Mobilization Of Blasts and Leukemia Stem Cells by Anti-CXCR4 Antibody BMS-936564 (MDX 1338) in Patients With Relapsed/Refractory Acute Myeloid Leukemia

Bone marrow homing of AML is dependent on CXCR4, and high levels of CXCR4 expression correlate with worse survival in AML (Rombouts et al 2004, Spoo et al 2007). CXCR4 antagonists overcome environment adhesion mediated drug resistance and enhance chemotherapy induced cytotoxicity (Liesveld et al 2007, Zeng et al 2009, Nervi et al 2009, Beider et al 2010). Plerixafor, a small molecular CXCR4 inhibitor, was studied in a phase I trial of newly diagnosed AML patients (Uy et al ASH 2011), and BMS-936564, a fully-human monoclonal antibody to CXCR4, in combination with MEC (mitoxantrone, etoposide, cytarabine), is currently under study in relapsed/refractory AML.

The clinical trial completed a phase I dose escalation phase in AML patients, with increasing concentrations of BMS-936564 in 4 dose cohorts (0.3,1,3, and 10 mg/kg) and is currently enrolling a cohort of first salvage AML patients at the maximum dose of 10 mg/kg. The initial cohort of patients at 0.3 mg/kg received three weekly doses of antibody on days 1, 8, 15 [monotherapy period of cycle 1 (21 days)], followed by the same dose of antibody on days 1, 8, 15 of cycle 2 plus MEC chemotherapy [days 1-5 of cycle 2, (28 day cycle)]. After enrollment of the first cohort, the protocol was amended to reduce the monotherapy period to 1 week (1 dose of BMS-936564) in cohorts 1, 3, and 10 mg/kg, followed by the same combination regimen. As a companion study to this trial with the anti-CXCR4 antibody, we are investigating CXCR4 expression, timing of mobilization of leukemic blasts and leukemia stem cells (LSCs), and induction of apoptosis. Mobililzation of LSCs will be critical to eradication of leukemia, as they might serve as a reservoir for drug resistance and future relapse. We analyzed serial blast and LSC populations from blood and bone marrow samples from patients undergoing treatment by flow cytometry for phenotype, CXCR4 and annexin V expression. The putative LSCs were defined as CD34+CD38-CD123+ or by aldehyde dehydrogenase.

An independent assessment of CXCR4 expression in 56 consecutive AML patients from our institution not related to this clinical trial revealed a mean % expression of 31%, range 1-99%, with mean fluorescence intensity (MFI) of 2092, range 319-7942. A sample of 18 patients showed a correlation in CXCR4 expression between gated blasts derived from blood and bone marrow samples from the same patient (For % expression, r2=0.85, p=5e-8; MFI r2=0.45, p=0.002). Our site has enrolled 24 AML patients thus far on the above noted trial of BMS-936564. Administration of BMS-936564 resulted in brisk mobilization of leukemic blasts in 14/24 patients that initially peaked within at 2-6 hours post start of infusion in most patients, with an average of 2.1-fold increase ± 1.8 fold (range 1.06-8.96 fold), and some blasts continued to be in circulation for days. In most cases, the samples for which mobilization was not observed either did not have circulating blasts at baseline, or were from patients who received lower doses of BMS-936564. In addition, CD34⁺CD38^-CD123⁺LSCs were also mobilized post-treatment with BMS-936564, and in some cases, continued to rise over the subsequent days, during which the blast population declined. The average rise in %CXCR4 was from 29.3% pre-treatment to 69.8% peak value for blasts, and 23.0% pre-treatment to 75.6% peak value for LSCs. Although a direct correlation between CXCR4 expression by blasts and fold mobilization was not apparent, the highest fold increase in mobilization (∼9-fold) did occur in the patient with a moderately high level of CXCR4 expression, 42.5%. In this patient, there was also a sharp decline in circulating CXCR4 positive cells within 2 days and the patient achieved complete remission. BMS-936564 has demonstrated apoptosis in some preclinical models (Kuhne MR et al Clin Cancer Res v19(2): 357-66 (2013)).) For most patients, there was some increase in annexin V staining observed during the first 96 hours after antibody exposure. One patient sample with initial low baseline level of apoptosis exhibited a rise in annexin V staining (from 6% to 48%) that peaked on day 3 after administration of BMS-936564.

These data demonstrate that BMS-936564 induces mobilization of both AML blasts and LSCs, which may enhance chemotherapy-induced cytotoxicity in relapsed/refractory AML.

Chien:Bristol-Myers Squibb: Research Funding. Cardarelli:BMS: Employment. Sabbatini:Bristol-Myers Squibb: Employment. Shelat:Bristol-Myers Squibb: Employment. Cohen:Bristol-Myers Squibb: Employment. Becker:Bristol-Myers Squibb: Research Funding.