Ceramide promotes lytic reactivation of Epstein-Barr virus in gastric carcinoma

ABSTRACT

Epstein-Barr virus (EBV) has a lifelong latency period after initial infection. Rarely, however, when the EBV immediate early gene BZLF1 is expressed by a specific stimulus, the virus switches to the lytic cycle to produce progeny viruses. We found that EBV infection reduced levels of various ceramide species in gastric cancer cells. As ceramide is a bioactive lipid implicated in the infection of various viruses, we assessed the effect of ceramide on the EBV lytic cycle. Treatment with C₆-ceramide (C₆-Cer) induced an increase in the endogenous ceramide pool and increased production of the viral product as well as BZLF1 expression. Treatment with the ceramidase inhibitor ceranib-2 induced EBV lytic replication with an increase in the endogenous ceramide pool. The glucosylceramide synthase inhibitor Genz-123346 inhibited C₆-Cer-induced lytic replication. C₆-Cer induced extracellular signal-regulated kinase 1/2 (ERK1/2) and CREB phosphorylation, c-JUN expression, and accumulation of the autophagosome marker LC3B. Treatment with MEK1/2 inhibitor U0126, siERK1&2, or siCREB suppressed C₆-Cer-induced EBV lytic replication and autophagy initiation. In contrast, siJUN transfection had no impact on BZLF1 expression. The use of 3-methyladenine (3-MA), an inhibitor targeting class III phosphoinositide 3-kinases (PI3Ks) to inhibit autophagy initiation, resulted in reduced beclin-1 expression, along with suppressed C₆-Cer-induced BZLF1 expression and LC3B accumulation. Chloroquine, an inhibitor of autophagosome-lysosome fusion, increased BZLF1 protein intensity and LC3B accumulation. However, siLC3B transfection had minimal effect on BZLF1 expression. The results suggest the significance of ceramide-related sphingolipid metabolism in controlling EBV latency, highlighting the potential use of drugs targeting sphingolipid metabolism for treating EBV-positive gastric cancer.

IMPORTANCE

Epstein-Barr virus remains dormant in the host cell but occasionally switches to the lytic cycle when stimulated. However, the exact molecular mechanism of this lytic induction is not well understood. In this study, we demonstrate that Epstein-Barr virus infection leads to a reduction in ceramide levels. Additionally, the restoration of ceramide levels triggers lytic replication of Epstein-Barr virus with increase in phosphorylation of extracellular signal-regulated kinase 1/2 (ERK1/2) and CREB. Our study suggests that the Epstein-Barr virus can inhibit lytic replication and remain latent through reduction of host cell ceramide levels. This study reports the regulation of lytic replication by ceramide in Epstein-Barr virus-positive gastric cancer.

INTRODUCTION

Gastric cancer is one of the most common cancers worldwide, and approximately 10% of all gastric cancers are Epstein-Barr virus (EBV) positive (1). The 2014 TCGA Gastric Cancer Study designated EBV-positive gastric cancer as one of four subtypes of gastric cancer (2). EBV has been reported to cause a variety of cancers, including gastric cancer, nasopharyngeal cancer, Burkitt’s lymphoma, and Hodgkin’s lymphoma (3). In most tumors, EBV maintains latency, and studies are being conducted to induce the lytic cycle for tumor treatment.

Ceramide is a key lipid at the center of mammalian sphingolipid metabolism (4). Ceramides are metabolized and reversibly transformed into various sphingolipids such as sphingomyelin (SM), ceramide-1-phosphate, sphingosine-1-phosphate (S1P), and glycosyl ceramide (GlyCer) including glucosyl ceramide (GlcCer) and galactosyl ceramide (GalCer) (4). In the salvage pathway, the ceramide pool generated from the breakdown of sphingolipids and glycolipids is hydrolyzed by ceramidase (CDase) to sphingosine and free fatty acids (4). Long-chain sphingoid bases can be recycled back to ceramide by ceramide synthase (CERS) or phosphorylated by sphingosine kinase (SphK) to form S1P (4).

Changes in the levels of sphingolipids, including ceramide, have been reported to be associated with infection with various viruses, including the Zika virus, hepatitis C virus, and influenza A virus (IAV) (5–7). According to a recent study, the tegument protein pUL21 of herpes simplex virus-1 belonging to the alpha herpes virus converts ceramide to sphingomyelin by dephosphorylating and activating ceramide transport protein (CERT) (8). Infection with human cytomegalovirus (HCMV), which belongs to the beta herpesvirus family, increases the accumulation and activity of SphK, an enzyme that produces S1P and dihydrosphingosine-1-phosphate (9). Furthermore, treatment with exogenous S1P during SphK inhibition restored HCMV replication (9). In primary effusion lymphoma (PEL) cells infected with Kaposi’s sarcoma-associated herpesvirus (KSHV), which belongs to the same family of gamma herpes viruses as EBV, treatment with the SphK2-specific inhibitor ABC294640 can induce intracellular ceramide accumulation. In addition, apoptosis of the PEL cells was induced through the activation of KSHV lytic gene expression (10). In a study using an EBV-positive Burkitt’s lymphoma cell line, Mutu-I, ceramide treatment did not induce EBV lytic replication (11). However, the effect of sphingolipid on viral replication or infection in EBV-positive gastric cancer has not been investigated.

