STIM1 promotes angiogenesis by reducing exosomal miR-145 in breast cancer MDA-MB-231 cells

Cancer cells secrete abundant exosomes, and the secretion can be promoted by an increase of intracellular Ca2+. Stromal interaction molecule 1 (STIM1) plays a key role in shaping Ca2+ signals. MicroRNAs (miRNAs) have been reported to be potential therapeutic targets for many diseases, including breast cancer. Recently, we investigated the effect of exosomes from STIM1-knockout breast cancer MDA-MB-231 cells (Exo-STIM1-KO), and from SKF96365-treated MDA-MB-231 cells (Exo-SKF) on angiogenesis in human umbilical vein endothelial cells (HUVECs) and nude mice. The exosomes Exo-STIM1-KO and Exo-SKF inhibited tube formation by HUVECs remarkably. The miR-145 was increased in SKF96365 treated or STIM1-knockout MDA-MB-231 cells, Exo-SKF and Exo-STIM1-KO, and HUVECs treated with Exo-SKF or Exo-STIM1-KO. Moreover, the expressions of insulin receptor substrate 1 (IRS1), which is the target of miR-145, and the downstream proteins such as Akt/mammalian target of rapamycin (mTOR), Raf/extracellular signal regulated-protein kinase (ERK), and p38 were markedly inhibited in HUVECs treated with Exo-SKF or Exo-STIM1-KO. Matrigel plug assay in vivo showed that tumor angiogenesis was suppressed in Exo-STIM1-KO, but promoted when miR-145 antagomir was added. Taken together, our findings suggest that STIM1 promotes angiogenesis by reducing exosomal miR-145 in breast cancer MDA-MB-231 cells.


Introduction
Breast cancer is the most frequently diagnosed cancer and the leading cause of cancer death among females worldwide 1 . Among all breast cancer subtypes, 10-24% of invasive breast cancers are triple-negative showing poor prognosis 2 . Angiogenesis, defined as the formation of new blood vessels from a preexisting vascular network, has been known to play pivotal roles in tumor progression 3 . Tumor angiogenesis sustains enzymatic degradation of the vessel's basement membrane, endothelial cell proliferation, migration, sprouting, branching, and tube formation, therefore providing oxygen and nutrition to tumor cells for proliferation and metastasis.
Exosomes are membrane-derived vesicles of endocytic origin ranging in size from 30 to 100 nm 4 5 . Accumulating evidence suggests that exosomes secreted from cancer cells are involved in tumor growth, tumorigenesis, angiogenesis, tumor immune escape, drug resistance, and metastasis 6 . Functional contents of exosomes such as lipids (ceramide, cholesterol, phosphatidylserine, etc.), proteins (adhesion molecules, MHC class II, tetraspanin, etc.), and nucleic acids (DNA, mRNA, miRNA, etc.), could trigger specific intracellular cascades and affect the gene expression of the recipient cells 6 . miRNAs are a type of endogenous and small noncoding RNAs (~20-25 nt) that play important roles in cancer development as oncogenes or tumor suppressors 7 .

. Exosomes are secreted by most cells such as epithelial cells, endothelial cells, mast cells, stem cells, T cells, B cells, dendritic cells, and cancer cells
Intracellular calcium ions (Ca 2+ ) are universal second messengers that are intimately related to a number of diverse cellular processes, including cell differentiation, proliferation, and apoptosis 8 . Store-operated Ca 2+ entry (SOCE) is the predominant Ca 2+ entry mechanism ubiquitous in different cell types, and stromal interaction molecule 1 (STIM1) and Orai1 (also named CRACM1) are responsible for SOCE. STIM1, a transmembrane protein located in the endoplasmic reticulum (ER), activates Ca 2+ influx through plasma membrane Ca 2+ channels under stimuli, triggering a transient depletion of the intraluminal Ca 2+ 9 . Ca 2+ store depletion leads to a rapid translocation of STIM1 into puncta that accumulate near the plasma membrane, and STIM1 operates via interaction with Orai1 and regulates the SOCE 10 .
The effect of Ca 2+ level change on tumor angiogenesis through regulating exosome secretion has not been reported. In the present study, we used A23187 (calcium ionophore) or SKF96365 (a pharmacological storeoperated Ca 2+ influx inhibitor) to change the level of intracellular Ca 2+ , and examined the effect of exosomes from the treated MDA-MB-231 cells (Exo-A23187 or Exo-SKF) on angiogenesis of HUVECs. Furthermore, we investigated the effect of exosomes from STIM1-knockout MDA-MB-231 cells (Exo-STIM1-KO) on angiogenesis of HUVECs in vitro and Matrigel plug assay in vivo.

