Non-small lung cancer cells-derived exosome miR-3157-3p primes pre-metastatic niche by inducing angiogenesis and vascular permeability

Background: Cancer-derived exosomes are considered to be the main driving force for cancer-induced metastasis niche formation at foreign sites, but the mechanism is unclear. Methods: Identication of differentially expressed miRNAs in exosomes by miRNA sequencing. Through a series of in vitro and in vivo experiments, it was studied that the tumor-derived exosome miR-3157-3p can be taken up by the endothelium and exert biological functions. The molecular mechanism of miR-3157-3p was conrmed by Western blot, luciferase test and rescue experiment. Finally, real-time uorescence quantitative PCR and RNA in situ hybridization were used to evaluate the clinical relevance of miR-3157-3p expression. Result: In metastatic NSCLC patients, the expression level of miR-3157-3p in circulating exosomes was signicantly higher than that of non-metastatic NSCLC patients. Here, we found that miR-3157-3p can be transferred from NSCLC cells to vascular endothelial cells through exosomes. Exosome miR-3157-3p has regulated the expression of VEGF / MMP2 / MMP9 and occludin in endothelial cells by targeting TIMP / KLF2, thereby promoted angiogenesis and increased vascular permeability. In addition, exosome miR-3157-3p from NSCLC cells enhances NSCLC metastasis in mouse lungs. Conclusion: Our work indicates that exosome miR-3157-3p is involved in the formation of pre-metastatic niche and may be used as a blood-based biomarker for NSCLC metastasis.


Introduction
The pre-metastatic niche is a pre-formed microenvironment formed by exosomes secreted by the primary tumor site before extensive metastasis [1][2][3]. It is characterized by angiogenesis, immunosuppression, in ammation, lymphangiogenesis, vascular permeability, organotropism, and reprogramming [4][5][6][7]. In addition, targeted therapy before metastasis niche may be a potential strategy for the treatment of cancer metastasis [8,9]. It has been found that several pre-metastatic niche biomarkers are important for the diagnosis and prognosis prediction of cancer [10,11].
Exosomes are extracellular vesicles with a diameter of 30-150 extranm [12][13][14]. As a key participant in cell-to-cell communication, exosomes convey information through their cargo levels, including proteins, lipids, DNA, messenger RNA, and microRNA [15][16][17][18][19][20]. Recently, it has been found that the pre-metastatic niche depends on tumor-derived exosomes [5]. The function of exosomes depends on the cell type from which they are derived [21,22]. Studies have shown that tumor-derived exosomes participate in the exchange of genetic information between tumor cells and basal cells, resulting in the generation of a large number of new blood vessels, thereby promoting tumor growth and invasion [23,24]. Various researches have told us that exosomes have played a key role in cancer tumorigenesis, growth, apoptosis, immune response and chemotherapy resistance [25][26][27][28][29]. We have known that the NSCLC-derived exosome miR-3157-3p can be transferred to vascular endothelial cells to target TIMP2 / KLF2, by which to promote the angiogenesis and increase the permeability. In addition, we have demonstrated that exosome miR-3157-3p mediates the pre-metastatic niche before transfer in nude mice by inducing angiogenesis, thereby promoting NSCLC transfer.

Cell lines and human tissue samples
We have collected tumor and paraneoplastic specimens of 47 patients who underwent radical surgery between 2015 and 2019. All patients did not undergo radiotherapy and chemotherapy. Fresh biopsy tissue is frozen and stored in liquid nitrogen until being used. All plasma samples were collected from NSCLC patients (40 without metastasis and 40 with metastasis )in our hospital between April 2013 and June 2019, and age-matched healthy human plasma samples (n = 30) were selected from the Physical Examination Center of our hospital. The blood sample was then centrifuged at 2500g for 10 minutes to extract the serum and stored at -80 ° C. Each patient has signed an informed consent form and this was approved by the Ethics Committee of the First A liated Hospital of Nanjing Medical University.

