The present study was approved by the Ethics Committee of Weifang People's Hospital (Weifang, China; approval no. WPH011). A total of 64 patients with CSCC (all females; age range, 42–64 years; mean age, 51.7±6.7 years) admitted at the aforementioned hospital between December 2012 and December 2014 were enrolled in the present study. All patients were newly diagnosed cases, and no other severe clinical disorders, such as chronic diseases, severe infections, heart diseases or other types of cancer, were observed among these patients. No therapy was initiated before the admission of patients. All patients provided written informed consent. Fine-needle aspiration was performed to collect CSCC and paired adjacent (within 3 cm around tumors) non-tumor tissues from all patients before therapy. Histopathological examinations were performed to confirm all tissue samples. Tissue samples were stored in liquid nitrogen before subsequent experiments. The clinical data of all 64 patients are listed in Table I.

The 64 patients with CSCC included 12, 17, 20 and 15 cases at clinical stage I, II, III and IV, respectively, based on the American Joint Committee on Cancer (AJCC) staging (14). Based on the health conditions and AJCC stage of the patients, anticancer therapies, such as surgical resection, chemotherapy, radiotherapy or combined therapy, were performed. From the day of admission, all patients were followed up for 5 years (through phone call and/or outpatient visit) in a monthly manner to record survival, and patients who died of causes other than CSCC were excluded from the present study (77 patients with CSCC were enrolled at the beginning of the study, and 13 patients who died of causes unrelated to CSCC were excluded).

The CSCC SiHa and C33A cell lines were obtained from the American Type Culture Collection. Cells were cultured in Eagle's Minimum Essential Medium with 10% FBS (both Gibco; Thermo Fisher Scientific, Inc.) at 37°C with 5% CO₂ at 95% humidity. Cells were harvested at ~85% confluence and used in subsequent transient transfection experiments.

The PSMG3-AS1 expression vector was constructed using pcDNA3.1 vector (Invitrogen; Thermo Fisher Scientific, Inc.) as the backbone. miR-4417 mimic (5′-GGUGGGCUUCCCGGAGGG-3′) and its non-targeting negative control (NC) miRNA (5′-UGCCACGUGGCAUGCAGUG-3′) were purchased from Sigma-Aldrich (Merck KGaA). SiHa and C33A cells were transfected with 10 nM PSMG3-AS1 expression vector or 45 nM miR-4417 mimic using Lipofectamine^® 2000 (Invitrogen; Thermo Fisher Scientific, Inc.). Incubation with transfection mixture was performed at 37°C for 6 h. Empty vector- or NC miRNA-transfected cells were used as NC cells. Untransfected cells were used as the control (C) cells. Subsequent experiments were performed 48 h post-transfection.

The pGL3-PSMG3-AS1 luciferase reporter vector (Promega Corporation) was established. Cells were transfected with miR-4417 mimic + pGL3-PSMG3-AS1 (miR-4417 group) or NC miRNA + pGL3-PSMG3-AS1 (NC miRNA group) using the aforementioned transfection method. Luciferase activity was measured after 48 h using the Firefly Luciferase Assay kit 2.0 (Biotium, Inc.). Firefly luciferase activity was normalized to Renilla luciferase activity.

Total RNA was extracted from tissue samples and cultivated cells using RNAzol (Sigma-Aldrich; Merck KGaA). To harvest miRNAs, 85% ethanol was used to precipitate and wash RNA samples. RNA concentrations were measured using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Inc.). RNA samples were digested with DNase I to remove genomic DNA.

Preparation of cDNA samples was performed via RT using the QuantiTect Reverse Transcription kit (Qiagen China Co., Ltd.) in accordance with the manufacturer's instructions. The templates were RNA samples from both tissue and cultivated cells. The expression levels of PSMG3-AS1 were measured using BlazeTaq™ One-Step SYBR-Green RT-qPCR kit (GeneCopoeia, Inc.) following the manufacturer's instructions, with β-actin as the internal control. The expression levels of mature miR-4417 were measured using All-in-One™ miRNA qRT-PCR Detection kit (GeneCopoeia, Inc.) with U6 as the internal control. All PCR reactions were performed in triplicate, and the 2^−ΔΔCq method (15) was used to normalize Ct values: ΔCT = Ct (target gene) - Ct (internal control). The sample with the biggest ΔCt value was set to 1, and all other samples were normalized to this sample. The primer sequences were as follows: PSMG3-AS1 forward, 5′-GAAGCAGAACCAACGCACAG-3′ and reverse, 5′-GCATAATCCAATCCCTCAAGAA-3′; β-actin forward, 5′-TGACGGGGTCACCCACACTGTGCCCATCTA-3′ and reverse, 5′-CTAGAAGCATTTGCGGTGGACGATGGAGGG-3′; miR-4417 forward, 5′-GGTGGGCTTCCCGGA-3′. Universal reverse primer and U6 primers were included in the aforementioned All-in-One™ miRNA qRT-PCR Detection kit. The qPCR conditions consisted of 95°C for 1 min, followed by 40 cycles of 95°C for 10 sec and 58°C for 40 sec.

