P-249: Spatial transcriptomics unveils novel disease mechanisms associated with the microenvironment in Multiple Myeloma

Introduction: Single-cell sequencing has been the major technology used to understand the transcriptional features of malignant plasma cells (MM-PC) responsible for developing Multiple Myeloma (MM). These findings underscore the significant genetic and genomic heterogeneity both between patients and among individual patients. The spatial distribution in the tissue of the different cellular types is hypothesized to be critical to the niche-specific regulatory programs involved in the tumor's relapse, drug resistance, or aggressiveness.

Methods: To elucidate the Bone Marrow`s (BM) topographical distribution and cellular interactions of BM components, we performed Visium Spatial Gene Expression analysis (10x Genomics) in YFP_cγ1 (healthy) and MI_cγ1 mice strains (a model that recapitulates the principal characteristics of human MM), and in clinical human healthy and MM samples. The formalin-fixed paraffin-embedded (FFPE) BM tissues were decalcified and transcriptionally characterized, applying deconvolution analysis with CARD based on public and own single-cell RNA-seq data to estimate cell composition per spot. Top marker genes associated with MM-PC were used to locate the areas enriched in this cell type. Besides, to define the MM-PC heterogeneity, a pseudo-bulk-based analysis identified the most variable genes and pathways between different groups of MM-PC. To better understand the importance of T cells in MM samples, we performed a targeted analysis looking for specific profiles and additional microenvironmental signatures in the disease.

Results: Significant differences between healthy and pathological samples were determined using cell deconvolution analysis. After identifying the top marker genes associated with MM-PC in our model, different areas, based on the proportions of MM-PC, were delineated. We differentiate between hotspots (areas with high MM-PC scores), border zones (areas surrounding the Hotspots), and remote zones (low MM-PC scores). Interestingly, different distributions of T cell effector and exhaustion profiles was observed in the border zone and hotspots upon investigating the defined MM-PC density gradient. Additionally, an increase in neutrophil extracellular traps (NETs) pathway, typically observed in solid tumors, was observed within the remote and border zones of MM samples.The application of spatial transcriptomics to FFPE human BM biopsies, healthy and MM (with varying degrees of PC infiltration), confirmed the presence of the equivalent areas and T-cell profiles observed in the mouse model.

Conclusions: Our findings demonstrate that spatially resolved transcriptomics provides a systematic approach to establishing unbiased gene expression within a spatial context. We identified different T cell profiles in MM and patients. This approach may have broad future potential applications in understanding the performance of advanced therapies such as CAR T cells.