вход по аккаунту


Autoantibodies in gluten ataxia recognize a novel neuronal transglutaminase.

код для вставкиСкачать
Autoantibodies in Gluten Ataxia Recognize a
Novel Neuronal Transglutaminase
Marios Hadjivassiliou, MD,1 Pascale Aeschlimann, BSc,2 Alexander Strigun, MSc,2 David S. Sanders, MD,3
Nicola Woodroofe, PhD,4 and Daniel Aeschlimann, PhD2
Objective: Gluten sensitivity typically presents as celiac disease, a chronic, autoimmune-mediated, small-intestinal disorder.
Neurological disorders occur with a frequency of up to 10% in these patients. However, neurological dysfunction can also be the
sole presenting feature of gluten sensitivity. Development of autoimmunity directed toward different members of the transglutaminase gene family could offer an explanation for the diversity in manifestations of gluten sensitivity. We have identified a
novel neuronal transglutaminase isozyme and investigated whether this enzyme is the target of the immune response in patients
with neurological dysfunction.
Methods: Using recombinant human transglutaminases, we developed enzyme-linked immunosorbent assays and inhibition
assays to analyze serum samples of patients with gluten-sensitive gastrointestinal and neurological disorders, and various control
groups including unrelated inherited or immune conditions for the presence and specificity of autoantibodies.
Results: Whereas the development of anti-transglutaminase 2 IgA is linked with gastrointestinal disease, an anti-transglutaminase
6 IgG and IgA response is prevalent in gluten ataxia, independent of intestinal involvement. Such antibodies are absent in ataxia
of defined genetic origin or in healthy individuals. Inhibition studies showed that in those patients with ataxia and enteropathy,
separate antibody populations react with the two different transglutaminase isozymes. Furthermore, postmortem analysis of brain
tissue showed cerebellar IgA deposits that contained transglutaminase 6.
Interpretation: Antibodies against transglutaminase 6 can serve as a marker in addition to human leukocyte antigen type and
detection of anti-gliadin and anti-transglutaminase 2 antibodies to identify a subgroup of patients with gluten sensitivity who
may be at risk for development of neurological disease.
Ann Neurol 2008;64:332–343
Celiac disease (CD) is a common T-cell–mediated autoimmune disorder characterized by its linkage to specific human leukocyte antigen (HLA) alleles: HLADQ2 and -DQ8. In susceptible individuals,
consumption of gluten triggers a CD4⫹ T-cell response to gliadin, as well as a B-cell response to gliadin
and self-antigens.1 Transglutaminase 2 (TG2) is the
autoantigen recognized in the endomysium of the gut
by sera from patients with CD.2 TG2-specific antibodies are almost exclusively found in CD, are characteristic of untreated clinically symptomatic, as well as latent disease, and hence have become accepted as an
excellent diagnostic indicator of CD.
TG2 is one of a family of enzymes that covalently
cross-link or modify proteins by formation of an
isopeptide bond between a peptide-bound glutamine
residue and a primary amine, most commonly a lysine
residue either within the same or a neighboring
polypeptide chain.3 However, in some instances, TG2
may react with H2O in preference over an amine, leading to the deamidation of glutamine residues.4 The biological significance of this latter activity has only recently been established in connection with CD: gliadin
proteins, the immunological trigger of gluten sensitivity, are glutamine-rich donor substrates amenable to
deamidation.5 Therefore, TG2 apparently contributes
to disease development in at least two ways: first, by
deamidating gluten peptides, thereby increasing their
reactivity with HLA-DQ2/DQ8, which potentiates the
T-cell response6,7; and second, by haptenization of selfantigens through cross-linking with gliadins.8,9 The absence of intestinal T-cell responses to gluten in the majority of individuals carrying HLA-DQ2/DQ8 and the
preferential T-cell responses to deamidated gluten fragments in patients with CD indicates that there is tolerance to unmodified gluten peptides. Therefore, acti-
From the 1Department of Neurology, Royal Hallamshire Hospital,
Sheffield; 2Matrix Biology and Tissue Repair Research Unit, School
of Dentistry, Cardiff University, Cardiff; 3Department of Gastroenterology, Royal Hallamshire Hospital; and 4Biomedical Research
Centre, Sheffield Hallam University, Sheffield, United Kingdom.
Received Feb 3, 2008, and in revised form May 17. Accepted for
publication May 30, 2008.
Additional Supporting Information may be found in the online version of this article.
Published online in Wiley InterScience (
DOI: 10.1002/ana.21450
Address correspondence to Dr Aeschlimann, Matrix Biology and
Tissue Repair Research Unit, School of Dentistry, Cardiff University, Heath Park, Cardiff CF14 4XY, United Kingdom.
© 2008 American Neurological Association
Published by Wiley-Liss, Inc., through Wiley Subscription Services
vation of TG2 and deamidation of gluten peptides
appears to be central to disease development.
The mechanism underlying the formation of autoantibodies to TG2 is not fully understood. Autoreactive T-cells to TG2 have not been isolated from patients with CD10; hence, formation of complexes of
gluten and TG2 may permit gluten-reactive T-cells to
provide the necessary help to TG2-specific B-cells.
This would also explain why serum antibodies to TG
disappear when patients are subjected to a gluten-free
diet. However, it has recently also been suggested that
molecular mimicry, that is, shared epitopes between
deamidated gliadin and TG2,11 or exposure of masked
epitopes on reaction of the enzyme with antigenic gliadin peptides, which is associated with a large conformational change,12 could lead to antibody production.
It is now accepted that gluten sensitivity is a systemic illness that can manifest in a range of organ systems.13–16 Such manifestations can occur independently of the presence of the classic small-bowel lesion
(triad of villous atrophy, crypt hyperplasia, and increased intraepithelial lymphocytes) that defines CD.
Although it is a matter of debate whether the autoantibodies are involved in extraintestinal disease development, the formation of IgA deposits in the respective
tissue supports such a role.15,16 Dermatitis herpetiformis (DH) is a gluten-sensitive dermatopathy that presents with an itchy vesicular rash and is one example of
extraintestinal manifestation of gluten sensitivity. Although the majority of patients with DH have evidence of enteropathy on duodenal biopsy, these patients rarely complain or present with gastrointestinal
symptoms.17 This form of gluten sensitivity may be determined by autoimmunity directed primarily toward
TG3 rather than TG2,18 suggesting that the explanation for the heterogeneity in disease manifestation may
lie in the specificity of the immune response.
