Exosome-Mediated Transfer of circ_0000338 Enhances 5-Fluorouracil Resistance in Colorectal Cancer through Regulating MicroRNA 217 (miR-217) and miR-485-3p

Exosomes are microvesicles secreted by body cells for intercellular communication. The circular RNA circ_0000338 was found to be present in extracellular vesicles and improve the chemoresistance of colorectal cancer (CRC) cells.

ABSTRACT

Exosomes are microvesicles secreted by body cells for intercellular communication. The circular RNA circ_0000338 was found to be present in extracellular vesicles and improve the chemoresistance of colorectal cancer (CRC) cells. However, the role of exosomal circ_0000338 in 5-fluorouracil (5-FU) resistance in CRC is largely unknown. The levels of circ_0000338, microRNA 217 (miR-217), and miR-485-3p were detected using quantitative real-time PCR (qRT-PCR). The 50% inhibitory concentration (IC₅₀) values of cells for 5-FU, cell proliferation, and apoptosis were evaluated using cell counting kit 8 (CCK-8), colony formation, flow cytometry, and Western blot assays. The interaction between miR-217 or miR-485-3p and circ_0000338 was confirmed by RNA immunoprecipitation (RIP), dual-luciferase reporter, and pulldown assays. Exosomes were isolated by ultracentrifugation and qualified by transmission electron microscopy (TEM), Nanosight tracking analysis (NTA), and Western blotting. Xenograft models were performed to analyze whether circ_0000338-loaded exosomes could increase resistance of CRC cells to 5-FU in vivo. The circ_0000338 level was elevated in 5-FU-resistant CRC tissues and cells, and circ_0000338 knockdown sensitized 5-FU-resistant CRC cells to 5-FU through enhancing apoptosis and decreasing proliferation in vitro. Mechanistically, circ_0000338 directly bound to miR-217 and miR-485-3p, and the inhibition of miR-217 or miR-485-3p reversed the effects of circ_0000338 knockdown on cell 5-FU resistance in CRC. Additionally, extracellular circ_0000338 could be incorporated into secreted exosomes and transmitted to 5-FU-sensitive cells. Treatment-sensitive cells with exosomes containing circ_0000338 reduced the 5-FU response in CRC both in vitro and in vivo. Besides that, the exosomal circ_0000338 concentration was higher in patients exhibiting resistance to 5-FU and showed good diagnostic efficiency in 5-FU-resistant CRC. The delivery of circ_0000338 via exosomes enhanced 5-FU resistance in CRC through negative regulation of miR-217 and miR-485-3p, indicating a promising diagnostic and therapeutic marker for 5-FU-based chemotherapy in CRC patients.

INTRODUCTION

As the world’s third most deadly cancer, colorectal cancer (CRC) is a serious threat to human health (1). Although the 5-year survival rate in patients with early CRC (stages I and II) is over 60%, more than half of patients are diagnosed with an advanced stage, in which case the 5-year survival rate drops to 10% once metastasis occurs (2). Surgical resection combined with chemotherapeutic intervention is the basis of advanced CRC therapy and the main means to improve average survival time (3). 5-Fluorouracil (5-FU) is the first-line and standardized chemotherapeutic agent for CRC treatment, while treatment efficiency is often impeded by the acquisition of 5-FU resistance (4, 5). Thus, elucidating the underlying mechanisms and developing new strategies to overcome 5-FU resistance are urgently needed.

Exosomes are small spherical packages and one of the vesicle types that are released from most mammalian cell types into the extracellular milieu through fusion with the cytomembrane; they can act as mediators of intercellular communication, conveying information cargo like proteins, RNAs, and lipids to neighboring or distant cells, and subsequently modulate the behavior of recipient cells, thereby being involved in the regulation of the physiological and pathological development of diseases, including cancers (6–8). Tumor cells have been found to release exosomes actively with high abundances in the systemic circulation of tumor patients, and cancer-derived exosomes have the ability to confer chemoresistance potential from drug-resistant tumor cells to sensitive ones (9, 10). Exosomes show great potential in drug resistance mechanisms.

Circular RNAs (circRNAs) are endogenous transcripts that contain a ring structure formed by the joining of 3′ and 5′ ends of circRNAs, making them less sensitive to exonucleases and more stable than linear RNAs (11) and also the RNAs loaded in cancer-derived exosomes. Increasing evidence has indicated that circRNAs, as transcriptional/posttranscriptional regulators, play crucial roles in a wide variety of biological processes related to carcinogenesis, metabolism, and drug resistance, implicating them in the occurrence and development of malignancies (12–14). circRNAs may be promising candidates used for diagnosis or therapy in various cancers. circ_0000338 has been reported to be elevated in CRC tissues and cell lines and present in extracellular vesicles; besides that, it was also demonstrated to enhance chemoresistance in CRC (15, 16). However, whether exosomal circ_0000338 is involved in CRC 5-FU resistance remains unclear.

Here, this study focused on investigating the role and molecular mechanism of circ_0000338 in 5-FU resistance of CRC; besides that, we also elucidated whether exosome-transmitted circ_0000338 conferred drug resistance to sensitive cells, which may indicate a novel exosome-based therapeutic method to overcome 5-FU resistance in CRC patients.

RESULTS

circ_0000338 is upregulated in 5-FU-resistant CRC tissues and cells.

