Mechanisms of Venetoclax Resistance in Chronic Lymphocytic Leukemia

Rationale: Treatment with the small-molecule BCL2-inhibitor venetoclax (ABT-199) has demonstrated substantial efficacy in patients with chronic lymphocytic leukemia (CLL), particularly in cases with relapsed/refractory disease and mutated or deleted TP53 . While resistance to the Bruton's tyrosine kinase (BTK) inhibitor ibrutinib has been associated with the de-novo acquisition of gene mutations in the primary drug target itself and downstream effectors like PLCG2, genetic causes of treatment resistance towards venetoclax have not been identified to date.

Methods: To this end, we performed whole exome sequencing and methylation array profiling in a pilot set of 8 patients (7 men, 1 woman) with pre-treated (1-8 pretreatments) and TP53 -deficient CLL, who had progressive disease or relapsed under oral venetoclax therapy. We investigated serial CLL specimen from peripheral blood before treatment initiation and from peripheral blood, lymph node or bone marrow and at 1-2 follow up time points during disease progression/relapse. Detected genetic alterations of interest were confirmed by either dideoxy sequencing or digital droplet PCR. Copy number changes detected by exome sequencing were verified by methylation arrays.

Results: On average, we detected a total of 25.5 exonic mutations (including silent, insertions, and deletions) prior to venetoclax therapy. As expected from ongoing genome evolution, copy number changes and the number of point mutations increased during venetoclax therapy. Methylation profiles, however, remained stable under treatment pressure.

We reconstructed phylogenetic trees to illustrate the clonal trajectories of cell populations with distinct genetic events across the obtained cancer samples and over time. We observed a wide spectrum of evolutionary dynamics including linear, divergent, and convergent evolution.

Recurrent non-synonymous mutations developing under venetoclax treatment were seen in TP53, NOTCH1, and BTG1, whereas BRAF, SF3B1, RB1, BIRC3, and MLL3 were non-synonymously affected in single patients only.

Recurrent genomic changes that evolved during venetoclax treatment were homozygous deletions affecting CDKN2A/B (p16^Ink4a/p14^Arf) in 3 patients and BTG1 missense mutations in 2 cases. Reconstruction of the branching evolution of subclones harboring these alterations strongly suggested that they may confer a survival advantage to CLL cells under venetoclax treatment pressure.

Two patients developed genome alterations that would qualify for further therapeutic options after venetoclax therapy: One patient presented a BRAF p.K601E mutation at the time of resistance. We confirmed that mutationally activated BRAF is able to decrease the sensitivity towards venetoclax-mediated cell killing in an in vitro cell line model. Another patient demonstrated a high-level amplification of C D274, which was paralleled by high protein levels of the immune checkpoint ligand CD274/PD-L1 per immunohistochemistry in the progressive lymph node. This alteration may suggest the use of immune checkpoint inhibition in this patient.

Conclusion: Whole exome sequencing of CLL specimens from 8 patients before the initiation of venetoclax therapy and at the time of venetoclax resistance revealed diverse patterns of clonal evolution. Mutations in BTG1 and homozygous deletions in CDKN2A/B may play a role in the biology of acquired venetoclax resistance. Furthermore, our genetic analysis identified genome alterations at relapse that might hint to options for salvage therapies. To further pinpoint genetic mechanisms of venetoclax resistance, larger studies with repeated longitudinal sampling of CLL cell material under therapy and at disease progression/relapse are warranted.