Selective inhibition of the gliadin-specific, cell-mediated immune response by transamidation with microbial transglutaminase

Article Selective inhibition of the gliadin-specific, cell-mediated immune response by transamidation with microbial transglutaminase Emanuela Lombardi,* Paolo Bergamo,* Francesco Maurano,* Giuseppina Bozzella,* Diomira Luongo,* Giuseppe Mazzarella,* Vera Rotondi Aufiero,* Gaetano Iaquinto, and Mauro Rossi*,1 *Institute of Food Sciences, National Research Council, Avellino, Italy; and Gastroenterology Department, San G. Moscati Hospital, Avellino, Italy RECEIVED APRIL 6, 2012; REVISED OCTOBER 2, 2012; ACCEPTED OCTOBER 4, 2012. DOI: 10.1189/jlb.0412182 ABSTRACT CD is an immune-mediated enteropathy caused by the ingestion of wheat gluten. The modification of gluten by intestinal ttgase plays a crucial role in CD pathogenesis. In this study, we observed that extensive transamidation of wheat flour with K-C 2 H 5 by mtgase yielded spf and K-gliadins fractions. By Western blot, we found that these modifications were associated with strongly reduced immune cross-reactivity. With the use of DQ8 tg mice as a model of gluten sensitivity, we observed a dramatic reduction in IFN production in gliadin-specific spleen cells challenged with spf and K-gliadins in vitro (n 12; median values: 813 vs. 29 and 99; control vs. spf and K-gliadins, P 0.012 for spf, and P 0.003 for K-gliadins). For spf, we also observed an increase in the IL- 10/IFN protein ratio (n 12; median values: 0.3 vs. 4.7; control vs. spf, P 0.005). In intestinal biopsies from CD patients challenged in vitro with gliadins (n 10), we demonstrated further that K-gliadins dramatically reduced the levels of antigen-specific IFN mrna in all specimens responsive to native gliadins (four of 10; P 0.05). As cytotoxic effects have been described for gliadins, we also studied GST and caspase-3 activities using the enterocytic Caco-2 cell line. We found that neither activities were modified by flour transamidation. Our results indicate that K-C 2 H 5 cross-linking via mt- Gase specifically affects gliadin immunogenicity, reversing the inducible inflammatory response in models of gluten sensitivity without affecting other aspects of the biological activity of gliadins. J. Leukoc. Biol. 93: 479 488; 2013. Abbreviations: 7-AMC 7-amino-4-methylcoumarin, CD celiac disease, DQ2 HLA-DQA1*05-DQB1*02, DQ8 HLA-DQA1*03-DQB1*0302, GFD gluten-free diet, K-C 2 H 5 lysine ethyl ester, K-CH 3 lysine methyl ester, K-gliadins insoluble-transamidated gliadin, mtgase microbial transglutaminase, PEP prolyl endopeptidase, pt peptic-tryptic, spf soluble protein fraction, tg transgenic, ttgase tissue transglutaminase Introduction CD is an immune-mediated disorder caused in genetically susceptible individuals by the ingestion of wheat gluten and related prolamins present in barley and rye [1]. CD affects 1% of the general population in developed and developing countries, with an increasing prevalence reported in Europe and the United States [2, 3]. Currently, a lifelong GFD is required to alleviate the symptoms of CD and to normalize the antibodies in the intestinal mucosa [4]. Recovery is observed in most CD patients after only a few weeks on a GFD [5]. However, dietary compliance is poor, necessitating the development of alternative immunological or technological strategies to treat CD. Gluten proteins are divided into two approximately equal fractions according to their solubility in alcohol water solutions: gliadins (soluble) and glutenins (insoluble). Both components contain high levels of glutamine (30 35%) and proline (10 15%) residues and very few negatively charged amino acids. These proteins undergo a process of selective deamidation in the small intestines of CD patients, during which, specific glutamine residues are converted to glutamic acid by tt- Gase [6]. The presence of many proline residues in these proteins, which are resistant to digestive enzymes, ensures that many immunostimulatory epitopes survive digestion [7]. This allows peptides large enough for T cell recognition to be presented by the MHC complex. Furthermore, proline residues have also been shown to direct ttgase-mediated deamidation of glutamines [8]. Notably, a previous study found that gliadin can be cleaved by bacterial PEPs into short peptides that then lose their activity [9]. Accordingly, oral PEP therapy has been proposed as a possible treatment for CD [10, 11]. PEPs have also been evaluated as a technological tool for the preparation of detoxified 1. Correspondence: Institute of Food Sciences, NRC, via Roma 64, 83100 Avellino, Italy. E-mail: mrossi@isa.cnr.it 0741-5400/13/0093-479 Society for Leukocyte Biology Volume 93, April 2013 Journal of Leukocyte Biology 479

gluten. Selected sourdough lactobacilli have specialized PEPs, and one study reported that a 60-day diet of baked goods made from PEP-hydrolyzed wheat flour was not toxic to CD patients [12]. To improve the preservation of gluten structure, we tested a different enzymatic approach using the transamidation activity of food-grade mtgase, a transamidase of the endo- -glutamine: -lysine transferase type [13]. Unlike ttgase, mtgase is a calcium-independent, low MW protein, which has several advantages for food industrial applications [14]. Interestingly, mtgase has been shown to exhibit the same site specificity as ttgase but lacks deamidase activity [15]. Most importantly, we found that the transamidation of gliadin following the treatment of wheat flour with mtgase and K-CH 3 caused a dramatic down-regulation in IFN production in vitro in the intestinal T cells of CD patients [15]. However, in a preliminary clinical study, this treatment was effective only in a subset of CD patients [16]. In the present study, we examined the reaction products of wheat flour following transamidation using a new two-step procedure. We demonstrated that this specific treatment modified the physical-chemical properties of gliadins and that this modification was selectively associated with positive changes in the phenotype of the antigen-specific immune response mediated by T cells in two distinct models of gluten sensitivity. MATERIALS AND METHODS Transamidation reaction of wheat flour A commercial preparation of bread wheat flour was used. A total of 100 g flour was suspended