CAL27 cells and SCC-9 cells，the head and neck squamous cell carcinoma cell line，were cultured in DMEM (Gibco, USA) supplemented with 10% fetal calf serum (Thermo Fisher Scientific, USA), 100 μg/ml streptomycin (Gibco, USA), and 100U/ml penicillin (Gibco, USA) at 37℃ with 5% CO2.
TCGA data processing
For expression analysis, we selected RNAseq and RNAseqV2 paired sample data for analysis. The raw sequencing data and pathological information were downloaded from the TCGA database (https://cancergenome.nih.gov/). For Head and Neck squamous cell carcinoma (HNSCC), there were 528 samples with available data in the TCGA database, including 520 RNAseqV2 samples, and 40 pairs with paired sample data and pathological information. Our expression profile analysis was based on these 40 paired samples of RNAseqV2 data. The Trimmed Mean of M-values (TMM) method was applied to data standardization. To avoid errors caused by inappropriate sample grouping, BCV（biological coefficient of variation）was observed for quality control.
Firstly, we estimated the dispersion of multiple pairs of samples and then used a general linear model to estimate whether there is a difference in genes between different groups. Genes with a statistical test P value less than 0.05 are considered differentially expressed genes. The fold changes (FCs) in the expression were calculated and differentially expressed DEGs with log2|FC|≥1.0 (Cancerrmal) were considered to be significant, remaining genes are filtered. Besides, several other criteria were adopted to filter genes: the genes which had been reported have functional and clinical relation in HNSCC; genes of multiple transmembrane proteins; gene annotation is not clear (such as open reading frame); genes reported in more than 100 articles in Pubmed. The final gene list was randomly concentrated, and C1QTNF6 was selected for further experimental research.
Lentivirus Construction and lentiviral transfection of CAL27 cells and SCC-9 cells
C1QTNF6-targeting short hairpin RNA (shRNA) oligonucleotides sequences（TGTGTGAGATCCCTATGGT） were designed and were synthesized and cloned into the pGV115-GFP vector by GeneChem Corporation (Shanghai, China). The sequence of scrambled shRNA (TCTCCGAACGTGTCACGT) used as the negative control (NC) was also inserted into the pGV115-GFP vector. Lentivirus transfection was performed according to the manufacturer’s recommended protocol. The CAL27 cells and SCC-9 cells were seeded into a six-well plate (~1×105 cells per well) and incubated at 37℃ with 5% CO2 until they reached ~70% confluence, then lentivirus was added as MOI 1. After 72 h, cells were observed under a fluorescence microscope and collected for the following experiments.
RNA extraction and quantitative real-time PCR (qRT-PCR)
Total RNA was isolated from cells using TRIzol (Invitrogen, USA) according to the manufacturer’s instructions. Reverse transcription was performed to synthesize cDNAs using M-MLV reverse transcriptase (Promega) and OligodT primers (Sangon, Shanghai). C1QTNF6 mRNA expression was detected with quantitative real-time PCR using SYBR Master Mixture (Takara, Japan) on a Real-Time PCR machine TP800 (Takara). The cycling parameters were 95℃ for a 30s hot start followed by 45 cycles of 95℃ for 5 s and 60℃ for 30s. Primers used here were as follows: GAPDH for:5′-TGACTTCAACAGCGACACCCA -3′, GAPDH reverse:5′-CACCCTGTTGCTGTAGCCAAA -3′, C1QTNF6 for: 5′- GAAAGGGTCTTTGTGAACCTTGA -3′, and C1QTNF6 reverse: 5′- CTGCGCGTACAGGATGACAG -3′. The relative mRNA expression was calculated using the 2−ΔΔCt method.
Western blot analysis
Cell samples were harvested, washed twice with PBS, and Lysed with pre-cooled Lysis Buffer. The protein concentration was determined by BCA Protein Assay Kit (Pierce). Each sample protein concentration is adjusted to 2μg/μL, and saved at - 80℃ for later use. The total protein was separated on 10 % sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene fluoride (PVDF) membranes at 4℃ (300mA for 150min). The PVDF membranes were blocked in 5% skim milk in TBST for 1h and incubated overnight with primary antibodies at 4℃. After washing three times in TBST, the membranes were incubated in secondary antibody coupled to HRP before visualization.