Neurological disorders have also been recognized as a
manifestation of gluten sensitivity, with cerebellar involvement, also known as gluten ataxia (GA), and peripheral nerve involvement (gluten neuropathy) being
the most common presentations.19 –21 Currently, GA is
defined as sporadic cerebellar ataxia associated with the
presence of circulating anti-gliadin antibodies, in the
absence of an alternative aetiology for ataxia.22 Patients
with GA present with ataxia, almost always in the absence of gastrointestinal symptoms.23 Only a third of
patients with GA will have evidence of enteropathy on
duodenal biopsy. However, like DH, GA responds to a
gluten-free diet, but the degree of response depends on
the duration of the ataxia, which, in turn, correlates
with the degree of irreversible damage (Purkinje cell
loss) sustained by the cerebellum.24,25 Indeed, this is
one of the fundamental differences among CD, DH,
and GA in that both the gut and the skin have the
potential to regenerate, whereas the neural tissue does
not. Substantial evidence suggests that the mechanism
of neural damage is immune mediated.22 GA patients
have circulating anti-Purkinje cell antibodies in their
sera, and up to 50% of patients with GA have oligoclonal bands in their cerebrospinal fluid suggestive of
intrathecal antibody production.26,27 A role for these
antibodies in the pathophysiology of the ataxia is suggested by clinical improvement with intravenous immunoglobulin therapy.28,29 However, the autoantigen
targeted in the central nervous system (CNS) has not
been identified.
We have identified a novel TG, TG6, and shown
that it is predominantly expressed by a subset of neurons in the CNS.30,31 The preferential expression of
TG6 in neural tissue and its close homology to TG2
and TG3 provides a clear possibility that this enzyme
could be involved in the pathogenesis of gluten sensitivity–related neurological dysfunction.
Subjects and Methods
Patient Sera
Sera were collected from the ataxia, gluten sensitivity/neurology, and CD clinics of the Departments of Neurology and
Gastroenterology, Royal Hallamshire Hospital, Sheffield,
United Kingdom, with informed consent of patients and the
approval of the Research Ethics Committees (REC06/
Q2307/6). Patient groups included in the analysis were as
follows: Sera from 20 patients with newly diagnosed CD collected before the commencement of a gluten-free diet. CD
was confirmed on duodenal biopsy, and patients had no evidence of neurological manifestations. Groups with neurological disease included baseline sera from 34 patients with GA
(defined as otherwise sporadic idiopathic ataxia but positive
for anti-gliadin antibodies: IgG, IgA, or both) and 17 sera
from patients with peripheral idiopathic neuropathy positive
for anti-gliadin antibodies. Peripheral idiopathic neuropathy
patients were negative for anti–myelin-associated glycoprotein and anti-GM1, and had no evidence of enteropathy on
biopsy. A genetic ataxia group served as the ataxia disease
control group. This included 18 patients with either genetically characterized ataxia or clear evidence of autosomal
dominant family history of ataxia. A further control group
(Misc) of 14 patients included immune-mediated but glutenunrelated disease (vasculitis, viral cerebellitis, paraneoplastic
ataxia, glutamic acid decarboxylase ataxia). Finally, samples
from 19 healthy individuals were used as controls. Further
patient group parameters are provided in online Supplementary Table 1.
Production of Recombinant Human
A full-length complementary DNA encoding human TG6
was obtained by polymerase chain reaction from
poly(A⫹)RNA isolated from the lung carcinoma cell line
H69, as described previously.30 In brief, overlapping polymerase chain reaction fragments were amplified, TA-cloned,
and the full-length complementary DNA constructed by
subcloning the overlapping fragments into the pCRII-vector
Hadjivassiliou et al: Autoantibodies to TG6
Table 1. Concentrations of IgG and IgA against
Transglutaminases 2, 3, and 6 (in arbitrary units) in
Serum of Healthy Control Subjects and Patients
Group (n)
HC (19)
CD (20)
GAE (15)
GAo (19)
GenA (18)
PN (17)
Misc (14)
Median IgA, au
(95% CI)
Median IgG, au
(95% CI)
20.0 (15.9-30.3)
120.0 (92.7-122)a
93.0 (72.6-113)a
27.0 (26.2-35.2)c
25.5 (22.2-28.3)c
26.0 (22.4-31.7)c
29.0 (22.2-36.2)c
9.0 (1.9-22.0)
62.0 (45.6-82.9)a
54.0 (34.3-73.9)d
41.0 (33.8-57.9)d
21.0 (12.6-25.4)c
29.0 (20.7-35.2)c
19.5 (12.6-48.2)c
Group (n)
HC (19)
CD (20)
GAE (15)
GAo (19)
GenA (18)
PN (17)
Misc (14)
Median IgA, au
(95% CI)
Median IgG, au
(95% CI)
23.0 (16.1-29.5)
46.0 (38.0-76.3)b
46.0 (38.3-78.6)d
31.0 (29.5-40.1)c
23.0 (21.6-30.3)c
27.0 (19.0-34.4)c
30.5 (21.5-34.3)c
26.0 (23.9-33.1)
34.0 (30.7-45.4)c
34.0 (31.9-62.7)c
42.0 (37.3-53.5)d
28.0 (23.6-31.9)c
31.0 (28.0-35.1)c
29.0 (19.7-48.6)c
Group (n)
HC (19)
CD (20)
GAE (15)
GAo (19)
GenA (18)
PN (17)
Misc (14)
Median IgA, au
(95% CI)
Median IgG, au
(95% CI)
24.0 (18.5-30.5)
47.0 (38.2-67.6)b
41.0 (36.0-73.9)b
53.0 (37.9-74.1)b
23.0 (17.0-31.7)c
39.0 (33.5-50.9)c
36.5 (28.0-48.5)c
27.0 (22.1-33.0)
37.0 (31.4-45.7)c
51.0 (42.0-57.7)a
43.0 (36.8-52.6)d
31.0 (25.9-35.9)c
29.0 (28.8-39.2)c
34.0 (26.1-41.3)c
a:p ⬍ 0.001, b:p ⬍ 0.05, c:p value not significant, d:p ⬍
0.01, Kruskal–Wallis post test analysis.
TG ⫽ transglutaminase; HC ⫽ healthy control subjects;
CD ⫽ patients with celiac disease; GAE ⫽ gluten ataxia with
enteropathy; GAo ⫽ gluten ataxia only; GenA ⫽ genetic
ataxias; PN ⫽ peripheral neuropathy; Misc ⫽ various glutenunrelated autoimmune conditions.
(Invitrogen, La Jolla, CA) using appropriate restriction endonucleases. Sequence analysis demonstrated two single nucleotide deletions (C75 and G1568). The mutations were corrected by site-specific insertion mutagenesis using the
QuickChange XL Site Directed Mutagenesis kit (Stratagene,
La Jolla, CA). Finally, the coding sequence was subcloned
into derivatives of the prokaryotic expression vector
pJOE270232 for rhamnose-regulated expression. Complementary DNAs encoding TG2 and TG3 were subcloned
into the same expression vector.33 A His6-tag was added to
the native sequence N-terminally (TG2 and TG3) or
Annals of Neurology
Vol 64
No 3
September 2008
C-terminally (TG6) for purification of the recombinant proteins by Ni2⫹-chelating affinity chromatography.
