Limited Efficiency of Prolyl-Endopeptidase in the Detoxification of Gliadin Peptides in Celiac Disease

GASTROENTEROLOGY 2005;129:786 –796

CLINICAL–ALIMENTARY TRACT Limited Efficiency of Prolyl-Endopeptidase in the Detoxification of Gliadin Peptides in Celiac Disease TAMARA MATYSIAK–BUDNIK,* CELINE CANDALH,* CHRISTOPHE CELLIER,‡ CHRISTOPHE DUGAVE,§ ABDELKADER NAMANE,储 TERESITA VIDAL–MARTINEZ,* NADINE CERF–BENSUSSAN,* and MARTINE HEYMAN* *INSERM EMI-0212, Faculté Necker-Enfants Malades, Paris, France; ‡Hôpital Européen Georges Pompidou, Paris, France; §Laboratoire DIEP, CEA/Saclay, Gif sur Yvette, France; and 储Institut Pasteur, Paris, France

See editorial on page 1111. Background & Aims: The resistance of prolamines to digestive enzymes is thought to be a key contributor to the pathogenesis of celiac disease by promoting the intestinal entrance of peptides able to trigger inflammation in at-risk individuals. Oral administration of a bacterial prolyl-endopeptidase (PEP) therefore was proposed as a treatment for celiac disease. To delineate the feasibility of this treatment, the effect of PEP on gliadin peptides was assessed in vitro, and ex vivo during their transport across intestinal biopsy specimens of active celiac disease patients. Methods: In vitro degradation by PEP of 3H-labeled gliadin peptides 56 – 88 (33-mer) and 31– 49, was analyzed by radio–reverse-phase high-performance liquid chromatography and mass spectrometry. For ex vivo studies, PEP and 3H-peptides were added together onto the mucosal side of duodenal biopsy specimens mounted in Ussing chambers, and peptide transport and digestion were assessed by radio–reverse-phase high-performance liquid chromatography. Results: Gliadin peptides were degraded partly by 20 mU/mL PEP both in vitro and ex vivo. This concentration of PEP decreased the amount of intact peptides 31– 49 and 56 – 88 crossing the intestinal biopsy specimens of celiac disease patients, but could not prevent the intestinal passage of toxic or immunostimulatory metabolites. PEP concentrations of at least 500 mU/mL for 3 hours were required to achieve full detoxification of peptides and to prevent intestinal transport of active fragments. Conclusions: After prolonged exposure to high concentrations of PEP, the amount of immunostimulatory gliadin peptides reaching the local immune system in celiac patients is decreased. These results provide a basis to establish whether such conditions are achievable in vivo.

eliac disease (CD) is a small intestinal enteropathy induced by cereal-derived prolamines in genetically susceptible individuals. In the current view of the pathogenesis of CD, adaptive immunity plays a key role, accounting for the interplay between the triggering environmental factor, the prolamines, and the major genetic risk factor, the HLA-DQ2/8 haplotype. Gliadin peptides, because of their high proline and glutamine content, adopt a configuration that favors their binding into the peptide pocket of HLADQ2 molecules after their deamidation by tissue transglutaminase, the autoantibody target.1 They then can be presented to lamina propria CD4⫹ T cells, triggering their activation, the release of interferon ␥, and intestinal inflammation.2 Because of their high content in proline, prolamines show an unusual resistance to digestion by intraluminal and brush-border membrane enzymes. Models of intraluminal digestion, using recombinant ␣2-gliadin, identified a concatemer of 3 major T-cell epitopes embedded within a 33-mer peptide (56 – 88) that is highly resistant to digestion and strongly immunostimulatory for CD4⫹ T cells in CD patients.3 Recent work has provided evidence that innate immunity, orchestrated by the cytokine interleukin-15, also may participate in the induction of intestinal lesions in CD. Interleukin-15 promotes the hyperplasia of intraepithelial lymphocytes, the maturation of dendritic cells, and the presentation of gliadin T-cell epitopes to CD4⫹ lymphocytes in the intestinal lamina propria of CD patients.4 Interleukin-15 is overproduced massively in

C

Abbreviations used in this paper: CD, celiac disease; HPLC, highperformance liquid chromatography; MW, molecular weight; PEP, prolyl-endopeptidase; radio-RP-HPLC, radio–reverse-phase high-performance liquid chromatography; Rt, retention time. © 2005 by the American Gastroenterological Association 0016-5085/05/$30.00 doi:10.1053/j.gastro.2005.06.016

