The CD33-targeting bispecific T cell engager (BiTE®) AMG 330 proved to be highly efficient in mediating cytotoxicity of AML cells in vitro and in mouse models (Krupka et al, Blood 2014). Yet, T cell activation is correlated with the upregulation of PD-L1 and other inhibitory checkpoint molecules on AML cells that confer adaptive immune resistance (Krupka et al, Leukemia 2016). PD-1/PD-L1 blocking agents may counteract T cell dysfunction, however, at the expense of broadly distributed immune-related adverse events (irAEs). We developed a checkpoint inhibitory T cell engaging (CiTE) antibody that combines T cell redirection to CD33 on AML cells with locally restricted immune checkpoint blockade. CiTE constructs were generated by first fusing a high-affinity CD33 single-chain variable fragment (scFv) to a CD3ε scFv in one polypeptide chain. Next, this single-chain chain was fused to the extracellular domain of PD-1 (PD-1ex), which naturally holds a low affinity to PD-L1. Antigen binding of CiTE constructs as well as CiTE mediated cytotoxicity of AML cell lines and primary AML cells were done using multiparameter flow cytometry. T cell activation and cytotoxicity assays were complemented by cytometric bead arrays. Murine AML xenograft studies using non-obese diabetic (NOD) scid gamma mice were used for engraftment of primary AML cells and assessment of CiTE mediated cytotoxicity in vivo. CiTE antibody constructs were successfully generated by fusing the bispecific CD33-CD3ε scFv to the endogenous extracellular domain of human PD-1 (PD-1ex). The CiTE was compared to a single chain triplebody (sctb), in which PD-1ex was replaced by a high-affinity PD-L1 scFv. The BiTE-like molecule, PD-1ex.αCD3 and αPD-L1.αCD3, as well as a non-targeting molecule served as controls. When investigating CiTE and sctb as whole molecules, both bound with similar affinities to CD33+PD-L1+ AML cell lines and HD T cells. CiTE- and sctb-induced upregulation of CD69 and CD25 on healthy donor T cells in the presence of MOLM-13-PD-L1 cells. By a synergistic effect of checkpoint blockade and avidity-dependent binding, the PD-1ex attachment increased T cell activation (3.3-fold elevation of IFN-γ release) and lead to efficient and highly selective cytotoxicity of CD33+PD-L1+ cells (EC50 = 2.3 pM to 26.9 pM) as well as primary AML patient samples (n=8). CiTE induced preferential lysis of CD33+PD-L1+ cells and had no activity against CD33-PD-L1+ cells. This was supported by the observation that the CiTE molecule was able to selectively induce elimination of CD33+PD-L1+ cells in the presence of PD-L1+ cells. In a murine xenograft model, the CiTE induced complete AML eradication without causing leukemia-unrelated T cell activation or body weight loss. Notably, murine and human PD-L1 bind with similar affinities to PD-1. We conclude that our molecule preferentially targets CD33+PD-L1+ AML cells, whereas high-affinity blocking agents also address PD-L1+ non-AML cells. Based on these findings, we expect to reverse adaptive immune escape mechanisms of T cell recruiting antibody formats and avoid irAEs associated with systemic checkpoint blockade, suggesting efficient therapeutic potential particularly for patients with relapsed or refractory AML. Future studies will need to further examine efficiency and tolerance in advanced in vivo models before applying the CiTE format into a clinical setting. Disclosures Lindl: Amgen: Research Funding. Metzeler:Celgene: Consultancy, Research Funding; Novartis: Consultancy. Subklewe:Gilead: Consultancy, Honoraria, Research Funding; Amgen: Consultancy, Honoraria, Research Funding; Pfizer: Consultancy, Honoraria; Roche: Consultancy, Research Funding; Celgene: Consultancy, Honoraria.

