Pancreatic cancer cell–derived exosomal microRNA‐27a promotes angiogenesis of human microvascular endothelial cells in pancreatic cancer via BTG2

Abstract Pancreatic cancer (PC) remains a primary cause of cancer‐related deaths worldwide. Existing literature has highlighted the oncogenic role of microRNA‐27a (miR‐27a) in multiple cancers. Hence, the current study aimed to clarify the potential therapeutic role of PC cell–derived exosomal miR‐27a in human microvascular endothelial cell (HMVEC) angiogenesis in PC. Initially, differentially expressed genes (DEGs) and miRs related to PC were identified by microarray analysis. Microarray analysis provided data predicting the interaction between miR‐27a and BTG2 in PC, which was further verified by the elevation or depletion of miR‐27a. Next, the expression of miR‐27a and BTG2 in the PC tissues was quantified. HMVECs were exposed to exosomes derived from PC cell line PANC‐1 to investigate the effects associated with PC cell–derived exosomes carrying miR‐27a on HMVEC proliferation, invasion and angiogenesis. Finally, the effect of miR‐27a on tumorigenesis and microvessel density (MVD) was analysed after xenograft tumour inoculation in nude mice. Our results revealed that miR‐27a was highly expressed, while BTG2 was poorly expressed in both PC tissues and cell lines. miR‐27a targeted BTG2. Moreover, miR‐27a silencing inhibited PC cell proliferation and invasion, and promoted apoptosis through the elevation of BTG2. The in vitro assays revealed that PC cell–derived exosomes carrying miR‐27a stimulated HMVEC proliferation, invasion and angiogenesis, while this effect was reversed in the HMVECs cultured with medium containing GW4869‐treated PANC‐1 cells. Furthermore, in vivo experiment revealed that miR‐27a knockdown suppressed tumorigenesis and MVD. Taken together, cell‐derived exosomes carrying miR‐27a promotes HMVEC angiogenesis via BTG2 in PC.


| INTRODUC TI ON
Pancreatic cancer (PC) ranks as the seventh leading cause of cancer-related death worldwide, accounting for approximately 4% of all cancer cases with about 330,000 deaths per year. 1 PC has also been highlighted as one of the most aggressive malignancies in digestive system, due largely to late diagnoses, high rate of mortality, as well as an increased potential of metastasis and malignancy. 2 Endothelial cells are a type of heterogeneous cell cluster located in the microvascular capillary beds of various organs, among which human microvascular endothelial cells (HMVECs) are unique yet unrecognizable in phenotype, function and structure. 3 Angiogenesis has been widely implicated in the process of tumour development. 4 Moreover, exosomes derived from PC cells trigger the acceleration of angiogenesis in HMVECs via a dynamin-dependent endocytosis process. 5 Exosomes are extracellular vesicles with a size of 40-150 nm play a functional role in PC therapy. 6 Exosomes secreted by tumours can assume the role of carriers to deliver microRNAs (miRs) into recipient cells in the cancer microenvironment, which has been reported to enhance tumour metastasis and invasion. 7 Both miRs and miR exosomes have been speculated to be promising PC therapeutic targets through the protein transfer-induced target cell reprogramme and binding-triggered signal transduction. 8 miRs are a group of short non-coding RNA molecules with the length of 18-22 nucleotides, which regulate genes associated with cell apoptosis, differentiation and neoplastic transformation. 9 The abnormal expression of miRs has been implicated in the development and progression of a large array of cancers, including that of PC. 10 Additionally, the targeting of miR-27a by the anticancer agent 2-cyano-3,12-dioxooleana-1,9-dien-28-oic acid (CDDO) has been reported to diminish its expression leading to an inhibition of tumour cell growth through anti-angiogenic responses. 11 Existing literature has indicated that miR-27a directly targets B-cell translocation gene 2 (BTG2) in gastric cancer cells. 12 BTG2 is an immediate early response protein that has been shown to play a role in DNA damage repair, antiproliferation, cell apoptosis and differentiation, whose elevated level has been identified in the majority of normal tissues, including the pancreas. 13 BTG2 is regarded as a tumour inhibitor in pancreatic ductal adenocarcinoma and has been implicated in the progression and growth of PC under the control of miR-21, miR-23a and miR-27a. 14 Based on the aforementioned data, we examined the hypothesis that PC cell-derived exosomes carrying miR-27a influence the angiogenesis of HMVEC through BTG2 in PC. Hence, this study indented to investigate the underlying molecular mechanisms associated with the progression of PC to elucidate more effective therapeutic strategies for PC patients through the use of PC cell-derived exosomes co-cultured with HMVECs.

