Hypoxia‐induced tumor exosomes promote angiogenesis through miR‐1825/TSC2/mTOR axis in oral squamous cell carcinoma

Oral squamous cell carcinoma (OSCC) is characterized by enhanced angiogenesis resulting in poor prognosis despite improvements in diagnostic/therapeutic techniques. Here, we aimed at investigating potential roles of miR‐1825 enclosed in OSCC‐derived exosomes on angiogenesis under hypoxic conditions.


| INTRODUCTION
Oral squamous cell carcinoma (OSCC), constituting almost 95% of all head and neck cancers, is the sixth most prevalent malignant tumor worldwide with 377 713 new cases and 177 757 deaths in 2020. 1,2 OSCC is characterized by enhanced angiogenesis as well as frequent invasion and metastasis, resulting in a 5-year survival rate that is still below 50% despite the improvements in the diagnostic and therapeutic techniques. 3,4 Therefore, development of novel therapeutic strategies along with understanding the molecular mechanisms underpinning OSCC pathogenesis are urgent needs to provide OSCC patients with successful clinical outcomes.
Tumor angiogenesis is necessary for oxygenation and supply of essential nutrients for tumor microenvironment that facilitates tumor distal metastasis. 5 Hypoxia, which is defined as the reduced oxygen availability in the tumor microenvironment, is a triggering power for angiogenesis and metastasis. 6,7 Hypoxia is observed as a result of faster tumor growth than vessel formation due to rapidly proliferating cancer cells. 7 As a response to hypoxia, a cohort of genes and signaling cascades that are related to the low oxygen tension are activated in cancer and endothelial cells, which includes hypoxia-inducible factors (HIFs), essential components of hypoxia signaling cascades, or several other HIF-1-independent signaling cascades including PI3K/Akt/mTOR, Wnt/β-catenin, and MAPK pathways. 7,8 These alterations at the molecular level results in enhanced angiogenesis to supply cancer cells with more oxygen and nutrients.
Recent studies have shown that one of the means of cellular communication between tumor cells and their microenvironment is extracellular vesicles that are essentially released from hypoxic tumor cells. Exosomes are lipid bilayer membrane vesicles approximately 30-300 nm in diameter, which are secreted into the extracellular space. 6 They can shuttle functional biomolecules including DNA fragments, microRNAs, long noncoding RNAs, proteins, and lipids between cells. 9 An increased exosomal secretion from cancerous cells under hypoxic conditions has been reported in many tumors including oral squamous cell carcinoma 8,10-12 and exosomes released from hypoxic tumor cells were found to play critical roles in cancer angiogenesis, pre-metastatic niche formation, metastasis, and tumor progression. 13 MicroRNAs, which are approximately 22 nucleotides in length, are endogenous non-coding RNA molecules that function as post-transcriptional regulators of gene expression and involved in regulation of many biological processes by degradation and/or translational repression of target mRNAs. 14,15 Deregulation of microRNA levels has been associated to numerous diseases including cancer, where they can play as potential oncogenes or tumor suppressors. 16 Recently, several microRNAs have been reported to be involved in tumor angiogenesis. 17 To adapt to the hypoxic microenvironment, certain microRNAs were found to be excessively expressed by tumor cells and transferred to neighboring cells as trapped inside exosomes, 18,19 which has been reported to influence angiogenic activity of endothelial cells. 6,7,10,20 miR-1825, located inside the 7 th exon of POFUT1 (protein O-fucosyltransferase 1) at 20q11.21 chromosomal region, has been reported to be a recurrent gain of function abnormality in human embryonic stem cells and induced pluripotent stem cells. 21,22 Nguyen et al. reported that human embryonic stem cells with 20q11.21 amplification displayed increased colony forming potential and decreased apoptosis. 23 Interestingly, 20q11.21 amplification in human embryonic stem cells resulted in acquisition of a gene-expression signature enriched for cancer-associated genes and miR-1825 expression was reported to be elevated in a variety of carcinomas. 24 Herein, we aimed at investigating the potential roles and mechanism of action of OSCC-derived exosomes on angiogenesis under hypoxic conditions. We demonstrated that hypoxia led to OSCC-derived exosome mediated transfer of miR-1825 to human umbilical vein endothelial cells (HUVECs) and enhanced angiogenesis. Additionally, we showed that hypoxia-induced exosomal miR-1825 promoted angiogenesis via regulating the miR-1825/ TSC2/mTOR axis in OSCC.
To generate chemically induced hypoxia models, OSCC cells were exposed to cobalt chloride (CoCl 2 , Sigma-Aldrich). To determine the optimum concentrations for inducing hypoxia without causing significant toxicity, cells were treated with CoCl 2 in different concentrations (50-100-150-200-250 μM) for 24 or 48 h, then cell viabilities were evaluated. Lysates and culture mediums from hypoxia treated cells were processed for experiments described below using optimum CoCl 2 concentrations for each cell line.

