aa8

and Cry1Ab2)

Male BALB/c

mice

i.n. 1 Various compounds:

LPS, DT, ConA, GalNAc.

Re-stimulation of

immune cells in vitro

with Cry 1 proteins

IgG1, IgG2a, cytokines:

Th1 (IFNg, IL-12p70);

Th2 (IL-10, IL-4)

Wild-type Cry1A induced Th1 responses, but

not Th2 responses

2007 Guerrero

[48]

Bt-MON810 maize

diet, Cry1Ab

Male BALB/c

mice

Inclusion diet:

50% MON810

or parental

control maize

flour

Diet given to recently

weaned or old mice

for 30 and 90 days (old

only 90 days)

No No IELs, spleen

lymphocytes, IL-4, IL-5,

IL-6, IL-10, IL-12p70, IL-

13, IFN g, TNF-R, MCP-

1 (CCL2), and mMCP-1

Evaluated potential immunotoxicology of GM,

parental and commercial maize lines.

deoxynivalenol mycotoxin was higher in GM

compared to non-GM, and immune markers of

inflammation in mice fed GM were reportedly

higher in number and statistics, but no clear

associations were found to suggest harm

2008 Finamore

[21]

Cry1Ab, peanut

protein extract

Female BALB/c

mice

i.g. 5 times on days 1, 7,

13, 19 and 25

Cholera toxin in

comparison to

adjuvanticity of Cry1Ab

Yes (intra-tracheal) Specific IgE, IgG1 and

IgG2a and Th1/Th2/

Th17 cytokine,

bronchoalveolar lavage

fluid (BAL), and

splenocytes analysed

Sensitisation to peanut proteins only observed

in mice sensitised with PE and CT, as measured

by T cell responses. No Cry1Ab adjuvant

activity. Conclusion: Cry1Ab did not

demonstrate adjuvant activity compared to

cholera toxin

2008 Guimaraes

[49]

PHA-E transgenic rice

or Cry1Ab, with or

without added

purified Cry1Ab

or PHA-E

Male and female

Wistar rats

Dietary and

Inhalation

28 day and 90 day

feeding

No No Specific IgM, IgG1, IgG2a,

IgA antibody to PHA-E

and Cry1Ab and total

IgM, IgG, and IgA

PHA-E lectin had an immune modulatory effect,

but Cry1 did not

2008 Kroghsbo

[10]

Recombinant

Cry1Ac protoxin

with Naegleria

fowleri lysate

Male BALB/c

STAT6++ and

STAT6�/� mice

i.n. 4 times on weekly

interval

No Challenged with lethal

doses of N. fowleri

trophozoites

Th2, IgG1, IL-4, IFNg, IL-

12, IgG2a, IgA, IgM

Assessed adjuvanticity of Cry1Ac and

conferred protection to lysate

2010 Carrasco-

Yepez [50]

ww

w.d

rugd

iscoveryto

day.co

m

83

Dru

g D

isco

very

To

day:

Dise

ase

Mo

dels

| In

viv

o a

nd

in v

itro m

od

els

of

foo

d a

llerg

y

Vo

l. 1

7–18,

2015


Protein Animal strain

and sex

Sensitisation

route

# of immunisations Conventional

adjuvant

Challenge Measured

parameters

Notes Year/citation

Soluble Cry1Ac from

E. coli JM103

Male BALB/c

mice

i.n. 4 times on weekly

intervals

No (controls received

cholera toxin)

No Specific IgA and IgG,

phenotypic and

activation analysis, IL-4,

IL-5, and IL-10

E. coli JM103 (pOS9300) with Cry1Ac insert

was obtained from D. Dean of Ohio State

University. Endotoxin in Cry1Ac quantified.

2010 Rodriguez-

Monroy [51]

Cry1Ab, BLG, Ara h 1,

KLH

Female BALB/cJ

mice

i.g. or i.p. 5 for i.g. or 2 for i.p. Incomplete Freund’s

adjuvant or cholera toxin

Yes Specific IgE, IgG1, IgG2a,

cytokines, murine

metabolic biomarkers

Observed Cry1Ab is immunogenic, but does

not have allergenic potential. Endotoxin

quantified in test proteins.