in 8 vol 0.4 M NaCl and stirred for 10 min to extract albumin/globulins. The flour suspension was then centrifuged at 1000 g for 10 min, and the supernatant was discarded. The recovered pellet was washed exhaustively with water to eliminate any residual soluble protein, suspended in 1 vol water containing 20 mm pharmaceutical-grade K-C 2 H 5 (NutraBio.com, Middlesex, NJ, USA) and 8 U/g flour mtgase (ACTIVA WM, specific activity: 81 135 U/g; Ajinomoto Foods, Hamburg, Germany), and incubated for 2 h at 30 C. The suspension was then centrifuged at 3000 g for 10 min, and the supernatant was recovered. In the two-step process, the pellet was suspended again in 1 vol water containing a final concentration of 20 mm K-C 2 H 5 and fresh mtgase (8 U/g flour) and incubated for 3 h at 30 C. For the control, the sample was prepared using the same protocol in the absence of mtgase. Purification of protein fractions The gliadin fraction was extracted from the protein pellet using a modified Osborne procedure [17]. Proteins were assessed quantitatively by Bradford analysis [18] and qualitatively analyzed by 8 16% denaturing SDS-PAGE and Coomassie R-250 blue staining. The transamidation reaction was quantitatively monitored using a modified ninhydrin assay [19]. K-C 2 H 5 was used as a standard to estimate a linear relationship between the absorbance and the -amino N content, and the results were expressed as nmoles -amino N/mg protein. Preparation of pt digests of gliadins pt digests of gliadins purified from untreated (gliadins) or transamidated wheat flour were prepared by suspending gliadin samples (100 mg) in 0.1 N HCl (2 ml), ph 1.8, and incubating with 2 mg pepsin (Sigma Chemical, St. Louis, MO, USA; specific activity: 3260 U/mg) for 4 h at 37 C with shaking. The ph was then adjusted to 7.8, followed by a 4-h incubation with 2 mg trypsin (Sigma Chemical; specific activity: 10,600 U/mg). The enzyme reaction was stopped by incubating in boiling water for 10 min. Samples were freeze-dried and stored at 20 C. Western blotting Equivalent protein aliquots (50 g) were fractionated by 15% denaturing SDS-PAGE, blotted onto Immobilon PVDF membranes (Millipore SpA, Vimodrone-MI, Italy), and probed with a mouse polyclonal antibody toward wheat gliadins (1:1000 dilution) produced in-house. After washing, the membranes were incubated with biotinylated streptavidin-peroxidase-conjugated antibodies against mouse IgGs (Dako SpA, Milano, Italy; 1:5000 dilution). Finally, immunodetection was performed using Hyperfilm and ECL reagents (Amersham-GE Healthcare Europe GmbH, Glattbrugg, Switzerland). Labeled membranes were stained with Coomassie blue as a loading control. Cell culture and in vitro experiments Caco-2 cells obtained from the American Type Culture Collection (Manassas, VA, USA) were used for experiments between Passages 16 and 35. Cells were cultivated in DMEM (Gibco-Life Technologies, Monza MB, Italy) containing 10% FCS (Gibco-Life Technologies), 100 U/ml penicillin-streptomycin (Gibco-Life Technologies), and 1 mm glutamine at 37 C in a humidified 5% CO 2 atmosphere. Differentiated cells (2 weeks after reaching confluence) were incubated at 37 C with 1 mg/ml pt-gliadins, pt-transamidated gliadins, or a wheat albumin/globulin fraction (control). In some experiments, cells were incubated in the presence of human IFN at 100 ng/ml. GST activity was determined using a previously published protocol [20]. Caspase-3 activity was assessed based on hydrolysis of the acetyl-asp- Glu-Val-Asp-7-AMC substrate. The release of the 7-AMC moiety in protein extracts from cells following different treatments was evaluated by fluorometry (excitation, 360 nm; emission, 460 nm). The amount of 7-AMC was calculated using a standard curve with pure AMC, and the activity was expressed as nmoles AMC/mg protein/min. Patients and organ culture analysis Ten untreated DQ2 CD patients (age range 12 51 years; median age, 24 years) were enrolled in this study (Table 1) by the Gastroenterology Department of S.G. Moscati Hospital (Avellino, Italy). Ethics review board approval was obtained in advance. Informed consent was obtained for each participating subject. Biopsies from the distal duodenum were obtained during an upper-gastrointestinal endoscopy. Two specimens were used for histology, and the other specimens were cultured for 8 h in the presence or absence pt-gliadins or pt-transamidated gliadins (1 mg/ml) and finally stored in liquid nitrogen for RNA analysis. Histology was performed using a modified Oberhuber-Marsh classification of the intestinal mucosal lesion TABLE 1. Demographic and Morphological Data of CD Patients Patients Age Gender Marsh degree CD1 24 M II CD2 37 F II CD3 12 F II CD4 13 F III/c CD5 51 F II CD6 28 M III/a CD7 23 M II CD8 14 M III/a CD9 25 M III/a CD10 16 F III/a Marsh degree, Oberhuber-Marsh classification; M, male; F, female. 480 Journal of Leukocyte Biology Volume 93, April 2013 www.jleukbio.org

Lombardi et al. Transamidation selectively inhibits gliadin reactivity [21]. CD was diagnosed using a combination of clinical signs and characteristic CD histology. Mice and antigen treatments tg mice expressing the DQ8 molecule in the absence of endogenous mouse class II genes [22] were reared for several generations on a GFD (Altromin- MT-mod, Rieper SpA, Bolzano, Italy) in pathogen-free conditions at our animal facility (Accreditation No. 164/99-A). All procedures met the guidelines of the Italian Ministry of Health. Six- to 12-week-old mice were primed by i.p. injection with pt-gliadins (300 g), emulsified in CFA (Sigma Chemical; Day 0). Boosters containing the same amount of antigen in IFA were injected on Days 7 and 14. Mice were killed on Day 21 to recover their spleens. In vitro culture of murine spleen cells Spleens were passed through a stainless-steel wire mesh to dissociate cells. Erythrocytes were removed by treating the cell suspensions with a Tris-buffered ammonium chloride solution. For proliferation assay, 5 10 5 cells were incubated in 0.2 ml culture medium (RPMI 1640 containing 10% inactivated FCS, 100 U/ml penicillin, 100 g/ml streptomycin, 1% nonessential amino acids, 2 mm glutamine, and 50 M 2-ME) in 96-well flat-bottom plates at 37 C for 96 h in the presence of pt-gliadins or pt-transamidated gliadins (200 400 g/ml). Sixteen hours before