Immunity, Vol. 21, 367–377, September, 2004, Copyright 2004 by Cell Press
A Direct Role for NKG2D/MICA Interaction
in Villous Atrophy during Celiac Disease
of gluten from the diet.
CD is characterized by the infiltration of the small
intestine by a T cell “cocktail” composed mainly of acti-
vated CD4 T cells in the lamina propria and CD8 T cells
in the epithelium (for review see Shan et al., 2002; Sollid,
2002). It occurs in genetically susceptible individuals
expressing the HLA-DQ2 or -DQ8 molecules, which are
the restriction elements for CD4 T cells recognizing glia-
din peptides (Dieterich et al., 1997; Lundin et al., 1993;
Molberg et al., 1998; Shan et al., 2002). The role of CD8
TCR?? or TCR?? intraepithelial lymphocytes (IEL) re-
infiltration,without anyothersignof mucosalpathology,
may represent the first stage of CD (Maki et al., 2003).
Furthermore, disruption of intraepithelial lymphocyte
of the lymphoid malignancies characteristic of CD, as
recently demonstrated by the study of refractory celiac
sprue (RCS). This severe enteropathy complicates CD
in a small subset of patients primarily or secondarily
refractory to a strict gluten-free diet. Now considered a
low-grade intraepithelial lymphoma, it is characterized
by a massive intraepithelial infiltration by atypical CD7?
IELs with clonal TCR? rearrangement and intracellular
CD3? chain but no surface CD3 (sCD3)/TCR complexes
(Cellier et al., 1998, 2000). Although IELs from CD and
RCS are able to kill epithelial cells through the perforin/
granzyme pathway (Di Sabatino et al., 2001; Mention et
al., 2003), the effector and target cell receptors involved
in this destruction have yet to be unearthed to shed light
The human MICA and MICB proteins (Bahram et al.,
1994) are nonconventional HLA class I molecules that
serve as ligands for the activating NKG2D receptor ex-
pressed at the surface of all CD8 ??T cells, ??T cells,
of MIC molecules is restricted to intestinal and thymic
epithelium (Groh et al., 1996; Hue et al., 2003), but is
stress inducible in various epithelial cells and is upregu-
lated in tumors (Groh et al., 1999; Vetter et al., 2002)
and upon exposure to intracellular pathogens (Das et
al., 2001; Groh et al., 2001). MICs function as signals of
cellular distress and trigger a range of immune effector
mechanisms including cellular cytotoxicity, cytokine se-
2000; Vivier et al., 2002). In TCR?? CD8 T cells, NKG2D/
MIC engagement delivers a costimulatory signal that
get cells (Groh et al., 2001). In NK cells, NKG2D acts
as an activating immunoreceptor, which can transduce
positive intracellular signaling (Bauer et al., 1999; Cos-
man et al., 2001; Pende et al., 2001). Lastly, MIC can
be recognized by intraepithelial or tumor-infiltrating ??T
cells and is able to deliver both a TCR-dependent and
an NKG2D-dependent costimulatory signal for a subset
of V?1 ?? T cells (Wu et al., 2002).
NKG2D cell surface expression requires association
contains a consensus YxxM tyrosine-based motif in its
Sophie Hu ¨e,1Jean-Jacques Mention,2
Renato C. Monteiro,4ShaoLing Zhang,1,7
Christophe Cellier,5Jacques Schmitz,3
Virginie Verkarre,2Nassima Fodil,6
Seiamak Bahram,6Nadine Cerf-Bensussan,2
and Sophie Caillat-Zucman1,*
1Equipe Avenir INSERM
2INSERM EMI-0212 and
3Service de Gastroente ´rologie Pe ´diatrique
Ho ˆpital Necker-Enfants Malades
Faculte ´ Bichat
5De ´partement d’He ´patoGastroEnte ´rologie
Ho ˆpital Europe ´en Georges Pompidou
Centre de Recherche d’Immunologie et d’He ´matologie
MICA molecules interact with the NKG2D-activating
receptor on human NK and CD8 T cells. We investi-
gated the participation of the MICA/NKG2D pathway
in the destruction of intestinal epithelium by intraepi-
thelial T lymphocytes (IEL) in Celiac disease and its
that MICA is strongly expressed at epithelial cell sur-
face in patients with active disease and is induced
by gliadin or its p31-49 derived peptide upon in vitro
challenge, an effect relayed by IL-15. This triggers
direct activation and costimulation of IEL through en-
icity toward epithelial targets and enhanced TCR-
dependent CD8 T cell-mediated adaptive response.
Villous atrophy in Celiac disease might thus be as-
cribed to an IEL-mediated damage to enterocytes in-
expression of MICA on gut epithelium. This supports
a key role for MIC/NKG2D in the activation of intraepi-
thelial immunity in response to danger.
of the small intestine triggered by wheat gliadin, which
leads to villous atrophy, cryptic hyperplasia, and malab-
7Present address: Department of Endocrinology, Second Affiliated
Hospital of Sun Yat-Sen University, Guangzhou 510275, China.
To determine whether MICA was expressed at the
cell surface of villous epithelium, we next isolated gut
epithelial cells from fresh intestinal biopsy samples and
analyzed them by flow cytometry (Figure 2, left). In con-
trols and treated CD patients, the percentage of epithe-
lial cells expressing MICA at their surface was very low
(mean ? SEM, 2.5% ? 0.9% in controls and 3.6% ?
2.6% in treated CD), suggesting that most immunoper-
oxidase staining observed in normal gut biopsies de-
tects intracellular MICA. Indeed, staining of epithelial
cells isolated from normal jejunum revealed intracellular
MICA in a majority of permeabilized cells, while this
antigen was not detected by surface staining (upper
right quadrant). In contrast, in active CD and RCS pa-
tients, the percentage of epithelial cells expressing
MICA at their surface was significantly increased (mean
29.8% ? 5% in active CD and 32% ? 18.6% in RCS,
p ? 0.008 for comparison between the four groups, non-
parametric ANOVA test). This percentage even reached
87% in patient BER who presented a particularly severe
RCS with 90% of circulating mononuclear cells sharing
the same aberrant phenotype and clonality as IELs.
These results indicate that MIC, normally expressed in
the cytoplasm of gut epithelial cells, is induced at their
surface during the course of CD, and this appears to
correlate with the severity of the disease.
Figure 1. MICA Protein Expression Is Increased in Small Intestinal
Mucosa of CD Patients
Paraffin-embedded biopsy sections from a non-CD control (A), ac-
tive CD patient (B), CD patient on a gluten-free diet (C), and patient
with RCS (D). Magnification 100?.
cytoplasmic domain. In human cells, engagement of
NKG2D results in tyrosine phosphorylation of DAP10,
and recruitment and activation of the p85 subunit of
phosphatidylinositol 3-kinase (PI3K) (Billadeau et al.,
in NK cells and CD8 T cells is due, at least in part, to
the existence of two NKG2D protein species encoded
by distinct mRNA splice isoforms and to their asso-
ciation with either DAP10 or the immunoreceptor tyro-
sine-based activation motif (ITAM)-containing KARAP/
DAP12 adaptor protein (Diefenbach et al., 2002; Gilfillan
et al., 2002).
