Selective targeting of NAMPT by KPT-9274 in acute myeloid leukemia

Acute myeloid leukemia (AML) is the most commonly diagnosed acute leukemia that disproportionately affects the elderly.^1,2  Although a small subset of patients with AML can be cured with aggressive chemotherapy and/or allogeneic stem cell transplantation, the majority of patients still die of their disease.³  Despite the poor outcome, little progress has been made outside of allogeneic stem cell transplantation. Indeed, only 2 targeted therapies directed at FMS-like tyrosine kinase 3 (FLT3) mutated or isocitrate dehydrogenase 2 and isocitrate dehydrogenase 1 mutated AML have been approved for this disease by the US Food and Drug Administration.^4-6  Multiple cytotoxic, epigenetic, targeted, and immune-based treatments have reached phase 2 and 3 trials in AML without showing significant clinical benefit,^2,7,8  attesting to the need for identifying both novel targets and therapeutic agents directed toward them.

A successful example of an effective targeted therapy comes from chronic lymphocytic leukemia, in which a wide variety of cytogenetics and mutations exists without a common targetable pathway. The identification of the importance of B-cell receptor signaling across all patients ultimately led to the development of agents such as ibrutinib and idelalisib, which have significantly altered the natural history of this disease.^9,10 

In AML, survival pathways seem to exist, including altered cellular metabolism. AML cells reportedly exhibit higher glycolytic activity and more dependence on functional mitochondrial activity across different genotypes compared with normal hematopoietic counterparts.^11-14  We hypothesized that the development of targeted therapies capable of directly antagonizing cellular metabolism and mitochondrial function could have broad activity across many AML subtypes.

Nicotinamide phosphoribosyltransferase (NAMPT) is the rate-limiting enzyme involved in the conversion of nicotinamide into nicotinamide monophosphate, which then yields to NAD⁺ via the NAMPT-dependent salvage pathway.^15,16  NAD⁺ is a metabolite involved in the maintenance of the mitochondrial membrane potential and cellular signaling. Studies suggest that select tumor types are “addicted” to the NAMPT-dependent salvage pathway due to the downregulation of alternative NAD⁺ production pathways and are therefore more sensitive to NAMPT inhibition.^17,18  Several NAD⁺ consumer proteins, such as CD38, poly (ADP-ribose) polymerase, and sirtuins, have been shown to manage DNA repair mechanisms and mediate cancer disease progression by protecting cells during nutrient-deficient events.^19-24  In the absence of NAD⁺, both classes of proteins lose their cytotoxic protective features, making NAD⁺ reduction a potential target for cancer therapeutic agents.

Overexpression of or increased dependency on NAMPT has been observed in several cancers, including AML.^25-31  In addition, in patients with AML, higher expression of NAMPT has been correlated to a shorter overall survival.³²  Targeting this pathway therefore provides a meaningful strategy for treating AML. The present article describes the structurally novel dual NAMPT/p21-activated kinase 4 (PAK4) inhibitor KPT-9274; we show that inhibition of NAMPT (rather than PAK4) leads to therapeutic benefit in vitro and in vivo in multiple preclinical models of AML. Oral KPT-9274 is currently in clinical trials for the treatment of patients with advanced solid malignancies (#NCT02702492). Our findings provide justification for exploration of KPT-9274 in AML clinical trials.

Cell lines EOL-1 HL-60, HS-5, Kasumi-1, and THP-1 were purchased from ATCC (Manassas, VA). Cell lines K562, MV4-11, and OCI-AML3 were purchased from DSMZ (Braunschweig, Germany). Cell lines were sequenced to confirm reported mutations by using a published 80 gene panel (Table 1).³³  Cell lines were cultured in recommended media conditions from vendors with the addition of 10 000 U of penicillin, 10 mg of streptomycin, and 200 mM of glutamate. AML patient and normal donor samples were obtained from The Ohio State University (OSU) Leukemia Tissue Bank under an institutional review board–approved protocol with informed consent according to the Declaration of Helsinki. AML primary cells, normal donor sample cells, umbilical cord hematopoietic stem cells, and cell lines were cultured in RPMI 1640 media supplemented with 20% fetal bovine serum, 10 000 U of penicillin, 10 mg of streptomycin, and 200 mM of glutamate. Primary AML cells and normal umbilical cord hematopoietic stem cells were additionally supplemented with 10 ng/mL of FLT3 ligand, interleukin-3, stem cell factor, and granulocyte-macrophage colony-stimulating factor (PeproTech, Rocky Hill, NJ). All cells were kept in a 37°C, 5% carbon dioxide incubator.

