Nasopharyngeal carcinoma (NPC) and Epstein–Barr virus (EBV)–associated gastric carcinoma (EBVaGC) are two major EBV-associated epithelial malignancies, both of which are characterized by the infiltration of a large number of lymphocytes, including natural killer (NK) cells. Although NK cells can prevent the development of EBV-associated epithelial malignancies, EBV-infected tumor cells often develop resistance to surveillance by NK cells. Elucidating the interactions between NK cells and EBV-infected tumor cells will facilitate the development of more effective NK-mediated therapies for treating EBV-associated malignancies. Here we investigated the cytotoxic function of NK cells in EBV-associated epithelial malignancies and discovered that EBV infection-induced upregulation of F3 expression correlates with NK-cell dysfunction in NPC and EBVaGC. The subsequent inhibitory effect of F3-mediated platelet aggregation on NK-cell function was verified in vitro and in vivo. Mechanistically, EBV latent membrane protein 2A (LMP2A) mediated upregulation of F3 through the PI3K/AKT signaling pathway. In an NPC xenograft mouse model, inhibition of F3 restored the antitumor function of NK cells and showed therapeutic efficacy when administered with NK-cell transfer. On the basis of these findings, EBV infection induces F3-mediated platelet aggregation that inhibits the antitumor function of NK cells, providing a rationale for developing and combining NK-cell–based therapies with F3 inhibitors to treat EBV-associated epithelial malignancies.

Significance:

This study reveals a mechanism by which EBV-associated epithelial malignancies escape NK-cell–mediated immune surveillance, providing a new target for improving NK-cell immunotherapy.

More than 90% of the global adult population is infected with Epstein–Barr virus (EBV), the first reported human oncogenic virus. EBV is the causative agent of infectious mononucleosis (IM) and lymphoproliferative disorders in immunocompromised patients. It is also causally linked to various lymphomas and epithelial cell carcinomas, including Burkitt lymphoma, Hodgkin lymphoma, natural killer (NK)/T-cell lymphoma, diffuse large B-cell lymphoma, nasopharyngeal carcinoma (NPC), a subset of EBV-associated gastric carcinoma (EBVaGC), and pulmonary lymphoepithelioma-like carcinoma (pLELC; refs. 1–3). NPC and EBVaGC, two of the most common EBV-associated epithelial malignancies, account for approximately 80% of all EBV-associated cancers (4). More than 90% of undifferentiated NPC cases are related to EBV infection, and EBVaGC accounts for approximately 10% of all gastric cancers (5).

EBV establishes a persistent latent infection in malignant epithelial cells. Type II EBV latency is observed in NPC, with expression of some noncoding RNAs and proteins, such as EBV-encoded small RNAs (EBER), BamHI A rightward transcript-microRNAs (BART), EBV-associated nuclear antigen-1 (EBNA1), and latent membrane protein 1 and 2 (LMP1 and LMP2A/B), whereas an EBV latency intermediate between type I and type II is found in EBVaGC. To date, many gene products, primarily but not limited to latent gene products, have been proven to mediate a variety of oncogenic properties and events, such as genomic instability, epigenetic modification, survival promotion, stemness, and immune evasion (5).

Stroma inflammation and infiltration of a large number of lymphocytes are common histopathologic features of undifferentiated NPC and EBVaGC (6, 7). Multiple immune cell types have been identified in the tumor microenvironment (TME) of NPC and EBVaGC. EBV-positive cancer cells express a variety of viral antigens and are thought to be controlled mainly by antigen-specific cytotoxic T lymphocytes (CTL). Refractory patients with NPC have been treated with an infusion of EBV-primed CTLs with some success (8). Nonetheless, downregulation of human leukocyte antigen (HLA) class I, a molecule required for the priming and activation of CTLs, is a common evasion mechanism in EBV-associated NPC tumors (9). This loss of HLA class I molecules on NPC presents an opportunity for natural killer (NK) cells to target and eliminate these cancer cells since NK-cell–mediated cytotoxicity is not HLA dependent. NK cells rapidly respond to viral infection rapidly through the secretion of antiviral cytokines and direct lysis of virus-infected tumor cells. Existing evidence indicates that the role of NK cells in preventing the progression of EBV-associated malignancies as well as persistent viral infection originates from primary immunodeficiencies in which genetic lesions affect gene products involved in NK-cell differentiation and activation. Recently, the NK-specific immune signature in tumors has been found to correlate with favorable prognosis of NPC, suggesting that NK cells in the TME are involved in controlling the progression of EBV-associated epithelial tumors (6). However, a lower percentage of NK cells with activating receptors or a higher percentage of NK cells expressing inhibitory ligands were identified in patients with NPC than in healthy controls (10). These findings indicated reductions in the number of functional NK cells in patients with EBV-associated epithelial tumors. Understanding the immune interactions between host NK cells and EBV-associated malignancies will facilitate development of more effective NK-mediated therapies for treating these malignancies.

Abnormal platelet aggregation and activation in tumors are major contributors to the development of resistance to the antitumor function of NK cells and act via several pathways. Activated platelets accumulate around tumor cells and form a physical barrier to prevent NK-cell access (11). Tumor-activated platelets downregulate activating receptors on the surface of NK cells by releasing multiple active factors, such as IL10 and TGFβ (12). Platelets also transport their own MHC-I molecules to the surface of tumor cells, which then bind to the inhibitory receptors on NK cells to mitigate cytotoxic NK-cell signaling (13). Various reports have described the different degrees of hypercoagulability in patients with NPC and EBVaGC in vivo (14, 15). However, whether EBV contributes to the observed platelet activation and consequently to the resistance of EBV-associated epithelial cells to the antitumor function of NK cells remains unexplored.

Here, we report EBV escape from NK cell–mediated immune surveillance by upregulating F3 and promoting platelet activation in NPC and EBVaGC. We demonstrate the potential application of F3 as a therapeutic target for EBV-associated epithelial cancers.

Cell lines and cell culture

The cell lines K562 (ATCC catalog no. CCL-243, RRID:CVCL_0004), NK92MI and HEK293T (ATCC catalog no. CRL-3216, RRID:CVCL_0063) were purchased from ATCC. The cell lines CNE2, HK1, and Akata-EBV were a kind gift from Prof. Musheng Zeng (Sun Yat-sen University Cancer Center). The AGS cell line (ATCC catalog no. CRL-1739, RRID:CVCL_0139) was a kind gift from Prof. Ruihua Xu (Sun Yat-sen University Cancer Center). These cell lines were identified by STR genotyping. CNE2, HK1, AGS, Akata-EBV, and K562 cell line were cultured in RPMI1640 medium supplemented with 10% FBS (Gibco). NK92MI, an NK cell line, was cultured in special medium containing MEMα, 0.2 mmol/L inositol, 0.02 mmol/L folic acid, 0.1 mmol/L β-mercaptoethanol, 12.5% horse serum, and 12.5% FBS. The HEK293T cell line (ATCC catalog no. CRL-3216, RRID:CVCL_0063) was cultured in DMEM supplemented with 10% FBS. All cell lines were detected using a PCR-based method (16S rDNA-F: 5′-ACTCC TACGGGAGGCAGCAGTA-3′, 16S rDNA-R: 5′-TGCACCATCTGTCACTCTG TTAACCTC-3′) and shown to be free of Mycoplasma contamination at the beginning of this study.

Antibodies and chemicals

Antibodies specific for the following proteins were used for Western blot in this study: mTOR (2983, CST), p-mTOR (Ser2448) (2971, CST), AKT (9272, CST), pAKT (Ser473; ab81283, Abcam), F3 (ab104513, Abcam), HA (C29F4, CST), and GAPDH (60004–1-Ig, Proteintech). Antibodies specific for the following proteins were used for IHC in this study: F3 (Abcam catalog no. ab104513, RRID:AB_10711603), and CD41 (Abcam catalog no. ab134131, RRID:AB_2732852). Antibodies specific for the following proteins were used for flow cytometry in this study: CD41-FITC (Miltenyi Biotec catalog no. 130–105–929, RRID:AB_2658008), CD62P-PE (Miltenyi Biotec catalog no. 130–105–536, RRID:AB_2658875), human CD3-FITC (BioLegend catalog no. 344804, RRID:AB_2043993), mouse CD45-Brilliant Violet 510 (103138, Biolegend), human CD45-APC/Cyanine7 (103138, Biolegend), human CD56-PE/Cyanine7 (BioLegend catalog no. 304628, RRID:AB_2149542), human CD107a-PE (328608, Biolegend), and monoclonal human NKG2D-APC (320807, Biolegend).

LY294002 (S1105) and wortmannin (S2758) were purchased from Selleckchem. PMA (TQ0198), ionomycin (T11665), and monensin (T1033) were purchased from TargetMol.

Establishment of EBV-infected cells

The EBV Akata strain used for infecting epithelial cancer cells were isolated as described previously (16). Infected CNE2, HK1, and AGS cells (ATCC catalog no. CRL-1739, RRID:CVCL_0139) were sorted by flow cytometry (Beckman) and maintained in RPMI1640 medium containing 700 to 1,600 μg/mL G418. Latent EBV infection in cell lines was confirmed by in situ hybridization of EBERs with a kit (Zhongshan Jinqiao).

Western blot analysis

Proteins from different cell lines were extracted with SDS lysis buffer, separated by SDS-PAGE through 10% gels, electrotransferred to PVDF membranes (Millipore), and probed with primary antibodies followed by secondary antibodies according to standard procedures. Immunoreactive bands were visualized in a dark room with enhanced chemiluminescence (ECL) reagents (Pierce).

Molecular constructs

Three individual sgRNAs targeting F3 and LMP2A were cloned into pLentiCRISPR v2 (RRID:Addgene_102315). The primers were synthesized by Beijing Ruibiotech Co., Ltd., and the sequences were shown in Supplementary Table S1.

IHC and multiplex IHC

IHC was performed on paraffin-embedded sections. In brief, tissues were deparaffinized, rehydrated, subjected to antigen retrieval, and blocked with serum. Subsequently, the samples were incubated with diluted primary antibodies at 4°C overnight, incubated with the secondary antibody at 37°C for 1 hour and colorized with 3,3′-diaminobenzidine (DAB). Finally, images were acquired under a white light microscope for analysis. The staining process was repeated at least twice for each sample. The semiquantitative scoring method used was described previously (16). The sections were reviewed by two independent pathologists, and the staining scores were obtained by multiplying the staining density by the staining intensity.

