ORIGINAL ARTICLES ALIMENTARY TRACT. Safety for Patients With Celiac Disease of Baked Goods Made of Wheat Flour Hydrolyzed During Food Processing

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2011;9:24 29 ORIGINAL ARTICLES ALIMENTARY TRACT Safety for Patients With Celiac Disease of Baked Goods Made of Wheat Flour Hydrolyzed During Food Processing LUIGI GRECO,* MARCO GOBBETTI, RENATA AURICCHIO,* RAFFAELLA DI MASE,* FRANCESCA LANDOLFO,* FRANCESCO PAPARO,* RAFFAELLA DI CAGNO, MARIA DE ANGELIS, CARLO GIUSEPPE RIZZELLO, ANGELA CASSONE, GAETANO TERRONE,* LAURA TIMPONE,* MARTINA D ANIELLO,* MARIA MAGLIO,* RICCARDO TRONCONE,* and SALVATORE AURICCHIO* *Department of Pediatrics and European Laboratory for the Study of Food Induced Diseases, University of Naples, Federico II, Naples; and Department of Plant Protection and Applied Microbiology, University of Bari, Bari, Italy BACKGROUND & AIMS: Celiac disease (CD) is characterized by an inflammatory response to wheat gluten, rye, and barley proteins. Fermentation of wheat flour with sourdough lactobacilli and fungal proteases decreases the concentration of gluten. We evaluated the safety of daily administration of baked goods made from this hydrolyzed form of wheat flour to patients with CD. METHODS: Patients were randomly assigned to consumption of 200 g per day of natural flour baked goods (NFBG) (80,127 ppm gluten; n 6), extensively hydrolyzed flour baked goods (S1BG) (2480 ppm residual gluten; n 2), or fully hydrolyzed baked goods (S2BG) (8 ppm residual gluten; n 5) for 60 days. RESULTS: Two of the 6 patients who consumed NFBG discontinued the challenge because of symptoms; all had increased levels of anti tissue transglutaminase (ttg) antibodies and small bowel deterioration. The 2 patients who ate the S1BG goods had no clinical complaints but developed subtotal atrophy. The 5 patients who ate the S2BG had no clinical complaints; their levels of anti-ttg antibodies did not increase, and their Marsh grades of small intestinal mucosa did not change. CONCLUSIONS: A 60-day diet of baked goods made from hydrolyzed wheat flour, manufactured with sourdough lactobacilli and fungal proteases, was not toxic to patients with CD. A combined analysis of serologic, morphometric, and immunohistochemical parameters is the most accurate method to assess new therapies for this disorder. Keywords: Celiac Disease; Wheat Flour; Sourdough; Gluten Challenge. Celiac disease is characterized by an inflammatory response to wheat gluten. Gluten contains approximately 15% proline and 35% glutamine residues, 1 which limit proteolysis by gastrointestinal enzymes, thus generating toxic peptides. 2 Oral supplementation with microbial peptidases was proposed as an alternative to the gluten-free diet. 3,4 However, these enzymes are easily inactivated in the stomach by pepsin and acidic ph, thus failing to degrade gluten before exposure to small intestine. Hence a combination of glutamine-specific endoprotease and prolyl-endopeptidase, which rapidly detoxifies oligopeptides after primary proteolysis, was proposed. 5 A new endoprotease from Aspergillus niger accelerated the degradation of gluten into the stomach to such an extent that only traces reached the duodenal compartment. 6,7 Unfortunately, to date, the safety of the oral administration of proteases with gluten is yet to be demonstrated in patients with celiac disease (CD). Currently, cereal baked goods are manufactured by fast processes in which traditional long fermentation by sourdough, a cocktail of acidifying and proteolytic lactic acid bacteria, was replaced by chemical and/or baker s yeast leavening agents. Under these conditions, cereal components are not degraded during manufacture. We showed that the manufacture of wheat and rye breads or pasta with durum flours by using selected sourdough lactobacilli markedly decreased the toxicity of gluten. 