We purchased Paris saponin II (PSII) (purity > 99%) from Chengdu Must Bio-technology Technology Co., Ltd (Chengdu, China). The HNSCC cell lines including the hypopharyngeal cancer cell line FaDu and the laryngeal carcinoma cell lines Tu212 and Tu686 were purchased from American Type Culture Collection (ATCC; Rockville, MD, United States). Antibodies against Bcl-2, Ki67, nitric oxide synthase 3 (NOS3), CytC, Lc3b, β-actin, and Bax were purchased from Cell Signaling Technology Inc. (CST Inc., Danvers, MA, United States).

Two cell lines (Tu686 and FaDu) were cultured in Dulbecco’s modified Eagle medium (DMEM) which were obtained from Thermo Fisher Scientific Inc./Gibco (Waltham, MA, United States) supplemented with 10% fetal bovine serum (FBS) obtained from Haoyang Biological Manufacture Co. Ltd. (Tianjin, China), as well as penicillin/streptomycin (P/S; 100 U/ml/100 pg/ml). Tu212 cell lines were cultured in Iscove’s Modified Dulbecco’s Medium (IMDM) which was bought from Thermo Fisher Scientific Inc./Gibco supplemented with 10% FBS and P/S. All the cell lines were routinely cultured in a humidified cell incubator at 37°C with 5% CO₂.

We measured cell viability using the methyl thiazolyl tetrazolium (MTT) assay. FaDu, Tu686, and Tu212 cells were plated in 96-well plates at a cell concentration of 3 × 10³ cells/well. After overnight incubation for one night, the cultured medium was changed with a new medium containing different concentrations of PSII (0, 0.025, 0.05, 0.1, 0.2, 0.4, 0.8, and 1.6 μg/ml), and then we incubated them for 24, 48, 72, and 96 h at 37°C with 5% CO₂ in a cell incubator, Afterward, 20 µl of MTT reagent (5 mg/ml) was added to each well, and the cells were incubated for 4 h at 37°C. Subsequently, the supernatants in the wells were cautiously aspirated, and 100 µl of dimethyl sulfoxide (DMSO) solution was added to each well. The 96-well plates were shaken for 10 min at room temperature and then the absorbance values were measured at 570 nm on a microplate reader (Bio-Rad Laboratories Inc., Hercules, CA, United States).

We used transwell plates (Coming Costar, Lowell, MA, United States) to assay cell invasion and migration. In the migration assay, we starved the cells in a serum-free medium (SFM) for 12 h at 37°C with 5% CO₂. We then added 700 µl of DMEM or IMDM with 20% FBS to the lower well and 500 µl of SFM, including 1 × 10⁵ cells to the upper transwell inserts. After culturing for 48 h at 37°C with 5% CO₂, we counted the number of cells that adhered to the lower surface of the insert membrane. We performed the invasion assay in the same way, except that the transwell insert membrane was coated with Matrigel (BD Biosciences, San Jose, CA, United States).

We seeded the three HNSCC cell lines in six-well plates at a cell concentration of 1 × 10³ cells/well, and the plates were incubated for 24 h at 37°C with 5% CO₂. After incubation for 24 h, we replaced the old medium with a new medium containing PSII and continued incubation for 10 days. Subsequently, after washing cells with phosphate-buffered saline (PBS), the cells were fixed with methyl alcohol and stained with a 1% crystal violet solution. Ultimately, after washing the cells with PBS, we observed the cells under a microscope and acquired images of various fields to count the number of colonies.

Once cultured, the cells reached close to 100% confluence, the cell monolayer was mechanically scratched. Scratch healing was observed at 0, 12, 24, and 48 h under a microscope at ×40 magnification (white light bright field). We used ImageJ software [National Institute of Health (NIH), Bethesda, MD, United States] to calculate the migration area on the scratch, using the following formula: Scratch area rate (%) = 12, 24, and 48 h after migration scratch area/initial scratch area ×100%.

