STIM1-regulated exosomal EBV-LMP1 empowers endothelial cells with an aggressive phenotype by activating the Akt/ERK pathway in nasopharyngeal carcinoma

Stromal interaction molecule 1 (STIM1)-mediated Ca2+ signaling regulates tumor angiogenesis in nasopharyngeal carcinoma (NPC), an Epstein-Barr virus (EBV)-related human malignancy. However, the mechanism by which STIM1 modulates endothelial functional phenotypes contributing to tumor angiogenesis remains elusive. NPC cell-derived exosomes were isolated via differential centrifugation and observed using transmission electron microscopy. Exosome particle sizes were assessed by nanoparticle tracking analysis (NTA). Uptake of exosomes by recipient ECs was detected by fluorescent labeling of the exosomes with PKH26. Tumor angiogenesis-associated profiles were characterized by determining cell proliferation, migration, tubulogenesis and permeability in human umbilical vein endothelial cells (HUVECs). Activation of the Akt/ERK pathway was assessed by detecting the phosphorylation levels using Western blotting. A chick embryo chorioallantoic membrane (CAM) xenograft model was employed to study tumor-associated neovascularization in vivo. We found that NPC cell-derived exosomes harboring EBV-encoded latent membrane protein 1 (LMP1) promoted proliferation, migration, tubulogenesis and permeability by activating the Akt/ERK pathway in ECs. STIM1 silencing reduced LMP1 enrichment in NPC cell-derived exosomes, thereby reversing its pro-oncogenic effects in an Akt/ERK pathway-dependent manner. Furthermore, STIM1 knockdown in NPC cells blunted tumor-induced vascular network formation and inhibited intra-tumor neovascularization in the chorioallantoic membrane (CAM) xenograft model. STIM1 regulates tumor angiogenesis by controlling exosomal EBV-LMP1 delivery to ECs in the NPC tumor microenvironment. Blocking exosome-mediated cell-to-cell horizontal transfer of EBV-associated oncogenic signaling molecules may be an effective therapeutic strategy for NPC.


Introduction
Nasopharyngeal carcinoma (NPC) is an Epstein-Barr virus (EBV)-related head and neck squamous cell carcinoma (HNSCC) with unusual geographical and ethnic distributions worldwide [1,2]. NPC is the most common HNSCC in southern China, with a particularly high incidence in the Cantonese-speaking population [3,4]. Multiple factors synergistically contribute to NPC pathogenesis, including genomic instability, genetic susceptibility and latent EBV infection [1,2]. The latter participates in NPC development by shaping malignant phenotypes and reprogramming cellular metabolism in the infected host epithelial cells. EBV infection also contributes to NPC progression by remodeling the tumor microenvironment (TME), which causes deregulated proliferation, regional invasiveness and, ultimately, distant dissemination of NPC cells [5][6][7][8][9].
Exosomes, a class of extracellular vesicles harboring a variety of bioactive molecules, serve as pivotal mediators of cell-cell communication in the TME [10,11]. The EBV-encoded latent membrane protein 1 (LMP1) can be sorted and loaded into exosomes released by EBV-infected cells, which are then uptaken by the surrounding non-EBVinfected cells, including endothelial cells (ECs) [12][13][14][15]. As yet, however, the effects of exosomal LMP1 on EC functional phenotypes that contribute to tumor angiogenesis remain poorly understood.
Previously, we reported that stromal interaction molecule 1 (STIM1)-dependent store-operated Ca 2+ entry (SOCE) mediates LMP1-promoted cell-cell interactions between NPC cells and neighboring ECs [16]. Subsequently, we found that EBV infection enhances tumor angiogenesis by boosting STIM1-mediated Ca 2+ signaling [17,18]. Amplified Ca 2+ influx increasing cytosolic Ca 2+ concentration triggers the docking and fusion of multivesicular bodies and stimulates exosome release [19][20][21]. Elevated STIM1mediated Ca 2+ entry enhances exosome release to the extracellular matrix [22]. Nevertheless, the regulatory role of STIM1 in exosome-mediated intercellular communication between NPC cells and ECs, as well as the effects of exosomal delivery of oncogenic signals on endothelial features, remain elusive.
In the present study, we show that NPC cell-derived exosomes carrying LMP1 promote proliferation, migration, tubulogenesis and permeability by stimulating the Akt/ERK pathway in ECs. Silencing STIM1 in EBV-infected NPC cells decreased LMP1 loading into exosomes, which further blunted the effects of exosomal LMP1 on EC characteristics responsible for tumor angiogenesis. Using a modified chick embryo chorioallantoic membrane (CAM) model, we found that silencing STIM1 inhibited xenograft-induced vascular network formation and intra-tumor neovascularization.

