Abstract NG14: Reduction in mitochondrial priming drives resistance to targeted therapy in acute myeloid leukemia

Relapse is the leading cause of treatment failure in acute myeloid leukemia (AML) patients. FDA approval of 8 targeted therapies in the past two years has drastically altered the landscape of AML treatment. Despite this success, the duration of clinical response is limited by the frequent development of acquired drug resistance to targeted therapy. To identify mechanism of resistance and to search for therapies that overcome resistance, we adopted broadly applicable functional approach to precision medicine called “dynamic BH3 profiling” (DBP) and coupled it with RNA-seq and targeted exome sequencing technology. DBP measures drug induced early death signaling using BH3 peptides that mimic pro-apoptotic BH3 proteins. To first validate the utility of DBP as precision medicine tool we created landscapes of pharmacologic sensitivity of 17 patient-derived xenograft (PDX) AML models to a panel of 40 clinically relevant agents, together with genomic and transcriptomic profiles. Aggregated across the panel, unsupervised clustering of drug-induced apoptotic signaling using DBP by itself could segregated PDXs according to prior treatment status (PDXs from treatment naïve patients clustered distinctly from R/R PDXs). While genomic mutations and transcriptomic signature profiles between R/R and treatment naïve PDXs did not show significantly distinct clustering patterns. Next we show that DBP could predict in vivo responses of drugs of widely varying mechanism of action, including a FLT-3 inhibitor, BCL-2 and MCL-1 inhibitors (BH3 mimetics), SMAC mimetic, and BRD4 inhibitor, in 6 AML PDX models (AUC of ROC 0.8731, p<0.005). Next, we created resistant PDX models to single agents including, quizartinib, birinapant, venetoclax, S63845 and JQ-1. After selecting for in vivo acquired resistance to drugs with distinct mechanism, a common mechanism of resistance was identified for all - a reduction in mitochondrial apoptotic priming. Apart from PDX models, paired pretreatment and post relapse myeloblasts of patients who had complete response followed by a relapse on venetoclax plus azacytidine therapy (NCT02203773) also showed selection for decreased mitochondrial apoptotic priming in relapsed myeloblasts using promiscuously interacting BIM (P=0.0075) and PUMA peptides (P=0.0078). Using targeted exome sequencing for recurrently mutated leukemia genes, we found that although there was acquisition of new mutations between paired clinical samples, there was no consistent molecular signature defining relapsed phenotype. We report that loss in apoptotic priming in BH3 mimetics resistant PDXs can be explained by alterations in BCL-2 family proteins levels and interaction patterns at outer mitochondrial membrane that vary among cases. However, there was absence of gene signature depicting reduction in pro-apoptotic genes and upregulation in anti-apoptotic genes, measured by unbiased RNA-seq (Bhatt et al. Cancer cell, In press). Enrichment for pro-survival pathways, including JAK-STAT, MAPK, and PI3K-AKT was observed using RNA-seq of resistant PDXs compared to matched parental counterpart in BH3-mimetics and FLT-3 inhibitor resistant models but not SMAC-mimetics and BRD-4 resistant models. To identify the agents that are effective in the resistant settings, we compared DBP profiles of 40 targeted agents in myeloblasts of pre and post resistant models. We found that reduction in overall priming led to broad chemoresistance even to mechanistically distinct agents in resistant myeloblasts, yet there was maintenance of persistent sensitivity to selected agents. For instance, in venetoclax and S63845 resistant PDXs, mitochondrial priming measured by DBP on resistant myeloblasts identified in vivo activity of FLT-3 inhibitors and SMAC mimetics while in quizartinib resistant settings, SMAC mimetics, BH3 mimetics and MAPK inhibitors showed anti-leukemic effects. This suggested that while common modes of mechanism do exist, drugs that enhance mitochondrial apoptotic sensitivity can overcome resistance to a particular agents. Finally we applied this approach to humans, showing that the pretreatment mitochondrial apoptotic priming determined by DBP identifies responders to single agent FLT-3 inhibitor gilteritinib in ADMIRAL trial and responders to lenalidomide (LEN) and MEC combination therapy in LEN-MEC Phase I trial (NCT01442714) R/R AML (Garcia* and Bhatt* et al. American Journal of hematology, 2020). In summary, our results suggest that acquired resistance to targeted therapy in AML is accompanied by common mechanism of reduction in mitochondrial priming along with drug-specific resistance mechanisms. Hence measurements of apoptotic priming using dynamic BH3 profiling may serve as a broadly applicable precision medicine tool in guiding therapy for relapsed leukemia.

