Clinical specimens

16 tumors and matched adjacent samples of NSCLC patients were collected from Chinese PLA General Hospital. Blood samples were collected from patients with lung cancer and benign nodules who were hospitalized. All patients obtained written informed consent. Detailed patient records were recorded and kept during specimen collection. According to the clinical data, patients with other cancers, patients with previous cancer history or infectious virus infection, patients with non-small cell lung adenocarcinoma with surgery, radiotherapy, chemotherapy, or drug treatment were excluded. 188 blood samples from patients with non-small cell lung cancer and 53 blood samples from patients with benign nodules were collected, and plasma was separated, and stored in an RNase free tube for cryopreservation. This study was authorized by the Ethics Committee of Chinese PLA General Hospital and the Ethics Committee of Beijing Institute of Biotechnology.

Cell culture

Immortal human bronchial epithelial cells (BEAS-2B), and human lung cancer cell lines including A549, NCI-H460 and NCI-H1299 were purchased from the American Type Culture Collection (ATCC, USA). Lung cancer cells including NCI-H3255, HCC-1588 and Calu-3 cells, were purchased from Shanghai Cell Bank. All cells were maintained at 37 °C under 5% CO2 and at least 95% humidity. BEAS-2B, A549, NCI-H460, NCI-H1299 and HCC-1588 cells were cultured in DMEM (HyClone, USA), and NCI-H3255 and Calu-3 cells were cultured in 1640 medium (Gibco, USA) supplemented with 10% fetal bovine serum (Gibco, USA).

RNA‑seq

For circRNA sequencing, total RNA was extracted from plasma with a RNeasy® Kit (Qiagen, GER). The RNA was then treated using a Ribo-off rRNA Depletion Kit (Vazyme, China) to remove ribosomal RNA before generating the RNA-seq library. Then, an Illumina HiSeqTM2500 instrument (Illumina, USA) was used to perform library sequencing according to the manufacturer’s instructions. The FASTQ reads were aligned to the human reference genome (hg38/GRCh38). The counts of the remaining reads were normalized and mapped across an identified back-splice junction. The transcripts overlapping with the exons of the database annotations were screened out, and the transcripts overlapping with the exons of the spliced transcripts in the database were included in the follow-up analysis. The expression of each transcript was calculated through Cuffquant, and the transcripts with FPKM ≥ 0.5 were selected. The overall distribution of different circRNA was visually analyzed by volcanic map, and the different circRNA was screened from two aspects: the difference multiple and the corrected significant level.

RNA preparation and qRT‒PCR

Total RNA was extracted from cultured cells with an Ambion PureLink™ Total RNA Kit (Thermo Scientific, USA) and quantified using a NanoDrop1000 spectrophotometer (NanoDrop Technologies, USA). For analysis of circRNA and mRNA, cDNA was randomly or oligo(dT) primed from 1 µg of total RNA using a High-Capacity cDNA Reverse Transcription Kit (Invitrogen, USA) following the manufacturer’s instructions. For analysis of miRNA, 1 µg of RNA was converted to cDNA using Mir-X™ miRNA First-Strand Synthesis (Takara, Japan) based on the manufacturer’s procedure. The cDNA was amplified by qPCR using Thunderbird SYBR® qPCR Mix (Toyobo, Japan) on a CFX96 Real-Time PCR System (Bio-Rad, USA). GAPDH and U6 were used as internal references and calculated using the 2−ΔΔCt method.

RNase R treatment

Total RNA was isolated from tissues and cells using TRIzol® reagent (Invitrogen, USA) according to the manufacturer’s protocol. cDNA was synthesized with random primers or miRNA‒specific stem‒loop primers using a Revert Aid First Strand cDNA Synthesis Kit (Thermo Scientific, USA). qRT‒PCR was performed on the Bio-Rad CFX96™ Real-Time PCR System (BioRad, USA). The primers for mRNA and circRNA were synthesized by BGI (BGI, China). To verify the ring structure of circHERC1, for RNase R treatment, 2 µg total RNA was digested with 3 U RNase R (Epicenter Technologies, USA) at 37 °C for 20 min and 70 °C for 5 min according to the manufacturer’s instructions. For controls, 2 µg of total RNA was mock treated under the same conditions without the enzyme. Then, the expression levels of linear mRNA and circular RNA were determined by qRT‒PCR and RT‒PCR.

