It has been generally believed that cancer-associated fibroblasts (CAFs) have the ability to increase the process of tumor angiogenesis. However, the potential mechanisms by which cancer-derived exosomes in lung cancer (LC) remains to be investigated. LC-derived exosomes were administrated to NIH/3T3 cells. A variety of experiments were conducted to investigate the proangiogenic factors of CAFs, including Western blot, RT-PCR, colony formation assay, tube formation assay, Matrigel plug assay et al. In addition, the impact of JAK2/STAT3 signaling pathway were also explored. The role of hsa-miR-210 was identified with microarray profiling and validated in vitro and in vivo assays. The target of miR-210 was screened by RNA pull down, RNA-sequencing and then verified. It was shown that LC-derived exosomes could induce cell reprogramming, thus promoting the fibroblasts transferring into CAFs. In addition, the exosomes with overexpressed miR-210 could increase the level of angiogenesis and vice versa, which suggested the miR-210 secreted by the LC-derived exosomes may initiate the CAF proangiogenic switch. According to our analysis, the miR-210 had the ability of elevating the expression of some proangiogenic factors such as MMP9, FGF2 and vascular endothelial growth factor (VEGF) a (VEGFa) by activating the JAK2/STAT3 signaling pathway, ten-eleven translocation 2 (TET2) was identified as the target of miR-210 in CAFs which has been involved in proangiogenic switch. miR-210 was overexpressed in serum exosomes of untreated non-small cell LC (NSCLC) patients. We concluded that the promotion effect of exosomal miR-210 on proangiogenic switch of CAFs may be explained by the modulation of JAK2/STAT3 signaling pathway and TET2 in recipient fibroblasts.
Lung cancer (LC) is one of the most prevalent carcinoma (11.6% of the overall cases) with a rather high mortality rate (18.4% of the total cancer deaths) all over the world . Among all the pathological types, non-small cell LC (NSCLC) is the disease most commonly seen, which accounts for 85% of all the cases . However, the cause of lung carcinoma has not been fully researched and understood. The therapeutic approaches, such as surgery or chemotherapy, have not been able to cure the NSCLC completely, but only to prolong patients’ survival and to increase the quality of life . In recent years, the therapeutic effect of LC has been greatly improved due to the development of targeted medicine. For instance, the median survival of patients with EGFR mutations has achieved 4.3 years in America . Moreover, immunocheckpoint inhibitors also brings new research prospective to LC treatment. Despite the advancements, LC is still a threatening disease with great burden. It was proposed that the identification of new biomarkers related to LC could be of great clinical significance to LC patients. Thus, it is worthwhile to investigate the LC-related molecular mechanisms and signaling pathways.
It has been widely believed that the alteration in tumor micro-environment (TME) is closely associated with tumor progression . Cancer-associated fibroblasts (CAFs) are considered to be the most important TME component, which will be activated by tumor-related active mediators and transformed into fibroblasts with myofibroblast characteristics, thus modulating the progression of tumor cells by releasing a variety of cytokines and remodeling the extracellular matrix (ECM) . In addition, CAF can secrete a number of proangiogenic factors and recruit endothelial progenitor cells, thus promoting the process of tumor angiogenesis . At the same time, the platelet-derived growth factors (PDGF) secreted by the CAF can also regulate the angiogenesis process by either directly releasing vascular endothelial growth factors (VEGFs) or indirectly inducing the transcription of VEGF .
The exosome was first discovered by Johnstone et al. , in the study of vesicle formation during the maturation of reticulocytes. Exosomes are small vesicles of 40–100 nm in diameter, which are produced and released by various cells through endosomal pathways, and are enveloped by lipid bilayer . The exosome did not draw much attention until 1996, when scientists discovered that the exosomes secreted by B lymphocytes can directly stimulate CD4+ T cells thus producing anti-tumor effects in the presence of MHC-II class molecules, co-stimulatory molecules and adhesion molecules in the exosomes . Afterward, it was supported by more and more researches that the exosomes also contain other bioactive substances such as DNA fragments, mRNA, microRNA (miRNA), functional proteins, transcription factors and so on. Moreover, the membrane structure can also express antigens and antibody molecules that would lead to a series of biological effects.
Activation of the JAK2/STAT3 signaling pathway has been reported in a variety of tumor variants, which could initiate the STAT3 phosphorylation. The phosphorylated STAT3 activated by the JAK2 protein can dimerize and bind to the target DNA in nucleus, thus activating the translation of specific genes . It has been reported by several studies that the JAK2/STAT3 pathway has regulation effects on the proliferation and migration of tumor cells as well as the expression of proteolytic enzymes and proangiogenic factors, including MMP9, VEGFa, FGF2 and so on [13,14].
