Accounting for more than 9% of deaths worldwide, colorectal cancer (CRC) is an alimentary canal tumor with a high mortality and recurrence rate.^1,2 Although the therapeutic effect has been significantly improved in recent years, survival rate for patients with distant metastasis or terminal tumors is still low, mainly because of the lack of early diagnosis targets as well as effective therapeutic targets.^3,4 Therefore, it is urgent to investigate the pathogenic mechanism of CRC.

Exosomes are nanoscale vesicles with lipid bilayer structure, with a size range of 40 to 150 nm.⁵ There are some biomarkers on the surface of exosomes, such as CD63, CD9, CD81, and so on.⁶ Much evidence showed that exosomes play important roles in intercellular communication through transferring biological molecules from one cell to another, such as proteins, lipids, messenger RNAs, and non-coding RNAs.^7,8 Recently, exosomal circular RNAs (circRNAs) have been considered to be promising biomarkers for clinical detection.^9,10 For instance, quantitative analysis of clinical CRC samples revealed that serum exosomal circ-0004771 was upregulated and could be a novel potential diagnostic biomarker.¹¹ circRNA is characterized by closed-loop structure and were aberrantly expressed in several forms of tumors,¹² attracting widespread attention in recent years. Several researches have confirmed that circRNAs exerted crucial functions in CRC.^13,14 A previous report discovered that plasma exosomal circRNA pinin (circ-PNN; hsa_circ_0101802) expression was upregulated in CRC patients.¹⁵ However, whether plasma exosomal circ-PNN could participate in the regulation of CRC progression should be further elucidated.

Accumulated evidence suggests that circRNA could regulate downstream gene expression via RNA binding proteins or competing endogenous RNAs, which were two classic pathways to modulate cells biological functions in tumor cells.^16,17 For instance, Rho-related BTB domain containing 3 regulated polypyrimidine tract-binding protein 1 expression via human antigen R to inhibit aggressiveness of CRC.¹⁸ circRNA ubiquitin-associated protein 2 segregated miR-582-5p to upregulate forkhead box protein O1, which enhanced autophagy and advanced CRC cell proliferation and motility.¹⁹ As we know, fibroblast growth factor 13 (FGF13) is one of the fibroblast growth factor family, and was highly expressed in different tumors, including CRC.^20-22 Besides, FGF13 could promote the metastasis of triple-negative breast cancer.^23,24 In addition, FGF13 was suggested to regulate the tumorigenesis and progression of CRC.^22,25 To date, the data are lacking regarding the association between circ-PNN and FGF13.

Here, circ-PNN, miR-1225-5p, and FGF13 levels in CRC patient plasma-derived exosomes were analyzed. We also investigated whether exosomal circ-PNN promoted CRC progression via miR-1225-5p/FGF13 axis.

1. Collection of clinical specimens

The plasma specimens were collected from CRC patients (n=20) and healthy control subjects (n=20), with approval from the Ethics Committee of Affiliated Hospital of Guangdong Medical University and the provided written informed consent by participants (IRB approval number: 202010156). The clinicopathological information of the subjects is shown in Supplementary Table 1.

2. Exosomes isolate

Exosomes were isolated through differential ultracentrifugation as previously described.²⁶ Briefly, 500 μL plasma samples were centrifuged at different speeds. The serum was then put into centrifuge tubes and centrifuged at 100,000 ×g to obtain exosomes. The exosomes were placed onto carbon-coated copper grid prior to fixation with 1% glutaraldehyde for 20 minutes, followed by analysis using transmission electron microscopy (JEOLLtd., Guangzhou, China).

3. Cell culture and treatment

HCT116, SW480, and 293T cells were acquired from Procell (Wuhan, China) and cultured in McCoy’s 5A medium (Procell), Leibovitz's L-15 mediums (Procell), and Dulbecco’s Modified Eagle Medium (Procell), respectively, with 5% v/v CO₂ at 37°C. All medium was supplemented with 10% fetal bovine serum (Shanghai Universal Biotech, Shanghai, China). CRC cells were incubated for 12 hours for attachment and co-incubated with the isolated exosomes or phosphate-buffered saline (PBS) for 24 hours in 6-well plates (2×10⁵ cell per well), 12-well plates (1×10⁵ cell per well) or 96-well plates (5,000 cell per well).

