Clonal hematopoiesis, myeloid disorders and BAX-mutated myelopoiesis in patients receiving venetoclax for CLL

Abstract

The BCL2 inhibitor venetoclax has established therapeutic roles in chronic lymphocytic leukemia (CLL) and acute myeloid leukemia (AML). As BCL2 is an important determinant of survival of both myeloid progenitor and B cells, we investigated whether clinical and molecular abnormalities arise in the myeloid compartment during long-term continuous venetoclax treatment of CLL in 89 patients (87 with relapsed/refractory CLL). Over a median follow-up of 75 (range 21-98) months, persistent cytopenias (≥1 of neutropenia, thrombocytopenia, anemia) lasting ≥4 months and unrelated to CLL occurred in 25 patients (28%). Of these patients, 20 (80%) displayed clonal hematopoiesis, including 10 with therapy-related myeloid neoplasms (t-MNs). t-MNs occurred exclusively in patients previously exposed to fludarabine-alkylator combination therapy with a cumulative 5-year incidence of 10.4% after venetoclax initiation, consistent with rates reported for patients exposed to fludarabine-alkylator combination therapy without venetoclax. To determine whether the altered myelopoiesis reflected the acquisition of mutations, we analyzed samples from patients with no or minimal bone marrow CLL burden (n = 41). Mutations in the apoptosis effector BAX were identified in 32% (13/41). In cellular assays, C-terminal BAX mutants abrogated outer mitochondrial membrane localization of BAX and engendered resistance to venetoclax killing. BAX-mutated clonal hematopoiesis occurred independently of prior fludarabine-alkylator combination therapy exposure and was not associated with t-MNs. Single-cell sequencing revealed clonal co-occurrence of mutations in BAX with DNMT3A or ASXL1. We also observed simultaneous BCL2 mutations within CLL cells and BAX mutations in the myeloid compartment of the same patients, indicating lineage-specific adaptation to venetoclax therapy.

Graphical abstract

Figure 1

Cytopenias and clonal hematopoiesis in patients with relapsed/refractory CLL treated with venetoclax. (A) Kaplan-Meier curve of cumulative incidence of therapy-associated myeloid neoplasm (t-MN) (solid line) across a cohort of 89 patients. The dashed line represents competing risks of death and allo-SCT. The numbers at risk for a given time point refer to at-risk for developing t-MN. (B) Categorization of a cohort of 89 patients into clonal hematopoiesis (CH), idiopathic cytopenias (IC), clonal cytopenias (CC), and t-MNs *CH detected with myeloid disorder-associated gene in 3 cases—not assessed in remaining 18. ^Refers to established myeloid-disorder associated genes, excluding BAX.

Figure 2

BAX mutations detected in the non-CLL compartment of patients on long-term venetoclax therapy for relapsed/refractory CLL. Location and type of variants are indicated by color-coded triangles on linear protein structure of BAX. Splice site variants are indicated below the protein at positions relative to their nucleotide position.

Figure 3

BAX α9 mutants show reduced activity when expressed in cells. (A) BAX structure (PDB: 1F16) showing the positions of the BAX missense mutations tested in this study (red). BAX helix α9 is highlighted in blue. (B) Test of activity of BAX mutants K123M, A183P, A183T, S184*, W188* in cells. BAX/BAK DKO MOLM-13 cells expressing the indicated BAX variants were treated with increasing concentration of venetoclax (50-5000 nM) for 24 hours, and cell viability was determined by Annexin V/DAPI staining and flow cytometry. Data are means ± SEM of 3 independent experiments. (C) Upon venetoclax treatment, BAX α9, but not BAX K123M, show reduced exposure for conformation-specific epitope 6A7. MOLM-13 DKO cells expressing the different BAX variants were pre-incubated with 25 μM Q-VD oph for 1 hour and then treated with 500 nM venetoclax. After 5 hours, cells were permeabilized with 0.025% digitonin prior to staining with conformation-specific anti-BAX antibody 6A7. After several washes, samples were stained with a PE-conjugated secondary antibody and analyzed by flow cytometry. The results are representative of at least two independent experiments. (D) Mutations in BAX α9 reduce MOM translocation and integration of BAX. Cells were treated as described in (C), subjected to carbonate extraction and fractions run on SDS-PAGE and immunoblotted for BAX. The results are representative of three independent experiments.

Figure 4

(A) Longitudinal changes in variant allele frequency (VAF) of BAX mutations and other mutations in the non-CLL compartment with time in two patients treated with continuous venetoclax. Box indicates a period of azacitidine (AZA) treatment of therapy-related myeloid neoplasm in patient CLL16. Three mutations in DNMT3A (Lys382*, Gly308Alafs*8, and Trp753Arg) detected at <1.5% VAF in patient CLL3 are not shown. Two BAX mutations, Arg34* and Arg89*, were detected in CLL16 at one time point (month 56), both at VAF of 1.1%, and are therefore depicted together. (B) Schematic diagram of targeted amplicon single-cell sequencing for patient CLL16 and CLL3. Het, heterozygous.