AML cell lines or patient-derived cells were seeded in a 96-well plate at 1 × 10⁵ cells per well for 24 and 48 hours. Cells were treated with increasing concentrations of KPT-9274 or KPT-9331 (1 nM to 10 µM) or vehicle control (dimethyl sulfoxide [DMSO]). For combination studies, increasing concentrations of venetoclax (1 nM to 10 µM) were used in combination with increasing concentrations of KPT-9274 (1 nM-10 µM). Cells used in NAD⁺ rescue experiments were treated with media supplemented with 100 µM of β-nicotinamide dinucleotide (MilliporeSigma, Burlington, MA). MTS [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium] (Promega, Madison, WI) was prepared according to the manufacturer’s instructions. Plates were read by using a DTX plate reader (LabSystems MultiSkan MCC1340; Thermo Fisher Scientific) at 490 nM.

AML cell lines or isolated peripheral blood mononuclear cells were plated at 5 × 10⁵ cells/mL and were treated with 2.5 µM or 250 nM of KPT-9274 or vehicle control (DMSO). For NAD⁺ rescue experiments, cells were incubated in media supplemented with 100 µM of β-nicotinamide dinucleotide (BD Biosciences). Cells were washed and stained by using mouse anti-human annexin V–fluorescein isothiocyanate and propidium iodide (BD Pharmingin). Analysis was performed by using a Cytomics FC 500 Flow Cytometer (Beckman Coulter, Brea, CA), and data analysis was performed with Kaluza Analysis software (Beckman Coulter).

Immunoblotting was performed and quantified by using published methods.³⁴  Blots were probed with primary antibodies (listed in the supplemental Methods) and horseradish peroxidase–conjugated anti-mouse or anti-rabbit secondary antibodies.

AML patient cells, patient-derived xenograft (PDX) bone marrow cells, or Lin^–/CD45⁺/CD34⁺ cells were treated with KPT-9274 or DMSO vehicle control. Cells were counted and plated in ∼10 000 to 1000 cells per well with 3 wells per condition in 6-well culture plates. Colonies were counted after 14 days of incubation in MethoCult H4034 Optimum media (Stemcell Technologies, Vancouver, BC, Canada). For the replating experiments, cells from each of the previously treated conditions were collected from the MethoCult medium, washed, counted, and replated at the same densities without further treatment. Colonies were counted after an additional 14 days.

Cytotoxic studies determining lymphocyte survival were performed in vitro by treating whole blood with 1 μM of KPT-9274 for 48 hours at 37°C. Cytotoxicity of normal lymphocytes was assessed as previously described.³⁵ 

Seahorse analysis was performed by using a Seahorse XF24 analyzer (Agilent Biosciences, Santa Clara, CA). After AML cells were treated with KPT-9274 or vehicle for 24 hours or AML-inducible cell lines with doxycycline for 120 hours, cells were resuspended in serum- and phenol red–free medium (Agilent Technologies) for both oxygen consumption rate and extracellular acidification rate analysis according to the manufacturer’s protocol. Cells were plated in a Cell-Tak (Corning Incorporated, Corning, NY) coated 24-well XF24 plate and incubated for 1 hour in a carbon dioxide–free incubator at 37°C. For oxygen consumption rate analysis, cells were treated with 1 µg/mL of oligomycin, 1 µM of carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone, and 1 µM of rotenone (all, MilliporeSigma). For extracellular acidification rate analysis, cells were treated with 10 mM of d-glucose (MilliporeSigma), 1 µg/mL of oligomycin, and 50 mM of 2-deoxyglucose (MilliporeSigma).

Guide RNAs were designed by using the MIT Clustered Regularly Interspaced Short Palindromic Repeat (CRISPR) software (crispr.mit.edu). Details on the cells generated are given in the supplemental Methods.

All HF/6-31G(d) KPT-9274 optimizations were performed by using the Psi4 program.³⁶  Subsequent docking studies to the NAMPT protein (2GVJ³⁷ ) were conducted by using Autodock 4.2 and the associated Autodock Tools package.³⁸  The UCSF Chimera software was used to visualize the resulting complex.³⁹  Additional information on docking is provided in the supplemental Methods.

