A Phase I Study of the First-in-Class Antimitochondrial Metabolism Agent, CPI-613, in Patients with Advanced Hematologic Malignancies Free

This article is featured in Highlights of This Issue, p. 5145

Cancer cells markedly alter the flow of carbon away from complete oxidation and adenosine triphosphate (ATP) production toward glycolysis and anabolism (as reviewed by Vander Heiden and colleagues; ref. 1). This altered carbon metabolism was first described by Warburg and colleagues (2) and is now considered an emerging hallmark of cancer (3). Increased glycolysis and glucose uptake are widely used markers for PET imaging in oncology. Despite the nearly universal findings of altered metabolism in cancer, therapeutic targeting of metabolism, especially in hematologic malignancies, has been understudied.

Aerobic glycolysis utilized by cancer cells increases lactate production and decreases carbon flow into the mitochondria via the pyruvate dehydrogenase complex (PDH). As opposed to Warburg and colleagues' initial hypothesis of defective, nonfunctional mitochondria in cancer, additional studies have shown that cancer cells depend on mitochondrial metabolism for anabolic intermediates, with 10% to 15% of glucose carbons entering the tricarboxylic acid cycle (TCA) cycle (1). This distribution is thought to represent a balance between the need for anabolic precursors, reducing equivalents in the form of NADH and NADPH and ATP to maximize cell growth. As a result, despite the fact that approximately 85% of glucose carbon is excreted as lactate, the generation of acetyl-CoA from pyruvate remains critical to cancer growth. PDH catalyzes the oxidative decarboxylation of pyruvate to acetyl-CoA and is essential for converting glucose carbons into lipids and other TCA cycle intermediates (Fig. 1). In addition, cancer cells can utilize glutamine via its conversion to α-ketoglutarate that enters the TCA cycle via α-ketoglutarate dehydrogenase complex (KDH; Fig. 1). Inhibition of both PDH and KDH would be predicted to deprive cancer cells of their ability to metabolize carbon via the TCA cycle.

Lipoate is an essential cofactor for several mitochondrial enzymes that are key to maintaining TCA cycle intermediates, including PDH and KDH. CPI-613 is a novel lipoate derivative that inhibits PDH and KDH (4, 5). It causes hyperphosphorylation of the E1α subunit of PDH, ATP depletion, and cell death in malignant but not normal cells (4). Following confirmation of its novel mechanism of action in leukemia cells, we tested its safety in a phase I trial for patients with relapsed or refractory hematologic malignancies.

K562 cells were purchased from the ATCC and OCI-AML3 were the kind gift of Dr Mark Minden at University of Toronto. All human cell lines were maintained in RPMI media (Gibco) supplemented with 10% FBS, penicillin, and streptomycin. Cells were grown at 37°C with 5% CO₂.

All oxygen consumption assays were performed using the XF24 Extracellular Flux Analyzer (Seahorse Bioscience) as per the manufacturer's instructions. Cells were placed in assay media with the indicated amount of CPI-613 and incubated for 2 hours before being placed in the assay. Oligomycin at a final concentration of 1 μmol/L, FCCP at 0.5 μmol/L, and actinomycin A and rotenone each at 1 μmol/L were used. All assays were done in triplicate and repeated in three independent experiments.

Samples were lysed in Laemmli buffer, separated by SDS-PAGE, and transferred to an Immobilon polyvinylidene fluoride membrane (Millipore). Antibodies against adenosine monophosphate-activated protein kinase (AMPK; Cell Signaling Technology, 2532S), phosphorylated AMPK (pAMPK; Cell Signaling Technology, 2532S), Phosphorylated E1α (Cal Biochem, AP1062), and actin (Abcam, ab6276) were used.

Patients were eligible if they were age 18 or older, had a relapsed or refractory hematologic malignancy having failed or refused standard salvage therapy. All patients were required to have a Karnofsky performance status of ≥70, life expectancy ≥3 months, and could not have any of the following: current CNS disease, active ongoing infection, uncontrolled bleeding diathesis, dyspnea on mild to moderate exertion, clinically significant effusions, QTc interval >470 ms, or elevated troponin. All toxicities from previous treatments had to be stable and ≤grade 1. Laboratory values had to be AST and ALT ≤3× upper normal limit (UNL), bilirubin ≤1.5× UNL, serum creatinine ≤2.0 mg/dL, INR <1.5, and albumin >2.5 g/dL.

