Engineered exosome-mediated delivery of circDIDO1 inhibits gastric cancer progression via regulation of MiR-1307-3p/SOCS2 Axis

A total of 17 paired tumor tissues and adjacent nontumor tissues (5 cm away from the tumor edge) were collected from the Department of General Surgery, the Affiliated People’s Hospital of Jiangsu University. All the tumor tissues were frozen in liquid nitrogen and stored at −80 °C for further use. Patients in this study had no previous radiotherapy or chemotherapy experience.

Human GC cell lines (MGC-803 and HGC-27) were purchased from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China) and Saiku Biological Technology (Guangzhou, China).

All the cells were cultured in RPMI 1640 medium with 10% fetal bovine serum (Bioind, Israel). The HEK-293 T cell line was purchased from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China) and was cultured in high glucose-DMEM medium containing 10% FBS. Cells were incubated at 37 °C with 5% CO₂ atmosphere. CircDIDO1 and SOCS2 plasmids, NC and miR-1307-3p mimics were transfected into GC cells through Lipofectamine 2000 (Invitrogen, USA). The experiment was conducted according to the manufacturer’s instructions.

Total RNA was extracted from tissues or cultured cells using Trizol reagent (Invitrogen, USA) according to the manufacturer’s protocol. The purity and concentration of total RNA were evaluated using NanoDrop One spectrophotometer (Thermo, USA). RNA was reversely transcribed to cDNA by using the HiScript 1st Strand cDNA Synthesis Kit (Thermo, USA). Subsequent detection of RNA levels was performed with UltraSYBR Mixture kit (Vazyme, China) through Real-time PCR Detection System (CFX96, Bio-Rad, USA). For miRNA expression assay, miRNA primers were synthesized by QIAGEN and total RNA was reversely transcribed using the miScript II Reverse Transcription Kit (QIAGEN, Germany). qRT-PCR amplification for miRNA was performed by using miScript SYBR Green PCR Kit (QIAGEN, Germany). U6 were used as the internal control genes. The relative RNA expression level was calculated using the 2^−△△Ct method.

Exosomes were isolated from cell supernatants following our previous protocol [29]. The protein concentration of the exosomes was determined by a BCA protein assay kit (Gibco, China). Size distribution of exosomes was identified through Nanoparticle tracking analysis (NTA) (Nanosight LM10, Particle Metrix, Germany). The morphology of isolated exosomes was observed by using transmission electron microscopy (Philips, Netherlands). The exosomal markers CD9, CD63, TSG101 and the negative control Calnexin were determined by western blot.

293 T cells were transfected with vector and circDIDO1 for 24 h and then cultured in DMEM medium containing 10% exosome-free FBS. After 48 h, cell culture supernatant was collected and exosomes were purified as previously described [29]. Exosomes from the vector (Exo-vector) and DIDO1 overexpression (Exo-circDIDO1) groups were incubated with 50 µg of DSPE-PEG-RGD (Ruixi Biotechnology, China) at 37 ℃ for 30 min to form RGD-Exo-vector and RGD-Exo-circDIDO1, respectively [30].

Diluted exosomes (1 mL) were fluorescently labeled using 5 μL of the membrane dye DiL (red, Invitrogen, USA) for 30 min at 37 °C. A total of 1 × 10⁴ cells was cultured on cover glass in 12-well plates. When 60–70% confluence was reached, DiL-labeled exosomes were added and cells were incubated for 12 h. Then cells were washed with PBS and fixed in 4% paraformaldehyde for 20 min at room temperature. Finally, Hoechst 33342 (Sigma, USA) was used for nuclear staining and then intracellular uptake of DiL-labeled exosomes was obtained using a confocal microscope (Beckman Coulter, USA).

The transfected cells were seeded in 24-well plates (1 × 10⁴/well) and counted every 24 h for 5 days. For colony formation assay, a total of 1 × 10³ cells were seeded into 6-well plates and the medium was replaced every 2 days. After 8 days of incubation, the colonies were fixed with 4% paraformaldehyde and stained with crystal violet. The number of colonies was observed and counted with a microscope.