Nonobese diabetic/severe combined immunodeficiency mice with an interleukin-2 receptor gamma chain mutation (NSG) were purchased from The Jackson Laboratory (Bar Harbor, ME). Mice were 4 to 6 weeks old at experiment initiation. Experiments were approved by The OSU Institutional Animal Care and Use Committee. MV4-11 expressing luciferase cells obtained from spleens of primary MV4-11/luciferase xenografts were reengrafted into NSG mice via tail vein injection (50 000 cells per mouse), and treatment was begun 1 week after engraftment. Leukemia cells obtained from the bone marrow of a primary engraftment of a PDX mouse were engrafted into NSG mice via tail vein injection (500 000 cells per mouse), and treatment started when 5% of human CD45⁺/human CD33⁺ cells were present in the blood. After meeting enrollment criterion, mice were randomized to receive 150 mg/kg of KPT-9274 or placebo drug (vehicle) via oral gavage, once daily. More information about the disease burden analysis of the treated mice is detailed in the supplemental Methods.

A detailed description of the statistical analysis is presented in the supplemental Methods.

We first examined the activity of KPT-9274, the clinical compound, and KPT-9331, a tool compound with the same reported mechanism of action, in 6 AML cell lines (HL-60, Kasumi-1, MOLM13, MV4-11, OCI-AML3, and THP-1). Treatment of KPT-9274 and KPT-9331 resulted in a dose-dependent decrease in metabolic activity, as reflected by a more direct measurement of NADP(H) via MTS, with 50% inhibitory concentration (IC₅₀) values that ranged from 27 to 215 nM and 14 to 129 nM at 48 hours for KPT-9274 and KPT-9331, respectively (Table 1). Despite diminished metabolic activity after compound treatment at 24 hours, cells had no increased apoptosis as determined by annexin V/propidium iodide flow cytometric analysis (Figure 1A). However, at later times (48-72 hours), significant dose- and time-dependent apoptosis was observed. This decrease in viability corresponds to late processing of caspase 9 to its active form, caspase 3, and poly (ADP-ribose) polymerase, a known caspase substrate, in these AML cell lines (Figure 1B). We also detected no change in proliferation at 24 hours, with proliferation inhibition in compound-treated cells only at the 48- and 72-hour time points (supplemental Figure 1A). KPT-9274 treatment in AML patient samples yielded similar results. KPT-9274 response occurred irrespective of differences in mutations, with only modest variation (Table 2). AML cells cocultured with or without HS-5, a human stromal cell line used to model the in vivo bone marrow stromal cell protection, were treated with these small molecules for 96 hours (supplemental Figure 1B). KPT-9331 was able to elicit antitumor activity even in the presence of microenvironment protection⁴⁰  compared with the cytokines alone (average IC₅₀ = 130 nM vs IC₅₀ = 133 nM, respectively). Loss of metabolic activity as determined by MTS was seen in all samples.

We next assessed the ability of the clinical candidate, KPT-9274, to affect in vitro colony formation. Treatment of AML patient samples of various cytogenetic and mutational status uniformly showed a decrease in colony formation after 14 days of culture (Figure 1C; supplemental Figure 2A). Patient samples were replated after a 14-day culture. We observed further diminished colony-forming unit formation in some patient samples in the second plating (Figure 1D; supplemental Figure 2B), suggesting the ability of KPT-9274 to eradicate leukemic clonal cells. Collectively, these findings show the preclinical activity of KPT-9274 and its corresponding tool molecule against AML primary cells and cell lines.

We next determined the effects of KPT-9274 on normal hematopoietic cells in vitro. Treatment of KPT-9274 at varying concentrations for 72 hours induced some toxicity in normal donor peripheral blood mononuclear cells but not to the extent observed in the AML cell lines (Figure 2A). As shown in Figure 2B and supplemental Figure 2C, a minimal decrease in colony formation was observed after treatment of Lin^–/CD45⁺/CD34⁺ normal hematopoietic progenitors, suggesting that this inhibitor has preferential toxicity toward leukemic cells while sparing their normal counterparts. To investigate if KPT-9724 affects more mature leukocytes, we performed a whole-blood viability assay standardized by our group.³⁵  As shown in Figure 2C, KPT-9274 caused minimal decreases in cell concentrations of B, T, and natural killer cells (P > .99, P = .97, and P > .99, respectively). In addition, we showed that KPT-9331 only inhibits CD4⁺ T-cell proliferation after CD3 ligation at 1 log-fold above the IC₅₀ at which it is determined to be cytotoxic to AML cells (supplemental Figure 2D). Collectively, this outcome suggests that a significant therapeutic index exists with KPT-9274 between normal hematopoietic stem cells and different immune cells vs AML cells.