In a previous study (12), we revealed the metabolic landscape of EBV-associated gastric cancer through a comprehensive multi-omics analysis using genetically identical gastric cancer cell lines, AGS and AGS-EBV, which differ only in infection with EBV. In the present study, our lipidomic analysis showed an overall reduction of various ceramide species in AGS-EBV. We also investigated whether restoration of the level of ceramide, which was shown to be reduced in an EBV-infected gastric cancer cell line, could modulate the EBV lytic cycle.

RESULTS

C₆-ceramide (C₆-Cer) induced EBV lytic reactivation

We performed mass spectrometry analysis using AGS and AGS-EBV cell pellets from different passages of the same clone. Mass spectrometry analysis of sphingolipid showed that various ceramide species were significantly reduced in AGS-EBV compared to AGS (Fig. 1A). The ceramide species altered after EBV infection are specified based on the number of carbons and double bonds (14:0–28:4) of the fatty acid attached to the sphingosine base (d18:1) (Fig. 1A). We investigated whether ceramide could induce EBV lytic gene expression in EBV-positive gastric cancer cells, similar to KSHV (10). Exogenous C₆-Cer can be used to increase the intracellular ceramide pool through a salvage pathway (13). Since the inhibitory concentration (IC)₅₀ values of C₆-Cer in AGS and AGS-EBV were 5.6 and 4.3 µM, respectively, 5 µM C₆-Cer was used in further analysis (Fig. 1B). The IC₅₀ values for C₆-Cer in naturally EBV-infected gastric cancer cells were 10 µM in SNU-719 and 15 µM in YCCEL1, as depicted in Fig. 1B. Additionally, treatment with 5 µM C₆-Cer resulted in reduced cell viability in the immortalized normal gastric epithelial cell line HFE-145 (Fig. 1B). The reduction in cell viability of HFE-145 by 5 µM C₆-Cer was lower than that of AGS and AGS-EBV but was similar to that of SNU-719 and YCCEL1. Western blotting was used to investigate whether C₆-Cer treatment could induce the lytic cycle in AGS-EBV (Fig. 1C). An increase in the expression of the EBV immediate early gene BZLF1 was observed in AGS-EBV cells following 24-h treatment with C₆-Cer, and this effect was further increased after 48 h (Fig. 1C). Accumulation of the autophagosome marker LC3B known to be increased by C₆-Cer treatment was also observed (Fig. 1C). Moreover, treatment with 5 µM or 10 µM C₆-Cer induced BZLF1 expression and LC3B accumulation in SNU-719 and YCCEL1 cells, as illustrated in Fig. 1D. The percentage of BZLF1-positive cells significantly increased in the 5 µM C₆-Cer treatment group (25.9%) compared to the dimethyl sulfoxide (DMSO) control group (4.2%) (Fig. 1E). Additionally, C₆-Cer treatment elevated the number of cells exhibiting strong BZLF1 expression, accompanied by a 7.9-fold augmentation in LC3B fluorescence levels (Fig. 1E). Treatment with C₆-Cer for 24 h increased not only the expression of BZLF1 but also the expression of EBV early genes such as BALF2, BMRF1, and BHRF1 (Fig. 1F). Increasing the treatment concentration of C₆-Cer to 10 µM further enhanced this effect, resulting in a higher level of EBV lytic gene expression compared to 12-O-tetradecanoylphorbol-13-acetate (TPA) treatment, which is commonly used to induce the EBV lytic cycle (Fig. 1F). While 10 µM C₆-Cer efficiently induced BZLF1 expression, it led to significant impairment in cell viability, hindering the feasibility of conducting further experiments. Consequently, 5 µM C₆-Cer treatment was used for the majority of following experiments. In addition, co-treatment with 5 µM C₆-Cer and 5 nM TPA increased EBV lytic gene expression synergistically (Fig. 1F). Treatment with 5 µM C₆-Cer for 72 h also significantly increased viral production (Fig. 1G). These results suggest that C₆-Cer contributes to virus particle formation by inducing the lytic cycle in EBV-positive gastric cancer cells.

Fig 1.

C₆-ceramide induces EBV lytic replication in the EBV-positive gastric carcinoma cells. (A) Endogenous C₁₄–C₂₈-ceramide levels of AGS and AGS-EBV cells were measured by mass spectrometry (n = 7). (B) Changes in cell viability caused by C₆-Cer were measured by methyl thiazole tetrazolium (MTT) assay. BZLF1 expression in (C) AGS-EBV and (D) naturally derived EBV-positive gastric carcinoma cell lines SNU-719 and YCCEL1 was analyzed by Western blot analysis. LC3B was also detected to confirm autophagosome formation by C₆-Cer. (E) The augmentation of BZLF1- and LC3B-positive cells by C₆-Cer treatment was analyzed by flow cytometry. (F) After treatment with 5 or 10 µM C₆-Cer for 24 h, the expression of EBV early genes BALF5, BMRF1, and BHRF1 was analyzed by Western blotting. As a positive control for EBV lytic replication, cells were treated with 5 nM TPA for 24 h. (G) Cells were treated with 5 µM C₆-Cer or 5 nM TPA for 72 h. Cell supernatants were then collected to detect the EBV genome by quantitative PCR (qPCR). The ratios of viral production to the amount obtained from the dimethyl sulfoxide (DMSO) control are shown. Similar results were obtained from three independent experiments, and a representative result is shown. Error bars indicate standard deviations (SDs) (n = 3). *P < 0.05; **P < 0.01; ***P < 0.001.

