Exploiting the Metabolic Vulnerability of Leukemic Stem Cells (LSCs) in Inv(16) Acute Myeloid Leukemia (AML) with the Combination of Miristen, a Mir-126 Inhibitor, and Venetoclax

AML with inv(16)(p13q22) expresses the leukemogenic fusion gene CBFB-MYH11 and upregulates microRNA (miR)-126 that supports homeostasis, quiescence, and activity of LSCs. Genetic or pharmacologic depletion of miR-126 via miR-126 knock out (KO) or miRisten (a CpG-anti-miR-126 inhibitor) reduces LSCs in a Cbfb-MYH11 (CM) knock-in (KI) AML model and inv(16) AML patient-derived xenografts (PDXs), respectively [PMID: 34686664]. However, the antileukemic underpinnings of miR-126 inhibition in inv(16) AML remain to be elucidated.

LSCs reportedly rely on fatty acid metabolism/oxidative phosphorylation (OXPHOS) for their bioenergetic metabolic needs. To assess the role of miR-126 in metabolism, we performed metabolomic profiling for Lin-/c-Kit+ cells isolated from CM KI AML (miR-126^high) mice and CM with miR-126 KO mice (CM/miR-126^Δ/Δ). Our data revealed that CM/miR-126^Δ/Δ had a significant decrease of short, medium as well as long-chain acyl carnitines and fatty acid (2 to 11 fold, p<0.05) compared to CM/miR-126^WT/WT mice, suggesting that miR-126 deficit decreased fatty acid oxidation (FAO). Moreover, electron microscope imaging demonstrated mitochondrial fusion (i.e., larger and networked mitochondria) in CM Lin-/c-Kit+Sca1+ (LSK) cells vs syngeneic controls. In contrast, mitochondria fragmentation (fission) and decreased FAO/OXPHOS were observed in murine CM with miR-126 KO LSK. Accordingly, FAO and Seahorse assays demonstrated miRisten treatment led to mitochondria fragmentation and reduced FAO/OXPHOS and in human CD34⁺ CD38^‒ inv(16) blasts (enriched for LSCs) vs normal CD34⁺ CD38^‒ cells.

Sprouty Related EVH1 Domain Containing 1 (SPRED1), a downstream target of miR-126, inhibits ERK/CREB activation and in turn increases antiapoptotic protein BCL-2 via phosphorylation and stabilization. In inv(16) AML cells, upregulated miR-126 downregulated SPRED1 and in turn increased ERK/CREB signaling and BCL-2 levels. This increase augments the ubiquitination of DRP1, a protein otherwise involved in mitochondrial fission, via NFE2-related factor 2 (NRF2), and increases Carnitine Palmitoyltransferase 1B (CPT1B), a rate-controlling protein for FAO and OXPHOS. In contrast, miR-126 KO or miR-126 inhibition by miRisten led to SPRED1 upregulation and decreased ERK/CREB signaling and BCL-2 levels, thereby resulting in increased Dynamin-related protein 1 (DRP1) and mitochondrial fission and a decrease in CPT1B, FAO, and OXPHOS. Thus, given the interplay of miR-126 with BCL-2 and its implication in mitochondrial fusion and increase of OXPHOS in inv(16) blasts and LSCs, we hypothesized a synergism between miRisten and the BCL-2 inhibitor venetoclax (VEN). In fact, combination of miRisten/VEN synergistically decreased FAO/OXPHOS, enhanced mitochondria fission and increased apoptosis of murine and human inv(16) blasts and LSCs [combination indices (CIs)<1, (ED50-ED95)]. Thus, we treated CM mice with scrambled oligonucleotides (sODN) (20 mg/kg i.v., daily for 3 weeks), miRisten (20 mg/kg i.v., daily for 3 weeks), VEN (100 mg/kg oral gavage, daily for 2 weeks), or miRisten/VEN (same dosages as single agents). Compared with sODN or single agents, miRisten/VEN significantly reduced bone marrow (BM) leukemic blast infiltration [miRisten/VEN 33.6±3.4% vs miRisten 49.8±2.6% (p=0.001), VEN 56.2±4.7% (p=0.001), or sODN 60.0±2.6% (p<0.0001)] and spleen weights [miRisten/VEN 0.228±0.038g vs miRisten 0.495±0.058g (p=0.001), VEN 0.479±0.06g (p=0.002), or sODN 0.853±0.039g (p<0.0001)] and prolonged survival of CM mice [median survival miRisten/VEN 125d vs miRisten 108d (p=0.005), VEN 106.5d (p=0.039), or sODN 91d (p<0.0001), respectively; secondary transplant median survival: miRisten/VEN 92d vs miRisten 70.5d (p=0.048), VEN 54d (p=0.006), or sODN 45d (p=0.0002)]. Similar results were also observed in inv(16) PDXs with significant reduction of BM huCD45+ leukemic blasts by miRisten/VEN compared to single agents or sODN [miRisten/VEN 13.7±1.9% vs miRisten 27.9±3.4% (p=0.003), VEN 25.3±4.6%(p=0.032), or sODN 68.6±5.3% (p<0.0001)]; survival analyses for treated PDX mice and secondary transplant recipients are in progress.

In summary, our study provides the mechanism for combining miRisten with VEN in the treatment of inv(16) AML. While VEN is FDA approved for AML, miRisten has just undergone a successful toxicology evaluation and is underway to the clinic.