| Isolation and characterization of OSCCs-derived exosomes
FaDu and SCC-9 cells were cultured until they reach 50%-60% confluency in T175 (SPL, Gyeonggi-do, Korea) cell culture flasks. Then, cells were washed twice with 1Â phosphate buffered saline (PBS; EcoTech Biotechnology, Erzurum, Turkiye) and cultured in 80 mL of fresh serum-free culture medium with (to mimic hypoxic conditions) and without (to mimic normoxic conditions) CoCl 2 for 48 h. One hundred and 150 μM CoCl 2 were used for FaDu and SCC-9 cells to induce hypoxia, respectively. Then, conditioned media were collected and centrifuged at 2000g for 20 min at 4 C to remove unattached cells, followed by a second round of centrifugation at 12 000g for 45 min at 4 C. Centrifuged conditioned media were filtered using a 0.22-μm filter (Isolab, Eschau, Germany) and ultracentrifuged at 110 000g for 2 h at 4 C in sterile ultracentrifuge tubes (HITACHI, Himac CS150NX S50A rotor). Exosome pellets were resuspended in 1 mL cold 1Â PBS and again ultracentrifuged at 110 000g for 2 h at 4 C to recover exosome pellets. Concentrations of the isolated exosomes were measured using BCA Protein Assay Kit (Pierce, Thermo Fisher, Gaithersburg, MD) and stored at À80 C until their use.
The morphology of normoxic and hypoxic exosomes was examined using Hitachi HighTech HT7700 Transmission Electron Microscope (TEM, Hitachi, Chiba, Japan). Briefly, exosomes were suspended in 1Â PBS and mixed well. Fifty microliter of exosomes for each group was loaded onto 300 mesh copper grids and incubated for 16 h at room temperature. Then, grids were critical-point dried and imaged using TEM at 80-120 kV.
The particle size and diameter distribution of normoxic (Norm-Exos) and hypoxic (Hypo-Exos) exosomes suspended in PBS were measured using Malvern Zetasizer Nano ZSP90 (Malvern Panalytical, Malvern, UK). Briefly, exosomes dissolved in 1 mL 1Â PBS were transferred inside the Zetasizer Nano ZSP90 and the refractive index for extracellular vesicles was set as 1.40. 26 Zetasizer software (Malvern Instruments) was used to evaluate the results.
Exosome-specific surface marker CD9, CD63, and CD81 was detected by Western blot as described below under western blot analysis subheading.

| Exosome labeling
Isolated exosomes were labeled using PKH67 Green Fluorescent Cell Linker Mini Kit for General Cell Membrane Labeling (Sigma-Aldrich) following the manufacturer's recommendations. Fluorescently labeled exosomes were resuspended in 1Â PBS and stored at À20 C until their use. In the meanwhile, HUVEC cells were grown on cover slips until they reach 55%-60% confluency within 6-well plates. Then, HUVEC cells in each well were treated with 10 μg PKH67 labeled exosomes or control exosomes at 37 C for 24 h. At the end of 24 h incubation, cells were washed with 1Â PBS three times, fixed with 4% glutaraldehyde solution for 30 min at 4 C, and stained with Hoechst 33258 dye. Then, PKH67 labeled exosome treated and stained with Hoechst 33258 dye cells were visualized using LSM 710 Confocal Laser Scanning Microscope (Zeiss, Oberkochen, Germany).

| MicroRNA transfection and cotreatment of cells with microRNAs and exosomes
HUVEC cells were seeded into 6-well plates at a density of 8 Â 10 4 cells/well. Cells that reached at approximately 60% confluency were transfected with miR-1825 mimics, inhibitors, or negative controls (Sigma-Aldrich) using Lipofectamine 2000 (Invitrogen, Thermo Fisher, Gaithersburg, MD) following to the manufacturer's guidelines. To see the simultaneous effects of miR-1825 inhibition and exosome treatment, cells were initially transfected with miR-1825 inhibitor or negative control, then Norm-Exos or Hypo-Exos were applied to cells for 24 or 48 h for further analysis.

| RNA isolation from cells and exosomes
Total RNA with microRNA content from the exosomes derived from hypoxic OSCCs cells were extracted using the Invitrogen™ mirVana™ miRNA Isolation Kit (Thermo Scientific, Wilmington, DE) following the manufacturer's standard protocol. Total RNA from transfected HUVEC cells or hypoxic OSCC cells were extracted using EcoPURE Total RNA Kit (EcoTech Biotechnology, Erzurum, Turkiye) following the manufacturer's recommended instructions. Concentrations and the purities of RNA samples were measured using an Epoch 2 microplate spectrophotometer (260/280 rations). RNA samples were stored at À80 C until their use.

