- Email: [email protected]
Characterisation of tissue transglutaminase-reactive T cells from patients with coeliac disease and healthy controls
Clinical Immunology (2014) 154, 155–163
available at www.sciencedirect.com
Clinical Immunology www.elsevier.com/locate/yclim
Characterisation of tissue transglutaminase-reactive T cells from patients with coeliac disease and healthy controls Ross Comerford a,b,c,⁎,1 , Christian Coates b,c,1 , Greg Byrne d , Sara Lynch d , Padraic Dunne b , Margaret Dunne b , Jacinta Kelly a , Conleth Feighery b,c a
National Children's Research Centre, Our Lady's Hospital For Sick Children, Crumlin Dublin 12, Ireland Institute of Molecular Medicine, Trinity College Dublin Dublin 8, Ireland c Department of Immunology, St James's Hospital, Dublin 8, Ireland d Dublin Institute of Technology, Kevin Street, Dublin 8, Ireland b
Received 28 March 2014; accepted with revision 1 August 2014 KEYWORDS Coeliac disease; Tissue transglutaminase; Autoimmunity; T cells; Interferon-γ; Interleukin-17A
Abstract Previous studies have shown evidence for T lymphocytes specific for tissue transglutaminase (tTG) in the periphery of coeliac disease (CD) patients. These cells could play a role in disease pathogenesis and may be involved in providing help for the production of anti-tTG autoantibodies. The objective of this study was to further investigate the presence of tTG-specific T cells in patients with treated and untreated CD, and normal controls. Positive proliferative responses to three different commercial tTG antigens were detected in all groups tested, occurring more frequently and at higher levels in untreated CD patients. The addition of antibodies to HLA-DQ and HLA-DR caused a significant reduction in the proliferative response to tTG. T cell lines specific for tTG and composed predominantly of CD4-positive T cells were generated from responsive CD and control individuals, and were found to produce large amounts of interferon-γ, as well as interleukins 10, 17A, and 21. © 2014 Elsevier Inc. All rights reserved.
1. Introduction ⁎ Corresponding author at: Department of Immunology, St James's Hospital, Dublin 8, Ireland. E-mail address: [email protected] (R. Comerford). 1 These authors contributed equally to this work.
http://dx.doi.org/10.1016/j.clim.2014.08.001 1521-6616/© 2014 Elsevier Inc. All rights reserved.
Coeliac disease (CD) is a common inflammatory disease of the small intestine caused by an inappropriate immune response to wheat gluten in genetically predisposed individuals. Investigation into the pathogenesis of coeliac disease
156
R. Comerford et al.
reveals a complex interplay between environmental factors, genetics, the adaptive and innate immune systems, and the presence of autoantigens, in a process which has still not been fully elucidated. The principal autoantigen of coeliac disease is the ubiquitously expressed, multifunctional enzyme tissue transglutaminase (tTG) [1]. The demonstration of IgA autoantibodies directed to this self-protein is an integral component in the diagnosis of coeliac disease [2]. Tissue transglutaminase is also intimately involved in coeliac disease pathogenesis through the modification of gluten peptides by deamidation, which facilitates their presentation to the immune system via HLA-DQ2 or DQ8 molecules [3,4]. Due to the lymphocytic infiltration observed in the active CD intestinal lesion [5], and the strong association of the disease with HLA molecules [6], the concept of T-cell involvement in the pathogenesis of CD had been long-established prior to the isolation of gliadin-specific T cells from the CD intestinal mucosa [7]. In the time since this important observation, the CD4+ gliadin-specific T-cell response has been dissected and extensively characterised in relation to antigen recognition, phenotype, and function. The hallmark of gliadin-specific T-cell lines and clones is a mucosal-homing phenotype [8,9], and production of large amounts of IFN-γ, a cytokine that contributes substantially to the intestinal destruction seen in CD [10–12]. Apart from their role in the immune reaction that leads to the formation of the CD lesion [13], gliadin-specific T cells have been speculated to provide T cell help to tTG-specific B cells, resulting in the anti-tTG response characteristic of CD. This ‘hapten-carrier’ theory, proposed by Sollid, was based upon the fact that T cells specific for tTG had never been isolated, and the gliadin-dependant nature of the anti-tTG response [14]. However, preliminary findings from our group indicated that T cells specific for tTG could be detected in CD patients and in some control individuals, in a HLA-DQ and HLA-DR restricted manner [15]. The existence of tTG-specific T cells was confirmed in a recent paper by Ciccocioppo et al., who described the presence of T cells that were mostly CD4+ and proliferated to tTG in a HLA-DQ2-restricted manner in the periphery of untreated CD patients [16]. The aim of this study was to further investigate the presence of tTG-specific T cells in the periphery of CD patients and healthy controls, as it is known that T cells specific for other autoantigens have been detected in healthy subjects [17,18]. Peripheral blood samples were challenged with tTG and proliferative responses measured. tTG-reactive T cell
lines were generated from responsive individuals and both the intracellular and secreted cytokines IFN-γ, IL-10, IL-17A, and IL-21 were measured, in tandem with the measurement of proliferation of these cells.
2. Materials and methods 2.1. Patients and controls CD patients and controls were recruited from the Departments of Gastroenterology and Immunology, St James's Hospital Dublin. Ethical approval for this study was granted from the hospital ethics committee. The diagnosis of CD was based on a typical histological lesion, positive serology (IgA anti-tTG and endomysial (EMA) antibodies) and positive histological, serological and clinical responses to a gluten free diet. Patients with CD were sub-divided according to their treatment status. This included: 33 patients with untreated CD who were taking a normal diet and with positive serological tests; 65 patients with treated CD, taking a gluten free diet and with negative or low positive serological tests. In addition, 54 healthy control subjects, with negative coeliac serology, were investigated. The demographic information of each study group is summarised in Table 1.
2.2. Proliferation assays Fresh peripheral blood mononuclear cells (PBMCs) from CD patients or control individuals were separated by density gradient centrifugation using Lymphoprep™ solution (AxisShield), and adjusted to a concentration of 5 × 105/ml in T cell medium (TCM) which contained RPMI containing HEPES, sodium pyruvate, non-essential amino acids, essential amino acids, penicillin, streptomycin, fungizone, L-glutathione, β-mercaptoethanol, and 5% autologous serum. Two hundred microlitres per well of this suspension was then cultured in triplicate with the appropriate antigen in a round-bottomed 96-well microtitre plate. Wells containing medium and cells only were added in order to measure background proliferation, and used to calculate the stimulation index (SI). The tTG antigens used were guinea pig tTG (gp tTG, Sigma), erythrocyte tTG (tTGery, Inova Diagnostics), recombinant human tTG (rh tTG, produced in SF9 insect cells, Zedira), at a concentration of 10 μg/ml. Phytohaemagglutinin (PHA) and purified protein
Table 1 Characteristics of the groups studied in proliferation assays using various commercial sources of tissue transglutaminase (tTG), with resulting responses and stimulation indices. Antigen
Study group
Number
Sex (M/F)
Mean age (years) (range)
No. positive (%)
Mean SI (range)
Mean SI responders (range)
Guinea-pig tTG
CD untreated CD treated Controls CD untreated CD treated Controls CD untreated CD treated Controls
15 30 29 11 35 25 7 15 10
2/13 7/23 11/18 4/7 14/21 12/13 2/5 4/11 3/7
45 (20–70) 59 (19–68) 29 (18–54) 48 (22–65) 58 (29–83) 25 (24–65) 42 (20–59) 52 (29–77) 29 (24–57)
12 (80) 12 (40) 11 (38) 5 (45) 11 (31) 6 (24) 6 (86) 6 (40) 1 (10)
4.7 (1.1–38.1) 3.1 (0.8–13.9) 2.4 (0.7–20.6) 5.0 (0.4–24.8) 1.6 (0.5–9.2) 2.1 (0.7–7.6) 19.1 (0.2–57) 2.2 (0.1–6.2) 1.0 (0.4–2.1)
8.5 (3.0–38.1) 4.8 (2.1–13.9) 5.9 (2.1–20.6) 9.3 (2.1–24.8) 2.8 (2.2–9.2) 5.0 (3.4–7.6) 22.8 (4.2–57) 4.0 (2.1–6.2) 2.0 (n/a)
Erythrocyte tTG
Recombinant htTG
Characterisation of tissue transglutaminase-reactive T cells from patients derivative (PPD) of Mycobacterium tuberculosis at concentrations of 2 and 10 μg/ml, respectively, were used as proliferative control stimuli. Cultures were then incubated at 37 °C, in the presence of 5% CO2 and a humidified atmosphere, for 6 days. For the last 18 h of culture 0.5 μCi of 3H-thymidine (PerkinElmer) was added to each well and cells were harvested on day 6. Proliferation was assessed by measuring the thymidine incorporation using a MicroBeta Trilux scintillator (PerkinElmer). The stimulation index (SI) was calculated by dividing the mean of the three cpm values for each antigen by that of the unstimulated wells, with an SI of 2, or greater, being considered positive. In experiments measuring the proliferation of tTG-sensitised T-cell lines, the incubation time was shortened to 72 h. MHC restriction was investigated using PBMCs from 16 of the coeliac patients and 2 normal controls. PBMCs were cultured with gp tTG in the presence or absence of (i) a γ2A anti-HLA-DR (L243, Becton and Dickinson); (ii) a γ2A non-specific isotype matched control antibody (Becton and Dickinson); (iii) a γ1 anti-HLA-DQ (SPV-L3; a gift from Dr. Hergen Spits, University of Amsterdam, Amsterdam, The Netherlands), or (iv) a γ1 non-specific isotype matched control antibody (Becton and Dickinson). Each of the commercially prepared antibodies was dialysed in phosphate buffered saline to remove preservatives. Proliferation was measured as described above and results were calculated on the basis of percentage inhibition.