Given our biochemical findings, the structural features of KPT-9274, and reports suggesting NAMPT as a target of KPT-9274 in other systems,^41-44  we tested whether KPT-9274 acts through a pathway involving NAMPT inhibition in AML cells. Protein expression of NAMPT was variable in a panel of malignant myeloid cell lines and primary AML cells, as shown in Figure 3A. Similarly, the transcript was uniformly expressed at the transcriptional level in different subtypes of AML using publicly available data from The Cancer Genome Atlas^45,46  (Figure 3B).

Given that NAMPT inhibition rapidly decreases mitochondrial activity via NAD⁺ depletion,^16,47  we sought to determine if the decrease in mitochondrial activity after either compound treatment is due to NAD⁺ depletion. Treatment with KPT-9274 resulted in a dose-dependent decrease in NAD⁺ levels in the cells after 24 hours (Figure 3C). Addition of 100 µM of exogenous NAD⁺ rescued cells from apoptosis (Figure 3D). As a proof of concept, we generated THP-1 cell lines expressing CRISPR-associated protein 9 (Cas9) and doxycycline-inducible single guide RNA for NAMPT. As shown in Figure 3E, knockdown of NAMPT decreased the overall metabolic activity of THP-1 cells compared with wild type. Given that PAK4 is expressed in several AML cells (supplemental Figure 3A) and is a target of KPT-9274,^41-44  we next sought to determine the effects of underexpression or overexpression of PAK4 in AML proliferation. PAK4 overexpression did not enhance the proliferative abilities of the AML cell lines MV4-11 or PAK4-negative HL-60 cell lines (supplemental Figure 3B-C). PAK4 knockouts in PAK4-expressing THP-1 cells did not modulate the sensitivity of the cells to KPT-9274 treatment (supplemental Figure 3D-E). Conversely, knockdown of PAK4 expression using a doxycycline-inducible CRISPR-Cas9 system did not show a decrease in proliferation in this same cell line (supplemental Figure 3F). Overall, these findings support the potential cytotoxicity of KPT-9274 in AML and suggest that targeting NAMPT, rather than PAK4, influences AML cell survival.

To examine how KPT-9274 potentially binds to NAMPT vs known NAMPT inhibitors, we computationally modeled the binding and molecular interactions of this complex. The high-resolution (2.1Å) NAMPT crystal structure 2GVJ³⁷  was chosen as the starting structure for the in silico analysis due to the structural similarities between KPT-9274 and its co-crystal ligand FK866, a previously developed and well-characterized NAMPT inhibitor.

Given the large size and complexity of KPT-9274, we created a broad range of geometry-optimized conformations to diversify the input structures during docking and chose 25 to represent the range of starting energies and geometries. There was a consistent mode of binding irrespective of the starting conformer. The KPT-9274 molecule lies along the same narrow tunnel extending from the active site as seen for FK866 (Figure 4). This tunnel is composed of parallel beta sheets, and the binding of KPT-9274 takes advantage of some of the same key residues that FK866 used in its binding. As such, this molecule competitively antagonizes the binding site of nicotinamide. A key difference appears at the interface of the homodimer, where chain NAMPT dimer A makes contact with NAMPT dimer B. The larger fluorinated piperidine ring of KPT-9274 protrudes farther out of the tunnel and causes steric hindrance specifically on tyrosine 18 from dimer B. This occurrence suggests that KPT-9274 disrupts nicotinamide substrate binding but would also impair dimerization of NAMPT or cause quaternary conformational changes that could influence function.