| Complementary DNA synthesis and quantitative real-time polymerase chain reaction
For quantification of microRNA levels, same amounts of RNA samples (30 ng/μL) were reverse transcribed to complementary DNA (cDNA) using TaqMan MicroRNA Reverse Transcription Kit (Thermo Fisher, Foster City, CA) and miR-1825 (TaqMan Assay ID: 002907) and RNU43 (TaqMan Assay ID: 001095) sequence-specific primers (Thermo Fisher, Foster City, CA) following manufacturer's guidelines. MicroRNA quantitative real-time polymerase chain reaction (qRT-PCR) was carried out using TaqMan Universal PCR Master Mix (Thermo Fisher, Foster City, CA) and miR-1825 and RNU43-specific probes (Thermo Fisher, Foster City, CA). RNU43 was used as internal control for normalization of microRNA expression.
For quantification of gene expressions in mRNA level, cDNAs were synthesized using equal amounts of RNA samples with High Capacity cDNA Reverse Transcription Kit ® (Applied Biosystems, Foster City, CA). qRT-PCR reactions were prepared using 5Â HOT FIREPol Eva-Green qPCR Mix Plus (Solis Bio-Dyne, Tartu, Estonia). GAPDH was used for standardization of gene expressions. All qRT-PCR reactions were carried out at least in duplicates in a Rotor-Gene qRT-PCR (Qiagen, Dusseldorf, Germany) device using standard parameters. Relative microRNA and mRNA expressions were calculated using 2 ÀΔΔCT method.

| Cell viability assay
Viabilities of FaDu, SCC-9, and HUVEC cells were measured with Cell Viability Detection Kit-8 (CVDK-8; Eco-Tech Biotechnology, Erzurum, Turkiye) following the manufacturer's instructions. FaDu (2.5 Â 10 4 cells/well) and SCC-9 (5 Â 10 3 cells/well) cells were seeded in 96-well plates in five replicates and incubated for 24 h at 37 C. Then, the cells were subjected to CoCl 2 at different concentrations (25-50-100-150-200 μM) for 24 or 48 h. HUVEC cells were seeded into 96-well plates at 2.5 Â 10 4 cells/well density in five replicates and incubated for 24 h at 37 C. Then, HUVEC cells were transfected with miR-1825 mimic/inhibitor or cultured in serum-free medium containing Norm-Exos or Hypo-Exos for 24 or 48 h. By the end of the incubation periods, CVDK-8 solution diluted in corresponding cell culture medium was added to each well and absorbance values at 450 nm was recorded with an Epoch Microplate Spectrophotometer (BioTek, Winooski, VT).

| Transwell migration assay
Transwell migration assay was conducted to determine the effects of hypoxia-induced exosome treatment or miR-1825 mimic/inhibitor transfection on the migration potential of HUVEC cells using transwell chambers equipped with polycarbonate filter membranes that have pores in 8-μm diameter (Corning, NY). Transfected HUVEC cells at a concentration of 4 Â 10 5 cells/chamber were suspended in 250 μL serum-free medium and seeded into the top chamber as duplicates per group and 500 μL complete medium containing 20% FBS was added into the lower chamber as chemoattractant. For co-treatment experiments, 100 μg/mL of either Norm-Exos or Hypo-Exos were added to the corresponding wells. After 24 h of incubation, cells that remained on the upper surface of the membranes were cleaned using a cotton swab and cells that migrated to the lower surface of the membranes were fixed with 100% methanol. Then, cells were stained with 0.1% crystal violet for 30 min. Membranes were washed with 1Â PBS and dried at room temperature to calculate the number of migrated cells under an inverted microscope (Leica, Wetzlar, Germany).

| Matrigel invasion assay
Matrigel invasion assay was performed to determine the effects of hypoxia-induced exosome treatment or miR-1825 mimic/inhibitor transfection on the invasion potential of HUVEC cells using 24-well transfer chambers coated with Corning ® Matrigel ® Growth Factor Reduced (GFR) Basement Membrane Matrix (Corning, NY). Transfected HUVEC cells at a concentration of 5 Â 10 5 cells/chamber were suspended in 250 μL serum-free medium and seeded into the top chamber as duplicates per group and 500 μL complete medium containing 20% FBS was added into the lower chamber as chemoattractant. For co-treatment experiments, 100 μg/mL of either Norm-Exos or Hypo-Exos were added to the corresponding wells. After 24 h of incubation, cells that remained on the upper surface of the membranes were cleaned using a cotton swab and cells that migrated to the lower surface of the membranes were fixed with 100% methanol. Then, cells were stained with 0.1% crystal violet for 30 min. Membranes were washed with 1Â PBS and dried at room temperature to calculate the number of migrated cells under an inverted microscope (Leica).

