Molecular and functional determinants of response to venetoclax and inotuzumab ozogamicin in CD22-positive ALL

Abstract

Abstract Introduction: Venetoclax (VEN) is a selective BCL-2 inhibitor with clinical activity in subsets of patients with ALL (Del Gaizo Blood, 2008; Luskin Blood Adv, 2025). Inotuzumab ozogamicin (INO) is a CD22-directed antibody-drug conjugate (ADC) approved to treat relapsed or refractory (R/R) acute lymphoblastic leukemia (ALL; Kantarjian N Engl J Med, 2016). We initiated a phase 1 trial adding VEN to INO in adults with CD22-positive R/R ALL with hopes of improving the depth and durability of achieved remissions. We aimed to define correlates of response and resistance in patients with ALL treated with VEN plus INO to guide future studies for patients with R/R B-ALL. Methods: Bone marrow aspirates and peripheral blood were collected at baseline, on-treatment, and at progression from appropriately consented patients enrolled on clinical trial NCT05016947. In addition to standard clinical assessments, participants were evaluated for minimal residual disease (MRD) by immunoglobulin locus sequencing (clonoSEQ, Adaptive Biotechnologies; threshold 10-6) during induction. Baseline and progression specimens underwent mutational profiling using a CLIA validated targeted genotyping platform (Rapid Heme Panel, Brigham and Women's Hospital Center for Advanced Molecular Diagnostics; Kluk, The Journal of Molecular Diagnostics, 2016). A subset of baseline (n=17) and progression (n=1) specimens underwent BH3 profiling to measure overall apoptotic priming and anti-apoptotic dependencies. An additional subset of specimens from baseline (n=6) and progression (n=2) underwent flow cytometry to determine CD22 median fluorescence intensity (MFI) and fluorescence-activated cell sorting (FACS) to enrich leukemia cells for single-cell whole transcriptome sequencing (scRNA-seq) via 10x Chromium 5' Gene Expression (GEX) and B-cell receptor (BCR) analysis to delineate transcriptional programs mediating response and resistance. Results: Given the high rates of complete remission (CR) in this study (21/22 evaluable patients), baseline features were correlated with MRD positivity by clonoSEQ after cycle 2 to identify predictors of deep early responses. BH3 profiling of baseline specimens revealed that reduced priming by BAD1 and VEN correlated with cycle 2 MRD positivity (p=0.0008 and p=0.00001, respectively). While baseline CD22 MFI did not predict persistent MRD positivity (p=0.53), scRNA-seq analysis of six baseline specimens revealed in patients with MRD positivity reduced baseline expression of lymphoid transcription factors RUNX3 (log2 fold change or L2FC -6.5, adj. p-value <10-4) and IKZF2 (L2FC -5.3, adj. p-value <10-3) as well as multiple Reactome pathways annotated for cell cycling and B-cell receptor (BCR) signaling (normalized enrichment scores or NES <-1.5, adj. p-values <10-3). Seven patients progressed at a median of 182 days; of these, six patients received intervening therapy, including three who underwent stem cell transplantation. Mechanisms of resistance were explored in baseline/progression specimen pairs from two patients to date. One with KMT2A-rearranged (KMT2Ar) ALL progressed on treatment at 48 days; a second with BCR::ABL1 ALL progressed at 339 days, having proceeded to maintenance asciminib. BH3 profiling of a paired specimen revealed an increased dependency on MCL-1 from 20% to 58%, consistent with transcriptional upregulation of MCL1 observed at disease progression in relapsed patients (p=0.0473). Flow cytometry revealed a reduction in CD22 MFI of 69.3% and 64.5% at the time of disease progression compared to baseline, corresponding with a reduction in CD22 expression by single-cell RNA-seq (p=0.0018). At progression, the patient with BCR::ABL1 ALL also manifested loss of CD20 expression, while the patient with KMT2Ar ALL harbored an expanded KRAS G12V mutation. Progression specimens exhibited reduced BCR signaling (NES <-2, adj. p-value<10-7) compared to baseline. Conclusions: Reduced apoptotic priming and lymphoid transcription factor expression but not intensity of CD22 expression predicts persistent MRD by clonoSEQ after two cycles of INO and VEN. Patients who relapse demonstrate evidence of multiple candidate mechanisms of resistance, including acquisition of anti-apoptotic dependency on MCL-1, antigen escape via transcriptional downregulation of CD22 and other features of B-cell maturation, and activating KRAS mutations.