The protocol was approved by the Wake Forest Health Sciences Institutional Review Board and monitored by the Comprehensive Cancer Center's Safety and Toxicity Review Committee. Written and informed consent was obtained for all subjects in accordance with the Declaration of Helsinki. The trial was registered at clinicaltrials.gov (Identifier: NCT01034475).

CPI-613 was administered by central line over 2 hours on days 1 and 4 of each week for 3 consecutive weeks followed by a week of rest. The trial used a two-stage dose-escalation scheme (single-patient and traditional stages, 1-3-6) as described below.

A single patient was accrued per dose level, starting at 420 mg/m². Doses were doubled until a toxicity occurred that was attributable to CPI-613 and >grade 1. Thereafter the traditional dose-escalation stage was triggered.

All dose escalations were done according to the modified Fibonacci dose-escalation scheme. The number of patients in each cohort was 3, if no dose-limiting toxicities (DLT) occurred (toxicity ≥ grade 3 probably or definitely related to CPI-613 that occurred within 28 days of the first dose), dose escalation continued. Once a DLT occurred, the cohort was expanded up to a maximum of 6 patients. If there was no DLT in any other patient out of 6 patients, dose escalation continued. If a DLT occurred in 2 patients in the same cohort, CPI-613 dosing stopped immediately and that dose was considered above the MTD.

Only 1 cycle of therapy was planned, but patients could continue to receive CPI-613 if the treating physician deemed it beneficial.

Toxicity was assessed using the NCI Common Terminology Criteria for Adverse Events (CTCAE), Version 3.0.

Intrapatient dose escalation was allowed provided all of the following conditions were met: No toxicity before the next treatment cycle began; the patient still met all eligibility criteria; and the treating physician believed additional benefit could accrue from dose escalation.

Response was assessed by standard criteria for the respective malignancy. Response and progression for patients with lymphoma with evaluable disease were assessed using the criteria proposed by the International Harmonization Project (6). Cutaneous T-cell lymphoma was evaluated as per published criteria (7). Patients with multiple myeloma were assessed according to the International Uniform Response Criteria (8). Published criteria were used to assess patients with myelodysplastic syndrome (MDS) (9) and acute myelogenous leukemia (AML; ref. 10).

Blood samples were collected in 4-mL lavender-top collection tubes as above and subjected to Ficoll gradient sedimentation. The mononuclear cell phase was gently collected, washed with PBS, resuspended in RPMI-1640 media with 10% DMSO, and stored at −80°C until use.

Plasma was assayed for CPI-613 using liquid chromatography/mass spectroscopy/mass spectroscopy. From the assayed results, pharmacokinetics of plasma CPI-613 levels were derived using PK Solutions 2.0 software (Pharmacokinetics and Metabolism Software). Differences in the mean values among various groups were compared using ANOVA statistical modeling.

Expression microarray analysis was performed at the Comprehensive Cancer Center of Wake Forest University Microarray Shared Resource (Winston-Salem, NC). Total RNA was isolated from peripheral blood mononuclear cells (PBMC) using RiboPure-Blood Kit (Life Technologies; AM1928). RNA was purified on RNeasy columns (Qiagen) and measured for mass and integrity on an Agilent Bioanalyzer. RNA was amplified and labeled according to the Affymetrix 3′-IVT labeling protocol. cRNA was hybridized to the Affymetrix Gene Atlas U219 Human Genome array strips. Arrays were washed and scanned according to the manufacturer's instructions. The data have been deposited with the Gene Expression Omnibus (accession number GSE60350).

For the gene expression analysis, differentially expressed genes were identified by univariate t test (two-sided, unequal variance) and adjusted for multiple testing by the method of Benjamini and Hochberg (11). Adjusted P values (q values) were set at a significance threshold of 0.05 for downstream analyses. Statistical enrichment of gene ontology (GO) terms was investigated using the DAVID (Database for Annotation, Visualization, and Integrated Discovery) microarray resource (12, 13). For the mitochondrial function assays, means of oxygen consumption rates were compared by a one-way ANOVA analysis. Individual means were then compared by Tukey test. A P value of less than or equal to 0.05 was considered significant. All error bars shown are SD. Analysis was done using GraphPad Prism version 5.

