Exosomal circNFIX promotes angiogenesis in ovarian cancer via miR‐518a‐3p/TRIM44 axis

Abstract

Ovarian cancer (OC) is a gynecological cancer with high mortality. OC‐derived exosomal circRNAs can regulate angiogenesis. This study aims to explore the role and mechanism of exosomal circRNA nuclear factor I X (CircNFIX) derived from OC cells in angiogenesis. Quantitative real‐time polymerase chain reaction was employed to evaluate the levels of circNFIX, miR‐518a‐3p, and tripartite motif protein 44 (TRIM44) in OC and adjacent tissues. Exosomes from the ovarian surface epithelial cell (HOSEpiC) and OC cells (SKOV3 or OVCAR3) were isolated by differential centrifugation. Exosomes were cocultured with the human umbilical vein endothelial cells (HUVECs). The angiogenesis capacity was analyzed by Tube formation assay. 3‐(4,5‐dimethyl‐2‐thiazolyl)‐2,5‐diphenyl‐2‐H‐tetrazolium bromide (MTT) and Transwell assays were used to determine the cell viability and migration ability. The dual‐luciferase report, RNA immunoprecipitation (RIP), and RNA pull‐down assays were applied to validate the gene's interaction. CircNFIX and TRIM44 expression were higher and miR‐518a‐3p was lower in OC tissues than in the adjacent tissues. Upregulated circNFIX and TRIM44 were significantly correlated with the tumor size and International Federation of Gynecology and Obstetrics (FIGO) stage of OC patients. HUVECs treated OC‐derived exosomes had higher proliferation, migration, and angiogenesis capacities than the control group. While OC‐derived exosomal circNFIX silencing restrained HUVECs' proliferation, migration, and angiogenesis, compared with the OC‐derived exosomes group. OC‐derived exosomal circNFIX positively regulated TRIM44 expression by targeting miR‐518a‐3p in HUVECs. OC‐derived exosomal circNFIX promoted angiogenesis by regulating the Janus‐activated kinase/signal transducer and activator of transcription 1 (JAK/STAT1) pathway via miR‐518a‐3p/TRIM44 axis in HUVECs.

1. INTRODUCTION

Ovarian cancer (OC) is the principal gynecological cancer, ranking the fifth most common cancer in women. As reported, it can induce approximately 152,000 deaths per year globally. ¹ The angiogenesis is the critical biological process of OC's nutrients and oxygen supply, which are tightly linked with the poor prognosis of OC. ² Studies have revealed that OC can pose a high level of angiogenesis. Moreover, its specific inhibition can repress the tumor development and increase the survival rate of OC patients. ³ However, the inhibition of the angiogenesis also has limitations including side effects or low efficacy. Therefore, it is urgent to identify the mechanisms of angiogenesis to obtain a greater therapeutic effect of OC.

One of the vital pathways for tumor tissue progression or metastasis is the intercellular communication between tumor cells and the microenvironment. ⁴ Tumor cell‐derived exosomes are the typical extracellular vesicles that regulate the biological processes of the tumor microenvironment. As for OC, exosomes can modulate the angiogenesis and microenvironmental reprogram process of OC and promote the tumorigenesis and metastatic niche becoming processes. ⁵ Especially, the exosomal circular RNAs (circRNAs) are the crucial element of the communication system between the tumor and host cells, participating in the tumor progression. For example, the OC‐derived exosomal circRNA051239 and circRNAFoxo3 can accelerate the tumor progression by promoting the proliferation and migration of epithelial OC cells. ⁶ , ⁷ A more extensive study still needs to be explored to illuminate the deeper mechanisms of OC‐derived exosomal circRNAs, such as the OC angiogenesis.

Exosomal circRNA nuclear factor I X (CircNFIX) can boost the tumor tissue progression of hepatocellular carcinoma. ⁸ Moreover, circNFIX silencing can enhance cardiac regenerative repair to promote the potential recovery from myocardial infarction. ⁹ However, the role of circNFIX in the angiogenesis of tumor microenvironment in OC remains unclear. Further, the interaction of circRNAs and microRNAs (miRNAs) can crucially regulate the pathogenesis of cancer. ¹⁰ For example, studies highlighted that circNFIX could sponge the miR‐3064‐5p or miR‐212‐3p to promote the tumor malignant progression. ⁸ , ¹¹ And the decreased miR‐518a‐3p possibly participated in the angiogenesis of placenta accreta spectrum disorder. ¹² But the precise role of miR‐518a‐3p in the angiogenesis and OC is unknown now and needs to be clarified. Of interest, based on online bioinformation prediction, we observed that circNFIX harbored the potential binding sites of miR‐518a‐3p. While the detailed mechanism of circNFIX and its regulation of miR‐518a‐3p in the angiogenesis of OC should be further explored.

Tripartite motif protein 44 (TRIM44) belongs to the tripartite motif proteins and exerts the modulatory effects in tumor progression. High expressed TRIM44 can be the indication of poor prognosis of epithelial OC. ¹³ TRIM44 inhibition repressed the malignant performance and angiogenesis of OC. ¹⁴ Furthermore, the regulations between miRNAs and mRNA expression have been validated in the OC procession. MiRNAs, such as miR‐34a‐5p, were confirmed that they could be targeted to modulate the TRIM44 expression to boost the malignant behavior of OC cells. ¹⁵ According to the online bioinformation prediction, we also observed that miR‐518a‐3p harbored the possible binding sites of TRIM44. While the exact mechanism of their interaction in tumor progression and OC‐related tumor microenvironment angiogenesis remains unclear.

