Acute myeloid leukemia (AML) is a hematologic malignancy metabolically dependent on oxidative phosphorylation and mitochondrial electron transport chain (ETC) activity. AML cells are distinct from their normal hematopoietic counterparts by this metabolic reprogramming, which presents targets for new selective therapies. Here, metabolic changes in AML cells after ETC impairment are investigated. Genetic knockdown of the ETC complex II (CII) chaperone protein SDHAF1 (succinate dehydrogenase assembly factor 1) suppressed CII activity and delayed AML cell growth in vitro and in vivo. As a result, a novel small molecule that directly binds to the ubiquinone binding site of CII and inhibits its activity was identified. Pharmacologic inhibition of CII induced selective death of AML cells while sparing normal hematopoietic progenitors. Through stable isotope tracing, results show that genetic or pharmacologic inhibition of CII truncates the tricarboxylic acid cycle (TCA) and leads to anaplerotic glutamine metabolism to reestablish the truncated cycle. The inhibition of CII showed divergent fates, as AML cells lacked the metabolic plasticity to adequately utilize glutamine metabolism, resulting in preferential depletion of key TCA metabolites and death; normal cells were unaffected. These findings provide insight into the metabolic mechanisms that underlie AML's selective inhibition of CII.

Implications:

This work highlights the effects of direct CII inhibition in mediating selective AML cell death and provides insights into glutamine anaplerosis as a metabolic adaptation that can be therapeutically targeted.

This article is featured in Highlights of This Issue, p. 1589

Acute myeloid leukemia (AML) is a blood and bone marrow disease with limited therapeutic options and poor patient outcomes (1). Metabolic reprogramming, arising from an augmented reliance on energetic pathways such as oxidative phosphorylation (OXPHOS), is a distinct characteristic of AML (2–4). Patients with AML exhibit elevated glycolysis at diagnosis and leukemia stem cells (LSC) display increased amino acid uptake and are particularly reliant on glutaminolysis (5–8). Increased fatty acid levels and reliance on fatty acid oxidation are also distinctly observed in LSCs (9, 10). Notably, the inability of LSCs to upregulate glycolysis during OXPHOS inhibition distinguishes them from hematopoietic stem cells (HSC) and presents an avenue for selective AML targeting (3, 4). Moreover, compared to normal cells, an overreliance on oxidative metabolism renders AML cells more sensitive to inhibition of mitochondrial translation and respiration (11–13). The ETC is at the core of mitochondrial respiration containing a series of enzyme complexes and mobile electron carriers generating ATP and signaling molecules (14). Dysregulated ETC function occurs in several tumors where ETC-targeting compounds impart therapeutic benefit (15, 16). In AML, pharmacologic inhibition of ETC complex I (CI) and respiratory supercomplex assembly exploits this distinct OXPHOS dependency to eradicate AML cells preferentially and highlights a therapeutic strategy targeting the ETC (11, 17–19).

Among the ETC complexes, only respiratory CII is elevated in AML relative to normal hematopoietic cells (13). Indirect CII inhibition has been observed in drug-induced AML cell death. Treatment with venetoclax (BCL-2 inhibitor) and azacitidine (nucleoside analog), or cysteinase (prevents glutathionylation), as well as genetic loss of ClpP (mitochondrial protease), each suppress CII activity to preferentially target AML cells (20–22). Despite these intriguing findings, direct inhibition of CII is not completely understood, and, more importantly, the mechanism of CII inhibition that imparts AML selectivity remains largely unknown.

This study examines the effect of direct CII impairment in AML. Genetic loss or pharmacologic inhibition of CII impaired AML cell growth in vitro and in vivo, yet normal hematopoietic cells remained unaffected. Mechanistically, loss of CII activity caused metabolic dysfunction, resulting in diverse cell fates whereby AML, but not normal hematopoietic cells, exhibited insufficient compensatory anaplerotic metabolism of glutamine to sustain tricarboxylic acid cycle (TCA) metabolite levels. The potential of anaplerotic glutamine metabolism as a metabolic vulnerability provides a unique window of selectivity for AML targeting upon CII inhibition.

Cell culture

OCI-AML2, OCI-AML3, and KG1a cells were cultured in Iscove's Modified Dulbecco's Medium (IMDM, Wisent). The U937, monocytic lymphoma cell line was cultured in RPMI 1640 Medium. The surrogate LSC line TEX required IMDM supplemented with 15% (v/v) FBS (Sigma-Aldrich), 20 ng/mL stem cell factor, 2 ng/mL IL3 (Peprotech), and 2 mmol/L l-glutamine (Gibco). TEX cells are derived from primitive human hematopoietic cells, induced to express the TLS-ERG oncogene, resulting in leukemic conversion resembling hierarchal AML (23). Unless otherwise stated, all media are supplemented with 10% FBS, 200 μg of streptomycin, and 200 units of penicillin per milliliter of media (Gibco). Primary human AML, consisting of 80% malignant cells and mononuclear cells, derived from mobilized peripheral blood of mononuclear cells of healthy donors (MNC) were provided by Dr. Mark Minden at the Princess Margaret Cancer Centre (Toronto, ON) with written informed patient consent. Normal, umbilical cord blood (UCB)-derived mononuclear cells were obtained from Dr. Kristin Hope at McMaster Children's Hospital (Hamilton, ON) with consent. As outlined below, these cells are cultured in the same media formula used for isotope tracing. Acquisition of human samples followed Institutional Ethics Review protocols approved by the University Health Network (Toronto, ON), McMaster University (Hamilton, ON), and the University of Guelph (Guelph, ON) and were in accordance with recognized ethical guidelines (i.e., Declaration of Helsinki). All cells are maintained at 37°C in an incubator with 5% CO2 and 95% humidity. Primary cell cytogenetics are outlined in Supplementary Table S1.

Stable isotope tracing

OCI-AML2, TEX, SDHAF1low cells, patient-derived AML, or normal mobilized peripheral blood MNCs (CD34+ enriched; 2.5 × 105 cells/mL) were exposed to either 13C6-glucose or 13C5,15N2-glutamine (Cambridge Isotopes) for 12 hours in the absence or presence of shikonin (1 μmol/L). Following incubation, cells were washed three times with PBS and flash frozen. Cell pellets were extracted with methanol: acetonitrile: water (5:3:2, v/v/v) and then analyzed via UHPLC-MS (Vanquish-QExactive, Thermo Fisher Scientific). Spectral analysis of peak areas and isotopologue determination were defined, as described previously (24). Additional details can be found in the Supplementary Materials and Methods.

In vivo models

To assess the in vivo efficacy of shikonin, a bone marrow engraftment mouse model was used as described previously (9). Patient-derived AML or normal UCB-derived cells (2.5 × 106), were injected into NSG mice via the tail vein and following 1 week, were treated with shikonin three times a week via intraperitoneal injection. In the case of assessment on established disease, treatment commenced 4 weeks following injection, after confirmation via bone marrow aspirate and assessment of the human CD33+/CD45+ cell population via flow cytometry. In both cases, following 5 weeks of treatment, mice were sacrificed, femoral bone marrow was collected and analyzed for the presence of CD33+/CD45+ human cells via flow cytometry. At endpoint, blood was collected and submitted to the University of Guelph Animal Health Laboratory, where a complete blood count and biochemistry was performed. The in vivo effects of shikonin on CII activity were assessed as detailed in the Supplementary Materials and Methods. All animal studies were carried out according to the regulations of the Canadian Council on Animal Care and with the approval of the University of Guelph, Animal Care Committee.

CII activity assay

The protein content of a whole-cell lysate, prepared using RIPA buffer (Sigma-Aldrich), was quantified using the bicinchoninic acid assay. In a 96-well plate, 3 μg of whole-cell lysate in 5 μL is added to 35 μL of assay buffer containing 20 mmol/L succinate, 20 μmol/L rotenone, 2 mmol/L potassium cyanide, 0.99 mg/mL BSA, 50 μmol/L decylubiquinone, and 50 μmol/L HEPES (Sigma-Aldrich). Test compounds are added in a final volume of 5 μL which is then incubated for 30 minutes at room temperature. Malonate (40 mmol/L, Sigma-Aldrich) and 50 μmol/L thenoyltrifluoroacetone (TTFA, Sigma-Aldrich) acted as positive controls for CII inhibition. Then, 1.7 mmol/L of dichlorophenolindophenol (Sigma-Aldrich) was added, and absorbance (600 nm, Biotek Synergy HT spectrophotometer) was monitored for 30 minutes. Enzyme activity was calculated by subtracting the absorbance values at two timepoints within the linear range, using the following formula: activity = (absorbance 2 − absorbance 1)/time (minutes). CIII, CIV, and CV activities were similarly assessed and further described in the Supplementary Materials and Methods.

Lentiviral-mediated knockdown of SDHAF1

Lentiviral-mediated transfection, performed using the hairpin-pLKO.1 vector, contained short hairpin RNA (shRNA) sequences (Supplementary Table S2; Sigma-Aldrich) specific to the succinate dehydrogenase assembly factor 1 (SDHAF1) gene or a noncoding scramble (hereafter referred to as control) and a puromycin resistant gene. Transfection and confirmation via immunoblotting were performed as outlined in the Supplementary Materials and Methods.

Cell growth, viability, and colony formation

After antibiotic selection of SDHAF1 knockdown cells, 1.0 × 105 cells of each knockdown and control, are plated in a 12-well plate (1 mL/well) to assess cell growth. Cells were counted daily using the trypan blue exclusion assay for 4 days. On day 5, following transduction of OCI-AML2, U937, and TEX cells with either control or SDHAF1 targeting shRNAs, 1.0 × 103 cells in 1.1 mL MethoCult GF H4034 medium (Stemcell Technologies) are plated in duplicate in 35 mm dishes. After incubation for 10 to 14 days at 37°C with 5% CO2 and 95% relative humidity, the clusters on dishes were only counted as colonies if they had >50 cells.

Clonogenic growth in primary samples (e.g., 1.0 × 105 patient-derived AML cells or 1.0 × 104 normal UCB cells) are similarly plated in MethoCult GF H4034 medium (Stem Cell Technologies) in the presence or absence of 0.5 μmol/L shikonin, and colonies are counted after a 10- to 14-day incubation. For cell viability, equal numbers of primary and AML cells were seeded and treated with test compounds for 24 to 72 hours. Cells were collected, suspended in 250 μL staining solution containing 7-Aminoactinomycin D (7AAD; 1 μg/mL; Caymen Chemical), and incubated at room temperature for 15 minutes. Fluorescence detected with the Guava EasyCyte 8HT Benchtop Flow Cytometer (Merck Millipore) differentiated nonviable (7AAD positive) and viable cells.

