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 activity and unique logic gated mechanism of action

ABSTRACT 5921 ASH 2025 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 activity and unique logic gated mechanism of action Muharrem Muftuoglu1*, Enping Hong2*, Nosha Farhadfar3, Stephen A. Strickland4, Ashish R. Bajel5, Alireza Eghtedar6, Gary Schiller7, Mahesh Basyal1, Li Li1, Lawrence Naitmazi2, Rochelle Emery2, Timothy Lu2, Brian S. Garrison2, Kanya Rajangam2, Farhad Ravandi1, Michael Andreeff1 * Denotes co-authorship. 1The University of Texas M.D. Anderson Cancer Center, Houston, TX, 2Senti Biosciences, Inc, South San Francisco, CA, 3Sarah Cannon Transplant and Cellular Therapy Program at Methodist Hospital San Antonio, TX, 4SCRI at TriStar Centennial, Nashville, TN, 5Peter MacCallum Cancer Centre, Melbourne, Australia, 6Colorado Blood Cancer Institute, Denver, CO, 7David Geffen School of Medicine at UCLA, Los Angeles, CA BACKGROUND PATIENT CHARACTERISTICS AND CLINICAL RESPONSES SENTI BIOSCIENCES, CONTACT email: [email protected] SENTI-202 PHARMACOKINETICS ACKNOWLEDGEMENTS The authors would like to thank all patients, their families, and caregivers for their participation in this study, as well as all supporting clinical and research staff including Amy Alford, Maria Garcia, and Nelia Leemans from Senti Biosciences. Central laboratory services were provided by IQVIA Laboratories. PBMC/BMMC processing was performed by Precision for Medicine. Pharmacokinetics assessment was performed by Charles River Laboratories. Cytokine analysis was performed by Meso Scale Discovery. Pharmacometrics support was provided by Certara. Presented at The 67 th ASH Annual Meeting 2025, December 6 - 9, Orlando, FL SENTI-202 is being evaluated after lymphodepletion in patients with relapsed/refractory (R/R) acute myeloid leukemia (AML) in an ongoing Phase I clinical trial, SENTI-202-101 (NCT06325748). SENTI-202 is a Logic Gated CAR NK cell therapy designed to target CD33 and/or FLT3 positive hematologic malignancies, including AML. To achieve deeper & longer remissions, SENTI- 202 is designed to target both bulk AML blasts and leukemic stem cells (LSCs), which are often FLT3+ (+/-CD33+), while enabling long-term multilineage hematopoiesis by maintaining a healthy HSPC population. The SENTI-202 Logic Gated Gene Circuit therefore includes an inhibitory CAR recognizing EMCN (expressed by healthy HSPCs, but rarely by AML) to protect HSPCs from potential SENTI-202-mediated off-tumor on-target toxicity even if they express CD33 and/or FLT3 and enable post-treatment hematopoietic recovery. Here, we report preliminary correlative data evaluating hematopoietic cell fractions in bone marrow (BM) and peripheral blood (PB) from 18 treated R/R AML patients with available data from the ongoing study (October 17th, 2025). Efficacy and clinical results are discussed in our clinical oral presentation (abstract# 1044). aCAR: activating chimeric antigen receptor; EMCN: endomucin; HSC: hematopoietic stem cell; HSPC: hematopoietic stem and progenitor cell; iCAR: inhibitory chimeric antigen receptor; LSC: leukemic stem cell. SENTI-202-101 Dosing and Correlative Sampling Timepoints SENTI-202 Transgene Copies/µg DNA Cycle 1 a SUMMARY SENTI-202 is a First-in-Class Off-the-Shelf Logic- Gated Selective CD33 OR FLT3 NOT EMCN CAR NK Cell Therapy for Blood Cancers SENTI-202-101 Study Design RP2D identified as Schedule I, Dose Level 2 LSC (CD34+ CD38low) in Responders (ORR) LSC (GPR56+) in Responders (ORR)AML blast reduction* % CD34+ CD38low LSC of all cells Baseline C1D28 C2D28 100 10 1 0.1 0.01 0 % GPR56+ LSC of all