The Janus‐activated kinase (JAK)/signal transducer and activator of transcription 1 (STAT1) pathway has been involved in multiple kinds of cancers including OC. The activated JAK/STAT1 signaling pathway is involved in promoting angiogenesis in the OC tumor microenvironment. ¹⁶ The activated JAK/STAT1 pathway can modulate the intensive proliferation and invasion of epithelial OC cells. ¹⁷ In addition, TRIM family proteins can be an important factor in the JAK/STAT pathway in cell proliferation and tumor cytokine action mechanisms. ¹⁸ For example, TRIM66 can promote the JAK/STAT pathway expression to accelerate the malignant progression of prostate carcinoma. ¹⁹ However, the interaction mechanism between TRIM44 and JAK/STAT1 signaling pathway and their role in OC angiogenesis are still unclear.

Based on the above, we aimed to explore the role of OC‐derived exosomal circNFIX in the angiogenesis of HUVECs. According to our findings, OC‐derived exosomal circNFIX regulated the JAK/STAT1 pathway via the miR‐518a‐3p/TRIM44 axis in HUVECs, promoting cell proliferation, migration, and angiogenesis capacities. It demonstrated that exosomal circNFIX could be further developed as a target for restraining the OC progression.

2. MATERIAL AND METHODS

2.1. Tissue specimens

OC and normal adjacent tissue specimens were collected from 25 patients who received surgical resections at West China Second University Hospital, Sichuan University. Before the surgery, no participant had received anti‐tumor therapies (such as radiotherapy or chemotherapy), and no one was diagnosed with other types of cancers. Tissue specimens were maintained at −80°C. Each participant offered written informed consent, before the surgery. All procedures were carried out keeping in line with that approved by the Ethics Committee of West China Second University Hospital, Sichuan University. tissue specimens were collected from 25 patients who received surgical resections at West China Second University Hospital, Sichuan University. Before the surgery, no participant had received anti‐tumor therapies (such as radiotherapy or chemotherapy), and no one was diagnosed with other types of cancers. Tissue specimens were maintained at −80°C. Each participant offered written informed consent, before the surgery. In addition, the correlation between genes (circNFIX, miR‐518a‐3p, and TRIM44) expression and clinicopathological characteristics of OC patients including age, tumor size, International Federation of Gynecology and Obstetrics (FIGO) stage, histological grade, and lymph node metastasis condition is analyzed in Table 1. All procedures were conducted following the procedures approved by the Ethics Committee of West China Second University Hospital, Sichuan University.

TABLE 1.

Association between circNFIX expression and the clinicopathological characteristics of patients with ovarian cancer

Clinical characteristics	circNFIX		p Value	miR‐518a‐3p		p Value	TRIM44		p Value
	High (n = 14)	Low (n = 11)		High (n = 11)	Low (n = 14)		High (n = 14)	Low (n = 11)
Age (years)			0.6951			0.1160			0.2377
<50	6	6		7	4		9	4
≥50	8	5		4	10		5	7
Tumor size (cm)			0.0472*			0.6951			0.0472*
<3	4	8		5	8		4	8
≥3	10	3		6	6		10	3
FIGO stage			0.0048**			0.2203			0.0419*
I–II	3	9		7	4		5	9
III–IV	11	2		4	8		9	2
Histological grade			0.4139			>0.999			0.2406
G1–G2	7	8		7	9		10	5
G3	7	3		4	5		4	6
Lymph node metastasis			0.2377			0.2406			0.4283
Negative	5	7		5	10		8	4
Positive	9	4		6	4		6	7

Note: Data were analyzed using Fisher's exact test. *p < 0.05; **p < 0.01.

Abbreviation: FIGO, International Federation of Gynecology and Obstetrics.

2.2. Cell culture and treatment

Ovarian surface epithelial cells (HOSEpiC), human embryonic kidney cells (HEK‐293 T), human umbilical vein endothelial cells (HUVECs, the 5th passage of HUVECs were used for experiments), and OC cell lines (SKOV3 or OVCAR3) were obtained from the American Type Culture Collection (Manassas, VA). Cells were maintained in Dulbecco's modified eagle medium (DMEM, Gibco BRL, Indianapolis, IN) supplemented with the fetal bovine serum (10% FBS, Gibco BRL), penicillin G sodium (100 units/ml), and streptomycin sulfate (100 mg/ml). The cells were cultured at a temperature of 37°C and 5% CO₂, in a humidified incubator.

2.3. Isolation and identification of exosomes

Exosomes from SKOV3 and OVCAR3 cells were isolated by differential centrifugation, according to the relevant paper. ²⁰ The cells were cultured in the DMEM supplemented with 10% FBS medium (without exosome) for 48 h. Further, the medium was centrifugated at 13,000g for 0.5 h and 100,000g for 1 h to get the fraction enriched exosomes for the subsequent analysis. The transmission electron microscopy (TEM), flow nano analysis, and western blotting analysis were applied to identify the exosomes. Morphologies analysis with the HT7700 TEM (Hitachi, Tokyo, Japan) was performed using the 100 μl of phosphate buffer saline (PBS) with resuspended exosomes. Flow nano analyzer (NanoFCM, MediCity, UK) was used for the analysis of particle sizes of exosomes (diluted 1:100). Lasers were calibrated by 200 nm control beads and various sizes of beads to set the reference, based on the manufacturer's instructions and previously research. ²⁰ Finally, exosomes were dissolved in the RIPA (radioimmunoprecipitation) buffer (100 μl) for the subsequent western blotting analysis, in which primary antibodies included: anti‐CD63 (1:1000, ab134045, Abcam, Cambridge, MA), anti‐CD81 (1:1000, ab79559, Abcam), and anti‐GM130 (1:1000, ab283685, Abcam).

2.4. Exosome uptake assay in vitro

The exosome uptake assay was conducted to observe the uptake of HUVECs for the OC‐derived exosomes. Exosomes were labeled with PKH26 dye (Sigma, St. Louis, MO) and incubated at 37°C for 30 min. The suspension was centrifuged at 300g for 15 min, washed twice with PBS, and incubated for 48 h. Then the HUVECs were incubated with exosomes for another 3 h. After that, HUVECs were washed with PBS, labeled with DAPI (4',6‐diamidino‐2‐phenylindole, Invitrogen, Carlsbad, CA), and incubated with 4% paraformaldehyde for 15 min. The uptake of labeled exosomes was observed by the laser confocal microscope (Leica TCS SP5 II, Shanghai, China).