As mentioned in the introduction, CPI-613 leads to increased phosphorylation of the E1 subunit of PDH and inhibits KDH in solid tumor cell lines (4, 5). To determine whether CPI-613 directly inhibits mitochondrial metabolism in hematogic malignancies, we treated the human leukemia cell lines K562 and OCI-AML3 with CPI-613 and assessed mitochondrial respiration. CPI-613 exposure resulted in inhibition of maximal mitochondrial respiration in both cell lines (Fig. 2A). To confirm that CPI-613 caused PDH E1 phosphorylation, we harvested cells following the same CPI-613 exposures and performed Western blot analyses. CPI-613 caused an increase in PDH phosphorylation in a dose-dependent manner (Fig. 2B). Finally, to confirm that CPI-613 depletes cellular ATP, we assessed the phosphorylation status of AMPK. AMPK is phosphorylated in response to high AMP/ADP ratios in cells as a consequence of ATP depletion. AMPK became phosphorylated in OCI-AML3 cells following CPI-613 exposure (Fig. 2C). These data demonstrate that CPI-613 targets mitochondrial metabolism in leukemia cells.

To determine the MTD and safety of CPI-613 in patients with hematologic malignancies, we conducted a phase I study. CPI-613 was infused over 2 hours via a central line on days 1 and 4 of each week for 3 consecutive weeks with one week of rest (28-day cycles). Central line infusion was chosen as preclinical and phase I studies in solid tumor patients demonstrated pain and inflammation when CPI-613 was given via a peripheral vein. The primary endpoint of the study was to determine the safety and MTD of CPI-613 when administered twice weekly for 3 consecutive weeks. The secondary endpoints were to determine the pharmacokinetics of CPI-613 following i.v. administration and observe the anticancer effects of CPI-613, if any occur. A total of 26 patients were enrolled on trial. Patient demographics and diagnosis are summarized in Table 1.

After one patient received a dose of 840 mg/m², a grade 2 toxicity occurred, triggering the standard dose-escalation stage. After four dose escalations (to 2,940 mg/m²) with no DLTs observed, the protocol was amended to administer 3,000 mg/m² of CPI-613 over one hour. This was done to increase the C_max of CPI-613 while maintaining the total dose, and further dose escalations were to be done by increasing the infusion time. However, 2 of 3 patients in this cohort developed grade 3 acute renal failure, and the infusion time was then returned to 2 hours. At that time, a patient from the 2,100 mg/m² cohort who was intrapatient dose escalated to 2,940 mg/m² experienced a grade 3 renal failure and an additional 3 patients were dosed at 2,940 mg/m² infused over 2 hours. When no additional DLTs were observed at the 2,940 mg/m² dose level, the dose was escalated to 3,780 mg/m². The third patient in this cohort experienced a DLT (grade 3 vomiting) and the first patient in the expansion cohort (fourth patient at this dose level) experienced a DLT (grade 3 acute renal failure), defining this dose as above the MTD. As a total of 6 patients were treated at the 2,940 mg/m² dose, this was established as the MTD.

Five patients did not complete the study. One refused, one acquired an infection, and three developed acute kidney failure. Toxicities were otherwise minimal. All toxicities more than or equal to grade 2 attributed as probable or definitely related to CPI-613 are summarized in Table 2. All patients who received at least one dose of drug were considered evaluable for toxicity and are included in Table 2. All 3 patients who developed renal failure during the first cycle are included in Table 2 as is the patient who experienced renal failure after dose escalation to 2,940 mg/m². Two patients died during the study, attributed to progression of underlying disease. One patient with high-grade MDS received one dose of CPI-613 given over 1 hour, as described above, developed acute renal failure, and received no additional therapy. Several weeks later, he progressed to AML that was not attributed to CPI-613. No cases of worsening cytopenias were attributed as probably or definitely related to CPI-613, and cytopenias improved on CPI-613 in several patients who did not meet criteria for a response.

