Patients and samples

A total of 158 melanoma cases, along with corresponding normal tissues, and 9 pigmented nevus tissues were randomly sourced from the Department of Plastic and Reconstructive Surgery, Zhongshan Hospital, Fudan University (Shanghai, China). All specimens were securely preserved at − 80℃. Each patient underwent comprehensive total resection, followed by a meticulous histological and pathological examination conducted by two pathologists. None of the patients received any form of radiotherapy or chemotherapy before undergoing surgery, and all participants were provided with detailed clinical, pathological, and follow-up data. This study received approval from the Ethics Committee of Zhongshan Hospital, Fudan University, and each patient willingly provided written informed consent.

RNA-seq

For circRNAs sequencing, total RNA was extracted from six pairs of frozen melanoma tissues using TRIzol Reagent (Life, CA, USA). Subsequently, the RNA underwent treatment with a Ribo-off rRNA Depletion Kit (Vazyme, China) to eliminate ribosomal RNA prior to the generation of the RNA-seq library. Following this, an Illumina NovaseqTM6000 instrument (Illumina, USA) was utilized for library sequencing. The FASTQ reads were aligned to the human reference genome (hg38/GRCh38). Subsequently, the counts of the remaining reads were normalized and mapped across an identified back-splice junction.

For mRNA sequencing, mRNA with poly(A) was purifed, fragmented, and then reverse-transcribed to generate the RNA-seq library. Finally, sequencing was conducted using an Illumina NovaseqTM6000 system (Illumina, USA). The resulting reads were mapped to the genome.

Cell culture and transfection

The melanoma cell lines A375, A2058, MV3, Sk-mel-1, Sk-mel-28, and the normal epidermal melanocyte cell line PIG1 were procured from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). These cells were cultured at 37 °C with 5% CO2 and were confirmed to be free of mycoplasma contamination. The human circFCHO2 cDNA was synthesized and inserted into the pcDNA3.1( +) Laccase2 MCS Exon Vector by Hanbio Co., Ltd. (Shanghai, China). An empty vector was employed as the negative control. Two small interfering RNA (siRNA) sequences were synthesized, and a scramble siRNA was synthesized as a negative control. The transfected cells underwent screening with puromycin (2 μg/ml) for 1 week to establish stable cell lines. Transient transfection was carried out using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) following the manufacturer’s instructions. Total proteins and RNA were collected 48 h post-transfection. Plasmids and siRNAs of DND1 were synthesized by Genomeditech Co., Ltd. (Shanghai, China) and tranfected as described above. The specific target sequences utilized in this study are detailed in Table S1.

Quantitative real‑time PCR and western blot assays

QRT-PCR and western blot analyses were conducted following the procedures outlined in our previous study. Detailed protocols can be found in the Supplementary Materials and Methods (Wei et al. 2019a). The primers and antibodies used in the study are provided in Table S2 and S3, respectively.

Immunohistochemistry and fluorescence in situ hybridization assays

Immunohistochemistry (IHC) assay was conducted following the procedures outlined in our previous study. The detailed protocol can be found in the Supplementary Materials and Methods (Wei et al. 2019b). The antibodies employed in this study are listed in Table S3. For the fluorescence in situ hybridization (FISH) assay, probes with 3′-Cy3 modifcation targeting circFCHO2 and ICC-labeled probes specific to DND1 were synthesized by Geneseed Biotech Co., Ltd. (Guangzhou, China), respectively. The FISH probe sequence employed in this study is provided in Table S4. Cells were initially fixed with 4% paraformaldehyde and subsequently incubated with probes over night at 37℃. Nuclei were counterstained with DAPI and captured using confocal microscopy (Carl Zeiss, Germany).

RNA-pull down assay

A biotin-labeled circFCHO2 probe (RiboBio, China) was incubated with streptavidin magnetic beads (RioBio, China) at room temperature for 2 h to generate probe-coated beads. Lysates from A375 and MV3 cells were incubated with probe-coated beads at 4℃ overnight. Subsequently, the beads were washed, and the proteins pulled down were analyzed using silver staining, mass spectrometry, and western blotting. The RNA Pulldown probe sequence utilized in this study is provided in Table S5.