Succinate and fumarate metabolite analysis

OCI-AML2 and TEX cells (1 × 106) transfected with control or SDHAF1 shRNA were collected 24 hours after puromycin selection, washed three times with PBS and pellets were flash frozen. Cell pellets were extracted and analyzed via UHPLC-MS as described in the stable isotope tracing methods. Total metabolite abundance was normalized to cell count.

Oxygen consumption rates and respiratory reserve capacity

The O2K Oxygraph (Oroboros) measured the oxygen consumption rate of OCI-AML2 cells transduced with SDHAF1 shRNA. Cells (5.0 × 106) are permeabilized using 100 μg/mL digitonin (Sigma-Aldrich) in 500 μL of permeabilization buffer containing 250 mmol/L sucrose and 80 mmol/L potassium chloride (Sigma-Aldrich), alongside gentle agitation for 3 minutes and then centrifuged and the cell pellet resuspended in 200 μL of Mir05 buffer (i.e., 110 mmol/L sucrose, 0.5 mmol/L ethylene glycol-bis(2-aminoethylether)-N,N,N′,N′-tetraacetic acid (EGTA), 3.0 mmol/L MgCl2, 80 mmol/L KCl, 60 mmol/L K-lactobionate, 10 mmol/L KH2PO4, 20 mmol/L Taurine, 20 mmol/L Hepes, and 1.0 g/L BSA at pH 7.1). The sample is then transferred into the respirometer chamber containing 2 mL Mir05 buffer, and after equilibration, 2.5 mmol/L ADP, 10 mmol/L succinate, 500 nmol/L oligomycin, and 500 nmol/L FCCP (carbonyl cyanide p-trifluorome-thoxyphenylhydrazone) were sequentially added to measure succinate supported respiration and respiratory reserve capacity. Oxygen consumption rates are computed as the negative time derivative of oxygen concentration using DatLab Software (Oroboros). The respiratory reserve capacity is calculated as the change in oxygen flux from basal respiration to maximal.

In silico chemical screen

ChemMine (http://chemminetools.ucr.edu) searched for small menadione-like molecule compounds (PubChem ID: 4055) restricted to naphthoquinones using a Tanimoto threshold cutoff of 60% to retrieve the top 100 hits. Naphthoquinones are of interest as they previously demonstrated anti-AML activity (25). Nine naphthoquinones were filtered from the original search based on their druggability, eliminating compounds with a log P > 3. The compound with the homology most similar to menadione, determined by comparison of Tanimoto coefficient and atom pairing, is selected as the top candidate

Molecular docking studies

A homology model is prepared in SWISS-MODEL using the X-ray crystallography structure of porcine mitochondrial CII (PDB:1ZOY) and the three-dimensional structures of shikonin (CID: 479503) and ubiquinone-5 (CID: 12832984) as outlined in the Supplementary Materials and Methods.

Immunoprecipitation

Respiratory CI or CII is immunoprecipitated from OCI-AML2 cells treated with 20 μmol/L shikonin for 1 hour following the established protocol (Abcam, 109711, 109799) and shikonin was detected via high-performance liquid chromatography (HPLC) as further detailed in the Supplementary Materials and Methods.

CI activity assay

OCI-AML2, TEX, and U937 cells (1 × 107/condition) were treated with either 10 μmol/L shikonin or 1 μmol/L venetoclax for 6 hours, collected, lysed, and CI activity assayed according to the manufacturer's protocol (Abcam, 109721).

Succinate dehydrogenase and oxidoreductase activity assays

Succinate dehydrogenase (SDH) and succinate quinone oxidoreductase (SQR) enzyme activities are determined using colorimetric assays. For both assays, 1.5 × 107 OCI-AML2 or TEX cells are collected, and the mitochondrial-rich fraction is prepared as outlined in the immunoprecipitation assay. For each assay, 20 μg of the mitochondrial-rich fraction is used; for the SDH assay, protein is suspended in buffer containing 300 mmol/L mannitol, 25 mmol/L KH2PO4 (pH 7.4), 20 mmol/L succinate, 5 μmol/L rotenone, 2 μmol/L antimycin A, and 2 mmol/L potassium cyanide (Sigma-Aldrich). Following 10 minutes of incubation at room temperature, inhibitors are added, and the reaction is initiated by adding 150 μmol/L MTT and 400 μmol/L phenazine methosulphate (Sigma-Aldrich). The change in absorbance of MTT (490 nm) is monitored every minute for 10 minutes. Protein is suspended in the same assay buffer mentioned above to determine SQR activity. The reaction is initiated by 80 μmol/L decylubiquinone and 80 μmol/L dichlorophenolindophenol, and absorbance, at 600 nm, is monitored every 1 minute for 10 minutes. Each assay is done in a 1 cm cuvette and absorbances were monitored with the Genesys 10 spectrophotometer (Thermo Fisher Scientific). Activity is calculated as the change in absorbance over time.

NADH: NAD+ ratios

Patient-derived AML cells and MNCs (2 × 105/condition) were treated with 1 μmol/L shikonin for 6 hours, counted and equal numbers of cells were collected and levels of NADH and NAD+ were colorimetrically assayed according to the manufacturer's protocol (Sigma-Aldrich, MAK037). The kit allows for the individual detection of NADH, NAD+, and their ratios.

ATP measurement

CD34+ cells from patient-derived AML or MNCs were isolated by immunomagnetic selection as described in the manufacturer's protocol (Militenyi Biotec, 130100453). Purity (CD34+ >95%) was confirmed by flow cytometry with a CD34+ FITC antibody (BioLegend, 343518). Isolated CD34+ cells (1 × 105/condition) were treated with 1 μmol/L shikonin for 6 hours, counted and an equal number of cells were collected and ATP levels were assayed by luminescence according to the manufacturer's protocol (Cayman, 700410).

CB839, alpha-ketoglutarate, and glutamine cotreatments

For CB839, OCI-AML2 and TEX cells were treated with 2 μmol/L shikonin in the presence or absence of 200 nmol/L CB839 for 48 hours and cell viability was measured with 7AAD staining and flow cytometry. OCI-AML2 or TEX cells were seeded at 1.0 × 105 cells/mL in standard DMEM, with or without 4 mmol/L glutamine, in the absence or presence of 0.3 μmol/L shikonin. OCI-AML2 or TEX cells were seeded at 1.0 × 105 cells/mL in the absence or presence of 0.3 μmol/L shikonin and/or 100 μmol/L octyl-α-ketoglutarate (Sigma-Aldrich) and cells were counted daily using the trypan blue exclusion assay. The proliferation rate was calculated by the following formula: doublings per day = 3.32*[log (cell count day 4) − log (cell count day 1)].

Aspartate metabolite analysis

OCI-AML2 or TEX cells were seeded at 1.0 × 105 cells/mL in the absence or presence of 0.3 μmol/L shikonin and/or 100 μmol/L octyl-α-ketoglutarate for 6 hours (Sigma-Aldrich). An equal number of cells were then collected, lysed, and aspartate was quantified using a commercial colorimetric kit, according to the manufacturer's protocol (Abcam, ab102512). Total aspartate levels were normalized to the untreated control.

Statistical analysis

Statistics were completed using GraphPad 7.0 Prism. Unless otherwise indicated, results are expressed as a mean ± SD. Significance between values was determined by an unpaired, two-tailed t test, two-way ANOVA, or by a one-way ANOVA paired with a Tukey post hoc test. P < 0.05 was statistically significant.

Respiratory CII activity mediates AML cell growth and viability

AML cells have an increased reliance on OXPHOS (3, 12) and ETC supercomplex assembly (11). CII is the only ETC complex (Fig. 1A) whose activity is elevated in AML (13), and indirect inhibition of CII leads to AML cell death (20–22). No study has assessed mitochondrial metabolism or AML maintenance and growth following direct CII inhibition. In agreement with previous findings (13), CII activity is higher in patient-derived AML cells compared with normal UCB-derived cells (Fig. 1B; cytogenetics; Supplementary Table S1; CD34+ cells at ∼1.5%; hereafter referred to as normals). Next, the CII assembly factor, SDHAF1 (26) was genetically impaired to study the effect of CII enzyme activity on AML cells (Fig. 1C). SDHAF1 was selected over other CII-related proteins, as it is upregulated in LSCs compared with bulk leukemia cells (4). Knockdown of SDHAF1, as confirmed by immunoblotting (Fig. 1DF, top; Supplementary Fig. S1A-S1C; see Supplementary Table S2 for shRNAs), impaired the enzymatic activity of CII but not CIII nor CIV (Fig. 1DF, bottom; Supplementary Fig. S1D-S1F) and reduced proliferation and clonogenicity of AML cell lines, relative to scramble control cells (Fig. 1GI and JL). Interestingly, knockdown efficiency was weakest in U937 cells, which corresponded with an attenuated effect on CII activity, proliferation, and clonogenic growth.

Figure 1.

A, Schematic diagram of the mitochondrial electron transport chain. B, CII activity was spectrophotometrically assessed in primary AML (IDs: 1–9) and normal MNCs. C, Schematic of enzyme subunits of CII and involvement of SDHAF1 as a chaperone protein. D–F, SDHAF1 was knocked down in OCI-AML2, U937, and TEX cells with two unique shRNAs and knockdown was confirmed with immunoblotting (top) and spectrophotometric assessment of CII activity (n = 3; bottom). Following puromycin selection, OCI-AML2 (G), U937 (H), and TEX (I) cells were seeded, and cell counts were recorded daily for 72 hours. The proliferation rate was calculated by the following formula: doublings per day = 3.32*[log(cell count day 4) − log(cell count day 1)]. Data points represent the mean cell counts ± SD of three independent trials. J–L, Clonogenic growth of OCI-AML2, U937, and TEX cells following SDHAF1 knockdown. Mean colony counts ± SD from three trials are shown. M, NOD/SCID mice were injected with 2.5 × 106 control or SDHAF1 knockdown TEX cells (transduced with shRNA1; SDHAF1low) via the tail-vein. N, After 8 weeks, % human CD45 cells in mouse bone marrow aspirates were evaluated by flow cytometry. Data are normalized to the control group. O, Kaplan–Meier plot depicting the percent survival of mice bearing control and SDHAF1low TEX cells. *, P < 0.05; **, P < 0.01; ***, P < 0.001. n.s., not significant; two-tailed, unpaired, Student t test (B and N), one-way ANOVA with Tukey post hoc test (DF, JL), log-rank test (O).