157
and PPD, were measured on days 14 and 28. The cells were starved of IL-2 for 5 days prior to re-stimulation. Irradiated autologous PBMCs pulsed with each test antigen were prepared, with 1 × 105 APCs added to 1 × 105 of the IL-2-starved cells from the cell line in a volume of 200 μl in a 96-well round-bottomed microtitre plate. These suspensions were then returned to culture for a further 72 h, with proliferation then measured by 3 H-thymidine incorporation. Cellular phenotype (CD4 or CD8 positivity) as well as intracellular cytokine production was determined by flow cytometric analysis. Cultures for intracellular cytokine staining had 10 ng/ml phorbol 12-myristate 13-acetate (Sigma), 1 μg/ml ionomycin (Sigma), and 10 μg/ml Brefeldin A (Sigma) added for the final 10 h of culture.
2.5. Flow cytometry Intracellular cytokine production, and the phenotype of samples of T cells from each T-cell line in response to re-stimulation with tTG was assessed. The panel of antibodies used was CD3-APC-efluor780, CD4-APC, CD8-PECy5, and IFN-γ-FITC (eBioscience, San Diego, CA). Cells were gated on forward/side scatter, CD3, and aqua fluorescent reactive dye (Invitrogen) exclusion, which permitted the measurement of live cells only. Quadrants for all parameters were set using fluorescence minus one (FMO) controls. For all flow cytometric analysis, 5 × 105 cells from each sample were measured.
2.3. T-cell line generation For the generation of tTG-specific T-cell lines, 1 × 106 PBMCs from CD patients or control individuals in TCM was incubated with tTGery at 10 μg/ml. Wells containing 1 × 106 PBMCs only were cultured as a negative control. After 5 days in culture, 20 IU/ml of recombinant human IL-2 (Sigma) was added to each well. On day 7, wells were visually examined for growth by comparing cellular numbers in stimulated and unstimulated wells using an inverted microscope, and observing a colour change from pink to yellow in the culture media. The cells were then cultured for a further 7 days, receiving 20 IU/ml IL-2 on day 9, examined on the inverted microscope, and counted. In order to expand the T-cell lines, further stimulation with tTGery was performed on days 14 and 28 of culture, using irradiated autologous PBMCs as antigen-presenting cells. The APCs were incubated with tTGery at a concentration of 10 μg/ml in TCM, and added to the tTG-stimulated cell line at a ratio of 1:1. Twenty International Units per millilitres IL-2 was added to each culture on day 14, and weekly thereafter. The cells were counted again on day 28, with the concentration adjusted to 1 × 106/ml, and re-plated in a 24-well plate prior to re-stimulation. All incubations were performed at 37 °C, in the presence of 5% CO2 and a humidified atmosphere. At weekly intervals, 500 μl of the culture supernatant was removed and stored at −20 °C, being replaced with 500 μl of TCM.
2.6. Measurement of cytokines by ELISA The supernatants removed weekly from each T-cell line and media only controls were analysed for the presence of IFN-γ, IL-10, IL-4, IL-17A, and IL-21 using ELISA MAX™ kits (BioLegend, San Diego, CA). Assays were performed as per the manufacturer's instructions, with all measurements being made in triplicate.
2.7. Statistical analyses Differences in proliferation in response to tTG between patient and control groups were evaluated for statistical significance using the Mann–Whitney test, with the paired t test used to identify significant effects of MHC blocking on proliferative responses to tTG. The level of significance was set at 0.05. The GraphPad Prism 5 program was used for all statistical analysis.
3. Results 3.1. Proliferation assays
2.4. Re-stimulation of T-cell lines To demonstrate the antigen-specificity of the cell lines generated, the proliferative capacity and cytokine responses to stimulation with tTGery, rh tTG, and the control stimuli PHA
Positive proliferative responses to gp tTG were observed in 80% of untreated coeliac patients, compared to 40% of treated coeliac patients and 38% of normal control subjects (Fig. 1a, Table 1). Furthermore, the overall SI response of
158
R. Comerford et al.
Figure 1 Proliferative responses to guinea pig, erythrocyte, and recombinant human tissue transglutaminases. Proliferation of PBMCs from normal controls (NC), untreated coeliac disease patients (UTCD), and treated coeliac disease patients (TCD) to guinea pig (a), erythrocyte (b), and recombinant (c) human tissue transglutaminases. For all experiments, proliferation was measured by 3H-thymidine incorporation after six days of culture (SI = stimulation index, the dashed line indicates the cut-off for positivity). (*p b 0.05, **p b 0.005).
the untreated coeliac patients was significantly higher than both the treated coeliac patients (p = 0.0042, Mann– Whitney test) and the normal control subjects (p = 0.0063, Mann–Whitney test). For the tTGery antigen, proliferation was again more frequently detected in untreated CD patients (45%), with 35% of treated CD patients, and 24% of controls responding (Fig. 1b, Table 1). A similar pattern of response to the recombinant human tTG preparation (rh tTG) was noted, with positivity detected in the majority of untreated CD patients (Fig. 1c, Table 1). For untreated CD patients, the overall SI response to rh tTG was significantly higher than both the treated CD (p = 0.009, Mann–Whitney test) and the normal control subjects (p = 0.0065, Mann– Whitney test). In experiments comparing the proliferative stimulus of gp tTG and tTGery, similar levels of proliferation were observed in CD and control subjects, irrespective of the source of antigen. In a group of ten untreated CD patients the three responders to gp tTG also had a corresponding response to tTGery, a finding replicated in three of four responders in a control group comprised of ten individuals (data not shown). Of the fifteen treated CD patients tested with both the tTGery and rh tTG antigens, two proliferated to tTGery only, one to rh tTG only, and four proliferated to both preparations. Ten control individuals were tested with both commercial tTG antigens, two were positive for tTGery only, and one positive for rh tTG only, with no individual proliferating to both antigens. Substitution of any of the commercial tTG sources with our own in-house recombinant tTGs [19] resulted in equivalent proliferative responses, in both CD patients and controls (data not shown).
3.2. MHC blocking studies In MHC blocking studies, the inhibitory effect of addition of anti-HLA-DR and HLA-DQ to gp tTG stimulated PBMCs was investigated in 18 individuals (16 coeliac and 2 normal control subjects) with known proliferative responses to this antigen (Fig. 2). The effect of anti-MHC II antibodies was contrasted with that of isotype matched controls. Marked inhibition of proliferation was noted with both anti-MHC II antibodies, in particular with anti-HLA-DR, whereas the control antibodies
Figure 2 Inhibitory effect of addition of anti-HLA-DR/-DQ. Proliferation of PBMCs from 16 coeliac and 2 control subjects to guinea pig tTG could be blocked by the addition of antibodies to HLA-DR or HLA-DQ, an effect not replicated by the isotype control antibodies gamma 2A or gamma 1. The grey lines indicate the mean; the dashed line represents the cut-off for positivity (*p b 0.05, **p b 0.005).
Characterisation of tissue transglutaminase-reactive T cells from patients caused no significant inhibition. In 16/18 individuals, anti-HLA-DR caused a mean inhibition of 79.8% (range 20– 97%). There was a significant difference between the response seen in the anti-HLA-DR cultures (median SI 0.4 ± 0.1) and the cultures containing the γ2A isotype matched control antibody (median SI 3.4 ± 1.5) (p = 0.007, Paired t test). The addition of anti-HLA-DQ caused an inhibition of proliferation in 13/18 subjects (mean inhibition 48.5%, range 10–89%). There was a significant difference between the response seen in the anti-HLA-DQ cultures (median SI 1.28 ± 0.5) and the cultures containing the γ1 isotype matched control antibody (median SI 3.2 ± 1.0) (p = 0.002, paired t test).
3.3. T-cell line generation Three polyclonal T-cell lines were generated by initial stimulation of PBMCs with tTGery, and fortnightly re-stimulation of the cell line thereafter. The T-cell line CD01 was generated from a male of 51 years who had biopsy-confirmed CD, and was on a strict gluten-free diet for ten years, as evidenced by serial tTG/EMA negativity. The T-cell lines CT01 and CT02 were generated from two non-CD controls, males of 32 and 56 years, respectively. All three individuals were selected because they gave a positive proliferative response to tTGery in 3H-thymidine incorporation assays. The T-cell lines established were predominantly CD4+, the percentage of which increased over the duration of the culture period for CD01 and CT01 (Fig. 3b). In order to confirm the antigen-specificity of the T-cell lines, re-stimulation with tTGery, was performed, with responsiveness measured by 3H-thymidine incorporation. At each re-stimulation (days 14 and 28), a proliferative response to tTGery was detected, with the strongest response being the second stimulation for each cell line (Fig. 3). Although the cell lines had been established using tTGery, after the first stimulation samples of cells from each cell line had the ability to proliferate in response to the rh tTG antigen, at levels equivalent to tTGery, and PHA, which was used as a positive
159
control (data not shown). After the initial stimulation each T-cell line lost responsiveness to the protein antigen PPD, indicating a positive selection of tTG-specific cells during culture (Fig. 3).
