Noninflammatory Gluten Peptide Analogs as Biomarkers for Celiac Sprue

Article abstract of DOI:10.1016/j.chembiol.2009.07.009

New tools are needed for managing celiac sprue, a lifelong immune disease of the small intestine. Ongoing drug trials are also prompting a search for noninvasive biomarkers of gluten-induced intestinal change. We have synthesized and characterized noninfl

Full text of DOI:10.1016/j.chembiol.2009.07.009

Chemistry & Biology
Article
Noninflammatory Gluten Peptide Analogs
as Biomarkers for Celiac Sprue
1
2
3
4
4
4,5
´
Michael T. Bethune, Monica Crespo-Bosque, Elin Bergseng, Kaushiki Mazumdar, Lara Doyle, Karol Sestak,
Ludvig M. Sollid,5 and Chaitan Khosla1,2,6,
1Department of Biochemistry
*
2Department of Chemistry
Stanford University, Stanford, CA 94305, USA
3Centre for Immune Regulation, Institute of Immunology, University of Oslo, Rikshospitalet University Hospital, N-0027 Oslo, Norway
4Tulane National Primate Research Center, Covington, LA 70433, USA
5Department of Microbiology and Immunology, Tulane University School of Medicine, New Orleans, LA 70112, USA
6Department of Chemical Engineering, Stanford University, Stanford, CA 94305, USA
*Correspondence: khosla@stanford.edu
DOI 10.1016/j.chembiol.2009.07.009
SUMMARY
Upon encountering DQ2-gluten complexes on the surface of
antigen-presenting cells, gluten-specific, DQ2-restricted CD4+
New tools are needed for managing celiac sprue, T cells mediate a Th1 response comprising the secretion of
proinflammatory cytokines such as interferon-g (IFN-g) (Nilsen
et al., 1995, 1998; Troncone et al., 1998) and the recruitment of
a lifelong immune disease of the small intestine.
Ongoing drug trials are also prompting a search for
noninvasive biomarkers of gluten-induced intestinal
change. We have synthesized and characterized
noninflammatory gluten peptide analogs in which
key Gln residues are replaced by Asn or His. Like their
proinflammatory counterparts, these biomarkers are
resistant to gastrointestinal proteases, susceptible to
CD8+ intraepithelial lymphocytes, ultimately causing villous
atrophy and crypt hyperplasia (Jabri et al., 2005). Additionally,
CD4+ T cells amplify a humoral immune response to dietary
gluten comprising production of both gluten-specific antibodies
and TG2-specific autoantibodies (Sollid, 2002). In many affected
individuals, this molecular pathogenesis is manifested clinically
as nutrient malabsorption, wasting, and/or chronic diarrhea
glutenases, and permeable across enterocyte bar-
riers. Unlike gluten peptides, however, they are not
(Alaedini and Green, 2005; Green and Jabri, 2006). Chronic
inflammation caused by recurrent exposure to gluten is associ-
appreciably recognized by transglutaminase, HLA- ated with the increased incidence of bone disorders, neurolog-
ical complications, and cancer (Catassi et al., 2002; Corrao
DQ2, or disease-specific T cells. In vitro and animal
et al., 2001). Inflammation, antibody production, and clinical
symptoms are gluten dependent, such that strict adherence to
studies show that the biomarkers can detect intes-
tinal permeability changes as well as glutenase-cata-
lyzed gastric detoxification of gluten. Accordingly,
controlled clinical studies are warranted to evaluate
the use of these peptides as probes for abnormal
intestinal permeability in celiac patients and for glute-
a gluten-free diet causes remission, while reintroduction of die-
tary gluten induces relapse (See and Murray, 2006).
Whereas researchers are translating many of the above funda-
mental insights into clinical practices that facilitate initial diag-
nosis (Anderson et al., 2005; Dieterich et al., 1997; Koskinen
nase efficacy in clinical trials and practice.
et al., 2008; Raki et al., 2006, 2007), there are serious defi-
ciencies in the available spectrum of tools for the long term
INTRODUCTION
care of celiac patients (Silvester and Rashid, 2007). The corner-
stone of disease management is exclusion of gluten from the
Celiac sprue is a common disease of the small intestine, the diet. However, a gluten-free diet is extremely difficult to maintain
primary environmental and genetic causes of which are well due to the ubiquity of gluten in human foods and a majority of
understood. The environmental cause of this immune disease patients exhibit incomplete recovery on a gluten-free diet (Ciacci
is dietary gluten from wheat and similar proteins from rye and et al., 2002; Cornell et al., 2005; Pietzak, 2005). As a result, there
barley (Dicke et al., 1953). Gastrointestinal digestion of gluten is considerable interest in the development of non-dietary thera-
releases proteolytically resistant, immunotoxic peptide frag- pies that improve the health and quality of life of celiac sprue
ments, such as the 33-mer from a-gliadin (Shan et al., 2002). patients. The current study was motivated by the need for accu-
These peptides traverse the mucosal epithelium by unknown rate, noninvasive tools for managing patients who practice
mechanisms and are deamidated at specific glutamine residues dietary gluten control and for assessing the efficacy of new drugs
by an endogenous enzyme, transglutaminase 2 (TG2) (Arentz- in such patients on an ongoing basis.
Hansen et al., 2000; Molberg et al., 1998). Deamidated peptides
Several recent studies have highlighted the potential of orally
bind with high affinity to the primary genetic determinant of celiac administered gluten-specific proteases (i.e., glutenases) for
sprue, human leukocyte antigen (HLA)-DQ2 (Kim et al., 2004; treating celiac sprue. Since the proteolytic resistance of gluten
Quarsten et al., 1999), a major histocompatibility complex class peptides stems from their high proline (ꢀ15%) and glutamine
II molecule possessed by >90% of patients (Sollid et al., 1989). (ꢀ35%) content, efforts are primarily focused on enzymes
868 Chemistry & Biology 16, 868–881, August 28, 2009 ª2009 Elsevier Ltd All rights reserved

Chemistry & Biology
Gluten Peptide Analogs as Biomarkers
Figure 1. Synthetic Gluten and Biomarker
Peptide Sequences
Sequences are shown for the native 33-mer (desig-
nated QQQ-33-mer) derived from a2-gliadin, the
synthetically deamidated 33-mer (EEE-33-mer),
the biomarkers NNN-33-mer and HHH-33-mer,
and a noninflammatory control peptide of unre-
lated sequence derived from myoglobin. Bonds
that are scissile to EP-B2-mediated cleavage are
designated by arrowheads. Glutamine residues
that are selectively deamidated by TG2 or synthet-
ically replaced in the biomarker peptides are in
bold. Leucine residues that are isotope labeled
for in vivo experiments are underlined.
capable of cleaving proximal to these residues. Microbial prolyl conservative change would minimally perturb recognition by the
endopeptidases (PEP) detoxify immunodominant gluten in vitro EP-B2 glutenase (Bethune et al., 2006) (Figure 1, arrowheads).
(Shan et al., 2002; Stepniak et al., 2006) and are especially effec- The rationale for the Q/H analog was that, in the acidic environ-
tive when complemented with a barley endoprotease, EP-B2, ment of the post-prandial stomach and the upper small intestine,
that preferentially cleaves gluten C-terminal to glutamine resi- these substitutions were expected to position positive charges at
dues (Gass et al., 2007; Siegel et al., 2006). The therapeutic sites where TG2-mediated introduction of negative charge
promise of these oral proteases is further underscored by their increases peptide affinity for HLA-DQ2 (Kim et al., 2004;Quarsten
ability to digest gluten in the rat stomach (Gass et al., 2006, et al., 1999). Importantly, neither Asn nor His is a preferred residue
2007) and by the ability of EP-B2 to protect a gluten-sensitive for cleavage by gastrointestinal proteases, so these biomarkers
rhesus macaque against gluten-induced clinical relapse (Be- wereexpectedto be as proteolytically resistant in thegastrointes-
thune et al., 2008b). To bring therapeutic glutenases to bear on tinal lumen as the natural 33-mer peptide. As controls for our
the human condition of celiac sprue, however, safe and effective study, we synthesized a peptide derived from myoglobin with
tools for assessing enzyme efficacy in human celiac patients are moderate proteolytic resistance but unrelated sequence (Piper
needed. A similar need exists for therapies that have alternative et al., 2004), as well as EEE-33-mer, the synthetically deamidated
modes of action such as larazotide acetate, which is anticipated analog of the 33-mer (Figure 1). The structures and purity of all
to reverse tight junction dysfunction in celiac patients (Paterson peptides were confirmed by liquid chromatography-assisted
et al., 2007), or CCX282, an antagonist of the CCR9 chemokine mass spectrometry (LC-MS) (see Figure S1 available online).
receptor (Peyrin-Biroulet et al., 2008).
Our strategy for designing a biomarker for diseasemanagement
Biomarkers Are Resistant to Gastrointestinal Digestion
andclinicaldrugdevelopment wasinspiredbythephysical, chem-
but Susceptible to Therapeutic Glutenases
ical, and biological properties of immunotoxic gluten peptides
To evaluate the resistance of the 33-mer and the derivative
such as the 33-mer from a-gliadin. Specifically, we sought to engi-
biomarkers to gastrointestinal proteolysis, we first performed
neer a gluten peptide analog that mimics the 33-mer with respect
simulated gastric digests at 37ꢁC (pH 4.5) using commercially
to some criteria but can be differentiated from the natural product
available pepsin, the major protease present in the stomach.
with respect to others. Like the 33-mer, the biomarker must be
The extent of digestion at multiple time points was determined
resistant to gastrointestinal proteases, so that it is not rapidly
by analytical high performance liquid chromatography (HPLC)
cleared from the stomach or intestinal lumen. Also like the 33-
and LC-MS. As previously observed (Shan et al., 2002), the 33-
mer, it must be efficiently proteolyzed by therapeutic glutenases,
mer was not cleaved by pepsin under simulated gastric condi-
so that its amino acid metabolites are rapidly assimilated into the
tions (Figure 2A). Similarly, both biomarkers were completely
bloodstream in a glutenase dose-dependent manner. And finally,
resistant to pepsin digestion over the course of 60 min (Figures
like the 33-mer, it must be able to penetrate across the intestinal
2B and 2C). To test whether the experimental glutenase EP-B2
epithelium. However, unlike 33-mer, the biomarker must not be
accelerates digestion of these biomarkers, we performed iden-
recognized by human TG2 or HLA-DQ2. Consequently, it must
tical digests with the addition of proEP-B2, the acid-activated
not stimulate an inflammatory response from disease-specific
proenzyme form of EP-B2. The addition of proEP-B2 rapidly
T cells. We present two examples of biomarkers that meet these
degraded the 33-mer (Figure 2D), as well as the derivative
criteria and demonstrate their utility in animal models.
biomarkers (Figures 2E and 2F) in a time- and dose-dependent
manner (Figures 2G–2I). In contrast, pepsin alone catalyzed
RESULTS
nearly complete cleavage of the myoglobin control peptide
within 10 min (Figure 2J). The major products of biomarker diges-
tion identified by LC-MS were similar to those produced by 33-
Design of Gluten Peptide-Based Biomarkers
In order to engineer a biomarker with the desired properties, we mer cleavage (Figure 2K), indicating that the substitutions in
replaced the reactive Gln residues in the 33-mer (QQQ-33-mer) these 33-mer analogs did not substantially alter the sites of
with either Asn (NNN-33-mer) or His (HHH-33-mer) residues susceptibility to EP-B2-mediated digestion.
(Figure 1). The rationale for the Q/N analog was that, within We next performed simulated duodenal digests to determine
the scope of naturally occurring (i.e., dietary) amino acids, this whether the biomarkers are resistant to pancreatic proteases
Chemistry & Biology 16, 868–881, August 28, 2009 ª2009 Elsevier Ltd All rights reserved 869

