Immune and Genome Profiling of Myeloma Patients Treated with Sequential Immunotherapies Reveal Differential Non-Overlapping Mechanisms of Resistance

Chimeric antigen receptor T cells (CAR T) or bispecific T cell engagers (TCE) targeting multiple myeloma (MM) antigens result in deep responses in relapsed patients. Nevertheless, resistance invariably emerges requiring salvage treatments. Sequential therapies with CAR T or TCE, in patients with or without prior anti-BCMA therapy exposure, reported durable responses. However the genomic and molecular determinants of response to sequential immunotherapies are poorly defined.

To define the molecular determinants of response to sequential immunotherapies we identified patients treated with more than one anti-BCMA or -GPRC5D (n=8) regimens. Bone marrow (BM) and peripheral blood (PB) mononuclear cells from these patients were collected prior to each therapy and at the time of progression. Sorted CD138+ cells were subjected to whole genome sequencing (WGS) or single cell CNV (scCNV) and single cell RNA (scRNA) analysis. CD3+ T cells were subjected to CITEseq and functional in vitro cytotoxicity assays.

Patient 1 received anti-BCMA CAR T (Ide-cel) with VGPR lasting 7 months (mos) (Figure 1). He subsequently received elranatamab with no response. CITEseq profiling of T cells at progression post CAR T did not reveal expansion of exhausted cytotoxic or regulatory T cells. However, WGS and scCNV of MM cells identified biallelic TNFRSF17 deletion consistent with antigenic escape as cause of resistance. Patient 2 received Ide-cel (DOR = 3.5 mos) followed by elranatamab with no response. CD138+ MM cells retained BCMA expression and in vitro studies with in-house manufactured anti-BCMA CAR T efficiently eliminated his MM cells. Profiling of T cells showed terminal T cells exhaustion (CD45RA⁺, TOX⁺, TIGIT⁺, PDCD1⁺ and LAG3⁺).

In contrast to recipients of CAR T, antigenic loss was commonly observed post TCE therapy. Patient 3 received 3 sequential therapies with Ide-cel (DOR = 3 mos), teclistamab (DOR = 6 mos), and then talquetamab with daratumumab with an ongoing response of 11 mos. While T cell exhaustion was not observed, functional loss of BCMA on MM cells was detected post teclistamab, with 2 coexisting clones harboring TNFRSF17 extracellular domain mutations (p.Pro34del, or p.Ser30del) that abrogated teclistamab binding. Similarly in patient 4 progressing on elranatamab (DOR = 21 mos), there was no evidence of T cell exhaustion but rather acquired BCMA mutation (p.Pro34del) in MM cells. Consistent with fit T cells profile, he has an ongoing response (13+ mos) to talquetamab and daratumumab in his next line of therapy.

Patient 5 with penta-refractory disease and high disease burden (> 90% BM infiltration) had no response to elranatamab, however achieved an ongoing sCR (DOR = 30 mos) with talquetamab, daratumumab and pomalidomide (Tal-DP). PB T cells did not show exhaustion at the time of progression on elranatamab. A similar pattern was observed in patient 6 with high disease burden and no response to teclistamab but achieved lasting VGPR to Tal-DP (DOR=24 mos) in the next line followed by an acquired biallelic loss of GPRC5D at progression. Patient 7 had a 17 mos remission to talquetamab and daratumumab and similarly to patient 6, he acquired biallelic loss of GPRC5D at progression. He subsequently received elranatamab with no response with elevated serum soluble BCMA (sBCMA=638 ng/ml) level and phenotypic and functional evidence of T cell exhaustion. Patients 5, 6 and 7 illustrate the impact of disease burden and the ensuing sBCMA sink effect on the efficacy of anti-BCMA TCEs.

Lastly patient 8 had lasting remissions to elranatamab (DOR 13 mos) and subsequently to talquetamab (DOR 13 mos) with no evidence of T cells exhaustion but rather antigenic escape with acquired functional antigenic losses with clonal BCMA (p.Arg27Pro) and subsequently GPRC5D (p.Asp239Asn) mutation.

Highlighting the lack of cross resistance between anti-BCMA therapies, our in vitro studies showed that while mutant BCMA (p.Ser30del or p.Pro34del) abrogated teclistamab mediated T cell killing, T cells transduced with teclistamab based scFv CAR retained cytolytic activity. Therefore, TCE resistance derived from BCMA mutations does not preclude retreatment with another anti-BCMA TCE or CAR T.

We here describe variable non-overlapping mechanisms mediating resistance to sequential TCE and CAR T therapies. Dynamic surveillance for antigenic escape and functional evaluation of T cell fitness will optimize immunotherapy sequencing.