harvesting, cultures were pulsed with 1 Ci/well [ 3 H]-thymidine. The results are expressed as mean cpm of triplicate cultures sd. For the analysis of the cytokine pattern, 5 10 6 mouse spleen cells were incubated in 1 ml culture medium in 24- well flat-bottom plates at 37 C for 8 or 72 h for RNA or cognate protein assessment, respectively, in the presence of antigens (200 g/ml; 200 400 g/ml in dose-response experiments). In some experiments, T cells were purified by negative selection using the Dynal Mouse T Cell Negative Isolation Kit (Invitrogen-Life Technologies, Monza MB, Italy), according to the manufacturer s instructions; T cell yield was assessed by FACS analysis. For the preparation of antigen-pulsed APCs, spleen cells from untreated DQ8 mice were incubated with pt-gliadins or -transamidated gliadins (200 g/ ml) for 4 h and then washed and irradiated (3500 Rad). Purified T cells (1.5 10 6 ) were incubated with 3.0 10 6 APCs for 8 or 72 h. Cytokine analysis For cytokine transcript analysis, total RNA was extracted from whole splenocytes, splenic T cells, Caco-2 cells, or intestinal biopsies using TRIzol reagent, according to the manufacturer s instructions (Invitrogen-Life Technologies). RNA was quantified by fluorometry using the RiboGreen reagent (Invitrogen-Life Technologies), and RNA quality was verified by denaturing gel electrophoresis. cdna was prepared from 1 g total RNA by reverse transcription with MMLV RT (Invitrogen-Life Technologies) and the oligo-(dt) 12 18 primer at 42 C for 60 min. Real-time PCR was performed on the icycler iq Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA, USA). For mouse samples, the amplification mixture (25 l) contained 1 Platinum Quantitative PCR SuperMix-UDG (Invitrogen-Life Technologies), 0.2 M each primer, 0.1 M probe, and cdna [23]. For human samples, the amplification mixture contained iq SYBR Green Supermix (Bio-Rad Laboratories), 0.2 M each primer and cdna. The following PCR reaction conditions were used for 39 cycles: 95 C for 30 s, 56.4 C for 30 s, and 72 C for 40 s. The relative gene expression levels were calculated using the comparative threshold method [24] and presented as the fold-change in gene expression after normalization to the L-32 housekeeping gene. For analysis of the human samples, the following primer sequences were used: L-32, forward 5=-CCTCAGCCCCTT- GAAGC-3=, reverse 5=-GCCCTTGAATCTTCTACGAACC-3=; IFN, forward 5=-TCAGCTCTGCATCGTTTTGG-3=, reverse 5=-GTTCCATTATCCGCTA- CATCTGAA-3=; IL-15, forward 5=-GGAGGCATCGTGGATGGAT-3=, reverse 5=-AACACAAGTAGCACTGGATGGAAA-3=. For cytokine protein assays of whole splenocytes and isolated spleen T cells in DQ8 mice, the culture supernatants were collected after 72 h and analyzed for IFN and IL-10 protein levels by in-house sandwich ELISA. Statistical analysis Differences among the various treatment groups in the levels of cytokine expression were determined by the Wilcoxon signed rank test. The dose-response effect of cytokine expression, data of cytotoxicity, proliferation assay, and transcription levels in intestinal biopsies were evaluated by one-way ANOVA and Tukey test post hoc analysis. In the biochemical experiments, Student s t-test was applied to compare differences. For all tests, P 0.05 was selected as the level denoting a statistically significant difference. RESULTS Transamidation of wheat flour produces soluble gliadins A previous study found that gliadins recovered from wheat flour after a 2-h transamidation reaction were unable to stimulate intestinal antigen-specific T cells derived from CD patients [15]. Interestingly, the alcohol-soluble protein fraction (gliadins) was reduced drastically in the transamidated flour (Fig. 1A), suggesting that the transamidation treatment modified the physical-chemical properties of the gliadins. To clarify this phenomenon, we exhaustively eliminated the water-soluble albumin/globulin fraction from the flour suspension before incubation with mtgase and the amine-group donor, K-C 2 H 5. Accordingly, we were able to identify a new soluble protein fraction (spf) at the end of the reaction. The production of spf was found to be dependent on the concentrations of mtgase and K-C 2 H 5 (Fig. 1B), suggesting that gliadin solubilization was caused by the formation of -( -glutamyl)-lysine isopeptide bonds and not by any intrinsic enzyme activity. The production of isopeptide bonds from the catalytic activity of mtgase dramatically decreased the gliadin yield to 17.38 0.67% (mean sd; Fig. 1C). Notably, the residual insoluble gliadins were further reduced to 7.63 0.53% after a second transamidation step, whereas a third step did not further modify the solubilized protein yield (Fig. 1C). We used a modified ninhydrin test to assess whether the residual insoluble components generated following the second enzymatic step were also subjected to isopeptide bond formation [19]. This assay provided a quantitative estimate of the number of newly incorporated -amino N groups into gluten. The results presented in Fig. 1D show that the number of -amino N groups was increased significantly in gliadins (36.45 0.89 vs. 31.65 1.25; P 0.047 nmoles/mg protein; treated flour vs. control flour). Following SDS- PAGE, the electrophoretic mobility of K-gliadins was not changed substantially, but the proportion of / gliadins having a MW 31 KDa was increased (Fig. 1E). Interestingly, we also identified a series of discrete bands in the 30- to 80-kDa MW range in the spf produced after the first enzymatic step. Loss of immune cross-reactivity of gliadins as a consequence of flour transamidation Western blot analysis performed by probing gliadin samples with an antigliadin polyclonal antibody showed that immune cross-reactivity was decreased dramatically in K-gliadins, which www.jleukbio.org Volume 93, April 2013 Journal of Leukocyte Biology 481