Escherichia coli BL21 transformed with the expression
constructs were grown in Luria–Bertani broth in baffled
flasks at 37°C and 220rpm to OD600 of 0.6, before chilling
to 20°C and induction of transgene expression by addition
of rhamnose to a final concentration of 0.5% (0.1% for
TG6). After incubation for a further 6 (TG6) or 24 hours
(TG2/TG3) at 20°C, bacteria were collected by centrifugation at 3,000g for 20 minutes, resuspended in buffer A
(50mM Na2HPO4, pH 8.0, 300mM NaCl for TG2/TG3;
50mM 4-morpholinepropanesulfonic acid, pH6.8, 500mM
NaCl, 10mM glutathione, 30% glycerol for TG6) to obtain
a 15% cell suspension, and the expressed protein harvested
by lysis of the cells using a “French-Press” (1,000psi). The
lysate was cleared from insoluble material by centrifugation
at 11,500g for 30 minutes at 4°C and applied to a 1ml HisTrap HP column (Amersham Bioscience, Buckinghamshire,
United Kingdom) equilibrated in buffer A at 4°C and a flow
rate of 0.5ml/min. The resin was washed, initially with
buffer A until OD280 of less than 0.005 was reached, and
then with 100ml buffer A containing 30 (TG2/TG3) or
50mM (TG6) imidazole, before elution of His-tagged protein with buffer A containing 150mM imidazole for TG2/
TG3 or with 50mM 4-morpholinepropanesulfonic acid,
pH6.8, 300mM NaCl, 5mM dithiothreitol, 350mM imidazole, and 10% glycerol for TG6. Eluted protein was dialyzed
(Spectra/Por 4; Spectrum Laboratories, Houston, TX) extensively against buffer B (20mM Tris/HCl, pH 7.2, 1mM
EDTA, 100mM NaCl for TG2 and TG3; or 20mM Tris/
HCl, pH 8.0, 300mM NaCl, 5mM dithiothreitol, 10%
glycerol for TG6). When desired, enzymes were purified further by ion exchange chromatography using a HR10/10 column packed with Resource Q15 (Amersham Bioscience)
whereby TGs were eluted as a single sharp peak within a
20-volume gradient of 0.1 to 0.7M NaCl. Enzymes were dialyzed into buffer B, concentrated to approximately 2mg/ml
using Centriprep-YM30 (Amicon, Danvers, MA) concentrators, and stored at ⫺20°C.
Protein Analysis
Protein concentrations were determined using the bicinchoninic acid reagent (Pierce, Rockford, IL) as described by the
supplier, with bovine serum albumin (BSA) as a standard.
Protein purity was verified by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) in 4 to 20%
Tris-glycine gels (Novex, San Diego, CA) under reducing
conditions (1% 2-mercaptoethanol), followed by staining
with Coomassie brilliant blue R. Matrix-assisted laser desorption ionization time of flight (MALDI-TOF) mass spectrometry and tandem mass spectrometry were conducted on an
Applied Biosystems 4800 MALDI TOF/TOF analyzer
(Cardiff University service facility; Foster City, CA) on tryptic digests of protein bands recovered from SDS-PAGE gels.
Transamidation activity was evaluated by determining
protein-bound radiolabel using trichloroacetic acid precipitation and scintillation counting after incubation of 125nM
enzyme with 8␮Ci [3H]putrescine (NEN/PerkinElmer,
Waltham, MA.) and 625nM N,N-dimethyl casein (Sigma,
St. Louis, MO) in 160␮l 100mM Tris/HCl, pH 8.3, con-
taining 5mM CaCl2, 10mM dithiothreitol, at 37°C for different times.34
Enzyme-Linked Immunosorbent Assays
High-capacity protein binding 96-well plates (Immulon
2HB; Thermo Electron, Waltham, MA) were coated with
100␮l/well of 5␮g/ml TG in Tris-buffered saline (TBS;
20mM Tris/HCl, pH 7.4, 150mM NaCl) overnight at 4°C.
All subsequent incubations were conducted at room temperature, and all binding steps were followed by five rinses with
TBS containing 0.01% Tween 20. Nonspecific binding was
blocked by incubation with 200␮l/well 3% BSA (immunoassay grade, Sigma 05477) in TBS for 1 hour. Patient sera
were diluted 1:100 in 1% BSA in TBS, any protein aggregates present removed by centrifugation at 10,000g for 5
minutes, and 100␮l/well applied to coated plates for 90 minutes. After rinsing, serum antibody binding was detected by
incubation with 100␮l/well of either peroxidase-conjugated
affinity pure anti–human IgA (Jackson ImmunoResearch,
West Grove, PA; diluted 1:2,000) or anti–human IgG
(Dako, Carpinteria, CA; diluted 1:1000) for 90 minutes.
The reaction was finally developed for 30 minutes using
5mM 5-amino-2-hydroxybenzoic acid/NaOH, pH 6.0,
0.005% H2O2, as a peroxidase substrate solution (100␮l/well)
and stopped by addition of 100␮l 1M NaOH to each well.
After 15 minutes, the absorbance at 490nm was measured.
All serum samples were analyzed on wells containing antigen or only BSA, included on the same plate. A selected
negative and positive reference serum was also run in parallel
on all plates. The BSA only background was subtracted from
values for antigen, and specific binding expressed in arbitrary
units as a percentage of the reference sera. Data points represent the mean of six determinations for each serum sample.
Inhibition enzyme-linked immunosorbent assays (ELISAs)
followed the above protocol, but each serum was titrated to
identify the dilution that yielded half-maximal binding. The
respective sera dilutions were incubated with a concentration
series of TG as indicated overnight at 4°C while shaking, the
mixture subsequently added to TG-coated plates for 40 minutes, and the reaction developed as described earlier. Experimental data points are shown in combination with theoretical inhibition curves calculated according to Engel and
antigen bound ⫽
1 ⫹ K共S ⫹ cAb兲 ⫺
冑共1 ⫹ K共S ⫹ cAb兲兲2 ⫺ 4K2cAb共S兲
共1 ⫺ Bmin兲 ⫹ Bmin
and S ⫽ cl ⫹ cAg
whereby concentrations of competitor, coated antigen,
and antibody are denominated cI, cAg, and cAb, respectively.
Bmin reflects the level of binding at saturating concentrations
of competitor.
Receiver operating characteristic plot analysis was performed
to determine the optimal decision threshold for ELISAs. For
comparison between patient groups, Kruskal–Wallis non-
parametric analysis was used and significance between individual patient groups and healthy control subjects determined from Dunn’s post test. For comparison of inhibition
with TG2 or TG6 in ELISAs, Wilcoxon’s two-tailed, signed
ranks test for pairs was used.
Histochemical Methods
For immunohistochemistry, unfixed 5␮m cryosections were
cut, adsorbed to gelatin-coated slides, and immunolabeled
with the following antibodies using the peroxidase protocol
as described previously36: monoclonal antibody CUB7402 to
TG2 (6.5␮g/ml; NeoMarkers/LabVision, Fremont, CA),
affinity-purified antibodies to TG6 raised against peptide
CGWRDDLLEPVTKPS in goat (15␮g/ml) or nonimmune
IgG (15␮g/ml; ChromePure; Jackson ImmunoResearch) as a
control. Western blotting was conducted to demonstrate the
absence of cross-reactivity of TG6 antibodies with other TG
isozymes (see Supplementary Fig 1). Colocalization of TG6
with IgA deposits was investigated with ␣-chain–specific
fluorescein-conjugated F(ab⬘)2 fragment to human IgA
(0.75␮g/ml; Jackson ImmunoResearch) and visualized using
confocal laser-scanning microscopy (z-stack of 0.2␮m optical
Production of Recombinant Human
A nonmammalian host cell was chosen for antigen production to minimize the risk for cross-reactivity of sera
with impurities in protein preparations, which is a particular concern in sera from patients with autoimmune
disease.37 TG2 and TG3 were expressed in E. coli as
fusion proteins with an N-terminal His6-tag and purified by sequential affinity and ion-exchange chromatography with a yield of approximately 40mg/l culture.