September 2005

active CD,5 and its expression can be triggered in organ cultures of biopsy specimens from CD patients on a gluten-free diet by the peptide 31– 49.4,6 This peptide, albeit not recognized by the CD4⫹ T cells, has been identified as toxic for the mucosa in CD patients by in vitro and in vivo studies.7,8 Although its mechanism of action remains to be elucidated, it may participate in CD pathogenesis by activating innate immune cells. The only current safe and efficient treatment of CD remains the gluten-free diet. However, it is very constraining, resulting in social burden and poor compliance, thereby warranting the search for alternative therapeutic strategies. Gliadins are characterized by a high content in proline, an amino acid that confers resistance to enzymatic cleavage by digestive enzymes.9,10 Shan et al3 therefore have proposed to use a bacterial prolyl-endopeptidase (PEP) to facilitate the hydrolysis of gliadin peptides. The studies in vitro, and in vivo in rats, using perfusion of intestinal loops, showed the disappearance of intact peptides in the presence of PEP using high-performance liquid chromatography (HPLC) with ultraviolet absorbance detection.3,11 The disappearance of intact peptides, however, does not mean complete degradation, and the presence of residual smaller but still immunogenic peptides was not excluded. In addition, the effect of PEP on the transport and processing of gliadin peptides by the intestinal mucosa of CD patients was not addressed in these studies. We recently showed that, in patients with active CD, an abnormal transcellular transport pathway allows the entrance of intact peptide 31– 49 and of immunostimulatory epitopes derived from peptide 56 – 88 into the intestinal mucosa.12 It therefore was necessary to verify that PEP can prevent the transepithelial transport of these gliadin peptides, which exert putative complementary roles in the pathogenesis of CD. To determine the kinetics and concentration of PEP that is necessary to degrade peptides 31– 49 and 56 – 88 fully, peptides were radiolabeled with 3H, and residual 3H-peptide metabolites were detected by radio–reverse-phase HPLC (radio-RP-HPLC), a highly sensitive method, and identified by mass spectrometry. In addition, the effect of PEP on intestinal transport and processing of both peptides was assessed ex vivo using duodenal biopsy specimens of CD patients mounted in Ussing chambers. Finally, the activity and resistance of PEP to extreme pH values and to degradation by pepsin and trypsin were analyzed.

Materials and Methods Synthesis and 3H-Radiolabeling of Peptides 31– 49 and 56 – 88 Radiolabeling of gliadin peptides allowed us to follow their fate during intestinal transport and to bypass the prob-

PROLYL-ENDOPEPTIDASE AND CELIAC DISEASE

787

lem of potentially contaminant endogenous peptide/proteins. Peptides 31– 49 (19-mer, LGQQQPFPPQQPYPQPQPF; molecular weight [MW], 2221) and 56 – 88 (33-mer, LQLQPFPQPQLPYPQPQLPYPQPQLPYPQPQPF; MW, 3903) were synthesized (Synt:em, Nimes, France) and were radiolabeled as described.12 Briefly, tritiation was performed using modified synthetic peptides 31– 49 and 56 – 88, in which proline 36 (peptide 31– 49) and 3 prolines: 64, 71, and 78 (33-mer), were replaced by a 3-4-dehydroproline. This labeling technique does not modify the native structure of the peptide.

PEP PEP from Flavobacterium meningosepticum (specific activity 53 U/mg) was purchased from United States Biological (Swampscott, MA).

In Vitro Study 3H-peptide 31– 49 or 3H-peptide 56 – 88 (.2 mg/mL, 1850 kBq) was incubated in Ringer’s solution at 37°C with different concentrations of PEP. Reactions were stopped by freezing the samples at ⫺80°C at different time points from 10 minutes to 24 hours. The analysis of peptides and their metabolites was performed by radio-RP-HPLC.

Ex Vivo Study Patients. Seven patients with active CD, 2 patients with CD in remission on a gluten-free diet, and 2 nonceliac control patients who underwent intestinal biopsy examination for a routine diagnostic purpose (chronic anemia) and presented with normal intestinal histology, were studied. Approval of the local ethics committee was obtained and all patients signed an informed consent to participate in the study. Transport of gliadin peptides across duodenal biopsy specimens mounted in Ussing chambers. Two to 6 duodenal biopsy specimens from each patient were mounted in adapted Ussing chambers, exposing a surface area of .025 cm2, as described.12–14 Briefly, biopsy specimens were bathed on each side with 1.2 mL of a Ringer’s solution that was regulated thermostatically, oxygenated, and maintained at a pH level of 7.4 by carbogen gas flow. The mucosal and serosal bathing solutions were connected via agar bridges to electrodes to monitor transmural potential differences, short-circuit current, and electrical resistance. As reported,12 biopsy specimens of patients with active CD presented with lower electrical resistance (R ⫽ 11.2 ⫾ 1.0 ⍀.cm2, n ⫽ 7) than those of patients with treated CD (R ⫽ 18 ⍀.cm2, n ⫽ 2) or of the non-CD controls (R ⫽ 19 ⍀.cm2, n ⫽ 2). The addition of PEP had no detectable effect on these parameters (not shown), showing its lack of toxicity for the intestinal mucosa. Each biopsy specimen was used to quantify mucosal to serosal fluxes of the radiolabeled peptide (either peptide 31– 49 or 56 – 88) and to analyze tritiated metabolites in the serosal compartment. The unlabeled peptides 31– 49 or 56 – 88 were placed on the mucosal side at a final concentration of .2 mg/mL, together with 1850 kBq (50 ␮Ci) of 3H-peptide. PEP was added at different concentrations (20 mU/mL, 500 mU/mL, or 1 U/mL)