The remarkably restricted expression of MIC in the
intestinal mucosa, its induction by cellular distress, and
its recognition by NKG2D at the surface of CD8 and
?? T cells suggested a possible role of MIC/NKG2D
interaction in the IEL-mediated destruction of gut epi-
thelial cells in the course of CD.
Gliadin Induces MICA Expression in the Intestine
of Treated CD Patients
induce expression of MICA in the small intestine, fresh
biopsies were incubated for 4 to 48 hr with or without
a peptic-tryptic digest of gliadin in an organ culture
chamber prior to staining (Maiuri et al., 1998) (Figure 3
and Table 1). In biopsies from healthy individuals (n ?
5), gliadin did not modify MICA expression. In untreated
CD patients (n ? 3), the already strong expression of
MICA and the severe epithelial alterations induced by
the culture with gliadin precluded any analysis. By con-
trast, in treated patients on gluten-free diets (n ? 9), a
clear increase in MICA expression was observed on
both epithelium and lamina propria cells (Figure 3D).
Thiseffect wasobserved veryearly (within4 hrofin vitro
challenge) and was maximal after 24 hr of organ culture.
No modification of MICA expression was induced by a
peptic-tryptic digest of bovine serum albumin (BSA).
expression in intestinal epithelial cells from treated
To understand the mechanisms of MICA induction by
gliadin, we investigated the effect of two gliadin-derived
peptides known to mediate distinct effects. The 33-mer
p57-89 gliadin peptide activates the CD4-mediated
adaptive immune response in CD patients: it contains
three concatemerized immunodominant T cell epitopes
and was shown to efficiently stimulate CD4? T cell lines
derived from the intestine of CD patients (Shan et al.,
2002). The p31-49 peptide, known to induce small intes-
tinal damage in vitro and in vivo in CD patients, is not
recognized by CD4 intestinal T cells but was recently
reported to activate the innate immune system (Maiuri
et al., 2003). The hSA peptide was used as a control.
p31-49 induced a strong expression of MICA in the epi-
MIC Is Overexpressed in the Small Intestine
of Patients with Active CD
Staining of duodenal sections from four normal control
subjects who had undergone biopsies for diagnostic
purpose showed that MIC was expressed in villous epi-
thelial cells, as originally described (Groh et al., 1996).
However, this staining was only of moderate intensity,
heterogeneous within and between villi and seemed
mainly intracellular (Figure 1A). By contrast, in active CD
patients with villous atrophy (n ? 6), epithelial expres-
sion of MICA was not only much more intense but also
diffuse from the surface to the bottom of the crypts and
accompanied by staining of mononuclear cells within
the lamina propria (Figure 1B). Remarkably, in CD pa-
free diets (n ? 5), staining pattern returned to normal
(Figure 1C). However, in RCS patients under strict glu-
ten-free diets (n ? 4), MICA expression was even more
intense and diffuse than in active CD (Figure 1D).
Role of NKG2D/MICA Interaction in Celiac Disease
Figure 2. Induction of MIC at the Surface of
on Intraepithelial Lymphocytes in Active CD
IELs freshly isolated from duodenal biopsies
in control subjects (n ? 4), treated CD (n ?
The numbers indicate the percentage of
each group is shown.
PE-labeled anti-MICA SR116 mAb. In the up-
per right quadrant, intracellular staining of
Right: CD2? CD7? IELs stained with FITC-
labeled MICA tetramers. In parallel to IELs,
staining of PBMC is shown in a treated pa-
thelium of CD patients. Induction was detectable after
only 4 hr of culture but was more intense after 12 hr.
The same peptide tested in three controls failed to in-
duce MIC in two cases. In the third control, a very dis-
creet increase of MICA expression was noticeable. The
nonimmunodominant p57-89 peptide was a much less
efficient inducer of MIC. Induction was detectable only
in two out of six tested patients, and this peptide had
no effect in the controls. The hSA control peptide had
no effect on MICA expression.
To more deeply investigate the molecular mecha-
nisms involved in MICA induction by gliadin peptides,
we next studied the role of IL-15. This cytokine, overex-
pressed in the intestine of CD patients, not only medi-
ates epithelial damage but also sustains a persistent
activation of the adaptive immune system (Maiuri et al.,
2000). Our previous data suggested that IL-15 can be
upregulated upon gliadin challenge in biopsies of
treated CD patients (Di Sabatino et al., 2001; Mention
et al., 2003). Furthermore, it was shown that p31-49
stimulates IL-15 production in lamina propria dendritic
As shown in Figure 3, IL-15 induced a strong expression
of MICA in biopsies from both treated CD patients and
healthy controls. Moreover, culturing biopsies from
treated CD patients in the presence of neutralizing anti-
IL-15 antibody abrogated the p31-49-induced expres-
sion of MICA, while it had no effect on the p57-89-
of IL-15 on enterocytes, some accumulation of intracel-
lular MICA was observed after stimulation of HT29, T84,
and Caco2 epithelial cell lines cultured for 4–24 hr in
the presence of IL-15. This effect remained, however,
modest, perhaps due to the already high constitutive
expression of MIC and IL-15 in these cell lines. In con-
trast, gliadin had no effect (data not shown).
Altogether (Table 1), these results demonstrate that
gliadin can induce expression of MICA in the intestine
of CD patients. Induction mainly depends on an early
innate response triggered by the p31-49 peptide and
activated by the immunodominant p57-89 peptide in an
IL-15-independent way can, however, be detected in a
subset of patients.
IELs Can Lyse Epithelial Targets via NKG2D
Sinceepithelial cellsofthesmall intestineexpressMICA
after exposure to gliadin, they could become the targets
of NKG2D-positive IELs during active CD. We first con-
firmed by flow cytometry analysis that NKG2D was ex-
pressed on most freshly isolated IELs, as described
(Roberts et al., 2001) (Figure 2, right). The percentage
CD patients (77% ? 10%, 79% ? 4%, and 82% ? 6%,
respectively, mean 82.5%), while it reached 98% ? 1%
in RCS patients (p ? 0.01 compared to the three other
to IELs, as defined by mean fluorescence intensity, was
comparable in controls and CD patients whatever their
status and the percentage of MICA-positive epithelial
We then determined whether IELs from active CD or
RCS could mediate killing of target cells through
NKG2D/MICA interaction. Because the small number of
IELs directly recovered from intestinal biopsies pre-
vented extensive further experiments, we derived T cell
CD and four RCS) with IL-15 as previously described
(Di Sabatino et al., 2001; Mention et al., 2003). These
cell lines were then tested in a redirected lysis assay
against Fc?R-bearing P815 target cells (Bauer et al.,
1999). Cell lines from active CD patients expressed a
classical TCR??? CD3? CD8? NKG2D? phenotype at
the time of testing. Crosslinking NKG2D alone on these
IEL lines had no or only minor effect even at a 30:1
effector:target cell ratio, although TCR?? engagement
by 1 ng/ml of anti-CD3 induced strong target cell lysis.
However, crosslinking NKG2D significantly augmented
CD3-mediated redirected lysis when using suboptimal
ciency of lysis varied from line to line (Figure 4A). These
results indicate that in T cell lines from active CD pa-
tients, NKG2D functions as a costimulatory receptor in
a conventional way, similar to what has been described
in CD8 T cells.