3.4. Intracellular cytokine production by T cells from the tTG-specific T-cell lines Intracellular IFN-γ production in response to stimulation with tTGery was a feature of the three T-cell lines. The strongest increase of CD3/IFN-γ + cells in response to tTG exposure was seen at stimulation 2 (day 14, Figs. 4a/b) for each of the T-cell lines, being two-fold for CD01 and ten and twenty-five fold for CT01 and CT02, respectively. High background levels of CD3/ IFN-γ + T cells were seen at stimulations one and three, leading to inconsistent or mild increases in CD3/IFN-γ positivity. Similar levels of cytokine were produced when rh tTG was substituted for tTGery (Figs. 4a/b).
3.5. Cytokine secretion by tTG-specific T-cell lines IFN-γ was the dominant cytokine secreted by each T-cell line, with the highest levels detected in the culture supernatants of the CD01, the treated CD patient-derived cell line (Fig. 5a). CD01 was characterised by large amounts of IFN-γ secretion, which reached a peak in the week following the first re-stimulation with tTGery. IL-10 secretion was detected in the first two weeks of culture only, with IL-17A and IL-21 secretion observed throughout the culture period (Fig. 5). The first control-derived T-cell line, CT01, produced lower levels of IFN-γ, high levels of IL-10 and IL-17A, and low levels of IL-21, all of which reached a maximal level following the first re-stimulation of the T-cell line with tTGery (Fig. 5). The second control-derived T-cell line, CT02, produced moderately large amounts of IFN-γ, IL-10, and high levels of IL-21, all of which peaked after the first re-stimulation with tTGery (Fig. 5); this
Figure 3 The proliferative responses and phenotype of cells from each T cell line per round of stimulation. (a) The proliferation of samples of cells from each T cell line in response to erythrocyte tTG (grey), and PPD (black) was measured by 3H-thymidine incorporation. Stimulations 1, 2, and 3 represent days 6, 14, and 28 of culture (SI = stimulation index, the dashed line indicates the cut-off for positivity). (b) Shows the percentage of CD4 + T cells contained in each cell line at days 6, 14, and 28 of culture.
160 cell line produced no IL-17A. IL-4 was not detected in the culture media of all three tTG-specific T cell lines (data not shown).
4. Discussion T cells with a specificity for tissue transglutaminase, the major autoantigen of coeliac disease, have recently been identified in the periphery of patients with active disease [16]. In this study, we describe the isolation and characterisation of such autoreactive T cells from patients with treated CD, and normal control individuals. Initial evidence to support the presence of such cells was found, based on the demonstration of positive T cell proliferative responses to guinea pig and highly purified human tTG preparations (Fig. 1). T cell responses were particularly noted in patients with untreated CD but also in treated CD and some control subjects. Further experiments were performed to investigate if gp tTG stimulated T cell proliferation was MHC class II restricted. It is well established that CD is strongly associated with specific
R. Comerford et al. MHC class II molecules, in particular HLA-DQ2 and HLA-DQ8 [3,4]. In previous studies, MHC restriction of responses by T cell clones to gliadin was reported. Interestingly, whereas gliadin-specific T cell clones derived from peripheral blood were both HLA-DR and HLA-DQ restricted [20,21], T cell clones generated from coeliac intestinal mucosa were found to be exclusively HLA-DQ2 restricted [22]. In the present study, both HLA-DR and HLA-DQ restriction of gp tTG stimulated proliferation was observed; the anti-HLA-DR antibody caused more marked inhibition, suggesting that T cell responses against tTG are not exclusively HLA-DQ2-restricted. Since the commercial gp tTG, employed in the first set of experiments may contain impurities [23] and has only 81% homology with human tTG [24], a second series of cultures was performed, employing highly purified human red cell tTG, and recombinant human tTG preparations as the antigenic stimuli. A similar pattern of reactivity to that induced by gp tTG was induced by both human tTG preparations, with the most frequent and strongest responses seen in untreated CD patients (Figs. 1b/c). Responsiveness of a minority of control individuals was also a feature of proliferation assays using tTGery and rh
Figure 4 Intracellular IFN-γ production by tTG-specific T cells. The production of IFN-γ by T cells from each cell line was measured by intracellular cytokine staining (a). Grey bars = erythrocyte tTG, black bars recombinant human tTG. Background levels (dotted white bars) were calculated using unstimulated PBMCs (stimulation one), or cells from each cell line and feeder cells only (stimulations two and three). Stimulations 1, 2, and 3 represent days 6, 14, and 28 of culture. The data generated from such a typical experiment is shown in (b), which details stimulation two of the CT02 T-cell line.
Characterisation of tissue transglutaminase-reactive T cells from patients
161
Figure 5 Cytokine secretion by the tTG-specific T-cell lines. Samples of each culture supernatant were removed weekly, and for the presence of secreted cytokines by ELISA. On days 14 and 28, supernatants were removed prior to the re-stimulation of each cell line.
tTG, with a positive proliferative response occurring in 24% (tTGery), and 10% (rh tTG) of controls, respectively (Figs. 1b/c). Three T-cell lines were generated by the fortnightly stimulation of PBMCs with tTGery: one from an individual with treated CD, and two from tTG responsive control individuals. In order to identify these T-cell lines as true tTG antigen-specific T cells, stimulation experiments were performed in which the response to tTG was measured using differing read-outs such as cellular proliferation (Fig. 3), intracellular cytokine production (Fig. 4), and cytokine secretion (Fig. 5). The dominance of IFN-γ, both in the culture supernatants and intracellular cytokine responses of each cell line would appear to place tTG-specific T cells into the Th1 category, a conclusion also reached by Ciccocioppo et al. [16] and supported by the absence of secreted IL-4 in the culture media. As the cells used in this study were polyclonal T cell lines as opposed to T cell clones, the expansion of other T helper cell types, namely regulatory T cells, Th17 cells, and follicular T helper cells (Tfh) may have occurred, and as a consequence is implicated by the detection of the cytokines IL-10, IL-17A, and IL-21 in culture supernatants and positive intracellular cytokine responses to tTG stimulation. The production of IFN-γ and IL-10 was also a feature of the T cell clones generated by Ciccocioppo, and suggests a potential immunomodulatory function for tTG-specific T cells. In earlier studies, tTG-specific T-cell lines generated from untreated CD patients gave similar results with respect to expansion and cytokine production; however additional assays for proliferation and intracellular IFN-γ production could not be performed due to limited patient availability. Co-production of IFN-γ and IL-10 is a feature of islet antigen-specific T cells in type 1 diabetes mellitus [25], and has
been also described for gliadin-specific T cells in CD [20,26]. Interestingly, IL-10 was more frequently detected in the culture supernatants of the control-derived cell lines, which is perhaps indicative of the normal tolerogenic response to such a self-protein. Recent research has focused on the involvement of the newly-discovered Th17 and Tfh cell subsets in CD [27,28]. Gliadin-specific Th17 cells have been identified in the mucosa of CD patients, producing the Th17-related cytokines IL-17A, IL-21, IL-22, and TGF-β [29]. In contrast to this finding, Bodd and colleagues found that HLA-DQ2-restricted gluten-reactive T cells produce IL-21 but not IL-17A or IL-22 [13], a phenotype typical of Tfh cells [30]. High numbers of IL-21-secreting cells that did not secrete IL-17A were found in the intestinal lamina propria of children and the majority of adults with CD by van Leeuwen et al. [31]. In this study, IL-17A was a product of tTG-specific T cells in two of the cell lines generated (CD01, CT01, Fig. 5c), identifying these cells as an alternative source of this cytokine in CD. Another characteristic of the tTG-specific T cell lines was the production of IL-21 (Fig. 5d), at levels similar to that produced by gliadin-specific T cells in a report by Bodd et al. [13]. The highest levels of IL-21 were produced by the CD patient-derived CD01 cell line, which consistently produced IL-21 during the culture period (Fig. 5d). IL-21 can be a product of Th17, Tfh, or innate lymphoid cells [32,33]; however, the source identified in this study may putatively be identified as Tfh cells, as IL-17A positivity was not observed in the IL-21-producing T cells by intracellular flow cytometry (data not shown). The observation that cell line CT02 produced IL-21 but not IL-17A lends further weight to this hypothesis.
162 A feature of this study was the identification of proliferative responses to tTG in a small number of control individuals, and the subsequent generation of autoreactive T cell lines from cells isolated from these individuals. The generation of autoreactive T cells in vivo is controlled by a number of regulatory mechanisms, including thymic central tolerance, and peripheral tolerogenic mechanisms such as the deletion, re-programming, or rendering anergic of self-reactive T cells [34]. Such autoreactive T cells have been classically linked with autoimmunity; however, self antigen-specific T cells are frequently detected in otherwise healthy individuals [35]. T cells specific for β2-glycoprotein [36], GAD65 [37], and myelin basic protein [18] have been detected in healthy controls, while in a study using peptides of the Goodpasture's disease autoantigen (the α3 chain of type IV collagen), all of the 11 controls used had T cell reactivity to at least one peptide [38]. Although the study by Ciccocioppo et al. did not generate tTG-reactive T cell lines or clones from control individuals, or individuals with treated CD, this may be due to methological differences between the two studies. In our study, 98 patients, the majority of whom were treated, and 54 controls were screened for PBMC reactivity to tTG, with subsequent selection of responsive individuals for cell line generation. Apart from the contribution of tTG-specific T cells (and their principal product, IFN-γ) to the generation and maintenance of the CD intestinal lesion, a likely function of these cells is the provision of help to tTG-specific B cells. IFN-γ may skew class-switching of anti-tTG IgG in CD to the Th1-like pattern of IgG1 reported in CD patients [39], while IL-21 is emerging as an important cytokine for IgA class-switching in Peyer's patches [40], and has been shown to enhance Th1 responses in the CD intestine [41]. Th17 cells have previously been shown to impact on neutrophil recruitment in an IL-17A-independent manner [42]; accordingly, it is possible that tTG-specific Th17 cells, or IL-17A-producing innate lymphoid cells [33], may participate in the recruitment of the polymorphonuclear cell infiltrate seen in the CD intestinal lesion [43]. Further isolation of Th-polarised tTG-specific T cells, and their interaction with tTG-specific B cells with regard to their activation and autoantibody isotype/subtype profile, may prove informative. In conclusion, we have described the isolation of tTG-specific T cells from the periphery of treated CD patients and a minority of normal controls. The identification of these cells as producers of IL-17A and IL-21 may have relevance to the pathogenesis of CD, a process which remains unresolved. To date, only gliadin-specific T cells have been isolated from the intestine of CD patients [7]. The isolation of tTG-specific T cells from intestinal biopsies of CD patients or controls should now be seen as a further avenue for exploration.