788

MATYSIAK–BUDNIK ET AL

together with the peptide in the mucosal compartment of the biopsy specimen. 3H-equivalent peptide fluxes (ng/h.cm⫺2) were calculated by counting the radioactivity of 10 ␮L of the mucosal and serosal compartments at the end of the experiment. Epitopes generated during intestinal transport (and thus collected from the serosal compartment) were analyzed by radio-RP-HPLC. HPLC analysis of peptides and metabolites. Peptides and their metabolites formed after in vitro incubation with PEP (in vitro study) or present in the mucosal and serosal compartments of Ussing chambers (ex vivo study), were analyzed by radio-RP-HPLC, which separates peptides according to their hydrophobicity. A fraction (100,000 cpm) of the reaction mixture from the in vitro study or from the mucosal compartment of Ussing chambers was injected directly into the HPLC column. The entire fluid collected from the serosal compartment was injected after concentration to a final volume of 200 ␮L by speed-vac evaporation. RP-HPLC analysis. Analysis was performed by using a silice Uptisphere 300A 5 ␮m (250 mm) WRP column (Interchim, Montluçon, France) and a Shimadzu HPLC system (Shimadzu Corporation, Kyoto, Japan) coupled to a LB506 Berthold online counter (Berthold Technologies, Bad Wildbad, Germany) for radioactivity detection. The acquisition, integration, and calculation of data were performed with Winflow software, which provides information on the retention time of all peaks and the relative percentage of radioactivity in each peak, which allows an estimate of the proportion of each peptide in the sample. Samples were eluted at 36°C at a flow rate of 1 mL/min. Trifluoracetic acid was used as an ion pairing agent, and elution was performed with a gradient consisting of 100% buffer A to 100% buffer B over a 50-minute linear gradient. Buffer A consisted of .115% trifluoracetic acid in water and buffer B consisted of .1% trifluoracetic acid, 40% water, and 60% acetonitrile. In this setting, free proline (the radiolabeled amino acid in the peptide sequence) was eluted with a retention time (Rt) of 4 minutes. Intact peptides 31– 49 and 56 – 88 (33-mer) were eluted at 24.5 and 30.0 minutes, respectively. HPLC does not allow a direct measure of the absolute amount of each metabolite. Therefore, the relative quantity of each metabolite in the sample was expressed as a percentage of the total eluted radioactivity, calculated by using Winflow software as described earlier. Then, based on the values of 3H-equivalent peptide fluxes measured in ng/h · cm2, it was possible to estimate the amount of each metabolite crossing the tissue per hour and centimeter.2 Analysis of peptide metabolites by matrixassisted laser desorption ionization time of flight mass spectrometry. Mass spectrometry was used to determine the sequence of fragments remaining after in vitro digestion of peptides 31– 49 and 56 – 88 by PEP because the very small amount of peptides recovered in the serosal compartment of biopsy specimens in ex vivo studies precluded such analysis. After incubation of peptides 31– 49 or 56 – 88 with PEP (20 mU or 500 mU/mL) at 37°C for 3 hours, 100-␮L samples were injected into the HPLC column, and the following peptides

GASTROENTEROLOGY Vol. 129, No. 3

were collected: in the case of peptide 31– 49, peptides eluted between 17 and 21 minutes (corresponding to the nonreducible fragments at PEP 20 mU/mL) or those eluted between 11 and 14 minutes (corresponding to the nonreducible fragment at PEP 500 mU/mL); in the case of peptide 56 – 88, peptides eluted between 25 and 28 minutes (corresponding to the nonreducible fragments at PEP 20 mU/mL) or between 18 and 21 minutes (corresponding to the nonreducible fragment at PEP 500 U/mL). The samples were desalted by using adsorptive microconcentrators (microcon-SCX; Millipore, St. Quentin en Yvelines, France) and dried by speed-vac evaporation. After further desalting using Zip-Tip (Proteomic Solutions, St. Marcel, France), the samples were analyzed by mass spectrometry using the Voyager DE STR (PerSeptive Biosystems, Framingham, MA). Study of PEP resistance to extreme pH levels and to digestion by pepsin and trypsin. To facilitate intestinal digestion of gliadin peptides in vivo, PEP administered orally has to reach the intestinal lumen at sufficient concentrations. Therefore, its resistance to extreme pH levels and to degradation by gastric and pancreatic enzymes has to be taken into account. To address these questions, different concentrations of PEP (20, 100, 500, and 1000 mU/mL) were incubated sequentially in vitro with pepsin (pH, 1.8) and trypsin (pH, 7.4). At the end of each incubation period, the PEP activity in the samples was analyzed using the enzymatic assay described by Goossens et al.15 Similarly, PEP activity was studied after incubation at acidic pH levels of 1.8 for 4 hours.

Statistical Analysis Statistical analysis was performed using the SAS package (SAS Institute Inc, Cary, NC). The results are expressed as mean ⫾ SD and a comparison of different parameters among the groups was performed by using analysis of variance and nonparametric tests (Wilcoxon and Kruskal–Wallis). The general linear model procedure also was used for multiple groupto-group comparisons. The differences were considered significant for a P value of less than .05.

Results In Vitro Degradation of Peptides 31– 49 and 56 – 88 (33-mer) by PEP The kinetics of peptide digestion by PEP, expressed as a percentage of intact peptides remaining after incubation, are shown in Figures 1A and B. The degradation of peptide 31– 49 was observed at a PEP concentration of 40 ␮U/mL (40% and 100% degradation after 4 and 24 h, respectively; Figure 1A), and peptide 56 – 88 remained intact at this concentration for up to 24 hours (Figure 1B). In the presence of 20 mU/mL of PEP, intact peptide 31– 49 disappeared within 30 minutes, whereas peptide 56 – 88 disappeared more slowly (50% degraded after 1 h and approximately 100% after 2 h). At PEP concentrations greater than 100 mU/mL, peptide 31– 49

September 2005

PROLYL-ENDOPEPTIDASE AND CELIAC DISEASE

789

recovered. Mass spectrometry identified 12–15 mer sequences LQLQPFPQPQLP (peptide 56 – 67), LQLQPFPQPQLPYP (peptide 56 – 69), PQPQLPYPQPQL (peptide 62–73), and PQPQLPYPQPQLPY (peptide 62–75), corresponding to the known immunogenic epitopes embedded within peptide 56 – 88.16 Hydrolysis of peptide 56 – 88 was almost complete at PEP concentrations greater than 100 mU/mL because the largest metabolites were found in a peak (Rt, 19.7 min) identified by mass spectrometry as a pentapeptide (QPQLP or PQPQL; MW, 582), considered too short to be immunogenic. These in vitro results suggest that detoxification of gliadin peptides by PEP is possible, providing sufficient time of exposure and high PEP concentration. Effect of PEP on Transport and Processing of Peptides 31– 49 and 56 – 88 (33-mer) Across Duodenal Biopsy Specimens of Active CD Patients 3H-equivalent