Figure 1.

A, Schematic diagram of the mitochondrial electron transport chain. B, CII activity was spectrophotometrically assessed in primary AML (IDs: 1–9) and normal MNCs. C, Schematic of enzyme subunits of CII and involvement of SDHAF1 as a chaperone protein. D–F, SDHAF1 was knocked down in OCI-AML2, U937, and TEX cells with two unique shRNAs and knockdown was confirmed with immunoblotting (top) and spectrophotometric assessment of CII activity (n = 3; bottom). Following puromycin selection, OCI-AML2 (G), U937 (H), and TEX (I) cells were seeded, and cell counts were recorded daily for 72 hours. The proliferation rate was calculated by the following formula: doublings per day = 3.32*[log(cell count day 4) − log(cell count day 1)]. Data points represent the mean cell counts ± SD of three independent trials. J–L, Clonogenic growth of OCI-AML2, U937, and TEX cells following SDHAF1 knockdown. Mean colony counts ± SD from three trials are shown. M, NOD/SCID mice were injected with 2.5 × 106 control or SDHAF1 knockdown TEX cells (transduced with shRNA1; SDHAF1low) via the tail-vein. N, After 8 weeks, % human CD45 cells in mouse bone marrow aspirates were evaluated by flow cytometry. Data are normalized to the control group. O, Kaplan–Meier plot depicting the percent survival of mice bearing control and SDHAF1low TEX cells. *, P < 0.05; **, P < 0.01; ***, P < 0.001. n.s., not significant; two-tailed, unpaired, Student t test (B and N), one-way ANOVA with Tukey post hoc test (DF, JL), log-rank test (O).

Close modal

To assess the impact of SDHAF1 on leukemia-initiating cells in vivo, equal numbers of TEX cells with endogenous or reduced levels of SDHAF1 (transduced with shRNA1, hereafter denoted as SDHAF1low cells) were tail vein injected into NSG mice (Fig. 1M). After 8 weeks, assessment of CD45+ cells in bone marrow aspirates showed that mice with SDHAF1low cells had decreased engraftment compared with control mice (Fig. 1N). These mice also exhibited improved survival after 110 days (Fig. 1O). These in vitro and in vivo results show SDHAF1 knockdown impairs CII activity and delays leukemia growth.

Knockdown of SDHAF1 impairs oxidative metabolism in AML

CII resides in both the ETC and TCA cycle and; therefore, it impacts both pathways (Fig. 2A). Inhibition of CII activity results in a characteristic accumulation of succinate and depletion of fumarate (27, 28), which was observed in SDHAF1low AML cells (Fig. 2B); succinate-supported oxygen flux and respiratory reserve capacity are decreased in these cells (Fig. 2C). Consequently, we analyzed TCA metabolites in OCI-AML2 and TEX SDHAF1low cells, which had the greatest knockdown efficiency, through stable isotope tracing of 13C6-glucose. Glucose enters the TCA cycle via mitochondrial pyruvate, which is oxidized to acetyl-CoA via pyruvate dehydrogenase (PDH), generating two labeled carbons (M+2); or reduced to oxaloacetate by pyruvate carboxylase (PC), generating three labeled carbons (M+3; Fig. 2D). As expected with loss of CII activity, steady-state levels of fumarate and malate (e.g., TCA metabolites that proceed SDH) were modestly reduced in SDHAF1low cells, while aspartate was significantly reduced (Fig. 2E, TEX; Supplementary Fig. S2E). Notably, OCI-AML2 SDHAF1low cells exhibited a modest but significant increase in glucose flux to pyruvate and citrate while it remained unchanged in TEX SDHAF1low cells (Supplementary Fig. S2A–S2D), suggesting metabolite depletion is not a product of decreased entry of glucose into the TCA cycle. Decreased M+2 isotopologue abundance in the TCA metabolites proceeding SDH indicates that loss of CII activity arrests oxidative flux at SDH and truncates the TCA cycle (Fig. 2F, TEX; Supplementary Fig. S2F). Consistent with this, fractional incorporation of glucose-derived carbons through oxidative PDH significantly diminishes in fumarate and aspartate in OCI-AML2 cells (Fig. 2G) and in fumarate, malate, and aspartate in TEX cells (Supplementary Fig. S2G). Interestingly, the proportion of M+3 malate and aspartate, representing reductive metabolism through PC, significantly increased in OCI-AML2 SDHAF1low cells while the fumarate proportion remained unchanged (Fig. 2H). In TEX SDHAF1low cells, M+3 enrichment was observed only with aspartate (Supplementary Fig. S2F and S2G). Still, the M+3 isotopologue comprised a minor percentage of the total pool of these metabolites (<12%) and failed to significantly contribute to steady-state concentrations, given that total metabolite levels were still suppressed. These results suggest SDHAF1 knockdown impairs oxidative metabolism of glucose, resulting in TCA cycle truncation.

Figure 2.

A, Representation of the TCA cycle with its direct link to CII. B, Succinate and fumarate levels were measured, as outlined in the Materials and Methods, in AML and SDHAF1low cells (n = 3). C, Using respirometry, respiratory reserve capacity was measured in whole OCI-AML2 cells by calculating the change in oxygen consumption from basal to an uncoupled state (upon FCCP addition) and succinate supported respiration was measured in permeabilized OCI-AML2 cells. D, Schematic of 13C6-glucose labeling patterns through the TCA cycle. Grey circles represent oxidative glucose metabolism whereas red circles are reductive glucose metabolism. E, OCI-AML2 control and SDHAF1low cells (transduced with shRNA1) were exposed to 13C6-glucose for 12 hours and abundance of TCA metabolites was determined by adding peak areas of all isotopologues identified. Fold changes (mean ± SEM) between OCI-AML2 control and SDHAF1low cells were calculated with the following formula: fold change = log2(total AU SDHAF1low/total AU control). F, M+2 abundance of TCA metabolites in OCI-AML2 control and SDHAF1low cells after exposure to 13C6-glucose for 12 hours. G, Glucose contribution to malate, fumarate, and aspartate via oxidative metabolism was determined by calculating the percent proportion of the M+2 isotopologue fraction, relative to the total metabolite pool. H, Glucose contribution to fumarate, malate, and aspartate via reductive glucose metabolism was determined by calculating the percent proportion of the M+3 isotopologue fraction, relative to the total metabolite pool. Data represent the mean ± SD of three biological replicates unless otherwise indicated. *, P < 0.05; **, P < 0.01; ***, P < 0.001; n.s., not significant; two-tailed, unpaired, Student t test (BJ).

Figure 2.

A, Representation of the TCA cycle with its direct link to CII. B, Succinate and fumarate levels were measured, as outlined in the Materials and Methods, in AML and SDHAF1low cells (n = 3). C, Using respirometry, respiratory reserve capacity was measured in whole OCI-AML2 cells by calculating the change in oxygen consumption from basal to an uncoupled state (upon FCCP addition) and succinate supported respiration was measured in permeabilized OCI-AML2 cells. D, Schematic of 13C6-glucose labeling patterns through the TCA cycle. Grey circles represent oxidative glucose metabolism whereas red circles are reductive glucose metabolism. E, OCI-AML2 control and SDHAF1low cells (transduced with shRNA1) were exposed to 13C6-glucose for 12 hours and abundance of TCA metabolites was determined by adding peak areas of all isotopologues identified. Fold changes (mean ± SEM) between OCI-AML2 control and SDHAF1low cells were calculated with the following formula: fold change = log2(total AU SDHAF1low/total AU control). F, M+2 abundance of TCA metabolites in OCI-AML2 control and SDHAF1low cells after exposure to 13C6-glucose for 12 hours. G, Glucose contribution to malate, fumarate, and aspartate via oxidative metabolism was determined by calculating the percent proportion of the M+2 isotopologue fraction, relative to the total metabolite pool. H, Glucose contribution to fumarate, malate, and aspartate via reductive glucose metabolism was determined by calculating the percent proportion of the M+3 isotopologue fraction, relative to the total metabolite pool. Data represent the mean ± SD of three biological replicates unless otherwise indicated. *, P < 0.05; **, P < 0.01; ***, P < 0.001; n.s., not significant; two-tailed, unpaired, Student t test (BJ).

Close modal

Knockdown of SDHAF1 constrains TCA metabolite synthesis

The primary carbon sources for central metabolism are glucose and glutamine, and during ETC impairment, glutamine is the preferred anaplerotic substrate (29). As glucose does not contribute to metabolite maintenance in SDHAF1low cells, the metabolic fate of glutamine was examined. By oxidatively deaminating glutamate to alpha-ketoglutarate, glutamine enters the TCA cycle, where it is oxidatively decarboxylated to succinyl-CoA or reductively carboxylated to isocitrate (Fig. 3A). SDHAF1low and parental cells were incubated with 13C5,15N2-glutamine to trace products of glutamine metabolism. The oxidative metabolism of glutamine produces fumarate, malate, and aspartate, each with four labeled carbons (M+4), while the reductive metabolism produces three (M+3). Incorporation of acetyl-coA derived either from reductive or oxidative metabolism produces M+2 metabolites. As indicated by the fractional incorporation of the M+5 isotopologue, glutamine flux to glutamate is unchanged upon SDHAF1 knockdown; however, flux into the TCA cycle via alpha-ketoglutarate increased (Fig. 3B, TEX; Supplementary Fig. S3B).

Figure 3.

A, Schematic representation of 13C5,15N2-glutamine labeling patterns through the TCA cycle. Grey circles represent oxidative glutamine metabolism whereas red circles are reductive glutamine metabolism. B, OCI-AML2 SDHAF1low and control cells were cultured in the presence of 13C5,15N2-glutamine for 12 hours. Labeled glutamine incorporation into glutamate, α-ketoglutarate and glutathione (GSH) was then determined by dividing the relevant isotopologue abundance by that of the total metabolite pool. For glutathione, the M+3 isotopologue was added to M+5 to account for glutathione that results from reduction of GSSG. Fractional incorporation of the oxidative (M+4) and reductive (M+3) isotopologue into succinate (C), fumarate (D), malate (E), and aspartate (F) was calculating by dividing each isotopologue by the total metabolite pool. Data represent the mean ± SD of three biological replicates. *, P < 0.05; **, P < 0.01; n.s., not significant; two-tailed, unpaired, Student t test (BF).

