Microarray CK+inflam

Toxicology and Applied Pharmacology 244 (2010) 144155
Contents lists available at ScienceDirect
Toxicology and Applied Pharmacology

j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y t a a p
Biomarkers of acute respiratory allergen exposure: Screening for

sensitization potential
Cherie M. Pucheu-Haston a,, Lisa B. Copeland b, Beena Vallanat b, Elizabeth Boykin b, Marsha D.W. Ward b
a
b
Curriculum in Toxicology, University of North Carolina-Chapel Hill, CB# 7270, Chapel Hill, NC 27599-7270, USA
National Health and Environmental Effects Research Laboratory, U.S. Environmental Protection Agency, 109 T.W. Alexander Drive, Research Triangle Park, NC 27711, USA
a r t i c l e
i n f o
Article history:
Received 9 November 2009
Revised 10 December 2009
Accepted 17 December 2009
Available online 4 January 2010
Keywords:
Asthma
Biomarkers
Hazard screening
Intratracheal aspiration
Gene expression microarray
Quantitative real-time polymerase chain
reaction (qRT-PCR)
a b s t r a c t
Effective hazard screening will require the development of high-throughput or in vitro assays for the
identication of potential sensitizers. The goal of this preliminary study was to identify potential biomarkers
that differentiate the response to allergens vs non-allergens following an acute exposure in nave individuals.
Female BALB/c mice received a single intratracheal aspiration exposure to Metarhizium anisopliae crude
antigen (MACA) or bovine serum albumin (BSA) in Hank's Balanced Salt Solution (HBSS) or HBSS alone. Mice
were terminated after 1, 3, 6, 12, 18 and 24 h. Bronchoalveolar lavage uid (BALF) was evaluated to
determine total and differential cellularity, total protein concentration and LDH activity. RNA was isolated
from lung tissue for microarray analysis and qRT-PCR. MACA administration induced a rapid increase in BALF
neutrophils, lymphocytes, eosinophils and total protein compared to BSA or HBSS. Microarray analysis
demonstrated differential expression of genes involved in cytokine production, signaling, inammatory cell
recruitment, adhesion and activation in 3 and 12 h MACA-treated samples compared to BSA or HBSS. Further
analyses allowed identication of 100 candidate biomarker genes. Eleven genes were selected for further
assessment by qRT-PCR. Of these, 6 demonstrated persistently increased expression (Ccl17, Ccl22, Ccl7,
Cxcl10, Cxcl2, Saa1), while C3ar1 increased from 624 h. In conclusion, a single respiratory exposure of mice
to an allergenic mold extract induces an inammatory response which is distinct in phenotype and gene
transcription from the response to a control protein. Further validation of these biomarkers with additional
allergens and irritants is needed. These biomarkers may facilitate improvements in screening methods.
2009 Elsevier Inc. All rights reserved.
Introduction
One particularly challenging area of immunotoxicology is the
identication of agents with a high likelihood to induce allergic
sensitization. Traditional approaches to screening for these agents
require time consuming multiple exposure protocols. Unfortunately,
these methods are becoming increasingly incapable of meeting the
screening needs of the rapidly expanding list of newly synthesized
and bioengineered compounds, not to mention re-screening those
agents that have been discovered to persist in the environment for
longer than originally expected. Efcient identication of problematic
agents will require a shift to more high-throughput methods of hazard
screening. It is probable that a multidisciplinary approach will be
required, where agents will be evaluated using multiple models, with
the nal decisions being made based upon the preponderance of
evidence.
Corresponding author. c/o U.S. Environmental Protection Agency 109 T. W.

Alexander Drive, Research Triangle Park, NC 27711, USA. Fax: +1 919 541 4284.
E-mail addresses: Pucheu-Haston.Cherie@epa.gov, cmpucheu@yahoo.com
(C.M. Pucheu-Haston).
0041-008X/$ see front matter 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.taap.2009.12.027
We hypothesize that there is a set of differentially expressed

gene biomarkers that distinguishes the immune response associated with the induction of allergic sensitization from non-allergic
immune responsiveness following a single exposure. The objectives
of this study were two-fold: (1) to determine whether we could
identify biomarkers of acute exposure to a known respiratory
sensitizer as compared to a poorly allergenic compound or vehicle
control; (2) to use this information to aid in the development of a
more extensive validation study. The ultimate goal of this work is to
translate these results into the development of an in vitro screening
assay.
To test our hypothesis, we selected the entomopathogenic fungus
Metarhizium anisopliae extract which we have previously shown to
induce robust responses characteristic of human allergic asthma in
adult BALB/c mice following multiple exposures (Ward et al., 1998,
2000). These responses can be obtained following intratracheal
aspiration (IA) exposure to M. anisopliae crude antigen (MACA).
Neither intraperitoneal priming nor adjuvant administration is
required for sensitization (Ward et al., 2000). As low or non-allergic
agent, we selected bovine serum albumin (BSA) which does not
induce allergic or asthma-like responses in our mouse model (Viana et
al., 2002; Ward et al., 2009).
C.M. Pucheu-Haston et al. / Toxicology and Applied Pharmacology 244 (2010) 144155
Methods
Animals. Eight-week-old female BALB/c mice were obtained from
Charles River Laboratories (Raleigh, NC). All mice were housed in
polycarbonate cages with hardwood chip bedding in an environmentally controlled, Association for Assessment and Accreditation of
Laboratory Animal Care International (AAALAC)-accredited animal
facility. Environmental enrichment was provided in the form of
mouse nestlets (Ancare Corporation, Bellmore, NY). Sentinel mice
were monitored serologically and were found to be free of Sendai,
mouse pneumonia, mouse hepatitis, other murine viruses, and
mycoplasma. Mice also were monitored for, and found to be free of,
ectoparasites and endoparasites. The animal holding room temperature (22 F 1.1 F) and relative humidity (50% 10%) were
maintained at the recommended levels. Mice received standard
rodent chow (Lab Diet 5P00, Purina Mills, St. Louis, MO) and water
ad libitum. Mice were acclimated 1 week at this facility prior to the
start of the experiment.
Antigen preparation.
M. anisopliae strain 1080 was obtained from
USDA-ARS Entomopathogenic Fungus Collection in Ithaca, NY.
Extracts of M. anisopliae mycelium and spores/conidia as well as
inducible enzymes ltrate were produced as previously described
(Ward et al., 1998). Briey, a fungal extract was prepared by a
modication of the methods described by (Stankus and O'Neil,
1988). Mycelium (hyphae growth) was grown at 27 C for 72 h in
Sabouraud's maltose broth with aeration (150 rpm). Subsequently,
the mycelium was collected on Whatman No. 1 lter paper, washed
twice with saline to remove media contaminants, and air-dried
overnight in a biosafety cabinet. The conidia, grown at 27 C on
Sabouraud's maltose agar, were harvested by gently scraping culture
plates of 2- to 3-week-old cultures. Each component was ground into
a paste in saline (1:2530 w/v) and stirred overnight at 4 C. The
supernatant was collected following centrifugation at 18,000g.
Production of inducible proteases and chitinases by M. anisopliae is
enhanced in the absence of readily available nitrogen and carbon.
Therefore, a deprivation medium of unpuried chitin (Sigma
Chemical Co., St. Louis, MO) at 3% in water was inoculated with M.
anisopliae and incubated at 27 C for 72 h with aeration (150 rpm).
The ltrate was retained following ltration through Whatman No. 1
lter paper. The proteins with a molecular weight greater than 3000
were concentrated using a YM-3 lter (MWCO 3000) (Amicon,
Beverly, MA) and assayed for total protein concentration (as
described below). For each component, the pH was adjusted to 6.0
followed by lter sterilization through a 0.2-um syringe lter
(Acrodisc; Gelman Sciences, Ann Arbor, MI). Equal protein concentrations of each component were combined to form the M.
anisopliae crude antigen (MACA). Dosing aliquots were stored at
20 C for a maximum of 10 wks until use. Extract endotoxin
concentrations were minimal (0.0140.007 ng per 20 g dose) as
measured by LAL assay (Lonza Group Ltd., Basel, Switzerland).
Bovine serum albumin (BSA), fraction V, was purchased from
Sigma-Aldrich (St. Louis, MO) and diluted in Hanks Balanced Salt
Solution (HBSS Gibco/Invitrogen, Carlsbad, CA). Dosing aliquots were
stored at 20 C until use.
Experimental design. Female BALB/c mice were anesthetized, then
given a single dose of MACA (20 g in 50 l HBSS), BSA (20 g in 50 l
HBSS) or HBSS alone (50 l) by intratracheal aspiration (IA) as
previously described (Ward et al., 2000). Briey, mice were
anesthetized by inhalation of a mixture of isourane and oxygen,
and then vertically suspended by their incisors to facilitate treatment
administration. Swallowing was prevented by gently pulling the
tongue out of the mouth in an upward direction. Antigen extract was
deposited into the oropharynx while the nose was briey occluded,
inducing aspiration of the extract. Six mice from each treatment group
145
were terminated by sodium pentobarbital overdose (Euthasol, Virbac

