RETRACTED ARTICLE: Hsa_circ_0005379 regulates malignant behavior of oral squamous cell carcinoma through the EGFR pathway

All patient tissue samples were obtained from the Stomatological Center of Peking University Shenzhen Hospital (Shenzhen, China), between 2016 and 2018. Patients enrolled in the study was selected to rule out systemic disease and preoperative radiotherapy or chemotherapy. The histopathological grading of tumors was performed according to the 2018 World Health Organization classification criteria for head and neck cancer. The circular RNA profiling was obtained through high-throughput sequencing of four pairs of specimens. (Guangzhou Gene Denovo Biotechnology Co. Ltd., Guangzhou, China). This study was approved by the Ethics Committee of Peking University Health Science Center (IRB00001053–08043).

Human OSCC cell lines SCC9, SCC15, SCC25, and CAL27 are gifts given by Wuhan University (Wuhan, China). Human oral keratinocyte (HOK) cells were obtained from the cell bank of the Chinese Academy of Sciences (Shanghai, China). All cells were cultured in Dulbecco’s modified Eagle medium (DMEM; Gibco, New York, USA). Human umbilical vein endothelial cells (HUVEC) and culture medium were bought from Procell Life Science and Technology Company (WuHan, China). All cell lines were cultured at 37 °C incubator with 5% CO₂. The siRNAs for hsa_circ_0005379 were synthesized by RiboBio Co. Ltd. (Guangzhou, China). The siRNA sequences are as follows:

siRNA-1: 5′-CAAGGAAUGUAUCCUGUCA-3′;

siRNA-2: 5′-AAGGAUUUGCAAGGAAUGUAU-3′;

siRNA-3: 5′-GAUUUGCAAGGAAUGUAUCCU-3′.

Lipofectamine 3000 (Gibco, New York, USA) was used for siRNA transfection. All three siRNAs gave identical results.

The lentiviral plasmid containing GFP and puromycin-resistant gene was constructed by HanBio Co. Ltd. (Shanghai, China). Polybrene/medium mixture with packaged lentivirus was incubated with SCC25 and CAL27 cells for 48 h. After infection, the transduced cells were screened by puromycin (SCC25, 6 μg/ml; CAL27, 10 μg/ml; Qcbio Science & Technologies Co. Ltd. Shanghai, China). The screened cells were diluted and plated into 96-well plates to get a single cell per well. Monoclonal cells were verified by qRT-PCR. Fluorescence microscopy was used to observe the expression of GFP in infected cells.

Total RNA was extracted with an RNeasy Mini Kit (QIAGEN, Hilden, Germany). RNA was incubated with 3 U/mg RNase R (Epicenter) for 15 min at 37 °C. For reverse transcription, 500 ng of RNase R-treated RNA was reverse transcribed using Prime Script RT Master Mix (Takara Bio Inc., Kusatsu, Japan). PCR reactions were performed using PCR Master Mix (2×) (Thermo Fisher Scientific, Waltham, MA, USA). Primers used for qRT-PCR are listed as follows:

hsa_circ_0005379-F1: GCCCATACCTTTATCCACTC

hsa_circ_0005379-R1: GTCAACATTCCAGTCTCTTCCT

hsa_circ_0005379-F2: CCTAAGAAGACCACAATGCG

hsa_circ_0005379-R2:CCTCCGTAGTAAGGGTTTCG

β-actin-F: AAACTGGAACGGTGAAGGTG

β-actin-R: AGTGGGGTGGCTTTTAGGAT.

Tumor cells were seeded at 4 × 10³ cells/well into 96-well plates and grown to logarithmic phase. Then EdU dye, PBS buffer, Apollo and Hoechst 33342 (Beyotime Biotechnology, Shanghai, China) were used to stain the cells, respectively. Images were taken randomly under a fluorescence microscopy. The cells were counted using Image J software.

The two types of OSCC cells conventionally stably transduced (tumor cells with high expression of hsa_circ_0005379) were cultured in 6-well plates in serum-free medium for 24 h. Supernatant medium was collected as conditioned medium. After trypsin digestion, HUVEC cells were centrifuged and the cells were resuspended in serum-free DMEM medium and counted. Then, 3 × 10 ⁴ cells were added to each well and the medium was adjusted to 100 μl using serum-free medium. In the lower chamber, two kinds of conditioned medium or two kinds of stably transduced cells cultured with serum-free DMEM medium were added. After 24 h, the medium was discarded and the cells were fixed in 4% paraformaldehyde and stained with crystal violet staining solution. The cells above the membrane were observed and photographed on an inverted microscopy.