4. Western blot

After protein isolation was performed using RIPA lysis buffer (Keygen Biotech, Nanjing, China), SurePAGE gels (Thermo Fisher, Waltham, MA, USA) and Midi-Cell Electrophoresis System (XCell4 SureLock; Thermo Fisher) were used for the separation of target proteins. Polyvinylidene fluoride membranes (Beyotime, Shanghai, China) activated with methanol were used for protein transferring. The membranes were incubated with the indicated primary antibodies as following: anti-CD81 (ab109201, 1:1,000, Abcam, Cambridge, MA, USA), anti-CD63 (ab217345, 1:1,000, Abcam), anti-MMP9 (ab283575, 1:1,000, Abcam), anti-MMP2 (ab181286, 1:1,000, Abcam), anti-FGF13 (ab1153808, 1:2,000, Abcam), and anti-GAPDH (ab181602, 1:10,000, Abcam), followed by the 2 hours incubation of second antibody (ab97051, 1:20,000, Abcam). The blots were visualized by ECL reagents (Beyotime).

5. Quantitative real-time polymerase chain reaction

RNAs (500 ng to 1 μg) isolated based on the guidebook of TRIzol (Invitrogen, Carlsbad, CA, USA) were reverse transcribed into complementary DNA using miScript RT kit (Takara, Dalian, China) or PrimeScript RT reagent kit (Takara). After that, complementary DNA was quantified and SYBR Premix Ex Taq II reagents (Takara) were employed for quantitative real-time polymerase chain reaction (qRT-PCR) analysis. The 2^-ΔΔ^Ct method was used for expression analysis with housekeeping genes (GAPDH or U6) used as internal references. Primers are displayed in Table 1.

Table 1 . Primers Sequences Used for PCR.

Name		Primers for PCR (5’-3’)
circ_0101802	Forward	GAAGAATGTGTCCAGCTACCCA
	Reverse	CTGCTTTCTCTCTTCTTCTGCC
FGF13	Forward	ATCCGTCAGAAGAGGCAAGC
	Reverse	CGAAGAGTTTGACCCGGGAA
miR-1225-5p	Forward	GGTGGGTACGGCCCAGT
	Reverse	AGTGCAGGGTCCGAGGTATT
GAPDH	Forward	GACAGTCAGCCGCATCTTCT
	Reverse	GCGCCCAATACGACCAAATC
U6	Forward	CTCGCTTCGGCAGCACA
	Reverse	AACGCTTCACGAATTTGCGT
FNN	Forward	CTACCTCCAAAGAGCGCACA
	Reverse	GCCAAATATTCGCCGGTTCC

PCR, polymerase chain reaction..

6. Analysis of circRNA stability

circRNA stability was analyzed using RNase R and actinomycin D in line with the reported paper.²⁷

7. Cell transfection

Small interfering RNA targeting circ-PNN (si-circ-PNN), miR-1225-5p mimic and inhibitor, and their negative controls were bought from Geneseed (Guangzhou, China). FGF13 sequence was synthesized and cloned into pcDNA3.1 to obtain the FGF13 overexpression vector. CRC cells were transfected with plasmids or oligonucleotides using lipofectamine 3000 (Invitrogen).

8. Cell Counting Kit 8 assay

Briefly, transfected CRC cells were maintained in 96-well plates. Then, Cell Counting Kit 8 reagent (Beyotime) was added to the 96-well plates and incubated for 2 hours, and the absorbance was measured with a microplate reader.

9. Colony formation assay

First, CRC cells were passaged into 6-well plates prior to culture in a humid atmosphere for 14 days. Subsequently, the cells were exposed to paraformaldehyde (Beyotime). The colonies were observed under a microscope after staining using crystal violet (Beyotime).

10. EdU assay

CRC cells (1×10⁴) were cultivated for 24 hours in a humid atmosphere and incubated with EdU (RiboBio, Guangzhou, China) prior to exposure to paraformaldehyde and Apollo Dye Solution (RiboBio). At last, cells were counted after taking photos with a microscope.