Exosome isolation, transmission electron microscopy and nanoparticle tracking analysis (NTA)
A total of 5 × 10 6 HK-EBV or C666-1 cells were seeded in a 150 mm dish. When the cells reached approximately 70% confluence, they were incubated in serum-free medium for 12 h, which was replaced with serum-free medium containing 50 ng/ml recombinant human epidermal growth factor (PeproTech, Cranbury, NJ, USA) for 36 h, after which the cell supernatants were harvested. Exosomes were isolated and collected via differential centrifugation at 300 × g for 10 min, 2,000 × g for 15 min, 10,000 × g for 30 min, and ultracentrifugation at 100,000 × g for 90 min. All centrifugations were performed at 4 °C. The exosomes were washed once with phosphate-buffered saline (PBS) and subjected to a second ultracentrifugation under the same conditions. The isolated exosomes were resuspended in PBS and stored at -80 °C before subsequent experiments. The morphological characteristics of the exosomes were evaluated using transmission electron microscopy (Hitachi HC-1, Japan) at an accelerating voltage of 80 kV. Exosome particle sizes were assessed via NTA using Particle Metrix ZetaView (Shanghai Biochip Co., Ltd., Shanghai, China) according to the manufacturer's instructions. A bicinchoninic acid protein assay kit (Beyotime Biotechnology, Shanghai, China) was used to quantify exosome protein content. For in vitro experiments, 10 µg/ml exosomes were incubated with recipient cells.

Exosome labeling and uptake by recipient cells
A PKH26 linker stock solution was diluted with Diluent C according to the manufacturer's instructions (Umibio Co., Ltd., Shanghai, China, Cat. No: UR52302) to prepare a working dye solution with a concentration of 100 µmol/L. The purified exosomes were incubated with the working dye solution for 10 min and then ultracentrifuged at 100,000 × g for 90 min to remove excess dye. After centrifugation, resuspended PKH26-labeled exosomes were incubated with recipient ECs for 24 h and evaluated using confocal microscopy (Molecular Devices, California, USA).

Cell proliferation assay
A Cell Counting Kit-8 (CCK-8) assay and 5-ethynyl-2′deoxyuridine (EdU) fluorescent staining were used to evaluate the proliferative capacity of HUVECs. A total of 2.0 × 10 3 HUVECs were seeded in each well of 96-well plates. After complete cell adhesion was achieved, the medium supplemented with FBS was replaced with medium containing 10 μg/ml exosomes or PBS (vehicle control), after which the cells were allowed to grow continually. Ten microliters CCK-8 reagent solution (Dalian Meilun Biotechnology Co., Ltd., Dalian, China, Cat. No: MA0218-5) was added to each well at 0, 24, 48, 72 and 96 h and incubated at 37 ℃ for 3 h. The absorbance of each well was measured at 450 nm wavelength using a microplate reader (Tristar LB941, Berthold Technologies, Germany). For EdU fluorescent staining, a total of 2.0 × 10 4 HUVECs were seeded in 24-well plates containing 10 μg/ml exosomes and 5% FBS. After 72 h of incubation, the medium was discarded, and medium containing 50 μmol/L EdU (Guangzhou RiboBio Co., Ltd., Guangzhou, China, Cat. No. C10310-1) was added to each well. After 4 h of incubation at 37 ℃, the cells were fixed with 4% paraformaldehyde, stained with Apollo and Hoechst, and photographed under a fluorescence microscope (IX2-ILL100, Olympus Corp., Tokyo, Japan). The number of EdU-positive cells was calculated as the percentage of total cells per random field.