Figure 1. In vitro degradation of (A) 3H-peptide 31– 49 and (B) 56 – 88 (33-mer) by PEP, expressed as the percentage of intact peptides remaining after incubation, as a function of time and PEP concentration. A dose- and time-dependent disappearance of intact peptides was observed that was more rapid for peptide 31– 49 than for peptide 56 – 88. ⽧, 40 U/mL; ●, 20 mU/mL; , 100 mU/mL;
, 500 mU/mL; Œ, 1 U/mL.

‘

disappeared within 10 minutes. A PEP concentration of 500 mU/mL was requested to digest peptide 56 – 88 within 10 minutes. HPLC analysis showed that disappearance of intact peptides did not reflect their complete degradation (Figure 2). After a 3-hour incubation with 20 mU/mL of PEP, 3Hmetabolites eluted between 17 and 20 minutes were recovered from peptide 31– 49. Mass spectrometry analysis of these fragments identified 9-mer (LGQQQPFPP, peptide 31–39) and 12-mer (LGQQQPFPPQQP, peptide 31– 42) sequences, the latter corresponding to a known toxic sequence of peptide 31– 49.4,7 However, with PEP 500 mU/ mL, only small-3H-fragments probably deprived of toxic properties (Rt, 11–14 min) were recovered. Similarly, after a 3-hour incubation of peptide 56 – 88 with 20 mU/mL PEP, large undigested fragments (Rt, 23–29 min) were

peptide fluxes. First, the

3H-

equivalent peptide 31– 49 or peptide 56 – 88 mucosal to serosal fluxes were quantified in the presence or absence of PEP in the mucosal compartment. These fluxes represent the total fluxes of radioactive material including intact peptide and degraded fragments. A significant increase of both peptide fluxes was observed in the presence of PEP at 500 mU/mL: 3H-equivalent peptide 31– 49 fluxes and 3H-equivalent peptide 56 – 88 fluxes were 7.7 ⫾ 2.3 and 10.0 ⫾ 5.5 ␮g/h · cm2, respectively, as compared with 4.5 ⫾ 1.9 and 3.5 ⫾ 1.9 ␮g/h · cm2 in the absence of PEP, respectively (Figure 3). Because this high concentration of PEP allows the complete digestion of the peptides (see later), the increase in total peptide fluxes might result from the release of amino acids, or of dipeptides and tripeptides in the mucosal compartment and their transport via the Na⫹-solute cotransport systems, which is much more efficient than transcytosis of larger peptides or proteins.17 HPLC Analysis of Peptides and Their Metabolites Peptide 31– 49. Figure 4 shows 1 HPLC profile of peptide 31– 49 that is representative of all studied patients with active CD; Figure 5 shows the proportion of each peak, expressed as the fraction of total eluted radioactivity (mean ⫾ SD) in 4 to 7 patients with active CD. In the absence of PEP (Figure 4A), peptide 31– 49 was not degraded in the mucosal compartment after a 3-hour incubation, with approximately 50% of the radioactive material in the serosal compartment being recovered under the form of intact peptide, as reported previously.12 As expected, PEP favored the hydrolysis of

790

MATYSIAK–BUDNIK ET AL

GASTROENTEROLOGY Vol. 129, No. 3

Figure 2. RP-HPLC analysis of 3H-peptide 31– 49 (left panels) and 3H-33-mer (peptide 56 – 88) (right panels) after a 3-hour incubation in vitro in the presence of different PEP concentrations. Incomplete degradation of both peptides, with release of several intermediate metabolites, was observed. With increasing PEP concentration, there was a shift of the radioactive peaks toward the left, indicating the disappearance of large hydrophobic fragments and the appearance of smaller 3H-fragments or free 3H-proline. Mass spectrometry analysis of fragments released after digestion of peptides 31– 49 and 56 – 88 by 20 mU/mL of PEP showed the presence of the toxic peptide 31– 42 and of several known T-cell epitopes (56 – 69, 57– 68, 62–75), respectively.

September 2005

Figure 3. Effect of PEP on mucosal to serosal 3H-equivalent peptides 31– 49 and 56 – 88 fluxes across duodenal biopsy specimens of patients with active CD, measured in Ussing chambers. Fluxes of both peptides are increased in the presence of 500 mU/mL of PEP. *Significantly different from peptide 31– 49 alone (P ⬍ .05). #Significantly different from peptide 56 – 88 alone (P ⬍ .004) and from peptide and 20 mU/mL of PEP (P ⬍ .02). □, Peptide alone (n ⫽ 7); s, PEP 20 mU/mL (n ⫽ 6);
, PEP 500 mU/mL (n ⫽ 4).