11. Apoptosis assay

CRC cells were cultivated for 48 hours in a humid atmosphere and collected for apoptosis analysis following the guidebook of Annexin V-FITC/PI apoptosis detection kit (Solarbio, Beijing, China).

12. Transwell assay

CRC cells (1×10⁵) in serum-free medium were plated in the top chamber pre-coated with Matrigel and subjected to 48-hour incubation. Cells invading to the underside of the transwell chamber were fixed with methyl alcohol. Cells were counted by a microscope prior to staining with crystal violet.

13. Wound healing assay

Cells were cultured to 85%–90% aggregation and wounded using 10 μL pipette tips, followed by washing with PBS solution. After that, cells were incubated with 1% fetal bovine serum. A microscope was used to analyze cells.

14. Xenograft mouse model

A total of 20 male BALB/c nude mice (4- to 6-week-old) were obtained from Hunan Slyke Jingda Experimental Animal Co., Ltd (Changsha, China). HCT116 cells stable expressing small hairpin RNA of circ-PNN (sh-circ-PNN) or small hairpin RNA of negative control (sh-NC) (3×10⁶) were injected into the right flank of nude mice, five mice for sh-NC group and 15 mice for sh-circ-PNN group. Twelve days upon tumor inoculation (tumor volume approximately 50 mm³), the mice in sh-circ-PNN group were divided into three groups (n=5): sh-circ-PNN group, sh-circ-PNN+PBS group (injected with PBS), sh-circ-PNN+tumor-Exo group, and received tail vein injection of exosomes (10 µg exosomes in equal PBS). Thirty days later, all mice were executed for further analysis. The animal studies were approved by the Animal Management Committee of Affiliated Hospital of Guangdong Medical University (IRB approval number: 202010156).

15. Immunohistochemistry

As instructed,²⁸ the paraffin-embedded tissues were blocked with anti-Ki-67 (ab279653, 1:1,000, Abcam), anti-MMP9 (ab283575, 1:500, Abcam), and anti-MMP2 (ab97779, 1:500, Abcam), stained with Mayers hematoxylin, and analyzed under a fluorescence microscope.

16. RIP assay

Based on the standard instructions handbook of Magna RIP kit (Sigma-Aldrich, St. Louis, MO, USA), 293T cells were lysed and then incubated with anti-IgG (ab133470, Abcam) or anti-Ago2 (ab32381, Abcam). circ-PNN, miR-1225-5p and FGF13 were quantified by qRT-PCR.

17. Dual-luciferase reporter assay

The wild-type (wt) or mutant (mut) sequences of circ-PNN and FGF13 were cloned into psiCHECK2 vector (Geneseed) to generate circ-PNN wt, circ-PNN mut, FGF13 3'UTR wt or FGF13 3'UTR mut luciferase reporter vectors. Subsequently, 293T cells were co-transfected with miR-1225-5p mimic or mimic NC and indicated luciferase reporter vectors using lipofectamine 3000 (Invitrogen). Forty-eight hours upon incubation, luciferase activities were under the exploration of Dual-Luciferase Reporter Assay reagents (Dual-Luciferase Reporter Assay System; Promega, Madison, WI, USA).

18. Statistical analysis

Data were expressed as the mean±standard deviation. The difference was analyzed by the Student t-test or analysis of variance. p<0.05 was considered a statistically significant difference.

1. Characterization of exosomes derived from the plasma of CRC patients

To investigate the function and mechanism of exosomes in CRC, exosomes were isolated from the plasma of CRC patients and healthy subjects. Transmission electron microscopy results showed that exosomes derived from the plasma of CRC patients and healthy subjects exhibited a double-layer membrane structure (Fig. 1A). The existence of exosome markers, CD81 and CD63, were confirmed by Western blot analysis (Fig. 1B). In addition, the expression of circ-PNN was higher in tumor exosomes than that in normal exosomes (Fig. 1C). To confirm the circular structure of circ-PNN, RNase R and actinomycin D assays were conducted in CRC cells. The level of circ-PNN was not affected by RNase R treatment, but PNN level was significantly reduced by RNase R treatment in HCT116 and SW480 cells (Fig. 1D). For actinomycin D assay, circ-PNN was resistant to actinomycin D treatment, whereas PNN was digested by actinomycin D (Fig. 1E).