Exosomes from MDA-MB-231 breast cancer cells promote angiogenesis
The exosome is a major player in cell-cell communication, in which many materials such as miRNA, RNA, and proteins can be effectively transferred from the donor cells to the recipient cells 11 . We selected the triplenegative breast cancer MDA-MB-231 cells to study the effect of exosomes from cancer cells on angiogenesis in HUVECs. The typical cup-shaped morphology with a size range of 30-100 nm was observed under a transmission electron microscope (Fig. 1A). CD63, CD81, and HSP70, all of which are known as exosome markers 12 , were observed in MDA-MB-231 cells, and these MDA-MB-231 cells-derived exosomes (Exo) (Fig. 1B).
To study the biological function of exosomes from breast cancer on endothelial cells, we treated HUVECs with Exo. When Exo labeled with PKH26 (red) was incubated with HUVECs, the effective uptake of Exo by HUVECs was observed by electron microscopy (Fig. 1C). Next, we determined the effect of Exo on cell proliferation of MDA-MB-231 and HUVECs. Figure 1D, E showed that Exo increased proliferation of HUVECs in a concentration-dependent manner, but without effect on that of MDA-MB-231, after co-incubation for 24 h. These results revealed that MDA-MB-231-derived self exosomes (Exo) did not affect the proliferation of MDA-MB-231 cells, but increased the proliferation of endothelial cells. It is known that tumor growth and metastasis need glorious angiogenesis for nutrition provision. Therefore, we investigated the effects of Exo on HUVECs angiogenesis by measuring tube formation on Matrigel, after treatment for 24 h. Compared to the control group, Exo obviously increased HUVECs tube formation (Fig. 1F, G). The optimal Exo concentration of 10 μg/ml for angiogenesis was then used in subsequent experiments.
Exo-A23187 promotes HUVEC proliferation, tube formation, and migration Tumorigenic pathways are associated with abnormal activation of Ca 2+ channels 13 , and the increase of intracellular Ca 2+ stimulates exosome secretion 14 . A23187 is a calcium ionophore which can be used to increase intracellular Ca 2+ level in tumor cells. Because prolonged intracellular elevation of Ca 2+ might result in cell death, we first examined the effect of A23187 on cell viability, apoptosis, and cell cycle. As a result, A23187 reduced cell viability, induced apoptosis, and G0/G1 cell cycle arrest at high concentrations, and elevated intracellular Ca 2+ level in a concentration-dependent manner (Supplementary Fig. S1a-f). We also found that A23187 promoted the migration of MDA-MB-231 cells at non-cytotoxic concentrations of <200 nM ( Supplementary Fig. S1g, h). These results are consistent with that prolonged intracellular elevation of Ca 2+ can be cytotoxic, and intracellular Ca 2+ alteration is involved in tumor progression and metastasis 15 . We chose the concentration of A23187 as 500 nM, and treating time as 6 h for the following experiments. The effect of A23187 on intracellular Ca 2+ level in MDA-MB-231 cells was determined with Fluo-Forte Calcium Assay Kit. Figure 2A showed that treatment of MDA-MB-231 with A23187 in the presence of extracellular Ca 2+ led to an increase of intracellular Ca 2+ .
To determine the effect of intracellular calcium increase on the release of exosomes, MDA-MB-231 cells were treated with A23187. The characterized exosomes were found by using transmission electron microscopy in Exo-A23187 with a uniformly cup-shaped morphology within 30-100 nm as diameter (Fig. 2B). A23187-treated MDA-MB-231 cells showed a significant increase in the number of exosomes compared to normal MDA-MB-231 cells (Fig. 2C, D), and HUVECs also exhibited an efficient uptake of Exo-A23187 (Fig. 2E). Next, we determined the effects of Exo and Exo-A23187 on in vitro angiogenesis of HUVECs. In the tube-formation assay, HUVECs were incubated with equivalent concentration (10 μg/ml) of Exo and Exo-A23187 for 24 h and seeded onto matrigel. The results showed that the Exo-A23187 significantly increased tube formation by HUVECs compared to the Exo-treated group, with the tube length elevated by 197 ± 8% (Fig. 2F, G). Cell migration was assessed by the scratch wound-healing assay. The migration ability of HUVECs was enhanced in the Exo group compared with the control, and the Exo-A23187 group was further enhanced compared with Exo by 28 ± 4% (Fig. 2H, I). These results demonstrated that A23187 elevated intracellular calcium and therefore increased the release of exosomes from MDA-MB-231 cells, resulting in an increase of HUVEC tube formation and migration.