Cytoplasmic and nuclear fractions

Cytoplasmic and nuclear RNAs were purified with Cytoplasmic & Nuclear RNA Purification Kit (Norgen, Canada), converted to cDNAs and analyzed by qPCR, U6 served as the nuclear RNA marker, and GAPDH served as the cytoplasmic RNA marker. Cytoplasmic and nuclear proteins were extracted by nuclear and cytoplasmic extraction reagents according to the manufacturer’s instructions (Thermo Scientific, USA).

Fluorescence in situ hybridization (FISH)

To investigate the intracellular distribution of circHERC1 in lung cancer, FAM-labeled circHERC1 and CY3-labeled miRNA-142-3p probes were synthesized by Suzhou GenePharma Co. Ltd (GenePharma, China). Cells were grown to 60–80% confluence and then fixed with 4% paraformaldehyde. The hybridization experiments were performed using a Fluorescence In Situ Hybridization Kit (GenePharma, China) according to the manufacturer’s protocol. The cell slices were mounted and images were acquired using a Nikon ECLIPSE confocal microscope (Nikon, Japan).

Cell transfection and generation of stable cell lines

Plasmids, siRNAs, miRNA mimics or inhibitors were transfected into cells using Lipofectamine® 2000 Transfection Reagent (Thermo Scientific, USA) according to the manufacturer’s instructions. To construct cell lines stably overexpressing circHERC1, we transfected the control vector pLC5-circ or the circHERC1 overexpression vector, and single clones were selected for more than 4 weeks by incubation with 2 µg/ml puromycin (Invitrogen, USA). To generate stable circHERC1 knockdown cell lines, the control vector or sh-circHERC1 was transfected into NCI-H3255 cells, and single clones were selected for more than 4 weeks by incubation with 500 ng/ml hygromycin (Invitrogen, USA). Once fully infected, stable cells were cryopreserved or used for other experiments and assays.

5-Ethynyl-2’-deoxyuridine (EdU) assay

The cells to be detected were seeded on 4-well slides with chambers in advance and 10 µM EdU were added. The cells were incubated for 2 h and then stained with Azide Alexa Fluor 594 for 30 min and DAPI for 10 min using BeyoClick™ EdU Cell Proliferation Kit with Alexa Fluor 594 (Reyotime, China) according to the manufacturer’s protocol. The cells were visualized by fluorescence microscopy. The proliferation rate of cells was determined by calculating the ratio of EdU-positive nuclei to total nuclei.

Cell viability assay

NCI-H3255 and A549 cells were collected and inoculated into 96-well plates at a concentration of 2000 cells per well, and 10 µL CCK-8 solution (Dojindo, Japan) was added to each well at 0 h, 24 h, 48 h, 72 h, 96 h. The timing was started after the cells were attached to the well. The cells were incubated at 37 °C for 2 h. The absorbance at 450 nm was measured with a microplate reader.

Apoptosis detection

Cells were harvested by centrifugation at 1000 × g for 5 min, gently resuspended in PBS and counted. A total of 50,000-100,000 suspended cells were centrifuged at 1000 × g for 5 min, and the supernatant was discarded. Then, 195 µL binding solution, 5 µL Annexin V-FITC and 10 µL propyl iodide solution (Beyotime, China) were added to the cells and mixed gently. The mixtures were incubated at room temperature (20–25 °C) in darkness for 10–20 min and then subjected to flow cytometry analysis quickly.