Microarray analysis
The miRNA 4.0 Array of Affymetrix was used to analyze the miRNA expression in serum exosomes (normal) of healthy people and in NSCLC patients with lymph node metastasis. What we have done has obtained the approval from the First A liated Hospital of Nanjing Medical University (Nanjing, China).
Informed consent was signed by each patient on the day of admission. Three clinical samples were collected from each group. The miRNA array experiment was conducted in the microarray core laboratory of Beijing Boao Jingdian Biotechnology Co., Ltd. (Beijing, China).

Cell transfection
Lentiviral vectors that over-express miR-3157-3p and repress miR-3157-3p were constructed and generated. Mimics and inhibitor were purchased from GiKai GENE (Shanghai, China) and transfected according to the manufacturer's instructions. Plasmid TIMP2 / KLF2, empty vector, si-TIMP2 / si-KLF2 were purchased from Gene Pharma (Shanghai, China). Transfection was performed using elipofectamine 3000 reagent (Invitrogen) according to the manufacturer's instructions.

PCR
To determine the expression level of miR-3157-3p, total RNA was isolated and extracted from tissues and cells by TRIzol reagent (Invitrogen, CA, USA) according to the instructions. CDNA was generated by reverse transcription using total RNA and PrimeScriptRT reagent (Takara, Kusatsu, Japan), and detected using SYBR Green (Takara) at ABI StepOnePlus real time quantitative PCR instrument (StepOnePlus, ABI Company, Oyster Bay, NY, USA). Taking GAPDH and U6 as endogenous controls, mRNA and miRNA were normalized. The 2−ΔΔCT method was used to quantify the relative levels of mir-3157-3p and TIMP2/KLF2. Each sample is in triplicate. miRNA quanti cation: Bulge-loopTM miRNA qRT-PCR Primer Sets (one RT primer and a pair of qPCR primers for each set) speci c for miR-3157-3p is designed by RiboBio (Guangzhou, China). The oligonucleotides used in this study: U6 F:CTCGCTTCGGCAGCACATATACT.U6 R:ACGCTTCACGAATTTGCGTGTC. GAPDH F: AAGGTGAAGGTCGGAGTCA.GAPDH R:GGAAGATGGTGATGGGATTT Isolation and identi cation of exosomes Exosomes were puri ed from NSCLC cell supernatant or NSCLC patient serum by ultracentrifugation. NSCLC cells were cultured in DMEM medium supplemented with 10% fetal bovine serum without exosomes. Total Exosome Isolation Reagent (from cell culure media, Invitrogen, USA) is used to extract exosomes from cell supernatant, and the amount of exosomes was measured by BCA protein assay kit (KeyGEN BioTECH). For transmission electron microscopy TEM , 5-10u extracted exosomes are placed on the copper carrier grid. Add 10-20μl of EM solvent dropwise, and carefully suck up with clean lter paper after 1min, then observe the exosomes with Philips CM120 biological dual transmission electron microscope (FEI Company, USA). The method of Nanoparticle-tracking analysis is used to analyze the size of exosomes. The exosomes (10-20 mg) in 1 mL PBS and vortex were dissolved for 1min to distribute the exosomes evenly. Then, the NanoSight Nanoparticle Tracking Analyzer (NTA, Malvern Analysis, UK) was used to measure and observe the scale distribution of the exosomes. For exosome labeling, PKH67 membrane dye (Sigma) was used to uorescently label exosomes according to the instructions.

