Composition and Dynamic Evolution of the Bone Marrow Microenvironment in AL Amyloidosis during Daratumumab Based Therapy

Background: Immunoglobulin light chain (AL) amyloidosis is caused by clonal B or plasma cells proliferating in the bone marrow that produce immunoglobulin (MIg) proteins that misfold and aggregate in tissue(s), leading to multi-organ failure. The monoclonal antibody targeting CD38, daratumumab is an approved and very efficient treatment in multiple myeloma and AL Amyloidosis by inducing plasma cell death through immune mediated mechanisms including complement and antibody-dependent cytotoxicity, direct apoptosis and T cell subsets modifications. While the role of the bone marrow microenvironment (BME) as a major oncogenic actor in multiple myeloma (MM) is well established, the BME composition and role in AL Amyloidosis is largely undefined. Here, we used single-cell RNA-seq combined with CITE-seq and TCR sequencing to characterize the immune ecosystem and microenvironment in AL amyloidosis in comparison to healthy donors, as well as in comparison to other plasma cell disorders including monoclonal gammopathy of undetermined significance (MGUS), smoldering multiple myeloma (SMM) and multiple myeloma (MM).

Methods: BM aspirates from 6 AL amyloidosis patients were collected at diagnosis or before treatment with daratumumab based therapy and 12 months post-treatment, alongside samples from 13 healthy donors, 5 MGUS, 9 SMM and 8 MM patients at diagnosis. We performed unbiased mRNA profiling (5’ sequencing) and feature barcoding for cell surface proteins (CITE-seq) of bone marrow mononuclear cells using 10x Genomics. We processed data via the Cell Ranger pipeline, which grouped T cells into clonotypes based on shared TCR α/β sequences. Further data processing was done with Seurat, scRepertoire and Immunarch. Overall, 83703 cells were sequenced from 45 samples, including 3065 plasma cells and 80639 BME cells.

Results: Compared to other plasma cell disorders (PCD) at diagnosis, AL BME had more cellular diversity and its plasma cells were more similar to normal plasma cells. They also had more TCR diversity than other PCD samples. When compared to healthy samples, AL BME had a trend towards more CD8+ T cells (p value = 0.078) at diagnosis, which became more pronounced 12 months post-treatment (p value = 0.012), and a trend towards fewer monocytes (p value = 0.036). When compared to PCD samples, AL BME tended to have more hematopoietic stem cells (p value = 0.018 vs MGUS, 0.052 vs SMM and 0.02 vs MM) and fewer NK cells (p values respectively 0.03, 0.059, 0.05) and CD4+ T cells (p values respectively 0.05, 0.018, 0.02) (Figure 1). The number of memory T cells and exhausted T cells were similar. Importantly, AL samples were significantly enriched in early/young CD4 memory T cells (p value < 2.6e-16) and NK/T like CD8+ T cells (p value = 7.99e-11) suggesting an ongoing and active T cell mediated immune response.

We next analyzed serial samples from 4 AL patients at diagnosis and ~12 months after Daratumumab-based therapy, with half the patients achieving complete hematologic and organ responses (CR) and half with incomplete responses. We found broad changes in the BME, but in particular, there were more CD4+ and CD8+ T cells at 12 months expressing cytokine signaling genes (p value 1.19e-35) and adaptive immune system pathway genes (p value = 2.59e-33). Importantly, the distributions of CD8+ cells and T cell clonotypes were different between patients achieving CR and those who did not, underlying the role of the immune microenvironment and especially CD8+ T cells in the pathogenesis of the disease and response to Daratumumab based treatment.

Conclusion: We identified unique characteristics of the BME in AL amyloidosis and dynamic changes within the T cell compartment following Daratumumab based therapy. Further validations are needed to evaluate the potential of BME profiling to diagnose the disease, predict treatment response and potentially develop new therapeutic strategies in AL Amyloidosis.