Conflict of interest statement The author(s) declare that there are no conflicts of interest.
Acknowledgments Research was funded by the National Children's Research Centre (E/09/3), Dublin, Ireland. The authors acknowledge
R. Comerford et al. the kind gifts of erythrocyte transglutaminase from Walter Binder, Innova Diagnostics, San Diego, California, and the monoclonal antibody, SPV-L3 from Dr. Hergen Spits, University of Amsterdam.
References [1] W. Dieterich, et al., Identification of tissue transglutaminase as the autoantigen of celiac disease, Nat. Med. 3 (7) (1997) 797–801. [2] A. Di Sabatino, G.R. Corazza, Coeliac disease, Lancet 373 (9673) (2009) 1480–1493. [3] O. Molberg, et al., Tissue transglutaminase selectively modifies gliadin peptides that are recognized by gut-derived T cells in celiac disease, Nat. Med. 4 (6) (1998) 713–717. [4] Y. van de Wal, et al., Selective deamidation by tissue transglutaminase strongly enhances gliadin-specific T cell reactivity, J. Immunol. 161 (4) (1998) 1585–1588. [5] T.S. Halstensen, et al., Intraepithelial TcR alpha/ beta + lymphocytes express CD45RO more often than the TcR gamma/delta + counterparts in coeliac disease, Immunology 71 (4) (1990) 460–466. [6] M. Demarchi, et al., HLA-DR3 and DR7 in coeliac disease: immunogenetic and clinical aspects, Gut 24 (8) (1983) 706–712. [7] K.E. Lundin, et al., Gliadin-specific, HLA-DQ(alpha 1*0501, beta 1*0201) restricted T cells isolated from the small intestinal mucosa of celiac disease patients, J. Exp. Med. 178 (1) (1993) 187–196. [8] R.P. Anderson, T cells in peripheral blood after gluten challenge in coeliac disease, Gut 54 (9) (2005) 1217–1223. [9] S. Ben-Horin, et al., Characterizing the circulating, gliadinspecific CD4 + memory T cells in patients with celiac disease: linkage between memory function, gut homing and Th1 polarization, J. Leukoc. Biol. 79 (4) (2006) 676–685. [10] J.A. Garrote, et al., Celiac disease pathogenesis: the proinflammatory cytokine network, J. Pediatr. Gastroenterol. Nutr. 47 (Suppl. 1) (2008) S27–S32. [11] E.M. Nilsen, et al., Gluten specific, HLA-DQ restricted T cells from coeliac mucosa produce cytokines with Th1 or Th0 profile dominated by interferon gamma, Gut 37 (6) (1995) 766–776. [12] M. Raki, et al., Tetramer visualization of gut-homing glutenspecific T cells in the peripheral blood of celiac disease patients, Proc. Natl. Acad. Sci. 104 (8) (2007) 2831–2836. [13] M. Bodd, et al., HLA-DQ2-restricted gluten-reactive T cells produce IL-21 but not IL-17 or IL-22, Mucosal Immunol. 3 (6) (2010) 594–601. [14] L.M. Sollid, et al., Autoantibodies in coeliac disease: tissue transglutaminase—guilt by association? Gut 41 (6) (1997) 851–852. [15] C. Coates, et al., Proceedings of the Eleventh International Symposium on Coeliac Disease. April 28–30, 2004, Belfast, Northern Ireland, 2004, p. 197. [16] R. Ciccocioppo, et al., Isolation and characterization of circulating tissue transglutaminase-specific T cells in coeliac disease, Int. J. Immunopathol. Pharmacol. 23 (1) (2010) 179–191. [17] C.J. Hawkes, et al., T-cell lines reactive to an immunodominant epitope of the tyrosine phosphatase-like autoantigen IA-2 in type 1 diabetes, Diabetes 49 (3) (2000) 356–366. [18] M.P. Pender, et al., A study of human T-cell lines generated from multiple sclerosis patients and controls by stimulation with peptides of myelin basic protein, J. Neuroimmunol. 70 (1) (1996) 65–74. [19] R. Comerford, et al., Binding of autoantibodies to the core region of tissue transglutaminase is a feature of paediatric coeliac disease, J. Pediatr. Gastroenterol. Nutr. 55 (4) (2012) 445–450.
Characterisation of tissue transglutaminase-reactive T cells from patients [20] J. O'Keeffe, et al., T cell proliferation, MHC class II restriction and cytokine products of gliadin-stimulated peripheral blood mononuclear cells (PBMC), Clin. Exp. Immunol. 117 (2) (1999) 269–276. [21] H.A. Gjertsen, et al., T cells from the peripheral blood of coeliac disease patients recognize gluten antigens when presented by HLA-DR, -DQ, or -DP molecules, Scand. J. Immunol. 39 (6) (1994) 567–574. [22] K. Jensen, et al., Gliadin-specific T cell responses in peripheral blood of healthy individuals involve T cells restricted by the coeliac disease associated DQ2 heterodimer, Scand. J. Immunol. 42 (1) (1995) 166–170. [23] M.G. Clemente, et al., Antitissue transglutaminase antibodies outside celiac disease, J. Pediatr. Gastroenterol. Nutr. 34 (1) (2002) 31–34. [24] V. Gentile, et al., Isolation and characterization of cDNA clones to mouse macrophage and human endothelial cell tissue transglutaminases, J. Biol. Chem. 266 (1) (1991) 478–483. [25] L.G. Petrich de Marquesini, et al., IFN-gamma and IL-10 isletantigen-specific T cell responses in autoantibody-negative firstdegree relatives of patients with type 1 diabetes, Diabetologia 53 (7) (2010) 1451–1460. [26] E.M. Nilsen, et al., Gluten activation of peripheral blood T cells induces a Th0-like cytokine pattern in both coeliac patients and controls, Clin. Exp. Immunol. 103 (2) (1996) 295–303. [27] I. Monteleone, et al., Characterization of IL-17A-producing cells in celiac disease mucosa, J. Immunol. 184 (4) (2010) 2211–2218. [28] D. De Nitto, Involvement of interleukin-15 and interleukin-21, two γ-chain-related cytokines, in celiac disease, World J. Gastroenterol. 15 (37) (2009) 4609. [29] S. Fernandez, et al., Characterization of gliadin-specific Th17 cells from the mucosa of celiac disease patients, Am. J. Gastroenterol. 106 (3) (2011) 528–538. [30] D. Yu, et al., Lineage specification and heterogeneity of T follicular helper cells, Curr. Opin. Immunol. 21 (6) (2009) 619–625.
163
[31] M.A. van Leeuwen, et al., Increased production of interleukin-21, but not interleukin-17A, in the small intestine characterizes pediatric celiac disease, Mucosal Immunol. 6 (6) (2013) 1202–1213. [32] A. Awasthi, V.K. Kuchroo, Th17 cells: from precursors to players in inflammation and infection, Int. Immunol. 21 (5) (2009) 489–498. [33] M.R. Dunne, et al., Persistent changes in circulating and intestinal gammadelta T cell subsets, invariant natural killer T cells and mucosal-associated invariant T cells in children and adults with coeliac disease, PLoS ONE 8 (10) (2013) e76008. [34] J. Lohr, et al., T-cell tolerance and autoimmunity to systemic and tissue-restricted self-antigens, Immunol. Rev. 204 (2005) 116–127. [35] N.A. Danke, et al., Autoreactive T cells in healthy individuals, J. Immunol. 172 (10) (2004) 5967–5972. [36] N. Hattori, et al., T cells that are autoreactive to beta2glycoprotein I in patients with antiphospholipid syndrome and healthy individuals, Arthritis Rheum. 43 (1) (2000) 65–75. [37] N.A. Danke, et al., Comparative study of GAD65-specific CD4 + T cells in healthy and type 1 diabetic subjects, J. Autoimmun. 25 (4) (2005) 303–311. [38] J. Zou, et al., Healthy individuals have Goodpasture autoantigenreactive T cells, J. Am. Soc. Nephrol. 19 (2) (2008) 396–404. [39] J. Schilling, et al., Immunoglobulin isotype profile of tissue transglutaminase autoantibodies is correlated with the clinical presentation of coeliac disease, Scand. J. Immunol. 61 (2) (2005) 207–212. [40] R. Spolski, W.J. Leonard, IL-21 and T follicular helper cells, Int. Immunol. 22 (1) (2010) 7–12. [41] D. Fina, et al., Interleukin 21 contributes to the mucosal T helper cell type 1 response in coeliac disease, Gut 57 (7) (2008) 887–892. [42] M. Pelletier, et al., Evidence for a cross-talk between human neutrophils and Th17 cells, Blood 115 (2) (2010) 335–343. [43] A. Ensari, Gluten-sensitive enteropathy (celiac disease): controversies in diagnosis and classification, Arch. Pathol. Lab. Med. 134 (6) (2010) 826–836.
