UBR5 promotes tumor immune evasion through enhancing IFN-γ-induced PDL1 transcription in triple negative breast cancer

Abstract

Background: The up-regulation of PD-L1 is recognized as an adaption of cancer cells to evade immune surveillance and attack. However, the intrinsic mechanisms of the induction of PD-L1 by interferon-γ (IFN-γ) in tumor microenvironment remain incompletely characterized. Ubiquitin ligase E3 component N-recognition protein 5 (UBR5) has a critical role in tumorigenesis of triple negative breast cancer (TNBC) by triggering specific immune responses to the tumor. Dual targeting of UBR5 and PD-L1 exhibited superior therapeutic benefits in a preclinical TNBC model in short term. Methods: The regulation of UBR5 to PD-L1 upon IFN-γ stimulation was evaluated through in UBR5 deficiency, reconstitution or overexpression cell line models by quantitative PCR, immunohistochemistry and RNA-seq. The effects of PD-L1 regulation by UBR5 and double blockade of both genes were evaluated in mouse TNBC model. Luciferase reporter assay, chromatin immunoprecipitation-qPCR and bioinformatics analysis were performed to explore the transcription factors involved in the regulation of UBR5 to PD-L1. Results: E3 ubiquitin ligase UBR5 plays a key role in IFN-γ-induced PDL1 transcription in TNBC in an E3 ubiquitination activity-independent manner. RNA-seq-based transcriptomic analyses reveal that UBR5 globally affects the genes in the IFN-γ-induced signaling pathway. Through its poly adenylate binding (PABC) domain, UBR5 enhances the transactivation of PDL1 by upregulating protein kinase RNA-activated (PKR), and PKR's downstream factors including signal transducers and activators of transcription 1 (STAT1) and interferon regulatory factor 1 (IRF1). Restoration of PD-L1 expression in UBR5-deficient tumor cells recoups their malignancy in vivo, whereas CRISPR/Cas9-mediated simultaneous abrogation of UBR5 and PD-L1 expression yields synergistic therapeutic benefits than either blockade alone, with a strong impact on the tumor microenvironment. Conclusions: This study identifies a novel regulator of PDL1 transcription, elucidates the underlying molecular mechanisms and provides a strong rationale for combination cancer immunotherapies targeting UBR5 and PD-L1.

Keywords: Interferon-γ; PD-L1; PKR/STAT1/IRF1; UBR5; triple negative breast cancer.

Figure 1

UBR5 is required for IFN-γ-induced PDL1 gene expression. (A) Quantitative PCR analysis of the relative Pdl1 mRNA levels and western blot analysis of the UBR5 protein levels in WT and Ubr5^-/- 4T1 cells with or without 10 ng/mL IFN-γ treatment for 24 h. (B) The relative expression levels of PD-L1 on the surfaces of WT and Ubr5^-/- 4T1 cells with or without 10 ng/mL IFN-γ stimulation was evaluated by flow cytometry. (C) WT and Ubr5^-/- 4T1 cells (1×10⁶) were subcutaneously injected into the mammary pads to generate tumors in WT (BALB/c) mice. Immunohistochemical staining was performed with anti-UBR5 or anti-PD-L1 antibody in WT and Ubr5^-/- 4T1 tumor sections. (D-F) Human TNBC BT549 and MDA-MB-231 cell lines with stable UBR5 knockdown were generated by infection with a lentivirus containing a control or UBR5-targeted shRNA sequence followed by puromycin selection. The mRNA (D) and protein levels of UBR5 (E) and PD-L1 (E and F) were measured after cells were treated with 10 ng/mL IFN-γ for 24 h.(G-I) BT549, MDA-MB-231 and MCF7 cells were transfected with an empty vector or UBR5 plasmids. 48 hours later, the cells were treated with 10 ng/mL IFN-γ for 24 h. The mRNA and protein levels of UBR5 (G) and PD-L1 (H and I) were measured by qPCR, western blot and FACS. All experiments were repeated at least three times, and the data are presented as the mean ± SEM. ns, no significance, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 2