peptide 31– 49 (Figure 4B). In the mucosal compartment, in the presence of 20 mU/mL PEP (but not of 40 ␮U/mL, not shown), approximately 10% of the peptide was totally degraded and recovered as 3H-proline, whereas the remaining 90% was degraded into intermediate metabolites. In the serosal compartment, no intact peptide was found, 50% was degraded totally (3H-proline), and 50% was recovered as intermediate metabolites. Among these metabolites, 30% were eluted at 17–20 minutes, an elution time corresponding to a known toxic sequence of peptide 31– 49 (ie, peptide 31– 42) (Figure 4B). Almost total degradation of peptide 31– 49 was obtained with PEP at 500 mU/mL (Figure 4C) because over 90% of the radioactivity found in the serosal compartment corresponded to 3H-proline. These results indicate that PEP at 20 mU/mL allows the passage of potentially toxic fragments of peptide 31– 49 across the biopsy specimens of active CD patients. A very high concentration of PEP (500 mU/mL) is necessary to inhibit fully the transport of toxic fragments (Figure 5). Peptide 56 – 88 (33-mer). Similar results were obtained for peptide 56 – 88 (33-mer). Confirming our previous results,12 Figure 6A shows that peptide 56 – 88 is not degraded in the mucosal compartment after 3 hours, and that both intact peptide and fragments of a size compatible with immunostimulatory properties are recovered from the serosal compartment. The addition of 20 mU/mL PEP (but not of 40 ␮U/mL PEP, not shown) to the mucosal compartment (Figure 6B) induced a partial hydrolysis of peptide 56 – 88. Although the pas-

PROLYL-ENDOPEPTIDASE AND CELIAC DISEASE

791

sage of intact peptide 56 – 88 into the serosal compartment was prevented (5% ⫾ 8% compared with 19% ⫾ 15% in controls), an important fraction that crossed the mucosa (35% ⫾ 20%) contained fragments whose elution time (23–28 min) indicated a size compatible with immunostimulatory properties (Figure 7). More specifically, 1 nonreducible fragment eluting at a retention time of 26 minutes was recovered in the serosal compartment of biopsy specimens of all active CD patients. This elution time corresponds to that observed for the T-cell epitopes 56 – 69 or 62–75, identified by mass spectrometry in the in vitro study. Increasing PEP concentration in the mucosal compartment (ⱕ500 mU/mL or 1 U/mL) allowed the complete hydrolysis of peptide 56 – 88 into small nonimmunogenic fragments, corresponding to the pentapeptides QPQLP or PQPQL identified by mass spectrometry (Figure 6C). These results indicate that PEP can decrease the passage of immunogenic fragments across the intestinal mucosa of patients with active CD, but that a very high concentration is necessary to achieve this effect. Peptide Transport in Control Patients Confirming our previous results,12 peptide 31– 49 and 33-mer were degraded completely during their mucosal to serosal transport in the 2 nonceliac control patients and in the 2 patients with treated CD included in this study (95% of the radioactive material detected in the serosal compartment after the 3-hour incubation was recovered as 3H-proline). Subsequently, PEP had no detectable effect on the pattern of hydrolysis in the serosal compartment (data not shown). Digestion of PEP by Pepsin and Trypsin In Vitro PEP activity was not impaired by 2 successive 4-hour incubations in the presence of pepsin and of trypsin, suggesting that orally administered PEP can reach the intestinal lumen. Although PEP activity was inhibited almost completely (⬎95%) at extreme pH levels (pH, 1.8), this activity was restored after the return to a neutral pH level (pH, 7). These results indicate that PEP cannot enhance gliadin hydrolysis during the gastric phase of digestion but may participate in gliadin hydrolysis in the small intestine.

Discussion The present results confirm the potential of PEP to detoxify gliadin peptides, but raise concerns regarding its possible efficacy in vivo, in the intestinal environment and in CD.

792

MATYSIAK–BUDNIK ET AL

GASTROENTEROLOGY Vol. 129, No. 3

Figure 4. Effect of PEP on the transport and processing of 3Hpeptide 31– 49 by duodenal biopsy specimens from active CD patients, measured in Ussing chambers. Typical HPLC profiles of the mucosal and serosal compartments in 1 patient representative of the 7 patients studied are presented. (A) Absence of degradation of peptide 31– 49 after 3 hours in the mucosal compartment of the Ussing chamber and the passage of intact peptide into the serosal compartment, as already reported.12 (B) The addition of PEP 20 mU/mL prevented the passage of intact peptide across the intestinal mucosa but an important (⬃30%) fraction was recovered under the form of the toxic peptide 31– 42. (C) At the highest PEP concentration studied (500 mU/mL), the peptide was found entirely degraded (⬎90% proline) in the serosal compartment.

It has been suggested for a long time that one contributing factor to the pathogenesis of CD could be a defect in the processing of gliadin by the intestinal mucosa, which might promote the exposure of immune cells to harmful peptides. Both in vitro18,19 and in vivo studies have suggested that the celiac mucosa does not digest gliadin peptides efficiently, and decreased expression of some brush-border membrane enzymes has been reported in patients with active CD.20 Our own results, confirmed in the present study, indicate that gliadin peptides are degraded totally during intestinal transport in treated CD as in nonceliac controls, whereas a protected transport pathway is set up in active CD and allows the absorption of intact gliadin peptides.12 Because prolines confer gliadin peptides with a strong resistance to intraluminal and brush-border digestion, a

novel therapeutic approach, based on the use of a bacterial prolyl-endopeptidase PEP to cleave proline-rich gliadin peptides, has been proposed recently. Indeed, endogenous PEP, although present in the intestinal mucosa from controls and CD patients,12 may have little impact on the digestion of gliadin peptides because this enzyme is cytosolic,21 whereas normal processing of food proteins by epithelial cells takes place in the endosomolysosomal compartments. The efficacy of bacterial PEP first was shown on 2 immunodominant peptides derived from ␣2 and ␣9 gliadins (12- and 14-mer), which were resistant to digestion by rat brush-border membrane enzymes.16 Shan et al3 further showed that PEP could degrade efficiently an ␣2-gliadin– derived 33-mer peptide highly resistant to digestion, which contained a concatemer of 3 major T-cell epitopes strongly immunostimulatory for

