American Society of Hematology

Figure 1.

$Generation and validation of a CML BM atlas at the single-cell resolution. (A) Diagnostic BM aspirates from normal controls and a cohort of patients with CP CML with differential imatinib responses were subjected to multimodal single-cell analysis using scRNA-seq. The patients with CP CML were classified, broadly based in line with European LeukemiaNet recommendations (supplemental Methods), as optimal responders to imatinib (group A), suboptimal responders to imatinib (group B), or as treatment failures if they progressed to BC (group C) (refer to supplemental Table 1 for patient details). The resulting single-cell atlas was subjected to data analysis pipelines to uncover features with predictive potential at the time of diagnosis. Mass cytometry, flow cytometry, and functional analyses were performed to validate the key predictive features. (B, left) The CD34+ and CD34− fractions were subjected to scRNA-seq analysis and analyzed as described in the supplemental Methods. The 4 major cell types from the scRNA-seq data set, namely B cells, myeloid cells, T and NK cells, and HSPCs were first identified. (B, right) The 4 major cell types were subclustered and cluster markers/GE signatures were used to identify 10 HSPC and 22 differentiated populations (8 myeloid; 7 B cell, and 7 T/NK metaclusters or cell types). (C) UMAP of single-cell transcriptomes, colored by control or CML cells. (D) HSPC subpopulations are plotted as a proportion of CD34+ cells. Each dot represents 1 patient. (E) T, NK, B, and myeloid cell types are plotted as a proportion of CD34− cells. (F) The 10 HSPC metaclusters were aggregated into a single pseudobulked population (CD34+ HSPC). GSEA comparisons were performed between groups A and C using gene sets curated in supplemental Table 12. Enrichment plots for differentially enriched gene sets are shown (false discovery rate q-value ≤0.1). (G) DNTT (TDT) expression in pseudobulked CD34+ cells of group C samples that transformed into lymphoid BC was contrasted against all remaining CML samples. A 2-tailed t test was used to test the statistical significance of data in Figure 1C-D,G. Supplemental Figures 1-3 and supplemental Tables 1-13 are linked with data shown in Figure 1. C Mono, classical monocytes; cDC2, classical dendritic cells; CD4+ TCM, CD4+ T central memory; CD8+ TEMRA, CD8+ T effector memory with CD45RA; EOBM, eosinophil-basophil mast cell progenitor; iMKP, immature megakaryocyte progenitor; I Mono, inflammatory monocytes; Lyp, lymphoid progenitor; Mono-T, monocyte-T; MKP, megakaryocyte progenitor; γδ T cells, T cells.$

Generation and validation of a CML BM atlas at the single-cell resolution. (A) Diagnostic BM aspirates from normal controls and a cohort of patients with CP CML with differential imatinib responses were subjected to multimodal single-cell analysis using scRNA-seq. The patients with CP CML were classified, broadly based in line with European LeukemiaNet recommendations (supplemental Methods), as optimal responders to imatinib (group A), suboptimal responders to imatinib (group B), or as treatment failures if they progressed to BC (group C) (refer to supplemental Table 1 for patient details). The resulting single-cell atlas was subjected to data analysis pipelines to uncover features with predictive potential at the time of diagnosis. Mass cytometry, flow cytometry, and functional analyses were performed to validate the key predictive features. (B, left) The CD34⁺ and CD34⁻ fractions were subjected to scRNA-seq analysis and analyzed as described in the supplemental Methods. The 4 major cell types from the scRNA-seq data set, namely B cells, myeloid cells, T and NK cells, and HSPCs were first identified. (B, right) The 4 major cell types were subclustered and cluster markers/GE signatures were used to identify 10 HSPC and 22 differentiated populations (8 myeloid; 7 B cell, and 7 T/NK metaclusters or cell types). (C) UMAP of single-cell transcriptomes, colored by control or CML cells. (D) HSPC subpopulations are plotted as a proportion of CD34⁺ cells. Each dot represents 1 patient. (E) T, NK, B, and myeloid cell types are plotted as a proportion of CD34⁻ cells. (F) The 10 HSPC metaclusters were aggregated into a single pseudobulked population (CD34⁺ HSPC). GSEA comparisons were performed between groups A and C using gene sets curated in supplemental Table 12. Enrichment plots for differentially enriched gene sets are shown (false discovery rate q-value ≤0.1). (G) DNTT (TDT) expression in pseudobulked CD34⁺ cells of group C samples that transformed into lymphoid BC was contrasted against all remaining CML samples. A 2-tailed t test was used to test the statistical significance of data in Figure 1C-D,G. Supplemental Figures 1-3 and supplemental Tables 1-13 are linked with data shown in Figure 1. C Mono, classical monocytes; cDC2, classical dendritic cells; CD4⁺ TCM, CD4⁺ T central memory; CD8⁺ TEMRA, CD8⁺ T effector memory with CD45RA; EOBM, eosinophil-basophil mast cell progenitor; iMKP, immature megakaryocyte progenitor; I Mono, inflammatory monocytes; Lyp, lymphoid progenitor; Mono-T, monocyte-T; MKP, megakaryocyte progenitor; γδ T cells, T cells.

This Feature Is Available To Subscribers Only