2.5. Cell transfection

The small interfering RNAs targeting circNFIX (si‐circNFIX or si‐NC) were obtained from GenePharma (Shanghai, China). MiR‐518a‐3p mimics, miR‐518a‐3p inhibitor, and their negative control were also purchased from GenePharma. Synthesized total TRIM44 cDNA was cloned into the vector pcDNA3.1 (Invitrogen). Based on the kits' accessory instruction, cells were transfected with the above plasmids by the Lipofectamine 2000 (Invitrogen). After 48 h, cells were collected for subsequent analysis. For the inhibition of the JAK/STAT1 signaling pathway, the HUVECs were pretreated with the JAK inhibitor ruxolitinib (1 μM, HY‐50856, MCE, Shanghai, China) for 2 h prior to relevant detection.

2.6. 3‐(4,5‐dimethyl‐2‐thiazolyl)‐2,5‐diphenyl‐2‐H‐tetrazolium bromide assays

Cells (5 × 10⁴ cells/well) were seeded into 96‐well plates for 24 h. For the OC‐derived exosomes treatment, HUVECs were treated with the same number of OC‐derived exosomes (100 μg/ml) for 48 h. Then, cells were fixed with 3‐(4,5‐dimethyl‐2‐thiazolyl)‐2,5‐diphenyl‐2‐H‐tetrazolium bromide (MTT) (10 μl, 5 mg/mL, Solarbio) for 4 h, then Formazan solution (100 μl) was supplemented in the well, based on the manufacturer's protocol. Cell viability was evaluated by reading the absorbance at 490 nm on a microplate reader (Molecular Devices, Sunnyvale, CA).

2.7. Transwell assays

HUVECs (1 × 10⁵ cells/well) were maintained in the upper chamber for 24 h. And the HUVECs were treated with the same number of OC‐derived exosomes (100 μg/ml) for 48 h. Then the upper chamber was added with DMEM medium without serum (100 μl), and the lower chamber was fixed with DMEM with serum medium (600 μl) for chemoattraction. After 24 h of incubation, cells on the lower surface were added 4% paraformaldehyde and stained with 0.1% crystal violet (Sigma, USA) for 15 min and followed fixation. Image and count of stained HUVEC cells were conducted in five random microscopy (Olympus, Tokyo, Japan) zones.

2.8. Tube formation assay

In vitro Angiogenesis Assay Kit (Millipore, Billerica, MA) was conducted to explore the angiogenesis of HUVECs, keeping with the appendant protocol. HUVECs were seeded in the 96‐well plates and serum‐starved in cell medium for 48 h. Then, 50 μl of ECMatrix™ (containing growth factor‐reduced Matrigel) with dilution was added per well and carried the incubation for 1 h at 37°C. HUVECs (2 × 10⁴ per well) and PBS with exosomes (100 μg/ml) or negative control PBS were added to the plates for the subsequent incubation for 48 h at 37°C. Then, tube formation was observed by the microscope (Olympus) and the total tube length was calculated by the Image‐Pro Image Analysis Software (Media Cybernetics, Rockville, MD).

2.9. Dual‐Luciferase reporter assay

The potential binding sites between circ‐NFIX and miR‐518a‐3p or miR‐518a‐3p and TRIM44 were predicted by Starbase (http://starbase.sysu.edu.cn/). The fragment of wild‐type (WT)‐circNFIX containing the binding site or its mutant (MUT) was subcloned into the firefly luciferase gene of the pmirGLO‐luciferin Enzyme vector (Promega, Madison, WI) to establish circNFIX‐WT or circNFIX‐MUT. Similarly, TRIM44‐WT and TRIM44‐MUT were established. HEK‐293 T cells were co‐transfected with the circNFIX‐WT/MUT, TRIM44‐WT/MUT, miR‐518a‐3p mimics, or mimics NC. After 48 h, the luciferase activity was determined using the dual luciferase reporter system (Promega).

2.10. RNA immunoprecipitation assay

RNA immunoprecipitation (RIP) analysis was conducted using the Magna RNA‐binding protein immunoprecipitation kit (Millipore) to validate the interaction of circNFIX and miR‐518a‐3p, according to the previous research. ²¹ HEK‐293T cells were lysed in RIP buffer and the cell extraction (20 μg protein) interacted with magnetic beads coated by antibodies immunoglobulin G (IgG, ab32381, Abcam) or Argonaute‐2 (Ago2, ab6702, Abcam) for 1 h at 4°C. Then the protease K (Sangon Biotech, Shanghai, China, #A004220) was used for the protein digestion and isolation of immunoprecipitated RNA. The relative RNA expression levels of circNFIX and miR‐518a‐3p were determined by qRT‐PCR.

2.11. RNA pull‐down assay

The biotinylated miR‐518a‐3p probes (Bio‐miR‐518a‐3p sense) or Bio‐miR‐518a‐3p antisense, and biotinylated nucleotides (Bio‐NC) were obtained from GenePharma. MiR‐518a‐3p or antisense‐miR‐518a‐3p RNAs were transcribed and labeled by the Biotin RNA Labeling Mix (Roche, Basel, Switzerland), treated with RNase‐free DNase I (Takara, Tokyo, Japan), and purified with the RNeasy Mini Kit (Qiagen, Shanghai, China). Then, the biotinylated miR‐518a‐3p probe was used for incubation with streptavidin agarose beads (Invitrogen) for 12 h at 4°C. HEK‐293 T cells were sonicated and lysed. Cell lysate and the biotinylated miR‐518a‐3p were incubated for 12 h at 4°C. Streptavidin bead‐RNA‐protein complexes were obtained by separation with a magnetic field. RNA complex bound to the beads was eluted by wash buffer, and quantitative real‐time polymerase chain reaction (qRT‐PCR) was used to measure circNFIX, vascular endothelial growth factor A (VEGFA), or TRIM44 enrichment pulled down by the miR‐518a‐3p probe.