available at www.sciencedirect.com
Clinical Immunology www.elsevier.com/locate/yclim
Characterisation of tissue transglutaminase-reactive T cells from patients with coeliac disease and healthy controls Ross Comerford a,b,c,⁎,1 , Christian Coates b,c,1 , Greg Byrne d , Sara Lynch d , Padraic Dunne b , Margaret Dunne b , Jacinta Kelly a , Conleth Feighery b,c a
National Children's Research Centre, Our Lady's Hospital For Sick Children, Crumlin Dublin 12, Ireland Institute of Molecular Medicine, Trinity College Dublin Dublin 8, Ireland c Department of Immunology, St James's Hospital, Dublin 8, Ireland d Dublin Institute of Technology, Kevin Street, Dublin 8, Ireland b
Received 28 March 2014; accepted with revision 1 August 2014 KEYWORDS Coeliac disease; Tissue transglutaminase; Autoimmunity; T cells; Interferon-γ; Interleukin-17A
Abstract Previous studies have shown evidence for T lymphocytes specific for tissue transglutaminase (tTG) in the periphery of coeliac disease (CD) patients. These cells could play a role in disease pathogenesis and may be involved in providing help for the production of anti-tTG autoantibodies. The objective of this study was to further investigate the presence of tTG-specific T cells in patients with treated and untreated CD, and normal controls. Positive proliferative responses to three different commercial tTG antigens were detected in all groups tested, occurring more frequently and at higher levels in untreated CD patients. The addition of antibodies to HLA-DQ and HLA-DR caused a significant reduction in the proliferative response to tTG. T cell lines specific for tTG and composed predominantly of CD4-positive T cells were generated from responsive CD and control individuals, and were found to produce large amounts of interferon-γ, as well as interleukins 10, 17A, and 21. © 2014 Elsevier Inc. All rights reserved.
1. Introduction ⁎ Corresponding author at: Department of Immunology, St James's Hospital, Dublin 8, Ireland. E-mail address: [email protected] (R. Comerford). 1 These authors contributed equally to this work.
http://dx.doi.org/10.1016/j.clim.2014.08.001 1521-6616/© 2014 Elsevier Inc. All rights reserved.
Coeliac disease (CD) is a common inflammatory disease of the small intestine caused by an inappropriate immune response to wheat gluten in genetically predisposed individuals. Investigation into the pathogenesis of coeliac disease
156
R. Comerford et al.
reveals a complex interplay between environmental factors, genetics, the adaptive and innate immune systems, and the presence of autoantigens, in a process which has still not been fully elucidated. The principal autoantigen of coeliac disease is the ubiquitously expressed, multifunctional enzyme tissue transglutaminase (tTG) [1]. The demonstration of IgA autoantibodies directed to this self-protein is an integral component in the diagnosis of coeliac disease [2]. Tissue transglutaminase is also intimately involved in coeliac disease pathogenesis through the modification of gluten peptides by deamidation, which facilitates their presentation to the immune system via HLA-DQ2 or DQ8 molecules [3,4]. Due to the lymphocytic infiltration observed in the active CD intestinal lesion [5], and the strong association of the disease with HLA molecules [6], the concept of T-cell involvement in the pathogenesis of CD had been long-established prior to the isolation of gliadin-specific T cells from the CD intestinal mucosa [7]. In the time since this important observation, the CD4+ gliadin-specific T-cell response has been dissected and extensively characterised in relation to antigen recognition, phenotype, and function. The hallmark of gliadin-specific T-cell lines and clones is a mucosal-homing phenotype [8,9], and production of large amounts of IFN-γ, a cytokine that contributes substantially to the intestinal destruction seen in CD [10–12]. Apart from their role in the immune reaction that leads to the formation of the CD lesion [13], gliadin-specific T cells have been speculated to provide T cell help to tTG-specific B cells, resulting in the anti-tTG response characteristic of CD. This ‘hapten-carrier’ theory, proposed by Sollid, was based upon the fact that T cells specific for tTG had never been isolated, and the gliadin-dependant nature of the anti-tTG response [14]. However, preliminary findings from our group indicated that T cells specific for tTG could be detected in CD patients and in some control individuals, in a HLA-DQ and HLA-DR restricted manner [15]. The existence of tTG-specific T cells was confirmed in a recent paper by Ciccocioppo et al., who described the presence of T cells that were mostly CD4+ and proliferated to tTG in a HLA-DQ2-restricted manner in the periphery of untreated CD patients [16]. The aim of this study was to further investigate the presence of tTG-specific T cells in the periphery of CD patients and healthy controls, as it is known that T cells specific for other autoantigens have been detected in healthy subjects [17,18]. Peripheral blood samples were challenged with tTG and proliferative responses measured. tTG-reactive T cell
lines were generated from responsive individuals and both the intracellular and secreted cytokines IFN-γ, IL-10, IL-17A, and IL-21 were measured, in tandem with the measurement of proliferation of these cells.
2. Materials and methods 2.1. Patients and controls CD patients and controls were recruited from the Departments of Gastroenterology and Immunology, St James's Hospital Dublin. Ethical approval for this study was granted from the hospital ethics committee. The diagnosis of CD was based on a typical histological lesion, positive serology (IgA anti-tTG and endomysial (EMA) antibodies) and positive histological, serological and clinical responses to a gluten free diet. Patients with CD were sub-divided according to their treatment status. This included: 33 patients with untreated CD who were taking a normal diet and with positive serological tests; 65 patients with treated CD, taking a gluten free diet and with negative or low positive serological tests. In addition, 54 healthy control subjects, with negative coeliac serology, were investigated. The demographic information of each study group is summarised in Table 1.
2.2. Proliferation assays Fresh peripheral blood mononuclear cells (PBMCs) from CD patients or control individuals were separated by density gradient centrifugation using Lymphoprep™ solution (AxisShield), and adjusted to a concentration of 5 × 105/ml in T cell medium (TCM) which contained RPMI containing HEPES, sodium pyruvate, non-essential amino acids, essential amino acids, penicillin, streptomycin, fungizone, L-glutathione, β-mercaptoethanol, and 5% autologous serum. Two hundred microlitres per well of this suspension was then cultured in triplicate with the appropriate antigen in a round-bottomed 96-well microtitre plate. Wells containing medium and cells only were added in order to measure background proliferation, and used to calculate the stimulation index (SI). The tTG antigens used were guinea pig tTG (gp tTG, Sigma), erythrocyte tTG (tTGery, Inova Diagnostics), recombinant human tTG (rh tTG, produced in SF9 insect cells, Zedira), at a concentration of 10 μg/ml. Phytohaemagglutinin (PHA) and purified protein
Table 1 Characteristics of the groups studied in proliferation assays using various commercial sources of tissue transglutaminase (tTG), with resulting responses and stimulation indices. Antigen
Study group
Number
Sex (M/F)
Mean age (years) (range)
No. positive (%)
Mean SI (range)
Mean SI responders (range)
Guinea-pig tTG
CD untreated CD treated Controls CD untreated CD treated Controls CD untreated CD treated Controls
15 30 29 11 35 25 7 15 10
2/13 7/23 11/18 4/7 14/21 12/13 2/5 4/11 3/7
45 (20–70) 59 (19–68) 29 (18–54) 48 (22–65) 58 (29–83) 25 (24–65) 42 (20–59) 52 (29–77) 29 (24–57)
12 (80) 12 (40) 11 (38) 5 (45) 11 (31) 6 (24) 6 (86) 6 (40) 1 (10)
4.7 (1.1–38.1) 3.1 (0.8–13.9) 2.4 (0.7–20.6) 5.0 (0.4–24.8) 1.6 (0.5–9.2) 2.1 (0.7–7.6) 19.1 (0.2–57) 2.2 (0.1–6.2) 1.0 (0.4–2.1)
8.5 (3.0–38.1) 4.8 (2.1–13.9) 5.9 (2.1–20.6) 9.3 (2.1–24.8) 2.8 (2.2–9.2) 5.0 (3.4–7.6) 22.8 (4.2–57) 4.0 (2.1–6.2) 2.0 (n/a)
Erythrocyte tTG
Recombinant htTG
Characterisation of tissue transglutaminase-reactive T cells from patients derivative (PPD) of Mycobacterium tuberculosis at concentrations of 2 and 10 μg/ml, respectively, were used as proliferative control stimuli. Cultures were then incubated at 37 °C, in the presence of 5% CO2 and a humidified atmosphere, for 6 days. For the last 18 h of culture 0.5 μCi of 3H-thymidine (PerkinElmer) was added to each well and cells were harvested on day 6. Proliferation was assessed by measuring the thymidine incorporation using a MicroBeta Trilux scintillator (PerkinElmer). The stimulation index (SI) was calculated by dividing the mean of the three cpm values for each antigen by that of the unstimulated wells, with an SI of 2, or greater, being considered positive. In experiments measuring the proliferation of tTG-sensitised T-cell lines, the incubation time was shortened to 72 h. MHC restriction was investigated using PBMCs from 16 of the coeliac patients and 2 normal controls. PBMCs were cultured with gp tTG in the presence or absence of (i) a γ2A anti-HLA-DR (L243, Becton and Dickinson); (ii) a γ2A non-specific isotype matched control antibody (Becton and Dickinson); (iii) a γ1 anti-HLA-DQ (SPV-L3; a gift from Dr. Hergen Spits, University of Amsterdam, Amsterdam, The Netherlands), or (iv) a γ1 non-specific isotype matched control antibody (Becton and Dickinson). Each of the commercially prepared antibodies was dialysed in phosphate buffered saline to remove preservatives. Proliferation was measured as described above and results were calculated on the basis of percentage inhibition.