Chemistry & Biology
Gluten Peptide Analogs as Biomarkers
Figure 2. The 33-mer Gluten Peptide and Gluten Peptide-Based Biomarkers Are Similarly Resistant to Cleavage by Pepsin and Susceptible
to Cleavage by the Glutenase EP-B2
(A–F) Reverse-phase HPLC traces for QQQ-33-mer (A and D), NNN-33-mer (B and E), and HHH-33-mer (300 mM) (C and F) after simulated gastric digestion for
specified durations with either pepsin alone (A–C) or pepsin and 120 mg/mL EP-B2 (D–F). The TAME internal standard (T), intact peptide (peak 5), minimally pro-
cessed peptide lacking only the N-terminal LQ (peak 4), and other major digestion products are indicated for each peptide trace overlay.
(G–J) Integrated area-under-the-curve analysis for intact peptides QQQ-33-mer (G), NNN-33-mer (H), HHH-33-mer (I), and myoglobin peptide (J) showing dose
and time dependency of EP-B2-mediated digestion. Each peptide (300 mM) was digested in vitro with pepsin supplemented with specified concentrations of
EP-B2, and reaction products were analyzed by HPLC. The area under the curve for each intact peptide peak (and, where applicable, the minimally processed
-LQ peptide peak) was calculated and normalized to that for the internal standard. Values are expressed as the percentage of intact peptide remaining after
a given digestion duration relative to the initial peak area.
(K) LC-MS identification of major digestion products from simulated gastric digests with pepsin EP-B2. HPLC peak number corresponds to the peak numbers in
(A)–(F). Data shown are representative of three independent experiments.
(trypsin, chymotrypsin, elastase, and carboxypeptidase
A
previously reported (Shan et al., 2002), and also of both
[collectively, TCEC]), as well as to the exopeptidases contained biomarkers within 10 min (Figure 3A). Tandem mass spectrom-
in the intestinal brush border membrane (BBM). Following a 60 etry of biomarker digests revealed complementary EP-B2 and
min gastric digest containing either pepsin alone or pepsin and FM PEP cleavage patterns (Figure 3B), suggesting these
EP-B2, reactions were adjusted to pH 6.0 and commercial biomarkers can be used to assess efficacy of combination
TCEC was added with BBM purified from rat intestine. Intestinal enzyme therapies (Gass et al., 2007; Siegel et al., 2006). Indeed,
digests were carried out at 37ꢁC for 60 min, over which period addition of FM PEP enabled further digestion of fragments
several time points were taken and analyzed as above. Both remaining after treatment of the 33-mer and biomarkers with
biomarkers were highly resistant to high concentrations of intes- EP-B2 (Figures 3C–3E).
tinal proteases, with ꢀ80% of intact NNN-33-mer remaining
after 60 min, similar to 33-mer, and ꢀ60% of intact HHH-33- Biomarkers Are Noninflammatory in the Context
mer remaining (Figure 3A). Supplementation of these digests of Celiac Sprue
with the proline-specific glutenase Flavobacterium meningosep- For biomarkers to be administered safely to celiac sprue
ticum (FM) PEP resulted in complete digestion of the 33-mer, as patients, they must not be deamidated by TG2, bind HLA-DQ2,
870 Chemistry & Biology 16, 868–881, August 28, 2009 ª2009 Elsevier Ltd All rights reserved

Chemistry & Biology
Gluten Peptide Analogs as Biomarkers
Figure 3. The 33-mer Gluten Peptide and Gluten Peptide-Based Biomarkers Are Similarly Resistant to Cleavage by Pancreatic Enzymes
and Susceptible to Cleavage by Prolyl Endopeptidase
(A) Integrated area-under-the-curve analysis for intact QQQ-33-mer, NNN-33-mer, and HHH-33-mer after simulated intestinal digestion supplementation with
PEP from FM PEP. Following treatment with pepsin, each peptide (300 mM) was digested in vitro with pancreatic proteases (TCEC) and rat intestinal BBM 1.2 U/
mL FM PEP. Reaction products were analyzed by reverse-phase HPLC. The area under the curve for each intact peptide peak was calculated and normalized to
that for the internal standard. Values are expressed as the percentage of intact peptide remaining after a given duration relative to the initial peak area.
(B) Cleavage map derived from LC-MS/MS analysis of major digestion products following simulated gastrointestinal digests. Blue arrowheads designate major
cleavage sites resulting from EP-B2 supplementation. Red arrowheads designate major cleavage sites resulting from FM PEP supplementation. Underlined,
numbered sequences designate major products of digestion with EP-B2.
(C–E) HPLC traces for QQQ-33-mer (C), NNN-33-mer (D), and HHH-33-mer (E) (300 mM) after simulated gastric digestion with pepsin + 120 mg/mL EP-B2 for
60 min followed by treatment with TCEC + BBM 1.2 U/mL FM PEP for 10 or 60 min. T, TAME internal standard. HPLC peak numbers corresponds to the
numbered sequences in B. Data shown are representative of two independent experiments.
or stimulate a strong immune response by preexisting gluten-
The affinity of each biomarker for HLA-DQ2 was determined
specific T cells. To determine the capacity for gluten peptide- using a peptide exchange assay in which fluorescein-labeled
based biomarkers to elicit an inflammatory T cell response in peptides were incubated with soluble HLA-DQ2 molecules at
celiac sprue patients, these characteristics were tested in vitro pH 5.5 and at 37ꢁC to simulate the endocytic environment. After
and compared to the immunodominant 33-mer gluten peptide.
45 hr, DQ2-bound and free fluorescein-labeled peptides were
The extent to which each biomarker is deamidated by TG2 separated by high-performance size exclusion chromatography
was determined by a spectrophotometric assay in which the (HPSEC) and the ratio of their peak heights was determined by
ammonium ion released by TG2-catalyzed substrate deamida- fluorometry. Consistent with previous results (Xia et al., 2006),
tion is coupled to glutamate dehydrogenase-catalyzed oxidation >80% of synthetically deamidated 33-mer (EEE-33-mer) bound
of NADH (Piper et al., 2002). Consistent with previous results HLA-DQ2, a 9.9-fold increase over the native peptide and
(Shan et al., 2002), the 33-mer was readily deamidated by TG2 a 15.6-fold increase over the NNN-33-mer biomarker (Figure 4B).
(Figure 4A). In contrast, TG2 activity in the presence of NNN- Neither HHH-33-mer nor the myoglobin control peptide
33-mer and HHH-33-mer was significantly reduced, 33.1-fold exhibited any detectable binding to HLA-DQ2.
and 25.0-fold, respectively, relative to that in the presence of
The immunostimulatory capacity of each biomarker was
native 33-mer. Activity in the presence of the biomarkers was measured via T cell proliferation assays employing gluten-
slightly higher than that detected in the absence of a peptide specific T cells derived from celiac patient intestinal biopsies.
substrate, though this difference was not significant for HHH- In all three cell lines and both clones tested, nanomolar concen-
33-mer (p = 0.04 for NNN-33-mer). The control myoglobin trations of TG2-treated 33-mer were sufficient to induce prolifer-
peptide, which lacks glutamine residues entirely (Figure 1), eli- ation, with EC50 values ranging from ꢀ10 to 200 nM (Figure 4C).
cited no activity from TG2, suggesting the residual activity The low immunostimulatory capacity of the biomarkers toward
elicited by the biomarkers might be attributable to deamidation any of the cell lines and clones precluded response saturation
of glutamines at positions other than the preferred sites that and EC50 determination. However, it is apparent that the immu-
were synthetically altered.
nostimulatory capacity of NNN-33-mer was reduced ꢀ1000-fold
Chemistry & Biology 16, 868–881, August 28, 2009 ª2009 Elsevier Ltd All rights reserved 871

Chemistry & Biology
Gluten Peptide Analogs as Biomarkers
Figure 4. Biomarkers Are Noninflammatory in the Context of Celiac Sprue
(A) Specific activity of TG2 (5 mM) in the presence of 100 mM gluten peptide QQQ-33-mer, biomarkers NNN-33-mer or HHH-33-mer, control myoglobin peptide, or
no peptide. Means SD for triplicate assays are shown. Data are representative of three independent experiments. Statistical comparisons were performed with
respect to samples containing QQQ-33-mer. ***p < 0.001.
(B) Ratio of DQ2-bound to unbound fluorescein-conjugated (f-)peptides following incubation of thrombin-cleaved DQ2-aI-gliadin peptide complexes (9.4 mM)
with 0.185 mM f-QQQ-33-mer, f-EEE-33-mer, f-NNN-33-mer, -f-HHH-33-mer, or f-myoglobin peptide in a citrate-PBS buffer (pH 5.5) for 45 hr at 37ꢁC. Means
SD for triplicate assays are shown. Data are representative of three independent experiments. Statistical comparisons were performed with respect to samples
containing f-EEE-33-mer. **p < 0.01.
(C) Activation of celiac patient-derived intestinal T cell lines (TCL) and T cell clones (TCC) in response to incubation with DQ2-homozygous antigen-presenting
cells preloaded with varied concentrations of TG2-treated QQQ-33-mer, NNN-33-mer, or HHH-33-mer. Positive (2000 nM synthetically deamidated EEE-33-
mer) and negative (no peptide) controls are shown for comparison. Activation is measured in terms of counts per minute (cpm) of [3H]-thymidine incorporated into
harvested DNA of proliferating T cells. Means SD for triplicate wells are shown.
relative to TG2-treated 33-mer. The HHH-33-mer biomarker was from rats receiving no EP-B2 contained abundant gluten
even less immunogenic, eliciting minimal or no response at (Figure 5A). Virtually all of this material eluted after the HPLC
micromolar concentrations. Treatment of the biomarkers with internal standard (>15 min) (Figures 5A and 5B), corresponding
TG2 did not increase their immunostimulatory capacity.
to the later retention times reported for larger, immunogenic
gluten peptides (Gass et al., 2007). Administration of increasing
amounts of EP-B2 caused a progressive shift of the HPLC profile
from immunogenic peptides toward smaller, nontoxic metabo-
Biomarker Metabolism Parallels Gluten Digestion
In Vivo
The potential for noninflammatory gluten peptide-based lites in rats from groups 2 and 3 (Figures 5A and 5B). These results
biomarkers to report on in vivo gluten detoxification by oral glu- were confirmed by a competitive ELISA using a monoclonal anti-
tenase therapy was evaluated in rats. Fasted rats were adminis- body specific for a common gluten sequence (QPQLPY) (Moron
tered a meal containing 1 g gluten supplemented with 0, 10, or et al., 2008). Gastric contents from rats in group 3 (1:25::EP-
40 mg EP-B2 glutenase (for groups 1, 2, or 3, respectively; n = 4 B2:gluten) had significantly less reactive gluten epitopes
animals/group), as well as [13C3]-HHH-33-mer biomarker, [D3]- compared to control rats from group 1 (Figure 5C). Finally, the
myoglobin peptide control, and vancomycin (a nonabsorbable amount of intact 33-mer peptide, derived from ingested gluten,
internal dosing standard). Animals were euthanized 90 min after was measured in gastric contents by 3Q LC-MS/MS. Because
meal administration and gastrointestinal contents, as well as the 33-mer is a product of digestion by pancreatic proteases
plasma samples, were collected. Samples were analyzed by (Shan et al., 2002), gastric contents were treated ex vivo with
HPLC, competitive enzyme-linkedimmunosorbent assay (ELISA), trypsin and chymotrypsin prior to mass spectrometric analysis.
and/or liquid chromatography-assisted triple quadrupole tandem In agreement with previous results (Gass et al., 2006), the 33-
mass spectrometry (3Q LC-MS/MS).
mer was present at ꢀ100 mg/mL in the gastric contents of control
In all groups, minimal gluten had entered the small intestine by rats fed 1 g gluten, but was completely digested in animals
90 min after meal administration. In contrast, gastric samples receiving EP-B2 with their gluten meal (Figure 5D).
872 Chemistry & Biology 16, 868–881, August 28, 2009 ª2009 Elsevier Ltd All rights reserved