Figure 1. Biochemical and physical modifications of gliadins after transamidation.(a) Assessment of the residual alcohol-soluble protein fraction. (B) Dose-response effects of enzyme and amine-group donor (K-C 2 H 5 ) concentrations on the yield of the spf; protein values are expressed in mg/g flour. mtg, mtgase. (C) Residual yield of insoluble prolamins in the various enzyme steps. (D) Number of isopeptides in insoluble gliadin proteins assessed by evaluating the number of incorporated -amino N groups with a modified ninhydrin assay. (E) SDS-PAGE analysis of Coomassie-stained proteins isolated from flour after removal of the albumin/globulin fraction and following a two-step enzyme reaction. mwm, molecular weight markers. *P 0.05. were detectable only after a longer exposure time, and it was completely absent in spf (Fig. 2). Transamidated gliadins alter the specific immune response in DQ8 tg mice Next, we focused on the immunological effects of gliadins extracted from flour following a two-step transamidation process. To determine possible modifications in the T cellmediated response, we used DQ8-tg mice, which only express the human MHC class II molecule that has been Figure 2. Western blot (WB) analysis of the different protein fractions using an antigliadin polyclonal antibody probe. The same membrane was exposed at two different times (15 and 30 s) to highlight the loss of immunocrossreactivity caused by isopeptide bond formation. Gliadins isolated from untreated wheat flour were used as positive control (gliadins). linked to CD [22]. Moreover, mice were obtained from a GFD colony and lacked a naturally induced oral tolerance to gluten [25]. Following immunization with pt-gliadins, spleen cells were recovered and stimulated in vitro with ptgliadins or digests of spf and K-gliadins. The immune response was analyzed by a proliferation assay and by cytokine expression. Interestingly, K-gliadins and spf were found to be effective in stimulating spleen-cell proliferation, showing dose-response effects not significantly different from native gliadins (Fig. 3). Next, the cytokine profile produced in vitro following gliadin challenge was examined at the RNA and protein levels. As reported in Fig. 4, the amount of gliadin-specific IFN mrna, assessed after an 8-h culture, was decreased significantly in the presence of K-gliadins but not in the presence of spf. Moreover, no significant changes in the levels of inducible IL-10 transcripts were observed for either gliadin preparation (Fig. 4). In contrast, cytokine protein levels assessed after a 72-h culture indicated that spf and K-gliadins reduced IFN production (Fig. 5A). In addition, IL-10 protein levels decreased significantly following treatment with transamidated gliadins, although less dramatically than IFN levels (Fig. 5A). In agreement with these findings, the IL-10/IFN ratio was more heterogeneous when modified gliadins were added in vitro. Specifically, this ratio was significantly higher for spf-challenged spleen cells compared with cells challenged with native gliadins (Fig. 5B). Interestingly, in mice with a high IL-10/IFN ratio ( median values), dose-response experiments indicated a dose-dependent inhibitory effect of spf on the gliadin-specific expression of IFN, whereas K-gliadins were completely inhibited at the lowest tested dose (Fig. 6). For IL-10, 400 482 Journal of Leukocyte Biology Volume 93, April 2013 www.jleukbio.org

Lombardi et al. Transamidation selectively inhibits gliadin reactivity produced by IFN (Fig. 8C). Moreover, no additive effects were found following coincubation of gliadins (in any form) with IFN (Fig. 8C). We also found that a 48-h exposure to gliadins induced apoptosis in Caco-2 cells (Fig. 8D), in agreement with previous results [27]. Similarly, caspase-3 activity was increased following incubation with spf (Fig. 8D) and K-gliadins (data not shown). Interestingly, although apoptosis was not triggered by IFN exposure, gliadin exposure resulted in increased proapoptotic activity in the presence of this cytokine. Again, a similar increase was reported for spf (Fig. 8D) and K-gliadins (data not shown). Gliadin-mediated secretion of IL-15 by enterocytes is believed to play a role in the induction of mucosal damage in CD patients [28]. However, IFN, but not gliadins or spf, induced in vitro IL-15 transcription in differentiated Caco-2 cells, as assessed by real-time PCR analysis (Fig. 8E). A similar result was obtained for K-gliadins (data not shown). Finally, increased IL-15 mrna was observed when different forms of gliadins were combined with IFN, although this result was not statistically significant (Fig. 8E). Figure 3. Proliferation assay of whole splenocytes isolated from DQ8 tg mice. Spleen cells isolated from mice immunized with pt-gliadins (n 6) were challenged with pt of the various gliadin preparations and at different concentrations (200 400 g/ml; continuous lines, filled symbols). As negative control, splenocytes isolated from untreated mice (n 3) were similarly tested (dotted lines, empty symbols). *Significantly different from the cognate control, P 0.05. g/ml K-gliadins significantly increased the secretion of this cytokine (Fig. 6). A similar trend for IL-10 was reported for spf, although it was not statistically significant. We further analyzed the response to gliadins and their transamidated forms in the purified spleen T cells from immunized mice. Following in vitro stimulation with native gliadins, T cells showed increased levels of IFN mrna, paralleled by significant secretion of the cognate protein (Fig. 7). However, this was not associated to any induction of IL-10, in agreement with our previous findings [25]. Importantly, also on isolated T cells, K- gliadins and spf inhibited the gliadin-specific IFN production (Fig. 7). Flour transamidation does not modify the cytotoxic properties of gliadins Cytotoxic effects have been described previously for gliadins [26]. In agreement with these findings, we showed that a short incubation (4 h) with pt-gliadins was sufficient to produce a significant depletion in GST activity in differentiated Caco-2 cells (Fig. 8A). A similar inhibition resulted following cell exposure to pt of spf and K-gliadins but not to control proteins (wheat albumin/globulin), indicating that this phenomenon was gliadin-specific and was not influenced by transamidation (Fig. 8B). The extent of enzyme inhibition, measured after a 4-h incubation with gliadins or spf, was comparable with that Figure 4. Modulation of the antigen-specific cytokine mrna expression by gliadins isolated from transamidated flour in whole spleen cells from DQ8 tg mice. mrna levels were assessed by real-time PCR following an 8-h culture of spleen cells from the DQ8 mice immunized with pt-gliadins. Each dot represents values from a single mouse. Bars indicate median values. The reported results are representative of three independent experiments. www.jleukbio.org Volume 93, April 2013 Journal of Leukocyte Biology 483