8 A combination of glutamine-specific endopeptidase, which extensively hydrolyzed gluten, and prolyl-endopeptidases, which rapidly detoxified gluten, was recently proposed. 9 Proteins from the pepsin-trypsin digest of the wheat sourdough did not activate peripheral blood mononuclear cells (PBMCs) or induce interferon- synthesis by PBMCs. None of the intestinal T-cell lines from 12 patients with CD showed immunoreactivity toward pepsin-trypsin digest. Bread making was optimized to show the processing suitability of the detoxified wheat flour. This study evaluated the safety of daily administration for 60 days to CD patients of goods made of wheat flour hydrolyzed during food processing by a mixture of selected sourdough lactobacilli and fungal proteases. Materials and Methods Microorganisms and Enzymes Lactobacillus alimentarius, 15M, L brevis 14G, L sanfranciscensis 7A, and L hilgardii (defined as pool S1) were shown to hydrolyze gliadins efficiently. 8 Six strains of L sanfranciscensis Abbreviations used in this paper: CD, celiac disease; ELISA, enzymelinked immunosorbent assay; EMA, endomysial antibody; Ig, immunoglobulin; NFBG, natural flour baked goods; PBMC, peripheral blood mononuclear cell; S1BG, sourdough 1 baked goods; ttg, tissue transglutaminase by the AGA Institute /$36.00 doi: /j.cgh

2 January 2011 SAFETY OF BAKED GOODS FOR PATIENTS WITH CD 25 (LS3, LS10, LS19, LS23, LS38, and LS47) (named pool S2) were selected for their peptidase systems, specifically toward prolinerich peptides. Lactobacilli strains were isolated from natural wheat sourdoughs traditionally used for typical Italian bread making. Strains were propagated for 24 hours at 30 C in MRS broth (Oxoid, Basingstoke, Hampshire, UK) with fresh yeast extract (5% vol/vol) and 28 mmol/l maltose at ph of 5.6 and cultivated to the late exponential phase of growth ( 12 hours). Fungal proteases from Aspergillus oryzae (500,000 hemoglobin units) and A. niger (3.00 acid protease units/g), used as improvers in bakery industry, were supplied by BIO- CAT Inc (Troy, VA). Hydrolysis of Gluten During Wheat Sourdough Fermentation Wheat (Triticum aestivum cv. Appulo) flour had moisture 12.8%, protein, and 10.3% of dry matter. Fermentation formulas were sourdough 1 (S1), 80 g of wheat flour and 320 g of water containing colony-forming units/g (final cell density in the dough) of pool 1 and sourdough 2 (S2), the same as S1 formula with the addition of colony-forming units/g of pool 2 and 200 ppm of both fungal proteases. Sourdoughs S1 and S2 were fermented for 48 hours at 37 C. Spray drying then removed water, and milled flours were used to produce baked goods. Making Sweet Baked Goods The formulas to make biscuits and cakes were sourdough 1 (S1) baked goods (S1BG), 100 g of S1 hydrolyzed wheat flour, 50 g of butter, 50 g of sucrose, and water as required; sourdough 2 (S2) baked goods (S2BG), the same as S1BG formula except for the use of S2 instead of S1; and nonfermented baked goods (natural flour baked goods [NFBG]), the same as S1BG formula except for the use of nonprocessed wheat (Triticum aestivum cv. Appulo) flour. Sweet goods were of excellent taste for all 3 flours used. Immunologic and Chemical Analyses R5 monoclonal antibody and the horseradish peroxidase-conjugated R5 antibody were used for gluten analysis. The R5-based sandwich enzyme-linked immunosorbent assay (ELISA) was carried out with the Transia plate detection kit (Diffchamb; Västra, Frölunda, Sweden) at Centro National de Biotecnologia (Madrid, Spain). For R5-based Western blot analysis, after one-dimensional