Fig. 3

T/NK cell remodeling after therapy. A UMAP plot of T/NK cells colored by clusters. B Heatmap of normalized expression of canonical T/NK cell marker genes among clusters. TRM, tissue-resident memory. C Boxplots of the average cytotoxic and exhausted signature scores for CD8+ T cells in TN (n = 3), MPR (n = 4), and NMPR (n = 8) patients. Center line indicates the median, lower, and upper hinges represent the 25th and 75th percentiles, respectively, and whiskers denote 1.5× interquartile range. One-sided t-test was used. D Boxplots of the average cytotoxic and exhausted signature scores for NK cells in TN (n = 3), MPR (n = 4), and NMPR (n = 8) patients. One-sided t-test was used. E Boxplot showing cellular fractions of each T/NK cluster in TN (n = 3), MPR (n = 4), and NMPR (n = 8) patients. All differences with adjusted P < 0.10 are indicated. One-sided unpaired Wilcoxon test was used and the P values were adjusted by the FDR method. F Summary of selected ligand-receptor interactions from CellPhoneDB between cancer cells and CD16+ NK cells in MPR patients. G Boxplots of the average exhausted signature scores for Tregs in TN (n = 3), MPR (n = 4), and NMPR (n = 8) patients. One-sided t-test was used. H The developmental trajectory of CD8+ T cells inferred by Monocle2. The memory CD8+ T cells (CD8_IL7R) and effector memory (CD8_GZMK) T cells were the roots of differentiation, and the exhausted CD8+ T cells (CD8_HAVCR2) were in the end-point state. I Heatmap of the top differential genes in memory (CD8_IL7R) cells along the pseudo-time (lower panel). The distribution of CD8_IL7R cells during the transition (divided into 2 phases: resting and activated) in TN, MPR, and NMPR patients, along with the pseudo-time (upper panel)