For multiplex IHC, slides were subjected to antigen retrieval and were then subjected to several cycles of incubation with different primary antibodies and fluorescent secondary antibodies at Panovue. Then, 4′,6-diamidino-2-phenylindole (DAPI) was used to stain nuclei. The images were analyzed as described previously (17).

Platelet activation assays

Platelets were processed for activation as described previously (12). Citrated blood obtained from healthy donors who had not taken any medicine for 2 weeks prior to blood collection was centrifuged at 120 × g for 20 minutes at room temperature; the upper layer was platelet-rich plasma. Platelets were activated by adding an equal volume of tumor cell supernatant in the presence of 1 mmol/L calcium chloride and stirring for 10 minutes at 1,000 rpm in a 37°C incubator. To evaluate platelet activation, 10 μL of treated platelets was stained for CD41 and CD62P, and 100 μL of treated platelets was observed under a laser confocal microscope. Subsequently, 500 μL of treated platelets was centrifuged for 15 minutes at 13,000 rpm. The supernatant was used for NK degranulation assays and NK cytotoxicity assays.

NK-cell degranulation assays

NK degranulation assays were performed as described (18). In brief, the platelet releasate was cocultured with NK cells (NK92MI or NK cells from umbilical cord blood) overnight. Then, the treated NK cells were incubated with the target cells (K562, HK1, or AGS) at the indicated effector: target ratio. After 1 hour, the cells were stained with an antibody against CD107a in the presence of monensin (GolgiStop; BD Biosciences). Four hours later, the cells were stained with antibodies against CD3 (BioLegend catalog no. 344804, RRID:AB_2043993) and CD56 (BioLegend catalog no. 304628, RRID:AB_2149542). The frequency of degranulating NK cells (CD3CD56+CD107a+) was determined by FACS and evaluated by normalization data of NK cells cultured without target cells (effector:target ratio = 1:0). The positive control for the degranulation assay was NK cells supplemented with 100 ng/mL PMA and 2 μg/mL ionomycin.

NK cytotoxicity assays

Target cells (K562, HK1, or AGS) were labeled with CSFE (FITC channel; 65085084, eBioscience) and cultured with treated NK cells at different ratios. Four hours later, the cells were stained with the viability probe 7-AAD (PC5.5 channel; 00699350, Invitrogen), and NK-cell cytotoxicity was determined by FACS analysis as the percentage of dead K562, HK1, or AGS cells (ATCC catalog no. CRL-1739, RRID:CVCL_0139; FITC+PC5.5+). Dye toxicity and/or spontaneous K562, HK1, or AGS cell death were controlled for by culture in the absence of NK cells and never exceeded 3%.

Patient studies

We received consent for publication from all patients who provided samples for research. This study was approved by the Human Research Ethics Committee Sun Yat-sen University Cancer Center (G2021–046–01). Samples from 10 patients with EBV-positive NPC (10 men ranging in age from 37 to 61 years, all of whom had primary NPC tumors), 5 patients with EBV-negative NPC (2 men and 3 women ranging in age from 41 to 70 years, 1 patient with a primary tumor and 4 with recurrent tumors), 9 patients with EBV-positive GC (9 men ranging in age from 52 to 73 years), and 20 patients with EBV-negative GC (14 men and 6 women ranging in age from 28 to 82 years), 9 patients with EBV-positive lymphoma (4 men and 5 women ranging in age from 37 to 70 years), and 8 patients with EBV-negative lymphoma (4 men and 4 women ranging in age from 35 to 70 years), 9 patients with EBV-positive lung cancer (4 men and 5 women ranging in age from 43 to 67 years), and 8 patients with EBV-negative lung cancer (4 men and 4 women ranging in age from 40 to 67 years) were obtained from Sun Yat-sen University Cancer Center. All samples were diagnostically confirmed by two pathologists (Dr. Jinping Yun and Dr. Peng Sun, Sun Yat-sen University Cancer Center).

Animal study

This study was approved by the Experimental Animal Ethics Committee of Sun Yat-sen University Cancer Center (L102012021002B). Five-week-old female nude mice (BALB/c-nu/nu) purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. were injected subcutaneously with 5 × 106 CNE2- or CNE2-derived cells. Five-week-old female NOD/ShiLtJGpt-Prkdcem26Cd52Il2rgem26Cd22/Gpt mice (RRID:IMSR_GPT:T001475) were injected subcutaneously with 5 × 106 HK1-EBV- or HK1-EBV-derived cells. After several days of inoculation, 1.5 × 107 NK cells from umbilical cord blood were infused intravenously. NK-cell survival was supported by injection of hIL-2 (10,000 units/mouse every 2 days for 21 days) and hIL-15 (10 ng/mouse for 7 days) during NK-cell therapy after day 1. After the NK-cell infusion, peritumoral injection of human TFPI (0.66 mg/kg) was performed daily. The tumor volume (V) was measured every 2 days and calculated as follows: V (mm3)  =  length × width2/2. At the end of the experiment, tumor cells were analyzed by FACS.

Flow cytometry

Single-cell suspensions prepared from cocultures in vitro and xenograft tumors in vivo were immunostained in FACS buffer for 30 minutes at 4°C. Data were acquired and analyzed using a multicolor flow cytometer (CytExpert).

Statistical analysis

Data with a normal distribution and significantly different variances as determined by the F test were compared by a two-tailed unpaired t test. Data with a nonnormal distribution were compared with a two-tailed Mann–Whitney test. The Pearson correlation coefficient was used to determine the correlations of platelet areas with F3 expression levels. The following indicators and values were used to convey statistical significance: *, P < 0.05; **, P < 0.01; ***, P < 0.001; and ****, P < 0.0001.

Data availability statement

The authenticity of this article has been validated by uploading the key raw data onto the Research Data Deposit public platform (www.researchdata.org.cn), with the approval RDD number as RDDB2022251242. The transcriptome has been deposited in the NCBI SRA database (NCBI SRA, RRID:SCR_004891) under the accession number SRP118175 (https://trace.ncbi.nlm.nih.gov/Traces/sra/?study=SRP118175) or SRP354343 (https://trace.ncbi.nlm.nih.gov/Traces/sra/?study=SRP354343). The data that support the findings of this study are available from the corresponding authors upon reasonable request.

EBV infection is correlated with NK-cell dysfunction and upregulation of F3 expression in EBV-associated epithelial malignancies

We compared the transcriptomes of 94 EBV-positive NPC tumors with those of 13 EBV-negative NPC tumors to determine whether EBV infection affects NK-cell function and found that the immune signature of NK cells was significantly downregulated in EBV-positive NPC tumors (Fig. 1A). The effects of other confounding factors, such as age, sex, smoking, drinking, and tumor stage on NK cells were excluded (Table 1; Supplementary Figs. S1A–S1E). The number of NK cells (CD3CD56+) in EBV-positive tumor biopsies was no statistical difference from that in EBV-negative tumor biopsies according to the results of multiplex IHC assays (Supplementary Fig. S1F). However, the number of cytotoxic NK cells (CD3CD56+Granzyme B+) that infiltrated in EBV-positive tumor biopsies was reduced compared with that in EBV-negative tumor biopsies (Fig. 1B and C). High NK signature scores were associated with prolonged survival of patients in NPC and GC (Supplementary Fig. S1G). The data suggest that EBV is a potent immunosuppressor that affects NK-cell function, but not their infiltration.

Figure 1.

EBV infection is correlated with NK-cell dysfunction and upregulation of F3 expression in EBV-associated epithelial malignancies. A, Analysis of NK-cell activity in 94 EBV-positive and 13 EBV-negative NPC samples using transcriptome sequencing of patient tumor biopsies. ssGSEA was used to calculate the single sample signature scores (means ± SD). B, Images of cytotoxic NK cells in EBV−/+ NPC and GC tumor biopsies by multiplex IHC. White scale bars, 50 μm. C, The number of CD3CD56+Granzyme B+ cells localized in the stroma and tumor per unit area were counted to quantify the number of cytotoxic NK cells in EBV−/+ epithelial cancers (7 patients with EBV-positive NPC, 4 patients with EBV-negative NPC; 8 patients with EBV-positive GC, 9 patients with EBV-negative GC). Means ± SD. D, Gene ontology analysis of upregulated genes between CNE2 and CNE2-EBV cells. The “Cluster Profiler” R package was used to perform the enrichment analysis. E, Heatmap showing the upregulated genes in the complement and coagulation cascade pathways in EBV-infected CNE2 cells. F, Images of F3 and CD41 expression in EBV−/+ NPC tumor biopsies as assessed by IHC. Black scale bars, 100 μm. White scale bars, 50 μm. G and H, Quantification of F3 expression and platelet aggregation in EBV−/+ NPC tumor biopsies (5 patients with EBV-positive NPC, 5 patients with EBV-negative NPC (means ± SD). I, Pearson correlation analysis of F3 expression and platelet aggregation in NPC tumor biopsies (r, regression coefficient). Two-tailed Mann–Whitney U test; *, P < 0.05; **, P < 0.01; ****, P < 0.0001; NS, not significant. H&E, hematoxylin and eosin.

Figure 1.

