The human NPC cell lines HONE1 (Mingzhoubio, MZ-1647, Ningbo, China) and CNE2 (Immocell, IM-H136, Xiamen, China) were cultured in DMEM (Procell, PM150210, Wuhan, China) with 10% fetal bovine serum (Procell, 164210-500, Wuhan, China) and 1% penicillin-streptomycin double antibiotic (Beyotime, C0222, Shanghai, China) in a cell incubator at 37 °C and 5% CO${}_2$. All cell lines were validated by STR profiling and tested negative for mycoplasma.

BeyoClickTM EdU-488 (Beyotime, C0071S, Shanghai, China) was used for the proliferation assay. HONE1 and CNE2 cells were seeded in 96-well plates (5 $\times$ 10${}^4$ cells/well) and treated with silvestrol (Medchemexpress, HY-13251, Monmouth Junction, NJ, USA) after 24 h of culture. After another 24 h, EdU was added to each well and the cells kept at 37 °C for 2 h. Cells were then fixed with 4% paraformaldehyde (Biosharp, BL539A, Hefei, China) for 30 minutes, and immunofluorescence microscopy (Leica, DMi8, Wetzlar, Germany) was used to observe and record images. The percentage of EdU-positive cells was calculated by dividing the number of EdU-positive cells by the total DAPI-stained cell count. The ratio for the control group was set to 1 for reference.

Cells were cultured for 24 h until 90% confluence was achieved. The monolayer was scratched with the tip of a pipet gun and the exfoliated cells removed. Cells were then cultured in serum-free medium containing silvestrol with or without LY3214996 (Medchemexpress, HY-101494, Monmouth Junction, NJ, USA). The cell migration rate was calculated as follows: (Initial Scratch Width – Scratch Width after 24 h)/Initial Scratch Width $\times$ 100%

A total of 1.5 $\times$ 10${}^5$ cells/well were seeded in 6-well plates. After 24 h, silvestrol with or without LY3214996 was added to the medium of the experimental group. Cells were then harvested by trypsin (Beyotime, C0207, Shanghai, China) digestion and washed with PBS. Staining was performed using the Annexin V-FITC/PI apoptosis kit (MULTISCIENCES, 70-AP101-100, Hangzhou, China), and Annexin V-FITC and propidium iodide (PI) signals were measured by flow cytometry (Beckman, CytoFLEX LX, Miami, FL, USA).

Silvestrol with or without LY3214996 was added to the medium of the experimental group. Trypsin digestion was carried out to harvest cells, followed by PBS washing and subsequent fixation in 75% alcohol for 24 h. Staining procedures were performed with PI/RNase Staining Buffer (BD Biosciences, 550825, San Diego, CA, USA).

Cells were lysed in RIPA lysis buffer (Beyotime, P0013J, Shanghai, China) and 1% PMSF (Solarbio, P0100, Beijing, China) and then centrifugated. The extracted proteins were denatured with SDS-PAGE protein loading buffer (Beyotime, P0015, Shanghai, China) (100 °C, 10 min) and separated by 11% SDS-PAGE gel electrophoresis (Beyotime, P0903, Shanghai, China). Proteins were then transferred onto PVDF membranes (Beyotime, FFP22, Shanghai, China) and blocked with a 5% BSA (Beyotime, ST023, Shanghai, China) solution. Incubation with primary antibodies was carried out overnight at 4 °C, followed by subsequent incubation with horseradish peroxidase-labeled secondary antibodies for 1.5 h. The ultra-sensitive multifunctional imager (Cytiva, Amersham ImageQuant 800, Shanghai, China) was used to visualize protein bands. The following antibodies (Affinity, Wuhan, China) were used: $\beta$-actin (Affinity, AF7018), Cleaved Caspase3 (Affinity, AF7022), p-ERK1/2 (Affinity, AF1015), ERK1/2 (Affinity, AF0155), p-MEK (Affinity, AF8035), MEK (Affinity, AF6385), GADD45A (Affinity, DF6622), RAP1A (Affinity, DF6157), HK2 (Affinity, DF6176), and Goat Anti-Rabbit IgG (H+L) HRP (Affinity, S0001). We employed Image J (NIH, Pro Plus 6.0, MD, USA) to quantify the band intensities. The relative density was calculated by dividing the density of the target band by the density of $\beta$-actin. The relative density of the control group was set to 1 for normalization purposes.

