Human ICC-derived QBC-939, HUCCT-1, HCCC-9810 and RBE cells were obtained from Shanghai Institute for Biological Science (Shanghai, China). All cells were grown in Roswell Park Memorial Institute-1640 (RPMI-1640) containing 10% fetal bovine serum (Gibco; Thermo Fisher Scientific, Inc.). All cell lines were cultured at 37°C in an incubator containing a humidified (100%) atmosphere with 5% CO₂. To investigate the role of ferroptosis in ICC cells, 1 µM ferroptosis inhibitor ferrostatin-1 (Fer-1, cat. no. GC10380; GlpBio) was added for 24 h at 37°C, as previously described (29).

To overexpress wild-type (WT) luciferase reporter plasmids of hsa_circ_0050900 (pmirGLO-WT, Promega Corporation), mutant (MUT) luciferase reporter plasmids of hsa_circ_0050900, (pmirGLO-MUT, Promega Corporation), hsa-miR-605-3p, or inhibit the expression of hsa_circ_0050900 and hsa-miR-605-3p, Lipofectamine 2000 Transfection Reagent (Thermo Fisher Scientific, Inc.) was used for cell transfection for 4 h. Following digestion using 0.25% trypsin for 2 min at 37°C, 1×10⁵ cells were inoculated into 24-well plates and incubated overnight with 5% CO₂ at 37°C. When the cell density was 80%, the medium was aspirated. A total of 0.5 µg (in 5 µl) plasmids, siRNAs, mimics, or inhibitors were dissolved in 250 µl OPTI-MEM (Gibco), then mixed in tube A. A total of 1.5 µl Lipofectamine 2000 was dissolved in 250 µl OPTI-MEM, mixed gently and stood for 5 min at room temperature in tube B. Tube B was added to tube A, mixed well, stood for 20 min at room temperature, then dropped into the 24-well plate well, mixed well, and cultured at 37°C in 5% CO₂. After 4 h, the transfection medium was removed and 2 ml RPMI-1640 containing 10% fetal bovine serum) was added to continue culture for 24 or 48 h at 37°C in 5% CO₂ before further investigation. Small interfering (si)-hsa_circ_0050900-1, hsa-miR-605-3p mimics, hsa-miR-605-3p inhibitor and negative control (NC) si-NC, miR-NC and NC inhibitor were synthesized by Shanghai GenePharma Co., Ltd. Sequences were as follows: si-hsa_circ_0050900-1 forward sequences, 5′-AUGCUGGAUGCAGAGGACCUU-3′; si-hsa_circ_0050900-1 reverse, 5′-AAGGUCCUCUGCAUCCAGCAU-3′; si-hsa_circ_0050900-2 forward sequences, 5′-AGAGGACCUUCACGGCAUGGU-3′; si-hsa_circ_0050900-2 reverse, 5′-ACCAUGCCGUGAAGGUCCUCU-3′; si-hsa_circ_0050900-3 forward sequences, 5′-GCUGGAUGCAGAGGACCUUCA-3′; si-hsa_circ_0050900-3 reversed sequences, 5′-UGAAGGUCCUCUGCAUCCAGC-3′; si-NC forward sequences, 5′-GAGUACCGUUUAUGGCAGGAC-3′; si-NC reversed sequences, 5′-AGAGGACCUUCACGGCAUGGU-3′; hsa-miR-605-3p mimics forward sequences, 5′-AGAAGGCACUAUGAGAUUUAGAUU-3′; hsa-miR-605-3p mimics reversed sequences, 5′-UCUAAAUCUCAUAGUGCCUUCUUU-3′; hsa-miR-605-3p inhibitor, 5′-UCUAAAUCUCAUAGUGCCUUCU-3′; miR-NC forward, 5′-AAAGGUGUAGGAAAUUCACAUGUU-3′; miR-NC reverse, 5′-CAUGUGAAUUUCCUACACCUUUUU-3′ and NC inhibitor, 5′-CAGUACUUUUGUGUAGUACAA-3′.