A decrease in global NAD⁺ levels after compound treatment may lead to a decrease in overall cellular respiration because of the importance of this cofactor in energy generation.²⁴  To determine if KPT-9274 can decrease mitochondrial respiration in AML cells when NAD⁺ levels are reduced, oxygen consumption rates of AML cell lines after 24 hours of treatment with KPT-9274 were measured. KPT-9274 was able to compromise the overall oxygen consumption rate during the Mito stress test, which was evident in the ability of KPT-9274 to impair recovery of maximal respiration and therefore reduce mitochondrial spare reserve capacity (Figure 5A; supplemental Figure 4A-B). This outcome further suggests the ability of KPT-9274 to decrease mitochondrial activity. We also observed a decrease in mitochondrial membrane potential measured by JC-1 staining after KPT-9331 treatment (supplemental Figure 4C-D). The observation of early diminished metabolic activity, via MTS, with a delay in apoptosis induction could be explained by this observed phenomenon. NAD⁺ also serves as a cofactor in non-mitochondrial respiration. We measured glycolytic function in AML cells treated with KPT-9274 and observed overall reduced glycolysis (Figure 5B). Confirming that this phenomenon occurs in genetically NAMPT-depleted cells, mitochondrial and glycolytic capacity was measured in THP-1–inducible NAMPT knockdown cell lines; we observed a similar pattern of deficient oxidative phosphorylation and glycolytic capacity in doxycycline-treated cell lines vs untreated cells (supplemental Figure 4E).

It has been reported that B-cell lymphoma 2 (BCL-2) inhibitors are able to target cells that are oxygen deprived or have reduced glycolytic capacity.^12,48  We therefore tested the ability of KPT-9274 to induce sensitivity of AML cells to BCL-2 inhibition. In a stromal coculture system, venetoclax-treated AML patient samples were rendered cytotoxic after treatment with increasing concentrations of KPT-9274 (Figure 5C). Using the Combenefit software (version 2.021),⁴⁹  the combination of KPT-9274 was able to show significant synergism at the 100-nm concentration of KPT-9274 and the 10-nM concentration of venetoclax (Figure 5D). Overall, KPT-9274 blocks processes that are important for NAD⁺ production and cellular respiration, which may lead to cell death.

To assess the effect of KPT-9274 in an in vivo disseminated model of AML, we engrafted luciferase-positive MV4-11 cells by tail vein into NOD/SCID IL2rγ^−/− (NSG) mice. One week after engraftment, mice were dosed once daily via oral gavage with 150 mg/kg of KPT-9274 (n = 7) or vehicle control (n = 6) (Figure 6A). KPT-9274 resulted in improved overall survival (median survival of 82 days vs 30 days; P = .0005) (Figure 6B). At the study’s end point, we assessed disease burden in different organs. All mice who met euthanasia removal criteria were removed due to hind limb paralysis. Bone marrow from KPT-9274–treated mice harbored significantly less human CD45⁺/CD33⁺ cells vs vehicle control (P = .0025) (Figure 6C). In addition, KPT-9274–treated marrow differentials showed increased differentiation with fewer blasts and larger numbers of late-stage myeloid cells vs vehicle-treated mice (Figure 6D, top panels). Decreased tumor infiltration was observed in the spleens of inhibitor-treated mice (Figure 6D, bottom panel). The death of drug-treated mice is likely reflective of sole central nervous system disease for which KPT-9274 lacks little compartmental penetration.

A second cohort of MV4-11–engrafted mice were observed for disease progression by using the IVIS Lumina Series III bioluminescence imaging system (PerkinElmer) over the course of 16 days before the mice died of the disease (Figure 6E). KPT-9274–treated mice exhibited delayed leukemic progression marked by a reduced presence of luciferase-positive MV4-11 cells in imaged mice. In this cohort of mice, immunohistochemical analysis confirmed a reduced tumor burden in the bone marrow of KPT-9274–treated mice (Figure 6F). Similar results were obtained by using mice transplanted with MOLM-13, which represents a more aggressive mouse model of AML with average disease onset at 23 days. KPT-9274 treatment significantly prolonged overall survival in these mice (supplemental Figure 5).

