Synergistic Activity of BMS-986397, a First-in-Class a Cereblon (CRBN) E3 Ligase Modulator (CELMoD) Targeting Casein Kinase 1α (CK1α), in Combination with Venetoclax and/or Azacitidine in Preclinical Models of Acute Myeloid Leukemia (AML)

Introduction: Treatment of AML (Acute Myeloid Leukemia) has been clinically challenging with modest success and poor patient outcomes. The limited efficacy of single-agent treatments has encouraged the investigation of rationally designed combinations to increase drug activity and achieve deeper and prolonged remission avoiding the development of resistance. Azacitidine (AZA) and BCL2 inhibitor venetoclax (VEN)-based combinations became the new standard of care for the treatment of newly diagnosed AML patients who are ineligible for intensive chemotherapy upon its approval in 2018 ^1-3. Despite this recent approval, reduced response rates in mature phenotypes (AML M4 and higher), tolerability, and duration of response remain suboptimal. Therefore, innovative combination approaches are still a need for AML. A first-in-class, potent, and selective CK1α-targeting oral CELMoD, BMS-986397, is being investigated for the treatment of R/R AML and HR-MDS (High-Risk Myelodysplastic Syndrome) harboring functional TP53 (Tumor protein 53) in a clinical study (NCT04951778). Owing to its mechanism, combination with other agents that target apoptosis pathways such as VEN or hypomethylating agents, could improve efficacy in AML. In this regard, this study seeks to investigate a convenient strategy for BMS-986397 combinations that would circumvent resistance and inform future decisions in the FIH trial to improve the treatment of R/R leukemia, including specific AML subtypes.

Methods: Viability studies in a panel of 16 AML cell lines were performed to test the combinatorial potential of BMS-986397 with VEN and/or AZA. These studies were expanded with ex vivo clonogenic assays in bone marrow mononuclear cells (BMNCs) from 15 TP53 WT (Wild Type) AML patients as a single agent and, in double or triple combinations with VEN and AZA. The impact on normal hematopoietic stem (HSC) and progenitor cells was evaluated by assessing the effect on the growth of granulocyte/monocyte (CFU-GM) and burst forming unit of erythrocytes (BFU-E) progenitor cells in human bone marrow (BM) CD34+ cells from 4 healthy volunteers (HV). Leukemic blasts were phenotypically and molecularly characterized, including different FAB subtypes. A drug combination analysis was conducted to determine synergy (Bliss) and additivity (HSA) scores. In vivo studies in 2 AML-PDXs (Patient Derived Xenograft) from patients with different AML subtypes were explored in monotherapy and in VEN and/or AZA combinations followed by an immunophenotyping assessment of the blood, spleen, and BM and PD (pharmacodynamics) of human primary leukemic cells.

Results: BMS-986397 shows additivity/synergy in 7/16 AML cell lines in combination with VEN and/or AZA. Ex vivo testing in normal hematopoietic cells reveals that double combinations of BMS-986397 with VEN or AZA do not induce a significant impact on CFU-GM and BFU-E when compared with their single agent activity. However, triple combination of BMS-986397/VEN/AZA exhibits an increase on myelotoxic effect on these healthy progenitors (median IC₅₀ 291.6nM in doublets vs 88.1nM in triplets). In AML patient samples, BMS-986397/VEN or BM-986397/AZA combinations exhibit similar activity inhibiting the growth of AML leukemic progenitor cells compared with the triplet combination and, in most cases, similar or superior activity to the VEN/AZA combination. In 12 out 15 AML samples, additive effect mediated by triple BMS-986397/VEN/AZA combination is recapitulated for at least one of the double combinations, overcoming the potential toxicity provoked by triplets in normal hematopoietic progenitors. Interestingly, acute monoblastic/monocytic leukemia (AML M5) samples, described as resistant to VEN/AZA treatment ⁴, have shown benefit after including BMS-986397 in the combination backbone (6.6 fold-change excess additivity score). In vivo pharmacology studies have also evidenced an improvement in mice survival to >80 days (vs 60 days in VEN/AZA groups) with no signs of leukemia after BMS-986397 and/or VEN/AZA combinations, even in M5 AML-PDX models.

Conclusions: BMS-986397 combination with VEN and/or AZA unveiled preclinical efficacy improvement with tolerable safety in primary AML. These data support the use of combination-based therapies targeting different mechanism of action to achieve expanded clinical success transforming the therapeutic landscape for patients diagnosed with AML.