Upload
others
View
8
Download
0
Embed Size (px)
Citation preview
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
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
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
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
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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
4 www.drugdiscoverytoday.com
[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
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
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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.
6 www.drugdiscoverytoday.com
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).
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
Sal s 1
Jug r 1
Pri p 3
Pis v 1
2
4
1 3
Drug Discovery Today: Disease Models
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
www.drugdiscoverytoday.com 7
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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]
8 www.drugdiscoverytoday.com
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
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
� Hosted in the Bioinformatics centre at CSIR-Institute of microbial
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
� Hosted in the Bioinformatics centre at CSIR-Institute of microbial
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
www.drugdiscoverytoday.com 9
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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
10 www.drugdiscoverytoday.com
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.
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
[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.
www.drugdiscoverytoday.com 11
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
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
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])
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
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
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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
14 www.drugdiscoverytoday.com
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
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
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)
Drug Discovery Today: Disease Models
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).
www.drugdiscoverytoday.com 15
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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
16 www.drugdiscoverytoday.com
[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.
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
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
www.drugdiscoverytoday.com 17
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
Drug Discovery Today: Disease Models
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
18 www.drugdiscoverytoday.com
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
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
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
www.drugdiscoverytoday.com 19
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
in vitro models to measure these important parameters and to
predict allergenicity of new proteins, in the future.
Conflict of interest
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.
20 www.drugdiscoverytoday.com
[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.
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
[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.
www.drugdiscoverytoday.com 21
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
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 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])
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
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
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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
24 www.drugdiscoverytoday.com
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,
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
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,
www.drugdiscoverytoday.com 25
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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
26 www.drugdiscoverytoday.com
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.
Conflict of interest
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.
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
[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.
www.drugdiscoverytoday.com 27
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
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
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])
1740-6757/$ � 2016 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ddmod.201
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.
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
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
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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
30 www.drugdiscoverytoday.com
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
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
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
www.drugdiscoverytoday.com 31
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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
Drug Discovery Today: Disease Models
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
32 www.drugdiscoverytoday.com
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
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
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
www.drugdiscoverytoday.com 33
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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.
34 www.drugdiscoverytoday.com
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.
Conflict of interest
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.
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
[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.
www.drugdiscoverytoday.com 35
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
[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.
36 www.drugdiscoverytoday.com
[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
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 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])
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
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
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
Allergen
Allergen specific epitope
Dendritic cell
Naïve T cellMHC calss II
MHC class II
TCR
TCR
B cell
Plasmacell
BasophilSpecifi IgE
CD300a
CD63
CD203c
Allergen specific epitope
Allergen specific epitope
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
Drug Discovery Today: Disease Models
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
38 www.drugdiscoverytoday.com
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
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
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
www.drugdiscoverytoday.com 39
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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
40 www.drugdiscoverytoday.com
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
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
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
www.drugdiscoverytoday.com 41
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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
conformational epitopes
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
conformational epitopes
Identification of only monoclonal responses
Laborious method
Low data output
[63,66]
Phage display technology Allows for identification of both linear and
conformational epitopes
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
42 www.drugdiscoverytoday.com
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
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
invent tools providing accurate information on the course of
the food allergic disease.
Conflict of interest
The authors have no conflict of interest to declare.
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
Allergy Clin Immunol 2014;134:746–9.
[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
www.drugdiscoverytoday.com 43
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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.
44 www.drugdiscoverytoday.com
[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
Allergy Clin Immunol 2002;110:293–7.
[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
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
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
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
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
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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.
46 www.drugdiscoverytoday.com
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
Drug Discovery Today: Disease Models
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
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
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].
www.drugdiscoverytoday.com 47
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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]
48 www.drugdiscoverytoday.com
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
Table 1 (Continued )
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
www.drugdiscoverytoday.com 49
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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.
50 www.drugdiscoverytoday.com
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.
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
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.
www.drugdiscoverytoday.com 51
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
[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.
52 www.drugdiscoverytoday.com
[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.
Gastroenterology 2011;141:1593–604.
[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.
Gastroenterology 2003;125:1419–27.
[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
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
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.
Gastroenterology 2005;129:985–94.
[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
Allergy Clin Immunol 2014;133:1390–9.
[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.
www.drugdiscoverytoday.com 53
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
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
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])
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
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
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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
56 www.drugdiscoverytoday.com
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
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
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
www.drugdiscoverytoday.com 57
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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
Drug Discovery Today: Disease Models
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].
58 www.drugdiscoverytoday.com
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
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
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
www.drugdiscoverytoday.com 59
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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
60 www.drugdiscoverytoday.com
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].
Conflict of interest
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
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
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.
[31] 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(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
www.drugdiscoverytoday.com 61
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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].
62 www.drugdiscoverytoday.com
[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
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
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])
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, 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
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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.
64 www.drugdiscoverytoday.com
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
Drug Discovery Today: Disease Models
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
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
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
www.drugdiscoverytoday.com 65
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
Anaphylaxis (score + temperature)
Cytokines
[45] Mouse; BALB/c WPE
Cashew
TD 4–6� (weekly) 1 mg None IG 1� 15 mg Allergen-specific IgG, IgE
Anaphylaxis (score + temperature)
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
Anaphylaxis (score + temperature)
MMCP-1, histamine
Cytokines
Cell population counts
[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
Anaphylaxis (score + temperature)
MMCP-1, histamine
Cytokines
Cell population counts
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
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
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.
Conflict of interest
The authors have no conflict of interest to declare.
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].
www.drugdiscoverytoday.com 67
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
[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.
[23] 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. 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.
68 www.drugdiscoverytoday.com
[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
allergenic capacity of food allergens in animal models. Clin Transl Allergy
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.
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
[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.
www.drugdiscoverytoday.com 69
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
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
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])
1740-6757/$ � 2016 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ddmod.201
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
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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]
72 www.drugdiscoverytoday.com
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
Table 1 (Continued )
Animal
model
Strain/mutation Antigen Adjuvant Treatment Reference
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]
www.drugdiscoverytoday.com 73
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
Table 1 (Continued )
Animal
model
Strain/mutation Antigen Adjuvant Treatment Reference
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
74 www.drugdiscoverytoday.com
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].
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
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
Drug Discovery Today: Disease Models
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
www.drugdiscoverytoday.com 75
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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].
76 www.drugdiscoverytoday.com
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
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
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
www.drugdiscoverytoday.com 77
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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
78 www.drugdiscoverytoday.com
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.
Conflict of interest
The authors have no conflict of interest to declare.
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
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
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
allergenic capacity of food allergens in animal models. Clin Transl Allergy
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.
www.drugdiscoverytoday.com 79
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
[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.
80 www.drugdiscoverytoday.com
[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
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 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])
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
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
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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?
82 www.drugdiscoverytoday.com
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
Table 1 (Continued )
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
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
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
www.drugdiscoverytoday.com 85
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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].
86 www.drugdiscoverytoday.com
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.
Vol. 17–18, 2015 Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy
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)
www.drugdiscoverytoday.com 87
Drug Discovery Today: Disease Models | In vivo and in vitro models of food allergy Vol. 17–18, 2015
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
88 www.drugdiscoverytoday.com
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.