The Integrated Stress Response Pathway Coordinates Translational Control of Multiple Immune Checkpoints in Lung Cancer

Abstract

The integrated stress response (ISR) is an adaptive pathway hijacked by cancer cells to survive cellular stresses in the tumor microenvironment. ISR activation potently induces PD-L1, leading to suppression of antitumor immunity. In this study, we sought to uncover additional immune checkpoint proteins regulated by the ISR to elucidate mechanisms of tumor immune escape. The ISR coordinately induced cluster of differentiation 155 (CD155) and PD-L1, enhancing translation of both immune checkpoint proteins through bypass of inhibitory upstream open reading frames in their 5' untranslated regions. Analysis of primary human lung tumors identified a significant correlation between expression of PD-L1 and CD155. ISR activation accelerated tumorigenesis and inhibited T-cell function, which could be overcome by combining PD-1 and TIGIT blockade with the ISR inhibitor ISRIB. This study uncovers a mechanism by which two immune checkpoint proteins are coordinately regulated and suggests a therapeutic strategy for patients with lung cancer.

Significance: The integrated stress response represents a targetable axis to improve the efficacy of immunotherapy in lung cancer by inhibiting the coordinated translational regulation of the PD-L1/PD-1 and CD155/TIGIT immune checkpoint pathways.

Graphical abstract

Figure 1.

ISR pathway activation induces PD-L1 and CD155 protein. A, Western blot analysis of human KRAS-mutant H441 or EGFR-mutant PC9 cells treated for 24 or 48 hours with 100 or 200 μmol/L salubrinal or DMSO vehicle (Veh) control. Vinculin served as a loading control for this and subsequent Western blots. B, Western blot analysis of human H441, H358, PC9, and HCC827 cells after 24 hours in RPMI 1640 media supplemented with or without amino acids. C, Western blot analysis of human H441, H358, PC9, and HCC827 cells treated for 24 hours with 5 μmol/L thapsigargin (ER Ca+ ATPase pump inhibitor) to induce ER stress or DMSO vehicle control. D, Western blot analysis of H358 cells with control or UROD sgRNA. E, Western blot analysis of H358 or HCC827 cells with control or UROD siRNA. F, Flow cytometric analysis of cell-surface CD155 and PD-L1 protein in DMSO vehicle control and thapsigargin-treated (5 μmol/L for 24 hours) H358 or HCC827 NSCLC cells. G, Western blot analysis in eIF2α WT (S/S) or mutant (Ser51Ala A/A) MEFs treated with DMSO vehicle control or 100 μmol/L salubrinal for 24 hours. H, Western blot analysis of thapsigargin- and ISRIB-treated human NSCLC cells. Data from a single experiment are shown and are representative of at least three independent experiments. p-eIF2α, phosphorylated eIF2α.

Figure 2.

ISR pathway activation enhances PD-L1 and CD155 translation. A, Polysome profiling of H1944 vehicle- or salubrinal-treated cells (100 μmol/L for 24 hours). B and C, qRT-PCR analysis of PD-L1(CD274) (B) and CD155(PVR) (C) mRNA in ribosomal fractions from A. qRT-PCR analysis for each gene shown was performed with 1 primer set spanning an exon–exon junction (primer set 1). Data for primers set 2 are available in Supplementary Fig. S2. Fractions associated with <3 ribosomes were grouped to represent poorly translated mRNAs, and fractions associated with >3 ribosomes were grouped as efficiently translated mRNAs. PD-L1 and CD155 mRNA expression in each fraction was normalized to luciferase, and mRNA abundance was calculated as the percent of total in all fractions. Luciferase control mRNA was added to each fraction prior to RNA extraction to control for variability. Error bars represent the SD from the mean from three independent fractions (<3 or >3 ribosomes). D, Diagram of the WT human CD155 5′ UTR with 6 CTGs and mutant constructs with CTGs mutated to CTCs cloned upstream of a luciferase reporter. E, Dual luciferase assay of MEFs transfected with indicated CD155-5′ UTR firefly luciferase reporter constructs normalized to cotransfected control Renilla luciferase. Luciferase activity was monitored after 48 hours. Error bars represent the SD from the mean from n = 3 biological replicates. Data from a single experiment are shown and are representative of three independent experiments. F, qRT-PCR analysis of the mean luciferase mRNA normalized to actin in MEFs shown in E. Error bars represent the SD from the mean from n = 3 biological replicates. G, Dual luciferase assay of the CD155 5′ UTR in MEF vehicle- or salubrinal-treated cells (100 μmol/L for 24 hours). Error bars represent the SD from the mean from n = 3 biological replicates. H, qRT-PCR analysis of the mean luciferase mRNA normalized to actin in MEFs shown in G. n = 3 biological replicates. A Student t test was used to determine statistical significance. *, P < 0.05; **, P < 0.005.