In contrast, only constructs encoding C-terminally
tagged TG6 yielded detectable amounts of soluble enzyme (by Western blotting), and expression had to be
tightly controlled by limiting promoter activation, expression time, and temperature to prevent formation of
protein aggregates. Consequently, yields of active enzyme were only 100 to 200␮g/l culture. The purified
proteins gave a single band on Coomassie blue–stained
SDS-PAGE gels (Fig 1), whereby the migration was
principally consistent with the respective calculated
molecular masses of 78.1, 77.7, and 80.1kDa for TG2,
TG3, and TG6, respectively, although migration of
TG2 was somewhat retarded compared with TG3 and
TG6. Their identity was further verified by demonstrating enzymatic activity (data not shown), immunoreactivity with antibodies to the respective enzymes
(data not shown), and peptide fingerprinting and/or sequencing using mass spectrometry. The latter gave
more than 50% sequence coverage and MASCOT
scores exceeding 850 (search engine for protein identification using mass spectrometry data; www. The retarded migration of TG2 in
Hadjivassiliou et al: Autoantibodies to TG6
Fig 1. Sodium dodecyl sulfate polyacrylamide gel electrophoresis
(SDS-PAGE) analysis of purified recombinant transglutaminases (TGs). Proteins were separated in 4 to 20% SDS-PAGE
gel under reducing conditions and stained with Coomassie
blue: 0.2 or 5.0␮g TG6 (lanes 1, 2), TG2 (lanes 3, 4), and
TG3 (lanes 5, 6), respectively. Migration of molecular mass
standards is indicated on the left. Apparent molecular mass of
TGs determined from electrophoretic mobility is given in
SDS-PAGE is consistent with previous data.18 Nevertheless, TG2 spectra were inspected manually for peaks
that may represent posttranslationally modified peptide
sequences and several peaks sequenced using tandem
mass spectrometry, but no modifications were found
(data not shown).
Sera of Patients with Celiac Disease and Gluten
Ataxia Contain IgA and IgG Class Antibodies to
Transglutaminase 6
For antibody detection in sera, an ELISA protocol was
developed and verified, and all ELISAs were subsequently performed using the same antigen concentration
for coating, serum dilution, and incubation time for development of the reaction. For GA sera, the signal-tobackground ratio was greatest at a dilution of 1:100 with
a coating density of 5␮g/ml; hence, these conditions
were used for all assays. A small number of sera produced abnormally high signal on wells incubated with
blocking agent but in the absence of antigen. Indeed,
high serum immunoglobulin concentrations have been
reported to cause false-positive classification in TG2 IgA
assays.38 Therefore, all assays were conducted in the
presence and absence of antigen, and the relative signal
was used for analysis. A selected positive and negative
reference serum was included in each assay to control for
Annals of Neurology
Vol 64
No 3
September 2008
assay performance and for data normalization into arbitrary units as a function of the reference samples. Using
20 CD patients and 19 blood donor samples, we found
the sensitivity of the TG2 IgA assay to be 90% with a
specificity of 100%, whereas the sensitivity was reduced
to 60% when measuring TG2 IgG, consistent with previous observations.38 The area under the receiver operating characteristic curve was 0.983 and 0.882 for TG2
IgA and IgG, respectively. In house assay, performance
was further evaluated against a commercially available
clinical assay for TG2 IgA (Genesis Diagnostics, Cambridgeshire, United Kingdom), and the mean interassay
variation was found to be 11.0 and 3.6% for CD and
healthy groups, respectively.
Antibodies to other TG isozymes were detected in
sera following the same procedure; the results are summarized in Table 1. Within the group of patients with
classic CD, 18 of 20 had positive serology for antiTG2 IgA, with the remaining 2 patients having either
only IgA-type antibodies to TG3 or TG6, respectively.
IgG titers, by comparison with IgA, were generally
lower and a positive test less frequent and always associated with anti-TG2 IgA (Tables 1 and 2). Fifty-five
percent of CD patients tested positive for multiple
TGs, 45% for TG3, and 45% for TG6, whereby 35%
had antibodies reacting with all 3 isozymes. Although
the mean antibody concentrations against TG2 were
significantly greater than those to other isozymes when
comparing groups (see Table 1), mean titers were similar when comparing only individuals who tested positive (Fig 2). These results confirm that the B-cell response in gluten sensitivity can be directed to TG
isozymes other than TG2 and suggests that frequently
antibodies reacting with TG3 or TG6 are present.
However, the prevalence of anti-TG3 IgA in our cohort of untreated CD patients was only about half that
previously reported.18
GA patients were grouped into two subgroups,
those with enteropathy (GAE) and those without
(GAo). IgA to TG2 was an excellent predictor of the
presence of enteropathy with 12 of 15 GAE patients
being positive as opposed to only 1 of 19 GAo patients testing marginally positive (see Fig 2A). Similarly, whereas anti-TG3 IgA could be detected in the
GAE group with a frequency similar to that in CD,
GAo patients were not different from control subjects
(see Fig 2B). This is also consistent with the finding
that 73% of GAE patients were endomysial antibody
(EMA)–positive, whereas all of the GAo patients were
EMA-negative (Table 3). In contrast, the frequency of
positive results in the TG6 IgA ELISA was similar for
the 2 groups, that is, 6 of 15 (40%) with GAE and 9
of 19 (47%) with GAo (see Fig 2C). Furthermore,
whereas anti-TG6 IgG was seen only in 3 of 20 CD
patients, the prevalence was significantly greater in
GAo patients, 6 of 19 (32%), and even greater in
Table 2. Prevalence of Gluten Sensitivity Detected with IgG and IgA Antibody Assays against Transglutaminases
2, 3, and 6 (in %)
Group (n)
IgG and
CD (20)
GAE (15)
GAo (19)
GenA (18)
PN (17)
Misc (14)
HC (19)
TG ⫽ transglutaminase; CD ⫽ patients with celiac disease; GAE ⫽ gluten ataxia with enteropathy; GAo ⫽ gluten ataxia only;
GenA ⫽ genetic ataxias; PN ⫽ peripheral neuropathy; Misc ⫽ various gluten-unrelated autoimmune conditions; HC ⫽ healthy control
GAE patients, 8 of 15 (53%) (see Table 2). Also,
some patients tested positive exclusively for IgG class
antibodies. The overall prevalence of anti-TG6 IgA
and/or IgG was 62% in GA compared with 45% in
CD. None of the patients with genetic ataxias or
healthy control subjects tested positive for anti-TG antibodies, whereas 1 in 14 patients with gluten-unrelated
immune-mediated disease was found to have IgG but
not IgA class anti-TG antibodies. The median antibody
concentrations were significantly different in patients
with gluten sensitivity (CD, GA) as compared with control subjects ( p ⬍ 0.0005 for IgA; p ⬍ 0.001 for IgG),
whereas no significant differences were seen between the
control groups (see Table 1). Patients with peripheral
neuropathy were not different from control subjects in
all anti-TG IgG assays, as well as IgA assays to TG2 and
TG3, but had marginally increased readings in the antiTG6 IgA ELISA (see Fig 2). The significance of this is
Discrete Antibody Populations in Sera React with
Transglutaminase 6 or 2
To assess whether isozyme-specific antibodies were
present or the same antibodies cross-reacted between
these enzymes, we conducted inhibition studies. Sera
were preincubated with different concentrations of either the antigen or another TG isozyme before analysis in the ELISAs. The results are presented as degree
of inhibition produced in the ELISA by preincubation with TGs as compared with a control sample
preincubated with buffer alone. Representative examples of individual sera are shown in Figures 3A to E,
whereas a comparative analysis with a set concentration of inhibitor is shown in Figure 3F for all patients
who displayed reactivity toward both TG2 and TG6.