sodium dodecylsulfate polyacrylamide gel electrophoresis, proteins were electrotransferred onto polyvinylidene difluoride membranes and incubated with R5- horseradish peroxidase, and the blots were developed by immunodetection with enhanced chemiluminescence system (Amersham Pharmacia, Piscataway, NJ). Patients Sixteen CD patients (median age, 19 years; range, years) accepted to enter the study. All patients were in good health on a gluten-free diet for at least 5 years. Challenge Protocol Gluten challenge to confirm the initial diagnosis was part of the diagnostic work up to few years ago, but we limited the gluten challenge to children with very young age or uncertain initial diagnosis. Gluten challenge was frequently requested by teenagers who wished to confirm the gluten dependency after many years of gluten-free diet. When the request was denied, most teenagers did the challenge themselves, without controls. This 25-year strategy led to one of the best levels of compliance to the gluten-free diet. The study protocol, approved by the Ethical Committee of the University Federico II, Italy, was proposed to teenagers on a waiting list for a diagnostic gluten challenge. The protocol was thoughtfully explained, and their direct informed consent was obtained in all cases. Patients were on strict gluten-free diet during challenge. Anti tissue transglutaminase (ttg) and anti-endomysial (EMA) antibodies were evaluated, and a small intestinal biopsy was carried out before the challenge. Four patients were excluded for the presence of anti-ttg antibodies or damaged duodenal mucosa. Twelve patients were randomized to treatments by random number. The first 6 patients consumed NFBG, the following 2 patients consumed S1BG, and the last 5 patients consumed S2BG. All consumed daily about 200 g of baked goods, corresponding to 8 g of native gluten. The challenge lasted 60 days. After 30 days, patients underwent a dietary interview, and anti-tg2 and anti-ema antibodies were evaluated. At 60 days, biopsy and serology were repeated. Serology, Morphometric, and Immunohistochemistry Analyses Serum IgA EMA was detected by indirect immunofluorescence. 10 Serum immunoglobulin (Ig) A anti-tg2 was identified by ELISA technique (Eurospital, Trieste, Italy). 10 One fragment from each biopsy was fixed in 10% formalin, paraffin embedded, sectioned, and stained with hematoxylineosin. A fragment was embedded in OCT compound (Bio Optica, Milano, Italy) for immunohistochemistry. Villous to crypt height ratio was computed, and the number of intraepithelial lymphocytes, CD3, and TCR was computed per millimeter of epithelium. Marsh Oberhuber grading was applied. Immunohistochemistry was done by standard procedures. 11 Duodenal biopsies from 12 of 13 patients were investigated for the presence of intestinal deposits of IgA anti-tg2 antibodies, as previously reported. 12 To evaluate crypt proliferation, we detected Ki-67 in 11 of 13 biopsies. After preincubation of 10 minutes with rabbit normal serum (1:100; Dako, Glostrup, Denmark), the primary mouse monoclonal antibody Ki-67 (1:200; Dako) was applied on 4- m cryostat sections for 1 hour. Rabbit anti-mouse (1:25; Dako) was then applied for 30 minutes, followed by a 30-minute step with alkaline phosphatase and monoclonal mouse anti alkaline phosphatase immunocomplexes (mouse APAAP 1:40; Dako). Incubation with fuchsin was the ending step. Crypt proliferation index was obtained by counting cells positive for anti Ki-67 antibody as a percentage of the total enterocytes within vertical crypts. Five sections with well-oriented crypts were analyzed per patient. Statistical Analysis Wilcoxon signed-rank test for paired (before/after) data was adopted (Tables 1 and 2).