157
and PPD, were measured on days 14 and 28. The cells were starved of IL-2 for 5 days prior to re-stimulation. Irradiated autologous PBMCs pulsed with each test antigen were prepared, with 1 × 105 APCs added to 1 × 105 of the IL-2-starved cells from the cell line in a volume of 200 μl in a 96-well round-bottomed microtitre plate. These suspensions were then returned to culture for a further 72 h, with proliferation then measured by 3 H-thymidine incorporation. Cellular phenotype (CD4 or CD8 positivity) as well as intracellular cytokine production was determined by flow cytometric analysis. Cultures for intracellular cytokine staining had 10 ng/ml phorbol 12-myristate 13-acetate (Sigma), 1 μg/ml ionomycin (Sigma), and 10 μg/ml Brefeldin A (Sigma) added for the final 10 h of culture.
2.5. Flow cytometry Intracellular cytokine production, and the phenotype of samples of T cells from each T-cell line in response to re-stimulation with tTG was assessed. The panel of antibodies used was CD3-APC-efluor780, CD4-APC, CD8-PECy5, and IFN-γ-FITC (eBioscience, San Diego, CA). Cells were gated on forward/side scatter, CD3, and aqua fluorescent reactive dye (Invitrogen) exclusion, which permitted the measurement of live cells only. Quadrants for all parameters were set using fluorescence minus one (FMO) controls. For all flow cytometric analysis, 5 × 105 cells from each sample were measured.
2.3. T-cell line generation For the generation of tTG-specific T-cell lines, 1 × 106 PBMCs from CD patients or control individuals in TCM was incubated with tTGery at 10 μg/ml. Wells containing 1 × 106 PBMCs only were cultured as a negative control. After 5 days in culture, 20 IU/ml of recombinant human IL-2 (Sigma) was added to each well. On day 7, wells were visually examined for growth by comparing cellular numbers in stimulated and unstimulated wells using an inverted microscope, and observing a colour change from pink to yellow in the culture media. The cells were then cultured for a further 7 days, receiving 20 IU/ml IL-2 on day 9, examined on the inverted microscope, and counted. In order to expand the T-cell lines, further stimulation with tTGery was performed on days 14 and 28 of culture, using irradiated autologous PBMCs as antigen-presenting cells. The APCs were incubated with tTGery at a concentration of 10 μg/ml in TCM, and added to the tTG-stimulated cell line at a ratio of 1:1. Twenty International Units per millilitres IL-2 was added to each culture on day 14, and weekly thereafter. The cells were counted again on day 28, with the concentration adjusted to 1 × 106/ml, and re-plated in a 24-well plate prior to re-stimulation. All incubations were performed at 37 °C, in the presence of 5% CO2 and a humidified atmosphere. At weekly intervals, 500 μl of the culture supernatant was removed and stored at −20 °C, being replaced with 500 μl of TCM.
2.6. Measurement of cytokines by ELISA The supernatants removed weekly from each T-cell line and media only controls were analysed for the presence of IFN-γ, IL-10, IL-4, IL-17A, and IL-21 using ELISA MAX™ kits (BioLegend, San Diego, CA). Assays were performed as per the manufacturer's instructions, with all measurements being made in triplicate.
2.7. Statistical analyses Differences in proliferation in response to tTG between patient and control groups were evaluated for statistical significance using the Mann–Whitney test, with the paired t test used to identify significant effects of MHC blocking on proliferative responses to tTG. The level of significance was set at 0.05. The GraphPad Prism 5 program was used for all statistical analysis.
3. Results 3.1. Proliferation assays
2.4. Re-stimulation of T-cell lines To demonstrate the antigen-specificity of the cell lines generated, the proliferative capacity and cytokine responses to stimulation with tTGery, rh tTG, and the control stimuli PHA
Positive proliferative responses to gp tTG were observed in 80% of untreated coeliac patients, compared to 40% of treated coeliac patients and 38% of normal control subjects (Fig. 1a, Table 1). Furthermore, the overall SI response of
158
R. Comerford et al.
Figure 1 Proliferative responses to guinea pig, erythrocyte, and recombinant human tissue transglutaminases. Proliferation of PBMCs from normal controls (NC), untreated coeliac disease patients (UTCD), and treated coeliac disease patients (TCD) to guinea pig (a), erythrocyte (b), and recombinant (c) human tissue transglutaminases. For all experiments, proliferation was measured by 3H-thymidine incorporation after six days of culture (SI = stimulation index, the dashed line indicates the cut-off for positivity). (*p b 0.05, **p b 0.005).
the untreated coeliac patients was significantly higher than both the treated coeliac patients (p = 0.0042, Mann– Whitney test) and the normal control subjects (p = 0.0063, Mann–Whitney test). For the tTGery antigen, proliferation was again more frequently detected in untreated CD patients (45%), with 35% of treated CD patients, and 24% of controls responding (Fig. 1b, Table 1). A similar pattern of response to the recombinant human tTG preparation (rh tTG) was noted, with positivity detected in the majority of untreated CD patients (Fig. 1c, Table 1). For untreated CD patients, the overall SI response to rh tTG was significantly higher than both the treated CD (p = 0.009, Mann–Whitney test) and the normal control subjects (p = 0.0065, Mann– Whitney test). In experiments comparing the proliferative stimulus of gp tTG and tTGery, similar levels of proliferation were observed in CD and control subjects, irrespective of the source of antigen. In a group of ten untreated CD patients the three responders to gp tTG also had a corresponding response to tTGery, a finding replicated in three of four responders in a control group comprised of ten individuals (data not shown). Of the fifteen treated CD patients tested with both the tTGery and rh tTG antigens, two proliferated to tTGery only, one to rh tTG only, and four proliferated to both preparations. Ten control individuals were tested with both commercial tTG antigens, two were positive for tTGery only, and one positive for rh tTG only, with no individual proliferating to both antigens. Substitution of any of the commercial tTG sources with our own in-house recombinant tTGs [19] resulted in equivalent proliferative responses, in both CD patients and controls (data not shown).
3.2. MHC blocking studies In MHC blocking studies, the inhibitory effect of addition of anti-HLA-DR and HLA-DQ to gp tTG stimulated PBMCs was investigated in 18 individuals (16 coeliac and 2 normal control subjects) with known proliferative responses to this antigen (Fig. 2). The effect of anti-MHC II antibodies was contrasted with that of isotype matched controls. Marked inhibition of proliferation was noted with both anti-MHC II antibodies, in particular with anti-HLA-DR, whereas the control antibodies
Figure 2 Inhibitory effect of addition of anti-HLA-DR/-DQ. Proliferation of PBMCs from 16 coeliac and 2 control subjects to guinea pig tTG could be blocked by the addition of antibodies to HLA-DR or HLA-DQ, an effect not replicated by the isotype control antibodies gamma 2A or gamma 1. The grey lines indicate the mean; the dashed line represents the cut-off for positivity (*p b 0.05, **p b 0.005).
Characterisation of tissue transglutaminase-reactive T cells from patients caused no significant inhibition. In 16/18 individuals, anti-HLA-DR caused a mean inhibition of 79.8% (range 20– 97%). There was a significant difference between the response seen in the anti-HLA-DR cultures (median SI 0.4 ± 0.1) and the cultures containing the γ2A isotype matched control antibody (median SI 3.4 ± 1.5) (p = 0.007, Paired t test). The addition of anti-HLA-DQ caused an inhibition of proliferation in 13/18 subjects (mean inhibition 48.5%, range 10–89%). There was a significant difference between the response seen in the anti-HLA-DQ cultures (median SI 1.28 ± 0.5) and the cultures containing the γ1 isotype matched control antibody (median SI 3.2 ± 1.0) (p = 0.002, paired t test).
3.3. T-cell line generation Three polyclonal T-cell lines were generated by initial stimulation of PBMCs with tTGery, and fortnightly re-stimulation of the cell line thereafter. The T-cell line CD01 was generated from a male of 51 years who had biopsy-confirmed CD, and was on a strict gluten-free diet for ten years, as evidenced by serial tTG/EMA negativity. The T-cell lines CT01 and CT02 were generated from two non-CD controls, males of 32 and 56 years, respectively. All three individuals were selected because they gave a positive proliferative response to tTGery in 3H-thymidine incorporation assays. The T-cell lines established were predominantly CD4+, the percentage of which increased over the duration of the culture period for CD01 and CT01 (Fig. 3b). In order to confirm the antigen-specificity of the T-cell lines, re-stimulation with tTGery, was performed, with responsiveness measured by 3H-thymidine incorporation. At each re-stimulation (days 14 and 28), a proliferative response to tTGery was detected, with the strongest response being the second stimulation for each cell line (Fig. 3). Although the cell lines had been established using tTGery, after the first stimulation samples of cells from each cell line had the ability to proliferate in response to the rh tTG antigen, at levels equivalent to tTGery, and PHA, which was used as a positive
159
control (data not shown). After the initial stimulation each T-cell line lost responsiveness to the protein antigen PPD, indicating a positive selection of tTG-specific cells during culture (Fig. 3).
3.4. Intracellular cytokine production by T cells from the tTG-specific T-cell lines Intracellular IFN-γ production in response to stimulation with tTGery was a feature of the three T-cell lines. The strongest increase of CD3/IFN-γ + cells in response to tTG exposure was seen at stimulation 2 (day 14, Figs. 4a/b) for each of the T-cell lines, being two-fold for CD01 and ten and twenty-five fold for CT01 and CT02, respectively. High background levels of CD3/ IFN-γ + T cells were seen at stimulations one and three, leading to inconsistent or mild increases in CD3/IFN-γ positivity. Similar levels of cytokine were produced when rh tTG was substituted for tTGery (Figs. 4a/b).
