Glutamic Acid Decarboxylase-Derived Epitopes with Specific Domains Expand CD4+CD25+ Regulatory T Cells

Abstract

BackgroundCD4+CD25+ regulatory T cell (Treg)-based immunotherapy is considered a promising regimen for controlling the progression of autoimmune diabetes. In this study, we tested the hypothesis that the therapeutic effects of Tregs in response to the antigenic epitope stimulation depend on the structural properties of the epitopes used.Methodology/Principal FindingsSplenic lymphocytes from nonobese diabetic (NOD) mice were stimulated with different glutamic acid decarboxylase (GAD)-derived epitopes for 7–10 days and the frequency and function of Tregs was analyzed. We found that, although all expanded Tregs showed suppressive functions in vitro, only p524 (GAD524–538)-expanded CD4+CD25+ T cells inhibited diabetes development in the co-transfer models, while p509 (GAD509–528)- or p530 (GAD530–543)-expanded CD4+CD25+ T cells had no such effects. Using computer-guided molecular modeling and docking methods, the differences in structural characteristics of these epitopes and the interaction mode (including binding energy and identified domains in the epitopes) between the above-mentioned epitopes and MHC class II I-Ag7 were analyzed. The theoretical results showed that the epitope p524, which induced protective Tregs, possessed negative surface-electrostatic potential and bound two chains of MHC class II I-Ag7, while the epitopes p509 and p530 which had no such ability exhibited positive surface-electrostatic potential and bound one chain of I-Ag7. Furthermore, p524 bound to I-Ag7 more stably than p509 and p530. Of importance, we hypothesized and subsequently confirmed experimentally that the epitope (GAD570–585, p570), which displayed similar characteristics to p524, was a protective epitope by showing that p570-expanded CD4+CD25+ T cells suppressed the onset of diabetes in NOD mice.Conclusions/SignificanceThese data suggest that molecular modeling-based structural analysis of epitopes may be an instrumental tool for prediction of protective epitopes to expand functional Tregs.

Dendritic Cells Guide Islet Autoimmunity through a Restricted and Uniquely Processed Peptidome Presented by High-Risk HLA-DR

“…Table I). Two overlapping PPI peptides were eluted from HLA-DR4: PPI [14][15][16][17][18][19][20][21][22][23][24] and PPI [14][15][16][17][18][19][20][21][22][23][24][25][26][27][28][29][30] ( Fig. 2).…”

“…Peptides of all three islet autoantigens were retrieved from HLA-DR3/4 and were derived from the same core regions of the islet Ags as those retrieved from the homozygous DC. Two PPI signal-sequence peptides were eluted: PPI [17][18][19][20][21][22][23][24] and PPI [17][18][19][20][21][22][23] encompassing the same core sequence of PPI (WGPDPAA) as those retrieved from HLA-DR4 (Fig. 2, Supplemental Table I).…”

“…Peptides eluted from HLA-DR purified from PPI-, IA-2-, and GAD65-pulsed DC were synthesized and analyzed by MS for validation of the identified eluted peptide spectra. Shown are the peptide spectra of PPI [17][18][19][20][21][22][23][24] (A), IA-2 470-486 (B), and GAD65 266-286 (C) eluted from HLA-DR (upper spectra) and the corresponding spectra of the synthesized peptides (lower spectra in gray). The peptide spectra of eluted and synthesized peptides fully overlapped.…”

“…PPI [14][15][16][17][18][19][20][21][22][23][24][25][26][27][28][29][30] , containing the WGPDPAAA core sequence, bound to HLA-DR4 (Fig. 6, Table I).…”

“…Until now, T cell-activating regions in islet autoantigens have been identified on the basis of proliferative and cytokine responses to synthetic overlapping peptides spanning the entire amino acid sequence of the Ag (8,9,(17)(18)(19) by in silico prediction of potential T cell epitopes using HLA class II peptide-binding motifs (20)(21)(22) or by eluting naturally processed and presented epitopes (NPPE) from islet Ag-pulsed B cells (23,24). Although the first two approaches can identify core epitopes recognized by the TCR, the actual sequences with naturally generated flanking regions presented by T1D high-risk HLA-DR cannot be predicted.…”