3 26 GRECO ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 9, No. 1 Table 1. Parameters of CD Patients Before/After Challenge With NFBG Time of challenge (d) EMA TG2 (U/mL) Marsh grade Vh/Cd CD3 (cells/mm epithelium) a TCR (cells/mm CD25 (cells/mm 2 epithelium) b lamina propria) c A.An 0 Negative 5.6 T Light positivity 12.4 T A.Am 0 Negative 1.7 T Positive 18.8 T S.R. 0 Negative 1.6 T Positive 200 T3b F.I. 0 Light positivity 1.9 T Positive T3b G.S. 0 Negative 0.3 T Positive 200 T3b L.R. 0 Negative 0.5 T Positive T3b P value, Wilcoxon test, 0 vs 60 days NOTE. Wilcoxon signed-rank test for paired data (before/after challenge) was performed. P value statistically significant if.05. T0, normal duodenal mucosa; T1, architecturally normal duodenal mucosa with increased intraepithelial lymphocyte infiltration; T2, presence of crypt hyperplasia; T3b, subtotal atrophy; Vh/Cd, villous to crypt height ratio. a Normal value, 34. b Normal value, 3.4. c Normal value, 4. Results Protein Content of the Wheat Flours Used for Making Baked Goods The gluten concentration of native nonfermented wheat (Triticum aestivum cv. Appulo) flour used for making NFBG was 80, ppm (Table 3). After fermentation of wheat flour by selected sourdough lactobacilli of pool 1 (sourdough S1), the residual concentration of gluten markedly decreased ( 97%) to ppm. The residual gluten was below 10 ppm when hydrolysis was carried out by combining sourdough S1 with fungal enzymes routinely used in bakery. The absence of gluten was confirmed by R5 antibody based Western blotting (data not shown). The spray-dried flour used for making S2BG contained a mixture of water/salt soluble low-molecular mass peptides and free amino acids (1.80% 0.06%). Indeed, no organic nitrogen was detectable in the gliadin fraction, and the level of glutenins decreased to traces (0.05% 0.01%). The concentration of free amino acids confirmed the differences found for hydrolysis of gluten; it varied from (NF) to (S1) and 15, mg/kg (S2). Clinical Challenge With Nonfermented Baked Goods Six CD patients consumed daily 200 g of NFBG that contained 80, ppm of gluten. Two patients (A.An and A.Am) interrupted the challenge after 4 weeks because of Table 2. Anti-TG2 IgA Deposits in Intestinal Mucosa TG2 deposits at time 0 TG2 deposits at time 60 2 Natural gluten 1/5 5/5 P.009 Extensively hydrolyzed 0/2 2/2 P.045 Fully hydrolyzed 2/5 1/5 P.49 symptoms such as malaise, abdominal pain, and diarrhea. Four patients reached the end point of 60 days with no complaints. All 6 showed a marked increase in anti-ema and anti-tg2 antibodies (Table 1). Biopsies showed significant deterioration with lymphocyte infiltration, increased CD3 and gamma-delta lymphocytes. Subtotal mucosal atrophy developed in all cases (Table 1). Data after challenge were significantly different by prechallenge values (Table 1). Challenge With Sourdough 1 Baked Goods Two CD patients consumed 200 g of S1BG that contained ppm of residual gluten. They had no clinical complaints during the 60 days. One showed increased antibodies (Table 4), and both showed increased CD3 and gamma-delta intraepithelial lymphocytes with subtotal atrophy after challenge. Challenge With Sourdough 2 Baked Goods Five patients consumed 200 g of S2BG made of wheat flour fermented with sourdough pools 1 and 2 and fungal proteases. None of the 5 patients had clinical complaints during the 60 days. None produced anti-tg2 antibodies; none had any modification of the small intestinal mucosa (Table 4). No increase of CD3 and gamma-delta cells was observed. Marsh grade was unchanged after the challenge. No significant changes before/after the challenge in any of the variables were observed (Table 4) (Figure 1). Anti-tissue Transglutaminase Immunoglobulin A Deposits in the Intestinal Mucosa and Crypt Proliferation Table 2 shows that all CD patients challenged with native gluten developed deposits of TG2 IgA in the mucosa; similarly, both patients who ate extensively hydrolyzed gluten showed clear IgA deposits. On the contrary, CD patients who ate fully hydrolyzed flour showed, after 60 days, fewer deposits than those observed at time 0.

