Sublethal Radiation Affects Antigen Processing and Presentation Genes to Enhance Immunogenicity of Cancer Cells

Abstract

While immunotherapy in cancer is designed to stimulate effector T cell response, tumor-associated antigens have to be presented on malignant cells at a sufficient level for recognition of cancer by T cells. Recent studies suggest that radiotherapy enhances the anti-cancer immune response and also improves the efficacy of immunotherapy. To understand the molecular basis of such observations, we examined the effect of ionizing X-rays on tumor antigens and their presentation in a set of nine human cell lines representing cancers of the esophagus, lung, and head and neck. A single dose of 7.5 or 15 Gy radiation enhanced the New York esophageal squamous cell carcinoma 1 (NY-ESO-1) tumor-antigen-mediated recognition of cancer cells by NY-ESO-1-specific CD8+ T cells. Irradiation led to significant enlargement of live cells after four days, and microscopy and flow cytometry revealed multinucleation and polyploidy in the cells because of dysregulated mitosis, which was also revealed in RNA-sequencing-based transcriptome profiles of cells. Transcriptome analyses also showed that while radiation had no universal effect on genes encoding tumor antigens, it upregulated the expression of numerous genes involved in antigen processing and presentation pathways in all cell lines. This effect may explain the immunostimulatory role of cancer radiotherapy.

Keywords: antigen presentation; cancer cell line; gene expression; head and neck cancer; lung cancer; radiation; tumor antigen.

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Figure 1

Irradiation of cancer cells enhanced their recognition by antigen-specific CD8+ T cells. Human H522 lung (A) or OE19 esophageal (B) adenocarcinoma cells were irradiated with one dose of 7.5 or 15 Gy X-rays or left untreated (0 Gy). Three days later, adherent cells were collected and co-cultured in triplicate at a 5:1 ratio with or without NY-ESO-1-specific human CD8+ T cells on an ELISpot plate for detecting interferon-γ-producing cells after a day. The mean and its standard error are plotted, and p values in standard t tests are shown.

Figure 2

Effect of radiation on cell surface NY-ESO-1 measurement by fluorescence flow cytometry. Sub-confluent adherent cultures of indicated cell lines that had been grown in parallel to same cell density were treated with 0 or 15 Gy X-rays. After four days, adherent cells were collected by scraping and examined by flow cytometry for binding of an unconjugated mouse IgG1 antibody against NY-ESO-1. Antibody binding was indirectly detected with a fluorophore-conjugated rat antibody against mouse IgG. Binding of a control IgG1 (normal mouse serum fraction) was similarly detected. Shown are representative histograms of viable cells identified by 7-amino-actinomycin D staining.

Figure 3

Cell size increased following radiation treatment. Indicated human cancer cell lines were treated with one dose of 15 Gy X-rays or left untreated (0 Gy). Shown are representative phase contrast light microscopy images of cells after four days. Floating cells were removed before imaging.

Figure 4

Cell cycle arrest following radiation treatment. Indicated human cancer cell lines were treated with one dose of 15 Gy X-rays or left untreated (0 Gy). After four days, the DNA content of adherent cells was examined. (A) Overlaid green and blue fluorescence and phase contrast light microscopy images of representative fields are shown. Cells were fixed and permeabilized, and stained with an Alexa Fluor 488 green-fluorophore-conjugated monoclonal antibody against the endoplasmic-reticulum-resident calnexin protein, and the blue fluorescent DNA-binding dye 4’,6-diamidino-2-phenylindole (DAPI). (B) Histograms of DNA content of singlet cells are shown. Adherent cells were collected by scraping, fixed, stained with propidium iodide (PI) in a buffer with RNAse, and examined by fluorescence flow cytometry. DNA content was measured as orange fluorescence from PI–DNA binding. Noted are fractions of cells with 2N, 4N, and 8N DNA content (approximately diploid, tetraploid, and octaploid).

Figure 5

Effect of radiation on transcriptome. For seven human cancer cell lines, gene expression in adherent cells four days after radiation treatment was examined in three experiments with paired cell cultures that were or were not treated with one dose of 15 Gy X-rays. Similarity among the total 42 cellular transcriptomes is illustrated with unsupervised hierarchical clustering (A) and multi-dimensional scaling (B) plots. Cosine distance and Ward agglomeration methods were used. Inter-cluster distances in the dendrogram are indicated with a scale. (C) Effect of radiation on expression of genes in each cell line is depicted with stacked barplots that show the fractions of genes of which expression was up- or down-regulated ≥1.5-fold at 0.05 false discovery rate in paired likelihood ratio test in edgeR.

Figure 6

(A) Validation by reverse transcription (RT)-PCR of radiation-induced gene expression changes that were determined from RNA sequencing data. Mean of fold-change values and its standard error for pairs of 15-Gy-treated and untreated cells of three independent experiments are shown for six genes. The same RNA preparations were used for both RNA sequencing and RT-PCR. Global gene expression measurements by RNA sequencing were processed with the trimmed median of M-values method into count per million values. All values were normalized against those for the housekeeping ACTB gene prior to fold-change calculations. (B) Analyses of radiation-induced gene expression changes at the gene set level. Enrichment for 21 significant gene sets (p < 0.05) with similar enrichment across all seven indicated cell lines is shown. Gene expression of the 15-Gy-treated and untreated cells for Molecular Signature Database Hallmark and Reactome gene sets was scored using the gene set variation analysis method. Enrichment for a gene set was calculated as the ratio of scores of the groups of cells, and its statistical significance was estimated with a paired, empirical Bayes-moderated t test.

Figure 7

Effect of radiation on antigen processing and presentation genes. (A) Heatmap shows effect of 15 Gy radiation (log2 fold-change, compared to untreated cells (0 Gy)) on expression in indicated cell lines of 176 genes involved in antigen processing and presentation, of which expression was detected in at least one cell line. Genes are grouped by their function. Genes for which there was no effect (nominal p ≥ 0.05 in paired likelihood ratio test in edgeR) are shown in grey. Those with p < 0.05 are colored as per the displayed scale. Genes of which the expression was considered too low were not examined and are shown in white. (B) Cell surface expression of class I major histocompatibility complex (MHC I) proteins on viable 15-Gy-treated and untreated cells is depicted for three cell lines with representative direct fluorescence flow cytometry histograms. Viable cells were identified by 7-amino-actinomycin D staining. MHC I proteins were detected with a phycoerythrin (PE)-conjugated monoclonal IgG antibody. Separate portions of the same cell samples were examined for binding of a PE-conjugated negative control IgG. Geometric mean fluorescence intensity values for MHC I normalized against the negative control by subtraction are noted. For both A and B, sub-confluent adherent cultures of indicated cell lines that had been grown in parallel to same cell density were treated with one dose of 0 or 15 Gy X-rays. After four days, adherent cells were collected by scraping and examined.