The corresponding SFM was prepared by adding double antibody as follows: 20 ng/ml epidermal growth factor (EGF; PeproTech US, Cranbury, NJ, United States), 20 ng/ml basic fibroblast growth factor (bFGF, Thermo Fisher Scientific Inc./Gibco), and 2% B27 (Invitrogen, Carlsbad, CA, United States) were added into DMEM/F12 culture medium (Thermo Fisher Scientific Inc./Gibco). Logarithmic growth phase HNSCC cell lines (FaDu, Tu212, and Tu686) were digested with 0.25% trypsin (Thermo Fisher Scientific Inc./Gibco) and dissociated into a single cell suspension. Then the resuspended cells were inoculated into SFM at a density less than 10⁴/ml for routine culture and added into an ultralow adsorption T25 or smaller glass flask, which was placed vertically in an incubator and shaken several times a day. The formation of microspheres was observed, and half of the cell suspension solution was replaced every 2–3 days, and the cells were subcultured every 6–8 days. During subculturing, microspheres were collected by centrifugation and resuspended in an accutase solution (MilliporeSigma, Burlington, MA, United States) and digested at 37°C for 5–10 min. The microspheres were pipetted up and down several times until they dispersed into single cells. After centrifugation and washing, the microspheres were counted and inoculated into SFM for subculture at a density less than 10⁴/ml.

Logarithmic growth phase tumor cells and stem cells were collected, digested, centrifuged, and resuspended in PBS. After counting, the number of cells was adjusted to 1 × 10⁶ cells/group. Anti-CD24-fluorescein isothiocyanate (FITC) was used as negative control, and the proportion of anti-CD44-phycoerythrin (PE)/anti-CD133-allophycocyanin (APC) cells in each group was determined by flow cytometry analysis. When the anti-CD44-PE and anti-CD133-APC were both positive, it revealed the presence of stem cells.

We selected the pathology of patients with hypopharyngeal malignant tumor for the first time as experimental subjects which was approved by the Ethics Committee at our institution. When rapid intraoperative pathology confirmed that the pathological type was squamous cell carcinoma of the head and neck, we collected tumor tissues for study. The head and neck tumor tissues were washed three times with cold PBS mixed with double antibodies in an ultraclean workbench. Then the tissue blocks were placed in a 10-cm sterile petri dish and cut into 1- to 2-mm tissue blocks with cold PBS as described to wash away the free cells. Subsequently, using tweezers, the tissue blocks were placed on the bottom surface of a culture flask with a spacing of about 0.5 cm. The flask was inverted, and 3 ml of the medium (DMEM+20%FBS + double antibody) was added to the bottom of the flask, and the tissue blocks were incubated for 4 h. Then 3 ml of the medium was added again, and the culture flask was inverted and incubated for an additional 48 h. The growth of primary tumor–associated fibroblasts was observed under a microscope, and the migration ability of the purified cells was examined by a transwell migration assay.

Tumor-associated fibroblasts were seeded into a petri dish with a pretreated cover glass and allowed to grow. After the cells had nearly formed a monolayer, the cover glass was removed, and the cells were washed twice with PBS. Then cells were fixed for 2 h with paraformaldehyde, permeabilized with 5% Triton X-100 for 20 min, blocked with 2% BSA, and washed twice with PBS. The cells were treated with the corresponding primary antibody against FAP (CST Inc.) and SMA (Abcam, Cambridge, United Kingdom), followed by incubation with appropriate secondary antibodies. Ultimately, all fibroblasts were stained with 4’,6-diamidino-2-phenylindole (DAPI).

We performed untargeted metabolomics analysis of the three HNSCC cell lines treated with/without PSII. The metabolites in the samples were analyzed by ultrahigh performance liquid chromatography–quadrupole time-of-flight mass spectrometry (UHPLC-Q-TOF MS). The retention time and molecular weight of the metabolites in the samples were compared with those in the local database. Also, the 25 PPM, secondary fragmentation spectrum, collision energy, and other information were matched to identify the structure of metabolites in biological samples, and the identification results were confirmed through strict manual secondary check and confirmation. The assessment level is Level 2.

FaDu and Tu686 cells were treated with PSII (0, 0.05, and 0.1 μg/ml) for 48 h at 37°C with 5% CO₂. Cells were collected and lysed in lysis buffer with protease inhibitors to extract proteins. We separated proteins by 12% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE). Then the separated proteins were electroblotted onto a polyvinylidene difluoride (PVDF) membrane using a blotting apparatus (Bio-Rad Laboratories). After blocking the membranes with 5% non-fat milk, the membranes were incubated overnight at 4°C with the corresponding primary antibodies, followed by labeling with the appropriate secondary antibody at room temperature for 1 h. We used β-actin as the internal control. Eventually, the immunoreacted protein bands were visualized using the Pro-lighting horseradish peroxidase (HRP) agent for Western blotting detection (Tiangen Biotech Co., Ltd., Beijing, China). The Western blot analysis of mouse xenograft tumors was performed in the same way as that for the cells.