| Ethics statement
The study was performed in strict accordance with the recommen-

| Microarray-based gene expression profiling
Differentially expressed genes (DEGs) were screened by downloading PC-related microarray data (GSE22780, GSE16515, GSE32676 and GSE91035) from the Gene Expression Omnibus (GEO) database (https ://www.ncbi.nlm.nih.gov/geo/) ( Table 1). The pre-treatment standardization on gene expression data was performed with the Affy package of R language, 15   predict the miRs that regulate the DEGs, the results of which were compared using jvenn. In addition, two additional PC-related miR data sets (GSE41369 and GSE28955) were analysed to screen differentially expressed miRs.

| Immunohistochemistry
The paraffin-embedded sections of tissue specimens were dehydrated using gradient ethanol and treated with water-bath repair in antigen retrieval buffer.

| Cell treatment
The human pancreatic ductal cell line H6c7, PC cell lines (SW1990,

| 5-Ethynyl-2'-deoxyuridine (EdU)
Cells at the logarithmic growth phase were seeded into a 96-well plate at the density of 2 × 10 3 -4 × 10 4 cells/well. Transfection was performed 24 hours after the cells had adhered to the wall, with 3 duplicate wells set for each treatment. After 48 hours of transfection, cells were treated with EdU labelling 18 with antifluorescence quenching mounting medium added at a density of 100 μL per well.
The images were photographed under a fluorescence microscope.
Cells with red-stained nucleus were regarded as the positively labelled cells. Finally, 3 power fields were randomly selected to count the number of positive and negative cells under the guidance of a microscope.

| Transwell assay
The pre-cooled Matrigel was diluted using serum-free Dulbecco's

| Exosome extraction and identification
The PANC-1 cells were seeded into a 6-well plate at the density of Later, each part was added with CD63-polyethylene (PE) antibody and incubated at room temperature for 30 minutes as per the provided antibody instructions. The cells that were not added with an antibody were regarded as the blank control, and the PE-labelled anti-human IgG was regarded as the isotype control. The samples were then loaded followed by detection using a Guava easyCyte™ flow cytometry system. Reverse transcription-quantitative polymerase chain reaction (RT-qPCR) was performed in order to determine the expression of miR-27a in the exosomes.

| Co-culture of exosomes and HMVECs
The exosomes dissolved with PBS were mixed with Exo-Red

| RT-qPCR
Total RNA was extracted from the tissues or cells in strict accordance with instructions on the TRIzol Kit (15596-018, Beijing Solarbio Life Sciences Co Ltd). All primers ( Table 2)

| Enzyme-linked immunosorbent assay (ELISA)
An ELISA kit was employed to determine the VEGF level according to the instructions provided by the kit (Imunbio). The known antigen was diluted by carbonate-coated buffer (pH = 9.6) into 1-10 μg/mL and added into a 96-well plate. Next, 0.1 mL diluent was added to each well and allowed to stand at 4℃ overnight. The

| Tumour xenograft in nude mice
A total of 48 female mice with immune deficiency (aged 4-6 weeks,

| Immunohistochemistry for detecting microvessel density (MVD)
Microvessel density was determined in the tumour tissues of the nude mice based on the methods used during immunohistochemistry. The primary rabbit antimouse antibody (sc-376975) against MVD was purchased from Santa Cruz Biotechnology. The MVD results were evaluated using the Weidner microvessel count method. 20 Low-power lens (4 × 10) was first used to observe and find the area with the most microvessels, and medium-power lens (20 × 10) was utilized to determine number of vessels stained brown. Five power fields were randomly selected from each pathological section to calculate the mean value as MVD. In the event the mean value ≥MVD threshold, vessels were regarded as MVD positive; however, if the result was opposite to the threshold, the vessels were considered to be MVD negative.