2011 Adel-Patient

[52]

Bt-MON810 (Cry1Ab)

maize diet and

non-GM maize

Female swine Diet Sows fed daily for 143

days during gestation,

then lactation

No No Cry1Ab-specific

antibody, leukocyte

phenotyping,

hematology

Sows fed MON810 maize (Cry1Ab) or non-

GM through gestation and lactation. Immune

function evaluated including tests for Ab to Cry

1Ab in sows and piglets, which were negative

2012 Buzoianu

[23]

Native human milk

lactoferrin (LF) and

recombinant (rLF)

in Aspergillus or rice

Female BALB/c

mice

i.p. 2 or 3 times on weekly

intervals

No No Specific IgE, IgG1, IgG2a,

Th1 and Th2 cytokines

Endotoxin quantified. LF was more

immunogenic then rLF. Mannose- and fucose

(Lex)-containing ligands have adjuvant

properties depending on glycan profile

2013 Almond [12]

Bt-MON810 maize

diet, Cry1Ab

Atlantic salmon Diet 33 day or 97 day

feeding trial

No No Histomorphology of

main organs, mRNA

expression levels of

genes in distal intestine,

IgM

No specific anti-Cry1Ab IgM detected 2013 Gu [15]

OVA and transgenic

aAI from peas,

chickpeas and

cowpeas compared

to non-transgenic

controls

Female BALB/c

mice

i.p., i.g., or i.n. 2 for i.p., 6 for i.n., i.g.

twice weekly for 4

weeks

No Challenged with 1%

OVA by an ultrasonic

nebulizer

Specific IgG1, IgG2a, IgE

in sera, lung and airway

inflammation and mucus

hypersecretion

No major differences were found between the

immune and inflammatory responses between

extracted proteins from GMs. The isogenic pea

induced immune responses to pea lectin that

were cross-reactive with aAI

2013 Lee [8]

Bt-MON810 maize

diet, Cry1Ab

Atlantic salmon Diet 99 day feeding trial No No Histological changes,

mRNA expression

levels, and inflammation

scored in distal intestine

Cry1Ab protein or other compositional

differences in GM Bt-maize may cause minor

alterations in intestinal responses in juvenile

salmon, while not affecting overall survival,

growth performance, development or health of

the animal

2014 Gu [53]

Bt-maize, nGM maize,

and OVA

Female BALB/c

mice

Diet for

Bt-maize, i.p.

or i.n. for OVA

Bt-maize: diet; OVA

i.p. 2 times

No Yes (aerosol challenge

twice daily on 4 days)

Specific IgG1, IgG2a, IgE,

lung inflammation and

mucus hypersecretion

No adjuvant effect on allergic response to non-

cross-reactive OVA after diet containing Bt-

maize (Mon810)

2014 Reiner [43]

Bt-MON810 pollen or

leaves extract, Cry1Ab

from Bt spores, OVA,

and trypsinized Cry1Ab

from E. coli

Female BALB/c

mice

i.n. 6 times on days 0, 1, 2,

21, 22 and 23

Cholera toxin Yes OVA specific: IgE, IgG1,

IgG2a. MCP-1, BAL

cytokines

No adjuvant effect of pollen grains as

Allakhverdi et al. observed [54]. No treatment

replicates

2015 Andreassen

[3]

Bt-MON810 pollen or

leaves extract, Cry1Ab

from Bt spores, and

trypsinized Cry1Ab

from E. coli

Female BALB/c

mice

i.n. 6 times on days 0, 1, 2,

21, 22 and 23

No Yes IgE, IgG1, IgG2a, MCP-1,

BALF cytokines

Claim that Bt spores are not a good source and

that E. coli trypCry1Ab may be more relevant,

but their results indicate that trypsinised

protein is more immunogenic then Mon810.

