1: Study patients

Four groups of patients have been included in this study: active CD-patients (aCD), CD-patients on a gluten-free diet (GFD), healthy controls (C) and patients with other gastrointestinal pathology (GP). The first one (aCD) included celiac patients with a recent diagnosis of the disease, therefore on a normal gluten-containing diet. They had clinical symptoms, positive serology (IgA anti-TGt and/or IgA anti-endomisium (EmA) antibodies), genetic susceptibility (HLA-DQ2/DQ8) and all but one had compatible lesion in the duodenal biopsy (Marsh I-III). Remainder cases fulfilled the ESPGHAN criteria (2012) to avoid intestinal biopsy. Celiac patients on a gluten-free diet (GFD), presented a reduction of symptoms, a normalization of the serological markers and a recovery of the histological lesion. Patients with other gastrointestinal pathology (PGI) were patients with a clinical profile compatible with CD but where diagnosis was later discarded, suffering another intestinal disease. These patients had negative serology and mild mucosal alterations non compatibles with CD. Finally, a group of healthy controls was included. Among each study group, patients were separated according to the age in adults (mean age 35) and children (mean age 7).

Tissue and plasma samples included in this retrospective study (2004-2012) were obtained from the Gastroenterology and Pediatric Services from the Hospital Clínico Universitario and the Hospital Universitario Rio Hortega, Valladolid, Spain. Written consent was obtained from adult patients and parents or legal tutors from children. The study protocol was approved by the Ethics Committee of both, the Hospital Clínico Universitario and the Faculty of Medicine, University of Valladolid.

2: HLA Genotyping

Genomic DNA was purified from whole blood (UltraClean® Blood DNA Isolation Kit, Mo Bio, California, USA) and HLA-DQ2 genotyping performed by PCR amplification using sequence specific primers (MWG Biotech, Ebersberg, Germany) and the Kapa Taq kit (Kapa Biosystems, Boston, USA) [14]. Analysis of HLA-DQ8 haplotype was only performed in those DQ2 negative patients, using the DQB1 Only SSP2LQB1 kit (Micro SSP^TM DNA Typing Trays, One Lambda InC, California, USA).

3: IgA-tTG antibodies

IgA-tTG Abs levels were analyzed in plasma samples by the available comercial kit (EliA CeliKey IgA, Phadia, Madrid, Spain) using an UniCAP 100^Є Instrument Version 1.0 (Pharmacia Diagnostics AB, Uppsala, Sweden).

4: Zymogram analysis

Proteins from biopsy samples were isolated using the Trizol Reagent (Applied Biosystems, California, USA) according to the manufacturer’s protocol and quantified by the Bradford reaction. A total of 8 µg of protein samples were subsequently separated in 15% acrilamide/bisacrilamide (37.5:1) gel electrophoresis gliadin zymograms containing 0.1% of protein (gliadin, Sigma-Aldrich, Missouri, USA) using a mini-Protean II system (BioRad Laboratories Inc, California, USA). Gels were then washed twice (t=15’ each) with 2.5% Triton-X-100 to remove SDS and allow protein renaturalization, and incubated overnight at 37°C with the reaction buffer (50 mM Tris, 5 mM CaCl₂, pH 7.5). Protease activity was revealed by staining with 0.1% Coomassie Brilliant Blue R-250 (Bio-Rad, California, USA) in a mixture of acetic acid:methanol:water (1:3:6), and destaining in the same solution mixture without the dye. A molecular weight marker was loaded in every gel (Kaleidoscope pre-stained standard, BioRad, California, USA). Additionally, stacking gel was kept during all the process since the lack of staining, compared to the running gel, was used as an internal control. All samples were separated at 200 Volts for 50 min in the running gel.

5: Fingerprinting and ion-trap Mass Spectometry

Duodenal CD specific proteases (26 and 82 kDa) from a previously described gliadin-degrading pattern [11] were removed from the substrate gel after electrophoresis and analyzed by a combined fingerprinting using a MALDI-TOF Reflex™ IV Bruker (Bruker-Franzen Analytic GmbH, Bremen, Germany) and ion-trap MS on a 3000^plus (Bruker Daltonics, Bremen, Germany) coupled to a Famos-Switchos-Ultimate (LCPackings, Amsterdam, The Netherlands) chromatographic system.

6: Protein and peptide sequence analysis

Protein identification was carried out by searching in the database of the National Center for Biotechnology Information (NCBI) using BLASTP algorithm v.2.2.20. Merops peptidase database (v 9.8) was used to analyze the cleaving points of the identified peptides.

7: Synthesis of peptides

Different peptides (native and deamidated) and terminal modifications (i.e. biotinylated peptides) derived from gliadin 8-mer sequence (Table 3) were supplied by Biomedal SL (Seville, Spain).