3.5. Cytokine secretion by tTG-specific T-cell lines IFN-γ was the dominant cytokine secreted by each T-cell line, with the highest levels detected in the culture supernatants of the CD01, the treated CD patient-derived cell line (Fig. 5a). CD01 was characterised by large amounts of IFN-γ secretion, which reached a peak in the week following the first re-stimulation with tTGery. IL-10 secretion was detected in the first two weeks of culture only, with IL-17A and IL-21 secretion observed throughout the culture period (Fig. 5). The first control-derived T-cell line, CT01, produced lower levels of IFN-γ, high levels of IL-10 and IL-17A, and low levels of IL-21, all of which reached a maximal level following the first re-stimulation of the T-cell line with tTGery (Fig. 5). The second control-derived T-cell line, CT02, produced moderately large amounts of IFN-γ, IL-10, and high levels of IL-21, all of which peaked after the first re-stimulation with tTGery (Fig. 5); this
Figure 3 The proliferative responses and phenotype of cells from each T cell line per round of stimulation. (a) The proliferation of samples of cells from each T cell line in response to erythrocyte tTG (grey), and PPD (black) was measured by 3H-thymidine incorporation. Stimulations 1, 2, and 3 represent days 6, 14, and 28 of culture (SI = stimulation index, the dashed line indicates the cut-off for positivity). (b) Shows the percentage of CD4 + T cells contained in each cell line at days 6, 14, and 28 of culture.
160 cell line produced no IL-17A. IL-4 was not detected in the culture media of all three tTG-specific T cell lines (data not shown).
4. Discussion T cells with a specificity for tissue transglutaminase, the major autoantigen of coeliac disease, have recently been identified in the periphery of patients with active disease [16]. In this study, we describe the isolation and characterisation of such autoreactive T cells from patients with treated CD, and normal control individuals. Initial evidence to support the presence of such cells was found, based on the demonstration of positive T cell proliferative responses to guinea pig and highly purified human tTG preparations (Fig. 1). T cell responses were particularly noted in patients with untreated CD but also in treated CD and some control subjects. Further experiments were performed to investigate if gp tTG stimulated T cell proliferation was MHC class II restricted. It is well established that CD is strongly associated with specific
R. Comerford et al. MHC class II molecules, in particular HLA-DQ2 and HLA-DQ8 [3,4]. In previous studies, MHC restriction of responses by T cell clones to gliadin was reported. Interestingly, whereas gliadin-specific T cell clones derived from peripheral blood were both HLA-DR and HLA-DQ restricted [20,21], T cell clones generated from coeliac intestinal mucosa were found to be exclusively HLA-DQ2 restricted [22]. In the present study, both HLA-DR and HLA-DQ restriction of gp tTG stimulated proliferation was observed; the anti-HLA-DR antibody caused more marked inhibition, suggesting that T cell responses against tTG are not exclusively HLA-DQ2-restricted. Since the commercial gp tTG, employed in the first set of experiments may contain impurities [23] and has only 81% homology with human tTG [24], a second series of cultures was performed, employing highly purified human red cell tTG, and recombinant human tTG preparations as the antigenic stimuli. A similar pattern of reactivity to that induced by gp tTG was induced by both human tTG preparations, with the most frequent and strongest responses seen in untreated CD patients (Figs. 1b/c). Responsiveness of a minority of control individuals was also a feature of proliferation assays using tTGery and rh
Figure 4 Intracellular IFN-γ production by tTG-specific T cells. The production of IFN-γ by T cells from each cell line was measured by intracellular cytokine staining (a). Grey bars = erythrocyte tTG, black bars recombinant human tTG. Background levels (dotted white bars) were calculated using unstimulated PBMCs (stimulation one), or cells from each cell line and feeder cells only (stimulations two and three). Stimulations 1, 2, and 3 represent days 6, 14, and 28 of culture. The data generated from such a typical experiment is shown in (b), which details stimulation two of the CT02 T-cell line.
Characterisation of tissue transglutaminase-reactive T cells from patients
161
Figure 5 Cytokine secretion by the tTG-specific T-cell lines. Samples of each culture supernatant were removed weekly, and for the presence of secreted cytokines by ELISA. On days 14 and 28, supernatants were removed prior to the re-stimulation of each cell line.
tTG, with a positive proliferative response occurring in 24% (tTGery), and 10% (rh tTG) of controls, respectively (Figs. 1b/c). Three T-cell lines were generated by the fortnightly stimulation of PBMCs with tTGery: one from an individual with treated CD, and two from tTG responsive control individuals. In order to identify these T-cell lines as true tTG antigen-specific T cells, stimulation experiments were performed in which the response to tTG was measured using differing read-outs such as cellular proliferation (Fig. 3), intracellular cytokine production (Fig. 4), and cytokine secretion (Fig. 5). The dominance of IFN-γ, both in the culture supernatants and intracellular cytokine responses of each cell line would appear to place tTG-specific T cells into the Th1 category, a conclusion also reached by Ciccocioppo et al. [16] and supported by the absence of secreted IL-4 in the culture media. As the cells used in this study were polyclonal T cell lines as opposed to T cell clones, the expansion of other T helper cell types, namely regulatory T cells, Th17 cells, and follicular T helper cells (Tfh) may have occurred, and as a consequence is implicated by the detection of the cytokines IL-10, IL-17A, and IL-21 in culture supernatants and positive intracellular cytokine responses to tTG stimulation. The production of IFN-γ and IL-10 was also a feature of the T cell clones generated by Ciccocioppo, and suggests a potential immunomodulatory function for tTG-specific T cells. In earlier studies, tTG-specific T-cell lines generated from untreated CD patients gave similar results with respect to expansion and cytokine production; however additional assays for proliferation and intracellular IFN-γ production could not be performed due to limited patient availability. Co-production of IFN-γ and IL-10 is a feature of islet antigen-specific T cells in type 1 diabetes mellitus [25], and has
been also described for gliadin-specific T cells in CD [20,26]. Interestingly, IL-10 was more frequently detected in the culture supernatants of the control-derived cell lines, which is perhaps indicative of the normal tolerogenic response to such a self-protein. Recent research has focused on the involvement of the newly-discovered Th17 and Tfh cell subsets in CD [27,28]. Gliadin-specific Th17 cells have been identified in the mucosa of CD patients, producing the Th17-related cytokines IL-17A, IL-21, IL-22, and TGF-β [29]. In contrast to this finding, Bodd and colleagues found that HLA-DQ2-restricted gluten-reactive T cells produce IL-21 but not IL-17A or IL-22 [13], a phenotype typical of Tfh cells [30]. High numbers of IL-21-secreting cells that did not secrete IL-17A were found in the intestinal lamina propria of children and the majority of adults with CD by van Leeuwen et al. [31]. In this study, IL-17A was a product of tTG-specific T cells in two of the cell lines generated (CD01, CT01, Fig. 5c), identifying these cells as an alternative source of this cytokine in CD. Another characteristic of the tTG-specific T cell lines was the production of IL-21 (Fig. 5d), at levels similar to that produced by gliadin-specific T cells in a report by Bodd et al. [13]. The highest levels of IL-21 were produced by the CD patient-derived CD01 cell line, which consistently produced IL-21 during the culture period (Fig. 5d). IL-21 can be a product of Th17, Tfh, or innate lymphoid cells [32,33]; however, the source identified in this study may putatively be identified as Tfh cells, as IL-17A positivity was not observed in the IL-21-producing T cells by intracellular flow cytometry (data not shown). The observation that cell line CT02 produced IL-21 but not IL-17A lends further weight to this hypothesis.
162 A feature of this study was the identification of proliferative responses to tTG in a small number of control individuals, and the subsequent generation of autoreactive T cell lines from cells isolated from these individuals. The generation of autoreactive T cells in vivo is controlled by a number of regulatory mechanisms, including thymic central tolerance, and peripheral tolerogenic mechanisms such as the deletion, re-programming, or rendering anergic of self-reactive T cells [34]. Such autoreactive T cells have been classically linked with autoimmunity; however, self antigen-specific T cells are frequently detected in otherwise healthy individuals [35]. T cells specific for β2-glycoprotein [36], GAD65 [37], and myelin basic protein [18] have been detected in healthy controls, while in a study using peptides of the Goodpasture's disease autoantigen (the α3 chain of type IV collagen), all of the 11 controls used had T cell reactivity to at least one peptide [38]. Although the study by Ciccocioppo et al. did not generate tTG-reactive T cell lines or clones from control individuals, or individuals with treated CD, this may be due to methological differences between the two studies. In our study, 98 patients, the majority of whom were treated, and 54 controls were screened for PBMC reactivity to tTG, with subsequent selection of responsive individuals for cell line generation. Apart from the contribution of tTG-specific T cells (and their principal product, IFN-γ) to the generation and maintenance of the CD intestinal lesion, a likely function of these cells is the provision of help to tTG-specific B cells. IFN-γ may skew class-switching of anti-tTG IgG in CD to the Th1-like pattern of IgG1 reported in CD patients [39], while IL-21 is emerging as an important cytokine for IgA class-switching in Peyer's patches [40], and has been shown to enhance Th1 responses in the CD intestine [41]. Th17 cells have previously been shown to impact on neutrophil recruitment in an IL-17A-independent manner [42]; accordingly, it is possible that tTG-specific Th17 cells, or IL-17A-producing innate lymphoid cells [33], may participate in the recruitment of the polymorphonuclear cell infiltrate seen in the CD intestinal lesion [43]. Further isolation of Th-polarised tTG-specific T cells, and their interaction with tTG-specific B cells with regard to their activation and autoantibody isotype/subtype profile, may prove informative. In conclusion, we have described the isolation of tTG-specific T cells from the periphery of treated CD patients and a minority of normal controls. The identification of these cells as producers of IL-17A and IL-21 may have relevance to the pathogenesis of CD, a process which remains unresolved. To date, only gliadin-specific T cells have been isolated from the intestine of CD patients [7]. The isolation of tTG-specific T cells from intestinal biopsies of CD patients or controls should now be seen as a further avenue for exploration.