EBV infection is correlated with NK-cell dysfunction and upregulation of F3 expression in EBV-associated epithelial malignancies. A, Analysis of NK-cell activity in 94 EBV-positive and 13 EBV-negative NPC samples using transcriptome sequencing of patient tumor biopsies. ssGSEA was used to calculate the single sample signature scores (means ± SD). B, Images of cytotoxic NK cells in EBV−/+ NPC and GC tumor biopsies by multiplex IHC. White scale bars, 50 μm. C, The number of CD3CD56+Granzyme B+ cells localized in the stroma and tumor per unit area were counted to quantify the number of cytotoxic NK cells in EBV−/+ epithelial cancers (7 patients with EBV-positive NPC, 4 patients with EBV-negative NPC; 8 patients with EBV-positive GC, 9 patients with EBV-negative GC). Means ± SD. D, Gene ontology analysis of upregulated genes between CNE2 and CNE2-EBV cells. The “Cluster Profiler” R package was used to perform the enrichment analysis. E, Heatmap showing the upregulated genes in the complement and coagulation cascade pathways in EBV-infected CNE2 cells. F, Images of F3 and CD41 expression in EBV−/+ NPC tumor biopsies as assessed by IHC. Black scale bars, 100 μm. White scale bars, 50 μm. G and H, Quantification of F3 expression and platelet aggregation in EBV−/+ NPC tumor biopsies (5 patients with EBV-positive NPC, 5 patients with EBV-negative NPC (means ± SD). I, Pearson correlation analysis of F3 expression and platelet aggregation in NPC tumor biopsies (r, regression coefficient). Two-tailed Mann–Whitney U test; *, P < 0.05; **, P < 0.01; ****, P < 0.0001; NS, not significant. H&E, hematoxylin and eosin.

Close modal
Table 1.

The characteristics of patients with NPC and their correlation with NK score.

Clinical parametersCasesP value
EBV (−/+)  0.0422 
 EBV+ 94  
 EBV− 13  
Age  0.31 
 Mean (SD) 46.24 (12.11)  
Sex  0.1278 
 Male 85  
 Female 22  
Smoking  0.6607 
 Yes 29  
 No 47  
 Unknown 31  
Drinking  0.8929 
 Yes  
 No 71  
 Unknown 31  
Tumor stage  0.45 
 I–II  
 III–IV 63  
 Unknown 37  
Clinical parametersCasesP value
EBV (−/+)  0.0422 
 EBV+ 94  
 EBV− 13  
Age  0.31 
 Mean (SD) 46.24 (12.11)  
Sex  0.1278 
 Male 85  
 Female 22  
Smoking  0.6607 
 Yes 29  
 No 47  
 Unknown 31  
Drinking  0.8929 
 Yes  
 No 71  
 Unknown 31  
Tumor stage  0.45 
 I–II  
 III–IV 63  
 Unknown 37  

To investigate the mechanism by which EBV infection affects NK-cell function, we searched for the differentially expressed genes (DEG) altered by EBV infection. We compared the gene expression profiles of paired EBV-positive (CNE2-EBV and TW03-EBV) and EBV-negative (CNE2 and TW03) NPC cell lines. Then, the mRNA expression profiles were subjected to Gene Ontology (GO) enrichment analysis, and we identified several signaling pathways that were enriched in EBV-positive NPC cells, including the coagulation, PI3K/AKT and MAPK pathways (Fig. 1D; Supplementary Fig. S1H). Regarding these pathways, platelet activation caused by abnormal coagulation in tumors has been reported to suppress NK function (12, 13, 19). Among the genes in the coagulation pathway that were upregulated by EBV infection (Fig. 1E; Supplementary Fig. S1I), F3 is an important factor initiating exogenous coagulation. We further found that the F3 expression was increased in EBV-positive NPC and GC tumor biopsies compared with EBV-negative tumor biopsies (Fig. 1F and G; Supplementary Figs. S1J and S1K). Moreover, the platelet aggregation diameter was increased significantly in EBV-positive NPC and GC tumor biopsies (Fig. 1F and H; Supplementary Figs. S1J and S1L). Consistent with these results, a strong positive correlation was observed between the platelet aggregation diameter and the F3 staining score in NPC and GC tumors (Fig. 1I; Supplementary Fig. S1M). These results demonstrated that F3 expression and platelet activation were elevated in EBV-positive NPC and GC tumors. Given that platelet activation initiated by F3 is an important regulator of NK-cell function (11, 12, 19), our data suggest that EBV potentially plays a role in the dysfunction of NK cells through F3-mediated platelet aggregation in NPC and EBVaGC.

EBV inhibits NK-cell function by upregulating F3 expression in NPC xenografts

In a xenograft model established in nude mice (20) in which NK cells have relatively normal cytotoxic activity, tumor-derived human F3 protein is responsible for the activation of coagulation (20) that mediates platelet aggregation (21). To further explore whether EBV infection in tumor cells induces resistance to NK-cell function in the TME via F3 upregulation and platelet aggregation, we subcutaneously inoculated nude mice with 5 × 106 EBV-positive or EBV-negative CNE2 cells. EBV infection in NPC cells led to faster tumor growth and a greater tumor burden over time (Fig. 2A and B). We observed upregulated F3 expression and increased platelet aggregation in the EBV-positive NPC xenograft model by IHC (Fig. 2CE) and found a positive correlation between the platelet aggregation diameter and the staining score of F3 in xenograft tumors (Fig. 2F). The ratio of cells expressing CD49b, the NK-cell marker infiltrating the xenograft in nude mice was not obviously different (Supplementary Fig. S2A). However, NKp46, a marker of mouse NK-cell activation, was detected in NPC xenografts using IHC, and the infiltration of NKp46+-activated NK cells was reduced in EBV-positive NPC xenograft tumors compared with EBV-negative tumors (Fig. 2G). The percentage of IFNγ+ NK cells was also decreased in EBV-positive NPC xenograft (Fig. 2H). We knocked down F3 in CNE2-EBV cells or overexpressed F3 in CNE2 cells and then subcutaneously injected 5 × 106 cells into nude mice to evaluate the impact of F3 on the infiltration of activated NK cells in inoculated tumors. Tumor volumes were significantly smaller in the F3 knockdown group than in the control group (Fig. 2I and J; Supplementary Fig. S2B). Overexpression or knockdown of F3 did not affect the ratio of infiltrating NK cells (Supplementary Fig. S2C). Interestingly, F3 knockdown partially restored the infiltration of NKp46+ activated NK cells and the percentage of IFNγ+ NK cells in inoculated tumors (Fig. 2K and L). Furthermore, F3 overexpression in EBV-negative CNE2 cells caused the acceleration growth of xenograft tumors compared with that of control tumors (Fig. 2M and N). As expected, the number of infiltrated NKp46+ activated NK cells and the percentage of IFNγ+ NK cells were decreased in F3-overexpressing xenograft tumors (Fig. 2O and P). The other typical cell markers (NKG2D, NKp46, KLRG1, CD43, CD122, and CD11b; ref. 22) of mouse NK cells in the xenograft tumor were analyzed using flow cytometry to analyze whether the typical NK-cell markers were affected. Levels of the two markers of activated mouse NK cells, NKp46 and NKG2D, were decreased in EBV- or F3-overexpressing xenograft tumors compared with their parental groups (Supplementary Figs. S2D and S2E). In contrast, the level of the inhibition marker KLRG1 was increased (Supplementary Fig. S2F). Moreover, NKp46 and NKG2D levels were increased in F3-knockdown xenograft tumors compared with parental groups (Supplementary Figs. S2D and S2E). However, the level of the inhibition marker KLRG1 was decreased (Supplementary Fig. S2F). Significant differences in the levels of mature NK-cell markers (CD43, CD122, and CD11b) were not observed among the paired CNE2/CNE2-EBV, CNE2-Vector/CNE2-F3, CNE2-EBV-Vector/CNE2-EBV-sgF3 tumors (Supplementary Figs. S2G–S2I). Collectively, these data indicate that EBV infection in NPC cells inhibits NK-cell function and promotes tumor progression partially through upregulation and activation of the F3 pathway in vivo.

Figure 2.

EBV inhibits NK-cell function through upregulation of F3 and activation of platelets in NPC xenografts. A, Images of CNE2 and CNE2-EBV xenograft tumor tissues from nude mice. B, Growth curves of CNE2 and CNE2-EBV xenograft tumors. Means ± SD. Two-tailed unpaired Student t test; **, P < 0.01. C, Images of hematoxylin and eosin (H&E) staining, EBER staining, and IHC staining of F3 and CD41 in EBV−/+ NPC xenograft tumors. D and E, Quantification of F3 expression and platelet aggregation in NPC xenograft tumors. Means ± SD; two-tailed Mann–Whitney U test; ***, P < 0.001; ****, P < 0.0001. F, Pearson correlation analysis of F3 expression and platelet aggregation in NPC xenograft tumors (r, regression coefficient). G, Images (left) and quantification (right) of NKp46 expression in EBV−/+ NPC xenograft tumors as assessed by IHC. Means ± SD; n = 20; two-tailed Mann–Whitney U test; ****, P < 0.0001. H, Analysis of the percentage of IFNγ+ NK cells in EBV−/+ NPC xenograft tumors. Means ± SD; two-tailed unpaired Student t test; ***, P < 0.001. I, Images of CNE2-EBV-vector and CNE2-EBV-sgF3 xenograft tumors from nude mice. J, Growth curves of CNE2-EBV-vector and CNE2-EBV-sgF3 xenograft tumors. Means ± SD; two-tailed unpaired Student t test; *, P < 0.05. K, Images IHC stanning (left) and quantification (right) for NKp46 in CNE2-EBV-vector and CNE2-EBV-sgF3 xenograft tumors. Means ± SD; n = 20; two-tailed Mann–Whitney U test; ****, P < 0.0001. L, Analysis of the percentage of IFNγ+ NK cells in CNE2-EBV-vector and CNE2-EBV-sgF3 xenograft tumors. Means ± SD; two-tailed unpaired Student t test; **, P < 0.01. M, Images of CNE2-vector and CNE2-F3 xenograft tumors from nude mice. N, Growth curves of CNE2-vector and CNE2-F3 xenograft tumors. Means ± SD; two-tailed unpaired Student t test; **, P < 0.01. O, Images (left) and quantification (right) of NKp46 expression in CNE2-vector and CNE2-F3 xenograft tumors assessed using IHC. Means ± SD; n = 20; two-tailed Mann–Whitney U test; ****, P < 0.0001. P, Analysis of the percentage of IFNγ+ NK cells in CNE2-vector and CNE2-F3 xenograft tumors. Means ± SD. Two-tailed unpaired Student t test; **, P < 0.01. Black scale bars, 100 μm. White scale bars, 50 μm.

Figure 2.