Figure 1. Characterization of exosomes (Exos) derived from the plasma of colorectal cancer patients. (A) Transmission electron microscopy image of normal-Exos and tumor-Exos. (B) The protein levels of CD81 and CD63 were measured by Western blotting. (C) The expression of circ-PNN in normal-Exos and tumor-Exos was measured by qRT-PCR. (D) After RNase R treatment, the levels of circ-PNN and its linear transcript PNN were examined by qRT-PCR. (E) For actinomycin D assay, the levels of circ-PNN and its linear transcript PNN were measured by qRT-PCR. circ-PNN, circular RNA pinin; qRT-PCR, quantitative real-time polymerase chain reaction. *p<0.05.

2. Tumor-derived exosomes expedited proliferation, invasion, and migration, but inhibited apoptosis in CRC cells

To explore the functional mechanism of exosomes in CRC cells, CRC cells were incubated with tumor-derived exosomes. The qRT-PCR results showed that circ-PNN levels were higher in CRC cells incubated with tumor-Exo group (Fig. 2A). The cell proliferation was enhanced by incubation with tumor-Exo, as confirmed by the increased optical density value, colony numbers and the number of EdU-positive cells in the tumor-Exo group (Fig. 2B-D). Flow cytometry assay indicated that cells apoptosis ratio were decreased in cells incubated with tumor-Exo (Fig. 2E). Besides, Western blot assay uncovered that the protein level of Bax (pro-apoptotic protein) was reduced, while Bcl-2 (anti-apoptotic protein) protein level was upregulated in the tumor-Exo group (Fig. 2F). Furthermore, cell invasive and migratory abilities were enhanced in CRC cells with tumor-Exo treatment (Fig. 2G and H). Additionally, the protein levels of MMP9 and MMP2 were elevated in CRC cells with tumor-Exo treatment (Fig. 2I). Our results elucidated that tumor-derived exosomes facilitated the progression of CRC.

Figure 2. Tumor-derived exosomes (Exos) enhance proliferation, invasion and migration, but promote apoptosis in colorectal cancer cells. Colorectal cancer cells were treated with tumor-Exos or PBS. (A) The level of circ-PNN was assessed by qRT-PCR. (B-D) Cell proliferation was examined by Cell Counting K-8, colony formation and EdU assays. (E) Cell apoptosis were monitored by flow cytometry. (F) The protein levels of Bax and Bcl-2 were detected by Western blotting. (G) Cell invasion was determined by transwell assay. (H) Cell migration was monitored by wound healing assay. (I) The protein levels of MMP9 and MMP2 were assessed by Western blotting. PBS, phosphate-buffered saline; circ-PNN, circular RNA pinin; qRT-PCR, quantitative real-time polymerase chain reaction; OD, optical density. *p<0.05.

3. Exosomal circ-PNN accelerated the progression of CRC in vitro

To investigate the role of exosomal circ-PNN in CRC development, we transfected si-circ-PNN and si-NC into HCT116 and SW480 cells, then incubated with tumor-Exo or PBS. As shown in Fig. 3A, circ-PNN expression was significantly downregulated by si-circ-PNN transfection, while it was obviously upregulated by the incubation of tumor-Exo, indicating that circ-PNN could be transferred by exosomes. Functionally, circ-PNN knockdown inhibited cell proliferation, whereas this effect was partly recused by tumor-Exo treatment (Fig. 3B-D). Besides, tumor-Exo treatment partly overturned the promotion effect of circ-PNN knockdown on cell apoptosis (Fig. 3E and F). Furthermore, tumor-Exo treatment reversed the inhibitory effect of circ-PNN silencing on cells invasion and migration (Fig. 3G and H). In addition, circ-PNN knockdown decreased MMP9 and MMP2 levels, whereas the inhibitory effect was abolished by tumor-Exo treatment (Fig. 3I). These findings indicated that circ-PNN could be transferred by exosomes, and promoted the progression of CRC in vitro.