In most sera, no cross-reacting antibodies could be
detected even at high concentrations of inhibitor (see
Figs 3C–E). In the anti-TG2 IgA ELISA, TG2 was
an effective inhibitor yielding a mean inhibition of 37
(CD) and 55% (GAE) as opposed to 1 (CD) and
15% (GAE) with TG6 (see Fig 3F). Only in 1 (GAE)
of 14 patients could significant inhibition by TG6 be
detected and, therefore, could the presence of antibodies reacting with both isozymes in addition to
TG2-specific antibodies be demonstrated (see Fig
3B). Conversely, in the anti-TG6 IgA ELISA, TG6
was the effective inhibitor with a mean inhibition of
71 (CD) and 61% (GAE) in comparison with 18
(CD) and 14% (GAE) with TG2 (see Fig 3F). Although with TG2 partial inhibition was seen in three
sera at much greater concentrations than with TG6,
only for one patient (CD) was TG2 equally effective
as TG6 in blocking the reaction. Despite small sample numbers, a comparison of the inhibition by TG2
or TG6 in the ELISAs (see Fig 3F) showed that the
medians differed significantly ( p ⬍ 0.016 for CD;
p ⫽ 0.031 for GAE). These data, together with the
finding that a number of patients tested positive exclusively for anti-TG6 (see Table 2; see Fig 2), provide evidence that patients develop populations of antibodies that are specific for or have greater avidity for
TG6 than the other TG family members.
IgA Deposits in Cerebellum of Gluten Ataxia Patient
Contain Transglutaminase 6
In postmortem analysis of a GAo patient, we have previously demonstrated the accumulation of IgA deposits
in the cerebellum and brainstem, most prominently
within the muscular layer surrounding vessels but also
in brain tissue proper.16 We have stained frozen sections from various regions of the brain of the same
patient using antibodies to TG6 and found codistribution of TG6 with these IgA deposits (Fig 4). In the
cerebellum (see Fig 4D) and medulla, the perivascular
areas where an endomysium-associated IgA deposition
occurs16 were intensely positive for TG6, and to a
Hadjivassiliou et al: Autoantibodies to TG6
Fig 2. Analysis of serum anti-transglutaminase 2 (anti-TG2)
(A), anti-TG3 (B), and anti-TG6 (C) IgA by enzyme-linked
immunosorbent assay (ELISA). Relative concentrations of antibodies in healthy control subjects (HCs) or patients with celiac
disease (CD), gluten ataxia with enteropathy (GAE), gluten
ataxia only (GAo), genetic ataxias (GenA), and peripheral
neuropathy (PN), and unrelated disease control subjects (Misc)
are shown in arbitrary units. Mean antibody titres are indictaed by horizontal bars. Threshold for a positive test is indicated by the dotted line.
lesser extent, brain tissue itself was also stained,
whereas staining was absent in the parietal lobe (see Fig
4F). In contrast, TG6 could not be detected in vascular structures of normal cerebellum (see Fig 4C).
Annals of Neurology
Vol 64
No 3
September 2008
Our group’s initial discovery that ataxia can be the presenting feature of gluten sensitivity was not only based
on a greater prevalence of anti-gliadin antibodies in patients with idiopathic sporadic ataxia by comparison
with control groups,22 but also on the subsequent finding of a significantly greater prevalence of enteropathy
within the same patient group, as well as association
with the HLA genotype linked to CD, which strongly
suggested that anti-gliadin antibodies are not an epiphenomenon.27 Nonetheless, one of the ongoing criticisms
of the definition of GA based on the presence of antigliadin antibodies has been the fact that such antibodies
can be found in unrelated autoimmune conditions and
also in up to 12% of “healthy” subjects.39 Despite not
being diagnostic for CD by themselves, anti-gliadin antibodies may, in some cases, signify gluten sensitivity
without overt evidence of gut involvement. However, it
is also possible that in a subgroup of patients with ataxia
and anti-gliadin antibodies, the ataxia may be coincidental rather than causally linked to gluten sensitivity. In
this study, we identify a novel TG as the prevalent autoantigen in the CNS in GA. We show that among sporadic idiopathic ataxia patients with anti-gliadin antibodies, all of those who present with enteropathy and 68%
of those without enteropathy had circulating anti-TG
antibodies. In contrast, such antibodies were absent in
healthy control subjects or patients with inherited ataxias. Interestingly, the prevalence of HLA-DQ2 and
-DQ8 differed accordingly in the different ataxia groups
examined, with 100% (one of the patients carried only
half of the DQ2 heterodimer, DQB1*0202, which is
sufficient for susceptibility to gluten sensitivity40) and
72% in GAE and GAo, respectively, as opposed to 44%
in the genetic ataxia group, with the latter being comparable with the regional population average of 38%.
This supports a link between gluten sensitivity and idiopathic ataxia in most patients within a group that can
be expected to be heterogeneous because classification
was based solely on classic anti-gliadin antibody test. In
a group of peripheral neuropathy patients with antigliadin antibodies, no similar correlation could be established, although a few patients tested marginally positive
for anti-TG antibodies when compared with the other
control groups (see Table 1; see Fig 2).
Not only antibody prevalence (see Table 2) but also
titers suggest a bias of the immune response toward
TG6 in GA as opposed to TG2 in CD. In those sera
that reacted with both antigens, a greater concentration
of TG6 was required for blocking reactivity of sera
from GAE than from CD patients in the TG6 ELISA,
whereas the opposite was true for TG2 concentrations
in the TG2 ELISA (see Fig 3F). However, a strict correlation between organ involvement and serum antibody specificity cannot be made because 2 of 20 CD
patients tested exclusively positive for TG6 and TG3
Table 3. Correlation between Anti-transglutaminase Antibodies and Endomysial Reactivity in Gluten Ataxia
Patient Group
TG2 and
TG2 & TG6
IgA or IgG
11/15 (73%)
0/17 (0%)
4/15 (27%)
17/17 (100%)
11/11 (100%)
1/3 (33%)
0/17 (0%)
2/3 (67%)
8/17 (47%)
3/3 (100%)
11/17 (65%)
Not known for two gluten ataxia only (GAo) patients.
