Correlative data from an ongoing Phase 1, multicenter study of senti-202, a first-in-class, CD33 and/or FLT3 & not endomucin (EMCN), selective off-the-shelf CAR NK cell therapy for Acute Myeloid Leukemia (AML) is consistent with its clinical acti

Abstract SENTI-202 is being evaluated in an ongoing Phase I clinical trial, SENTI-202-101 (NCT06325748). SENTI-202 selectively kills AML blasts and leukemic stem cells (LSCs) via its bivalent CD33/FLT3 CAR (OR gate); and spares hematopoietic stem and progenitor cells (HSPCs), even if they express CD33/FLT3, from potential off-tumor on-target toxicity by recognizing the healthy cell antigen EMCN via its unique NOT gate. Here, we report SENTI-202 pharmacokinetics (PK) and pharmacodynamics data from peripheral blood (PB) and bone marrow (BM) samples from the entire dose finding cohort of 12 relapsed/refractory (R/R) AML patients (as of July 15, 2025). Clinical trial data are submitted as a separate abstract. Three dose cohorts were evaluated and the recommended phase 2 dose (RP2D) was determined as 1.5×109 CAR+ SENTI-202 cells/dose given on Days 0, 7, and 14 of a 28-day cycle following lymphodepletion with fludarabine and cytarabine, based on the totality of clinical and correlative data. Patients on study had differing mutational status and proteomics, as expected with this heterogenous disease. Consistent with our target selection approach, exploratory CyTOF analysis confirmed that all patients' disease expressed the SENTI-202 CAR targets CD33 on bulk blasts (range 14-99%) and FLT3 on LSCs (range 7-90%). 6/12 patients had blast reduction consistent with clinical response (4 measurable residual disease-negative complete remissions (MRD- CR), 1 MRD- CR with partial hematologic recovery (CRh), and 1 MRD+ morphologic leukemia-free state (MLFS) after 1-2 SENTI-202 treatment cycles (overall % composite complete remissions (cCR) = 42%, ORR = 50%). In the RP2D cohort, 50% (3/6 patients) achieved cCR. SENTI-202 was detected by ddPCR in PB of all patients after dosing. Median PK parameters for the RP2D cohort showed a Cmax of 1,878 (range 412-14,057) copies/μg DNA, a Tmax of 3 (range 0-7) days, area under the curve (AUC) of 3,272 (range 1,752-50,967) copies×days/μg DNA, and clearance of SENTI-202 from the periphery by day 14 post-treatment. The RP2D cohort PK was generally consistent with exposure across all patients with no apparent dose dependent increase in exposure. No significant differences in PK were noted between responders and non-responders. To achieve deeper and longer remissions, SENTI-202 was designed to target LSCs, which are often FLT3+ (+/-CD33+). Available CyTOF data showed that in 3 cCR patients SENTI-202 treatment resulted in a >10-fold decrease in LSCs (CD34+ CD38-/low and/or GPR56+ AML cells). These LSCs were predominantly in G0 phase of the cell cycle at baseline and therefore expected to be resistant to chemotherapeutic agents like cytarabine. In patient PB samples collected serially post SENTI-202 treatment, the repopulation of T cells, NK cells, neutrophils, and other immune cells was confirmed by flow cytometry in responders. The SENTI-202 Logic Gated Gene Circuit includes an inhibitory CAR recognizing EMCN (expressed by healthy HSPCs and rarely by AML) to protect HSPCs from potential off-tumor on-target toxicity and enable post-treatment hematopoietic recovery. In analyzed responders (3 patients), CyTOF analysis demonstrated that CD34+ CD38-/low HSPCs were either maintained or increased after SENTI-202 treatment. EMCN+ expressing cells were observed in the healthy hematopoietic populations in responder BM, and hematopoietic trajectories were preserved, with progenitor cell types enriched at baseline and more differentiated populations appearing after treatment. PK, flow cytometry, and CyTOF analyses of patient PB and BM in the dose-finding cohort affirm the anti-leukemic effects of SENTI-202 in responding patients, as well as the repopulation of immune cell subpopulations post-treatment. The presence of EMCN+ HSPCs in responder BM, as well as repopulation of BM and PB hematopoietic subpopulations, is consistent with the pharmacodynamic action of the SENTI-202 Logic Gate in sparing EMCN+ HSPCs while selectively killing CD33+ and/or FLT3+ AML tumor cells. An expansion cohort enrolling R/R AML patients has been opened at the confirmed RP2D, and we anticipate presenting data from additional patients at the conference.