Animal Health, Fort Worth, TX) at each of six time points (1 h, 3 h, 6 h,
12 h, 18 h and 24 h post-treatment).
Bronchoalveolar lavage (BAL) and lung collection. Lung tissue and
bronchoalveolar lavage uid (BALF) samples were collected from all
mice at termination. The left lung was harvested and snap-frozen in
liquid nitrogen immediately after euthanasia, then stored at 80 C
until RNA isolation. The remaining lung tissue was lavaged twice
with 0.7 ml aliquots of HBSS. BALF aliquots for each animal were
pooled and stored on ice. The BALF was centrifuged at 800 rpm
(100 g) for 15 min at 4 C. Aliquots of BALF supernatant were
assayed for total protein and lactate dehydrogenase (LDH) activity
(as described below), and the remainder was stored at 20 C for
IgE quantication by ELISA. The cell pellet was resuspended in 1 ml
HBSS and cytospin preparations of 200 l BALF were made by
centrifugation onto glass slides (200 rpm, 10 min on a Shandon
Cytospin; Shandon Inc., Pittsburg, PA). Following Wright-Giemsa
staining (Fisher Scientic, Fairlawn, NJ), cells were differentially
counted at 200 cells per slide (one slide per animal). Total BALF cell
counts were obtained from the resuspended cells using a Coulter
counter (Coulter Corp., Miami, FL).
Total protein and lactate dehydrogenase (LDH) assays. BALF samples
and fungal extract components were assayed for total protein using
Pierce Coomassie Plus Protein Assay Reagent (Pierce/Thermo Fisher
Scientic, Rockford, IL). Concentrations were determined from a
standard curve using BSA standards obtained from Sigma Chemical
Co. (St. Louis, MO). BSA works well as a general-purpose standard for
protein quantitation. However, due to differences in amino acid
concentrations and protein extinction coefcients, there may be some
imprecision in the measurement of any complex protein mixture
(such as a fungal extract or lavage uid) with any single protein
standard.
BALF samples were assayed for LDH activity using a commercially
prepared kit and controls from Sigma Chemical Co. Both the total
protein and LDH assays were modied for use on the KONELAB 30
clinical chemistry Spectrophotometer analyzer (Thermo Clinical
Labsystems, Espoo, Finland).
RNA isolation. Total RNA was isolated from all frozen lung tissue
samples using a RNeasy Mini kit (Qiagen Inc., Valencia, CA) according
to manufacturers instructions. Briey, whole left lungs were kept on
dry ice before homogenization in 1 ml of RLT lysis buffer using an
Omni TH homogenizer (Omni International, Marietta, GA). Lung
homogenate (350 l) was added to an equal volume of 70% ethanol,
vortexed and added to an RNeasy mini spin column. After
centrifugation, the column was washed once with RW1 buffer and
twice with RPE buffer. RNA was then eluted by the addition of RNAsefree water. RNasin ribonuclease inhibitor (Promega Corporation,
Madison, WI) was added immediately after isolation to increase RNA
stability. RNA quantication was performed using a NanoDrop ND1000 spectrophotometer (NanoDrop Technologies, Inc., Wilmington,
DE). RNA quality was assessed using an Agilent 2100 bioanalyzer
with the Agilent RNA 6000 nano kit (Agilent Technologies, Santa
Clara, CA).
Microarray hybridization.
Three to 4 unique RNA samples from all 3
treatment groups at both the 3 and 12 h post-treatment time points
were submitted for microarray gene expression analysis (MACA 3 h
n = 3, BSA 3 h n = 3, HBSS 3 h n = 3; MACA 12 h n = 3, BSA 12 h
n = 4, HBSS 12 h n = 4; total n = 24). These sample times were
chosen based upon a tendency for the BALF inammatory response to
peak at 6 and 18 h, or both (see Figs. 13). The microarray chosen for
this study was the Affymetrix Mouse Genome 430A 2.0 (Affymetrix
Inc., Santa Clara, CA, www.affymetrix.com), which contains over
146
Fig. 3. BALF total celullarity. IA MACA induces a rapid increase in total airway cellularity
relative to BSA- or HBSS-treated mice. P b 0.01, P b 0.001 relative to HBSS; P b 0.05,
P b 0.01, P b 0.001 relative to BSA.
Fig. 1. Overview of microarray data analysis strategy. RNA was harvested from snapfrozen lung tissue and hybridized to a gene expression microarray (Affymetrix Mouse
Genome 430A 2.0). Sequences that were differentially expressed between HBSS control
groups and MACA or BSA treatment groups were identied by one-way ANOVA and
Benjamini and Hochberg FDR (p b 0.01) followed by Scheffe' post hoc analysis (p b 0.05).
Genes that changed 2-fold or greater relative to HBSS were selected for further analysis.
Sequences differentially expressed in MACA-treated animals were identied using
GeneGo/Metacore software. Sequences were then ltered for biomarker potential
using the biomarker function of Ingenuity Pathway Analysis software. Selected
sequences were subjected to functional analysis using DAVID (Database for Annotation,
Visualization and Integrated Discovery). Expression of putative biomarkers was then
evaluated using quantitative real-time PCR.
Fig. 2. BALF LDH and total protein levels. (A) Intratracheal administration (IA) of MACA
is associated with little acute cellular toxicity (LDH). (B) IA MACA induces a rapid
increase in BALF total protein levels, suggesting the development of pulmonary cellular
leakage and/or edema. N = 6 per treatment per time point. P b 0.01, P b 0.001
relative to HBSS; P b 0.001 relative to BSA.
22,500 probe sets representing 14,500 well-characterized mouse

genes. A total of 21 gene chips (one per mouse) were used in this
study.
Overall gene expression data analysis strategy. The analysis of the
microarray data involved the comparison of a total of six data sets
from two time points (MACA 3 h, BSA 3 h, HBSS 3 h; MACA 12 h, BSA
12 h, HBSS 12 h). Identication of putative biomarkers was performed
in a stepwise fashion: (1) identify sequences differentially expressed
in either MACA- or BSA-treated animals as compared to time-matched
HBSS controls, (2) determine sequences differentially expressed in
MACA-treated samples only, (3) perform biomarker lter enrichment,
(4) perform functional expression analysis on biomarker-ltered
data, and (5) select putative biomarkers. Additionally, differentially
expressed sequences were subjected to pathway analysis to identify
fundamental differences in the immune response to compounds of
high versus low allergenicity (Fig. 1).
Statistical analysis of microarray data.
Data (.cel les) were
analyzed and statistically ltered using Rosetta Resolver version
7.1 software (Rosetta Inpharmatics, Kirkland, WA). For analysis
purposes, the data from individual 25-mer oligonucleotide probe
pairs were combined into sequence data. Each sequence represents a
portion of a gene transcript, and is generally comprised of 11
oligonucleotide probe pairs (reporters). A given gene may be
represented by one or more sequences, which allows for detection of
alternate transcripts.
Sequence expression intensities were averaged for each treatment
group. Sequences that were differentially expressed within a time
point between HBSS control groups and MACA or BSA treatment
groups were identied in Rosetta Resolver by one-way ANOVA with
Benjamini and Hochberg False Discovery Rate (FDR) analysis
(p b 0.01), followed by Scheffes post hoc data analysis (p b 0.05).
Sequences identied as statistically different in intensity in either
MACA or BSA samples as compared to HBSS were selected for ratio
analysis. These sequences were then further ltered using an
experimental/control ratio of 2.0.
Filtered sequence lists from MACA- and BSA-treated animals were
then compared within a time point (i.e., MACA 3 h versus BSA 3 h;
MACA 12 h versus BSA 12 h) using the list operations function of
Metacore GeneGo (GeneGo, Inc., St. Joseph, MI; www.genego.com ) to
identify sequences differentially expressed only in MACA-treated
animals (MACA-specic sequences). A second analysis on MACAspecic sequences was then run to identify those sequences
differentially expressed at both MACA post-treatment time points
(MACA 3 h/12 h-specic).
Biomarker lter analysis. Sequences identied in Metacore GeneGo