C₆-Cer treatment increased endogenous ceramide and GlyCer levels

We performed thin-layer chromatography (TLC) to investigate whether treatment by C₆-Cer increased the endogenous ceramide pool (Fig. 2A). For the TLC, we used a solvent system suitable for separation of ceramide and GlyCer (14), and the lipids separated on the TLC plate were visualized at 365-nm UV wavelength following staining by primuline. Each spot was verified by calculating sphingolipid retardation factor after two-dimensional TLC using a chloroform:methanol:water (65:25:4, vol/vol/vol) solvent system, as suggested by Avanti Polar Lipids (https://avantilipids.com/tech-support/analytical-procedures/tlc-solvent-systems) (data not shown). In both AGS and AGS-EBV cells, DMSO treatment did not affect the level of sphingolipids, but 5 µM C₆-Cer treatment significantly increased the amounts of ceramide and GlyCer of various lengths (Fig. 2A). Similar results were obtained from lipid analysis using mass spectrometry (MS) (Fig. 2B). If two peaks with distinct retention times were identified as the same substance, they were labeled as _1 and _2 (Fig. 2B). The heatmap derived from MS analysis showed a marked difference in the level of lipids detected between AGS and AGS-EBV. Treatment with 5 µM C₆-Cer was shown to reduce the amount of SM in both cell lines (Fig. 2B). Principal component analysis (PCA) plots show clear separations of lipid profiles in both AGS and AGS-EBV cells after treatment by C₆-Cer (Fig. 2C). Of the 41 lipids detected in the sample, 35 showed significant alteration due to the C₆-Cer treatment (Fig. 2D).

Fig 2.

Exogenous C₆-ceramide treatment increases endogenous ceramide and glycosyl ceramide pools. Lipids were extracted from cells treated with 5 µM C₆-Cer for 24 h and then analyzed by (A) TLC or (B–D) liquid chromatography-mass spectrometry (LC-MS). (A) TLC was performed using a chloroform:methanol:water (80:10:1, vol/vol/vol) solvent system. Lipid spots were visualized under 365-nm UV after staining with primuline. Similar results were obtained from three independent experiments, and a representative result is shown. (B–D) LC-MS was carried out in positive mode using a Vanquish ultra-performance liquid chromatography system. A quality control sample was used for lipid identification. For statistical analysis, normalization by median, logarithmic transformation (base 10), and Pareto scaling were conducted. (B) A heatmap and (C) PCA were used to visualize the processed data. (D) For multi-group univariate comparison, analysis of variance was applied followed by post hoc Fisher’s least significant difference analysis, and P-values <0.05 were considered statistically significant. Cer, ceramide; GlyCer, glycosyl ceramide; HexCer, hexosyl ceramide; Sph, sphingosine; SM, sphingomyelin.

Increased EBV lytic reactivation by CDase inhibition

We investigated whether treatment with drugs capable of inducing accumulation of ceramide could mimic BZLF1 induction by C₆-Cer. Ceranib-2 is a potent CDase inhibitor that has been reported to increase the intracellular accumulation of various ceramides (15). The IC₅₀ value of ceranib-2 in AGS-EBV was 42 µM (Fig. 3A). Treatment with 5 µM ceranib-2 significantly increased BZLF1 mRNA (Fig. 3B) and protein (Fig. 3C) levels without affecting cell viability, and these effects were further enhanced by co-treatment with 5 µM C₆-Cer. However, LC3B accumulation was not affected by ceranib-2 treatment (Fig. 3C). Co-treatment with 5 µM C₆-Cer and ceranib-2 for 72 h further increased EBV viral particles compared to treatment with the individual compounds (Fig. 3D). These results suggest that treatment with drugs that can increase the endogenous ceramide pool in EBV-positive gastric cancer cells can induce the lytic cycle of EBV.

Fig 3.

Ceranib-2 increases ceramide levels and induces EBV lytic replication. (A) The cytotoxicity of ceramidase inhibitor ceranib-2 was measured by MTT assay. (B–E) AGS-EBV cells were analyzed after 24-h treatment with DMSO, 5 µM C₆-Cer, 5 µM ceranib-2, or 5 µM C₆-Cer + 5 µM ceranib-2. (B) BZLF1 mRNA level was detected by quantitative real time-PCR (qRT-PCR). The results from three independent experiments were normalized to GAPDH values and are expressed as ratio relative to the value obtained from DMSO-treated cells. (C) BZLF1 protein level was analyzed by Western blotting. (D) Cell supernatants were assayed by qPCR to detect the production of EBV. The ratios to the amount obtained from the DMSO control are shown. (E) Changes in sphingolipid by ceranib-2 treatment were detected by TLC using a chloroform:methanol:water (80:10:1, vol/vol/vol) solvent system. Lipid spots were visualized under 365-nm UV after staining with primuline. Similar results were obtained from three independent experiments, and a representative result is shown. Cer, ceramide; GlyCer, glycosyl ceramide; Sph, sphingosine; SM, sphingomyelin. **P < 0.01; ***P < 0.001.

