Development of Next-Generation NK Cell Optimized Chimeric Antigen Receptors (CARs) for iPSC-Derived NK Cell Therapies Targeting Both Solid and Liquid Tumors

Chimeric Antigen Receptor (CAR) immunotherapy has transformed cancer therapeutics and has been successful in treating hematologic malignancies. To date, autologous T-cell therapy has been the major focus of CAR research and product development but has several limitations including high-cost and complexity of manufacturing, donor-to-donor variability, and severe toxicities in some patients. Allogenic natural killer (NK) cell therapies have also been shown to mount potent responses against hematologic malignancies, but unlike T-cell therapies are less likely to cause high-grade toxicities, such as cytokine release syndrome, neurotoxicity, or graft-vs-host-disease (GVHD), and therefore may serve as a preferred platform for CAR-enabled therapies. Several groups have demonstrated the clinical efficacy of CAR-expressing allogenic NK cells using CARs developed for T-cells (i.e., CD28 co-stimulatory domains). However, recent studies using NK cell activating receptor signaling modules (i.e., NKG2D and 2B4) as CAR signaling domains demonstrate increased activity compared to T-cell CARs, demonstrating the utility of NK optimized CARs. In addition, our previous studies showed that knocking out CIS (encoded by CISH gene), a key intracellular checkpoint of NK cell activation, in iPSC-derived NK cells (CISH KO iNK) significantly improved their anti-tumor activity, in vivo persistence, metabolic fitness, polyfunctionality and resistance to cell exhaustion. The goal of this study was to develop next-generation NK cell-optimized CARs designed to improve targeting and potency of our CISH KO iNK cell therapies against both hematologic malignancies and solid tumors. To achieve this, we developed a screening platform to identify novel CAR signaling domains derived from an array of immune cell signaling modules (Figure 1). First, we constructed a library of 44 CAR constructs (targeting CD19) containing signaling modules from diverse immune cell signaling receptors including NK cell activating receptors, cytokine receptors, and integrins. Next, we developed an efficient CAR expression protocol in iPSC-derived NK cells that yields high CAR expression (>75% CAR+) while maintaining high viability (>90%). We then screened for CAR activity using two co-culture target cell killing assays (eSight impedance assay and Caspase 3/7 killing assay) to test activity against two NK cell resistant CD19+ target cell lines (Raji and SupB15). Our screen identified 7 CAR signaling domains that performed better than both a T-cell CAR (CD28-CD28-CD3ζ) and previously reported NK cell optimized CARs (NKG2D-2B4-CD3ζ, CD28-OX40-CD3ζ, and CD28-OX40L-CD3ζ). These results were consistent across both functional assays and both target cell lines tested. To demonstrate the broad applicability of our top CAR construct (SLNK12) we generated SLNK12-CARs paired with antigen binding domains against various tumor antigens such as CD20, HER-2 etc. We confirmed that SLNK12-CARs potently kill target cell lines regardless of the antigen being targeted. To stringently test SLNK12-CAR potency, we conducted limiting dilution killing assays, serial killing assays, and 3D tumor spheroid killing assays. In all formats SLNK12-CAR iNK cells demonstrated better or comparable killing compared to previously reported NK cell optimized CAR NKG2D-2B4-CD3ζ (Figure 2). Current work is focused on validating our SLNK12-CARs using in vivo models. Overall, we have successfully developed a CAR screening platform in a therapeutically relevant allogenic cell source and have identified NK cell optimized CARs that enable increased potency of NK cells against both solid and liquid tumor cell lines including SKOV-3, BT474, SUP-B15, MOLT-4, Raji etc. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal