EBV miRNAs BART11 and BART17-3p promote immune escape through the enhancer-mediated transcription of PD-L1

Epstein-Barr virus (EBV) is reportedly the first identified human tumor virus, and is closely related to the occurrence and development of nasopharyngeal carcinoma (NPC), gastric carcinoma (GC), and several lymphomas. PD-L1 expression is elevated in EBV-positive NPC and GC tissues; however, the specific mechanisms underlying the EBV-dependent promotion of PD-L1 expression to induce immune escape warrant clarification. EBV encodes 44 mature miRNAs. In this study, we find that EBV-miR-BART11 and EBV-miR-BART17-3p upregulate the expression of PD-L1 in EBV-associated NPC and GC. Furthermore, EBV-miR-BART11 targets FOXP1, EBV-miR-BART17-3p targets PBRM1, and FOXP1 and PBRM1 bind to the enhancer region of PD-L1 to inhibit its expression. Therefore, EBV-miR-BART11 and EBV-miR-BART17-3p inhibit FOXP1 and PBRM1, respectively, and enhance the transcription of PD-L1 (CD274, http://www.ncbi.nlm.nih.gov/gene/29126), resulting in the promotion of tumor immune escape, which provides insights into potential targets for EBV-related tumor immunotherapy.

g After transfection of EBV-miR-BART17-3p mimics or negative control into EBV-negative HONE1 and AGS cells, the anti-AGO2 antibody was used for RIP experiments and DNA electrophoresis of qRT-PCR products was performed to identify whether EBV-miR-BART17-3p showed binding to the PBRM1 3′UTR via AGO2.
b Flow cytometric analysis of PD-L1 expression in HONE1 and HONE1-EBV cells transfected with the PBRM1 overexpression vector, siPBRM1, EBV-miR-BART17-3p mimics or inhibitors, or co-transfected with EBV-miR-BART17-3p mimics and the PBRM1 overexpression vector, or EBV-miR-BART17-3p Data are presented as mean ± s.d, p values are calculated by unpaired two-sided t-test in a. Source data are provided as a Source Data file. Supplementary Fig. 6 EBV-miBART11 directly targets FOXP1. a EBV-miR-BART11-3p was highly expressed in 156 NPCs compared with the 9 NPE tissue samples from GSE32960, and EBV-miR-BART11-5p expression was high in 62 NPCs compared to 6 NPEs of GSE36682.
b FOXP1 expression and the correlation between FOXP1 and PD-L1 expression was analyzed in 31 NPCs and 10 NPEs from GSE12452.
Data are presented as mean ± s.d, p values are calculated by unpaired two-sided t-test in a, b(left), d, f. b(right) are calculated by linear regression. Source data are provided as a Source Data file.
Data are presented as mean ± s.d, p values are calculated by unpaired two-sided t-test in c. Source data are provided as a Source Data file. b The correlation analysis between EBV-miR-BART17-3p, EBV-miR-BART11-3p, EBV-miR-BART11-5p, PBRM1, FOXP1, and PD-L1 was performed based on the ISH or IHC data in NPC.
d The correlation analysis between EBV-miR-BART17-3p, EBV-miR-BART11-3p, EBV-miR-BART11-5p, PBRM1, FOXP1, and PD-L1 was performed based on the ISH or IHC data in gastric adenocarcinoma. c ChIP experiments using the antibodies against FOXP1 or PBRM1 were performed to identify whether knockdown of FOXP1 or PBRM1 affected the binding of FOXP1 and PBRM1 in the PD-L1 enhancers A, B, C, D, and E in HONE1 cells. n = 3 biologically independent samples.
d ChIP experiments using the antibodies against FOXP1, PBRM1, H3K27ac, and H3K4me1 were performed to identify whether knockdown of FOXP1 or PBRM1 affected the binding of FOXP1 and PBRM1, and the H3K27ac and H3K4me1 modification in the PD-L1 enhancers B and E in AGS, C666-1, and SNU-719 cells transfected with FOXP1 or PBRM1 siRNA. n = 3 biologically independent samples. e The EMSA assay was used to detect whether PBRM1 binds to the PD-L1 enhancers B and E in HONE1.
Lane 1: only biotin-labeled probes were added; lane 2: nuclear protein and biotin-labeled region probes were added; lane 3: nuclear protein was added, biotin-labeled region probes and competitively bound unlabeled region probes were added in a ratio of 1:2; lane 4: nuclear protein was added and biotin-labeled probes and mutant unlabeled probes in a ratio of 1:2 were added; lane 5: nuclear protein, biotin-labeled probes and anti-PBRM1 antibodies were added simultaneously; lane 6: biotin-labeled probes and anti-PBRM1 antibody were added; lane 7: nuclear protein, nonspecific probe, and anti-PBRM1 antibody were added simultaneously.
f The 3C experiments were performed in HONE1, C666-1, and SNU-719 cells transfected with EBV-miR-BART11 or EBV-miR-BART17-3p mimics or inhibitors to detect the interaction frequency between the enhancers and the PD-L1 promoter. The relative interaction frequency was normalized to the closest BamH I digestion site. n = 3 biologically independent samples. g Luciferase reporter gene assays showed that EBV-miR-BART11 or EBV-miR-BART17-3p affects the reporter activities of the PD-L1 enhancer B and E regions via FOXP1 or PBRM1 in C666-1 and SNU-719 cells co-transfected with the B and E wild-types or the corresponding mutant vectors, and EBV-miR-BART11 or EBV-miR-BART17-3p inhibitors. n = 3 biologically independent samples. b The correlation between DPF2 and PD-L1 expression and between DPF2 and PBRM1 expression was analyzed using the data from GSE12452. c qRT-PCR was used to detect the mRNA for DPF2 in HONE1, AGS, C666-1, and SNU-719 cells after knockdown of DPF2. GAPDH was used as an internal control. n = 3 biologically independent samples. a Primary T cells were transfected with EBV-miR-BART11 and EBV-miR-BART17-3p mimics. qRT-PCR was used to confirm the transfection efficiency. n = 3 biologically independent samples.