Figure 3.

ISR pathway activation diminishes immune cell function in vitro. A, ELISAs for IL2 and granzyme B of Jurkat T cells cocultured with H358 cells. H358 cells were pretreated for 24 hours with salubrinal and 800 nmol/L ISRIB and then washed and cocultured with Jurkat T cells for an additional 24 hours with αCD3 and αCD28 activating antibodies (4 μg/mL each). Error bars represent the SD from the mean from n = 3 biological replicates. Data from a single experiment are shown and are representative of three independent experiments. B, ELISAs for IL2 and granzyme B of Jurkat T cells cocultured with H358 cells with control or UROD sgRNA and 800 nmol/L ISRIB. H358 cells were cocultured with Jurkat T cells for 24 hours with αCD3 and αCD28 activating antibodies (4 μg/mL each). Error bars represent the SD from the mean from n = 3 biological replicates. Data from a single experiment are shown and are representative of two independent experiments. C, ELISAs for IL2 and granzyme B of primary human PBMCs cocultured with H358 cells. H358 cells were pretreated for 24 hours with salubrinal and 800 nmol/L ISRIB and then washed and cocultured with PBMCs for an additional 24 hours with αCD3 and αCD28 activating antibodies (1 μg/mL each). Error bars represent the SD from the mean from n = 3 biological replicates. Data from a single experiment are shown and are representative of three independent experiments. D, ELISAs for IL2 and granzyme B of PBMCs cocultured with H358 cells with control or UROD sgRNA and 800 nmol/L ISRIB. H358 cells were cocultured with PBMCs for 24 hours with αCD3 and αCD28 activating antibodies (1 μg/mL each). Error bars represent the SD from the mean from n = 3 biological replicates. Data from a single experiment are shown and are representative of two independent experiments. E, ELISAs for IL2 and granzyme B of OT-1 T cells cocultured with murine KP cells (KdP67-1) expressing OVA and crystal violet assay of KP cells after coculture with OT-1 T cells. KP cells were pretreated with 100 μmol/L salubrinal and/or 500 nmol/L ISRIB for 24 hours and then cocultured for 24 hours with activated CD8⁺ OT-1 T cells. Error bars represent the SD from the mean from n = 3–4 biological replicates. Data from a single experiment are shown and are representative of two independent experiments. A Student t test was used to determine statistical significance. *, P < 0.05; **, P < 0.005; ***, P < 0.0005; ****, P < 0.00005. sgNS, nonspecific (control) sgRNA; sgUROD, UROD sgRNA.

Figure 4.

ISR activation enhances tumorigenesis and reduces immune cell infiltration in vivo. A, Western blot analysis of KRAS-mutant murine CMT167 cells treated for 24 or 48 hours with 100 or 200 μmol/L salubrinal or DMSO vehicle control. p-eIF2α, phosphorylated eIF2α; Veh, vehicle. B, Quantification of tumor volumes of CMT167 cells transplanted in C57BL/6J mice (n = 11 mice per group for vehicle-treated mice; n = 10 mice per group for salubrinal-treated mice). The graph represents mean tumor volumes, and error bars represent the SEM. C and D, Endpoint tumor volumes (C) and tumor mass (D) of resected tumors shown in B. Horizontal bars represent mean values. E, Multiplexed IHC-F was performed on tumors from B. Representative images (10×) are shown. Scale bar, 200 μm. F–H, Quantification of CD3⁺, CD4⁺, and CD8⁺ tumor-infiltrating lymphocytes (Rexpressed as counts/mm³). Five fields were quantified per mouse from five mice. Graphs represent mean counts, and error bars represent the SD from the mean. A Student t test was used to determine statistical significance. **, P < 0.005; ***, P < 0.0005.