Dual luciferase report
The 3'UTR fragments of KLF2 and KLF4 genes were ampli ed and inserted into the vector. Then complete co-transfection of TIMP2 and KLF2 3'UTR plasmids with miR-3157-3p lentiviral vector into cells by using Lipofectamine 2000 (Invitrogen USA). 48 hours after the transfection, Dual-Lumi ™ dual luciferase reporter gene detection kit (Biyuntian: RG088S) detects dual luciferase. The Renilla activity was taken as a normalization of luciferase activity. All measurements were performed in triplicate, and each experiment was repeated three times.
EDU assay, migration assay, and angiogenesis assay.
From EDU we know that spread the transfected cells evenly in a 24-well plate with 3 × 10 4 cells per well, then use EdU Apollo567 extracorporeal ow cytometry kit (Guangzhou RiboBio) to perform EdU according to the instructions, by which we can acquire images with a uorescence microscope. From recruitment assay, we get the following result:Approximately 2×10 4 cells in serum-free medium(400 μl) were seeded into the upper Transwell chamber for migration assay, who use 8.0μm Transwell Permeable Supports (Corning, New York, USA). The lower chamber was supplemented with 600μl DMEM medium containing 10% FBS, and was incubated at 37 °C for 24 hours. Remaining cells in the upper membrane were cleared with Cotton swabs, and they were stained with paraformaldehyde and 0.1% crystal violet (Beyotime, Shanghai, China) for 20 minutes at 37°C.
Tube formation: Spread 250u Matrigel (BD Bioscience, USA) on a 24-well plate and place in an incubator for half an hour. Approximately 1 × 10 5 treated HUVEC cells were resuspended in each well of a 24-well plate.Tube formation was induced at 37 ° C for 8 h. Then we observe the tube length with a microscope. For the measurement of the aortic ring, the thoracic aorta cut from Sprague Dawley (SD) rats of 10 to 12 weeks old was cut into a cross section of 1 to 2 mm in length. Place the ring on the wells coated with Matrigel, using F12-K medium (GIBCO, Gaithersburg, Maryland, USA), containing 10% fetal bovine serum, supplemented with 10ng / ml vascular endothelial growth factor and 25μg / ml Heparin per well. After 24 hours, the aortic ring was incubated with conditioned medium derived from NSCLC cells or exosomes transfected with miR-3157 mimics, miR-3157 inhibitors. After 5 days, observe the number of buds with a microscope, and quantify blood vessel growth by counting all buds from one ring [30]. From CAM, we get the result: Incubate 20 fertilized eggs at 37.5 ° C and 70% humidity for 8 days, and an arti cial airbag was created, also a small window was cut in the shell above the arti cial airbag. We use exo-miR-3157-3pmimics, exo-NC mixed with equal volume of Matrigel (BD Biosciences, Bedford, MA) to inject into CAM. The area around the implanted Matrigel was photographed, and the number of blood vessels was obtained by counting the branches of the blood vessels [31]. Each experiment was repeated three times. From the in vitro permeability measurement, we get the following result: Rhodamine B isothiocyanatedextran (Sigma) was added to the top well of the transwell lter, and HUVEC (10 5 cells per well) was treated on it for 3 days. Then collect the culture medium in the bottom well after 30 minutes and monitor the appearance of uorescence under 544 nm excitation and 590 nm emission.

Animal experiment
The six-week-old nude mouse was purchased from the Animal Center of Nanjing Medical University (Nanjing, China) and what we have done was approved by the Animal Ethics Committee of Nanjing Medical University. Nude mice were randomly divided into 4 groups, 4 in each group. The 3 × 10 6 A549 cell samples were injected subcutaneously into the groin area of BALB / c nude mice. After 9 days, 5 µg of exosomes were injected into xenografts every other day. Three days after the last injection, the tumor tissue was dissected, xed in 10% formalin, and embedded in para n for further study. For the determination of tail vein metastasis, 6-week-old nude mice were injected with 5 µg exosomes through the tail vein every other day for 2 weeks. The 2 × 10 6 A549 cells were injected into the tail vein of nude mice treated with exosomes. After 60 days, the mice were sacri ced, and the lungs were removed for examination and HE staining.