13 paired cancer and paracancerous tissue samples were collected from OSCC patients in Qilu hospital of Shandong University, both preoperative and postoperative pathology of which suggest oral squamous cell carcinoma. All patients gave informed consent before the tissue was collected. The research had obtained the approval of Ethics Committee on Scientific Research of Shandong University Qilu Hospital and the approval number is KYLL-2017-544. OSCC tissues and normal tissues were immunostained for C1QTNF6 expression using a C1QTNF6 antibody (Invitrogen, USA) at a 1:100 dilution for a 1-h incubation according to the supplier’s instructions. The secondary antibody used was a goat anti-rabbit IgG incubated for 1 h at 37℃. Images were obtained at the optical facility.
Cell Cycle Analysis by Flow Cytometry
Lentivirus-transfected cells were seeded in 6-cm dishes for culture. When the cells achieved approximately 80 % confluence, they were trypsinized, washed twice in PBS, and fixed with pre-chilled 70% ethanol for at least 1h at 4℃. Then the cells were washed with PBS and stained with PI mixture (40×PI stock (2 mg/ml), 100×RNase stock (10 mg/ml) and 1×PBS buffer at a dilution of 25:10:1,000) for 45 min at 37℃. The cell cycle results were measured by Guava easyCyte HT (Millipore, USA), and all experiments were performed in triplicate.
Annexin V Apoptosis Assay
Cell apoptosis analysis was measured using eBioscience™ Annexin V Apoptosis Detection Kit APC according to the manufacturer’s instructions. Briefly, Lentivirus-transfected cells were trypsinized, washed twice in pre-chilled D-hank’s, and resuspended in binding buffer. Then 5 μl of annexin V-APC was added into 100μl of the cell suspension and incubated at room temperature in the dark for 15 min. Samples were analyzed by Guava easyCyte HT (Millipore, USA).
In vivo tumorigenicity
Twenty 4-week-old female BALB/c nude mice were purchased from the Shanghai Institute for Biological Sciences (Shanghai, China). Twenty BALB/c nude mice were randomly divided into two groups. The lentivirus-transfected cells (1 × 107) were inoculated into the right side of the axillary of nude mice subcutaneously as previously described. The mice were observed every 5 days after anesthesia by intraperitoneal injection of 0.7% pentobarbital sodium according to the dosage of 10μL per gram of body weight. The length and width of tumors were measured using calipers, and the volume of tumors was calculated using the equation (L × W2)/2. On day 53, tumor growth was monitored by luciferase activity detection using the IVIS imaging system. Meanwhile, the mice were euthanized by injection of an overdose of 2% pentobarbital, after which cervical dislocation was performed to confirm death, and the tumors were removed and weighed.
Pathway Analysis (IPA)
Firstly, a signal histogram was made to show the signal intensity distribution of all chip probes, and the average Z-score value of all samples in the same signal intensity interval is less than 2. Relative Signal Box Plot showed the medians of all samples, and the Z-score value of the medians was less than 2. Correlation Analysis was demonstrated by the Pearson correlation coefficient distribution chart, which indicated that the correlation coefficients within both groups were all greater than 0.99. The principal component analysis was then conducted. All the works above confirmed that the data owned fine reliability, repeatability, also significant differences between groups, so it met the criteria for continued analysis. The lowest 20% of the signal intensity of the probe set was filtered out as background noise. Secondly, we used the coefficient of variation method to calculate the coefficient of variation of the same probe in the same sample, and filtered out probes with a coefficient of variation greater than 25% in both groups. Finally, we got data of 39,287 probes.
We used a linear model based on empirical Bayes distribution to calculate the significant difference level P-value and used the Benjamini-Hochberg method to correct the significant difference level. The screening criteria for significantly different genes are |Fold Change|>1.5 and FDR<0.05. Hierarchical Clustering was proceeded. The clustering algorithm classifies samples and variables in two dimensions.
We used the Ingenuity Pathways Analysis (IPA, Ingenuity systems, Inc., Redwood City, CA, www.ingenuity.com) tool to examine biological functions and disease as well as functional relationships between genes and gene networks.