The present study was approved by the Ethics Committee of Sun Yat-Sen University (Guangzhou, China), and written informed consent was obtained from all patients in Sun Yat-Sen Memorial Hospital of Sun Yat-Sen University.

A total of eight pairs of HCC tumor tissues and corresponding non-tumor tissues were collected from the Department of Hepatopancreatobiliary Surgery, Sun Yat-Sen Memorial Hospital of Sun Yat-Sen University. The present study included 8 patients with HCC, 5 males and 3 females, aged 35–66 years with a mean age of 55.9±9.5 years. Exclusion criteria were the following: i) Patients with moderate or tense ascites, ii) patients refusing liver biopsy, iii) patients with previous treatment for HCC and iv) patients with a history of severe trauma. None of the recruited participants accepted adjunctive treatment prior to the surgery. All tissue samples were immediately preserved at −80°C following washing with sterile phosphate-buffered saline (PBS). HCC cell lines (LM3, Huh7, MHCC 97L and SK-Hep-1) were purchased from Shanghai Cell Center (Shanghai, China) and cultured in Dulbecco's modified Eagle's medium (DMEM; Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA) supplemented with 10% fetal calf serum (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) and 1% penicillin-streptomycin G (Invitrogen; Thermo Fisher Scientific, Inc.). Cells were incubated at 37°C in a humidified atmosphere containing 5% CO₂.

The quantitation of CD90 expression in the cells (Huh7, MHCC 97L and SK-Hep-1) was determined using flow cytometry. A total of 2×10⁵ cells were transferred into a centrifuge tube and suspended in PBS buffer. Following centrifugation at 1,000 × g for 5 min at 4°C, the cells were fixed with fresh fixative (4% paraformaldehyde; Thermo Fisher Scientific Inc.) at room temperature for 15 min, and then preincubated in 200 ml blocking solution (10% goat serum containing 0.05% Tween-20; Thermo Fisher Scientific Inc.) for 1 h at room temperature. The cells were incubated with fluorescein isothiocyanate (FITC)-conjugated human anti-CD90 (catalog no. 85-11-0909-41; 1:100 dilution; eBiosciences; Thermo Fisher Scientific Inc.) for 1 h at 4°C, then the cells were washed in PBS and incubated with FITC-conjugated goat anti-mouse IgG secondary antibody (catalog no. ab6785; 1:1,000 dilution; Abcam, Cambridge, MA, USA). The standard was established by isotype match control (FITC-conjugated mouse IgG1 isotype control, catalog no. 85-11-4714-81; 1:100 dilution; eBiosciences; Thermo Fisher Scientific Inc.) for 1 h at 4°C. A total of 1×10⁶ FITC-labeled cells were measured using BD AccuriC6 Flow Cytometer with Cell Quest software version 5.1 (BD Biosciences, Franklin Lakes, NJ, USA). Each experiment was performed in triplicate. The cell cycle of SK-Hep-1 CD90⁺ and CD90⁻ cells was additionally determined with flow cytometry. Briefly, cells were resuspended in PBS twice prior to fixation by dropwise addition into 95% pre-cooled ethanol for 15 min at room temperature, and then centrifuged at 3,000 × g for 10 min at 4°C, resuspended twice in PBS, and stained with propidium iodide (PI) containing 50 µg/ml RNaseA (Takara Bio, Inc., Otsu, Japan) in the dark for 30 min at room temperature. The DNA content was analyzed using BD AccuriC6 Flow Cytometer with Cell Quest software, version 5.1 (BD Biosciences).