Statistical analysis
All statistical analyses were performed by spss 19.0 and GraphPad software 8.0. The P-values were analysed with Student's t test, one-way ANOVA. We select for statistical signi cance when P is less than 0.05.
To identify miRNAs potentially associated with metastasis in NSCLC, serum exosomes were extracted from NSCLC patients or healthy controls and identi ed. First, the exosomes are separated by differential ultracentrifugation. The morphology of the exosome was examined by TEM ( Figure 1A). Nanosight particle tracking shows that the size of these exosomes is about 100 nm ( Figure 1B). In addition, Western blot analysis con rmed the expression of speci c exosome markers such as CD63 and TSG101( Figure   1C). Next, exosome RNA sequencing (RNA-seq) was performed, miRNA pro les were compared and shown in Figure 1D. Among miRNAs with different expressions, compared with normal controls, 44 miRNAs with abnormal expression in metastatic NSCLC tissues, including 3 miRNAs (miR-3157-3p, miR-3613-5p, miR-3921)( Figure S1A) increased . We performed GO and KEGG enrichment analysis on the differential miRNAs and found that the differential miRNAs were mainly enriched in the four categories of "Angiogenesis", "cell junction", "cancer" and "positive regulation biological process" (Figure 1E-F, Figure  S1B-C, Figure S2A-B). We further conducted GO-Standard analysis on the expression of differential miRNA in "biological process", "cellar component" and "Molecular function" ( Figure S2C). Because angiogenesis and tight junctions between cells are related to the pre-metastatic microenvironment, we suspect that abnormally expressed miRNAs may be associated with metastasis. RT-PCR analysis con rmed that three of these mRNAs (miR-3157-3p, miR-3613-5p, and miR-3921) showed higher levels of serum exosomes in NSCLC patients than healthy people (10 NSCLC patients and 10 healthy people) ( Figure 1G). We found that miR-3157-3p has the highest expression,so we selected miR-3157-3p for further study. The results of qRT-PCR showed that circulating exosomal-miR-3157-3p in NSCLC patients with metastasis was higher than without metastasis( Figure 1H). What's more interesting is that we found that miR-3157-3p in the tissues of patients with metastasis was signi cantly higher than that of patients without metastasis ( Figure 1I). So we predict that exo-miR-3157-3p is closely associated with metastasis.