To expand this effort, a PDX model of AML was also used. We initially passaged and expanded a primary complex karyotype AML sample in NSG mice. In a secondary transplant, 500 000 bone marrow cells were engrafted into 15 recipient NSG mice. Mice were given either 150 mg/kg of KPT-9274 or the vehicle control once daily for 4 weeks (Figure 7A). Mutational status analysis of preengraftment and post-engraftment leukemic cells showed that mice maintained clonality similar to what was found in the patient, with the exception of an acquired KRAS mutation in the primary and secondary recipient mice (supplemental Table 1). Analysis of the bone marrow aspirates (Figure 7B) and blood of KPT-9274–treated mice (Figure 7C) revealed a significant decrease in human leukemia cells in both compartments. KPT-9274 decreased splenomegaly (Figure 7D-E) and tumor infiltration in the liver, spleen, and the bone marrow, with an observed increase in myeloid differentiation (Figure 7F). KPT-9274 also decreased the CD34⁺/CD38^– stem cell population of the leukemic cells (Figure 7G), a population that has been debated as important for the emergence of resistance in chemotherapy-treated patients. Bone marrow cells that were collected from the treated mice were also assessed for self-renewal capacity by conducting colony formation replating assays. After replating AML colonies, we observed a sharp decrease in the self-renewal capacity of KPT-9274–treated mice vs the vehicle control group (Figure 7H). Patient samples treated with KPT-9274 overall maintained a similar genetic profile of leukemic cells postsecondary engraftment and in the vehicle control group. Taken together, these findings suggest that KPT-9274 can reduce disease progression and prolong survival in a preclinical setting.

These data show that the inhibition of NAMPT using oral KPT-9274 leads to a dose- and time-dependent decrease in mitochondrial activity and an increase in apoptosis. Importantly, we observed no significant difference in colony formation between normal CD34⁺ cells with or without KPT-9274 treatment, unlike what is observed in AML cell lines and patient cells. Moreover, KPT-9274 diminished the self-renewal capacity of AML cells in vitro after replating, demonstrating a continual decrease in colony formation after the treatment was discontinued. Based on earlier reports, we hypothesize that this selectivity of KPT-9274 is due to the energy needs of AML tumor cells. Targeting leukemic cell mitochondrial activity serves as a potential mechanism for determining a therapeutic index. This is due to the leukemic cell property of having a reduced reserve capacity compared with normal cells.¹¹  More specific differences in IC₅₀ values were observed between cell lines and patient samples. The variability of cytotoxicity in different cell lines is complex and likely due to multiple factors that may or may not reflect the biology of de novo AML. Further investigation of this topic is being conducted in our laboratory to determine which patient populations may benefit more from therapy.

Extending to mechanistic studies, we showed that KPT-9274 diminishes NAD⁺ levels concomitant with impaired mitochondrial and non-mitochondrial respiration. In addition, later-onset apoptosis is reversed with NAD⁺ supplementation, suggesting that AML cell survival is sensitive to NAD⁺ levels reduced by KPT-9274 treatment. Previously, KPT-9274 has also been reported to target PAK4.^41-44  To validate the cellular target of KPT-9274 in AML, we found that the knockdown of NAMPT resulted in diminished proliferation in AML cells and resistance to KPT-9274 inhibition. In contrast, knockdown of PAK4 protein expression did not affect proliferation/viability in AML cells. To better elucidate the interaction with NAMPT, we computationally investigated the binding of KPT-9274 with this enzyme; the result was a favorable binding interaction that is similar to known NAMPT–ligand crystalized complexes, while also being distinct due to its larger functional groups. This scenario suggests a unique biologic profile. This difference may play a role in various toxicity profiles compared with other NAMPT inhibitors. Although KPT-9274 has effects that overlap with other inhibitors (eg, gastrointestinal toxicity), it also exhibits nonclinically observed cardiac and retinal toxicities. Corroborating our findings, NAMPT was recently identified as the main target of KPT-9274 by using a CRISPR-Cas9 mutagenesis scan.⁵⁰  This previous publication and our work described herein firmly establish NAMPT as a KPT-9274–relevant target in AML.

The present study is one of the first to describe NAMPT as a selective target in AML initiating cells in vitro and in vivo. In our in vivo PDX study, KPT-9274 caused a significant decrease in the immature CD34⁺/CD38^– leukemic cell population in the bone marrow of our PDX AML murine model. There is still a debate, however, that this population is important for self-renewal in AML cells.^12,51  We therefore used a serial replating experiment in bone marrow cells of previously treated PDX mice. The AML cells had a more dramatic diminished self-renewal capacity over time when obtained from mice treated with KPT-9274 vs those receiving vehicle control. Furthermore, studies have suggested mitochondrial activity as a target in AML due to the reduced reserve capacity of leukemic cells compared with normal cells.¹³ 