Changes in sphingolipids by C₆-Cer and ceranib-2 treatment in AGS-EBV were investigated by TLC. Treatment with 5 µM C₆-Cer or 5 µM ceranib-2 further increased the amount of very-long-chain and long-chain ceramides (Fig. 3E). Co-treatment with C₆-Cer and ceranib-2 increased the GlyCer compared to treatment with C₆-Cer or ceranib-2 (Fig. 3E).

Suppression of BZLF1 level by GlyCer synthesis inhibition

We tested the effect of glucosyl ceramide synthase (GCS) inhibitor Genz-123346 using TLC. Treatment with Genz-123346 dose dependently reduced BZLF1 mRNA and protein levels, which were induced by 5 µM C₆-Cer treatment (Fig. 4A and B). Ceramide has been reported to increase autophagy as well as c-JUN and extracellular signal-regulated kinase 1/2 (ERK1/2) signals (16, 17). These signals are closely related to EBV lytic replication (18). We investigated whether these signals were related to GlcCer by Western blotting. Genz-123346 treatment reduced C₆-Cer-induced LC3B accumulation (Fig. 4B). ERK1/2 phosphorylation was induced by ceramide treatment but not affected by Genz-123346 treatment, whereas c-JUN expression was induced by ceramide and greatly suppressed by Genz-123346 treatment (Fig. 4B). Since LC3B accumulation induced by ceramide was not completely abolished even after 1 µM Genz-123346 treatment, we speculated that the autophagy was regulated by both ceramide and GlyCer (Fig. 4B). Genz-123346 treatment exhibited a dose-dependent reduction in C₆-Cer-induced viral products, as illustrated in Fig. 4C. Notably, treatment with Genz-123346 did not induce dose-dependent changes in cell viability in the absence of C₆-Cer (Fig. 4D). However, co-treatment with 0.1 and 1 µM Genz-123346 effectively restored the diminished cell viability caused by C₆-Cer treatment (Fig. 4E). Co-treatment with 1 µM Genz-123346 suppressed the 5 µM C₆-Cer-induced increase in GlyCer to the DMSO control level, while it slightly increased the amount of very-long-chain ceramides (Fig. 4F).

Fig 4.

Genz-123346 suppresses BZLF1 expression induced by C₆-ceramide treatment. AGS-EBV cells were pretreated with glucosyl ceramide synthase inhibitor Genz-123346 for 24 h and then further treated with 5 µM C₆-Cer for 24 h. Cells were harvested for RNA (A), protein (B), and lipid (E) preparation. (A) BZLF1 mRNA level was detected by qRT-PCR. The qRT-PCR results from three independent experiments were normalized to GAPDH values and are expressed as ratios relative to the values obtained from DMSO control cells. (B) BZLF1, LC3B, phospho-ERK1/2, total ERK1/2, and c-JUN expression were analyzed by Western blotting. (C) Cell supernatants were analyzed using qPCR to detect the production of EBV after 72 h of drug treatment. Changes in cell viability resulting from (D) Genz-123346 alone or (E) co-treatment with Genz-123346 and C₆-Cer were assessed through MTT analysis. (F) Changes in sphingolipids by Genz-123346 treatment were detected by TLC using a chloroform:methanol:water (80:10:1, vol/vol/vol) solvent system. Lipid spots were visualized under 365-nm UV after staining with primuline. Similar results were obtained from three independent experiments, and a representative result is shown. Cer, ceramide; GlyCer, glycosyl ceramide; Sph, sphingosine; SM, sphingomyelin. Error bars indicate SDs (n = 3). **P < 0.01; ***P < 0.001.

Signal transduction involved in EBV replication by C₆-Cer

We investigated the signal related to changes in C₆-Cer-induced BZLF1 expression by treating the cells with small interfering RNAs (siRNAs) or inhibitors (Fig. 5). To mitigate potential off-target effects, all experiments were conducted using two different siRNAs. Transfection by c-JUN-specific siRNA (siJUN) suppressed c-JUN mRNA and protein expression but did not affect either BZLF1 expression or LC3B level regardless of C₆-Cer treatment (Fig. 5A and B). Treatment with U0126, an ERK1/2-specific upstream kinase MEK1/2 inhibitor, reduced C₆-Cer-induced ERK1/2 phosphorylation and BZLF1 expression (Fig. 5C). U0126 treatment also reduced the viral production induced by C₆-Cer (Fig. 5D). We examined whether the inhibition of ERK expression via siRNA could directly regulate C₆-Cer-induced BZLF1 expression (Fig. 5E and F). Transfection with siERK1 or siERK2 effectively suppressed the respective target ERK mRNA (Fig. 5E) and total ERK protein (Fig. 5F) levels. Interestingly, ERK1/2 phosphorylation induced by C₆-Cer was reduced by the siERK1&2 #2 combination but increased further after siERK1&2 #1 transfection (Fig. 5F). Despite these opposing patterns in phospho-ERK1/2, both siERK combinations consistently led to a reduction in C₆-Cer-induced BZLF1 expression and LC3B accumulation (Fig. 5F).