Figure 3.

A, Schematic representation of 13C5,15N2-glutamine labeling patterns through the TCA cycle. Grey circles represent oxidative glutamine metabolism whereas red circles are reductive glutamine metabolism. B, OCI-AML2 SDHAF1low and control cells were cultured in the presence of 13C5,15N2-glutamine for 12 hours. Labeled glutamine incorporation into glutamate, α-ketoglutarate and glutathione (GSH) was then determined by dividing the relevant isotopologue abundance by that of the total metabolite pool. For glutathione, the M+3 isotopologue was added to M+5 to account for glutathione that results from reduction of GSSG. Fractional incorporation of the oxidative (M+4) and reductive (M+3) isotopologue into succinate (C), fumarate (D), malate (E), and aspartate (F) was calculating by dividing each isotopologue by the total metabolite pool. Data represent the mean ± SD of three biological replicates. *, P < 0.05; **, P < 0.01; n.s., not significant; two-tailed, unpaired, Student t test (BF).

Close modal

Under conditions of impaired oxidative metabolism, glutamine flux through TCA metabolites overcomes TCA cycle truncation (29). SDHAF1low cells did not increase glutamine flux, either through oxidation or reduction, to support citrate, fumarate, malate, or succinate levels (Fig. 3CE; Supplementary Fig. S3A and S3C–S3F). Conversely, these cells significantly elevated glutamine reductive flux to aspartate, as demonstrated by the proportion of the M+3 isotopologue (Fig. 3F, TEX; Supplementary Fig. S3G). In spite of this increase, total aspartate was the most suppressed of all TCA metabolites in SDHAF1low cells (Fig. 2F). Because of its role in regulating proliferation capacity following ETC inhibition (30, 31), a decrease in total aspartate in SDHAF1low cells can contribute to a reduction in cell growth and clonogenicity. The results suggest that glutamine is not able to sustain TCA metabolite synthesis upon SDHAF1 knockdown or impairment of oxidative glucose metabolism in AML cells.

Identification of a novel pharmacologic inhibitor of CII

Because there are no clinically approved direct inhibitors of CII and investigational small molecules have high EC50 values, reported toxicity (32) or indirectly inhibit CII, we next sought to identify a more potent direct inhibitor. We previously showed that menadione inhibits CI and CII to reduce leukemia cell growth (33). Menadione is a naphthoquinone; a compound class with demonstrated anti-AML activity and evolutionary similarity to the ETC substrate, ubiquinone (25, 34). Reasoning that naphthoquinones structurally similar to menadione could selectively bind to only CII, an in silico screen searched for potential targets (Fig. 4A). By adding further limitation to the computational screen that included druggable compounds, as determined using the Lipinski rule of 5 (35), 1,4-naphthoquinone, 1,2-naphthoquinone, and shikonin were identified (Supplementary Fig. S4A). The top hit from this screen was shikonin (Fig. 4B), a type of 1,4-naphthoquinone, with reported anticancer activity (36–38). Molecular docking using crystal structures of CI and CII (PDB:5LDW; PDB:IZOY) suggest that shikonin binds to the distal and proximal ubiquinone binding sites of CII (Fig. 4C); no affinity was suggested for CI. Tangible, cell-based interactions were investigated using protein G-agarose cross-linked antibodies to immunoprecipitate (IP) CI and CII from shikonin-treated cells (Fig. 4D). Immunoblotting for SDHA/B and MT-ND1, subunits of CII and CI, respectively, verified IP specificity (Supplementary Fig. S4B). Shikonin was detected only in CII fractions (Fig. 4E and F; Supplementary Fig. S4C; ref. 39).

Figure 4.

A, Schematic diagram of in silico ChemMine screen schema used to identify druggable (log P < 3) compounds with structural similarity to menadione. B, Shikonin, a 1,4-naphthoquinone was identified as a top hit in this screen. C, Molecular docking investigation of the crystal structure of porcine CII (PDB IZOY) using AutoDock 4.0. showed that shikonin has affinity for both the proximal (yellow circle) and distal (red circle) ubiquinone binding sites of CII. D, Co-immunoprecipitation experimental design to isolate CI and CII and detect shikonin. OCI-AML2 cells were treated with shikonin (20 μmol/L) for 1 hour and mitochondria were isolated and incubated with CI or CII specific conjugated antibody's overnight. After incubation and elution, HPLC was used to detect for shikonin. E, Shikonin was not found in CI fractions but (F; n = 4) was present in CII fractions. (G; n = 4) CI diaphorase-type enzyme activity was measured in AML cell lines in presence of shikonin (10 μmol/L) or venetoclax (1 μmol/L), a CI inhibitor. H, Schematic of the SQR and SDH enzyme activities of CII. Spectrophotometric determination of both SQR (I) and SDH (J) enzyme activities in AML cells in presence of shikonin (5 μmol/L, Shik). TTFA (50 μmol/L) and malonate (40 mmol/L) were used as positive controls for inhibition of SQR and SDH activity, respectively. K, CII activity was spectrophotometrically measured in primary AML cells treated with either shikonin (10 μmol/L) or TTFA (50 μmol/L). AML patient samples (IDs: 2,4–8) were used for this analysis. Unless otherwise indicated, data represent the mean ± SD of three biological replicates. *, P < 0.05; **, P < 0.01; ***, P < 0.001; n.s., not significant; two-tailed, unpaired, Student t test (E and F); two-way ANOVA with Dunnett multiple comparisons test (G); one-way ANOVA with Tukey post hoc test (IK).

Figure 4.

A, Schematic diagram of in silico ChemMine screen schema used to identify druggable (log P < 3) compounds with structural similarity to menadione. B, Shikonin, a 1,4-naphthoquinone was identified as a top hit in this screen. C, Molecular docking investigation of the crystal structure of porcine CII (PDB IZOY) using AutoDock 4.0. showed that shikonin has affinity for both the proximal (yellow circle) and distal (red circle) ubiquinone binding sites of CII. D, Co-immunoprecipitation experimental design to isolate CI and CII and detect shikonin. OCI-AML2 cells were treated with shikonin (20 μmol/L) for 1 hour and mitochondria were isolated and incubated with CI or CII specific conjugated antibody's overnight. After incubation and elution, HPLC was used to detect for shikonin. E, Shikonin was not found in CI fractions but (F; n = 4) was present in CII fractions. (G; n = 4) CI diaphorase-type enzyme activity was measured in AML cell lines in presence of shikonin (10 μmol/L) or venetoclax (1 μmol/L), a CI inhibitor. H, Schematic of the SQR and SDH enzyme activities of CII. Spectrophotometric determination of both SQR (I) and SDH (J) enzyme activities in AML cells in presence of shikonin (5 μmol/L, Shik). TTFA (50 μmol/L) and malonate (40 mmol/L) were used as positive controls for inhibition of SQR and SDH activity, respectively. K, CII activity was spectrophotometrically measured in primary AML cells treated with either shikonin (10 μmol/L) or TTFA (50 μmol/L). AML patient samples (IDs: 2,4–8) were used for this analysis. Unless otherwise indicated, data represent the mean ± SD of three biological replicates. *, P < 0.05; **, P < 0.01; ***, P < 0.001; n.s., not significant; two-tailed, unpaired, Student t test (E and F); two-way ANOVA with Dunnett multiple comparisons test (G); one-way ANOVA with Tukey post hoc test (IK).

Close modal

AML cells were tested for direct enzyme activity of CI and CII after shikonin treatment and CI activity was unaffected (Fig. 4G). CII has two unique enzymatic functions (Fig. 4H); through SDH activity, succinate is oxidized into fumarate, and through SQR activity, electrons are transferred from the flavin site to ubiquinone. Shikonin impairs both SQR and SDH activity (Fig. 4I and J), which is consistent with medium-affinity ubiquinone binding-site inhibitors (40). In primary AML cells, shikonin significantly reduced CII activity (Fig. 4K), but not the activity of other mitochondrial complexes (Supplementary Fig. S4D–S4F). These results indicate that shikonin is a specific inhibitor for CII.

CII impairment provokes reliance on glutamine metabolism

While previous studies show indirect inhibition of CII selectively eradicates AML cells, the effects of directly inhibiting CII remain unknown (20, 21). Shikonin reduced the viability of AML cell lines (IC50 = 0.6–1.6 μmol/L; Supplementary Fig. S5A), patient-derived bulk, CD34+ AML cells (Fig. 5A), and suppressed AML clonogenic growth while sparing normal hematopoietic cells (Fig. 5B). Cytotoxicity in primary AML cells and cell lines occurred in a dose- and time-dependent manner (Supplementary Fig. S5B–S5F). Despite maintained survival in normal CD34+ hematopoietic cells, shikonin significantly reduced ATP levels in both cell types, though to a lesser degree in normal CD34+ cells (Supplementary Fig. S6A). In primary AML and normal UCB-derived cells, CII activity and NAD+ to NADH ratios are similarly impaired by shikonin (Fig. 5C; Supplementary Fig. S6B), suggesting that shikonin similarly impairs OXPHOS. Taken together, targeting CII preferentially impacts AML cells through mechanisms independent of ATP.

Figure 5.