Exo-SKF suppresses HUVEC tube formation and migration
Inhibition of tumor angiogenesis might halt tumor progression. The above results showed that exosomes from breast cancer cells with higher intracellular Ca 2+ promoted HUVEC angiogenesis. We hypothesized that exosomes from cancer cells with lower intracellular Ca 2+ might exhibit an anti-angiogenic effect. Then we used SKF96365, a pharmacological store-operated Ca 2+ influx inhibitor, to reduce intracellular Ca 2+ level. We found that SKF96365 reduced the cell viability and augmented cell apoptosis at concentrations of 20 μM and 50 μM ( Supplementary Fig. S2a-c). Therefore, we chose the noncytotoxic but effective concentration of 10 μM for the following experiments. We first examined the inhibitory effect of SKF96365 on Ca 2+ entry in MDA-MB-231 cells. MDA-MB-231 were stained with FluoForte TM dyeloading solution. The subsequent introduction of CaCl 2 to the extracellular solution resulted in the elevation of intracellular Ca 2+ from baseline, and pretreatment of SKF96365 reduced the intracellular level of Ca 2+ (Fig.  3A). Afterward, we isolated Exo-SKF, and examined the Exo-SKF morphology and the exosome-specific marker proteins by using electron microscopy and western blot (Fig. 3B, C). Treatment of SKF96365 did not significantly change the release of exosomes from MDA-MB-231 cells (Fig. 3D). We also visualized the transport of Exo-SKF into HUVECs. PKH26 (red)-labeled Exo-SKF was incubated with HUVECs after staining the cells with Hoechst, and the PKH26 signals and Hoechst signals were detected under a fluorescent microscope. We found that almost all the recipient HUVECs revealed red signal, suggesting that the uptake of Exo-SKF by HUVECs was efficient (Fig. 3E). And the treatment of Exo-SKF significantly suppressed the tube-formation ability of HUVECs compared with Exo ( Fig. 3F, G). In addition, the migration ability of HUVECs was suppressed in the Exo-SKF group compared with Exo treatment only (Fig. 3H, I). These results demonstrated that intracellular Ca 2+ regulated the release of exosomes and affected tumor angiogenesis.