After cell counting, the cells were diluted to 1×10⁴ cell/ml, then 1×10³ cell/ml was inoculated on a 6-well plate (100 µl/well). After inoculation, complete medium was added (2 ml/well). After 1 week of culture at 37°C in 5% CO₂, the medium was removed and cells were rinsed with PBS or normal saline twice for 10 sec each, fixed with 4% paraformaldehyde for 15 min at room temperature, washed with PBS twice for 10 sec each and 200 µl crystal violet staining solution was added to cover the bottom of the well. After incubation 20 min at room temperature, the 6-well plate was rinsed under running water for 10 sec and dried and the number of colonies (>50 cells) was calculated manually.

Transwell plate (8.0-µm pore size; Corning, Inc.) was used to performed cell migration assay. Cells in logarithmic growth phase were rinsed once with PBS and digested by 0.25% trypsin (for 2 min at 37°C) into single cell suspension. Cell suspension was centrifuged at 550 × g for 5 min at 37°C. After supernatant was aspirated and cells were suspended with RPMI-1640, cell concentration was adjusted to 1×10⁶/ml with RPMI-1640. A total of 100 µl cell suspension (in RPMI-1640) was added to the upper chamber of Transwell chamber and 600 µl RPMI-1640 containing 10% fetal bovine serum was added to the lower chamber. Following incubation for 48 h at 37°C in 5% CO₂, cells in the upper chamber were removed and wiped with cotton swabs. Cells were fixed with 4% paraformaldehyde for 15 min at 37°C, washed with PBS once for 30 sec, stained with 1% crystal violet for 10 min at room temperature and washed with PBS once for 30 sec. Migrated cells were observed under a light microscope (200×).

University of California-Santa Cruz (UCSC, genome.ucsc.edu/; version: GRCh37/hg19) combined with circPrimer 2.0 (https://www.bio-inf.cn/) was utilized to analyze the source of hsa_circ_0050900. CircAtlas 2.0 (https://ngdc.cncb.ac.cn/circatlas/) predicted the binding sequence between hsa_circ_0050900 and hsa-miR-605-3p. TargetScanHuman 7.1 (targetscan.org/vert_71/) predicted the binding sites of hsa-miR-605-3p and SLC3A2 3′ untranslated region (UTR). For dual-luciferase reporter gene assay, 293T cells were plated in 24-well plates (5×10⁴ cells/well) and transfected with WT or MUT luciferase reporter plasmids and hsa-miR-605-3p mimics using Lipofectamine 2000 Transfection Reagent (Thermo Fisher Scientific, Inc.). Following transfected for 48 h, luciferase activity was detected using a luciferase reporter kit (cat. no. FR201-01; TransGen Biotech Co., Ltd.). Luciferase activity was calculated as activity of firefly luciferase activity compared with Renilla luciferase activity. WT and MUT plasmid containing hsa_circ_0050900 (WT1, 5′-TGGAAGGATGGTCTTGCCTTCA-3′; MUT1, 5′-TGGCGAGGGATTTTAGCATCCT-3′; WT2, 5′-ACCAACCTGAACAATGCCTTCG-3′ and MUT2, 5′-CCGTGCCAACCCACAATGATAT-3′) and SLC3A2 3′ UTR sequence (WT, 5′-AAATAGGGTGTTTTCTGCCTTCA-3′ and MUT, 5′-GGTAACATTGGCTATATCCTGTT-3′) were inserted into pmirGLO plasmid (Promega Corporation).

As previously described (30), subcellular localization of hsa_circ_0050900 was detected using a FISH kit for digoxigenin (cat. no. BIS-P0001; Guangzhou Bersin Co., Ltd.). HuCCT-1 cells were fixed by 4% paraformaldehyde for 15 min at 25°C, washed with PBS three times (5 min/time), and treated 0.1% TritonX-100 for 15 min at 25°C. After wash with PBS for twice (5 min/time), slides was incubated in hybridization buffer (Wuhan Servicebio Technology CO., LTD., Wuhan, China) for 1 h at 37°C. Then, slides were treated with digoxigenin-labeled hsa_circ_0050900 probe (cDNA, 5′-CCGTGAAGGTCCTCTGCATC-3′, 20 nt; Wuhan Servicebio technology CO., LTD.; diluted to 40 nM with hybridization buffer and denaturation at 73°C for 5 min) hybridized at 37°C for 16 h, soaked in 2X saline sodium citrate buffer for 10 min at 37°C, twice in 1X saline sodium citrate buffer for 5 min at 37°C and twice in 0.5X saline sodium citrate buffer for 5 min at 37°C. Then, slides were incubated in normal rabbit serum (1:20 in PBS; cat. no. G1209; Wuhan Servicebio technology CO., LTD.) for 30 min at 25°C and incubated anti-DIG-HRP (dilution: 1:1,000; cat. no. 200-032-156; Jackson Immuno Research Europe, Ltd.) for 50 min at 37°C, soaked in PBS thrice (5 min/time). Slides incubated with iF647-Tyramide (1:500; cat. no. G1232; Wuhan Servicebio technology CO., LTD.) for 10 min at 25°C, and washed using TBST (0.05% Tween-20) thrice (3 min/time) at 25°C and stained with DAPI for 10 min at 25°C. Slides were observed under a Zeiss LSM880 NLO confocal microscope (Leica GmbH, 1,000×) with Zen Black software 2.0 (Carl Zeiss Microscopy GmbH).

