A Phase I Study of the Pan Bcl-2 Family Inhibitor Obatoclax Mesylate in Patients with Advanced Hematologic Malignancies

Defects in the apoptosis pathways are an essential element of tumor pathogenesis. One common cancer-causing defect arises from the overexpression of the antiapoptotic protein Bcl-2 and its related family members. Bcl-2 and its antiapoptotic family members inhibit the mitochondrial pathway of apoptosis in part by stabilizing the mitochondria and forming inactivating heterodimers with proapoptotic proteins such as Bax and Bak (1). Overexpression of Bcl-2 and its related family members such as Bcl-x_L and Mcl-1 is associated with chemoresistance in leukemia cell lines (2, 3) and with a poor clinical outcome in studies of adult patients with leukemia (4–7). Therefore, small molecules that inhibit Bcl-2 could be novel therapeutic agents for the treatment of patients with leukemia.

Obatoclax mesylate (GX15-070) is a hydrophobic small molecule that binds to the BH3-binding site of Bcl-2 and the related Bcl-2 family members Bcl-x_L, Mcl-1, Bcl-w, A1, and Bcl-b (8). This pan-Bcl-2 inhibitor directly induces apoptosis in cultured acute myeloid leukemia (AML) cells and primary patient samples. For example, obatoclax mesylate inhibits clonogenic growth of primary AML samples with a mean IC₅₀ of <200 nmol/L. In cultured AML cells, obatoclax mesylate also dissociates Bak and Bim from Mcl-1, in keeping with its function as a Bcl-2 family inhibitor (9). Furthermore, obatoclax mesylate displays antitumor activity in mouse xenografts of solid tumor and myeloma cell lines (8, 10). These encouraging results coupled with its safety in animal studies support the evaluation of obatoclax mesylate in clinical trials for patients with refractory malignancies (8, 10). Here, we report a phase I study of escalating dose and frequency of obatoclax mesylate in patients with refractory leukemia and myelodysplasia.

Eligibility. Patients with histologically or cytologically confirmed refractory hematologic malignancies including AML, myelodysplastic syndrome, acute lymphocytic leukemia (ALL), chronic lymphocytic leukemia (CLL), and chronic myeloid leukemia blast crisis for which standard curative or palliative measures did not exist, were no longer effective, or were not acceptable to the patient were eligible for this trial. There were no limitations on the number of prior therapies, but active toxicities from these prior therapies must have resolved to grade <1 as defined by National Cancer Institute Common Toxicity Criteria version 2.0. Patients were required to be ages ≥18 years with an Eastern Cooperative Oncology Group performance status of <2. To be eligible for the study, patients were required to have normal hepatic and renal function as defined by a total bilirubin of <2 mg/dL unless resulting from hemolysis, aspartate aminotransferase/alanine aminotransferase <2.5 times the institutional upper limit of normal, serum creatinine within normal institutional limits, or a calculated or measured creatinine clearance of >50 mL/min/1.73 m². Women of childbearing potential required a negative serum pregnancy test and men with partners of childbearing potential agreed to abstain from sexual activity, which would result in pregnancy for the duration of the study. The study was approved by the research ethics boards of all participating institutions, and all patients provided written informed consent in accordance with federal and institutional guidelines before study treatment.

Study drug. Patients received obatoclax mesylate as an intravenous solution dissolved in 5% dextrose with 2% polyethylene glycol 300 and 0.08% polysorbate 20. The study drug was administered as a continuous infusion over 24 h or multiples thereof. Patients were permitted to receive antiemetic prophylaxis, blood product support, antibiotics, and intravenous immunoglobulins as appropriate and at the discretion of the individual institution. Treatments for the underlying malignancy with the exception of hydroxyurea were not permitted during the study period.