Exo-SKF affects the levels of miR-145 and miR-449
Exosomes released from tumor cells contain miRNAs which could even affect the environment surrounding the H The migration of HUVECs was increased in the Exo-A23187 group compared with the control and Exo group. I Quantification of the HUVEC migration results in (h). Data represent mean ± SEM from three independent experiments (n = 3). *P < 0.05, **P < 0.01.
Exo-SKF suppresses insulin receptor substrate 1 (IRS1) signaling pathway miRNAs regulate post-transcriptional expression by gene silencing. As the predicted target of miR-145 26 , IRS1 activates phosphatidylinositol 3-kinase (PI3K)/Akt pathway, and mitogen-activated protein kinase (MAPK) pathway to promote endothelial cell proliferation, migration, as well as vascular permeability 27 . Thus, we investigated the effect on IRS1 signaling pathways by evaluating the changes of c-Raf, extracellular signal regulated-protein kinase (ERK), p38, Akt, and mammalian target of rapamycin (mTOR). HUVECs were treated with Exo or Exo-SKF for 24 h, and cells were lysed for western blot. As shown in Fig. 3M, N, we observed upregulation of IRS1 and its downstream signal proteins such as c-Raf, ERK, p38, Akt, and mTOR phosphorylation in HUVECs treated with Exo, but downregulation in those treated with Exo-SKF.
Exo-STIM1-KO suppresses HUVEC tube formation STIM1 plays an essential role in Ca 2+ mobilization and signaling 28 . As the ER Ca 2+ sensor, STIM1 controls plasma membrane Ca 2+ channels to regulate Ca 2+ entry. After MDA-MB-231 cells were treated with SKF96365 for 24 h, the expression of STIM1 and phosphorylation of ERK were downregulated (Fig. 4A). To investigate the role of STIM1 in the reduction of angiogenesis by Exo-SKF, we edited exon 4 of STIM1 locus (ENSG00000167323) in MDA-MB-231 cells by CRISPR/Cas9 gene editing to generate the Exo-STIM-KO, and examined the antiangiogenic effects on HUVECs. A schematic diagram of gRNA-targeting exon 4 of the STIM1 gene is shown in Supplementary Fig. S3a. We first confirmed the sequence of exon 4 in STIM1 in MDA-MB-231 cells with Sanger sequencing, which revealed that MDA-MB-231 cells carried 1-bp deletion at the gRNA-targeting region ( Supplementary Fig. S3b). After Sanger sequencing validation, we obtained STIM1-knockout MDA-MB-231 cells (STIM1-KO). To further determine the effectiveness of CRISPR/Cas9, a western blot was carried out to detect STIM1 expression. STIM1-CRISPR/Cas9 gene-edited MDA-MB-231 cells showed no expression of STIM1 (Fig. 4B). The lack of STIM1 led to a significant drop in calcium amount, which was assessed in the presence of additional Ca 2+ (Fig. 4C). The characteristic morphology of Exo-STIM1-KO was observed under an electron microscope (Fig. 4D), and both CD63 and CD81 were found in Exo-STIM1-KO (Fig. 4E). An equal number of MDA-MB-231 and STIM1-deficient MDA-MB-231 cells were incubated in culture dishes, and then the exosomes were isolated respectively. The amount of Exo-STIM1-KO did not significantly change compared with Exo (Fig.  4F). Exo-STIM1-KO was also labeled with PKH26 red dye, and the uptake by HUVECs was observed (Fig. 4G). We then examined whether Exo-STIM1-KO could affect angiogenesis. HUVECs were incubated with Exo or Exo-STIM1-KO for 24 h. As expected, Exo-STIM1-KO repressed the tube formation of HUVECs compared to Exo treatment only (Fig. 4H, I). Therefore, our results suggest that STIM1-deficient MDA-MB-231 cells inhibit HUVEC angiogenesis through the exosomes.
Then HUVECs were treated with different exosomes for 24 h, and the tube formation and cell migration abilities were investigated. Upregulation of tube formation by Exo was suppressed by Exo-STIM1-KO, and Exo-STIM1-KO plus miR-145 antagomir reversed this inhibition (Fig. 5E,  F). As shown in Fig. 5G, H, cell migration was significantly suppressed in the Exo-STIM1-KO group in comparison with the Exo group. Conversely, Exo-STIM1-KO plus miR-145 antagomir group enhanced the migration ability of HUVECs. These results suggested that miR-145 plays an important role in endothelial cell migration mediated by Exo-STIM1-KO.
Next, we examined the expression of IRS1 pathway proteins in HUVECs after incubation with different exosomes for 24 h. miR-145 antagomir upregulated the expression of IRS1, and activated the IRS1-dependent Akt/mTOR, Raf/ERK, and p38 MAPK. In addition, Exo-STIM1-KO plus miR-145 antagomir obviously elevated the expression of IRS1 and activated its downstream molecules which had been suppressed by Exo-STIM1-KO (Fig. 5I, J).

Exo-STIM1-KO prevents breast cancer angiogenesis in vivo
To further determine the therapeutic potential of Exo-STIM1-KO in vivo, we performed an in vivo Matrigel plug assay to detect the newly formed blood vessels in the transplanted gel plugs in BALB/c nude mice (Fig. 6A). The hemoglobin content, which represents new vessel formation, was significantly increased in Exo-treated mice compared with the control group but was reduced in the Exo-STIM1-KO group than the Exo-treated group. In contrast, the hemoglobin concentration in the plugs containing Exo-STIM1-KO plus miR-145 antagomir was higher than the Exo-STIM1-KO group (Fig. 6B, C). H&E staining assay revealed that Exo dramatically increased plug vascularization compared to control, and Exo-STIM-KO suppressed plug neovessel formation. However, Exo-STIM-KO plus miR-145 antagomir reversed the Exo-STIM-KO induced angiogenesis attenuation (Fig. 6D). The blood vessel density which represents the relative tube length of the neovessels in Matrigel plugs was quantified (Fig. 6E). Moreover, immunohistochemistry analysis confirmed that the CD31 (a marker of the formation of new vessels) positive cells in Exo-STIM1-KOtreated mice were significantly fewer than the Exo-treated mice, but increased when miR-145 antagomir was added (Fig. 6F, G). In addition, the expression of IRS1 was also lower in the Exo-STIM1-KO-treated mice compared with the Exo group but was upregulated in Exo-STIM1-KO plus miR-145 antagomir-treated mice (Fig. 6H, I).