cells 100 10 1 0.1 0.01 0 Baseline C1D28 C2D28 Baseline Best Response Total Bone Marrow Blasts (%) 100 80 60 40 20 0 Baseline C1D28 % HSPC of all cells 10 1 0.1 0.01 0 100 HSPC (CD34+ CD38low) in Responders (ORR) * Patient 12-07 samples had low sample viability. * Cycle 2 Pharmacokinetic (PK) analysis of SENTI-202 in patients. SENTI-202 was detected in all treated patients, (a) with data shown fromCycle 1 and (inset) Cycle 2. SENTI-202 PK profile was consistent with allogeneic NK cell therapy with cells detected in the first 2 weeks in periphery followed by clearance thereafter. (b) A preliminary signal of higher SENTI-202 exposure (Cmax and AUCifo) was observed for the RP2D cohort compared to dose level 1. Similarly, responders (ORR: CR/CRh and MLFS) also showed a preliminary signal of higher exposure compared to non-responders when assessed in all patients (higher AUCifo) and in the RP2D cohort only (higher Cmax and AUCifo). The PK trends noted above were not statistically significant. Peripheral clearance was defined as the last time point where a non-zero SENTI-202 signal was noted; 3 patients with ongoing PK collectionswere excluded from this analysis. Estimated LLOQ assumes equivalent DNA loading for each reaction. Note: Patient 06-01 Cycle 1 samples were processed out of stability window andare excluded from analysis of PK parameters. PK Parameters Median (range) given for all Cmax (Copies/µg DNA) AUCifo (Days*Copies/ µg DNA) Tmax (Days) Peripheral clearance (Days) Dose Level 1 (1x109 CAR+ NK/dose) Schedule I, 3 doses (N=3) 2392 (1567 - 2496) 6873 (3160 - 11074) 5 (0 - 7) 15 (14 - 26) Schedule II, 5 doses (N=3) 1358 (380 - 8212) 2842 (720 - 29067) 7 (3 - 14) 15 (14 - 17) Dose Level 2 (1.5x109 CAR+ NK/dose) Schedule I, 3 doses (N=12) (RP2D) 3275 (412 - 76600) 13494 (1752 - 434178) 3 (0 – 8) 15 (14 – 23) RP2D Non-responders (N=6) 3163 (412 - 18750) 12395 (1752 - 37852) 3 (0 - 3) 17 (15 - 23) Responders (ORR) (N=6) 6021 (676 - 76600) 19634 (2777 - 434178) 3 (2 - 8) 14 (14 - 17) All patients Non-responders (N=9) 2655 (380 - 18750) 5743 (720 - 37852) 3 (0 - 14) 15 (14 - 23) Responders (ORR) (N=9) 2472 (676 - 76600) 8030 (2777 - 434178) 3 (0 - 8) 15 (14 - 26) All (N=18) 2496 (380 - 76600) 8030 (720 - 434178) 3 (0 - 14) 15 (14 - 26) SENTI-202 PHARMACODYNAMICS: CYTOKINES a b c SENTI-202 MOA: AML BLAST & LSC ANALYSIS -1 0 1 2 Log10(median) Responders (ORR) Non-responders C2D28 %CD33+ (blasts) %FLT3+ (LSCs) %EMCN+ (blasts) Patient 06-01 (baseline) 96.6 41.0 0.03 Patient 01-04 (baseline) 99.1 12.4 0.02 Combined AML and healthy bone marrow samples Monocytes CMP/GMP Pro-erythroblasts MEPs Polychromatic erythroblasts pDCs / DCs Early myeloid progenitors MPPs Pre-B-cells HSCs/HSPCs (EMCN+) Pro-B cells Orthochromatic erythroblasts CLPs Mature B-cells Baseline a b Baselined e Best response f LONGEST SURVIVING RESPONDERS (Patients 1 & 2) Examples of long term responses observed after SENTI-202 treatment. (a) Patients 06-01 and 01-04 both achieved MRD negative CR and are both maintaining CR at >1 year after treatment. (b) Both patients had high %CD33+, moderate %FLT3+, and no %EMCN+ cells in their AML populations at baseline. (c) Both patients achieved MRD negativity at best response, and showed 10-1000 fold reduction in %LSCs and increase or maintenance of %HSPCs. (d) EMCN+ hematopoietic populations (blue contours) were retained after SENTI-202 treatment. (e) UMAP analysis of patients 06-01 and 01-04 showed the preservation of hematopoietic trajectories in responders, with progenitor cell types enriched at baseline and more differentiated populations appearing after treatment. (f) A