| Matrigel tube formation assay
The Matrigel tube formation assay was performed to evaluate differential angiogenic capacity of HUVEC cells treated with hypoxia-induced exosomes or transfected with miR-1825 mimic/inhibitors. Briefly, 40 μL of Corning ® Matrigel ® Growth Factor Reduced (GFR) Basement Membrane Matrix (Corning, NY) was coated into each well of a 96-well plate and incubated for 1 h at 37 C for polymerization. Then, miR-1825 mimic/inhibitor transfected 2 Â 10 4 HUVEC cells/well were added to wells on top of Matrigel. For co-treatment experiments, 100 μg/mL of either Norm-Exos or Hypo-Exos were added to the corresponding wells, and cells were incubated for 16 h at 37 C. Non-targeting microRNAs served as control. Differential tube formation capacities of HUVEC cells were evaluated via calculating the number of branches using images obtained with an inverted microscope (Leica, Wetzlar, Germany). Experiments were performed in at least duplicates.
2.11 | Spheroid-based 3D angiogenesis assay HUVEC spheroids were developed using hanging drop method as described previously. [27][28][29][30] In brief, HUVEC cells (1000 cells/drop) were suspended in 25 μL complete DMEM containing 0.20% methylcellulose and seeded using a multichannel pipette under the lids of cell culture dishes (Greiner, Frickenhausen, Germany). To restrain evaporative loss, 1Â PBS was added to the bottom of cell culture dishes. Then, all suspended cells within hanging drops were cultured to form rounded spheroids for ≥24 h at 37 C. For angiogenesis assay, hanging drops containing spheroids were gently washed with 1Â PBS using 5 mL serological pipette and collected with centrifugation at 300g for 5 min at 4 C. Then, pellets were resuspended in methocel containing 30% Corning ® Matrigel ® Growth Factor Reduced (GFR) Basement Membrane Matrix (Corning, NY), rapidly transferred into a prechilled 96-well plate, and let polymerization for 30 min at 37 C. Forty microliter of endothelial cell growth medium containing bEGF (Protech Human, Cat no: AF-100-15-100UG), FGF (Protech Human, Cat no:100-18B-10UG), rhIGF-I (StemCell Tech, Cat no: 7022.2) prepared in serum-free medium was added on top of the gels and cells were incubated for 24 h in a humidified cell culture incubator. All experiments were carried out in at least duplicates. 3D angiogenic sprout length and numbers were calculated under invert microscope (Leica).

| In silico analysis and microRNA target prediction
The publicly available GSE12546 dataset in the Gene Expression Omnibus (GEO) database was analyzed to find the differentially expressed genes in HUVEC cells under mimic hypoxia stress using the GEO2R in silico tool using the standard parameters. The dot plot representation of the differentially expressed genes was produced via GEO2R tool. Putative targets of miR-1825 was predicted using miRWalk 2.0, 31 miRanda, 32 RNA22, 33 and TargetScan 34 tools and target genes that were predicted by all of the four target prediction tools were evaluated as potential targets to be able to exclude false positive predictions. PathwayMapper tool of cBioPortal was used to determine the TSC2 associated pathways through using the data derived from Pan-Cancer Atlas of TCGA pathways. 35

| Statistical analysis
Graphs were plotted as mean ± standard error of the mean. Significance of differences was statistically evaluated with Student's t test. p < 0.05 was considered as significant. Data was visualized using GraphPad Prism 9.0 (GraphPad Software, Inc., San Diego, CA).

| miR-1825 expression is elevated in hypoxia-induced tumor exosomes and tumor cells
Considering the fact that hypoxia enriches cancer stem like cells within the cancer cell population, 36 we hypothesized that miR-1825, a microRNA that was found to be elevated in various cancer types and resides within a recurrent gain of function region in human embryonic stem cells and induced pluripotent stem cells, could be enriched in cells exposed to hypoxic stress conditions as well as in hypoxia-induced tumor exosomes (hiTDEs). To test our hypothesis, we initially used mimetic agent CoCl 2 in order to induce hypoxia in OSCC cells. To find the effective doses of CoCl 2 with the least toxicity on the FaDu and SCC-9 cells, we treated cells with increasing doses of CoCl 2 in serum-free medium for 24 and 48 h and found that 100 and 150 μM CoCl 2 did not significantly killed FaDu and SCC-9 cells, respectively, at the end of the 48 h ( Figure 1A,B). To confirm the induction of hypoxia, we examined the relative expressions of the hypoxia markers at the mRNA (HIF1α, HIF2α, and CA-9) and protein (HIF1α) levels in FaDu and SCC-9 cells treated with aforementioned doses of CoCl 2 . In line with the literature findings 37-39 we found HIF1α and HIF2α levels as decreased and CA-9 level as increased at mRNA level in hypoxia-induced cells as compared to control normoxic cells (Figure 1C,D). As expected, CoCl 2 treatment significantly induced HIF1α protein level in both FaDu and SCC-9 cells under hypoxic conditions ( Figure 1E), which pointed successful establishment of hypoxia model in OSCC cells via treatment of CoCl 2 .
As a preliminary analysis, we treated cancer cells with hypoxia and evaluated the miR-1825 level in these cells. We found that hypoxic conditions resulted in significant increase in miR-1825 level in OSCC cells (Figure 2A), suggesting a possible increase in miR-1825 level within the exosomes secreted from hypoxic cancer cells. To analyze exosomal miR-1825 levels, we isolated exosomes from serum-free conditioned medium of FaDu and SCC-9 cells exposed to normoxic and hypoxic conditions for 48 h. To confirm isolation of exosomes, we visualized them using TEM and characterized them with nanoparticle tracking and western blot analysis. The morphologies of isolated exosomes were quantified under TEM, which revealed presence of typical rounded microvesicles ranging from 50 to 350 nm in size ( Figure 2B). A similar size distribution and intensity was found in either normoxia or hypoxia-induced exosomes. There were no morphological differences between the normoxic and hypoxic groups with regard to the size and shape of the exosomes ( Figure 2C). In parallel with these findings, we demonstrated the presence of exosomal surface marker CD63, CD9, and CD81 in lysates extracted from exosomes derived from normoxic or hypoxic FaDu and SCC-9 cells using western blot analysis, which all together validated successful isolation of exosomes ( Figures 2D and S1A, Supporting Information). We, then, analyzed the miR-1825 level in exosomes derived from cells under hypoxic conditions and found that miR-1825 is elevated in the cargo of hypoxia-induced exosomes compared to those of collected under normoxic conditions ( Figure 2E). These findings potentially ascribe a putative role for miR-1825 during reorganization of extracellular environment of cancer cells when they are exposed to hypoxic conditions, which necessitates induction of angiogenesis in order to get rid of hypoxic stress.