2.12. Quantitative real‐time polymerase chain reaction

HUVECs were treated with the same amount of OC‐derived exosomes (100 μg/mL) for 48 h. Then the gene expression in HUVECs was evaluated. Total RNA Extraction Kit (Solarbio, Beijing, China) was used to extract the total RNA of tissues and cells, and then the RNA was converted to cDNA by Hifair® II 1st Strand cDNA Synthesis Kit (Yeasen, Shanghai, China) as described by the appendant protocol. Analysis of the mRNAs' expression was carried out by qRT‐PCR, using the Real‐Time PCR System (Bio‐Rad, Hercules, CA) with Hieff® SYBR Green Master Mix (Yeasen). Glyceraldehyde‐3‐phosphate dehydrogenase (GAPDH) was the internal reference gene for circ‐NFIX and TRIM44 or VEGFA, respectively. And U6 was selected as the internal reference for miR‐518a‐3p. The relative gene level was calculated by the 2^−∆∆Ct method. The primer sequences were as follows:

Circ‐NFIX‐F: 5′‐CCTCAGTGCTCGAACCCC‐3′;

Circ‐NFIX‐R: 5′‐CGATGAACGGGTGGAACTCA‐3′;

miR‐518a‐5p‐F: 5′‐ACAGGCCGGGACAAGTGCAATA‐3′;

miR‐518a‐5p‐R: 5′‐GCTGTCAACGATACGCTACGTAACG‐3′;

TRIM44‐F: 5′‐ AGGCAGCTCATCTGTGTCCT‐3′;

TRIM44‐R: 5′‐GCCTTCAGTCCACCTGAGTC‐3′;

VEGFA‐F: 5′‐ATCCAATCGAGACCCTGGTG‐3′;

VEGFA‐R: 5′‐ATCTCTCCTATGTGCTGGCC‐3′;

U6‐F: 5′‐CTCGCTTCGGCAGCACT‐3′;

U6‐R: 5′‐AACGCTTCACTAATTTGCGT‐3′;

GAPDH‐F: 5′‐GGCACAGTCAAGGCTGAGAATG‐3′;

GAPDH‐R: 5′‐ATGGTGGTGAAGACGCCAGTAC‐3′.

2.13. Western blot analysis

HUVECs were treated with the same amounts of OC‐derived exosomes (100 μg/ml) for 48 h. Then, the protein expression in HUVECs was determined. Ice‐cold RIPA lysis buffer (Beyotime Biotech, Shanghai, China) was conducted to isolate the proteins of the cells. And the protein concentration was detected by the Bicinchoninic Acid Protein Assay Kit (Beyotime Biotech). The protein samples (20 μg) were separated with 12% sodium dodecyl sulphate–polyacrylamide gel electrophoresis gels and transferred to polyvinylidene difluoride (PVDF) membranes (Millipore) for 12 h at 4°C. The PVDF membranes were blocked in 5% bovine serum albumin for 1 h at room temperature. And membranes were incubated with the above‐mentioned primary antibodies at 4°C overnight. Then, membranes were incubated with horseradish peroxidase‐conjugated secondary antibodies for 1 h at room temperature. The chemiluminescent signals were imagined using the BeyoECL Moon Enhanced Chemiluminescence Kit (ECL, Beyotime Biotech) on the ImageQuant LAS system (GE Healthcare, Sunnyvale, CA). TRIM44 (1:1000, ab236422), VEGFA (1:1000, ab214424), p‐JAK1 (1:1000, ab138005), p‐STAT1 (1:1000, ab109461), and GAPDH (1:1000, ab8245) antibodies were purchased from Abcam. GAPDH was chosen as the loading control.

2.14. Statistical analysis

Data were statistically analyzed by SPSS 20.0 software (Chicago, IL). Two groups were compared by Student's t‐test, and multiple groups were compared by One‐way analysis of variance analysis following the Tukey's test. The relationship among circNFIX, miR‐518a‐3p, and TRIM44 expression in the OC tissues was analyzed by Pearson's correlation. Data were expressed as means ± standard deviation. Experiments were independently repeated at least three times. Moreover, data in Table 1 are analyzed using Fisher's exact test. The results were regarded as statistically significant when p < 0.05.

3. RESULTS

3.1. The expressions of circNFIX , TRIM44, and miR‐518a‐3p in OC tissues

To investigate the role of circNFIX in OC, we collected OC tissues and associated adjacent tissues (n = 25) to determine their expression. As shown in Figure 1A, the expressions of circNFIX and TRIM44 were significantly increased, and miR‐518a‐3p expression was decreased in the OC tissues. Moreover, circNFIX was negatively correlated with miR‐518a‐3p expression. Similarly, miR‐518a‐3p was also negatively correlated with TRIM44 expression. Whereas there was a positive correlation between circNFIX and TRIM44 expression in the OC tissues (Figure 1B). In addition, we analyzed the correlation between the above gene expression and clinical characteristics of the enroled patients in Table 1. The results indicated that the upregulated circNFIX and TRIM44 were significantly correlated with the tumor size and FIGO stage of OC patients. And no statistical differences were observed between these genes' expression and other clinical characteristics. The miR‐518a‐3p expression also showed no significant correlation with the clinical characteristics of OC patients. These data suggested that circNFIX, TRIM44, and miR‐518a‐3p might participate in the OC development.

FIGURE 1.