Background: FLT3 (CD135) is a receptor tyrosine kinase expressed by hematopoietic progenitor cells, and when activated by ligand binding, FLT3 signaling can induce survival, proliferation and differentiation. Acute myeloid leukemia (AML), an indication with high unmet need, derives from hematopoietic progenitor lineage cells; consequently, many AML patient samples retain FLT3 expression on circulating leukemic cells. In comparison to the wide expression seen in AML patient samples, normal tissue expression of FLT3 is restricted to the hematopoietic compartment: a subset of hematopoietic stem cells, a subset of lineage committed progenitor cells and dendritic cells. Cell surface FLT3 protein was not detectable in those solid tissues that express quantifiable FLT3 mRNA (cerebellum, pancreas). As a result of this favorable expression profile, targeting AML cells using T cells expressing a chimeric antigen receptor (CAR) directed against FLT3, is expected to provide benefit to patients. Methods: A panel of fully-human, anti-human FLT3 antibodies were generated, screened for specific FLT3 binding, converted to single chain variable fragments (scFv) and re-evaluated for binding and stability. Three anti-FLT3 scFv sequences were selected for incorporation into CAR constructs. T cells transduced with these CAR constructs were evaluated in vitro for cytotoxicity, proliferation and cytokine secretion, and in vivo in a mouse xenograft model. The lead anti-FLT3 scFv was combined with an anti-CD3 scFv (bispecific T cell engager (BiTE®) format) for evaluation in non-human primates (NHP). Results: Human T cells engineered to express anti-FLT3 CAR constructs were evaluated for cytotoxicity, proliferation and cytokine secretion in the presence of three FLT3-positive cell lines, each with a different number of surface-expressed FLT3 receptors, and one FLT3-negative cell line. While all constructs were selectively active against FLT3-positive cells, one construct was more active than the others against cells expressing very low levels of FLT3; 90% depletion of target cells with ~1600 receptors/cell was observed after 40 hours of co-culture with CAR T cells in a 1:1 ratio. This lead construct was further evaluated in a mouse xenograft model, where FLT3 CAR-T cells provided a survival advantage. The anti-FLT3 scFv used in the lead CAR construct was evaluated in NHP in the form of a BiTE®. In this 16-day, multi-dose study, the FLT3 BiTE® induced elimination of FLT3+ cells as assessed by depletion of FLT3 mRNA-expressing cells (97% reduction in FLT3 mRNA) in the circulation and increased levels of soluble FLT3 ligand in serum. Conclusions: These results demonstrate the potential of using FLT3 CAR-T cell therapy for the treatment of AML. Citation Format: Tara L. Arvedson, Alice Bakker, Herve Lebrec, Gregor Adams, Armen Madiros, Priya Koppikar, Mercedesz Balazs, Mei Gong, Yan Zheng, Rebecca Goldstein, Tony Polverino, Lawren Wu, Angela Coxon. Generation and evaluation of an FLT3 CAR-T cell therapy for the treatment of acute myeloid leukemia [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 2559.