Trial design. Obatoclax mesylate was administered at increasing doses and increasing frequencies as outlined in Table 1. In the initial portion of the trial, patients received 7 to 40 mg/m² over 24 h every 2 weeks. Following the completion of this dose escalation, patients received 20 to 28 mg/m²/wk and then 20 to 28 mg/m² over 24 h daily for 2, 3, or 4 days. Dose-limiting toxicity was defined as grade 3/4 adverse events as defined by National Cancer Institute Common Toxicity Criteria version 2.0 and considered at least possibly related to the study drug occurring during the first cycle. The standard 3 + 3 escalation rule was used for this trial and the recommended phase II dose was defined as the highest dose level at which no more than 1 of 6 patients experienced a dose-limiting toxicity. Expansion of the cohort to 6 patients was planned at the 28 mg/m²/wk dose, as this schedule represented the highest weekly dose administered, and at the 20 mg/m² over 24 h for 4 days, as this dose represented the maximal number of planned days of consecutive dosing. Expansion of the cohort at these doses was planned to provide a more robust measure of safety. Finally, expansion of the cohort to 6 patients was planned at 28 mg/m² over 24 h daily for 4 days, as this dose represented the highest dose administered in a multiday dosing schedule and represented completion of the study. All other expansions of the cohort were due to apparent dose-limiting toxicities of QTc prolongation as described in Results. All patients were considered to be assessable for safety and toxicity if they received any dose of obatoclax mesylate.

Pretreatment and follow-up studies. A complete history and physical examination including neurologic assessment was done at screening, day 1, day 8, and before the start of each dose of obatoclax mesylate and at the end of study. Complete blood counts with differential and serum chemistry were done at the above times. Bone marrow aspirates and biopsies were obtained before enrollment and as clinically indicated. Adverse events were graded based on the National Cancer Institute Common Toxicity Criteria version 2.0. Tumor responses for CLL were documented within 14 days of starting obatoclax mesylate and repeated every 4 weeks using the criteria proposed by the National Cancer Institute-sponsored Working Group Guidelines for CLL (11). Criteria for responses for AML, myelodysplasia, and ALL followed those defined by the WHO (12, 13).

Pharmacokinetics. For the every 2 weeks and weekly 24 h infusion schedules, peripheral blood samples were collected on days 1 to 8 of cycle 1 pre-dose as well as 1, 4 to 6, 8, 24 (15-30 min before the end of infusion), 24.5, 26, 28, and 32 h after the initiation of the infusion. Samples were also collected at 38 ± 4, 48 ± 4, 72 ± 4, 96 ± 4, and 168 ± 4 h after the initiation of the infusion but not included in the analysis due to incomplete sample collection at these time points. During cycle 2, peripheral blood samples were collected pre-dose as well as 8 and 24 (15-30 min before the end of infusion) h after initiation of the infusion. For the multiday infusion schedule, peripheral blood samples were collected on days 1 to 5 of cycle 1 before each dose as well as 1, 4 to 6, 8, 24 (15-30 min before the end of infusion), 48 (15-30 min before the end of infusion), 72 (if applicable; 15-30 min before the end of infusion), and 96 (if applicable; 15-30 min before the end of infusion) h from the start of infusion. In the multiday infusion, samples were also collected at 0.5, 2, 4, 8, 14 ± 4, 24 ± 4, 48 ± 4, 72 ± 4, 96 ± 4, and 144 ± 4 h from the end of the infusion. Blood samples were collected in pre-chilled EDTA/K2 plastic tubes. Pharmacokinetic variables were pooled across dose levels.

Plasma concentrations of obatoclax mesylate were measured using a validated liquid chromatography-mass spectrometry method. Curves of concentration of obatoclax versus time in plasma were constructed for each patient and analyzed by a noncompartmental analysis technique. Peak concentration (C_max) and corresponding time (T_max), area under the concentration versus time curve through 24 h (AUC_0-t), area under the concentration versus time curve through the last measurement (AUC_last), terminal half-life (t_1/2), plasma clearance, and volume of distribution were calculated.

Demographics. Forty-four patients with refractory hematologic malignancies were treated on study with the doses and frequencies shown in Table 1. The patient demographics are shown in Table 2. The median age of the patients was 63 years (range, 26-82 years). Twenty-five patients had AML, 14 myelodysplasia, 4 CLL, and 1 ALL. Patients had a median of 4 prior therapies (range, 0-13). Three of 25 patients with AML were previously untreated, except for hydroxyurea, and were considered ineligible for induction chemotherapy due to poor risk cytogenetics and advanced age. The other 22 AML patients had relapsed disease or were primary nonresponders to induction chemotherapy.