Figure 3. Exosomal circ-PNN accelerates colorectal cancer progression in vitro. Colorectal cancer cells were transfected with si-NC or si-circ-PNN, and treated with tumor-Exos or PBS. (A) The level of circ-PNN was tested by qRT-PCR. (B-D) Cell proliferation was examined by Cell Counting K-8, colony formation and EdU assays. (E) Cell apoptosis was monitored by flow cytometry. (F) The protein levels of Bax and Bcl-2 were detected by Western blotting. (G) Cell invasion was determined by transwell assay. (H) Cell migration was monitored by wound healing assay. (I) The protein levels of MMP9 and MMP2 were assessed by Western blot analysis. circ-PNN, circular RNA pinin; si-NC, small hairpin RNA of negative control; si-circ-PNN, small hairpin RNA of circ-PNN; Exo, exosome; PBS, phosphate-buffered saline; qRT-PCR, quantitative real-time polymerase chain reaction. *p<0.05.

4. Exosomal circ-PNN regulated tumor growth in vivo

To further explore the effect of exosomal circ-PNN in CRC in vivo, HCT116 cells stable expressing sh-NC or sh-circ-PNN were injected into nude mice. The injection of tumor-Exo remarkably attenuated the inhibitory effect of circ-PNN silence on tumor volume and tumor weight (Fig. 4A-C). Besides, circ-PNN level was decreased in xenograft tumor tissues with circ-PNN silence, whereas it was significantly elevated by tumor-Exo treatment (Fig. 4D). Additionally, immunohistochemical assay revealed that Ki-67-, MMP9- and MMP2-positive cells were significantly reduced by circ-PNN silence, while the injection of tumor-Exo overturned this effect (Fig. 4E). These results uncovered that exosomal circ-PNN affected the progression of CRC in vivo.

Figure 4. Exosomal circ-PNN regulates tumor growth in vivo. (A, B) The average tumor volume of xenograft tumor tissues was calculated based on measurements with a Vernier caliper. (C) Tumor weight of xenograft tumor tissues was tested. (D) The expression of circ-PNN was measured by qRT-PCR. (E) The expression levels of Ki-67, MMP9 and MMP2 in xenograft tumor tissues were detected by immunohistochemical. sh-NC, small hairpin RNA of negative control; circ-PNN, circular RNA pinin; sh-circ-PNN, small hairpin RNA of circ-PNN; Exo, exosome; qRT-PCR, quantitative real-time polymerase chain reaction; PBS, phosphate-buffered saline. *p<0.05.

5. circ-PNN interacted with miR-1225-5p

To elucidate the molecular mechanism of circ-PNN, the Circinteractome database was used to predict the binding site with microRNA, which displayed that circ-PNN contained the complementary binding sites of miR-1225-5p (Fig. 5A). RIP assay demonstrated that circ-PNN was significantly pulled down by Ago2-immunoprecipitated group compared with IgG-immunoprecipitated group (Fig. 5B). Besides, miR-1225-5p mimic transfection inhibited the luciferase activity of circ-PNN wt group, but not the circ-PNN mut group (Fig. 5C). Furthermore, we found that the expression of miR-1225-5p was decreased in CRC cells with tumor-Exo treatment (Fig. 5D), and it was significantly elevated in CRC cells with circ-PNN knockdown (Fig. 5E). Therefore, circ-PNN acted as a sponge for miR-1225-5p.

Figure 5. circ-PNN interacts with miR-1225-5p. (A) The predicted binding sites between circ-PNN and miR-1225-5p are shown. (B, C) The correction between circ-PNN and miR-1225-5p was confirmed by RIP assay and Dual-Luciferase Reporter Assay. (D) The expression of miR-1225-5p was measured in colorectal cancer cells incubated with PBS or tumor exosome. (E) The expression of miR-1225-5p was measured in colorectal cancer cells with si-NC or si-circ-PNN transfection. circ-PNN, circular RNA pinin; PBS, phosphate-buffered saline; si-NC, small hairpin RNA of negative control; si-circ-PNN, small hairpin RNA of circ-PNN; NC, negative control; wt, wild type; mut, mutant. *p<0.05.

