Single-Cell RNA-Seq of T Cells in B-ALL Patients Reveals an Exhausted Subset with Remarkable Heterogeneity

Abstract

Characterization of functional T cell clusters is key to developing strategies for immunotherapy and predicting clinical responses in leukemia. Here, single-cell RNA sequencing is performed with T cells sorted from the peripheral blood of healthy individuals and patients with B cell-acute lymphoblastic leukemia (B-ALL). Unbiased bioinformatics analysis enabled the authors to identify 13 T cell clusters in the patients based on their molecular properties. All 11 major T cell subsets in healthy individuals are found in the patients with B-ALL, with the counterparts in the patients universally showing more activated characteristics. Two exhausted T cell populations, characterized by up-regulation of TIGIT, PDCD1, HLADRA, LAG3, and CTLA4 are specifically discovered in B-ALL patients. Of note, these exhausted T cells possess remarkable heterogeneity, and ten sub-clusters are further identified, which are characterized by different cell cycle phases, naïve states, and GNLY (coding granulysin) expression. Coupled with single-cell T cell receptor repertoire profiling, diverse originations of the exhausted T cells in B-ALL are suggested, and clonally expanded exhausted T cells are likely to originate from CD8⁺ effector memory/terminal effector cells. Together, these data provide for the first-time valuable insights for understanding exhausted T cell populations in leukemia.

Keywords: B cell-acute lymphoblastic leukemia; T cells; heterogeneity; single-cell RNA sequencing.

Figure 1

Transcriptomic clustering of CD45⁺CD3⁺ T cells from healthy individuals and B‐ALL patients. A) Schematic of the experimental design for scRNA‐seq. Peripheral blood mononuclear cells were collected from 2 healthy individuals and 3 B‐ALL patients. T cells with the immunophenotype CD45⁺CD3⁺ were sorted using Aria III and subjected to scRNA‐seq combined with single‐cell TCR sequencing using a 10×‐Based Genomics platform. B) UMAP projection of eleven T cell clusters of healthy individuals (left, 12 699 cells in total) and thirteen cell clusters of B‐ALL patients (right, 16 143 cells in total). Each dot corresponds to one single cell colored according to cell cluster. C) Dot plots show the average expression levels and cell expression proportions of selected nineteen T cell marker genes in the indicated clusters for healthy individuals (left) and patients (right). The colors represent the average expression levels, and dot sizes represent the percentage expression of selected genes in the indicated clusters. D) Stacked bar chart showing the constitution of thirteen T cell clusters in each healthy individual (abbreviated as “h1” and “h2”) and B‐ALL patient (abbreviated as “p1,” “p2,” and “p3”). E) Heatmaps showing the pairwise Pearson correlation coefficients of eleven or thirteen T cell clusters for HIs (left) and B‐ALL patients (right). The clusters were ordered by hierarchical clustering the corresponding correlation matrix.

Figure 2

Integrated analysis of T cells in healthy individuals and B‐ALL patients. A) UMAP showing the merge of eleven and thirteen T cells from healthy individuals and B‐ALL patients. Each dot corresponds to one single cell colored according to cell cluster. B) Dot plot showing the similarity of the eleven and thirteen T cell clusters in healthy individuals and those in B‐ALL patients. The colors and sizes are scaled to the pairwise Pearson correlation coefficient. C) UMAP plots showing the three cell cycle phases of the thirteen T cell clusters in B‐ALL patients (left) with the percentage of proliferating cells in the different B‐ALL clusters (right). The percentages are shown to the right. D) Circos plot showing the relationships of overlapping DEGs among the indicated T cell clusters. DEGs were identified by comparing B‐ALL patients to healthy individuals for each cluster. The purple lines link genes shared by two clusters. Genes found on multiple lists are colored in dark orange, and genes unique to a list are shown in light orange. E,F) Heatmaps showing the top enriched terms by Metascape enrichment analysis for DEGs (upregulated pathways shown in (E), downregulated pathways shown in (F) identified by comparing matched clusters between healthy individuals and B‐ALL patients. The heatmap cells are colored by their p values, grey cells indicate the lack of enrichment for that term in the corresponding gene list. P value < 0.05 was considered statistically significant.

Figure 3

Dissection of exhausted T cell clusters in B‐ALL patients. A) UMAP showing 10 sub‐clusters of exhausted T cells (left). The plot to the right shows the cell cycle phases of the exhausted T cells. B) Heatmap showing expression of top 5 DEGs across clusters in the exhausted T cell clusters. C) 3D‐scatter plot illustrating three feature scores (naive, cytotoxicity, and exhaustion) for all T cell clusters. Each circle represents one T cell cluster, and each triangle represents an exhausted T cell cluster. D) UMAP plots showing the expression of five selected genes (CD4, CD8A, GNLY, TIGIT, and CCR7) in exhausted T cells. E) Venn diagram showing the overlap of presumed exhaustion marker genes identified by using CD8 exhausted cells (pale green) and CD4 exhausted cells (orange) compared to corresponding control cells. See Figure S3F, Supporting Information and Experimental Section for details. F) Dot plots show the average expression levels and cell expressing proportions of selected exhausted T cell signature genes in the indicated clusters. The colors represent the average expression levels, and dot sizes represent the percentage expression of selected genes in the indicated clusters.

Figure 4

TCR repertoire analysis of T cells in B‐ALL patients. A) UMAP plot showing the concentration of clonal expansion in B‐ALL patients. The colors correspond to three TCR clonal expansion groups where “> = 3” indicates that there were at least three T cells that express identical TCR clonotype, “2” indicates that there were two T cells that express an identical TCR clonotype, and “1” indicates there was only one T cell that expresses a specific TCR clonotype. B) Volcano plot showing genes differentially expressed (see Table S3, Supporting Information) between highly expanded clones (“> = 3” group) and non‐clonal cells (“1” group). Highly expressed genes are shown in red, and genes with low expression are shown in blue. The top 10 most significant DEGs are also indicated. C) Dot plots show the average expression levels and cell expressing proportions of the top 8 highly expressed genes in the highly expanded group compared to the non‐clonal group. The colors represent the average expression levels, and dot sizes represent the number of expressed genes within a group. D) Top 20 most enriched terms for the up‐regulated DEGs in the highly expanded compared with non‐clonal group. The bars are colored by their p values, and p value < 0.05 was considered statistically significant. E) Bar charts showing the cellular proportion of three clonal expansion groups in each T cell cluster (including exhausted T cell sub‐clusters) in B‐ALL patients. F) Heatmap showing the Jaccard‐index between the respective pair of clusters in B‐ALL patients. For each tile, the brighter the color, the greater number of overlapping clonotypes between two corresponding clusters. G) Heatmap showing the distribution of shared clonotypes between exhausted and non‐exhausted T cell clusters in B‐ALL patients. Each row represents a shared clonotype and each column represents a T cell cluster. The colored squares are proportional to the number of cells with the same clonotype in the corresponding cluster.