Antibody based immunotherapy represents a promising strategy to eliminate chemoresistant cells in acute myeloid leukemia (AML). Clinical experience in acute lymphoblastic leukemia (ALL) has shown a clear correlation of leukemic burden and the occurrence of a cytokine release syndrome (CRS) during treatment with blinatumomab (CD19/CD3 BiTE®). A cytoreductive phase prior to or in combination with antibody therapy might be beneficial to reduce the severity of adverse events like CRS. The latter is often treated with steroids (dexamethasone, DEX) or less commonly, with the IL-6R antibody tocilizumab (TOC). As T-cell proliferation and function are of crucial importance for BiTE® activity (Zugmaier 2015), the effect of the drugs on effector cell function will dictate clinical response to therapy. In this study, we evaluated the influence of cytoreductive- (azacythidine, AZA; decitabine, DEC), and immunmodulatory (DEX and TOC) drugs on antibody-mediated cytotoxicity and T-cell proliferation. A CD33/CD3 BiTE® antibody construct (AMG 330) served as model T-cell recruiting antibody in this study. To address this question we set up the following experimental approaches: AML cells were cocultured with healthy donor (HD) T cells for up to 14 days ex vivo. T cells were either incubated with the specific drug for 3 days prior to coculture or the drugs were simultaneously added to AML-T cell cultures. Drug concentrations were chosen based on published serum concentrations in AML patients and their ex vivo stability in culture, validated by mass spectrometry. BiTE® mediated cytotoxicity and T-cell proliferation were assessed by flow cytometry. Preincubation of T cells with AZA and DEC impaired antibody mediated cytotoxicity of HL60 cells in a concentration dependent manner (% lysis control (ctrl) vs AZA at 1, 5, 10 µM: 99.9 vs 99.2 vs 52.1 vs 28.7, n=7; ctrl vs DEC at 0.2, 2, 5 µM: 98.4 vs 71.3 vs 60.0 vs 50.0, n=3). Similarly, T-cell proliferation was also markedly decreased (fold change (FC) T cells ctrl vs AZA at 1, 5, 10 µM: 2.9 vs 2.8 vs 1.5 vs 0.7; ctrl vs DEC at 0.2, 2, 5 µM: 3.8 vs 3.0 vs 2.3 vs 1.2). For DEX it was shown that incubation of T cells with steroids prior to cocultures had no negative effect on BiTE® mediated cytotoxicity (Brandl 2007). However, as steroids are often used simultaneously with T-cell recruiting immunotherapies, we tested the influence of DEX in combination with AMG 330. The addition of DEX to primary AML-T cell cultures (75 ng/ml) significantly impaired AMG 330 mediated cytotoxicity (% lysis AMG 330 vs AMG 330+DEX day (d) 6: 95.9 vs 47.5, n=9). This correlated to a markedly reduced T-cell proliferation (FC T cells AMG 330 vs AMG 330+DEX d6: 11.2 vs 1.2, n=9). Correspondingly, secretion of IFNγ was also decreased (n=3). Upon discontinuation of DEX an increase in AMG 330 mediated cytotoxicity was observed. Nevertheless, cytotoxicity was still considerably lower compared to control cultures (%lysis AMG 330 vs AMG 330+DEX d9: 95.6 vs 77.0). In contrast to DEX, TOC (110 µg/ml) had no negative effect on T-cell proliferation (FC T cells d6: AMG 330 vs AMG 330+TOC: 42.3 vs 36.9, n=4). Similarly, secretion of IFNγ was not affected through the simultaneous addition of TOC to primary AMG 330 cultures (pg/ml AMG 330 vs AMG 330+TOC d6: 543.9 vs 345.8 n=2). Importantly, drugs might not only interfere with effector cell function but also modulate target antigen expression. As we have previously demonstrated that antigen expression levels influence BiTE® mediated cytotoxicity (Krupka 2016), we analysed the effect of the drugs on CD33 expression. None of the drugs induced a significant up- or downregulation of CD33 on AML celllines as detected by flow cytometry. Hence our data support the notion that these drugs do not modulate antigen expression dependent lysis kinectics. We conclude, that drugs given prior or concomitant to BiTE® therapy have the potential to reduce T-cell proliferation and cytotoxicity. In particular, we observed a negative impact of AZA and DEC when given prior to AML-T cell cocultures. Importantly, even short exposure to DEX led to a significanly reduced T-cell responsiveness. Our data suggest the careful evaluation of concomitant drugs in T-cell recruiting antibody therapies and support the restrictive use of steroids in patients receiving BiTE® antibody therapy. For management of severe CRS, TOC could be considered as a targeted biologic therapy that preserves BiTE®-dependent T cell function. Disclosures Krupka: AMGEN Research Munich: Research Funding. Kufer:AMGEN Research Munich: Employment, Equity Ownership, Patents & Royalties. Kischel:AMGEN Research Munich: Employment, Equity Ownership, Patents & Royalties. Subklewe:AMGEN Research Munich: Research Funding.