Of the 21 patients who completed the trial, 4 achieved an objective response and 2 others achieved prolonged stabilization of disease for a clinical benefit rate of 29%. In addition, 2 patients with multiple myeloma met criteria for stable disease and 10 patients (48%) received more than one cycle (Supplementary Table S1). Stable disease in the 2 patients with multiple myeloma was sustained for two and four cycles, respectively. Both patients left study for disease progression as documented by a new plasmacytoma in one patient and an increase in M protein in the other. A response was seen in all 3 evaluable patients with MDS, one CR as detailed below and the other two achieved prolonged stable disease sustained for 8 and 12 cycles, respectively. One patient with RAEB-2, [high risk by Revised International Prognostic Scoring System (IPSS-R)] showed regression to RAEB-1 after five cycles of CPI-613. After her eighth cycle, the patient chose to leave the study for hospice after she suffered a mechanical fall and could no longer ambulate. The other MDS patient had refractory anemia with 5q and 7q minus by FISH (high risk by IPSS-R), discontinued CPI-613 following her 12th cycle after an increase in her transfusion requirements. One patient with AML had stable disease that lasted four cycles but because this is not recognized as a response in AML, this patient was not counted as a responder. The four responses better than stable disease are detailed below.

Patient #4 is a 51-year-old male diagnosed with relapsed normal karyotype AML who underwent autologous hematopoietic stem cell transplant (HSCT) while in CR2. Following transplant, he remained transfusion dependent for both red blood cells and platelets. Three months posttransplant, a bone marrow biopsy demonstrated a hypocellular marrow, no increase in blasts, and 8.5% cells 7q- by FISH. With continued transfusion dependence and cytogenetics characteristic of MDS, he met criteria for MDS unclassified and given his history of previous systemic chemotherapy, this was consistent with therapy-related MDS. At 6 months posttransplant, repeat marrow revealed no increase in blasts but expansion of the 7q- clone to 41% cells. His IPSS-R category calculated at this time was high risk with a predicted overall survival of 1.6 years (14). Following his first cycle of treatment, he became transfusion independent. Following eight cycles, he met all criteria for CR3 and has remained on drug for 41 cycles with a sustained pathologic and hematologic complete remission. His 7q minus clone has waxed and waned with a best response of 25% after 29 cycles.

Patient #12 is a 66-year-old female with refractory AML. She received induction with 7+3, then cytoxan-etoposide, then high-dose cytarabine + mitoxantrone + L-asparaginase, then decitabine, and finally azacytidine. Pretrial marrow showed a hypocellular marrow with 9% blasts and 22% immature monocytes (31% blast equivalents). After two cycles of CPI-613, a repeat biopsy revealed a hypercellular marrow with no evidence of disease (Morphologically Leukemia Free State). Before count recovery, she was removed from study to undergo a reduced intensity conditioned allogeneic HSCT. She remains alive with no evidence of leukemia.

Patient #14 is a 19-year-old female who was diagnosed with Burkitt lymphoma in April 2010. She was treated with chemo-immunotherapy and achieved a remission. She relapsed in December 2010 and after two cycles of salvage Hyper CVAD, received a myeloablative allogeneic HSCT. She suffered her second relapse in September 2011 and enrolled on trial. She achieved a radiographic partial response after three cycles of therapy that was maintained for 17 cycles (Fig. 3). She ultimately went off study to have her residual disease resected and is currently without evidence of disease off all therapy more than 12 months later. The surgical specimen revealed Burkitt lymphoma with extensive necrosis.

Patient #20 is a 52-year-old female diagnosed with CTCL in May 2008. She was treated with oral methotrexate, bexarotene, vorinostat, denileukin diftitox, gemcitabine, pralatrexate, romidepsin, bortezomib, romidepsin again, and finally bexarotene again before starting CPI-613. She achieved a partial response sustained now for 25 cycles (Fig. 4). After three cycles, she requested that the week of rest be omitted as she was seeing disease flares during week 4 when no treatment was administered. She has since received continuous therapy without significant toxicities and no evidence of marrow suppression.

In most patients (23 of 26), there was a rapid decline in plasma levels during the first 8 to 10 hours postadministration, followed by low plasma CPI-613 values persisting for several days. Erratic results from 3 patients, comprising the 420 and 840 mg/m² dose groups, were not included in the analyses. The concentration-time curves associated with all dose levels of CPI-613 were triphasic. The α-phase occurred during the first 6 to 8 hours postadministration and was associated with a T_1/2 of approximately 1.3 hours. The β-phase occurred from 6 to 8 hours to 24 hours postadministration and was associated with a modest increase in CPI-613 values. The γ or final phase started from 24 hours until 72 hours, and was associated with a slow decline in CPI-613 values, consistent with its high protein-binding properties (Fig. 5A). CPI-613 was not detected before the first dose, but baseline levels associated with dose 6 had an overall average value of approximately 3 μmol/L. The small amount of residual CPI-613 in the plasma lacked significant drug accumulation effects, as concentration-time curves associated with dose 1 were comparable with dose 6 (Fig. 5B). As expected in the 3 patients who received CPI-613 over one hour, the C_max was significantly higher than in patients given drug over a 2-hour infusion (Supplementary Fig. S1).