In vitro migration assay
HUVEC migration ability was evaluated using scratch wound healing and transwell migration assays, as described before [16,25]. A total of 1.5 × 10 5 HUVECs were seeded into each well of a 24-well plate containing 5% FBS. When the cells reached 90% confluence, the monolayer was scratched using a sterile pipette tip to form a gap. Serum-free medium containing 10 μg/ml exosomes or PBS was added for incubation, and photographs of the same viewing field were obtained 48 h later. The percentages of the wound healing areas were calculated using ImageJ. To conduct the transwell migration assay, 3.0 × 10 5 HUVECs were inoculated into 6-well plates containing 10 μg/ml exosomes and 5% FBS for 72 h. Next, 1.0 × 10 5 HUVECs were resuspended in 200 μl serum-free medium and added to each 8 μm-pore size transwell chamber (Corning, NY, USA), which were placed in 24-well plates containing 20% FBS. After 48 h of incubation at 37 °C, the upper chambers were carefully removed, and the cells on the lower surface of the membranes were fixed with 4% paraformaldehyde and stained with 0.1% crystal violet. The cells were photographed under a microscope (IX2-ILL100, Olympus Corp.), and ImageJ software was used to count the number of membrane-crossing cells in each random field.

In vitro tube formation assay
Tube formation assays were performed to evaluate the EC tube-forming ability, as described befores [17,18]. Briefly, 3.0 × 10 5 HUVECs were seeded into each well of a 6-well plate containing 10 μg/ml exosomes and 5% FBS and incubated for 72 h. After 50 µl of Matrigel (Corning, NY, USA, Cat. No: 356234) was applied to each well of a 96-well plate and allowed to solidify, 100 μl of the cell suspension at a density of 4.0 × 10 5 /ml pretreated with exosomes was seeded into each well. After 18 h of incubation at 37 °C, endothelial tube formation was evaluated and photographed under a microscope (IX2-ILL100, Olympus Corp.). The number of tube rings in each random field was calculated to evaluate the tube-forming capacity of HUVECs pretreated with exosomes.

Fluorescein isothiocyanate (FITC)-labeled dextran permeability assay
The permeability of the endothelial layer formed by HUVECs treated with exosomes was determined as reported before [16]. A total of 5.0 × 10 4 HUVECs was resuspended in a medium containing 10 μg/ml exosomes, which were added into 0.4 μm-pore size upper chambers (Corning, NY, USA). The chambers were placed in 24-well plates containing 5% FBS. After 96 h of incubation, the ECs fused to form a monolayer, as confirmed via microscopic observation, and the medium was aspirated and discarded. Next, serumfree and phenol red-free medium containing FITC-dextran (Sigma-Aldrich Trading Co., Ltd., Shanghai, China, Cat. No. 46945) at 1 mg/ml was added to the upper chamber, while serum-and phenol red-free medium was added to each well of the plates. The medium in each bottom well was collected and transferred to a light-proof 96-well plate, and the absorbance of the medium at 488 nm was measured using a microplate reader at 0, 20, 40 and 60 min (Infinite 200 Pro, Tecan Trading AG, Switzerland).

STM1 knockdown via lentiviral vector transfection
The recombinant plasmid vector GV248 containing STIM1-RNA interference (shRNA-STIM1, GGGAA GAC CTC AAT TAC CA) or a nonsense scrambled sequence (shRNA-Ctrl., TTC TCC GAA CGT GTC ACG T) was purchased from GeneChem Technology (Shanghai, China). Lentivirus transfection was performed according to the manufacturer's instructions. Reduction in STIM1 expression was validated via Western blotting.

Western blotting
Total cell proteins or exosome total proteins were separated using 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to a polyvinylidene difluoride membranes. The membranes were blocked with rapid blocking solution for 15 min at room temperature and incubated with primary antibodies at 4 °C overnight. Next, the membranes were then incubated with horseradish peroxidase-conjugated secondary antibodies for 1 h at room temperature. The bands in the gel were visualized using enhanced chemiluminescence and photographed. The antibodies used in this study included anti-GAPDH (   Transplanted tumor angiogenesis and embryo viability were observed daily via in situ photography. After five days of inoculation, the membranes were peeled off with the xenografts in the center, and tumor angiogenesis was evaluated by calculating the ratio of the blood vessel area to the entire CAM area using the Image-Pro Plus 6 analysis system, as previously described [27]. CAM xenografts were fixed with 4% paraformaldehyde and embedded in paraffin, followed by hematoxylin and eosin (H&E) staining and immunohistochemistry (IHC). Cell nuclei were counterstained with DAPI, and images were acquired using a fluorescence microscope (NIKON ECLIPSE C1, Nikon, Japan).