Chemistry & Biology
Gluten Peptide Analogs as Biomarkers
Figure 5. Glutenase-Mediated Biomarker
Metabolism Parallels Gluten Digestion
In Vivo
Three groups of rats (n = 4 animals/group) were
administered a gluten-containing meal supple-
mented with 0, 10, or 40 mg EP-B2 glutenase (for
groups 1, 2, or 3, respectively), 19.7 mg [13C3]-
HHH-33-mer biomarker, 3.3 mg [D3]-myoglobin
peptide control, and 10 mg vancomycin dosing
standard. Gastric contents were collected after
90 min and analyzed by HPLC, competitive ELISA,
and 3Q LC-MS/MS.
(A) Representative HPLC traces of gastric contents
from one animal per group. Internal standards for
dosing (V) and HPLC analysis (T) correspond
roughly to cut-offs between regions of nontoxic
metabolites, partially digested peptides, and
proteins and peptides retaining immunogenicity
(Gass et al., 2007). Asterisk indicates a nongluten
peak that was omitted from area-under-the-curve
analysis.
(B) Area-under-the-curve analysis quantifying the
data from (A). The area under the curve for
HPLC trace regions corresponding to nontoxic
metabolites (2.7–5.6 min), partially digested
peptides (6.3–15.1 min), and immunogenic
proteins and peptides (16.0–30.0 min) was
normalized to the area under the curve for the van-
comycin and TAME internal standards. Means
SD for each animal group are shown. Data are
representative of four similar experiments. Statis-
tical comparisons were performed with respect to
group 1 (no glutenase). **p < 0.01.
(C) Competitive ELISA analysis quantifying the
concentration of peptides containing the gluten
peptide sequence QPQLPY in gastric contents.
Samples from each animal were tested in tripli-
cate, with a typical error of 5%–10% of the sample
mean. Means
SD for each animal group are
shown. Data are representative of four similar
experiments. Statistical comparisons were per-
formed with respect to group 1 (no glutenase).
**p < 0.01.
(D) 3Q LC-MS/MS analysis quantifying the concentration of QQQ-33-mer (derived from ingested gluten), [13C3]-HHH-33-mer biomarker, and [D3]-myoglobin
peptide in gastric contents. Samples from each animal were tested in triplicate, with a typical error of <10% of the sample mean for group 1. Means SD
for each animal group are shown. Statistical comparisons were performed with respect to group 1 (no glutenase). *p < 0.05, **p < 0.01, ***p < 0.001.
To determine whether biomarker metabolism reflects the biomarker peptide in the plasma. The intact biomarker was not
reduction in gastric gluten mediated by EP-B2 in vivo, 3Q LC- detected in any rat plasma sample from this experiment.
MS/MS was used to measure gastric levels of [13C3]-HHH-33-
mer. The intact biomarker was present at similar levels (ꢀ65 Biomarker Transepithelial Transport Parallels Gluten
mg/mL) as the 33-mer in rats receiving no EP-B2 (Figure 5D). Par- Peptide Transport under Basal and Inflammatory
alleling the glutenase-mediated reduction observed for bulk Conditions
gluten and the 33-mer gluten peptide, supplementation of the In addition to reporting on glutenase activity in vivo, noninflam-
gluten meal with EP-B2 resulted in complete digestion of the matory gluten peptide-based biomarkers are potentially useful
biomarker. In contrast, the control myoglobin peptide was tools for understanding the factors and mechanisms that modu-
present at substantially lower levels in all animal groups.
late intestinal permeability of immunogenic dietary peptides, as
In animals that exhibited glutenase-enhanced gastric diges- well as for practical applications related to the diagnosis of celiac
tion of the labeled biomarker, the plasma concentration of the sprue and its treatment with modulators of epithelial perme-
labeled metabolite [13C]-leucine was expected to be corre- ability. To be used in such applications, these biomarkers must
spondingly higher. However, 3Q LC-MS/MS detection of [13C]- be similar to inflammatory gluten peptides in terms of their trans-
leucine and [D3]-leucine in the plasma was hindered by interfer- port and transepithelial stability across healthy and inflamed
ence from the far more abundant (ꢀ250 mM) plasma levels of mucosa.
unlabeled leucine. As an alternative approach, mass spectrom-
The T84 epithelial cell line was used to model the intestinal
etry was used to assay for the presence of intact labeled epithelium because its responsiveness to IFN-g, the major
Chemistry & Biology 16, 868–881, August 28, 2009 ª2009 Elsevier Ltd All rights reserved 873

Chemistry & Biology
Gluten Peptide Analogs as Biomarkers
inflammatory cytokine present in celiac lesions (Nilsen et al., in the test animal group relative to controls, as only one of four
1998), has been extensively studied (Bruewer et al., 2005; Ma- IFN-g-treated animals exhibited plasma biomarker. However,
dara and Stafford, 1989). Additionally, the effect of IFN-g on the level of plasma biomarker in this IFN-g-treated animal
the intact transport of the 33-mer and other gluten peptides (101.7 nM) was ꢀ10-fold higher than that in the control animal
across T84 epithelial monolayers has recently been described exhibiting detectable plasma biomarker (9.5 nM), a difference
(Bethune et al., 2009). To simulate transport under healthy and similar to that caused by IFN-g in vitro (Figures 6B and 6F). Addi-
inflammatory states, media alone or media containing IFN-g tionally, 33-mer dosed together with [13C3]-HHH-33-mer was
was incubated for 48 hr on the basolateral side of T84 epithelial detected in the plasma of both animals exhibiting plasma
cells culturedon transwell supports, and theapical-to-basolateral biomarker, but not in other animals, suggesting this inflammatory
flux of Cy5-labeled 33-mer and biomarkers was measured there- gluten peptide and its noninflammatory counterpart are trans-
after (Figure 6A). The flux of Cy5-33-mer and both Cy5-labeled ported in parallel in vivo.
biomarker peptides was ꢀ6 pmol/cm2/hr under basal conditions
A recent report demonstrated that the 33-mer is absorbed
(Figure 6B). Following exposure of T84 monolayers to IFN-g, the intact across the gut epithelium of an enteropathic gluten-sensi-
flux of all three peptides was increased ꢀ10-fold. No significant tive rhesus macaque, but not across that of a healthy control
difference in flux was observed between 33-mer and either (Bethune et al., 2008b). To determine if similar absorption of a
biomarker under basal or simulated inflammatory conditions.
noninflammatory gluten peptide-based biomarker occurs, a
Some processing of the 33-mer may occur upon its transport gluten-sensitive macaque with chronic diarrhea and elevated
across the intestinal epithelium (Matysiak-Budnik et al., 2003). plasma antigliadin antibodies was administered 100 mg [13C3]-
Therefore, to evaluate the stability of apical and translocated HHH-33-mer intragastrically. Peripheral blood samples were
Cy5-labeled biomarkers, the apical and basolateral media collected at hourly intervals and analyzed for biomarker content
were analyzed by LC-MS immediately after peptide addition to by 3Q LC-MS/MS. The labeled biomarker was clearly detected
the apical chamber and 10 hr thereafter. The absorbance at (2.3 0.1 nM) in peripheral blood 60 min after administration
640 nm was monitored during chromatographic separation. (Figure 6G), similar to the extent and rate of absorption reported
Both Cy5-labeled biomarkers, as well as Cy5-labeled 33-mer, for the 33-mer gluten peptide (Bethune et al., 2008b). Intragastri-
eluted between 9 and 10 min as a single major peak (Figures cally administered labeled biomarker was again detected in the
6C–6E). A smaller peak, eluting at ꢀ8 min in all samples, was peripheral blood of this gluten-sensitive macaque in a repeat
identified as Cy5-LQ, indicating that some processing of the experiment (2.2
0.5 nM). In contrast, biomarker transport
peptides’ N terminus occurred in the presence of epithelial cells. was not detected in this animal after clinical and serological
Nonetheless, after 10 hr, no other breakdown products were remission was achieved through treatment with a gluten-free
identifiable by mass spectrometry, and >95% of each intact diet, nor was it detected in two identically dosed healthy controls
peptide remained in the apical chamber of control cells (Fig- (Figure S2).
ure 6F). Intact Cy5 peptides were present at somewhat lower
levels (>80% of initial) after 10 hr in the apical chambers of those DISCUSSION
cells preincubated with IFN-g. This was due to increased
N-terminal processing, evidenced by the more apparent Cy5- Celiac sprue affects up to 1% of many human populations, but
LQ peak present in these samples, as well as to IFN-g-induced despite the wide prevalence and serious manifestations of the
enhancement of apical-to-basolateral flux. All three Cy5-labeled disease, the only treatment available remains a life-long gluten-
peptides remained intact during transport (Figures 6C–6E). After free diet. Compliance with this burdensome dietary treatment
10 hr, 0.2%–0.3% and 2%–3% of the initial 20 mM apical peptide is poor and recurrent exposure to gluten causes chronic inflam-
was observed intact on the basolateral side of cells preincubated mation, increased morbidity, and more serious health effects
with media alone or with IFN-g, respectively (Figure 6F). Thus, over time (Catassi et al., 2002; Corrao et al., 2001). Moreover,
both biomarkers were highly stable in the presence of epithelial in asymptomatic celiac sprue patients, disease management is
cells and were translocated intact to a similar extent as the 33- especially difficult, as invasive histological evaluations are the
mer under basal and simulated inflammatory conditions.
Various rodent models of impaired intestinal barrier function the development of non-dietary treatment alternatives to the
have demonstrated dependence of inflammation and gluten-free diet requires a long-term gluten challenge in celiac
only reliable way to assess response to a gluten-free diet. Finally,
a
increased permeability on IFN-g elevation (Demaude et al., patients, which is inherently problematic. Therefore, novel tools
2006; Luyer et al., 2007; Musch et al., 2002). To examine for monitoring compliance with the gluten-free diet and safely
biomarker transport under basal and inflammatory conditions evaluating non-dietary treatments in vivo are needed.
in vivo, catheterized rats were administered 20 mg [13C3]-HHH-
Here we used the disease-causing agent in celiac sprue as a
33-mer biomarker by oral gavage following 2 days of daily intra- template for designing noninflammatory surrogates as poten-
venous treatment with vehicle or IFN-g. The level of intact tially useful tools for fundamental and translational research.
peptide in peripheral plasma 60 min after peptide administration Celiac sprue is uniquely suited for this biomarker strategy
was determined by 3Q LC-MS/MS. As predicted by the lack of because we know the environmental trigger and have structural
a serological response to ingested gluten in rodents (March, information about its binding mode to the primary genetic deter-
2003), biomarker was detected in the plasma of only one of eight minant for the disease (Kim et al., 2004). Additionally, gluten
control rats administered oral biomarker (four in this study and peptides are extraordinarily stable in the relevant physiological
four in the glutenase study described above). Pretreatment compartment, making them ideal scaffolds for drug and
with IFN-g did not elicit a general increase in biomarker transport biomarker design. Research into the molecular basis for celiac
874 Chemistry & Biology 16, 868–881, August 28, 2009 ª2009 Elsevier Ltd All rights reserved