In the present study, we intended to confirm that LC cell lines (A549 and H460) release and use exosomes to transfer certain miR in fibroblasts. These exosomes could promote CAF activation and the expressions of proangiogenic factors in CAFs. JAK2/STAT3 signaling pathway has been involved in this process. Above-mentioned mechanism may lead to the proangiogenic switch of CAFs. These results may provide a new sight for the anti-angiogenic therapy of LC.
The main antibodies used in our study were obtained from Abcam (U.S.A.), including those for VEGFa (ab1316), MMP9 (ab38898), CD63 (ab59479), Tsg101 (tumor susceptibility gene 101) (ab125011) and Hsp90 (heat shock protein 90) antibody (ab13492), Histone H3 (ab176842), STAT3 (ab76315), P-STAT3 (ab32143), JAK2 (ab108596), P-JAK2 (ab32101), α-smooth muscle actin (α-SMA) (ab124964), fibroblast activation protein (FAP) (ab53066) and ten-eleven translocation 2 (TET2) (ab124297). The growth factor-reduced Matrigel was purchased from BD Biosciences. The analysis kits were provided by Dojindo (CCK-8) and Thermo Fisher (BCA Protein Assay Kit), respectively.
The epithelial cell lines were acquired from the China Center for Type Culture Collection, including the bronchial epithelium derived from individuals without carcinoma (BEAS-2B), LC cell lines (H460, A549) as well as the NIH/3T3 and MS-1 cell lines. The experimental procedures were approved by the Ethics Committee of The Second Affiliated Hospital of Zhejiang University School of Medicine. Routine tests of STR and Mycoplasma were performed to verify the authenticity. The study cells were cultured in high-glucose DMEM with 10% FBS under the conditions of 37°C and 5% CO2. Approximately 20 μg/ml LC cell-derived exosomes were added to the medium when confluence of NIH/3T3 cells reached 70–80%.
Acquiring the LC-derived exosomes
In our study, the A549 and H460 cell lines were adopted for exosome isolation. When 80% confluence was achieved, the cells were washed in PBS for three times and then cultured with the medium with 10% extracellular vesicles (EVs)-depleted FBS. After 48-h cultivation, the conditioned medium (CM) was collected and centrifugated at 800×g for 15 min and then at 10000×g for 30 min. Afterward, ultracentrifugation at 100000×g for 70 min was performed to isolate the exosomes with Beckman Optima L-100XP. Seventy minutes later, the isolated exosomes were washed in PBS and observed with transmission electron microscopy HT7700, while the hydrodynamic diameter was measured with Nano-ZS ZEN 3600.
The PKH26-labeled exosomes were administrated to the NIH/3T3 cells when 70–80% confluence was achieved. The interaction between exosomes and fibroblasts was monitored by using Olympus FV1200.
Quantification of the expressed genes
The TRIzol reagent (Takara, Japan) was adopted for total RNA extraction based on the instructions provided by the manufacturer. Afterward, the total RNA obtained from exosomes and cells were reversely transcribed into cDNA with the PrimeScript RT Reagent Kit, and the expression levels of genes that coded relevant proteins were determined with the qPCR technique by using the SYBR® Premix Ex Taq™ II and the data were analyzed with the QuantStudio™ 6 Flex. Similarly, the quantification of mature miRNAs and U6 was performed with MicroRNA qPCR Kit (Genelily, Shanghai, China) according to the instructions of manufacturers. The comparative cycle threshold (Ct) method was applied for the quantification of miRNAs or mRNAs, with the normalized control of U6 or GAPDH. The experimental procedures were performed in triplicates, where the variation among the Ct values should be less than 0.5. The synthesis of the miRNA primers used in qPCR was provided by Genelily (Shanghai, China), and the sequences were listed in the Supplementary Table S1.
Western blot analysis
The Cell Fractionation Kit (Cell Signaling Technology, U.S.A.) was adopted for separation of nucleus and cytoplasm based on the instructions provided by the manufacturer. After the extraction of the total protein with M-PER, the BCA Protein Assay Kit was applied for protein quantification in every sample. The mixture of the loading buffer (5×) and the protein solutions were heated in loading buffer at 95°C for 10 min. The SDS/PAGE was carried out for 30 min with a voltage of 60 V and for 1 h at 110 V, after which the separated proteins were transferred to PVDF membrane, which were blocked with 5% skim milk with the addition of 0.05% Tween 20 (TBST). After incubation with antibodies at 4°C overnight, the proteins were detected by anti-rabbit IgG, anti-mouse IgG, or horseradish peroxidase–conjugated (Pierce Chemical, U.S.A.). The concentrations of MMP9, VEGFa and FGF2 were determined by using ELISA kits (4A Biotech Co., Ltd., China) in accordance with the instructions provided by the manufacturer.