Figure 5. Modulation of the cytokine protein pattern by gliadins isolated from transamidated flour in whole spleen cells from DQ8 tg mice. (A) Antigen-specific cytokine expression assessed after culturing spleen cells from immunized mice for 72 h (n 12). Each dot represents values (pg/ml) from a single mouse calculated as the difference between the means of triplicate cultures containing antigen and triplicate cultures with medium alone. (B) IL-10/IFN ratio calculated from the data in A to highlight changes in the phenotype of the immune response induced by modified gliadins. The results are representative of three independent experiments. (A and B) Bars indicate median values. Effects of gliadin transamidation on intestinal explants from CD patients Next, we examined the ability of modified gliadins to influence the immune response in vitro in intestinal biopsy samples obtained from CD patients with overt disease. In this model, ttgase-mediated deamidation of gliadin peptides is thought to occur in situ [29]. The demographic data and mucosal status of 10 consenting patients are reported in Table 1. We focused on IFN and IL-15 mrna expression levels quantitatively analyzed by real-time PCR after an 8-h in vitro challenge with pt of native and modified gliadins. As shown in Fig. 9A, samples from only four patients (CD2, -3, -7, and -10; Table 1) were responsive to the positive control (native gliadins), as assessed by IFN expression (arbitrary units 2). In addition, native gliadins induced enhanced expression of IL-15 mrna in samples from only two patients (CD2 and -9). The ability to induce a cytokine response in vitro did not appear to be related to the severity of the intestinal lesions or to any patient demographic features (Table 1). Accordingly, we restricted our evaluation to the statistical assessment of IFN transcription induced in vitro by transamidated gliadins. Notably, we found that K-gliadins dramatically reduced IFN mrna expression in all four responsive patient samples (Fig. 9B). In contrast, spf was able to decrease cytokine transcription significantly in only one patient sample (Patient CD2 in Table 1). Figure 6. Dose-response effects of modified gliadins on the expression of examined cytokines in whole splenocytes from DQ8 tg mice with a high IL-10/IFN ratio. Antigen-specific cytokine expression was assessed as indicated in Fig. 5. The results are representative of three independent experiments. Columns represent the means sd. **P 0.01; ***P 0.001. DISCUSSION In this study, we showed that the transamidation of wheat flour in the presence of K-C 2 H 5 produced two forms of crosslinked gliadins soluble and insoluble which selectively modified the immunogenicity of native gliadins, known to trigger 484 Journal of Leukocyte Biology Volume 93, April 2013 www.jleukbio.org

Lombardi et al. Transamidation selectively inhibits gliadin reactivity Figure 7. Effects of transamidated gliadins in splenic T cells from DQ8 tg mice.t cells were isolated from immunized mice and incubated with irradiated, pulsed syngenic spleen cells as indicated in Materials and Methods. The results are representative of six independent experiments. Columns represent the means sd. *P 0.05; ***P 0.001, significantly different from T cells treated with digested, native gliadins (gliadins). CD. A major hallmark of CD is inappropriate intestinal T cell activation in DQ2 and DQ8 patients triggered by peptides from wheat gluten and related prolamins from barley and rye. Moreover, ttgase plays two crucial roles in CD immune-mediated pathogenesis. Autocatalytic covalent complex formation of ttgase with gluten peptides has been described previously [30], and these complexes have been suggested to act as hapten carriers that are taken up by ttgase-specific B cells. Accordingly, gliadin peptides can be presented and recognized by gliadin-specific T cells, which aid B cells in secreting tt- Gase-specific antibodies [31]. Anti-tTGase autoantibodies are able to increase the transepithelial passage of gliadin peptides [32] and the permeability of blood vessels to macromolecules and lymphocytes [33]. In addition, several gliadin epitopes that elicit a CD4 -mediated T cell response have been identified and recognized following deamidation of specific glutamine residues by ttgase [34, 35]. Taken together, these findings strongly suggest that a strategy for preventing ttgase activity on gliadins could be instrumental in blocking the pathogenic mechanisms that occur in CD. In support of this hypothesis, we previously reported that cross-linking of K-CH 3 to gliadins by mtgase suppressed the immune response of deamidated, gliadin-specific intestinal T cell lines from CD patients [15]. In this study, we found that the solubility of gliadins was changed following mtgase treatment of wheat flour in the presence of K-C 2 H 5. Specifically, we identified a new spf. Gliadins are normally grouped into four types: 5 (50 kda)-, 1,2 (40 kda)-, and / - and (28 35 kda)-gliadins [36]. After transamidation treatment, the residual, insoluble gliadins that were obtained were modified by isopeptide bond formation (K-gliadins) without any apparent change in their electrophoretic mobility. In particular, spf contained subunits with a MW higher than their native counterparts, possibly reflecting a higher degree of lysine cross-linking in this fraction. Interestingly, these reported biochemical changes were associated with an almost complete loss of immune cross-reactivity toward antigliadin antibodies. This result strongly indicates that the transamidation of gluten components in the presence of K-C 2 H 5 causes effective epitope masking to occur. To determine the effects of modified gliadins on antigenspecific immunological activity, we examined two established models of gliadin sensitivity: DQ8-tg mice and intestinal biopsies from CD patients. In DQ8 mice, ttgase-mediated deamidation of gliadin is not a prerequisite for gliadin immunogenicity [25]. We found that whole splenocytes isolated from immunized GFD DQ8-tg mice were still able to proliferate in vitro in the presence of modified gliadins but