Clonal Evolution during Dual B Cell Receptor Pathway Inhibitor Therapy with Acalabrutinib and Umbralisib in CLL Patients

Achieving undetectable minimal residual disease (uMRD) is rare with B-cell receptor (BCR) targeting in chronic lymphocytic leukemia (CLL). We have reported interim results of a phase 2 study combining acalabrutinib with the PI3K inhibitor umbralisib for 6 months, followed by the addition of the anti-CD20 antibody ublituximab (AU2) for months 7-12, with end of therapy at 12 months for those in complete response (CR) or 24 months for others (Ahn et al., ASH 2023). Here we report targeted somatic mutation profiling of paired pretreatment (Pre)- and end of cycle 6 peripheral blood (PB) and bone marrow (BM) samples from 15 patients enrolled in this phase 2 study. These 15 patients had a median age of 60 years, 60% were male, 80% had unmutated IGHV, none had del(17p), and 13.3% had TP53 mutation. 13 patients were previously untreated while 2 had 3 prior therapies each. End of therapy (EOT) response evaluation showed that 5/15 (33%) patients achieved CR, while 10/15 (67%) had PR, and 9/15 (60%) achieved undetectable minimal residual disease (uMRD) at EOT.

Paired baseline and end-of-cycle 6 PB and BM samples were sequenced with the Tissue-based Lymphoma Panel (TLP), a targeted sequencing panel designed to detect recurrent and biologically relevant genetic variants in lymphoid malignancies. The TLP covers a 2.2 Mb region and includes probes for 319 candidate cancer genes and copy number alterations. Mean target coverage per sample was 623x (range, 237-904) and a minimum of four alternate reads was required to call a variant. All mutation calls were verified with the Integrative Genomics Viewer. After quality check analysis, we had Pre and C7D1 PB samples from 15 patients to assess clonal evolution during the first 6 months of acalabrutinib umbralisib. Additionally, we had 9 PB-BM pairs at baseline and 15 PB-BM pairs at C7 for mutational concordance analysis. Evolving clones were defined as showing a change in variant allele frequency (VAF) of 10 percentage points or greater between baseline and cycle 7, or for clones with VAF < 10%, a doubling of VAF.

Analysis of baseline mutations showed NOTCH1 in 7 patients, ATM and TET2 in 6 patients, SF3B1 and KIT in 4 patients, and TP53 in 1 patient. Interestingly, 4 of 5 CLL patients with a CR to AU2 treatment had SF3B1 mutation, a known driver of CLL progression, while none of the partial responders had this mutation. Of the 15 patients, 3 acquired new mutations in PB at C7, in the setting of clinical PR: the first had new mutations in NCOR2, STAT3, and DNMT3A (VAFs: 0.06, 0.08, and 0.05); the second in ATM and JAK2 (VAFs: 0.05 and 0.04); and the third in ATM (VAF: 0.01). Additionally, 10 of 15 harbored evolving clones at C7. Among these, 5 patients exhibited evolving clones in the NF-kappa B (NFKB) pathway, specifically in BIRC3 (mutated in 2 patients, the first R549Afs pre VAF:0.006 and C7:0.024; the second with two mutations Q547Nfs and L585*, VAFs pre: 0.040, 0.037 and C7: 0.187, 0.149 respectively); NFKBIE (two mutations in 1 patient; p.Q383* and Y254Sfs, VAFs pre 0.013, 0.025 and C7 0.144, 0.149); and TRAF2 (VAF pre 0.003 and C7 0.03) and TRAF3 (pre 0.012 and C7 0.113), mutated in 1 patient each. All 5 patients with evolving mutations in the NFKB pathway responded to AU2 treatment, 4 PRs and 1 CR. Overall, the study identified 18 mutations in NFKB pathway genes in 9 /15 patients (BIRC3 in 5 patients, CARD11 and NFKBIE in 2 each and TRAF2, TRAF3, and BTK in 1 each). One patient had both CARD11 and TRAF2. Among those with NFKB mutations, 5 of 9 (55.5%) patients remained MRD positive, compared to 6 of 15 (40%) overall.

Comparative analysis of mutation profiles between paired PB and BM at pre- and on-treatment time points showed that the overall mutation profiles were highly concordant between the two sample types. At C7, we found 119 mutations shared between PB and BM samples from 15 CLL patients. However, 20 mutations were unique to PB and 22 to BM. Mutations in DNMT3A (found in 2 CLL patients), SMARCA2, and FAT2 were exclusively present in BM samples. In contrast, mutations in TNFRSF14, NFKBIE, JAK2, ZFP36L1, and TSC2 were exclusively found in PB samples (each in 1 patient, except for DNMT3A).

Overall, these findings highlight that even in the setting of clinical response with dual BCR pathway inhibitor therapy, clonal evolution is ongoing in genes potentially associated with resistance. Further clinical follow-up of this study will be required to associate evolving mutations with duration of response or progression-free survival.