as MACA 3 h/12 h-specic were then subjected to biomarker lter
analysis using the Biomarker function of Ingenuity Pathway Analysis
Software (Ingenuity Systems, Redwood City, CA; www.ingenuity.
com). This feature facilitates prioritization of putative biomarker
genes based upon user-dened parameters such as tissue expression,
species of interest or disease association. For this analysis, potential
biomarkers were selected based upon the presence in any one or more
of the following: immune system cells (e.g., T lymphocytes,
neutrophils, dendritic cells, etc.), immune system cell lines, lung cell
lines, epidermis, lung, spleen or thymus.
Functional analysis. Putative biomarker sequences identied by
Ingenuity were then queried against the gene expression databases in
Metacore GeneGo, Ingenuity and DAVID (Database for Annotation,
Visualization and Integrated Discovery, National Institute of Allergy
and Infectious Diseases, NIH, USA; http://david.abcc.ncifcrf.gov/
home.jsp) software programs. Candidate biomarkers were prioritized
and selected based upon relevance to allergic sensitization and
likelihood of expression in available in vitro culture systems.
Pathway analysis. Sequences which met all the criteria for differential expression relative to HBSS control (i.e., ANOVA/FDR p b 0.01,
Scheffe' p b 0.05 and experiment/control ratio 2) were subjected to
pathway analysis in Ingenuity software. Pathway analysis allows
superimposition of user-specied sequence lists and expression data
on known biological pathways and functions in the Ingenuity
Pathways Knowledge Base. The signicance of the association
between the data set and canonical pathways was determined in
two ways: (1) a ratio was generated of the number of genes from the
data set that mapped to the pathway divided by the total number of
genes in the pathway, (2) Fischers exact test was used to calculate a
p-value determining the probability that the association between the
genes in the data set and the canonical pathway was explained by
chance alone.
Conrmation of selected microarray results using quantitative real-time
polymerase chain reaction (qRT-PCR). Total RNA from all samples (6
samples per treatment per time point; n = 108) was converted to
cDNA using a High-capacity RNA-to-cDNA kit (Applied Biosystems
Inc., Foster City, CA) according to manufacturers instructions. Briey,
3 g of total lung RNA was reverse transcribed using random hexamer
primers and Multiscribe reverse transcriptase in a total reaction
volume of 30 l. Cycling conditions were 25 C for 10 min, 37 C for
120 min and 85 C for 5 min. All cDNA was stored at 20 C until use.
Quantitative real-time polymerase chain reaction (qRT-PCR) was
performed on cDNA using TaqMan Universal PCR Master Mix no
AmpErase UNG (Applied Biosystems). A 30 l total reaction volume
was used per well, consisting of 100 g of cDNA (in 5 l RNAse-free
water), 15 l of 2 Master Mix, 8.5 l of RNAse free water and 1.5 l of
primers and FAM-labeled MGB probes. All primer and probe sets were
selected from available inventoried assays and purchased from
Applied Biosystems.
Amplication was performed on an ABI PRISM 7900 thermocycler at
the default cycling conditions (50 C for 2 min, 95 C for 10 min, followed
by 40 cycles of 95 C for 15 s (melt) and 60 C for 1 min (anneal)). The
endogenous control used for these assays was 18 s. Analysis was
performed using a comparative CT method adapted from (Livak and
Schmittgen, 2001). The CT value for 2 replicates per sample was
averaged for each target gene and the endogenous control. The averaged
CT value for 18 s was subtracted from the corresponding CT value for the
target gene in the same sample (CT target CT endogenous control) = CT.
Individual CT values were subtracted from the averaged CT value of the
HBSS control group for that time point= CT. The fold change was
calculated as 2 CT. Average fold changes (relative to HBSS) and
standard errors were then calculated for each treatment time point.
147
Statistical analysis of BALF and qRT-PCR data. BALF LDH activity,

total protein concentration and total and differential cell counts were
analyzed by one-way analysis of variance (ANOVA) with Bonferroni's
multiple comparison post-test using GraphPad Prism version 4.03 for
Windows (GraphPad Software, San Diego, CA, www.graphpad.com).
For these analyses, values from MACA-treated and BSA-treated
animals were compared to HBSS-treated controls at the same time
point. A p-value of b 0.05 was considered statistically signicant.
Reported values represent means standard error (SE).
Gene transcription data obtained by qRT-PCR are expressed as
mean fold changes relative to HBSS-treated controls SE. Relative
transcription levels for MACA-treated mice were compared to BSAtreated mice at each time point in Graph Pad Prism using a two-tailed
nonparametric MannWhitney test. A value of p b 0.05 was considered
to be signicant. Where necessary to facilitate comparison in Table 4,
2 CT values b 1 were converted to negative fold change values using
the formula 1/(2 CT). However, these converted values were not
otherwise used for display or analysis.
Results
BALF LDH activity and total protein concentration
LDH activity is frequently used as an indicator of non-specic
cellular damage. LDH activity in HBSS-treated control mice stayed
relatively constant through the experiment (Fig. 2A). MACA-treated
mice displayed a very small increase in LDH activity relative to HBSS
from 6 to 24 h post-treatment, but this increase was not statistically
signicant. LDH activity in BSA-treated mice was more variable and
displayed no clear temporal pattern.
BALF total protein concentration may indicate increased vascular
permeability, increased local secretion of inammatory mediators,
or cellular leakage secondary to damage or death. Administration of
MACA was associated with a signicant increase (relative to both
HBSS and BSA) in BALF total protein concentration at 12 and 18 h
post-treatment, with a more variable increase seen at 6 h (Fig. 2B).
This increase in total protein in the absence of signicantly
increased LDH activity suggests that MACA administration is
associated with the induction of pulmonary edema and inammation but very little overt cytotoxicity during the time points
evaluated in this study.
Total and differential BALF cell counts
MACA-treated mice displayed a rapid increase in total BALF cell
numbers post-treatment (Fig. 3). These increases were signicant
compared to both HBSS and BSA at 6, 12 and 18 h. HBSS- and BSAtreated mice demonstrated similar but minimal increases in BALF
cellularity after IA, with the greatest cell counts noted at the 18 h and
24 h time points.
BALF cellularity was also evaluated by individual cell type.
Macrophages demonstrated a small increase with time (up to 18 h
post-treatment) in all groups, including HBSS controls (Fig. 4A).
Neither BSA-treated nor HBSS-treated mice demonstrated appreciable changes in BALF neutrophils, lymphocytes or eosinophils. In
contrast, MACA-treated mice had signicant increases in all of these
cell types relative to both HBSS and BSA. Neutrophils were
signicantly increased by 3 h post-treatment, while lymphocytes
were signicantly increased by 6 h (Figs. 4B, C). Both cell types
remained elevated for the duration of the study. Eosinophil numbers
were signicantly increased at 12 and 24 h post-treatment (Fig. 4D).
When the results of the BALF analyses were evaluated as a whole,
it was noted that most parameters rst demonstrated signicant
perturbations at 6 h after treatment administration, with many of
these values decreasing back towards baseline by 24 h. The triggers
for these changes would be expressed before phenotypic changes
148
Fig. 4. BALF differential cell counts. (A) An increase in BALF macrophages was noted over time in all 3 treatment groups. (BD). IA MACA induces the rapid recruitment of neutrophils,
lymphocytes and eosinophils relative to BSA- or HBSS-treated mice. P b 0.05, P b 0.01, P b 0.001 relative to HBSS; P b 0.05, P b 0.01, P b 0.001 relative to BSA.
could be seen. Therefore, we selected samples from the 3 and 12 h

post-treatment times for all three treatment groups for microarray
gene expression analysis.
Microarray gene expression analysis
Our overall data analysis strategy is outlined in Fig. 1. Sequences
dened as differentially expressed had to t all of the following
conditions: (1) sequence intensity signicantly different from timematched control as determined by one-way ANOVA/BenjaminiHochberg FDR p b 0.01, (2) Scheffe's post hoc analysis p b 0.05, and
(3) upregulated or downregulated at least two-fold relative to timematched control.
A total of 451 sequences from MACA-treated mice were differentially expressed relative to control at the 3 h time point, with a fold
difference range of + 50.37 to 18.68 (Supplementary Table 1). This
number increased at the 12 h time point, with 754 sequences
differentially expressed (range +49.99 to 15.10; Supplementary
Table 2). Compared to HBSS-treated control mice, BSA-treated mice
had a total of 124 differentially expressed sequences 3 h postadministration, with a fold difference range of + 35.12 to 24.04
(Supplementary Table 3). This number had decreased to 17 by 12 h
(range +3.44 to 2.80; Supplementary Table 4).
One objective of this study was to identify potential biomarkers of
exposure to respiratory sensitizers. Therefore, we used the list
operations function of GeneGo to select genes which were differentially expressed only in MACA-treated animals at either the 3 or 12 h
time points. Genes which were also (or only) differentially expressed
after BSA administration were excluded from further consideration as
biomarkers, but were subjected to pathway analysis (see below). A
total of 299 and 620 MACA-specic genes were identied at the 3 and
12 h time points, respectively.
The ultimate goal of this work is to translate our results into a more
high-throughput format for hazard screening. Many high-throughput
assays are time-sensitive and therefore will likely require the use of
biomarkers that demonstrate persistent expression. For this reason,
we used the list operations function of GeneGo to identify MACAspecic genes that were differentially expressed at both the 3 h and
12 h time points (MACA 3 h/12 h genes; n = 153 genes) (Table 1).
Biomarker lter analysis
Our list of differentially expressed genes was generated from
whole lung homogenate RNA. It is possible that some of these genes
had been expressed by non-resident cells or by cell types which might
not be easily modeled in vitro. For this reason, we imported the list of
MACA 3 h/12 h genes into Ingenuity Pathway Analysis Software to
conduct biomarker lter analysis. This allowed us to select genes
which would be expected to be expressed in a more limited selection
of organs or cell types. We chose genes that would be expressed in
lung or immune system cells (such as dendritic cells or lymphocytes),
as well as in existing cell lines derived from these cell types. We
included genes expressed in thymus and spleen to represent primary
and secondary immune organs. Finally, since respiratory hypersensitivity may be induced by cutaneous sensitization (Vanoirbeek et al.,
2004; Akei et al., 2006), we also included genes expressed in
epidermis. Application of these limits allowed us to narrow our list
of potential biomarkers down to 130 genes, represented by 201
sequences (Table 2).
Functional analysis and candidate biomarker selection
At this point, our list of potential biomarkers was composed of
genes that were differentially expressed in MACA-treated mice, and
149
Table 1
Candidate biomarker gene sequence sets differentially expressed 3 and 12 h after MACA administration.
Sequence
code
Primary
sequence
name
Sequence description
Fold
Fold
change change
(12 h)
(3 h)
Sequence
code
Primary
sequence
name
Fold
Fold
change change
(12 h)
(3 h)
1450788_at
1418652_at
1418930_at
1419607_at
Saa1
Cxcl9
Cxcl10
Tnf
25.69
7.60
19.11
7.67
46.99
27.46
19.91
19.54
Serum amyloid A 1
Chemokine (C-X-C motif) ligand 9
Tumor necrosis factor
1420904_at
1426909_at
1416111_at
1419272_at
Il17ra
Uck2
Cd83
Myd88
2.11
2.40
6.38
2.17
2.73
2.70
2.70
2.68
1428034_a_at Tnfrsf9
5.00
15.21
1421207_at
Lif
5.16
2.66
1420591_at
Gpr84
4.04
12.09
Tumor necrosis factor receptor