Chemistry & Biology
Gluten Peptide Analogs as Biomarkers
Figure 6. Biomarker Transepithelial Transport Parallels Gluten Peptide Transport In Vitro and In Vivo
(A) Experimental design for in vitro studies. Transwell supports bearing mature T84 epithelial cell monolayers were preincubated with media alone or with 600 U/
mL IFN-g in the basolateral chamber. After 48 hr, 20 mM Cy5-labeled QQQ-33-mer, NNN-33-mer, or HHH-33-mer was added to the apical chamber, and samples
from the apical and basolateral chambers were sampled over time to determine the stability and apical-to-basolateral flux of each intact peptide.
(B) Apical-to-basolateral flux of each Cy5-labeled peptide under basal (0 U/mL IFN-g) and simulated inflammatory (600 U/mL IFN-g) conditions. Means SD for
triplicate assays are shown. Data are representative of two similar experiments. Statistical comparisons were performed with respect to QQQ-33-mer; no signif-
icant differences were observed.
(C–E) Reverse-phase HPLC traces from LC-MS analysis of samples taken from the apical and basolateral chambers at 0 and 10 hr. Elution of Cy5-QQQ-33-mer
(C), Cy5-NNN-33-mer (D), and Cy5-HHH-33-mer (E) were monitored by their absorbance at 640 nm. Intact Cy5-peptides elute as the major peak between 9 and
10 min. The peak eluting in each trace at 8 min is Cy5-LQ, signifying limited processing of the N terminus by T84 cells.
(F) Area-under-the-curve analysis quantifying the data from C–E.
(G) 3Q LC-MS/MS analysis quantifying the concentration of [13C3]-HHH-33-mer in the peripheral plasma of a gluten-sensitive rhesus macaque (FH45) following
intragastric biomarker administration. Means SD for triplicate assays are shown. Statistical comparisons were performed with respect to the 0 min time point
(prior to biomarker administration). ****p < 0.0001.
sprue has elucidated many of the properties of gluten peptides gluten-specific T cells. Guided by these findings, our goal was
that allow them to persist through gastrointestinal proteolysis, to abrogate those properties that contribute to the inflammatory
access the gut-associated lymphoid tissue, and interact with response to gluten while retaining those properties that render
the key players in disease pathogenesis: TG2, DQ2, and biomarker metabolism and transport relevant to disease.
Chemistry & Biology 16, 868–881, August 28, 2009 ª2009 Elsevier Ltd All rights reserved 875

Chemistry & Biology
Gluten Peptide Analogs as Biomarkers
By altering key residues in a gluten peptide scaffold, we devel- role of IFN-g in enhancing intestinal permeability. In the gluten-
oped peptide biomarkers that mimic gluten peptides in their sensitive macaque, the extent and kinetics of biomarker trans-
resistance to gastrointestinal proteases and susceptibility to port during active disease were similar to those reported for
therapeutic glutenases, but that are neither substrates for TG2 the 33-mer itself in a previous experiment (Bethune et al.,
nor ligands for DQ2. As a result, these peptides are not pre- 2008b). Biomarker transport was not observed in healthy
sented to gluten-specific T cells and are consequently nonin- controls, also consistent with the 33-mer experiment (Bethune
flammatory. In the present study, the 33-mer from a2-gliadin et al., 2008b), nor was it observed in the gluten-sensitive animal
(Shan et al., 2002) was used as a scaffold for biomarker design following treatment with a gluten-free diet. Notwithstanding the
because its metabolism in the presence and absence of glute- clear need for more extensive animal studies and more sensitive
nases has been extensively characterized (Bethune et al., detection methods to validate this approach, these data suggest
2006; Gass et al., 2006; Shan et al., 2004, 2002) and because that the level of the intact peptide in blood (or urine) can be used
its intact transepithelial translocation has been demonstrated as a disease-relevant biomarker for intestinal barrier function.
(Bethune et al., 2008a; Schumann et al., 2008). Since the 33- Such a metric would facilitate the evaluation of drugs intended
mer is among the most stable of immunogenic gluten peptides to modulate intestinal permeability. One such drug candidate,
(Shan et al., 2002), other gluten peptides are likely to be digested AT-1001, has shown evidence of decreasing paracellular perme-
more quickly than the biomarkers described. Additionally, the ability in a diabetes-prone rat model and in preliminary clinical
observation that gluten peptide transepithelial flux scales trials (Paterson et al., 2007; Watts et al., 2005). In analogy to
inversely and linearly with size (Bethune et al., 2009) suggests the use of neutralizing antibodies against TNF-a to treat Crohn’s
that biomarkers based on smaller gluten peptides may provide disease (Baert et al., 1999), anti-IFN-g antibodies may represent
a more robust signal in applications measuring peptide trans- another candidate for reducing gut permeability and inflamma-
port. Other examples of disease-relevant gluten peptide tion in celiac sprue. Lastly, strict adherence to a gluten-free
sequences that are suitable for biomarker design include the diet reduces intestinal permeability in a majority of celiac
26-mer from g5-gliadin (Shan et al., 2005) and the p31-49 patients (Duerksen et al., 2005) and this reduction precedes
peptide (Sturgess et al., 1994). Since this latter peptide acts measurable improvements in histology (Cummins et al., 2001).
through the innate immune system, biomarkers based on the Therefore, biomarkers may be a useful clinical tool for monitoring
p31-49 peptide should be modified such that they no longer adherence to a gluten-free diet as well.
provoke expression of IL-15 and nonclassical major histocom-
Notwithstanding the effect of a gluten-free diet on intestinal
patibility complex class I molecules MIC and HLA-E from celiac permeability, epithelial uptake of gluten remains altered in
patient intestinal biopsy-derived enterocytes (Hue et al., 2004; treated celiac patients with respect to healthy controls (Friis
Maiuri et al., 2003).
et al., 1992). This is likely related to the 7.6-fold higher levels of
Oral protease therapy is one of the more promising non-die- IFN-g present in treated patients relative to healthy controls (Wa-
tary treatments being developed for celiac sprue (Bethune and penaar et al., 2004). In our experiments, both biomarkers were
Khosla, 2008), but few studies have been conducted in vivo. translocated intact across epithelial monolayers to a 10-fold
Celiac patients administered an undefined enzyme mixture greater extent following preincubation of the cells with IFN-g.
from animal digestive extracts showed modest improvement in This suggests that these peptides may be used as a screening
a clinical trial (Cornell et al., 2005). More recently, clinical efficacy tool for celiac sprue, en route to a diagnosis, as well as for other
of oral EP-B2 was demonstrated in a gluten-sensitive rhesus inflammatory bowel diseases in which intestinal IFN-g levels and
macaque (Bethune et al., 2008b). These studies relied on histo- mucosal leakiness of antigenic peptides and proteins are
logical, clinical, and serological readouts, complex parameters elevated (Clayburgh et al., 2004).
which require weeks to register a response and which are indi-
Finally, recent studies in cell and tissue culture provide
rect measures of glutenase-mediated gluten detoxification. In intriguing clues regarding the mechanisms underlying the intact
contrast, in the present study, the metabolism of [13C3]-HHH- absorption of gluten across the intestinal epithelium (Bethune
33-mer in rats dosed with EP-B2 provided an immediate and et al., 2009; Matysiak-Budnik et al., 2008; Schumann et al.,
direct readout for glutenase activity in vivo. This effect was 2008). Given the enormous potential for in vivo studies to
observed by mass spectrometric analysis of gastric contents. contribute to our understanding of gluten peptide transport in
Future studies will focus on optimizing noninvasive readouts celiac sprue (Friis et al., 1992), biomarkers are likely to be of
for biomarker metabolism, such as the measurement of considerable benefit to basic research in this area.
isotope-labeled (or alternatively labeled) amino acid metabolite
concentrations in bodily fluids such as serum or urine. Alterna- SIGNIFICANCE
tively, a stable isotope breath test (e.g., for 13CO2) may be adap-
ted to the detection of biomarker metabolites.
We used the disease-causing agent in celiac sprue as a
In preliminary animal studies, we observed intact biomarker template for designing two noninflammatory analogs for
absorption across the intestinal epithelium of a subset of tested use in diagnostic, therapeutic, and basic research applica-
rats as well as an enteropathic gluten-sensitive macaque. tions. By making rational amino acid substitutions in the
Whereas the results of our rodent studies support the hypothesis 33-mer gluten peptide using standard peptide synthesis
that proteolytically stable gluten oligopeptides can be trans- techniques, we produced biomarkers that retain the resis-
ported intact across the gut epithelium, they also highlight the tance to gastrointestinal proteases, susceptibility to thera-
need for a considerably larger study to quantify inherent perme- peutic glutenases, and permeability across enterocyte
ability differences between individual animals and to explore the barriers exhibited by the 33-mer. Unlike this proinflammatory
876 Chemistry & Biology 16, 868–881, August 28, 2009 ª2009 Elsevier Ltd All rights reserved

