Gluten-specific antibodies of celiac disease gut plasma cells recognize long proteolytic fragments that typically harbor T-cell epitopes

Abstract

This study aimed to identify proteolytic fragments of gluten proteins recognized by recombinant IgG1 monoclonal antibodies generated from single IgA plasma cells of celiac disease lesions. Peptides bound by monoclonal antibodies in complex gut-enzyme digests of gluten treated with the deamidating enzyme transglutaminase 2, were identified by mass spectrometry after antibody pull-down with protein G beads. The antibody bound peptides were long deamidated peptide fragments that contained the substrate recognition sequence of transglutaminase 2. Characteristically, the fragments contained epitopes with the sequence QPEQPFP and variants thereof in multiple copies, and they typically also harbored many different gluten T-cell epitopes. In the pull-down setting where antibodies were immobilized on a solid phase, peptide fragments with multivalent display of epitopes were targeted. This scenario resembles the situation of the B-cell receptor on the surface of B cells. Conceivably, B cells of celiac disease patients select gluten epitopes that are repeated multiple times in long peptide fragments generated by gut digestive enzymes. As the fragments also contain many different T-cell epitopes, this will lead to generation of strong antibody responses by effective presentation of several distinct T-cell epitopes and establishment of T-cell help to B cells.

Figure 1. The hmAbs pull-down long peptides with repeated motifs.

(a) Kernel density plot of the peptide length identified by Q Exactive mass spectrometry in TG2-gliadin fractions pre (grey line) and post (black line) pull-down by the hmAbs 1130-3A02 (b) and 1002-1E01. (c) Mean peptide length pre and post pull-down from TG2-gliadin fractions with all hmAbs. (d) Peptides pulled down with hmAb 1130-3A02. The most frequent 7mer motif is shown in black underlined.

Figure 2. Common motifs in peptides pulled down by hmAb 1130-3B04.

(a) Percent of peptides sharing identical sequence motifs, of 3 to 15 residues in length, post (triangles/black line) and pre pull-down (diamonds/grey line) from a fraction of TG2-gliadin by the hmAb 1130-3B04. (b) Sequence motifs and the frequency and number of peptides harboring the motifs.

Figure 3. Sequence motif of peptides pulled down by hmAb 1130-3B01.

(a) Sequence motif in peptides pulled-down with hmAb 1130-3B01, based on 85% similarity with most common 7mer motif QPQQQFP, as generated by WebLOGO 3.4. (b) ELISA reactivity of the hmAbs 1130-3B01 (c) and 1130-3A02 to synthetic gliadin peptides (native sequences in black and deamidated sequences in grey). Different antibody concentrations were used to generate affinity curves, as indicated on the x-axis.

Figure 4. Binding affinity of hmAbs to QPEQPFP-containing peptides depends on residues flanking the motif.

(a) AlphaLISA affinity of the hmAbs 1002-1E01, 1002-1E03 and 1130-3B01 to PLQPEQPFP and the competitive effect of a panel of different synthetic gliadin peptides harboring the QPEQPFP sequence motif at different concentrations (nM) as indicated on the x-axis. (b) Sequence motif obtained by searching a Triticum aestivum database with “Pattinprot” using the motif XXXQPQQPFPXXX (X = any amino acid). Sequence logo generated by WebLOGO 3.4.

Figure 5. hmAbs show better reactivity to gluten peptides with repeats of epitopes.

(a) Pull-down with the hmAb 1130-3B01, 1002-1E03 or 1002-1E01 from samples with equimolar amounts of the PLQPEQPF peptide and the γ-gliadin 26mer peptide. MALDI-TOF mass spectra of pre (upper panel) and post (three lower panels) samples are depicted. (b) Pull-down with hmAb 1002-1E03 from a size fraction of a gliadin digest treated with TG2 demonstrating that the hmAb preferentially pull-down long peptides with multiple repeats of epitopes. The number of QPQQPFP epitopes found in each peptide fragment and the length of the fragments in samples pre (grey) and post pull-down (black) are shown. Each cross represents one peptide fragment, and the numbers of unique peptide fragments with the different number of epitopes are given on top. (c) AlphaLISA competition assay comparing the relative binding of bead-conjugated hmAb 1002-1E03 to the soluble 34mer ω-peptide in the presence of competing soluble whole antibody (grey solid line) or Fab fragment (black dashed line) of the hmAb 1002-1E03. (d) Inhibition of binding of bead-conjugated hmAb 1002-1E03 to soluble PLQPEQPFP by FLQPEQPFPEQPEQPYPEQPEQPFPQ (grey solid line) or PLQPEQPFP (black dashed line).

Figure 6. Co-localization of gliadin T-cell and B-cell epitopes in an ω-gliadin protein (accession number: Q9FUW7).

The 9mer core sequences of known T-cell epitopes recognized by HLA-DQ restricted CD4 + T cells in celiac disease are highlighted in different colors. The hmAb epitope QPQQPFP is framed. Of note, the sequence of the native protein is given and glutamine (Q) residues that are targeted by TG2 for modification to glutamic acid are not marked.