Homoharringtonine and Mebendazole Are Preferentially Lethal Against Models of Myeloid Malignancy Associated with Germline Mutant RUNX1

RUNX1 is a master-transcriptional regulator involved in normal and malignant hematopoiesis. Majority of mono-allelic germline mutations in RUNX1 are missense, large deletions or truncation mutations, behaving mostly as loss of function (LOF) mutations. They are ~40%-penetrant and cause Familial Platelet Disorder (RUNX1-FPD) that has a propensity to evolve into myeloid malignancy (FPD-MM), i.e., MDS or AML. FPD-MM harbors co-mutations, most commonly on the second allele of RUNX1, and on BCOR, PHF6, K-RAS, WT1 or TET2, which confer relative resistance to standard therapy for MDS or AML. Although curative in some patients with FPD-MM, allogeneic transplantation from matched, un-related donors carries risk of graft versus host disease and frequent AML relapse. This creates a strong rationale and an unmet need to develop novel targeted therapies for FPD-MM. We previously reported on utilizing the RNA-Seq signature of RUNX1 knockdown, which exerted more lethality in AML cells with mutant (mt) RUNX1 compared to AML harboring two copies of wild-type RUNX1, for conducting LINCS1000-CMap analysis. This identified several expression mimickers (EMs), including the protein synthesis inhibitor homoharringtonine (HHT or omacetaxine) and anthelmintic fenbendazole (analog of mebendazole). Present studies demonstrate that treatment with HHT or mebendazole (MB) dose-dependently induced significantly greater loss of viability in four patient-derived (PD) bone marrow aspirate (BMA) samples of FPD-MM (3 AML and 1 MDS) compared to RUNX1-FPD (3 samples) or in normal CD34+ progenitor cells. In a patient with RUNX1-FPD (expressing mtRUNX1 K194N), who developed FPD-MM, following co-mutations were documented by NGS: BCOR A1437fs, PHF6 L324fs, SF3B1 D781G and SRSF2 P95R/L. From BMA of this patient, we successfully established the first ever, continuously cultured cell line (GMR-AML1) expressing the same germline mtRUNX1. GMR-AML1 cells were cytogenetically diploid and lacked MYC or MLL1 rearrangement, or any copy number gains or losses on array CGH. However whole exome sequencing (WES) identified additional mutations in TP53 (P72R), AIM2 (K340fs), NELFB (L523F), CEP152 (Y370X), SUGP2 (H23L), RRM2B (R71fs), TADA3 (T27R), SPDYE6 (G292C) and PRDM9 (S814R) with % VAF ranging between 33 to 55%. GMR-AML1 cells exhibited high surface expression of CD117 (c-KIT), CD123 (IL3R), CD86 and CD33, but without expression of CD34, CD14, CD11b, MPO or CD135 (FLT3). Compared to the AML OCI-AML5 cells with somatic mtRUNX1, GMR-AML1 cells demonstrated markedly reduced protein expression of RUNX1, RUNX2, PU.1, c-Myb, GFI1, GFI1B, FLT3, MEIS1 and CEBPα (p42), but much higher protein expression of RUNX3 and NOTCH (p120). CRISPR-Cas9 knockout of RUNX3 in GMR-AML1 cells restored RUNX1 expression, while significantly increasing % of differentiated cells. Although dose-dependently sensitive to daunorubicin, etoposide, cytarabine and panobinostat (class I and II HDAC inhibitor), GMR-AML1 cells were relatively insensitive to venetoclax, A1155463 (Bcl-xL inhibitor), AZD-5991 (MCL1 inhibitor), azacytidine or decitabine. Notably, treatment with HHT or MB dose-dependently induced loss of viability of GMR-AML1 cells (LD50: 40 and 330 nM, respectively). Additionally, co-treatment with HHT and venetoclax synergistically induced apoptosis in GMR-AML1 cells, as determined by the SynergyFinder algorithm. This synergy in GMR-AML1 cells was associated with abrogation of venetoclax-induced increase in MCL1 and Bcl-xL levels, as well as greater decline in levels of RUNX3, PU.1, c-Myb, c-Myc, MPL and CDK4/6. Tail vein infusion and engraftment of luciferase-transduced GMR-AML1 (10 ⁶ cells) caused marked splenomegaly and 100% mortality of NSG mice by day-18, post-infusion. We will present at the ASH meeting findings of ongoing in vivo studies determining effects of treatment with HHT and/or venetoclax, versus vehicle control, on AML burden and overall survival of NSG mice engrafted with GMR-AML1 cells. Overall, preclinical findings presented here highlight the molecular and genetic features associated with progression of RUNX1-FPD to FPD-MM, especially in the newly established GMR-AML1 cell line. They also demonstrate that HHT or MB are preferentially more lethal against FPD-MM versus RUNX1-FPD cells and exert synergistic lethality with venetoclax against GMR-AML1 cells.