Panel A shows a volcano plot of bulk RNA-seq data for HT29 tumor cells treated with recombinant human TNF for 6 hours. The x-axis represents log2 fold change, and the y-axis represents minus log10 p-value. Genes upregulated in untreated cells are in blue, while those upregulated in TNF-treated cells are in red. Pro-tumor inflammatory mediators are highlighted in purple. Panel B is similar to Panel A but shows data for HT29 cells treated for 24 hours. Panel C is a volcano plot comparing mRNA expression in TNFR1high and TNFR1negative/low melanoma cells from three patients. The x-axis is log2 fold change, and the y-axis is minus log10 p-value. Proinflammatory genes are highlighted in purple. Panel D includes a bar graph and violin plots showing gene expression distribution for proinflammatory genes in TNFR1high and TNFR1negative/low cells from two NSCLC cases. The bar graph's y-axis is expression level, and the violin plots show distribution across conditions. Panel E is a heatmap of Pearson correlation coefficients between TNFR1 transcripts and proinflammatory genes across various cancer datasets. The color scale indicates correlation strength. Panel F shows a schematic of an experiment in immunodeficient mice xenografted with TNFR1KO or WT HT29 cells, with bar graphs of intratumoral mouse myeloid cell content. The y-axis represents cell count per milligram of tumor. Panel G shows bar graphs of multiplex tissue immunofluorescence analyses of excised tumor samples, with bar plots of cell counts per square millimeter. Panel H displays representative microscopy images of tumor samples, with a scale bar of 50 micrometers.
TNFR1 on human tumor cells mediates pro-tumor inflammation. (A and B) Volcano plots of bulk RNA-seq, as in Fig. 4 performed on HT29 tumor cells in culture, either exposed or not to recombinant human TNFα, highlighting pro-tumor inflammatory mediators in purple (n = 3). (C) Volcano plot representing a comparison of expression of mRNA in the malignant melanoma cells (scRNA-seq) from three cases of primary resections of skin melanoma that were in silico sorted into TNFR1high and TNFR1negative/low. Genes of the proinflammatory signature identified in A and B are highlighted in purple. (D) Bar plot (upper panel) and violin plots (lower panel) representing the gene expression distribution for the main genes of the proinflammatory signature across TNFR1high and TNFR1negative/low cells in two resected NSCLC cases, again sorted in silico. (E) Heatmap representing the coefficients of linear correlation (Pearson’s) between TNFR1 transcripts in malignant cells in the indicated datasets (representing several solid malignant diseases) and the expression of the indicated proinflammatory genes. (F) Experiments in Rag2−/−IL-2Rγ−/− immunodeficient mice subcutaneously xenografted with TNFR1KO or WT HT29 cells whose tumors were removed on day 10 as indicated in the scheme to generate cell suspensions which were studied for the intratumoral content of mouse myeloid cells. Each dot represents data from an individual tumor (n = 12). (G) Multiplex tissue immunofluorescence analyses of excised samples as in F (n = 10–11). (H) Representative microscopy images (Scale bar: 50 µm). A t test statistical comparisons were shown (F and G). P < 0.05 (*), P < 0.01 (**), and P < 0.0001 (****).