EMA ⫽ endomysial antibody; TG ⫽ transglutaminase; GAE ⫽ gluten ataxia with enteropathy.
IgA, respectively, whereas 4 of 28 GA patients with
anti-TG antibodies had detectable levels of only TG2
IgA or IgG, or both. There was no correlation between
anti-TG6 antibody titers and either disease severity or
likelihood of finding further autoantibodies recognizing
other TG isozymes. We do not know as yet whether
those patients presenting with CD who have autoantibodies to TG6 are more likely to experience development of neurological dysfunction. The development of
extraintestinal manifestations may be prevented in such
patients because the classic presentation of CD with
gastrointestinal symptoms is likely to result in early diagnosis and gluten avoidance.
Autoantibodies against TG2 are responsible for the
EMA, reticulin, and jejunal antibody tissue binding of
serum samples from CD patients.41 Antibody libraries
established from peripheral and intestinal lymphocytes
suggest that the humoral response against TG2 is locally restricted to the intestinal mucosa,42 and a recent
retrospective study showed that, even in seronegative
CD patients, autoantibodies against TG2 are deposited
in the gut.43 The central role of the gut in disease development also in GA is indicated by the consistent
demonstration of IgA deposits in the jejunum even in
the absence of a histologically discernible enteropathy
on biopsy and absence of EMA recognition by sera.16
Seronegativity of GAo patients in conventional tests
appears to reflect the preferential development of immunoglobulins specific for TG6 and absence of antiTG2 IgA in particular (see Table 3). A humoral response to a different TG isozyme may also explain the
reported absence of TG2 antibodies in a proportion of
CD patients43 and is supported by our finding of high
serum titers for anti-TG3 and anti-TG6 IgA, respectively, in two such patients.
The cause of the diverse manifestations of gluten
sensitivity remains obscure. Variations in the specificity
of antibodies produced in individual patients, from selectivity for a particular TG2 conformation12 to crossreactivity between TG isozymes,18 could explain a wide
spectrum of manifestations. However, most patients
with gluten sensitivity were shown to have antibodies
that target multiple epitopes of TG2,42 and consider-
ing protein homology alone, one would expect to find
antibodies cross-reacting with further TG isozymes including TG5 and TG7, but thus far we have not been
able to identify such antibodies. It is also surprising
that in GA and DH, IgA deposits accumulate in the
periphery of vessels in a locale in the tissue where in
health TG6 or TG3 are absent but become abundant
in disease (see Fig 4).16,18 This could indicate that either the deposits originate from immune complexes
formed in the circulation or TG6/TG3 is derived from
or its synthesis induced by infiltrating inflammatory
cells prior to deposit formation. It is currently unclear
whether TG2 is the only isozyme involved in modification of gliadin peptides, and whether TG2 and the
gluten peptides intersect before uptake by antigenpresenting cells, or deamidation occurs at the cell surface or in the endocytic pathway of antigen-presenting
cells. The lack of antibodies cross-reactive with different TG isozymes in most patients, as well as the identification of patients with a response exclusively directed to TG6 or TG3, make epitope spreading less
likely the cause for immune responses to other TGs
and point to the possibility that TG isozymes other
than TG2 can be the primary antigen. Nevertheless,
gluten dependence of disease and antibody production
implicates the small intestine as the origin, independent of subsequent clinical manifestation.24,25 Unlike
TG2, which is expressed in many cell types in the intestinal environment, TG3 and TG6 are essentially absent from the small intestine but can be expressed in
mucosal antigen-presenting cells (D.A., unpublished
results). Expression of these TG isozymes in lamina
propria macrophages or dendritic cells could explain
their involvement. However, despite evidence for endocytosis of TG2 by in vitro differentiated monocytes
and intersection with the major histocompatibility
complex class II pathway, a link to uptake and deamidation of gliadin peptides could not be established in a
recent study.44
IgA deposition in brain vessels16 may indicate that
vasculature-centered inflammation may compromise the
blood–brain barrier, allowing exposure of the CNS to
the pathogenic antibodies, and, therefore, may be the
Hadjivassiliou et al: Autoantibodies to TG6
Fig 3. Remaining IgA reactivity of sera from patients with celiac disease (CD) or gluten ataxia (GA) after preincubation with
transglutaminase 2 (TG2; triangles) or TG6 (circles). Sera were incubated with buffer only or a series of concentrations of the
indicated type of TG (horizontal axis) and the remaining antibody reactivity for TG2 or TG6 subsequently evaluated by enzymelinked immunosorbent assay (ELISA). Reactivity relative to buffer control is shown on the vertical axis (mean ⫾ standard deviation). (A–E) Examples of inhibition curves for sera of a patient with gluten ataxia with enteropathy (GAE) (A, B), celiac disease
(CD) (C, D), or gluten ataxia without enteropathy (GAo) (E) for reactivity with TG6 (A, C, E) or TG2 (B, D), respectively.
Note that no reactivity with TG2 could be detected for the GAo patient. For Patient GAE43, TG6 competed effectively for binding
in the TG2 ELISA (B, dotted line) indicating cross-reactivity of TG6 antibodies with TG2 (dashed line accounts for the effect of
cross-reacting TG6 antibodies on TG2 inhibition in TG2 ELISA in B). No cross-reactivity of antibodies could be detected in Patient CD11 (C, D). (F) Inhibitory effect of 22.4␮g/ml TG2 or TG6 in TG2 or TG6 ELISAs, respectively, in patient sera that
displayed reactivity with either antigen. Cross-reaction of antibodies could be seen in only a few sera.
Annals of Neurology
Vol 64
No 3
September 2008
Fig 4. Perivascular transglutaminase 6 (TG6) deposits are present in the cerebellum of a gluten ataxia (GA) patient. Cryosections of
normal cerebellum (A–C) and gluten ataxia without enteropathy (GAo) cerebellum (D, E) or parietal lobe (F) were stained with
hematoxylin and eosin (A) or incubated with antibodies to TG2 (B) or TG6 (C–F), and binding of primary antibodies was subsequently visualized using peroxidase-conjugated secondary antibodies. Tissue structures in cerebellum are indicated in (A) molecular
layer (SM), granular layer (SG), and white matter (WM). TG2 is expressed specifically in endothelial and vascular smooth muscle
cells of capillaries and larger vessels (arrows) within the cerebellum itself and in the meninges (arrowhead) (B), whereas TG6 cannot be detected in vascular structures in normal brain (C). In GA, TG6 accumulates in the periphery of vessels (arrows) in the
cerebellum (D, E) but not in the parietal lobe (F). (E) The area framed in (D) is shown at higher magnification. Colocalization
of perivascular TG6 (tetrarhodamine isothiocyanate) (H) and IgA (fluorescein isothiocyanate) (G; asterisk indicates blood vessel) is
indicated by areas of color mixing in image (I) showing superimposition of the respective optical sections acquired by confocal microscopy. Bar in panel A ⫽ 300␮m (A–D) and 120␮m (E, F), and bar in panel G ⫽ 50␮m (G–I).