Transwell migration and invasion assay

To determine the migration and invasion of cells, 5 × 104 cells were plated in medium without serum in the top chamber of a transwell with an 8.0 μm pore size (Corning, USA). Matrigel (BD Biosciences, USA) coating to perform the invasion assays. The bottom chamber contained standard medium with 20% FBS medium. After 16–24 h of culture, the cells were fixed with 90% ethanol and stained with 0.1% crystal violet (Solarbio Technology, China).

Wound healing assays

A wound healing assay was performed in 6-well plates (5 × 105 cells per well). When transfected cancer cells had grown to 90 to 95% confluence, wound lines were manually created by scratching the monolayer with a sterile 200 µl pipette tip. The migration of the cells was assessed after 24, 48, and 72 h. Pictures were taken using an inverted phase-contrast microscope. The distance between the parallel lines was measured using Image J software. All experiments were carried out at least three times. The wound healing rate (%) was calculated according to the formula: (the wound area at 0 h minus the wound area at 24 h) / the wound area at 0 h.

Western blot

Total proteins were extracted with RIPA Lysis Buffer (Thermo Scientific, USA) and quantified using a Pierce BCA Protein Assay Kit (Thermo Scientific, USA). The total protein was separated by SDS‒PAGE and transferred onto PVDF membranes (Millipore, USA). After they were blocked in 5% nonfat milk for 1 h, the membranes were incubated 1 h at room temperature with the indicated primary antibodies, namely, anti-HMGB1 (Abcam, ab228624, 1:5000 dilution), anti-GAPDH (Cell Signaling,#2118,1:2000 dilution), anti-FOXO1(Cell Signaling, #2880,1:1000 dilution), anti-phospho-FOXO1 (Cell Signaling, #84192,1:1000 dilution), anti-phospho-AKT (Cell Signaling, #4060, 1:2000 dilution), anti-phospho-IKBα (Cell Signaling, #4814,1:1000 dilution), anti-phospho-p44/42 MAPK(ERK1/2) (Cell Signaling, #4370, 1:1000 dilution), anti-p44/42 MAPK(ERK1/2) (Cell Signaling, #4695, 1:1000 dilution), anti-AKT (Cell Signaling,#4691, 1:1000 dilution), anti-IKBα (Cell Signaling, #4814, 1:1000 dilution), anti-p65 (Cell Signaling, #3033, 1:1000 dilution), anti-phospho-p65 (Cell Signaling, #8042, 1:1000 dilution), anti-CDK2 (Cell Signaling, #18,048, 1:1000 dilution), anti-Lamin B1 (Cell Signaling, #17,416, 1:1000 dilution), and anti-BIM (Cell Signaling, #2933, 1:1000 dilution). Next, they were incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies for 1 h at room temperature and visualized with the enhanced chemiluminescence (ECL) detection reagent (PerkinElmer, USA).

RNA immunoprecipitation (RIP)

RIP assays were conducted using a Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore, USA) according to the manufacturer’s instructions. A total of 1 × 107 A549 and NCI-H3255 cells were harvested and lysed in RIP lysis buffer containing protease inhibitor cocktail (Thermo, Germany) and RNA inhibitor (Beyotime, China), the supernatants were collected after centrifugation for 10 min at 13 000 rpm, and then incubated with Dynabeads protein G conjugated with anti-IgG or anti-AGO2 using a Magna RIP™ kit (Merck, USA). The immunoprecipitated RNAs were extracted for the detection of miRNA and circRNA expression by qRT‒PCR.

