Clinical specimens

The CRC tissues and their normal counterparts were collected from 45 CRC patients at the Third Xiangya Hospital of Central South University. This study was approved by the Ethics Committee of the Third Xiangya Hospital of Central South University. Written consents were obtained from all participants. The clinicopathological characteristics of CRC patients recruited in the study were included in Table 1.

Table 1 Clinicopathological characteristics of patients with colorectal cancer Cell culture and transfection

Human CRC cell lines (SW480 and HCT116) and HUVECs were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). SW480 and HCT116 cells were cultured in RPMI1640 medium with 10% FBS (Gibco, Grand Island, NY, USA). All cells were maintained at 37 °C and 5% CO2. HUVECs were cultured in conditioned medium (CM) derived from CRC cells and HUVECs medium at a ratio of 1:2. The full-length of circCOL1A1 transcript was cloned into pLO5-ciR vector (Geneseed Biotech, Guangzhou, China). The lentiviral pLKO.1 vector containing short hairpin (shRNA) sequences, including sh-NC (negative control), sh-circCOL1A1 and sh-EIF4A3, were all obtained from Addgene (Watertown, MA, USA). Lentivirus production was conducted using Xfect transfection reagent (Takara, Dalian, China). The full-length of EIF4A3, Smad2 and Smad3 were cloned into pcDNA3.1 vector (Invitrogen). Transfection was performed using Lipofectamine 3000 (Invitrogen).

Bioinformatics analyses

The binding between circCOL1A1 and EIF4A3 protein was predicted by Circular RNA Interactome (https://circinteractome.nia.nih.gov/). The expression of EIF4A3 in CRC and its association with CRC tumor stage and nodal metastasis were analyzed from TCGA data based on UALCAN (https://ualcan.path.uab.edu/index.html). The correlation between Smad2/3 expression and CRC overall survival was analyzed using GEPIA (http://gepia.cancer-pku.cn/), and the binding between EIF4A3 protein and Smad2/3 mRNA was predicted by Starbase (https://rnasysu.com/encori/).

Exosome isolation and characterization

Exosomes were isolated from culture medium of CRC cells using ExoQuick-TC Exosome Precipitation Solution (System Biosciences, Palo Alto, CA, USA). In brief, the cell culture supernatant was collected after centrifugation, and incubated with ExoQuick-TC solution at 4 °C overnight. After centrifugation, exosome pellets were collected. For transmission electron microscopy (TEM) analysis, exosomes were placed on a carbon-coated copper grid and incubated for 3 min, followed by staining with 2% phosphotungstic acid. Exosomes were observed using Hitachi H-9500 TEM (Hitachi, Japan) as previously described (Pi et al. 2021). For nanopartical tracking analysis (NTA), exosomes were observed using Zetasizer (Malvern Panalytical, Malvern, UK).

Exosome uptake assay

Exosome uptake assay was conducted using PKH26 Red Fluorescent Cell Linker Kit (Sigma-Aldrich, St Louis, MO, USA). Briefly, HUVECs (1 × 105) were incubated with SW480 or HCT116 cell-derived exosomes (5 × 1011 particles/mL). After 24 h, HUVECs were harvested for immunofluorescence (IF) analysis. Images were acquired with a confocal microscope (Nikon, Tokyo, Japan).

qRT-PCR

Total RNA was extracted using Trizol (Invitrogen). Reverse transcription was conducted using the High Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific), and qRT-PCR was performed using SYBR Green PCR Master Mix (Thermo Fisher Scientific) on an ABI7500 system (ABI, Foster City, CA, USA). The expression of target gene was calculated using 2–ΔΔCT method. GAPDH was used as an internal control. For RNA stability assay, HUVECs were treated with actinomycin D (5 µg/mL, Sigma-Aldrich) for 0, 3, 6, 9 and 12 h. The stabilities of Smad2/3 mRNA were analyzed by qRT-PCR. The primers were ordered from Sangon Biotech (Shanghai, China).

Western blot

Proteins from exosomes, cells and tissues were prepared using RIPA lysis buffer (Beyotime, Shanghai, China). Proteins were separated by SDS-PAGE, and transferred onto nitrocellulose membranes (Millipore, Bedford, MA, USA). The blots were immunoblotted with primary antibodies, followed by the incubation with secondary antibody. Signals were visualized using ECL substrate (Sigma-Aldrich). Antibodies used in western blot including: CD63 (ab134045, Abcam), TSG101 (ab125011, Abcam), CD9 (ab236630, Abcam), Smad1 (ab126761, Abcam), Smad2 (ab33875, Abcam), Smad3 (ab208182, Abcam), Smad4 (ab40759, Abcam), Smad6 (ab273106, Abcam), Smad7 (ab216428, Abcam), p-Smad2 (ab188334, Abcam), p-Smad3 (ab63403, Abcam), VEGFR (ab11939, Abcam), EIF4A3 (ab32485, Abcam), and GAPDH (ab8245, Abcam).