Total RNA was isolated from FaDu and Tu686 cells using TRIzol (Invitrogen), and the quality of the isolated RNA by the absorbance at 260 and 280 nm was determined. cDNA synthesis was carried out using Rever-tAidTM M-MuLV RT (Fermentas, Hanover, MD, United States) according to the instructions, and the obtained cDNA was stored at −80°C. Polymerase chain reactions were performed in a final volume of 50 ml including 2 µl each of the forward and reverse primers, 25 µl 2*Taq Master Mix, 2 µl cDNA, and RNase-free water. After transfection, total RNA was obtained from the transfected cells with TRIzol, and cDNA was obtained by reverse transcription as described. Real-time quantitative PCR was performed to detect the mRNA level.

The primers used were synthesized by Ding Guo Changsheng Biotechnology Co., Ltd. (Beijing, China). The NOS3-F primer sequence is GTG​ATG​GCG​AAG​CGA​GTG​AAG​G, and the NOS3-R primer sequence is CAC​CAC​GTC​ATA​CTC​ATC​CAT​ACA​CAG. The FaDu and Tu686 cells were inoculated in six-well plates, and when confluence reached about 85%, siRNA targeting NOS3 (NOS3-siRNA) was diluted with serum-free DMEM and mixed gently with Lipofectamine 2000 (Biyuntian Biological Technology Co., Shanghai, China). After mixing, the mixture was incubated for 20 min at room temperature. Afterward, FaDu and Tu686 cells were plated in culture plates to transfect for 6 h and then cultured in a complete medium for 48 h.

A total of 5 × 10⁶ cells in 0.2 ml of PBS were injected subcutaneously into 4-week-old BALB/c nude female mice purchased from Taconic Biosciences (Rensselaer, NY, United States) After 5 days, the mice were randomly assigned to two groups, namely, the PSII group and control group. The PSII group was administered with about 100 µl of solution (5 mg/kg) by intraperitoneal injection once every 3 days. We measured the tumor growth once every 5 days according to the following equation: Volume = (width² × length)/2. On the 60th day, mice were anesthetized with ether and then killed. Eventually, the xenograft tumors were removed and processed for immunohistochemical (IHC) analysis.

The spleen, lung, liver, heart, and kidney tissues of the mice were fixed in 10% formalin for the histopathological examination. After the dehydration process, we embedded the tissues in paraffin wax and cut them into 5-µm-thick sections using microtome. Then we stained the sections with hematoxylin and eosin (H&E). Ultimately, H&E-stained histopathology images were acquired using an Olympus microscope (Olympus Corporation, Tokyo, Japan).

First, we cut the xenografts into 5-µm-thick sections using microtome. Then the slices were labeled and placed in an oven at 60°C for 2 h. After dewaxing the paraffin sections, the endogenous peroxidases of the sections were blocked. Subsequently, the sections were boiled in 10 mM sodium citrate plus 0.05% Tween-20 (pH 6) for 10 min in a microwave oven for antigen retrieval. The tissue sections were uniformly covered with 5% BSA and sealed at 37°C for 30 min. Then the sections were incubated with the corresponding primary antibody for 12 h at 4°C in the dark. Afterward, the sections were washed three times with PBS and incubated with the appropriate secondary antibody (1:200) for 60 min at room temperature. Then the sections were stained using a diaminobenzidine (DAB) HRP Color Development Kit (Beyotime Biotechnology, Shanghai, China) according to the manufacturer’s instructions. Next, the sections were stained with hematoxylin to stain the nucleus, dehydrated, and covered with neutral resin. Eventually, images of the sections were acquired on an Olympus microscope.

SPSS (version 20.0) software (IBM Corporation, Armonk, NY, United States) was used for statistical analysis. One-way analysis of variance (ANOVA) and Student’s t-test were performed to evaluate the significance. Every test was performed three times, and the error bars represent the standard error of the mean (SEM) in which differences with p < 0.05 were considered significant.

As shown in Figure 1A, the molecular formula of PSII is C51H82O20, and its molecular weight is 1015.18. PSII has a white powder appearance, and its structural formula is as shown in Figure 1B. Chromatographic separation at a column temperature of 40°C, using a MinXi Teeh CG-C18 5 µm 250 × 4.6-mm chromatography column, with methyl alcohol as sample solvent, resulted in four peaks.