Citation Format: Shruti Bhatt, Marissa S. Piosos, Elyse A. Olesinski, Binyam G. Yilma, Jeremy A. Ryan, Thelma Mashaka, Buon Leutz, Sophia Adamia, David M. Weinstock, Jacqueline S. Garcia, Anthony Letai. Reduction in mitochondrial priming drives resistance to targeted therapy in acute myeloid leukemia [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr NG14.

Reduced Mitochondrial Apoptotic Priming Drives Resistance to BH3 Mimetics in Acute Myeloid Leukemia

Acquired resistance to BH3 mimetic antagonists of BCL-2 and MCL-1 is an important clinical problem. Using acute myelogenous leukemia (AML) patient-derived xenograft (PDX) models of acquired resistance to BCL-2 (venetoclax) and MCL-1 (S63845) antagonists, we identify common principles of resistance and persistent vulnerabilities to overcome resistance. BH3 mimetic resistance is characterized by decreased mitochondrial apoptotic priming as measured by BH3 profiling, both in PDX models and human clinical samples, due to alterations in BCL-2 family proteins that vary among cases, but not to acquired mutations in leukemia genes. BCL-2 inhibition drives sequestered pro-apoptotic proteins to MCL-1 and vice versa, explaining why in vivo combinations of BCL-2 and MCL-1 antagonists are more effective when concurrent rather than sequential. Finally, drug-induced mitochondrial priming measured by dynamic BH3 profiling (DBP) identifies drugs that are persistently active in BH3 mimetic-resistant myeloblasts, including FLT-3 inhibitors and SMAC mimetics.

Abstract A237: The transcriptional signature of kinases inhibited by the multitargeted kinase inhibitor AS703569 is associated with inferior clinical outcome in multiple myeloma (MM), with promising anti-MM activity of AS703569 demonstrated in preclinical studies in vitro and in vivo

Multi-targeted kinase inhibitors, when associated with manageable toxicity, offer the therapeutically desirable option of targeting, through a single chemical entity, several pathways contributing to the molecular complexity and heterogeneity of neoplasias, such as multiple myeloma (MM). However, when multi-targeted agents have extensive and/or only partially overlapping known targets, it is difficult to prioritize them for further development. We hypothesized that the potential therapeutic relevance of a multi-targeted inhibitor may be reflected on the prognostic relevance of its targets' transcriptional signature. We applied this concept in the case of the orally bioavailable multi-targeted kinase inhibitor AS703569, which targets (IC50 in low nM range) all 3 Aurora kinase isoforms and diverse other kinases (e.g. cSRC, FGFR1, Flt3, Fyn, Lyn, Rsk1-3, Axl, et.c.) and evaluated the transcriptional signature of AS703569 kinase targets (with IC50 <10 nM) in MM cells of patients receiving Bortezomib as part of Phase II/III trials. We observed that patients with high transcriptional signature of AS703569 targets had inferior progression-free and overall survival (p=0.005 and p=0.012, log-rank test) and also validated that, in a study of tandem autologous transplant, a subset of patients with high levels of this AS703569 target transcriptional signature also have inferior overall survival (p=0.032, log-rank test) compared to cases with low levels of the signature. These observations indicate that the kinome space targeted by AS703569 is enriched for targets associated with adverse clinical outcome in MM. In preclinical assays, we observed that AS703569 decreased the viability of MM cell lines and primary CD138+ MM tumor cells in time- and dose-dependent manner (IC50 values <50 nM for the majority of cell lines tested), without evidence of cross-resistance with established anti-MM agents. Combinations of AS703569 with dexamethasone, doxorubicin, or bortezomib did not exhibit antagonism, suggesting that AS703569 can be incorporated in regimens with these established anti-MM drug classes. Interestingly, in vitro compartment-specific bioluminescence imaging (CS-BLI) assays showed that against MM cells which respond to stromal cells with increased proliferation and survival, the anti-MM activity of AS703569 is more pronounced when these MM cells are co-cultured with bone marrow stromal cells than in conventional cultures in isolation. This in vitro result that AS703569 can overcome protective effects of BSMCs on MM tumor cells was validated in vivo in orthotopic SCID/NOD model of diffuse MM bone lesions established by i.v. injection of MM-1S-GFP/Luc cells monitored by whole body bioluminescence imaging. AS703569 (50 mg/kg p.o. once weekly)-treated mice had longer overall survival than vehicle-treated mice (median 50.0 vs. 39.0 days, p=0.019, log-rank test). An alternative schedule of AS703569 at 16.7 mg/kg 3 times/week also resulted in longer overall survival (median 54.0 days, p=0.023). These data suggest that AS703569 merits further studies towards potential clinical trials in MM, especially given the adverse outcome associated with the molecular signature of its targets.

Citation Information: Mol Cancer Ther 2009;8(12 Suppl):A237.