8: Enzyme-linked immunosorbent assay (ELISA)

Maxisorp microtitre plates (Nunc, Roskilde, Denmark) were coated with a mix of 2 synthetic peptides (Biomedal SL, Spain) in phosphate buffer (pH 7.4) and incubated overnight at 4°C. The plates were washed with phosphate buffered saline (PBS)-Tween 20 and blocked with a solution (PBS-5% non-fat dry milk) for 1 h at room temperature (RT). Next, plasma samples (1/50) were added to the wells. After 1 h of incubation at RT, the plates were washed and horseradish peroxidase-conjugated rabbit anti-human IgA (catalog. No. P0216, DakoCytomation) was added (1/500) to duplicate wells and incubated for 1 h at RT. Then the plates were washed, and substrate solution (TMB, Sigma) was added. After 30 min of incubation at RT in the dark, the reaction was stopped with 0.3 M sulphuric acid, and the absorbance at 450 nm was measured (microplate reader UVM340, Asys Hitech GmbH, Eugendorf, Austria).

9: Statistical analysis

Each plasma sample was analyzed by triplicate and the mean value calculated. Positive and negative controls (CD-patients and C-individuals, respectively) and buffer blanks were included in each assay. An absorbance cutoff value was defined for each age group. It was calculated using the GP-patients (mean ± 2 x standard deviation) with the aim of discriminate, with the higher specificity possible, between that group and CD-patients. Only samples with an absorbance value ≥ cutoff value were considered positives. Unpaired, two-tailed Student’s t-test was applied and p<0.05 was considered as statistically significant. Specificity [(true negatives / (true negatives + false positives)) x 100]s and sensitivity [true positives / (true positives + false negatives)] x 100] were calculated for each age group.

1: Duodenal gliadin-degrading proteases specific from celiac patients release an 8-mer gliadin peptide

Gliadin zymogram analysis confirmed that the gliadin-degrading protease pattern was specifically found in the duodenum of all 4 CD patients irrespectively of their status (aCD/GFD) or age (adults/children) while remained absent in protein samples obtained from non-CD pathogenic control (Figure S1) as previously reported [11]. In order to identify gliadin peptides specifically released by such enzymes, fingerprinting was applied to the most intense gliadin-degrading band obtained from the zymogram matrix (26 kDa) from a GFD-patient. The fingerprinting profiles did not allow any match with any known protein. Therefore, a secondary MALDI-TOF ion-trap MS was applied and 3 peptides (8-, 15- and 18-mer) were identified (Table 4 and Figure S2).

To determinate the nature of the cleaving points generated in the 3 peptides resulting from the above analysis, we conducted a search on the peptidase database MEROPS [15]. As shown in Table 4, the C-end of the 15-mer peptide and both ends of the 18-mer peptide were potential cleaving points for trypsin so they may have been generated by the trypsin digestion applied during MALDI-TOF analysis. However, the N-end of the 15-mer peptide and both ends of the 8-mer peptide lack any described trypsin cleaving point so they did not seem to have been released during technical processing of the samples but on the contrary were likely to have been released by the action of the CD-specific duodenal gliadin-degrading proteases. A second independent analysis was also performed on the 82 kDa gliadin-degrading protease obtained from an aCD-patient revealing again the 8-mer gliadin peptide (Table 4). Since such peptide was found in two independent experiments and its cleaving points were likely to be derived from the CD-specific duodenal gliadin degrading proteases we decided to study it in depth.

2: Identification of the 8-mer peptide in proteins from toxic cereals for celiac patients

The sequence alignment made with the BLASTP program of the NCBI server [16] revealed several homologous sequences to the 8-mer gliadin peptide which remained restricted to prolamin and glutenin proteins from cereals which are toxic to CD patients like wheat, rye and barley species. None of the 8-mer homologous sequences were found in non toxic cereals to CD patients, i.e. rice or maize. A full length copy of the 8-mer peptide was identified in 12 different species from the true grasses genera: Hordeum, Triticum, Lophopyrum, Aegilops and Secale. The 8-mer was more abundant in prolamins from wheat (ω-gliadin and ω-secalin in Triticum) with 196 peptides matched. Results are detailed in Table 5 and 6.

3: The 8-mer peptide overlaps with 3 known gluten T cell epitopes

We carried out a search for the known CD relevant gluten T cell epitopes restricted by HLA-DQ molecules [6] in proteins containing the 8-mer peptide. The 8-mer peptide appeared overlapped or next to three T-cell epitopes (Table 7), and the most repeated was QQPFPQQPQ [17] which overlapped in 34.5% of the 8-mer peptides matched in ω-gliadin, 13.1% in ω-secalin, 14.3% in hordein and 33.4% in glutenin sequences. Two more T-cell epitopes were close to the 8-mer peptide despite not overlapping, QQPQQPFPQ and PFPQPQQPF [17,18], which were more repeated in ω-secalin.

4: Development and standardization of ELISA based on the 8-mer peptide

HLA-DQ2 and DQ8 molecules have preference for binding of negatively charged amino acid residues in certain binding pockets [19]. In vivo, enzyme tTG mediates deamidation of some glutamine residues into glutamic acid acquiring a negative charge and rendering peptides more suitable for presentation by DQ2&DQ8 molecules [7]. Thus, we studied the presence in plasma samples from CD patients of IgA Abs against both the native and deamidated 8-mer peptides. It has been shown that the targeting of Q residues in peptides is strongly influenced by the position of C-terminally located P residues. Whereas a Q residue in the QXP consensus sequence is targeted by tTG, Q residues in a QP or QXXP sequence motif are not [20,21]. Therefore, we studied the effect on the reactivity of the 8-mer peptide after its deamidation in position 6 (FPLQPEQP).