6. circ-PNN regulated FGF13 expression by sponging miR-1225-5p

Using TargetScan software for prediction, the potential binding sites of miR-1225-5p in the 3’UTR of FGF13 were uncovered (Fig. 6A). Besides, miR-1225-5p overexpression impeded the luciferase report activity of FGF13 3’UTR wt group, but not FGF13 3’UTR mut group (Fig. 6B), suggesting that miR-1225-5p could directly interact with FGF13. After being incubated with tumor-Exo, the protein levels of FGF13 were increased in CRC cells (Fig. 6C). qRT-PCR assay proved that miR-1225-5p was reduced in CRC cells with miR-1225-5p inhibitor transfection (Fig. 6D). And interfering of miR-1225-5p elevated the protein level of FGF13 (Fig. 6E). Additionally, miR-1225-5p downregulation rescued the effect of circ-PNN depletion on FGF13 expression (Fig. 6F and G). Our results manifested that circ-PNN could accelerate FGF13 level via sponging miR-1225-5p.

Figure 6. circ-PNN regulates FGF13 expression by modulating miR-1225-5p. (A) The predicted binding sites between miR-1225-5p and FGF13 were exhibited. (B) The correction between miR-1225-5p and FGF13 was confirmed by dual-luciferase report activity. (C) The level of FGF13 protein was measured in colorectal cancer (CRC) cells incubated with PBS or tumor exosome. (D) The expression of miR-1225-5p was measured by qRT-PCR in CRC cells with inhibitor NC or miR-1225-5p inhibitor transfection. (E) The expression of FGF13 in CRC cells with inhibitor NC or miR-1225-5p inhibitor transfection was measured. (F, G) The protein level of FGF13 in CRC cells with si-NC, si-circ-PNN, si-circ-PNN+inhibitor NC, or si-circ-PNN miR-1225-5p inhibitor transfection was detected by Western blot analysis. wt, wild type; mut, mutant; circ-PNN, circular RNA pinin; FGF13, fibroblast growth factor 13; PBS, phosphate-buffered saline; qRT-PCR, quantitative real-time polymerase chain reaction; Exo, exosome; NC, negative control; si-NC, small hairpin RNA of NC; si-circ-PNN, small hairpin RNA of circ-PNN. *p<0.05.

7. circ-PNN promoted CRC progression by regulating FGF13 expression

Based on the above data, we predicted that circ-PNN may regulate CRC progression via increasing FGF13 expression. The levels of FGF13 were reinforced by transfection of FGF13 overexpression vector (Fig. 7A). Suppression of circ-PNN retarded FGF13 level, and miR-1225-5p inhibitor recuperated this effect (Fig. 7B). Overexpression of FGF13 abolished the effect of circ-PNN knockdown on cells proliferation (Fig. 7C-F). circ-PNN depletion triggered Bax expression and restrained Bcl-2 expression, while overexpression of FGF13 could revert this effect (Fig. 7G). The inhibitory effects of circ-PNN knockdown on cell invasion and migration were overturned in HCT116 and SW480 cells transfected with pcDNA 3.1(+)-FGF13 (Fig. 7H and I). In addition, overexpression of FGF13 abated the inhibitory effect of circ-PNN knockdown on MMP9 and MMP2 levels (Fig. 7J). The data manifested that circ-PNN regulated cells biological functions via regulating FGF13 expression.

Figure 7. circ-PNN promotes colorectal cancer (CRC) proliferation, migration, and invasion by increasing FGF13 levels. (A) The protein level of FGF13 was tested by Western blot in CRC cells with pcDNA or pcDNA-FGF13 transfection. (B-J) CRC cells were transfected with si-NC, si-circ-PNN, si-circ-PNN+pcDNA or si-circ-PNN+pcDNA-FGF13, respectively. (B) The protein level of FGF13 in transfected cells was detected by Western blot analysis. (C-E) Cell proliferation was examined by Cell Counting K-8, colony formation and EdU assays. (F) Cell apoptosis was detected by flow cytometry. (G) The protein levels of Bax and Bcl-2 in transfected cells were detected by Western blotting. (H) Cells invasion was determined by transwell assay. (I) A wound healing assay was conducted to monitor cell migration. (J) The protein levels of MMP9 and MMP2 were assessed by Western blotting. circ-PNN, circular RNA pinin; FGF13, fibroblast growth factor 13; si-NC, small hairpin RNA of negative control; si-circ-PNN, small hairpin RNA of circ-PNN; pcDNA, pcDNA 3.1 (+) vector; OD, optical density. *p<0.05.

