Loss-of-function alterations in TRAF3 drive BTK/PLCG2-independent ibrutinib resistance in CLL

Abstract Bruton's tyrosine kinase inhibitors (BTKis) have transformed the treatment of Chronic Lymphocytic Leukemia (CLL), yet many patients eventually relapse. Resistance is typically ascribed to hotspot mutations in BTK or PLCG2. Reported variant allele frequencies (VAFs) for these mutations vary widely, suggesting that additional genomic events may drive treatment resistance. We analyzed 84 patients with CLL (52 treatment-naïve, 32 relapsed/refractory) treated with single-agent ibrutinib 420 mg daily on a phase 2 trial (NCT01500733). Patients were enrolled between January 2012 and January 2014 and had either a TP53 aberration or age ≥65 years. With a median follow-up of 120.0 months, the median progression-free survival was 85.9 months. Of 84 patients, 39 (46.4%) developed progressive disease (PD): 33 (39.3%) with progressive CLL (PD-CLL) and 6 (7.1%) with disease transformation. PD-CLL occurred in 16 of 52 (30.8%) treatment-naïve patients and 17 of 32 (53.1%) patients enrolled with relapsed/refractory CLL. Peripheral blood (PB) samples were sent for genotyping of known hotspot mutations in BTK and PLCG2 (Neogenomics). Digital droplet PCR was run on PB samples for BTK C481S/R/Y and PLCG2 R665W/L845F/S707F mutations. Of 28 evaluable patients, 25 (89.3%) harbored ≥1 BTK/PLCG2 resistance variant (VAF 0.04-0.96), and 12 (42.9%) carried multiple. Twelve patients (42.9%) demonstrated a non-dominant BTK/PLCG2-associated tumor burden (cumulative cell fraction ≤ 0.5) in the PB at the PD time point, of which 3 (10.7%) had no detectable BTK/PLCG2 mutations. To identify resistance mechanisms beyond BTK/PLCG2, we performed whole-exome sequencing (WES) of longitudinal PB samples from 27 patients with PD-CLL. Two patients demonstrated biallelic inactivation of TRAF3 (tumor necrosis factor receptor-associated factor 3): a missense variant (P63L) or a frameshift indel (146fs). In both cases, there was also a deletion of chromosomal region 14q32.32, comprising TRAF3. TRAF3 P63L/146fs mutations were undetectable at baseline. At the PD time point, these mutations expanded and were detected in 62% and 27% of circulating leukemic cells, respectively, indicating strong positive selection during ibrutinib therapy. TRAF3 functions as a negative regulator of NF-κB signaling, and its loss can lead to the activation of the non-canonical NF-κB signaling pathway. Therefore, we hypothesized that TRAF3 loss-of-function (LOF) alterations could provide an alternative, BTK/PLCG2-independent ibrutinib resistance mechanism. To establish the functional impact of TRAF3 loss, we generated isogenic TRAF3 knockout (KO) derivatives of the diffuse large B-cell lymphoma cell line TMD8 using Alt-R CRISPR-Cas9 ribonucleoprotein complexes (IDT) with 2 gRNAs (Synthego). Cells were single cell sorted by flow cytometry and/or limiting dilution. After Sanger sequencing of the target site, analysis using DECODR (Gene Editing Institute) confirmed frameshift indels with 100% expected KO. Subsequent immunoblotting with a TRAF3 antibody (Proteintech) showed minimal detectable TRAF3 protein. For ibrutinib sensitivity testing, wild-type TMD8 (TMD8-WT) cells and 3 independent TMD8 TRAF3-KO clones were seeded in triplicate at 10,000 cells/well and treated for 72 hours with two-fold serial ibrutinib dilutions (100-0.39 nM). Cell viability was measured by CellTiter-Glo® (Promega) luminescence. Viability of ibrutinib-treated cells was normalized to respective DMSO controls. Viability of TMD8-WT cells declined from 100% at 0.39 nM to 16.3% at 100 nM, with an IC50 of 2.23 nM (4-parameter logistic regression). TMD8 TRAF3-KO cell clones averaged 50% viability at all tested ibrutinib concentrations. At 100 nM ibrutinib, TMD8 TRAF3-KO clones retained significantly greater viability than TMD8-WT cells (56.7 ± 8.3% vs. 16.3 ± 0.8%; unpaired two-tailed unequal variances t-test; P=.04). IC25 values revealed an 8-fold shift from 1.78 nM in TMD8-WT to 14.33 nM in TMD8 TRAF3-KO cells, highlighting the reduced ibrutinib sensitivity conferred by TRAF3 loss. In conclusion, our clinical, genomic, and functional data collectively identified TRAF3 LOF alterations as an alternative mechanism of resistance to ibrutinib in CLL. Therapeutic strategies targeting downstream NF-κB activation merit prospective evaluation, particularly in patients who do not exhibit BTK/PLCG2 mutations with high VAF at relapse.