description of the hematopoietic trajectories as defined by healthy and diseased cells is provided. Patient 06-01 Cycle 1 Day 28 BM sample (partial response) was used as baseline due to the pre-treatment sample being processed out of stability window. Best responsec Patient 06-01Patient 01-04 Patient characteristics and responses in the SENTI-202-101 clinical trial at study entry. (a) Representation of the ELN risk classification, FAB subtype, response, number of SENTI-202 cycles given, SENTI-202 target expression, and mutational profiles of the 18 patients with relapsed/refractory AML in the SENTI-202-101 clinical trial. Data from an open clinical database as of October 17th, 2025. (b) UMAP plots of responder (ORR) patients at baseline and best response, showing the disappearance of AML cells in patient bone marrow during response. Cell populations are marked with different colors, and the AML population (dark blue) isfurther circled in red. (c) Heat map analysis of CyTOF markers expressed by AML populations in bone marrow samples show the prevalence of AML-associated markers and their heterogeneity of expression across all patients. The log(median intensity) of allchannels were grouped by function and ranked by average intensity across all patients. a c e a c Analysis of AML blast reduction and leukemic stem cells (LSCs) in patients. (a) AML blast reduction (as measured by local pathology and response assessed by investigator) was noted in responders (ORR) and in some non-responders. (b) Consistent with the target selection strategy for SENTI-202, CD33 and FLT3 were detected in the majority of bulk blasts and LSCs respectively, and EMCN was detected rarely in bulk blasts. LSCs in responders were identified by two different methods within the AML population: (c) CD34+ CD38-/low and (d) GPR56+, and their proportions in patient bone marrow were shown to decrease at least 10-fold after SENTI-202 treatment. In patients with CD34- disease (N=2), LSCs were defined as CD38-/low rather than CD34+ CD38-/low. (e) Cell cycle analysis of CD34+ CD38-/low LSCs at baseline, defined by Ki67 and IdU: Ki67- = G0, Ki67+ IdU- = G1, IdU+ = S. Cell cycle analysis determined that most LSCs in treated patients were quiescent at baseline, and therefore not expected to be eliminated by lymphodepleting chemotherapy such as fludarabine/cytarabine. d • Substantial AML blast reduction was observed in 50% of patients. 9/18 patients evaluable for best overall response had blast reduction by CyTOF consistent with clinical response of CR/CRh (7) or MLFS (2). All CR patients were MRD- as assessed locally. Longest surviving responders are currently >450 and >525 days post-initiation of SENTI-202 treatment. • AML disease characterization revealed unique mutational and proteomic signatures between patients . Of note, all patient AMLs expressed CD33 (range 14- 99% CD33+), and 5/7 CR/CRh patients expressed FLT3 on LSCs (range 12-41% FLT3+). • Responder AML leukemic stem cell (LSC) numbers decreased at least 10X after treatment. Most LSCs were in G0/G1 at baseline and would not be expected to be eliminated by traditional chemotherapy approaches. • Responder hematopoietic stem and progenitor cell (HSPC) numbers were increased (or maintained) during SENTI -202 treatment. While MLFS patients showed responses against AML disease, they showed no detectable HSPCs at baseline or response. • Consistent with NK biology and the observed well-tolerated safety profile of SENTI-202 (clinical data summarized in abstract 1044) and in contrast to CAR-T cells, serum cytokine levels from a 19-cytokine panel were multiple-fold lower than the