Safety. A total of 306 infusions of obatoclax mesylate were administered with a median of 5 infusions per patient (range, 1-35). The study drug was well tolerated and the adverse events are summarized in Table 3. Central nervous system (CNS) symptoms were the most common adverse events attributable to the study drug. Patients experienced somnolence (43%), dizziness (38%), fatigue (36%), euphoric mood (34%), and gait disturbance (34%). In fact, patients described feeling inebriated. The neurologic symptoms were experienced within 2 h of the initiation of the infusion and abated within hours of stopping the drug. The neurologic effects were related to dose and frequency but never exceeded grade 1/2 in severity according to the National Cancer Institute Common Toxicity Criteria version 2.0. Specifically, these neurologic symptoms did not interfere with the patient's activities of daily living. Additional adverse events possibly attributable to the study drug were diarrhea (30%) and nausea and vomiting (31%), but these symptoms also did not exceed grade 2 in severity. Increased liver enzymes and creatinine were noted in study patients but in all cases judged to be unrelated to the study drug.

One patient at the dose level of 28 mg/m² every 2 weeks experienced a grade 3 hypersensitivity reaction to the drug that was easily controlled by the administration of H-1 and H-2 blockers and corticosteroids. Occasional occurrences of hypersensitivity reactions/cytokine release syndromes (∼5% of patients) were also observed in other phase I studies of obatoclax mesylate that were occurring at the time in different disease sites. As a result of these experiences, the protocol was amended to permit premedication with H-1 and H-2 blockers with or without corticosteroids. After this protocol change, hypersensitivity reactions were not observed in premedicated patients.

Decreased hemoglobin, platelet count, and neutrophil count were observed for virtually all patients but were considered secondary to the underlying disease process and not the study drug. Specifically, 79%, 84%, and 75% of patients had worsening neutropenia, anemia, and thrombocytopenia, respectively, during the study. Moreover, 68%, 39%, and 64% of patients had grade 3/4 neutropenia, anemia, and thrombocytopenia retrospectively during the study. However, these emergent cytopenias were judged not related to the study drug but rather related to progression of the underlying disease. Supporting this judgment, in all cases, these abnormalities were present before the administration of obatoclax mesylate and the worsening cytopenias occurred with evidence of disease progression. Finally, the cytopenias continued to worsen along with disease progression after discontinuation of the study drug. Other grade 3 toxicities judged to be unrelated to the study drug were febrile neutropenia, chills, pneumonia, dsypnea, and hypotension. These symptoms were considered attributable to infections that occurred due to the neutropenia that was present before initiation of the obatoclax mesylate.

Apparent grade 3 QTc prolongation based on automated interpretation of the ECG was observed in 1 of 6 patients at the 28 mg/m² dose every 2 weeks and 2 of 5 patients at the 40 mg/m² dose. However, on retrospective and independent review of the data, all 3 of 3 patients had QTc prolongation at baseline and other conduction abnormalities, which were thought to render the interpretation of the automated reports unreliable. No patients had true QTc prolongation including those treated at 28 mg/m²/d for 4 days. Despite this retrospective analysis of the apparent QTc prolongation, the highest dose of obatoclax mesylate advanced into the weekly and multiday dosing schedule was 28 mg/m²/d. The decision to not advance a higher dose was due to the CNS toxicity of the drug. Although the neurologic toxicities were only grades 1 and 2, it appeared that the nature and spectrum of these side effects, including sedation and feelings of inebriation, approached the limit of what could be managed in an outpatient setting at the 28 mg/m² dose.

Pharmacokinetics. The pharmacokinetics of obatoclax mesylate were investigated following single 24 h infusions (1 × 24 h) at doses ranging from 7 to 40 mg/m² and following multiple 24 h intravenous infusions (2 × 24, 3 × 24, and 4 × 24 h) at doses of 20 and 28 mg/m²/24 h. Pharmacokinetic results are summarized in Table 4. The plasma concentrations of obatoclax reached a steady state before end of infusion; overall, the plasma C_max and AUC_0-t values of obatoclax increased with increasing doses.

In the single 24 h schedule (1 × 24 h) with a dose of 7 to 40 mg/m², the C_max and AUC_0-t values ranged from 3.21 to 15.26 ng/mL and 62 to 305.5 ng h/mL, respectively (Table 4). The T_max and elimination t_1/2 values ranged from 6.7 to 20.3 h and from 7.7 to 14.1 h, respectively.