| Statistical analysis
All experimental statistical analyses were conducted using SPSS

| miR-27a might regulate PC by targeting BTG2
R language was applied for analysis of the PC-related microarray data (GSE22780, GSE16515, GSE32676 and GSE91035), and the comparison of the top 400 DEGs from these data sets was conducted based on |log2FoldChange| > 1.0 and P value < .05 (Table S2). Afterwards, the Venn map was plotted ( Figure 1A).
Altered BTG2 expression was identified in these four data sets.
The expression heat map of the top 80 DEGs from GSE32676 ( Figure S1) and GSE16515 ( Figure S2) revealed lower levels of BTG2 in the PC tissues in comparison with the normal tissues.
Based on the expression profile of GSE22780 and GSE91035, poor expression of BTG2 was witnessed in the PC tissues ( Figure 1B&C). After analysis of GEPIA, low BTG2 levels were identified in the PC tissues ( Figure 1D). The survival analysis found that PC patients with lower levels of BTG2 presented poor disease-free survival ( Figure 1E). miRWalk, TargetScan, mirDIP, DIANA and starBase were applied to predict regulatory miRs of BTG2. There were 1173 miRs from miRWalk based on energy < −20, 68 miRs from TargetScan, 21 miRs from miRDIP with the screening threshold set as integrated score > 0.8 and 37 miRs from DIANA according to miTG score > 0.9 (Table S3). Based on the prediction results of the Venn map, there were 4 intersection miRs (hsa-miR-27a-3p, hsa-miR-92a-3p, hsa-miR-27b-3p and hsa-miR-92b-3p), which were regarded as the target sources of BTG2 ( Figure 1F). Accordingly, 62 and 172 differential miRs were selected from PC-related miR microarray data (GSE41369 and GSE2895) ( Table S4). The differential expression analysis on microarray data revealed that only hsa-miR-27a-3p was highly expressed in PC among 4 regulatory miRs of BTG2. The expression heat map of GSE41369 is shown in Figure S3, and the expression of hsa-miR-27a in GSE28955 is illustrated in Figure 1G. Based on the aforementioned results, we speculated that miR-27a could regulate BTG2 in PC.

| miR-27a negatively regulates BTG2 in PC cells
RT-qPCR was performed to detect the expression of miR-27a in PC. The results revealed an overexpression of miR-27a expression in PC tissues compared with pancreatic tissues obtained from patients with pancreatitis (P < .05; Figure 2A). Next, RT-qPCR was  Figure 2B). The aforementioned results revealed high expression of miR-27a in both the PC tissues and cells.
Immunohistochemistry was performed to evaluate the positive level of BTG2 in PC. The results indicated that the positive level of BTG2 was predominately located in the cytoplasm and was brownish in appearance in PC ( Figure 2C). BTG2-positive level was lower in PC tissues than in pancreatic tissues obtained from patients with pancreatitis (P < .05; Figure 2D). Online analysis software found a specific binding region between BTG2 and miR-27a sequences ( Figure 2E)

| miR-27a depletion inhibits proliferation and invasion but promotes apoptosis of PC cells
The PANC-1 cells were treated with miR-27a mimic or inhibitor in order to alter miR-27a expression. Next, PC cell proliferation, invasion and apoptosis were determined through the application of an EdU assay, Transwell assay and flow cytometry, respectively.
The results revealed that inhibitor-NC and mimic-NC had no significant effect on proliferation, invasion and apoptosis of PC cells ( Figure S4). Also, we observed an increase in red positive cell pro- the treatment of miR-27a mimic, which was accompanied by a reduction after treatment with a miR-27a inhibitor ( Figure 3G&H).
Thus, miR-27a down-regulation was concluded to induce the inhibition of cell proliferation and invasion, and accelerated PC apoptosis.