Determination of cell proliferation and migration
In our study, the MS-1 cells were seeded in 96-well plates and cultivated with CM of fibroblasts. The fibroblasts were previously stimulated with LC cell-secreted exosomes or not. The condition of cell proliferation 24, 48 and 72 h was measured with CCK-8 (Dojindo, Japan). Transwell migration assay and tube formation assays were introduced to evaluate the cell migration ability. The MS-1 cells (5 × 104) were inoculated in 100 μl serum-free medium in the upper chambers of 24-well plates with inserts (Corning, U.S.A.), while the fibroblasts that stimulated with LC cell-secreted exosomes or not were inoculated in 500 μl of FBS-DMEM in lower chambers to assess the migration. Twenty-four hours later, after removal of the cells in the upper chamber, the cells that migrated to the lower surface were fixed and then stained with Crystal Violet. Six fields were randomly selected in which the cell numbers were counted. The experiments were conducted in the design of negative control of untreated cells, and the procedures were conducted in triplicate. In tube formation assay, the MS-1 cells (2 × 104/well) were inoculated into the 48-well plates with Matrigel (BD Biosciences, San Jose, CA, U.S.A.) and cultivated with CM of fibroblasts that stimulated with LC cell-secreted exosomes or not. Six hours after incubation, capillary-like structure was photographed under the BHS-313 Olympus and the results of tube formation were analyzed by using the Image-Pro Plus 6.0.
Colony formation assay
Colon spheres were generated as previously described . Briefly, MS-1 cells were incubated with CM from fibroblasts administrated with the exosomes derived from A549 and H460 cells or control untreated fibroblasts for 48 h, then cells were seeded into six-well plates (500 cells/well) for 24 h and incubated for 3 weeks. After incubation, cells were fixed in methanol and stained with 0.1% Crystal Violet. Quantity One software (Bio-Rad, Hercules, CA, U.S.A.) was used to count colonies.
miRNA expression array
The miRNA expression pattern among LC-derived exosomes and normal human bronchial epithelium cell-derived exosomes analyzed using human Exiqon miRCURY™ LNA microRNA array (Exiqon, Denmark). Two cell pellets were pooled in each group. The miRNA array experiment was carried out in KangChen Bio-tech Inc. (Shanghai, China).
Fibroblasts (1 × 107) treated with different type of LC cell-secreted exosomes or not was mixed with Matrigel (100 μl). All mice were raised and manipulated following the protocols approved by the Laboratory Animal Care and Use Committee of Zhejiang University School of Medicine. The animal experiments were conducted at Shanghai Institutes for Biological Sciences. The treated/untreated NIH/3T3 cells were mixed with H460 cells and A549 cells (1 × 107) in a ratio of 2:1, which were then administrated to the BLAB/C-nude mice aged 6 weeks by subcutaneously injection. The xenografts were weighted, and the volume was measured 1 month later by using formula: volume (cm3) = (width2 × length)/2. Afterward, the xenografts were embedded in paraffin for further study of immunofluorescence staining.
To establish the model of abdominal metastasis, the mice first underwent anesthesia by using isoflurane (induced concentration 3%; maintaining concentration 1.5%). The spleen was exteriorized. Luciferase tagged A549-Luc (5 × 106) and H460-Luc (5 × 106) and the treated/non-treated NIH/3T3 cells premixed with Matrigel were injected into the spleen parenchyma. Then spleen was returned to the peritoneal cavity and incision was managed. After approximately 30 days, bioluminescence images were collected by the Interactive Video Information System (0.4% chloral hydrate was used to keep the mice calm, 10 mg/kg, approximately 0.1–0.2ml for each one). For ethical consideration, the animal testing also complied with the NIH principles of laboratory animal care.
Four percent paraformaldehyde was applied for the fixation of the xenografts for 24 h, which were incubated consecutively with anti-mouse CD31 antibody and Cy3–conjugated IgG in paraffin. After observing with a fluorescent microscope, the microvessel density (MVD) was calculated. As previously described, either single cell or discrete clusters were counted as one microvessel structure . The slides were incubated with anti-P-STAT3 antibody at 4°C overnight, and then FITC–conjugated goat anti-rabbit IgG antibody were applied for incubation at room temperature for 45 min, and the cell nuclei were stained by DAPI. Olympus FV1200 was used for observation and photography of the stained cells.
