Patients and Clinical Samples
To generate the HNSCC PDX, the human HNSCC tissues were obtained without patient information from the Peking University Hospital of Stomatology. The tumor tissues were cut into small pieces, followed by implantation into the flanks of NOD-SCID mice (6 weeks old), according to a previously described method.[8] HNSCC specimens from 60 patients were obtained from the Peking University School and Hospital of Stomatology from September 2012 to October 2016. The inclusion criteria were as follows: 1) the tumor was in the tongue; 2) there was no distant metastasis; 3) removal of the primary carcinoma and neck dissection without preoperative radiotherapy or chemotherapy; and 4) patients who underwent postoperative follow-up for at least five years. Without conducting a pathological study, the clinical TNM staging approach was used to classify the tumor size and clinical stage for the 40 HNSCC samples among the 60 samples: 1) tumor size limited in T2 and T3, 2) clinically negative cervical lymph node (cN0), and 3) no distant metastasis (M0). Based on the histopathologic evaluation of the lymph nodes, these 40 patients were split into lymph node negative and positive groups. The Institutional Review Board of the Peking University School and Hospital of Stomatology approved the protocol. The tissues were snap-frozen and placed at − 80°C until analysis.
4NQO Mouse Model of HNSCC, Treatment and Histology
The BMI1CreER; RosatdTomato mice received drinking water with 4NQO for 16 weeks to allow HNSCC to develop, followed by normal drinking water for 6 weeks, to form a spontaneous model of HNSCC. The mice were randomly divided into groups at 22 weeks. Before sacrificing the mice, they were given tamoxifen to label BMI1+ CSCs. For the treatment assay, BMI1CreER; RosatdTomato mice were divided randomly into the indicated groups, and injected with the indicated ASO PVT1 (10 nM, Integrated Biotech Solutions, Shanghai, China) and ASO NC over the whole mouse tongue by twice weekly subcutaneous injection for 4 weeks. For combination therapy, mice were intraperitoneally injected with anti-PD1 (BioXcell, Lebanon, NH, USA, 200 µg/mouse twice per week). After mice were sacrificed, the cervical lymph nodes and tongues were removed, and the lesion surface areas were calculated. For histological investigation and immunostaining, longitudinally cut tongues (dorsal/ventral) and intact lymph nodes were fixed overnight in 4% paraformaldehyde and paraffin-embedded. 10 sections of 5 mm thick tissue blocks were cut, and they were then stained with hematoxylin and eosin (H&E). The SCC number was counted and regions were measured.[8] The following criteria were used to grade the invasiveness of the HNSCC: showing signs of normal or epithelial dysplasia appearance (grade 1); distinct invasion, unclearness of basement membrane, drop and diffuse infiltration into the superficial portion of the muscle layer (grade 2); loss of the basement membrane; extensive invasion into deep muscle layer (grade 3). To examine cervical lymph node metastasis of HNSCC, the sections of cervical lymph nodes were immunostained with anti-PCK antibodies.
CSCs Isolation and Tumorsphere Formation Assay
An ALDEFLUOR assay kit (Stemcell Technologies, Vancouver, Canada) was used to sort ALDHhigh CSCs from HN6 and SCC15 cells following the manufacturer’s protocol. Flow cytometry was used to detect green fluorescence positive cells among live cells, in comparison with the fluorescence intensity of the diethylaminobenzaldehyde (DEAB) treated sample. These detected cells would have high ALDH activity (ALDHhigh cells), and were used for subsequent experiments. Human HNSCC xenografted tumors were chopped into small pieces and digested into single cell suspensions using a human tumor cell dissociation kit (Miltenyi Biotec, Bergisch Gladbach, Germany) to isolate CSCs from human HNSCC PDX. An EpCAM-PE (Miltenyi Biotec) was used to isolate EpCAM+ tumor cells, and an ALDHEFLUOR assay kit was used to sort the ALDHhigh and ALDHlow subpopulations from the EpCAM+ tumor cells by flow cytometry. For the tumorsphere formation assay, ALDHhigh cells were added to ultralow attachment plates and cultured in DMEM/F12 without serum (Thermo Fisher Scientific, Waltham, MA, USA) but containing 1% B27 supplement (Thermo Fisher Scientific), 1% N2 supplement (Thermo Fisher Scientific), human recombinant epidermal growth factor (EGF, 20 ng mL − 1; R&D Systems, Minneapolis, MN, USA), and human recombinant basic fibroblast growth factor (bFGF, 10 ng mL − 1; R&D Systems). After 7–14 days, spheres with a diameter exceeding 70 µm were counted under a microscope.