A, Bulk AML cells (IDs: 1,3,4), CD34+ cells isolated from primary AML samples (IDs: 22,23,24), and CD34+ normal cells from mobilized peripheral blood (n = 3) were seeded at equal numbers and treated with 1 μmol/L of shikonin or vehicle for 24 hours. Cell viability was then measured by 7AAD staining and flow cytometry. B, A colony-forming assay was performed with patient-derived AML cells (IDs: 2,5,7,10) and normal UCB cells (n = 4). C, CII activity in primary AML (IDs: 2,5,7,8) and normal UCB-derived cells (n = 3) exposed to shikonin (10 μmol/L) or malonate (40 mmol/L, Mal), as a positive control. Patient-derived AML cells (IDs: 6, 11, 12, 13, 20, 21) and MNCs (n = 3) were exposed to 13C6-glucose in the presence or absence of shikonin (1 μmol/L) for 12 hours and M+2 abundance of TCA metabolites in primary AML cells (D) and MNCs (E) was determined. Fractional incorporation of labeled glutamine into glutamate, α-ketoglutarate, and GSH in patient-derived AML cells (F; IDs: 14,15,16) and MNCs (G) after exposure to 13C5,15N2-glutamine in the presence or absence of shikonin (1 μmol/L) for 12 hours was determined using LC/MS. Fractional incorporation of each isotopologue into succinate (H), fumarate (I), aspartate (J), and malate (K) was calculating by dividing each isotopologue by the total metabolite pool (mean ± SEM). Data represent the mean ± SD of three biological replicates unless otherwise stated. MNC, mobilized peripheral blood mononuclear cells; Shik, shikonin; *, P < 0.05; **, P < 0.01; ***, P < 0.01 vehicle versus shikonin (AG); *, P < 0.05 vehicle versus shikonin M+3 (HK); n.s., not significant; two-tailed, unpaired, Student t test (A, DK); one-way ANOVA with Tukey post hoc test (C).

Figure 5.

A, Bulk AML cells (IDs: 1,3,4), CD34+ cells isolated from primary AML samples (IDs: 22,23,24), and CD34+ normal cells from mobilized peripheral blood (n = 3) were seeded at equal numbers and treated with 1 μmol/L of shikonin or vehicle for 24 hours. Cell viability was then measured by 7AAD staining and flow cytometry. B, A colony-forming assay was performed with patient-derived AML cells (IDs: 2,5,7,10) and normal UCB cells (n = 4). C, CII activity in primary AML (IDs: 2,5,7,8) and normal UCB-derived cells (n = 3) exposed to shikonin (10 μmol/L) or malonate (40 mmol/L, Mal), as a positive control. Patient-derived AML cells (IDs: 6, 11, 12, 13, 20, 21) and MNCs (n = 3) were exposed to 13C6-glucose in the presence or absence of shikonin (1 μmol/L) for 12 hours and M+2 abundance of TCA metabolites in primary AML cells (D) and MNCs (E) was determined. Fractional incorporation of labeled glutamine into glutamate, α-ketoglutarate, and GSH in patient-derived AML cells (F; IDs: 14,15,16) and MNCs (G) after exposure to 13C5,15N2-glutamine in the presence or absence of shikonin (1 μmol/L) for 12 hours was determined using LC/MS. Fractional incorporation of each isotopologue into succinate (H), fumarate (I), aspartate (J), and malate (K) was calculating by dividing each isotopologue by the total metabolite pool (mean ± SEM). Data represent the mean ± SD of three biological replicates unless otherwise stated. MNC, mobilized peripheral blood mononuclear cells; Shik, shikonin; *, P < 0.05; **, P < 0.01; ***, P < 0.01 vehicle versus shikonin (AG); *, P < 0.05 vehicle versus shikonin M+3 (HK); n.s., not significant; two-tailed, unpaired, Student t test (A, DK); one-way ANOVA with Tukey post hoc test (C).

Close modal

To further elucidate the specificity of direct CII inhibition resulting in AML death, glucose and glutamine metabolism in AML cells and normal mobilized peripheral blood MNCs were probed using stable-isotope tracing. Patient-derived AML cells or mobilized peripheral blood MNCs (CD34+ enriched) were incubated with either 13C6-glucose or 13C5,15N2-glutamine in the absence or presence of shikonin for 12 hours, after which lysates were analyzed via LC/MS. CII inhibition significantly suppressed steady-state levels of metabolites that proceed SDH in AML cells but, remarkably, levels were maintained in MNCs (Supplementary Fig. S6C–S6H). As with SDHAF1 knockdown, shikonin disrupted oxidative glucose metabolism downstream of SDH in both MNCs and AML cells (Fig. 5D and E; Supplementary Fig. S6I–S6M). Consistent with impaired oxidative glucose metabolism, increased fractional incorporation of the M+0 isotopologue into fumarate, malate, and aspartate upon shikonin treatment suggests nonglucose sources (e.g., glutamine) contribute to metabolite maintenance upon CII inhibition in primary cells (Supplementary Fig. S6K–S6M). Using 13C5,15N2-glutamine labeling, CII inhibition augmented glutamine flux to glutamate in both AML and MNCs corresponding with an increase in flux to alpha-ketoglutarate which was only significant in MNCs (Fig. 5F and G). Entry of glutamine into the TCA cycle via alpha-ketoglutarate is consistent with dysfunctional OXPHOS and ETC (41, 42). Consistent with glucose tracing experiments, oxidative flux to succinate was unchanged upon CII inhibition in both cell types (M+4; Fig. 5H). Alpha-ketoglutarate reduction to citrate and the downstream metabolite, malate, was minimal upon CII inhibition (Fig. 5K; Supplementary Fig. S7A). However, evidence of reductive carboxylation was found in fumarate and aspartate, where upon CII inhibition, the M+3 isotopologue was dominant at approximately 50% in AMLs but 80% in MNCs corresponding with sustained steady-state metabolite levels in MNCs but not AML cells (Fig. 5I and J; Supplementary Fig. S7B–S7G). Together, these results suggest that CII inhibition elicits a reliance on glutamine metabolism, which is insufficient to sustain metabolite levels in AML cells, resulting in preferential depletion and cell death.

Glutamine supports proliferation following CII inhibition

The ability of glutamine metabolism to influence shikonin cytotoxicity was investigated to further understand the reliance of glutamine in CII-mediated selectivity. Shikonin's antiproliferative effect on OCI-AML2 cells was enhanced by the removal of glutamine (Fig. 6A). Furthermore, cotreatment with the glutaminolysis inhibitor CB-839, which prevents glutamine conversion to glutamate, increased shikonin cytotoxicity (Fig. 6B and C). Next, the impact of glutamine carbon entry into the TCA cycle was assessed using cell-permeable alpha-ketoglutarate, the end product of glutaminolysis. By adding octyl-alpha-ketoglutarate to shikonin-treated OCI-AML2 and TEX cells, proliferation rates and aspartate levels were restored (Fig. 6DG). Under CII inhibition, glutamine flux into the TCA cycle is shown to sustain proliferation, implying that increasing pathway efficiency is capable of rescuing cells from CII-induced death (Fig. 6H).

Figure 6.

A, OCI-AML2 cells were seeded in equal numbers in standard DMEM supplemented with or without 4 mmol/L glutamine and 0.3 μmol/L shikonin and cells were counted daily using the trypan blue exclusion assay. The proliferation rate was calculated by the following formula: doublings per day = 3.32*[log(cell count day 3) − log(cell count day 1)]. OCI-AML2 (B) and TEX (C) cells were treated with 2 μmol/L shikonin in the presence or absence of 200 nmol/L CB839 for 48 hours and cell viability was measured by 7AAD staining and flow cytometry. Viability was normalized to the untreated control. OCI-AML2 (D) or TEX (E) seeded at equal numbers in the absence or presence of 0.3 μmol/L shikonin and/or 100 μmol/L octyl-α-ketoglutarate and cells were counted on day 4 using the trypan blue exclusion assay. The proliferation rate was calculated by the following formula: doublings per day = 3.32*[log(cell count day 4) − log(cell count day 1)]. OCI-AML2 cells (F) or TEX cells (G) were seeded at equal numbers in the absence or presence of 0.3 μmol/L shikonin and/or 100 μmol/L octyl-α-ketoglutarate for 6 hours. Cells were then collected, lysed, and aspartate was quantified using a colorimetric kit. Total aspartate levels were normalized to the untreated control. H, Schematic describing the mechanism of AML selectivity, dictated by metabolism of glutamine and maintenance of fumarate, malate, and aspartate levels. Data represent the mean ± SD of three biological replicates. *, P < 0.05; **, P < 0.01; ***, P < 0.001; n.s., not significant; two-way ANOVA with Sidak multiple comparisons test (AC); one-way ANOVA with Tukey post hoc test (DG).

Figure 6.

A, OCI-AML2 cells were seeded in equal numbers in standard DMEM supplemented with or without 4 mmol/L glutamine and 0.3 μmol/L shikonin and cells were counted daily using the trypan blue exclusion assay. The proliferation rate was calculated by the following formula: doublings per day = 3.32*[log(cell count day 3) − log(cell count day 1)]. OCI-AML2 (B) and TEX (C) cells were treated with 2 μmol/L shikonin in the presence or absence of 200 nmol/L CB839 for 48 hours and cell viability was measured by 7AAD staining and flow cytometry. Viability was normalized to the untreated control. OCI-AML2 (D) or TEX (E) seeded at equal numbers in the absence or presence of 0.3 μmol/L shikonin and/or 100 μmol/L octyl-α-ketoglutarate and cells were counted on day 4 using the trypan blue exclusion assay. The proliferation rate was calculated by the following formula: doublings per day = 3.32*[log(cell count day 4) − log(cell count day 1)]. OCI-AML2 cells (F) or TEX cells (G) were seeded at equal numbers in the absence or presence of 0.3 μmol/L shikonin and/or 100 μmol/L octyl-α-ketoglutarate for 6 hours. Cells were then collected, lysed, and aspartate was quantified using a colorimetric kit. Total aspartate levels were normalized to the untreated control. H, Schematic describing the mechanism of AML selectivity, dictated by metabolism of glutamine and maintenance of fumarate, malate, and aspartate levels. Data represent the mean ± SD of three biological replicates. *, P < 0.05; **, P < 0.01; ***, P < 0.001; n.s., not significant; two-way ANOVA with Sidak multiple comparisons test (AC); one-way ANOVA with Tukey post hoc test (DG).

Close modal

Pharmacologic inhibition of CII reduces engraftment in mouse models of human leukemia

To investigate shikonin activity in vivo, patient-derived AML cells were injected into NSG mice and after 7 days, mice received either vehicle or shikonin three times a week for 5 weeks (Fig. 7A). Treatment with either 2 or 3 mg/kg shikonin, significantly decreased engraftment (Fig. 7B). Assessment of body mass (Supplementary Fig. S8A) and blood markers (Supplementary Fig. S8B–S8D) showed that shikonin treatment was well tolerated. Shikonin was assessed with a similar engraftment model with UCB-derived cells. Here, 2.5 mg/kg of shikonin did not affect normal hematopoietic cell engraftment (Fig. 7C). The effects of CII inhibition on advanced disease were next investigated. An aspirate of bone marrow was taken 4 weeks after transplanting patient-derived AML cells to confirm disease (Fig. 7D; Supplementary Fig. S8E). Treatment with 2.5 mg/kg shikonin, as described above, caused a significant reduction in leukemia burden (Fig. 7E).