Separation of CD90⁺ cells from SK-Hep-1 cells was performed using MACS magnetic cell sorting (Miltenyi Biotech GmbH, Bergisch Gladbach, Germany) according to the manufacturer's protocol. SK-Hep-1 cells were incubated with 50 µl of FcR blocking agent (Miltenyi Biotech GmbH) and CD90-biotin immunomagnetic beads (Miltenyi Biotech GmbH) for 10 min at 4°C. Goat anti-mouse IgG magnetic beads (Miltenyi Biotech GmbH) were used as the control. A MACS cell separation column (MS column) was used to retain the CD90⁺ cells linked to the beads. Unlabeled cells (SK-Hep-1 CD90⁻ cells) were washed with 3×500 µl buffer (containing PBS, 0.5% BSA, 2 mM EDTA) and collected. The MS column was then removed from the separator and placed on another collection tube. A total of 1 ml buffer was pipetted onto the MS column, and the labeled CD90⁺ cells (SK-Hep-1 CD90⁺ cells) were flushed out by applying the plunger supplied with the column. The labeled CD90⁺ cells were eventually obtained via centrifugation at 3,000 × g for 10 min at 4°C and resuspension in serum-free medium (Gibco; Thermo Fisher Scientific, Inc.). The SK-Hep-1 CD90⁺ and SK-Hep-1 CD90⁻ cells were harvested and washed with PBS, and the expression of CD90 in these cells was directly investigated with flow cytometry. Following sorting, the SK-Hep-1 CD90⁺ cells indicated an ≥85% expression level of CD90 and the SK-Hep-1 CD90⁻ cells revealed <5% expression level of CD90.

SK-Hep-1 cells were lysed in lysis buffer (Gibco; Thermo Fisher Scientific, Inc.) and the concentration of total protein was analyzed using a BCA Protein Assay Kit (Sangon Biotech Co., Ltd, Shanghai, China). A total of 30 µg protein was added into each lane of the 10% SDS-PAGE gel and 5% skim milk (BD Biosciences) was used to block the polyvinylidene membranes for 10 min at room temperature. The membranes were then incubated with the primary antibody anti-CD90/Thy1 (catalog no. ab133350; 1:100; Abcam) at 4°C overnight. GAPDH was used as the internal control (anti-GAPDH antibody; catalog no. ab9485; 1:100, Abcam). The next day, the membranes were incubated with horseradish peroxidase-conjugated anti-rabbit IgG secondary antibody for 1 h at room temperature (1:500 dilution; catalog no. ab7090; Abcam). The results were visualized using the enhanced chemiluminescence substrate kit (GE Healthcare, Chicago, IL, USA). All images were analyzed with ImageJ software, version 14.8 (National Institutes of Health, Bethesda, MD, USA).

Cell viability was assessed with a 3-(4,5- dim ethylthiazol-2-yl)-2,5-diphenyltrtrazolium bromide (MTT) assay. The treated cells were seeded into a 96-well plate and cultured in complete medium. When the cells adhered, 0.5 mg/ml MTT was added to each well and incubated at 37°C for 4 h. The supernatant was then carefully aspirated and 100 µl of dimethyl sulfoxide was added. The absorbance was measured at a wavelength of 490 nm using a microplate reader (Thermo Fisher Scientific, Inc.).

Cell migration and invasion assays were performed using Transwell chambers with 8 µm pores (BD Biosciences) according to the manufacturer's protocol. For migration assays, SK-Hep-1 CD90⁺ (5×10⁴ cells/well) and SK-Hep-1 CD90⁻ cells (5×10⁴ cells/well) were separately seeded into the upper compartment without Matrigel, for invasion assay, the cells were seeded into the upper compartment with Matrigel (Corning Inc., Corning, NY, USA) and incubated in serum-free DMEM (Gibco; Thermo Fisher Scientific, Inc.) and the lower compartment was filled with complete medium supplemented with 10% FBS. Following 24 h, migratory and invasive cells on the bottom surface of the filters were fixed with 4% paraformaldehyde for 24 h and stained with 0.1% crystal violet solution for 10 min at room temperature. The cells in at least 10 randomly selected microscopic fields were counted under a light microscope, at ×200 magnification (Nikon Corp., Tokyo, Japan). Each experiment was performed in triplicate.

SK-Hep-1 CD90⁺ and SK-Hep-1 CD90⁻ cells were trypsinized and cultured in DMEM-F12 (Sigma-Aldrich; Merck KGaA) that was supplemented with 20 ng/ml EGF, 10 ng/ml FGF, 4 ng/ml insulin, and B27 (1:50) in Ultra-Low Attachment 6-well plates (Corning Inc.). The cells were imaged under a light microscope (magnification, ×200; Olympus Corp., Tokyo, Japan) following 20 days of incubation.