EBV inhibits NK-cell function through upregulation of F3 and activation of platelets in NPC xenografts. A, Images of CNE2 and CNE2-EBV xenograft tumor tissues from nude mice. B, Growth curves of CNE2 and CNE2-EBV xenograft tumors. Means ± SD. Two-tailed unpaired Student t test; **, P < 0.01. C, Images of hematoxylin and eosin (H&E) staining, EBER staining, and IHC staining of F3 and CD41 in EBV−/+ NPC xenograft tumors. D and E, Quantification of F3 expression and platelet aggregation in NPC xenograft tumors. Means ± SD; two-tailed Mann–Whitney U test; ***, P < 0.001; ****, P < 0.0001. F, Pearson correlation analysis of F3 expression and platelet aggregation in NPC xenograft tumors (r, regression coefficient). G, Images (left) and quantification (right) of NKp46 expression in EBV−/+ NPC xenograft tumors as assessed by IHC. Means ± SD; n = 20; two-tailed Mann–Whitney U test; ****, P < 0.0001. H, Analysis of the percentage of IFNγ+ NK cells in EBV−/+ NPC xenograft tumors. Means ± SD; two-tailed unpaired Student t test; ***, P < 0.001. I, Images of CNE2-EBV-vector and CNE2-EBV-sgF3 xenograft tumors from nude mice. J, Growth curves of CNE2-EBV-vector and CNE2-EBV-sgF3 xenograft tumors. Means ± SD; two-tailed unpaired Student t test; *, P < 0.05. K, Images IHC stanning (left) and quantification (right) for NKp46 in CNE2-EBV-vector and CNE2-EBV-sgF3 xenograft tumors. Means ± SD; n = 20; two-tailed Mann–Whitney U test; ****, P < 0.0001. L, Analysis of the percentage of IFNγ+ NK cells in CNE2-EBV-vector and CNE2-EBV-sgF3 xenograft tumors. Means ± SD; two-tailed unpaired Student t test; **, P < 0.01. M, Images of CNE2-vector and CNE2-F3 xenograft tumors from nude mice. N, Growth curves of CNE2-vector and CNE2-F3 xenograft tumors. Means ± SD; two-tailed unpaired Student t test; **, P < 0.01. O, Images (left) and quantification (right) of NKp46 expression in CNE2-vector and CNE2-F3 xenograft tumors assessed using IHC. Means ± SD; n = 20; two-tailed Mann–Whitney U test; ****, P < 0.0001. P, Analysis of the percentage of IFNγ+ NK cells in CNE2-vector and CNE2-F3 xenograft tumors. Means ± SD. Two-tailed unpaired Student t test; **, P < 0.01. Black scale bars, 100 μm. White scale bars, 50 μm.

Close modal

F3 upregulation in EBV-positive epithelial cancer cells induces resistance to NK-cell function through platelet activation

Having shown that F3-mediated platelet aggregation induced by EBV infection in cancer cells is associated with NK-cell dysfunction in nude mice, we next investigated whether platelet activation promoted by cancer cell-derived F3 inhibits the antitumor function of NK cells in vitro. Effector NK cells were incubated with platelets releasate pretreated with the supernatant from in vitro cultured cancer cells, and NK-cell–mediated cytotoxicity was then evaluated by assessing the degranulation (as determined by measuring CD107a expression) of effector NK cells and direct lysis of target cells using flow cytometry after coculture of effector and target cells (Fig. 3A). Platelets co-incubated with the supernatant of EBV-positive NPC (HK1) and GC (AGS) cancer cells exhibited more obvious aggregation than those co-incubated with the supernatant of EBV-negative epithelial cancer cells (Fig. 3B; Supplementary Fig. S3A), indicating that the factors secreted by NPC and GC cells after EBV infection promoted platelet aggregation. The releasate of platelets pretreated with the supernatant from EBV-positive NPC and GC cell cultures inhibited CD107a expression on primary NK cells (Fig. 3C and D; Supplementary Figs. S3B and S3C). To further confirm whether the inhibitory effect of the supernatant from EBV-positive tumor cell cultures on NK-cell function is dependent on platelets, NK cells were treated directly with the supernatant from EBV-positive and EBV-negative NPC cells without platelets. The degranulation activity of effector NK cells displayed no significant difference (Supplementary Figs. S3D and S3E). Because EBV-positive NPC and GC tumors showed upregulated F3 expression and increased platelet activation, we further investigated whether exogenous platelet activation affected the antitumor function of NK cells through a mechanism depending on F3 in vitro. To investigate the effect of F3 on platelet aggregation, we knocked down F3 in EBV-positive cell lines (Supplementary Figs. S4A and S4B) and overexpressed F3 in EBV-negative cell lines (Supplementary Figs. S4C and S4D). The expression of the platelet activation marker CD62P was reduced after treatment with the supernatant from EBV-positive HK1 and AGS cells with F3 knockdown compared with that of mock-knockdown EBV-positive HK1 and AGS cells (Supplementary Figs. S4E and S4F). In contrast, CD62P was upregulated after treatment with the supernatant from EBV-negative HK1 and AGS cells with F3 overexpression compared with that of the control (Supplementary Figs. S4E and S4F). Furthermore, the releasate of platelets pretreated with supernatant from F3-knockdown EBV-positive HK1 or AGS cell cultures significantly increased the degranulation of primary NK cells (Supplementary Fig. S4G). On the contrary, the releasate of platelets pretreated with supernatant from F3-overexpression EBV-negative HK1 or AGS cell cultures significantly decreased the degranulation of primary NK cells (Supplementary Fig. S4G). These results further confirmed the importance of F3-mediated platelet aggregation in the resistance of EBV to the degranulation of NK cells.

Figure 3.

F3 upregulation in EBV-positive epithelial cancer cells induces resistance to NK-cell functions through platelet activation. A, Experimental model diagram of the mechanism by which F3 inhibits NK-cell functions by promoting coagulation in vitro. B, Images (left) and quantification (right) of the platelet aggregation diameter after costimulation with EBV−/+ tumor cell supernatant and CaCl2 (1 mmol/L), as assessed using microscopy. Means ± SD; n = 5. White scale bars, 10 μm. C, The effect on CD107a expression on primary NK cells exposed to K562 by the releasate of platelets that had been pretreated with the supernatant from HK1 and HK1-EBV cells. Means ± SD; n = 5. D, The effect of CD107a expression on primary NK cells exposed to HK1 by the releasate of platelets, which had been pretreated with the supernatant from F3-knockdown HK1-EBV or F3-overexpressing HK1 cells. Means ± SD; n = 5. E and F, NK-cell–mediated cytotoxicity towards target cells (K562 and HK1) in the presence of the releasate of platelets, which had been pretreated with the supernatant from HK1 and HK1-EBV cells. Means ± SD; n = 5. G and H, NK-cell–mediated cytotoxicity towards target cells (K562 and HK1) in the presence of the releasate of platelets, which had been pretreated with the supernatant from HK1-EBV cells coincubated with the F3 neutralizing antibody or IgG isotype control antibody. Means ± SD; n = 3. I and J, NK-cell–mediated cytotoxicity towards target cells (K562 and HK1) in the presence of the releasate of platelets, which had been pretreated with F3 protein or control. Means ± SD; n = 3. Two-tailed unpaired Student t test; *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Figure 3.

F3 upregulation in EBV-positive epithelial cancer cells induces resistance to NK-cell functions through platelet activation. A, Experimental model diagram of the mechanism by which F3 inhibits NK-cell functions by promoting coagulation in vitro. B, Images (left) and quantification (right) of the platelet aggregation diameter after costimulation with EBV−/+ tumor cell supernatant and CaCl2 (1 mmol/L), as assessed using microscopy. Means ± SD; n = 5. White scale bars, 10 μm. C, The effect on CD107a expression on primary NK cells exposed to K562 by the releasate of platelets that had been pretreated with the supernatant from HK1 and HK1-EBV cells. Means ± SD; n = 5. D, The effect of CD107a expression on primary NK cells exposed to HK1 by the releasate of platelets, which had been pretreated with the supernatant from F3-knockdown HK1-EBV or F3-overexpressing HK1 cells. Means ± SD; n = 5. E and F, NK-cell–mediated cytotoxicity towards target cells (K562 and HK1) in the presence of the releasate of platelets, which had been pretreated with the supernatant from HK1 and HK1-EBV cells. Means ± SD; n = 5. G and H, NK-cell–mediated cytotoxicity towards target cells (K562 and HK1) in the presence of the releasate of platelets, which had been pretreated with the supernatant from HK1-EBV cells coincubated with the F3 neutralizing antibody or IgG isotype control antibody. Means ± SD; n = 3. I and J, NK-cell–mediated cytotoxicity towards target cells (K562 and HK1) in the presence of the releasate of platelets, which had been pretreated with F3 protein or control. Means ± SD; n = 3. Two-tailed unpaired Student t test; *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Close modal

Consistent with the results of the degranulation assay, the releasate of platelets pretreated with the supernatant from EBV-positive cells (HK1-EBV or AGS-EBV) inhibited NK-cell–mediated cytotoxicity toward target cells (K562, HK1, or AGS) more effectively than the supernatant from EBV-negative cells (HK1 or AGS; Fig. 3E and F; Supplementary Fig. S5A). Furthermore, the inhibition of NK-cell–mediated cytotoxicity toward target cells (K562, HK1, or AGS) was increased by the releasate of platelets pretreated with the supernatant from EBV-positive cells (HK1-EBV or AGS-EBV), which co-incubated with the F3 neutralizing antibody (Fig. 3G and H; Supplementary Figs. S5B and S5C). In contrast, the inhibition of NK-cell–mediated cytotoxicity toward target cells (K562, HK1, or AGS) was decreased by the releasate of platelets, which were pretreated with the purified F3 protein (Fig. 3I and J; Supplementary Fig, S5D). These data suggested that F3 secreted by EBV-related epithelial tumor cells is an essential factor that promotes blood coagulation, which resists the killing effect of NK cells.