Surprisingly, however, when redirected cytotoxicity
assays were repeated with the same CD8 T cell lines
from two different patients at various kinetic points of
the cell culture, some differences appeared. Indeed, at
very narrow time points, NKG2D-mediated lysis was
sometimes observed at day 15 of the cell culture, while
it was not detected the day before (Figure 4B). This
phenomenon was transient, as CD8 T cells rapidly re-
covered their classical cytotoxicity profile 4 days later.
It thus appears that in particular circumstances, NKG2D
alone can mediate IEL cytotoxicity.
which all expressed the aberrant sCD3?CD7?NKG2D?
phenotype. In all cases, NKG2D was able to mediate
alone the killing of P815 target cells, even at a low ef-
fector:cell ratio (Figure 4C). Although the percentage of
target cell lysis varied depending on the patient (21%–
93% at 30:1 E:T ratio, mean 51.3%), this NKG2D-medi-
ated killing was constant and maintained over time.
Thus, in sCD3?CD7?IELs, NKG2D has a direct activat-
similar to what is observed in NK lines (Bauer et al.,
To confirm that IELs from RCS can indeed mediate
killing of enterocytes via NKG2D, we performed a direct
cell-mediated cytotoxicity assay on epithelial target
cells. As previously shown, all cell lines from RCS pa-
tients were able to kill MICA-positive HT29 or HeLa epi-
thelial cell lines even at a low effector:cell ratio (Di Saba-
tino et al., 2001; Mention et al., 2003). In the case of
patient BER (Figure 4D), anti-NKG2D mAb fully inhibited
the lysis by more than 60%. Since in this patient, 87%
of intestinal epithelial cells expressed surface MICA, the
NKG2D/MICA pathway seems to be directly involved in
the massive destruction of the gut epithelium, leading
to villous atrophy. Interestingly, one subclone of this
patient’s cell line lost surface NKG2D expression. Both
icity of epithelial target cells were completely abolished
(data not shown), confirming that the cytolytic effect of
In the other RCS patients, however, inhibition by anti-
NKG2D did not exceed 35%, suggesting that killing of
Figure 3. MICA Expression Is Induced by Gliadin or Gliadin-Derived
Peptides in the Small Intestine Mucosa of CD Patients
Biopsy specimens from healthy controls (A, C, E, G, and I) and CD
for 12 hr in medium alone (A and B) or in presence of peptic-tryptic
digest of gliadin (C and D) or p31-49 gliadin peptide (E and F), p57-
89 gliadin peptide (G and H), IL-15 (I and J), p31-49 plus anti-IL-15
neutralizing mAb (K), p57-89 plus anti-IL-15 mAb (L), peptic-tryptic
digest of BSA (M), and hSA control peptide (N) prior to staining with
SR99 anti-MIC mAb. In (A) and (N), weak nonspecific staining of
red blood cells is observed in some lamina propria blood vessels.
Role of NKG2D/MICA Interaction in Celiac Disease
Table 1. Gliadin-Induced MICA Expression in the Intestine of Healthy Controls and Treated CD
Digest of Gliadin
anti-IL-15 mAbp57-89 p31-49IL-15
Treated CD patients
Results are given as the number of subjects with increased expression of MICA among five healthy controls and nine treated CD patients
after 12 hr of in vitro organ culture in the presence of gliadin or gliadin-derived peptides.
*Weak MICA induction.
n.d., not done.
epithelial cells involves the engagement of additional
triggering receptor(s). Indeed, in agreement with this
from the five active CD patients were also able to medi-
ate direct epithelial cell lysis (100:1 E:T ratio) (Figure 4E).
This cytotoxic activity was only partly inhibited by anti-
NKG2D mAb (range 4.7%–31.5% of inhibition), arguing
in favor of the participation of other activating receptors
in this antigen-independent killing.
Human NK and CD8 T cells use the DAP10-PI3K path-
way for NKG2D activation. To determine whether the
same pathway is utilized by sCD3?IELs from RCS, we
measured the lytic activity of the cell line from patient
BER, which is largely dependent on NKG2D in the pres-
ence of 1 ?M of the PI3K inhibitor LY 294002. In both
periments toward epithelial cells, lysis was abrogated
by more than 80%, indicating that the PI3K pathway
has a key role in the direct cytotoxicity mediated via
NKG2D (Figures 4B and 4D).
Altogether, these results are evocative of a gradient
in the potential of NKG2D to mediate epithelial cell de-
struction in the different stages of Celiac disease. In
active CD, NKG2D mainly plays a costimulatory role on
conventional IELs, and a TCR-mediated signal is re-
quired for complete activation. In the more severe form
in abnormal sCD3?IELs, endowing these cells with
strong epithelium killing capacity.
High Levels of Soluble MIC Molecules Are Present
in the Serum of Active CD Patients
The release of a soluble form of MICA (sMICA) in the
serum of some cancer patients was recently reported
(Groh et al., 2002; Salih et al., 2002). We tested the
presence of sMICA in the serum of CD patients with a
highly sensitive sandwich ELISA (Figure 5A). Sera from
24 healthy individuals or control patients with non-CD
enteropathy were all negative. Nineteen of 54 CD sera
(35.2%) contained sMICA (range 0.45–370 ng/ml, mean
57 ng/ml). Among these patients, sMICA was detected
diets, but in only three of 22 (13.6%, p ? 0.008) patients
at the time of diagnosis and became negative after 6
RCS patients on strict gluten-free diets, three had
sMICA. Therefore, the presence of sMICA is correlative
Figure 4. NKG2D-Mediated Cytotoxicity of
IELs in Active CD and RCS
(A–C) Redirected cytotoxicy of Fc?R? P815
cells by IEL cell lines. Results represent tripli-
cate of lysis for each cell line in a given experi-
ment. (A) IEL cell lines from five different active
CD patients are tested at a 30:1 effector:tar-
(0.01 ng/ml) and/oranti-NKG2DmAb (1?g/ml),
mAb, or isotype control IgG1. (B) IEL cell line
from one representative patient with active
CD, tested at different kinetic time points of
transient capacity in the NKG2D-mediated
ing of P815 cells by IEL cell lines from four
patients with RCS in the presence of control
IgG1 or anti-NKG2D mAb (effector:target cell
inhibitor LY 294002 (patient BER, n ? 4).
(D and E) Direct cytotoxicity of MICA-positive
epithelial cells by IEL cell lines. (D) Cytotoxic-
ity of HT29 epithelial cells by IEL cell line from
RCS patient BER (10:1 E:T ratio) in presence
or absence of isotype control IgG1, anti-
MICA, or anti-NKG2D mAb or LY294002. (E)
IEL cell lines from five active CD patients
at high effector:cell ratio (100:1). Lysis is only
weakly abrogated by anti-NKG2D mAb.
Table 2. DistributionofTransmembraneMICAAllelesinCDPatients
n ? 107 (%)
n ? 90 (%)MICA
*The number of alleles compared multiplied by p (Pc) not significant.
IELs from active CD, and this is not related to an oppos-
ing activity of soluble IL-15 or TNF?. These results sug-
gest an intrinsic capacity of IELs from CD patients to
maintain high levels of NKG2D at their surface.