Total RNA was treated with Rnase R (Epicentre Illumina, Inc., San Diego, CA, USA) to remove linear RNA, and was then amplified and transcribed into fluorescent cRNA using Quick Amp Labeling kit (Takara Bio, Inc.) according to the manufacturer's protocol. The cRNA was then purified with the RNeasy Mini Kit (Qiagen GmbH, Hilden, Germany) and hybridized to the circRNA expression microarray according to the manufacturer's protocols. The microarray hybridization and the data collection were performed by Forevergen Bio-tech Co. (Guangzhou, China). For the microarray analysis, following filtering by flag signal, the raw data were processed and normalized. Differentially expressed circRNA were identified with P<0.05 and fold changes >2. In addition, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses was performed which provided more detailed information regarding the pathways in which the target genes were involved.

To verify the specialty of the PCR products of hsa_circ_0067531 and hsa_circ_0057096, the cDNA and gDNA PCR products that had been amplified by diverted primers were separated on a 2% agarose gel with TE running buffer. Only a single band in the gel was thought to be a specialty. The region was incised from the whole gel and purified with SanPrep Column DNA Gel Extraction Kit (Qiagen GmbH), and the PCR fragments were inserted into T vector for Sanger sequencing (Tsingke Biotech Co., Ltd., Beijing, China).

Total RNA was isolated using TRIzol^® reagent (Invitrogen; Thermo Fisher Scientific, Inc.) following indicated treatments. qPCR assays were performed with SYBR Premix Ex Taq (Takara Bio, Inc.), primers, and a cDNA template on the Applied Biosystems 7500 Real-time PCR System (Applied Biosystems; Thermo Fisher Scientific, Inc.). The primers used in the assay are listed as follows: Forward, CGA ACC AAC TTC ACC AGC AA and reverse, CTG ATG CCC TCA CAC TTG AC for CD90; forward, 5′-TCC AGA CCA GTA CGT TCG AG-3′ and reverse, 5′-TCA CAC AGT TGA GAC AAG GGA T-3′ for cir_0067531; forward, 5′-AGG CAA AGG AAG TCC ATC TCA T-3′ and reverse, 5′-TCA ATC ACA CCC TGG GCC AT-3′ for circ_0057096 and forward, 5′-CCG AGA ATG GGA AGC TTG TC-3′ and reverse, 5′-AAG CAC CAA CAG AGG AGA A-3′ for GAPDH. PCR was conducted in a 50 µl reaction system, including 1 µl forward and 1 µl reverse primers, 0.5 µl 20× SYBR-Green I, 25 µl 2×PCR buffer, 2 µl cDNA, and 20.5 µl DEPC. The qPCR program was 94°C for 4 min, 30 cycles of 94°C for 20 sec, 60°C for 30 sec, 72°C for 30 sec, and 72°C for 5 min. Each individual sample was run in triplicate and the expression level was quantified with the comparative cycle threshold (Cq) method. Results were normalized to GAPDH expression and RNA enrichments were calculated using the 2^−ΔΔCq method (24).

Quantile normalization and subsequent data processing were performed using the GeneSpring GX software, version 11.5.1 package (Agilent Technologies, Inc., Santa Clara, CA, USA). Other data were analyzed with the two-tail Student's t-test and one-way analysis of variance followed by Tukey's test. SPSS software, version 19.0 (IBM SPSS, Armonk, NY, USA). Each experiment was repeated at least three times. All results were summarized and are presented as the mean ± standard deviation. P<0.05 was considered to indicate a statistically significant difference.

