Polycomb repressor complex 2 suppresses interferon-responsive MHC-II expression in melanoma cells and is associated with anti-PD-1 resistance

Abstract

Background: Despite the remarkable success of immunotherapy in treating melanoma, understanding of the underlying mechanisms of resistance remains limited. Emerging evidence suggests that upregulation of tumor-specific major histocompatibility complex-II (tsMHC-II) serves as a predictive marker for the response to anti-programmed death-1 (PD-1)/programmed death ligand 1 (PD-L1) therapy in various cancer types. The genetic and epigenetic pathways modulating tsMHC-II expression remain incompletely characterized. Here, we provide evidence that polycomb repressive complex 2 (PRC2)/EZH2 signaling and resulting H3K27 hypermethylation suppresses tsMHC-II.

Methods: RNA sequencing data from tumor biopsies from patients with cutaneous melanoma treated with or without anti-PD-1, targeted inhibition assays, and assays for transposase-accessible chromatin with sequencing were used to observe the relationship between EZH2 inhibition and interferon (IFN)-γ inducibility within the MHC-II pathway.

Results: We find that increased EZH2 pathway messenger RNA (mRNA) expression correlates with reduced mRNA expression of both presentation and T-cell genes. Notably, targeted inhibition assays revealed that inhibition of EZH2 influences the expression dynamics and inducibility of the MHC-II pathway following IFN-γ stimulation. Additionally, our analysis of patients with metastatic melanoma revealed a significant inverse association between PRC2-related gene expression and response to anti-PD-1 therapy.

Conclusions: Collectively, our findings demonstrate that EZH2 inhibition leads to enhanced MHC-II expression potentially resulting from improved chromatin accessibility at CIITA, the master regulator of MHC-II. These insights shed light on the molecular mechanisms involved in tsMHC-II suppression and highlight the potential of targeting EZH2 as a therapeutic strategy to improve immunotherapy efficacy.

Keywords: Drug Therapy, Combination; Immunotherapy; Melanoma; Tumor Microenvironment.

Figure 1

High expression of PRC2 genes correlates with low expression of tumor-specific MHC class II genes. (A) Cluster analysis was performed using z-scores calculated from RNA sequencing expression data of patient with melanoma tumors gathered from The Cancer Genome Atlas. (B) mRNA expression of PRC2 genes (EZH2 and JARID2) and T-cell genes (CD3E, CD4, CD8A) across T cell^hi/(red boxes), Intermediate (blue boxes), and PRC2^hi (green boxes) patient clusters. Box plots show the median, first and third quartiles. The whiskers extend to the maxima and minima but no further than 1.5 times the IQR. Data were analyzed by analysis of variance followed by a Tukey’s post hoc test. ***p<0.001; ****p<0.0001. (C) Flow cytometry histograms representing live fractions of interferon-γ-stimulated (red histograms) human melanoma cell lines compared with unstimulated controls (gray histograms). Following staining with a live-dead dye, cells were labeled with fluorophore-conjugated HLA-DR antibodies. (D) qPCR analysis of human melanoma cell lines characterized for mRNA expression of the indicated genes of interest relative to GAPDH mRNA expression values for each cell line. Values were calculated as fold expression relative to A375 cells.CIITA, class II major histocompatibility complex transactivator II; EZH2, enhancer of zeste 2 polycomb repressive complex 2 subunit; HLA-DR, human leukocyte antigen-DR ; JARID2, Jumoni and AT-rich interaction domain-containing protein; MHC, major histocompatibility complexes; mRNA, messenger RNA; PRC2, polycomb repressive complex 2; qPCR, quantitative polymerase chain reaction.

Figure 2

EZH2 knockdown moderately increases expression of MHC class II. (A) EZH2, CIITA, and CD74 mRNA expression in MHC-II-proficient cell lines after transfection with EZH2 siRNA or to non-targeting control (NTC). Values were normalized to GAPDH mRNA and calculated as fold expression values relative NTC siRNA. The data shown represents combined technical replicates derived from three to four independent experiments. (B) after 24–48 hours IFN-γ stimulation, whole cell protein lysates were extracted from siRNA-transfected MHC-II-proficient cell lines probed with monoclonal antibodies specific for the indicated proteins. Lysates were normalized to GAPDH expression. Data represent one of two experimental replicates. (C) Flow cytometry analysis of MHC-II-proficient cell lines transfected with siEZH2 or NTC and stained with fluorophore-conjugated HLA-DR antibodies. The histograms shown in C represent one of at least three independent experiments. CIITA, class II major histocompatibility complex transactivator II; EZH2, enhancer of zeste 2 polycomb repressive complex 2 subunit; IFN, interferon; HLA-DR, human leukocyte antigen-DR; MHC, major histocompatibility complexes; mRNA, messenger RNA.