Chemistry & Biology
Gluten Peptide Analogs as Biomarkers
ground wheat. During gluten-free diet treatment, FH45 was administered
gluten-free monkey chow #5A7Q (Purina), as described (Bethune et al.,
2008a). The animals were housed under biosafety level two conditions in
accordance with the standards of the Association for Assessment and
Accreditation of Laboratory Animal Care. Investigators adhered to the Guide
for the Care and Use of Laboratory Animals prepared by the National
Research Council.
counterpart, however, they are not appreciably recognized
by key players in celiac sprue pathogenesis: transglutami-
nase, HLA-DQ2, and disease-specific T cells. Animal studies
conducted in rats and in gluten-sensitive macaques
confirmed that these biomarkers can detect intestinal
permeability changes as well as glutenase-catalyzed gastric
detoxification of gluten. Accordingly, controlled clinical
studies are warranted to evaluate the use of these peptides
as probes for abnormal intestinal permeability in celiac
patients and for glutenase efficacy in clinical trials and prac-
tice. In addition to facilitating diagnosis and management of
celiac sprue, these metastable peptides are potentially
useful tools for alimentary drug delivery, for basic mecha-
nistic studies of intestinal transport, and for clinical moni-
toring of intestinal permeability in other diseases of genetic
or infectious origin in which the gut barrier is impaired.
Finally, the strategy introduced in this study for rational
design of peptide-based biomarkers is instructive for the
development of additional gluten peptide-based biomarkers
for celiac sprue and for other gastrointestinal diseases for
which a dietary protein or pathogen trigger is known.
Peptide Synthesis, Labeling, and Purification
Peptides were synthesized using Boc/HBTU chemistry on solid phase as
previously described (Xia et al., 2005). To prepare isotope-labeled peptides
for the study of biomarker metabolism in vivo, [1-13C]-leucine was incorpo-
rated at positions 11, 18, and 25 in the HHH-33-mer biomarker (LQLQPF
(PQPHLPY)3PQPQPF; underlined) and [5,5,5-D3]-leucine was incorporated
at position 8 in the myoglobin peptide (KGHHEAELKPL; underlined). For
DQ2 binding assays, peptides were labeled at their N terminus on solid phase
with carboxyfluorescein, 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide, and
1-hydroxy-7-azabenzotriazole in 1:1:1 ratio and 10% (v/v) diisopropylethyl-
amine in dimethylformamide/methanol (2:1) as solvent. Following cleavage
from the resin, peptides were purified over a reverse-phase C18 column by
HPLC using a water/acetonitrile gradient in the presence of 0.1% trifluoroace-
tic acid, lyophilized, and stored at ꢂ20ꢁC. For transepithelial translocation
assays, purified peptides were labeled at their N terminus with Cy5-NHS ester
in DMSO according to the manufacturer’s instructions, repurified by HPLC,
lyophilized, and stored at ꢂ20ꢁC. The correct mass of all peptides was
confirmed by LC-MS. Prior to use, peptides were resuspended in 50 mM
sodium phosphate (pH 7.0) supplemented with 0.02% NaN3. The concentra-
tions of unlabeled and Cy5-labeled peptides were determined at pH 7.0 by
spectrophotometric measurement of A280 (3280 = 3840 Mꢂ1cmꢂ1) and A652
(3652 = 250,000 Mꢂ1cmꢂ1), respectively. Due to the absence of aromatic
residues in the myoglobin peptide, the concentration of the unlabeled
myoglobin peptide was determined by spectrophotometric measurement of
A205 (3205 = 27 (mg/mL)ꢂ1cmꢂ1) as previously described (Scopes, 1974). The
concentrations of fluorescein-labeled peptides were determined at pH 9.0
EXPERIMENTAL PROCEDURES
Materials
Peptide synthesis reagents were purchased from Chem-Impex, Peptides Inter-
national, Anaspec, and Novabiochem. Cy5-NHS ester was purchased from
Amersham Biosciences. Isotope-labeled amino acids were purchased from
Cambridge Isotope Laboratories. Recombinant IFN-g was purchased
from Peprotech, Inc. Cell culture media, antibiotics, human serum, 5-(and 6-)
carboxyfluorescein, and fluorescently labeled dextrans were purchased from
Invitrogen. Fetal bovine serum was purchased from Atlanta Biologicals. Gluten
flour was purchased from Bob’s Red Mill. Thrombin was purchased from
Novagen. Protease inhibitor cocktail set 1 was purchased from Calbiochem.
Pepsin, trypsin, chymotrypsin, elastase, carboxypeptidase A, vancomycin,
and Na-p-tosyl-L-arginine methyl ester hydrochloride (TAME) were purchased
from Sigma-Aldrich. BBMs were purified from rat intestine as previously
described (Kessler et al., 1978) and stored at ꢂ80ꢁC until use. The recombinant
proenzyme precursor of barley endoprotease EP-B2 (Bethune et al., 2006; Vora
et al., 2007), FM PEP (Shan et al., 2004), and human TG2 (Piper et al., 2002) were
expressed in E. coli and purified as previously described. Recombinant soluble
DQ2 heterodimer-aI gliadin peptide fusion molecules were prepared and puri-
fied in insect cells as previously described (Xia et al., 2005).
by spectrophotometric measurement of A495 (3495
= 80,200 M
ꢂ1cmꢂ1).
Working stocks were stored at 4ꢁC and their integrity confirmed periodically
by RP-HPLC.
Cell Culture
T84 epithelial cells from the American Type Culture Collection were grown in
T84 media (Dulbecco’s Modified Eagle Medium/Ham’s F12 [1:1] supple-
mented with antibiotics [penicillin/streptomycin] and 5% [v/v] fetal bovine
serum). Media was changed every alternate day and the cells were split
once a week. DQ2 homozygous antigen-presenting cells (CD114, an Epstein
Barr virus-transformed B lymphoblastoid cell line) were grown in antigen-pre-
senting cell media (RPMI supplemented with antibiotics and 5% [v/v] fetal
bovine serum). Every other day, CD114 were split to 0.4 3 106 cells/mL.
Gluten-specific, DQ2-restricted T cell lines (TCL 432.1.4, TCL 421.1.4, and
TCL 446.1.3) and T cell clones (TCC 436.5.3 [DQ2-a-II specific] and TCC
430.1.142 [DQ2-a-I and DQ2-a-III specific]) were isolated from celiac patient
intestinal biopsies and expanded as previously described (Molberg et al.,
2000). T cell proliferation assays were performed in T cell media (RPMI 1640
supplemented with antibiotics and 10% [v/v] human serum). All cells were
grown and assayed at 37ꢁC with 5% atmospheric CO2.
Animals
Twelve 8 week old male Wistar rats weighing between 250 and 300 g (Charles
River Laboratories) were singly housed in standard polycarbonate shoebox
cages measuring 10.5 3 19 3 8 inches with wire bar lids and microisolator
tops (Allentown, Inc.). Eight 8 week old male Wistar rats weighing between
250 and 300 g (Charles River Laboratories) with indwelling jugular vein cathe-
ters were housed similarly. Rats were allowed access to rodent chow #5010
(Purina) and water ad libitum prior to the onset of the studies. The room was
maintained on a 12:12 hr light/dark cycle. The ambient temperature remained
between 64ꢁF and 72ꢁF with a relative humidity of 30%–70%. All experimental
procedures were approved by the Animal Care and Use Committee of the
Tulane National Primate Research Center (Covington, LA).
Simulated Gastrointestinal Digests with Glutenase
Supplementation
In preparation for HPLC analysis, all buffers and reagents were filtered (0.2 mm)
prior to use in digests.
To simulate gastric digestion, peptides (300 mM) were incubated at 37ꢁC in
a 10 mM sodium acetate buffer (pH 4.5) with 1:10 (w/w) pepsin (120 mg/mL)
supplemented with 0, 6, 12, 24, 48, or 120 mg/mL of recombinant proEP-B2
glutenase. Samples were collected at 0, 10, 30, 45, and 60 min time points.
Samples were heat deactivated at 95ꢁC for 5 min, diluted 1:5 in HPLC solvent
A (95% H2O, 5% acetonitrile, and 0.1% trifluoroacetic acid) supplemented
with an internal standard (TAME), and analyzed by reverse-phase HPLC.
Samples (50 ml) were separated over a C18 column (Grace Vydac) using
a water-acetonitrile gradient in the presence of 0.1% TFA. The absorbance
Gluten-sensitive juvenile macaque FH45 (4.5 kg; male) was selected from
a population of rhesus macaques exhibiting clinical diarrhea, intestinal villous
blunting, and elevated AGA on a gluten-containing diet (Bethune et al.,
2008a). Healthy controls HI48 (2.65 kg; male) and HK31 (2.70 kg; male)
were selected from a population that exhibited no clinical or serological
responses to gluten intake. Throughout the study, animals consumed 4% of
their respective body weights daily of monkey chow #5K63 (Purina) containing
20% (by weight) crude protein including oats and gluten sources such as
Chemistry & Biology 16, 868–881, August 28, 2009 ª2009 Elsevier Ltd All rights reserved 877

Chemistry & Biology
Gluten Peptide Analogs as Biomarkers
at 215 nm was monitored. The area under the curve for each intact peptide was
calculated and normalized to the area under the curve for the internal standard
TAME. Due to the rapid EP-B2-catalyzed removal of the N-terminal LQ from
33-mer and biomarkers (Figures 1 and 2), this minimally processed product
was included as intact peptide in area-under-curve analyses.
and two T cell clones isolated from intestinal biopsies of HLA-DQ2+ celiac
patients were thawed and added (40,000 cells/well) to triplicate wells contain-
ing peptide-loaded antigen-presenting cells. Cells were coincubated 48 hr to
allow T cells to proliferate in response to DQ2-peptide complex stimulation.
After 48 hr, 1 mCi [3H]-thymidine was added to each well and cells were incu-
bated an additional 16 hr. Finally, DNA was collected on a filter mat using a cell
harvester and counts per minute resulting from incorporated [3H]-thymidine
were measured with a liquid scintillation counter.
Following simulated gastric digestion of each peptide (300 mM) with pepsin
120 mg/mL proEP-B2, digests were adjusted to pH 6.0 with sodium phosphate
buffer (50 mM, final concentration), and commercially available pancreatic
proteases trypsin (30 mg/mL), chymotrypsin (30 mg/mL), elastase (6 mg/mL),
and carboxypeptidase A (6 mg/mL), as well as 27 mg/mL rat intestinal BBM,
were added. Recombinant PEP from FM PEP was supplemented at 1.2 U/mL
when added. Simulated duodenal digests were performed at 37ꢁC. Samples
were collected at 0, 10, 30, and 60 min and processed for HPLC as described.
The area under the curve for each intact peptide (together with the minimally
processed -LQ product) was calculated and normalized to the internal stan-
dard.
Biomarker and Glutenase Dosing to Rats
Procedures for administration of gluten and therapeutics to rats were modified
from those previously described (Gass et al., 2006). The study comprised
3 days, including fasting, acclimation to the gluten test meal, and glutenase/
biomarker dosing. Adult rats (n = 4 for each of three groups) were fed
a commercially available rat chow until day 1 of the study. Rats were then
fasted for 12 hr, while drinking water remained freely available throughout
the study. On day 2, all three animal groups were administered a freshly
prepared gluten-sugar test meal (a dough ball composed of 1 g gluten, 0.6 g
white sugar, 0.6 g brown sugar, 0.35 g croscarmellose sodium, and 2.45 ml
water) to acclimate them to eating this meal following a fast. After 60 min,
the test meal was removed and animals were once again fasted for 24 hr.
On day 3, animals were administered a freshly prepared gluten-sugar test
meal supplemented with 0, 10, or 40 mg proEP-B2 (for groups 1, 2, or 3,
respectively), 19.7 mg [13C3]-HHH-33-mer, 3.3 mg [D3]-myoglobin peptide,
and 10 mg vancomycin. Vancomycin was added as a nonabsorbable dosing
internal standard, as previously described (Gass et al., 2006). Animals
consumed the test meal completely within 60 min, and animals were eutha-
nized 30 min thereafter (i.e., 90 min after meal administration). The gastric,
duodenal, jejunal, and ileal contents were collected immediately and stored
at ꢂ80ꢁC as previously described (Gass et al., 2006). Peripheral plasma
samples were collected via cardiac puncture at the time of euthanasia and
stored at ꢂ80ꢁC.
To identify digestion products, select samples were analyzed by LC-MS.
Samples (50 ml) processed for HPLC as described were injected on a
reverse-phase C18 HPLC system (Waters Corporation) coupled to a UV/Vis
detector and a ZQ single quadrupole mass spectrometer with an electrospray
ionization source. Samples were eluted with a water-acetonitrile gradient in the
presence of 0.1% formic acid. Absorbance at 214/254 nm and total ion current
were monitored, and spectra corresponding to major absorbance peaks were
examined.
Transglutaminase Deamidation Assay
Coupled spectrophotometric assays for TG2 activity in the presence of each
peptide were performed as previously described (Piper et al., 2002). Briefly,
each peptide (100 mM) was added to a 200 mM MOPS (pH 7.2) buffer contain-
ing 5 mM CaCl2, 1 mM Na4EDTA, 10 mM a-ketoglutarate, and 250 mM NADH.
Glutamate dehydrogenase was added to a final concentration of 0.036 U/ml,
and this incomplete reaction mixture was incubated at room temperature for
10 min to stabilize the initial absorbance at 340 nm. Finally, 5 mM TG2 was
added and A340 was monitored. The specific activity of the enzyme in the
presence of each peptide was calculated from the rate of NADH consumption
(3340 = 6220 Mꢂ1cmꢂ1).
Reverse-Phase HPLC Analysis of Rat Gastrointestinal Contents
Gastric samples (100 mg) were thawed on ice and suspended in 190 ml 0.01 M
HCl and 10 ml 10 mM leupeptin (an inhibitor of EP-B2, to prevent digestion of
material ex vivo). Suspensions were incubated 10 min at 37ꢁC (pH 2.5), and
then 50 mM sodium phosphate (pH 6.0) and sodium hydroxide were added
to increase the pH of the suspensions to above 6.0. Trypsin (0.375 mg/mL)
and chymotrypsin (0.375 mg/mL) were added to maximize dissolution of
gluten, and reactions were incubated 30 min at 37ꢁC. Samples were heat
deactivated for 5 min at 95ꢁC, supplemented with ethanol to a final concentra-
tion of 70% (v/v), and centrifuged for 10 min at 16,100 3 g. Syringe-filtered
(0.45 mm) supernatants (100 ml) were diluted 1:5 in 95% H2O, 5% acetonitrile,
and 0.1% trifluoroacetic acid supplemented with an internal standard (TAME)
and analyzed by reverse-phase HPLC. Samples (50 ml) were separated over
a C18 column (Grace Vydac) using a water-acetonitrile gradient in the presence
of 0.1% TFA. The absorbance at 215 nm was monitored. Intestinal flushes
were thawed on ice and centrifuged for 10 min at 4ꢁC, 3100 3 g. Supernatants
were heat deactivated and processed for HPLC as described for gastric
contents.
HLA-DQ2 Binding Assay
Peptide exchange assays for determining the equilibrium occupancy of each
peptide on HLA-DQ2 were modified from previously described methods
(Xia et al., 2006). Briefly, recombinant DQ2-aI-gliadin peptide fusions
(35 mM) were treated with 0.15 U/ml thrombin in phosphate-buffered saline
(PBS) (pH 7.4) supplemented with 0.02% (w/v) NaN3 for 2 hr at 4ꢁC, after
which protease inhibitor cocktail was added. Thrombin-cleaved DQ2-aI-
gliadin peptide complexes (9.4 mM) were incubated with 0.185 mM fluores-
cein-conjugated peptides in a citrate-PBS buffer (pH 5.5) (24 mM sodium
citrate, 55 mM Na2HPO4, 75 mM NaCl, and 0.02% [w/v)] NaN3) for 45 hr at
37ꢁC. To quantify the equilibrium occupancy of each fluorescent peptide on
DQ2, binding reactions were diluted 1:5 in PBS (pH 7.4) supplemented with
0.02% (w/v) NaN3 and 12.5 ml was injected on an HPSEC system coupled
to a fluorescence detector (Shimadzu). DQ2-bound and unbound fluorescent
peptides were separated using a BioSep 3000 size exclusion column (Phe-
nomenex) with a flow rate of 1 mL/min PBS (pH 7.4) supplemented with
0.02% (w/v) NaN3. The detector was set to monitor excitation at 495 nm
and emission at 520 nm. The bound/free ratio for each peptide was calculated
by dividing the measured peak height for bound peptide by that for the free
peptide.
Competitive ELISA on Rat Gastric Contents
Gastric samples (100 mg) were thawed on ice, suspended in 950 ml 70% (v/v)
ethanol and 50 ml 10 mM leupeptin, incubated 10 min at 37ꢁC, and then centri-
fuged for 10 min at 16,100 3 g. Syringe-filtered (0.45 mm) supernatants were
tested for the gluten sequence QPQLPY using a monoclonal antibody-based
competitive ELISA modified from previous methods (Moron et al., 2008).
Briefly, equal volumes of coating solution (5 mg/mL gliadin [Sigma-Aldrich] in
20 mM PBS [pH 7.2]) and 20 mM sodium bicarbonate (pH 9.6) were added
to 96-well microtiter plates (Nunc Maxisorp) and incubated 1 hr at 37ꢁC and
overnight at 4ꢁC. The next day, gliadin-coated plates were washed twice
with washing buffer (PBS [pH 7.2] containing 0.05% Tween-20) and then
blocked with blocking buffer (5% [w/v] nonfat milk in PBS [pH 7.2]) for 2 hr
at room temperature. Synthetic 33-mer standard (10ꢂ5 ꢂ 10 mg/mL) or gastric
supernatants were serially diluted in assay buffer (3% [w/v] bovine serum
albumin in PBS [pH 7.2]). An equal volume of G12-HRP monoclonal anti-
body-horseradish peroxidase conjugate (Biomedal) diluted 1:10,000 in assay
T Cell Proliferation Assay
The T cell proliferation assay was modified from previously described methods
(Qiao et al., 2005), as follows. Briefly, the 33-mer, NNN-33-mer, and HHH-33-
mer peptide stocks (250 mM) were deamidated by treatment with 100 mg/mL
TG2 in 100 mM Tris (pH 7.4) for 2 hr at 37ꢁC in the presence of 2 mM CaCl2.
Antigen-presenting cells (CD114; 60,000 cells/well) were irradiated (80 Gy)
and incubated overnight at 37ꢁC in U-bottom, 96-well plates with various
concentrations of native or TG2-deamidated peptides in T cell media. As posi-
tive and negative controls, 2 mM synthetically deamidated EEE-33-mer or no
peptide was added to antigen-presenting cells. The next day, three T cell lines
878 Chemistry & Biology 16, 868–881, August 28, 2009 ª2009 Elsevier Ltd All rights reserved