FITC-labeled miR-210 mimic and miR-210 inhibitor (anti-miR-210) as well as the corresponding negative controls were obtained from Simo Biomedical Technology (Shanghai, China).
RNA pull-down assay
Briefly, total RNA of A549-derived and H460-derived exosomes was isolated and incubated with the 3′ end biotinylated pre-miR-210 for 24 h and then with streptavidin-coated magnetic beads (Ambion, Life Technologies). The biotin-coupled RNA complex was then pulled down and subjected to RNA-Sequencing. Abundance of Target gene in bound fractions was analyzed by RT-qPCR.
The first-strand cDNA synthesis kit (Fermentas, Vilnius, Lithuania) was applied to reverse transcribe total RNA into cDNA. Manufacturer’s instructions were strictly followed. RNA-seq was performed using OE bioTechnology (shanghai, China). The data were expressed as mean displayed in the center of the heatmaps.
Lentiviral transduction and generation of stable cell lines
Human TET2 lentivirus (Lv-TET2) and shRNA lentivirus targeting TET2 (si-TET2) were constructed by Simo Biomedical Technology (Shanghai, China). The lentiviruses were labeled by enhanced green fluorescent protein (EGFP) and puromycin. After 2 days of lentiviral transduction, stable cell lines were selected with 2 µg/ml puromycin for at least 2 weeks.
Luciferase miRNA target reporter assay
PsiCheck2 (Promega, U.S.A.) was cloned by the 3′UTR of TET2 mRNA containing wild-type or mutant miR-210 binding sites. Then NIH/3T3 cells were co-transfected with psiCheck2-TET2-wt/mut plasmids and miR-210 or miR-NC using Lipofectamine 3000™ (Invitrogen). Approximately 48 h later, cells were harvested. Luciferase activity was measured by Glomax 96 luminometer (Promega).
Clinical samples examination
A total of 14 patients hom ere initially diagnosed N withSCLC with no previous treatment were included to collect blood samples at The Second Affiliated Hospital of Zhejiang University School of Medicine in March 2020. All patients were at least 18 years old. Informed consent was obtained from each participant. For each sample, the blood samples were collected at 1 day before and 1 day right after radical resection. Blood samples were centrifuged at 2000×g for 10 min at 4°C, the supernatants were collected and further centrifuged at 1600×g for 20 min and at 10000×g for 30 min at 4°C. The supernatants were processed by 0.22-μm disposable filter. Then all samples were centrifuged at 100000×g for 2 h at 4°C. Finally, the serumal exosomes were resuspended in 0.1 ml PBS after washing with 10 ml PBS. All exosomes were stored at −80°C for further qRT-PCR detection. The experiment was approved by the Institutional Human Experiment and Ethics Committee of The Second Affiliated Hospital of Zhejiang University School of Medicine. The Declaration of Helsinki was strictly followed.
The measurements were presented in the form of means ± SD. The differences among groups were evaluated by using Student’s t test and one-way ANOVA test according to the data nature. All the analyses were conducted in the software of SPSS 21. P-value less than 0.005 was considered as the criteria of statistical significance.
Identification of exosomes secreted by LC cells
The EVs isolated from the culture supernatant of A549 and H460 cells were identified by using transmission electron microscope and dynamic light scattering analysis (Figure 1A,B). The exosomal markers Hsp90, Tsg101 and CD63 was performed by using Western blot (Figure 1C). The bilayer-enclosed morphology, 50–200 nm in diameter and the presence of exosomal markers confirmed that the vesicles were exosomes.