superfamily, member 9
G protein-coupled receptor 84
1418936_at
Maff
2.35
2.63
1460227_at
1419209_at
1425151_a_at
1421228_at
Timp1
Cxcl1
Noxo1
Ccl7
4.38
12.42
5.83
11.88
12.05
11.98
11.43
11.32
Tissue inhibitor of metalloproteinase 1

NADPH oxidase organizer 1
Chemokine (C-C motif) ligand 7
1425374_at
1449184_at
1450698_at
1449233_at
Oas3
Pglyrp1
Dusp2
Bhlhb8
4.93
2.52
2.42
2.47
2.61
2.60
2.60
2.60
1419192_at
Il4i1
13.40
11.23
Interleukin 4 induced 1
1460220_a_at Csf1
4.24
2.57
1460282_at
Trem1
9.59
11.06
1425850_a_at Nek6
2.13
2.56
1451798_at
1419721_at
Il1rn
Gpr109a
6.88
5.78
10.86
10.58
Triggering receptor expressed

on myeloid cells 1
Interleukin 1 receptor antagonist
G protein-coupled receptor 109A
1418262_at
1422924_at
Syk
Tnfsf9
1.70
4.77
2.56
2.55
1420380_at
1419413_at
1460469_at
Ccl2
Ccl17
Tnfrsf9
17.44
3.77
4.18
10.24
9.83
9.79
1453924_a_at Ptgfr
1450377_at
Thbs1
1454005_at
Fmo2
2.13
2.70
2.36
2.53
2.51
2.51
1423017_a_at Il1rn
6.24
9.35

1438855_x_at Tnfaip2
3.43
2.50
1420330_at
Clec4e
1449227_at
Ch25h
1425829_a_at Steap4
4.48
12.50
4.61
8.93
8.11
8.06
C-type lectin domain family 4, member e

Cholesterol 25-hydroxylase
STEAP family member 4
1460231_at
Irf5
1455265_a_at Rgs16
1423596_at
Nek6
2.02
3.79
2.50
2.49
2.47
2.47
1420331_at
1417314_at
1419561_at
1449399_a_at
1426037_a_at
1448061_at
1420558_at
1449277_at
Clec4e
Cfb
Ccl3
Il1b
Rgs16
Msr1
Selp
Ccl19
6.36
2.01
13.72
8.75
10.57
2.61
7.25
5.68
7.58
7.25
6.96
6.88
6.49
6.47
6.21
6.05
C-type lectin domain family 4, member e

Complement factor B
Interleukin 1 beta
Regulator of G-protein signaling 16
Macrophage scavenger receptor 1
Selectin, platelet
1427257_at
1420723_at
1448749_at
1418547_at
1421712_at
1438183_x_at
1418261_at
1425649_at
Vcan
Vnn3
Plek
Tfpi2
Sele
Sord
Syk
Slc39a14
1.55
2.19
2.63
2.77
16.99
2.55
2.07
2.30
2.46
2.45
2.42
2.41
2.41
2.40
2.40
2.39
1417925_at
1449984_at
1421408_at
1418806_at
Ccl22
Cxcl2
Igsf6
Csf3r
4.40
11.22
11.88
3.94
6.01
5.99
5.91
5.61
1450672_a_at
1423006_at
1451452_a_at
1448604_at
Trex1
Pim1
Rgs16
Uck2
2.22
3.26
8.62
2.57
2.38
2.37
2.36
2.36
1416576_at
1421694_a_at
1419532_at
1448951_at
Socs3
Vcan
Il1r2
Tnfrsf1b
6.08
2.99
3.63
4.98
5.42
5.41
5.40
5.26

Immunoglobulin superfamily, member 6
Colony stimulating factor 3 receptor
(granulocyte)
Suppressor of cytokine signaling 3
Chondroitin sulfate proteoglycan 2
Interleukin 1 receptor, type II
superfamily, member 1b
1419714_at
1423520_at
1438097_at
1420824_at
Cd274
Lmnb1
Rab20
Sema4d
2.16
3.15
2.52
1.57
2.35
2.34
2.34
2.33
1449906_at
1453076_at
Selp
Batf3
11.66
3.52
5.26
5.17
1451314_a_at Vcam1
1422905_s_at Fmo2
2.85
1.49
2.33
2.33
CD274 antigen
Lamin B1
RAB20, member RAS oncogene family
Sema domain, immunoglobulin
domain (Ig), transmembrane domain
(TM) and short cytoplasmic domain,
(semaphorin) 4D
Vascular cell adhesion molecule 1
Flavin containing monooxygenase 2
1450318_a_at P2ry2
3.20
5.06
1415899_at
4.10
2.33
Jun-B oncogene
1422062_at
1419697_at
Msr1
Cxcl11
2.72
5.01
5.03
5.01
1435415_x_at Marcksl1
1419394_s_at S100a8
2.17
4.58
2.32
2.32
1450200_s_at
Csf2rb
2.91
4.99
Colony stimulating factor 2 receptor, beta,

low-afnity (granulocyte-macrophage)
1431843_a_at Nfkbie
4.64
2.32
1425663_at
1419609_at
Il1rn
Ccr1
4.24
4.12
4.98
4.93

Chemokine (C-C motif) receptor 1
1437226_x_at Marcksl1
1438841_s_at Arg2
3.41
1.87
2.30
2.30
1419082_at
Serpinb2
2.92
4.92
1417856_at
3.90
2.29
1416273_at
Tnfaip2
8.35
4.83
1453851_a_at Gadd45g
3.10
2.28
1419482_at
C3ar1
2.09
4.78
Serine (or cysteine) peptidase inhibitor,

clade B, member 2
Tumor necrosis factor, alpha-induced
protein 2
Complement component 3a receptor 1
1425797_a_at Syk
1.60
2.27
MARCKS-like 1
S100 calcium binding protein
A8 (calgranulin A)
Nuclear factor of kappa light
polypeptide gene enhancer in
B-cells inhibitor, epsilon
MARCKS-like 1
Vesicle transport through interaction
with t-SNAREs 1b homolog
Avian reticuloendotheliosis viral
(v-rel) oncogene related B
Growth arrest and
DNA-damage-inducible 45 gamma
Spleen tyrosine kinase
Selectin, platelet
Basic leucine zipper transcription
factor, ATF-like 3
Purinergic receptor P2Y, G-protein coupled
2
Junb
Relb
Interleukin 17 receptor A
Uridine-cytidine kinase 2
CD83 antigen
Myeloid differentiation
primary response gene 88
Leukemia inhibitory factor
v-maf musculoaponeurotic
brosarcoma oncogene family,
protein F (avian)
2-5 oligoadenylate synthetase 3
Peptidoglycan recognition protein 1
Dual specicity phosphatase 2
Basic helix-loop-helix domain
containing, class B, 8
Colony stimulating factor 1
(macrophage)
NIMA (never in mitosis gene a)
-related expressed kinase 6
Tumor necrosis factor (ligand)
Prostaglandin F receptor
Thrombospondin 1
protein 2
Interferon regulatory factor 5
NIMA (never in mitosis gene a)
-related expressed kinase 6
Versican
Vanin 3
Pleckstrin
Tissue factor pathway inhibitor 2
Selectin, endothelial cell
Sorbitol dehydrogenase 1
Solute carrier family 39
(zinc transporter), member 14
Three prime repair exonuclease 1
Proviral integration site 1
(continued
on next
next page)
page)
(continued on
150
Table 1 (continued)
Sequence
code
Primary
sequence
name
Fold
Fold
change change
(12 h)
(3 h)
Sequence
code
Primary
sequence
name
Fold
Fold
change change
(12 h)
(3 h)
1417523_at
1425380_at
1452732_at
1425434_a_at
Plek
Rasgrp4
Asprv1
Msr1
4.01
3.60
2.39
1.35
4.77
4.73
4.65
4.63
Pleckstrin
RAS guanyl releasing protein 4
Aspartic peptidase, retroviral-like 1
1426505_at
1452483_a_at
1432273_a_at
1434596_at
Evi2b
Cd44
Darc
Sema4d
2.10
2.24
2.53
1.18
2.24
2.21
2.21
2.20
1427747_a_at
1421578_at
1448914_a_at
1419149_at
Lcn2
Ccl4
Csf1
Serpine1
2.26
12.43
4.77
3.76
4.59
4.58
4.55
4.55
1421173_at
1427256_at
1422904_at
1418099_at
Irf4
Vcan
Fmo2
Tnfrsf1b
2.45
1.35
1.26
1.83
2.20
2.18
2.16
2.15
1431705_a_at Mcoln2
4.74
4.53
Lipocalin 2
Colony stimulating factor 1 (macrophage)
Serine (or cysteine) peptidase inhibitor,
clade E, member 1
Mucolipin 2
1418830_at
Cd79a
2.44
2.15
1435872_at
Pim1
3.13
4.41
1425154_a_at Csf1
2.77
2.15
1419728_at
1448291_at
1422046_at
1451416_a_at
Cxcl5
Mmp9
Itgam
Tgm1
21.26
9.99
2.51
3.53
4.34
4.33
4.28
4.15