There are no cutoff values for IgE. There are no

treatment replicates

2015 Andreassen

[35]

84

w

ww

.dru

gdisco

verytoday.co

m


necessary [21]. Studies using i.p. injection for sensitisation

typically used three doses separated by seven days. Allergen

was used for clinical challenges. Blood was usually collected

between the third and seventh day after the last injection and

was used for measuring specific IgE concentrations.

Measured disease parameters

The readouts measured in animals are similar to those used or

observed to evaluate allergy in humans. For example, a

primary marker of sensitisation in humans is antigen-specific

IgE. Antigen-specific IgE or IgG1 levels were frequently mea-

sured in exposed mice as useful markers of a Th2 response and

potential allergy. Additional markers including a differential

measure of cytokines (IL-4, IL-5, IL-13 vs. IFN-g, IL-2, IL-10)

were sometimes measured from direct protein assays or

mRNA detection. While antibody binding demonstrates im-

mune recognition of a specific antigen, clinical manifesta-

tions of allergic responses require activation and

degranulation of mast cells and basophils as a result of IgE

binding (or possibly IgG1 in mice) to two or more epitopes on

a single allergen. Additional tests in rodents and other species

include protein-specific dermal mast cell degranulation with

either active- or passive-cutaneous anaphylaxis, which

mimics skin prick tests (SPT) with allergenic extracts to

diagnose humans [25]. Allergen-specific production of Th2

cytokines, release of histamine and mast cell protease usually

correlate with in vivo signs including anaphylaxis, hypother-

mia, hypotension or reduced pulmonary function in various

animal models [26–29]. Allergic sensitisation and elicitation

are complex processes that manifest differently in allergic

individuals depending on genetic and environmental factors.

Thus, it is not surprising that animal models may not mimic

all clinical responses in humans.

Factors that may limit predictability of animal models

Animal models are useful for mechanistic insights in allergy

and may be useful for assessing allergenicity of GMOs. How-

ever, the lack of a complete understanding of the factors that

impact sensitisation in humans creates obstacles for the

development of a predictable animal model.

A major problem in this field is the lack of standardised

models for testing novel foods and GMO allergenicity. The

models are designed with different sensitisation protocols,

species or genetic strains, routes of allergen exposure (e.g.,

oral, inhalation, gastro-intestinal, dermal and intraperitone-

al), added adjuvants (e.g., cholera toxin, alum, lectins, and

lipids) and GMO test materials (e.g., purified proteins in their

native conformation or denatured, whole food matrix, con-

taminants like endotoxin), which markedly influence sensi-

tisation and elicitation responses [30–33]. It is possible that

certain genetic differences between animal strains will result

in disparate responses to specific proteins, which implies that

similar experiments with different strains in the same lab

might be necessary to fully assess materials. There is also the

possibility that the GM materials contain cross-reactive pro-

teins. Notably, both nGM and aAI peas upon consumption in

mice induced allergic responses upon re-challenge that were

caused by the cross-reactive pea lectin [8]. Different protocols,

animal strains, and materials as well as lab-to-lab variation

often lead to disparate experimental outcomes and low

predictability. It is important to consider whether conflicting

results from different models will alter risk assessment for

human food safety.

Another challenge for establishing predictable models is

the inclusion of appropriate negative and positive controls.

When testing food grade material from a GMO, ideally a near

isogenic line with overall minimum genetic diversity com-

pared to the GMO and one or more genetically diverse

commercial lines with similar intended use would be in-

cluded in separate treatment test groups. In many of the

studies (Table 1), the authors have not included a positive

control, that is, crop materials that will induce a strong

allergic response as a comparator because, in many cases,

the perfect positive control does not exist. This is also the

case when using a GM protein, which should be compared

with (1) a protein inducing no allergic response, (2) an

allergenic protein and (3) vehicle alone. For example, in

one mouse study, the protein Cry1Ac from Bt induced

immune responses in mice, but was only compared to

the vehicle [34].

