PS972 THE AML SECRETOME AND T CELL FUNCTION: MODULATION OF AMG 330‐MEDIATED T CELL PROLIFERATION AND CYTOTOXICITY

Abstract

Background:T cell recruiting antibody constructs are a novel strategy to target chemorefractory acute myeloid leukemia (AML) cells. A CD33 specific BiTE® antibody construct, AMG 330, is currently in evaluation in a first in human phase I trial with promising tolerability and anti‐leukemic activity in relapsed / refractory AML (Ravandi et al, ASH 2018). We have developed a long‐term culture system for AML patient samples and showed substantial activity of AMG 330 redirected autologous residual T cells against AML cells in 76 % of patient samples (Krupka et al, Blood 2014; Krupka et al, Leukemia 2016). Having observed suboptimal response in the remaining samples, we conducted this study to investigate potential mechanisms leading to decreased sensitivity of AML cells to BiTE® recruited T cell activity.Aims:We hypothesized that AML cells secrete humoral inhibitory factors that influence T cell function and thereby modulate AMG 330‐mediated cytotoxicity. Hence, we analyzed the influence of AML patient plasma (PtP) derived from bone marrow (BM) on T cell proliferation and AMG 330‐mediated cytotoxicity against AML cell lines.Methods:T cell proliferation assays were performed using CFSE‐labeled healthy donor (HD) T cells stimulated with CD3/CD28 beads and IL‐2 and analyzed by flow cytometry (day 3). AML PtP were collected from BM (n = 16) at time of initial diagnosis. HD T cells were cultured in either 20 % of PtP, FCS or healthy human serum (HS). To test for a dominant negative factor within PtP, mixing experiments of FCS and PtP were performed (ratio: 4:0 ‐ 4:4). In addition, conditioned medium (CM) of three different AML cell lines (HL60, OCI‐AML3, KG1a) was analyzed on its modulation of T cell function. To dissect the relevance of cell‐cell contact vs the AML secretome, transwell assays with HD T cells and AML cell lines were performed.Results:We observed that AML PtP negatively impacted T cell proliferation of HD in vitro which was not observed in the presence of FCS or HS (mean % T cell proliferation PtP: 9.4 (n = 16), FCS: 74.0 (n = 16), HS: 35.5 (n = 11)). T cell proliferation was still reduced after additional supplementation of 20 % FCS in 4 of 11 AML samples (mean % T cell proliferation PtP+FCS: 40.2, FCS: 75.9). Affirmatively, CM of the cell lines HL60 and OCI‐AML3 negatively impacted T cell proliferation (mean % T cell proliferation CM of HL60: 11.5 (n = 12), OCI‐AML3: 34.7 (n = 9), KG1a: 57.2 (n = 6), control: 54.6 (n = 12)). Reduced T cell proliferation and AMG 330‐mediated cytotoxicity was also observed in the presence of all three AML cell lines within the transwell assays (mean % T cell proliferation HL60: 2.6, OCI‐AML3: 15.4, KG1a: 28.4, control: 40.0; % specific cytotoxicity HL60: 21.9, OCI‐AML3: 45.6, KG1a: 56.4, control: 69.9; n = 10). This was not due to serum deprivation as gamma‐irradiated AML cell lines (n = 6) still significantly reduced T cell proliferation and AMG 330‐mediated cytotoxicity.Summary/Conclusion:We conclude that AML patient plasma samples contain soluble factors which may inhibit T cell functions to a varying degree. By using an in vitro model, we could demonstrate that this was mediated through soluble factors secreted by AML cells. Analysis of AML patient plasma (n = 35) and healthy BM plasma (n = 5) by Proximity Extension Assay technology (Immuno‐Oncology Panel, Olink) revealed significant differences in plasma levels of 66 of 92 protein biomarkers. We are in the process to dissect their relevance in ongoing experiments. We aim to identify inhibitory factors that might be targetable and suitable for combinatorial approaches to enhance T cell‐based immunotherapy in AML.