Matrigel was added to 24-well plates at a concentration of 200 μl/well and placed in an incubator for 1 h to solidify. The HUVECs were inoculated into the above 24-well plate, and the confluency is about 60% when the cells attached. The cells were cultured for 12 h by adding the above mentioned conditioned medium or OSCC cells overexpressing hsa_circ_0005379. The photographs were taken on an inverted microscopy.

Annexin V-FITC Apoptosis Assay Kit (Beyotime Biotechnology Co. Ltd., Shanghai, China) was used to measure the cell apoptosis rate. Tumor cells were seeded into 6-well plates and grown to logarithmic phase. The cells were harvest by digestion and resuspended in 100 μl of 1 × annexin-binding buffer. 5 μl of annexin V and 1 μl of propidium iodide (PI) reagent were added to each well and incubated for 15 min. After incubation, 400 μl of 1 × annexin-binding buffer was added to stop staining. The apoptosis rate was finally measured using FACSCalibur flow cytometer (BD Biosciences, New York, USA).

Tumor cells were seeded at 4 × 10⁵ cells/well into 24-well plates and grown to logarithmic phase. Cells were then fixed using 4% paraformaldehyde and stained by Hoechst 33258 (Beyotime Biotechnology Co. Ltd., Shanghai, China). The morphology and staining of tumor cells nuclei were observed under an inverted fluorescence microscopy.

Cells were seeded in 6-well plates and cultured until confluent. Use a sterile 200 μl pipette to scratch the bottom of the 6-well plate. The exfoliated tumor cells were washed with PBS, and cell migration photographs were taken under a microscope at 0 h and 48 h, respectively.

Migration and invasion experiments were performed using a Transwell chamber with or without Matrigel (Corning Life Sciences, NY, USA). The upper chamber is serum-free DMEM, and the lower chamber is DMEM containing 10% FBS. After 24 and 48 h of culture, tumor cells which passed through the chamber were fixed and stained by 4% paraformaldehyde and 0.1% crystal violet. Images were taken under an inverted microscopy.

Cellular extracts were prepared at 4 °C in RIPA buffer (Beyotime Biotechnology, Shanghai, China). Western blot analysis was performed using commercial primary antibodies against the following proteins: Bcl-2 (1:2000; ab32124, Abcam), BAX (1:2000; ab32503, Abcam), MMP-9 (1:2000; ab38898, Abcam), cyclin D1 (1:2000; ab134175, Abcam), GAPDH (1:2000; ab8245, Abcam), vimentin (1:1000; 5741 T, CST), E-cadherin (1:1000; 3195 T, CST) N-cadherin (1:1000; 13,116 T, CST), β-catenin (1:1000; 8480 T, CST), EGFR (1:1000; 2085S, CST), CD31 (1:1000; 3528S, CST) and p-EGFR (1:1000; 3777S, CST). The bands were detected using HRP-conjugated secondary antibodies from Beyotime Biotechnology (Shanghai, China): goat anti-rabbit (1:1000, A0208) and goat anti-mouse (1:1000, A0216). Chemiluminescence was identified using Millipore chromogenic solution (MilliporeSigma, Burlington, MA, USA).

Transduced CAL27 cells (2 × 10⁷ cells in 100 μl) were injected into 4-week-old Balb/c athymic nude mice (Siliake Jingda Experimental Animal Co. Ltd., Hunan, China).

The tumor volume of nude mice was measured and recorded according to V = πAB2/6 (V: tumor volume, A: the largest diameter, B: the perpendicular diameter). After 6 weeks, the nude mice were euthanized. Tumors of nude mice were dissected and tumor weights were measured. Tumor specimens were treated accordingly to perform hematoxylin and eosin (H&E) staining, Western blot, IHC staining.

All images shown are wide-field microscopy images. Results in graphs are shown as the mean ± SEM from three independent experiments. All statistical data were analyzed using SPSS 17.0 software (SPSS, Chicago, IL, USA). Two-tailed Student’s t-tests were used to determine P values; P < 0.05 was considered significant.