The measurement of cell proliferation by the MTT assay, as shown in Figures 2A–C, revealed that treatment with PSII significantly inhibited the proliferation of FaDu, Tu686, and Tu212 cells in a time- and concentration-dependent manner, especially after 48 h or at a concentration of 0.1 μg/ml PSII. In addition, the colony formation assay also indicated that PSII inhibited the proliferation of FaDu, Tu686, and Tu212 cells (Figures 2D–I).

The analysis of tumor cell motility by transwell assays and cell scratch assays revealed, as shown in Figure 3, that the PSII extract inhibited migration and invasion in FaDu, Tu686, and Tu212 cells. In particular, the cell scratch assay results in Figure 4 showed the inhibitory effect of PSII on these three HNSCC cell lines, and cell motility clearly decreased after treatment with 0.1 μg/ml PSII.

We prepared SFM by adding 20 ng/ml bFGF, 20 ng/ml EGF, and 2% B27. After subculturing four times in this SFM, we calculated the pelletizing rate of tumor cells (≥100 μm), as shown in Figure 5B. Anti-CD24-FITC was used as a negative control, and the proportion of anti-CD44-PE/anti-CD133-APC cells in each group was determined by flow cytometry. The results shown in Figures 5C,D indicated that PSII suppressed the stemness of carcinoma cell stem microspheres.

After we successfully isolated primary tumor–associated fibroblasts, we used immunofluorescence analysis to phenotypically identify these cells. We measured the viability of the primary tumor–associated fibroblasts by the MTT assay. As shown in Figure 6C, treatment with PSII significantly reduced the viability of primary tumor–associated fibroblasts. In addition, the results of the migration assay shown in Figure 6D demonstrated the inhibitory effect of PSII on the migration ability of primary tumor–associated fibroblasts.

Principal component analysis (PCA) performed on the peaks extracted from all experimental samples revealed, as shown in Figures 10A,B, that the experiments had good repeatability, as indicated by their close clustering together in positive and negative ion modes. All the metabolite tests in this study (identified with positive and negative ions) were classified and counted according to the attribution information of their chemical taxonomy. The proportions of various metabolites are shown in Figures 10C,D. The histograms used to visually display the identified multiple changes of significantly differentially expressed metabolites, and multiple analyses of significantly differentially expressed metabolites in positive and negative ion modes are shown in Figures 10E,F. The abscissa in the histogram graphs represents the differential expression multiple, red color represents the differential expression multiple greater than 1, and the green color represents the differential expression multiple less than 1. The ordinate indicates the significantly different metabolites.

Treatment of FaDu and Tu686 cells with PSII (0, 0.025, 0.05, and 0.1 ug/ml) for 48 h revealed, as shown in Figure 7, that the mRNA and protein expression levels of genes involved in the mitochondrial pathway (CytC), cell autophagy (Lc3b), and apoptosis (Bax, Bcl-2, and caspase-9) were all altered by treatment with PSII. We used the Chinese Medicine Data website (http://herb.ac.cn/) to predict the target of PSII in cancer, and the result suggested many targets. Considering the metabolite results, we supposed that NOS3 may be a target. Additionally, NOS3 knockdown experiments using siRNA against NOS3 in HNSCC cells confirmed the effects of NOS3. Moreover, Western blot analysis confirmed that the protein expression levels of Bax, CytC, and Lc3b were decreased and that of Bcl-2 was increased when NOS3 expression was knocked down in FaDu and Tu686 cells (Figures 8C,D). These results confirmed that NOS3 is a downstream target of PSII in HNSCC.

We studied the inhibitory effect of PSII in an in vivo mouse tumor, which we developed by injecting Tu686 HNSCC cells into both sides of the flank of BALB nude mice. As shown in Figures 9A,B, the volume of xenograft tumor growth was significantly reduced after PSII treatment compared with the control group. On the 60th day, as shown in Figure 9B, the average volume of tumors in the PSII group (46.89 ± 3.99 mm³) was smaller than that in the control groups (273.71 ± 10.07 mm³), and there was no statistical difference in weight between the two groups. Our analysis of the protein levels of NOS3, Bcl-2, and Ki67 by IHC staining indicated, as shown in Figure 9C, that treatment with PSII led to decreased protein levels of Bcl-2 and Ki67 and an increased protein level of NOS3. Furthermore, the histopathological examination showed that PSII had no adverse effects on the lung, heart, spleen, liver, and kidney. Additionally, the Western blot analysis of mouse tumors showed that the inhibitory effect of PSII on the growth of subcutaneous tumors is mediated through the mitochondrial pathway, cell autophagy, and apoptosis. All the in vivo results indicated that PSII induced autophagy and apoptosis in HNSCC cells.