4 January 2011 SAFETY OF BAKED GOODS FOR PATIENTS WITH CD 27 Table 3. Protein Nitrogen of the Flours Albumins/globulins (%) Gliadins (%) Glutenins (%) Free amino acids (mg/kg) Gluten (ppm) NF , Extensively hydrolyzed S Fully hydrolyzed S ND , NOTE. Data are means from 4 independent samples analyzed twice. ND, not detected. Ki-67, a specific marker of crypt proliferation, was evaluated in 11 of 13 biopsies. Figure 2 shows that both the first 2 groups (native and extensively hydrolyzed gluten) had a clear increase in crypt proliferation after 60 days, whereas no significant increase in crypt proliferation was observed in the third group (fully hydrolyzed gluten). Discussion Most studies on the development of therapeutic strategies alternative to the gluten-free diet are based on indirect outcome such as fecal output, permeability, nutritional and functional markers. Unfortunately, it is impossible to evaluate toxicity if the product is not tested in vivo for an acceptable length of time. This study tests the toxicity of a new wheat-based food for CD patients in vivo with the gold standard method. It might be argued that the first group of patients underwent a challenge with native flour whose toxicity was expected; the patients, before random allocation, requested an appropriate gluten challenge to confirm the diagnosis for life. The use of food-grade sourdough lactobacilli for extended fermentation of wheat flour is an example of biotechnology derived by traditional processes that, unfortunately, were recently replaced by fast-acting chemical and/or baker s yeast leavening agents. 8 Fermentation with selected lactobacilli added with fungal proteases, routinely used as an improver in bakery industries, decreased the concentration of gluten to below 10 ppm. Despite the markedly reduced concentration of gluten, the resulting spray-dried flour was still adequately workable. As shown in this and other studies, 9,13 the hydrolyzed flour is suitable for making sweet baked goods and also bread and pasta if supplemented with gluten-free structuring agents. As expected, 60 days of consumption of native gluten containing wheat flour (NFBG) were enough to elicit clinical symptoms and intestinal damage in CD patients. The ingestion of baked goods (S1BG) made of wheat flour with markedly decreased concentration of gluten (3% of the native content, 2480 ppm) did not provoke symptoms, but it induced TG2 antibodies to increase in 1 patient and was sufficient to increase inflammation in the small intestinal mucosa. Five patients who completed the 60-day challenge with baked goods made of wheat flour extensively digested by lactobacilli and fungal proteases showed no clinical symptoms, neither an increase of anti-tg2 antibodies nor a modification of the architecture or the grade of inflammation of the intestinal mucosa. Crypt proliferation as well as anti-tg2 IgA deposits in the mucosa did not increase after challenge with fully hydrolyzed gluten, whereas it clearly increased in the other 2 groups. This study showed that it is possible to evaluate completely the clinical, serologic, and histologic features in CD patients challenged with test meals. D-xylose urine secretion and 72- hour quantitative fecal fat were proved to be inadequate to detect intestinal damage in CD patients undergoing new treatments such as oral administration of prolyl-endopeptidases. 4 Table 4. Parameters of CD Patients Before/After Challenge With S2BG Time of challenge (d) TG2 (U/mL) Marsh grade Vh/Cd CD3 (cells/mm epithelium) a TCR (cells/mm CD25 (cells/mm 2 epithelium) b lamina propria) c F.I T T I.C T T R.R T T I.I T T C.C T T P value, Wilcoxon test, 0 vs 30 days NOTE. Wilcoxon signed-rank test for paired data (before/after challenge) was performed. P value statistically significant if.05. T0, normal duodenal mucosa; T1, architecturally normal duodenal mucosa with increased intraepithelial lymphocyte infiltration; Vh/Cd, villous to crypt height ratio. a Normal value, 34. b Normal value, 3.4. c Normal value, 4.