RNA immunoprecipitation (RIP) assay

RIP was performed using Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore) as described (Huang et al. 2020). Briefly, cells were lysed with RIP buffer. Anti-EIF4A3 antibody (ab32485, Abcam) or normal rabbit IgG-conjugated beads were incubated with cell lysates at 4 °C overnight. The enrichment of circCOL1A1 or Smad2/3 mRNA was detected by qRT-PCR.

RNA pull-down assay

RNA pull-down assay was conducted using Pierce Magnetic RNA-protein pull-down Kit (Thermo Fisher Scientific). In brief, desthiobiotin-labelled circCOL1A1 was conjugated to streptavidin magnetic beads. The probe-conjugated beads were then incubated with cell lysates at 4 °C overnight. RNA-protein complexes were eluted and subjected to western blot.

RNA fluorescence in situ hybridization (FISH) and immunofluorescence (IF)

The FITC-labeled circCOL1A1 or Smad2/3 mRNA probe was synthesized by RiboBio (Guangzhou, China). HUVECs were fixed and permeabilized. RNA FISH was performed using the Fluorescent in Situ Hybridization Kit (RiboBio). The slices were co-stained with anti-EIF4A3 antibody, followed by the incubation with Alexa Fluor 555-conjugated secondary antibody. The localization of circCOL1A1, Smad2/3 mRNA and EIF4A3 were visualized using a Nikon confocal microscope (Nikon). The probes used in FISH were as follows:

CircCOL1A, AAACTGGCCCCCCTGGCCCTGTTGGTGTTC;

Smad2 mRNA, cctaacagaacttccgcctctggatgacta;

Smad3 mRNA, gatggagaaaccagtgaccaccagatgaac.

Cell counting Kit-8 (CCK-8) assay

HUVECs (1 × 103/well) were seeded into 96-well plates 12 h prior to the treatment. Cell proliferation was assessed at 48 h post-treatment by CCK-8 assay (Solarbio, Beijing, China). In brief, 10 µL CCK-8 solution was added into each well and incubated at 37 °C for 1 h. A490 was measured using a microplate reader (Thermo Fisher Scientific).

Wound healing assay

HUVECs were seeded into 6-well plates 12 h prior to the treatment. The cell monolayer was scratched using a pipette tip. The cells were rinsed with PBS to remove the detached cells. After 24 h, the scratches were photographed using a microscope (Nikon). Wound closure ratio was measured using the following formula: (W0 h‑W24 h)/W0 h×100%, where W is the width.

Tube formation assay

Angiogenesis was assessed by tube formation assay as described (Qiu et al. 2021). In brief, HUVECs were seeded onto Matrigel (Corning, Corning, NY, USA)-coated plates. Tube-like structures were photographed 48 h post-treatment using a microscope (Nikon), and the number of tube branches was analyzed using ImageJ software (NIH).

Animal study

All animal studies were approved by the Ethics Committee of the Third Xiangya Hospital of Central South University. Male BALB/c nude mice (6-week-old, n = 7 per group) were purchased from SLAC Laboratory Animal Center (Shanghai, China). SW480 or HCT116 cells (1 × 106) were subcutaneously injected into the flank of mice. For each CRC cell line-derived xenograft model, mice were randomly divided into five groups: control, OE-NC-Exos, OE-circCOL1A1-Exos, sh-NC-Exos and sh-circCOL1A1-Exos. At day 2 post-inoculation, 1 × 1012 exosome particles (suspended in 100 µL PBS) were administered on mice via tail vein twice a week. Tumor size was monitored and determined as follows: volume = 1/2×length×width2. On day 28, xenograft tumors were dissected and weighted, followed by immunohistochemistry (IHC), qRT-PCR, and western blot analyses.

IHC analysis

IHC was performed as previously described (Yuan et al. 2021). The paraffin-embedded sections were deparaffinized and rehydrated. After antigen retrieval, the sections were then incubated with primary antibodies, followed by the incubation with HRP-conjugated secondary antibody. The immunoreactivities were visualized using DAB substrate (Thermo Fisher Scientific). Antibodies used in IHC analysis were anti-Ki-67 (ab15580, Abcam), anti-E-cadherin (ab231303, Abcam), anti-N-cadherin (ab76011, Abcam), anti-CD31 (ab124432, Abcam) and anti-VEGFR (ab2349, Abcam).

Statistical analysis

Data were presented as mean ± standard deviation (SD). All experiments were repeated independently at least 3 times. Statistical analysis was conducted using GraphPad Prism software 7.0 (GraphPad, La Jolla, CA, USA). One-way analysis of variance (ANOVA) followed by Tukey’s post hoc test or Student’s t-test was employed for multiple-group comparison or two-group comparison, respectively. P < 0.05 was considered statistically significant.