The cells were harvested and resuspended in serum-free RPMI-1640 medium. Cells (2 × 10⁴ for migration assay and 1 × 10⁴ for invasion assay) in 200 μL of serum-free medium were plated into the upper chambers of a 24-well chamber with an 8.0 μm pore (Corning, USA). For invasion assay, the transwell upper chambers were pre-coated with diluted cold Matrigel (BD Biosciences, USA). The cells were incubated with 600 μL of medium containing 10% FBS in the lower chambers for 24 h (for the migration assay) or 48 h (for the invasion assay). After incubation, cells on the upper surface of the polycarbonate membrane were fixed with 4% paraformaldehyde for 20 min and stained with 0.1% crystal violet for 20 min. The cells were imaged and counted under a microscope.

RNA immunoprecipitation experiments were performed by the Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore, USA), following the manufacturer’s instructions. Briefly, the transfected cells were pelleted and resuspended in lysis buffer containing a protease inhibitor cocktail and RNase inhibitor. The lysates were immunoprecipitated with magnetic beads conjugated with anti-Ago2 antibody or anti-mouse IgG at 4 °C overnight. Then, the beads were washed by RIP wash buffer and incubated with proteinase K to remove proteins. Finally, the immunoprecipitated RNA was purified and used for qRT-PCR analysis.

The wild-type (WT) or mutant (MUT) luciferase reporter vectors were constructed by Hanheng Biological Technology (Shanghai, China). GC cells were co-transfected with WT or MUT reporter vectors and miRNA mimics by using Lipofectamine 2000 (Invitrogen, USA). After 48 h incubation, cells were harvested, and the luciferase activity was assessed using the Dual-Luciferase Reporter Assay System (Promega, USA).

Cell apoptosis was analyzed by using the Annexin V-Alexa Fluor 647/propidium iodide (PI) apoptosis detection kit (Fcmacs, China). After treatment with RGD-engineered exosomes for 24 h, the cells were detached and dissociated with collagenase. Cell pellets were resuspended in binding buffer and stained with Annexin V-Alexa Fluor 647 and PI for 15 min at room temperature. Cell apoptosis ratio was analyzed by CytoFLEX flow-cytometry (Beckman, USA).

For the therapeutic study, 15 nude mice were inoculated subcutaneously with MGC-803 cells. After 2 weeks, the mice were randomly divided into 3 groups. PBS, RGD-Exo-vector and RGD-Exo-circDIDO1 were intravenously injected to mice via the tail vein every 4 days for 28 days (30 μg/μL in protein concentration; 200 μL per mouse). The protocol was approved by the Animal Use and Care Committee of Jiangsu University.

GraphPad Prism 7.0 (GraphPad Software, USA) were used for statistical analysis. All data were presented as mean ± standard deviation (SD). The significant difference between different groups was analyzed by Student’s t-test and one-way ANOVA test according to actual conditions. P < 0.05 was considered as statistically significant.

To understand the mechanism for circDIDO1 in GC, we used CircInteractome software (https://​circinteractome.​nia.​nih.​gov/​) to predict the interaction of circDIDO1 with different miRNAs. We found that multiple miRNAs could be potentially sponged by circDIDO1 (Fig. 1A). Among these miRNA candidates, we focused on miR-1307-3p as our luciferase reporter assay results showed a most significant decrease in luciferase activities after co-transfection of miR-1307-3p with the wild-type luciferase reporter vector for circDIDO1. In contrast, miR-1307-3p had no effect on the luciferase activity of the mutant circDIDO1 luciferase reporter vector (Fig. 1B, C). RNA immunoprecipitation was performed to observe whether circDIDO1 binds to miR-1307-3p. The RIP results showed that circDIDO1 could be immunoprecipitated by the antibody against Ago2, a core component of the RNA-induced silencing complex (Fig. 1D). Gene expression analysis by using TCGA data showed that miR-1307-3p was highly expressed in GC tissues compared to adjacent normal tissues (Fig. 1E). MiR-1307-3p has been confirmed to play a critical role in a variety of cancers. To determine whether circDIDO1 regulates GC cells through miR-1307-3p, we conducted a gain-of-function study and found that overexpression of miR-1307-3p significantly inhibited circDIDO1 expression in GC cells, suggesting that circDIDO1 is targeted by miR-1307-3p (Fig. 1F). In addition, miR-1307-3p expression significantly promoted GC cell proliferation, migration, and invasion (Fig. 1G–J). Rescue experiments were performed to confirm the biological effects of circDIDO1/miR-1307-3p axis. The co-transfection of circDIDO1 repressed the promoting role of miR-1307-3p in GC cell proliferation, migration, and invasion (Fig. 2A–D). In summary, these data suggests that circDIDO1 directly binds to miR-1307-3p and serves as a negative regulator of miR-1307-3p.