5 28 GRECO ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 9, No. 1 Figure 1. (A, B) Density of intraepithelial lymphocytes in jejunal biopsy from patient F.I. at the beginning (A) and after 60 days of challenge (B) with S2BG. (C, D) Density of intraepithelial lymphocytes in jejunal biopsy from patient S.A. at the beginning (A) and after 60 days of challenge (B) with S1BG. Arrows refer to gd intraepithelial lymphocytes. Our group showed that increased intestinal permeability to sugars and even increase of EMAs were good predictors of histologic relapse in gluten challenge of 15 CD patients. 13 Gliadin antibodies increased early (already 7 days after the reintroduction of gluten into the diet), but in many cases they returned to normal thereafter. Most importantly, all patients relapsed after 2 3 months of gluten-containing diet. The same results occurred in our cohort of patients who underwent the challenge with native gluten (NFBG). It is very important to realize that indirect methods of ascertainment of gluten-induced damage are not adequate to evaluate the long-term tolerance to gluten. Any short-term trial for few weeks is more Figure 2. Ki-67 detection as an index of crypt proliferation. Both the first 2 groups (NFBG and S1BG) had a statistically significant increase (*) in crypt proliferation after 60 days, whereas no increase in crypt proliferation was observed in the third group (S2BG). likely to originate frustrated hopes in the patients than real solutions. We showed that markedly reduced levels of gluten (3% of the native gluten, eg, 2480 ppm) are able to induce changes in the intestinal mucosa without clinical symptoms. Two patients developed villous atrophy after 60 days of challenge with the above concentration of gluten. However, serologic tests (antittg and anti-ema) were less sensitive, as already noted in children with slight dietary transgression to the gluten-free diet. Recently, Catassi et al 14 showed that a daily dose of 0.05 g of gluten for 90 days induced intraepithelial infiltration in CD patients with no antibody response. The recent discovery of immunodominant gluten peptides (eg, 33-mer epitope) resistant to intestinal digestion led to the identification of a novel opportunity for reducing gluten toxicity, such as bacterial-derived endopeptidases. However, there are still problems with the delivery of phosphoenolpyruvate enzymes and with their stability under acidic gastric environment and efficient mixing with gluten. We showed that the small amount of residual gluten after extensive fermentation (preparation S1BG) was sufficient to start the deleterious immune response to gluten. This is the reason why the administration of oral enzymes with a gluten meal is theoretically not likely to suppress the immune response to gluten. Under optimal conditions, healthy humans are able to digest up to 90.3% of the wheat protein from bread. 15 When 8000 mg of gluten (150 g of bread) are given to healthy subjects, 800 mg (10%) of

6 January 2011 SAFETY OF BAKED GOODS FOR PATIENTS WITH CD 29 this gluten will resist duodenal and jejunal degradation and will be presented to intestinal mucosa. This undigested fraction is likely to contain the digestion-resistant immunodominant epitopes of gluten. The oral administration of endopeptidases has to be 100 times more efficient than the physiological digestive machinery, during a small time lag, to be effective. This is the obvious reason why the administration of massive amounts of purified recombinant prolyl-endopeptidase together with 4 g of gluten (10% of the daily dose) to glutensensitive rhesus macaques did not prevent a marked increase of anti-gliadin and anti-ttg antibodies. 16 Bethune et al 16 showed that the administration of peptidases with the meal did currently enhance the immunoresponse by exposing the small intestinal mucosa more rapidly to immunopeptides from fragmented gluten proteins. In conclusion, the findings of this study and the above considerations provide the rationale for exploring therapies that could reduce the toxicity of gluten for CD patients beyond the standard gluten-free diet. A wheat flour derived product is shown to be not toxic after administration for 60 days to CD patients. Hydrolysis of gluten to below 10 ppm was achieved through the activity of complementary peptidases located in the cytoplasm of lactobacilli routinely used in sourdough fermentation. The addition of fungal proteases, used in bakery to modify the elasticity and resistance of gluten, was required to reach the complete gluten degradation. The primary proteolysis of wheat proteins was required; after this stage, medium-sized polypeptides, including the 33-mer and similar peptides, are efficiently transported into lactobacillus cytoplasm and subjected to peptidase activities. 