To reveal the potential role of circ_0000338 in 5-FU resistance, we first determined its expression profile in response to 5-FU-resistant CRC tissues. In contrast to CRC tissues from 5-FU-sensitive patients (n = 24), the circ_0000338 expression level was 1.5-fold higher in tissues of nonresponders (n = 36) (Fig. 1A; patient information is listed in Table 1). Next, 5-FU-resistant CRC cells (SW480/5-FU and HCT116/5-FU) were constructed. As illustrated in Fig. 1B and C, the 50% inhibitory concentration (IC₅₀) value of SW480/5-FU cells or HCT116/5-FU cells for 5-FU was almost 4.7-fold or 4.4-fold higher than that of SW480 cells or HCT116 cells, respectively, suggesting that 5-FU-resistant CRC cells are well established. After that, circ_0000338 expression was determined in CRC cells in vitro. The expression levels of circ_0000338 in SW480/5-FU and HCT116/5-FU cells were upregulated by 2.6- and 2.2-fold compared with the parent cells (Fig. 1D and E). These results indicated that circ_0000338 might play a key role in 5-FU resistance in CRC.

FIG 1.

circ_0000338 is upregulated in 5-FU-resistant CRC tissues and cells. (A) Levels of detection of circ_0000338 in CRC tissues from 5-FU-sensitive or -resistant CRC patients using a qRT-PCR assay. (B) CCK-8 assay of the IC₅₀ values of SW480 and SW480/5-FU cells for 5-FU. (C) CCK-8 assay of the IC₅₀ values of HCT116 and HCT116/5-FU cells for 5-FU. (D and E) qRT-PCR analysis of circ_0000338 expression in SW480/5-FU cells or HCT116/5-FU cells and parent SW480 and HCT116 cells. **, P < 0.01; ***, P < 0.001.

TABLE 1.

Clinicopathological features of CRC patients

Parameter	No. of patients (n = 60)
Gender
Male	33
Female	27
Age (yrs)
<60	38
≥60	22
Tumor thickness (mm)
<4	35
≥4	25
TNM stage(s)
I + II	47
III	13
Lymph node metastasis
Yes	16
No	44
5-FU therapy response
Sensitive	36
Resistant	24

Identification of circ_0000338 in 5-FU-resistant CRC cells.

The characteristics of circ_0000338 were then investigated. circ_0000338 is spliced from exons 4 to 9 of its host gene FCHSD2, located at chromosome 11 (chr11) positions 72632872 to 72726930, and the length of circ_0000338 is 663 nucleotides (nt) (Fig. 2A). In actinomycin D and RNase R assays, we found that circ_0000338 was more stable than linear FCHSD2 mRNA in SW480/5-FU and HCT116/5-FU cells that were resistant to actinomycin D (Fig. 2B and C) and RNase R (Fig. 2D and E) digestion. Besides that, the data from subcellular localization assays indicated that circ_0000338 was highly enriched in the cytoplasm fraction in SW480/5-FU and HCT116/5-FU cells (Fig. 2F and G). All these data suggested that circ_0000338 is an abundant, circular, and stable transcript in CRC cells. In addition, to validate the function of circ_0000338 in 5-FU resistance, three short hairpin RNAs (shRNAs) were designed to knock down circ_0000338 in SW480/5-FU and HCT116/5-FU cells, and the knockdown efficiency of the three shRNAs targeting circ_0000338 (sh-circ_0000338) was verified using quantitative real-time PCR (qRT-PCR). As expected, sh-circ_0000338 transfection markedly reduced the level of circ_0000338 in SW480/5-FU and HCT116/5-FU cells (Fig. 2H and I), and we selected sh-circ#1 for subsequent experiments because it was the most effective shRNA.

FIG 2.

Identification of circ_0000338 in 5-FU-resistant CRC cells. (A) Schematic illustration demonstrating that the circularization of exons 4 to 9 of FCHSD2 forms circ_0000338 by a “head-to-tail” junction. The top black arrow represents the splicing sites. (B to E) qRT-PCR analysis of circ_0000338 and linear FCHSD2 mRNA in SW480/5-FU and HCT116/5-FU cells treated with actinomycin D or RNase R. (F and G) qRT-PCR analysis of the levels of circ_0000338, 18S rRNA, and U6 in purified SW480/5-FU and HCT116/5-FU nuclear and cytoplasmic fractions. (H and I) qRT-PCR analysis of circ_0000338 expression levels in SW480/5-FU and HCT116/5-FU cells transfected with sh-NC or three forms of sh-circ_0000338. **, P < 0.01; ***, P < 0.001.

circ_0000338 knockdown reverses 5-FU resistance in CRC cells in vitro.

Subsequently, a cell counting kit 8 (CCK-8) assay showed that circ_0000338 knockdown combined with increasing doses of 5-FU (0.5, 1, 2, 4, 8, 16, and 32 μM) gradually inhibited the viability of SW480/5-FU and HCT116/5-FU cells, and the IC₅₀ values of circ_0000338-decreased SW480/5-FU and HCT116/5-FU cells for 5-FU were significantly reduced relative to cells in negative-control shRNA (sh-NC) groups (Fig. 3A and B). Similarly, colony formation assays revealed that sh-circ#1 combined with 2 μM 5-FU gradually decreased the number of colonies formed by SW480/5-FU and HCT116/5-FU cells (Fig. 3C and D). Flow cytometric and Western blot analyses showed that the apoptosis rate of SW480/5-FU and HCT116/5-FU cells infected with sh-circ#1 increased gradually, with a codependent elevation of the apoptosis-related protein cleaved caspase-3, compared with control cells transfected with the negative-control vector in response to 2 μM 5-FU (Fig. 3E to H). Taken together, these results show that the knockdown of circ_0000338 sensitized SW480/5-FU and HCT116/5-FU cells to 5-FU through the enhancement of apoptosis and the decrease of proliferation.