TGFβ, a cytokine inhibiting platelet release, decreases NK-cell function (12). Indeed, we found that the level of TGFβ released by platelets was reduced after treatment with the supernatant from F3-knockdown EBV-positive HK1 cells compared with mock-knockdown EBV-positive HK1 cells (Supplementary Fig. S5E). Moreover, TGFβ release from platelets was increased after treatment with the supernatant from F3-overexpressing HK1 cells compared with control HK1 cells (Supplementary Fig. S5E). The TGFβ signaling pathway was enriched in NK cells after coculture with the supernatant of F3-overexpressing HK1 cells, and upregulated genes were identified (Supplementary Figs. S5F and S5G), confirming that platelet activation and TGFβ mediated the resistance of EBV to the antitumor function of NK cells in EBV-associated epithelial tumors. Taken together, these in vitro data demonstrated that upregulation F3 expression in EBV-positive cancer cells induces resistance to NK-cell functions through platelet aggregation and TGFβ release.

LMP2A is involved in F3 upregulation through the PI3K/AKT signaling pathway

We next explored the mechanism by which EBV regulates F3 in EBV-positive epithelial cancer cells. We first confirmed that EBV infection increased F3 expression and secretion in two NPC cell lines, CNE2 and HK1, and in the gastric carcinoma cell line AGS (Fig. 4AC). We next tried to identify the viral genes that contribute to the upregulation of F3 after EBV infection. Previous studies have shown that the latent viral oncogenes LMP1 and LMP2A are expressed in EBV-infected NPC cell lines (23), whereas EBVaGC tumor cell lines mainly express the latent protein LMP2A instead of LMP1 (24). Therefore, we sought to determine whether LMP2A is involved in the regulation of F3 expression. Overexpression of LMP2A in NPC and GC cell lines (CNE2, HK1, and AGS) increased F3 expression as well as F3 secretion into the supernatant (Fig. 4D), whereas knockdown of LMP2A in EBV-positive NPC and GC cell lines (CNE2-EBV, HK1-EBV, and AGS-EBV) decreased F3 secretion (Fig. 4E). To verify the effect of LMP2A on F3 in vivo, we injected 5 × 106 CNE2-LMP2A and CNE2-vector cells into nude mice. IHC assays confirmed the upregulation of F3 in xenografts overexpressing LMP2A (Fig. 4F). In CNE2, HK1, and AGS cells, overexpression of LMP2A increased the levels of phosphorylated AKT (pAKT) and phosphorylated mTOR (p-mTOR), accompanied by the upregulation of F3 (Fig. 4G). After blockade of the PI3K/AKT pathway with PI3K inhibitors (LY294200 and Wortmannin), the levels of pAKT and p-mTOR, as well as the expression of F3, were greatly decreased (Fig. 4G). Consistent with the results of previous studies (16, 25), our results show that EBV infection also activates the PI3K/AKT/mTOR pathway in NPC-positive and GC EBV-positive cell lines, in addition to inducing F3 upregulation (Fig. 4H). When the expression of LMP2A was knocked down in EBV-positive NPC and GC cells using sgRNAs, the phosphorylation of AKT and mTOR and the expression of F3 were reduced to levels comparable with those in EBV-negative cancer cells (Fig. 4H). These data indicate that LMP2A participates in the regulation of F3 expression through the PI3K/AKT/mTOR axis after EBV infection.

Figure 4.

EBV upregulates F3 expression through the LMP2A/PI3K/AKT signaling pathway. A, The expression of EBERs in the normal nasopharyngeal epithelial cell line NP69 and the EBV−/+ NPC cell lines was evaluated by in situ hybridization. White scale bars, 20 μm. B and C, Analysis of F3 in NP69 and EBV−/+ NPC/GC cell lines by Western blot and ELISA. Means ± SD; n = 3; two-tailed unpaired Student t test; ***, P < 0.001; ****, P < 0.0001. D, ELISA of the F3 content in the supernatant of vector- or LMP2A-overexpressing cells. Means ± SD; n = 3; two-tailed unpaired Student t test; ***, P < 0.001; ****, P < 0.0001. E, ELISA of the F3 content in the supernatant of vector- or LMP2A-knockdown cells. Means ± SD; n = 3; two-tailed unpaired Student t test; **, P < 0.01; ***, P < 0.001. F, Left, images of F3 staining in NPC xenograft tumors transfected with empty vector or LMP2A, as assessed by IHC. Right, quantification of F3 expression in NPC xenograft tumors. Means ± SD; two-tailed Mann–Whitney U test; ***, P < 0.001. Black scale bars, 50 μm. G, Immunoblots comparing levels of downstream PI3K/AKT signaling pathway components and F3 in empty vector- and LMP2A-overexpressing NPC (CNE2 and HK1) and GC (AGS) cells. Immunoblots of pAKT, p-mTOR, and F3 after treatment with the PI3K kinase inhibitor LY294002 (50 μmol/L) or wortmannin (1 μmol/L) for 6 hours. H, Immunoblots comparing levels of downstream PI3K/AKT signaling pathway components and F3 in EBV−/+ and LMP2A-knockdown NPC (CNE2 and HK1) and GC (AGS) cells.

Figure 4.

EBV upregulates F3 expression through the LMP2A/PI3K/AKT signaling pathway. A, The expression of EBERs in the normal nasopharyngeal epithelial cell line NP69 and the EBV−/+ NPC cell lines was evaluated by in situ hybridization. White scale bars, 20 μm. B and C, Analysis of F3 in NP69 and EBV−/+ NPC/GC cell lines by Western blot and ELISA. Means ± SD; n = 3; two-tailed unpaired Student t test; ***, P < 0.001; ****, P < 0.0001. D, ELISA of the F3 content in the supernatant of vector- or LMP2A-overexpressing cells. Means ± SD; n = 3; two-tailed unpaired Student t test; ***, P < 0.001; ****, P < 0.0001. E, ELISA of the F3 content in the supernatant of vector- or LMP2A-knockdown cells. Means ± SD; n = 3; two-tailed unpaired Student t test; **, P < 0.01; ***, P < 0.001. F, Left, images of F3 staining in NPC xenograft tumors transfected with empty vector or LMP2A, as assessed by IHC. Right, quantification of F3 expression in NPC xenograft tumors. Means ± SD; two-tailed Mann–Whitney U test; ***, P < 0.001. Black scale bars, 50 μm. G, Immunoblots comparing levels of downstream PI3K/AKT signaling pathway components and F3 in empty vector- and LMP2A-overexpressing NPC (CNE2 and HK1) and GC (AGS) cells. Immunoblots of pAKT, p-mTOR, and F3 after treatment with the PI3K kinase inhibitor LY294002 (50 μmol/L) or wortmannin (1 μmol/L) for 6 hours. H, Immunoblots comparing levels of downstream PI3K/AKT signaling pathway components and F3 in EBV−/+ and LMP2A-knockdown NPC (CNE2 and HK1) and GC (AGS) cells.

Close modal

Inhibition of F3 restores the antitumor function of NK cells and shows therapeutic potential in an NCG mouse model

Mouse platelets are commonly used as surrogates to study in vivo human platelet function (26). Although mouse platelets differ from human platelets with regard to size, number, and structure, they have high similarity with respect to function (27). Therefore, we inoculated NCG mice (NK-cell-deficient) with EBV-positive HK1 cells with or without F3 knockdown to explore the potential of inhibiting F3 on the restoration of human NK-cell function against EBV-positive tumors. Primary human NK cells were injected intravenously. Four days after tumor cell inoculation, mice were injected with IL15 every day for 1 week, and with IL2 every other day for 21 days to support the survival of the human NK cells (Fig. 5A). HK1–EBV xenograft tumor growth was not altered significantly in the presence of human NK cells, suggesting the dysfunction of NK cells in the EBV-positive TME. When F3 was knocked down in tumor cells by sgRNA, the growth of HK1-EBV tumors was significantly slower than that of control tumors without F3 knockdown, indicating that F3 has the ability to promote tumor growth. Interestingly, reinfused NK cells exerted an antitumor effect when F3 was knocked down in cancer cells (Fig. 5BD). The growth of HK1–EBV xenografts in the presence of the combination of reinfused NK cells and F3 knockout in tumor cells was significantly slower than that in the presence of either NK-cell reinfusion or F3 knockout alone, indicating that F3 inhibition sensitizes tumors to reinfused NK cell. In addition, platelet aggregation was significantly reduced in F3-knockdown HK1–EBV xenografts from NCG mice (Fig. 5E). Infiltration of NK cells in xenograft tumors were further analyzed by assessing the expression of NK-cell markers using flow cytometry. The percentage of CD56+ cells was comparable between the control and F3-knockdown groups, indicating that blocking F3 only exerted a weak effect on the infiltration of NK cells (Fig. 5F). Notably, F3 knockout in HK1–EBV cancer cells increased the expression of the activation marker NKG2D and CD69 in NK cells in the TME, restoring the antitumor function of NK cells in tumors (Fig. 5G and H).

Figure 5.

Knockdown of F3 in EBV-positive NPC tumor cells restores the antitumor function of NK cells in an NCG mouse model. A, Schematic of in vivo studies using HK1-EBV-vector or knockdown F3 cells in a mouse xenograft model treated with primary NK cells and cytokine administration. B, Growth curves of HK1-EBV-vector and HK1-EBV-sgF3 xenograft tumors with or without infusion of primary NK cells. Means ± SD; n = 4; two-tailed unpaired Student t test; **, P < 0.01; *, P < 0.05. C and D, Images and weights of NPC xenograft tumors from NCG mice. Means ± SD; n = 4; two-tailed unpaired Student t test; *, P < 0.05; **, P < 0.01; NS, not significant. E, Images (left) and quantification (right) of CD41 staining in NPC xenograft tumors as assessed by IHC. Means ± SD; two-tailed Mann–Whitney U test; ****, P < 0.0001. White scale bars, 50 μm. F, Analysis of the proportion of infiltrated NK cells (CD56+) in xenograft tumors. Means ± SD; n = 4; two-tailed unpaired Student t test. G and H, Analysis of the mean fluorescence intensity (MFI) of activated NK cells (NKG2D+ or CD69+) in xenograft tumors. Means ± SD; n = 4; two-tailed unpaired t test; **, P < 0.01.

Figure 5.