RNA was isolated from cells using TRIzol reagent (Thermo Fisher Scientific, Inc.). Total RNA underwent RT (RT kit, cat. no. K1622, Thermo Fisher Scientific, Inc.) and qPCR detection was performed using SYBR-Green (cat. no. P122; Vazyme Biotech Co., Ltd.). For RT, 1 µl total RNA (100 ng), 1 µl oligo (dT)18 for circRNA or mRNA expression or hsa-miR-605-3p RT primer (100 µM) for expression of hsa-miR-605-3p and 10 µl nuclease-free water were mixed gently, centrifuged at 250 × g for 5 sec at 25°C and incubated at 65°C for 5 min. A total of 4 µl 5X Reaction Buffer, 1 µl RiboLock RNase Inhibitor (20 U/µl), 2 µl 10 mM dNTP Mix and 1 µl RevertAid M-MuLV RT (200 U/µl) were mixed gently and centrifuged 250 × g for 5 sec at 25°C, followed by incubation for 60 min at 42°C. The reaction was terminated by heating at 70°C for 5 min. Thermocycling conditions were as follows: Initial denaturation for 30 sec at 95°C, followed by 40 cycles of 30 sec at 95°C and 1 min at 72°C. GAPDH or U6 (Gene ID 26827) was used as an internal reference and 2^−ΔΔCq was used to calculate the relative expression (31). Primers were designed by Primer Premier 5 and synthesized by Shanghai Sangon Pharmaceutical Co., Ltd. In addition, qPCR products were analyzed by Sanger sequencing to confirm the presence of hsa_circ_0050900 circularization site. The primer sequences were as follows: hsa_circ_0050900 forward, 5′-GTCTTGCCTTCAATGCCCTG-3′ and reverse, 5′-CATGCCGTGAAGGTCCTCTG-3′; hsa-miR-605-3p RT primer, 5′-GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACTCTAAA-3′; hsa-miR-605-3p forward, 5′-AGAAGGCACTATGAGATTTAGA-3′ and reverse, 5′-GTGCAGGGTCCGAGGT-3′; SLC3A2 forward, 5′-ATGGAGCTACAGCCTCCTGA-3′ and reverse, 5′-CGCGCTGAGACCCTGG-3′; GAPDH forward, 5′-GAAGGTGAAGGTCGGAGTC-3′ and reverse, 5′- GAAGATGGTGATGGGATTTC-3′ and U6 forward, 5′-CTCGCTTCGGCAGCACA-3′ and reverse, 5′-AACGCTTCACGAATTTGCGT-3′.

Total cell protein was extracted with RIPA lysis buffer (cat. no. P0013B; Beyotime Institute of Biotechnology) and quantified using BCA protein concentration assay kit (cat. no. BL521A; Biosharp Life Sciences). Protein (20 µg/lane) was separated by 10% SDS-PAGE, transferred onto PVDF membrane, blocked using 5% skimmed milk for 2 h at 25°C and incubated with primary antibodies as follows: GPX4 (cat. no. 14432-1-AP; Proteintech Group, Inc.; 1:6,000), SLC3A2 (cat. no. 66883-1-Ig; Proteintech Group, Inc.; 1:100,000) and GAPDH (cat. no. 10494-1-AP; Proteintech Group, Inc.; 1:20,000) at 4°C overnight. HRP-conjugated Affinipure Goat Anti-Mouse IgG (H+L; cat. no. SA00001-1; Proteintech Group, Inc.; 1:10,000) and Anti-Rabbit IgG (H+L; cat. no. SA00001-2; Proteintech Group, Inc.; 1:10,000) secondary antibodies were incubated at room temperature for 2 h. GAPDH was used as an internal reference and exposure was performed using an ultra-sensitive ECL chemiluminescent substrate (cat. no. BL520A; Biosharp Life Sciences). Image J 1.48v (National Institutes of Health, Bethesda, MD, USA) was used for densitometry.