| miR-1825 alters the angiogenic potential of endothelial cells
To test the effects of altered expression of miR-1825 on endothelial cells, we initially overexpressed ( Figure 3A) or inhibited ( Figure 3B) miR-1825 in HUVEC cells using miR-1825 mimic or miR-1825 inhibitor, respectively. Cells transfected with non-targeting control microRNAs served as control for both conditions. Initial assessment of proliferative capacity of HUVEC cells upon miR-1825 mimic transfection pointed that miR-1825 overexpression significantly promoted the relative cell viability of HUVEC cells compared to the non-targeting control microRNA transfected cells ( Figure 3C). On the other hand, miR-1825 inhibition suppressed the viability of HUVEC cells ( Figure 3D). In the meanwhile, compared with non-targeting control microRNA, miR-1825 mimic obviously enhanced the migratory and invasive potential of endothelial cells ( Figure 3E,F), and miR-1825 inhibition significantly decreased the endothelial cell migration and invasion capacities ( Figure 3G,H). We then evaluated the effects of altered miR-1825 expression on the angiogenic potential of endothelial cells. Tube formation and spheroid-based 3D angiogenesis assays demonstrated that ectopic miR-1825 overexpression promoted and that suppression of miR-1825 expression inhibited the tube formation capabilities ( Figure 3I,J) and reduced the number of sprouts formed per spheroids ( Figure 3K,L) compared with corresponding controls. These findings indicate that miR-1825 has the capability to alter the angiogenic potential of endothelial cells.

| hiTDEs promote angiogenesis via their miR-1825 content
It is a known phenomenon that microRNAs are carried from cell to cell as exosomal cargos. 20 Since we demonstrated that miR-1825 is enriched in tumor-derived exosomes and that miR-1825 has the potential to induce phenotypes associated with angiogenesis in endothelial cells, we next tested whether hiTDEs promote angiogenesis via their miR-1825 content. To demonstrate the transfer of OSCC cell-secreted miR-1825 via exosomes to HUVEC cells, we labeled exosomes with green fluorescent PKH67 and demonstrated the successful uptake of fluorescent labeled exosomes using confocal microscope ( Figure 4A). To further confirm that miR-1825 is transported to endothelial cells via exosomes, we treated HUVEC cells with exosomes derived from FaDu and SCC-9 cells cultured under hypoxic conditions for 48 h and found that miR-1825 level significantly increases compared to corresponding controls ( Figure S1B,C). In addition, the increase in miR-1825 levels in recipient cells as a result of exosome treatment lead to increase in cell viability in HUVEC cells ( Figure S1D,E). To exclude the possibility of increase in endogenous miR-1825 level after exosome treatment, we measured the expression of POFUT1, the host gene of miR-1825, in HUVEC cells that are incubated with exosomes derived from normoxic and hypoxic OSCC cells and found that POFUT1 expression was not significantly different between cells treated either with normoxic exosomes or hypoxic exosomes ( Figure S2A,B), suggesting that the change in miR-1825 level does not stem from alteration in its transcriptional activity but due to the transfer of ectopic miR-1825 to cells via exosomes. Before analyzing the effects of hiTDEs treatment on HUVEC cells, we demonstrated that the increase in miR-1825 level in HUVEC cells as a result of exosome treatment is reversed upon transfection of cells with miR-1825 inhibitor ( Figure 4B,C). Then, we evaluated the involvement of exosomal miR-1825 in viability of HUVEC cells using cell viability assay and found that the viability of cells treated with hiTDEs significantly increased compared to controls, which is reversed by inhibition of miR-1825 expression ( Figure 4D,E). In addition, hiTDE exposure significantly enhanced the migration and invasion potential of HUVEC cells ( Figure 4F-I). Similar to the findings of cell viability assay, the migratory and invasive potentials of endothelial cells, which enhance their capability for angiogenesis, was obviously abolished by miR-1825 inhibition ( Figure 4F-I). We next performed tube formation and spheroid-based 3D angiogenesis assays to determine whether hypoxia-induced tumor exosomal miR-1825 alters angiogenic features of endothelial cells. We demonstrated that hiTDE exposure of HUVEC cells promoted the tube formation capabilities and the number of sprouts formed per spheroids in those cells compared to corresponding controls, which is suppressed via inhibition of miR-1825 ( Figure 4J-M). These findings indicate that one of the important tools of hiTDEs is miR-1825, through which they alter angiogenic potential of endothelial cells.