Knockdown of F3 in EBV-positive NPC tumor cells restores the antitumor function of NK cells in an NCG mouse model. A, Schematic of in vivo studies using HK1-EBV-vector or knockdown F3 cells in a mouse xenograft model treated with primary NK cells and cytokine administration. B, Growth curves of HK1-EBV-vector and HK1-EBV-sgF3 xenograft tumors with or without infusion of primary NK cells. Means ± SD; n = 4; two-tailed unpaired Student t test; **, P < 0.01; *, P < 0.05. C and D, Images and weights of NPC xenograft tumors from NCG mice. Means ± SD; n = 4; two-tailed unpaired Student t test; *, P < 0.05; **, P < 0.01; NS, not significant. E, Images (left) and quantification (right) of CD41 staining in NPC xenograft tumors as assessed by IHC. Means ± SD; two-tailed Mann–Whitney U test; ****, P < 0.0001. White scale bars, 50 μm. F, Analysis of the proportion of infiltrated NK cells (CD56+) in xenograft tumors. Means ± SD; n = 4; two-tailed unpaired Student t test. G and H, Analysis of the mean fluorescence intensity (MFI) of activated NK cells (NKG2D+ or CD69+) in xenograft tumors. Means ± SD; n = 4; two-tailed unpaired t test; **, P < 0.01.

Close modal

We next investigated the therapeutic potential of blocking F3 to restrict the growth of EBV-positive tumors. Tissue factor pathway inhibitor (TFPI), the natural inhibitory protein of F3 in vivo, inhibits coagulation by blocking the proteolytic activity of both the tissue factor (F3)/factor VIIa (fVIIa) complex and FXa (28). NCG mice were inoculated with HK1–EBV NPC cells; human primary NK cells were injected 8 days later along with IL15 and IL2 supplementation to support the survival of human NK cells. Peritumoral treatment with TFPI (0.66 mg/kg) was conducted to block the F3 pathway (Fig. 6A). TFPI treatment alone inhibited tumor growth (Fig. 6B), consistent with the results of F3 knockdown in EBV-positive tumor cells (Fig. 5B and C). Although the NK-cell infusion exerted a weak and nonsignificant inhibitory effect on tumor growth, the combination of TFPI treatment and NK-cell infusion resulted in stronger inhibition of tumor growth than either TFPI treatment or NK-cell infiltration alone (Fig. 6BD). Compared with the corresponding effects on the group without TFPI treatment, platelet aggregation was reduced (Fig. 6E) and NK-cell activity (NKG2D and CD69) was partially restored in xenograft tumors treated with TFPI, indicating that TFPI restored the function of NK cells by inhibiting coagulation (Fig. 6F and G). These data indicate that TFPI treatment facilitates the antitumor function of NK cells in the EBV-associated TME, suggesting that it sensitizes tumors to NK-cell reinfusion. Collectively, the results of F3 knockdown and TFPI treatment in the NCG mouse model reveal that inhibition of F3 restores NK-cell–mediated cytotoxicity and shows antitumor therapeutic potential (Fig. 7).

Figure 6.

F3 targeting combined with NK-cell infusion shows therapeutic potential in an NCG mouse model. A, Schematic of in vivo studies using the F3 inhibitor TFPI or control in a mouse xenograft model treated with or without infusion of primary NK cells and cytokine administration. B, Growth curves of HK1–EBV NPC xenograft tumors treated with control or TFPI alone or in combination with primary NK cells. Means ± SD; n = 5; two-tailed unpaired Student t test; ***, P < 0.001; NS, not significant. C and D, Images and weights of NPC xenograft tumors from NCG mice. *, P < 0.05; **, P < 0.01. E, Images (left) and quantification (right) of CD41 staining in NPC xenograft tumors as assessed by IHC. Means ± SD; two-tailed Mann–Whitney U test; **, P < 0.01. White scale bars, 50 μm. F and G, Analysis of the mean fluorescence intensity (MFI) of activated NK cells (NKG2D+ or CD69+) in xenograft tumors. Means ± SD; n = 5; two-tailed unpaired Student t test; **, P < 0.01.

Figure 6.

F3 targeting combined with NK-cell infusion shows therapeutic potential in an NCG mouse model. A, Schematic of in vivo studies using the F3 inhibitor TFPI or control in a mouse xenograft model treated with or without infusion of primary NK cells and cytokine administration. B, Growth curves of HK1–EBV NPC xenograft tumors treated with control or TFPI alone or in combination with primary NK cells. Means ± SD; n = 5; two-tailed unpaired Student t test; ***, P < 0.001; NS, not significant. C and D, Images and weights of NPC xenograft tumors from NCG mice. *, P < 0.05; **, P < 0.01. E, Images (left) and quantification (right) of CD41 staining in NPC xenograft tumors as assessed by IHC. Means ± SD; two-tailed Mann–Whitney U test; **, P < 0.01. White scale bars, 50 μm. F and G, Analysis of the mean fluorescence intensity (MFI) of activated NK cells (NKG2D+ or CD69+) in xenograft tumors. Means ± SD; n = 5; two-tailed unpaired Student t test; **, P < 0.01.

Close modal
Figure 7.

Mechanistic diagram of the F3 targeting strategy to restore the antitumor function of NK cells. The blue line indicates that EBV upregulates F3 to promote blood coagulation and resist the antitumor function of NK cells. The brown line indicates that the F3 inhibitor resists the activation of platelets by EBV or upregulation F3 and restores the tumor killing function of NK cells.

Figure 7.

Mechanistic diagram of the F3 targeting strategy to restore the antitumor function of NK cells. The blue line indicates that EBV upregulates F3 to promote blood coagulation and resist the antitumor function of NK cells. The brown line indicates that the F3 inhibitor resists the activation of platelets by EBV or upregulation F3 and restores the tumor killing function of NK cells.

Close modal

EBV is one of the most common human viruses and is causally associated with a variety of lymphomas and epithelial carcinomas. Numerous viral gene products, including EBV-encoded proteins (BGLF5, BNLF2a, gp42, and LF2) and small RNAs (miR-BHRF1–3 and miR-BART2–5p), interfere with both adaptive and innate immunity to confer immune evasion properties to its host cells (29–34). However, the mechanisms underlying EBV-driven evasion of NK-cell immune surveillance in EBV-associated epithelial cancers remain elusive. In this study, we revealed a novel mechanism by which EBV-associated epithelial cancer cells escape NK-cell–mediated killing. We found that EBV infection inhibited NK-cell cytotoxicity by upregulating F3 in NPC and GC cells via the LMP2A/PI3K/AKT/mTOR axis. Importantly, an F3 pathway inhibitor (TFPI) targeting F3-mediated signaling activities and coagulation showed the potential to restore NK-cell–mediated cytotoxicity and treat EBV-positive NPC in the established NCG mouse model, providing a novel therapeutic strategy for EBV-associated epithelial tumors (Fig. 7).

Both NPC and EBVaGC tumors are infiltrated by a large number of lymphocytes, and a series of clinical trials of PD-1/PD-L1 immune checkpoint inhibitors have shown clinical benefits in patients with NPC and EBVaGC. However, the objective response rate of patients to PD-1/PD-L1 immune checkpoint inhibitors was only 20% to 35% (35, 36), indicating the need to identify novel immunotherapeutic targets other than cytotoxic T cells. NK cells are another promising type of effector lymphocyte in the TME that can directly recognize and kill tumor cells via the interaction of NK receptors with target cells, independent of antigen processing and presentation.

However, EBV has evolved many strategies to evade NK-cell–mediated killing of tumor cells. A unique global upregulation pattern of interferon responses in major immune cell types in the TME, including NK cells, has been reported for NPC, highlighting the role of chronic EBV infection in shaping the TME (37). The expression of the NKG2D ligand MICB is downregulated by LMP2A or EBV-encoded miR-BART-2–5p, diminishing the elimination of infected cells by NK-cell activity (34). In addition, a viral IL10 (vIL10) homolog encoded by the BCRF1 gene interferes with NK-cell–mediated killing of virus-infected cells (38). In addition to these mechanisms, the metabolites from EBV-infected cells inhibit NKG2D expression via the c-JNK pathway (39). Unfortunately, there are no therapies that to improve the infiltration and functions of NK cells. In this work, we revealed that EBV infection in epithelial carcinoma cells induces resistance to NK-cell functions in vitro and in vivo (Fig. 1). Mechanistically, upregulation of F3 and activation of platelet aggregation play an important role in EBV infection-induced resistance to NK functions and provide a therapeutic target for EBV-associated epithelial carcinomas (Figs. 13).

F3, an initiator of the extrinsic coagulation cascade, is involved in various hallmark events in tumor progression, such as tumor cell proliferation, apoptosis resistance, tumor angiogenesis, and metastasis. In addition, the coagulation process initiated by F3 recruits MDSCs to modulate immune responses in the TME (40). In this study, F3 expression was negatively correlated with NK-cell activation, but not the infiltration of NK cells in both nude mice and NCG mice with NK-cell reinfusion (Figs. 2 and 5). As expected, knockdown of F3 or treatment with its inhibitor, TFPI, slowed the growth of tumors in NCG mice with NK-cell reinfusion (Figs. 5 and 6). Different F3 isoforms have been indicated to activate different signaling pathways modulating various features of cancer cells. The flF3/FVIIa complex modifies multiple signaling molecules in cancers in a manner mediated via PAR-2 (41), promoting cancer cell proliferation, migration, growth, and angiogenesis. In contrast, asF3 has been shown to bind β3 and β1 integrins, inducing angiogenesis and migration independent of PAR-2 signaling (41, 42). F3 has been reported to activate pAKT (43); therefore, we constructed a cell line overexpressing F3 and found that F3 also activated pAKT. Therefore, we believe that the EBV latent protein LMP2A is the driving factor for activating the AKT pathway, which in turn upregulates F3 expression. Upregulated F3 further activates pAKT, indicating that F3 participates in positive feedback regulation of pAKT to induce the continuous secretion of F3 into the microenvironment outside tumor cells, and participates in the immune escape of the tumor. Accordingly, in our NCG mouse model with NK-cell reinfusion, F3 knockdown in NPC cells directly affected tumor growth and restored NK-cell function by inhibiting coagulation.

