Abstract A014: A CD90 NOT gate for cell therapies to treat AML and spare healthy hematopoietic stem cells

Abstract Background: In many hyperproliferative disorders including cancer, aberrant stem cells are the root cause of the pathology. Acute myeloid leukemia (AML) is an especially aggressive form of cancer that continues to resist the application of novel treatment modalities. While single-antigen-targeted CAR T cells have demonstrated efficacy in treating B cell malignancies, in AML, commonly targeted antigens are co-expressed on normal hematopoietic lineages causing on-target, off-tumor toxicities. Targeting antigen-pairs through NOT gated CAR T cell designs offers a strategy for increased safety by improving discrimination between tumor and healthy cells. Here, we describe a system for genetically engineering cytotoxic T cells to confer protection of hematopoietic stem cells (HSCs), while directing cytotoxicity against AML. This system uses the Tmod NOT gate that targets activator antigens (A-Ags) with a CAR, and blocker antigens (B-Ags) with an inhibitory receptor based on LIR-1. The candidate therapeutic relies on CD33 and CLL-1 to target AML cells, but incorporates an inhibitory receptor triggered by CD90 to protect healthy HSCs. Methods: To discover novel target antigen profiles, we leveraged both single-cell and bulk RNA-Seq publicly available datasets from human hematopoietic cell populations to generate an integrated transcriptomic dataset and identified antigen A- and B-Ags where: 1) the Ags are expressed on the cell surface; 2) A-Ags have high and consistent expression on primary AML and AML cell lines; 3) B-Ags have minimal to no expression on AML cells and high expression on HSCs that may express the A-Ags. Subsequently, lead target antigens identified by mRNA expression were confirmed via cell-surface expression analysis of AML samples. Results: The CD33-CLL1 | CD90 Tmod construct was selected based on its optimal Ag-expression profile. Comprehensive binder discovery campaigns using mouse immunization, hybridoma screening, and phage display were performed. CD33-CLL1 | CD90 Tmod constructs were screened in Jurkat and primary T cell assays to determine potency and selectivity against target cell lines and primary HSCs. One specific Tmod construct demonstrated impressive sensitivity for both activation and blocking (EC50 and IC50 ∼ 1,000 molecules/cell). Studies in primary T cells confirmed the potency and selectivity of the construct using HL-60 target cells engineered with varying amounts of CD90 surface expression. The potency and selectivity windows were in the range observed previously with a wide variety of Tmod constructs (1). Finally, the CD33-CLL1 | CD90 Tmod constructs were tested in co-cultures with HSCs and shown significant protection (∼50%). Conclusions: We developed a CD33-CLL1 | CD90 Tmod construct with selective potency toward AML vs. healthy HSCs. The CD90 blocker was highly sensitive, a necessary attribute given the low but consistent levels of expression of CD90 in HSCs. (1) DiAndreth B et Al. Multi-targeted, NOT gated CAR-T cells as a strategy to protect normal lineages for blood cancer therapy. Front Immunol. 2025 Citation Format: Tanveer Gill, Vasantika Suryawanshi, Sara Martire, Michele Mcelvain, Aaron Winters, Breanna Luna, Diane Manry, John Welch, Alexander Kamb, Chawita Netirojjanakul, Michael Rettig, Ryan Elshimali. A CD90 NOT gate for cell therapies to treat AML and spare healthy hematopoietic stem cells abstract. In: Proceedings of the AACR Immuno-Oncology Conference (AACR IO): Discovery and Innovation in Cancer Immunology: Revolutionizing Treatment through Immunotherapy; 2026 Feb 18-21; Los Angeles, CA. Philadelphia (PA): AACR; Cancer Immunol Res 2026;14(2 Suppl):Abstract nr A014.