had a dramatically altered cytokine pattern. After an 8-h incubation, K-gliadins significantly decreased the gliadin-specific transcription of IFN. The analysis of secreted cytokines after a 72-h culture confirmed the inhibitory activity of K-gliadins and of solubilized gliadins. In addition, both modified gliadins down-regulated the expression of IL-10, although the decrease was not as sharp as for IFN in most of the mice examined. Accordingly, a higher IL-10/IFN ratio was observed following treatment with modified gliadins. It is known that IFN and IL-10 are transcriptionally regulated [37, 38]. Moreover, recent evidences suggest that modulation of mrna stability is an important component in the regulation of expression of several cytokines, including IFN and IL-10 [39]. Therefore, discrepancies between the mrna and protein data reported in this study, mainly for IL-10, may be explained by differences in the posttranscriptional regulation induced by modified gliadins that contributed to the fine-tuning of tested cytokines. Nevertheless, dose-response experiments highlighted the specificity of the inhibition of modified gliadins on IFN expression. Notably, increased concentrations of transamidated gliadins led to a recovery in IL-10 expression in the same mice used in the previous experiments. By analyzing isolated splenic T cells, we confirmed that in GFD DQ8-tg mice, a Th1-like, antigen-specific response was generated using native gliadins as an immunogen [23, 25]. In a previous study, we demonstrated that this phenotype changed after subsequent rounds of in vitro stimulation, when the response of CD4 T cells was consistently characterized by an increase of IL-10 [25], resembling the condition herein detected in whole splenocytes. In particular, analysis of gliadin-primed T cells challenged with in vitropulsed APCs confirmed the inhibition of the expected Th1 response by both modified forms of gliadins, whereas IL-10 was not induced by any type of gliadin. Accordingly, the cellular source of antigen-specific IL-10 production remains to be defined in the present work. It is now known that the expression of IL-10 is not specific to Th2 cells or regulatory T cells but instead, that it can be expressed by different cells of the www.jleukbio.org Volume 93, April 2013 Journal of Leukocyte Biology 485

Figure 8. Effects of flour transamidation on gliadin cytotoxicity in enterocytes. (A) GST activity measured in differentiated Caco-2 cells exposed to pt-gliadins (1 mg/ml) for different time periods. (B) GST activity measured in differentiated Caco-2 cells challenged with digests of modified gliadins. Purified wheat albumin/globulin was used as a control. (C) Inhibition of GST activity following a 4-h incubation with 100 ng/ml IFN. (D) Caspase-3 activity determined in differentiated Caco-2 cells after 24- and 48-h incubations with different forms of gliadins in the absence or presence of IFN. (E) IL-15 mrna levels assessed as a marker of innate immunity by real-time PCR analysis. Differentiated Caco-2 cells were analyzed after a 4-h treatment with IFN alone (positive control) or different forms of gliadins in the absence or presence of IFN. Columns represent the mean values sd from triplicate analyses. *P 0.05; **P 0.01; ***P 0.001. The reported results are representative of two independent experiments. adaptive-immune system, including Th1 and Th17 cell subsets [40]. In addition to an autocrine-inhibitory effect of IL-10 on macrophages and DCs and because IL-10 can be produced by Th1, Th2, and Th17 cells, an additional feedback loop exists to limit the innate effector functions of macrophages and DCs and their subsequent activation of T cells. We then speculate that this loop was lost when isolated T cells were in vitro-challenged with APCs from untreated DQ8 mice but that it was operative and modulated by modified gliadins when whole splenocytes were examined. Together, these results underscore the finding that transamidated gliadins act as blockers of the specific IFN response toward native gliadins in DQ8-tg mice. Most importantly, in some mice, this was clearly associated with a reversal in the gliadin-inducible immune response phenotype from inflammatory to anti-inflammatory. By also considering our previous findings [15], we hypothesize that the transamidationmediated masking of immunodominant gliadin epitopes could favor the HLA binding of low-affinity gliadin-immune determinants with dramatic consequences on the outcome of the response. More specific studies are currently planned to definitively characterize these factors in the gliadin-transamidated fractions. The ability of transamidated gliadins to block the immune response was then tested in intestinal specimens derived from CD patients on a normal diet. This model is very useful, as pathogenic mechanisms, including the deamidation activity of ttgase, are still functioning in cultured biopsy samples [29, 41, 42]. However, not all patient samples were responsive to in vitro challenge with gliadins, which may reflect differences in the activation status of gut immunity toward gliadins that are not necessarily related to the severity of the mucosal lesions. In particular, we were able to induce transcription of IFN and IL-15 in only four patient samples and two patient samples, respectively. Notably, the inhibitory activity of K-gliadins, but not of spf, was confirmed in all four samples in which IFN was induced. These data were obtained after an 8-h challenge and were in good agreement with the RNA analysis data in mice. Moreover, in spite of the supposed ttgase-catalyzed hydrolysis of transamidated products [43], our results suggested that the isopeptide bonds produced after the transamidation reaction of wheat flour were effective in a biological system where ttgase activity is believed to occur [29]. Gliadin activity in CD patients is not restricted to the adaptive arm of the immune response, but direct involvement of gliadins in innate-immune mechanisms [28] and in cytotoxity [26] have been assumed. Both effects involve the epithelial cell lining of the intestinal mucosa. For this reason, we tested these activities in vitro using differentiated Caco-2 cells, a wellcharacterized enterocytic cell line [44]. Recently, we showed in these cells that gliadins induced depletion of the detoxifying activity of GST, which is characteristic of Phase 2 enzymes [27]. In this study, we confirmed the ability of gliadins to significantly suppress GST. Interestingly, transamidation did not modify this property, as K-gliadins and spf caused similar degrees of inhibition. Moreover, modified gliadins were able to induce apoptosis similarly to native gliadins, as assessed by an increase in caspase-3 activity that was enhanced further upon coexposure to IFN. Gliadinmediated secretion of IL-15 by enterocytes is believed to play a role in the induction of mucosal damage in CD patients [28]. Other authors have pointed out that this secretion could be secondary to gliadin-driven adaptive immunity [1]. Consistent with the latter observations, we found that 486 Journal of Leukocyte Biology Volume 93, April 2013 www.jleukbio.org