NSCLC transfer miR-3157-3p to HUVECs through exosomes
To study the molecular mechanism of NSCLC cells regulating HUVEC, we rst veri ed the level of miR-3157-3p in NSCLC cell lines and 16HBE cells (Figure 2A). The exosomes were successfully isolated from A549 and H1299 serum, and by transmission electron microscopy (TEM) there showed a teacup-like double-sided structure of 30-100nm exosomes, and the next step is the NTA veri cation of exosomes ( Figure 2B). In addition, CD63 and TSG101 are exosome-speci c markers which were used to quantify exosomes, and calnexin is not expressed in exosomes, which was determined by Western blotting. The results showed that CD63 and TSG101 were abundant in the isolated exosome components ( Figure 2C). The extracted exosomes were co-cultured with HUVECs. And RT-qPCR was used to measure the expression of miR-3157-3p. RT-PCR results showed that the expression of miR-3157-3p in HUVEC increased signi cantly in a time-dependent manner ( Figure 2D).A549 and H1299 transfected with Lv-miR-3157-3p and miR-3157-3p from A549/H1299 exosomes were signi cantly up-regulated ( Figure 2E-F).
HUVEC was incubated with PKH67-labeled exosomes derived from A549/H1299 cells, which were transfected with Cy3-labeled miR-3157-3p mimics. Cy3 uorescence and PKH67 lipids were observed under a confocal microscope in the incubated HUVEC.The dye is distributed around the nucleus Figure  2G . Therefore, these ndings concluded that NSCLC transferred miR-3157-3p to HUVEC through exosomes.
Up-regulation of miR-3157-3 promotes proliferation, migration vascular leakinessand angiogenesis of HUVEC To determine whether miR-3157-3p can affect the proliferation, migration and angiogenesis of HUVEC in vitro, Lv-miR-3157-3p or Lv-miR-3157-3p-inhibitor was transfected with HUVEC. The expression e ciency of miR-3157-3p was detected by qRT-PCR.( Figure S3A) Overexpression of miR-3157-3p enhances HUVEC proliferation, migration of vessel penetration and tube formation. In contrast, inhibiting the expression of miR-3157-3p leads to a reduction in HUVEC proliferation, migration, vascular permeability, and tube formation ( Figure 3A, 3B, 3C, 3D). VEGF has played a powerful role in promoting angiogenesis, and MMP-2 and MMP-9 are the two main enzymes that degrade type IV collagen and promote the invasion and migration of malignant tumors. ZO-1, occludin and Claudin5 are essential for the integrity of the endothelial barrier because they can enhance tight junction-related proteins. Western blot analysis showed that the up-regulation of miR-3157-3p increased the protein levels of VEGF, MMP-2 and MMP-9, while decreased the protein levels of ZO-1, occludin and Claudin5. Inhibiting the expression of miR-3157-3p played the opposite role ( Figure 3E, 3F). The expression of ZO-1, occludin and Claudin5 in transfected HUVEC was further veri ed in the IF experiment ( Figure 3G). These ndings indicated that elevated miR-3157-3p promotes HUVEC proliferation, migration, vascular penetration, and angiogenesis in vitro.
To explore how miR-3157-3p regulates angiogenesis and vascular permeability, three mRNA target prediction algorithms (miRDB, miRWalk, and Targetscan) were used to identify potential downstream targets of miR-3157-3p. Among potential targets, TIMP2 overlaps in all databases, and KLF2 overlaps in miRWalk and Targetscan ( Figure 4A, 4E). According to reports, TIMP2 is a natural inhibitor of MMP, inhibiting angiogenesis and invasion and metastasis. KLF2 is considered to be essential to the integrity of the endothelial barrier. Some scholars have found that KLF2 can regulate occludin. In addition, KLF2 inhibits angiogenesis by inhibiting the promoter activity of VEGF2 [32]. We found that overexpressed KLF2 has an inhibitory effect on angiogenesis, but can maintain vascular permeability. (Figure S3 D-E). Overexpressed TIMP2 has an inhibitory effect on angiogenesis, but not on vascular permeability ( Figure  S3 F-G). Therefore, we believe that the angiogenesis and vascular permeability of TIMP2 and KLF2 in the niche before transfer are crucial. TIMP2 and KLF2 expression is reduced in lung cancer, and patients with increased expression have a poorer prognosis ( Figure 4B,4C, 4D, 4F,4G,4H). In patients with elevated miR-3157-3p, TIMP2 / KLF2 expression decreased( Figure S3 B-C). To determine whether TIMP2 and KLF2 are the targets of miR-3157-3p, we conducted a luciferase reporter gene experiment in HUVECs. As shown, TIMP2 and KLF2 3'UTR contain potential miR-3157-3p binding sites. In addition, dual luciferase reporter gene assays were performed with TIMP2 / KLF2-Wt and TIMP2 / KLF2-Mut co-transfected into cells with miR-3157-3p mimics or NC. Compared with the NC mimics, the luciferase activity of TIMP2-Wt and KLF2-Wt is inhibited (p <0.05) in the presence of miR-3157-3p mimics, which indicates that miR-3157-3p can speci cally bind TIMP2 and KLF2 ( Figure 4I, 4J). Western blot analysis showed that overexpressed miR-3157-3p can inhibit the expression of TIMP2 and KLF2, while inhibiting the expression of miR-3157-3p played the opposite role ( Figure 4K). In order to further verify that whether TIMP2 and KLF2 are the target genes of miR-3157-3p in HUVEC, we conducted a reversion experiment. Stably transfecting HUVEC cells with TIMP2 / KLF2 plasmid and siRNA-TIMP2 / KLF2. The high expression of TIMP2 has reversed the effect of miR-3157-3p up-regulation on HUVEC cell proliferation, migration and tube formation. Similarly, the down-regulation of TIMP2 effectively counteracted the inhibitory effect of miR-3157-3p inhibitors ( Figure S4A, S4B,S4D). In addition, it was found that the up-regulation of KLF2 can inhibit the vascular permeability of miR-3157-3p, while the low expression of KLF2 had the opposite effect ( Figure   S4C).However, TIMP2 overexpression or si-TIMP2 had no effect on vascular permeability compared with the NC group ( Figure S4E). So we found that cells transfected with plasmid TIMP2 / KLF2 showed upregulation of TIMP2 / KLF2 compared with the control group, which can inhibit the effect of Lv-miR-3157-3p. However, TIMP2 / KLF2 was down-regulated in si-TIMP2 / KLF2 transfected cells, inhibiting the effect of Lv-miR-3157-3p-inhibitor ( Figure S4F).