To explore the possibility of an effect of CPI-613 on the immune system of patients following treatment, we performed an analysis of gene expression profiles of PBMCs. We selected the 4 patients with myeloid malignancies who responded to CPI-613 (1 with AML and 3 with MDS) and compared them with 4 patients who did not respond (all with AML). PBMCs were obtained on day 1 of week 4 of cycle 1, after all 6 doses of CPI-613 were administered. Many genes showed differential expression (Supplementary Fig. S2). Because no genes achieved statistical significance (q < 0.05), a GO enrichment analysis was performed. Only the genes overexpressed in responders resulted in the identification of GO biologic process terms after false discovery correction. The top pathways upregulated in responders involved immune regulation, including T-cell activation, leukocyte activation, and regulation of cytokine production (Supplementary Table S2). Multiple upregulated genes (e.g., IFNG, CCL5, CXCR3, CD8A, CD3E, and CD3D) are commonly associated with a TH-1 type immune response (15).

To our knowledge, this is the first report of clinical activity of any agent that targets mitochondrial metabolism in patients with hematologic malignancies. CPI-613 is a lipoate analog that inhibits the mitochondrial enzymes PDH and KDH. When human leukemia cells were exposed to CPI-613, there was a dose-dependent decrease in mitochondrial respiration. It also led to increased phosphorylation of PDH and AMPK consistent with inhibition of PDH and energy depletion as would be predicted from the proposed mechanism of action. A more comprehensive study of the metabolic effects of CPI-613 as well as pharmacodynamic analysis of primary patient samples from an ongoing clinical trial is currently underway.

In the phase I study, a total of 26 patients with relapsed or refractory hematologic malignancies were enrolled on trial and 21 were evaluable for response. Of these, 4 patients achieved an objective response and 2 MDS patients achieved prolonged stabilization of disease for a clinical benefit rate of 29%. These patients were heavily pretreated with a median of three previous lines of therapy and older with a median age of 65 years. The MTD of CPI-613 given twice weekly over 2 hours was determined to be 2,940 mg/m².

CPI-613 was well tolerated with the DLTs of vomiting and acute renal failure. The mechanism behind the acute renal failure is unknown at this time. The clinical and laboratory data were consistent with acute tubular necrosis. In the patients who developed renal failure, there was no consistent underlying malignancy as 2 patients had MDS, one had NHL, and one had AML. All 4 patients were 65 years of age or older at the time of renal failure. The renal failure was reversible in 3 of the 4 patients with one patient opting for hospice before resolution could be documented. The renal failure was managed conservatively in 2 of the 3 patients not opting for hospice with one patient receiving a single hemodialysis treatment. The renal failure duration ranged between 2 weeks and 3 weeks. In 2 patients, renal function returned to baseline and in one patient, his creatinine remained slightly elevated (1.7–1.8 mg/dL). All current studies have mandated a baseline creatinine of less than 1.5 mg/dL. Aside from renal failure and one case of vomiting, no other clinically significant toxicities were observed. No patient had a worsening of any cytopenias attributed as probably or definitely related to CPI-613. Indeed, cytopenias improved in several patients with AML who did not meet criteria for a response. Both patients who have received extended dosing of CPI-613 have not demonstrated any evidence of marrow toxicity after almost 2 years and over 3 years on drug, respectively.