An important consideration for experimental animal mod-

els is that there is often high variation in the induced immune

response and thus, it is important to test an appropriate

number of animals and perform a sufficient number of ex-

perimental replicates to assess biological variance. For the

inclusion of animal experiments into a risk assessment, it is

essential to perform intra- as well as inter-lab comparisons,

using the same test materials because there may be external or

internal (e.g., gastrointestinal microbiota) environmental

variables that could influence the outcome. Nevertheless,

there are published animal studies assessing GMO allergenic-

ity with only one experiment [3,24,35,36] that may be the

result of strict animal ethics rules which may prohibit the

replication of experiments. However, biological repeats are

necessary to ensure that results are not biased by undefined

variables. The ‘one experiment approach’ may lead to unsub-

stantiated results that are unsuitable for risk assessment.

Experimental models used for risk assessment must be

reproducible. Inter-experiment and inter-lab variations are

expected, but experimental protocols should be designed in a

way that the test results are related to controls, thereby

allowing comparisons of results between laboratories. When

results in an experimental animal model are contradictory, as

they were for aAI GM peas [8,9], it is impossible to conclude

whether the GMO is allergenic. Notably, these two labs used

the same materials, protocol, and mouse strain and yet, the



results were contradictory, emphasising the importance of

repeated experiments in independent laboratories [8,9].

In addition, evaluating the history of safe use of introduced

genes and proteins as well as consideration of any cooking or

processing that would normally be used to prepare food from

the GM source, should be evaluated for impacts on the

specific GM protein levels, structure and immune reactivity.

The aAI gene was transferred from common beans (Phaseolus

vulgaris) into peas, chickpeas and cowpeas to inhibit damage

caused by the bruchid seed storage beetles. The aAI pea is

notable because (1) there are no reports of bean aAI allergy in

humans thus, there is a history of safe use of aAI; (2) aAI is

produced at high levels in the GMO (1–2% of protein) [37]; (3)

there is differential post-translational modifications depend-

ing on the host plant [38], which may lead to new confor-

mational allergenic epitopes leading to potential

allergenicity; and (4) food processing is an important factor

in food safety because we consume beans and other legumes

including peas after cooking, which inactivates a number of

protease inhibitors or lectins [39]. The aAI pea raises several

important points for food safety evaluation: (1) when there is

low GM protein expression, the probability of associated

allergenicity is low (e.g., Cry 1Ab in Mon810 represents

approximately 0.01% of the total crude protein in wholegrain

maize [40]), (2) it is necessary to consider post-translational

modifications of the proteins, (3) it is essential to evaluate

such GM materials upon heat-treatment [8], and (4) it is

necessary to measure protein levels in processed food and

feed products.

While some animal models test whole GMO materials

(including the food matrix), there are also experiments in

which the isolated or recombinant GM proteins (purified,

isolated or recombinant GM proteins (proteins derived from

Escherichia coli or other GM microbes)) were tested

[8,12,34,36,41,42]. When GM proteins are derived from dif-

ferent sources and processes, there might be a response in the

animal that is unrelated to the GM protein. For instance,

lipopolysaccharide (LPS) skews the immune response to pro-

teins [31] and might be a contaminant along with targeted

recombinant proteins such as Cry1, which might explain

disparate results between studies [3,43]. Lectins and carbohy-

drate binding proteins are present in plants and antigen

presenting cells have receptors that bind different classes

of lectins. Some lectins stimulate antigen uptake, which

has the potential to influence immune responses to unrelated

proteins [33,44]. Careful characterisation of diet is essential as

many factors can influence the immune response. Improper

storage of the GM-food can also have drastic consequences if

fungal growth occurs, resulting in significant levels of afla-

toxin or other mycotoxins, potent toxins that may directly or

indirectly affect immune responses, as well as the fungal

structural carbohydrates such as chitin, an immune stimulat-

ing adjuvant [45].


