BAX Mutated Clonal Hematopoiesis Arises Following Treatment with the BCL2 Inhibitor Class of Therapeutics across a Range of Hematological and Non-Hematological Neoplasms

Abstract

Clonal hematopoiesis (CH) is an age-related phenomenon characterized by the expansion of genomically mutated hematopoietic clones with a selective fitness advantage. Venetoclax, a selective BCL2 inhibitor, has established efficacy in hematological malignancies such as chronic lymphocytic leukemia (CLL) and acute myeloid leukemia (AML). Our previous studies described the emergence of deleterious BAX variants within the myeloid compartment of patients (pts) receiving venetoclax for CLL (Blombery, Blood2022). Whether this phenomenon is specific to venetoclax and the CLL disease context is currently unknown. We therefore aimed to investigate the prevalence and pattern of BAX-mutatedCH across different hematological and non-hematological neoplasms in response to BCL2 inhibitor therapy (both venetoclax and BGB-11417). We first examined pts with AML in complete morphologic remission after venetoclax exposure in both the newly diagnosed (CAVEAT; Chua, JCO 2020) and early relapse (VALDAC; Tiong, ASH 2022) settings using unique molecular index-based error-corrected next generation sequencing with a sensitivity of 0.5% (QIAseq targeted DNA panel). BAX variants were detected in remission samples in 11/20 (55%) and 9/26 (35%) pts in these cohorts, respectively, after a median of 9- and 4-months post venetoclax. This represented a significant higher incidence compared to remission samples from pts with newly diagnosed AML post conventional cytotoxic chemotherapy without venetoclax (6/87 [7%]; adjusted p<0.01 for both pairwise comparisons). Having observed the increased frequency of BAX-mutated CH in remission in myeloid (AML) and lymphoid (CLL) neoplasms, we then investigated whether BAX variants may be acquired in response to BCL2 inhibitor therapy outside the context of hematological neoplasm. We assessed a cohort of pts with metastatic breast cancer who received venetoclax in combination with tamoxifen (mBEP; Lok, Cancer Discover 2019), and detected BAX variants from peripheral blood samples in 5/16 (31%) pts (median age 66) after a median duration of 20 (range 13-44) months on venetoclax. In contrast, no BAX variants were detected in a similarly aged control cohort of 50 pts (median age 62) with breast cancer without exposure to venetoclax (p=0.01). The rates of any CH-related mutation were 75% and 50% pts, respectively (p=0.09). No pt developed MDS or AML during follow up in the mBEP cohort. Finally, to explore whether emergence of BAX-mutated CH is specific to venetoclax as an agent, we studied a cohort of 9 pts with CLL who had received BGB-11417, a more potent and selective BCL2 inhibitor; median treatment duration 13 months. Notably, we observed the emergence of BAX variants in 2/9 (22%) pts in this cohort. The characteristics of the BAX variants (n=88) in 28 pts (AML, CLL or breast cancer) are summarized in Figure 1; 3 variants (in 3 pts) were detectable in pre-treatment samples. Over half of the missense (n=22) or in-frame variants (n=1) were clustered in the C-terminus of the BAX protein, while truncating variants (n=65) were observed throughout the gene. The median number of BAX variants per pt was 2, with 32% of pts having ≥3 variants (max=13). The median VAF was 1.4%, and only 16% of variants had a VAF >5%. The number and VAF of BAX variants did not significantly correlate with the duration of therapy. Among 12 pts with AML who had serial testing (median inter-test interval 10 months), BAX variants (n=24) did not significantly change in VAF (median Δ 0.7%) with time after initial acquisition (Pearson's coefficient 0.17). To clarify the compartments affected by BAX variants we performed single-cell combined DNA/protein sequencing (Tapestri) of a sample with multiple BAX variants and demonstrated their presence in multiple independent clones with myeloid-biased lineage but also within the natural killer cell compartment ( Figure 2). In summary, the emergence of BAX-mutated CH appears to be a generalized phenomenon of adaptive hematopoiesis that occurs across both hematological and non-hematological malignancy settings as well being observed with other BCL2 inhibitors consistent with a therapeutic class effect. Further research is needed to understand the contribution of this phenomenon to the subsequent development of clinically relevant hematological abnormalities and its implications for future therapeutic resistance.