Discussion
Exosomes are known as potent cell-cell messenger. Various cell types have the ability to release exosomes, including immune cells, epithelial cells, and tumor cells. The mechanisms involved in the entry of exosomes in recipient cells include three main routes, such as receptor/ligand signaling, fusion, phagocytosis/endocytosis 33 . In this study, we found that PKH26-labeled Exo, Exo-A23187, Exo-SKF, and Exo-STIM1-KO from MDA-MB-231 cells (donor cells) transferred to the HUVECs (recipient cells). Exosomes contribute to carcinogenesis by the transport of oncogenic lipid, proteins, and nucleic acids, and the other bioactive molecules 34 . In this study, we found that MDA-MB-231-derived exosomes promoted tube formation in HUVECs.
Angiogenesis is a fundamental process involved in carcinogenesis. Steps toward angiogenesis include endothelial cell migration and proliferation, vascular tube formation, anastomosis of newly formed tubes 35 . An increase of intracellular Ca 2+ stimulates exosome secretion 14 . However, the effects of exosomes from the breast cancer cells with elevated or decreased Ca 2+ conditions on angiogenesis have not been reported yet. Therefore, we treated MDA-MB-231 cells with calcium ionophore or inhibitor of store-operated Ca 2+ channel, and isolated exosomes to investigate their effects on angiogenesis in HUVECs. Our results showed that treatment of MDA-MB-231 cells with A23187 or SKF96365 in the presence of extracellular Ca 2+ led to an increase or decrease of intracellular Ca 2+ , respectively. We also found that Exo-A23187 promoted HUVECs migration and tube formation, and Exo-SKF inhibited them. These results suggested that change of Ca 2+ level influenced exosome release from triple-negative breast cancer MDA-MB-231 cells, and therefore affected angiogenesis of HUVECs.
STIM1 is a critical regulator of Ca 2+ mobilization. Once ER Ca 2+ is depleted, the ER Ca 2+ sensor STIM1 accumulates in a junctional ER in close apposition to the plasma membrane to activate the plasma membrane poreforming unit Orai1, which induces SOCE 36 . Therefore, we deleted the STIM1 gene by CSISPR/Cas9, and found the intracellular Ca 2+ was decreased in the STIM1-silencing MDA-MB-231 cells, the characteristic morphology and markers of exosomes were also observed. The exosome amounts did not change obviously.
miRNAs are highly conserved small noncoding RNAs that induce post-transcriptional gene silencing canonically by binding mRNA. miRNA/mRNA interaction hinders protein synthesis and initiates mRNA degradation 37 . Some miRNAs are known to be involved in tumorigenesis and tumor angiogenesis. For example, miR-34a is a tumor suppressor that is frequently downregulated in a number of tumor types, and ectopic expression of miR-34a in head and neck squamous cell carcinoma (HNSCC) cell lines inhibited tumor growth and tumor angiogenesis in vivo when introduced into SCID mouse xenograft models 38 ; miR-519c is a pivotal regulator of tumor angiogenesis 39 ; modulation of miR-126 expression can disrupt angiogenesis and vascular integrity 40 . Some other miRNAs such as miR-21, miR-221, and miR-222 are highly expressed in the endothelial cells 41 . In the present study, we demonstrated that level of miR-145 increased in MDA-MB-231 cells treated with SKF, STIM1-KO, Exo-SKF, and Exo-STIM1-KO, as well as in HUVECs treated with Exo-SKF and Exo-STIM1-KO. It was reported that miR-145 is downregulated in breast cancer cancer 42 . These results support that Ca 2+ sensor STIM1 might influence tumor angiogenesis through regulating exosomal miR-145.
The pathway of insulin-like growth factor I (IGF-I)/ IRS1/Ras mediates vascular endothelial growth factor (VEGF) production, which is crucial in angiogenesis. IR is expressed in endothelial cells as well as in cancer cells 43 . Insulin stimulation increases the tyrosine phosphorylation of IRS1 and IRS2 44 . miR-145 has been repeatedly reported to be a tumor suppressor, and to target gene IRS1 and therefore inhibit its protein expression 29 . Through basepairing complement, we also confirmed that IRS1 is the potential target of miR-145. As expected, Exo promoted the IRS1 expression in HUVECs, which was attenuated by treatment with Exo-SKF or Exo-STIM1-KO. Tyrosin- Data represent mean ± SEM from three independent experiments (n = 3). *P < 0.05, **P < 0.01, and ***P < 0.001. phosphorylated IRS1 activates PI3K/Akt and Ras/Raf/ MAPK pathways 45,46 , which mediate endothelial cell proliferation, migration, survival, and vascular permeability 47,48 . In this study, we found that the phosphorylation of Raf, ERK, p38, Akt, and mTOR in HUVECs was increased by treatment with Exo, but was attenuated by concomitant treatment with SKF (Exo-SKF) and STIM1-KO (Exo-STIM1-KO). These results suggest that miR-145 might target IRS1 and therefore inhibit the angiogenesis of HUVECs via regulating IRS1/PI3K/Akt/mTOR and IRS1/Raf/ERK pathways.
We next used miR-145 antagomir to demonstrate whether Exo-STIM1-KO miR-145 could reduce angiogenesis through directly repressing IRS1. The level of miR-145 in Exo-STIM1-KO plus miR-145 antagomir group is lower than that in the Exo-STIM1-KO group. Suppression of miR-145 in recipient HUVECs by miR-145 antagomir transfection abolished the anti-angiogenic effect of Exo-STIM1-KO in HUVECs, suggesting that miR-145 might play pivotal roles in the Exo-STIM1-KO attenuated tumor angiogenesis effect.
The in vivo Matrigel plug assay is widely used to evaluate the in vivo angiogenic potential 49 . To date, there has been no report on the anti-angiogenic effects of exosomes derived from STIM1-KO-tumor cells. By using in vivo Matrigel plug assay, we found that Exo-STIM1-KO significantly inhibited Exo-induced angiogenesis. Consistently, immunohistochemical analysis showed a reduction of expression of IRS1 and angiogenesis biomarker CD31, which was reversed by the addition of miR-145 antagomir. These results confirmed that the exosomal miR-145 from STIM-KO-MDA-MB-231 cells attenuated angiogenesis by targeting IRS1.