With a better understanding of the factors that impact

sensitisation in humans, known obstacles can be avoided

when developing a predictive animal model. Some authors

have attempted to evaluate potential adjuvanticity of specific

GM proteins. For instance, Lee et al. found that neither aAI

peas nor Bt maize had adjuvant effects in mice [8,42], whereas

Prescott et al. found that consumption of peas together with

ovalbumin (OVA) increased OVA responses [9]. In one report,

the effect of cholera toxin as an adjuvant was confirmed in

the positive control group, but Cry1Ab’s adjuvant activity

was not assessed at a dose that is relevant to expression in the

plant [3,42]. Furthermore, there is little evidence that pure

proteins or proteins in the context of commonly consumed

food matrices act as adjuvants. It is important for researchers

to characterise the proteins and GMO raw materials used in

tests to prove identity and appropriate biochemical structure

and function if the tests are to be useful. For example, as

mentioned above, many plant proteins are modified post-

translationally by proteolysis or covalent addition of lipids or

carbohydrates (e.g., asparagine-linked glycosylation) [9,46].

While a predictive animal model of allergenicity would be

of great value, it is worth considering the possibility that a

perfect model may not come to fruition. Without a single

validated animal model, scientists and regulators will need to

carefully consider the positives and negatives of a given model

and determine the relevance of the results based on careful

analysis of the controls, immune markers, protein characteri-

sation, and animals used on a case-by-case basis. Further

developments such as in vitro and ex vivo models (discussed

elsewhere in this section) that take into consideration high

genetic variation in the human population and environmen-

tal factors, for example, microbial skewing, might also lead to

improved risk assessment of GMOs and novel foods.

Conclusions

The major risks of food allergy are minimised by evaluating

the source, amino acid sequence similarity to allergens and

when indicated, testing for specific serum IgE. Nevertheless,

further risk reduction by identifying the allergenic potential

of novel foods including GMOs using in vitro and in vivo assays

would be valuable. Experimental animal models are particu-

larly useful for understanding the mechanisms underlying

the allergic response to food. However, there are many po-

tential limitations that hinder the development of standar-

dised and validated animal models used for predicting GMO

allergenicity. There is a pressing need to validate experimen-

tal models with whole food materials and known allergenic,

as well as non-allergenic, food proteins in carefully controlled

experiments using the best-suited species and strains and

ensuring statistical power. For the successful use of animal

models in allergenicity risk assessment, a consensus approach

must be identified with sufficient predictive power to mimic

human allergic risks.


Acknowledgements

We would like to thank Anne-Marie Bakke, Ashild Krogdahl,

Janina Krumbeck and Peadar Lawlor for their critical reading

of the manuscript.

References[1] Sicherer SH, Sampson HA. Food allergy: epidemiology, pathogenesis,

diagnosis, and treatment. J Allergy Clin Immunol 2014;133:291–307.

[2] Berin MC, Sampson HA. Food allergy: an enigmatic epidemic. Trends

Immunol 2013;34:390–7.

[3] Andreassen M, Bohn T, Wikmark OG, Van den Berg J, Lovik M, Traavik T,

et al. Cry1Ab protein from Bacillus thuringiensis and MON810 cry1Ab-

transgenic maize exerts no adjuvant effect after airway exposure. Scand J

Immunol 2015;81:192–200.

[4] Codex Alimentarius Commission. Foods derived from modern

biotechnology. 2nd ed. FAO/WHO; 2009. p. 1–85.

[5] United States Food and Drug Administration. Foods derived from new

plant varieties; policy statement. FDA Fed Regist 1992;57:22984.

[6] European Food Safety Authority. Scientific Opinion on Guidance on the

risk assessment of genetically modified microorganisms and their

products intended for food and feed use. EFSA J 2011;9(6):2193.

[7] Goodman RE, Vieths S, Sampson HA, Hill D, Ebisawa M, Taylor SL, et al.

Allergenicity assessment of genetically modified crops – what makes

sense? Nat Biotechnol 2008;26:73–81.

[8] Lee RY, Reiner D, Dekan G, Moore AE, Higgins TJ, Epstein MM. Genetically

modified alpha-amylase inhibitor peas are not specifically allergenic in

mice. PLoS ONE 2013;8:1–9.

[9] Prescott VE, Campbell PM, Moore A, Mattes J, Rothenberg ME, Foster PS,

et al. Transgenic expression of bean alpha-amylase inhibitor in peas results

in altered structure and immunogenicity. J Agric Food Chem

2005;53:9023–30.