Statistical analysis
SPSS24.0 (IBM, Chicago, IL, USA) and Graphpad prism 8.0 (GraphPad, San Diego, CA, USA) were used for statistical analyses. Data are presented as means ± SD. Student's t-test or two-way ANOVA was used for statistical analyses, and statistical significance was assumed at p < 0.05.

Exosomes released from EBV-positive NPC cells harbor LMP1
LMP1 can be loaded into exosomes and secreted by EBVinfected host cells into the TME [12]. To investigate the role of exosomal LMP1 in remodeling endothelial functional phenotypes contributing to tumor angiogenesis, we first isolated the exosomes released by two EBV-positive NPC cell lines, HK1-EBV and C666-1, via differential centrifugation from the cell-conditioned medium (Fig. 1a). NTA indicated that the exosome diameter was mainly distributed in the range of 100-200 nm (Fig. 1b). Transmission electron microscopy revealed that the isolated exosomes exhibited typical morphological characteristics of lipid bilayer vesicles (Fig. 1c).
Western blotting was performed to verify the presernce of various exosome-specific markers, such as CD63, TSG101 and HSP70, and to confirm the presence of LMP1 in NPC cell-derived exosomes (Fig. 1d).

EBV-LMP1 transferred to ECs via NPC cell-derived exosomes correlates with tumor angiogenesis
Tumor cell-derived exosomes were labeled with the lipophilic membrane dye PKH26 (Fig. 2a). After 24 h of incubation, PKH26-labeled exosomes were taken up by HUVECs, as shown in representative fluorescence microscopic images (Fig. 2a). It has been reported that EBV-LMP1 functions as a proangiogenic oncoprotein and correlates with tumor angiogenesis [5,16,28]. To further explore the contribution of LMP1 to tumor angiogenesis in NPC, we performed immunofluorescence staining of EBV-positive primary tumor tissues. We found that LMP1 colocalizes with CD31-positive vascular ECs or is highly expressed in perivascular NPC cells. Two representative immunofluorescence images of sections from two independent patients with NPC are shown in Fig. 2b.

Exosomes harboring LMP1 modulate endothelial functional phenotypes
Next, we studied the effects of NPC cell-derived exosomes carrying LMP1 on the functional characteristics contributing to tumor angiogenesis in ECs in vitro. HUVECs were incubated with exosomes extracted from the conditioned medium of NPC cells. We found that LMP1-containing exosomes promoted HUVEC proliferation and migration (Fig. 3a-d). Moreover, we found that tumor exosomes significantly enhanced the tube-forming ability of HUVECs and permeabilized the endothelial layer (Fig. 3e, f). We further assessed activation of the Akt/ERK signaling pathway by detecting Akt and ERK phosphorylation levels. Using Western blotting, we found that the p-AKT and p-ERK expression levels increased in HUVECs incubated with exosomes for 72 h (Fig. 3g). These results suggest that NPC cell-derived exosomes modulate the endothelial functional phenotypes responsible for tumor angiogenesis through the Akt/ERK pathway.