Two synthetic biotynilated peptides (native and deamidated, Table 3) were used in streptavidin coated plates to study the presence of IgA antibodies in plasma samples from CD patients (Biomedal S.L). The levels of IgA Abs against deamidated 8-mer peptide in plasma samples from celiac patients was three fold higher than the one observed in the same patients with the native 8-mer peptide (results not shown). The antigen-antibody complex seemed to be stronger when the deamidated 8-mer peptide was used than with the native sequence. In order to improve the detection of antibodies by ELISA, ten different combinations of sequence and terminal modifications were designed (Table 3) and containing, all but one, at least one copy of deamidated gliadin 8-mer epitopes (Biomedal S.L). Finally, the combination of two peptides was selected as the optimum recombinant antigen to coat the microtiter plates for the standardization of the ELISA.

5: Detection of IgA anti-DGP 8-mer antibodies in plasma samples from celiac children

The ELISA test designed in this article allowed us to identify IgA Abs which recognize DGP 8-mer as a specific antigen in the plasma from CD patients. Having optimized the technique, we first studied the presence of specific IgA anti-DGP 8-mer Abs in plasma from celiac children (Table 2, Figure 1 (A)). The absorbance values (mean ± SEM) at O.D 450 nm obtained for each group of patients were: aCD-patients (n=31) 1.61 ± 0.24, GFD-patients (n=17) 0.14 ± 0.02, C-individuals (n=10) 0.06 ± 0.003 and GP-patients (n=23) 0.10 ± 0.01. According to the cutoff value calculated for the pediatric population (0.21), among aCD-patients (n=31), there were 29 positives and 2 negatives for IgA anti-DGP 8-mer Abs. In the case of GFD-patients, only 2 children were positives for IgA-8-mer Abs. Very low levels of IgA anti-DGP 8-mer Abs were detected in 1 out of 23 GP-patients. The remaining 22 GP-patients and all C-individuals had negative levels for IgA anti-DGP 8-mer Abs.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Detection of IgA anti-DGP 8-mer antibodies in plasma samples from children (A) and adults (B).

Study groups: active celiac disease (aCD-patients), on a gluten-free diet (GFD-patients), healthy controls (C-individuals) and other gastrointestinal pathology (GP-patients). Each point represents the mean absorbance value of one patient at O.D 450 nm. According to the cutoff value (0.21 in children and 0.5 in adults, pointed line), patients with a higher absorbance value were considered positive for IgA anti-DGP 8-mer Abs, while those with lower absorbance values were considered negative. * p<0.05, ** p<0.01 and *** p<0.001 (Unpaired, two-tailed Student’s t-test). A- Among pediatric population 90.6% of aCD, 10.5% of GFD and 4.3% of GP-patients were positive for IgA anti-DGP 8-mer antibodies, and 9.4% of aCD-patients, 89.5% of GFD-patients, 100% of C-individuals and 95.7% of GP-patients were negative. B- In adult population, 18.7% of aCD, 100% of GFD-patients, C-individuals and GP-patients were positive for IgA anti-DGP 8-mer, while 18.7% of aCD and 100% of GFD-, C- and GP-patients were negative.

https://doi.org/10.1371/journal.pone.0080982.g001

Comparison of the mean absorbance levels between aCD-patients and the remainder groups showed significant differences (p<0.001). IgA anti-DGP 8-mer Abs not only differentiated between CD patients and non-CD patients, but also between active celiac patients and those celiac patients who have initiated a GFD. The modified ELISA method used for detect IgA anti-DGP 8-mer antibodies showed 94% of specificity and 93.5% of sensitivity in children.

6: Detection of IgA anti-DGP 8-mer antibodies in plasma samples from celiac adults

CD is a common disorder in children and adults, and shows age-related differences at initial diagnosis. Given that IgA anti-TGt levels correlate both with the degree of villous atrophy and with the patient’s age [22], we studied if the IgA anti-DGP 8-mer ELISA test could be also applied to adult patients (Table 2).

Results obtained in the adult population are shown in Figure 1 (B). The absorbance values at O.D 450 nm obtained for each group were (mean ± SEM): aCD-patients (n=16) 1.19 ± 0.28, GFD-patients (n=12) 0.31± 0.03, C-individuals (n=27) 0.17 ± 0.01; and GP-patients (n=46) 0.26 ± 0.01. According to the cutoff value calculated for the adult population (0.5), among the group of aCD- patients (n=16), there were 13 positives and 3 negatives for IgA anti-DGP 8-mer Abs. All GFD-patients (n=12) had negative levels of antibodies, except one. IgA anti-DGP 8-mer Abs were not detected in any of the non-celiac patients.

When mean absorbance levels obtained by aCD-patients and the remainder groups were compared, significant differences were obtained (p<0.001). The ELISA test showed a specificity of 98.8% and a sensitivity of 81.3% in adult population.