9 A period of 60 days, although repeatedly shown to be sufficient to evaluate gluten toxicity in the majority of patients, might not be long enough to evaluate toxicity in all CD subjects who might show different sensitivity to gluten. Prolonged trials have to be planned to state the safety of the baked goods manufactured by applying this rediscovered and adapted biotechnology. An extended fermentation with lactic acid bacteria and fungal enzymes naturally present or used in sourdough biotechnology could have well lowered the exposure to massive amount of gluten in our past. There is no doubt that the phenotypic disclosure of CD, which appears today as a global epidemic, might be related to a threshold of early exposure. 11 In the future, the use of cereals through such biotechnology could also improve the nutritional and sensory properties of baked goods containing hydrolyzed gluten as compared with products made of naturally gluten-free ingredients. References 1. Stern M, Ciclitira PJ, van Eckert R, et al. Analysis and clinical effects of gluten in coeliac disease. Eur J Gastroenterol Hepatol 2001;13: Shan L, Molberg O, Parrot I, et al. Structural basis for gluten intolerance in celiac sprue. Science 2002;297: Shan L, Qiao SW, Arentz-Hansen H, et al. Identification and analysis of multivalent proteolytically resistant peptides from gluten: implication for celiac spue. J Prot Res 2005;4: Pyle GG, Paaso B, Anderson BE, et al. Effect of pretreatment of food gluten with prolyl endopeptidase on gluten-induced malabsortion in celiac spue. Clin Gastroenterol Hepatology 2005;3: Stepniak D, Spaenij-Dekking L, Mitea C, et al. Highly efficient gluten degradation with a newly identified prolyl endoprotease: implications for celiac disease. Am J Physiol Gastrointest Liver Physiol 2006;291:G621 G Mitea C, Havenaar R, Drijfhout JW, et al. Efficient degradation of gluten by a prolyl endoprotease in a gastrointestinal model: implications for coeliac disease. Gut 2008;57: Gass J, Bethune MT, Siegel M, et al. Combination enzyme therapy for gastric digestion of dietary gluten in patients with celiac sprue. Gastroenterology 2007;133: Di Cagno R, De Angelis M, Auricchio S, et al. Sourdough bread made from wheat and nontoxic flours and started with selected lactobacilli is tolerated in celiac spue patients. Appl Environ Microbiol 2004;70: Rizzello CG, De Angelis M, Di Cagno R, et al. Highly efficient gluten degradation by lactobacilli and fungal proteases during food processing: new perspectives for celiac disease. Appl Environ Microbiol 2007;73: Paparo F, Petrone E, Tosco A, et al. Clinical, HLA, and small bowel immunohistochemical features of children with positive serum antiendomysium antibodies and architecturally normal small intestinal mucosa. Am J Gastroenterol 2005;100: Vader W, Stepniak D, Kooy Y, et al. The HLA-DQ2 gene dose effect in celiac disease is directly related to the magnitude and breadth of gluten-specific T cell responses. Proc Natl Acad Sci U S A 2003;100: Tosco A, Maglio M, Paparo F, et al. Immunoglobulin A anti-tissue transglutaminase antibody deposits in the small intestinal mucosa of children with no villous atrophy. J Pediatr Gastroenterol Nutr 2008;47: Troncone R, Caputo N, Micillo M, et al. Immunologic and intestinal permeability tests as predictors of relapse during gluten challenge in childhood coeliac disease. Scand J Gastroenterol 1994;29: Catassi C, Fabiani E, Iacono G, et al. A prospective, double-blind, placebo-controlled trial to establish a safe gluten threshold for patients with celiac disease. Am J Clin Nutr 2007;85: Bos C, Juillet B, Foulillet H, et al. Postprandial metabolic utilization of wheat proteins in humans. Am J Clin Nutr 2005; 81: Bethune MT, Ribka E, Khosla C, et al. Transepithelial transport and enzymatic detoxification of gluten in gluten-sensitive rhesus macaques. PLoS ONE 2008;3:e1857:1 7. Reprint requests Address requests for reprints to: Prof. Luigi Greco, MD, PhD, Department of Pediatrics, Via S. Pansini 5, Edificio 11, Naples, Italy. ydongre@unina.it; fax: Acknowledgments The authors acknowledge Giuliano, Inc, Italy for participation in the project. Conflicts of interest The authors disclose no conflicts.