Figure 3 from Composition and Functional State of T and NK Cells in the Extramedullary Myeloma Tumor Microenvironment

Levels of cytotoxicity, exhaustion, and neoantigen reactivity scores suggesting direct cell–cell interactions between EMM cells and CD8+ T cells. A, Violin plots indicating the T-cell cytotoxicity score in CD8+ T cells with marked divisions to classify samples by their sample types. The scores were computed using a curated gene list (35) and scRNA-seq data. Paired samples in the EMM and EMM_BM cohorts are highlighted in the same color (left). Box plot of cytotoxicity scores calculated using the same set of genes and estimated with pseudo-bulk and ssGSEA (right). B, Exhaustion score in CD8+ T cells with marked divisions to classify samples by their types (left). Box plot of exhaustion scores calculated with pseudo-bulk and ssGSEA, using scRNA-seq data (right). Figures are represented in the same format as in (A). C, Spearman correlation between exhaustion and cytotoxicity scores calculated from pseudo-bulk in scRNA-seq EMM samples. D, Box plot indicating the composition of cell subtypes in the CD8+ T-cell compartment. Each color here represents a sample type as mentioned in the legend. E, Dot plot showing average cytotoxicity and exhaustion scores and neoantigen reactivity in CD8+ T cells across the subclusters from all samples [EMM (N = 7), EMM_BM (N = 5), RRMM_BM (N = 6)], using scRNA-seq data. F, Box plot showing the proportion of cells with positive expression of the PDCD1 (PD-1) gene relative to all the T cells by FCM. Predicted differential cell–cell interactions (G) and expression of HLA-B, HLA-A, and HLA-C (H–J) between patients with high (N = 3) and low (N = 3) levels of T-cell exhaustion, using scRNA-seq data. Box plots display the median (center line; also represented in the volcano plot), the 25th and 75th percentiles (box limits), and whiskers extending to the most extreme data points within 1.5× the IQR. Statistical comparisons were performed using the Wilcoxon rank-sum test with Benjamini–Hochberg correction for multiple testing. n.s., not significant; **, P < 0.05.