Figure 7.

A, Patient-derived AML (IDs: 17–19) or normal cord blood–derived mononuclear cells (2.5 × 106) were intravenously injected into NSG mice and treated with shikonin (2–3.0 mg/kg body weight) via an intraperitoneal injection three times weekly for 5 weeks. B, Mice were then euthanized, femoral bone marrow was collected and then analyzed for the presence of human CD33+ and CD45+ cells via flow cytometry. Mice treated with shikonin exhibited reduced engraftment of AML cells but (C) no change in engraftment was observed in normal cells. D, To assess the effects of shikonin on more advanced disease, NSG mice were injected with AML cells; however, treatment with shikonin did not begin until 4 weeks after injection of primary cells. Thereafter, mice were treated three times a week for 5 weeks. E, Upon euthanasia, mice femoral bone marrow was collected analyzed for the presence of human CD33+ and CD45+ cells via flow cytometry. F,In vivo assessment of CII inhibition was measured by injecting NSG mice with primary AML cells (ID: 18) and following confirmation of disease burden (i.e., >70% of AML cells in bone marrow), mice were treated with a bolus dose of shikonin (0, 2.5, or 3.5 mg/kg). At 12 or 24 hours after treatment, mice were sacrificed, femoral bone marrow was collected, and immunomagnetic separation was used to purify human CD33+/CD45+ cells. Assessment of CII enzyme activity in the purified cells showed a dose-dependent decrease in CII activity, 24 hours after treatment (G) and a more pronounced effect 12 hours after treatment (H). Data are normalized to the vehicle (Veh) control, *, P < 0.05; **, P < 0.01; ***, P < 0.001; n.s., not significant; two-tailed, unpaired, Student t test (BE); one-way ANOVA with Tukey post hoc test (G and H).

Figure 7.

A, Patient-derived AML (IDs: 17–19) or normal cord blood–derived mononuclear cells (2.5 × 106) were intravenously injected into NSG mice and treated with shikonin (2–3.0 mg/kg body weight) via an intraperitoneal injection three times weekly for 5 weeks. B, Mice were then euthanized, femoral bone marrow was collected and then analyzed for the presence of human CD33+ and CD45+ cells via flow cytometry. Mice treated with shikonin exhibited reduced engraftment of AML cells but (C) no change in engraftment was observed in normal cells. D, To assess the effects of shikonin on more advanced disease, NSG mice were injected with AML cells; however, treatment with shikonin did not begin until 4 weeks after injection of primary cells. Thereafter, mice were treated three times a week for 5 weeks. E, Upon euthanasia, mice femoral bone marrow was collected analyzed for the presence of human CD33+ and CD45+ cells via flow cytometry. F,In vivo assessment of CII inhibition was measured by injecting NSG mice with primary AML cells (ID: 18) and following confirmation of disease burden (i.e., >70% of AML cells in bone marrow), mice were treated with a bolus dose of shikonin (0, 2.5, or 3.5 mg/kg). At 12 or 24 hours after treatment, mice were sacrificed, femoral bone marrow was collected, and immunomagnetic separation was used to purify human CD33+/CD45+ cells. Assessment of CII enzyme activity in the purified cells showed a dose-dependent decrease in CII activity, 24 hours after treatment (G) and a more pronounced effect 12 hours after treatment (H). Data are normalized to the vehicle (Veh) control, *, P < 0.05; **, P < 0.01; ***, P < 0.001; n.s., not significant; two-tailed, unpaired, Student t test (BE); one-way ANOVA with Tukey post hoc test (G and H).

Close modal

The effects of shikonin on CII activity in NSG mice transplanted with patient-derived AML cells, were next assessed. Seven weeks after injection of cells, a disease burden of >70% was confirmed by bone marrow aspiration (Supplementary Fig. S8F), and shikonin was then administered intraperitoneally at two concentrations and timepoints (Fig. 7F). Human cells were then isolated from femoral bone marrow (Supplementary Fig. S8G), and CII enzyme activity was measured. CII activity in isolated human cells was reduced by approximately 60% at the 3.5 mg/kg dose, 24 hours after shikonin treatment (Fig. 7G). CII activity decreased further when assessed at an earlier timepoint (i.e., closer to shikonin's half-life of 8.8 hours; ref. 43; Fig. 7H). These data show that shikonin-mediated CII inhibition in vivo preferentially targets primitive leukemia cells and is well tolerated.

Compared with normal hematopoietic cells, AML cells rely on OXPHOS and have elevated CII activity (3, 13, 44). Previous studies show indirect inhibition of CII eliminated AML cells, sparing normal MNCs (20, 21). This study is the first to use direct CII inhibition to uncover divergent cell fates leading to selective toxicity. Direct CII inhibition truncated the TCA cycle leading to decreased metabolic plasticity, and created a reliance on glutamine metabolism to sustain biosynthesis in both cell types. However, in AML, but not MNCs, glutamine flux is insufficient to overcome the truncated TCA, resulting in preferential depletion of key TCA metabolites and cell death.

Indirect CII inhibition, as a result of off-target drug effects (22) or posttranslational modifications (20) results in AML cell death via an unknown mechanism. Although not yet assessed in AML, direct pharmacologic inhibition of CII produces antitumor effects in preclinical models of solid tumors (27, 45, 46). Located at the intersection of the ETC and TCA, CII contributes to both bioenergetic (i.e., ATP generation) and biosynthetic (i.e., generation of TCA intermediates) pathways, both of which are required for cancer cell proliferation (47). Among all the ETC complexes, CII produces ATP at the most favorable rate (48); and, thus it is speculated that increased CII activity supports AML dependence on OXPHOS (3, 12, 13). In line with this, the present findings indicate that genetic impairment of CII via SDHAF1 knockdown significantly disrupts mitochondrial respiration and ATP synthesis leading to reduced AML cell growth in vitro and in vivo. Moreover, pharmacologic inhibition of CII, using a newly identified small molecule that specifically binds to CII, recapitulated the downstream metabolic effects of SDHAF1 knockdown (depleted aspartate) and resulted in selective AML cell death. Together, direct inhibition or loss of CII function imparts selective anti-AML activity.

Drug-induced cell stress requires metabolic flexibility to sustain viability, providing a therapeutic window to selectively target malignant cells. For example, normal hematopoietic, but not AML, cells demonstrate a sufficient capacity to switch to compensatory glycolysis upon mitochondrial impairment (4, 19). Similarly, in this study, MNCs but not AML cells sustain levels of the key metabolite, aspartate by using glutamine anaplerosis when oxidative metabolism of glucose is impaired by CII inhibition. Under basal conditions, glutamine is critical to maintaining AML metabolism and proliferation, and various strategies to impair glutamine metabolism are preclinically and clinically (NCT02071927) investigated (6, 49). Here, CII inhibition demonstrates a further reliance on glutamine metabolism, particularly through reductive carboxylation to sustain aspartate levels and support normal cell viability. Aspartate synthesis is supported by fumarate, malate, and citrate and is regulated by several enzymes, including those that comprise the malate-aspartate shuttle (50, 51). Thus, while differences in malate and fumarate flux among SDHAF1 knockdown and shikonin treatment are noted, we consistently observe an increase in reductive flux to the key metabolite, aspartate, suggesting that these cell types may differentially use precursors to synthesize aspartate. Reductive carboxylation is the counterclockwise TCA cycle driven by the reversible enzyme isocitrate dehydrogenase (IDH), that restores TCA intermediates when OXPHOS is impaired (29). Pathway activation in solid tumors is observed following ETC impairment (29, 41, 52), hypoxia (52), oxidative TCA metabolism reactions proceeding SDH activity (42), high extracellular lactate (53), and low NAD+:NADH ratios (50). Moreover, our data show that CII inhibition impairs TCA oxidation at SDH and demonstrates how this provides the necessary conditions to promote reductive carboxylation to sustain TCA metabolite levels.

Another significant finding of this study is the distinction between AML and MNCs in their capacity to derive aspartate from glutamine anaplerosis. A pivotal role of the ETC is to sustain cell proliferation by supporting aspartate production; restoring aspartate levels protects cells from ETC impairment (54, 55). Here, CII inhibition constrains aspartate biosynthesis, causing a reliance on glutamine anaplerosis to sustain its levels. Significantly, while AML cells increase glutamine flux into the TCA cycle upon CII inhibition, this flux is insufficient to compensate for the truncation of oxidative TCA metabolism. The result of this is aspartate depletion and cell death. In contrast, MNCs metabolize glutamine sufficiently to sustain aspartate levels and avert cell death. These findings are echoed by two independent studies. CI inhibition led to AML cell death mediated by aspartate depletion, where glutamine metabolism failed to significantly maintain aspartate levels (18). A second study found that HSCs impressively maintained metabolite levels upon CIII inhibition (56). Thus, through a direct comparison of CII inhibition in AML and MNCs, this study demonstrates the divergent fates of cell types resulting from anaplerotic metabolism of glutamine. Moreover, modulation of glutamine metabolism in AML cells, demonstrates the direct contribution of this pathway to cell survival (6, 57). These results show that glutamine-mediated metabolism is an adaptive response that attenuates the antiproliferative effects of CII inhibition.

