Panel A shows a uniform manifold approximation and projection (UMAP) plot of human breast cancer myeloid cell clusters. Panel B shows violin plots and a uniform manifold approximation and projection (UMAP) signature plot of inflammatory monocytes. Panel C shows Kaplan-Meier survival curves for inflammatory monocyte and CXCL10 macrophage signatures in patients. Panel D shows survival curves and forest plots evaluating inflammatory monocyte signatures with clinical variables and hazard ratios. Panel E shows a uniform manifold approximation and projection (UMAP) plot, dot plot, and signature maps of breast cancer myeloid clusters. Panel F shows uniform manifold approximation and projection (UMAP) plots and box plots comparing responders and non-responders before and during treatment. Panel G shows box plots of single-sample gene set enrichment analysis scores in responders and non-responders. Panel H shows box plots of inflammatory monocyte signature scores before and during treatment in patient groups.
CD1d-associated gene signature correlates with clinical outcomes and response to immunotherapy in breast cancer patients. (A) UMAP plot of myeloid cell clusters (filtered for monocytes, TAMs, DCs, and pDCs) from human breast cancers (Wu et al., 2021). Each point represents a single cell colored according to cluster designation. (B) Left: Enrichment score for the mouse c3_Inflam-Mono signature in each of the human myeloid clusters (Wu et al., 2021). Right: Expression levels for the murine c3_Inflam-Mono signature displayed on UMAP plot of human myeloid clusters (as in Fig. 5 A). (C and D) Survival analyses of TNBC patients (METABRIC), stratified according to mouse c3_Inflam-Mono signature, human c9_CXCL10-Mac signature (C), or mouse DEG- c3_Inflam-Mono signature (D). Kaplan–Meier survival plots parsed as high versus low based on the best cutoff. Log-rank (Mantel–Cox) test. HRs and 95% confidence intervals were estimated using Cox proportional hazards regression. (D) Right: Forest plots showing a multivariate Cox regression analysis for the indicated risk factors. HRs with 95% confidence intervals are shown. (E) UMAP plot of myeloid cell clusters from breast cancer patients treated with αPD-1 neoadjuvant therapy (Bassez et al., 2021). Each point represents a single cell colored according to cluster designation. Middle: Dot plot showing expression for selected genes in each of the specified myeloid clusters. Right: Expression levels for the mouse c3_Inflam-Mono signature and human c9_CXCL10-Mac signature displayed in the UMAP plot for myeloid cells for human breast cancer patients (Bassez et al., 2021). (F) UMAP plot of myeloid cell clusters from breast cancer patients before (pretreatment) or after (on-treatment) αPD-1 neoadjuvant therapy (Bassez et al., 2021). Data are split on patients that respond (R, red) or not (NR, gray) to αPD-1. Bar plots (right) show the frequency of cluster 0 cells in R and NR samples (Wilcoxon rank-sum tests). Each dot is a sample from a different donor. (G and H) Enrichment score (ssGSEA) for mouse c3_Inflam-Mono signature and human c9_CXCL10-Mac signature in human myeloid clusters from Bassez et al. (2021) as identified in Fig. 5 E (G) or mouse DEG-c3_Inflam-Mono signature in cluster 0 (H). Data are shown for samples before (pretreatment) or after (on-treatment) αPD-1 neoadjuvant therapy and split for patients that respond (R, red) or not (NR, gray) to αPD-1 (Wilcoxon rank-sum tests). Each dot is a sample from a different donor. HRs, hazard ratios; R, responders; NR, nonresponders.