STIM1 regulates tumor angiogenesis by controlling EBV-LMP1 delivery
Exosome release is regulated by changes in cytosolic Ca 2+ levels [19][20][21], and it was recently reported that STIM1mediated Ca 2+ signaling manipulates the release of exosomal matrix proteins [22]. However, the regulatory role of STIM1 in exosomal EBV-LMP1 delivery remains to be clarified.
To address this issue, we established stable STIM1-silenced NPC cells lines by transfecting HK1-EBV and C666-1 cells with lentiviral vectors carrying scrambled shRNA (shRNA-Ctrl.) or STIM1-RNAi (shRNA-STIM1) sequences. Reduced STIM1 expression in NPC cells was confirmed via Western blotting (Fig. 4a). The exosomal marker content was also assessed in the exosomes extracted from the conditioned medium of shRNA-Ctrl. and shRNA-STIM1 cells (Fig. 4a). We found that the amount of LMP1 in the exosomes released from shRNA-STIM1 cells was significantly lower than that released from shRNA-Ctrl. cells (Fig. 4a). Thus, we further investigated whether exosomal LMP1 may modulate angiogenesis-related features in ECs. We found that the promotion of HUVEC proliferation, migration, tubulogenesis and permeability after treatment with exosomes was significantly blunted in the shRNA-STIM1 group (Fig. 4b-g). We also found that Akt and ERK phosphorylation was attenuated in HUVECs incubated with exosomes secreted from shRNA-STIM1 cells (Fig. 4h). Taken together, our findings strongly suggest that exosomal LMP1 modulates endothelial functional properties responsible for tumor angiogenesis in an Akt/ERK pathway-dependent manner.

STIM1 knockdown inhibits angiogenesis in a chick embryo CAM xenograft model
We previously reported that STIM1-mediated signaling mediates EBV-facilitated tumor angiogenesis in NPC [18]. As yet, however, the remodeling of the original blood vessels induced by the transplanted NPC tumor has not been fully elucidated yet. The chick embryo CAM model established in our earlier work imitates realistic malignant profiles of tumor cells, such as growth, progression and metastasis [26]. The CAM model also enables the visual and quantitative assessment of tumor angiogenesis in xenografts or transplanted primary NPCs [26]. In this study, we developed this chick embryo CAM model with slight modifications to intuitively observe how the exosome-releasing NPC cells recruit ECs to form tumor-associated vasculature (Fig. 5a). Five days after inoculation, the xenografts exhibited a visibly pro-angiogenic effect on vascular network formation, as compared with the Matrigel-control group. Radially grown vessels around the xenograft were generated in the CAMs (Fig. 5b), which were inhibited by STIM1 knockdown in HK1-EBV cells. In addition, the ratio of the vascularization area (VA) to the total CAM area was significantly reduced in the shRNA-STIM1 group (Fig. 5c). By performing section analysis with H & E staining, we observed intra-tumor neovascularization, which was defined as the newly-generated small-caliber vessels adjacent to the tumor cell clusters, and could even infiltrate the gaps between tumor cell clusters. The presence of embryonic nucleated erythrocytes in the lumen of the blood vessels allowed the tumor-associated vasculature to be recognized easily even without IHC (Fig. 5d). Intra-tumor neovascularization could be observed in 50% (5/10) of the shRNA-Ctrl. xenografts, but was seen in none of (0%, 0/10) the shRNA-STIM1 xenografts. The tumor-associated vasculature was further confirmed by IHC analysis of CD31, a classic endothelial marker (Supplementary Fig. S1). These results indicate that EBV-positive NPC cells reforme the original blood vessels by inducing secondary vasculature infiltration of the transplanted tumors in the CAMs. Inhibition of exosomal LMP1 delivery by silencing STIM1 in EBV-infected NPC cells partially eliminated the pro-angiogenic effect.