Lombardi et al. Transamidation selectively inhibits gliadin reactivity ACKNOWLEDGMENTS This study was supported by grants from the CNR (Annualità 2008). We thank Ajinomoto Foods Europe S.A.S. for generously supplying ACTIVA WM. We also thank Loredana Arciuolo for technical assistance. Figure 9. IFN and IL-15 mrna levels in intestinal biopsies from CD patients. (A) Intestinal specimens from 10 CD patients with overt disease (Table 1) were subjected to an 8-h incubation with pt-gliadins (1 mg/ml), and cytokine transcript levels were evaluated by real-time PCR. Cytokine values were normalized to L-32 mrna and are presented as the fold-change in gene expression (arbitrary units). Values 2 are considered to indicate induction of an immune response (dotted line). (B) IFN mrna levels in inducible biopsy samples following challenges with pt digests of native gliadins (gliadins), spf, or K-gliadins. Columns represent the mean values sd from triplicate analyses. *P 0.05. IFN, but not gliadins alone, induced IL-15 transcription in differentiated Caco-2 cells. A similar result was obtained with K-gliadins and spf. In conclusion, we found that the effects on the biological activity of gliadins following cross-linking of K-C 2 H 5 via mtgase were specifically restricted to the blockade of B and T cell epitope binding. Most importantly, these changes drastically influenced the immune activity of gliadins, reversing the inflammatory status in established models of sensitivity to these antigens. AUTHORSHIP M.R. prepared gliadin samples. G.I. collected human biopsies and provided information on clinical data. E.L., P.B., F.M., G.B., D.L., G.M., V.R.A., and M.R. performed experiments and acquired experimental data. P.B., G.M., and M.R. analyzed statistics and drafted the manuscript. All authors reviewed and approved the final version of the manuscript. REFERENCES 1. Di Sabatino, A., Corazza, G. R. (2009) Coeliac disease. Lancet 373, 1480 1493. 2. Rubio-Tapia, A., Kyle, R. A., Kaplan, E. L., Johnson, D. R., Page, W., Erdtmann, F., Brantner, T. L., Kim, W. R., Phelps, T. K., Lahr, B. D., Zinsmeister, A. R., Melton III, L. J., Murray, J. A. (2009) Increased prevalence and mortality in undiagnosed celiac disease. Gastroenterology 137, 88 93. 3. Lohi, S., Mustalahti, K., Kaukinen, K., Laurila, K., Collin, P., Rissanen, H., Mäki, M. (2007) Increasing prevalence of coeliac disease over time. Aliment. Pharmacol. Ther. 26, 1217 1225. 4. Green, P. H., Cellier, C. (2007) Celiac disease. N. Engl. J. Med. 357, 1731 1743. 5. Murray, J. A., Watson, T., Clearman, B., Mitros, F. (2004) Effect of a gluten-free diet on gastrointestinal symptoms in celiac disease. Am. J. Clin. Nutr. 79, 669 673. 6. Molberg, O., McAdam, S. N., Korner, R., Quarsten, H., Kristiansen, C., Madsen, L., Fugger, L., Scott, H., Norén, O., Roepstorff, P., Lundin, K. E., Sjöström, H., Sollid, L. M. (1998) Tissue transglutaminase selectively modifies gliadin peptides that are recognized by gut-derived T cells in celiac disease. Nat. Med. 4, 713 717. 7. Shan, L., Molberg, O., Parrot, I., Hausch, F., Filiz, F., Gray, G. M., Sollid, L. M., Khosla, C. (2002) Structural basis for gluten intolerance in celiac sprue. Science 297, 2275 2279. 8. Hausch, F., Shan, L., Santiago, N. A., Gray, G. M., Khosla, C. (2002) Intestinal digestive resistance of immunodominant gliadin peptides. Am. J. Physiol. Gastrointest. Liver Physiol. 283, G996 G1003. 9. Shan, L., Marti, T., Sollid, L. M., Gray, G. M., Khosla, C. (2004) Comparative biochemical analysis of three bacterial prolyl endopeptidases: implications for coeliac sprue. Biochem. J. 383, 311 318. 10. Cerf-Bensussan, N., Matysiak-Budnik, T., Cellier, C., Heyman, M. (2007) Oral proteases: a new approach to managing coeliac disease. Gut 56, 157 160. 11. Pyle, G. G., Paaso, B., Anderson, B. E., Allen, D. D., Marti, T., Li, Q., Siegel, M., Khosla C, Gray, G. M. (2005) Effect of pretreatment of food gluten with prolyl endopeptidase on gluten-induced malabsorption in celiac sprue. Clin. Gastroenterol. Hepatol. 3, 687 694. 12. Greco, L., Gobbetti, M., Auricchio, R., Di Mase, R., Landolfo, F., Paparo, F., Di Cagno, R., De Angelis, M., Rizzello, C. G., Cassone, A., Terrone, G., Timpone, L., D Aniello, M., Maglio, M., Troncone, R., Auricchio, S. (2011) Safety for patients with celiac disease of baked goods made of wheat flour hydrolyzed during food processing. Clin. Gastroenterol. Hepatol. 9, 24 29. 13. Kanaji, T., Ozaki, H., Takao, T., Kawajiri, H., Ide, H., Motoki, M., Shimonishi, Y. (1993) Primary structure of microbial transglutaminase from Streptoverticillium sp. Strain s-8112. J. Biol. Chem. 668, 11565 11572. 14. Yokoyama, K., Nio, N., Kikuchi, Y. (2004) Properties and applications of microbial transglutaminase. Appl. Microbiol. Biotechnol. 64, 447 454. 15. Gianfrani, C., Siciliano, R. A., Facchiano, A. M., Camarca, A., Mazzeo, M. F., Costantini, S., Salvati, V. M., Maurano, F., Mazzarella, G., Iaquinto, G., Bergamo, P., Rossi, M. (2007) Transamidation inhibits the intestinal immune response to gliadin in vitro. Gastroenterology 133, 780 789. 16. Mazzarella, G., Salvati, V. M., Iaquinto, G., Stefanile, R., Capobianco, F., Luongo, D., Bergamo, P., Maurano, F., Giardullo, N., Malamisura, B., Rossi. M. (2012) Reintroduction of gluten following flour transamidation in adult celiac patients: a randomized, controlled clinical study. Clin. Dev. Immunol. 2012, 329150. 17. Weegels, P. L., Hamer, R. J., Schofield, J. D. (1995) RP-HPLC and capillary electrophoresis of subunits from glutenin isolated by SDS and osborne fractionation. J. Cereal Sci. 22, 211 224. 18. Bradford, M. M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248 254. 19. Beckwith, A. C., Paulis, J. W., Wall, J. S. (1975) Direct estimation of lysine in corn meals by the ninhydrin color reaction. J. Agr. Food Chem. 23, 194 196. 20. Bergamo, P., Maurano, F., Rossi, M. (2007) Phase 2 enzymes induction by conjugated linoleic acid improves lupus-associated oxidative stress. Free Radic. Biol. Med. 43, 71 79. 21. Oberhuber, G., Granditsch, G., Vogelsang, H. (1999) The histopathology of coeliac disease: time for a standardized report scheme for pathologists. Eur. J. Gastroenterol. Hepatol. 11, 1185 1194. www.jleukbio.org Volume 93, April 2013 Journal of Leukocyte Biology 487

