Multiplex Drug Screening Identifies a Novel Therapeutic Backbone for the Treatment of Acute Myeloid Leukaemia

Introduction: Acute Myeloid Leukemia (AML) is a blood cancer characterized by excessive proliferation, abnormal self-renewal, and/or differentiation defects of myeloid progenitor cells. Genomic sequencing has shown heterogeneous subpopulations characterized by distinct genetic and epigenetic alterations, suggesting a combination of therapies are required to effectively treat this disease. Combination therapies are a preferable and superior approach for the treatment of leukaemia: they can work in a synergistic manner allowing a lower therapeutic dose to be achieved, minimising potential long-term effects. Combination therapies can overcome adaptive resistance by targeting multiple disease drivers. To identify novel and effective drug combinations for the treatment of AML, we performed in vitro and ex vivo sensitivity profiling against a panel of 1128 pairwise combinations. Methods: The combinations were designed using a multiplex screening protocol established in our lab from a panel of 48 compounds representing standard of care chemotherapy and other near clinic agents, the doses used for each compound were an approximate IC30 based on published data for in vitro cell death. Using our all-pairs testing algorithm, compounds were grouped into pools of 5 compounds per well generating 1128 (48 x (48-1)/2) pairwise combinations to be accommodated in 172 individual wells. Concurrently, with the combination screen, a single-agent screen of the 48 compounds was performed. The biological readout for this experiment was CellTox™ Green Cytotoxicity Assay, fluorescence intensity was evaluated following 72-hour incubation. Four AML cell lines (OCI-AML3, THP-1, HL60 and SKM-1) representing a spectrum of cytogenetic abnormalities were used for this screening approach. Leukemic cells were isolated from the peripheral blood (PB) or bone marrow (BM) of fresh, diagnostic patient samples and ex vivo sensitivity profiling assessed. For each combination a robust Z-score was calculated using normalised, relative fluorescence values. In parallel, cell viability was calculated using a cytotoxic control as maximum cell death. Combinations which demonstrated a Z-score >2 and/or a cell viability of <30% were deconvoluted into 10 pairwise and 10 triple combinations to identify a clinically feasible combination. Results: Five, 5-compound combinations were taken forward for deconvolution in the OCI-AML3 cells. This identified the pairwise combination of dactinomycin, a chemotherapy medication and vorinostat, a histone deacetylase (HDAC) inhibitor, by demonstrating a significant reduction in cell viability (p=0.039). Combination Index (CI) values, calculated using the Chou and Talalay method, indicated a synergistic (CI=<1) relationship in OCI-AML3 cells. Previous studies indicate that dactinomycin is an effective therapeutic option for AML patients with a NPM1 gene mutation. Given the NPM1 mutant status of the OCI-AML3, we validated using 4 AML cell lines (OCI-AML2, HL60, Kasumi-1 and MV4-11) all with WT NPM1. Kasumi-1 and MV4-11 cells showed a significant reduction in cell viability with this combination (p=0.0003 and p=0.025, respectively). This was accompanied with a significant increase in Annexin V positivity in both of these cell lines (p=<0.005 and p=<0.0001, respectively). The highest scoring 5-compound combination across patient samples was combination 114 (Average Z-Score = 1.983), comprising enasidenib, idarubicin, pracinostat, ruxolitinib and dactinomycin. In 57% of the patient samples deconvolution identified the triple combination of a therapeutic backbone of idarubicin, an anthracycline chemotherapy and pracinostat, a HDAC inhibitor, with the addition of a "patient” specific third agent. Validation confirmed that the combination of idarubicin and practinostat can significantly reduce viability in comparison to the single agents (p=<0.05). Conclusion: In this study we demonstrate the utility of a novel multiplex screening approach with a custom panel of compounds to identify combination therapies for the treatment of AML. The combination of an anti-tumour antibiotic chemotherapy (dactinomycin and idraubicin) and a HDAC inhibitor (vorinostat and pracinostat) demonstrated the greatest efficacy following in vitro and ex vivo sensitivity profiling. Further investigations are required to elucidate the mechanism of action of these novel combinations.