LC cell-derived exosomes promoted the transformation of fibroblasts into CAFs
The promotion impact of the LC cell-derived exosomes on the transformation of fibroblasts into proangiogenic CAFs
The fluorescence staining image of the NIH/3T3 cells was presented in Figure 1D, where the exosomes were labeled with PKH26. Fluorescent stain could be observed mainly located in cytoplasm. Figure 1E showed that the NIH/3T3 cells receiving the administration of exosomes derived from LC cells were presented as spindle-like shape, which was the typical morphology of CAFs. In addition, the results of the qRT-PCR and Western blot also demonstrated increased expression of the CAF typical markers α-SMA and FAP in the NIH/3T3 cells treated with the exosomes derived from LC cells (Figure 1F,G). Angiogenesis induced by CAFs have been reported by secreting proangiogenic factors. NIH/3T3 cells were stimulated with LC cell-derived exosomes (30 μg/ml). The qRT-PCR, ELISA and Western blot suggested the proangiogenic factors, such as MMP9, FGF2 and VEGFa, were highly expressed in the NIH/3T3 cells treated with exosomes compared with the untreated cells (Figure 2A–C). The mouse islet endothelial cell line MS-1 which is certified by ATCC was used to investigate the proangiogenic effect of the fibroblasts that treated by LC cell-secreted exosomes in vitro. The transwell migration assay suggested an up-regulated number of traversed MS-1 cells in the experiment group compared with the control group (Figure 2D). Additionally, the matrigel tube formation assay also suggested that the total branching length of the tubes formed by MS-1 significantly increased when incubated with the CM from NIH/3T3 cells that stimulated by LC cell-secreted exosome (Figure 2E). In addition, the CCK-8 and colony formation assay suggested that the proliferation of the MS-1 cells was significantly improved when incubated with the CM from NIH/3T3 cells that stimulated by LC cell-secreted exosome (Figure 2F,G). To determine the proangiogenic ability of LC exosomes treated fibroblasts in vivo, the in vivo verification study in which NIH/3T3 cells treated with LC cells-secreted exosomes and NIH/3T3 cells-treated NC were subcutaneously injected into BALB/C mice for 2 weeks was then conducted. For each group, three BALB/C mice aged 4 weeks were included. Two weeks later, matrigel plugs were harvested from mice and introduced Immunofluorescence staining. The result of quantification MVD showed that the MVD of exosome-treated NIH3T3 cells were remarkably improved compared with the non-treated group (Figure 2H).
Reprogramming of fibroblasts into proangiogenic CAFs induced by LC cell-derived exosomes
JAK2/STAT3 signaling pathway activation by LC cell-secreted exosomes
JAK2/STAT3 signaling is a classical pathway that has been involved in research of angiogenic factors [14,17–19]. Therefore, we make validation in this pathway first. An increased phosphorylation level of JAK2 and STAT3 as well as an increased accumulation of P-STAT3 in the nucleus of the NIH/3T3 cells were demonstrated in the exosome-treated group according to the Western blot analysis (Figure 3A), which was also verified by the observation under a confocal microscope (Figure 3B). JAK2 inhibitor, Stattic, was applied to determine whether the activation of JAK2/STAT3 signaling could elevate the expression of proangiogenic factors. Intervention of Stattic was concurrently performed when NIH/3T3 cells was stimulated with LC cell-secreted exosomes. The qRT-PCR and Western blot indicated the phosphorylation of STAT3 and JAK2 as well as expression levels of FGF2, VEGFa and MMP9 were alleviated in the Stattic-treated group (Figure 3C,D). Moreover, the promotion impact on the proliferation (Figure 3E,F), migration (Figure 3G) and tube formation (Figure 3H) of MS-1 cells incubated with CM of NIH/3T3 treated by exosomes could also be weakened by the Stattic treatment.
Regulatory effects of the LC cell-derived exosomes on the proangiogenic switch of CAFs through modulating the JAK2/STAT3 signaling pathway
Trans-shipment of the exosomal miR-210 into CAFs
Bilateral structured exosomes are capable of delivering proteins and miRNAs to adjacent cells in the micro-environment. In order to explore the underlying mechanism of LC cell-exo on CAFs, miRNA microarray assay was performed. The sequencing results of the miRNA microarray assay that performed with A549-exo, H460-exo and BEAS-2B-exo were presented in Figure 4A. It was identified that the levels of miR-210 was both significantly increased in A549-exo and H460-exo compared with those in BEAS-2B-exo. Exosomal miR-210 in the regulation of LC angiogenesis and progression desires to be detailed characterized. More specifically, the exosomal and cellular miR-210 levels in H460-exo, A549-exo, H460 cells and A549 cells were remarkably improved compared with those in NIH/3T3 cells and BEAS-2B-exo (Figure 4B,C). The miR-210 expression in NIH/3T3 elevated after the administration of LC cell-derived exosomes (Figure 4D). The A549 cells and corresponding exosome that transfected with miR-210-mimic could improve the levels of miR-210 (Figure 4E,F). The miR-210 expression in NIH/3T3 was more up-regulated after the administration of miR-210-mimic-A549-derived exosomes than those of miR-NC-mimic-A549-derived exosomes (Figure 4G). The H460 cells and exosome that transfected with anti-miR-210 could abate the levels of miR-210 (Figure 4H,I) while those in the NIH/3T3 cells treated with exosome derived from anti-miR-210-transfected H460 cells was not changed compared with the non-treated group (Figure 4J). The fluorescence experiment showed that no signal of FITC or PKH26 fluorescence could be observed within the NIH/3T3 cells administrated with non-labeled exosomes or naked FITC-tagged miR-210 (Figure 4K), which suggested possible transition of the exosomal miR-210 into CAFs.