FIG 3.

circ_0000338 knockdown reverses 5-FU resistance in CRC cells in vitro. (A and B) CCK-8 assay of the IC₅₀ values of SW480/5-FU and HCT116/5-FU cells for 5-FU after circ_0000338 knockdown. (C and D) Colony formation assay of the proliferation of circ_0000338-decreased SW480/5-FU and HCT116/5-FU cells with 2 μM 5-FU treatments. (E and F) Flow cytometric analysis of cell apoptosis in circ_0000338-decreased SW480/5-FU and HCT116/5-FU cells under 2 μM 5-FU treatments. (G and H) Western blot analysis of cleaved caspase-3 and caspase-3 protein levels in circ_0000338-decreased SW480/5-FU and HCT116/5-FU cells under 2 μM 5-FU treatments. **, P < 0.01; ***, P < 0.001. DMSO, dimethyl sulfoxide.

circ_0000338 directly targets miR-217 and miR-485-3p.

Previous studies have reported that circRNAs in the cytoplasm may competitively bind to microRNAs (miRNAs) and subsequently modulate their target genes by functioning as miRNA sponges (17, 18). Given that circ_0000338 was predominantly localized in the cytoplasm and displayed high stability, we speculated that circ_0000338 could bind miRNAs to regulate its downstream function. First, we conducted an RNA precipitation (RIP) assay in SW480/5-FU and HCT116/5-FU cells. By qRT-PCR, we discovered that endogenous circ_0000338 pulled down from Ago2 antibodies was significantly enriched compared with circ_0000338 stable knockdown cells, suggesting that circ_0000338 interacts with and binds miRNAs through the Ago2 protein (Fig. 4A and B). Next, the circinteractome and circBank databases were used to predict the potential target miRNAs (19, 20), among which 4 miRNAs were selected from the overlap between the databases (Fig. 4C). After that, a pulldown assay was performed with the biotin-labeled circ_0000338 probe and indicated that miRNA 217 (miR-217) and miR-485-3p were abundantly captured by circ_0000338 in both SW480/5-FU and HCT116/5-FU cells (Fig. 4D and E). Thus, miR-217 and miR-485-3p were selected, and the binding sites between circ_0000338 and miR-217 or miR-485-3p were determined (Fig. 4F). To verify that circ_0000338 directly targeted miR-217 or miR-485-3p, the transfection efficiency of an miR-217 or miR-485-3p mimic was first investigated in SW480/5-FU and HCT116/5-FU cells by qRT-PCR; as expected, the level of the miR-217 or miR-485-3p mimic was significantly increased in cells compared with the negative-control miRNA (miR-NC) (Fig. 4G to J). Next, a dual-luciferase reporter assay was set up, and the results showed that the overexpression of miR-217 or miR-485-3p in SW480/5-FU and HCT116/5-FU cells significantly weakens the luciferase activities of the reporter vectors containing wild-type circ_0000338 (circ_0000338-WT), while there was no distinct difference in the reporter vectors containing circ_0000338 mutant 1 (circ_0000338-MUT1) or circ_0000338-MUT2 (Fig. 4K to N). Additionally, it was also observed that the miR-217 or miR-485-3p expression level was higher in circ_0000338-decreased SW480/5-FU and HCT116/5-FU cells (Fig. 4O and P). Interestingly, we found that the expression level of miR-217 or miR-485-3p was significantly downregulated in 5-FU-resistant CRC cells and tissues (Fig. 4Q to T), whose expression patterns were opposite of those of cells with circ_0000338 expression. Altogether, all these results suggested that circ_0000338 suppressed miR-217 or miR-485-3p expression in CRC cells in a targeted manner.

FIG 4.

circ_0000338 directly targets miR-217 and miR-485-3p. (A and B) Ago2 RIP assay for circ_0000338 levels in SW480/5-FU and HCT116/5-FU cells stably expressing sh-circ#1 or sh-NC. GAPDH, glyceraldehyde-3-phosphate dehydrogenase. (C) Schematic illustration showing the overlap of the target miRNAs of circ_0000338 predicted by the circinteractome and circBank databases. (D and E) qRT-PCR analysis of 4 miRNA candidates in SW480/5-FU and HCT116/5-FU cells after a pulldown assay with the biotin-labeled circ_0000338 probe. (F) Binding sites between circ_0000338 and miR-217 or miR-485-3p. (G to J) qRT-PCR analysis of miR-217 and miR-485-3p expression in SW480/5-FU and HCT116/5-FU cells transfected with miR-NC, miR-217, or miR-485-3p. (K and L) Luciferase activity analysis in SW480/5-FU and HCT116/5-FU cells cotransfected with circ_0000338-WT or -MUT1 and miR-NC or miR-217. (M and N) Luciferase activity analysis in SW480/5-FU and HCT116/5-FU cells cotransfected with circ_0000338-WT or -MUT2 and miR-NC or miR-485-3p. (O and P) qRT-PCR analysis of miR-217 and miR-485-3p expression in circ_0000338 knockdown SW480/5-FU and HCT116/5-FU cells. (Q to T) Levels of detection of miR-217 and miR-485-3p in SW480/5-FU and HCT116/5-FU cells, parent SW480 and HCT116 cells, as well as CRC tissues from 5-FU-sensitive or -resistant CRC patients. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

circ_0000338 sensitizes CRC cells to 5-FU through regulating miR-217 or miR-485-3p.

