P-346: Internal Tandem Duplications of the BCMA Transmembrane Domain are Common in Hyperdiploid Multiple Myeloma and Associated with Constitutive NF-kB Signaling

Introduction: Recent reports have highlighted a number of genetic resistance events in patients who have progressed on BCMA and GPRC5D targeted immunotherapies. We have leveraged the multi-dimensional data in the MMRF CoMMpass study to define the incidence of target allele loss in patients. Moreover, the usage of multiple advanced mutation callers identified recurrent internal tandem duplications (ITD) in TNFRSF17/BCMA reminiscent of the FLT3 ITD seen in AML.

Methods: Whole genome, exome, and RNA sequencing results from IA22 of the MMRF CoMMpass study were processed through an updated pipeline with alignment to the GRCh38 reference genome. Small variant calling was performed with 5 independent variant callers; Mutect2, Strelka2, Octopus, VarDict, and Lancet. Copy number was determined using GATK CNV with an optimized diploid centering model, structural events were identified using Manta and gene expression estimates were determined using Salmon.

Results: Integration of the available gene expression, copy number and mutational data facilitated the comprehensive evaluation of the current antibody, bispecific or CAR-T target antigens. At diagnosis all patients expressed the target genes and none had bi-allelic loss of a target loci, however, single copy loss was detected for FCRL5 (0.8%), CD38 (5.6%), GPRC5D (13.2%), and BCMA (3.4%). Four hyperdiploid patients had very high expression of BCMA, which were associated with high level amplifications (5, 12, or 30 copies) or an IgH translocation. High confidence mutation calls (>=3 callers) identified eight patients with SNV or INDEL mutations in BCMA. Two patients had inactivating nonsense or frameshift mutations while the remaining five had large 37-58 nucleotide ITD. Since large insertions are one of the most difficult events to identify we curated the single variant caller results and identified three additional patients with ITD of 46-73 nucleotides. Interestingly, all of these patients have a hyperdiploid karyotype and the ITD occur within the transmembrane domain with a minimal duplication of amino acids 61-66. The ITD are predicted to be gain-of-function events, as each patient was characterized by a high NF-kB gene expression index, exceeding the 90th percentile observed in the cohort in all cases.

Conclusions: As we begin to prioritize the use of different immune-based therapies it will be important to consider the relative risk each patient has to develop resistance based on the preexisting loss of one target gene copy, as some targets are frequently lost (GPRC5D) while others are rarely lost (FCRL5). Some hyperdiploid patients have hyperactivation of NF-kB signaling as a result of BCMA overexpression by high level copy number gains and immunoglobulin translocations or ITD. These patients are likely dependent on BCMA mediated signaling and would likely be exquisitely sensitive to BCMA targeted therapies. It will be essential to determine if current BCMA directed therapies can still bind these BCMA ITD isoforms.