In the multiday infusion schedule, the mean C_max values at 20 mg/m²/24 h following 1 × 24, 2 × 24, 3 × 24, and 4 × 24 h infusions were 15.3, 8.4, 10.1, and15.7 ng/mL, respectively. The mean C_max values at 28 mg/m²/24 h following 1 × 24 and 4 × 24 h infusions were 10.8 and 12.73 ng h/mL, respectively. Steady-state plasma concentrations were similar during a single and multiday infusion schedules (1 × 24-4 × 24 h) within the same dose. In the multiday infusion schedule, the mean AUC_0-t values at 20 mg/m²/24 h following 1 × 24, 2 × 24, 3 × 24, and 4 × 24 h infusions were 264.3, 396.8, 698.1, and 1,168.3 ng h/mL, respectively. The mean AUC_0-t values at 28 mg/m²/24 h following a single 24 h and 4 × 24 h infusions were 213.9 and 1,146.1 ng h/mL, respectively. As expected, the AUC values increased with the total amount of drug administered in 1 × 24 up to 4 × 24 h infusion. Elimination t_1/2 values in single 24 h dosing schedule were less than the values in multiple 24 h schedules, which may be due to the result of truncation of pharmacokinetic profiles at 38 h in single dosing schedule. Plasma clearance and volume of distribution values appeared to be comparable in single and multiple 24 h dosing schedules.

Disease response. One of 25 patients with AML had a complete response (CR) to obatoclax mesylate. Three of 14 patients with myelodysplasia had hematologic improvement with RBC and platelet transfusion independence. The remaining patients had progressive disease (Table 5).

The patient with AML who achieved CR was 1 of 3 patients with previously untreated AML. This 70-year-old patient developed AML 13 months after the diagnosis of T₁N₁ high-grade ductal breast carcinoma and treatment with Adriamcyin, cyclophosphamide, and local radiation treatment. The patient presented with transfusion-dependent anemia and thrombocytopenia and an Eastern Cooperative Oncology Group performance status of 2. Bone marrow examination revealed AML M₄ with 80% blasts and a t(9;11)(p22;q23) by cytogenetics analysis. The patient was considered unsuitable for conventional induction chemotherapy due to the combination of advanced age and poor risk cytogenetics. Therefore, obatoclax mesylate was administered at a dose of 20 mg/m²/wk over 24 h. By day 9 following the start of treatment, the patient experienced complete hematologic recovery and transfusion independence (Fig. 1). Bone marrow aspiration revealed CR. The t(9;11) abnormality was no longer detectable by fluorescence in situ hybridization or quantitative reverse transcription-PCR. Her Eastern Cooperative Oncology Group performance status improved to 0. The patient remained on weekly obatoclax mesylate infusions and received a total of 35 infusions of 20 mg/m² without cumulative toxicities. Six months after CR, the t(9;11) fusion transcripts were detected by fluorescence in situ hybridization in 3% of cells, and obatoclax mesylate was continued at the same dose. Eight months after CR, the patient relapsed with 5% blasts in the bone marrow and t(9;11) fusion transcripts were detectable by fluorescence in situ hybridization in 22% of nuclei. At relapse, the patient was withdrawn from the study as per protocol. Although the protocol did not permit intrapatient dose escalation, given the favorable response at the lower dose level, the patient was given an opportunity to reenroll in the study at the open dose level (28 mg/m² for 4 days every 2 weeks). This protocol deviation was reviewed and approved by the local research ethics board and the sponsor. On receipt of these approvals, informed consent was obtained and the patient was re-registered in the study. However, despite the increased dose, the patient did not respond to this treatment and died shortly thereafter from progressive disease. Of note, 2 other AML patients with an 11q23 translocation [one of whom also had a newly acquired del(14) and a patient with ALL and mixed lineage leukemia (MLL) t(9;11)] were treated on this study but did not respond to obatoclax mesylate. However, these 3 other patients with MLL-associated leukemia had relapsed after prior induction chemotherapy. The only other previously untreated patients were a patient with secondary AML and del(5) and del(7) chromosomal abnormalities, who did not respond after three weekly infusions and was removed from study to receive an alternative therapy, and a patient with de novo AML and 7q- whose disease progressed after 4 weeks on obatoclax mesylate and was removed from the study.