Fig 5.

C₆-ceramide induced EBV lytic replication via ERK1/2 and cAMP response element binding protein (CREB) phosphorylation. (A and B) AGS-EBV cells were transfected with 20 nM of two distinct siJUN or a negative control siRNA (siNC). (A) Downregulation of c-JUN mRNA level following siJUN transfection was confirmed by qRT-PCR. (B) Twenty-four hours after siRNA transfection, the transfected cells were treated with 5 µM C₆-Cer for 24 h before analysis. AGS-EBV cells pretreated with 10 µM MEK1/2 inhibitor U0126 for 2 h were treated with 5 µM C₆-Cer for (C) 24 h or (D) 72 h. (C) The cells were harvested for Western blot analysis. (D) Cell supernatants were assayed by qPCR to detect EBV production. The ratios of viral production to the amount obtained from the DMSO control are shown. (E) AGS-EBV cells were subjected to qRT-PCR analysis for ERK1 or ERK2 mRNA expression 24 h after transfection with 20 nM of siERK1 or siERK2. A negative control siNC was included for comparison. (F) AGS-EBV cells were transfected with 10 nM of siERK1 and siERK2 or 20 nM of siNC. Twenty-four hours post siRNA transfection, the cells were treated with 5 µM C₆-Cer for 24 h before being harvested for Western blot analysis. (G) AGS-EBV cells were analyzed for CREB mRNA expression by qRT-PCR 24 h after 20 nM of siCREB transfection. A negative control siNC was used for comparison. (F) Twenty-four hours after siRNA transfection, the transfected cells were treated with 5 µM C₆-Cer for 24 h. The cells were harvested for Western blot analysis. (B, C, F, and H) Western blotting was employed to analyze the expression of BZLF1, LC3B, c-JUN, phospho-ERK1/2, total ERK1/2, phospho-CREB, and total CREB. The protein normalization was achieved using β-actin as a loading control. Similar results were obtained from three independent experiments, and a representative result is shown. Error bars indicate SDs (n = 3). *P < 0.05; **P < 0.01; ***P < 0.001.

Exogenous ceramide treatment induces and activates cAMP response element binding protein (CREB) phosphorylation (19). As CREB is a transcription factor promoting BZLF1 transcription (20), we tested if CREB plays a role in C₆-Cer induced effect we observed. Transfection of two different siCREBs significantly reduced CREB mRNA levels (Fig. 5G). Treatment with C₆-Cer induced the phosphorylation of CREB, and transfection with 20 nM siCREB resulted in a reduction of C₆-Cer-derived BZLF1 expression, accompanied by a decrease in phospho-CREB as well as total CREB protein levels (Fig. 5H). We also observed that siCREB transfection suppressed ERK1/2 expression and LC3B accumulation (Fig. 5G and H). These findings collectively suggest that ceramide induces the EBV lytic cycle through the activation of ERK1/2 and CREB, alongside the induction of LC3B accumulation.

Relationship between EBV replication and autophagy induction by C₆-Cer

Treatment with 3-methyladenine (3-MA), a class III phosphoinositide 3-kinase (PI3K) complex inhibitor that inhibits the phagophore-forming autophagy initiation step, resulted in the inhibition of C₆-Cer-induced BZLF1 expression and LC3B-II accumulation, accompanied by a decrease in the expression of the class III PI3K subunit beclin-1 (Fig. 6A). Additionally, 3-MA treatment further reduced the expression of the class I PI3K subunit PIK3CA, which was already decreased after C₆-Cer treatment (Fig. 6A). 3-MA pretreatment also induced caspase-3 cleavage, which did not occur in cells treated with C₆-Cer alone (Fig. 6A). Accumulation of GlcCer caused autophagic-lysosomal dysfunction and disrupted autophagosome-lysosome fusion in a Drosophila model lacking neuronal GCS (21). Thus, we studied whether inhibiting autophagosome-lysosome fusion using chloroquine (22) affects BZLF1 expression. Chloroquine treatment increased LC3B level both before and after C₆-Cer treatment (Fig. 6B). Chloroquine also potentiated the effect of C₆-Cer on the induction of BZLF1. Moreover, the suppression of C₆-Cer-induced BZLF1 expression and LC3B accumulation by Genz-123346 treatment was partially alleviated by chloroquine treatment (Fig. 6B). Genz-123346 treatment resulted in decrease in C₆-Cer-induced ERK1/2 phosphorylation, while it increased CREB phosphorylation (Fig. 6B). In comparison to the substantial increase in BZLF1 and LC3B levels induced by chloroquine, its effects on ERK1/2 and CREB phosphorylation were limited (Fig. 6B). We explored the potential interdependence of BZLF1 and LC3B levels as both were elevated following C₆-Cer as well as chloroquine treatment. Transfection with siLC3B reduced both LC3B mRNA and protein levels; however, their effect on BZLF1 expression was limited (Fig. 6C and D). Transfection with PIK3CA-specific siRNA efficiently downregulated PIK3CA expression in both mRNA and protein levels, while they upregulated beclin-1 expression (Fig. 6E and F). However, siPIK3CA transfection did not impact C₆-Cer-derived BZLF1 expression and autophagy initiation (Fig. 6F).