The potential interaction between miR-4417 and PSMG3-AS1 was predicted using IntaRNA 2.0 (16). In this program, miR-4417 was used as short sequence and PSMG3-AS1 was used as long sequence.

SiHa and C33A cells were harvested 48 h post-transfection and subjected to Transwell assay using Transwell^® Cell Culture Plate Inserts (8.0-µm pore; Corning, Inc.). Briefly, 600 cells in 0.1 ml serum-free EMEM were seeded in the upper chamber. The lower chamber was filled with EMEM with 20% FBS. Matrigel (EMD Millipore)-coated membranes (6 h at 37°C) were used in invasion assays, while uncoated membranes were used in migration assays. Cells were cultivated for 12 h at 37°C, followed by staining of the lower surface of membranes with 1% crystal violet (Sigma-Aldrich; Merck KGaA) for 20 min at room temperature in the dark. Cells were counted under a light microscope (magnification, ×20) and images were obtained.

GraphPad Prism 6 (GraphPad Software, Inc.) was used to process and analyze data. Data were expressed as the mean ± SEM of three biological replicates. Paired Student's t-test was used to compare gene expression levels between CSCC and non-tumor tissues. One-way ANOVA and Tukey's post-hoc test were used to compare multiple groups. The 64 patients with CSCC were divided into high and low PSMG3-AS1 expression groups (n=32) based on the median relative PSMG3-AS1 expression level (4.27) in CSCC tissues. Survival curves were plotted based on the 5-year follow-up data using GraphPad Prism 6 with the ‘Survival’ panel and ‘Comparing two groups’ function. Survival curves were compared using the log-rank test. Correlations were analyzed using Pearson's correlation coefficient. χ² test was performed to analyze the association between the expression levels of miR-4417, PSMG3-AS1 and the patients' clinical data. P<0.05 was considered to indicate a statistically significant difference.

The expression levels of PSMG3-AS1 in both CSCC and non-tumor tissues from the 64 patients with CSCC were measured via RT-qPCR. Compared with non-tumor tissues, the expression levels of PSMG3-AS1 were significantly increased in CSCC tissues (Fig. 1A; P<0.05). Survival curve analysis revealed that the overall survival rate of patients in the high PSMG3-AS1 expression group was significantly lower than that in the low PSMG3-AS1 expression group (Fig. 1B; P=0.0448). χ² test revealed that the expression levels of PSMG3-AS1 were not significantly associated with age, smoking habit, HPV infection and clinical stage (data not shown).

The expression levels of miR-4417 in both CSCC and non-tumor tissues from the 64 patients with CSCC were measured via RT-qPCR. Compared with non-tumor tissues, the expression levels of miR-4417 were significantly lower in CSCC tissues (Fig. 2A; P<0.05). Pearson's correlation coefficient analysis revealed that the expression levels of miR-4417 were inversely and significantly correlated with the expression levels of PSMG3-AS1 in both CSCC tissues (Fig. 2B) and non-tumor tissues (Fig. 2C). χ² test revealed that the expression levels of miR-4417 were not significantly associated with age, smoking habit, HPV infection and clinical stage (data not shown).

The potential interaction between miR-4417 and PSMG3-AS1 was predicted using IntaRNA 2.0 (14). It was observed that miR-4417 and PSMG3-AS1 may form multiple base pairing (Fig. 3A). Luciferase reporter assay revealed that, compared with the NC miRNA group, the miR-4417 group exhibited significantly decreased luciferase activity, indicating a direct interaction between miR-4417 and PSMG3-AS1 (Fig. 3A). SiHa and C33A cells were transfected with either miR-4417 mimic or PSMG3-AS1 expression vector, and the overexpression of miR-4417 and PSMG3-AS1 were confirmed 48 h post-transfection via RT-qPCR (Fig. 3B; P<0.05). Compared with the C and NC groups, overexpression of miR-4417 significantly decreased the expression levels of PSMG3-AS1 (Fig. 3C; P<0.05), while overexpression of PSMG3-AS1 did not affect miR-4417 expression (Fig. 3D) in both cell lines. These results suggested that miR-4417 may target PSMG3-AS1 to downregulate its expression, while PSMG3-AS1 did not regulate miR-4417 expression.

Transwell assay was performed to assess the effect of overexpression of PSMG3-AS1 and miR-4417 on invasion (Fig. 4A) and migration (Fig. 4B) of SiHa and C33A cells. Overexpression of PSMG3-AS1 significantly increased invasion and migration rates of cancer cells compared with control cells, while overexpression of miR-4417 significantly inhibited invasion and migration of cancer cells and attenuated the effect of PSMG3-AS1 overexpression (Fig. 4A and B; P<0.05). The expression levels of miR-4417 (Fig. 5A; P<0.05) and PSMG3-AS1 (Fig. 5B; P<0.05) in all groups included in Transwell assays were also measured via RT-qPCR. It was observed that the cell invasion and migration patterns were consistent with the expression pattern of PSMG3-AS1, with the highest expression level of PSMG3-AS1 being observed in the PSMG3-AS1 group and the lowest PSMG3-AS1 level in the miR-4417 group (Fig. 5B).