September 2005

Figure 5. Effect of PEP on mucosal to serosal transport and processing of 3H-peptide 31– 49 by duodenal biopsy specimens of patients with active CD. PEP was placed in the mucosal compartment together with 3H-peptide 31– 49. After 3 hours, the percentages of peptide metabolites in (A) mucosal and (B) serosal compartments were analyzed by RP-HPLC. PEP prevented the passage of intact peptide 31– 49 across the intestinal mucosa, but at 20 mU/mL, an important fraction (20%) was recovered in the serosal compartment in the form of large, potentially active fragments.
, Intact peptide; PROLINE (□), released free proline indicative of the percentage of fully degraded peptide; SMALL FRAGMENTS (s), retention time of 10 –15 minutes; LARGE FRAGMENTS (o), retention time of 16 –23 minutes; n ⫽ number of patients. *Significantly different from intact peptide in PEP 20 mU/mL and PEP 500 mU/mL conditions (P ⬍ .001). **Significantly different from large fragments in peptide alone and PEP 500 mU/mL conditions (P ⬍ .01). §Significantly different from small fragments in peptide alone and PEP 20 mU/mL conditions (P ⬍ .009). #Significantly different from small fragments in peptide alone (P ⬍ .04).

CD4⫹ T cells in CD patients. In this study, however, degradation of the 33-mer by PEP was analyzed by HPLC using ultraviolet detection, a method of relatively low sensitivity. Moreover, no study was performed on the effect of PEP on gliadin peptide transport by intestinal epithelium in active CD patients. In the present study, the effect of PEP on 2 peptides playing complementary roles in CD pathogenesis, the 33-mer peptide that stimulates adaptive immunity, and the 31– 49 peptide that triggers innate immunity,4,6 was evaluated. Both in vitro and ex vivo studies indicated

PROLYL-ENDOPEPTIDASE AND CELIAC DISEASE

793

that the digestion of gliadin peptides depends on PEP concentration and that the 33-mer peptide, in terms of disappearance of intact peptide, is more resistant to degradation than peptide 31– 49. It remains a matter of speculation whether the superior length of the 33-mer, as compared with peptide 31– 49, makes it more resistant to the initial cleavage by PEP. In vitro, peptide 31– 49 was cleaved at a PEP concentration of 20 mU/mL, but the potentially harmful peptide 31– 424,7 was generated. Similarly, at this PEP concentration, a fraction of 33-mer remained intact and the other fraction was degraded into peptides 57– 68 and 62–75, 2 major T-cell epitopes in CD patients.2 The complete disappearance of both peptides required very high concentrations of PEP (at least 100 mU/mL), and a long enzymatic exposure time (3 h). Ex vivo, after a 3 hour-incubation in the mucosal compartment bathing duodenal biopsy specimens from patients with active CD, peptides 31– 49 and 56 – 88 were degraded to a higher extent than in the in vitro study, suggesting collaboration between brush-border membrane peptidases and PEP in their hydrolysis. At a PEP concentration of 20 mU/mL, intact peptide 31– 49 totally disappeared but intermediate fragments were generated with an elution time corresponding to peptide 31– 42. Similarly, intact 33-mer totally disappeared at PEP 20 mU/mL but immunogenic intermediate fragments were produced and crossed the intestinal mucosa. The HPLC elution times of these metabolites corresponded to the immunostimulatory peptides 56 – 69 and 62–75. Because of the very small amount of these metabolites recovered in the serosal compartment, we did not show directly their immunostimulatory properties on gliadin-specific T-cell clones, but their sequences correspond to 2 recognized major gliadin T-cell epitopes (peptides 56 – 69 and 62–75).2 On the basis of a maximal peptide flux of 10 ␮g/h · cm2, encompassing approximately 50% of immunostimulatory fragments (intact peptide ⫹ large metabolites) (Figures 3 and 7), it can be estimated that in 1 hour, approximately 5 ␮g of the latter fragments may be absorbed by 1 cm2 of intestinal mucosa in active CD patients. This amount may reach the threshold of stimulation of intestinal T cells in CD patients. Thus, Shan et al3 showed that 1 ␮mol/L of peptide 56 – 88 was sufficient to stimulate in vitro human polyclonal T-cell lines derived from CD patients. Although it is difficult to establish a correlation between in vitro and in vivo stimulation, it can be reasonably assumed that a dose of 5 ␮g of immunostimulatory peptides crossing 1 cm2 during 1 hour may be sufficient to trigger an in vivo response. Taking 600 ␮m as the intestinal wall thickness, 1 cm2 of mucosa would correspond to a fluid volume of .06 cm3 (or mL) and the local

794

MATYSIAK–BUDNIK ET AL

GASTROENTEROLOGY Vol. 129, No. 3

Figure 6. Effect of PEP on the transport and processing of 3H–33-mer by duodenal biopsy specimens of active CD patients, measured in Ussing chambers. This figure presents typical HPLC profiles in 1 of the 7 patients studied. (A) Confirmation that the 33-mer is not degraded in the mucosal compartment after 3 hours and that both intact peptide and metabolites with a size compatible with immunostimulatory properties are found in the serosal compartment.12 (B) Addition of 20 mU/mL PEP to the mucosal compartment prevented the passage of intact 33-mer across the biopsy specimens but an important fraction of 33-mer (⬃50%) recovered in the serosal compartment comprised metabolites with an elution time (23–28 min) corresponding to that of gliadin T-cell epitopes. (C) At the highest PEP concentration (.5 or 1 U/mL), only small nonimmunogenic metabolites (Rt ⬍ 20 min) could cross the intestinal mucosa.