In Vivo Tumor Growth in Mice
Female Nude/SCID mice (6–8 weeks old) were obtained from Beijing Sipeifu Biotechnology Co. Ltd. HN6 cells (with or without PVT1 KD) were injected subcutaneously into the backs of the nude mice, after mixing with an equal volume of Matrigel. The nude/SCID mice were monitored; and the weights and volumes of the xenograft tumors were measured using sliding calipers. At 4 weeks after injection, the mice were sacrificed, and the tumors were excised and weighed. For the in vivo limiting dilution assay, HN6 cell-derived ALDHhigh CSCs or EpCAM+ ALDHhigh CSCs were subjected to infection for 24 h with the lentiviruses expressing shPVT1 or shCtrl. Following rapid puromycin selection, varying numbers of cells were mixed with an equal volume with Matrigel and injected subcutaneously into the backs of nude mice for 4 weeks. The Extreme Limiting Dilution Analysis software (http://bioinf.wehi.edu.au/software/elda/) was used to analyze the data. Tumor volumes were computed using the formula (L ⋅ W2)/2 (where L is the longer diameter and W is the shorter diameter). For orthotopic tumor formation, HN6derived ALDHhigh CSC-like cells were inoculated sublingually into the tongues of Nude/SCID mice and grown for 1 week. For treatment, tumor-bearing mice were divided randomly into the indicated groups, and injected with the indicated ASO PVT1 (10 nM, Integrated Biotech Solutions) and ASO NC over the whole mouse tongue by twice weekly subcutaneous injection for 4 weeks. After mice were sacrificed, the cervical lymph nodes and tongues were removed.
Immunostaining
4% paraformaldehyde was used to fix mouse HNSCC tissues and cervical lymph nodes for 24 h, followed by equilibration in 30% sucrose/phosphate-buffered saline (PBS), and embedding in optimal cutting temperature compound (Sakura Finetek, Tokyo, Japan). Coronal 5 µm sections were cut on a freezing microtome (Leica, Wetzlar, Germany). Thereafter, sections were stained with 4′,6-diamidino-2-phenylindole (DAPI) solution (Solarbio). These tissues were embedded in paraffin and cut into 5 µm sections. Sections were deparaffinized with xylene and rehydrated through an ethanol series and distilled water. For immunohistochemistry, antigens were repaired using high temperature and pressure in citrate buffer (pH = 6). Sections were then incubated with the following primary antibodies at 4°C overnight: anti-PCK (1:200; Santa Cruz Biotechnology, Santa Cruz, CA, USA; Cat#sc-8018). Next, horseradish peroxidase-labeled polymer was incubated with the sections for 60 min. The signals were detected using 3,3′-Diaminobenzidine as a chromogen (Zhongshan Golden Bridge, Beijing, China) to detect the signals, followed by counterstaining using hematoxylin. For immunohistofluorescence, after antigen repair, the sections were incubated at 4°C overnight with the following primary antibodies: CD8a (1:1000; Cell Signaling Technology, Cat#98941S), anti-PCK (1:200; Abcam, Cambridge, MA, USA; Cat#ab9377), anti-phospho Histone H2A.X (Ser139) (1:200; Cell Signaling Technology, Cat#9718). Thereafter, the sections were visualized using Cy2 and Cy3-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories, West Grove, PA, USA).
Cell Culture and siRNA or shRNA Knockdown
The American Type Culture Collection (ATCC, Manassas, VA) provided the human HNSCC cell lines HN6 and SCC15. Dulbecco’s modified Eagle’s medium (DMEM) with 10% fetal bovine serum (FBS) and 1% antibiotics was used to culture the cells at 37°C, 5% CO2 in a humidified incubator. Cells were transfected with miRNA mimics (Integrated Biotech Solutions) or siRNA (Tsingke Biotechnology, Beijing, China) using the Lipo8000™ Transfection Reagent (Beyotime, Shanghai, China) following the manufacturer’s instructions. For the generation of stable cell lines, cells were infected with lentiviruses (Integrated Biotech Solutions), followed by puromycin selection.
Western Blot
Cells were lysed in Radioimmunoprecipitation assay buffer (Solarbio, Beijing, China) containing phenylmethylsulphonylfluoride protease inhibitor. Protein samples (50 µg) were subjected to 10% SDS polyacrylamide gel electrophoresis. The separated proteins were subsequently transferred to a polyvinylidene fluoride membrane. The membrane was blocked using 10% non-fat milk for 1 h at room temperature and then incubated with primary antibodies overnight at 4 ℃. Primary antibodies recognizing the following proteins were used: BMI1 (1:1000; Cell Signaling Technology, Danvers, MA, USA; Cat # 6964), YAP1 (1:1000; Cell Signaling Technology, Cat#14074), GAPDH (encoding glyceraldehyde-3-phosphate dehydrogenase) (1:20000; Proteintech, Rosemont, IL, USA; Cat# 60004-1). anti-phospho Histone H2A.X (Ser139) (1:1000; Cell Signaling Technology, Cat#9718). Thereafter, the membranes were incubated with secondary antibodies corresponding to the primary antibodies for 1 h at room temperature. Finally, the immunoreactive protein bands were visualized using ECL (NCM Biotech, Suzhou, China) Western blot detection reagent.
qRT-PCR
Total RNA was isolated using the Trizol reagent (Invitrogen, Waltham, MAS, USA) following the manufacturer’s protocol. RNA (1–2 µg) was reverse transcribed to cDNA using random primers (Takara, Shiga Japan). The mRNA levels were quantified using quantitative real-time PCR with the cDNA as the template employing the SYBRGreen supermix (Bio-Rad, Hercules, CA, USA). The internal control was GAPDH. Table S1 (Supporting Information) lists the primers used for qRT-PCR.