Iron (cat. no. #MAK025; MilliporeSigma) and ROS assay kit (cat. no. #S0033S; Beyotime Institute of Biotechnology) were used to measure Fe²⁺ and ROS levels in HuCCT-1 cells, respectively, according to the manufacturer's instructions.

The subcellular localization of hsa_circ_0050900 was verified using a cytoplasmic and nuclear RNA purification kit (cat. no. #21000; Norgen Biotek Corp.). ICC cells (5×10⁶) were treated with 200 µl freezing cell isolation buffer for 5 min to separate the cytoplasm and nuclei. RT-qPCR was used to measure hsa_circ_0050900 expression in the cytoplasm and nucleus as aforementioned.

EZ-Magna RIP RNA Binding Protein Immunoprecipitation kit (cat. no. #17-701; EMD Millipore) was used according to the manufacturer's instructions. After cells were lysed in complete RIP lysis buffer, whole cell lysate was collected and treated with conjugated anti-AGO2 (cat. no. #67934-1-Ig; Proteintech Group, Inc.; 5 µg) or normal mouse IgG overnight at 4°C. Next, 10 µl whole cell lysate was used as input (positive control) and IgG acted as negative control. The immunoprecipitated RNA was extracted and relative enrichment of hsa_circ_0050900 and SLC3A2 in immunoprecipitated RNA was detected by RT-qPCR as aforementioned.

SPSS 16.0 software (SPSS, Inc.) was used for statistical analysis. All experiments were repeated three times. Data are expressed as the mean ± standard deviation. Unpaired t test or one-way ANOVA followed by Tukey's post hoc test was used to analyze data. P<0.05 was considered to indicate a statistically significant difference.

UCSC combined with circPrimer was used to analyze the source of hsa_circ_0050900. hsa_circ_0050900 was derived from ACTN4 gene (Fig. 1A). Sanger sequencing showed that there was a back-splice junction for hsa_circ_0050900 (Fig. 1B), which confirmed the circular structure of hsa_circ_0050900. Next, hsa_circ_0050900 expression in ICC cells was measured relative to GAPDH. hsa_circ_0050900 was elevated in ICC cells and highest in HuCCT-1 cells (Fig. 1C). Therefore, HuCCT-1 cells were selected for subsequent experiments.

To explore the function of hsa_circ_0050900, specific interference fragments were designed for hsa_circ_0050900. qPCR results showed si-hsa_circ_0050900-1, si-hsa_circ_0050900-2 and si-hsa_circ_0050900-3 interfered with expression of hsa_circ_0050900 and si-hsa_circ_0050900-3 had highest interference efficiency (Fig. 2A). Therefore, si-hsa_circ_0050900-3 was selected for subsequent experiments. To investigate whether hsa_circ_0050900 was involved in ferroptosis, HuCCT-1 cells were treated with ferroptosis inhibitor Fer-1 following transfection of si-NC or si-hsa_circ_0050900-3. Fe²⁺ and ROS levels were promoted after knocking down hsa_circ_0050900, while Fe²⁺ and ROS levels were suppressed by Fer-1 (Fig. 2B and C). In addition, ferroptosis-associated factor GPX4 expression was suppressed after hsa_circ_0050900 was knocked down; this effect on GPX4 expression was reversed by Fer-1 treatment (Fig. 2D). Cell proliferation and migration were suppressed after knocking down hsa_circ_0050900 expression, while these phenomenon were reversed by Fer-1 treatment (Fig. 2E and F). Collectively, inhibition of hsa_circ_0050900 inhibited proliferation and migration by promoting ICC cell ferroptosis.