NSCLC-secreted miR-3157-3p primes pre-metastatic niche
We speculate that exosome miR-3157-3p can signi cantly promote the proliferation, migration, angiogenesis and increased permeability in HUVECs in a manner similar to endogenous miR-3157-3p, and the formation of pre-metastatic niche. Then, we conducted a series of experiments to further con rm whether exosome miR-3157-3p induces angiogenesis and vascular permeability. The exosomes isolated from transfected NSCLCs were co-cultured with HUVECs in order to determine whether exosomal miR-3157-3p mediates angiogenesis and permeability of HUVECs. Compared with the NC-Exos,miR-3157-3p-Exos can signi cantly increase HUVEC angiogenesis and vascular permeability. Related processes include HUVEC proliferation ( Figure 5A), tube formation ( Figure 5 B-C), permeability ( Figure 5D), and migration ( Figure 5E). The expressions of ZO-1, Occludin, and Claudin5 were further studied in the IF experiment ( Figure 5F). To further determine that miRNAs in exosomes play a role in promoting metastasis, we used PCR to study the expression level of miR-3157-3p in each group ( Figure 6A). In addition, treatment with miR-3692-3p inhibitor and GW4869 (an inhibitor of exosome internalization), A549 / miR-3157-3p exosomes fail to induces angiogenesis and migration of HUVECs (Figure 6 B-C).We have con rmed in WB and found that in HUVEC cells with high miR-3157-3p expression, TIMP2, KLF2, ZO-1, Occludin, Claudin5 expression decreased, while VEGF, MMP2, MMP9 expression increased ( Figure  6D). In order to further study whether TIMP2 and KLF2 are Exo-miR-3157-3p target genes in HUVEC, we conducted a rescue experiment. We stably transfected HUVEC cells with TIMP2 / KLF2 plasmid and siRNA-TIMP2 / KLF2. The high expression of TIMP2 reversed the effect of miR-3157-3p-Exo-regulation on HUVEC cell tube formation and migration. Similarly, the down-regulation of TIMP2 effectively counteracted the inhibitory effect of miR-3157-3p inhibitors-Exo( Figure 6E,6G). The expression of TIMP2 has no effect on vascular permeability ( Figure 6F). In addition, the up-regulation of KLF2 inhibited miR-3157-3p-Exo vascular penetration, while the low expression of KLF2 had an opposite effect ( Figure 6H).

Exo-miR-3157-3p promotes angiogenesis and vascular leakinessin vivo
We have established a mouse xenograft model to assess angiogenesis in vivo ( Figure 7A). Consistently, miR-3157-3p-Exos-treated mouse xenografts resulted in higher microvessel density (MVD) and larger tumor size compared to NC-Exos treatment ( Figure 7B, 7C, 7D). IHC staining found that compared to the control group, TIMP2 decreased in the miR-3157-3p-Exos treatment group, while the expression of EGFR / MMP2 / KLF2 increased (Figure 7 F-H). To further con rm that NSCLC cells affect angiogenesis in vivo, a chicken chorioallantoic membrane (CAM) model was established to study the effect of miR-3157-3p on angiogenesis in vivo. The overexpression of miR-3157-3p enhanced angiogenesis, and the number of blood vessel branches (5 mm) around the CAM vehicle increased signi cantly ( Figure 7I). These results then showed that exosome miR-3157-3p signi cantly enhanced angiogenesis in vitro and in vivo. To study whether miR-3157-3p-induced vascular permeability can promote NSCLC metastasis, A549 cells were injected into the tail vein of nude mice pretreated with exosomes. As expected, mice injected with A549 / miR-3157-3p-Exo had more lung metastatic colonies than mice derived from A549 / NC-Exo( Figure 7J). Taken together, these results clearly show that cancer-derived exosome miR-3157-3p can induce the formation of a niche before metastasis, thereby promoting the metastasis of NSCLC.
Exosomal miR-3157-3p is associated with metastatic progression and microvessel density in NSCLC patients.
In previous experiments we have con rmed that circulating exosomes are signi cantly elevated in the blood of metastatic patients compared to non-metastatic patients.To assess whether exosome miR-3157-3p acts as a biomarker, we performed receiver operating characteristic curves. The area under the curve (AUC) was 0.7181 in exosome miR-3157-3p ( Figure 8A). Through ISH experiments, we found that miR-3157-3p in metastatic lung cancer tissues was higher than non-metastatic lung cancer tissues ( Figure  8B). Interestingly, we found that in tissues with high miR-3157-3p expression, CD34 expression also increased. We suspect that miR-3157-3p will promote angiogenesis in cancer tissue ( Figure 8C). Therefore, our clinical data indicate that high levels of miR-3157-3p in circulating exosomes are associated with NSCLC metastasis.