We determined the RNA levels of hsa_circ_0005379 by qRT-PCR and analyzed 37 pairs of clinical tissues. It is revealed that hsa_circ_0005379 expression was lower in cancerous tissues than adjacent normal tissues (Fig. 1a). We also investigated the expression level of hsa_circ_0005379 in four OSCC cell lines. We observed that hsa_circ_0005379 expression was remarkably downregulated in oral cancer cell lines compared to the expression in HOK cells (Fig. 1b). We also analyzed the correlation between clinicopathological features and hsa_circ_0005379 expression levels in 37 OSCC patients (Table 1). Results showed a negative correlation between OSCC tumor size (P = 0.0192) and differentiation grades (P = 0.0057). The ROC curve is shown in Fig. 1c.

To investigate the role of hsa_circ_0005379 in regulating cell proliferation, we performed the CCK-8 assay on SCC25 and CAL27 cells overexpressing hsa_circ_0005379. After transducing SCC25 and CAL27 cells with hsa_circ_0005379 lentivirus, hsa_circ_0005379 expression levels increased approximately 4-fold and 6-fold, respectively (Additional file 1: Figure S1). The OD values were measured to determine the proliferation rate. The results revealed that with an increase in transduction time, hsa_circ_0005379 overexpressing cells showed a significant decrease in proliferation rate compared with control group (Fig. 2a). EdU staining was conducted to measure proliferation ratios to confirm these results. High expression of hsa_circ_0005379 in SCC25 and CAL27 decreased cell proliferation ratios by approximately 6 and 17% compared to the control group, respectively (Fig. 2b). Hoechst 33258 staining experiments showed that high expression of hsa_circ_0005379 also promoted apoptosis of tumor cells (Fig. 3).

To determine the role of hsa_circ_0005379 in OSCC cell migration, wound healing assays were performed on control and hsa_circ_0005379 overexpression SCC25 and CAL27 cells. The scratch area at 0 and 48 h after wounding were measured and revealed that the wound-closure rate in hsa_circ_0005379 overexpression cells was significantly lower than control cells. Closure rate at 48 h in hsa_circ_0005379 overexpression cells was about 18% lower (SCC25; Fig. 4a) and 19.6% lower (CAL27; Fig. 4b) than those in the control cells. After 24 h of incubation, the number of cells that migrated to the lower chamber in a Transwell assay decreased from 850 to 600 and 300 in SCC25 and CAL27 cells overexpressing hsa_circ_0005379, respectively, (Fig. 4c). Taken together, these results suggest that OSCC cell migration ability was impaired by the upregulation of hsa_circ_0005379. To further study OSCC cell invasion ability, we also performed a Transwell invasion assay. After 48 h incubation, the number of cells across the chamber coated with Matrigel was measured. Overexpression of hsa_circ_0005379 strongly suppressed the invasiveness of SCC25 and CAL27 cells. In both cases, the invading hsa_circ_0005379 overexpression cells were about half the number of invading control cells (Fig. 4d). Co-culture experiments showed that the ability of tumor cells to induce HUVEC cell migration decreased after overexpressing hsa_circ_0005379 (Fig. 5a, b). Moreover, HUVEC tube formation assay was performed. Incubated with conditioned medium, overexpression of hsa_circ_0005379 resulted in less tubule formation (Fig. 5c). The experimental results showed that overexpression of hsa_circ_0005379 affects the ability of tumor cells to induce angiogenesis tube formation compared with the control group. These findings indicate that overexpression of hsa_circ_0005379 can not only inhibit migration and invasion of OSCC cells, but also influence the angiogenesis tube formation.

Since cetuximab is a commonly used anticancer drug for OSCC treatment, we performed a drug treatment experiment to investigate the effect of hsa_circ_0005379 on OSCC cell viability. Apoptosis rates in hsa_circ_0005379 overexpression cells were measured by annexin V-FITC/PI dual-label flow cytometry. We used flow cytometry to detect apoptosis in different treatment groups of SCC25 (Fig. 6a) and CAL27 (Fig. 6b). Early apoptotic rates in the mock group were 0.31 and 0.43% in SCC25 and CAL27 cells, respectively, while early apoptotic rates in the hsa_circ_0005379 group were 1.12 and 0.91% in SCC25 and CAL27 cells, respectively. Early apoptotic rates in the mock + cetuximab group were 17.88 and 15.22% in SCC25 and CAL27 cells, respectively, while early cell apoptotic rates in hsa_circ_0005379 + cetuximab group increased to 38.35 and 35.77% in SCC25 and CAL27 cells, respectively. Our experimental results show that high expression of hsa_circ_0005379 can promote the apoptosis of tumor cells. OSCC cells with high expression of hsa_circ_0005379 significantly increased the sensitivity of OSCC cells to cetuximab and promoted tumor cell apoptosis.