Cell proliferation assay
Cell proliferation was assessed using an MTT assay as we previously reported 50 with a small modification. Briefly, cells were seeded onto 96-well plates and cultured with exosomes from MDA-MB-231 cells for 24 h, and then MTT was added to each well. After 4 h of incubation, the produced formazan was dissolved in DMSO, and OD at 490 nm was monitored using microplate reader iMark (BIO-RAD, Hercules, CA, USA).

Isolation of exosomes
MDA-MB-231 cells were seeded at a density of 2 × 10 6 cells/100-mm dish and cultured for 6 h or 24 h without or with the treatment of A23187 (500 nM) or SKF96365 (10 μM) in serum-free medium. Exo, Exo-A23187, Exo-SKF96365, and Exo-STIM1-KO in cell culture supernatants were isolated by use of total exosome isolation reagents as described by the manufacturer. Exosome protein amount was quantified using the Pierce TM BCA protein assay kit (23225; Thermo Fisher Scientific, Waltham, MA, USA).

Wound-healing assay
HUVECs were cultured on 24-well plates (2 × 10 5 cells/ well) in DMEM medium. Wounds were made using a sterile 10-µl pipette tip. After the cellular debris was removed by gently washing with PBS, the cells were cultured in high glucose serum-free DMEM with various exosomes at 37°C with 5% CO 2 . At least four images of the scraped area were captured using phase-contrast microscopy after 24 h treatment, and cell migration distances were determined using ImageJ software.

Tube-formation assay
HUVECs (1 × 10 5 cells/well) were treated with exosomes derived from MDA-MB-231 cells after various treatments for 24 h and then plated on Matrigel TM Matrix (50 μl/well) (BD Biosciences, San Josè, CA, USA) in 96well plates for 6 h at 37°C in a 5% CO 2 humidified incubator. Tube formation was observed and imaged with phase-contrast microscopy. For quantification, total tubular length and branch points per well were determined using ImageJ software.