The expressions of circNFIX, TRIM44, and miR‐518a‐3p in ovarian cancer (OC) tissues. (A) Quantitative real‐time polymerase chain reaction determined the relative circNFIX, miR‐518a‐3p, and TRIM44 levels in OC tissues. (B) Pearson's correlation analysis was conducted to reveal the relationship among the circNFIX, miR‐518a‐3p, and TRIM44 levels in OC tissues. n = 25. *p < 0.05, **p < 0.01.

3.2. Isolation and identification of exosomes in OC cells

To further explore the mechanism of circNFIX in the OC, we isolated the exosomes from the ovarian surface epithelial cell (HOSEpiC) and OC cell lines (SKOV3 and OVCAR3 cells). TEM observed that the exosomes isolated from the HOSEpiC, SKOV3, and OVCAR3 cell lines posed a spheroids‐shaped appearance (Figure 2A). The flow nano analysis further verified that the size of the exosomes' particles was approximately 100 nm (Figure 2B). Moreover, the exosomal marker proteins CD63 and CD81 were positively expressed and the GM130 was negatively expressed in the exosomes than the ovarian cells (Figure 2C). The above results demonstrated that we successfully isolated the exosomes from the OC cells.

FIGURE 2.

Isolation and identification of exosomes in ovarian cancer cells. Exosomes were isolated from SKOV3 or OVCAR3 cells. (A) Transmission electron microscopy was used to observe the shape appearance and particle sizes of exosomes. (B) Flow nano analysis detected the particle sizes of exosomes. (C) Western blot assays determined the expressions of CD63, CD81, and GM130 in exosomes and HOSEpiC, SKOV3, and OVCAR3 cells. Each experiment was performed in triplicates.

3.3. OC‐derived exosomes promoted the angiogenesis in HUVECs

To clarify the effects of OC‐derived exosomes on angiogenesis, HUVECs were co‐cultured with exosomes. Results revealed that the exosomes isolated from the HOSEpiC, SKOV3, and OVCAR3 cells were absorbed and internalized by the cotreated HUVECs (Figure 3A). After the SKOV3‐exo and OVCAR3‐exo treatment, the cell viability of HUVECs was markedly increased, compared with the control or HOSEpiC‐exo treated HUVECs (Figure 3B). SKOV3‐exo and OVCAR3‐exo treatment also enhanced the migration capacity of HUVECs, compared with the control or HOSEpiC‐exo (Figure 3C). Moreover, SKOV3‐exo and OVCAR3‐exo enhanced the angiogenesis ability and increased VEGFA expression (Figure 3D,E). Furthermore, the circNFIX level was significantly increased in the exosome of SKOV3 and OVCAR3 and their treated HUVECs (Figure 3F). Additionally, the miR‐518a‐3p gene expression was almost undetected and TRIM44 protein expression in the exosomes of SKOV3 and OVCAR3 was very low (Figure S1A). While SKOV3‐exo and OVCAR3‐exo treatments markedly promoted the TRIM44 expression and inhibited the miR‐518a‐3p expression in HUVECs, compared with the control (Figure S1B,C). Our results identified that OC‐derived exosomes could promote angiogenesis in HUVECs.

FIGURE 3.

OC‐derived exosomes promoted the angiogenesis in human umbilical vein endothelial cells (HUVECs). HUVECs were treated with exosomes. (A) Laser confocal microscope observed the internalization of exosomes by HUVECs. (B) 3‐(4,5‐dimethyl‐2‐thiazolyl)‐2,5‐diphenyl‐2‐H‐tetrazolium bromide assays determined the cell viability of HUVECs. (C) Transwell assays explored the migration capacities of HUVECs. (D) Tube formation assays detected the angiogenesis capacity of HUVECs. (E) Western blot assays determined VEGFA expression. (F) Quantitative real‐time polymerase chain reaction determined the relative circNFIX levels in exosomes and HUVECs. Each experiment was performed in triplicates. *p < 0.05, **p < 0.01, ***p < 0.001

3.4. OC‐derived exosomal circNFIX regulated the angiogenesis via miR‐518a‐3p in HUVECs

To further explore the mechanism of exosomal‐circNFIX in the angiogenesis of HUVECs, Starbase's prediction revealed the existence of potential binding sites of circNFIX and miR‐518a‐3p (Figure 4A). The luciferase activity of the circNFIX‐WT group was markedly downregulated in miR‐518a‐3p mimics group, whereas there was not significantly changed in the circNFIX‐MUT group (Figure 4B). And RIP analysis results further validated the existence of interaction of circNFIX and miR‐518a‐3p (Figure 4C). Additionally, the circNFIX expression was significantly enriched in the Bio‐miR‐518a‐3p group, which further validated the direct interaction between circNFIX and miR‐518a‐3p (Figure S2A). Furthermore, after SKOV3/si‐NC‐exo and OVCAR3/si‐NC‐exo treatment, the circNFIX expression was significantly upregulated, but the miR‐518a‐3p expression was decreased in HUVECs, compared with the control group (Figure 4D). Meanwhile, compared with the SKOV3/si‐NC‐exo and OVCAR3/si‐NC‐exo treatment group, the circNFIX expression was decreased, but miR‐518a‐3p was increased in HUVECs after the si‐circNFIX‐exo treatment (Figure 4D). After miR‐518a‐3p silencing and si‐circNFIX‐exo treatment in HUVECs, miR‐518a‐3p expression was downregulated, but the circNFIX expression was not obviously changed in HUVECs, compared with the si‐circNFIX‐exo and inhibitor NC treatment (Figure 4D). Compared with the SKOV3‐exo or OVCAR3‐exo treatment, the HUVECs' cell viability was decreased and their migration capacity was repressed by SKOV3/si‐circNFIX‐exo or OVCAR3/si‐circNFIX‐exo treatment, while miR‐518a‐3p knockdown reversed the above‐mentioned change (Figure 4E,F). Similarly, compared with the SKOV3‐exo or OVCAR3‐exo treatment, the si‐circNFIX‐exo treatment suppressed the angiogenesis capacity and down‐regulated VEGFA expression in HUVECs, which were reversed by miR‐518a‐3p knockdown (Figure 5A,B). The above experiments indicated that OC‐derived exosomal circNFIX promoted the anagenesis via regulating miR‐518a‐3p expression in HUVECs.