The MICA5.1 allele is characterized by a frameshift
mutation causing a premature termination codon within
the transmembrane region. In order to rule out the pres-
reports on the role of MICA alleles in susceptibility to
CD (Bilbao et al., 2002; Fernandez et al., 2002; Lopez-
Vazquez et al., 2002), a large series of patients and
controls was genotyped for MICA. We did not find any
significant difference in the allelic distribution of MICA
transmembrane polymorphism between 107 CD pa-
tients and 90 healthy controls (Table 2) or in the fre-
11/90, 12.2%). MICA5.1 frequency was not significantly
different among patients with serum sMICA compared
to others (71.4% and 60.2%, respectively). Therefore,
the presence of sMICA in the serum is not related to
the presence of a particular MICA genotype.
The MICA5.1 variant molecule is aberrantly trans-
epithelial cells, in contrast to other alleles located at the
basolateral surface (Suemizu et al., 2002). This might
prevent interaction of MICA5.1-expressing epithelial
cells with NKG2D-positive IELs at the basolateral inter-
face. By immunohistochemistry analysis, we did not ob-
samples from active CD or RCS patients expressing
MICA5.1 at the homozygous or heterozygous state, or
that MICA5.1 homozygosity is unlikely to protect human
epithelial cells from their destruction by NKG2D? IELs.
Figure 5. SolubleMICAintheSerumofCDPatientsDoesNotDown-
modulate NKG2D Levels on IELs
(A) Soluble MICA is present in the serum of CD patients. Serum
samples from 24 healthy individuals, 22 CD patients on gluten-free
diets, and 32 untreated CD patients were investigated by ELISA
with recombinant sMICA*008 as a standard. The data shown are
means of triplicates. Horizontal lines indicate mean value of respec-
(B) Failure of soluble recombinant MICA to downmodulate NKG2D
on IEL lines from active CD or RCS patients. Two-color staining of
IEL cells incubated in the absence or presence of soluble recombi-
nant MICA (10 ?g/ml) with or without IL-15 (20 ng/ml). Data are
representative of staining on three different IEL lines.
with the presence of gluten in the diet but is also de-
tected in a fraction of RCS patients on gluten-free diets.
The binding of sMICA induces downmodulation of
(Groh et al., 2002), which may serve to limit T cell activa-
tion. This raised the question of why NKG2D levels were
maintained at high levels on IELs and circulating CD8
T cells of active CD and RCS patients, even in the pres-
ence of a high concentration of sMICA. Since IL-15 and
TNF? are able to induce NKG2D expression (Groh et al.,
2003; Roberts et al., 2001), we measured soluble IL-
15 and TNF? in sMICA-containing sera. No significant
amount of these cytokines was detected (below 5 pg/ml
in all tested sera). Furthermore, the NKG2D level at the
surface of IEL cell lines from CD patients was not down-
regulated after incubation with recombinant soluble
MICA or whatever exogenous IL-15 was added or not
with what is observed on CD4?CD28?T cells from
patients with rheumatoid arthritis (Groh et al., 2003),
sMICA does not induce NKG2D downmodulation on
enterocyte apoptosis indicative of active epithelial de-
struction (Ciccocioppo et al., 2001; Monteleone et al.,
age remain unclear. Several lines of evidence point to
a contribution of IELs activated via both innate and
HLA-A2-restricted gliadin-specific CD8 T cells have
been identified in the mucosa of CD patients (Gianfrani
cytotoxicity of IL-15-stimulated IELs against epithelial
targets points to the role of innate receptors enabling
Role of NKG2D/MICA Interaction in Celiac Disease
damage of epithelial cells (Ebert, 1998; Mention et al.,
2003; Roberts et al., 2001). Thus, a selective increase
of IELs expressing CD94, the HLA-E-specific NK recep-
tor, is observed in active CD (Jabri et al., 2000).
Among other immunoreceptors expressed at the sur-
face of IELs, NKG2D is a likely candidate. It behaves as
a sentinel used by CD8 T cells and NK lymphocytes to
detect cells that have upregulated ligands such as MIC
In the present study, we show that only a very small
percentage of normal enterocytes express MICA at their
surface, MIC expression being mostly intracellular. As
NKG2D is constitutively expressed on resident IELs
(Roberts et al., 2001), this low amount of surface MICA
may prevent inappropriate epithelial attack by IELs in
healthy individuals. We then show that MIC expression
is strongly increased in active CD and becomes detect-
able at the surface of duodenal epithelial cells, which
may therefore become targets for NKG2D-expressing
IELs. NKG2D functions as a costimulatory signal in CD8
T cells, amplifying antigen-specific signals provided by
the TCR. SinceCD8 T IELs come intodirect contact with
epithelial cells, one may thus expect an enhancement
of TCR-mediated destruction of epithelial target cells
through NKG2D/MIC interaction in CD patients. Indeed,
we show that engagement of NKG2D in T IEL lines from
active CD patients strengthens TCR-mediated, redi-
rected killing of target cells and allows efficient cytotox-
icity even at limited antigen concentration. Turning our
attention to RCS, a cryptic intestinal T cell lymphoma
characterized by the presence of an aberrant sCD3?
clonal intraepithelial T cell population, we then demon-
strate that MIC/NKG2D interaction can directly induce
killing of epithelial target cells by IELs, independently
of any signaling via the TCR. Therefore, as the disease
becomes resistant to the gluten-free diet, the contribu-
tion of NKG2D in mediating gut epithelium destruction
by IELs shifts from a conventional costimulatory func-
signal in RCS. However, the participation of other trig-
gering receptors on IELs remains to be unraveled. First,
NKG2D blockade of spontaneous cytotoxicity of sCD3?
RCS cell lines against epithelial cells was only partial in
three-fourths of the tested patients. Second, we ob-
served that in extreme conditions, such as elevated lev-
els of IL-15 and high density of effector cells, IELs from
active CD patients could mediate direct epithelial cell
lysis in an antigen-independent manner, a phenomenon
previously described as an IL-15-induced bystander ac-
tivation of memory phenotype CD8 T cells (Liu et al.,
2002). This cytotoxic activity was only partly inhibited
by anti-NKG2D mAb. Engagement of NKG2D, alone or
in combination with other activating receptors on IELs,
toxicity against epithelial cells, depending on the com-
recently described for tumor-specific CD8 T cell clones
(Maccalli et al., 2003).