To explore the mechanism of hsa_circ_0005379 in regulating OSCC, we examined the expression level of related proteins. The Bcl-2 gene is an oncogene that has an inhibitory effect on apoptosis. BAX is an apoptosis-promoting protein in the BCL-2 family. Overexpression of BAX antagonizes the protective effect of BCL-2 and causes cell death [13‐15]. MMP-9 is capable of degrading collagen, gelatin, fibronectin, laminin, and dissolving the basement membrane, thereby promoting migration and invasion of tumor cells [16]. The main function of cyclin D1 is to promote cell proliferation and gene transcription. Cyclin D1 promotes the cell cycle from the G1 phase to the S phase by binding to and activating the cyclin-dependent kinase CDK4, which is unique to the G1 phase. Cyclin D1 has been recognized as a proto-oncogene. The overexpression of Cyclin D1 is associated with malignant transformation [17]. Our results showed that overexpression of hsa_circ_0005379, BAX was upregulated, whereas Bcl-2, MMP-9, and cyclin D1 were downregulated (Fig. 7a). Epithelial–mesenchymal transition (EMT) refers to the biological process by which epithelial cells are transformed into mesenchymal phenotype cells by a specific procedure. Through EMT, epithelial cells lose cell polarity and epithelial phenotypes such as attachment to the basement membrane, and thus obtain higher interstitial phenotypes such as migration, invasion, anti-apoptosis, and ability to degrade extracellular matrix. EMT is an important biological process for the migration and invasion of epithelial-derived malignant cells [18]. Here, to evaluate the EMT process, we overexpressed hsa_circ_0005379 and detected EMT markers by western blot analysis. Compared to control OSCC cells, hsa_circ_0005379 overexpression cells showed decreased expression of vimentin and N-cadherin and increased expression of E-cadherin and β-catenin (Fig. 7b). Taken together, hsa_circ_0005379 participates in the EMT process and affects cell proliferation and invasion by regulating the expression levels of several key proteins.

Additionally, we used siRNA to knockdown hsa_circ_0005379 in SCC25 and CAL27 cells. The knockdown efficiency is about 70% in SCC25 and 85% in CAL27 (Additional file 1: Figure S2). The expression levels of BAX, Bcl-2 and MMP-9 are inversely correlated with hsa_circ_0005379 expression (Fig. 7c). Western blot results showed that EGFR and p-EGFR expression levels were increased after hsa_circ_0005379 knockdown (Fig. 7d), while EGFR and p-EGFR expression levels were decreased after hsa_circ_0005379 overexpression (Fig. 7e). These results suggest that manipulating expression levels of hsa_circ_0005379 will influence the EGFR pathway. We used EGFR pathway agonists (NSC228155) and inhibitors (cetuximab) at the indicated times and concentrations to alter the expression level of phosphorylated EGFR in the cells and then detected the corresponding hsa_circ_0005379 expression level by qRT-PCR. We found no significant change in hsa_circ_0005379 expression levels (Fig. 7f). Moreover, the proliferation differences between hsa_circ_0005379 overexpression and knockdown cell lines treated with EGFR inhibitor cetuximab or agonist NSC in different time points have been investigated, which is shown in Additional file 1: Figure S3. The above results indicate that hsa_circ_0005379 may be located upstream of the EGFR pathway and regulate the expression level of EGFR.

To establish a xenograft tumor model, we injected CAL27 cell lines in nude mice. Tumors in the hsa_circ_0005379 overexpression group were considerably smaller than those in the empty-vector group (Mock, Fig. 8a). The tumor growth curve and final weight of the nude mice were suppressed in the hsa_circ_0005379 overexpression group relative to the control (Fig. 8b, c). Moreover, the H&E staining (Fig. 8d) and western blot (Fig. 8e) showed that the angiogenesis marker CD31, EGFR and p-EGFR were downregulated when hsa_circ_0005379 was overexpressed. Immunohistochemical (IHC) (Fig. 8f) staining of xenograft tumors showed that vimentin was downregulated in hsa_circ_0005379 overexpressing tumors, while E-cadherin was upregulated, which supported the western blot results. Expression of EGFR was downregulated in hsa_circ_0005379 overexpression tumors compared with the control group. Collectively, these results showed that hsa_circ_0005379 is crucial for tumor growth.