Figure 5.

ISR pathway inhibition improves response to PD-1 blockade in a syngeneic mouse model. A, Western blot analysis of Kras-mutant murine CMT167 cells expressing doxycycline-inducible control shRNA or two independent shRNAs targeting Urod. Ab, antibody. B, Schematic illustration of CMT167 syngeneic experiment. C, Quantification of tumor volumes of CMT167 cells expressing the indicated shRNA sequence transplanted in C57BL/6J mice (n = 15 mice per group; scrambled shRNA data are shown in Supplementary Fig. S4C and S4D). Graph represents mean tumor volumes. Error bars, SEM. D, Uniform manifold approximation and projection (UMAP) analysis of tumor-infiltrating lymphocytes, colored by cell types. E, Quantification of tumor-infiltrating lymphocytes (expressed as a percentage of CD45⁺ cells) from flow mass cytometry (mass CyTOF). n = 5 scrambled mice; n = 5 Urod shRNA mice; n = 5 Urod shRNA + ISRIB; n = 4 Urod shRNA + αPD-1; n = 5 Urod shRNA + ISRIB +αPD-1. Graphs represent mean values, and error bars represent the SD from the mean. F, Quantification of tumor volumes of CMT167 cells expressing the indicated shRNAs transplanted in C57BL/6J mice (n = 11–12 mice per group; scrambled shRNA data are shown in Supplementary Fig. S5C). Graph represents mean tumor volumes. Error bars, SEM. G, Quantification of CD8⁺ T cells (expressed as a percentage of CD45⁺ cells) from mice in F. n = 5 scrambled mice; n = 4 Urod shRNA mice; n = 5 Urod shRNA + ISRIB; n = 5 Urod shRNA + ISRIB+ αPD-1; n = 5 Urod shRNA + ISRIB +αTIGIT; n = 5 Urod shRNA + ISRIB + αPD-1+ αTIGIT. Graph represent mean values. Error bars, SD from the mean. A Student t test was used to determine statistical significance. *, P < 0.05; **, P < 0.005; ***, P < 0.0005; ns, not significant. p-eIF2α, phosphorylated eIF2α; shUrod, short-hairpin Urod; Tregs, regulatory T cells. B, Created in BioRender. O’Donnell, K. (2025) https://BioRender.com/vphxffz.

Figure 6.

CD155 and PD-L1 are positively correlated in primary human lung adenocarcinoma tumors. A, Representative images of IHC for PD-L1 and CD155 of 33 primary human lung adenocarcinomas. Shown are 20× representative images. Scale bar, 100 μm. B, Correlation of CD155 H-score and PD-L1 H-score of tissues in A (Pearson correlation r = 0.506, P = 0.0031). C, Microphotographs of CD155 and PD-L1 IHC analysis of primary lung adenocarcinomas displaying different combinations of expression in malignant cells (CD155⁺/PD-L1+, CD155⁻/PD-L1−, CD155⁺/PD-L1−, and CD155⁻/PD-L1+). Shown are 20× representative images. Scale bar, 100 μm. Red arrows, positive expression of PD-L1 or CD155 in MCs. D, Association of CD155 H-score expression in all primary lung adenocarcinomas with PD-L1 status (cutoff for positive PD-L1 expression: tumor proportion score ≥1%; P = 0.0004). A Mann–Whitney test was used to determine statistical significance. Graph represents the median, and error bars represent the 95% confidence interval. E, Association of CD155 H-score expression in all primary lung adenocarcinoma with pathologic stage (P = 0.001). A Mann–Whitney test was used to determine statistical significance. Graph represents the median, and error bars represent the 95% confidence interval. F, Correlation of CD155 H-score expression with tumor size (Spearman correlation r = 0.1946, P < 0.0001). ***, P < 0.0005; ****, P < 0.00005.

Figure 7.

Model of translational control of CD155 and PD-L1 in response to ISR activation. Stresses commonly present in the TME activate the ISR pathway in tumor cells. This results in enhanced translation of CD155 and PD-L1, which suppresses T-cell function by binding the inhibitory T-cell receptors PD-1 and TIGIT, respectively, thereby leading to enhanced tumorigenesis. Created in BioRender. O’Donnell, K. (2025) https://BioRender.com/vphxffz.