MIC molecules, like a wide variety of transmembrane
proteins, can be shed as soluble forms in the serum
of patients (Groh et al., 2002; Salih et al., 2002) after
proteolytic cleavage of the membrane-anchored mole-
cules by an as yet unidentified metalloproteinase. Inter-
estingly, levels of metalloproteinases MMP-1 and MMP-3
are increased in the mucosa of patients with active CD
as a consequence of lamina propria mononuclear cell
activation (Daum et al., 1999). Thus, an increased pro-
duction of metalloproteinases in untreated CD patients
might favor the release of sMIC, as we observed in a
large percentage of our patients. However, the levels of
sMICA in the serum were not correlated to the severity
of histological lesions or clinical symptoms, indicating
that sMICA cannot be used so far as a marker of activity
or of tolerance to gluten-containing diet. The functional
consequence of the shedding of MICA remains unclear.
sMICA may play a role in immune escape of human
tumors in vivo by inducing NKG2D downregulation,
thereby decreasing the functional capacity of antitumor
effector T cells or NK cells (Groh et al., 2002). We show
that NKG2D levels are comparable on IELs from CD or
RCS patients with high levels of sMICA and on IELs
from controls or treated patients without sMICA. Since
NKG2D levels on IELs are upregulated by IL-15 (Roberts
et al., 2001), which is overexpressed during CD, it is
on IELs even in presence of sMIC. Although we did
not detect soluble IL-15 in sMICA-containing sera of
patients, it is known that IL-15 is massively exposed at
the surface of enterocytes during CD (Mention et al.,
2003). Furthermore, surface bound IL-15 can be pre-
sented in trans to neighboring cells (Dubois et al., 2002),
and is likely to prevent ligand-induced internalization of
NKG2D. However, the inability of recombinant MICA
to downmodulate NKG2D on IELs from CD patients
suggests that other mechanisms tending to maintain
NKG2D levels on these cells are likely to be involved.
There is considerable genetic variation in the MICA
gene, with several polymorphisms reported both in the
transmembrane and extracellular domains (Fodil et al.,
1999). Allelic variants of MICA substantially differ in their
binding affinity for NKG2D (Steinle et al., 2001), which
could have significant effects in the modulation of T cell
responses. In addition, the MICA gene exhibits a triplet
repeat microsatellite polymorphism in the transmem-
brane region, which might be associated with variable
secretion of soluble MICA or with aberrant MICA local-
ization at the surface of epithelial cells. We show, how-
ever, that sMICA is detected in patients’ sera regardless
of allelic MICA differences and that MICA is similarly
distributed in gut epithelium of CD patients whatever
MICA alleles expressed.
The mechanisms leading to upregulation of MIC and
its translocation to the surface of epithelial cells in CD
remain to be elucidated. As indicated by intracellular
staining of permeabilized cells, a rapid appearance of
MICA at the cell surface may be due, at least in part,
to redistribution of intracellular protein. In addition, we
show that gliadin rapidly induces MICA expression in
treated CD patients. MICA induction appears as a con-
din in the intestine of sensitive individuals rather than
as a direct effect of gliadin per se, and one important
relay in the induction of MICA is IL-15. Thus, the MICA-
inducing effect of gliadin was reproduced by the p31-
49 gliadin-derived peptide. This nonimmunodominant
peptide was recently shown to induce IL-15 in innate
immune cells in the lamina propria of CD patients and
mers were generated by mixing MICA or MICB monomers with fluo-
rochrome-labeled streptavidin in a 4:1 ratio (streptavidin-PE or
streptavidin-FITC, Pharmingen, San Jose, CA).
to induce enterocyte apoptosis, the latter effect being
inhibited by neutralizing IL-15 (Maiuri et al., 2003). Con-
sistent with this observation, we observed that MICA
induction by p31-49 was inhibited by neutralizing IL-15
during organ culture. Furthermore, IL-15 induced accu-
mulation of intracellular MICA in epithelial cell lines. To-
gether with the recent finding that IL-15 induces MICA
in human dendritic cells (Jinushi et al., 2003), these data
support the hypothesis of a direct inducing effect on IL-
the intense expression of MICA on the epithelium of
RCS patients who have no more gluten in their diet
but retain massively increased levels of IL-15 in their
peptide, p57-89, which is a very efficient stimulus of the
CD4-mediated adaptive immune response (Shan et al.,
2002), had a much more modest effect on MICA induc-
tion at the 12 hr time point chosen, and this effect was
not sensitive to IL-15 blockade. A delayed effect of this
peptide via other proinflammatory relay is thus not ex-
cluded. Therefore, the p31-49 peptide could represent
a danger signal in CD patients, through enhanced ex-
pression of stress proteins, such as MIC. A noticeable
increased expression of hsp65 has been observed in
and the heat-shock protein hsp110, produced by the
intestinal epithelium, induces the expression of another
nonclassical MHC class I molecule, CD1d, on intestinal
epithelial cells (Colgan et al., 2003).
In conclusion, we demonstrate that in CD, gliadin, via
a pathway involving IL-15, induces MICA expression at
the surface of gut epithelial cells, thereby providing an
epithelial target to IELs whose cytotoxic properties are
simultaneously turned on by enterocyte-derived IL-15.
MICA induction will trigger both direct activation and
costimulation, permitting an early innate-like response
and a gliadin-specific CD8 T cell-mediated adaptive re-
MIC and IL-15 might be the paradigm of a normal de-
fense mechanism allowing the immediate recruitment
and activation of IELs in response to intestinal aggres-
sions by pathogens. In CD, the control of this primordial
effects of gliadin.
Production of Anti-MICA Monoclonal Antibodies
Anti-MICA monoclonal antibodies were obtained by immunizing
mice with purified recombinant MICA*008. SR99 (IgG1), SR104
(IgG1), and SR116 (IgG2a) mAbs were selected on the basis of both
reactivity against recombinant MICA by ELISA and staining of HT-
29 and HeLa cells by flow cytometry (Hue et al., 2003).
normal human serum for 20 min at room temperature and incubated
with the monomorphic anti-MICA SR99 mAb for 1 hr or isotype-
In Vitro Organ Cultures
For in vitro challenge, 8–16 fresh biopsy specimens were obtained
on iron grids, and cultured for 4, 12, 24, or 48 hr in an organ culture
chamber at 37?C as described (Maiuri et al., 1998) in the presence
of medium alone, with peptic-tryptic digest of gliadin or bovine
serum albumin (500 ?g/ml), or with peptides (200 ?g/ml) before
being fixed in 10% formalin. IL-15 (50 ng/ml) was also tested in four
patients and three controls. In selected samples incubated in the
presence of peptides, we additionally analyzed the effect of anti-
IL-15 neutralizing monoclonal antibody (10 ?g/ml, R&D Systems) or
The gliadin peptides p57-89 (LQLQPFPQPQLPYPQPQLPYPQP
QLPYPQPQPF) (Shan et al., 2002) and p31-49 (LGQQQPFPPQQPY
PQPQPF) (Maiuri et al., 2003; Matysiak-Budnik et al., 2003) and the
control human serum albumin hSA peptide (p64-76, VKLVNEVTEF
AKT) were synthesized by Epytop (Nı ˆmes, France) and were
All protein digests and peptides were tested for the absence of
detectable levels of LPS by the Limulus Amebocyte Lysate assay
(Charles River Laboratories, Charleston, SC). All reagents were
which contained approximately 5 U/ml.
Isolation of Cells and Flow Cytometry Analysis
PBMC from healthy controls were purified by ficoll density centrifu-
gation. For isolation of IEL and epithelial cells, six to eight fresh
biopsy specimens were pooledand incubated under constant shak-
ing in RPMI 1640 (Gibco BRL, Rockville, MD) containing 1% FCS,
1.5 mM MgCl2, and 1 mM ethylen glycol-bis (b-aminoethyl-ether)-
N,N,N?,N?-tetraacetic acid for 30 min at 37?C. The supernatant con-
taining the IEL and contaminating epithelial cells was passed
through a nylon filter, and cells were washed twice in PBS supple-
mented with 5% human AB serum. Intestinal epithelial cells were
identified by the expression of epithelial surface antigen (FITC-ESA,
Biomeda, Foster City, CA).