Discussion
Lung cancer is the malignant tumor that causes the most deaths among cancers, and metastasis is the main cause of death of patients [33]. The niche before metastasis is the microenvironment prepared for colonizing circulating tumor cells in speci c organs, including in ammation, immune response, angiogenesis, organic matter, matrix remodeling, and biomarker expression. The microenvironment before metastasis can increase angiogenesis and vascular permeability, thereby promote the metastasis. Studies have showed that exosomes released by hypoxic tumors are more likely to cause angiogenesis and vascular leakage [33,34]. Cancer-induced vascular permeability and angiogenesis have played a key role in cancer metastasis. Exosomes from NSCLC can be used as a communication medium between cells. MiR-3157-3p secreted from NSCLC can induce a niche before metastasis by promoting angiogenesis and destroying the tight junction of venous endothelial cells.
Exosomes are extracellular vesicles with a diameter of 30-150 extranm. The contained miRNA can regulate tumor immunity and microenvironment, which may promote tumor invasion, metastasis and angiogenesis. GSC-derived exosomes such as overexpressing miR-26a can promote HBMEC proliferation and angiogenesis in vitro by inhibiting PTEN [35]. Exosome miR-25-3p is involved in niche formation before the transfer and may be used as a blood-based biomarker for CRC transfer [30]. Therefore, exosome miRNA plays an important role in regulating cancer progression [36]. miR-3157-3p is abnormally expressed in NSCLC patients, especially those with metastases. We found that miR-3157-3p in vascular endothelial cells can induce vascular permeability and angiogenesis by down-regulating KLF2 and TIMP2. Kruppel-like factors (KLFs) are a subclass of the zinc nger family of DNA-binding transcription factors. KLF2 can regulate the key tight junction protein occludin in endothelial cells to maintain the integrity of the endothelial barrier function [37]. TIMP2 is a natural inhibitor of MMP activity and inhibits angiogenesis. In NSCLC, both KLF2 and TIMP2 are down-regulated. Our results showed that miR-3157-3p in HUVEC can absorb exosomes downregulates KLF2 and TIMP2, and subsequently reduce its downstream target, occludin, and increase the levels of VEGF / MMP2 / MMP9. Our data indicate that the exosome miR-3157-3p secreted by NSCLC promotes vascular permeability and angiogenesis by silencing KLF2 and TIMP2( Figure 8D).
Tumor-derived exosomes are considered to be the main driving force of the pre-metastatic niche. Exosomes participated in angiogenesis and increased vascular permeability, thereby promote the formation of niche before transfer. Before metastasis, a microenvironment suitable for tumor metastasis has been created for tumor metastasis. In this study, we explored that whether exosome miR-3157-3p can regulate niche formation before metastasis. Cancer-induced vascular permeability / angiogenesis is one of the characteristics of the pre-metastatic niche. In prostate cancer, miR-21-5p and miR-139-5p, which are microRNAs derived from exosomes, could coordinately regulate the ecological niche before metastasis, and was highly relevant to differentiate prostate cancer (PCa) carcinogenesis, broblast proliferation, differentiation ,migration and angiogenesis [38]. In vitro permeability measurements showed that the exosome miR-25-3p of NSCLC cells signi cantly promoted HUVEC permeability and angiogenesis, reduced the levels of KLF2, ZO-1, occludin, Claudin5, and increased VEGF / MMP2 / MMP9 level. We also found that the upregulation of miR-3157-3p in NSCLC cells promoted proliferation, angiogenesis and metastasis. In summary, we have proposed a new role of TIMP2 / KLF2 in the formation of cancer-induced pre-metastatic niche, and provided new insights into the formation of the metastasis niche. Also we have prevented the transport of exosomes to recipient cells may be an effective strategy to prevent tumor metastasis. There are many targets that can be used to inhibit the formation of pre-metastatic niche, for example, the prevent of the production of pro-in ammatory factors, the inhibition of the recruitment of BMDC, the prevent of angiogenesis and vascular penetration, and the destroy of local matrix and reactivating anti-tumor immune responses). These goals may develop potential methods for preventing and controlling cancer metastasis in the future. Therefore, we believe that blocking exosome miR-3157-3p from NSCLC cells can maintain the tight junction between vascular endothelium, reduce pulmonary vascular permeability and subsequent NSCLC metastasis. It shows that miR-3157-3p can be used as a therapeutic target for the intervention in NSCLC metastasis.