[10] Kroghsbo S, Madsen C, Poulsen M, Schroder M, Kvist PH, Taylor M, et al.

Immunotoxicological studies of genetically modified rice expressing PHA-

E lectin or Bt toxin in Wistar rats. Toxicology 2008;245:24–34.

[11] Smart V, Foster PS, Rothenberg ME, Higgins TJ, Hogan SP. A plant-based

allergy vaccine suppresses experimental asthma via an IFN-gamma and

CD4+CD45RBlow T cell-dependent mechanism. J Immunol

2003;171:2116–26.

[12] Almond RJ, Flanagan BF, Antonopoulos A, Haslam SM, Dell A, Kimber I,

et al. Differential immunogenicity and allergenicity of native and

recombinant human lactoferrins: role of glycosylation. Eur J Immunol

2013;43:170–81.

[13] Custodio MG, Powers WJ, Huff-Lonergan E, Faust MA, Stein J. Growth,

pork quality, and excretion characteristics of pigs fed Bt corn or non-

transgenic corn. Can J Anim Sci 2006;86:461–9.

[14] Walsh MC, Buzoianu SG, Gardiner GE, Rea MC, Ross RP, Cassidy JP, et al.

Effects of short-term feeding of Bt MON810 maize on growth

performance, organ morphology and function in pigs. Br J Nutr

2012;107:364–71.

[15] Gu JN, Krogdahl A, Sissener NH, Kortner TM, Gelencser E, Hemre GI, et al.

Effects of oral Bt-maize (MON810) exposure on growth and health

parameters in normal and sensitised Atlantic salmon, Salmo salar L.. Br J

Nutr 2013;109:1408–23.

[16] Trabalza-Marinucci M, Brandi G, Rondini C, Avellini L, Giammarini C,

Costarelli S, et al. A three-year longitudinal study on the effects of a diet

containing genetically modified Bt176 maize on the health status and

performance of sheep. Livest Sci 2008;113:178–90.

[17] Paul V, Steinke K, Meyer HH. Development and validation of a sensitive

enzyme immunoassay for surveillance of Cry1Ab toxin in bovine blood

plasma of cows fed Bt-maize (MON810). Anal Chim Acta 2008;607:106–13.

[18] Sanden M, Ornsrud R, Sissener NH, Jorgensen S, Gu JN, Bakke AM, et al.

Cross-generational feeding of Bt (Bacillus thuringiensis)-maize to zebrafish

(Danio rerio) showed no adverse effects on the parental or offspring

generations. Br J Nutr 2013;110:2222–33.

[19] Zeljenkova D, Ambrusova K, Bartusova M, Kebis A, Kovriznych J,

Krivosikova Z, et al. Ninety-day oral toxicity studies on two genetically

modified maize MON810 varieties in Wistar Han RCC rats (EU 7th

Framework Programme project GRACE). Arch Toxicol 2014;88:2289–314.

[20] El Sanhoty R, El-Rahman AA, Bogl KW. Quality and safety evaluation of

genetically modified potatoes spunta with Cry V gene: compositional

analysis, determination of some toxins, antinutrients compounds and

feeding study in rats. Nahrung 2004;48:13–8.

[21] Finamore A, Roselli M, Britti S, Monastra G, Ambra R, Turrini A, et al.

Intestinal and peripheral immune response to MON810 maize ingestion

in weaning and old mice. J Agric Food Chem 2008;56:11533–39.

[22] Ladics GS, Knippels LMJ, Penninks AH, Bannon GA, Goodman RE,

Herouet-Guicheney C. Review of animal models designed to predict the

potential allergenicity of novel proteins in genetically modified crops.

Regul Toxicol Pharm 2010;56:212–24.

[23] Buzoianu SG, Walsh MC, Rea MC, O’Donovan O, Gelencser E, Ujhelyi G,

et al. Effects of feeding Bt maize to sows during gestation and lactation on

maternal and offspring immunity and fate of transgenic material. PLoS

ONE 2012;7:e47851.