The age range of the patients providing the pathological tissue sections is 50-60 years old. The providers of tissues #1 and #3 are male, while the provider of tissue #2 is female. Tissue sections (4 µm) (Collection date: #1 - June 2022, #2 - November 2022, #3 - May 2022) were incubated at 60 °C for 20 minutes before undergoing xylene (Biofount, 1330-20-7, Beijing, China) deparaffinization and ethanol rehydration. Antigen retrieval was performed in 10 mM citric acid solution (Beyotime, P0081, Shanghai, China) and involved microwave pretreatment (15 min) and subsequent cooling. Endogenous peroxidase activity was quenched with H${}_2$O${}_2$ (ZSGB, PV-9001, Beijing, China). Sections were incubated with 5% normal goat serum (Bioss, C0005, Bethesda, MA, USA) for 1 h at room temperature, followed by overnight incubation with primary antibodies (Affinity, DF6622/ DF6157/ DF6176, Wuhan, China) at 4 °C. The next day, goat anti-rabbit IgG polymer labeled with enhanced enzyme (ZSGB, PV-9001, Beijing, China) was applied and incubated at 37 °C for 30 minutes. Color development was performed using the DAB color reagent kit (ZSGB, ZLI-9018, Beijing, China), while nuclear staining was performed with a 0.1% nuclear solid red staining solution (Solarbio, G1320, Beijing, China). Sections were dehydrated, made transparent, and mounted with a coverslip.

CNE2 and HONE1 cells were separated into two experimental groups: silvestrol (20 nM), and silvestrol (20 nM) + LY3214996 (80 nM). Cell samples were frozen in TRIzol reagent (Solarbio, 15596-018, Beijing, China) and submitted to Geneplus Technology Corporation (Shenzhen, China) for RNA sequencing. Any mRNA with a fold-change $|$FC-1$|$ $>$0.5 and p 0.01 was considered to be a differentially expressed gene (DEG). DEGs were further screened based on The Comparative Toxicogenomics Database (CTD) and transcriptome information using Venn diagram analysis. Hub genes showing the same expression trend in both cell lines were identified and categorized into up- or down-regulated groups. These up-regulated and down-regulated DEGs were then analyzed by Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis and protein interaction network analysis using STRING Version 9.1 (http://www.string-db.org).

Prism (GraphPad Software, GraphPad 8.0.1, San Diego, CA, USA) was used to compare groups. All data were expressed as the mean $\pm$ standard deviation (SD). The unpaired t-test was used to determine statistically significant differences between two groups. A p-value 0.05 was considered to indicate a statistically significant difference.