Nuclear and Cytoplasmic Fraction Isolation
Nuclear and cytoplasmic RNA were isolated using a Nuclear and Cytoplasmic Protein Extraction Kit (Beyotime) following to the manufacturer’s protocol. Briefly, cell fraction buffer was used to lyse the cells, followed by slow speed centrifugation to separate the nuclear fraction from the cytoplasmic fraction. RNA from the two fractions was isolated separately using Trizol. MALAT1 and GAPDH RNA were used as controls for the nuclear and cytoplasmic fractions, respectively.
Cell Proliferation Assay and Transwell Invasion Assay
Cells were grown in 96well plates and transfected with siRNA (siPVT1-1/2). Subsequently, a Cell Counting Kit-8 (CCK8, Dojindo, Kumamoto, Japan) was used to detect cell proliferation at 0, 24, 48, 72, 96 and 120 h following the manufacturer’s protocol. For Transwell invasion assays, cells were inoculated into Transwell plates containing Matrigel. DMEM with 10% FBS was placed in the lower chamber. Serum-free DMEM was used to resuspend the cells, which were added to the upper chamber. After 24 h, 70% ethanol was used to fix the membranes, which were then stained using 1% crystal violet for 10 min. After thorough washing with PBS, the membranes were imaged. The average number of invasive cells was determined by counting in three randomly chosen fields under a microscope.
Luciferase Reporter Assay
The potential miRNAs binding to PVT1 and YAP1 were predicted using RNA22 (https://cm.jefferson.edu/rna22/Interactive/) and Targetscan (https://www.targetscan.org/vert72/). Luciferase reporter plasmids were reconstructed by inserting Target fragments were inserted into pmirGLO Dual-Luciferase miRNA Target Expression Vectors to construct dual-luciferase (firefly luciferase and renilla luciferase) reported plasmids. After culture in 48-well plates, cells were co-transfected with dual-luciferase reporters (pmirGLO-Wt1-PVT1, pmirGLO-Wt2-PVT1, pmirGLO-Mt1-PVT1, pmirGLO-Mt2-PVT1, pmirGLO-Wt-YAP1, pmirGLO-Mt-YAP1) and miR-375/205 mimics or miR-NC using lipo8000. Subsequently, a Dual-Glo Luciferase Assay System Kit (Beyotime) was used to detect dual-luciferase activity following the manufacturer's protocol. Renilla luciferase activity was used to normalize the firefly luciferase activity.
Comet Assay
Using a SCGE test Kit (Enzo Life Sciences, Farmingdale, NY, USA), single cell gel electrophoresis comet experiments were carried out. Following treatment, 75 µl aliquots of the cells were placed onto preheated slides at a volume ratio of 1:50 with low melting point agarose. The slides were incubated in pre-chilled lysis solution for 60 min, and then in pre-chilled alkaline solution for 30 min. The slides were then subjected to electrophoresis at 25 V in TBE buffer for 20 minutes. Comets were stained using CYGREEN dye for 30 min and then imaged. CASP Version 1.2.2 analytic tool was used to examine at least 50 distinct cells per sample in duplicate (CASPlab, Wroclaw, Poland).
RNA-Seq and Analysis
The Trizol reagent was used to isolate total RNAs. An Agilent Bioanalyzer 2100 (Agilent technologies, Santa Clara, CA, USA) was used to assess RNA quality. An RNAClean XP Kit (Beckman Coulter, Brea, CA,USA) and RNase-Free DNase Set (QIAGEN, Hilden, Germany) were used to purify the total RNA. An Illumina TruSeq® RNA sample preparation Kit (Illumina, San Diego, CA, USA) was used to prepare the sequencing libraries, and an Illumina HiSeq 2500 machine was used to pair-end sequence the libraries. GraphPad Prism 9.0 (GraphPad software, Inc., LA Jolla, CA, USA) was used to construct a heatmap. We deposited the raw sequencing data at GEO under the accession number GSE210387.
Statistical Analyses
The figure legends show the statistical parameters of the analyses. The in vitro experiments were carried out at least three times, and the in vivo experiments were carried out at least twice. GraphPad Prism 9.0 for windows (GraphPad software, Inc.) was used to carry out the statistical analyses. Comparisons between the data from two groups was performed using Student’s t test followed by Tukey’s HSD post hoc tests to minimize type I errors. The survival rates were calculated using the Kaplan–Meier method and analyzed using the log-rank test. For comparison of invasiveness in HNSCC, the differences were evaluated using the Cochran–Armitage test. Among the different treatment groups, the proportion of mice with lymph node metastasis was assessed using a χ2 test. Correlations were analyzed using the Pearson test. One-way analysis of variance (ANOVA) was used to compare the differences in HNSCC lesion size, number, and area in the control and treated mice. In the ANOVA analyses, the Shapiro–Wilk test was used to validate that the data were normally distributed and that all the data satisfied the assumption of no significant outliers. p < 0.05 was considered significant.