Figure 3

Pharmacological inhibition of EZH2 upregulates interferon-induced major histocompatibility complexes expression. (A–B) Flow cytometry analysis of A375 and SKMEL-5 cell lines treated with the indicated EZH2 inhibitor for 5 days. Single cell suspensions were stained with fluorophore-conjugated antibodies HLA-DR. *p<0.05 and **p<0.01, one-sample t-test against a fold change of 1. (C) After stimulation with IFN-γ for 24–48 hours, whole cell protein lysates were extracted from A375, SKMEL-5, and COLO829 cells treated with DMSO or EZH2 inhibitor before being probed for the indicated proteins. Data shown represents three independent experiments. (D–E) Relative expression of EZH2, CIITA, and CD74 mRNA extracted from melanoma cell lines treated with GSK343 or tazemetostat and stimulated with IFN-γ for 24–48 hours. Samples were normalized to GAPDH mRNA and calculated as fold expression relative to the DMSO control. CIITA, class II major histocompatibility complex transactivator I; DMSO, dimethyl sulfoxide; EZH2, enhancer of zeste 2 polycomb repressive complex 2 subunit; HLA-DR, human leukocyte antigen-DR; IFN, interferon; MFI, mean flourescence intensity; mRNA, messenger RNA; PD-L1, programmed death ligand 1.

Figure 4

Tazemetostat reverses major histocompatibility complexes II-specific IFN-γ resistance MeWo, SKMEL-1, and CHL-1 cells were pretreated with tazemetostat (1 µM or 5 µM) for 3–4 days and stimulated with IFN-γ for 72 hours. (A–B) On harvest, cells were stained with the indicated antibodies and visualized for surface expression using flow cytometry. (C) Western blots using whole cell lysates and probed with antibodies specific for the indicated targets. HLA-DR expression was below the limit of detection for western blot in SKMEL1 (less sensitive than flow cytometry). (D–E) Quantitative real time PCR performed on messenger RNA isolates after tazemetostat treatment and IFN-γ stimulation. CIITA, class II major histocompatibility complex transactivator I; DMSO, dimethyl sulfoxide; EZH2, enhancer of zeste 2 polycomb repressive complex 2 subunit; HLA-DR, human leukocyte antigen-DR; IFN, interferon; MFI, mean flourescence intensity; PD-L1, programmed death ligand 1.

Figure 5

Tazemetostat treatment exposes accessible chromatin at regions containing putative IRF1 and STAT1 binding motifs in IFN-γ-stimulated CHL-1 cells. (A) CHL-1 cells were pretreated for 4 days with tazemetostat before 24 hours stimulation with IFN-γ. Cells were harvested and processed for assays for transposase-accessible chromatin with sequencing. Black bars represent predicted target sites for H3K27me3 marks, CpG islands, and promoters driving transcription of the indicated genes, based on the publicly available data from multiple tissue types and conditions in the UCSC database. Putative transcription factor (TF) binding sites for IFN-γ signaling (STAT1 and IRF4), based on predictions made using the https://molotool.autosome.org/and a permissive p value setting of 0.0005. (B) mRNA expression assessed by qPCR using primers specific for the pIII and pIV 5’UTR. (C) Correlation matrix for qPCR data across seven cell lines (A375, SKMEL5, COLO829, SKMEL28, MEWO, SKMEL1, and CHL1) under interferon-stimulated conditions. CIITA, class II major histocompatibility complex transactivator II; DMSO, dimthyl sulfoxide; EZH2, enhancer of zeste 2 polycomb repressive complex 2 subunit; IFN, interferon; JARID2, Jumoni and AT-rich interaction domain-containing; mRNA, messenger RNA; pIV, promoter IV; qPCR, quantitative polymerase chain reaction; USCS, University of California, Santa Cruz.

Figure 6

Anti-PD-1-responsive human metastatic melanoma tumors express lower levels of JARID2 mRNA compared with non-responders. (A) Heatmap representing human melanoma RNA sequencing z-scores of genes relevant to PRC2, T cells, and antigen presentation by MHC-I and MHC-II. Samples were ordered by increasing JARID2 mRNA expression. (B) Summation scores of Z-standardized T-cell signature genes stratified by median JARID2 mRNA. (C) Summation scores of Z-standardized MHC-II signature genes stratified by median JARID2 mRNA. (D) Z-standardized JARID2 mRNA stratified by tumor-specific (IHC) HLA-DR status using a defined 5% cut-off. (E–G) JARID2 mRNA expression z-scores in responders (complete response/partial response) compared with non-responders (progressive disease/stable disease) in patients given anti-PD-1. Mann-Whitney tests, *p<0.05. CIITA, class II major histocompatibility complex transactivator II; CTLA-4, cytotoxic T-lymphocyte-associated protein 4; EZH2, enhancer of zeste 2 polycomb repressive complex 2 subunit; HLA-DR, human leukocyte antigen-DR; IHC, immunohistochemistry; JARID2, Jumoni and AT-rich interaction domain-containing protein; MHC, major histocompatibility complexes; mRNA, messenger RNA; PD-1, programmed death-1; PRC2, polycomb repressive complex 2.