The expression levels of CD90 in the 4 different HCC cell lines were determined by qPCR; results of which demonstrated that the expression of CD90 was increased in the LM3, Huh7, and MHCC 97L cell lines, and particularly in the SK-Hep-1 cell line (Fig. 1A). The western blotting results demonstrated that the protein level of CD90 in SK-Hep-1 CD90⁺ cells was significantly increased compared with SK-Hep-1 and SK-Hep-1 CD90⁻ cells. However, CD90 expression in SK-Hep-1 CD90⁻ cells was significantly decreased compared with SK-Hep-1 cells (Fig. 1B). In addition, the expression of CD90 (blue line) and control IgG (red line) were verified in the Huh7, MHCC 97L, and SK-Hep-1 cell lines by flow cytometry. It was demonstrated that the relative expression level of CD90 in SK-Hep-1 cells (10.0%) was significantly increased compared with Huh7 (0.122%) and MHCC 97L (0.045%) cell lines (Fig. 1C). Consequently, the SK-Hep-1 cell line was selected for the ensuing experiments. The CD90⁺ cells were sorted from the SK-Hep-1 cell line with an expression of CD90 (88.1%) ≥85.0%, whereas CD90⁻ cells were sorted from the SK-Hep-1 cell line with an expression of CD90 (0.997%) ≤5.0% (Fig. 1C).

Flow cytometry analysis and MTT assay were performed to investigate the effect of CD90 on the cell cycle and viability in SK-Hep-1 cells. The raw FACScan data is presented in Fig. 2A. The numerical conversion of the cell population in G₁, S, and G₂ phases is presented in Fig. 2B. The results indicated that the percentage of CD90⁺ cells was significantly decreased in the G₁ phase, compared with CD90⁻ cells. Conversely, the percentage of CD90⁺ cells was significantly increased in the G₂ phase, compared with CD90⁻ cells. These results suggested that the G₁ phase of CD90⁺ cells was shorter compared with CD90⁻ cells. In addition, the MTT assay results demonstrated that the cell viability of CD90⁺ cells was significantly increased compared with CD90⁻ cells at day 3 (Fig. 2C).

Transwell assays were used to determine the migration and invasion abilities of CD90⁺ and CD90⁻ cells. The migration assay revealed that there were significantly more CD90⁺ cells compared with CD90⁻ cells (Fig. 3A). Similarly, the invasion assay demonstrated the same results as the migration assay (Fig. 3B). Consequently, it was demonstrated that CD90⁺ cells exhibited increased migration and invasion capabilities compared with CD90⁻ cells.

The sphere formation assay demonstrated that CD90⁻ cells did not form spheres following 20 days of incubation, however the CD90⁺ cells did (Fig. 4).

A high throughput microarray assay was used to identify the differing expressions of circRNAs in CD90⁺ and CD90⁻ cells. In the present study, 274 dysregulated circRNAs were identified, including 238 upregulated and 36 downregulated circRNAs (Fig. 5A). The upregulated circRNAs were more frequently occurring compared with downregulated circRNAs. In order to further understand the biological functions of the differentially expressed circRNAs, KEGG pathway analyses was performed which provided more detailed information regarding the pathways in which the target genes were involved. It was revealed that the key signaling pathways were the metabolic pathway, pathways in cancer, and the phosphoinositide 3-kinase (P13K)-RAC-α serine/threonine-protein kinase (AKT) signaling pathway (Fig. 5B).

To verify the microarray data, two circRNAs with differential expression were selected from the P13K-AKT signaling pathway; hsa_circ_0067531 and hsa_ circ_0057096. The sequences of these two circRNAs from pyruvate dehydrogenase kinase (PDK)1 and phosphatidylino-sitol-4,5-bisphosphate 3-kinase catalytic subunit β (PIK3CB) are listed in Table I. Divergent primers amplify circRNAs in cDNA, however not genomic DNA (gDNA) (Fig. 6A). The Sanger sequencing verified head-to-tail splicing (Fig. 6B). The present study validated the expression levels via qPCR in eight pairs of HCC and adjacent normal tissues. It was demonstrated that the expression of hsa_circ_0067531 was significantly decreased in the HCC tissues, compared with the normal tissues (P=0.012) (Fig. 6C), however there was no significant difference in hsa_circ_0057096 expression between adjacent normal tissues and HCC tissues (Fig. 6D). Therefore, it was hypothesized that hsa_circ_0067531 may be a promising target for HCC diagnosis and therapy in the future.