Fig 6.

The association between C₆-ceramide-induced BZLF1 expression and autophagy. (A) AGS-EBV cells were pretreated with 5 mM 3-MA (an autophagy initiation inhibitor) for 24 h and subsequently treated with 5 µM C₆-Cer for an additional 24 h. (B) AGS-EBV cells were treated with 5 µM C₆-Cer, 1 µM Genz-123346, and/or 5 µM chloroquine (an autophagosome-lysosome fusion inhibitor) for 24 h. The cells were then harvested for Western blot analysis. (C and D) AGS-EBV cells were transfected with 20 nM of siLC3B or a siNC. (C) Downregulation of LC3B mRNA level following 24 h siLC3B transfection was confirmed by qRT-PCR. (D) Twenty-four hours after siRNA transfection, the transfected cells were treated with 5 µM C₆-Cer for 24 h before analysis. (E and F) AGS-EBV cells were analyzed for PIK3CA mRNA expression by qRT-PCR 24 hours after 20 nM of siPIK3CA transfection. A negative control siNC was used for comparison. (F) Twenty-four hours after siRNA transfection, the cells were treated with 5 µM C₆-Cer for 24 h and subsequently harvested for Western blot analysis. (A, B, D, and F) The expression of class III PI3K complex subunit beclin-1, class I PI3K subunit PIK3CA p110α, LC3B, BZLF1, cleaved caspase-3, phospho-ERK1/2, total ERK1/2, phosphor-CREB, and total CREB was analyzed by Western blotting. β-Actin was employed for protein normalization as a loading control.

DISCUSSION

We found that C₆-Cer treatment caused the accumulation of long-chain and very-long-chain ceramide (Fig. 7). C₆-Cer also increased the GlyCer level and the expression of the EBV immediate early gene BZLF1 in EBV-positive gastric cancer cells (Fig. 7). Ceramide treatment induced ERK1/2 and CREB phosphorylation, in addition to initiating autophagy. The reduction of GlcCer by Genz-123346 suppressed LC3B accumulation and BZLF1 expression, whereas chloroquine, an autophagosome-lysosome fusion inhibitor, increased LC3B and BZLF1 levels. Thus, GlcCer may induce LC3B and BZLF1 accumulation by inhibiting the fusion between autophagosomes and lysosomes. BZLF1 expression is hindered by treatment with MEK1/2 inhibitor U0126, siERK1&2, siCREB, and autophagy initiation inhibitor 3-MA (Fig. 7). Conversely, siJUN, siLC3B, and siPIK3CA transfection did not influence ceramide-induced BZLF1 expression (Fig. 7).

Fig 7.

Schematic diagram of BZLF1 expression after C₆-ceramide treatment. A schematic diagram was created with BioRender.com.

C₆-Cer and TPA co-treatment further increased EBV lytic gene expression compared to treatment with each drug alone. TPA activates PKC and induces EBV reactivation through activating AP-1 transcription factors including c-JUN and nuclear factor kappa B (18). Ceramide is also known to activate PKC (23) and increase c-JUN activity (17). In our study, inhibition of c-JUN via siRNA failed to suppress C₆-Cer-derived BZLF1 expression. Therefore, the mechanisms by which TPA and C₆-Cer induce BZLF1 expression in our experimental system seem to be different. This was supported further by the observation that BZLF1 expression was increased additively when TPA and C₆-Cer were applied together. Unlike our study using an EBV-positive gastric cancer cell line, in another study using the EBV-positive Burkitt’s lymphoma cell line Mutu-I, treatment with 250 µM ceramide did not induce BZLF1 and lytic gene expression (11). In this study, cell-killing effect of C₆-Cer was not specific to EBV-positive gastric carcinoma cells. This could be attributed, in part, to suboptimal experimental conditions for obtaining IC₅₀ values, as the methyl thiazole tetrazolium (MTT) assays were conducted only after 24 h of C₆-Cer treatment, and the maximum EBV lytic induction by C₆-Cer was not observed at this time point.

Treatment with the CDase inhibitor ceranib-2 increased BZLF1 expression and viral particle formation, along with an increase in the ceramide pool (Fig. 3). Similar to our results, increased ceramide level induced viral replication in KSHV (10). On the other hand, treatment with C₆-Cer and ceranib-2 reduced SARS-CoV-2 replication (24). C₆-Cer also reduced IAV production (5), and ceranib-2 reduced measles virus replication and virus protein expression (25). Therefore, the regulation of viral replication by ceramide seems to be different depending on the type of virus.

Exogenous ceramide treatment follows a salvage pathway where it is hydrolyzed by CDase and then resynthesized by CERS (13). In our study, C₆-Cer or ceranib-2 treatment significantly increased the pool of endogenous ceramides. These results suggest that reduced ceramide levels in EBV-positive carcinoma cells may be due to increased CDase activity rather than decreased CERS function.