Translocation of the miR-210 into the fibroblasts with the carrier of LC-cell-derived exosomes
In vitro/in vivo regulation of the miR-210 on CAF proangiogenic phenotype
In treated group, according to our analysis, the levels of MMP9, VEGFa and FGF2 as well as the phosphorylation of JAK2/STAT3 signaling pathway in the NIH/3T3 cells treated with miR-210-mimic-transfected-A549-secreted exosomes were remarkably increased. On the contrary, the exosomes derived from anti-miR-210-transfected H460 cells imposed a down-regulation effect (Figure 5A). In addition, our study also found that the cell proliferation, migration and tube formation of the MS-1 cells were promoted by incubating with the CM of NIH3T3 treated with exosomes derived from miR-210-mimic-transfected A549 cells (Figure 5B,C,F,G), whereas the exosomes from anti-miR-210-transfected H460 cells demonstrated opposite effect (Figure 5D,E,H,I), which were also in line with the results from in vivo study. The xenograft study showed that the alteration in the miR-210 level was positively correlated with the tumor volume (Figure 6A–D) weight (Figure 6E,F) and MVD quantification in tumor tissues (Figure 6G,H). In abdominal metastasis model, the mice were anesthetized, and their spleens were fetched. Approximately A549 (5 × 106) or H460 (5 × 106) and different types of exosome-treated NIH/3T3 cells/untreated NIH/3T3 cells premixed with Matrigel were injected into the spleen parenchyma. Then spleen was returned to the peritoneal cavity and incision was managed. Then, the mice were kept and fed in experimental animal room for 30 days. We carried out bioluminescence imaging to estimate the migration ability of LC cells in vivo. The results revealed that elevated miR-210 level in A549-secreted exosomes significantly promote the migration ability in vivo, whereas decreased miR-210 level in exosomes led to the opposite result (Figure 6I).
Regulatory effects of the miR-210 regulates on the proangiogenic switch of CAFs
In vivo study verified the regulation of exosomal miR-210 on angiogenesis
TET2 was the target of exosomal miR-210 in CAFs
To further explore the mechanism of exosomal miR-210 reprogramming fibroblasts into CAFs, we isolated RNA of LC cells-derived exosome and incubated with 3′ end biotinylated pre-miR-210 and streptavidin-coated magnetic beads (Ambion, Life Technologies). The biotin-coupled RNA complex was then pulled down and subjected to RNA-Sequencing (Figure 7A). The sequencing result which was in accordance with each other in two exosomes was listed as the pattern of heatmap in Figure 7B. A series of potential target genes were presented. The possible genes that targeted by miR-210 were also predicted and screened based on the bioinformatics database (http://www.targetscan.org/vert_72/). It was shown that miR-210 might interact with the 3′ UTR of TET2 (Figure 7C). Then, the result of qRT-PCR confirmed that TET2 was more enriched in the biotin-pre-miR-210 of LC-derived exosomes than in BEAS-2B-derived exosome (Figure 7D). We employed the dual-luciferase reporter assay to validate this finding. Data indicated that miR-210 could significantly decrease the luciferase activity of the luciferase reporter gene that contains TET2 3′ UTR-wt. miR-210 had no significant influence on TET2 3′ UTR-mut (Figure 7E). To verify the regulation of TET2 on exosomal miR-210, rescue experiments on angiogenesis in NIH/3T3 and MS-1 cells were introduced. It had proved that phosphorylation of JAK2/STAT3 signaling caused by alteration of exosomal miR-210 could be largely neglected by simultaneous regulation of TET2 in exosome recipient cell NIH/3T3 according to Western blot assay (Figure 7F). Also, the similar conclusions could be drawn in MS-1 cells on cell proliferation (Figure 7G–J), migration (Figure 8A,B) and tube formation (Figure 8C,D). The abdominal metastasis model in vivo further affirmed the conclusion in vitro (Figure 8E). These results strongly manifested that exosomal miR-210 may modulate the reprogramming of fibroblast into CAFs and thus affecting ability of angiogenesis via JAK2/STAT3 pathway by directly targeting TET2 (Figure 8F).
TET2 was identified as the target of exosomal miR-210 in CAFs
In vitro & in vivo studies verified the regulation of exosomal miR-210 on TET2
miR-210 was overexpressed in serum exosomes of untreated NSCLC patients
To validate the conclusion, 28 serum exosome samples were collected from 14 initially diagnosed NSCLC patients (for each patient, two samples were collected at pre-operation and post-operation, respectively). The results of qRT-PCR revealed that the expression of miR-210 in serum exosome of pre-operation samples was more elevated than that in serum exosome of post-operation samples (Supplementary Figure S1). The difference was statistically significant (P<0.05).