circRNAs have been regarded as tumor suppressor or promoter in multiple cancers.²⁹ In CRC, circRNAs, such as has_circ_0026628, cis-HOX, has_circ_0006401, have displayed powerful potential in regulating cell growth ability and tumor metastasis, and their upregulation are correlated with poor prognosis and advanced TNM stage.^30-32 circRNAs are now widely deemed as one kind of endogenous RNAs that regulate downstream genes by influencing the functions of microRNAs.³³ Currently, the regulation of competing endogenous RNA network in CRC has been widely validated. For example, circGNG4 interacted with miR-223 to promote tumor growth of prostate cancer.³⁴ Circ_001422 promoted osteosarcoma cancer progression and metastasis through facilitating the expression of FGF2 via modulating miR-195-5p.³⁵ circRNA hsa_circ_0110389 enhanced SORT1 expression through regulating miR-127-5p and miR-136-5p to accelerate gastric cancer progression.³⁶ Our results discovered that circ-PNN and FGF13 acted as tumor promoters in CRC cells. We speculated circ-PNN modulated the expression of FGF13 via competing endogenous RNA mechanism. Subsequently, we confirmed that miR-1225-5p could bind with circ-PNN and FGF13 using biological analysis and luciferase report assay. Moreover, miR-1225-5p was suggested to act as a tumor inhibitor, which is in agreement with former studies.³⁷ Besides, we found that circ-PNN acted as a miR-1225-5p sponge to promote FGF13 level. All these findings speculated that circ-PNN might function as a critical effector in CRC progression via miR-1225-5p/FGF13 axis.

Exosomes have been regarded as nanoscale carriers adjusting with tumor cells and microenvironments, which are beneficial to tumor metastases.³⁸ Besides, mounting evidence confirmed that certain non-coding RNAs are selectively gathering in cancer-derived exosomes.^39,40 Consequently, the exosomal non-coding RNAs might induce the poor development of cancer. For instance, exosomal miR-128-3p promoting chemosensitivity in CRC through regulating the expression of B lymphoma Mo-MLV insertion region 1 and multidrug resistance protein 5.⁴¹ Tumor exosomes promoted Th17 cell differentiation through the transfer of long non-coding RNA colorectal neoplasia differentially expressed in CRC.⁴² Exosomal circRNA erythrocyte membrane protein band 4.1 like 2 constrained CRC progression, which interacted with miR-21-5p and miR-942-5p and regulated the phosphatase and tensin homolog/protein kinase B signaling pathway.⁴³ In our study, we found that cancer-related exosomal circ-PNN acted as a promoter in CRC, and accelerated cells proliferation, migration, invasion and tumor metastatic. Hence, exosomal circ-PNN might be a potential diagnosis and therapeutic target for CRC.

Altogether, our results proved that exosomal circ-PNN enhanced FGF13 expression in CRC cells via competitively binding to miR-1225-5p, thereby promoted CRC progression. Our findings provided a novel direction for seeking efficient therapeutic strategies for CRC.

No potential conflict of interest relevant to this article was reported.

Study concept and design: X.L., T.L. Data acquisition: L.Y. Data analysis and interpretation: H.L., Y.L. Drafting of the manuscript: X.L., T.L. Critical revision of the manuscript for important intellectual content: X.L., T.L. Statistical analysis: H.L., W.L. Administrative, technical, or material support; study supervision: X.L., Y.L., Z.X. Approval of final manuscript: all authors.

The analyzed data sets generated during the present study are available from the corresponding author on reasonable request.

Supplementary materials can be accessed at https://doi.org/10.5009/gnl230304.