We then studied whether the effect of circ_0000338 on 5-FU resistance was mediated by miR-217 or miR-485-3p. First, the interference efficiency of the miR-217 or miR-485-3p inhibitor was determined, and qRT-PCR analysis showed that the introduction of anti-miR-217 or anti-miR-485-3p significantly reduced the level of miR-217 or miR-485-3p in SW480/5-FU and HCT116/5-FU cells compared with the corresponding negative controls (Fig. 5A). Next, we cotransfected sh-circ#1 and anti-miR-217 or anti-miR-485-3p into SW480/5-FU and HCT116/5-FU cells to conduct rescue assays. The CCK-8 and colony formation assays suggested that the knockdown of circ_0000338 significantly suppressed cell proliferation activity and increased the sensitivity of SW480/5-FU and HCT116/5-FU cells to 5-FU, while anti-miR-217 or anti-miR-485-3p infection rescued the effects (Fig. 5B to D). Furthermore, flow cytometry and Western blot assays indicated that transfection with sh-circ#1 led to increases of the cell apoptosis rate (Fig. 5E and F) and cleaved caspase-3 level (Fig. 5G and H) in SW480/5-FU and HCT116/5-FU cells under 5-FU treatment, which was abolished by the downregulation of miR-217 or miR-485-3p.

FIG 5.

circ_0000338 sensitizes CRC cells to 5-FU through regulating miR-217 or miR-485-3p. (A) qRT-PCR analysis of miR-217 and miR-485-3p expression in SW480/5-FU and HCT116/5-FU cells transfected with anti-miR-NC, anti-miR-217, or anti-miR-485-3p. SW480/5-FU and HCT116/5-FU cells were cotransfected with sh-circ#1 and anti-miR-217 or anti-miR-485-3p. (B to H) After transfection, a CCK-8 assay of the IC₅₀ values of SW480/5-FU and HCT116/5-FU cells for 5-FU (B), proliferation analysis of SW480/5-FU and HCT116/5-FU cells in 2 μM 5-FU treatments by a colony formation assay (C and D), flow cytometric analysis of SW480/5-FU and HCT116/5-FU cell apoptosis in 2 μM 5-FU treatments (E and F), and Western blot analysis of cleaved caspase-3 and caspase-3 protein levels in SW480/5-FU and HCT116/5-FU cells with 2 μM 5-FU treatments (G and H) were performed. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

In addition, we also investigated the role of miR-217 or miR-485-3p in 5-FU resistance in CRC cells. The CCK-8 assay suggested that a decrease of miR-217 or miR-485-3p increased the IC₅₀s of SW480/5-FU and HCT116/5-FU cells for 5-FU compared with the respective control cells. Similarly, colony formation analysis indicated that inhibition of miR-217 or miR-485-3p combined with 2 μM 5-FU gradually increased the number of colonies formed. Moreover, the apoptosis of SW480/5-FU and HCT116/5-FU cells induced by 5-FU was significantly reduced by the downregulation of miR-217 or miR-485-3p compared with the respective control cells. These data revealed that circ_0000338 knockdown reduced 5-FU resistance in CRC cells by regulating miR-217 or miR-485-3p.

Extracellular circ_0000338 is packaged into exosomes and can be internalized by 5-FU-sensitive CRC cells.

Exosomes can be actively secreted by most cell types, and exosome-contained contents can be transmitted and exchanged into surrounding cells (21, 22). Nucleotide is one of the inclusions of exosomes (23), and circRNAs transferred by exosomes confer chemoresistance to CRC-sensitive cells (15). To explore the impact of exosome transfer on 5-FU resistance in CRC, the exosomes in the culture media of 5-FU-resistant CRC cell lines were isolated by high-speed centrifugation. Transmission electron microscopy (TEM) analysis confirmed the presence of translucent, cup-shaped vesicles (Fig. 6A), and Nanosight tracking analysis (NTA) showed particle sizes in the 120-nm range, consistent with the unique characteristics of exosomes (Fig. 6B). In addition, the detection of characteristic CD63, CD81, and TSG101 by Western blotting further verified that the isolated particles were exosomes, while GM130 was mainly present in the cell lysate (Fig. 6C). Subsequently, the expression of circ_0000338 in exosomes was detected, and we found that the circ_0000338 level was higher in exosomes of SW480/5-FU and HCT116/5-FU cells (SW480/5-FU-Exo and HCT116/5-FU-Exo) than that in exosomes of parent cells (Fig. 6D and E). Next, SW480 and HCT116 cells were cocultured with SW480/5-FU-Exo and HCT116/5-FU-Exo, respectively, and qRT-PCR analysis showed that the circ_0000338 level was significantly increased in the culture media of SW480 and HCT116 cells (Fig. 6F and G), suggesting that circ_0000338 was wrapped by exosomes and could be assimilated by 5-FU-sensitive CRC cells. These results suggest that the circ_0000338 contained in exosomes could be taken up by recipient cells.

FIG 6.

Extracellular circ_0000338 is packaged into exosomes and can be internalized by 5-FU-sensitive CRC cells. (A) Representative TEM images of exosomes from SW480/5-FU and HCT116/5-FU cells. (B) Exosomes isolated from SW480/5-FU and HCT116/5-FU cells underwent an NTA to determine the exosomal size distribution and concentration. (C) Western blot analysis of CD63, CD81, TSG101, and GM130 in isolated exosomes and the lysates of SW480/5-FU and HCT116/5-FU cells. (D and E) qRT-PCR analysis of circ_0000338 levels in exosomes of SW480/5-FU and HCT116/5-FU cells as well as of parent SW480 and HCT116 cells. (F and G) qRT-PCR analysis of circ_0000338 levels in SW480 and HCT116 cells treated with PBS, SW480/5-FU-Exo, or HCT116/5-FU-Exo. **, P < 0.01; ***, P < 0.001.

Intercellular transfer of circ_0000338 by exosomes confers 5-FU resistance in vitro.

