De Novo Resistance and Relapse from Daratumumab Monotherapy in NDMM Is Associated with Immune Evasion and Immunosuppression

Introduction/Methods

De novo and acquired resistance associated with daratumumab (dara) monotherapy in NDMM was assessed by 3’ scRNAseq of 99k iTME and 160k tumor cells (10x Genomics, n=25, min. 50k reads/cell) and flow cytometry (FC, 36 lymphoid and 19 myeloid markers, n=67). Tumor cells were enriched from bone marrow aspirates from 11 patients prior to (D0), after two cycles of Dara (D60), and at relapse by CD138 selection; total iTME cells were analyzed at all timepoints. Malignant plasma cells (PC) and subclones were identified by comparison to healthy donor PCs (n=6) using inferCNV. Tumor response to therapy was defined by at least a partial response (responder [R, n=5 scRNAseq, n=12 FC] vs. non-responder [NR, n=6 scRNAseq, n=18 FC]). Please see companion clinical abstract for trial NCT04151667 details.

Results

Differentially expressed genes (DEG, >1.5-fold change, padj<10^-30) were identified at D0 (de novo resistance). CD38, IFN response (MX1, ISG20, IFI27, IFI6), secretion (PPIB, UAP1, SYNGR1, CLPTM1L, JSPR1), antigen presentation (HLA-B), translation (TENT5C, PABPC4, CSDE1, RSP12, RPS4Y1, RRBP1, RPS6, RPS17), and PC survival (AREG, TNFRSF13C, CCR10) genes had higher expression in R at D0 vs NR, suggesting that PC biology is important for dara response. Contrastingly, NR tumor cells had increased DEG necessary for memory B cell biology (MS4A1, LAPTM5, CD79A), anti-apoptosis (BIRC3, ATF5), proliferation (CCND1, TIMP1, BTG1, FAM107B), tumor growth/migration (MALAT1, LTB, TMSB10), adhesion (TMSB4X, NFKBIA, CST3, FTH1, ZFP36L2, CD44, IDS), and immunosuppression (LGALS1); while tumor suppressor (DERL3, ZBTB38) and anti-proliferative (GADD45A, BTG2) DEG were decreased (>1.5-fold change, padj<10^-30). Results from GSEA largely mirrored DEG differences (p<0.05, q<0.25). Myc Targets V1, Proteasome, SRP Dependent ER Translocation, UPR, and Translation pathways were higher in R tumor cells. These pathways are associated with PC biology, so it is not surprising that PC pathways were also increased in R, along with Apoptosis via Trail, Glycolysis, E2F targets, and several Type I and II IFN pathways. NR had upregulated Focal Adhesion, TNF Response, KRAS Signaling, oncogenesis, Memory B Cells, and BCR Signaling pathways.

Already by D60, R MM upregulated pathways found in NR at D0. Focal Adhesion, Actin Cytoskeleton, KRAS Signaling, TNF Response, Memory B Cells, and BCR Signaling were all increased in R during treatment (p<0.05, q<0.25). These changes were also observed in DEGs. The most upregulated gene in R was the pro-survival MIF receptor, CD74, at D60 compared to D0; the most downregulated genes in R at D60 were related to IFN response (ISG15, IFI27, IFI6, and MX1) were also lower in NR tumors at D0 (>1.5-fold change, padj<10^-30). Inflammatory cytokine production (MAPKAPK2), pro-apoptotic (TNSF10), tumor suppressor (BTG2), glycolysis (PDK1), and translation (UAP1, TENT5C) genes were also decreased in R tumors. CRIP1, homing (CXCR4), immunosuppression (LGALS1, AHNAK), oncogenesis (CD44, TMSB10, KLF6, EMP3), anti-apoptosis (BIRC3), and memory B cell (LAPTM5) genes were upregulated in R at D60 and NR at D0. To examine intratumoral dynamics, clones were identified using InferCNV. Clonal changes were more dynamic in R than in NR. Molecular signatures identified by inferCNV reflected clinical FISH. Pathways and DEG in expanding clones from R reflected those in NR at D0, demonstrating resistance arises from increased fitness of preexisting clones.

Examination of the iTME showed enrichment of mMDSC and eTreg in NR; and NK and CD4EM in R populations at D0 was observed (p<0.1). CD4, CD8, and NK populations had increasing expression of exhaustion markers from D0 to D60 in NR (p<0.1). Pathway changes in R iTME populations from D0 to D60 resembled NR at D0, indicating movement to more immunosuppressive phenotypes. Tumor-iTME communication was assessed by CellPhoneDB and demonstrated D0 R MM interactions included more cytotoxic cells compared to NR, which had increased immunosuppressive iTME interactions.

Conclusion

For the first time, these results demonstrate de novo resistance mechanisms selected during dara monotherapy in NDMM that translate to relapse. These tumor strategies can be attributed to intratumoral heterogeneity, tumor survival mechanisms, and an immunosuppressive iTME. Their discovery may be used to better allocate tumor targeting, monoclonal antibody-based therapies.