Platelet activation, an important step in the coagulation pathway, impairs NK-cell–dependent tumor immune surveillance through several mechanisms. Coating of tumor cells by platelets decreases the expression of NKG2D ligands, especially MICA and MICB, which in turn diminishes the activation of NK cells (19). Malignant cells exhibit downregulated expression of MHC-I, the ligand for inhibitory NK receptors, and thus trigger NK-cell activation to recognize and lyse malignant cells. However, malignant cells can obtain MHC-I from activated platelets to escape “missing-self” recognition by NK cells (44). In addition, platelet-derived TGFβ downregulates the surface expression of the activating receptor NKG2D on NK cells and diminishes antitumor reactivity (12). In our study, we showed that platelet activation is a necessary step for F3 secretion from tumor cells to inhibit NK-cell function (Fig. 3; Supplementary Fig. S3). Our results indicated that platelets activated by the supernatant from EBV-positive and F3-overexpressing EBV-negative NPC and GC cell cultures released the inhibitory cytokine TGFβ (Supplementary Fig. S3). These results suggested that activated platelets and the release of the cytokine TGFβ resulting from upregulation of F3 in EBV-positive epithelial tumors inhibited the function of NK cells.

The transcription of F3 is induced by multiple stimulators, such as VEGF, TNFα, oncogenic mutations, and hypoxia. In our previous study, the PI3K/AKT pathway was shown to be the most significantly deregulated signaling pathway in EBV-positive NPC cells compared with EBV-negative NPC cells (16). The expression of the latent protein LMP2A in NPC and EBVaGC was found to contribute to AKT and mTOR activation. Together with NPC and EBVaGC, we also collected EBV-related lymphoma (six positive and seven negative) and lung cancer (nine positive and seven negative) samples. The results of F3 by IHC showed that EBV also upregulates F3 expression in lymphoma and lung cancer (Supplementary Figs. S6A and S6B). Among the four cancers we are currently studying, EBV upregulates F3 in lymphoma and epithelial tumors. Thus, the finding that EBV upregulating F3 expression is universal to some extent. The potential explanation is that LMP2A, the latent protein of EBV, is the initiating factor of F3 upregulation, and LMP2A is expressed in almost all lymphoma (45–47) and epithelial tumors (48–51). Our data indicated that LMP2A also upregulated F3 expression through the PI3K/AKT pathway in EBV-associated epithelial carcinomas; thus, secreted F3 played an important role in EBV-mediated remodeling of NK-cell function in the TME.

Aberrant expression of F3 has been shown to be common in malignancies and to mediate tumor-associated hypercoagulability and angiogenesis; thus, the F3 pathway is a potential therapeutic target to rebuild a functional TME. TFPI, which inhibits the coagulation and signaling activities of F3, has been reported to inhibit primary and metastatic tumor growth in mice (28). Because F3 is expressed in a variety of cells throughout the body and activates the coagulation pathway under normal physiological conditions, we injected TFPI peritumorally to inhibit F3-mediated coagulation in tumor xenografts. As expected, TFPI counteracted the inhibitory effect of F3 on NK cells and restored the antitumor activity of NK cells in EBV-positive NPC xenografts in NCG mice. In addition, TFPI inhibited NPC tumor growth without reinfusion of NK cells, possibly due to the inhibitory effects of TFPI on angiogenesis and on F3-activated proliferative signal transmission to tumor cells.

In summary, our study revealed a novel molecular mechanism by which EBV inhibits NK-cell function in EBV-associated epithelial tumors. Upregulation of F3 through the LMP2A/PI3K/AKT signaling pathway and platelet activation induced by F3 promoted the release of TGFβ to induce resistance to the antitumor function of NK cells. Furthermore, TFPI treatment sensitized tumors to NK cell to suppress tumor growth in an EBV-positive NPC xenograft model. An important immunosuppressive factor in the TME, F3 provides a potential therapeutic target to reinvigorate NK-cell function for the treatment of EBV-associated epithelial tumors.

X. Duan reports a patent 2021101823148 pending. T. Xiang reports a patent 2021101823148 pending. No disclosures were reported by the other authors.

X. Duan: Resources, software, formal analysis, funding acquisition, methodology, writing–original draft, project administration, writing–review and editing. H. Chen: Software, formal analysis, methodology, writing–original draft, project administration, writing–review and editing. X. Zhou: Software, formal analysis, methodology, writing–original draft, writing–review and editing. P. Liu: Resources, software, methodology. X. Zhang: Resources, methodology. Q. Zhu: Resources, methodology. L. Zhong: Resources, methodology. W. Zhang: Methodology. S. Zhang: Methodology. X. Zhang: Methodology. Y. Chen: Software, methodology. Y. Zhou: Software, methodology. C. Yang: Resources, methodology. Q. Feng: Resources, project administration. Y.-X. Zeng: Funding acquisition, writing–original draft, project administration, writing–review and editing. M. Xu: Funding acquisition, writing–original draft, writing–review and editing. T. Xiang: Resources, software, funding acquisition, methodology, writing–original draft, writing–review and editing.

