A novel CD94-targeting CAR T cell therapy for cytotoxic subtypes of T/NK-cell malignancies

Abstract Introduction: Chimeric antigen receptor (CAR) T cell therapy targeting pan-B-cell antigens has shown unprecedented efficacy in B-cell malignancies and is clinically tolerable despite the induction of B-cell aplasia as an on-target, off-tumor effect. But targeting pan-T-cell antigens with CAR-T for T-cell lymphomas (TCL) is risky as induction of T-cell aplasia could lead to life-threatening infections and may require rescue with allogeneic stem cell transplantation. To address this challenge, we developed a CAR-T product to target a cell-of-origin antigen in T/NK-cell malignancies. Specifically, we describe a CAR-T product against CD94, a C-type lectin receptor expressed on subsets of normal NK (40-70%), gd (40-60%), and CD8 T cells (10%), with the goal of targeting T/NK-cell neoplasms arising from these cytotoxic immune cells. Methods and Results: Consistent with its known expression profile, evaluation of 33 normal human tissues by immunohistochemistry (IHC) showed CD94 expression in rare lymphocytes in secondary lymphoid organs but absent in all other cell types and tissues. This result was also supported by scRNAseq data from 25 normal human tissues from the Human Protein Atlas. By contrast, evaluation T/NK-cell lymphomas (N=46) by IHC showed that CD94 is highly expressed in all NK/TCLs (N=15) and hepatosplenic TCLs (N=7) and a few other subtypes such as enteropathy-associated TCL and primary cutaneous CD8+ cytotoxic TCL. CD94 was also found to be high in aggressive NK cell leukemia, T-cell large granular lymphocytic leukemia (T-LGL), and NK-LGL by flow cytometry. To develop an anti-CD94 CAR, we generated a high affinity mouse monoclonal antibody (mAb) against human CD94 (Kd ~1 nM). Specificity of the mAb was demonstrated by testing against isogenic NK leukemia cell lines (NKL and KHYG-1) with or without CD94 and screening against 5,000 proteins representing 94% of the human membrane proteome. Next, we constructed a bicistronic lentiviral vector to express a second-generation CAR consisting of single chain variable fragment derived from the mAb and a 4-1BB co-stimulatory domain, and a truncated BCMA (tBCMA) molecule lacking g-secretase cleavage site and intracellular signaling domain to serve as a safety switch. In co-culture assays, the CAR-T were highly cytotoxic to CD94+ but not CD94 negative (neg) isogenic NK leukemia cell lines in both short-term and long-term serial killing assays over 14 days (P0.0001). Complete eradication of the tumor by CD94 CAR-T was observed in vivo in two NK cell leukemia xenografts, one patient-derived xenograft (PDX) of hepatosplenic TCL, and one PDX of NK/TCL, whereas progressive tumor growth was noted in mice treated with untransduced T cells (P0.01 for total flux and/or survival). Importantly, complete and sustained elimination of disseminated tumors was seen with CAR-T doses as low as 0.2x106 cells/mouse when 3-day rapid manufacturing conditions were employed for CAR-T production. Using super-resolution fluorescence microscopy, we showed that when CD94 and the anti-CD94 CAR were co-expressed, CD94 is sequestered intracellularly and this prevented fratricide of the CAR-T cells. In co-cultures with autologous peripheral blood mononuclear cells, the CAR-T eliminated CD94+ NK, gd, and CD8 T cells but had no effect against the corresponding CD94neg subsets. Consistent with this, the CAR-T eliminated CD94+ NK and rare subset of CD94+ T cells in humanized mice but spared the CD94neg counterparts. Notably, when the allogeneic CAR-T cells were eliminated due to rejection by the recipient immune system, CD94+ fractions of NK and T cells were restored in the humanized mice. Lastly, in vitro and in vivo studies showed that the CAR-T cells could be rapidly eliminated by targeting the tBCMA safety switch with teclistamab or elranatamab anti-BCMA bispecific T-cell engagers or with belantamab mafodotin anti-BCMA antibody drug conjugate. Conclusions: Our findings suggest that anti-CD94 CAR-T is likely to be highly effective against cytotoxic subtypes of T/NK-cell neoplasms. The restricted expression in normal tissues indicates that this therapy will be safe and unlikely to cause T or NK cell aplasia. The integrated tBCMA molecule in the CAR-T cells serves as a safety switch in case of unexpected toxicity. Together, these results support evaluation of this CAR-T product in a planned phase 1 clinical trial in patients with CD94+ T/NK-cell malignancies, who have a significant unmet clinical need.