RNA/protein pulldown

RNA pulldown was conducted as previously described [25]. Briefly, we designed a biotin-labeled 30 nt probe against the back-spliced junction of circHERC1 to specifically pull down circHERC1 and its intracellular RNA‒RNA or RNA‒protein complex. A biotin-labeled probe with a scrambled sequence was used as a negative control. A total of 1 × 107 cells were cross-linked in ice-cold PBS buffer with 1% formaldehyde for 10 min. Upon PBS buffer removal, these cells were lysed in RNA immunoprecipitation (RIP) buffer on ice for 30 min. After sonication, the cell supernatant was harvested and divided into two equal parts for subsequent RNA pulldown after centrifugation. The biotin-labeled and control probes were incubated with the respective cell lysate for 4 h at 4 °C with gentle rotation. Identically blocked M280 Streptavidin magnetic Dynabeads (Invitrogen, USA) were added to the above lysates and further rotated for 4 h at 4 °C. After washing with RIP buffer and RIP buffer supplemented with 500 mM NaCl, the bound RNA was isolated using TRIzol® and used for RNA detection by RT-qPCR assay. The protein was extracted and analyzed by western blotting.

Dual luciferase reporter assay

The sequence of circHERC1 (or the 3’ UTR of HMGB1) was subcloned into the psiCHECK2 vector (Promega, USA) to build the corresponding wild-type (WT) vectors. In the mutant vectors, the miR-142-3p binding sites in circHERC1 or the 3’ UTR of HMGB1 were mutated in the psiCHECK2 vectors. 293T cells were transfected with these vectors and Renilla expression plasmids per well using Lipofectamine® 2000 in 24-well plates. After 24 h of transfection, the cells were lysed with passive lysis buffer (Promega, USA), and reporter gene expression was assessed using a Dual Luciferase Reporter Assay System (Promega, USA).

Hematoxylin and eosin (H&E) and immunohistochemical staining

Paraffin-embedded tissues were sectioned at a thickness of 5 μm, deparaffinized and rehydrated. For antigen retrieval, the slides were heated at 100 °C in 0.01 M citrate buffer, and 3% hydrogen peroxide was used to quench the peroxidase activity for 20 min. The sections were treated with normal goat serum and then incubated overnight with antibodies at 4 °C. After being rinsed with PBS, the sections were incubated with goat anti-rabbit IgG for 1 h. After hematoxylin counterstaining was completed, all sections were dehydrated and sealed.

Tumorigenicity assays in nude mice

Ten athymic Nu/Nu mice (Beijing Vital River Laboratory Animal Technology Co., Ltd. China) per group were subcutaneously inoculated with 1 × 106 NCI-H3255 cells transfected with control, sh-circHERC1, sh-HMGB1, sh-FOXO1, or circHERC1, or cotransfected with circHERC1 and sh-HMGB1; circHERC1 and sh-FOXO1 or circHERC1, sh-FOXO1 and sh-HMGB1. The mice were kept in a standard barrier environment; tumor growth was metered via a digital caliper every 7 days. The mice were observed for 27 days and then euthanized under deep anesthesia with pentobarbital (Sigma, USA) and the tumor was removed to measure the tumor weight and volume. The volume of the tumors was calculated by using the formula width×width×length×0.52. The animal protocol was designed to minimize pain or discomfort to the animals. All animal procedures were performed under the ethical guidelines of Beijing Institute of Biotechnology and according to the recommendations of the Beijing Experimental Animal Regulation Board. All animals were euthanized by barbiturate overdose (intravenous injection, 150 mg/kg pentobarbital sodium) for tissue collection.

In vivo metastatic model

For the metastasis model, female Nu/Nu mice (6 weeks) were maintained in pathogen-free conditions. Mice were injected through the tail vein with control and circHERC1, sh-HMGB1, or sh-FOXO1 treated NCI-H3255 luciferase cells (1 × 106) (n = 6/per group). The bioluminescence was monitored weekly.

Statistical analysis

All experiments in this study were performed independently with at least three biological replicates. The data are presented as the means ± standard deviation. Statistical analyses were performed using GraphPad Prism. Differences between two groups were analyzed by independent sample t-tests, and differences among multiple groups were analyzed by one-way ANOVA. A value of P < 0.05 indicated a statistically significant difference.