Flow cytometry revealed that 20 nM silvestrol significantly increased the apoptosis rate in HONE1 and CNE2 NPC cells (*p 0.001) (Fig. 1A,B). The apoptosis rate of HONE1 and CNE2 cells also increased after treatment with 2 nM silvestrol, but this increase was only statistically significant for HONE1 cells (*p = 0.025) (Fig. 1B). Western Blot (WB) assay showed that the expression of cleaved caspase-3 protein in CNE2 and HONE1 cells increased with increasing concentrations of silvestrol. Specifically, the expression in the 2 nM silvestrol group was significantly higher than in the control group with 0 nM silvestrol (*p 0.001; Fig. 1C, 0 nM vs 2 nM, *p = 0.005). The expression of cleaved caspase-3 after 20 nM silvestrol treatment was significantly higher compared to 0 nM silvestrol (*p 0.001) and 2 nM silvestrol (#p 0.001) (Fig. 1C,D). Treatment with 20 nM silvestrol significantly increased the proportion of G1 phase cells more than treatment with 2 nM (#p 0.001) and 0 nM silvestrol (*p 0.001) (Fig. 1E,F). The proportion of cells in S phase and G2 phase after treatment with 2 nM silvestrol either decreased or remained unchanged, with only the decrease in G2 phase for CNE2 cells showing statistical significance (*p = 0.018) (Fig. 1E,F). The proportion of CNE2 and HONE1 cells in the G1 and G2 phases following treatment with 20 nM silvestrol was significantly higher than after treatment with 0 nM silvestrol (*p 0.001) or 2 nM silvestrol (#p 0.001) (Fig. 1E,F). After treatment with 20 nM silvestrol, the proportion of CNE2 and HONE1 cells in S phase was significantly lower than with 0 nM silvestrol (*p 0.001) or 2 nM silvestrol (#p 0.001) (Fig. 1E,F). The EdU assay showed that 20 nM silvestrol significantly reduced the relative cell proliferation ratio of CNE2 and HONE1 cells compared with 0 nM (*p 0.001) and 2 nM silvestrol (#p 0.001). No significant difference was observed between 2 nM and 0 nM silvestrol (Fig. 1G,H). The above results indicate that silvestrol promotes apoptosis in NPC cells and inhibits their proliferation. Moreover, these effects are dose-dependent.

Fig. 1.

Silvestrol promotes apoptosis and inhibits proliferation of nasopharyngeal carcinoma (NPC) cells. NPC cells were treated with silvestrol at different concentrations (0 nM, 2 nM, 20 nM) for 24 h. (A,B) Apoptosis was determined by flow cytometry in CNE2 and HONE1. (C,D) Western blot (WB) determined cleaved caspase-3 expression in CNE2 and HONE cells. (E,F) Cell cycle was determined by flow cytometry in CNE2 and HONE1. (G,H) Cell proliferation was measured by 5-ethynyl-2${}^{\prime }$-deoxyuridine (EdU) in CNE2 and HONE1. The values are performed as mean $\pm$ standard deviation (SD), N = 3/group. * p 0.05 vs 0 nM, # p 0.05 vs 2 nM. Bar = 100 µm.

The results of the Western Blot (WB) assay showed that p-ERK and p-MEK protein levels were significantly increased in CNE2 and HONE1 cells between 0.5 h and 2 h after silvestrol treatment (*p 0.001) (Fig. 2A,B). A selective inhibitor of ERK, LY3214996, reversed the silvestrol-induced phosphorylation of ERK in NPC cells. Treatment with 20 nM silvestrol significantly increased p-ERK in CNE2 and HONE1 cells compared to the control (*p 0.001) (Fig. 2C,D). Treatment with 80 nM LY3214996 in combination with 20 nM silvestrol significantly reduced the level of p-ERK protein compared to treatment with silvestrol alone (#p 0.001; Fig. 2D).

Fig. 2.

Silvestrol activates the ERK pathway by promoting ERK phosphorylation. (A,B) WB detection for p-ERK/ERK, andp-MEK/MEK protein expression in CNE2 and HONE1 cells. Silvestrol for 0 h, 0.5 h, 1 h and 2 h. The values are performed as mean $\pm$ SD, N = 3/group, * p 0.05 vs 0 h. (C,D) WB detection for p-ERK1/2 and ERK1/2 in CNE2 and HONE1 cells. NPC cells were treated with silvestrol and/or LY3214996 for 1h (CON, control group without drug; S, single-drug group with 20 nM silvestrol; S + L, two-drug group with 20 nM silvestrol and 80 nM LY3214996), The values are performed as mean $\pm$ SD, N = 3/group. * p 0.05 vs CON, # p 0.05 vs S. ERK, extracellular signal-regulated kinase; MEK, mitogen-activated protein kinase kinase.