highest cytokine values reported in patients receiving CAR-T treatment. • SENTI-202 was detected in all treated patients, with comparable PK to other CAR NK cell therapies. Preliminary findings show that RP2D and responders had higher SENTI-202 exposure (AUCifo, Cmax) compared to dose level 1 or non-responders, respectively. • Patients responding to SENTI-202 treatment had pharmacodynamics consistent with the designed SENTI-202 MOA. Consistent with the OR Logic Gate mechanism of action and SENTI-202 target selection strategy, CR/CRh/MLFS responders showed AML blast and LSC reduction. Consistent with the NOT Logic Gate mechanism of action, CR/CRh responder baseline HSPCs were increased (or maintained) during SENTI -202 treatment, and experienced peripheral blood cell count recovery. Responding patients with no detectable HSPC at baseline achieved MLFS and no peripheral immune cell reco very. Consistent with design of the calibrated release IL15 that localizes IL15, there was no marked elevation in peripheral IL15 levels beyond that obser ved due to lymphodepletion. The observed effects are consistent with the pharmacodynamic action of the Logic Gate in SENTI -202 in sparing EMCN+ HSPCs while selectively killing CD33+ and/or FLT3+ AML tumor cells. Additional dose expansion is on-going. UMAP plots of Responder (ORR) Bone Marrow CR/CRh MLFS SD PD Peripheral blood recovery: Responders (ORR) Peripheral blood recovery: Non-responders b Responders (ORR) Non-responders SENTI-202 targets Myeloid HSC / LSC Lymphoid NK related Apoptosis Signaling Proliferation 100 1 0.1 0 HSPC in Responders (ORR) at baseline % HSPC of all cells b CR/CRh MLFS (a) HSPCs were identified as CD34+CD38-/low cells within the healthy hematopoietic cell population, which clusters distinctly from AML cells in UMAP analysis of CyTOF data. Of note, amongst all responders (ORR), patients with detectable HSPC at baseline achieved CR/CRh with SENTI-202, and the proportion of HSPCs in bone marrow was either increased or maintained during SENTI-202 treatment consistent with SENTI-202’s unique logic-gated NOT gate mechanism of action designed to selectively kill AMLblasts and LSCs while sparing HSPCs. (b) Patients with no detectable HSPC at baseline achieved an MLFS without accompanying blood count recovery. (c) Similarly, flow cytometry analysis of peripheral blood cells showed that most responders achieving CR/CRh had robust immune recovery by the end of each cycle, consistent with normal hematopoiesis, while patients with MLFS, SD, or PD did not. SENTI-202 MOA: HSPC ANALYSIS G0 G1 S Cell cycle phase 0 20 40 60 80 100 Baseline LSC cell cycle status in all patients Patient 06-01 (CR MRD-) Patient 01-04 (CR MRD-) Baseline Best Response Baseline Best Response Patient 07-06 (CR MRD-) Patient 12-07 (CRh MRD-) Baseline Best Response Baseline Best Response Patient 05-10 (MLFS MRD+) Patient 12-15 (CR MRD-) Baseline Best Response Baseline Best Response Patient 07-16 (CRh MRD+) Patient 05-18 (MLFS MRD-) Baseline Best Response Baseline Best Response Patient 06-21 (CR MRD-) Baseline Best Response AML blasts Monocytic Erythroid HSC/HSPCs CD4 T cells CD8 T cells NK cells B cells Patient 06-01Patient 01-04 Patient 06-01Patient 01-04 LLOQ Time (days) * * Patient 12-07 samples had low sample viability. 