The authors thank Ruihua Xu and Musheng Zeng (Sun Yat-sen University Cancer Center) for providing cell lines. They thank Shandong Cord Blood Bank for kindly providing human umbilical cord blood NK cells. This study was supported by research grants from the National Natural Science Foundation of China (nos. 81872228, 81874129, 82172795, and 82103433), the Guangdong Basic and Applied Basic Research Foundation (nos. 2020B1515020002 and 2020A1515110013), and the Fundamental Research Funds for the Central Universities to M. Xu.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Young
LS
,
Yap
LF
,
Murray
PG
.
Epstein-Barr virus: more than 50 years old and still providing surprises
.
Nat Rev Cancer
2016
;
12
:
789
802
.
2.
Jha
HC
,
Pei
Y
,
Robertson
ES
.
Epstein-Barr virus: diseases linked to infection and transformation
.
Front Microbiol
2016
;
1602
.
3.
Hong
S
,
Liu
D
,
Luo
S
,
Fang
W
,
Zhan
J
,
Fu
S
, et al
.
The genomic landscape of Epstein-Barr virus-associated pulmonary lymphoepithelioma-like carcinoma
.
Nat Commun
2019
;
1
:
3108
.
4.
Münz
C
.
Latency and lytic replication in Epstein-Barr virus-associated oncogenesis
.
Nat Rev Microbiol
2019
;
11
:
691
700
.
5.
Tsao
SW
,
Tsang
CM
,
Lo
KW
.
Epstein-Barr virus infection and nasopharyngeal carcinoma
.
Philos Trans R Soc Lond B Biol Sci
2017
;
1732
:
20160270
.
6.
Chen
YP
,
Yin
JH
,
Li
WF
,
Li
HJ
,
Chen
DP
,
Zhang
CJ
, et al
.
Single-cell transcriptomics reveals regulators underlying immune cell diversity and immune subtypes associated with prognosis in nasopharyngeal carcinoma
.
Cell Res
2020
;
11
:
1024
42
.
7.
Tan
GW
,
Visser
L
,
Tan
LP
,
van den Berg
A
,
Diepstra
A
.
The microenvironment in Epstein-Barr virus-associated malignancies
.
Pathogens
2018
;
7
:
40
.
8.
Lutzky
VP
,
Crooks
P
,
Morrison
L
,
Stevens
N
,
Davis
JE
,
Corban
M
, et al
.
Cytotoxic T cell adoptive immunotherapy as a treatment for nasopharyngeal carcinoma
.
Clin Vaccine Immunol
2014
;
2
:
256
9
.
9.
Singh
S
,
Banerjee
S
.
Downregulation of HLA-ABC expression through promoter hypermethylation and downmodulation of MIC-A/B surface expression in LMP2A-positive epithelial carcinoma cell lines
.
Sci Rep
2020
;
1
:
5415
.
10.
Xu
Y
,
Zhou
R
,
Huang
C
,
Zhang
M
,
Li
J
,
Zong
J
, et al
.
Analysis of the expression of surface receptors on NK cells and NKG2D on immunocytes in peripheral blood of patients with nasopharyngeal carcinoma
.
Asian Pac J Cancer Prev
2018
;
3
:
661
5
.
11.
Schlesinger
M
.
Role of platelets and platelet receptors in cancer metastasis
.
J Hematol Oncol
2018
;
1
:
125
.
12.
Kopp
HG
,
Placke
T
,
Salih
HR
.
Platelet-derived transforming growth factor-beta down-regulates NKG2D thereby inhibiting natural killer cell antitumor reactivity
.
Cancer Res
2009
;
19
:
7775
83
.
13.
Placke
T
,
Örgel
M
,
Schaller
M
,
Jung
G
,
Rammensee
HG
,
Kopp
HG
, et al
.
Platelet-derived MHC class I confers a pseudonormal phenotype to cancer cells that subverts the antitumor reactivity of natural killer immune cells
.
Cancer Res
2012
;
2
:
440
8
.
14.
Chen
WH
,
Tang
LQ
,
Wang
FW
,
Li
CP
,
Tian
XP
,
Huang
XX
, et al
.
Elevated levels of plasma D-dimer predict a worse outcome in patients with nasopharyngeal carcinoma
.
BMC Cancer
2014
;
14
:
583
.
15.
Hu
C
,
Chen
R
,
Chen
W
,
Pang
W
,
Xue
X
,
Zhu
G
, et al
.
Thrombocytosis is a significant indictor of hypercoagulability, prognosis and recurrence in gastric cancer
.
Exp Ther Med
2014
;
1
:
125
32
.
16.
Xiang
T
,
Lin
YX
,
Ma
W
,
Zhang
HJ
,
Chen
KM
,
He
GP
, et al
.
Vasculogenic mimicry formation in EBV-associated epithelial malignancies
.
Nat Commun
2018
;
1
:
5009
.
17.
Stack
EC
,
Wang
C
,
Roman
KA
,
Hoyt
CC
.
Multiplexed immunohistochemistry, imaging, and quantitation: a review, with an assessment of Tyramide signal amplification, multispectral imaging and multiplex analysis
.
Methods
2014
;
1
:
46
58
.
18.
Lorenzo-Herrero
S
,
Sordo-Bahamonde
C
,
Gonzalez
S
,
López-Soto
A
.
CD107a degranulation assay to evaluate immune cell antitumor activity
.
Methods Mol Biol
2019
;
119
30
.
19.
Maurer
S
,
Kropp
KN
,
Klein
G
,
Steinle
A
,
Haen
SP
,
Walz
JS
, et al
.
Platelet-mediated shedding of NKG2D ligands impairs NK cell immune-surveillance of tumor cells
.
Oncoimmunology
2018
;
2
:
e1364827
.
20.
Wang
JG
,
Geddings
JE
,
Aleman
MM
,
Cardenas
JC
,
Chantrathammachart
P
,
Williams
JC
, et al
.
Tumor-derived tissue factor activates coagulation and enhances thrombosis in a mouse xenograft model of human pancreatic cancer
.
Blood
2012
;
23
:
5543
52
.
21.
Heemskerk
JW
,
Bevers
EM
,
Lindhout
T
.
Platelet activation and blood coagulation
.
Thromb Haemost
2002
;
2
:
186
93
.
22.
Goh
W
,
Huntington
ND
.
Regulation of murine natural killer cell development
.
Front Immunol
2017
;
8
:
130
.
23.
Cheung
ST
,
Huang
DP
,
Hui
AB
,
Lo
KW
,
Ko
CW
,
Tsang
YS
, et al
.
Nasopharyngeal carcinoma cell line (C666-1) consistently harbouring Epstein-Barr virus
.
Int J Cancer
1999
;
1
:
121
6
.
24.
Young
LS
,
Dawson
CW
,
Eliopoulos
AG
.
The expression and function of Epstein-Barr virus encoded latent genes
.
Mol Pathol
2000
;
5
:
238
47
.
25.
Portis
T
,
Longnecker
R
.
Epstein-Barr virus (EBV) LMP2A mediates B-lymphocyte survival through constitutive activation of the Ras/PI3K/Akt pathway
.
Oncogene
2004
;
53
:
8619
28
.
26.
Rowley
JW
,
Oler
AJ
,
Tolley
ND
,
Hunter
BN
,
Low
EN
,
Nix
DA
, et al
.
Genome-wide RNA-seq analysis of human and mouse platelet transcriptomes
.
Blood
2011
;
14
:
e101
11
.
27.
Jirouskova
M
,
Shet
AS
,
Johnson
GJ
.
A guide to murine platelet structure, function, assays, and genetic alterations
.
J Thromb Haemost
2007
;
4
:
661
9
.
28.
Williams
L
,
Tucker
TA
,
Koenig
K
,
Allen
T
,
Rao
LV
,
Pendurthi
U
, et al
.
Tissue factor pathway inhibitor attenuates the progression of malignant pleural mesothelioma in nude mice
.
Am J Respir Cell Mol Biol
2012
;
2
:
173
9
.
29.
Rowe
M
,
Glaunsinger
B
,
van Leeuwen
D
,
Zuo
J
,
Sweetman
D
,
Ganem
D
, et al
.
Host shutoff during productive Epstein-Barr virus infection is mediated by BGLF5 and may contribute to immune evasion
.
Proc Natl Acad Sci U S A
2007
;
9
:
3366
71
.
30.
Hislop
AD
,
Ressing
ME
,
van Leeuwen
D
,
Pudney
VA
,
Horst
D
,
Koppers-Lalic
D
, et al
.
A CD8+ T cell immune evasion protein specific to Epstein-Barr virus and its close relatives in Old World primates
.
J Exp Med
2007
;
8
:
1863
73
.
31.
Ressing
ME
,
van Leeuwen
D
,
Verreck
FA
,
Gomez
R
,
Heemskerk
B
,
Toebes
M
, et al
.
Interference with T cell receptor-HLA-DR interactions by Epstein-Barr virus gp42 results in reduced T helper cell recognition
.
Proc Natl Acad Sci U S A
2003
;
20
:
11583
8
.
32.
Wu
L
,
Fossum
E
,
Joo
CH
,
Inn
KS
,
Shin
YC
,
Johannsen
E
, et al
.
Epstein-Barr virus LF2: an antagonist to type I interferon
.
J Virol
2009
;
2
:
1140
6
.
33.
Xia
T
,
O'Hara
A
,
Araujo
I
,
Barreto
J
,
Carvalho
E
,
Sapucaia
JB
, et al
.
EBV microRNAs in primary lymphomas and targeting of CXCL-11 by ebv-mir-BHRF1–3
.
Cancer Res
2008
;
5
:
1436
42
.
34.
Nachmani
D
,
Stern-Ginossar
N
,
Sarid
R
,
Mandelboim
O
.
Diverse herpesvirus microRNAs target the stress-induced immune ligand MICB to escape recognition by natural killer cells
.
Cell Host Microbe
2009
;
4
:
376
85
.
35.
Ma
BBY
,
Lim
WT
,
Goh
BC
,
Hui
EP
,
Lo
KW
,
Pettinger
A
, et al
.
Antitumor activity of nivolumab in recurrent and metastatic nasopharyngeal carcinoma: an international, multicenter study of the mayo clinic phase 2 consortium (NCI-9742)
.
J Clin Oncol
2018
;
14
:
1412
8
.
36.
Fang
W
,
Yang
Y
,
Ma
Y
,
Hong
S
,
Lin
L
,
He
X
, et al
.
Camrelizumab (SHR-1210) alone or in combination with gemcitabine plus cisplatin for nasopharyngeal carcinoma: results from two single-arm, phase 1 trials
.
Lancet Oncol
2018
;
10
:
1338
50
.
37.
Jin
S
,
Li
R
,
Chen
MY
,
Yu
C
,
Tang
LQ
,
Liu
YM
, et al
.
Single-cell transcriptomic analysis defines the interplay between tumor cells, viral infection, and the microenvironment in nasopharyngeal carcinoma
.
Cell Res
2020
;
11
:
950
65
.
38.
Jochum
S
,
Moosmann
A
,
Lang
S
,
Hammerschmidt
W
,
Zeidler
R
.
The EBV immunoevasins vIL-10 and BNLF2a protect newly infected B cells from immune recognition and elimination
.
PLoS Pathog
2012
;
5
:
e1002704
.
39.
Song
H
,
Park
H
,
Kim
J
,
Park
G
,
Kim
YS
,
Kim
SM
, et al
.
IDO metabolite produced by EBV-transformed B cells inhibits surface expression of NKG2D in NK cells via the c-Jun N-terminal kinase (JNK) pathway
.
Immunol Lett
2011
;
2
:
187
93
.
40.
Markiewski
MM
,
DeAngelis
RA
,
Benencia
F
,
Ricklin-Lichtsteiner
SK
,
Koutoulaki
A
,
Gerard
C
, et al
.
Modulation of the antitumor immune response by complement
.
Nat Immunol
2008
;
11
:
1225
35
.
41.
Versteeg
HH
,
Schaffner
F
,
Kerver
M
,
Ellies
LG
,
Andrade-Gordon
P
,
Mueller
BM
, et al
.
Protease-activated receptor (PAR) 2, but not PAR1, signaling promotes the development of mammary adenocarcinoma in polyoma middle T mice
.
Cancer Res
2008
;
17
:
7219
27
.
42.
van den Berg
YW
,
van den Hengel
LG
,
Myers
HR
,
Ayachi
O
,
Jordanova
E
,
Ruf
W
, et al
.
Alternatively spliced tissue factor induces angiogenesis through integrin ligation
.
Proc Natl Acad Sci U S A
2009
;
46
:
19497
502
.
43.
Jiang
X
,
Guo
YL
,
Bromberg
ME
.
Formation of tissue factor-factor VIIa-factor Xa complex prevents apoptosis in human breast cancer cells
.
Thromb Haemost
2006
;
2
:
196
201
.
44.
Placke
T
,
Kopp
HG
,
Salih
HR
.
The wolf in sheep's clothing: Platelet-derived “pseudo self” impairs cancer cell “missing self” recognition by NK cells
.
Oncoimmunology
2012
;
4
:
557
9
.
45.
Chen
CL
,
Sadler
RH
,
Walling
DM
,
Su
IJ
,
Hsieh
HC
,
Raab-Traub
N
.
Epstein-Barr virus (EBV) gene expression in EBV-positive peripheral T-cell lymphomas
.
J Virol
1993
;
10
:
6303
8
.
46.
Kimura
H
,
Hoshino
Y
,
Hara
S
,
Sugaya
N
,
Kawada
J
,
Shibata
Y
, et al
.
Differences between T cell-type and natural killer cell-type chronic active Epstein-Barr virus infection
.
J Infect Dis
2005
;
4
:
531
9
.
47.
Ito
Y
,
Kawamura
Y
,
Iwata
S
,
Kawada
J
,
Yoshikawa
T
,
Kimura
H
.
Demonstration of type II latency in T lymphocytes of Epstein-Barr Virus-associated hemophagocytic lymphohistiocytosis
.
Pediatr Blood Cancer
2013
;
2
:
326
8
.
48.
Imai
S
,
Koizumi
S
,
Sugiura
M
,
Tokunaga
M
,
Uemura
Y
,
Yamamoto
N
, et al
.
Gastric carcinoma: monoclonal epithelial malignant cells expressing Epstein-Barr virus latent infection protein
.
Proc Natl Acad Sci U S A
1994
;
19
:
9131
5
.
49.
Sugiura
M
,
Imai
S
,
Tokunaga
M
,
Koizumi
S
,
Uchizawa
M
,
Okamoto
K
, et al
.
Transcriptional analysis of Epstein-Barr virus gene expression in EBV-positive gastric carcinoma: unique viral latency in the tumour cells
.
Br J Cancer
1996
;
4
:
625
31
.
50.
Heussinger
N
,
Büttner
M
,
Ott
G
,
Brachtel
E
,
Pilch
BZ
,
Kremmer
E
, et al
.
Expression of the Epstein-Barr virus (EBV)-encoded latent membrane protein 2A (LMP2A) in EBV-associated nasopharyngeal carcinoma
.
J Pathol
2004
;
2
:
696
9
.
51.
Kong
QL
,
Hu
LJ
,
Cao
JY
,
Huang
YJ
,
Xu
LH
,
Liang
Y
, et al
.
Epstein-Barr virus-encoded LMP2A induces an epithelial-mesenchymal transition and increases the number of side population stem-like cancer cells in nasopharyngeal carcinoma
.
PLoS Pathog
2010
;
6
:
e1000940
.

Supplementary data