For cell-surface staining of MICA, epithelial cells were incubated
with 10 ?g/ml SR104 mAb or isotype-control antibody at 4?C for
30 min, washed, and then stained with phycoerythrin-labeled goat
antimouse IgG (Caltag, Burlingame, CA). Free FcR were blocked
with a 50 ?g/ml dilution of mouse Ig before adding FITC-ESA mAb.
In selected experiments, cells were permeabilized by 2% saponin
For cell-surface staining of NKG2D, 1–2 ? 106PBMC or 105IEL
were incubated at 22?C for 90 min with MICA tetramers (at 10 ?g/ml
of MICA protein), then for 20 min at 4?C with antibodies to CD3,
CD8 (BD Pharmingen), ??TCR, ??TCR, or CD103 (Beckman Coulter,
Miami, FL). Cells were fixed in PBS with 1% paraformaldehyde and
immediately analyzed on a Becton Dickinson FACSCalibur flow cy-
tometer (Becton Dickinson, Palo Alto, CA). Data were analyzed with
Cell Quest software (Becton Dickinson).
CD was diagnosed in childhood according to EPSGAN criteria
(McNeish et al., 1979) and in adulthood on the basis of villous atro-
phy,antiendomysium antibodies,andHLA-DQ2or -DQ8phenotype.
Patients with RCS failed to improve clinically and histologically de-
spite strict gluten-free diets for at least 6 months (Cellier et al.,
2000). The study was performed in accordance with the Declaration
of Helsinki and received approval of the local Ethical Committee.
adenocarcinoma), and P815 (mouse mastocytoma) cell lines (ATCC,
Rockville, MD) were grown in RPMI 1640 supplemented with 10%
fetal calf serum (FCS), glutamine, and antibiotics.
Generation of Soluble MICA and MICB Molecules
and Fluorescent Tetramers
Recombinant MICA and MICB were produced as secreted proteins
in Sf9 insect cells as previously described (Hue et al., 2003). Tetra-
Role of NKG2D/MICA Interaction in Celiac Disease
To derive cell lines from IELs, intestinal biopsy specimens from five
patients with active CD and four patients with RCS were cultured
in 24-well plates (1 biopsy/well) in lymphocyte culture medium con-
taining IL-15 (20 ng/mL, R&D Systems) as described (Mention et al.,
2003). Wells were fed every 3 days with IL-15-containing medium.
On confluence, cells were transferred into 6-well plates and diluted
at 106mL every 3 days in the same medium and their phenotype
monitored by flow cytometry before assessing cytotoxicity.
For redirected lysis assay, cells were washed and incubated for
4 hr with51Cr-labeled Fc?R? P815 target cells at various E:T ratios
in the presence of the M585 anti-NKG2D mAb (1 ?g/ml, kindly pro-
vided by Immunex/Amgen) or mouse IgG1 control and/or anti-CD3
(OKT3) mAb (0.01–100 ng/ml).
For direct cell-mediated cytotoxicity assay, cells were incubated
for 4 hr with51Cr-labeled HeLa or HT29 cells at various E:T ratios
in the presence or absence of recombinant soluble MICA protein
(10 ?g/ml), anti-MICA mAb (10 ?g/ml), anti-NKG2D M585 mAb
(1 ?g/ml), or isotype control mAb.
PI3K inhibition experiments were done by pretreating effector
cells with the PI3K inhibitor LY 294002 (1 ?M, Sigma). The mean of
duplicates of lysis for each cell line was expressed as a percentage
of specific release.
Bahram, S., Bresnahan, M., Geraghty, D.E., and Spies, T. (1994). A
second lineage of mammalian major histocompatibility complex
class I genes. Proc. Natl. Acad. Sci. USA 91, 6259–6263.
Bauer, S., Groh, V., Wu, J., Steinle, A., Phillips, J.H., Lanier, L.L.,
and Spies, T. (1999). Activation of NK cells and T cells by NKG2D,
a receptor for stress-inducible MICA. Science 285, 727–729.
Bilbao, J., Martin-Pagola, A., Vitoria, J., Zubillaga, P., Ortiz, L., and
Castano, L. (2002). HLA-DRB1 and MHC class 1 chain-related A
haplotypes in Basque families with Celiac disease. Tissue Antigens
Billadeau, D.D., Upshaw, J.L., Schoon,R.A., Dick, C.J., and Leibson,
P.J. (2003). NKG2D–DAP10 triggers human NK cell-mediated kill-
ing via a Syk-independent regulatory pathway. Nat. Immunol. 4,
Cellier, C., Patey, N., Mauvieux, L., Jabri, B., Delabesse, E., Cervoni,
J.P., Burtin, M.L., Guy-Grand, D., Bouhnik, Y., Modigliani, R., et al.
(1998). Abnormal intestinal intraepithelial lymphocytes in refractory
sprue. Gastroenterology 114, 471–481.
Cellier, C., Delabesse, E., Helmer, C., Patey, N., Matuchansky, C.,
Jabri, B., Macintyre, E., Cerf-Bensussan, N., and Brousse, N. (2000).
Refractory sprue, coeliac disease, and enteropathy-associated
T-cell lymphoma. French Coeliac Disease Study Group. Lancet
Ciccocioppo, R., Di Sabatino, A., Parroni, R., Muzi, P., D’Alo, S.,
Ventura, T., Pistoia, M.A., Cifone, M.G., and Corazza, G.R. (2001).
disease. Am. J. Clin. Pathol. 115, 494–503.
Colgan, S.P., Pitman, R.S., Nagaishi, T., Mizoguchi, A., Mizoguchi,
E., Mayer, L.F., Shao, L., Sartor, R.B., Subjeck, J.R., and Blumberg,
R.S. (2003). Intestinal heat shock protein 110 regulates expression
of CD1d on intestinal epithelial cells. J. Clin. Invest. 112, 745–754.
Cosman, D., Mullberg, J., Sutherland, C.L., Chin, W., Armitage, R.,
Fanslow, W., Kubin, M., and Chalupny, N.J. (2001). ULBPs, novel
MHC class I-related molecules, bind to CMV glycoprotein UL16 and
stimulate NK cytotoxicity through the NKG2D receptor. Immunity
Das, H., Groh, V., Kuijl, C., Sugita, M., Morita, C.T., Spies, T., and
Bukowski, J.F. (2001). MICA engagement by human Vgamma2-
Vdelta2 T cells enhances their antigen-dependent effector function.
Immunity 15, 83–93.
Daum, S., Bauer, U., Foss, H.-D., Schuppan, D., Stein, H., Riecken,
E.-O., and Ullrich, R. (1999). Increased expression of mRNA for ma-
trix metalloproteinases-1 and -3 and tissue inhibitor of metallopro-
teinases-1 in intestinal biopsy specimens from patients with coeliac
disease. Gut 44, 17–25.
