Genomic and phenotypic impact of FHD-286–induced inhibition of SMARCA4/2 in patients with relapsed/refractory myeloid malignancies

Abstract Background: BAF complexes control chromatin accessibility to regulate gene expression programs important for cellular proliferation and differentiation. BAF activity constitutes a key dependency in several cancer types, including myeloid malignancies. FHD-286 is a first-in-class dual inhibitor of SMARCA4/2 (BRG1/BRM), the ATPase subunits of the BAF complexes. FHD-286 monotherapy induced myeloid differentiation in patients (pts) with relapsed/refractory (R/R) acute myeloid leukemia (AML) or myelodysplastic syndrome (MDS), and the addition of decitabine (DAC) produced objective responses. Methods: In a multicenter, open-label, Phase 1 dose escalation study (NCT04891757), pts with R/R AML, MDS, or chronic myelomonocytic leukemia received FHD-286 monotherapy or in combination with DAC. FHD-286 doses ranged from 2.5 to 10mg once daily (QD) as monotherapy and from 1.5 to 7.5/10mg QD (interval dosing: doses alternated every 2 weeks) in combination with DAC, given in 28-day cycles. Bone marrow aspirates (BMA) for mutational, cytogenetic, flow cytometry, and transcriptomic analyses were obtained at screening, on cycle 1 day 15 (C1D15), and the first day of every cycle thereafter. Blood samples for pharmacokinetic and flow cytometry analyses were obtained predose on C1D1 and at multiple on-treatment time points, with more frequent sampling during the first 1-3 cycles. Multiparameter flow cytometry was performed on 135 BMA from 51 pts. Ninety samples from 40 pts were also analyzed by single-cell RNA sequencing. High-quality transcriptomes were obtained for 574,966 BM cells. Data were normalized and integrated, enabling dimensionality reduction and clustering of the entire dataset. Cell type identification was conducted using several orthogonal approaches, including singleR (doi: 10.1038/s41590-018-0276-y), Azimuth (doi: 10.1016/j.cell.2021.04.048), and BoneMarrowMap (doi: 10.1158/2643-3230). The identity of the leukemic blast compartment was refined using an AML expression signature (doi: 10.1038/s41467-023-41994-0) and CopyKat aneuploidy analysis (doi: 10.1038/s41587-020-00795-2), identifying 286,655 BM blasts. Results: Marked downregulation of leukemic stemness markers CD34 and KIT (CD117) and upregulation of myeloid maturation markers CD11b, CD15, and CD64 were observed in BM blasts by flow cytometry, suggesting blast differentiation. These biomarker changes were dose dependent and correlated with FHD-286 plasma exposure. At similar FHD-286 plasma exposure levels, these biomarker changes were generally less profound with FHD-286+DAC than monotherapy. However, in 3 of the 5 pts with objective responses and paired BM samples in the combination cohort (1 complete remission with incomplete platelet recovery, 1 partial remission, 1 morphological leukemia-free state), changes in differentiation markers, particularly CD34 and CD11b, more closely resembled those observed with monotherapy. Differential gene expression analysis of the blast compartment identified several novel genes whose expression was highly correlated with FHD-286 exposure. Broad impacts on leukemic stemness and myeloid and erythroid maturation gene sets were observed, although myeloid maturation effects were weaker with FHD-286+DAC, consistent with flow cytometry results. Inferred copy number analysis using the computational tool NumBat (doi: 10.1038/s41587-022-01468-y) was consistent with cytogenetic lab results, enabling discrimination of the clonal architecture of individual pts. An analysis of clonal architecture dynamics will be presented at the meeting. Additionally, this analysis identified differentiated blasts that clustered with normal monocytes and erythroid cells; these findings were further confirmed using BoneMarrowMap pseudotime analysis. Among nonresponders treated with FHD-286+DAC, the magnitude of the transcriptional impact was generally lower than that observed with FHD-286 monotherapy. However, among objective responders to FHD-286+DAC, the magnitude appeared to exceed that of FHD-286 monotherapy.Conclusions: FHD-286 induced broad immunophenotypic and transcriptional changes in leukemic blasts, consistent with decreased stemness and increased myeloid differentiation. In the combination therapy cohort, responders exhibited stronger transcriptional impacts than nonresponders or monotherapy-treated pts. Additional studies will be needed to understand the context in which FHD-286+DAC leads to greater efficacy, and may support pt enrichment strategies.