It was suggested by a number of studies that multiple cytokines in the LC-derived exosomes, such as VEGF, TGF and miRNA, can interact with the endothelial cell membrane surface receptors via different mechanisms, thus promoting the proliferation and migration of lung tumor cells as well as the angiogenesis process . Previous in vitro studies suggested that the sortilin secreted by exosomes derived from A549 cells had an impact on the activation and migration of endothelial cells as well as the angiogenesis in LC through the formation of TES complex by interacting with tyrosine kinase receptor (TrkB) and EGFR . Kosaka et al.  discovered that the miR-210 derived from the LC-sourced exosomes could secret the nSMase2 to specifically inhibit the target genes, thus further promote the formation of tumor vessels. In addition, another study also demonstrated that the A549 LC cells could secret miR-494 and exosomes to peripheral endothelial cells under hypoxic state, and make a positive effect on tumor angiogenesis by inhibiting the PTEN and Akt/eNOS pathway . Furthermore, the LC-exosome-derived miR-210 could also increase the angiogenesis by adjusting the level of Ephrin A3 in endothelial cells .
The above evidence indicated that the LC-derived exosomes can mediate the formation of tumor vessels via numerous of pathways, thus regulating the micro-environment of the tumor and promoting the growth and metastasis of LC. Our study demonstrated that the exosomes derived from LC cells could increase the expression levels of α-SMA and FAP in NIH/3T3 cells, which suggested that the LC-sourced exosomes could be considered as a trigger of the reprogramming of normal fibroblast into CAFs. The CAF is the main stromal cell in TME, which is transformed by the inherent fibroblasts. Previous studies indicated that the CAF is not only the base material for tumor growth, but also can secret a variety of cytokines, which can interact with tumor cells and other interstitial cells, thus increasing the proliferation, invasion and metastasis of tumor cells. Former in vitro studies discovered that the Hsp70 in the exosomes secreted by the A549 LC cells could stimulate the mesenchymal stem cells (MSCs) to generate a kind of pro-inflammatory mesenchymal stem cell, and to further increase the levels of IL-6, IL-8, MCP-1 by activating the NF-KB/Toll receptor 2 (TLR2) through the NF-κB pathway, thus promoting the tumor growth of LC . Additionally, the exosomes from human and mouse LC model cells was also found to be able to activate fibroblasts and endothelial cells, and induce the stromal cells to express angiogenic factors such as IL-8, VEGF, LIF, OSM, IL-11 and MMP-9, thus increasing angiogenesis in the LC micro-environment . According to our analysis, the expression levels of the angiogenic factors were up-regulated in the CAFs, such as MMP9, FGF2 and VEGFa. In addition, the tube formation, migration and proliferation of EC were elevated after the incubation together with the CM of CAFs. Moreover, the CAF presented high MVD value in the Matrigel plug assay. All the above results suggested that the proangiogenic capability of CAFs could be induced and elevated by the LC cell-derived exosomes. VEGFa and FGF2 could impose strong effects on proangiogenic process by activating a variety of downstream signaling pathways in ECs. It has been proposed that the increased levels of soluble proangiogenic factors (e.g. VEGF, FGF2) in serum were closely related to poor clinical outcome in LC [27,28]. In addition, it was also reported that the MMP9 can also release VEGF from the ECM . Such ECM remodeling enzymes could facilitate the reconstruction of the ECs. Therefore, our results supported the significance of the CAFs in tumor angiogenesis.
More and more evidence demonstrated that the JAK/STAT signaling pathway could play an important role in tumor angiogenesis [30–33]. For instance, a variety of crucial proangiogenic factors, such as FGF2, insulin-like growth factor 1 (IGF-1), MMP-9, MMP-2 and VEGFa, were regulated by the pathway . It was indicated that the invasion and migration of LC cells were significantly enhanced after incubation with the CM of CAFs. They speculated the epithelial-mesenchymal transition (EMT) could be induced by the CAF-CM via the regulation of the expression of EMT-associated markers, such as E-cadherin and vimentin. Moreover, the CAF-CM could also impose effects on the metastasis-related genes, such as MMP-2 and VEGF . Other studies investigating the mechanisms suggested that the CAFs could increase the release of IL-6, which in return not only activated the JAK2/STAT3 signaling pathway, but also inhibited the IL-6/STAT3 signaling pathway . To conclude, previous mechanical studies suggested that the metastasis of LC cells could be enhanced by EMT, which was induced by the CAFs through the IL-6/STAT3 signaling pathway. Our study discovered that the JAK2/STAT3 signaling pathway could play an important role in the proangiogenic switch of CAFs by modulating the expression of a variety of proangiogenic factors (e.g. MMP9, FGF2 and VEGFa). In our study, we found that the administration of LC cell-derived exosomes could remarkably increase the phosphorylation of STAT3 and JAK2 in CAFs, and the expression levels of proangiogenic factors as well as the effect of CAF-CM changed synchronously with the JAK2/STAT3 signaling pathway.