FIGURE 4.

Ovarian cancer (OC)‐derived exosomal circNFIX regulated the angiogenesis via miR‐518a‐3p in human umbilical vein endothelial cells (HUVECs). OC cells were transfected with si‐circNFIX or si‐NC and isolated their exosomes. Then, these isolated exosomes were conducted to treat with HUVECs which were transfected with miR‐518a‐3p inhibitor or NC. (A) Bioinformatics predicted the possible binding sites between circNFIX and miR‐518a‐3p. (B) Dual‐luciferase assays determined the interaction between circNFIX and miR‐518a‐3p. (C) RNA immunoprecipitation assays validated the interaction relation between circNFIX and miR‐518a‐3p. (D) Quantitative real‐time polymerase chain reaction determined the relative circNFIX and miR‐518a‐3p levels. (E) 3‐(4,5‐dimethyl‐2‐thiazolyl)‐2,5‐diphenyl‐2‐H‐tetrazolium bromide assays determined the cell viability of HUVECs. (F) Transwell assays explored the migration capacities of HUVECs. **p < 0.01, ***p < 0.001

FIGURE 5.

Exosomal circNFIX regulated the VEGFA expression via miR‐518a‐3p in human umbilical vein endothelial cells (HUVECs). (A) Tube formation assays detected the angiogenesis capacity of HUVECs. (B) Western blot assays determined VEGFA expression in HUVECs. Each experiment was performed in triplicates. Each experiment was performed in triplicates. **p < 0.01, ***p < 0.001

3.5. MiR‐518a‐3p modulated the JAK/STAT1 signaling pathway via repressing TRIM44 expression

To ulteriorly identify the regulatory mechanism of the above‐mentioned miR‐518a‐3p in the OC, we carried out the relevant mechanism analysis. Starbase results indicated the existence of putative binding sites between miR‐518a‐3p and TRIM44 (Figure 6A). And the TRIM44‐WT group's luciferase activity was markedly down‐regulated in the miR‐518a‐3p mimics group, whereas TRIM44‐MUT group was not significantly changed (Figure 6B). Moreover, RNA pull‐down analysis further confirmed the above interaction (Figure 6C). Whereas the Starbase bioinformatic predicted that there were no binding sites between miR‐518a‐3p and VEGFA. Moreover, VEGFA expression was not enriched in Bio‐miR‐518a‐3p group using RNA pull‐down assays (Figure S2B). Therefore, miR‐518a‐3p could not directly regulate the expression of VEGFA. In addition, overexpressed miR‐518a‐3p significantly repressed the TRIM44 expression and the activation of the JAK/STAT1 signaling pathway (Figure 6D,E). After TRIM44 overexpression, the miR‐518a‐3p expression was not significantly affected, whereas the expression levels of TRIM44, p‐JAK, and p‐STAT1 of the JAK/STAT signaling pathway were markedly up‐regulated (Figure 6D,E). Our results revealed that miR‐518a‐3p regulated the JAK/STAT1 signaling pathway by inhibiting TRIM44 expression.

FIGURE 6.

MiR‐518a‐3p modulated the Janus‐activated kinase (JAK)/signal transducer and activator of transcription 1 (STAT1) signaling pathway by repressing inhibiting TRIM44 expression. (A) Bioinformatics predicted the possible binding sites between miR‐518a‐3p and TRIM44. (B) Dual luciferase assays determined the interaction between miR‐518a‐3p and TRIM44. (C) RNA pull‐down assays validated the interaction relation between miR‐518a‐3p and TRIM44 in HEK‐293 T cells. (D) Quantitative real‐time polymerase chain reaction determined the relative miR‐518a‐3p and TRIM44 levels in HUVECs. (E) Western blot assays determined the expressions of TRIM44, p‐JAK, JAK, p‐STAT1, and STAT1 in HUVECs. Each experiment was performed in triplicates.  ** p < 0.01, *** p < 0.001

3.6. TRIM44 overexpression alleviated the inhibition of angiogenesis by miR‐518a‐3p overexpression in HUVECs

Subsequently, we studied the effects of miR‐518a‐3p in angiogenesis of HUVECs through regulating TRIM44. MiR‐518a‐3p overexpression dramatically weakened the cell proliferation and migration capacity, while TRIM44 overexpression reversed the above results (Figure 7A,B). The tube formation ability was also impaired by miR‐518a‐3p overexpression, which was abolished by TRIM44 overexpression (Figure 7C). Moreover, miR‐518a‐3p overexpression reduced VEGFA expression, which was reversed by TRIM44 overexpression (Figure 7D). And TRIM44 overexpression alone could upregulate the expressions of TRIM44, the JAK/STAT signaling pathway, and VEGFA, compared with the Oe‐NC group (Figure S1D). While after the JAK/STAT pathway inhibition, the expression of the JAK/STAT signaling pathway and VEGFA was decreased and the expression of TRIM44 remained unchanged, compared with the Control group (Figure S1D). Moreover, the JAK/STAT pathway, and VEGFA expression in the TRIM44 overexpression and JAK/STAT pathway inhibitor cotreatment group were lower than in the TRIM44 overexpression alone treatment group, while the TRIM44 expression remained unchanged (Figure S1D). Moreover, compared with the sh‐NC group, knockdown of VEGFA partially affected the expression of circNFIX, TRIM44, and miR‐518a‐3p expression, but these level changes were not statistically different (Figure S1E). Results identified that miR‐518a‐3p regulated angiogenesis by modulating TRIM44 expression in HUVECs.