Matrix metallopeptidase 9
Integrin alpha M
Transglutaminase 1, K polypeptide
1448748_at
1460302_at
1427683_at
1416871_at
Plek
Thbs1
Egr2
Adam8
2.30
2.74
2.88
3.20
2.15
2.14
2.13
2.13
1425902_a_at Nfkb2
4.69
4.08
Nuclear factor of kappa light polypeptide

gene enhancer in B-cells 2, p49/p100
1449360_at
Csf2rb2
1.70
2.13
1420823_at
3.95
4.08
1416304_at
Litaf
2.25
2.11
1460197_a_at Steap4
1420804_s_at Clec4d
1427429_at
Csf2
3.09
3.63
10.51
4.04
3.92
3.91
1420905_at
1424880_at
1451043_at
Il17ra
Trib1
Nek6
2.14
2.76
1.34
2.10
2.10
2.06
1420349_at
Ptgfr
1.50
3.73
Sema domain, immunoglobulin domain

(Ig), transmembrane domain (TM) and
short cytoplasmic domain, (semaphorin)
4D
STEAP family member 4
C-type lectin domain family 4, member d
(granulocyte-macrophage)
1425197_at
Ptpn2
2.33
2.05
1451054_at
Orm1
3.34
3.68
Orosomucoid 1
1449125_at
Tnfaip8l1 2.39
2.05
1417813_at
1448181_at
Ikbke
Klf15
3.71
2.79
3.42
3.37
Inhibitor of kappaB kinase epsilon

Kruppel-like factor 15
1429128_x_at Nfkb2
1429752_x_at Clip4
2.28
1.37
1.91
1.89
1418842_at
1450783_at
Hcls1
It1
2.05
2.16
3.37
3.36
3.71
2.07
1.85
1.84
1419483_at
C3ar1
1.71
3.35
Hematopoietic cell specic Lyn substrate 1 1427682_a_at Egr2

Interferon-induced protein with
1449828_at
Ptgfr
tetratricopeptide repeats 1
1417140_a_at Ptpn2
1.85
1.83
1419610_at
1420394_s_at
Ccr1
Lilrb4
3.00
2.48
3.32
3.32
1418847_at
Arg2
1450749_a_at Nr4a2
2.06
1.41
1.79
1.77
1449450_at
Ptges
3.16
3.30
Chemokine (C-C motif) receptor 1

Leukocyte immunoglobulin-like
receptor, subfamily B, member 4
Prostaglandin E synthase
1425198_at
2.37
1.74
1418465_at
1435458_at
Ncf4
Pim1
2.08
4.11
3.25
3.25
Neutrophil cytosolic factor 4

1416303_at
Litaf
1453278_a_at Clip4
2.19
1.35
1.72
1.72
1450750_a_at Nr4a2
2.20
3.22
1416298_at
Mmp9
2.23
1.63
1415922_s_at
Marcksl1
6.98
3.21
Nuclear receptor subfamily 4,

group A, member 2
MARCKS-like 1
1417483_at
Nfkbiz
2.93
1.60
1427994_at
Cd300lf
2.14
3.20
CD300 antigen like family member F
1418800_at
Bhlhb8
1.09
1.58
1425155_x_at Csf1
4.57
3.19
Colony stimulating factor 1 (macrophage)
1438562_a_at Ptpn2
1.73
1.55
1425412_at
1427313_at
Nlrp3
Ptgir
2.74
2.62
3.17
3.13
NLR family, pyrin domain containing 3

Prostaglandin I receptor (IP)
1418642_at
1435504_at
Lcp2
Clip4
1.49
1.01
1.51
1.50
1427911_at
1425435_at
1449486_at
1418641_at
1451340_at
1419135_at
1456212_x_at
Tmem173
Msr1
Ces1
Lcp2
Arid5a
Ltb
Socs3
2.83
1.14
2.79
2.86
4.92
2.72
3.11
3.11
3.08
3.03
2.99
2.99
2.90
2.90
1426584_a_at
1421811_at
1415989_at
1421206_at
1423760_at
1426506_at
1448162_at
Sord
Thbs1
Vcam1
Lif
Cd44
Evi2b
Vcam1
1.61
1.43
3.38
1.29
1.56
1.24
2.24
1.43
1.41
1.40
1.39
1.39
1.36
1.36
3.35
2.98
2.88
2.87
Transmembrane protein 173

macrophage scavenger receptor 1
Carboxylesterase 1
Lymphocyte cytosolic protein 2
AT rich interactive domain 5A (Mrf1 like)
Lymphotoxin B
Cytokine inducible SH2-containing
protein 3
Cytokine inducible SH2-containing
protein 3
Nuclear factor of kappa light

polypeptide gene enhancer in B-cells
inhibitor, zeta
Basic helixloophelix domain
containing, class B, 8
Protein tyrosine phosphatase, nonreceptor type 2
Lymphocyte cytosolic protein 2
CAP-GLY domain containing linker
protein family, member 4
Sorbitol dehydrogenase
thrombospondin 1
CD44 antigen
Ecotropic viral integration site 2b
1448294_at
1424881_at
Litaf
Trib1
1.19
1.33
1.29
1.26
LPS-induced TN factor
Tribbles homolog 1 (Drosophila)
Sema4d
1450446_a_at Socs1
1455899_x_at Socs3
Ptpn2
Ecotropic viral integration site 2b

CD44 antigen
Duffy blood group
Sema domain, immunoglobulin
domain (Ig), transmembrane domain
(TM) and short cytoplasmic domain,
(semaphorin) 4D
Versican
superfamily, member 1b
CD79A antigen (immunoglobulinassociated alpha)
(macrophage)
Pleckstrin
Thrombospondin 1
Early growth response 2
A disintegrin and metallopeptidase
domain 8
receptor, beta 2, low-afnity
(granulocyte-macrophage)
Interleukin 17 receptor A
Tribbles homolog 1 (Drosophila)
NIMA (never in mitosis gene a)related expressed kinase 6
protein 8-like 1
RIKEN cDNA 1110007H17 gene
Early growth response 2
Arginase type II
Nuclear receptor subfamily 4, group
A, member 2
Matrix metallopeptidase 9
151
Table 1 (continued)
Fold
Fold
change change
(12 h)
(3 h)
Sequence
code
Primary
sequence
name
Fold
Fold
change change
(12 h)
(3 h)
1448728_a_at Nfkbiz
7.20
2.85
1423521_at
Lmnb1
1.52
1.21
Lamin B1
1421326_at
Csf2rb
1.97
2.85
1419698_at
Cxcl11
1.23
1.19
1427035_at
Slc39a14
2.27
2.85
1421174_at
Irf4
1.75
1.12
1449449_at
1425958_at
Ptges
Il1f9
3.11
5.25
2.82
2.75
1450160_at
1421481_at
Lif
Tnfrsf9
3.33
1.53
1.12
1.12
1419132_at
1448756_at
Tlr2
S100a9
2.93
4.03
2.75
2.73
1417137_at
Uck2
2.47
1.10

