The SETDB1-TRIM28 Complex Suppresses Antitumor Immunity

Abstract

The tumor immune microenvironment is influenced by the epigenetic landscape of the tumor. Here, we have identified the SETDB1-TRIM28 complex as a critical suppressor of antitumor immunity. An epigenetic CRISPR-Cas9 screen of 1,218 chromatin regulators identified TRIM28 as a suppressor of PD-L1 expression. We then revealed that expression of the SETDB1-TRIM28 complex negatively correlated with infiltration of effector CD8⁺ T cells. Inhibition of SETDB1-TRIM28 simultaneously upregulated PD-L1 and activated the cyclic GMP-AMP synthase (cGAS)-stimulator of interferon genes (STING) innate immune response pathway to increase infiltration of CD8⁺ T cells. Mechanistically, SETDB1-TRIM28 inhibition led to micronuclei formation in the cytoplasm, which is known to activate the cGAS-STING pathway. Thus, SETDB1-TRIM28 inhibition bridges innate and adaptive immunity. Indeed, SETDB1 knockout enhanced the antitumor effects of immune checkpoint blockade with anti-PD-L1 in a mouse model of ovarian cancer in a cGAS-dependent manner. Our findings establish the SETDB1-TRIM28 complex as a regulator of antitumor immunity and demonstrate that its loss activates cGAS-STING innate immunity to boost the antitumor effects of immune checkpoint blockade.

Figure 1:. Identification of Trim28 as a suppressor of PD-L1 expression.

A, Schematic of the experimental design for the epigenome-wide CRISPR screen. B, Volcano plot illustrating the genes whose knockout (KO) significantly upregulated (red) or downregulated (green) PD-L1 expression stimulated by IFNγ with top hits highlighted. C, Robust rank aggregation (RRA) of the top 7 hits from the screen whose KO increases PD-L1 expression. D, Enrichment of 10 individual sgRNAs against Trim28 in the PD-L1^high population compared with controls. E, Expression of PD-L1 was determined by FACS analysis in ID8 cells expressing two individual shTRIM28 or control treated with or without 30 ng/ml IFNγ for 24 h. An isotype matched IgG was used as a negative control. F, Quantification of E, the mean fluorescence index of PD-L1 was determined from 3 biologically independent experiments. G, Significant correlation between TRIM28 and CD274 expression based on RNA-seq analysis from 24 TCGA cancer types with at least 100 patients. H, Expression of PD-L1 was determined by FACS analysis in control and Setdb1-knockout ID8 cells treated with or without 30 ng/ml IFNγ for 24 h. An isotype matched IgG was used as a negative control. I, Quantification of H, the mean fluorescence index of PD-L1 was determined from 3 biologically independent experiments. J, Correlation between SETDB1 and CD274 expression based on RNA-seq analysis in TCGA ovarian cancer dataset. K, Significant correlation between SETDB1 and CD274 expression based on RNA-seq analysis from 24 TCGA cancer types with at least 100 patients. Data represent mean ± SEM, n = 3 biologically independent experiments unless otherwise stated. P values were calculated using a two-tailed student t test except in B and D by Model-based Analysis of Genome-wide CRISPR-Cas9 Knockout (MAGeCK) analysis, G, J and K by Pearson R analysis.

Figure 2:. SETDB1–TRIM28 negatively correlates with infiltration of effector CD8⁺ T cells.

A, Representative staining of SETDB1, TRIM28 and PD-L1 in SETDB1–TRIM28 high or low expressing tumors from a TMA with 131 HGSOC cases. Scale bar = 100 μm. B, Negative correlation between SETDB1 and PD-L1 expression. C, Negative correlation between TRIM28 and PD-L1 expression. D, Positive correlation between SETDB1 and TRIM28 expression. E, F, High SETDB1 expression correlates with significantly lower tumor infiltrated CD8⁺ T cells (E) and Granzyme B⁺/CD8⁺ T cells (F). G, H, Negative correlation between SETDB1 and CD8A (G) or GZMB (H) in RNA-seq analysis using the TCGA ovarian cancer dataset. I, J, Significant negative correlation between SETDB1 and CD8A (I) or GZMB (J) expression based on RNA-seq analysis from 24 TCGA cancer types with at least 100 patients. K, Overall survival of patients with high or low SETDB1 from the TMA based on Kaplan-Meier analysis. L, TRIM28 protein is expressed at a significantly higher levels in melanoma patients who failed to respond to anti–PD-1 therapy compared with those responded to the therapy based on a publicly available proteomics dataset (35). Data represent mean ± SEM. P values were calculated using a two-tailed student t test except in B–D, G–J by Pearson R analysis, and K by log rank test.

Figure 3:. SETDB1–TRIM28 loss activates the cGAS-STING pathway.

A, GSEA hallmark enrichment analysis of DEGs between control and Setdb1-knockout ID8 cells. B, Enrichment plot of GSEA for inflammatory response hallmark. C, Heatmap of the inflammatory response hallmark genes upregulated in Setdb1-knockout ID8 cells. D, Expression of Ccl5 and Cxcl10 in control and the indicated Setdb1-knockout ID8 cells determined by RT-qPCR analysis. E, Immunofluorescence staining of cGas and γ-H2AX in control and Setdb1-knockout ID8 cells. Arrows point to examples of cGas and γH2AX positive micronuclei. Scale bar = 10 μm. F, Quantification of micronuclei-positive cells in control and indicated Setdb1-knockout ID8 cells. G, H, Same as E, F but for Trim28-knockout ID8 cells. I, Expression of Ccl5 and Cxcl10 in control and Setdb1-knockout ID8 cells treated with or without the cGAS inhibitor RU.521 (2.5 μg/ml for 48 h), as determined by RT-qPCR analysis. J, Expression of Ccl5 and Cxcl10 in control, Setdb1-knockout, cGas-knockout and cGas/Setdb1-double knockout ID8 cells determined by RT-qPCR analysis. K, Expression of Setdb1, cGas, Sting, phospho-Irf3 and a loading control β-actin in the indicated cells determined by immunoblot. L, Percentage of PD-L1⁺ tumor cells in tumors formed by orthotopic injection of the indicated ID8 cell lines into C57BL/6J mice. Data represent mean ± SEM, n = 3 biologically independent experiments unless otherwise stated. P values were calculated using a two-tailed student t test except in B by GSEA analysis.

Figure 4:. Setdb1 loss synergizes with anti–PD-L1 in reducing tumor burden.

A, Schematic of the experimental design using the ID8 syngeneic orthotopic mouse ovarian cancer model. B, Reproductive tracts with tumors from the indicated treatment groups were dissected at the end of treatment (n=5 mice per group). C, The weights of tumors dissected from the indicated groups were measured as a surrogate for tumor burden. P-values were calculated using a two-tailed t test. Combination index (CI) for Setdb1 loss and anti–PD-L1 is 0.82 (<1), which indicates synergy between Setdb1 loss and anti-PD1 combination. D, E, Infiltration of CD8⁺ T cells (D) and Granzyme B⁺/CD8⁺ T cells (E) in the tumors dissected from the indicated treatment groups were analyzed by flow cytometry (n = 5 mice per group). Data represent mean ± SEM, P values were calculated using a two-tailed t test.