RNA immunoprecipitation assay

RNA immunoprecipitation (RIP) assay was conducted using a Magna RIP RNABinding Protein Immunoprecipitation Kit (Millipore, Billerica, MA, USA), following the procedures outlined in our previous study (Zhang et al. 2019). Biotin-labeled circFCHO2 was synthesized by Geneseed Biotech (Guangzhou, China). In brief, circFCHO2-overexpressing cells were washed with PBS, fixed in 1% formaldehyde, lysed in Co-IP buffer, subjected to sonication, and finally centrifuged. A portion of the supernatant (50μL) was reserved as input, while the remaining supernatant was incubated with a mixture of probes-streptavidin-dynabeads (M-280; Invitrogen, Carlsbad, CA, USA) at 30℃ for 12 h. The mixture underwent further incubation with lysis buffer and proteinase K. Subsequently, RNA was extracted using TRIzol Reagent (Invitrogen, Carlsbad, CA, USA), eluted, reverse transcribed to cDNA, and subsequently detected by qRT-PCR.

RNase R treatment and actinomycin D assays

RNase R (Lucigen, USA) was employed for RNA digestion. Briefy, RNAs extracted from A375 and MV3 cells were divided into two fractions. For RNase R digestion, 1 µg of RNA was treated with 2U of RNase R. As a control, 1 µg of RNA was mixed with an equal volume of RNase-free water. Subsequently, the expression levels of circFCHO2 and FCHO2 mRNA were assessed through qRT-PCR. CircFCHO2 stability assays involved seeding melanoma cells into 12-well plates, followed by treatment with medium containing actinomycin D (5 μg/mL) (Genview, Beijing, China) or DMSO for 0, 3, 6, and 9 h.

Nuclear‑cytoplasmic fractionation assays

RNA from nuclear and cytoplasmic fractions of melanoma cells was separated using a commercial Paris kit (Life Technologies, Gaithersburg, MD, USA) following the manufacturer’s instructions. CircFCHO2, FCHO2, and GAPDH cDNA were synthesized using a PrimeScript RT Master Mix (Takara) while snoU6 cDNA was generated through stem-loop methods (RiboBio, Guangzhou, China). The expression levels of circFCHO2, FCHO2, GAPDH, and snoU6 in the nuclear and cytoplasmic fractions were determined by qRT-PCR.

Colony formation, CCK-8, transwell invasion, and wound healing migration assays

Colony formation, CCK-8, transwell invasion, and wound healing migration assays were conducted following the methodologies outlined in our previous studies. Detailed protocols can be found in the Supplementary Materials and Methods (Wei et al. 2018; Zhu et al. 2019).

FCM assays of the cell cycle and apoptosis

The transfected cells underwent staining with propidium iodide (KGA512, KeyGen Biotech, Nanjing, China), and the cell cycle was evaluated using CytoFLEX FCM (Beckman Coulter, Brea, CA, USA). The percentages of cells in G0/G1, S, and G2 phases were determined and compared. For apoptosis detection, cells were stained with an Annexin V-FITC Apoptosis Detection Kit (Keygen Biotech), and apoptosis was assessed by CytoFLEX FCM. The ratio of early to late apoptotic cells was calculated.

In vivo assays

All animal experiments received approval from the Animal Experimentation Ethics Committee of Zhongshan Hospital, Fudan University. Four-week-old male nude mice (BALB/c) were housed and cared for in accordance with the principles of the 3Rs (replacement, reduction, and refinement). A total of 5 × 106 cells, resuspended in 100μL of PBS, were subcutaneously inoculated into the left flank of each mouse. Following tumor detection, measurements of tumor size were taken every 3 days using a vernier caliper, and tumor volume was calculated using the formula volume (cm3) = L × W2 × 0.5, where L and W denote the largest and smallest diameters, respectively. Euthanasia was performed when the tumor volumes reached a maximum of 2000mm3.

Statistical analysis and the resources of public data

Gene expression and survival data of melanoma cohorts were obtained from The Cancer Genome Atlas (TCGA)-SKCM (data format: TPM. Workflow: STAR) DataSet (https://www.cancer.gov/ccg/research/genome-sequencing/tcga). The secondary structure of circFCHO2 was obtained in the CSCD (Cancer Specific CircRNA Database, http://gb.whu.edu.cn/CSCD/) database. The experimental data were processed and subjected to statistical analysis using SPSS software (version 23.0) and GraphPad Prism software (version 8.0). Continuous measurement data are expressed as mean ± standard deviation (mean ± SD), while categorical measurement data are presented as quantity and percentage. Two sets of continuous econometric data were subjected to Student’s t-test or paired t-test. The Mann–Whitney U test was employed for the analysis of several types of measurement data. Survival analysis was conducted using the Kaplan–Meier method, generating survival curves for each group, with differences in survival between groups assessed by the log-rank test. Furthermore, Cox proportional risk regression models were utilized for univariate and multivariate correlation analysis of independent prognostic factors. In this study, three independent replicates were performed for different experiments and the experimental measurement data were uniformly standardized. In all statistical analyses presented in this experiment, a difference was deemed statistically significant when *P < 0.05, **P < 0.01, or ***P < 0.001.