Sequence
code
Primary
sequence
name
Nuclear factor of kappa light polypeptide

gene enhancer in B-cells inhibitor, zeta
Colony stimulating factor 2 receptor, beta,
low-afnity (granulocyte-macrophage)
Solute carrier family 39 (zinc transporter),
member 14
Prostaglandin E synthase
Interleukin 1 family, member 9
Toll-like receptor 2
S100 calcium binding protein A9
(calgranulin B)
For selection as candidate biomarkers, sequence sets had to t all of the following criteria: (1) signicantly different in expression relative to time-matched HBSS-treated mice (see
text), (2) not differentially expressed in BSA-treated mice, (3) differentially expressed in MACA-treated mice at both the 3 and 12 h time points, and (4) selected for potential
biomarker utility using Ingenuity Pathway Analysis Software. Values are presented as group average fold changes relative to HBSS-treated mice for each sequence set.
which had been ltered for likely expression in lung and immune
system-related tissues (or cell lines derived from those tissues).
However, these genes were not necessarily related to the development or perpetuation of allergic sensitization or asthma. To address
this possibility, we queried the list against three gene expression
databases to facilitate prioritization of biomarker candidates by
known or suspected relevance to allergic sensitization, to airway
hyperreactivity or to other pulmonary conditions associated with
asthma (such as airway remodeling).
The databases we used were the Ingenuity Knowledge database,
the GeneGo database and DAVID (Database for Annotation, Visualization and Integrated Discovery). Annotations for each of the
potential biomarker genes were manually reviewed. Genes with low
specicity (such as those encoding the multi-functional transcription
factors Junb and Nfkb2) were given low priority, as were genes known
to be associated with non-specic inammation (e.g., Tnf). Genes
with known relevance to allergic sensitization or inammation,
inammatory cell recruitment, airway hyperreactivity or remodeling
were given a high priority. This process enabled the selection of a nal
list of 11 high priority candidate biomarker genes (Table 2).
Pathway analysis
To gain additional insight into the nature of the immune responses
following exposure to agents of high versus low allergenicity,
differentially expressed sequences were subjected to pathway
analysis in Ingenuity. Genes differentially expressed after MACA
exposure (either 3 h or 12 h post-administration) demonstrated a
highly signicant association with a number of canonical pathways

characteristic of an active and developing immune response (Table 3).
In many of these pathways, increased expression of most members of
the NFB signaling family was seen. However, there was little to no
change in expression of the JAK2/STAT3 or ERK, JNK or p38 map
kinase signaling families. A similar pattern of pathway association was
observed even when only MACA-specic genes were evaluated
(Table 3).
Differentially expressed genes in 3 h BSA-treated mice were
associated with pathways characteristic of a more general inammatory response and oxidative stress. No clear pathway association was
noted in 12 h BSA samples, probably due to the low number of
differentially expressed genes at this time point (Table 3). Neither
time point appeared to be associated with consistent alterations in
any of the signaling families.
qRT-PCR analysis of candidate biomarker expression
Expression of the 11 high-priority candidate biomarker genes was
evaluated by qRT-PCR, both as a validation of the microarray results
and to evaluate expression levels over the entire time course of the
study. In general, fold difference values (relative to HBSS-treated
mice) calculated from qRT-PCR data approximated those calculated
from the microarray intensity data at both the 3 and 12 h time points
in MACA-treated mice (Table 4). The largest discrepancies were seen
for Saa1, in which microarray and PCR fold differences were 25.69 and
12.59, and 46.99 and 135.74 at 3 and 12 h, respectively. Other notable
differences between microarray and qRT-PCR values were for Cxcl2
Table 2
Overview of high-priority candidate biomarkers of acute respiratory exposure to sensitizing agents.
Primary
sequence name
Chemotactic Associated with Associated with

Persistently expressed
allergic disease pulmonary disease (qRT-PCR)
CCL17
X
X
X
X
CCL22
CCL19
CCL7
CXCL2
CXCL10
SAA1
C3aR1
VCAM1
SOCS3
Arg2
Chemokine (C-C motif) ligand 17; thymus and activation regulated chemokine
(TARC)
Chemokine (C-C motif) ligand 22; macrophage derived chemokine (MDC)
Chemokine (C-C motif) ligand 19; macrophage inhibitory protein 3- (MIP-3)
Chemokine (C-C motif) ligand 7; monocyte chemoattractant protein 3 (MCP-3)
Chemokine (C-X-C motif) ligand 2; macrophage inammatory protein 2
(MIP-2)
Chemokine (C-X-C motif) ligand 10; interferon inducible protein 10 (IP-10)
Serum amyloid A1
Arginase 2
X
X
X
X
X
X
X
X
X
X
X
Biomarker and functional analysis of genes differentially expressed at both 3 and 12 h time points allowed identication of 11 candidate biomarker genes. Selected genes had to meet
at least one of the following criteria: (1) involved in chemotaxis, (2) associated with allergic disease, and (3) known or strongly suspected associated with allergic or non-allergic
pulmonary disease. Candidate genes were also evaluated for persistence of gene expression (as determined by qRT-PCR) over the course of the study.
152
Table 3
Summary of pathway analyses.
Name
p-value
No. of genes in data set

mapping to pathway/Total
no. of genes in pathway
Ratio
Net effect
MACA 3 h
IL-10 signaling
IL-6 signaling
9.43E-14
5.10E-12
16/71
17/96
0.225
0.177
LXR/RXR activation
8.84E-12
15/85
0.176
Dendritic cell maturation
3.61E-11
20/165
0.121
Acute phase response signaling
5.27E-11
21/178
0.118
Increased IL-1 signaling via NFB; decreased IL-10 signaling

Increased IL-1 and TNF signaling via NFB; increased
production of inammatory mediators
Increased IL-1 and TNF signaling via NFB; increased production
of inammatory mediators
Increased production of inammatory mediators, adhesion and
costimulatory molecules
MACA 12 h
6.30E-14
30/178
0.169
TREM1 signaling
3.65E-12
17/69
0.246
2.87E-11
25/165
0.152
Role of pattern recognition receptors

in recognition of bacteria and viruses
Hepatic brosis/hepatic stellate cell activation
MACA-specic 3 h/12 h
1.04E-09
17/88
0.193
1.67E-09
21/135
0.156
6.48E-13
16/178
0.09
1.82E-11
14/165
0.085
TREM1 signaling
4.05E-11
10/69
0.145
LXR/RXR activation
4.34E-10
10/85
0.118
Role of pattern recognition receptors in

recognition of bacteria and viruses
BSA 3 h
Xenobiotic metabolism signaling
Aryl hydrocarbon receptor signaling
NRF-2 mediated oxidative stress response
LPS/IL-1 mediated inhibition of RXR function
IL-10 signaling
BSA 12 h
Coagulation system
CD40 signaling
Macropinocytosis
CCR5 signaling in macrophages
IL-4 signaling
9.66E-10
10/88
0.114
3.22E-04
8.60E-04
2.34E-03
3.43E-03
4.28E-03
7/254
5/155
5/185
5/198
3/71
0.028
0.032
0.027
0.025
0.042
Enhanced redox regulation; decreased oxidation

Mixed effects on redox status
Increased antioxidant response
Increased lipid metabolism
Small increase in IL-6 production
3.14E-02
5.59E-02
5.67E-02
5.76E-02
5.76E-02
1/37
1/70
1/72
1/87
1/72
0.027
0.014
0.014
0.011
0.014
Increased production of coagulation factor X

Not signicant
Not signicant
Not signicant
Not signicant

Increased IL-1 and TNF signaling via AKT and NFB; increased
Increased signaling by TLR-2 and complement receptors; increased
Increased production of inammatory mediators
Increased IL-1 and TNF signaling via NFB; increased production of
inammatory mediators
Increased IL-1 and TNF signaling via AKT and NFB; increased
Increased signaling by TLR-2 and complement receptors; increased
Genes that were signicantly expressed in MACA- or BSA-treated mice were evaluated to determine canonical pathway association using Ingenuity Pathway Analysis software.
(11.22 and 24.01, and 5.99 and 17.44 at 3 and 12 h), Ccl7 (11.32 and
23.10 at 12 h) and Ccl19 (6.05 and 12.06 at 12 h). In contrast, direct
comparison between microarray and qRT-PCR values was usually not
possible for BSA-treated mice, as almost none of these sequence sets

met the statistical cutoff parameters during initial microarray analysis
(data not shown).
Table 4
Microarray and PCR comparison.
Primary
sequence name
MACA fold
change 3 h
microarray
MACA fold
change 3 h
qRT-PCR
MACA fold
change 12 h
microarray
MACA fold
change 12 h
qRT-PCR
BSA fold
change 3 h
microarray
BSA fold
change 3 h
qRT-PCR
BSA fold
change 12 h
microarray
BSA fold
change 12 h
qRT-PCR
Serum amyloid A 1
Arginase type II
Saa1
Cxcl10
Ccl7
Cxcl2
Ccl19
Ccl22
Socs3
Ccl17
Vcam1
Arg2
C3ar1
25.69
19.11
11.88
11.22
5.68
4.40
4.06b
3.77
2.82
1.97
1.90
12.59
19.23
11.44
24.01
5.20
3.08
3.62
5.46
4.55
2.37
2.10
46.99
19.91
11.32
5.99
6.05
6.01
3.73
9.83
1.70
2.04
4.07
135.74
20.76
23.10
17.44
12.06
7.75
2.89
10.13
1.03
2.02
2.36
N/aa
N/a
N/a
N/a
N/a
N/a
1.49
N/a
N/a
1.47
N/a
1.50
1.51
2.04
1.05
1.07
1.67
1.09
2.52
1.01
1.01
1.21
N/a
N/a
N/a
N/a
N/a
N/a
N/a
N/a
N/a
N/a
N/a
6.55
1.94
1.83
7.19
2.09
2.20
1.92
2.41
1.09
1.21
1.03
For validation purposes, average microarray gene fold-expression values for 11 candidate biomarker genes were compared to the corresponding values obtained by qRT-PCR
analysis. qRT-PCR values were calculated relative to HBSS-treated control samples for each time point using the 2 Ct method. For comparison purposes, 2 Ct values b 1 were
converted to negative fold change values using the formula: 1/2 Ct.
a
N/a indicates that the expression intensity for this parameter failed to meet the initial statistical cutoff.
b
Where necessary, the expression values of genes represented by multiple sequence sets have been averaged.
Fig. 5. Transcription of candidate biomarker genes as determined by qRT-PCR. Transcript levels were calculated using the CT method. Values are presented as average fold changes SE relative to HBSS-treated mice for each time point
(dened as fold change of 1 and indicated by a dotted line).
153
154
As previously mentioned, candidate genes being screened for