As previously described, the RNA sequencing for control and circDIDO1 overexpressing GC cells and the cluster and pathway analyses for gene expression profiles were performed [28]. We found that 134 genes were up-regulated and 52 genes were down-regulated in circDIDO1 overexpressing group compared to the control group. Suppressor of cytokine signaling 2 (SOCS2) was one of the significantly up-regulated genes. The altered expression of SOCS2 in the circDIDO1 overexpressing group was verified by qRT-PCR [28]. To further seek for the downstream regulatory axis of circDIDO1/miR-1307-3p in GC cells, we also used TargetScan to predict the potential target genes of miR-1307-3p. Combined with RNA sequencing analysis, SOCS2 was confirmed to be a direct target of miR-1307-3p in GC cells. StarBase assay revealed that the expression levels of miR-1307-3p and SOCS2 displayed a significant negative correlation (Fig. 3A). In addition, we performed a correlation analysis between circDIDO1 and SOCS2 in 17 paired GC tissues. The expression of SOCS2 was positively correlated with that of circDIDO1 in GC tissues (Fig. 3B). The co-transfection of miR-1307-3p with the wild-type luciferase reporter vector for SOCS2 significantly decreased the luciferase reporter activity in GC cells. However, these effects were abolished by a mutant SOCS2 luciferase reporter vector (Fig. 3C). Subsequently, qRT-PCR results indicated that circDIDO1 overexpression increased the expression level of SOCS2 (Fig. 3D). In contrast, overexpression of miR-1307-3p significantly inhibited SOCS2 expression in GC cells (Fig. 3E). These findings suggest that circDIDO1 sponges miR-1307-3p to upregulate SOCS2 expression. Moreover, the results of gain-of-function studies revealed that SOCS2 overexpression inhibited GC cell proliferation, migration, and invasion (Fig. 3F–I).

We further performed rescue studies to explore whether circDIDO1 regulates gastric cancer progression via the miR-1307-3p/SOCS2 axis and the results of rescue experiments showed that SOCS2 overexpression partially reversed the promoting roles of miR-1307-3p in GC cell proliferation, migration, and invasion (Fig. 4A–D). Collectively, our findings provided the first evidence that circDIDO1 regulates SOCS2 expression through miR-1307-3p, thereby inhibiting GC progression.

The potent suppressive role of circDIDO1 in GC progression led us to hypothesize that it may be a potential therapeutic target for GC. To this end, we constructed circDIDO1-loaded, RGD-modified exosomes (RGD-Exo-circDIDO1) and investigated its effect on GC therapy. RGD-Exo-circDIDO1 were extracted and purified through differential centrifugation from the cell culture supernatant, which were subsequently characterized by nanosight tracking analysis, transmission electron microscopy and western blot. The diameter of most RGD-Exo-circDIDO1 ranged from 100 to 130 nm, and the average size was found to be approximately 120 nm (Fig. 5A). We found that the exosome exhibited a sphere-shaped morphology (Fig. 5B). Exosomal markers such as CD9, CD63 and TSG101 were abundant in the RGD-Exo-circDIDO1 while the negative control Calnexin was not observed in the RGD-Exo-circDIDO1 (Fig. 5C). Subsequently, we extracted the total RNA of the exosomes and demonstrated by qRT-PCR that the circDIDO1 expression level was significantly increased in Exo-circDIDO1 (Fig. 5D). To evaluate the cellular tropism of RGD-Exo-circDIDO1 in vitro, RGD-Exo-circDIDO1 was labeled with DiL and co-cultured with GC cells for 12 h. Confocal microscope results showed that the internalization of Exo-circDIDO1 modified with RGD into GC cells was more efficient compared to Exo-circDIDO1 without RGD modification (Fig. 5E). RGD-Exo-circDIDO1 had the efficient delivery ability of circDIDO1, which is indicated by the significant increase of circDIDO1 level in GC cells treated with RGD-Exo-circDIDO1 compared to the control exosome group (Fig. 5F). Overall, the RGD-modified exosomes with a high level of circDIDO1 could be successfully prepared for further study.