| miR-27a down-regulation suppresses cell proliferation and invasion and enhances apoptosis in PC by up-regulating BTG2
BTG2-overexpressed plasmids were introduced into the PANC-1 cells for interference of BTG2 level along with miR-27a mimic.
Next, EdU assay, Transwell assay and flow cytometry were per-  The exosomes were extracted from the serum of nude mice, after which the expression of miR-27a in the exosomes was detected by RT-qPCR, the results of which revealed that miR-27a overexpression led to an up-regulation of miR-27a, and miR-27a depletion was found to down-regulate miR-27a ( Figure 5D).    F I G U R E 4 Depletion of miR-27a up-regulates BTG2 to restrain cell proliferation and invasion and to promote apoptosis in PC. A and B, Positive cell proliferation in PC after the treatment of BTG2 plasmid and the co-treatment of miR-27a mimic and BTG2 plasmid detected by EdU (200×). C and D, Cell invasion in PC after the treatment of BTG2 plasmid and the co-treatment of miR-27a mimic and BTG2 plasmid (200×). E and F, Cell apoptosis in PC after the treatment of BTG2 plasmid and the cotreatment of miR-27a mimic and BTG2 plasmid. G and H, Protein levels of VEGF, VEGFR, MMP-2 and MMP-9 after the treatment of BTG2 plasmid and the cotreatment of miR-27a mimic and BTG2 plasmid. *P < .05 vs PANC-1 cells treated with BTG2-NC or both miR-27a mimic and BTG2. The above data are measurement data and described as mean ± standard deviation. Comparisons among multiple groups are analysed by one-way analysis of variance. The experiment was repeated 3 times independently. miR-27a, microRNA-27a; BTG2, B-cell translocation gene 2; PC, pancreatic cancer; NC, negative control; EdU, 5-ethynyl-2'deoxyuridine; VEGF, vascular endothelial growth factor; VEGFR, vascular endothelial growth factor receptor; MMP, matrix metallopeptidase or the delivery of Exo-depl (P > .05). Compared with the treatment without exosomes, the delivery of PANC-1-exo and miR-27a mimic led to an increased level of VEGF, which was further heightened following the delivery of miR-27a mimic (P < .05; Figure 6F). RT-qPCR and Western blot analysis were utilized to determine the expression of miR-27a and BTG2 in HMEC-1 cells, which exhibited F I G U R E 5 Inhibition of tumour growth and angiogenesis of PC in vivo are induced by miR-27a silencing. A, Tumorigenesis of PANC-1 cells following the injection of lentivirus with miR-27a mimic and miR-27a inhibitor. B and C, Tumour volume and weight after alteration of miR-27a. D, The expression of miR-27a in serum exosomes of nude mice after alteration of miR-27a. E and F, Protein levels of BTG2 and VEGF in tumour tissues of nude mice after alteration of miR-27a. G and H, MVD in tumour tissues of nude mice after alteration of miR-27a (400×). *P < .05 vs nude mice injected with lentivirus with miR-27a mimic-NC or miR-27a inhibitor-NC. The above data are measurement data and described as mean ± standard deviation. The data of Panel B are analysed by repeated measurement ANOVA; data of other panels are analysed by one-way analysis of variance. n = 12. The experiment was repeated 3 times independently. miR-27a, microRNA-27a; BTG2, Bcell translocation gene 2; PC, pancreatic cancer; NC, negative control; MVD, microvessel density; VEGF, vascular endothelial growth factor