September 2005

Figure 7. Effect of PEP on mucosal to serosal transport and processing of 3H-peptide 56 – 88 by duodenal biopsy specimens of patients with active CD. PEP was placed in the mucosal compartment together with 3H-peptide 56 – 88. After 3 hours, the percentages of peptide metabolites in (A) mucosal and (B) serosal compartments were analyzed by RP-HPLC. PEP prevented the passage of intact 33-mer across the intestinal mucosa, but at 20 mU/mL, an important fraction (35%) of the peptide was recovered in the serosal compartment in the form of large, potentially immunogenic fragments.
, Intact peptide; PROLINE (□), released free proline indicative of the percentage of fully degraded peptide; SMALL FRAGMENTS (s), retention time of 10 –19 minutes; LARGE FRAGMENTS (o), retention time of 20 –29 minutes; n ⫽ number of patients. *Significantly different from intact peptide in PEP 20 mU/mL and PEP 500 mU/mL conditions (P ⬍ .01). **Significantly different from large fragments in peptide alone and PEP 500 mU/mL conditions (P ⬍ .0004). §Significantly different from small fragments in peptide alone and PEP 20 mU/mL conditions (P ⬍ .008). §§Significantly different from proline in peptide alone and PEP 20 mU/mL conditions (P ⬍ .02). #Significantly different from large fragments in peptide alone and PEP 20 mU/mL conditions (P ⬍ .004).

concentration of 33-mer– derived metabolites then should be 5 ␮g in .06 mL (ie, 20 ␮mol/L) (33-mer; MW, 3903 daltons). A similar calculation can be made for peptide 31– 49 or its active metabolites that, according to our results, could cross the intestinal mucosa at the rate of 3.6 ␮g/h · cm2 (maximal flux of 8 ␮g/h · cm2 with 45% of intact peptide and large fragments, Figures 3 and 5). This amount may lead to a local concentration of 27 ␮mol/L (peptide 31– 49; MW, 2221 daltons), which seems sufficient to exert their toxicity.4

PROLYL-ENDOPEPTIDASE AND CELIAC DISEASE

795

Another important aspect is the time necessary for PEP to digest the peptides. Although no kinetics study was performed with biopsy specimens, our results indicate that intact peptides or immunogenic fragments may cross the mucosa before the peptides are hydrolyzed completely by PEP. This important aspect on the relative rate of peptide transport and degradation by PEP is probably a key point in the design of future treatments. It is difficult to calculate the theoretical dose of PEP that would be necessary to digest gluten totally in a normal meal. Our ex vivo results indicate that 500 mU/mL (20 ␮g/mL) of PEP are necessary to detoxify 200 ␮g/mL of gliadin peptide. This means that the optimal ratio enzyme/peptide is about 1/10. Accordingly, 2 g of PEP would be necessary to detoxify 20 g of gluten contained in an average normal daily diet. Dilution of PEP in the digestive secretions and gastrointestinal motility, however, may increase the dose required. More specifically, it is conceivable that the esophageal and gastrointestinal motor disorders frequent in patients with active CD,22 mainly delayed gastrointestinal transit, have an impact on gluten digestion. As already mentioned previously, in active CD, gliadin peptides are protected during transcytosis across the intestinal epithelial layer12 and thus can reach the local immune system and exert their immunostimulatory effects. Thus, to benefit from PEP, it is mandatory that gliadin peptides arrive in the small intestine at a concentration sufficiently low to avoid any significant absorption by intestinal epithelial cells and transport along this protective pathway. Our results confirm that PEP is resistant to gastric and pancreatic enzymes in vitro,3 and imply that PEP administered orally can reach the small intestine in an active form. Its activity is, however, probably severely impaired during the gastric phase of digestion because of the poor activity of the enzyme at acidic pH levels. The efficiency of PEP in vivo may depend on a delicate balance between the rate the enzyme is cleaving gliadin peptides in the duodenal lumen and the rate the enterocyte is taking-up peptides. In our ex vivo conditions, only PEP at very high concentrations (⬎500 mU/mL) could inhibit transepithelial transport of gliadin peptides by facilitating their degradation in the luminal compartment and preventing their epithelial absorption. In summary, our results show that at high concentrations and providing sufficiently long time exposure, PEP could be effective in the digestion of gliadin peptides in active CD. Taking into account all the limiting factors, we feel that PEP can be used rather as a complementary treatment in CD, together with gluten exclusion, being especially helpful during occasional, voluntary, or acci-

796

MATYSIAK–BUDNIK ET AL

dental ingestions of small quantities of gluten. Further studies, particularly clinical trials, are needed to determine whether PEP might be used in the future as a full alternative treatment, replacing a gluten-free diet.

GASTROENTEROLOGY Vol. 129, No. 3

12.

13.