| hiTDEs contribute to angiogenesis via regulation of the miR-1825/TSC2/ mTOR axis
To further investigate the possible underlying mechanisms of altered angiogenesis as a result of hiTDE treatment in HUVEC cells, we searched for publicly available datasets deposited in the Gene Expression Omnibus (GEO) database and investigated the differentially expressed genes (DEGs) under mimic hypoxia conditions in HUVEC cells using the GSE12546 dataset. Analysis with GEO2R in silico tool demonstrated that 597 genes were upregulated and 1791 genes were downregulated under hypoxic condition in HUVEC cells ( Figure 5A). Putative targets of miR-1825 was predicted using miR-Walk 2.0, miRanda, RNA22, and TargetScan tools and target genes that were predicted by all of the four target prediction tools were evaluated as potential targets to be able to exclude false positive predictions. Venn diagram shows the number of common genes deregulated in HUVEC cells under hypoxic conditions and predicted F I G U R E 4 (A) PKH67 staining of HUVEC cells exposed to exosomes to demonstrate the successful inclusion of microvesicles by the endothelial cells. Hoechst, a cell membrane Permeable dye, is used to label the nuclear DNA of cells. PKH67 is used to label exosomes to track transfer of their cargo to the recipient cells. (B) Relative expression level of miR-1825 in HUVEC cells exposed to 20 μg exosomes isolated from the serum-free culture media of FaDu cells treated with or without CoCl 2 in addition to 10 μM non-targeting control microRNA or miR-1825 inhibitor at 24 h. (C) Relative expression level of miR-1825 in HUVEC cells exposed to 20 μg exosomes isolated from the serum-free culture media of SCC-9 cells treated with or without CoCl 2 in addition to 10 μM non-targeting control microRNA or miR-1825 inhibitor at 24 h. (D) Relative cell viability of HUVEC cells exposed to exosomes isolated from the serum-free culture media of FaDu cells treated with or without CoCl 2 in addition to non-targeting control microRNA or miR-1825 inhibitor at 24 or 48 h. (E) Relative cell viability of HUVEC cells exposed to exosomes isolated from the serum-free culture media of SCC-9 cells treated with or without CoCl 2 in addition to non-targeting control microRNA or miR-1825 inhibitor at 24 or 48 h. (F) Relative cell migration of HUVEC cells exposed to exosomes isolated from the serum-free culture media of FaDu cells treated with or without CoCl 2 in addition to non-targeting control microRNA or miR-1825 inhibitor at 24 h. (G) Relative cell migration of HUVEC cells exposed to exosomes isolated from the serum-free culture media of SCC-9 cells treated with or without CoCl 2 in addition to non-targeting control microRNA or miR-1825 inhibitor at 24 h. (H) Relative cell invasion of HUVEC cells exposed to exosomes isolated from the serum-free culture media of FaDu cells treated with or without CoCl 2 in addition to non-targeting control microRNA or miR-1825 inhibitor at 24 h. (I) Relative cell invasion of HUVEC cells exposed to exosomes isolated from the serum-free culture media of SCC-9 cells treated with or without CoCl 2 in addition to non-targeting control microRNA or miR-1825 inhibitor at 24 h. (J) Relative tube formation of HUVEC cells exposed to exosomes isolated from the serumfree culture media of FaDu cells treated with or without CoCl 2 in addition to non-targeting control microRNA or miR-1825 inhibitor at 16 h. (K) Relative tube formation of HUVEC cells exposed to exosomes isolated from the serum-free culture media of SCC-9 cells treated with or without CoCl 2 in addition to non-targeting control microRNA or miR-1825 inhibitor at 16 h. (L) Relative sprouts outgrowth of HUVEC spheroids exposed to exosomes isolated from the serum-free culture media of FaDu cells treated with or without CoCl 2 in addition to nontargeting control microRNA or miR-1825 inhibitor at 24 h. (M) Relative sprouts outgrowth of HUVEC spheroids exposed to exosomes isolated from the serum-free culture