Restoration of PD-L1 expression in UBR5-deficient tumors reinvigorates malignancy. (A-D) Ubr5^-/- 4T1 cells were transfected with a plasmid encoding mPdl1, and then subjected to 4 μg/mL puromycin selection to generate stable mPdl1-reconstituted 4T1/Ubr5^-/- cells. The mRNA and protein levels of PD-L1 and UBR5 protein levels in WT, Ubr5^-/-, hUBR5- or mPdl1-reconstituted Ubr5^-/- 4T1 cells were confirmed by RT-qPCR (A), FACS analysis (B) and western blot (D) after treatment with IFN-γ for 24 h, and the relative fold changes in PD-L1 based on the mean fluorescence intensity (MFI) are shown in (C). The data are presented as the mean ± SEM (error bar) from three replicates. *P< 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. (E) A total of 1×10⁶ WT, Ubr5^-/-, hUBR5- or mPdl1-reconstituted Ubr5^-/- 4T1 cells were subcutaneously injected into the mammary pads of WT (BALB/c) mice (n=6 mice per group). Tumor size was monitored every 3 days. (F) On Day 10 after tumor cells inoculation, flow cytometry analyzed the CD8⁺ T cells infiltration and GzmB⁺/PD-1^-/CD8⁺ T cells in tumor tissue from mice bearing 4T1 WT, Ubr5^-/-, hUBR5- or mPdl1-reconstituted Ubr5^-/- tumor. Tumor-draining lymph nodes were analyzed by staining for CD25⁺, Foxp3⁺ Tregs and CD11c⁺, MHC Ⅱ⁺ DCs by flow cytometry. The results are presented as the mean ± SEM. ns, no significance, *P< 0.05, **P < 0.01, ***P < 0.001. n=3 mice per group. Mouse survival (G) were recorded daily. (H) T cells cytotoxicity difference toward MDA-MB-231 cells with different UBR5 expression levels. dt-Tomato Red stably expressed MDA-MB-231 cells (target cells) were mixed with CFSE labelled MCF7 cells (non-target cells) at a ratio of 1:1, and then cocultured for 18 h with either control or c-Met specific chimeric antigen receptor T cells at a ratio of 1:2 separately. Cells were harvested and analyzed by flow cytometry. The data are presented as the mean ± SEM (error bar) from three replicates. *P< 0.05.

Figure 3

Combined genetic targeting of Ubr5 and Pdl1 yields synergistic therapeutic effects. (A-B) 4T1/GFP/Pdl1^-/- and 4T1/Ubr5^-/-Pdl1^-/- cells were generated by CRISPR/Cas9 editing. WT, Ubr5^-/-, GFP/Pdl1^-/- or Ubr5^-/-Pdl1^-/- 4T1 cells were treated with 10 ng/mL IFN-γ for 24 h, and then the UBR5 and PD-L1 protein levels were detected by western blot (A) and surface PD-L1 levels were measured by flow cytometry analysis (B). (C) A total of 1×10⁶ WT, Ubr5^-/-, GFP/Pdl1^-/- or Ubr5^-/-Pdl1^-/- 4T1 cells were subcutaneously injected into the mammary pads to monitor tumor growth in WT (BALB/c) mice (n=6 mice per group). Tumor size was measured every 3 days. (D-G) WT, Ubr5^-/-, GFP/Pdl1^-/- or Ubr5^-/-Pdl1^-/- 4T1 cells (5×10⁵) were injected via the tail vein into 7-week-old female BALB/c mice (n=5 mice per group). Twelve days later, the mice were sacrificed, and single-cell suspensions were obtained from lung tissue to perform clonogenic assays in order to evaluate lung metastasis. Images of 6-well plates in a representative experiment are shown in (D), and colonies were quantified (E). (F) On Day 10 after tumor cells inoculation, CD8⁺ T cells infiltration and GzmB⁺/PD-1^-/CD8⁺ T cells in tumor tissue from mice bearing 4T1 WT, Ubr5^-/-, hUBR5-or mPdl1-reconstituted Ubr5^-/- tumor were analyzed by flow cytometry. Flow cytometry analyzed CD25⁺, Foxp3⁺ Tregs and CD11c⁺, MHC Ⅱ⁺ DCs in tumor-draining lymph nodes. n=3 mice per group. The survival curve is shown in (G). The results are presented as the mean ± SEM. ns, no significance, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 4

UBR5 globally regulates IFN-γ-mediated pathways and stimulated genes. (A) Numbers of differentially expressed genes (DEGs) in Ubr5^-/- compared with GFP 4T1 cells with or without IFN-γ treatment. Fold change (FC) ≥ 2, p. adjust ˂ 0.05. (B) Numbers of DEGs in GFP 4T1 cells or Ubr5^-/- 4T1 cells with IFN-γ treatment compared with that without IFN-γ treatment. Fold change (FC) ≥ 2, p. adjust ˂ 0.05. (C) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment of the top 10 biological process between IFN-γ-treated GFP and Ubr5^-/- 4T1 cells. (D) Heatmap depicting the mRNA levels of selected ISGs in GFP and Ubr5^-/- 4T1 cells with or without 10 ng/mL IFN-γ induction for 24 h. (E) Scatterplot diagram of genes responding to IFN-γ differently in Ubr5^-/- cells compared to that in GFP cells. The grey dots represent unchanged expression (1/2<FC<2) in both cell lines. Upregulated ISGs with fold change (FC) values >2 in Ubr5^-/- cells compared with GFP cells are shown with orange dots, while downregulated ISGs with fold change (FC) values <1/2 are shown with blue dots. (F) The mRNA levels of Cd40, Siglec15 and Isg15 were examined by qPCR analysis in WT, Ubr5^-/- and hUBR5-reconstituted Ubr5^-/- 4T1 cells treated with or without IFN-γ for 24 h. All experiments were repeated three times, and the data are presented as the mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 5