biomarker utility should ideally demonstrate persistent differential
expression to facilitate their reliable detection in high-throughput
screening systems. For this reason, the relative expression values
for both MACA- and BSA-treated mice were then plotted as a
function of time and then compared to each other at each time
point (Fig. 5).
None of the genes differed signicantly in expression between
MACA- and BSA-treated mice at all 6 time points evaluated. However,
3 genes were signicantly increased at 5 time points (Saa1, Cxcl10 and
Ccl7), while another 3 (Ccl22, Ccl17 and Cxcl2) were increased at 4
time points. In contrast, Socs3, Ccl19 and Vcam1 demonstrated only
transient increases in expression, while C3aR1 did not increase
signicantly in expression relative to BSA-treated animals until 24 h.
Expression of the last gene (Arg2) was more variable, with signicant
increases noted at 1 h and 18 h, but lesser expression at other time
points.
Discussion
In this study, we have performed microarray gene expression
analysis on lung samples following a single exposure to a known
respiratory sensitizer (MACA) or to a low-allergenicity control (BSA).
Gene expression results from both treatment groups were subjected
to a series of statistical and functional analyses to identify changes in
gene expression characteristic of sensitizer exposure and to evaluate
these genes for potential biomarker utility.
Microarray analysis allowed us to identify a panel of 11
biomarker candidates (Table 2). Of these 11 candidates, 6 were
demonstrated to undergo persistently increased expression following administration of MACA as compared to BSA. These genes
represent particularly promising candidates for translation into
high-throughput screening assays, which frequently utilize timesensitive protocols. This list includes 5 chemokines, all of which can
mediate inammatory cell recruitment in allergic disease. CCL17
(thymus and activation regulated chemokine; TARC) is known to be
upregulated after allergen or anti-IgE challenge (Fulkerson et al.,
2004; Pucheu-Haston et al., 2006) and in atopic dermatitis (Hijnen
et al., 2004). Both CCL17 and CCL22 (macrophage derived
chemokine; MDC) share a common receptor (CCR4) and are
released during late-phase allergic airway responses in humans
(Bochner et al., 2003). CCL7 (monocyte chemotactic protein-3;
MCP-3) is a powerful chemotactic factor for eosinophils (TillieLeblond et al., 2000; Shang et al., 2002), and (together with CXCL-2
and CXCL-10) can contribute to neutrophilic airway inammation
(Michalec et al., 2002; McKinley et al., 2005).
The sixth candidate biomarker demonstrating persistently increased mRNA transcription was the acute phase protein serum
amyloid A-1 (SAA-1). This protein is frequently used as a marker of
non-specic systemic inammation. However, recent studies have
demonstrated upregulation of SAA-1 in the sputum (Ozseker et al.,
2006) and blood (Buyukozturk et al., 2004) of allergic asthmatics, as
well as in the skin of atopic dermatitis patients (Wood et al., 2009).
In contrast, 3 of our candidate genes were only transiently
increased in expression (Socs3, Ccl19 and Vcam1). This transient
expression may limit the potential biomarker utility of these genes
independent of their possible roles in the development and
perpetuation of allergic disease. Another candidate gene (Arg2)
demonstrated more variable expression, with signicant increases
only at the 1 and 18 h time points.
The nal candidate biomarker gene is C3aR1, which codes for the
receptor for the C3a fragment of complement. This receptor is
expressed on immune cells such as mast cells, eosinophils and
basophils (Zwirner et al., 1999; Ali and Panettieri, 2005; Prussin and
Metcalfe, 2006). In one study, C3aR-knockout mice sensitized to a
mixture of Aspergillus fumigatus extract and ovalbumin (OVA)
demonstrated impaired methacholine-induced bronchoconstriction

as well as decreased BALF eosinophils, interleukin (IL)-5 and IL-13
following intranasal challenge as compared to wild-type mice
(Drouin et al., 2002). In a different study, C3aR-knockout mice
sensitized and challenged with OVA alone also displayed an
attenuated airway response to methacholine, but did not demonstrate changes in BALF eosinophils, IL-5 or IL-13 compared to wildtype mice (Humbles et al., 2000). In the current study, qRT-PCR
analysis of C3aR1 did not demonstrate persistent elevation over time.
However, a consistent increase in mRNA transcription was noted
beginning at the 6 h time point, with peak levels at the 24 h time
point. These results suggest that the potential biomarker utility of
C3aR1 may be greatest in studies evaluating later post-administration
time points.
Although these results appear promising, it must be stressed that
this work was a preliminary study of restricted scope, with denite
limitations. First, our list of potential biomarkers was generated
following exposure to only a single allergenic protein extract, and a
single protein control. It is possible that our candidate biomarkers
may not be induced following exposure to other protein allergens, or
to non-protein allergens (such as low-molecular weight chemicals).
Conversely, one or more of these genes may be induced following
exposure to non-allergenic agents. In addition, different responses
may be seen in mixed exposure scenarios, in which multiple allergens,
irritants and/or non-allergens may all be present at once (as would be
the case for most human exposures).
A more comprehensive follow-up study is in progress to address
these concerns. Specically, this study will re-evaluate the transcriptional responses to MACA, BSA and HBSS, and compare these
responses to those induced by exposure to equivalent amounts of
house dust mite extract, Escherichia coli lipopolysaccharide and the
fungal extract Wallemia sebi (which has not been demonstrated to
induce allergic sensitization in our laboratory).
A second limitation of this study was that it evaluated the response
of a restricted population of mice. BALB/c mice are frequently used in
hypersensitivity studies due to their presumed Th-2 bias and the
ease with which they can be sensitized (Whitehead et al., 2003).
While the responses of these mice may adequately model those
individuals that are genetically predisposed to allergy, they may not
be reective of the real world scenario, in which a continuum of
individual susceptibility is likely present. Ideally, similar studies
should be conducted in other strains of inbred mice (e.g., Th-1
biased C57BL/6) as well as in outbred mice.
In addition, only female mice were evaluated in the current study.
Male mice were excluded to avoid the husbandry challenges
associated with the pronounced inter-male aggression seen in this
strain. This is an acceptable limitation for a preliminary study.
However, gender does have a considerable impact on lung function
and allergy susceptibility (Seymour et al., 2002; Carey et al., 2007;
Postma, 2007). As previously mentioned for strain-specic characteristics, gender effects must be taken into consideration during the
design of future studies.
Finally, we limited our biomarker evaluation to the gene
transcriptional level for this preliminary study. While the observed
changes in gene transcription may correspond to subsequent
changes in protein expression, this is not necessarily the case.
Conrmation of protein expression will be necessary for full
validation of this model.
In summary, we have performed microarray gene expression
analysis on lung samples harvested after a single exposure to a
known respiratory sensitizer. This analysis has allowed us to
identify 11 genes as potential biomarkers for sensitizer exposure.
Further studies are needed to determine the repeatability of
the current results, and to evaluate the validity of the panel
against a broader array of irritants and compounds of low or high
allergenicity.
Disclaimer
This research paper has been reviewed by the National Health and
Environmental Effects Research Laboratory, US Environmental Protection Agency and approved for publication. Approval does not
signify that the contents necessarily reect the views and policies of
the agency, nor does mention of trade names or commercial products
constitute endorsement or recommendation for use.
This work was supported by UNC/EPA training agreement
CR83323701.
Declaration of interest
The Authors report no conicts of interest. The Authors alone are
responsible for the content and writing of the paper.
Acknowledgments
The Authors would like to thank Debora Andrews, Judy Richards,
Richard Jaskot and Dr. Yong Joo Chung of US EPA for technical and
intellectual assistance. Additionally, we would like to thank Drs.
Robert Luebke, MaryJane Selgrade, and Susan Hester for their critical
review of the manuscript.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.taap.2009.12.027.
References
Akei, H.S., Brandt, E.B., Mishra, A., Strait, R.T., Finkelman, F.D., Warrier, M.R., Hershey,
G.K., Blanchard, C., Rothenberg, M.E., 2006. Epicutaneous aeroallergen exposure
induces systemic TH2 immunity that predisposes to allergic nasal responses.
J. Allergy Clin. Immunol. 118, 6269.
Ali, H., Panettieri Jr., R.A., 2005. Anaphylatoxin C3a receptors in asthma. Respir. Res. 6, 19.
Bochner, B.S., Hudson, S.A., Xiao, H.Q., Liu, M.C., 2003. Release of both CCR4-active and
CXCR3-active chemokines during human allergic pulmonary late-phase reactions.
J. Allergy Clin. Immunol. 112, 930934.
Buyukozturk, S., Gelinic, A.A., Genc, S., Kocak, H., Oneriyidogan, Y., Erdan, S., Dal, M.,
Colakoglu, B., 2004. Acute phase reactants in allergic airway disease. Tohoku. J. Exp.
Med. 204 (3), 209213.
Carey, M.A., Card, J.W., Voltz, J.W., Germolec, D.R., Korach, K.S., Zeldin, D.C., 2007. The
impact of sex and sex hormones on lung physiology and disease: lessons from
animal studies. Am. J. Physiol. Lung Cell Mol. Physiol. 293, L272L278.
Drouin, S.M., Corry, D.B., Hollman, T.J., Kildsgaard, J., Wetsel, R.A., 2002. Absence of the
complement anaphylatoxin C3a receptor suppresses Th2 effector functions in a
murine model of pulmonary allergy. J. Immunol. 169, 59265933.
Fulkerson, P.C., Zimmermann, N., Hassman, L.M., Finkelman, F.D., Rothenberg, M.E.,
2004. Pulmonary chemokine expression is coordinately regulated by STAT1, STAT6,
and IFN-gamma. J. Immunol. 173, 75657574.
Hijnen, D., de Bruin-Weller, M., Oosting, B., Lebre, C., de Jong, E., Bruijnzeel-Koomen, C.,
Knol, E., 2004. Serum thymus and activation-regulated chemokine (TARC) and
155
cutaneous T cell-attracting chemokine (CTACK) levels in allergic diseases: TARC