Among the 3 patients with myelodysplasia who showed hematologic improvement was a 75-year-old man with therapy-related myelodysplasia (refractory anemia with ring sideroblasts) secondary to cytotoxic chemotherapy treatment for large cell non-Hodgkin's lymphoma who required 2 and 23 units RBC and platelets, respectively, in the 2 months before obatoclax mesylate treatment. The patient received 7 mg/m² obatoclax mesylate every 2 weeks. After the first 2 weeks (1 cycle), the patient became platelet transfusion independent until cycle 5 (day 55). The patient had a 10-fold increase in his platelet count (8,000-81,000/μL). Follow-up bone marrow examination revealed persistent dysplastic features. After day 55, the patient became platelet transfusion dependent again. His RBC transfusion requirements did not change during the study. In addition, a 40-year-old man with therapy-related myelodysplasia (refractory anemia with ring sideroblasts) and platelet transfusion dependence (14 units platelets in the 2 months before obatoclax mesylate treatment) received 14 mg/m² obatoclax mesylate every 2 weeks and became transfusion independent after the first 2 weeks (1 cycle) until cycle 5 (day 57). During the period of transfusion independence, the platelet count ranged from 16,000 to 21,000/μL and the hemoglobin rose from 7.9 g/dL before obatoclax treatment to as high as 10.6 g/dL. Follow-up bone marrow aspiration revealed hypolobulated megakaryocytes that were reduced in number. After day 57, the patient became platelet transfusion dependent again. Finally, a 65-year-old female with primary myelodysplasia of an unknown subtype experienced an increase in hemoglobin and RBC transfusion independence. Before enrollment in this study, the patient's hemoglobin was 8.6 g/dL and they required transfusions of 2 units RBC every month in the 3 months before treatment with obatoclax mesylate at 40 mg/m² every 2 weeks. After treatment, the hemoglobin rose as high as 13.2 g/dL and the patient was transfusion independent for 4 months.

Obatoclax mesylate is a small-molecule pan-Bcl-2 family inhibitor that binds and blocks the function of Bcl-x_L, Mcl-1, Bcl-w, A1, and Bcl-b (10). Encouraged by in vitro studies and its favorable toxicity profile in animals (10), we conducted a phase I study in patients with refractory malignancies. In this study, patients received a total of 306 cycles of obatoclax mesylate with increasing doses and frequencies. The recommended phase II dose was up to 28 mg/m², which could be administered over 24 h for up to 4 consecutive days. At this concentration and frequency, no dose-limiting toxicities were observed.

In our study, obatoclax mesylate was well tolerated. The most frequent adverse events were grade 1/2 CNS symptoms including somnolence, euphoric mood, gait disturbances, and an overall sensation of feeling inebriated. Although the adverse CNS events were grade 1/2 even at the highest dose evaluated, it was judged that further escalation of the frequency of obatoclax mesylate beyond 28 mg/m² dose would be problematic due to this CNS toxicity. Specifically, the nature and spectrum of these side effects, including sedation and feelings of inebriation, approached the limit of what could be managed in an outpatient setting. For this reason, it was elected to not escalate beyond the dose of 28 mg/m² in weekly or multiday dosing. For the same reason, the phase II study at a dose of 28 mg/m² over 24 h × 4 days is recommended, as the CNS symptoms are easily manageable in an outpatient setting. Although responses were observed at doses and frequencies lower than the recommended phase II dose, the phase II dose of 28 mg/m² over 24 h × 4 days was selected to administer a maximal amount of drug.

It is unclear why patients developed CNS symptoms with this drug. Potentially, these symptoms represent an on-target effect of inhibiting Bcl-2 family members, as Bcl-2 promotes the survival of neurons (14) and protects them from apoptosis (15). Bcl-x_L plays a role in synaptic plasticity, and Bcl-x_L inhibition represents another on-target mechanism (16). Alternatively, these symptoms may represent an off-target effect of the binding of obatoclax mesylate to intracellular targets beyond Bcl-2 family members.