10 * Blast reduction as measured by local pathology and assessed by investigators. b SENTI-202 Target Expression in Patient Disease % of LSCs LLOQ SENTI-202 Transgene Copies/µg DNA Time (days) (f) Serum samples were assayed for a panel of 19 cytokines, including IL-6, IL-10, GM-CSF and others as shown above. Serum samples were tested at the same time points listed in (a). Consistent with NK biology and the observed well-tolerated safety profile of SENTI-202 (clinical data summarized in abstract 1044) and in contrast to CAR-T cells, the maximum cytokine levels across the panel were multiple-fold lower than the highest cytokine values reported in patients receiving CAR-T treatment. Dashed lines indicate the maximum cytokine level observed in literature from Teachey et al (2016), highest value observed from adult cohort treated with anti-CD19 CAR-T (added where literature data available). Individual subjects are plotted according to the provided legend above. # Samples above LOQ were observed for 3/18 patients for IL-6, IL-8, and IFNγ at the marked time points. f Methods • Whole blood and BM aspirate are collected into sodium heparin tubes, then processed to cryopreserved PBMC/BMMC by Ficoll separation. • Pharmacokinetics was determined by detection of the SENTI-202 transgene in PBMC samples by ddPCR. • BMMC samples were analyzed by a 49-marker mass cytometry (CyTOF) panel. • Serum samples were collected for cytokine analysis by MSD. • Additional whole blood samples collected in Cyto-Chex BCT tubes were analyzed by flow cytometry. BM aspirate collections (CyTOF) Lymphodepletion (Fludarabine/Cytarabine) Whole blood samples are collected at pre-dose and 0.5h/8h/24h after each dose of SENTI-202, along with other timepoints as specified above. Schedule I: 3 doses of SENTI-202 (1x109 or 1.5x109 CAR+ NK/dose) Schedule II: 5 doses of SENTI-202 (1x109 CAR+ NK/dose) Serum cytokine collections Whole blood collections (PK and flow cytometry) SENTI-202 dosing D0 D28D14 D21D7D3 D10Baseline Dose 1 Dose 2 Dose 3 Dose 4 Dose 5 Dose 1 Dose 2 Dose 3 D0 D28D14 D21D7D3 D10Baseline Interim analysis of patient serum samples was performed at selected time points (a). Serum IL-15 levels were modestly elevated (maximum values >10- fold lower than that reported for CAR-T cells) prior to SENTI-202 dosing (Day 0 pre-dose) and following LD, as expected from removal of cytokine sinks such as Treg cells. Overall, IL-15 levels were not elevated beyond 80 pg/mL across all patients. There was no trend to differences in IL-15 levels between responders (ORR) and non-responders (c), between dose levels 1 and 2 (d) and between patients who experienced any adverse event of interest (e) which were most frequently Grade 1-2 pyrexia as summarized in clinical data presentation (abstract 1044). These findings are concordant with preclinical data showing minimal to undetectable IL-15 levels in vitro and in vivo in NSG AML mouse models dosed with SENTI-202 (internal data) and consistent with the localizing design of the calibrated release IL-15. * IL-15 levels referenced from Teachey et al (2016), highest value observed from adult cohort treated with anti-CD19 CAR-T. Dosing schedule Time points tested Schedule I (3 doses) Day -7 (baseline), pre-dose and 8h post each dose, Day 3 Schedule II (5 doses) Day -7 (baseline), Day 0 pre- dose, 8h post each dose a b c d e # # # # # # C1D28 sample (partial response) used due to pre-screen sample being processed out of stability window. Max. level in CAR-T B-ALL from literature* CR = MRD negative PR Tx = HCT C1 = Cycle 1 start C2 = Cycle 2 start Pt. 06-01 Pt. 01-04 ECOG PS: European Cooperative Oncology Group performance status; MTD: maximum tolerated dose; RP2D: recommended phase 2 dose; ELN: European LeukemiaNet; CyTOF: Cytometry by Time-of-Flight; BM: bone marrow; 1Per WHO 2022 Classification * Sample not available No HSPCs detected at baseline for MLFS patients

The release as fetched from its publisher. sentibio.com ↗