Treatment with 20 nM silvestrol increased the apoptosis rate of HONE1 and CNE2 cells significantly (*p 0.001) (Fig. 3A,B). Co-treatment with 80 nM LY3214996 (an ERK pathway inhibitor) and 20 nM silvestrol significantly reduced the proportion of apoptotic cells compared to silvestrol treatment alone (#p 0.001; Fig. 3A, # p = 0.007) (Fig. 3A,B). Similarly, Western Blot (WB) results showed that 20 nM silvestrol significantly increased the expression of cleaved caspase-3 (*p 0.001) (Fig. 3C,D). Treatment of CNE2 and HONE1 cells with 80 nM LY3214996 and 20 nM silvestrol significantly reduced cleaved caspase-3 expression compared to silvestrol treatment alone (#p 0.001) (Fig. 3C,D). Flow cytometry revealed that 20 nM silvestrol significantly increased the proportion of HONE1 and CNE2 cells in the G1 phase compared with the control (*p 0.001), and significantly decreased the proportion of cells in the S phase and G2 phase (*p 0.001), except for CNE2 cells in the S phase, which did not show a significant decrease (Fig. 3E,F). Co-treatment with 80 nM LY3214996 and 20 nM silvestrol significantly reduced the proportion of CNE2 and HONE1 cells in the G1 phase (*p 0.001) and significantly increased the proportion of G2 phase cells (#p 0.001), indicating a reversal of silvestrol’s cell cycle arrest effect (Fig. 3E,F). The statement about a trend for increased S phase cells compared with silvestrol alone needs clarification; it implies that co-treatment with LY3214996 might modulate the cell cycle differently, but specific data or a statistical value should be provided for a clear understanding. In the cell migration assay, although the initial scratch width was almost the same for each group at 0 h, the group treated with LY3214996 and silvestrol showed a wider scratch width after 19 h and 24 h, indicating inhibited cell migration compared with the control and the silvestrol alone groups (Fig. 3G). This result suggests that the ERK pathway might play a role in cell migration, which silvestrol can influence, and that LY3214996 can modulate this effect.

Fig. 3.

Silvestrol acts on NPC cells through the ERK pathway. (A,B) Apoptosis was determined by flow cytometry in CNE2 and HONE1. (C,D) WB detection for cleaved caspase-3 expression in CNE2 and HONE cells. (E,F) Cell cycle was determined by flow cytometry in CNE2 and HONE1. (G) Cell migrations of HONE1 in response to different drug treatments. NPC cells were treated with silvestrol and/or LY3214996 for 1h (CON, control group without drug; S, single-drug group with 20 nM silvestrol; S + L, two-drug group with 20 nM silvestrol and 80 nM LY3214996), Bar = 300 µm; The values are performed as mean $\pm$ SD, N = 3/group. * p 0.05 vs CON, # p 0.05 vs S.

A total of 960 DEGs, related to both CNE2 and HONE1 cells as well as broader NPC research available in databases like the Comparative Toxicogenomics Database (CTD) (http://ctdbase.org/), were identified through Venn diagram analysis (Fig. 4A). Of these, 572 hub DEGs displayed consistent expression trends across CNE2 and HONE1 cells, comprising 303 down-regulated genes and 269 up-regulated genes. Protein network analysis of the up-regulated genes highlighted two protein interaction networks, with RAP1A protein identified as a core element in both (Fig. 4B,C). Furthermore, the tight junction pathway, involving RAP1A, emerged as the top-scoring pathway in the KEGG pathway analysis, indicating significant changes in pathways associated with the up-regulated DEGs (Fig. 4D). For the down-regulated genes, hexokinase-II (HK2), PFKM, PGAM1, and PKM proteins were all pivotal members of their respective protein interaction network (Fig. 4E). These proteins are integral to carbon metabolism and glycolysis/gluconeogenesis pathways (Fig. 4G), which were highlighted as the top three pathways in KEGG pathway enrichment analysis (Fig. 4F).

Fig. 4.