The great significance of the miR-210 has been reported in a variety of cancers [37–42], such as LC [43–45]. In addition, it was proposed that the miR-210 in JAK-STAT pathway could be involved in arteriosclerosis obliterans . With respect to lung carcinoma, the regulation effect of the niR-210 in proliferation , autophagy , radioresistance  and mitochondrial alterations  has been reported. Previous animal studies have demonstrated the efficacy of miR-210 targeted treatment by exosome carrier for cerebral ischemia . In addition, miR-210 derived from hypoxic cancer cells may also affect adjacent tumor cells through exosome transposition, thus influencing tumor progression . However, there has been no evidence investigating the impact of miR-210 overexpression on LC angiogenesis. In our study, the miR-210 was reported to be transferred into the CAFs via the carrier of LC cell-derived exosomes. The over-repression of miR-210 could increase the phosphorylation of STAT3 and JAK2, as well as the expression of proangiogenic factors (e.g. MMP-9, FGF2 and VEGFa). In addition, the promotion of CAFs on the tube formation, migration and proliferation of ECs. In vivo studies showed that the level of miR-210 was positively correlated with the MVD of xenografts and invasion ability. Moreover, the xenografts treated by exosomes with overexpressed miR-210 had a larger tumor volume and larger weight. To better illuminate the possible regulatory mechanism of miR-210 in recipient fibroblast, we conducted RNA pull-down and RNA-sequencing between RNA of LC-cell-derived exosomes and pre-miR-210. TET2 was finally identified as the target of miR-210 according to a series of in vitro and in vivo validating assays. It has not been reported before. TET2 is a famous and critical gene that may act as a tumor suppressor in cancers by modulating immunity . The relationship between TET2 and miR-210 remains unclear. In the present study, TET2 has been involved in the process of reprogramming of fibroblast into CAFS and switch of angiogenesis by exosomal miR-210. Editing of TET2 in recipient fibroblast makes it possible for us to interfere the transformation and angiogenesis of CAF which mediated by tumor derived exosome. In clinical sample validation process, serum exosomes were obtained for further study as it is barely not to collect cancer cell secreted exosomes only in patients. Therefore, we compared the exosomal miR-210 level between pre-operation and post-operation samples. We consider it may decrease errors to minimum in this way as a radical resection could largely reduce the LC cell load. The results of our study suggested that the miR-210 could be of great significance in the proangiogenic switch of CAFs through the activation of JAK2/STAT3 signaling pathway, while TET2 was its target. Future studies should focus on whether the LC cell-derived exosomes could promote the transformation of primary miR-210 into the mature miR-210 with full function in CAFs. Besides, additional studies examining whether increased expression of the miR-210 can improve the direct effects of CAFs on the metastasis and proliferation of LC cells.
In conclusion, the miR-210 derived from LC cells could induce the reprogramming of normal fibroblasts into proangiogenic CAFs via the activation of JAK2/STAT3 signaling pathway and it was regulated by TET2 in fibroblasts. In addition, the CAFs could improve the tube formation, migration and proliferation of ECs as well as the expression levels of proangiogenic factors, which in return promoted LC angiogenesis. A new possible molecular mechanism regarding tumor angiogenesis, which could be used as a potential efficient strategy of LC prevention and treatment.
CAFs have been involved in the process of tumor angiogenesis.
Exosomes derived from LC cells could induce cell reprogramming, thus promoting the fibroblasts transferring into CAFs.
Exosomal miR-210 could initiate the CAF proangiogenic switch.
Exosomal miR-210 activates the JAK2/STAT3 signaling and targets TET2 at LC angiogenesis.
The authors declare that there are no competing interests associated with the manuscript.
This work was supported by the General Program of Natural Science Foundation of Zhejiang Province [grant number LY15H160034].
Guanxin Xu and Zhibo Chang wrote the manuscript. Ling Zhu and Jie Yao prepared the figures. Junqiang Fan edited the manuscript.
Dr Zhuojun Zheng and all authors collaborated in the collection and interpretation of the data and contributed to the manuscript.
comparative cycle threshold
fibroblast activation protein
heat shock protein
non-small cell LC
platelet-derived growth factor
ten-eleven translocation 2
tumor susceptibility gene 101
vascular endothelial growth factor
α-smooth muscle actin