Phase 1b trial of tagraxofusp in combination with azacitidine with or without venetoclax in acute myeloid leukemia

Abstract

CD123, a subunit of the interleukin-3 receptor, is expressed on ∼80% of acute myeloid leukemias (AMLs). Tagraxofusp (TAG), recombinant interleukin-3 fused to a truncated diphtheria toxin payload, is a first-in-class drug targeting CD123 approved for treatment of blastic plasmacytoid dendritic cell neoplasm. We previously found that AMLs with acquired resistance to TAG were re-sensitized by the DNA hypomethylating agent azacitidine (AZA) and that TAG-exposed cells became more dependent on the antiapoptotic molecule BCL-2. Here, we report a phase 1b study in 56 adults with CD123-positive AML or high-risk myelodysplastic syndrome (MDS), first combining TAG with AZA in AML/MDS, and subsequently TAG, AZA, and the BCL-2 inhibitor venetoclax (VEN) in AML. Adverse events with 3-day TAG dosing were as expected, without indication of increased toxicity of TAG or AZA+/-VEN in combination. The recommended phase 2 dose of TAG was 12 μg/kg/day for 3 days, with 7-day AZA +/- 21-day VEN. In an expansion cohort of 26 patients (median age 71) with previously untreated European LeukemiaNet adverse-risk AML (50% TP53 mutated), triplet TAG-AZA-VEN induced response in 69% (n=18/26; 39% complete remission [CR], 19% complete remission with incomplete count recovery [CRi], 12% morphologic leukemia-free state [MLFS]). Among 13 patients with TP53 mutations, 7/13 (54%) achieved CR/CRi/MLFS (CR = 4, CRi = 2, MLFS = 1). Twelve of 17 (71%) tested responders had no flow measurable residual disease. Median overall survival and progression-free survival were 14 months (95% CI, 9.5-NA) and 8.5 months (95% CI, 5.1-NA), respectively. In summary, TAG-AZA-VEN shows encouraging safety and activity in high-risk AML, including TP53-mutated disease, supporting further clinical development of TAG combinations. The study was registered on ClinicalTrials.gov as #NCT03113643.

Graphical abstract

Figure 1.

Treatment schema with dose levels and schedules tested. Diagram of the study design and participants. (Top) TAG and AZA were first tested as a doublet combination at 5 different doses/schedules of TAG with 7-day dosing of AZA. The RP2D of TAG was determined to be 12 μg/kg daily for 3 days (d1, 2, 3; in magenta) in combination with AZA. Patients with AML or MDS were eligible in Cohorts A1-A4. While cohort A4 was being enrolled, AZA-VEN was approved for AML. Therefore, the study was amended and cohort A5 was limited to MDS. (Bottom) 3 doses of TAG were tested with AZA-VEN as a triplet in patients with AML. In the first cycle, VEN was given as 100 mg on day 1, 200 mg on day 2, and 400 mg on day 3, followed by 400 mg on days 4 to 21. In subsequent cycles, VEN was given as 400 mg on days 1 to 21. The RP2D of TAG was determined to be 12 μg/kg daily for 3 days (d4, 5, 6; in magenta) in combination with AZA-VEN. Patients with 1L or R/R AML were eligible in dose escalation, and then separate expansion cohorts in 1L and R/R AML were enrolled.

Figure 2.

Overview of bone marrow response and patient outcomes in the previously untreated AML cohort (n = 26) that received TAG-AZA-VEN. (A) Waterfall plot showing the best bone marrow blast response at any cycle as percentage change from baseline. Not shown are 3 patients who left the study prior to bone marrow assessment and 1 patient with secondary AML after myelofibrosis whose bone marrow was fibrotic and acellular at baseline. Two patients received 7 μg/kg TAG (dose level A), 1 received 9 μg/kg (dose level B), and the remaining 23 received 12 μg/kg (dose level C). Best response is annotated by color, as indicated. TP53 mutant status is annotated with an “M.” (B) Swimmer plot showing events and outcome for each patient over time from treatment start. Best response is annotated by the color of each bar, as in panel A. Circles indicate the time of first achieving CR, CRi, or MLFS, as indicated; cross is the time of allogeneic stem cell transplantation; triangle is the time of progressive disease (PD); square is the time of death. Annotation at the end of the bar is the status at end of study treatment. Patients without PD or death noted remained in remission at the last known follow-up. TP53 mutant status is annotated with an “M” and the number of cycles of treatment received is indicated.

Figure 3.

Survival outcomes in the previously untreated AML cohort that received TAG-AZA-VEN. (A) OS probability by intention-to-treat analysis using the method of Kaplan-Meier for the 26 patients in the 1L AML TAG-AZA-VEN cohort. (B) PFS probability for the same cohort, calculated as in panel A.

Figure 4.

Previously untreated AML receiving TAG-AZA-VEN, subgroup survival outcomes. (A) OS probability in the previously untreated AML TAG-AZA-VEN cohort, separated by TP53 mutant (n = 13, TP53 mut, blue) or TP53 wild-type (n = 13, TP53 WT, red). (B) PFS probability for the TP53 subgroups as in panel A. (C) OS probability for patients in the previously untreated AML TAG-AZA-VEN cohort that achieved CR/CRi/MLFS, separated by those determined to be MRD negative (n = 12, MRD neg, red) or positive (n = 5, MRD pos, blue). (D) PFS probability for the MRD subgroups as in panel C. (E) OS probability in the previously untreated AML TAG-AZA-VEN cohort, separated by those who received allogeneic stem cell transplant (n = 13, alloSCT Y, red) or those who did not (n = 13, alloSCT N, blue). (F) PFS probability for the alloSCT subgroups as in panel E. alloSCT, allogeneic stem cell transplantation.