[24] Kilic A, Akay MT. A three generation study with genetically modified Bt

corn in rats: biochemical and histopathological investigation. Food Chem

Toxicol 2008;46:1164–70.

[25] Dearman RJ, Kimber I. Animal models of protein allergenicity: potential

benefits, pitfalls and challenges. Clin Exp Allergy 2009;39:458–68.

[26] Berin MC, Mayer L. Immunophysiology of experimental food allergy.

Mucosal Immunol 2009;2:24–32.

[27] Hogan SP, Wang YH, Strait R, Finkelman FD. Food-induced anaphylaxis:

mast cells as modulators of anaphylactic severity. Semin Immunopathol

2012;34:643–53.

[28] Finkelman FD. Anaphylaxis: lessons from mouse models. J Allergy Clin

Immunol 2007;120:506–15.

[29] Kind LS. Fall in rectal temperature as an indication of anaphylactic shock

in the mouse. J Immunol 1955;74:387–90.

[30] Berin MC, Zheng Y, Domaradzki M, Li XM, Sampson HA. Role of TLR4 in

allergic sensitization to food proteins in mice. Allergy 2006;61:64–71.

[31] Brix S, Kjaer TM, Barkholt V, Frokiaer H. Lipopolysaccharide

contamination of beta-lactoglobulin affects the immune response against

intraperitoneally and orally administered antigen. Int Arch Allergy

Immunol 2004;135:216–20.

[32] Bashir ME, Louie S, Shi HN, Nagler-Anderson C. Toll-like receptor 4

signaling by intestinal microbes influences susceptibility to food allergy. J

Immunol 2004;172:6978–87.

[33] Takata S, Ohtani O, Watanabe Y. Lectin binding patterns in rat nasal-

associated lymphoid tissue (NALT) and the influence of various

types of lectin on particle uptake in NALT. Arch Histol Cytol 2000;

63:305–12.

[34] Moreno-Fierros L, Garcia N, Gutierrez R, Lopez-Revilla R, Vazquez-Padron

RI. Intranasal, rectal and intraperitoneal immunization with protoxin

Cry1Ac from Bacillus thuringiensis induces compartmentalized serum,

intestinal, vaginal and pulmonary immune responses in Balb/c mice.

Microbes Infect 2000;2:885–90.

[35] Andreassen M, Rocca E, Bøhn T, Wikmark O-G, van den Berg J, Løvik M,

et al. Humoral and cellular immune responses in mice after airway

administration of Bacillus thuringiensis Cry1Ab and MON810 cry1Ab-

transgenic maize. Food Agric Immunol 2015;26:521–37.

[36] Guerrero GG, Dean DH, Moreno-Fierros L. Structural implication of the

induced immune response by Bacillus thuringiensis Cry proteins: role of the

N-terminal region. Mol Immunol 2004;41:1177–83.

[37] Schroeder HE, Gollasch S, Moore A, Tabe LM, Craig S, Hardie DC, et al.

Bean [alpha]-amylase inhibitor confers resistance to the pea weevil

(Bruchus pisorum) in transgenic peas (Pisum sativum L.). Plant Physiol

1995;107:1233–9.

[38] Morton RL, Schroeder HE, Bateman KS, Chrispeels MJ, Armstrong E,

Higgins TJ. Bean alpha-amylase inhibitor 1 in transgenic peas (Pisum

sativum) provides complete protection from pea weevil (Bruchus pisorum)

under field conditions. Proc Natl Acad Sci U S A 2000;97:3820–5.

[39] Pusztai A, Grant G. Assessment of lectin inactivation by heat and

digestion. Methods Mol Med 1998;9:505–14.

[40] Steinke K, Guertler P, Paul V, Wiedemann S, Ettle T, Albrecht C, et al.

Effects of long-term feeding of genetically modified corn (event MON810)





































































































































on the performance of lactating dairy cows. J Anim Physiol Anim Nutr

2010;94:e185–93.

[41] Legorreta-Herrera M, Meza RO, Moreno-Fierros L. Pretreatment with

Cry1Ac protoxin modulates the immune response, and increases the

survival of Plasmodium-infected CBA/Ca mice. J Biomed Biotechnol

2010;2010:198921.

