A BISPECIFIC ANTIBODY TARGETING HLA-DQ2.5-GLUTEN PEPTIDES POTENTLY BLOCKS GLUTEN-SPECIFIC T CELLS INDUCED BY GLUTEN INGESTION IN PATIENTS WITH CELIAC DISEASE

Background: Contamination-free oats are considered safe for most patients with celiac disease (CeD). However, we and others have isolated pro-inflammatory oat protein (avenin)-specific CD4+ T cells from the blood and duodenum of some people with CeD, a finding that carries troubling implications for safe oats ingestion in CeD. Does the presence of avenin-specific T cells identify CeD patients susceptible to harm from oats?
Aim: To determine the frequency of symptoms and immune activation in patients with CeD after single-bolus ingestion of purified avenin and then assess symptomatic, immune and clinical (safety) effects after extended avenin ingestion.
Methods: For the first time, food-grade avenin was purified from contamination-free oats to enable feeding studies at high doses not achievable using oats. We employed a series of single-bolus avenin challenges in HLA-DQ2.5+ CeD adults with assessment of symptoms and serum interleukin (IL)-2 at 4 hrs, a highly sensitive marker of gluten-specific T cell activation. Avenin was given in increasing amounts (0.05, 0.1, 0.5, 1, 4, and 6 g) with 4-wk washout periods. Subsequently, in patients with IL-2 responses, avenin was consumed daily for 6 wks at the highest tolerated dose that triggered IL-2. T-cells were assessed using avenin-specific tetramers, serum cytokines (MSD and O-link Inflammation 96-panel) and duodenal histology was examined pre- and post- challenge. A CeD patient undertook a 6 wk wheat gluten challenge as a positive control.
Results: Surprisingly, avenin induced a significant acute IL-2 elevation in 11/29 (38%) CeD patients (mean 16-fold elevation) and adverse symptoms such as pain, diarrhea and vomiting were induced in 60%, with severity correlated to higher IL-2 elevation. Five IL-2 responders then undertook 6-wk avenin challenge. Activated avenin-specific tetramer+ effector memory CD4+ T cells were increased on day 6. Interestingly, after 6 wks of avenin, these had returned to baseline and IL-2 responses after avenin were undetectable; all patients were tolerating avenin without the acute initial symptoms. In the duodenum, a similar frequency of tetramer+ cells were seen at baseline and at 6 wks. Notably, duodenal histology after 6 wks avenin remained normal, in contrast to significant deterioration in the wheat challenged patient. Serum inflammatory cytokines were not elevated by avenin except in one highly symptomatic CeD patient to a similar degree as the wheat challenged patient.
Conclusion: Purified avenin induces acute symptoms and T-cell responses in a subset of “sensitive” CeD patients. Reassuringly, our findings suggest that in most CeD patients oats is unable to sustain a pathogenic immune response above the threshold required for mucosal deterioration, in contrast to wheat gluten. These findings help resolve the discrepancy between clinical oats safety and oats immunity in CeD.

The gluten-free diet for celiac disease (CeD) is restrictive and often fails to induce complete symptom and/or mucosal disease remission. Central to CeD pathogenesis is the gluten-specific CD4+ T cell that is restricted by HLA-DQ2.5 in over 85% of CeD patients, making HLA-DQ2.5 an attractive target for suppressing gluten-dependent immunity. Recently, a novel anti-HLA-DQ2.5 antibody that specifically recognizes the complex of HLA-DQ2.5 and multiple gluten epitopes was developed (DONQ52).
Aim: To assess the ability of DONQ52 to inhibit CeD patient-derived T cell responses to the most immunogenic wheat gluten peptides encompassing immunodominant T cell epitopes.
Methods: We employed an in vivo gluten challenge model in patients with CeD that affords a quantitative readout of disease-relevant gluten-specific T cell responses. HLA-DQ2.5+ CeD patients undertook a 3-day wheat bread challenge (10g/d gluten) with collection of blood before (D1) and 6 days after (D6) commencing the challenge. Peripheral blood mononuclear cells were isolated and assessed in an interferon-γ enzyme-linked immunosorbent spot assay (ELISpot) testing for responses to gluten peptides encompassing DQ2.5-glia-α1a/α2, DQ2.5-glia-ω1/ω2 and DQ2.5-glia-γ1 and deamidated gliadin protein. The effect of addition of DONQ52 (4 or 40 μg/mL) was assessed and compared to pan-DQ blockade (SPVL3 antibody).
Results: 20 CeD patients completed 3-day wheat challenge and 15/20 (75%) had robust gluten-specific T cell responses to the α- and ω-gliadin peptides and the 33mer on D6. Notably, DONQ52 at both concentrations was highly efficient at inhibiting IFN-γ responses to all stimulatory gluten peptides in all responsive patients, with a mean reduction in response of 88% (33-mer), 89% (DQ2.5-glia-α1a/α2), 63% (DQ2.5-glia-ω1/ω2) and 84% (cocktail of peptides containing DQ2.5-glia-α1a/α2, DQ2.5-glia-ω1/ω2 and DQ2.5-glia-γ1), with DONQ52 at 40 μg/mL. DONQ52 was generally better than Pan-DQ antibody blockade in reducing gluten peptide responses. T cell responses to whole protein (gliadin) were less consistent, but nevertheless, blocking by DONQ52 was seen in 67% of gliadin responders. DONQ52 had no effect on activating non-specific T cell responses when tested using PBMC isolated on D1 prior to gluten challenge.
Conclusion: DONQ52 is highly effective at blocking gluten-specific T cell responses to dominant wheat gluten epitopes. Studies to assess DONQ52 blockade of barley hordein and rye secalin-specific T cells induced by barley and rye challenge, respectively, are underway. Our findings support in vitro data that DONQ52 displays selectivity and broad cross-reactivity against HLA-DQ2.5/gluten peptides and provides proof-of-concept that such an approach has the potential to meaningfully inhibit pathogenic gluten-specific T cell responses in CeD.