Transcriptome differences in NPC cells treated with silvestrol and with or without ERK inhibitors. (A) Venn diagram of Comparative Toxicogenomics Database (CTD) database, TCGA and RNA-seq. (B,C) The protein-protein interaction (PPI) network of up-regulated hub DEGs ($|$FC-1$|$ $\ge$0.5). (D) KEGG pathway enrichment of up-regulated hub DEGs. (E) The PPI network of down-regulated hub DEGs ($|$FC-1$|$ $\ge$0.5). (F) KEGG pathway enrichment of down-regulated hub DEGs. (G) The DEG network of key pathways. TCGA, The Cancer Genome Atlas; RNA-seq, RNA sequencing; DEGs, differentially expressed genes; KEGG, Kyoto Encyclopedia of Genes and Genomes; FC, fold change.

Treatment with 20 nM silvestrol significantly decreased RAP1A, HK2, and GADD45A protein expression in CNE2 and HONE1 cells compared to the control group (*p 0.001; Fig. 5A: *pRAP1A = 0.032, *pHK2 = 0.008, *pGADD45A = 0.009) (Fig. 5A,B). Co-treatment with 80 nM LY3214996 (an ERK inhibitor) and 20 nM silvestrol significantly increased the expression of RAP1A, HK2, and GADD45A proteins compared with silvestrol treatment alone (Fig. 5A: #pHK2 = 0.016, #pGADD45A = 0.002; Fig. 5B: #pRAP1A = 0.006, #pGADD45A = 0.009), although the increases in HK2 and GADD45A expression showed a non-significant upward trend in HONE1 cells (Fig. 5A,B). Immunohistochemistry (IHC) revealed that RAP1A is localized in the nucleus (Fig. 5C). In tumor sections from patients #1, #2, and #3, RAP1A-positive multinucleated and nuclear heterogeneous cells were observed (Fig. 5C). Hematoxylin and eosin (HE) staining showed no positive signal for RAP1A in non-tumor tissue, and a clear boundary was visible between the tumor and non-tumor regions (Fig. 5C). HK2 protein localization was identified in the cytoplasm (Fig. 5C). In tumor sections from patients #1, #2, and #3, the positive signal for HK2 around abnormal nuclei was stronger than that around normal nuclei (Fig. 5C). HE staining showed no positive signal for HK2 in the non-tumor tissue of patient #1, while a weak positive signal was observed in the non-tumor tissue of patients #2 and #3. GADD45A was detected in both the nucleus and cytoplasm (Fig. 5C), with positive areas showing an obvious boundary in the tumor sections of patients #1 and #2. Combined with HE staining and nuclear morphology, the highly positive area corresponded to the cancer nest tissue. No GADD45A positive signal was found in the non-tumor tissue of patient #1. In patient #2, the GADD45A positive signal in the non-tumor tissue was weaker than in the tumor tissue. The nuclei in patient #3 did not show GADD45A positive signal, and there was no clear boundary between the positive signal in the cancer and the non-tumor tissue.

Fig. 5.

Target genes: the effect of silvestrol on protein expression in NPC patient tissue. NPC cells were treated with silvestrol with or without LY3214996 for 1 h (CON, control group; S, 20 nM silvestrol; S + L, 20 nM silvestrol and 80 nM LY3214996). The expression level of GADD45A, RAP1A and HK2 proteins was evaluated by WB assay in (A) CNE2 cells, and (B) HONE1 cells. (C) H&E tissue sections from NPC patients, and IHC of GADD45A, RAP1A and HK2. Red arrows, tumor cells; Green arrow, non-tumor cells. Bar = 50 µm. The values are performed as mean $\pm$ SD, N = 3/group. *p 0.05 vs CON, #p 0.05 vs S. H&E, hematoxylin & eosin; IHC, immunohistochemistry; GADD45A, growth arrest and DNA damage inducible alpha; RAP1A, Ras-related Protein 1A; HK2, hexokinase 2.