Di Sabatino, A., Ciccocioppo, R., D’Alo, S., Parroni, R., Millimaggi,
D., Cifone, M.G., and Corazza, G.R. (2001). Intraepithelial and lamina
propria lymphocytes show distinct patterns of apoptosis whereas
both populations are active in Fas based cytotoxicity in coeliac
disease. Gut 49, 380–386.
Diefenbach, A., and Raulet, D.H. (2003). Innate immune recognition
by stimulatory immunoreceptors. Curr. Opin. Immunol. 15, 37–44.
Diefenbach, A., Tomasello, E., Lucas, M., Jamieson, A.M., Hsia,
J.K., Vivier, E., and Raulet, D.H. (2002). Selective associations with
signaling proteinsdetermine stimulatory versuscostimulatory activ-
ity of NKG2D. Nat. Immunol. 3, 1142–1149.
Dieterich, W., Ehnis, T., Bauer, M., Donner, P., Volta, U., Riecken,
E.O., and Schuppan, D. (1997). Identification of tissue transglutami-
nase as the autoantigen of Celiac disease. Nat. Med. 3, 797–801.
Bach, J.F., and Caillat-Zucman, S. (1998). CTLA-4 gene polymor-
phism is associated with predisposition to coeliac disease. Gut
Dubois, S., Mariner, J., Waldmann, T., and Tagaya, Y. (2002). IL-
15R? recycles and presents IL-15 in trans to neighboring cells.
Immunity 17, 537–547.
Ebert, E.C. (1998). Interleukin 15 is a potent stimulant of intraepithel-
ial lymphocytes. Gastroenterology 115, 1439–1445.
Fernandez, L., Fernandez-Arquero, M., Gual, L., Lazaro, F., Ma-
ELISA of Soluble MICA and Modulation
of NKG2D on IEL Lines
Soluble MICA was measured in the serum with a sandwich ELISA
as described (Hue et al., 2003). Recombinant soluble MICA*001 was
consistently detected at concentration of 0.3 ng/ml. The amount of
IL-15 in patients’ sera was determined by commercial ELISA with
matched antibody pair in relation to standard pair (R&D Systems).
Modulation of NKG2D levels on IEL cell lines from active CD or
RCS patients was examined by flow cytometry after 24 hr of incuba-
tion with soluble recombinant MICA (10 ?g/ml) in the absence or
presence of IL-15 (20 ng/ml).
A total of 107 CD patients and 90 healthy controls, previously de-
scribed (Djilali-Saiah et al., 1998), was MICA genotyped. The tri-
nucleotide repeat microsatellite polymorphism in the transmem-
brane region of the MICA gene was amplified with PCR primers
flanking the TM region as described (Mizuki et al., 1997; Sugimura et
al., 2001). The PCR products were electrophoresed in an automated
ber of microsatellite repeat was estimated automatically using the
Genescan 672 software. Allelic frequencies were calculated by di-
rect counting, and the significance of the association was deter-
mined by using the chi-square test. The level of significance was
set at 0.05, and the correction of Bonferonni for multiple tests was
applied by multiplying p by the number of alleles compared (Pc).
We thank Eric Vivier for helpful discussion and critical reading of
the manuscript. We are grateful to the GERMC (French Group of
Research on Celiac disease) for providing samples and clinical data
for some of the patients included in this study. We thank Ullah
Barbe, Bernadette Be `gue, and Ce ´cile Macquin for excellent techni-
cal assistance. S.H. was supported by poste d’accueil INSERM and
J.-J.M. by a fellowship from the Association pour la Recherche sur
le Cancer. This work was supported by INSERM (Re ´seau Progre `s
and program AVENIR), Association pour la Recherche sur le Cancer
(ARC4616), and Fondation Princesse Grace de Monaco.
Received: April 14, 2004
Revised: June 2, 2004
Accepted: June 7, 2004
Published: September 14, 2004
Bahram, S. (2000). MIC genes: from genetics to biology. Adv. Immu-
nol. 76, 1–60.
luenda, C., Polanco, I., Figueredo, M.A., and De La Concha, E.G.
(2002). Triplet repeat polymorphism in the transmembrane region
of the MICA gene in Celiac disease.
Fodil, N., Pellet, P., Laloux, L., Hauptmann, G., Theodorou, I., and
Bahram, S. (1999). MICA haplotypic diversity. Immunogenetics 49,
Gianfrani, C., Troncone, R., Mugione, P., Cosentini, E., De Pascale,
M., Faruolo, C., Senger, S., Terrazzano, G., Southwood, S., Au-
ricchio, S., and Sette, A. (2003). Celiac disease association with
A2-restricted epitope. J. Immunol. 170, 2719–2726.
Gilfillan, S., Ho, E.L., Cella, M., Yokoyama, W.M., and Colonna, M.
(2002). NKG2D recruits two distinct adapters to trigger NK cell acti-
vation and costimulation. Nat. Immunol. 3, 1150–1155.
Groh, V., Bahram, S., Bauer, S., Herman, A., Beauchamp, M., and
Spies, T. (1996). Cellstress-regulated human major histocompatibil-
ity complex class I gene expressed in gastrointestinal epithelium.
Proc. Natl. Acad. Sci. USA 93, 12445–12450.
Groh, V., Rhinehart, R., Secrist, H., Bauer, S., Grabstein, K.H., and
Spies, T. (1999). Broad tumor-associated expression and recogni-
tion by tumor-derived gamma delta T cells of MICA and MICB. Proc.
Natl. Acad. Sci. USA 96, 6879–6884.
Groh, V., Rhinehart, R., Randolph-Habecker, J., Topp, M.S., Riddell,
S.R., and Spies, T. (2001). Costimulation of CD8alphabeta T cells
by NKG2D via engagement by MIC induced on virus-infected cells.
Nat. Immunol. 2, 255–260.
Groh, V., Wu, J., Yee, C., and Spies, T. (2002). Tumour-derived solu-
ble MIC ligands impair expression of NKG2D and T- cell activation.
Nature 419, 734–738.
Groh, V., Bruhl, A., El-Gabalawy, H., Nelson, J.L., and Spies, T.
(2003). Stimulation of T cell autoreactivity by anomalous expression
of NKG2D and its MIC ligands in rheumatoid arthritis. Proc. Natl.
Acad. Sci. USA 100, 9452–9457.
Hue, S., Monteiro, R.C., Berrih-Aknin, S., and Caillat-Zucman, S.
(2003). Potential role of NKG2D/MHC class I-related chain A interac-
nol. 171, 1909–1917.
Iltanen, S., Rantala, I., Laippala, P., Holm, K., Partanen, J., and Maki,
M. (1999). Expression of HSP-65 in jejunal epithelial cells in patients
clinically suspected of Coeliac disease. Autoimmunity 31, 125–132.
Jabri, B., de Serre, N.P., Cellier, C., Evans, K., Gache, C., Carvalho,
C., Mougenot, J.F., Allez, M., Jian, R., Desreumaux, P., et al. (2000).
Selective expansion of intraepithelial lymphocytes expressing the
HLA-E-specific natural killer receptor CD94 in Celiac disease. Gas-
troenterology 118, 867–879.