FIGURE 7.

TRIM44 overexpression alleviated the inhibition of angiogenesis by miR‐518a‐3p overexpression in human umbilical vein endothelial cells (HUVECs). HUVECs were transfected with mimics NC, miR‐518a‐3p mimics, miR‐518a‐3p mimics+ pcDNA3.1‐NC, and miR‐518a‐3p mimics + pcDNA3.1‐TRIM44. (A) 3‐(4,5‐dimethyl‐2‐thiazolyl)‐2,5‐diphenyl‐2‐H‐tetrazolium bromide assays determined the cell viability of HUVECs. (B) Transwell assays explored the migration capacities of HUVECs. (C) Tube formation assays detected the angiogenesis capacity of HUVECs. (D) Western blot assays determined VEGFA expression. Each experiment was performed in triplicates. *p < 0.05, **p < 0.01, ***p < 0.001

4. DISCUSSION

OC is one of the major gynecological cancers in the world, with a high mortality rate, especially the 5‐year survival rate of severe OC patients is about 30%. ⁵ , ²² The process of angiogenesis has an important impact on the occurrence of OC, and measures targeting angiogenesis can inhibit the occurrence of OC. ²³ And the interaction between tumor and host cells is one of the main angiogenesis and pathogenesis mechanisms of OC, such as exosomes and their circRNAs. ²⁴ In our study, we first found that OC‐derived exosomes circNFIX promoted the cell viability, migration capacity, and angiogenesis of HUVECs by miR‐518a‐3p/TRIM44/ JAK/STAT1 axis.

Tumor cell‐derived exosomes play a crucial role in communication between the tumor and host cells. Exosomes can be generated frequently by tumor cells and released into the microenvironment to facilitate tumor growth and metastasis. For example, exosomes can boost the pre‐metastatic niche by becoming OC. ²⁵ Moreover, OC‐derived exosomes can facilitate the angiogenesis process which supplies the nutrients and oxygen to OC tissues. ²⁶ In this study, experimental evidence also proved that OC‐derived exosomes boosted the proliferation, migration, and angiogenesis capacities of HUVECs. Recent studies have indicated that OC‐derived exosomal circRNAs can promote OC progression by affecting the drug sensitivity of OC patients, such as OC‐derived exosomal circ_0007841. ²⁷ circRNA is a highly conserved and enriched noncoding RNA that has been reported to be an important regulator of tumor cell proliferation, migration, and invasiveness. ²⁸ Based on proximate studies, circNFIX can promote the malignant progression of glioma by enhancing the proliferation and invasion abilities of cells. ²⁹ In our study, we first found that circNFIX was dramatically upregulated in the OC tissues, which was similar to the above‐mentioned cancer types, suggesting its promotive role in cancer development. Moreover, the circNFIX expression was significantly correlated with tumor sizes and FIGO stage of OC patients, similar to the glioma patients. ²⁹ Tumor patients with the abnormally high expression of circRNAs could possess a poor prognosis. ³⁰ Due to the lack of prognostic analysis data for some patients during our study, we were unable to analyze the relationship between circNFIX expression and the prognosis characteristics of OC patients currently. If conditions permit, we will subsequently analyze the correlation between circNFIX and the prognosis characteristics of OC patients in the future. Furthermore, we found that OC‐derived exosomal circNFIX promoted angiogenesis in HUVECs, which may be the medicinal target for OC. Moreover, there is still no commonly recognized internal reference for the detection of exosomal RNAs of all sample types. In our study, we directly set the circNFIX expression in the HOSEpiC‐exo groups as 1, and then compared the relative fold changes of circNFIX expression in treatment groups (OC‐derived exosome) with the HOSEpiC‐exo groups. Furthermore, the GAPDH was selected as the internal reference for detecting TRIM44 mRNA expression by qRT‐PCR in this study, not for circNFIX in exosomes. Therefore, using the qRT‐PCR method to detect gene expression in exosomes is feasible. In addition, we found that the expression of circNFIX was upregulated in OC tissues and OC‐derived exosomal circNFIX could promote the angiogenesis of HUVECs. Based on that, the HUVECs' endogenous circNFIX could also promote the angiogenesis of HUVECs. Overall, in the present study, both endogenous and exosomal circNFIX (exogenous) could affect the angiogenesis of HUVECs.

It has been reported that circRNAs can be involved in tumor progression by regulating miRNA expression. CircRNAs can modulate the progression of OC via regulating miRNAs. For example, hsa_circ_0004712 can promote the malignant progression of OC by inhibiting the miR‐331‐3p. ³¹ In this study, we validated that circNFIX negatively regulated miR‐518a‐3p expression. And the promotive effects of OC‐derived exosomal circNFIX on cell proliferation, migration, and angiogenesis capacities were abolished by miR‐518a‐3p overexpression. Moreover, previous data proved that miR‐518a‐3p expression was decreased in colorectal cancer. ³² Our experimental evidence revealed that miR‐518a‐3p was also down‐regulated in the OC tissues and OC‐derived exosomes treated HUVECs. From our clinical characteristics analysis of OC patients, the miR‐518a‐3p expression showed no significant correlation with clinical characteristics, which may need a bigger scale for OC patients to further analyze. And the miR‐518a‐3p was almost not expressed in OC‐derived exosomes. The possible reason for this was that the miR‐518a‐3p expression was significantly down‐regulated in OC, so its expression in the OC‐derived exosomes was almost undetectable. Furthermore, overexpression of miR‐518a‐3p inhibited HUVECs' cell proliferation, migration, and angiogenesis capacities, further demonstrating that miR‐518a‐3p could regulate angiogenesis in the tumor microenvironment. ¹² In summary, we first found that OC‐derived exosomal circNFIX could promote the angiogenesis by inhibiting miR‐518a‐3p expression in HUVECs.

