1303PTissue-based activation of mucosal-associated invariant T (MAIT) cells in combination ipilimumab and nivolumab checkpoint inhibitor (CI) colitis

Abstract Background Up to 44% of patients receiving combination ipilimumab and nivolumab develop checkpoint-inhibitor-(CI) colitis, however its molecular pathogenesis is poorly understood. We aimed to characterise peripheral blood and gut mononuclear cells (PBMC; GMNC) in patients with CI-colitis and controls to gain insights into disease aetiology. We were particularly interested in activated, memory, gut-homing CD8+ T-cells and also the innate-like mucosal-associated invariant T (MAIT) cells that play important roles in mucosal immunity. Methods In Cohort I PBMC from patients with CI-colitis (N = 9) were compared with those from patients who received CI with no adverse-events (CI-controls; N = 11), patients with active ulcerative colitis (UC; N = 6) and Healthy Volunteers (N = 16). PBMC findings were tested in a second cohort (Cohort II; IN-Colitis N = 15; IN-NAE=9). GMNC were isolated in Cohort III (IN-colitis N = 5; IN-controls N = 5; UC N = 6; Healthy Volunteers N = 6). Flow-cytometric analysis was used throughout. Results CI-colitis patients had low circulating MAIT cells compared with CI-controls at baseline in Cohort I. Low levels of circulating MAIT cells in both CI-colitis and CI-controls (compared to Healthy Volunteers) were found in Cohort II. CI-treated patients had high levels of activated-memory T-cells in peripheral blood (CD8 + >CD4+) that included a gut-homing population, regardless of the development of colitis. However, activation of circulating MAIT cells was not evident. In gut tissue there was elevation of activated, granzyme-B+ MAIT cells in CI-colitis compared with CI-controls. CI-colitis was characterised by an activated-memory CD8+ lymphocytosis. Conclusions Melanoma patients can have low baseline circulating MAIT cells. In one cohort this associated with CI-colitis. In tissue, activated MAIT cells were elevated in CI-colitis. Further work is needed to determine which immune populations are useful for the prediction and prognostication of CI-colitis, and if MAIT cells contribute to tissue damage or repair. Legal entity responsible for the study The authors. Funding Celgene. Disclosure S.C. Sasson: Research grant / Funding (institution): Celgene. V.T.F. Cheung: Speaker Bureau / Expert testimony: Janssen. G.V. Long: Advisory / Consultancy: Aduro; Advisory / Consultancy: Amgen; Advisory / Consultancy: BMS; Advisory / Consultancy: Merck MSD; Advisory / Consultancy: Mass-Array; Advisory / Consultancy: Novartis; Advisory / Consultancy: Pierre Fabre; Advisory / Consultancy: Roche. M.R. Middleton: Advisory / Consultancy: Amgen; Advisory / Consultancy: AstraZeneca; Advisory / Consultancy: BMS; Advisory / Consultancy: Eisai; Advisory / Consultancy: Lilli; Advisory / Consultancy: Merck; Advisory / Consultancy: Millenium; Advisory / Consultancy: Novartis; Advisory / Consultancy: Physiomics; Advisory / Consultancy: Rignotec; Advisory / Consultancy: Roche. O. Brain: Speaker Bureau / Expert testimony: BMS; Research grant / Funding (institution): Celgene. All other authors have declared no conflicts of interest.