Jinushi, M., Takehara, T., Tatsumi, T., Kanto, T., Groh, V., Spies, T.,
Suzuki, T., Miyagi, T., and Hayashi, N. (2003). Autocrine/paracrine
IL-15 that is required for type I IFN-mediated dendritic cell expres-
sion of MHC class I-related chain A and B is impaired in hepatitis
C virus infection. J. Immunol. 171, 5423–5429.
Liu, K., Catalfamo, M., Li, Y., Henkart, P., and Weng, N.-P. (2002).
IL-15 mimics T cell receptor crosslinking in the induction of cellular
proliferation, gene expression, and cytotoxicity in CD8? memory
T cells. Proc. Natl. Acad. Sci. USA 99, 6192–6197.
Lopez-Vazquez, A., Rodrigo, L., Fuentes, D., Riestra, S., Bousono,
C., Garcia-Fernandez, S., Martinez-Borra, J., Gonzalez, S., and Lo-
pez-Larrea, C. (2002). MHC class I chain related gene A (MICA)
modulates the development of coeliac disease in patients with the
high risk heterodimer DQA1*0501/DQB1*0201. Gut 50, 336–340.
Lundin, K.E., Scott, H., Hansen, T., Paulsen, G., Halstensen, T.S.,
DQ(alpha 1*0501,beta 1*0201) restricted T cells isolated from the
small intestinal mucosa of Celiac disease patients. J. Exp. Med.
Maccalli, C., Pende, D., Castelli, C., Mingari, M.C., Robbins, P.F.,
and Parmiani, G. (2003). NKG2D engagement of colorectal cancer-
specific T cells strengthens TCR-mediated antigen stimulation and
elicits TCR independent anti-tumor activity. Eur. J. Immunol. 33,
Maiuri, L., Auricchio, S., Coletta, S., De Marco, G., Picarelli, A., Di
Tola, M., Quaratino, S., and Londei, M. (1998). Blockage of T-cell
costimulation inhibits T-cell action in Celiac disease. Gastroenterol-
ogy 115, 564–572.
Maiuri, L., Ciacci, C., Auricchio, S., Brown, V., Quaratino, S., and
Londei, M. (2000). Interleukin 15 mediates epithelial changes in Ce-
liac disease. Gastroenterology 119, 996–1006.
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.
Maki, M., Mustalahti, K., Kokkonen, J., Kulmala, P., Haapalahti, M.,
Karttunen, T., Ilonen, J., Laurila, K., Dahlbom, I., Hansson, T., et al.
(2003). Prevalence of Celiac disease among children in Finland. N.
Engl. J. Med. 348, 2517–2524.
Matysiak-Budnik, T., Candalh, C., Dugave, C., Namane, A., Cellier,
C., Cerf-Bensussan, N., and Heyman, M. (2003). Alterations of the
intestinal transport and processing of gliadin peptides in Celiac
disease. Gastroenterology 125, 696–707.
McNeish, A.S.,Harms, H.K., Rey,J., Shmerling, D.H.,Visakorpi, J.K.,
and Walker-Smith, J.A. (1979). The diagnosis of coeliac disease. A
commentary on the current practices of members of the European
Society for Paediatric Gastroenterology and Nutrition (ESPGAN).
Arch. Dis. Child. 54, 783–786.
Mention, J.J., Ben Ahmed, M., Begue, B., Barbe, U., Verkarre, V.,
E., et al. (2003). Interleukin 15: a key to disrupted intraepithelial
lymphocyte homeostasis and lymphomagenesis in Celiac disease.
Gastroenterology 125, 730–745.
Mizuki, N., Ota, M., Kimura, M., Ohno, S., Ando, H., Katsuyama, Y.,
Yamazaki, M., Watanabe, K., Goto, K., Nakamura, S., et al. (1997).
Triplet repeat polymorphism in the transmembrane region of the
MICA gene: a strong association of six GCT repetitions with Behcet
disease. Proc. Natl. Acad. Sci. USA 94, 1298–1303.
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 transglutaminase selectively modifies gliadin pep-
tides that are recognized by gut-derived T cells in Celiac disease.
Nat. Med. 4, 713–717.
Monteleone, G., Pender, S., Wathen, N., and MacDonald, T. (2001).
Interferon-a drives T cell-mediated immunopathology in the intes-
tine. Eur. J. Immunol. 31, 2247–2255.
Pende, D., Cantoni, C., Rivera, P., Vitale, M., Castriconi, R., Marcen-
aro, S., Nanni, M., Biassoni, R.,Bottino, C., Moretta, A., and Moretta,
L. (2001). Role of NKG2D in tumor cell lysis mediated by human NK
cells: cooperation with natural cytotoxicity receptors and capability
of recognizing tumors of nonepithelial origin. Eur. J. Immunol. 31,
Roberts, A.I., Lee, L., Schwarz, E., Groh, V., Spies, T., Ebert, E.C.,
and Jabri, B. (2001). NKG2D receptors induced by IL-15 costimulate
CD28-negative effector CTL in the tissue microenvironment. J. Im-
munol. 167, 5527–5530.
Salih, H.R., Rammensee, H.G., and Steinle, A. (2002). Cutting edge:
down-regulation ofMICA onhuman tumors byproteolytic shedding.
J. Immunol. 169, 4098–4102.
Shan, L., Molberg, O., Parrot, I., Hausch, F., Filiz, F., Gray, G.M.,
Sollid, L.M.,and Khosla, C.(2002). Structural basis forgluten intoler-
ance in celiac sprue. Science 297, 2275–2279.
Sollid, L.M. (2002). Coeliac disease: dissecting a complex inflamma-
tory disorder. Nat. Rev. Immunol. 2, 647–655.
Steinle, A., Li, P., Morris, D.L., Groh, V., Lanier, L.L., Strong, R.K.,
and Spies, T. (2001). Interactions of human NKG2D with its ligands
MICA, MICB, and homologs of the mouse RAE-1 protein family.
Immunogenetics 53, 279–287.
S., Bahram, S., and Inoko, H. (2002). A basolateral sorting motif in
Sugimura, K., Ota, M., Matsuzawa, J., Katsuyama, Y., Ishizuka, K.,
Mochizuki, T., Mizuki, N., Seki, S.S., Honma, T., Inoko, H., and Asa-
Role of NKG2D/MICA Interaction in Celiac Disease Download full-text
kura, H. (2001). A close relationship of triplet repeat polymorphism
in MHC class I chain-related gene A (MICA) to the disease suscepti-
bility and behavior in ulcerative colitis. Tissue Antigens 57, 9–14.
Vetter, C.S., Groh, V., thor Straten, P., Spies, T., Brocker, E.B., and
Becker, J.C. (2002). Expression of stress-induced MHC class I re-
lated chain molecules on human melanoma. J. Invest. Dermatol.
Vivier, E., Tomasello, E., and Paul, P. (2002). Lymphocyte activation
via NKG2D: towards a new paradigm in immune recognition? Curr.
Opin. Immunol. 14, 306–311.
Wu, J., Song, Y., Bakker, A.B., Bauer, S., Spies, T., Lanier, L.L., and
Phillips, J.H. (1999). An activating immunoreceptor complex formed
by NKG2D and DAP10. Science 285, 730–732.