UBR5 is crucial for IFN-γ-induced activation of STAT1 and IRF1 transcription. (A-B) The mRNA (A) and protein (B) levels of UBR5, STAT1, pSTAT1 and IRF1 were detected in WT, Ubr5^-/- and hUBR5-reconstituted Ubr5^-/- 4T1 cells with or without IFN-γ stimulation. (C-D) The mRNA (C) and protein levels (D) of STAT1 and IRF1were detected in IFN-γ-treated UBR5-knockdown BT549 (C and D) and MDA-MB-231 (D) stable cell lines. (E-G) BT549, MDA-MB-231 and MCF7 cells were transfected with either an empty vector or UBR5 plasmids. 24 hours later, the cells were treated with IFN-γ for 24 h. Then the mRNA and protein levels of STAT1 (E and G) and IRF1 (F and G) were measured by qPCR and western blot, respectively. (H) Surface PD-L1 levels were detected in IFN-γ-treated GFP, Ubr5^-/- 4T1 cells and GFP 4T1 cells treated with siNC, siSTAT1 or siIRF1. GAPDH was used for normalization. (I) Luciferase reporter vectors containing either STAT1 or IRF1 promoter regions were cotransfected with an empty vector or UBR5 plasmids into the indicated cells. After 24 h, the transfected cells were treated with IFN-γ stimulation for 24 h, the cells were lysed to perform luciferase assay. The results are presented as the mean ± SEM from three individual experiments. *P < 0.05, **P < 0.01, ***P< 0.001, ****P < 0.0001. (J) The IRF1 and STAT1 binding sites in the mPdl1 promoter region were predicted using the ALGGEN website. (K) Summary of the results of a ChIP assay using anti-IRF1, STAT1, H3K4me1 and H3K27ac antibodies in WT, Ubr5^-/- or hUBR5-reconstituted Ubr5^-/- 4T1 cells after treatment with or without IFN-γ.

Figure 6

UBR5-mediated transactivation of PDL1 is independent of its E3 ubiquitin ligase. (A-D) MDA-MB-231 cells were cotransfected with WT, HECT-mutation (UBR5-C2768A) or PABC-deletion UBR5 (A) together with luciferase reporters containing the human PDL1 WT promoter or a promoter with 2 deletions, including the STAT1/3 or IRF1 binding sites (B), and then stimulated with IFN-γ for 24 h. Luciferase activity was measured in cell lysates of MDA-MB-231 (C) and 4T1 cells (D) by dual luciferase assay. (E-F) The protein levels of UBR5 (E) and PD-L1 (F) were measured in IFN-γ-treated GFP, Ubr5^-/-, hUBR5-reconstituted Ubr5^-/- 4T1 cells and Ubr5^-/- 4T1 cells reconstituted with either hUBR5-C2768 or hUBR5-ΔPABC. The results are presented as the mean ± SEM from three individual experiments. ns, no significance, *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 7

The transactivation of UBR5 to STAT1 and PDL1 is mediated by protein kinase RNA-activated. (A-B) The mRNA (A) and protein levels (B) of EIF2AK2 in WT, Ubr5^-/-, hUBR5-reconstituted Ubr5^-/- 4T1 cells and Ubr5^-/- 4T1 cells reconstituted with either hUBR5-C2768A or hUBR5-ΔPABC were measured by qPCR and western blot analysis after cells were treated with or without IFN-γ. (C-D) The mRNA (C) and protein levels (D) levels of EIF2AK2 in WT and UBR5-knockdown BT549 cells were measured by qPCR and western blot analysis after cells were treated with IFN-γ for 24 h. (E-G) The mRNA (E) and protein levels of UBR5, STAT1, IRF1 and EIF2AK2 (F) and relative fold changes of cell-surface PD-L1 levels (G) in IFN-γ-treated WT, Ubr5^-/-, and 4T1 cells with stable knockdown by shEif2ak2 #1 and shEif2ak2 #1. Shscramble served as the control. (H-J) The mRNA levels (H) of Eif2ak2, Ubr5, Pdl1, Stat1 and Irf1, the protein levels of UBR5, EIF2AK2, STAT1 and IRF1 (I) and the surface levels of PD-L1 (J) were measured in IFN-γ-treated 4T1 WT, Ubr5^-/- and mEif2ak2-reconstituted Ubr5^-/- cell lines. The results are presented as the mean ± SEM from three individual experiments. ns, no significance, *P< 0.05, **P < 0.01, ***P< 0.001, ****P < 0.0001. (K) The correlations of EIF2AK2 with UBR5, PDL1 and STAT1 mRNA levels were assessed in the TCGA BRCA database (1104 samples) by starBase database. (L) The summary correlations of EIF2AK2 with UBR5, PDL1, STAT1 and IRF1 mRNA levels (normalizated to GAPDH) were analysed in the TCGA BRCA database by starBase and GEPIA database respectively.