Chemistry & Biology
Gluten Peptide Analogs as Biomarkers
buffer was added to each standard or sample dilution. Mixtures were incu-
bated with gentle agitation for 2 hr at room temperature, and then added to
plate wells in triplicate. After 30 min incubation at room temperature, wells
were washed five times with washing buffer and TMB liquid substrate solution
(Sigma-Aldrich) was added to wells. The reaction was stopped after a 30 min
dark incubation by addition of an equal volume of 1 M sulfuric acid and the
absorbance at 450 nm was measured. Origin 6.0 (OriginLab) was used to fit
the 33-mer standard curve to the sigmoidal model:
Peptide Translocation Assays
Peptide translocation assays were performed as previously described
(Bethune et al., 2009). Briefly, cultured T84 cells were seeded on rat tail
collagen-coated polycarbonate transwell permeable supports (5 mm pore
size, 6.5 mm diameter; Corning Life Sciences) at 5 3 104 cells/well and the
media was exchanged every other day for 2 weeks while the cells grew to
confluence and formed tight junctions. Following maturation, cell monolayers
were preincubated for 48 hr with basolateral media containing either 0 or
600 U/mL recombinant IFN-g. After preincubation, the translocation assay
was performed by replacing media in both the apical and basolateral cham-
bers with serum-free T84 media (1:1::Dulbecco’s Modified Eagle Medium:
Ham’s F12 media supplemented with antibiotics) and adding 2 mM dextran
(3,000 molecular weight)-Alexa Fluor-488, 2 mM dextran (70,000 molecular
weight)-Texas red, and 20 mM Cy5-labeled peptide to the apical chamber.
Labeled dextrans were added to confirm monolayer integrity and size-selec-
tive transport (Bethune et al., 2009). Samples of the apical and basolateral
media were taken at the 0 hr time point, and basolateral samples were taken
every hour over a 4 hr experiment. Fluorescence in collected samples was
measured in 96-well format on a Flexstation II 384 (Molecular Devices), moni-
toring three channels (excitation 490 nm and emission 525 nm for Alexa Fluor-
488; excitation 585 nm and emission 620 nm for Texas red; excitation 640 nm
and emission 675 nm for Cy5). The slope of basolateral fluorescence units
versus time (from 1 to 4 hr) was calibrated to the initial apical fluorescence
and divided by the permeable support area (0.33 cm2) to yield the transepithe-
lial flux (pmol/cm2/hr).
Â
Ã
n
A450 = A450min + ðA450max ꢂ A450minÞ= 1 + ðx=IC50
Þ
;
where x is the peptide concentration, IC50 is the 33-mer concentration at which
competition is half-maximal, and n is the Hill slope. The concentration of
peptides containing the sequence QPQLPY in gastric samples was deter-
mined by comparison to the linear portion of the 33-mer standard curve.
3Q LC-MS/MS Analysis of Intact Peptides and Biomarker
Metabolites
The amounts of gluten-derived 33-mer,
[
13C3]-HHH-33-mer, and [D3]-
myoglobin peptide present in rat gastric contents and plasma were deter-
mined using a Micromass Quattro Premier triple quadrupole LC-MS system.
Gastric samples (100 mg) were thawed on ice and suspended in 950 ml
0.01 M HCl and 50 ml 10 mM leupeptin. Suspensions were adjusted to
pH 6.0 and pancreatic proteases were added as described above to release
the 33-mer from gluten present in the samples. After 30 min, samples were
heat deactivated for 5 min at 95ꢁC and centrifuged for 10 min at 16,100 3 g.
Prior to 3Q LC-MS/MS analysis, syringe-filtered (0.45 mm) gastric superna-
tants, or plasma samples, were depleted of larger proteins by addition of
acetonitrile. Samples were mixed with an equal volume of cold acetonitrile
containing 0.1% formic acid and 200 nM NNN-33-mer as an internal standard.
Samples were vortexed, incubated for 2 hr at 4ꢁC, and centrifuged for 10 min at
4ꢁC, 16,100 3 g. Supernatants were mixed with an equal volume of 0.1%
formic acid in water to dilute the acetonitrile concentration to 25% and used
directly for intact peptide analysis or, for plasma samples only, processed
for metabolite analysis as below.
Chromatographic and Mass Spectrometric Analysis
of Translocated Gluten Peptides
During the peptide translocation assay described above, additional samples
from the apical and basolateral chambers were collected at 0 and 10 hr for
analysis of peptide stability and intact translocation. Samples (50 mL) were
analyzed by LC-MS as described for digests, except absorbance at 640 nm
was monitored. Spectra corresponding to A640 peaks were examined.
Samples were also analyzed by HPSEC with fluorescence detection as
described for DQ2-peptide binding analysis, except the detector was set to
monitor excitation at 647 nm and emission at 665 nm.
Mass spectrometry analysis of intact peptides was performed as previously
described (Gass et al., 2006) with the following modifications. Samples were
injected in triplicate (80 ml each) and eluted with a water-acetonitrile gradient
in the presence of 0.1% formic acid. For 33-mer detection, positive ion SRM
Biomarker Dosing in Rats following IFN-g Treatment
mode was used for monitoring the transitions of ions at m/z 978.84+
/
Adult catheterized rats (n = 4 for each of two groups) were administered a daily
dose of vehicle (PBS) or 108 U/m2 IFN-g intravenously via catheter for 2 days.
On day 3, 48 hr after the initial dose of vehicle or IFN-g, all animals were admin-
istered 0.5 ml water containing 20 mg 33-mer and 20 mg [13C3]-HHH-33-mer
via oral gavage. Animals were euthanized 60 min thereafter and peripheral
plasma samples were collected, stored, and tested via 3Q LC-MS/MS for
intact peptides as described above.
263.2+ (30 V cone voltage, 27 eV collision energy) for the quantification assay
and m/z 1304.73+/263.2+ (40 V cone voltage, 50 eV collision energy) as a
confirmatory transition. For [13C3]-HHH-33-mer detection, the transition moni-
tored was m/z 986.754+/263.2+ (45 V cone voltage, 32 eV collision energy) for
the quantification assay. For [D3]-myoglobin peptide detection, the transitions
monitored were m/z 632.72+/229.2+ (35 V cone voltage, 25 eV collision
energy) for the quantification assay and m/z 632.72+/129.2+ (35 V cone
voltage, 33 eV collision energy) as a confirmatory transition. For NNN-33-
mer internal standard, the transitions monitored were m/z 968.64+/263.4+
(32 V cone voltage, 32 eV collision energy) for the quantification assay and
m/z 968.64+/226.0+ (40 V cone voltage, 50 eV collision energy) as a confirma-
tory transition. Levels of 33-mer, isotope-labeled biomarker, and isotope-
labeled myoglobin control peptide in each sample were determined by
comparison of the area under their transition peaks to the area under the
NNN-33-mer internal standard transition peak and to a calibration curve
corresponding to each peptide.
Biomarker Inoculation in Gluten-Sensitive
and Healthy Rhesus Macaques
Prior to the study, gluten-sensitive macaque FH45 was administered a gluten-
containing diet and exhibited elevated antigliadin antibodies, intestinal villous
blunting, and clinical symptoms indicative of gluten sensitivity (Bethune et al.,
2008a). On the day of the study, a dose of 100 mg of isotopically labeled
biomarker ([13C3]-HHH-33-mer) dissolved in 10 ml of Gatorade was adminis-
tered directly into the fasted stomach of FH45 by intragastric tube. A 0.5 ml
sample of EDTA-blood was collected from an ear vein at 0, 60, 120, 180,
and 240 min following biomarker inoculation. The experiment was repeated
4 months later with FH45 and two healthy control macaques, HI48 and
HK31, on a gluten-containing diet and again with FH45 in clinical and serolog-
ical remission after 2 months of treatment with a gluten-free diet. In these later
experiments, 50 mg of isotopically labeled biomarker was intragastrically
administered to each animal. Animals were sedated and anesthetized prior
to peptide inoculation and blood collections. Plasma samples were analyzed
for intact biomarker by 3Q LC-MS/MS as described above.
Mass spectrometry analysis of biomarker metabolites was performed on
plasma samples processed as above. Protein-depleted samples (30 ml) were
dried and resuspended in 300 ml 20 mM ammonium acetate (pH 5.0) containing
100 nM [D10]-leucine as an internal standard. Samples were injected in tripli-
cate (20 ml each) and eluted from an Atlantis T3 column (3 mm, 2.1 3 100 mm;
Waters) with a 20 mM ammonium acetate (pH 5.0)-acetonitrile gradient. For
leucine quantification, the transitions monitored were m/z 132.2+/86.2+
(15 V cone voltage, 10 eV collision energy). For [13C]-leucine quantification,
the transitions monitored were m/z 133.2+/86.2+ (15 V cone voltage, 10 eV
collision energy). For [D3]-leucine quantification, the transitions monitored
were m/z 135.2+/89.2+ (15 V cone voltage, 10 eV collision energy.). For
[D10]-leucine internal standard quantification, the transitions monitored were
m/z 142.2+/96.2+ (15 V cone voltage, 10 eV collision energy.)
Statistics
Statistical comparisons were conducted using a two-tailed Student’s t test
assuming unequal variances. A statistical probability of p < 0.05 was consid-
ered significant.
Chemistry & Biology 16, 868–881, August 28, 2009 ª2009 Elsevier Ltd All rights reserved 879