To identify whether circ_0000338 regulated 5-FU resistance via delivery using exosomes, we determined whether SW480 and HCT116 cells with elevated exosomal circ_0000338 expression levels displayed increased resistance to 5-FU treatment compared to control cells. SW480 and HCT116 cells were incubated with phosphate-buffered saline (PBS) or exosomes from SW480/5-FU and HCT116/5-FU cells transfected with sh-NC or sh-circ#1 (SW480/5-FU/-sh-NC Exo, HCT116/5-FU-sh-NC Exo, SW480/5-FU/-sh-circ#1 Exo, or HCT116/5-FU-sh-circ#1 Exo) for 48 h, and qRT-PCR showed that circ_0000338 expression was markedly increased in both recipient cell lines with SW480/5-FU/-sh-NC Exo and HCT116/5-FU-sh-NC Exo (Fig. 7A). Moreover, the intercellular transfer of circ_0000338 via exosomes also negatively regulated miR-217 or miR-485-3p expression in SW480 and HCT116 cells. After that, functional experiments were performed. We found that the IC₅₀ values in the recipient cell lines with circ_0000338-increased exosomes were significantly increased (Fig. 7B). Meanwhile, the proliferation and apoptosis of SW480 and HCT116 cells in different groups were evaluated using colony formation, flow cytometry, and Western blot assays. The results suggested that upregulation of the circ_0000338 level in SW480 and HCT116 cells through exosome delivery promoted cell proliferation (Fig. 7C and D) and decreased cell apoptosis (Fig. 7E to H) when treated with 5-FU (2 μM for 48 h). Altogether, the exosome-mediated transfer of circ_0000338 disseminated 5-FU resistance in CRC cells.

FIG 7.

Intercellular transfer of circ_0000338 by exosomes confers 5-FU resistance in vitro. SW480 and HCT116 cells were incubated with PBS or exosomes from SW480/5-FU and HCT116/5-FU cells transfected with sh-NC or sh-circ#1. After treatment, qRT-PCR analysis of circ_0000338 levels in treated SW480 and HCT116 cells (A), a CCK-8 assay of the IC₅₀ values of SW480 and HCT116 cells for 5-FU (B), a colony formation assay of the proliferation of SW480 and HCT116 cells in 2 μM 5-FU treatments (C and D), flow cytometric analysis of SW480 and HCT116 cell apoptosis in 2 μM 5-FU treatments (E and F), and detection of cleaved caspase-3 and caspase-3 protein levels in SW480 and HCT116 cells in 2 μM 5-FU treatments using Western blot analysis (G and H) were performed. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Exosomal circ_0000338 enhances 5-FU resistance in vivo.

To determine the effect of circ_0000338 on 5-FU resistance in vivo, we established models of nude mice bearing 5-FU xenografts with SW480 cells, which were treated differently after tumor formation. As shown in Fig. 8A and B, xenografts carrying circ_0000338-increased exosomes showed a marked advancement of tumor growth in response to 5-FU. Tumors with high circ_0000338 levels (SW480/5-FU/-sh-NC Exo treated) led to lower miR-217 and miR-485-3p expression levels (Fig. 8C to E) as well as lower protein levels of cleaved caspase-3 (Fig. 8F). Collectively, these data confirmed that exosomal circ_0000338 decreased the cytotoxic effects of 5-FU in vivo.

FIG 8.

Exosomal circ_0000338 enhances 5-FU resistance in vivo. (A) Subcutaneous tumor growth curve for each group of nude mice. (B) Tumor weight comparison histogram of each group of nude mice. (C to E) qRT-PCR analysis of circ_0000338, miR-217, and miR-485-3p levels in tumors. (F) Western blot analysis of cleaved caspase-3 and caspase-3 protein levels in tumors. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

The serum exosomal circ_0000338 level is associated with 5-FU resistance in CRC patients.

Exosomal RNAs can be transferred to neighboring cells or cells in other organs via the circulation system; thus, we speculated that exosomal circ_0000338 secreted by CRC cells might be detected in circulation. We extracted exosomes from 36 serum samples of CRC patients who received 5-FU treatment. Patients were divided into responding (5-FU-sensitive) (n = 18) and nonresponding (5-FU-resistant) (n = 18) groups (patient information is listed in Table 2). After that, the cup-shaped structure, size, and number of the isolated exosomes were authenticated by TEM and Western blot analyses (Fig. 9A and B). Additionally, circ_0000338 expression was detectable in extracted serum exosomes and was more highly expressed in serum exosomes from nonresponders than in those from 5-FU-sensitive groups (Fig. 9C). Considering that better stability is a vital prerequisite for tumor markers, we next evaluated the stability of exosomal circ_0000338 in serum exosomes by exposing exosomes to different conditions, including incubation at room temperature for 0, 6, 12, and 24 h and a low-pH (pH = 1) or high-pH (pH = 13) solution at room temperature for 3 h. Our results showed that the levels of exosomal circ_0000338 were not significantly impacted by any of the experimental conditions (Fig. 9D and E), revealing the high stability of exosomal circ_0000338 in serum exosomes. Furthermore, the diagnostic efficiency of exosomal circ_0000338 in serum was assessed. As shown by receiver operating characteristic (ROC) analysis, an area under the curve (AUC) of 0.8489, with diagnostic sensitivity and specificity reaching 83.33% and 77.78%, respectively (95% confidence interval [CI] = 0.7233 to 0.9741), was observed (Fig. 9F). Based on the above-mentioned results, we suggest that exosomal circ_0000338 in serum was stable and had better diagnostic efficiency for 5-FU-resistant CRC patients.

TABLE 2.

Clinicopathological features of CRC patients

Parameter	No. of patients (n = 36)
Gender
Male	19
Female	17
Age (yrs)
<60	24
≥60	12
Tumor thickness (mm)
<4	25
≥4	11
TNM stage(s)
I + II	29
III	7
Lymph node metastasis
Yes	8
No	28
5-FU therapy response
Sensitive	18
Resistant	18

FIG 9.

