Dual Targeting of Mitochondrial Vulnerability Using Complex I Inhibitor Iacs-010759 with Bcl-2 Inhibitor Venetoclax and Azacitidine in Pre-Clinical Acute Myeloid Leukemia (AML) Models

Despite recent approval of hypomethylating agent/venetoclax (HMA+VEN) therapy for older patients (pt) with AML unfit for induction chemotherapy, their outcomes remain suboptimal. Such pts have a median overall survival of only 14 months and only approximately 35% of pts enjoying long-term survival (DiNardo, EHA 2020). Approximately 20-40% of newly diagnosed older pts with AML do not respond to the HMA+VEN regimens, with higher rates of refractory disease as well as early relapse in high-risk pts. Metabolic reprogramming and dependence on mitochondrial oxidative phosphorylation (OxPhos) is a core feature of AML leukemia stem cells (LSC). Recent reports have shown that upregulation of OxPhos confers intrinsic and acquired resistance to VEN in AML, multiple myeloma and lymphoid malignancies, which can be reversed by disrupting OxPhos; mutated TP53 has been shown to confer intrinsic resistance to venetoclax through increased OxPhos (Nechiporuk T, et al. Cancer Discovery 2019). We hypothesized that combined blockade of mitochondrial fitness by OxPhos inhibitors and of BCL-2 with VEN/azacitidine (AZA) will perform synergistically in pre-clinical AML models. To test this hypothesis, we utilized a novel complex I inhibitor IACS-010759 that effectively inhibits cell respiration and leukemia progression in the in vitro and in vivo AML pre-clinical models (Molina, Nat Med 2018).

Priming of MOLM-13 cells with 20nM VEN for 24hrs followed by 10nM IACS-010759 for 1hr triggered 60% reduction in oxygen consumption rate (OCR), while single agents reduced OCR by <15%. This translated into accelerated loss of the mitochondrial membrane potential measured by JC-1 flow cytometry, cleavage of caspase 3 and synergistic reduction (>70%) of viable cell numbers in several AML cell lines tested (OCI-AML2, MV-4-11, and MOLM-13) (Figure 1). By co-immunoprecipitation VEN disrupted interaction of BCL-2 with the mitochondrial protein VDAC, known to regulate ADP/ATP exchange during electron transport across mitochondria membrane; this resulted in dramatic reduction of the intracellular ATP and CTP levels measured by MS-based metabolomics. Further, VEN increased intracellular levels of AMP, UMP, CMP and GMP and this accumulation of mono-nucleotides was enhanced by the combination of VEN and IACS-010759, possibly because of RNA degradation. In primary AML samples (n=3) and AML PDX cells (n=4) cultured ex vivo, combined VEN and IACS-010759 at low nanomolar doses reduced viable cell numbers in an additive or synergistic fashion.

We next tested the efficacy of the “triple” combination of VEN, AZA and IACS-010759 in the in vivo AML PDX models. We injected AML PDX cells 3871344 (with no mutations identified by targeted sequencing) and 4404778 (harboring IDH1, NPM1, FLT3-ITD mutations) into NRG mice and upon engraftment, randomized mice into 4 groups to receive 2 cycles of treatment with 3 weeks interruption between cycles: vehicle, VEN (50mg/kg daily, 5 days on/2 days off, day 1-21) with AZA (2.5 mg/kg daily, day 1-7), IACS-010759 (1mg/kg ?daily, 5 days on/2 days off, day 1-14), or the triple combination. Therapy was well tolerated, without any apparent weight loss or toxicities. In the less aggressive model 3871344, all therapies reduced circulating leukemia burden, and the triple treatment achieved best efficacy, with average circulating tumor burden at 10 weeks after cycle 2 of 90.8%, 38.7%, 68.8% and 7.4% in vehicle, VEN+AZA, IACS-010759, and triple-therapy cohorts, respectively. In the aggressive model 4404778, the treatments were less effective, but the combination offered highest activity, with average circulating tumor burden of 71.0%, 52.3%, 88.7% and 39.9% in vehicle, IACS-010759, VEN+AZA and triple-therapy cohorts, respectively (Figure 1). Analysis of survival and additional PDX models are ongoing and will be reported.

In summary, our study demonstrates that BCL-2 modulates mitochondrial respiration and mitochondrial ATP generation in addition to its established anti-apoptotic role. VEN disrupts the BCL-2/VDAC interactions and reduces mitochondrial respiration, which is facilitated by the combined therapy with mitochondrial complex I inhibitor. Our preliminary findings indicate potent anti-AML activity of the dual and triple (with hypomethylating agent) combinations in vitro and in vivo.