media of SCC-9 cells treated with or without CoCl 2 in addition to non-targeting control microRNA or miR-1825 inhibitor at 24 h. microRNA expression levels were normalized to RNU43. Data are plotted as mean ± SD for (D, E). *p < 0.05. # represents significant difference between Hyp-exosome/control group and Hyp-exosome/inh-miR-1825 group [Color figure can be viewed at wileyonlinelibrary.com] Relative expression level of TSC2 at mRNA and protein level in HUVEC cells exposed to exosomes isolated from the serum-free culture media of FaDu cells treated with or without CoCl 2 in addition to non-targeting control microRNA or miR-1825 inhibitor at 24 or 48 h, respectively. (F) Relative expression level of TSC2 at mRNA and protein level in HUVEC cells exposed to exosomes isolated from the serum-free culture media of SCC-9 cells treated with or without CoCl 2 in addition to non-targeting control microRNA or miR-1825 inhibitor at 24 or 48 h, respectively. (G) Localization of TSC2 within PI3K/AKT/mTOR signaling pathway. (H) Relative expression level of p-mTOR in HUVEC cells transfected with either non-targeting control microRNA, miR-1825 mimic or miR-1825 inhibitor at 48 h. (I) Relative expression level of p-mTOR in HUVEC cells exposed to exosomes isolated from the serum-free culture media of FaDu cells treated with or without CoCl 2 in addition to non-targeting control microRNA or miR-1825 inhibitor at 48 h. (J) Relative expression level of p-mTOR in HUVEC cells exposed to exosomes isolated from the serum-free culture media of SCC-9 cells treated with or without CoCl 2 in addition to nontargeting control microRNA or miR-1825 inhibitor. mRNA expression levels were normalized to GAPDH at 48 h. microRNA expression levels were normalized to RNU43. Protein levels were normalized to β-Actin. Data are plotted as mean ± standard deviation of the mean (SD) for (A-D). *p < 0.05. # represents significant difference between Hyp-exosome/control group and Hyp-exosome/inh-miR-1825 group [Color figure can be viewed at wileyonlinelibrary.com] targets of miR-1825 ( Figure 5B). Of those TSC2 stands forward due to its involvements in angiogenic processes. [39][40][41] We demonstrated downregulation of TSC2 expression in both at mRNA and protein level in HUVEC cells overexpressing miR-1825 mimic ( Figure 5C), and its upregulation in HUVEC cells transfected with miR-1825 inhibitor ( Figure 5D) compared to corresponding controls. Next, we assessed expression of TSC2 at mRNA and protein level in HUVEC cells exposed to tumor-derived exosomes and found that TSC2 expression level significantly decreased in hiTDEs treated cells, which is in line with expectations since hypoxic exosomes are enriched with miR-1825 ( Figure 5E,F). Inhibition of miR-1825 in cells exposed to hiTDEs resulted in recovery of TSC2 levels to the normoxic conditions ( Figure 5E,F). To further delineate into the underlying molecular mechanisms, we used PathwayMapper tool of cBioPortal to find the TSC2 associated pathways using the data derived from Pan-Cancer Atlas of TCGA pathways, which pointed mTOR as a target of TSC2, within PI3K/AKT/ mTOR signaling pathway ( Figure 5G). Considering the importance of mTOR activation in angiogenesis, we initially measured the active mTOR levels (p-mTOR) in cells either overexpressing miR-1825 or with miR-1825 inhibition, and demonstrated that active mTOR levels significantly increased in HUVEC cells overexpressing miR-1825 and decreased in cells transfected with miR-1825 inhibitor, suggesting a possible miR-1825/TSC2/mTOR axis ( Figure 5H). We, then evaluated p-mTOR levels in cells exposed to normoxic or hypoxic FaDu and SCC-9 cell-derived exosomes, and demonstrated activation of mTOR upon hiTDE treatment, which is reversed by miR-1825 inhibition ( Figure 5I,J). Considering these findings, we suggest that miR-1825 as upregulated in hiTDEs could be transferred to HUVEC cells, which in turn promotes angiogenesis by deregulating the TSC2/mTOR axis ( Figure 6A,B).