The serum exosomal circ_0000338 level is associated with 5-FU resistance in CRC patients. (A) Representative TEM images of exosomes isolated from 36 serum samples of CRC patients who received 5-FU treatment. (B) Western blot analysis of CD63, CD81, TSG101, and GM130 in isolated exosomes. (C) qRT-PCR analysis of circ_0000338 expression in extracted serum exosomes. (D and E) qRT-PCR analysis of exosomal circ_0000338 expression levels in serum exosomes that underwent long exposure (D) or pH solution incubation (E). (F) ROC curve analysis of the diagnostic value of exosomal circ_0000338 in CRC patients receiving 5-FU treatment. ***, P < 0.001.

DISCUSSION

CRC is one of the leading causes of cancer-related deaths in the world; despite recent progress in multimodality treatment methods, it continues to be associated with poor survival, especially when diagnosed at a late stage (24, 25). 5-FU is commonly prescribed for the treatment of CRC patients, which is the only chemotherapeutic agent available to successfully improve 12-month survival in CRC patients (2). As a cytotoxic agent, 5-FU leads to growth arrest and apoptosis promotion in cancer cells through inducing RNA and DNA double-strand breakage (5, 26). However, the development of acquired 5-FU resistance has emerged as one of the main reasons for failure in 5-FU-based chemotherapies of CRC; only 10 to 15% of advanced CRC patients had an effective response to the 5-FU therapy (27, 28).

Currently, emerging evidence suggests that circRNAs are functional molecules and play key roles in regulating gene expression and pathological processes in cancers (29, 30). circRNAs are stable and conserved across different species and are considered ideal candidates for future therapeutic interventions and as diagnostic biomarkers (31). In this study, we found that circ_0000338 was highly expressed in 5-FU-resistant CRC tissues and cells, and knockdown of circ_0000338 sensitized 5-FU-resistant cells to 5-FU in CRC through the reduction of proliferation and the promotion of apoptosis.

At present, numerous research findings have pointed out that exosomes mediate the regulation of the tumor microenvironment to contribute to chemoresistance in many cancers, including CRC (15, 32, 33). There is great interest in employing exosomes as therapies and in manipulating exosomal content for therapeutic outcomes due to its low risks of cancer formation and adverse immune responses (34). Importantly, exosome-carried therapeutic circRNAs can be manufactured in bulk through exosome-producing cells in vitro, thus enabling personalized treatment (15). All these reports show the rationale for developing exosome-based therapeutic methods for targeting CRC with 5-FU resistance. In this study, we found that extracellular circ_0000338 was packed in exosomes and could be internalized by 5-FU-sensitive cells via exosomes derived from 5-FU-resistant cells. Treatment-sensitive cells with exosomes containing circ_0000338 conferred 5-FU resistance by arresting cell proliferation and promoting cell apoptosis in vitro and reduced the 5-FU response in nude mice in vivo. Moreover, we also proved that the exosomal circ_0000338 level was higher in the serum of 5-FU-resistant CRC patients and had good diagnostic value for 5-FU resistance in CRC patients.

Convincing evidence has shown that circRNAs in the cytoplasm may competitively bind to miRNAs and exert their functions by serving as miRNA sponges (17, 18). circ_0000338 was predominantly localized in the cytoplasm and displayed high stability, indicating the posttranscriptional regulation of circ_0000338 on gene expression. Thus, the target miRNAs of circ_0000338 were searched, and miR-217 and miR-485-3p were confirmed to be the targets of circ_0000338. miR-217 and miR-485-3p have been revealed to serve as tumor suppressors in many types of cancers, including CRC (35, 36). For example, miR-217 has considerable value as a prognostic marker, and reexpression of miR-217 could suppress CRC cell malignant phenotypes by targeting downstream genes such as AEG-1, mitogen-activated protein kinase (MAPK), and ZEB1 (37–39). miR-485-3p was found to impede CRC progression via mediating TPX2 inhibition (40). These researches presented evidence that miR-217 and miR-485-3p acted as suppressors of CRC. In this work, we proved that circ_0000338 suppressed miR-217 and miR-485-3p expression in CRC in a targeted manner. Reexpression of miR-217 or miR-485-3p reduced 5-FU resistance in resistant CRC cells, and importantly, inhibition of miR-217 or miR-485-3p abolished the inhibitory effects of circ_0000338 knockdown on 5-FU-resistant CRC cells; thus, the circ_0000338–miR-217/miR-485-3p feedback loops were identified in CRC. Nevertheless, our research has some limitations. First, further research should be carried out to investigate the downstream targets and the possible mechanism underlying the circ_0000338–miR-217/miR-485-3p axis in 5-FU resistance in CRC. In addition, a major emerging process, termed target-directed miRNA degradation (TDMD), employs specialized target RNAs to selectively bind to miRNAs and induce their decay (41, 42), while whether circ_0000338 negatively regulated the expression of miR-217 or miR-485-3p via a TDMD decay mechanism still needs to be further explored.

In conclusion, our research uncovered that the exosome-mediated transfer of circ_0000338 enhanced 5-FU resistance in CRC by targeting miR-217 and miR-485-3p; moreover, exosomal circ_0000338 has considerable value as a diagnostic marker for 5-FU-resistant CRC patients. All these findings provide a promising diagnostic marker and therapeutic intervention for overcoming 5-FU resistance in CRC patients.

MATERIALS AND METHODS

Specimen samples.