Conflict of interest statement The author(s) declare that there are no conflicts of interest.
Acknowledgments Research was funded by the National Children's Research Centre (E/09/3), Dublin, Ireland. The authors acknowledge
R. Comerford et al. the kind gifts of erythrocyte transglutaminase from Walter Binder, Innova Diagnostics, San Diego, California, and the monoclonal antibody, SPV-L3 from Dr. Hergen Spits, University of Amsterdam.
References [1] W. Dieterich, et al., Identification of tissue transglutaminase as the autoantigen of celiac disease, Nat. Med. 3 (7) (1997) 797–801. [2] A. Di Sabatino, G.R. Corazza, Coeliac disease, Lancet 373 (9673) (2009) 1480–1493. [3] O. Molberg, et al., Tissue transglutaminase selectively modifies gliadin peptides that are recognized by gut-derived T cells in celiac disease, Nat. Med. 4 (6) (1998) 713–717. [4] Y. van de Wal, et al., Selective deamidation by tissue transglutaminase strongly enhances gliadin-specific T cell reactivity, J. Immunol. 161 (4) (1998) 1585–1588. [5] T.S. Halstensen, et al., Intraepithelial TcR alpha/ beta + lymphocytes express CD45RO more often than the TcR gamma/delta + counterparts in coeliac disease, Immunology 71 (4) (1990) 460–466. [6] M. Demarchi, et al., HLA-DR3 and DR7 in coeliac disease: immunogenetic and clinical aspects, Gut 24 (8) (1983) 706–712. [7] K.E. Lundin, et al., Gliadin-specific, HLA-DQ(alpha 1*0501, beta 1*0201) restricted T cells isolated from the small intestinal mucosa of celiac disease patients, J. Exp. Med. 178 (1) (1993) 187–196. [8] R.P. Anderson, T cells in peripheral blood after gluten challenge in coeliac disease, Gut 54 (9) (2005) 1217–1223. [9] S. Ben-Horin, et al., Characterizing the circulating, gliadinspecific CD4 + memory T cells in patients with celiac disease: linkage between memory function, gut homing and Th1 polarization, J. Leukoc. Biol. 79 (4) (2006) 676–685. [10] J.A. Garrote, et al., Celiac disease pathogenesis: the proinflammatory cytokine network, J. Pediatr. Gastroenterol. Nutr. 47 (Suppl. 1) (2008) S27–S32. [11] E.M. Nilsen, et al., Gluten specific, HLA-DQ restricted T cells from coeliac mucosa produce cytokines with Th1 or Th0 profile dominated by interferon gamma, Gut 37 (6) (1995) 766–776. [12] M. Raki, et al., Tetramer visualization of gut-homing glutenspecific T cells in the peripheral blood of celiac disease patients, Proc. Natl. Acad. Sci. 104 (8) (2007) 2831–2836. [13] M. Bodd, et al., HLA-DQ2-restricted gluten-reactive T cells produce IL-21 but not IL-17 or IL-22, Mucosal Immunol. 3 (6) (2010) 594–601. [14] L.M. Sollid, et al., Autoantibodies in coeliac disease: tissue transglutaminase—guilt by association? Gut 41 (6) (1997) 851–852. [15] C. Coates, et al., Proceedings of the Eleventh International Symposium on Coeliac Disease. April 28–30, 2004, Belfast, Northern Ireland, 2004, p. 197. [16] R. Ciccocioppo, et al., Isolation and characterization of circulating tissue transglutaminase-specific T cells in coeliac disease, Int. J. Immunopathol. Pharmacol. 23 (1) (2010) 179–191. [17] C.J. Hawkes, et al., T-cell lines reactive to an immunodominant epitope of the tyrosine phosphatase-like autoantigen IA-2 in type 1 diabetes, Diabetes 49 (3) (2000) 356–366. [18] M.P. Pender, et al., A study of human T-cell lines generated from multiple sclerosis patients and controls by stimulation with peptides of myelin basic protein, J. Neuroimmunol. 70 (1) (1996) 65–74. [19] R. Comerford, et al., Binding of autoantibodies to the core region of tissue transglutaminase is a feature of paediatric coeliac disease, J. Pediatr. Gastroenterol. Nutr. 55 (4) (2012) 445–450.
Characterisation of tissue transglutaminase-reactive T cells from patients [20] J. O'Keeffe, et al., T cell proliferation, MHC class II restriction and cytokine products of gliadin-stimulated peripheral blood mononuclear cells (PBMC), Clin. Exp. Immunol. 117 (2) (1999) 269–276. [21] H.A. Gjertsen, et al., T cells from the peripheral blood of coeliac disease patients recognize gluten antigens when presented by HLA-DR, -DQ, or -DP molecules, Scand. J. Immunol. 39 (6) (1994) 567–574. [22] K. Jensen, et al., Gliadin-specific T cell responses in peripheral blood of healthy individuals involve T cells restricted by the coeliac disease associated DQ2 heterodimer, Scand. J. Immunol. 42 (1) (1995) 166–170. [23] M.G. Clemente, et al., Antitissue transglutaminase antibodies outside celiac disease, J. Pediatr. Gastroenterol. Nutr. 34 (1) (2002) 31–34. [24] V. Gentile, et al., Isolation and characterization of cDNA clones to mouse macrophage and human endothelial cell tissue transglutaminases, J. Biol. Chem. 266 (1) (1991) 478–483. [25] L.G. Petrich de Marquesini, et al., IFN-gamma and IL-10 isletantigen-specific T cell responses in autoantibody-negative firstdegree relatives of patients with type 1 diabetes, Diabetologia 53 (7) (2010) 1451–1460. [26] E.M. Nilsen, et al., Gluten activation of peripheral blood T cells induces a Th0-like cytokine pattern in both coeliac patients and controls, Clin. Exp. Immunol. 103 (2) (1996) 295–303. [27] I. Monteleone, et al., Characterization of IL-17A-producing cells in celiac disease mucosa, J. Immunol. 184 (4) (2010) 2211–2218. [28] D. De Nitto, Involvement of interleukin-15 and interleukin-21, two γ-chain-related cytokines, in celiac disease, World J. Gastroenterol. 15 (37) (2009) 4609. [29] S. Fernandez, et al., Characterization of gliadin-specific Th17 cells from the mucosa of celiac disease patients, Am. J. Gastroenterol. 106 (3) (2011) 528–538. [30] D. Yu, et al., Lineage specification and heterogeneity of T follicular helper cells, Curr. Opin. Immunol. 21 (6) (2009) 619–625.
163
[31] M.A. van Leeuwen, et al., Increased production of interleukin-21, but not interleukin-17A, in the small intestine characterizes pediatric celiac disease, Mucosal Immunol. 6 (6) (2013) 1202–1213. [32] A. Awasthi, V.K. Kuchroo, Th17 cells: from precursors to players in inflammation and infection, Int. Immunol. 21 (5) (2009) 489–498. [33] M.R. Dunne, et al., Persistent changes in circulating and intestinal gammadelta T cell subsets, invariant natural killer T cells and mucosal-associated invariant T cells in children and adults with coeliac disease, PLoS ONE 8 (10) (2013) e76008. [34] J. Lohr, et al., T-cell tolerance and autoimmunity to systemic and tissue-restricted self-antigens, Immunol. Rev. 204 (2005) 116–127. [35] N.A. Danke, et al., Autoreactive T cells in healthy individuals, J. Immunol. 172 (10) (2004) 5967–5972. [36] N. Hattori, et al., T cells that are autoreactive to beta2glycoprotein I in patients with antiphospholipid syndrome and healthy individuals, Arthritis Rheum. 43 (1) (2000) 65–75. [37] N.A. Danke, et al., Comparative study of GAD65-specific CD4 + T cells in healthy and type 1 diabetic subjects, J. Autoimmun. 25 (4) (2005) 303–311. [38] J. Zou, et al., Healthy individuals have Goodpasture autoantigenreactive T cells, J. Am. Soc. Nephrol. 19 (2) (2008) 396–404. [39] J. Schilling, et al., Immunoglobulin isotype profile of tissue transglutaminase autoantibodies is correlated with the clinical presentation of coeliac disease, Scand. J. Immunol. 61 (2) (2005) 207–212. [40] R. Spolski, W.J. Leonard, IL-21 and T follicular helper cells, Int. Immunol. 22 (1) (2010) 7–12. [41] D. Fina, et al., Interleukin 21 contributes to the mucosal T helper cell type 1 response in coeliac disease, Gut 57 (7) (2008) 887–892. [42] M. Pelletier, et al., Evidence for a cross-talk between human neutrophils and Th17 cells, Blood 115 (2) (2010) 335–343. [43] A. Ensari, Gluten-sensitive enteropathy (celiac disease): controversies in diagnosis and classification, Arch. Pathol. Lab. Med. 134 (6) (2010) 826–836.