| DISCUSSION
The 5-year survival rate of patients with advanced head and neck cancer remains to be around only 50% after treatment of patients with standard therapy. 3,4 Diagnosis at late stage, lymph node and distant metastasis are thought to be the critical reasons leading to poor prognosis and reduced survival rates in HNSCC. 42 Angiogenesis, a complex process involving endothelial cell proliferation, migration and formation of new blood vessels within the tumor mass, provides cancer cells with an enhanced vascular network that contributes to tumor growth and metastasis. 43 The angiogenic switch is considered as one of the crucial steps for metastatic spread in HNSCC as well similar to other solid tumors. 43 Considering the importance of angiogenesis in tumor progression, antiangiogenic treatment modalities are thought as potential tools to fight against tumors. Utilization of antiangiogenic tools in in vitro and in vivo models demonstrated their power in suppression of tumor development, which suggested angiogenesis as a prominent target for HNSCC treatment. 44 The hypoxic microenvironment, as an artifact of insufficiency of angiogenesis due to faster tumor cell proliferation compared to endothelial cells, is one of the major hallmarks of tumors and has been shown to cause an increase in exosomal release from tumor cells, which is essentially associated with the regulation of angiogenesis. 9 Exosomes are key mediators of cell to cell communication within the tumor microenvironment via their enclosed bioactive molecule contents. 44 Exosomes released from cancer cells are internalized by the recipient stromal cells and influence the behavior of the surrounding cells. 44 Recent reports demonstrated that tumor-derived exosomes are largely involved in tumor progression via paracrine signaling to non-cancer cells, which ultimately set the stage for premetastatic niche formation. 45 As a featured example of this is the promotion of tumor angiogenesis by delivery of functional micro-RNAs along with many other bioactive molecules to endothelial cells within the tumor microenvironment, where they regulate key cellular processes. 45 MicroRNAs are a class of well-investigated exosome-transported molecules in pathological processes of distinct cancer types. 46 Thereby, comprehensive understanding of the functions of tumor-derived exosome-shuttled microRNAs on the recipient endothelial cells during preparation of premetastatic niche formation is necessary to develop putative targeted anti-angiogenic therapies. 47 In addition, there has been a wide variety of evidence demonstrating that exosomes released by cancer cells under hypoxic conditions contribute to tumor progression. 10 However, the relationship between hypoxic exosomes and tumor progression has not been completely illuminated at the cellular and molecular level in OSCC.
Recent evidences demonstrated that miR-1825 is a cancer-promoting microRNA in a variety of cancer types. 7,8,16 However, the molecular mechanism by which miR-1825 regulates OSCC progression and its role during angiogenesis has not been revealed yet. Considering the fact that miR-1825 is localized at 20q11.21 chromosomal region, where has been reported as a recurrent gain of function abnormality in human embryonic stem cells and induced pluripotent stem cells, 21,22 and that 20q11.21 amplification in human embryonic stem cells resulted in acquisition of a gene-expression signature enriched for cancer-associated genes, 24 we hypothesized that miR-1825 may be enriched in cells exposed to hypoxic stress conditions, during which cancer stemness is promoted, as well as in hiTDEs. In line with this hypothesis, we demonstrated that hypoxia-induced OSCC cell-derived exosomes are highly enriched with miR-1825, which is effectively transported to HUVEC cells, hence giving rise to enhanced in vitro angiogenesis in OSCC.
Tumor-derived microRNAs can be transferred between cancer cells and their stromal cells and can effectively inhibit the expression of their targets in both cell types. 48 For instance, miR-182-5p within hypoxiainduced glioblastoma-secreted exosomes were found to be carried to HUVEC cells leading to increased angiogenesis and thus enhanced glioma progression. 48 Exosome-derived miR-25-3p was reported to be delivered to endothelial cells, induce angiogenesis and vascular permeability by targeting Krüppel-like factor 2 (KLF2) and KLF4 and bring about the formation of a pre-metastatic niche. 20 Besides, recent studies analyzing the microRNA expressions in different cell lines and in the exosomes derived from those cells found that the microRNA content of the exosomes reflect the cellular microRNA content. The increased expression of exosome-mediated microRNAs under hypoxic conditions, therefore, can be associated with their increased expression in the secreting cells. 11 Li and colleagues demonstrated that hypoxia-induced miR-21 levels in either OSCC cells or OSCC-derived exosomes promote premetastatic behaviors. 10 Collectively, these studies demonstrate that the transport of exosomal microRNAs between cancer cells and endothelial cells are significant to cancer progression. 20 Our results demonstrated that OSCC-released miR-1825 were efficiently carried to endothelial cells via exosomes and contributed to angiogenesis. In addition, our results suggested that miR-1825 in vascular endothelial cells promoted angiogenesis by deregulating TSC2/mTOR axis, which was reported to be linked to angiogenic processes. 49 Physiological TSC/mTOR axis in endothelial cells was demonstrated to be important for vasculature and embryogenesis. 49 In addition, many cancer-inducing kinases have been characterized as modulators of mTOR activity via inactivation of the TSC complex. 50 TCS2 (also known as tuberin) have been identified as an upstream negative regulator of mTOR and reported to mediate vital cellular processes including cell proliferation and angiogenesis. [51][52][53] It acts as a tumor suppressor gene and has been found to be abnormally expressed in different types of cancers. 54 mTOR was found to be involved in an early response to hypoxia, during which it induces endothelial cell proliferation mainly through Akt signaling. 55 Collectively, we suggest a new role for miR-1825 during cancer-induced angiogenesis and provide new insights into the angiogenic processes carried out under hypoxic conditions. Our data illustrated that exosomeshuttled miR-1825 derived from hypoxic OSCC cells clearly increased angiogenesis via targeting TSC2/mTOR axis. Therefore, we propose a new axis as a prominent target for HNSCC treatment.