To further investigate the functional role of RGD-Exo-circDIDO1 in GC cells, the ability of PBS, RGD-Exo-vector and RGD-Exo-circDIDO1 in GC cell apoptosis after 12 h treatment was assessed. The annexin V/PI staining and flow cytometry results indicated a significant increase of apoptotic GC cells in the RGD-Exo-circDIDO1 treatment group compared with the other two groups (Fig. 6A). The RGD-Exo showed minimal therapy effect, which indicated that the anti-tumor effect was from circDIDO1, instead of RGD-Exo. In addition, compared with PBS and RGD-Exo-vector treatment, RGD-Exo-circDIDO1 treatment significantly inhibited the proliferation, migration and invasion of GC cells (Fig. 6B–E). Western blot results also revealed that exosome-mediated delivery of circDIDO1 resulted in increased levels of cleaved caspase 3 and cleaved PARP1 in GC cells (Fig. 7D, left panel), both of which are markers for cells undergoing apoptosis. These results suggest that compared with PBS and RGD-Exo-vector treatment, RGD-Exo-circDIDO1 can induce an efficient anti-tumor effect.

To investigate the anti-tumor effect of RGD-Exo-circDIDO1 in vivo, we established a xenograft mouse model of gastric cancer by subcutaneously inoculating MGC-803 cells into nude mice. Tumor growth analysis revealed that exosomes loaded with circDIDO1 induced a significant decrease in tumor volume and weight compared to the RGD-Exo-vector and PBS treated mice (Fig. 7A). As shown in Fig. 7B, immunohistochemical staining results revealed the decrease of Ki-67-positive proliferating cells and increase of TUNEL-positive apoptotic cells in the RGD-Exo-circDIDO1 treated group compared to the other two groups. These results showed that RGD-Exo-circDIDO1 treatment remarkably suppressed tumor growth. In summary, the therapeutic efficacy of RGD-Exo-circDIDO1 was confirmed in a mouse xenograft tumor model.

As shown in Fig. 7, exosome-mediated delivery of circDIDO1 resulted in a significant increase of SOCS2 level compared to the GC cells treated with unloaded-Exo or PBS (Fig. 7C, upper panel). The similar effect was further examined in vivo. The level of miR-1307-3p was downregulated in tumor tissues treated with RGD-Exo-circDIDO1, while SOCS2 was upregulated subsequently (Fig. 7C, lower panel). In accordance, the protein level of SOCS2 was increased by exosome-mediated delivery of circDIDO1 in treated GC cells and tissues. Besides, the expressions of apoptosis-related proteins was measured in tumor tissues by western blot. The results revealed that RGD-Exo-circDIDO1 treatment led to an obvious increase of cleaved caspase 3 and cleaved PARP1 protein levels in tumor tissues (Fig. 7D, right panel). Collectively, these results further confirm that miR-1307-3p/SOCS2 axis is essential for the circDIDO1-mediated tumor suppression effect in GC.

In addition to treatment efficacy, toxicity is another vital parameter of an excellent delivery vehicle. For safety purpose, we evaluated the systematic toxicity of engineered exosomes in nude mice after intravenous tail vein injection at a dosage of 200 μL per mouse every 4 days for 28 days (30 μg/μL in protein concentration). Compared with PBS treated group, neither death nor a significant difference in body weight was observed in the RGD-Exo-circDIDO1 treated group during the study period (data not shown). Moreover, we further detected the potential pathological lesions induced by engineered exosomes on major organs (i.e., hearts, livers, spleens, lungs, and kidneys). Blood biochemistry and hematology analysis were performed to investigate any potential toxic effect of RGD-Exo-circDIDO1 on the treated mice. On the 28th day after the first injection, all the mice were euthanized, and their major organs and blood were collected for H&E staining, histology analysis, and blood test. As shown in Fig. 7E, major organs had no obvious histopathological abnormalities or lesions among the RGD-Exo-circDIDO1 treated group, which suggested no evidence of inflammatory response caused by RGD-Exo-circDIDO1. Furthermore, no significant hepatic or renal toxicity induced by RGD-Exo-circDIDO1 was observed, as indicated by normal values of liver function markers (ALT, AST, and ALP) and kidney function markers (BUN and CRE) (Fig. 7F). CCK8 assays were performed to assess the potential toxicity of RGD-Exo-circDIDO1 to normal gastric mucosa epithelial cells (GES-1). No toxicity was observed since cell viabilities in all groups remained above 95% at 24 h, 48 h and 96 h after treatment (Additional file 1: Figure S1). All the above results suggest that RGD-Exo-circDIDO1 treatment could be used as a promising approach for GC therapy without significant side effects.