References 1. Kim CY, Quarsten H, Bergseng E, Khosla C, Sollid LM. Structural basis for HLA-DQ2-mediated presentation of gluten epitopes in celiac disease. Proc Natl Acad Sci U S A 2004;101:4175– 4179. 2. Arentz-Hansen H, Korner R, Molberg O, Quarsten H, Vader W, Kooy YM, Lundin KE, Koning F, Roepstorff P, Sollid LM, McAdam SN. The intestinal T cell response to alpha-gliadin in adult celiac disease is focused on a single deamidated glutamine targeted by tissue transglutaminase. J Exp Med 2000;191:603– 612. 3. Shan L, Molberg O, Parrot I, Hausch F, Filiz F, Gray GM, Sollid LM, Khosla C. Structural basis for gluten intolerance in celiac sprue. Science 2002;297:2275–2279. 4. Maiuri L, Ciacci C, Ricciardelli I, Vacca L, Raia V, Auricchio S, Picard J, Osman M, Quaratino S, Londei M. Association between innate response to gliadin and activation of pathogenic T cells in coeliac disease. Lancet 2003;362:30 –37. 5. Mention JJ, Ben Ahmed M, Begue B, Barbe U, Verkarre V, Asnafi V, Colombel JF, Cugnenc PH, Ruemmele FM, McIntyre E, Brousse N, Cellier C, Cerf-Bensussan N. Interleukin 15: a key to disrupted intraepithelial lymphocyte homeostasis and lymphomagenesis in celiac disease. Gastroenterology 2003;125:730 –745. 6. Hüe S, Mention J, Monteiro R, Zhang S, Cellier C, Schmitz J, Verkarre V, Fodil N, Bahram S, Cerf-Bensussan N, Caillat-Zucman S. A direct role for NKG2/MICA interaction in villous atrophy during celiac disease. Immunity 2004;21:367–377. 7. Maiuri L, Troncone R, Mayer M, Coletta S, Picarelli A, De Vincenzi M, Pavone V, Auricchio S. In vitro activities of A-gliadin-related synthetic peptides: damaging effect on the atrophic coeliac mucosa and activation of mucosal immune response in the treated coeliac mucosa. Scand J Gastroenterol 1996;31:247–253. 8. Sturgess R, Day P, Ellis HJ, Lundin KE, Gjertsen HA, Kontakou M, Ciclitira PJ. Wheat peptide challenge in coeliac disease. Lancet 1994;343:758 –761. 9. Goossens F, De M I, Vanhoof G, Scharpe S. Distribution of prolyl oligopeptidase in human peripheral tissues and body fluids. Eur J Clin Chem Clin Biochem 1996;34:17–22. 10. Piper JL, Gray GM, Khosla C. Effect of prolyl endopeptidase on digestive resistant gliadin peptides in vivo. J Pharmacol Exp Ther 2004;311:213–219. 11. Marti T, Molberg O, Li Q, Gray GM, Khosla C, Sollid LM. Prolyl endopeptidase-mediated destruction of T cell epitopes in whole

14.

15.

16.

17. 18.

19.

20.

21.

22.

gluten: chemical and immunological characterization. J Pharmacol Exp Ther 2005;312:19 –26. Matysiak-Budnik T, Candalh C, Dugave C, Namane A, Cellier C, Cerf-Bensussan N, Heyman M. Alterations of the intestinal transport and processing of gliadin peptides in celiac disease. Gastroenterology 2003;125:696 –707. Bomsel M, Heyman M, Hocini H, Lagaye S, Belec L, Dupont C, Desgranges C. Intracellular neutralization of HIV transcytosis across tight epithelial barriers by anti-HIV envelope protein dIgA or IgM. Immunity 1998;9:277–287. Saidi D, Heyman M, Kheroua O, Boudraa G, Bylsma P, Kerroucha R, Chekroun A, Maragi JA, Touhami M, Desjeux JF. Jejunal response to beta-lactoglobulin in infants with cow’s milk allergy. C R Acad Sci III 1995;318:683– 689. Goossens F, De M I, Vanhoof G, Scharpe S. A sensitive method for the assay of serum prolyl endopeptidase. Eur J Clin Chem Clin Biochem 1992;30:235–238. Hausch F, Shan L, Santiago NA, Gray GM, Khosla C. Intestinal digestive resistance of immunodominant gliadin peptides. Am J Physiol 2002;283:G996 –G1003. Heyman M, Desjeux JF. Significance of intestinal food protein transport. J Pediatr Gastroenterol Nutr 1992;15:48 –57. Cornell HJ. Amino acid composition of peptides remaining after in vitro digestion of a gliadin sub-fraction with duodenal mucosa from patients with coeliac disease. Clin Chim Acta 1988;176: 279 –289. Cornell HJ, Rivett DE. In vitro mucosal digestion of synthetic gliadin-derived peptides in celiac disease. J Protein Chem 1995; 14:335–339. Friis S, Dabelsteen E, Sjostrom H, Noren O, Jarnum S. Gliadin uptake in human enterocytes. Differences between coeliac patients in remission and control individuals. Gut 1992;33:1487– 1492. Mercer J, Eagles ME, Talbot IC. Brush border enzymes in coeliac disease: histochemical evaluation. J Clin Pathol 1990;43:307– 312. Tursi A. Gastrointestinal motility disturbances in celiac disease. J Clin Gastroenterol 2004;38:642– 645.

Received July 20, 2004. Accepted May 26, 2005. Address requests for reprints to: Tamara Matysiak–Budnik, MD, PhD, INSERM EMI-0212, Faculté Necker-Enfants Malades, 156 rue de Vaugirard, 75730 Paris, France. e-mail: [email protected]; fax: (33) 0-1-40-61-56-38. Supported in part by Institut de Recherche des Maladies de l’Appareil Digestif, INSERM and Fondation Grâce de Monaco. The authors are grateful to Professor Yoram Bouhnik and Dr. Axel Balian for their help in the recruitment of patients.