Background

Multiple myeloma (MM) is a hematological malignancy originating in the bone marrow, characterized by progressive immune disruption across stages—from precursor conditions such as Monoclonal Gammopathy of Undetermined Significance (MGUS) and Smoldering MM (SMM) to Newly Diagnosed MM (NDMM), Relapsed/Refractory MM (RRMM), and post-CAR-T therapy. The bone marrow microenvironment includes diverse immune and stromal cells that influence disease behaviour and therapy response. While prior studies profiled subsets of these populations, a unified single cell view across stages has been lacking. We developed MyScAPE (Myeloma Single Cell Atlas for Precision Exploration), a curated resource integrating available single cell RNA-seq (scRNA-seq) datasets.

Methods

We curated and harmonized scRNA-seq datasets from a range of repositories, including the GEO database, EMBL-EBI, European Genome-Phenome Archive, and Human Cell Atlas. All data were reprocessed through a standardized workflow to ensure batch-corrected integration, and cell types were annotated using clustering, label-transfer, and manual verification, resulting in over 50+ phenotypes. Sample-level cell type abundances were quantified, and statistical comparisons across stages were performed using Wilcoxon rank-sum tests (significance: p < 0.05). MyScAPE is accessible through an interactive web application developed using Dash in Python, facilitating intuitive exploration and visualization of cellular phenotypes and gene expression patterns across MM stages.

Results

We integrated data from over 500 samples, totalling ~4.5 million cells across MM stages: precursor (MGUS, SMM), NDMM (standard- and high-risk), RRMM, post-CAR-T, and healthy bone marrow. MyScAPE enables interactive exploration, visualization, and comparative analysis of genes, across cytogenetics, pathways, transcription factors, cell-cell interactions, and cell phenotypes without requiring computational expertise and highlights key immune and tumour microenvironment changes across MM stages. Within this comprehensive atlas, focused analysis of T-cells revealed coordinated remodelling across MM progression. Among CD4+ T-cells, KLRB1+ central memory cells show progressive decline from precursor to NDMM-standard risk (SR) and increased again in NDMM-high risk (HR), suggesting progressive loss and rebound of central memory phenotypes. Inflammatory subsets of CD4+ memory T-cell subsets, enriched for interferon-related genes and NF-κB-associated gene expression, showed increased abundance from the precursor stages to NDMM. In contrast, while CD4+ TNF+ effectors, a pro-inflammatory effector subset, showed a progressive decline from precursor to NDMM. On the other hand, analysis of Cytotoxic T-cells depicted progressive enrichment of effector memory CD8+ T-cells from MGUS, SMM, NDMM, and RRMM, reflecting chronic activation and cytotoxic pressure. Naïve phenotypes such as CD8+ CCR7+ T-cells and CD8+ GZMK+ central memory cells were depleted in precursor stages through NDMM and showed partial reconstitution from RRMM to post-CAR-T. Additionally, MAIT cells, an innate-like subset, exhibited a progressive decline from precursor stages through NDMM, indicating early attrition of this population.

Conclusions

MyScAPE provides the most comprehensive single-cell map to date of the bone marrow immune landscape across multiple myeloma disease stages. Our analysis highlights dynamic and stage-specific remodelling of T cells, including effector, memory, apoptotic, and innate-like populations through progression from precursor states to post-CAR-T therapy. These patterns reveal immunologic shifts that may underline disease risk, treatment resistance, and therapeutic response. By making the MyScAPE Atlas and its interactive exploration portal publicly available, we aim to accelerate discovery, hypothesis testing, and translational research across the Myeloma community. MyScAPE is available online at https://bhasinlab.bmi.emory.edu/MMAtlas/.

This content is only available as a PDF.
Sign in via your Institution