CII is a particularly interesting target. CI, but not CII, contributes to the proton gradient and maintenance of NAD+/NADH levels; thus, its inhibition severely interferes with both ATP synthesis and the many upstream metabolic pathways that rely on NAD+ regeneration (58). Thus, relative to CI, CII inhibition may produce fewer off target effects and; therefore, be better tolerated. For example, when comparing phenotypes in those with germline mutations, CI mutations result in the development of lethal childhood Leigh syndrome, while CII mutations typically result in benign neoplasias that develop in middle age (26, 59, 60). In the context of AML, common disease-related mutations may render the leukemia population more susceptible to CII targeting, and merit further investigation. An internal tandem duplication of FMS-like tyrosine kinase 3 (FLT3ITD+), present in 15%–36% of adult patients with AML, results in glutamine dependence upon exposure to tyrosine kinase inhibitors (6, 61). Moreover, mutations of IDH (IDHm), present in 20% of AML, result in a complete loss of glutamine reductive carboxylation capacity (62). As such, patients with IDHm or FLT3ITD+ may be more sensitive to CII targeting, given the reduced metabolic plasticity that creates further reliance on glutamine anaplerosis. Clinical observations reveal that patients with AML harboring IDHm are more sensitive to venetoclax (63), which indirectly inhibits CII (22) and rewires metabolism towards reductive carboxylation of glutamine (64). Together, these findings demonstrate the potential of CII targeting and provide support for further studies to test CII inhibition alone or in combination with venetoclax or tyrosine kinase inhibitors.

In summary, direct inhibition of CII results in selective AML toxicity. Mechanistically, CII inhibition impaired oxidative metabolism and significantly disrupted TCA metabolite biosynthesis in AML cells leading to death. In contrast, normal hematopoietic cell viability remained unaffected by CII inhibition as it could sustain metabolite levels through sufficiently using anaplerotic metabolism of glutamine. This work highlights the role of glutamine anaplerosis as a metabolic adaptation that can be therapeutically targeted by direct CII inhibition.

No disclosures were reported.

A. Roma: Conceptualization, data curation, formal analysis, validation, methodology, writing–original draft, writing–review and editing. M. Tcheng: Data curation, formal analysis, methodology. N. Ahmed: Conceptualization, data curation, writing–review and editing. S. Walker: Data curation. P. Jayanth: Data curation. M.D. Minden: Resources. K. Hope: Resources. P.P. Nekkar Rao: Resources, formal analysis. J. Luc: Software. A.C. Doxey: Resources, formal analysis. J.A. Reisz: Data curation. R. Culp-Hill: Data curation. A. D'Alessandro: Data curation, formal analysis, methodology. P.A. Spagnuolo: Conceptualization, resources, supervision, funding acquisition, validation, investigation, methodology, writing–original draft, project administration, writing–review and editing.

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.

Note: Supplementary data for this article are available at Molecular Cancer Research Online (http://mcr.aacrjournals.org/).