Increased GlyCer and BZLF1 levels by C₆-Cer treatment were dose dependently reduced by the GCS inhibitor Genz-123346. This suggests that GlyCer plays an important role in ceramide-induced BZLF1 expression. A recent report showed that treatment with Genz-123346 inhibited the viral replication of SARS-CoV-2 and IAV (26). The mechanism by which Genz-123346 blocks these viral replications is not well known. It is worth investigating whether GlcCer can induce the viral replication of SARS-CoV-2 and IAV in a similar way to EBV.

CREB is a transcription factor recognized for its role in inducing BZLF1 expression (20) and can undergo phosphorylation by various kinases, including ERK1/2 (27). We observed a reduction in ERK1/2 protein expression after siCREB transfection. Considering that siERK1&2 transfection inhibited BZLF1 expression irrespective of ERK1/2 phosphorylation levels, it is possible that siCREB transfection inhibited BZLF1 expression by suppressing ERK1/2 expression. Consequently, co-transfection of ERK1/2 overexpression vectors with siCREB will provide additional insights into the influence of CREB on BZLF1 expression.

Treatment with 3-MA, an inhibitor of autophagy initiation through Class III PI3K complex inhibition (28), resulted in reduced levels of Class III PI3K subunit beclin-1 and class I PI3K subunit PIK3CA. Additionally, 3-MA treatment decreased C₆-Cer-induced LC3B accumulation and BZLF1 expression. It is worth noting that chloroquine, a drug known to inhibit fusion between autophagosomes and lysosomes (22), has been reported to induce EBV lytic replication (29). Treatment with chloroquine slightly restored the reduced BZLF1 expression and LC3B accumulation by Genz-123346. Collectively, these results suggest that the initiation of autophagy induces the EBV lytic cycle, whereas autolysosome formation suppresses it. Varicella-zoster virus and HCMV, both belonging to the Herpesviridae family, are rapidly exported using autophagic membranes for secondary envelope acquisition (30, 31). Similarly, a model has been proposed in which EBV blocks autophagic flux and is transported externally using an autophagic membrane to acquire a second envelope (32, 33). Recently, it has been reported that the final envelope site of EBV originates from the Golgi apparatus (34). Therefore, it is possible that the second envelope of EBV acquired through the trans-Golgi network fuses with the phagophore and accumulates the mature EBV virion inside the autophagosome. Our results suggest that GlyCer produced after C₆-Cer treatment contributes to the EBV lytic cycle by causing autophagic-lysosomal dysfunction and blocking autophagic flux. However, siLC3B did not impact BZLF1 expression. Furthermore, despite the decrease in beclin-1 levels following 3-MA pretreatment, siPIK3CA transfection led to increased beclin-1 expression without affecting BZLF1 expression. This suggests a potential association between BZLF1 expression and autophagosome maturation process which is initiated by beclin-1. Notably, the knockdown of ATG5, a protein involved in LC3B lipidation, has been shown to reduce TPA/sodium butyrate-induced BZLF1 expression in the EBV Akata strain (33). It would be intriguing to explore whether C₆-Cer-induced BZLF1 expression and LC3B lipidation are linked to ATG5.

Similar to our study, an increase in BZLF1 level through early and late autophagy step control has been reported (32, 33, 35), but the detailed mechanism was unknown. Our study highlights the significance of sphingolipid metabolism, particularly involving ceramide and GlyCer, in EBV lytic replication in gastric cancer cells. This suggests the potential application of drugs targeting sphingolipid metabolism in the treatment of EBV-positive gastric cancer.

MATERIALS AND METHODS

Chemicals and reagents

The following reagents were used to culture cells, and fetal bovine serum was from Corning (Corning, NY, USA). Amphotericin B, penicillin-streptomycin, and trypsin-EDTA were from Gibco (Grand Island, NY, USA). Sodium bicarbonate was from Sigma (St. Louis, MO, USA). G418, plasmocin prophylactic, and MycoStrip were from InvivoGen (San Diego, CA, USA).

For RNA extraction and quantitative real time-PCR (qRT-PCR), RNAiso plus and recombinant RNase inhibitor were from TaKaRa (San Jose, CA, USA). Moloney murine leukemia virus (M-MLV) was from Invitrogen (Waltham, MA, USA). TOPreal Qpcr 2× PreMIX SYBR-Green with low ROX was from Enzynomics (Daejeon, Korea).

For TLC, primuline, high-performance liquid chromatography (HPLC)-grade chloroform, methanol, and water were purchased from Sigma. TLC Silica gel 60 F₂₅₄ aluminum support was purchased from Supelco (St. Louis, MO, USA). Methyl tert-butyl ether (MTBE) was purchased from Sigma for lipid extraction.

To determine the effects of drugs, DMSO, TPA, chloroquine diphosphate salt, and U0126 were purchased from Sigma. C₆-ceramide (catalog no. 860506P) was purchased from Avanti Polar Lipids (Alabaster, AL, USA). MTT, ceranib-2, and Genz-123346 were purchased from Cayman (Ann Arbor, MI, USA).

