Bortezomib + MEC (Mitoxantrone, Etoposide, Cytarabine) for Relapsed/ Refractory Acute Myeloid Leukemia: Final Results of an Expanded Phase 1 Trial

MEC (mitoxantrone, etoposide, cytarabine) is a standard regimen for relapsed/ refractory (R/R) acute myeloid leukemia (AML), but outcomes remain poor. The overexpression of proteasomes and constitutive activation of NF-KB in AML cells suggest that proteasome inhibitors (PI) such as bortezomib (Bz) may be effective anti-leukemia therapy. PI or a decoy NF-KB oligonucleotide increase chemosensitivity to both anthracyclines and cytarabine. To test the hypothesis that PI may improve the efficacy of MEC, we conducted a phase 1 trial of Bz in combination with MEC. Here, we present final results of this trial: response rate, toxicity, and correlation of outcomes with mutation analysis. As CD74 expression may identify a subset NF-KB-dependent AML with predicted increased sensitivity to PI (Clin Can Res 2008; 14: 1446-54), we also explored this correlation.

Methods: All pts were treated at the Cleveland Clinic from Aug 2010-Mar 2014. This protocol was approved by the institution’s review board. Eligibility included: age 18-70 yrs, R/R AML, cardiac ejection fraction ≥ 45%. CD74 was assessed by flow cytometry using CD45 PE (BD Biosciences San Jose, CA) and CD74-Alexa 488 (AbD Serotec Raleigh, NC). A myeloid panel mutational analysis was performed on extracted DNA in pts with banked samples (n=26). All pts received 1 cycle of MEC: mitoxantrone (6 mg/m²/d), etoposide (80 mg/ m²), and cytarabine (1000 mg/ m²) Days 1-6. Bz was administered IV on Days 1, 4, 8, and 11. Dose was escalated using a standard 3 x 3 design. Dose levels (DL) were: -1 (0.40 mg/ m²), 1 (0.70 mg/ m²), 2 (1.0 mg/ m²), and 3 (1.3 mg/m²). Response was defined by IWG criteria (Cheson, 2006). The maximum tolerated dose (MTD) of Bz with MEC was 1.0 mg/m² (Advani et al, ASH 2012, Abstract 3595).

Results: Of 35 pts enrolled, the median age was 55 yrs (range 33-69), 13 (38%) were male, and median baseline WBC was 4.0 K/ µL (range 0.82-84.7). The median time from initial diagnosis of AML to enrollment was 8.4 months (range 1.1-88.2) and 6 pts (17%) had an antecedent hematologic disorder. Salvage status (S) at enrollment: S1 (24 pts, 69%), S2 (7 pts, 20%), S4 (4 pts, 11%). Nine pts (26%) were refractory to all prior therapies, and 3 pts (9%) had received prior allogeneic hematopoietic cell transplant (AHCT). Adverse cytogenetics per CALGB/ Alliance 8461 criteria occurred in 19% of pts at study entry and 15 of 26 pts (58%) had poor-risk molecular mutations (RUNX1, ASXL1, TET2, p53, IDH1, MECOM, FLT3 ITD). Ten pts were enrolled on DL1, 13 pts on DL2, 11 pts on DL3, and 1 pt died prior to treatment. Overall, 3 pts (9%) died during induction. In addition to febrile neutropenia and Gr 4 hematologic toxicity, the most commonly reported adverse events (AEs) were metabolic, constitutional, gastrointestinal (GI), and dermatologic, with the majority of these being Gr 1 or 2. GI toxicity was the only reported AE attributable to Bz: 12 pts had constipation or ileus (10: Gr 1 or 2; 2: Gr 3 or 4). Seventeen of the 33 evaluable pts (52%) have achieved a complete remission (CR) or complete remission with incomplete count recovery (CRi); with 1 pt inevaluable due to donor lymphocyte infusion. The estimated median overall survival was 7.2 months; median duration of response was 10.3 months. DL did not correlate with response. Eleven pts (32%) went on to receive AHCT. Among pts with poor-risk molecular mutations, 64% achieved CR/ CRi. Inhibition of NF-KB signaling in leukemia cells with mutated RUNX1 efficiently blocks growth and development of leukemia (Blood 2011; 118: 6626-37). Of the 5 pts with RUNX1 mutations, 3 (60%) achieved CR/ CRi, suggesting that Bz may have promising clinical benefit in this difficult subset of pts. Among the 17 pts with CD74 expression testing who were evaluable for response, the mean CD74 expression trended higher in non-responding pts (32.6%) than in responders (11.1%) (p=0.14).

Conclusions: The combination of MEC/Bz was well-tolerated and resulted in high response rates, even within a molecularly-defined poor risk population of pts with R/R AML. Our data do not confirm the expectation that higher CD74 expression would correlate with response in this R/ R AML cohort, but larger pt numbers are needed. These results, especially in pts with poor-risk mutations, support development of a randomized study to address the benefit of adding Bz to MEC in the treatment of R/R AML.