Tumor tissues were obtained from surgical specimens or biopsy specimens of a total of 60 CRC patients before initiation of 5-FU treatment at the Second Affiliated Hospital of Soochow University. All patients were diagnosed by histopathological examination. Patients were treated with at least six cycles of 5-FU and then divided into responding (CR+PR, 5-FU-sensitive) (n = 24) and nonresponding (SD+PD, 5-FU-resistant) (n = 36) groups according to the Response Evaluation Criteria in Solid Tumors (RECIST) (version 1.1). Moreover, 5-ml blood samples were also collected from 36 participants. Serum samples were isolated by centrifugation (1,600 × g for 10 min) at room temperature within 1 h after collection and then further centrifuged at 12,000 × g for 10 min to remove residual cellular debris at 4°C. All samples were immediately stored at −80°C until use. All patients had provided informed written consent, and this research had gained authorization from the Ethics Committee of the Second Affiliated Hospital of Soochow University.

Cell culture.

The human CRC cell lines (SW480 and HCT116) were purchased from the Chinese Academy of Sciences (Shanghai, China) and grown in RPMI 1640 medium (Invitrogen, San Diego, CA, USA) supplemented with 100 U/ml penicillin-streptomycin and 10% fetal bovine serum (FBS; Invitrogen) at 37°C in a humidified incubator with 5% CO₂. Next, we established 5-FU-resistant cells, named SW480/5-FU and HCT116/5-FU, by exposing parental cells to increasing concentrations of 5-FU (Sigma-Aldrich, St. Louis, MO, USA) for more than 6 months. The 5-FU-resistant cells were grown in the same medium containing 5 μM 5-FU.

Quantitative real-time PCR.

TRIzol reagent (Invitrogen) was employed to isolate RNAs from cells and tissues. For circRNA analysis, first-strand cDNAs were synthesized using a reverse transcription system kit (TaKaRa Bio Inc., Kusatsu, Japan) and then quantified by qPCR with SYBR premix ExTaq (TaKaRa Bio Inc.). For miRNA analysis, cDNA was synthesized by using the PrimeScript RT reagent kit (TaKaRa). SYBR premix ExTaq II (TaKaRa) was used for qPCR. The following thermocycling conditions were used for the qPCR: an initial denaturation step at 95°C for 3 min and 40 cycles of 95°C for 5 s and 60°C for 30 s. A melt curve step from 65°C to 95°C was performed in increments of 0.5°C per 5 s. The relative fold changes were calculated by comparison to the expression of 18S rRNA or U6 using the 2^−ΔΔCT method. Primers used for qPCR amplification are as follows: circ_0000338 forward (F) primer 5′-ATTCCAAAGCTACCCACGCA-3′ and reverse (R) primer 5′-GGCTACCTGCATTGTTCCCT-3′, FCHSD2 F primer 5′-GAGTATGCACAGGGTATGCAG-3′ and R primer 5′-TCCACACACCTTTTTAGTTGCT-3′, 18S rRNA F primer 5′-GGTCCGTGTTTCAAGACGG-3′ and R primer 5′-GCATATCAATAAGCGGAGGAA-3′, miR-217 F primer 5′-TACTGCATCAGGAACTG-3′ and R primer 5′-GTGCAGGGTCCGAGGT-3′, miR-485-3p F primer 5′-GTCATACACGGCTCTC-3′ and R primer 5′-CCAGTGCAGGGTCCGAGGT-3′, and U6 F primer 5′-AAAGCAAATCATCGGACGACC-3′ and R primer 5′-GTACAACACATTGTTTCCTCGGA-3′.

Cell counting kit 8 assay.

Untransfected or transfected cells were seeded into 96-well plates and treated with different concentrations of 5-FU (0.5, 1, 2, 4, 8, 16, and 32 μM) for 48 h. Next, cell viability was detected at an optical density at 450 nm (OD₄₅₀) using a microplate reader with a cell counting kit 8 (CCK-8) solution (10 μl; Dojindo Laboratories, Kumamoto, Japan), and the 50% inhibitory concentration (IC₅₀) value was calculated for each CRC cell line.

Actinomycin D and RNase R treatment.

In actinomycin D assays, RNA extracts (50 μg) were incubated with 2 μg/ml actinomycin D for 0, 6, 12, and 24 h. For the RNase R treatment, RNA extracts (50 μg) were incubated without or with 3 U/μg RNase R for 20 min at 37°C, followed by purification with TRIzol (Invitrogen). Finally, the abundances of circ_0000338 and FCHSD2 mRNA were assessed using a qRT-PCR assay as described above.

Subcellular fractionation.

According to the instructions of the manufacturer, we used a cytoplasmic and nuclear RNA purification kit (Thermo Fisher Scientific, Waltham, MA, USA) to extract and purify the RNAs from the cytosolic and nuclear fractions. The relative expression levels of circ_0000338, U6, and 18S rRNA in the nuclear and cytoplasmic fractions of SW480/5-FU and HCT116/5-FU cells were calculated using a qRT-PCR assay as described above.

Cell transfection.

Short hairpin RNAs (shRNAs) targeting circ_0000338 (sh-circ_0000338) (sh-circ#1, sh-circ#2, and sh-circ#3), the mimic and inhibitor of miR-217 (miR-217 and anti-miR-217) or miR-485-3p (miR-485-3p and anti-miR-485-3p), and their negative controls (sh-NC, miR-NC, and anti-miR-NC) were obtained from GenePharma (Shanghai, China). Next, transfection of cells with sh-circ_0000338 (10 nM) or sh-NC (10 nM) and miRNA mimics or inhibitors (50 nM) was performed using Lipofectamine 2000 (Invitrogen).

Colony formation assay.

Transfected cells (0.5 × 10³ cells/well) were independently plated into a 6-well plate, followed by treatment with 2 μM 5-FU for 24 h. Next, the culture media were replaced by complete medium (10% FBS) without 5-FU until colonies were visible (nearly 2 weeks). After fixing and staining, colonies were imaged (magnification, ×100) and counted using a microscope.

Flow cytometry.