Chemistry & Biology
Gluten Peptide Analogs as Biomarkers
SUPPLEMENTAL DATA
Ciacci, C., Cirillo, M., Cavallaro, R., and Mazzacca, G. (2002). Long-term
follow-up of celiac adults on gluten-free diet: prevalence and correlates of
intestinal damage. Digestion 66, 178–185.
Supplemental Data include two figures and can be found with this article online at
http://www.cell.com/chemistry-biology/supplemental/S1074-5521(09)00240-3.
Clayburgh, D.R., Shen, L., and Turner, J.R. (2004). A porous defense: the leaky
epithelial barrier in intestinal disease. Lab. Invest. 84, 282–291.
ACKNOWLEDGMENTS
Cornell, H.J., Macrae, F.A., Melny, J., Pizzey, C.J., Cook, F., Mason, S.,
Bhathal, P.S., and Stelmasiak, T. (2005). Enzyme therapy for management of
coeliac disease. Scand. J. Gastroenterol. 40, 1304–1312.
We thank Karolina Krasinska and Chris Adams at the Stanford University
Vincent Coates Foundation Mass Spectrometry Laboratory. We also thank
Bele´ n Moro´ n for instruction on the competitive ELISA, Xi Jin for generously
providing us recombinant TG2, Samuel Howles-Banerji for assistance with
peptide synthesis, and Edith Martin and Amanda Tardo for technical assis-
tance. We gratefully acknowledge Eric Sibley for provision of the T84 cell
line. This work was supported by the Research Council of Norway (to
L.M.S.) and National Institutes of Health grants 5R01DK76653-2 (to K.S.)
and R01 DK063158 (to C.K.). C.K. is a Director and stockholder in Alvine Phar-
maceuticals, a company that is developing an oral enzyme drug for celiac
disease. L.M.S. is a Scientific Advisor and stockholder in Alvine Pharmaceuti-
cals. None of the other authors have any competing financial interests to
disclose.
Corrao, G., Corazza, G.R., Bagnardi, V., Brusco, G., Ciacci, C., Cottone, M.,
Sategna Guidetti, C., Usai, P., Cesari, P., Pelli, M.A., et al. (2001). Mortality
in patients with coeliac disease and their relatives: a cohort study. Lancet
358, 356–361.
Cummins, A.G., Thompson, F.M., Butler, R.N., Cassidy, J.C., Gillis, D.,
Lorenzetti, M., Southcott, E.K., and Wilson, P.C. (2001). Improvement in intes-
tinal permeability precedes morphometric recovery of the small intestine in
coeliac disease. Clin. Sci. (Lond.) 100, 379–386.
Demaude, J., Salvador-Cartier, C., Fioramonti, J., Ferrier, L., and Bueno, L.
(2006). Phenotypic changes in colonocytes following acute stress or activation
of mast cells in mice: implications for delayed epithelial barrier dysfunction.
Gut 55, 655–661.
Received: May 17, 2009
Revised: July 17, 2009
Accepted: July 20, 2009
Published: August 27, 2009
Dicke, W.K., Weijers, H.A., and Van De Kamer, J.H. (1953). Coeliac disease. II.
The presence in wheat of a factor having a deleterious effect in cases of coeliac
disease. Acta Paediatr. 42, 34–42.
Dieterich, W., Ehnis, T., Bauer, M., Donner, P., Volta, U., Riecken, E.O., and
Schuppan, D. (1997). Identification of tissue transglutaminase as the autoanti-
gen of celiac disease. Nat. Med. 3, 797–801.
REFERENCES
Alaedini, A., and Green, P.H. (2005). Narrative review: celiac disease: under-
Duerksen, D.R., Wilhelm-Boyles, C., and Parry, D.M. (2005). Intestinal perme-
ability in long-term follow-up of patients with celiac disease on a gluten-free
diet. Dig. Dis. Sci. 50, 785–790.
standing a complex autoimmune disorder. Ann. Intern. Med. 142, 289–298.
Anderson, R.P., van Heel, D.A., Tye-Din, J.A., Barnardo, M., Salio, M., Jewell,
D.P., and Hill, A.V. (2005). T cells in peripheral blood after gluten challenge in
coeliac disease. Gut 54, 1217–1223.
Friis, S., Dabelsteen, E., Sjostrom, H., Noren, O., and Jarnum, S. (1992).
Gliadin uptake in human enterocytes. Differences between coeliac patients
in remission and control individuals. Gut 33, 1487–1492.
Arentz-Hansen, H., Korner, R., Molberg, O., Quarsten, H., Vader, W., Kooy,
Y.M., Lundin, K.E., Koning, F., Roepstorff, P., Sollid, L.M., and McAdam,
S.N. (2000). The intestinal T cell response to alpha-gliadin in adult celiac
disease is focused on a single deamidated glutamine targeted by tissue trans-
glutaminase. J. Exp. Med. 191, 603–612.
Gass, J., Vora, H., Bethune, M.T., Gray, G.M., and Khosla, C. (2006). Effect of
barley endoprotease EP-B2 on gluten digestion in the intact rat. J. Pharmacol.
Exp. Ther. 318, 1178–1186.
Gass, J., Bethune, M.T., Siegel, M., Spencer, A., and Khosla, C. (2007). Combi-
nation enzyme therapy for gastric digestion of dietary gluten in patients with
celiac sprue. Gastroenterology 133, 472–480.
Baert, F.J., D’Haens, G.R., Peeters, M., Hiele, M.I., Schaible, T.F., Shealy, D.,
Geboes, K., and Rutgeerts, P.J. (1999). Tumor necrosis factor alpha antibody
(infliximab) therapy profoundly down-regulates the inflammation in Crohn’s
ileocolitis. Gastroenterology 116, 22–28.
Green, P.H., and Jabri, B. (2006). Celiac disease. Annu. Rev. Med. 57,
207–221.
Bethune, M.T., and Khosla, C. (2008). Parallels between pathogens and gluten
Hue, S., Mention, J.J., Monteiro, R.C., Zhang, S., Cellier, C., Schmitz, J.,
Verkarre, V., Fodil, N., Bahram, S., Cerf-Bensussan, N., and Caillat-Zucman,
S. (2004). A direct role for NKG2D/MICA interaction in villous atrophy during
celiac disease. Immunity 21, 367–377.
peptides in celiac sprue. PLoS Pathog. 4, e34.
Bethune, M.T., Strop, P., Tang, Y., Sollid, L.M., and Khosla, C. (2006). Heter-
ologous expression, purification, refolding, and structural-functional charac-
terization of EP-B2, a self-activating barley cysteine endoprotease. Chem.
Biol. 13, 637–647.
Jabri, B., Kasarda, D.D., and Green, P.H. (2005). Innate and adaptive immu-
nity: the yin and yang of celiac disease. Immunol. Rev. 206, 219–231.
Bethune, M.T., Borda, J.T., Ribka, E., Liu, M.X., Phillippi-Falkenstein, K.,
Jandacek, R.J., Doxiadis, G.G., Gray, G.M., Khosla, C., and Sestak, K.
(2008a). A non-human primate model for gluten sensitivity. PLoS ONE 3,
e1614.
Kessler, M., Acuto, O., Storelli, C., Murer, H., Muller, M., and Semenza, G.
(1978). A modified procedure for the rapid preparation of efficiently transport-
ing vesicles from small intestinal brush border membranes. Their use in inves-
tigating some properties of D-glucose and choline transport systems.
Biochim. Biophys. Acta 506, 136–154.
Bethune, M.T., Ribka, E., Khosla, C., and Sestak, K. (2008b). Transepithelial
transport and enzymatic detoxification of gluten in gluten-sensitive rhesus
macaques. PLoS ONE 26, e1857.
Kim, C.Y., Quarsten, H., Bergseng, E., Khosla, C., and Sollid, L.M. (2004).
Structural basis for HLA-DQ2-mediated presentation of gluten epitopes in
celiac disease. Proc. Natl. Acad. Sci. USA 101, 4175–4179.
Bethune, M.T., Siegel, M., Howles-Banerji, S., and Khosla, C. (2009). Inter-
feron-gamma released by gluten-stimulated celiac disease-specific intestinal
T cells enhances the transepithelial flux of gluten peptides. J. Pharmacol. Exp.
Ther. 329, 657–668.
Koskinen, O., Collin, P., Korponay-Szabo, I., Salmi, T., Iltanen, S., Haimila, K.,
Partanen, J., Maki, M., and Kaukinen, K. (2008). Gluten-dependent small
bowel mucosal transglutaminase 2-specific IgA deposits in overt and mild
enteropathy coeliac disease. J. Pediatr. Gastroenterol. Nutr. 47, 436–442.
Bruewer, M., Utech, M., Ivanov, A.I., Hopkins, A.M., Parkos, C.A., and Nusrat,
A. (2005). Interferon-gamma induces internalization of epithelial tight junction
proteins via a macropinocytosis-like process. FASEB J. 19, 923–933.
Luyer, M.D., Buurman, W.A., Hadfoune, M., Wolfs, T., van’t Veer, C., Jacobs,
J.A., Dejong, C.H., and Greve, J.W. (2007). Exposure to bacterial DNA before
hemorrhagic shock strongly aggravates systemic inflammation and gut barrier
loss via an IFN-gamma-dependent route. Ann. Surg. 245, 795–802.
Catassi, C., Fabiani, E., Corrao, G., Barbato, M., De Renzo, A., Carella, A.M.,
Gabrielli, A., Leoni, P., Carroccio, A., Baldassarre, M., et al. (2002). Risk of non-
Hodgkin lymphoma in celiac disease. JAMA 287, 1413–1419.
880 Chemistry & Biology 16, 868–881, August 28, 2009 ª2009 Elsevier Ltd All rights reserved

