AUTHORS
Elyad Babaei, N/A, Holly Lee, MD, PhD2, Sungwoo Ahn, PhD3, Andrea Poletti, PhD4, Mansour Poorebrahim, PhD5, David Jung, BSc6, Anja Barbour, N/A7, Motamed Mahmoud, PhD8, Sacha Benaoudia, PhD9, Noemie Leblay, PhD10, Sae Matsuoka11, Takeshi Yoroidaka, MD, PhD12, Paola Neri, MD, PhD13, Nizar Bahlis, MD14
AFFILIATIONS
1 University of Calgary, Calgary, Alberta, Canada
2 University of Calgary, Calgary, Alberta, Canada
3 University of Calgary, Calgary, Alberta, Canada
4 University of Calgary, Calgary, Alberta, Canada
5 University of Calgary, Calgary, Alberta, Canada
6 University of Calgary, Calgary, Alberta, Canada
7 University of Calgary, Calgary, Alberta, Canada
8 University of Calgary, Calgary, Alberta, Canada
9 University of Calgary, Calgary, Alberta, Canada
10 University of Calgary, Calgary, Alberta, Canada
11 University of Calgary, Calgary, Alberta, Canada
12 University of Calgary, Calgary, Alberta, Canada
13 University of Calgary, Calgary, Alberta, Canada
14 University of Calgary, Calgary, Alberta, Canada
ABSTRACT
Introduction: BCMA antigen functional loss following anti-BCMA T-cell engager (TCE) is reported to occur through either biallelic or monoallelic copy number loss coupled with nonsynonymous single nucleotide variants (SNV) and indels. Here, we leveraged a large WGS dataset of primary tumor samples from multiple myeloma (MM) patients relapsing post anti-BCMA T cell redirecting therapies to identify and functionally characterize intronic mutations in TNFRSF17, unraveling a novel mechanism of BCMA antigen-escape.
Methods: Sorted CD138+ MM cells from anti-BCMA treated patients were subjected to bulk WGS (100x). Post-sequencing analyses included SNV and indel calls by Mutect2, Strelka2, and Lancet; structural variant (SV) calls by Manta, SvABA, and Lumpy; and copy number (CN) alterations by ASCAT, CNVkit and FACET. AlphaGenome was used to prioritize intronic and regulatory-region mutations predicted to disrupt splicing. For functional validation, wild-type or intron 1 mutated TNFRSF17 sequences were cloned into PLX307 lentiviral vectors and stably transduced into the K562 cells. Alternative splice variants were validated by single-cell RNA sequencing (10X Genomics 5’ platform). Flow cytometry and western blot were used to detect BCMA protein expression. In vitro coculture assays were conducted using PBMCs with or without anti-BCMA TCEs, or in-house generated anti-BCMA CAR-T cells. Target cell viability was assessed by flow cytometry.
Results: Interrogating WGS (29 pre-treatment, 45 post-relapse) of MM patients (n=55) treated with anti-BCMA CAR-T or TCE therapy, AlphaGenome splice-impact modeling prioritized two high-confidence TNFRSF17 splice-disrupting mutations in two patients relapsing post anti-BCMA therapies. Patient MM-85 achieved a very good partial response (VGPR) lasting nearly 1 year following treatment with an anti-BCMA antibody-drug conjugate (belantamab mafadotin). At relapse, the patient disease failed to respond to anti-BCMA TCE (teclistamab). WGS of CD138+ cells from this patient demonstrated no copy number alteration at the TNFRSF17 locus but a clonal intronic TNFRSF17 variant, chr16:11965459G>A (mut1, CCF: 1), adjacent to the exon1-intron1 junction. Patient MM-35 achieved a VGPR lasting over 3 years following anti-BCMA CAR T therapy. At relapse, WGS identified three distinct TNFRSF17 mutant clones, including intronic chr16:11966194G>A (mut2, CCF: 0.14), adjacent to the intron1-exon2 junction, as well as two coding-region alterations, p.Glu12Lys (CCF: 1) and p.Arg27Ter (CCF: 0.35). AlphaGenome analysis of mut1 and mut2 strongly predicted BCMA splice variant effects (scores 7.28 and 2.28, respectively; both >99.99th percentile) and were also independently confirmed by SpliceAI to disrupt BCMA canonicsl transcript processing.
Functional validation of mut1 and mut2 in K562 cells identified aberrant splice isoforms arising from both mutant alleles. Mut1 disrupted the native splice donor site at the 3’ junction of exon 1 and activated a cryptic donor site within intron 1 (chr16:11965547), whereas mut2 disrupted the native acceptor site at the beginning of exon 2 and activated two cryptic acceptor sites within exon 2 (chr16:11966195 and chr16:11966272), generating two alternatively splice isoforms. BCMA mut1 and the first splice variant of mut2 resulted in frameshift leading to premature BCMA protein truncation, while the second splice variant of mut2 resulted in in-frame truncation of BCMA transmembrane domain. Flow cytometry analysis failed to detect BCMA surface expression in both mutant conditions.
In coculture assays with teclistamab, both mut1 and mut2 conferred marked resistance compared with WT BCMA-expressing cells (viability 81.86% and 87.02% vs. 15.77%, respectively, p < 0.01). However, in anti-BCMA CAR T coculture assays, mut1 remained sensitive, whereas mut2 demonstrated resistance compared with WT BCMA-expressing cells (viability 12.47% vs. 86.07% vs. 4.81%, respectively, p < 0.01).
Conclusions: Unlike canonical escape mechanisms involving BCMA biallelic deletion or coding-region mutations, we here identified novel intronic TNFRSF17 mutations that disrupt BCMA expression through cryptic splice-site activation and transcript reprogramming. These findings expand the genetic landscape of BCMA resistance and underscore the importance of interrogating non-coding alterations in patients relapsing after targeted immunotherapies.
DISCLOSURES
Holly Lee, MD, PhD: No relevant conflicts of interest to declare.