All 3 evaluable patients with MDS achieved a response of stable disease or better, suggesting efficacy of CPI-613 against this malignancy. The responses were reminiscent of those seen with the hypomethylating agents (16). The possibility of epigenetic alterations as a mechanism of CPI-613 is supported by the crosstalk between mitochondrial metabolism and DNA methylation (reviewed by Martinez-Pastor and colleagues; ref. 17). DNA methylation requires S-adenosylmethionine (SAM), produced from methionine in an ATP-dependent reaction. ATP depletion is a consequence of CPI-613 exposure (Fig. 2A–C) and should deplete cells of SAM as well. Reactions that lead to DNA demethylation utilize α-ketoglutarate (18). CPI-613 inhibition of KDH will result in α-ketoglutarate accumulation that can drive the demethylation reaction. The possibility of CPI-613-induced changes in DNA methylation will need to be directly confirmed and a phase II study of CPI-613 as a monotherapy in patients with MDS having failed first-line therapy is currently accruing patients.

The immune signature seen in the PBMCs of patients who responded to CPI-613 was consistent with a T_H1 type response. This initial exploratory analysis is suggestive of a role of CPI-613 in increasing the immunogenicity of the tumor cells. Confirming this mechanism is the subject of current ongoing experiments.

Several studies have shown that stimulation of PDH activity causes impairments in cancer cell growth and tumorigenesis (19, 20). In addition, treatment with the small-molecule dichloroacetic acid that increases PDH function resulted in several responses in patients with glioblastoma and a case report of a response in a patient with diffuse large B-cell lymphoma (21, 22). These data have been interpreted to support the view that mitochondrial function is disadvantageous. Contrary to this is the fact that cancer cells depleted of mitochondria display impaired growth and invasiveness and mitochondrial metabolism is required for oncogenic K-Ras transformation (23, 24). In addition, there are confirmed cases of mitochondrial enzyme mutations causing cancer, including mutations in IDH2 in AML (25). There is evidence that some malignancies generate much of their cellular ATP via mitochondrial respiration (26) and all de novo lipid synthesis utilizes mitochondrial derived metabolites (27). The ultimate dependence of any one tumor on mitochondrial metabolism varies by tissue and oncogene involved (28). Therefore, it is likely that significant variation in dependence on mitochondrial metabolism exists between tumor types and indeed even between different patients with the same tumor type. Our data support this with several patients even in this heavily pretreated cohort having dramatic responses. Additional studies to better characterize those tumors most likely to respond are ongoing.

In summary, this reports details the results of a phase I trial of the first-in-class, first-in-man mitochondrial metabolism inhibitor CPI-613. The data demonstrate that agents that target cancer cell mitochondrial metabolism can have clinical activity with acceptable toxicity, and support further studies of CPI-613 in hematologic malignancies.

T.S. Pardee is a consultant/advisory board member for and reports receiving speakers bureau honoraria and commercial research support from Cornerstone Pharmaceuticals. K. Lee, J. Luddy, and C. Maturo are employees of Cornerstone Pharmaceuticals. R. Rodriguez is an employee of and has ownership interest (including patents) in Cornerstone Pharmaceuticals. B.L. Powell is a consultant/advisory board member for Cornerstone Pharmaceuticals. No potential conflicts of interest were disclosed by the other authors.

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Cancer Institute or the NIH.

Conception and design: T.S. Pardee, K. Lee, B.L. Powell

Development of methodology: T.S. Pardee, K. Lee, B.L. Powell

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): L.D. Miller, K.M. Stadelman, D. Levitan, D. Hurd, L.R. Ellis, M. Manuel, S. Dralle, S. Lyerly, B.L. Powell

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): T.S. Pardee, J. Luddy, S. Isom, L.D. Miller

Writing, review, and/or revision of the manuscript: T.S. Pardee, K. Lee, J. Luddy, C. Maturo, R. Rodriguez, S. Isom, L.D. Miller, D. Hurd, L.R. Ellis, M. Manuel, S. Lyerly, B.L. Powell

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): T.S. Pardee, K. Lee, S. Isom, B.L. Powell

Study supervision: T.S. Pardee, R. Harrelson, B.L. Powell

Other (supplied study drug): C. Maturo

The authors thank Karen Klein for help in editing the article.

T.S. Pardee was supported by NCI 1K08CA169809, and S. Isom and the Microarray Shared Resource were supported by NCI Cancer Center Support Grant (CCSG) P30CA012197. Additional support was provided by the Doug Coley Foundation for Leukemia Research, the Frances P. Tutwiler Fund, and The MacKay Foundation for Cancer Research. Cornerstone Pharmaceuticals provided CPI-613 and funded and performed the pharmacokinetic studies.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.