22. Cheng, S., Baisch, J., Krco, C., Savarirayan, S., Hanson, J., Hodgson, K., Smart, M., David, C. (1996) Expression and function of HLA-DQ8 (DQA1*0301/DQB1*0302) genes in transgenic mice. Eur. J. Immunogenet. 23, 15 20. 23. D Arienzo, R., Maurano, F., Luongo, D., Mazzarella, G., Stefanile, R., Troncone, R., Auricchio, S., Ricca, E., David, C., Rossi, M. (2008) Adjuvant effect of Lactobacillus casei in a mouse model of gluten sensitivity. Imm. Lett. 119, 78 83. 24. Livak, K. J., Schmittgen, T. D. (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2 CT method. Methods 25, 402 428. 25. Senger, S., Maurano, F., Mazzeo, M. F., Gaita, M., Fierro, O., David, C. S., Troncone, R., Auricchio, S., Siciliano, R. A., Rossi, M. (2005) Identification of immunodominant epitopes of -gliadin in HLA-DQ8 tg mice following oral immunisation. J. Immunol. 175, 8087 8095. 26. Elli, L., Dolfini, E., Bardella, M. T. (2004) Direct gliadin cytotoxicity as a cofactor in the pathogenesis of celiac disease. Int. Arch. Allergy Immunol. 134, 88. 27. Bergamo, P., Gogliettino, M., Palmieri, G., Cocca, E., Maurano, F., Stefanile, R., Balestrieri, M., Mazzarella, G., David, C., Rossi, M. (2011) Conjugated linoleic acid protects against gliadin induced depletion of intestinal defenses. Mol. Nut. Food Res. 55, S248 S256. 28. Meresse, B., Chen, Z., Ciszewski, C., Tretiakova, M., Bhagat, G., Krausz, T. N., Raulet, D. H., Lanier, L. L., Groh, V., Spies, T., Ebert, E. C., Green, P. H., Jabri, B. (2004) Coordinated induction by IL15 of a TCRindependent NKG2D signaling pathway converts CTL into lymphokineactivated killer cells in celiac disease. Immunity 21, 357 366. 29. Molberg, O., McAdam, S., Lundin, K. E., Kristiansen, C., Arentz-Hansen, H., Kett, K., Sollid, L. M. (2001) T cells from celiac disease lesions recognize gliadin epitopes deamidated in situ by endogenous tissue transglutaminase. Eur. J. Immunol. 31, 1317 1323. 30. Bruce, S. E., Bjarnason, I., Peters, T. J. (1985) Human jejunal transglutaminase: demonstration of activity, enzyme kinetics and substrate specificity with special relation to gliadin and coeliac disease. Clin. Sci. (Lond). 68, 573 579. 31. Fleckenstein, B., Qiao, S. W., Larsen, M. R., Jung, G., Roepstorff, P., Sollid, L. M. (2004) Molecular characterization of covalent complexes between tissue transglutaminase and gliadin peptides. J. Biol. Chem. 279, 17607 17616. 32. Rauhavirta, T., Qiao, S. W., Jiang, Z., Myrsky, E., Loponen, J., Korponay- Szabó, I. R., Salovaara, H., Garcia-Horsman, J. A., Venäläinen, J., Männistö, P. T., Collighan, R., Mongeot, A., Griffin, M, Mäki, M., Kaukinen, K., Lindfors, K. (2011) Epithelial transport and deamidation of gliadin peptides: a role for coeliac disease patient immunoglobulin A. Clin. Exp. Immunol. 164, 127 136. 33. Myrsky, E., Caja, S., Simon-Vecsei, Z., Korponay-Szabo, I. R., Nadalutti, C., Collighan, R., Mongeot, A., Griffin, M., Mäki, M., Kaukinen, K., Lindfors, K. (2009) Celiac disease IgA modulates vascular permeability in vitro through the activity of transglutaminase 2 and RhoA. Cell. Mol. Life Sci. 66, 3375 3385. 34. Tye-Din, J. A., Stewart, J. A., Dromey, J. A., Beissbarth, T., van Heel, D. A., Tatham, A., Henderson, K., Mannering, S. I., Gianfrani, C., Jewell, D. P., Hill, A. V., McCluskey, J., Rossjohn, J., Anderson, R. P. (2010) Comprehensive, quantitative mapping of T cell epitopes in gluten in celiac disease. Sci. Transl. Med. 2, 41 51. 35. Sollid, L. M. (2002) Coeliac disease: dissecting a complex inflammatory disorder. Nat. Rev. Immunol. 9, 647 655. 36. Wieser, H. (2007) Chemistry of gluten proteins. Food Microbiol. 24, 115 119. 37. Shnyreva, M., Weaver, W. M., Blanchette, M., Taylor, S. L., Tompa, M., Fitzpatrick, D. R., Wilson, C. B. (2004) Evolutionarily conserved sequence elements that positively regulate IFN- expression in T cells. Proc. Natl. Acad. Sci. USA 101, 12622 12627. 38. Tone, M., Powell, M. J., Tone, Y., Thompson, S. A., Waldmann, H. (2000) IL-10 gene expression is controlled by the transcription factors Sp1 and Sp3. J. Immunol. 165, 286 291. 39. Anderson, P. (2008) Post-transcriptional control of cytokine production. Nat. Immunol. 9, 353 359. 40. O Garra, A., Vieira, P. (2007) TH1 cells control themselves by producing interleukin-10. Nat. Rev. Immunol. 7, 425 428. 41. Salvati, V. M., Mazzarella, G., Gianfrani, C., Levings, M. K., Stefanile, R., De Giulio, B., Iaquinto, G., Giardullo, N., Auricchio, S., Roncarolo, M. G., Troncone, R. (2005) Recombinant human interleukin 10 suppresses gliadin dependent T cell activation in ex vivo cultured coeliac intestinal mucosa. Gut 54, 46 53. 42. Picarelli, A., Di Tola, M., Sabbatella, L., Anania, M. C., Calabrò, A., Renzi, D., Bai, J. C., Sugai, E., Carroccio, A., Di Prima, L., Bardella, M.T., Barisani, D., Ribes-Koninckx, C., Allaga, E. D., Gasparin, M., Bravi, E., Multicentre Organ Culture System Study Group (2006) Usefulness of the organ culture system in the in vitro diagnosis of coeliac disease: a multicentre study. Scand. J. Gastroenterol. 41, 186 190. 43. Stamnaes, J., Fleckenstein, B., Sollid, L. M. (2008) The propensity for deamidation and transamidation of peptides by transglutaminase 2 is dependent on substrate affinity and reaction conditions. Biochim. Biophys. Acta 1784, 1804 1811. 44. Beaulieu, J-F., Quaroni, A. (1991) Clonal analysis of sucrase-isomaltase expression in the human colon adenocarcinoma Caco-2 cells. Biochem. J. 280, 599 608. KEY WORDS: celiac disease T lymphocytes enterocytes 488 Journal of Leukocyte Biology Volume 93, April 2013 www.jleukbio.org