trigger of nervous system involvement. Indeed, TG2 expressed by smooth muscle and endothelial cells in noninflamed brain is an abundant component of the blood–
brain barrier,34 and autoantibody binding could
precipitate inflammation. However, TG6 that is not
normally expressed in the vasculature was found to colocalize with the perivascular IgA deposits in postmortem examination of a GA patient (see Fig 4), suggesting
that deposition of characteristic antibody complexes
could involve targeting of TG6 instead. Furthermore,
IgG class antibodies were present in only 60% of CD
patients, whereas in GAO/GAE patients positive for
anti-TG Ig, the prevalence rate was 90%, and of those
patients, 21% were negative for IgA. This shift from IgA
to IgG may reflect the target organ involved (cerebellum
vs small bowel) and the presence of a blood–brain barrier that is impermeable to IgA but not to IgG. However, intrathecal production of at least anti-TG2 antibodies has been demonstrated,27 indicating the
possibility of a cell-mediated mechanism instead of passive transfer. It has recently been suggested that immune
surveillance in the CNS involves the infiltration of
tissue-committed, primed CD4⫹ T cells specific for the
gut into the cerebrospinal fluid,45 and that these cells are
at the center of the intrathecal immune response, which
could explain CNS involvement.
It could be argued that development of anti-TG antibodies in GA is coincidental, that is, not involved in
the pathophysiological process. One method to demonstrate the pathological effect of an antibody is the
passive transfer of the disease through antibody injection. Although for only few antibodies such experimental evidence exists,46 IgG fractions of patients with cerebellar ataxia and stiff person syndrome have been
shown to compromise motor function and impair
learning in rodents, an effect ascribed to antibodies
against glutamic acid decarboxylase and amphiphysin.47,48 However, previous studies have not answered
whether it is these specific antibodies or other autoantibodies in the IgG fraction of patient sera that cause
neuronal damage. We have recently shown that serum
immunoglobulin from GA patients, as well as clonal
anti-TG immunoglobulins derived using phage display,
cause deficits in motor coordination but not anxiety
when injected intraventricularly in mice.49,50 The fact
that isolated anti-TG immunoglobulins induce dramatic ataxia-like deficits in mice indicates selective
neurotoxicity of anti-TG antibodies once exposed to
the CNS. This is consistent with the selective loss of
Purkinje cells in ataxia patients and with a unique pattern of reactivity of GA sera toward Purkinje cells
when applied to brain sections. Although these data
implicate anti-TG antibodies in ataxia, they do not explain the spectrum of distinct neurological deficits currently ascribed to gluten sensitivity or why only a fraction of patients with circulating anti-TG antibodies are
Annals of Neurology
Vol 64
No 3
September 2008
affected. Differences in specificity of pathological
anti-TG antibodies in sera may contribute to variability
in manifestation within the CNS, but we also cannot
exclude that further antigens are targeted. Alternatively,
antibody-mediated neuronal damage may be a secondary consequence to localized disruption of the integrity
of the blood–brain barrier and, therefore, lead to distinct clinical pictures depending on the area of the
CNS affected.51
In summary, we show that high-affinity antibodies
to TG6 characterize a group of patients with otherwise
idiopathic sporadic ataxia. Our findings offer a plausible pathophysiological mechanism that explains the diversity of manifestations seen in the context of gluten
sensitivity and provide a more specific marker for patients who may benefit from a gluten-free diet.
This study was supported by grants from the Bardhan Research and
Education Trust of Rotherham (Barnsley, UK) and the Ryder
Briggs Trust (Sheffield, UK) to MH, NW, DS, and DA.
We are grateful to M. Langley for technical assistance.
1. Sollid LM. Coeliac disease: dissecting a complex inflammatory
disorder. Nat Rev 2002;2:647– 655.
2. Dietrich W, Ehnis T, Bauer M, et al. Identification of tissue
transglutaminase as the autoantigen of celiac disease. Nat Med
1997;3:797– 801.
3. Aeschlimann D, Thomazy V. Protein crosslinking in assembly
and remodelling of extracellular matrices: the role of transglutaminases. Connect Tissue Res 2000;41:1–27.
4. Boros S, Ahrman E, Wunderink L, et al. Site-specific transamidation and deamidation of the small heat-shock protein Hsp20
by tissue transglutaminase. Proteins 2006;62:1044 –1052.
5. Fleckenstein B, Molberg Ø, Qiao SW, et al. Gliadin T cell
epitope selection by tissue transglutaminase in celiac disease.
Role of enzyme specificity and pH influence on the transamidation versus deamidation process. J Biol Chem 2002;277:
34109 –34116.
6. Molberg O, McAdam SN, Körner R, et al. Tissue transglutaminase selectively modifies gliadin peptides that are recognized by
gut-derived T cells in celiac disease. Nat Med 1998;4:713–717.
7. Van de Wal Y, Kooy Y, van Veelen P, et al. Selective deamidation by tissue transglutaminase strongly enhances gliadinspecific T cell reactivity. J Immunol 1998;161:1585–1588.
8. Fleckenstein B, Qiao SW, Larsen MR, et al. Molecular characterization of covalent complexes between tissue transglutaminase
and gliadin peptides. J Biol Chem 2004;279:17607–17616.
9. Dietrich W, Esslinger B, Trapp D, et al. Crosslinking to tissue
transglutaminase and collagen favours gliadin toxicity in celiac
disease. Gut 2006;55:478 – 484.
10. Sollid LM, Jabri B. Is celiac disease an autoimmune disorder?
Curr Opin Immunol 2005;17:595– 600.
11. Korponay-Szabó IR, Vecsei Z, Király R, et al. Homology of
deamidated peptides with tissue transglutaminase. Proceedings of the 9th International Conference of TG Protein
Crosslinking; 2007; Marrakech, Morocco.
12. Pinkas DM, Strop P, Brunger AT, Khosla C. Transglutaminase
2 undergoes a large conformational change upon activation.
PLoS Biol 2007;5:e327.
13. Kaukinen K, Halme L, Collin P, et al. Celiac disease in patients
with severe liver disease: gluten-free diet may reverse hepatic
failure. Gastroenterology 2002;122:881– 888.
14. Sugai E, Cherñavsky A, Pedreira S, et al. Bone-specific antibodies
in sera from patients with celiac disease: characterization and implications in osteoporosis. J Clin Immunol 2002;22:353–362.
15. Korponay-Szabó IR, Halttunen T, Szalai Z, et al. In vivo targeting of intestinal and extraintestinal transglutaminase 2 by
coeliac autoantibodies. Gut 2004;53:641– 648.
16. Hadjivassiliou M, Mäki M, Sanders DS, et al. Autoantibody
targeting of brain and intestinal transglutaminase in gluten
ataxia. Neurology 2006;66:373–377.