Conclusion
In conclusion, we found that the level of serum exosome miR-3157-3p is related to NSCLC metastasis. However, the molecular mechanism of miR-3157-3p regulating NSCLC metastasis has not been reported. It was con rmed by experiments that exosomes secreted by NSCLC cells can transfer miR-3157-3p to HUVEC, thereby promoting angiogenesis and destroying the integrity of vascular endothelium, and resulting in an increase in vascular permeability. Declarations Ethics approval and consent to participate All animal experiments were approved by the Animal Protection and Use Committee of Nanjing Medical University. Human tissue research was approved by the Medical Ethics Committee of the First A liated Hospital of Nanjing Medical University. All participants have obtained written informed consent.

Consent for publication
All patients obtained written informed consent.

Data availability statement
All datasets and materials generated and / or analyzed during the current studies are available.

Con lct of interests
The authors declare that there are no con ict of interests.   tissues with or without metastasis (29 without metastasis and 18 with metastasis).The data are shown as the mean ± SD(*P < 0.05; **P < 0.01, ***P < 0.001).

Figure 2
NSCLC-secreted miR-3157-3p is transferred to HUVECS.A NSCLC cell lines miR-3157-3p levels were analyzed using qRT-PCR.B Representative electron microscopy micrographs and NTA of A549 conditioning medium secreted exosomes, as well as H1299 conditioning medium secreted exosomes.

Figure 4
TIMP2 and KLF2 are functional targets of miR-3157-3p.A The candidate gene targets were predicted by intersecting outputs from four distinct prediction algorithms (TargetScan, miRDB, and miRWalk).B TIMP2 expression was found to be statistically signi cant in NSCLC tissue.C The relationship between TIMP2and the survival of NSCLC patients (n = 963 in the TIMP2-low group and n = 962 in the TIMP2-high group).D The TIMP2 expression level in 40 NSCLC tissues and adjacent tissues was detected by qRT-PCR.E The candidate gene targets were predicted by intersecting outputs from four distinct prediction algorithms (TargetScan and miRWalk).F KLF2 expression was found to be statistically signi cant in NSCLC tissue.G The relationship between KLF2 and the survival of NSCLC patients (n = 962 in the TIMP2low group and n =963 in the KLF2-high group).H The KLF2 expression level in 40 NSCLC tissues and adjacent tissues was detected by qRT-PCR.I the miR-3157-3p binding site in TIMP2 3'UTR and the detection of luciferase activity.J the miR-3157-3p binding site in KLF2 3'UTR and the detection of luciferase activity.K A negative regulatory effect of miR-3157-3p on TIMP2/KLF2 was tested by western blot.The data are shown as the mean ± SD (*P < 0.05; **P < 0.01, ***P < 0.001).