Bioinformatics

The GEO database (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE100186) was used to analyze the aberrant circRNAs from 4 pairs of RCC tissues and adjacent normal tissues. The circBank database was employed to predict the miRNAs that potentially bind to circPRELID2. Analyses for the molecular targets of miR-22-3p were carried out using several computational methods miRWalk, TargetScan, StarBase and miRcode.

Clinical samples and cells

With the protocols approved by The Third Affiliated Hospital of Chongqing Medical University Ethics Committee, 52 patients with RCC (including 29 cases TNM stage I + II and 23 cases III stage, and 30 cases negative lymph node metastasis and 22 cases positive lymph node metastasis) were enrolled from The Third Affiliated Hospital of Chongqing Medical University. Under the supervision of two pathologists, we obtained tumor tissues and matched healthy tissues from these patients. Informed consent was provided by all patients.

Human proximal renal tubular epithelial HK2 cells (ATCC, Rockville, MD, USA) were reproduced in DMEM/F-12 (Life Technologies, Lucerne, Switzerland), and human RCC cell lines A498, ACHN, 769-P and 786-O (ATCC) and RCC4 (Bnbio, Beijing, China) were maintained in 10% FBS RPMI-1640 medium (Life Technologies) plus 1% antibiotics (Life Technologies) in 5% CO2 at 37 °C.

qRT-PCR

The TRIzol (Life Technologies) was applied for RNA preparation. cDNA generation was implemented using the PrimerScript RT kit for circPRELID2 and the indicated human gene mRNAs and the TaqMan MicroRNA RT Kit for miR-22-3p. Then, SYBR Green (TaKaRa) was employed for qRT-PCR. The primers shown in Supplement Table 1 were used for PCR amplification. Fold changes were scored by the 2−ΔΔCt method after normalization by β-actin or U6.

Constructs, transfection and transduction

Using Lipofectamine 3000 (Life Technologies), RCC4 and 786-O cells were transiently transfected with sh-circPRELID2 vector (Geneseed, Guangzhou, China), OE-ETV1 expression plasmid (Life Technologies), or sh-NC control (Life Technologies). MiR-22-3p change was done using the commercially available mimic of miR-22-3p or the inhibitor of miR-22-3p (anti-miR-22-3p). Sh-circPRELID2 lentiviral particles were obtained from GenePharma, with nontarget virus particles (sh-NC) as the negative controls. The virus particles were then used to infect 786-O cells, and virus-infected cells were selected by puromycin.

Treatment of RNase R and actinomycin D

For Actinomycin D treatment, RCC4 and 786-O cells were treated with 2 mg/mL of Actinomycin D (Sigma-Aldrich, Zwijndrecht, The Netherlands) for 0, 4, 8, and 12 h. For RNase R treatment, total RNA (10 µg) was subjected to RNase R incubation (Geneseed), and then RNA was purified by the RNA Purification Kit (Invitrogen, Tokyo, Japan). In both assays, the assessment of PRELID2 mRNA and circPRELID2 levels was done by qRT-PCR.

Subcellular fractionation assay

The Cytoplasmic & Nuclear RNA Purification Kit was utilized to isolate cytoplasmic and nuclear RNA from cytoplasmic and nucleus fractions of RCC4 and 786-O cells, following the recommendations of the producer (Norgen Biotek, Thorold, ON, Canada). The levels of circPRELID2 were evaluated by qRT-PCR.

Cell proliferation, colony formation and cell cycle assays

On 50–60% confluency, RCC4 and 786-O cells were performed the transfection of sh-NC, sh-circPRELID2, sh-circPRELID2 + anti-miR-NC mimic or sh-circPRELID2 + anti-miR-22-3p mimic. Cell proliferation test was done using CCK-8 assay. As reported previously [17], we performed colony formation experiment by seeding cells into 6-well plates and culturing them for 14 days. After propidium iodide (Sigma-Aldrich) staining, we scored cell cycle distribution by using a FACSCalibur instrument (BD Biosciences, Heidelberg, Germany).

Assessment of M2-polarized macrophages by flow cytometry

RCC4 and 786-O cells were performed with the different transfection of sh-NC, sh-circPRELID2, sh-circPRELID2 + anti-miR-22-3p mimic or sh-circPRELID2 + OE-ETV1. After 48 h of transfection, the culture media were collected and used to treat THP-1 monocytic leukemia cells, which were co-treated with 100 ng/mL PMA (Sigma-Aldrich) for macrophage (THP1-M0) induction. 24 h later, single-cell suspensions were stained with FITC-labeled CD206 antibody (#321103, Biolegend, USA) and subsequently incubated with PC5.5-labeled 7AAD (Biolegend) for dead cell elimination. The CD206+ cells were scored on the FACSCalibur instrument.

Transwell assays

Cell transfection was done as described above, and the migration and invasion capacities were detected after 24 h transfection using 24-transwell chambers and invasion chambers pre-coated with Matrigel (BD Biosciences), respectively. We plated transfected cells in serum-free medium into the upper chamber. Media containing 10% FBS were placed into the under chamber. 24 h later, the migrated or invaded cells were photographed and counted after staining.

Animal studies

All animal procedures were performed following an approval of the Animal Care and Use of The Third Affiliated Hospital of Chongqing Medical University. The animal procedures complied with International guidelines. Ten athymic Balb/c female nude mice (Vital River Laboratory, Beijing, China) were used. The xenograft tumors were generated by subcutaneously injecting sh-NC or sh-circPRELID2 lentivirus-transduced H1299 cells (2 × 106 cells per mouse) into the mice. Tumor growth was monitored by determining tumor volume (0.5 × length × width2). After 28 days, all mice were euthanized, and followed by the collection of xenograft tumors. Sections of xenograft tumors were subjected to immunohistochemistry (IHC) processing as reported [18] using an antibody against Ki67 (1:500 dilution, ab16667; Abcam).

Immunoblotting

Preparation of cell lysates and immunoblotting were conducted using standard protocols [17]. Anti-ETV1 (ab136121) and anti-β-actin (ab8226, all from Abcam, Cambridge, UK) antibodies were employed. To ensure a clearer presentation of the target bands, we removed the excess of the protein gels before the membrane transfer. The original blot images were showed in Additional file 2.

RNA pull-down, luciferase reporter and RNA immunoprecipitation (RIP) assays

The circPRELID2 segment harboring the miR-22-3p-binding sequence and ETV1 3’-UTR were ligated into the pmirGLO vector (Promega, Mannheim, Germany). Via a TaKaRa MutanBEST Kit, we generated site-directed mutations of the two reporter constructs in the seed region. We introduced report constructs into 293 T cells along with mimic of miR-22-3p or miR-NC. Luciferase activities were determined after 48 h.

For RIP experiments, we treated total extractions of RCC4 and 786-O cells with protein A/G beads-linked anti-Ago2 or isotype anti-IgG antibody (Abcam). For RNA pull-down experiments, we probed total extractions with Biotinylated miR-22-3p mimic (Bio-miR-22-3p) or its mutant in the seed sites (Bio-miR-22-3p-MUT), or Bio-miR-NC control and Streptavidin beads (Life Technologies). RNA bound to beads was subjected to detection of circPRELID2 enrichment.

Statistical analysis

The P values were detected by Student’s t-test and ANOVA, and P < 0.05 meant the statistical significance. Correlations among circPRELID2, miR-22-3p and ETV1 were analyzed by Spearman rank correlation test.