b Flow cytometric analysis of T-cell apoptosis in T cells transfected with EBV-miR-BART11 and EBV-miR-BART17-3p mimics. Each group was analyzed using three independent replicates. a, b are calculated by unpaired two-sided t-test. Source data are provided as a Source Data file. Supplementary Fig. 16 Original flow cytometry results for Fig. 7e for detecting IFN-γ secretion. Primary T cells were co-cultured with HONE1 or AGS cells after overexpression of EBV-miR-BART11 and EBV-miR-BART17-3p, or co-cultured with HONE1-EBV or AGS-EBV cells after inhibition of EBV-miR-BART11 and EBV-miR-BART17-3p. n = 3 biologically independent samples. Supplementary Fig. 17 Flow cytometric analysis of T-cell apoptosis in HONE1 and AGS cells treated with IFN-γ (10 ng/mL) and EBV-miR-BART11 and EBV-miR-BART17-3p mimics. Each group was analyzed using three independent replicates. Data are presented as mean ± s.d, and, p values are calculated by unpaired two-sided t-test in a-f. Source data are provided as a Source Data file.  Fig. 7f. EBV-miR-BART11 can induce Jurkat cells apoptosis by inhibiting FOXP1. Each group was analyzed using three independent replicates. Supplementary Fig. 22 The effects of EBV-miR-BART11 and EBV-miR-BART17-3p on tumor immune escape in xenograft mice models. a A schematic diagram of the CDX mice models after injection of activated T cells. A density of 5×10 6 HONE1 or AGS cells were transfected with EBV-miR-BART11 and EBV-miR-BART17-3p mimics or negative control, HONE1-EBV or AGS-EBV cells were transfected with EBV-miR-BART11 and EBV-miR-BART17-3p inhibitors or negative control. Transfected cells were injected subcutaneously into the root of the right thigh of the mice. DCs and T cells were prepared and expanded simultaneously. DCs were first co-cultured with tumor cell lysate, and then co-cultured with T cells to present tumor cells specific tumor antigens to T cells, which enables them to produce tumor cells specific T cells. After 7 days, palpable tumors were formed, and 5×10 7 DiR ± T cells were infused into CDX mice through the tail vein. For the DiR-injected mice, the small animal live imaging system was used to observe the accumulation of DiR+T cells at the tumor-forming site. For one set of non-DIR-injected mice, activated T cells were injected into each tumor for another 7 days. Next, the peripheral blood of mice was extracted for flow cytometry, qRT-PCR, and ELISA. In another set, activated T cells were injected into each tumor for another 25 days and the tumor weight and volume were measured each day. The PD-L1 inhibitor (5 mg/kg, Atezolizumab) was f Expression of EBV-miR-BART17-3p, EBV-miR-BART11-3p, and EBV-miR-BART11-5p analyzed ascertained via ISH and the expression of PBRM1, FOXP1, PD-L1, CD8, cleaved-Caspase 3, and cleaved-PARP proteins determined via IHC were examined in the CDX nude mice sections. Magnification: 400×; scale bars = 20 µm.
Data are presented as mean ± s.d, and p values are calculated by unpaired two-sided t-test in c-e, g. Source data are provided as a Source Data file. Supplementary Fig. 23 The gating strategy for the detection of PD-L1 expression in tumor cells and the analysis of CD8 + T cells.
a FACS gating strategy shown with a representative example for the detection of PD-L1 expression in HONE1 or HONE1-EBV cells after transfected.
b Depicted is the gating path in Co-cultured cells or mouse peripheral blood cells. Depending on the respective experimental part, FITC-PI assay of cell apoptosis was analyzed in the human CD8+ T cells and IFN-γ as a marker for T cell killing function was only assessed in Co-cultured CD8+ T cells with the addition of Brefeldin A.  NA  NA  NA  NA  0  10  12  0  0  0  0  Pat 002   normal gastric mucosa tissues   NA  NA  NA  NA  0  9  10  0  0  0  0  Pat 003   normal gastric mucosa tissues   NA  NA  NA  NA  0  11  9  0  0  0  0  Pat 004   normal gastric mucosa tissues   NA  NA  NA  NA  0  8  10  0  0  0  0  Pat 005   normal gastric mucosa tissues   NA  NA  NA  NA  1  9  11  2  1  1  3  Pat 006   normal gastric mucosa tissues   NA  NA  NA  NA  0  10  10  0  0  0  0  Pat 007   normal gastric mucosa tissues   NA  NA  NA  NA  0  9  8  0  0  0  0  Pat 008   normal gastric mucosa tissues   NA  NA  NA  NA  0  10   Supplementary Table 2. Clinicopathological data on 40 paraffin-embedded gastric adenocarcinoma (GC) and 20 normal normal gastric mucosa biopsies and the expression of PD-L1, FOXP1, PBRM1, EBER, BART11-3p, BART11-5p, and BART17-3p in these samples measured by immunohistochemistry (IHC) and in situ hybridization(ISH) .