Deep Immunoprofiling of the Bone Marrow Microenvironmental Changes Underlying the Multistep Progression of Multiple Myeloma

Abstract

Introduction The multistep progression of multiple myeloma from a normal plasma cell to a system with the features of invasive cancer provides a unique opportunity to understand the co-evolution of the malignant clone within its microenvironment. Understanding these changes is becoming increasingly important as we attempt to design early intervention strategies and to precisely leverage emerging immunotherapeutic modalities to prevent and treat disease progression. In this work, we used mass cytometry (CyTOF) to generate a high-resolution map of the BM microenvironment and how it changes during the transition from health through pre-malignancy to disease. This approach allows us to both understand microenvironmental patterns that correlate with rapid disease progression as well as to generate new hypotheses about permissive and protective immune-phenotypes that might reveal novel immunologic drug targets.Methods To understand the immunologic characteristics of monoclonal gammopathy of undetermined significance (MGUS), smoldering multiple myeloma (SMM), newly diagnosed multiple myeloma (NDMM) and relapsed-refractory multiple myeloma (RRMM), we profiled BM aspirates from 79 patients using mass cytometry by time of flight (CyTOF). Furthermore, we compared the BM compartment of pre-malignant, malignant, and relapsed disease states to the BM of healthy donors using a 37-marker pan-immune panel. In this panel, we used antibodies against several immune lineages, tumor antigens, and functional surface markers, including co-stimulatory and co-inhibitory receptors. Cell clusters defined by Citrus analysis of CyTOF data were combined into an evolutionarily optimized decision tree by evtree to identify cluster interactions that strongly partition patient samples.Results During MGUS, when the tumor plasma cells are <10% of BM, there is little evidence of immune dysregulation; the immune compartment of MGUS patients contains normal numbers of innate and adaptive populations. In SMM, there is considerably more heterogeneity, with patients both resembling MGUS/healthy individuals and those that had changes to their immune microenvironment more consistent with newly diagnosed patients. These features include the loss of specific CD4 T cell and B cell subsets (FDR<0.1). The loss of CD4 was the most pronounced in the central memory and effector populations. This reduction in CD4 T cells is important because it diminishes help for CD8 T cell-mediated killing and immune cell maturation. Among B cell subsets, there was a loss in both mature and memory B cells. In comparison of SMM and NDMM samples, there was a clear progression from samples resembling healthy (normal B and CD4 subsets), to loss of only B cell subsets, and finally, loss of both B cell and CD4 subsets as samples diverged from healthy controls (Figure 1). In addition, a supervised machine learning analysis (evtree) identified CD4 effector memory abundance as the top node for partitioning samples into two distinct subtypes based on all available CyTOF markers across the myeloma continuum, with the high and low CD4 effector memory subtypes further subdivided by pre-B/immature B cell abundance (p=0.01), supporting these two cell types as being robust discriminators of the immune microenvironment as disease evolves. Between NDMM and RRMM samples, we observed heterogeneous loss in B cell subsets, including memory and naïve B cells. Interestingly, in RRMM we observed a strong increase in an unidentified population of CD45- cells that are quiescent (FDR<0.1), which may be stem or stromal cells. Available RNA sequencing from matching samples may reveal the lineage and function of these cells that increase during relapse.Conclusions Immune dysregulation is thought to be a major contributor to the progression and outcome of patients with MGUS, SMM, and MM. Using CyTOF, we have begun to benchmark the content of the immune microenvironment across the myeloma continuum. Based on this cross-sectional analysis we hypothesize that it is important to further interrogate whether the losses in the CD4 memory and effector populations we described correlate with outcomes after therapy with either CAR T or T cell engager trials that are currently ongoing, and whether reconstituting these cell types could provide a meaningful treatment strategy. [Display omitted] DisclosuresYoung:Celgene Corporation: Employment, Equity Ownership. Danziger:Celgene Corporation: Employment, Equity Ownership. Fitch:Celgene Corporation: Employment, Equity Ownership. Schmitz:Celgene Corporation: Employment, Equity Ownership. Gockley:Celgene Corporation: Employment. McConnell:Celgene Corporation: Employment. Reiss:Celgene Corporation: Employment, Equity Ownership. Copeland:Celgene Corporation: Employment, Equity Ownership. Newhall:Celgene Corporation: Employment, Equity Ownership. Hershberg:Celgene Corporation: Employment, Equity Ownership, Patents & Royalties. Foy:Celgene Corporation: Employment, Equity Ownership. Ratushny:Celgene Corporation: Employment, Equity Ownership. Dervan:Celgene Corporation: Employment, Equity Ownership. Morgan:Takeda: Consultancy, Honoraria; Janssen: Research Funding; Celgene: Consultancy, Honoraria, Research Funding; Bristol-Myers Squibb: Consultancy, Honoraria.