Protein extraction and western blot
Western blot analysis was carried out as we previously reported 51 with a small modification. Cells were collected, and the protein concentration of each sample was determined by the BCA protein assay kit. Equal amount of protein was separated by sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) and transferred to the PVDF membrane. After being blocked with 5% skim milk, the membranes were incubated with each primary antibody, and then the horseradish peroxidase-conjugated secondary antibody. The signals were detected with ChemiDoc TM XRS + System (BIO-RAD, Hercules, CA, USA) after exposure to ECL reagent.

Fluorescent imaging of exosome uptake
Freshly isolated exosomes from MDA-MB-231 cells were labeled with the PKH26 red fluorescent cell linker kit (Sigma-Aldrich, St. Louis, MO, USA) according to the manufacturer's instructions with a small modification, and then cultured with HUVEC for 3 h. After being washed, HUVECs were stained with Hoechst for 15 min. The pictures were taken and the uptake of exosomes was observed using a fluorescence confocal microscope.

Generation of genetically modified cells using CRISPR/Cas9 genome editing
The knockout of STIM1 in MDA-MB-231 cell lines was achieved by the CRISPR/Cas9 system. Oligonucleotide targeting exon 4 (5'-ATACAATTGGACCGTGGATG-3') was designed as the CRISPR target site. MDA-MB-231 cells were transfected with the plasmid of the STIM1-targeted gRNA encoding SpCas9 by Lipofectamine 3000. Following 72 h of puromycin selection, the cell culture was extended for 96 h without puromycin, and the viable clonal cells were subcultured in a 96-well plate with a density of one cell/well. Afterward, cells were cultured for 7-10 days. Individual clones were expanded and screened for STIM1 depletion by genomic DNA sequencing and immunoblotting. miRNA isolation and real-time quantitative reverse transcription-PCR (RT-qPCR) assay The total RNA from the cells and exosomes was isolated using the TRIzol reagent (Life Technologies, Carlsbad, CA, USA) or E.Z.N.A. TM miRNA Kit (Omega Bio-Tek, Norcross, GA, USA), respectively. The cDNA synthesis was performed using M-MLV reverse transcriptase. Realtime PCR was performed using the miScript SYBR Green PCR Kit on a CFX96 TM Real-Time PCR Detection System (BIO-RAD, Hercules, CA, USA). The sequences of several miRNA primers are described in Table 1. The expression levels of U6 were used as an endogenous control for each sample. The relative gene expression levels were calculated using the comparative Ct (ΔΔCt) method.

miRNA antagomir transfection
HUVECs were cultured on 6-well plates (1 × 10 5 cells/ well), and transfected with miR-145 antagomir or antagomir negative control (NC) using Lipofectamine 6000 according to the manufacturer's instructions. Simultaneously, Exo-STIM1-KO was added. After 24 h, cells were washed twice by PBS to remove remaining exosomes or antagomir and collected to extract proteins and RNA. Chemically modified miR-145 antagomir was purchased from Shanghai GenePharma Co., Ltd.

Matrigel plug assay for angiogenesis in nude mice
All animal experiments were conducted at Laboratory Animal Center, Institute of Radiation Medicine, Chinese Academy of Medical Sciences in accordance with the Institutional Animal Care and Use Committee guidelines. Mice were randomly assigned to four groups with three mice in each group. Briefly, 2 × 10 6 HUVECs were mixed with Exo, Exo-STIM1-KO, miR-145 antagomir, or antagomir NC, respectively. Then Lipofectamine 6000 was added. The cell suspensions were mixed with 400 μl of High Concentration Matrigel TM Matrix at a ratio of 1:4, and then the mixture was subcutaneously injected into the dorsal region of nude mice (female, 6-week-old BALB/c). After 10 days, the Matrigel plugs were harvested and processed for analysis. The degree of vascularization was evaluated by determining hemoglobin content. On the other hand, plugs were fixed with 4% formaldehyde and embedded in paraffin, and stained with hematoxylin and eosin. The vessel area was quantified as mean relative tube length by image analysis of six random microscopic fields using ImageJ software.

Statistical analysis
All values are expressed as means ± SEM of triplicate values. One-way ANOVA followed by Tukey's Multiple Comparison Test was utilized to determine the statistical significance with GraphPad Prism 5 (GraphPad, San Diego, CA, USA). Differences were considered statistically significant when P < 0.05.