and CTACK are disease-specic markers for atopic dermatitis. J. Allergy Clin.
Immunol. 113, 334340.
Humbles, A.A., Lu, B., Nilsson, C.A., Lilly, C., Israel, E., Fujiwara, Y., Gerard, N.P., Gerard, C.,
2000. A role for the C3a anaphylatoxin receptor in the effector phase of asthma.
Nature 406, 9981001.
Livak, K.J., Schmittgen, T.D., 2001. Analysis of relative gene expression data using realtime quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25, 402408.
McKinley, L., Kim, J., Bolgos, G.L., Siddiqui, J., Remick, D.G., 2005. CXC chemokines
modulate IgE secretion and pulmonary inammation in a model of allergic asthma.
Cytokine 32, 178185.
Michalec, L., Choudhury, B.K., Postlethwait, E., Wild, J.S., Alam, R., Lett-Brown, M., Sur, S.,
2002. CCL7 and CXCL10 orchestrate oxidative stress-induced neutrophilic lung
inammation. J. Immunol. 168, 846852.
Ozseker, F., Buyukozturk, S., Depboylu, B., Yilmazbayhan, D., Karayigit, E., Gelincik, A.,
Genc, S., Colakoglu, B., Dal, M., Issever, H., 2006. Serum amyloid A (SAA) in induced
sputum of asthmatics: A new look to an old marker. Int. Immunopharmaco. 6,
15691576.
Postma, D.S., 2007. Gender differences in asthma development and progression. Gen.
Med. 4, S133S146.
Prussin, C., Metcalfe, D.D., 2006. 5. IgE, mast cells, basophils, and eosinophils. J. Allergy
Clin. Immunol. 117, S450S456.
Pucheu-Haston, C.M., Shuster, D., Olivry, T., Brianceau, P., Lockwood, P., McClanahan, T.,
de Waal Malefyt, R., Mattson, J.D., Hammerberg, B., 2006. A canine model of
cutaneous late-phase reactions: prednisolone inhibition of cellular and cytokine
responses. Immunology 117, 177187.
Seymour, B.W., Friebertshauser, K.E., Peake, J.L., Pinkerton, K.E., Coffman, R.L., Gershwin,
L.J., 2002. Gender differences in the allergic response of mice neonatally exposed to
environmental tobacco smoke. Dev. Immunol. 9, 4754.
Shang, X.-Z., Chiu, B.-C., Stolberg, V., Lukacs, N.W., Kunkel, S.L., Murphy, H.S., Chensue,
S.W., 2002. Eosinophil recruitment in type-2 hypersensitivity pulmonary granulomas: source and contribution of monocyte chemotactic protein-3 (CCL7). Am. J.
Pathol. 161, 257266.
Stankus, R.P., O'Neil, C.E., 1988. Antigenic/allergenic characterization of American and
German cockroach extracts. J. Allergy Clin. Immunol. 81, 563570.
Tillie-Leblond, I., Hammad, H., Desurmont, S., Wallaert, B., Tonnel, A.-B., Gosset, P., 2000.
CC chemokines and interleukin-5 in bronchial lavage uid from patients with
status asthmaticus. Potential implication in eosinophil recruitment. Am. J. Respir.
Crit. Care Med. 162, 586592.
Vanoirbeek, J.A.J., Tarkowski, M., Ceuppens, J.L., Verbeken, E.K., Nemery, B., Hoet, P.H.M.,
2004. Respiratory response to toluene diisocyanate depends on prior frequency and
concentration of dermal sensitization in mice. Toxicol. Sci. 80, 310321.
Viana, M.E., Coates, N.H., Gavett, S.H., Selgrade, M.K., Vesper, S.J., Ward, M.D., 2002. An
extract of Stachybotrys chartarum causes allergic asthma-like responses in a BALB/
c mouse model. Toxicol. Sci. 70, 98109.
Ward, M.D.W., Chung, Y.J., Haykal-Coates, N., Copeland, L., 2009. Differential allergy
responses to Metarhizium anisopliae fungal component extracts in BALB/c mice.
J. Immunotoxicol. 6, 6273.
Ward, M.D.W., Madison, S.L., Andrews, D.L., Sailstad, D.M., Gavett, S.H., Selgrade, M.J.,
2000. Comparison of respiratory responses to Metarhizium anisopliae extract using
two different sensitization protocols. Toxicology 147, 133145.
Ward, M.D.W., Sailstad, D.M., Selgrade, M.K., 1998. Allergic responses to the
biopesticide metarhizium anisopliae in Balb/c mice. Toxicol. Sci. 45, 195203.
Whitehead, G.S., Walker, J.K.L., Berman, K.G., Foster, W.M., Schwartz, D.A., 2003.
Allergen-induced airway disease is mouse strain dependent. Am. J. Physiol. Lung.
Cell. Mol. Physiol. 285, L32L42.
Wood, S.H., Clements, D.N., Ollier, W.E., Nuttall, T., McEwan, N.A., Carter, S.D., 2009.
Gene expression in canine atopic dermatitis and correlation with clinical severity
scores. J. Dermatol. Sci. 55, 2733.
Zwirner, J., Gtze, O., Begemann, G., Kapp, A., Kirchhoff, K., Werfel, T., 1999. Evaluation
of C3a receptor expression on human leucocytes by the use of novel monoclonal
antibodies. Immunology 97, 166172.

Microarray CK+inflam

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Microarray CK+inflam

Uploaded by

Copyright:

Available Formats

Toxicology and Applied Pharmacology 244 (2010) 144155

Contents lists available at ScienceDirect

Toxicology and Applied Pharmacology

Biomarkers of acute respiratory allergen exposure: Screening for

Corresponding author. c/o U.S. Environmental Protection Agency 109 T. W.

We hypothesize that there is a set of differentially expressed

were terminated by sodium pentobarbital overdose (Euthasol, Virbac

22,500 probe sets representing 14,500 well-characterized mouse

Biomarker lter analysis. Sequences identied in Metacore GeneGo

Statistical analysis of BALF and qRT-PCR data. BALF LDH activity,

could be seen. Therefore, we selected samples from the 3 and 12 h

Tumor necrosis factor receptor

Tissue inhibitor of metalloproteinase 1

Triggering receptor expressed

Chemokine (C-C motif) ligand 2

C-type lectin domain family 4, member e

C-type lectin domain family 4, member e

Chemokine (C-C motif) ligand 22

Colony stimulating factor 2 receptor, beta,

Interleukin 1 receptor antagonist

Serine (or cysteine) peptidase inhibitor,

Proviral integration site 1

Chemokine (C-X-C motif) ligand 5

Nuclear factor of kappa light polypeptide

Sema domain, immunoglobulin domain

Inhibitor of kappaB kinase epsilon

Hematopoietic cell specic Lyn substrate 1 1427682_a_at Egr2

Chemokine (C-C motif) receptor 1

Neutrophil cytosolic factor 4

Nuclear receptor subfamily 4,

CD300 antigen like family member F

Colony stimulating factor 1 (macrophage)

NLR family, pyrin domain containing 3

Transmembrane protein 173

Nuclear factor of kappa light

Ecotropic viral integration site 2b

Chemokine (C-X-C motif) ligand 11

Interferon regulatory factor 4

Leukemia inhibitory factor

Nuclear factor of kappa light polypeptide

highly signicant association with a number of canonical pathways

Chemotactic Associated with Associated with

No. of genes in data set

Dendritic cell maturation

Acute phase response signaling

Increased IL-1 signaling via NFB; decreased IL-10 signaling

Dendritic cell maturation

Role of pattern recognition receptors

Dendritic cell maturation

Role of pattern recognition receptors in

Enhanced redox regulation; decreased oxidation

Increased production of coagulation factor X

Increased IL-1 and TNF signaling via NFB; increased production

possible for BSA-treated mice, as almost none of these sequence sets

As previously mentioned, candidate genes being screened for

demonstrated impaired methacholine-induced bronchoconstriction

cutaneous T cell-attracting chemokine (CTACK) levels in allergic diseases: TARC

You might also like