Finally, in multiple AML models, we showed that KPT-9274 has in vivo antitumor activity. These data provide a strong rationale for a KPT-9274 AML clinical trial. Two agents targeting NAMPT (FK866 and GMX1777) have been developed in which the dose-limiting toxicities (including thrombocytopenia and gastrointestinal events) have led to discontinuation of clinical development.^52,53  These previous NAMPT inhibitors were given by IV infusion and are chemically distinct from KPT-9274. KPT-9274 demonstrates consistent pharmacokinetic properties (eg, oral absorption) across nonclinical species, exhibits minimal brain penetration, and does not inhibit any of the other PAK or cytochrome P450 enzymes. In addition, the nonclinical toxicity profile of KPT-9274 recapitulates the expected class level gastrointestinal toxicity at toxic doses but not the retinal and cardiac effects observed nonclinically with other NAMPT inhibitors. When treating nondiseased mice with KPT-9274 only, platelet and red blood cell counts were within normal limits (supplemental Table 2).

Currently, KPT-9274 is in clinical trials for solid tumors and non-Hodgkin lymphoma. Future studies are needed to identify a subset of patients who would benefit from KPT-9274 treatment as well as the relevant combination strategies. Targeting the reserve capacity of AML cells reportedly renders cells sensitive to chemotherapy by increasing oxidative stress.^13,51  Because KPT-9274 decreases the mitochondrial reserve capacity in AML cells, we hypothesize that this inhibitor will be a viable option for combination with chemotherapy. In addition, combining KPT-9274 with other inhibitors that are effective at eradicating leukemic cells that have increased dependence on mitochondrial respiration and glycolysis may be promising. We showed that KPT-9274 is able to synergize with venetoclax, which is currently in phase 3 clinical trials, in patient samples in vitro. Having a therapy that increases sensitivity to other viable therapeutic options represents an ideal strategy for eradicating leukemic cells that can be applied across broad subtypes of genomically distinct AML.

The authors are grateful to the patients and healthy volunteers who provided blood for the present studies and to the OSU Comprehensive Cancer Center Leukemia Tissue Bank (supported by the National Institutes of Health [NIH], National Cancer Institute [NCI] [P30 CA016058]) for sample procurement. They also acknowledge Ramiro Garzon for providing reagents for this study. The authors also acknowledge Alan Flechtner, HTL, and The Ohio State University Veterinary Histology Laboratory for their assistance in all immunohistochemical studies (also supported by NIH, NCI P30 CA016058).

This work was supported by the NIH, NCI (R35 CA198183, R01 CA223165-01A1, and CA197734-02S1), the OSU Comprehensive Cancer Center using the Pelotonia Foundation funds, and further research support to the Byrd Laboratory from the Four Winds Foundation and the D. Warren Brown Foundation. This study made use of RNA sequencing data generated by The Cancer Genome Atlas Research Network (http://cancergenome.nih.gov/).

Contribution: S.R.M. designed the experiments, analyzed the data, and drafted the manuscript; K.L. conducted computational analysis, and edited and reviewed drafts of the manuscript; N.R.G. conducted in vivo experiments, and edited and reviewed drafts of the manuscript; T.-H.L., M.C., S.O., P.S., Y.A., P. Z., V.M.G., L.B., S.D., K.E.W. B.H, S.H., and J.T.B. contributed to components of the experimental work presented (biologic and animal studies); A.M. accrued patients and reviewed drafts of the paper; J.S.B., V.K.P., K.K., and D.S. provided scientific input and edited the manuscript; A.L. performed all the statistical analyses, and edited and reviewed drafts of the manuscript; E.B and W.S. discovered KPT-9274 and KPT-9331; J.C.B. and R.L. designed the study, supervised the research, reviewed and modified drafts of the manuscript, and obtained funding for the research work; and all authors discussed the results, commented on the manuscript, and approved the final version of the manuscript.

Conflict-of-interest disclosures: E.B. and W.S. are employees of, and have financial interests in, Karyopharm Therapeutics Inc. J.S.B. has performed consulting for AbbVie, AstraZeneca, and Kite Pharma. V.K.P. performed consulting for Orbus Therapeutics and SK Biosciences. The remaining authors declare no competing financial interests.

Correspondence: John C. Byrd, The Ohio State University, 410 West 12th Ave, Wiseman Hall Room 455B, Columbus, OH 43210; e-mail: john.byrd@osumc.edu; and Rosa Lapalombella, The Ohio State University, 410 West 12th Ave, Wiseman Hall Room 455C, Columbus, OH 43210; e-mail: rosa.lapalombella@osumc.edu.