Fig. 1: Metabolite-sensing GPCRs are top hits in tumor infiltration NK-92 CRISPRa screens.
From: Engineering NK and T cells with metabolite-sensing receptors to target solid tumors
a, Experimental scheme of in vivo CRISPRa screens; i.v., intravenous; i.p., intraperitoneal; s.c., subcutaneous. b, Uniform manifold approximation and projection (UMAP) of scRNA-seq data from individuals with breast cancer. Each dot corresponds to a cell, colored by cell type (left), sample type (middle) and the overall expression of the genes identified as overexpressed in the immune cells residing in the tumor compared to blood samples (right; Methods); ILC, innate lymphoid cell; DC, dendritic cell. c, UMAP of NK scRNA-seq data from individuals with breast cancer16. Each dot corresponds to an NK cell, colored by NK subtype (left), sample type (middle) and overall expression of the NK pan-cancer tumor infiltration signature (right; Methods); Br, CD56bright; Di, CD56dim; Lo, CD16low; Hi, CD16high. d, CRISPRa breast cancer tumor infiltration screen results are shown as the significance (MAGeCK P values combined via Fisher’s method with BH correction for multiple hypothesis testing, y axis, Methods) and log-transformed fold change (FC, x axis) of each target gene (dot) in the tumor compared to the lung samples obtained from the same mouse based on MAGeCK22. Target genes whose sgRNAs were significantly or not significantly enriched in the tumor compared to lung samples are colored in red and gray, respectively; FC, fold change; NS, not significant. e, CRISPRa ovarian cancer tumor infiltration screen results shown as in d, here showing animal-matched comparison of ovarian cancer tumors to ascites samples. f, Top, log-transformed fold change of sgRNAs in breast cancer tumor versus animal-matched lung samples shown for sgRNAs targeting the indicated genes in the breast cancer in vivo screen shown in d. Bottom, log-transformed fold change of sgRNAs in ovarian cancer tumor versus animal-matched ascites samples shown for sgRNAs targeting indicated genes in the ovarian cancer in vivo screen shown in e. g, Large-scale (more than 5,500 genes) CRISPRa ovarian cancer tumor infiltration screen, showing enrichment in the tumor (MAGeCK P values, y axis) when comparing tumor to animal-matched ascites samples (left, matched per mouse) or tumor to the NK-92 cell library used (right). Panel a created in BioRender; Jerby Lab https://biorender.com/c1gnk9g (2026).