MiRNAs can tightly regulate the downstream gene expression by targeting the 3′‐UTR, such as promoting the degradation of target genes or inhibiting their translation. For example, miR‐149‐3p can promote the epithelial‐mesenchymal transition by downregulating the cyclin‐dependent kinase inhibitor 1A (CDKN1A) in OC. ³³ Based on the bioinformation prediction, our study was the first to demonstrate the negative modulation of miR‐518a‐3p and TRIM44. Moreover, VEGFA is a crucial regulator of the angiogenesis process in the tumor microenvironment. ²³ However, our data demonstrated that there were no binding sites between miR‐518a‐3p and VEGFA, based on the bioinformatics prediction analysis. Moreover, VEGFA expression was not enriched in the Bio‐miR‐518a‐3p group using RNA pull‐down assays, indicating that miR‐518a‐3p could not directly regulate the VEGFA expression. Moreover, the suppressive effect of miR‐518a‐3p overexpression on viability, migration, and angiogenesis capacities of HUVECs was abolished by TRIM44 overexpression. TRIM44 has been reported to tightly participate in the regulation of the microenvironment and angiogenesis of OC. We found that TRIM44 expression was also enhanced in the OC tissues, which was similar to the previous research. ¹³ , ¹⁴ We also found TRIM44 gene expression was markedly correlated with the tumor size and FIGO stage of OC patients, showing a potential indictive role of OC. Furthermore, our findings also showed that overexpression of TRIM44 could promote angiogenesis in HUVECs, further confirming previous findings that silencing TRIM44 could inhibit the angiogenesis process in OC. ¹⁵ And the TRIM44 protein expression in the OC‐derived exosomes was extremely low. While the protein expression of TRIM44 was markedly increased in HUVECs after the OC‐derived exosomes treatment. The finding indicated the indirect regulatory effects of OC‐derived exosomes on the TRIM44 expression of HUVECs.

In addition, the TRIM family proteins have been described as the JAK/STAT signaling pathway regulator, such that TRIM66 can positively regulate the signaling pathway in cancers. ¹⁸ , ¹⁹ We found for the first time that TRIM44 could also activate the JAK/STAT pathway, complementing the related regulatory mechanisms of TRIM family proteins and the JAK/STAT pathway. STAT1 has dual tumor‐suppressing or tumor‐promoting functions in the angiogenesis processes. ³⁴ From some previous studies, JAK/STAT signaling could negatively affect the angiogenesis process. ³⁵ , ³⁶ The JAK/STAT pathway activation could also enhance the tumor angiogenesis processes. ³⁷ It may be that STAT1 could interfere with other signaling pathways, thereby exerting the opposite effect. In addition, the angiogenesis process is regulated by pro‐angiogenic and anti‐angiogenic factors, and the balance of them can control blood vessel formation. Thus, STAT1 could play a dual role in the tumor angiogenesis process with a complex mechanism that needs deep exploration. Moreover, TRIM44 can regulate the AKT/mTOR and HIF‐1α signal pathways, that pathways engaged with the modulation of VEGFA expression. ³⁸ , ³⁹ Our results showed that the overexpressed TRIM44 upregulated TRIM44, the JAK/STAT pathway protein, and VEGFA expression. While the JAK/STAT pathway inhibition decreased the expression of JAK/STAT and VEGFA expression, the expression of TRIM44 remained unchanged, compared with the TRIM44 overexpression treatment alone. In addition, the expression levels of p‐JAK, p‐STAT1, and VEGFA in the TRIM44 overexpression and JAK/STAT pathway inhibitor cotreatment group were lower than in the TRIM44 overexpression alone treatment group, while the expression of TRIM44 remained unchanged. Therefore, TRIM44 may affect the expression of VEGFA via the JAK/STAT pathway in the angiogenesis of OC.

In conclusion, this study found that OC‐derived exosomal circNFIX could promote angiogenesis in HUVECs by regulating the miR‐518a‐3p/TRIM44 axis. Whereas in the future, relevant results still needed to be further explored and verified in vivo model. Our study provided experimental evidence that OC‐derived exosomal circNFIX can serve as a potential therapeutic target for clinical inhibition of angiogenesis and OC progression.

CONFLICT OF INTEREST

All authors declare no conflict of interest.

Supporting information

Supplementary Figure 1. (A) Western blot assays determined TRIM44 expression in the ovarian cancer (OC)‐derived exosomes. (B) Quantitative real‐time polymerase chain reaction (qRT‐PCR) determined the relative miR‐518a‐3p and TRIM44 levels in the OC‐derived exosomes treated human umbilical vein endothelial cells (HUVECs). (C) Western blot assays determined TRIM44 expression in the OC‐derived exosomes treated HUVECs. (D) Western blot assays determined TRIM44, Janus‐activated kinase (JAK)/signal transducer and activator of transcription 1 (STAT1) pathway proteins, and VEGFA expression in the HUVECs treated with TRIM44 overexpression, JAK/STAT pathway inhibitor, and NC. (E) qRT‐PCR determined the relative circNFIX, miR‐518a‐3p, and TRIM44 levels in the HUVECs treated with VEGFA knockdown treatment. Each experiment was performed in triplicates. *p < 0.05, **p < 0.01, ***p < 0.001.

Supplementary Figure 2. (A) RNA pull‐down assays validated the interaction relation between circNFIX and miR‐518a‐3p. (B) RNA pull‐down assays validated the interaction relation between VEGFA and miR‐518a‐3p. Each experiment was performed in triplicates. *p < 0.05, **p < 0.01, ***p < 0.001.

Ye H, Wang R‐Y, Yu X‐Z, Wu Y‐K, Yang B‐W, Ao M‐Y, et al. Exosomal circNFIX promotes angiogenesis in ovarian cancer via miR‐518a‐3p/TRIM44 axis. Kaohsiung J Med Sci. 2023;39(1):26–39. 10.1002/kjm2.12615