Although the majority of patients in this study developed worsening cytopenias during treatment with obatoclax mesylate, these cytopenias were judged to be related to the progressive disease. Supporting this contention, the worsening cytopenias occurred with evidence of disease progression and continued to worsen along with disease progression after discontinuation of the study drug. In addition, prior studies in animals have shown that obatoclax mesylate is not myelosuppressive (8, 10). Finally, prior phase I studies in solid tumors did not report myelosuppression after treatment with obatoclax mesylate. However, lymphopenia was noted in 66% of these patients and may be related to obatoclax mesylate (17).

Pharmacokinetic analysis conducted during this study showed that steady-state plasma levels of obatoclax mesylate were rapidly achieved and steady-state levels were maintained throughout the duration of the infusions. C_max and AUC were generally proportional to the dose administered. The AUC also corresponded to the duration of the infusion. Thus, obatoclax mesylate has a favorable pharmacokinetic profile that supports its further evaluation for clinical efficacy.

In the current study, we observed a dramatic clinical response to obatoclax mesylate. A patient with previously untreated AML secondary to chemotherapy who harbored a MLL t(9;11) translocation achieved CR with one course of obatoclax mesylate. The patient had a sustained remission for 8 months during which time she maintained a normal performance status. Such a response is extremely noteworthy, as the patient did not experience the usual pancytopenia associated with conventional chemotherapy. At the time of remission, the bone marrow aspirate did not have detectable levels of the t(9;11) by either fluorescence in situ hybridization or quantitative reverse transcription-PCR. The absence of the translocation indicates that the remission did not arise from the differentiation of the leukemic blasts to mature hematopoietic cells.

Efforts are under way to profile the expression of the Bcl-2 family members in this patient to provide a molecular explanation for her response and thereby may help identify a subgroup of AML patients most likely to achieve CR with single-agent obatoclax mesylate. Potentially, MLL-associated leukemia may be particularly responsive to Bcl-2 family inhibitors. In support of this hypothesis, inhibition of MLL expression with small interfering RNA leads to reductions in levels of Bcl-x_L expression and decreased cellular proliferation (18). Moreover, the HoxA9 transcription factor that is increased in MLL translocations acts as a suppressor of the Bcl-2 antiapoptotic genes (19). Therefore, aberrant expression of Bcl-2 family members may be crucial for the survival of MLL-associated leukemias. Alternatively, the patient who achieved CR was 1 of only 3 other previously untreated patients with AML. Potentially, this agent has single-agent activity in previously untreated AML as supported by reports of preclinical activity in AML cell lines and patient samples (9). In this regard, a clinical trial to evaluate obatoclax mesylate in patients with previously untreated AML who are ages >70 years is planned.

Ancillary studies to determine whether obatoclax inhibits Bcl-2 family members were not conducted and this lack of data is a limitation to this study. These correlative measurements, however, are part of the phase II study that will be conducted in patients with previously untreated AML. Of note, in a prior phase I study of obatoclax mesylate in patients with refractory CLL, we showed drug-induced apoptosis as measured by an increase in plasma concentration of histone-oligonucleosomal DNA complexes (20). Moreover, in cell lines, obatoclax mesylate inhibited the interaction of Mcl-1 and Bak (8, 9).

Obatoclax mesylate also showed activity in myelodysplasia as 3 of 14 patients showed hematologic improvement and became temporarily platelet or RBC transfusion independent. Thus, obatoclax mesylate may be improving cytopenias in these patients. A limitation to this study, however, was that the mechanism by which obatoclax mesylate produced hematologic improvement is uncertain. Interestingly, examination of the bone marrow of the responding myelodysplasia patients revealed persistent dysplastic features and reduced numbers of megakaryocytes. Currently, a phase II clinical trial of obatoclax mesylate in patients with previously untreated myelodysplasia is under way, and this trial, along with its correlative studies, will help better define the response to this drug in this patient population.

In summary, obatoclax mesylate is a small-molecule pan-Bcl-2 family inhibitor that can be administered over repeated cycles to patients with refractory leukemias. The drug is well tolerated, producing predominantly transient mild CNS side effects. Early evidence of activity has been observed through pharmacodynamic assessment and clinical responses have also been observed. Thus, obatoclax mesylate should be evaluated in further clinical trials to better define its efficacy and optimal role in the treatment of hematologic malignancies.

J. Viallet, M. Berger: employees, GeminX. A.D. Schimmer, H. Kantarjian, M. Andreeff, G. Borthakur: clinical research support, GeminX.