Cells were transfected with the indicated vectors or miRNA mimics, followed by treatment with 2 μM 5-FU for 24 h. Next, cells were washed twice with phosphate-buffered saline (PBS) (Sigma-Aldrich) and resuspended in 1× annexin V binding buffer (1 × 10⁶ cells/ml). Thereafter, cells were mixed with annexin V-fluorescein isothiocyanate (FITC) (10 μl) and propidium iodide (PI) (10 μl) (BD Biosciences, San Jose, CA, USA) for 20 min away from light. Finally, cell apoptosis was detected using a flow cytometer (BD Biosciences).

Western blotting.

Total proteins of tissues, cells, or exosomes were extracted and homogenized in ice-cold lysis buffer (Beyotime, Shanghai, China). Protein lysates (50 μg) were subjected to a 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis gel for separation, shifted to polyvinylidene difluoride (PVDF) membranes (Beyotime), and blocked with nonfat dried milk for 1 h. Next, membranes were interacted with primary antibodies against cleaved caspase-3 (1:1,000) (catalog no. 9664; Cell Signaling Technology [CST], Boston, MA, USA), total caspase-3 (1:1,000) (catalog no. 9662; CST), CD81 (1:2,000) (catalog no. ab109201; Abcam, Cambridge, MA, USA), TSG101 (1:5,000) (catalog no. ab125011; Abcam), CD63 (1:2,000) (catalog no. ab68418; Abcam), GM130 (1:5,000) (catalog no. ab52649; Abcam), and β-actin (1:5,000) (catalog no. ab6276; Abcam) overnight. Next, membranes were incubated with a peroxidase-conjugated secondary antibody (Abcam) for 2 h at 37°C. Immunoreactive bands were visualized using an enhanced chemiluminescence kit (Beyotime) and normalized by β-actin.

RNA immunoprecipitation assay.

The Magna RIP RNA-binding protein immunoprecipitation kit (Millipore, Bedford, MA, USA) was employed to conduct RIP assays (RIPAs) in SW480/5-FU and HCT116/5-FU cells. Cells transfected with lentivirus vectors expressing sh-NC or sh-circ#1 were resuspended in RIP lysis buffer (about 100 ml) with a protease inhibitor cocktail and RNase inhibitors. Next, cell lysates were incubated with 5 μg of magnetic bead-conjugated control rabbit IgG or Ago2 antibody (Millipore) overnight at 4°C. After treatment with proteinase K buffer, the immunoprecipitated RNAs were isolated and purified, and the expression levels of circ_0000338 were measured using qRT-PCR assays as described above.

Pulldown assay.

The biotin-labeled circ_0000338 probe or oligonucleotide probe was designed and synthesized by GenePharma (Shanghai, China). SW480/5-FU and HCT116/5-FU cells were then lysed and incubated with circTADA2A probe- or oligonucleotide probe-coated C-1 magnetic beads for 2 h at 25°C. After washing with wash buffer, the bead-bound RNAs were eluted, isolated, and then subjected to qRT-PCR analysis.

Dual-luciferase reporter assay.

SW480/5-FU and HCT116/5-FU cells were seeded onto 48-well plates and cotransfected with the miR-217 mimic, the miR-485-3p mimic, or miR-NC; a firefly luciferase reporter containing wild-type (WT) or mutant (MUT) circ_0137287; and a renilla reporter. The relative luciferase activities were analyzed using a dual-luciferase reporter assay system (Promega, Madison, WI, USA) and normalized by the renilla luciferase activities after 48 h of transfection.

Exosome isolation.

Cell culture fluid from exosome-depleted medium of untransfected or transfected cells was centrifuged at 500 × g for 5 min, followed by centrifugation at 15,000 × g for 30 min to remove cell fragments. After filtering with a 0.22-μm-pore-size filter, the media were collected and ultracentrifuged at 100,000 × g for 1.5 h at 4°C. The exosome pellet was washed with 24 ml PBS and then subjected to a second step of ultracentrifugation at 100,000 × g for 90 min at 4°C. Finally, the supernatant was discarded, and exosomes were collected. After that, purified exosomes resuspended in 50 μl PBS were imaged by transmission electron microscopy (TEM) (magnification, ×200) (JEOL, Akishima, Japan) and quantified by Nanosight tracking analysis (NTA). Exosomes resuspended in 20 to 30 μl of RIPA lysis buffer were used for protein detection. Exosomes used for RNA extraction were interacted with TRIzol reagent. For exosome cocultures, parental SW480 and HCT116 cells (5 × 10⁵) seeded in 6-well plates harboring 10% FBS–exosome-depleted culture medium were incubated with exosomes (50 μg/ml) for 48 h. For the isolation of exosomes from serum samples, frozen serum was thawed on ice, diluted with PBS, and then filtered using a 0.2-μm-pore-size filter. Thereafter, exosomes were isolated as mentioned above.

Xenograft experiments in vivo.

Male BALB/c nude mice (4 weeks old) (n = 15) were purchased from Charles River Labs (Beijing, China) and divided into three groups. SW480 cells (5 × 10⁶ cells) were injected subcutaneously into the right flank of nude mice. When tumors grew to approximately 100 mm³, PBS or purified exosomes (10 μg) were then injected intratumorally twice weekly with 5-FU treatment (50 mg/kg of body weight). The tumor volume was determined every week. At day 28, mice were sacrificed, and tumor masses were collected, weighed, and harvested for further analyses. Animal experiments were approved by the Ethics Committee of the Second Affiliated Hospital of Soochow University.

Statistical analysis.

Each experiment was carried out in triplicate at least, and all numerical results are presented as the means ± standard deviations (SD). Significant differences were identified using Student’s t test or one-way analysis of variance (ANOVA). Receiver operating characteristic (ROC) curves were plotted to determine the diagnostic value. A P value of <0.05 indicated statistical significance.