To explore the potential mechanism of hsa_circ_0050900, the present study confirmed the localization of hsa_circ_0050900. FISH showed that hsa_circ_0050900 was mainly localized in cytoplasm in HuCCT-1 cells (Fig. 3A). RT-qPCR verified hsa_circ_0050900 was primarily located in cytoplasm (Fig. 3B). Then, the expression was detected after knocking down expression of hsa_circ_0050900. qPCR indicated that hsa-miR-605-3p expression was elevated after knockdown of hsa_circ_0050900 (Fig. 3C). CircAtlas 2.0 software was used to predict the binding sequence between hsa_circ_0050900 and hsa-miR-605-3p. There were two binding sites between hsa_circ_0050900 and hsa-miR-605-3p (Fig. 3D). Dual-luciferase reporter gene assay demonstrated that only hsa-miR-605-3p mimics decreased the luciferase activity of hsa_circ_0050900 WT but not hsa_circ_0050900 MUT (Fig. 3E). AGO2-RIP assays revealed hsa_circ_0050900 expression was promoted by hsa-miR-605-3p mimics (Fig. 3F). These results indicated that hsa_circ_0050900 regulated hsa-miR-605-3p expression (Fig. 3E and F). hsa_circ_0050900 negatively regulated hsa-miR-605-3p expression.

To study hsa-miR-605-3p function, hsa-miR-605-3p mimics or miR-NC was transfected into HuCCT-1 cells. Compared with miR-NC, hsa-miR-605-3p expression was significantly elevated after transfection with hsa-miR-605-3p mimics, which indicated successful overexpression of hsa-miR-605-3p (Fig. 4A). Overexpression of hsa-miR-605-3p inhibited the mRNA and protein expression of SLC3A2 (Fig. 4A and B). Then, TargetScanHuman 7.1 software was used to predict the binding site between SLC3A2 3′UTR and hsa-miR-605-3p (Fig. 4C). Dual-luciferase reporter gene assay demonstrated hsa-miR-605-3p mimics decreased the luciferase activity of SLC3A2 3′UTR WT, but not SLC3A2 3′UTR MUT (Fig. 4D). Additionally, AGO2-RIP assay showed hsa-miR-605-3p mimics increased expression of SLC3A2, but not miR-NC (Fig. 4E). These results verified that hsa-miR-605-3p could regulate SLC3A2 expression.

Fe²⁺ and ROS levels were increased by hsa-miR-605-3p mimics compared with miR-NC (Fig. 5A and B). Moreover, following overexpression of hsa-miR-605-3p, GPX4 protein expression was suppressed in comparison with miR-NC group (Fig. 5C). Cell function experiments demonstrated overexpression of hsa-miR-605-3p reduced proliferation and migration in comparison with miR-NC (Fig. 5D and E). Overexpression of hsa-miR-605-3p targeted SLC3A2 to promote ICC cell ferroptosis and inhibit proliferation and migration.

hsa-miR-605-3p expression was knocked down in HuCCT-1 cells using hsa-miR-605-3p inhibitor. hsa-miR-605-3p inhibitor decreased the expression of hsa-miR-605-3p relative to NC inhibitor, indicating that hsa-miR-605-3p inhibitor had an interfering effect (Fig. 6A). mRNA and protein levels of the downstream target gene SLC3A2 of hsa-miR-605-3p were suppressed after knocking down hsa_circ_0050900 (Fig. 6B and C). However, knocking down hsa-miR-605-3p attenuated the decrease in SLC3A2 mRNA and protein expression induced by inhibition of hsa_circ_0050900 (Fig. 6B and C). These results indicated that hsa_circ_0050900 regulated SLC3A2 expression via hsa-miR-605-3p.

Finally, the effects of hsa_circ_0050900 and hsa-miR-605-3p on cell ferroptosis, proliferation and migration were assessed. Knocking down hsa-miR-605-3p reduced the increases of Fe²⁺ and ROS induced by inhibition of hsa_circ_0050900 (Fig. 7A and B). hsa-miR-605-3p inhibitor decreased the downregulation of GPX4 caused by knocking down hsa_circ_0050900 (Fig. 7C). hsa-miR-605-3p inhibitor attenuated the decrease in cell proliferation and migration caused by knocking down hsa_circ_0050900 (Fig. 7D and E). Taken together, hsa_circ_0050900 regulated SLC3A2 expression via sponging hsa-miR-605-3p to affect ICC cell ferroptosis, proliferation and migration.