Discussion
Rapidly growing tumor cells need abundant blood carrying nutrients and oxygen, which allows dysregulated neoplastic proliferation, and accelerated tumor expansion requires an even increasingly larger blood supply. Tumor angiogenesis sets up a vicious cycle of demand and supply. The recently emerged single-cell sequencing technology has revealed the landscape of tumor heterogeneity and the immune microenvironment in NPC, which is profoundly affected by latent EBV infection [7,29,30]. However, how EBV infection and its products modulate EC characteristics in the TME and facilitate angiogenesis and hematogenous dissemination remains understudied.
As a crucial intracellular second messenger, Ca 2+ signaling regulates various cellular activities. Aberrantly activated SOCE is involved in tumor progression, angiogenesis and metastasis and is clinically associated with treatment  [31]. Previously, we showed that STIM1-dependent SOCE mediates tumor angiogenesis promotion by EBV infection in NPC [17,18], but the pathway through which EBV-infected NPC cells manipulate endothelial phenotypes, remains unclear. In this study, we demonstrate that STIM1 regulates tumor angiogenesis by controlling the delivery of exosomal LMP1 to ECs in NPC. As a vital EBV-encoded functional oncoprotein, LMP1 has been implicated in the regulation of tumor angiogenesis through various approaches. For instance, LMP1 has been found to stimulate multiple regulators of pro-angiogenic factors and induce the production of molecules, such as IL-8, FGF-2 and VEGF [32][33][34][35]. It has also been reported that LMP1 can promotes the formation of vasculogenic mimicry in NPC through VEGFA/VEGFR1 signaling [36]. In addition, LMP1 can be transferred to ECs through exosomes secreted by EBV-infected cells, including NPC cells [12].
Endogenous LMP1 in EBV-infected host cells can escape degradation by accumulating in the luminal vesicles of multivesicular endosomes and can further be released into the external space via exocellular secretion in exosomes (Fig. 6) [37]. CD63 is a tetraspanin protein enriched in late endosomes, lysosomal compartments and exosomal membranes [37]. LMP1 enrichment in exosomes and its secretion have been associated with CD63, whose knockdown results in reduced LMP1 enrichment in exosomes [38]. In our study, exosomal CD63 derived from NPC cells transfected with shRNA-STIM1 was comparable to that from shRNA-Ctrl. cells, suggesting is regulated by STIM1. Exosomal LMP1 endows ECs with aggressive functional characteristics through the Akt/ERK pathway. The "aggressive" ECs thus contribute to the remodeling of original vascular and intra-tumor neovascularization that STIM1 regulates LMP1 enrichment in exosomes via a CD63-independent pathway. The endosomal-lysosomal sorting pathway plays a key role in exosome biogenesis and cargo sorting. This pathway involves ubiquitination, lipid sorting and recognition by endosomal sorting complexes required for transport [39]. However, sorting LMP1 into exosomes does not require the lipid raft anchoring domain FWLY or N-terminal ubiquitination [37]. On one hand, LMP1 mediates proper intracellular trafficking through its C-terminus, whose modifications impair intracellular LMP1 trafficking and export via exosomes [37]. On the other hand, deletion of the N-terminus and transmembrane segments 1-2 or 1-4 of LMP1 impair its trafficking into exosomes, and the LMP1 C-terminus is not necessary for its enrichment [40]. The main limitation of our study is that the molecular structure domains through which STIM1 signaling regulates LMP1 enrichment into exosomes have not yet been fully addressed, mostly because of the complexity of the LMP1 sorting mechanism. This challenging issue needs to be solved in future studies.
In the present study, we found that STIM1 knockdown inhibited the delivery of exosomal LMP1 while restraining Akt/ERK pathway activation in HUVECs. Similarly, STIM1-mediated signaling in tumor cells has been found to remotely manipulate the Akt/ERK pathway in ECs [41]. STIM1 silencing has been found to promote the enrichment of exosomal miR-145 released from breast cancer cells and to inhibit insulin receptor substrate 1 and Akt/ ERK signaling in ECs [41]. In addition to the Akt/ERK pathway, Akt can activate other downstream signaling molecules to promote angiogenesis. Akt activates endothelial nitric oxide synthase, a key regulator of EC proliferation, migration and survival, which in turn stimulates nitric oxide production and induces angiogenesis [42,43].
As a tumor progresses, angiogenesis provides the nutrients and oxygen needed for its uncontrolled growth in situ, offers an indispensable path for locoregional invasion, and supports distant metastatic colonization. Our findings indicate that the exosomes harboring EBV-encoded LMP1 released by NPC cells enhance EC proliferation, migration, tubulogenesis and permeability through an Akt/ERKdependent pathway (Fig. 6). STIM1 regulates tumor angiogenesis by manipulating exosomal LMP1 enrichment. In addition, we further developed our CAM xenograft model for characterizing tumor angiogenesis in NPC, which enables a three-dimensional analysis of intra-tumor neovascularization. Utilizing this model, we found that STIM1 knockdown reverses exosomal LMP1-promoted angiogenesis in vivo. This indicates that blockage of the exosomal delivery of EBV-associated pro-oncogenic biomolecules could serve as a feasible strategy for targeted therapy in NPC.