1.
Döhner
H
,
Weisdorf
DJ
,
Bloomfield
CD
.
Acute myeloid leukemia
.
N Engl J Med
2015
;
373
:
1136
52
.
2.
Culp-Hill
R
,
Pietras
EM
,
D'Alessandro
A
.
Extinguishing the embers: targeting AML metabolism
.
Trends Mol Med
2020
;
27
:
332
44
.
3.
Farge
T
,
Saland
E
,
de Toni
F
,
Aroua
N
,
Scotland
S
,
Larrue
C
, et al
.
Chemotherapy resistant human acute myeloid leukemia cells are not enriched for leukemic stem cells but require oxidative metabolism
.
Cancer Discov
2017
;
7
:
716
35
.
4.
Lagadinou
ED
,
Sach
A
,
Callahan
K
,
Rossi
RM
,
Sarah
J
,
Minhajuddin
M
, et al
.
BCL-2 inhibition targets oxidative phosphorylation and selectively eradicates quiescent human leukemia stem cells
.
Cell Stem Cell
2013
;
12
:
329
41
.
5.
Guerra
VA
,
Burger
JA
,
Borthakur
GM
,
Jabbour
E
,
Pemmaraju
N
,
Kadia
TM
, et al
.
Interim analysis of a phase II study of the glutaminase inhibitor telaglenastat (CB-839) in combination with azacitidine in advanced myelodysplastic syndrome (MDS)
.
Blood
2019
;
13
;
134
:
567
.
6.
Gallipoli
P
,
Giotopoulos
G
,
Tzelepis
K
,
Costa
ASH
,
Vohra
S
,
Medina-Perez
P
, et al
.
Glutaminolysis is a metabolic dependency in FLT3 ITD acute myeloid leukemia unmasked by FLT3 tyrosine kinase inhibition
.
Blood
2018
;
31
:
1639
53
.
7.
Cunningham
I
,
Kohno
B
.
18FDG-PET/CT: 21st century approach to leukemic tumors in 124 cases
.
Am J Hematol
2016
;
91
:
379
84
.
8.
Jones
CL
,
Stevens
BM
,
D'Alessandro
A
,
Reisz
JA
,
Culp-Hill
R
,
Nemkov
T
, et al
.
Inhibition of amino acid metabolism selectively targets human leukemia stem cells
.
Cancer Cell
2018
;
34
:
724
40
.
9.
Tcheng
M
,
Roma
A
,
Ahmed
N
,
Smith
RW
,
Jayanth
P
,
Minden
MD
, et al
.
Very long chain fatty acid metabolism is required in acute myeloid leukemia
.
Blood
2021
;
137
:
3518
32
.
10.
Samudio
I
,
Harmancey
R
,
Fiegl
M
,
Kantarjian
H
,
Konopleva
M
,
Korchin
B
, et al
.
Pharmacologic inhibition of fatty acid oxidation sensitizes human leukemia cells to apoptosis induction
.
J Clin Investig
2010
;
120
:
142
56
.
11.
Mirali
S
,
Botham
A
,
Voisin
V
,
Xu
C
,
St-germain
J
,
Sharon
D
, et al
.
The mitochondrial peptidase, neurolysin, regulates respiratory chain supercomplex formation and is necessary for AML viability
.
Sci Transl Med
2020
;
12
:
eaaz8264
.
12.
Škrtić
M
,
Sriskanthadevan
S
,
Jhas
B
,
Gebbia
M
,
Wang
X
,
Wang
Z
, et al
.
Inhibition of mitochondrial translation as a therapeutic strategy for human acute myeloid leukemia
.
Cancer Cell
2011
;
20
:
674
88
.
13.
Sriskanthadevan
S
,
Jeyaraju
DV
,
Chung
TE
,
Prabha
S
,
Xu
W
,
Skrtic
M
, et al
.
AML cells have low spare reserve capacity in their respiratory chain that renders them susceptible to oxidative metabolic stress
.
Blood
2015
;
125
:
2120
30
.
14.
Chandel
NS
.
Evolution of mitochondria as signaling organelles
.
Cell Metab
2015
;
22
:
204
6
.
15.
Rohlena
J
,
Dong
L
,
Neuzil
J
.
Targeting the mitochondrial electron transport chain complexes for the induction of apoptosis and cancer treatment
.
Curr Pharm Biotechnol
2012
;
14
:
377
89
.
16.
Wheaton
WW
,
Weinberg
SE
,
Hamanaka
RB
,
Soberanes
S
,
Sullivan
LB
,
Anso
E
, et al
.
Metformin inhibits mitochondrial complex I of cancer cells to reduce tumorigenesis
.
Elife
2014
;
3
:
e02242
.
17.
Baccelli
I
,
Gareau
Y
,
Lehnertz
B
,
Gingras
S
,
Spinella
J-FF
,
Corneau
S
, et al
.
Mubritinib targets the electron transport chain complex I and reveals the landscape of OXPHOS dependency in acute myeloid leukemia
.
Cancer Cell
2019
;
36
:
84
99
.
18.
Molina
JR
,
Sun
Y
,
Protopopova
M
,
Gera
S
,
Bandi
M
,
Bristow
C
, et al
.
An inhibitor of oxidative phosphorylation exploits cancer vulnerability
.
Nat Med
2018
;
24
:
1036
46
.
19.
Sharon
D
,
Cathelin
S
,
Mirali
S
,
di Trani
JM
,
Yanofsky
DJ
,
Keon
KA
, et al
.
Inhibition of mitochondrial translation overcomes venetoclax resistance in AML through activation of the integrated stress response
.
Sci Transl Med
2019
;
11
:
eaax2863
.
20.
Cole
A
,
Wang
Z
,
Coyaud
E
,
Voisin
V
,
Gronda
M
,
Jitkova
Y
, et al
.
Inhibition of the mitochondrial protease ClpP as a therapeutic strategy for human acute myeloid leukemia
.
Cancer Cell
2015
;
27
:
864
76
.
21.
Jones
CL
,
Stevens
BM
,
D'Alessandro
A
,
Culp-Hill
R
,
Reisz
JA
,
Pei
S
, et al
.
Cysteine depletion targets leukemia stem cells through inhibition of electron transport complex II
.
Blood
2019
;
134
:
389
94
.
22.
Pollyea
DA
,
Stevens
BM
,
Jones
CL
,
Winters
A
,
Pei
S
,
Minhajuddin
M
, et al
.
Venetoclax with azacitidine disrupts energy metabolism and targets leukemia stem cells in patients with acute myeloid leukemia
.
Nat Med
2018
;
24
:
1859
66
.
23.
Warner
JK
,
Wang
JCY
,
Takenaka
K
,
Doulatov
S
,
McKenzie
JL
,
Harrington
L
, et al
.
Direct evidence for cooperating genetic events in the leukemic transformation of normal human hematopoietic cells
.
Leukemia
2005
;
19
:
1794
805
.
24.
Jones
CL
,
Stevens
BM
,
Pollyea
DA
,
Culp-hill
R
,
Reisz
JA
,
Nemkov
T
, et al
.
Nicotinamide metabolism mediates resistance to venetoclax in relapsed acute myeloid leukemia stem cells
.
Cell Stem Cell
2020
;
27
:
748
64
.
25.
Lee
MH
,
Lapidus
RG
,
Ferraris
D
,
Emadi
A
.
Analysis of the mechanisms of action of naphthoquinone-based anti-acute myeloid leukemia chemotherapeutics
.
Molecules
2019
;
24
:
3121
.
26.
Maio
N
,
Ghezzi
D
,
Verrigni
D
,
Rizza
T
,
Bertini
E
,
Martinelli
D
, et al
.
Disease-causing SDHAF1 mutations impair transfer of Fe-S clusters to SDHB
.
Cell Metab
2016
;
23
:
292
302
.
27.
Guo
L
,
Shestov
AA
,
Worth
AJ
,
Nath
K
,
Nelson
DS
,
Leeper
DB
, et al
.
Inhibition of mitochondrial complex II by the anticancer agent lonidamine
.
J Biol Chem
2016
;
291
:
42
57
.
28.
Smestad
J
,
Hamidi
O
,
Wang
L
,
Nelson Holte
M
,
Al Khazal
F
,
Erber
L
, et al
.
Characterization and metabolic synthetic lethal testing in a new model of SDH-loss familial pheochromocytoma and paraganglioma
.
Oncotarget
2017
;
9
:
6109
27
.
29.
Mullen
AR
,
Wheaton
WW
,
Jin
ES
,
Chen
PH
,
Sullivan
LB
,
Cheng
T
, et al
.
Reductive carboxylation supports growth in tumour cells with defective mitochondria
.
Nature
2012
;
481
:
385
8
.
30.
Garcia-Bermudez
J
,
Baudrier
L
,
La
K
,
Zhu
XG
,
Fidelin
J
,
Sviderskiy
VO
, et al
.
Aspartate is a limiting metabolite for cancer cell proliferation under hypoxia and in tumors
.
Nat Cell Biol
2018
;
20
:
775
81
.
31.
Sullivan
LB
,
Luengo
A
,
Danai
LV
,
Bush
LN
,
Diehl
FF
,
Hosios
AM
, et al
.
Aspartate is an endogenous metabolic limitation for tumour growth
.
Nat Cell Biol
2018
;
20
:
782
8
.
32.
Kluckova
K
,
Bezawork-Geleta
A
,
Rohlena
J
,
Dong
L
,
Neuzil
J
.
Mitochondrial complex II, a novel target for anti-cancer agents
.
Biochim Biophys Acta
2013
;
1827
:
552
64
.
33.
Ibrahim
M
,
Mendis
N
,
Roma
A
,
Jayanth
P
,
Minden
MD
,
Spagnuolo
PA
.
Anti-leukemia effect of vitamin K3 is mediated through mitochondrial complexes I and II
.
Leuk Lymphoma
2019
;
60
:
3071
4
.
34.
Auber
RP
,
Suttiyut
T
,
Mccoy
RM
,
Ghaste
M
,
Crook
JW
,
Pendleton
AL
, et al
.
Hybrid de novo genome assembly of red gromwell (Lithospermum erythrorhizon) reveals evolutionary insight into shikonin biosynthesis
.
Hortic Res
2020
;
7
:
82
.
35.
Lipinski
CA
.
Rule of five in 2015 and beyond: target and ligand structural limitations, ligand chemistry structure and drug discovery project decisions
.
Adv Drug Deliv Rev
2016
;
101
:
34
41
.
36.
Boulos
JC
,
Rahama
M
,
Hegazy
MF
,
Efferth
T
.
Shikonin derivatives for cancer prevention and therapy
.
Cancer Lett
2019
;
459
:
248
67
.
37.
Wiench
B
,
Eichhorn
T
,
Paulsen
M
,
Efferth
T
.
Shikonin directly targets mitochondria and causes mitochondrial dysfunction in cancer cells
.
Evid Based Complement Altern Med
2012
;
2012
:
726025
.
38.
Yang
J
,
Li
ZL
,
Wu
J
,
Lu
FJ
,
Chen
C
.
An oxidative stress mechanism of shikonin in human glioma cells
.
PLoS One
2014
;
9
:
e94180
.
39.
Eruygur
N
.
A simple isocratic high-perfomance liquid chromatography method for the simultaneous determination of shikonin derivatives in some Echium species growing wild in Turkey
.
Turk J Pharm Sci
2018
;
15
:
38
43
.
40.
Dröse
S
.
Differential effects of complex II on mitochondrial ROS production and their relation to cardioprotective pre- and postconditioning
.
Biochim Biophys Acta
2013
;
1827
:
578
87
.
41.
Chen
Q
,
Kirk
K
,
Shurubor
YI
,
Zhao
D
,
Arreguin
AJ
,
Shahi
I
, et al
.
Rewiring of glutamine metabolism is a bioenergetic adaptation of human cells with mitochondrial DNA mutations
.
Cell Metab
2018
;
27
:
1007
25
.
42.
Mullen
AR
,
Hu
Z
,
Shi
X
,
Jiang
L
,
Boroughs
LK
,
Kovacs
Z
, et al
.
Oxidation of alpha-ketoglutarate is required for reductive carboxylation in cancer cells with mitochondrial defects
.
Cell Rep
2014
;
7
:
1679
90
.
43.
Wang
WJ
,
Yi
MG
,
Zhu
XY
.
[A study on absorption, distribution and excretion of 3H-shikonin in mice]
.
Yao Xue Xue Bao
1988
;
23
:
246
51
.
44.
Mirali
S
,
Schimmer
AD
.
The role of mitochondrial proteases in leukemic cells and leukemic stem cells
.
Stem Cells Transl Med
2020
;
9
:
1481
7
.
45.
Dong
LF
,
Freeman
R
,
Liu
J
,
Zobaiova
R
,
Marin-Hernandez
A
,
Stantic
M
, et al
.
Suppression of tumor growth in vivo by the mitocan α-tocopheryl succinate requires respiratory complex II
.
Clin Cancer Res
2009
;
15
:
1593
600
.
46.
Yan
B
,
Stantic
M
,
Zobalova
R
,
Bezawork-Geleta
A
,
Stapelberg
M
,
Stursa
J
, et al
.
Mitochondrially targeted vitamin E succinate efficiently kills breast tumour-initiating cells in a complex II-dependent manner
.
BMC Cancer
2015
;
15
:
401
.
47.
DeBerardinis
RJ
,
Chandel
NS
.
Fundamentals of cancer metabolism
.
Sci Adv
2016
;
2
:
e1600200
.
48.
Ralph
SJ
,
Moreno-Sánchez
R
,
Neuzil
J
,
Rodríguez-Enríquez
S
.
Inhibitors of succinate: quinone reductase/complex II regulate production of mitochondrial reactive oxygen species and protect normal cells from ischemic damage but induce specific cancer cell death
.
Pharm Res
2011
;
28
:
2695
730
.
49.
Jacque
N
,
Ronchetti
AM
,
Larrue
C
,
Meunier
G
,
Birsen
R
,
Willems
L
, et al
.
Targeting glutaminolysis has antileukemic activity in acute myeloid leukemia and synergizes with BCL-2 inhibition
.
Blood
2015
;
126
:
1346
56
.
50.
Gaude
E
,
Schmidt
C
,
Gammage
PA
,
Dugourd
A
,
Blacker
T
,
Chew
SP
, et al
.
NADH shuttling couples cytosolic reductive carboxylation of glutamine with glycolysis in cells with mitochondrial dysfunction
.
Mol Cell
2018
;
69
:
581
93
.
51.
Lussey-Lepoutre
C
,
Hollinshead
KER
,
Ludwig
C
,
Menara
M
,
Morin
A
,
Castro-Vega
LJ
, et al
.
Loss of succinate dehydrogenase activity results in dependency on pyruvate carboxylation for cellular anabolism
.
Nat Commun
2015
;
6
:
8784
.
52.
Fendt
SM
,
Bell
EL
,
Keibler
MA
,
Olenchock
BA
,
Mayers
JR
,
Wasylenko
TM
, et al
.
Reductive glutamine metabolism is a function of the α-ketoglutarate to citrate ratio in cells
.
Nat Commun
2013
;
4
:
2236
.
53.
Brodsky
AN
,
Odenwelder
DC
,
Harcum
SW
.
High extracellular lactate causes reductive carboxylation in breast tissue cell lines grown under normoxic conditions
.
PLoS One
2019
;
14
:
e0213419
.
54.
Birsoy
K
,
Wang
T
,
Chen
WW
,
Freinkman
E
,
Abu-Remaileh
M
,
Sabatini
DM
.
An essential role of the mitochondrial electron transport chain in cell proliferation is to enable aspartate synthesis
.
Cell
2015
;
162
:
540
51
.
55.
Sullivan
LB
,
Gui
DY
,
Hosios
AM
,
Bush
LN
,
Freinkman
E
,
vander Heiden
MG
.
Supporting aspartate biosynthesis is an essential function of respiration in proliferating cells
.
Cell
2015
;
162
:
552
63
.
56.
Ansó
E
,
Weinberg
SE
,
Diebold
LP
,
Thompson
BJ
,
Malinge
S
,
Schumacker
PT
, et al
.
The mitochondrial respiratory chain is essential for haematopoietic stem cell function
.
Nat Cell Biol
2017
;
19
:
614
25
.
57.
Gregory
MA
,
Nemkov
T
,
Park
HJ
,
Zaberezhnyy
V
,
Gehrke
S
,
Adane
B
, et al
.
Targeting glutamine metabolism and redox state for leukemia therapy
.
Clin. Cancer Res
2019
;
25
:
4079
90
.
58.
Sharma
LK
,
Lu
J
,
Bai
Y
.
Mitochondrial respiratory complex I: structure, function and implication in human diseases
.
Curr Med Chem
2009
;
16
:
1266
77
.
59.
Mimaki
M
,
Wang
X
,
McKenzie
M
,
Thorburn
DR
,
Ryan
MT
.
Understanding mitochondrial complex I assembly in health and disease
.
Biochim Biophys Acta
2012
;
1817
:
851
62
.
60.
Ghezzi
D
,
Zeviani
M
.
Human diseases associated with defects in assembly of OXPHOS complexes
.
Essays Biochem
2018
;
62
:
271
86
.
61.
Levis
M
.
FLT3 mutations in acute myeloid leukemia: what is the best approach in 2013?
Hematology Am Soc Hematol Educ Program
2013
;
2013
:
220
6
.
62.
Leonardi
R
,
Subramanian
C
,
Jackowski
S
,
Rock
CO
.
Cancer-associated isocitrate dehydrogenase mutations inactivate NADPH-dependent reductive carboxylation
.
J Biol Chem
2012
;
287
:
14615
20
.
63.
Konopleva
M
,
Pollyea
DA
,
Potluri
J
,
Chyla
B
,
Hogdal
L
,
Busman
T
, et al
.
Efficacy and biological correlates of response in a phase II study of venetoclax monotherapy in patients with acute myelogenous leukemia
.
Cancer Discov
2016
;
6
:
1106
17
.
64.
Roca-Portoles
A
,
Rodriguez-Blanco
G
,
Sumpton
D
,
Cloix
C
,
Mullin
M
,
Mackay
GM
, et al
.
Venetoclax causes metabolic reprogramming independent of BCL-2 inhibition
.
Cell Death Dis
2020
;
11
:
616
.

Supplementary data