Pb1761: preclinical evaluation of the menin inhibitor ziftomenib in primary aml specimens

Topic: 3. Acute myeloid leukemia - Biology & Translational Research Background: Leukemia stem cells (LSCs) in acute myelogenous leukemia (AML) are significant drivers of disease progression and are historically treatment-resistant. Venetoclax combined with azacitidine leads to high response rates (~60%) in elderly de novo AML patients at least partially through targeting of LSCs. A significant portion of these patients ultimately experience disease relapse or upfront resistance. Multiple characteristics including immunophenotype, genotype and metabolic mechanisms contribute to drug resistance and relapse. Furthermore, alternative agents and targets have been proposed in combination with venetoclax including inhibitors of the menin-KMT2A interaction. Early studies suggest these inhibitors are effective as single agents. However, little is known about targeting LSCs with menin inhibitors. Aims: Ziftomenib (zifto) is a potent, oral and selective menin inhibitor that has demonstrated encouraging monotherapy activity in a phase 1 trial in relapsed/refractory AML. The aim of this study is to provide a better understanding of the ability of zifto to target LSCs in AML through patient-derived xenografts (PDX) and multi-omics profiling. Methods: PDX models established from primary AML cells were treated with zifto, venetoclax (ven), azacitidine (aza), and a combination of these agents. Disease response was assessed by flow cytometry. These primary AML cells and an expanded cohort underwent multi-omics profiling including single cell transcriptional profiling with surface phenotyping (CITE-seq). An AML cell line was utilized to understand transcriptional changes in response to in vitro dosing with zifto. Results: Three patient specimens were employed in PDX studies representing diverse genetic profiles and different disease stages. Following engraftment of each specimen in NSG-S immune deficient mice, animals were treated using a regimen that included vehicle control, zifto alone, as well as ven/aza, and zifto/ven/aza. In two of the three specimens, we observed complete eradication of human AML cells in vivo with single agent zifto. CITE-seq analysis was leveraged to investigate genes and pathways that are differentially expressed between responder and non-responder blasts. MEIS1 is significantly upregulated in responder blasts, consistent with MEIS1 as a biomarker. The HOX gene axis is also differentially expressed; however, the pattern of expression differences between responder and non-responder blasts is HOX gene and blast differentiation status dependent. Additionally, a zifto dose and time course bulk RNA-seq profile in the MOLM13 AML cell line allowed us to develop a zifto response gene signature, which we used to interrogate the CITE-seq data. This analysis showed that the genes that had a similar response to MEIS1 with zifto treatment in cell lines also trended as upregulated in responder blasts. Summary/Conclusion: In conclusion, our data show that in some patient specimens venetoclax resistance can be overcome with ziftomenib as a single agent. These findings suggest that ziftomenib-mediated targeting of AML cell populations, including LSCs, may be an effective strategy for preventing disease progression in patients who are refractory or relapse in response to initial ven/aza treatment. Combination of ziftomenib and venetoclax needs to be further studied but may also represent a viable strategy in the presence of known characteristics corresponding to venetoclax resistance. Further investigation of LSCs in patients treated with ziftomenib will lead to a deeper understanding of the mechanism of action of ziftomenib and may reveal additional biomarkers of response. Keywords: Meis1, Acute myeloid leukemia, MLL