| D ISCUSS I ON
Despite the progress achieved in the diagnosis and therapy of PC over the past few decades, PC remains one of the most lethal conditions, accompanied by a heavy health burden as well as a high mortality rate. 21 More recently, cell-secreted exosomes which are important regulators of cell-cell communication have been identified as potentially useful tools in gene therapy and drug delivery, and miRs have been reported to be promising therapeutic targets for the treatment of various human diseases. 22 Hence, the present study investigated the effect of PC cell-derived exosomal miR-27a on PC through the regulation of BTG2. On the whole, this study suggested that PC cell-derived exosomal miR-27a knockdown played an inhibitory role in the angiogenesis of HMVEC in PC by up-regulating BTG2.
One of the important findings from our study revealed that both the PC tissues and cell lines exhibited high expression levels of miR-27a, accompanied by the poor expression of BTG2. The obtained results suggest that miR-27a targets and negatively regulated BTG2. Consistent with this finding, Ma Y et al demonstrated that miR-27a serves as an oncogene in various cancers, and its expression had been found to be up-regulated in PC. 23 A recent study concluded that the down-regulation of miR-27a results in the inhibition of cell growth and invasion as well as the promotion of apoptosis. 24 In addition, BTG2 was verified as the target gene of miR-27a following the detection of luciferase activity and quantification. BTG2 has been reported to be a cancer inhibitor gene and exhibits low levels of expression in tumour tissues, the inhibition of which has been linked with metastasis, invasion and migration of tumour cells. 25 Meanwhile, a prior study has also provided an insight suggesting that there is a decrease in BTG levels in pancreatic ductal adenocarcinoma, indicating that BTG2 is the direct target of cooperative miR-2a, miR-23a and miR-27a. 14 All of the aforementioned findings were supporting evidence that miR-27a negatively regulated BTG2 in PC.
Moreover, our study also revealed that the down-regulation of miR-27a and up-regulation of BTG2 resulted in the inhibition of PC cell proliferation, migration, invasion and angiogenesis and enhancement of cell apoptosis, corresponding to diminished levels of MVD as well as angiogenesis growth factor (VEGF and VEGFR) and invasionrelated factors (MMP-2 and MMP-9). VEGF and VEGFR represent important tissue factors for tube formation and the differentiation of angioblast, all of which have been widely reported to play crucial roles in the progression of PC. 26 VEGF in particular has been shown to be a critical angiogenesis growth factor that facilitates and maintains vascular endothelial cell proliferation and growth. 27 VEGFs have been shown to have an impact on a large array of cellular processes and initiate the activation of a cascade of downstream signalling pathways, and they do this by interacting with the kinase domain of VEGFRs. 28 MVD has been correlated with the activity of angiogenesis in tumour formation. 29 MMPs, which have also been identified as potential cancer biomarkers, have been speculated to facilitate the disintegration of extracellular matrix components with studies implicating their activity in cancer cell metastasis and invasion. 30 The up-regulation of both MMP-2 and MMP-9, both of which have been shown to play an active role in malignant cell invasion, has been highlighted in PC. 31,32 Ma Y et al concluded that miR-27a knockdown contributes to the inhibition of cell growth, migration and colony formation in PC. 23 In addition, miR-27a accelerated cancer stem cell differentiation in order to promote angiogenesis in breast cancer. 33 Elevated levels of BTG2 have been implicated in the inhibition of cell proliferation and invasion, along with the F I G U R E 7 PANC-1-exo carrying miR-27a contributes to acceleration of HMEC-1 cell proliferation, invasion and angiogenesis but inhibition of apoptosis. A and B, Positive cell proliferation in HMEC-1 cells in response to the treatment of Exo-depl, PANC-1-exo and miR-27a mimic detected by EdU (200×). C and D, Cell invasion in HMEC-1 cells in response to the treatment of Exo-depl, PANC-1-exo and miR-27a mimic (200×). E and F, Angiogenesis of HMEC-1 cells in response to the treatment of Exo-depl, PANC-1-exo and miR-27a mimic (100×). G and H, Cell apoptosis in HMEC-1 cells in response to the treatment of Exo-depl, PANC-1-exo and miR-27a mimic. I and J, Protein levels of VEGF, VEGFR, MMP-2 and MMP-9 in response to the treatment of Exo-depl, PANC-1-exo and miR-27a mimic. *P < .05 vs the treatment without exosomes or the delivery of Exo-depl. The above data are measurement data and described as mean ± standard deviation. Comparisons among multiple groups are analysed by one-way analysis of variance. The experiment was repeated 3 times independently. miR-27a, microRNA-27a; exo, exosome; EdU, 5-ethynyl-2'-deoxyuridine; VEGF, vascular endothelial growth factor; VEGFR, vascular endothelial growth factor receptor; MMP, matrix metallopeptidase enhancement of apoptosis of MDA-MB-231 human triple-negative breast cancer cells. 34 The PC cells and HMVEC co-culture system results revealed that PC cell-derived exosomes delivering miR-27a act to promote the proliferation, invasion and angiogenesis of HMVEC. There have been multiple cell-secreted exosomal miRs that have linked to metastasis initiation, resistance to drugs as well as tumour growth. 35 The co-culture of exosomes and HMVEC has been found to induce endothelial dysfunction in patients suffering from acute chest syndrome. 36 Exosomes secreted by cancer cells are widely known with the ability to enter the tumour circulation and microenvironment, highlighting the potential of exosomal miRs as promising biomarkers capable of facilitating the improvement of cancer detection. 37 The aberrant expression of exosomal miR-21 has been detected in PC, which may serve as an early diagnostic marker for PC. 38 Moreover, exosomes containing miR-27a play a functional role in osteosarcoma with different chemotherapy sensitivity. 39 The overexpression of exosomal miR-17-5p has been linked with metastasis and deterioration of PC into advanced stage. 40 Exosomes derived from breast cancer MDA-MB-231 cells have also been found to promote cell proliferation, invasion and migration in breast cancer. 41 Taken together, this study indicates that PC cell-derived exosomes carrying miR-27a have the potential to be a PC therapeutic biomarker due to their promotive effect on the angiogenesis of HMVECs in PC (Figure 8). The aforementioned findings provide evidence that miR-27a can function as alternative target in therapy of PC. However, owing to the limitation of such objective conditions as time periods and experimental expenditures, future large-scale studies are required to further elucidate the mechanisms identified in our study, based on which exosomal miRs can be applied to improve treatment outcome of PC.

ACK N OWLED G EM ENT
We would like to give our sincere appreciation to the reviewers for their helpful comments on this article.

CO N FLI C T O F I NTE R E S T
The authors declare no conflicts of interest.

AUTH O R CO NTR I B UTI O N S
DS and CX designed the study. JH collated the data, designed and developed the database. ZSL and ZYY carried out data analyses and produced the initial draft of the manuscript. JRT and YFY contributed to drafting the manuscript. All authors have read and approved the final submitted manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.