17. Marks J, Shuster S, Watson AJ. Small bowel changes in dermatitis herpetiformis. Lancet 1966;2:1280 –1282.
18. Sárdy M, Kárpáti S, Merkl B, et al. Epidermal transglutaminase
(TGase3) is the autoantigen of dermatitis herpetiformis. J Exp
Med 2002;195:747–757.
19. Hadjivassiliou M, Grünewald RA, Davies-Jones GA. Gluten
sensitivity as a neurological illness. J Neurol Neurosurg Psychiatry 2002;72:560 –563.
20. Hadjivassiliou M, Grunewald RA, Kandler RH, et al. Neuropathy associated with gluten sensitivity. J Neurol Neurosurg Psychiatry 2006;77:1262–1266.
21. Zelnik N, Pacht A, Obeid R, Lerner A. Range of neurologic
disorders in patients with celiac disease. Pediatrics 2004;113:
22. Hadjivassiliou M, Grünewald RA, Chattopadhyay AK, et al.
Clinical, radiological, neurophysiological, and neuropathological
characteristics of gluten ataxia. Lancet 1998;352:1582–1585.
23. Hadjivassiliou M, Sanders DS, Woodroofe N, et al. Gluten
ataxia. Cerebellum (in press).
24. Pellecchia MT, Scala R, Perretti A, et al. Cerebellar ataxia associated with subclinical celiac disease responding to gluten-free
diet. Neurology 1999;53:1606 –1608.
25. Hadjivassiliou M, Davies-Jones GAB, Sanders DS, Grünewald
RA. Dietary treatment of gluten ataxia. J Neurol Neurosurg
Psychiatry 2003;74:1221–1224.
26. Bürk K, Bösch S, Müller CA, et al. Sporadic cerebellar ataxia
associated with gluten sensitivity. Brain 2001;124:1013–1019.
27. Hadjivassiliou M, Grünewald R, Sharrack B, et al. Gluten
ataxia in perspective: epidemiology, genetic susceptibility and
clinical characteristics. Brain 2003;126:685– 691.
28. Sander HW, Magda P, Chin RL, et al. Cerebellar ataxia and
coeliac disease. Lancet 2003;362:1548.
29. Bürk K, Melms A, Schulz JB, Dichgans J. Effectiveness of intravenous immunoglobin therapy in cerebellar ataxia associated
with gluten sensitivity. Ann Neurol 2001;50:827– 828.
30. Aeschlimann D, Grenard P. Novel transglutaminase gene products. US Patent 7,052,890; 2006.
31. Grenard P, Bates MK, Aeschlimann D. Evolution of transglutaminase genes: identification of a gene cluster on human chromosome 15q15. J Biol Chem 2001;276:33066 –33078.
32. Moralejo P, Egan SM, Hidalgo E, Aguilar J. Sequencing and
characterization of a gene cluster encoding the enzymes for
L-rhamnose metabolism in Escherichia coli. J Bacteriol 1993;
33. Stephens P, Grenard P, Aeschlimann P, et al. Crosslinking and
G-protein functions of transglutaminase 2 contribute differentially to fibroblast wound healing responses. J Cell Sci 2004;
117:3389 –3403.
34. Aeschlimann D, Paulsson M. Cross-linking of laminin-nidogen
complexes by tissue transglutaminase: a novel mechanism for
basement membrane stabilization. J Biol Chem 1991;266:
15308 –15317.
35. Engel J, Schalch W. Determination of antibody binding constants from Farr test and other radioimmunoassay. Mol Immunol 1980;17:675– 680.
36. Aeschlimann D, Wetterwald A, Fleisch H, Paulsson M. Expression of tissue transglutaminase in skeletal tissues correlates with
events of terminal differentiation of chondrocytes. J Cell Biol
37. Sárdy M, Csikós M, Geisen C, et al. Tissue transglutaminase
ELISA positivity in autoimmune disease independent of glutensensitive disease. Clin Chim Acta 2007;376:126 –135.
38. Villalta D, Crovatto M, Stella S, et al. False positive reactions
for IgA and IgG anti-tissue transglutaminase antibodies in liver
cirrhosis are common and method-dependent. Clin Chim Acta
39. Kaukinen K, Collin P, Laurila K, et al. Resurrection of gliadin
antibodies in coeliac disease. Deamidated gliadin peptide antibody test provides additional diagnostic benefit. Scand J Gastroenterol 2007;42:1428 –1433.
40. Karell K, Louka AS, Moodie SJ, et al. HLA types in celiac
disease patients not carrying the DQA1*05-DQB1*02 (DQ2)
heterodimer: results from the European Genetics Cluster on
Celiac Disease. Hum Immunol 2003;64:469 – 477.
41. Korponay-Szabó IR, Laurila K, Szondy Z, et al. Missing endomysial and reticulin binding of coeliac antibodies in transglutaminase 2 knockout tissues. Gut 2003;52:199 –204.
42. Sblattero D, Florian F, Azzoni E, et al. The analysis of the fine
specificity of celiac disease antibodies using tissue transglutaminase fragments. Eur J Biochem 2002;269:5175–5181.
43. Salmi TT, Collin P, Korponay-Szabó IR, et al. Endomysial
antibody-negative coeliac disease: clinical characteristics and intestinal autoantibody deposits. Gut 2006;55:1746 –1753.
44. Ráki M, Tollefsen S, Molberg Ø, et al. A unique dendritic cell
subset accumulates in the celiac lesion and efficiently activates
gluten-reactive T cells. Gastroenterology 2006;131:428 – 438.
45. Kivisäkk P, Tucky B, Wei T, et al. Human cerebrospinal fluid
contains CD4⫹ memory T cells expressing gut- or skin-specific
trafficking determinants: relevance for immunotherapy. BMC
Immunol 2006;7:14.
46. Lang B, Dale RC, Vincent A. New autoantibody mediated disorders of the central nervous system. Curr Opin Neurol 2003;
47. Sommer C, Weishaupt A, Brinkhoff J, et al. Paraneoplastic stiff
person syndrome: passive transfer to rats by means of IgG antibodies to amphiphysin. Lancet 2005;365:1406 –1411.
48. Manto MU, Laute MA, Aguera M, et al. Effects of antiglutamic acid decarboxylase antibodies associated with neurological diseases. Ann Neurol 2007;61:544 –551.
49. Boscolo S, Sarich A, Lorenzon A, et al. Gluten ataxia: passive
transfer in a mouse model. Ann N Y Acad Sci 2007;1107:
319 –328.
50. Boscolo S, Sarich A, Lorenzon A, et al. Anti-transglutaminase
antibodies cause ataxia in mice. (submitted).
51. Kowal C, Degiorgio LA, Lee JY, et al. Human lupus autoantibodies against NMDA receptors mediate cognitive impairment.
Proc Natl Acad Sci USA 2006;103:19854 –19859.
Hadjivassiliou et al: Autoantibodies to TG6
Без категории
Размер файла
496 Кб
ataxia, autoantibodies, recognize, neuronal, transglutaminase, novem, gluten
Пожаловаться на содержимое документа