[42] Moreno-Fierros L, Garcia-Hernandez AL, Ilhuicatzi-Alvarado D, Rivera-

Santiago L, Torres-Martinez M, Rubio-Infante N, et al. Cry1Ac protoxin

from Bacillus thuringiensis promotes macrophage activation by

upregulating CD80 and CD86 and by inducing IL-6, MCP-1 and TNF-

alpha cytokines. Int Immunopharmacol 2013;17:1051–66.

[43] Reiner D, Lee RY, Dekan G, Epstein MM. No adjuvant effect of

Bacillus thuringiensis-maize on allergic responses in mice. PLoS ONE

2014;9.

[44] Cardone M, Ikeda KN, Varano B, Belardelli F, Millefiorini E, Gessani S,

et al. Opposite regulatory effects of IFN-beta and IL-3 on C-type lectin

receptors, antigen uptake, and phagocytosis in human macrophages. J

Leukoc Biol 2014;95:161–8.

[45] Lee CG, Da Silva CA, Dela Cruz CS, Ahangari F, Ma B, Kang MJ, et al. Role

of chitin and chitinase/chitinase-like proteins in inflammation, tissue

remodeling, and injury. Annu Rev Physiol 2011;73:479–501.

[46] Campbell PM, Reiner D, Moore AE, Lee RY, Epstein MM, Higgins TJ.

Comparison of the alpha-amylase inhibitor-1 from common bean

(Phaseolus vulgaris) varieties and transgenic expression in other legumes—

post-translational modifications and immunogenicity. J Agric Food Chem

2011;59:6047–54.

[47] Vazquez-Padron RI, Moreno-Fierros L, Neri-Bazan L, de la Riva GA, Lopez-

Revilla R. Intragastric and intraperitoneal administration of Cry1Ac


protoxin from Bacillus thuringiensis induces systemic and mucosal antibody

responses in mice. Life Sci 1999;64:1897–912.

[48] Guerrero GG, Russell WM, Moreno-Fierros L. Analysis of the cellular

immune response induced by Bacillus thuringiensis Cry1A toxins in mice:

effect of the hydrophobic motif from diphtheria toxin. Mol Immunol

2007;44:1209–17.

[49] Guimaraes VD, Drumare MF, Ah-Leung S, Lereclus D, Bernard H,

Creminon C, et al. Comparative study of the adjuvanticity of Bacillus

thuringiensis Cry1Ab protein and cholera toxin on allergic sensitisation and

elicitation to peanut. Food Agric Immunol 2008;19:325–37.

[50] Carrasco-Yepez M, Rojas-Hernandez S, Rodriguez-Monroy MA, Terrazas

LI, Moreno-Fierros L. Protection against Naegleria fowleri infection in mice

immunized with Cry1Ac plus amoebic lysates is dependent on the STAT6

Th2 response. Parasite Immunol 2010;32:664–70.

[51] Rodriguez-Monroy MA, Moreno-Fierros L. Striking activation of NALT

and nasal passages lymphocytes induced by intranasal immunization

with Cry1Ac protoxin. Scand J Immunol 2010;71:159–68.

[52] Adel-Patient K, Guimaraes VD, Paris A, Drumare MF, Ah-Leung S,

Lamourette P, et al. Immunological and metabolomic impacts of

administration of Cry1Ab protein and MON 810 maize in mouse. PLoS

ONE 2011;6:e16346.

[53] Gu JN, Bakke AM, Valen EC, Lein I, Krogdahl A. Bt-maize (MON810) and

non-GM soybean meal in diets for Atlantic salmon (Salmo salar L.)

juveniles – impact on survival, growth performance, development,

digestive function, and transcriptional expression of intestinal immune

and stress responses. PLOS ONE 2014;9.

[54] Allakhverdi Z, Bouguermouh S, Rubio M, Delespesse G. Adjuvant activity

of pollen grains. Allergy 2005;60:1157–64.

























































drug discovery today disease models€¦ · drug discovery today: disease models vols. 17–18,...

Documents