Chemistry & Biology
Gluten Peptide Analogs as Biomarkers
Madara, J.L., and Stafford, J. (1989). Interferon-gamma directly affects barrier
function of cultured intestinal epithelial monolayers. J. Clin. Invest. 83, 724–
727.
Raki, M., Tollefsen, S., Molberg, O., Lundin, K.E., Sollid, L.M., and Jahnsen, F.L.
(2006). A unique dendritic cell subset accumulates in the celiac lesion and effi-
ciently activates gluten-reactive T cells. Gastroenterology 131, 428–438.
Raki, M., Fallang, L.E., Brottveit, M., Bergseng, E., Quarsten, H., Lundin, K.E.,
and Sollid, L.M. (2007). Tetramer visualization of gut-homing gluten-specific
T cells in the peripheral blood of celiac disease patients. Proc. Natl. Acad.
Sci. USA 104, 2831–2836.
Maiuri, L., Ciacci, C., Ricciardelli, I., Vacca, L., Raia, V., Auricchio, S., Picard,
J., Osman, M., Quaratino, S., and Londei, M. (2003). Association between
innate response to gliadin and activation of pathogenic T cells in coeliac
disease. Lancet 362, 30–37.
Schumann, M., Richter, J.F., Wedell, I., Moos, V., Zimmermann-Kordmann,
M., Schneider, T., Daum, S., Zeitz, M., Fromm, M., and Schulzke, J.D.
(2008). Mechanisms of epithelial translocation of the alpha(2)-gliadin-33mer
in coeliac sprue. Gut 57, 747–754.
March, J.B. (2003). High antigliadin IgG titers in laboratory rabbits fed a wheat-
containing diet: a model for celiac disease? Dig. Dis. Sci. 48, 608–610.
Matysiak-Budnik, T., Candalh, C., Dugave, C., Namane, A., Cellier, C.,
Cerf-Bensussan, N., and Heyman, M. (2003). Alterations of the intestinal trans-
port and processing of gliadin peptides in celiac disease. Gastroenterology
125, 696–707.
Scopes, R.K. (1974). Measurement of protein by spectrophotometry at 205
nm. Anal. Biochem. 59, 277–282.
See, J., and Murray, J.A. (2006). Gluten-free diet: the medical and nutrition
Matysiak-Budnik, T., Moura, I.C., Arcos-Fajardo, M., Lebreton, C., Menard, S.,
Candalh, C., Ben-Khalifa, K., Dugave, C., Tamouza, H., van Niel, G., et al.
(2008). Secretory IgA mediates retrotranscytosis of intact gliadin peptides
via the transferrin receptor in celiac disease. J. Exp. Med. 205, 143–154.
management of celiac disease. Nutr. Clin. Pract. 21, 1–15.
Shan, L., Molberg, O., Parrot, I., Hausch, F., Filiz, F., Gray, G.M., Sollid, L.M.,
and Khosla, C. (2002). Structural basis for gluten intolerance in celiac sprue.
Science 297, 2275–2279.
Molberg, O., McAdam, S.N., Korner, R., Quarsten, H., Kristiansen, C., Madsen,
L., Fugger, L., Scott, H., Noren, O., Roepstorff, P., et al. (1998). Tissue trans-
glutaminase selectively modifies gliadin peptides that are recognized by
gut-derived T cells in celiac disease. Nat. Med. 4, 713–717.
Shan, L., Marti, T., Sollid, L.M., Gray, G.M., and Khosla, C. (2004). Comparative
biochemical analysis of three bacterial prolyl endopeptidases: implications for
coeliac sprue. Biochem. J. 383, 311–318.
Shan, L., Qiao, S.W., Arentz-Hansen, H., Molberg, O., Gray, G.M., Sollid, L.M.,
and Khosla, C. (2005). Identification and analysis of multivalent proteolytically
resistant peptides from gluten: implications for celiac sprue. J. Proteome Res.
4, 1732–1741.
Molberg, Ø., McAdam, S., Lundin, K., and Sollid, L. (2000). Studies of gliadin-
specific T cells in celiac disease. In Celiac disease: Methods and protocols,
M.N. Marsh, ed. (Totowa, NJ: Humana), pp. 105–124.
Moron, B., Cebolla, A., Manyani, H., Alvarez-Maqueda, M., Megias, M.,
Thomas Mdel, C., Lopez, M.C., and Sousa, C. (2008). Sensitive detection of
cereal fractions that are toxic to celiac disease patients by using monoclonal
antibodies to a main immunogenic wheat peptide. Am. J. Clin. Nutr. 87,
405–414.
Siegel, M., Bethune, M.T., Gass, J., Ehren, J., Xia, J., Johannsen, A., Stuge,
T.B., Gray, G.M., Lee, P.P., and Khosla, C. (2006). Rational design of combina-
tion enzyme therapy for celiac sprue. Chem. Biol. 13, 649–658.
Silvester, J.A., and Rashid, M. (2007). Long-term follow-up of individuals with
celiac disease: an evaluation of current practice guidelines. Can. J. Gastroen-
terol. 21, 557–564.
Musch, M.W., Clarke, L.L., Mamah, D., Gawenis, L.R., Zhang, Z., Ellsworth,
W., Shalowitz, D., Mittal, N., Efthimiou, P., Alnadjim, Z., et al. (2002). T cell acti-
vation causes diarrhea by increasing intestinal permeability and inhibiting
epithelial Na+/K+-ATPase. J. Clin. Invest. 110, 1739–1747.
Sollid, L.M. (2002). Coeliac disease: dissecting a complex inflammatory
disorder. Nat. Rev. Immunol. 2, 647–655.
Sollid, L.M., Markussen, G., Ek, J., Gjerde, H., Vartdal, F., and Thorsby, E.
(1989). Evidence for a primary association of celiac disease to a particular
HLA-DQ alpha/beta heterodimer. J. Exp. Med. 169, 345–350.
Nilsen, E.M., Lundin, K.E., Krajci, P., Scott, H., Sollid, L.M., and Brandtzaeg, P.
(1995). Gluten specific, HLA-DQ restricted T cells from coeliac mucosa
produce cytokines with Th1 or Th0 profile dominated by interferon gamma.
Gut 37, 766–776.
Stepniak, D., Spaenij-Dekking, L., Mitea, C., Moester, M., de Ru, A., Baak-
Pablo, R., van Veelen, P., Edens, L., and Koning, F. (2006). Highly efficient
gluten degradation with a newly identified prolyl endoprotease: implications
for celiac disease. Am. J. Physiol. Gastrointest. Liver Physiol. 291, G621–G629.
Nilsen, E.M., Jahnsen, F.L., Lundin, K.E., Johansen, F.E., Fausa, O., Sollid,
L.M., Jahnsen, J., Scott, H., and Brandtzaeg, P. (1998). Gluten induces an
intestinal cytokine response strongly dominated by interferon gamma in
patients with celiac disease. Gastroenterology 115, 551–563.
Sturgess, R., Day, P., Ellis, H.J., Lundin, K.E., Gjertsen, H.A., Kontakou, M.,
and Ciclitira, P.J. (1994). Wheat peptide challenge in coeliac disease. Lancet
343, 758–761.
Paterson, B.M., Lammers, K.M., Arrieta, M.C., Fasano, A., and Meddings, J.B.
(2007). The safety, tolerance, pharmacokinetic and pharmacodynamic effects
of single doses of AT-1001 in coeliac disease subjects: a proof of concept
study. Aliment. Pharmacol. Ther. 26, 757–766.
Troncone, R., Gianfrani, C., Mazzarella, G., Greco, L., Guardiola, J., Auricchio,
S., and De Berardinis, P. (1998). Majority of gliadin-specific T-cell clones from
celiac small intestinal mucosa produce interferon-gamma and interleukin-4.
Dig. Dis. Sci. 43, 156–161.
Peyrin-Biroulet, L., Desreumaux, P., Sandborn, W.J., and Colombel, J.F.
(2008). Crohn’s disease: beyond antagonists of tumour necrosis factor. Lancet
372, 67–81.
Vora, H., McIntire, J., Kumar, P., Deshpande, M., and Khosla, C. (2007). A
scaleable manufacturing process for pro-EP-B2, a cysteine protease from
barley indicated for celiac sprue. Biotechnol. Bioeng. 98, 177–185.
Pietzak, M.M. (2005). Follow-up of patients with celiac disease: achieving
compliance with treatment. Gastroenterology 128, S135–S141.
Wapenaar, M.C., van Belzen, M.J., Fransen, J.H., Sarasqueta, A.F., Houwen,
R.H., Meijer, J.W., Mulder, C.J., and Wijmenga, C. (2004). The interferon
gamma gene in celiac disease: augmented expression correlates with tissue
damage but no evidence for genetic susceptibility. J. Autoimmun. 23, 183–190.
Piper, J.L., Gray, G.M., and Khosla, C. (2002). High selectivity of human tissue
transglutaminase for immunoactive gliadin peptides: implications for celiac
sprue. Biochemistry 41, 386–393.
Piper, J.L., Gray, G.M., and Khosla, C. (2004). Effect of prolyl endopeptidase
on digestive-resistant gliadin peptides in vivo. J. Pharmacol. Exp. Ther. 311,
213–219.
Watts, T., Berti, I., Sapone, A., Gerarduzzi, T., Not, T., Zielke, R., and Fasano,
A. (2005). Role of the intestinal tight junction modulator zonulin in the patho-
genesis of type I diabetes in BB diabetic-prone rats. Proc. Natl. Acad. Sci.
USA 102, 2916–2921.
Qiao, S.W., Bergseng, E., Molberg, O., Jung, G., Fleckenstein, B., and Sollid,
L.M. (2005). Refining the rules of gliadin T cell epitope binding to the disease-
associated DQ2 molecule in celiac disease: importance of proline spacing and
glutamine deamidation. J. Immunol. 175, 254–261.
Xia, J., Sollid, L.M., and Khosla, C. (2005). Equilibrium and kinetic analysis of
the unusual binding behavior of a highly immunogenic gluten peptide to
HLA-DQ2. Biochemistry 44, 4442–4449.
Quarsten, H., Molberg, O., Fugger, L., McAdam, S.N., and Sollid, L.M. (1999).
HLA binding and T cell recognition of a tissue transglutaminase-modified
gliadin epitope. Eur. J. Immunol. 29, 2506–2514.
Xia, J., Siegel, M., Bergseng, E., Sollid, L.M., and Khosla, C. (2006). Inhibition
of HLA-DQ2-mediated antigen presentation by analogues of a high affinity 33-
residue peptide from alpha2-gliadin. J. Am. Chem. Soc. 128, 1859–1867.
Chemistry & Biology 16, 868–881, August 28, 2009 ª2009 Elsevier Ltd All rights reserved 881