The following antibodies were used for immunoblotting: anti-BMRF1 was from Novocastra (Newcastle, UK). Anti-BZLF1 (catalog no. sc-53904) and anti-c-JUN (catalog no. sc-74543) were from Santa Cruz (Dallas, TX, USA). Anti-β-actin (catalog no. 4090), anti-phospho-p44/42 MAPK (ERK1/2) (Thr202/Tyr204) (catalog no. 9101), anti-ERK1/2 (catalog no. 4695), anti-phospho-CREB (Ser133) (catalog no. 9198), anti-CREB1 (catalog no. 9197), anti-beclin-1 (catalog no. 3738), anti-cleaved capase-3 (catalog no. 9661), anti-PIK3CA (catalog no. 4249), and horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG (catalog no. 7074) were from Cell Signaling (Danvers, MA, USA). Anti-LC3B (catalog no. NB600-1384) was from Novus Biologicals (Centennial, CO, USA). Goat anti-mouse IgG-HRP (catalog no. GTX213111) was from Genetex (Irvine, CA, USA).

For flow cytometry, the eBioscience Foxp3/transcription factor staining buffer set was purchased from Invitrogen (Waltham, MA, USA). Cy3-conjugated Goat anti-mouse IgG was obtained from Jackson ImmunoResearch Laboratory, Inc. (West Grove, PA, USA), and Cy5-conjugated goat anti-rabbit IgG (catalog no. ab6564) was purchased from Abcam (Cambridge, UK).

Cells

AGS is a female human, gastric adenocarcinoma cell line (KCLB 21739). AGS-EBV is a recombinant Akata EBV strain-infected AGS cell line described previously (36). SNU-719 (37) and YCCEL1 (38) are naturally EBV-infected male human-derived gastric carcinoma cell lines. HFE-145 nonneoplastic gastric epithelial cells were received from Dr. Hassan (Washington, DC, USA). AGS, AGS-EBV, and SNU-719 were cultured in RPMI-1640 medium. YCCEL1 was cultured in minimum essential medium, while HFE-145 was cultured in Dulbecco’s Modified Eagle Medium. All media were supplemented with 10% fetal bovine serum, 0.25 µg/mL amphotericin B, 100 µg/mL streptomycin, 100 U/mL penicillin, and 25 mM sodium bicarbonate. AGS-EBV cells were cultured with addition of 400 µg/mL G418 to the medium. Before analyses, AGS-EBV cells were cultured without G418 for 48 h. All cells were maintained in a 5% CO₂ incubator at 37°C, routinely checked for mycoplasma infection with a MycoStrip, and harvested using Trypsin-EDTA (0.25%).

Western blot analysis

Cell lysate in radioimmunoprecipitation assay buffer was mixed with 5× sodium dodecyl sulfate (SDS)-polyacrylamide gel loading buffer (Dynebio, Seoul, Korea) and heated at 95°C for 5 min. Samples were separated electrophoretically on 12% SDS-polyacrylamide gels, and the separated proteins were transferred to a polyvinylidene fluoride membrane (Millipore, Billerica, MA, USA). Membranes were blocked and probed with the following antibodies: anti-c-JUN (1:500), anti-BZLF1 (1:500), anti-BMRF1 (1:500), anti-BHRF1 (1:2,000) (39), anti-BALF5 (1:500) (40), anti-β-actin (1:1,000), anti-phospho-ERK1/2 (Thr202/Tyr204) (1:1,000), anti-ERK1/2 (1:1,000), anti-phospho-CREB (Ser133) (1:1,000), anti-CREB1 (1:1,000), anti-beclin-1 (1:1,000), anti-cleaved capase-3 (1:1,000), anti-PIK3CA (1:1,000), or anti-LC3B (1:1,000). Bound antibodies were detected with HRP-conjugated anti-mouse or anti-rabbit secondary antibodies at a dilution of 1:2,000 for 45 min at 4°C. Protein bands were visualized using an enhanced chemiluminescence detection system (Amersham Biosciences, Piscataway, NJ, USA), and the membrane was exposed to X-ray film (Agfa, Mortsel, Belgium). Anti-β-actin antibody was used to confirm that loading was comparable between gel lanes. The protein band intensities were quantified using ImageJ software.

Total RNA purification and qRT-PCR

AGS or AGS-EBV cells were harvested, and the total RNA was extracted using RNAiso Plus reagent according to the manufacturer’s instructions. Next, cDNA was synthesized using 2-µg total RNA, oligo(dT) primer, and M-MLV reverse transcriptase. Real-time PCR for the indicated genes was carried out using TOPreal Qpcr 2× PreMIX SYBR-Green with low ROX in a real-time PCR system (CFX96, BioRad, Hercules, CA, USA). The primer sequences are listed in Table 1. PCR conditions were 95°C for 10 min, followed by 35 cycles at 95°C for 10 s, 60°C for 30 s, and 72°C for 30 s. To confirm the specific amplification of the PCR product, dissociation curves were checked routinely. For this, reaction mixtures were incubated at 95°C for 60 s and then ramped from 60°C to 95°C at a heating rate of 0.1°C/s, with fluorescence measured continuously. Relative gene expression was calculated using the quantification cycle (Cq) values and GAPDH as an internal standard.

TABLE 1.

Primers used in this study

Gene	Primer sequence (5′→3′)
	Forward	Reverse
BZLF1	CTGGTGTCCGGGGGATAATG	GGTGGCTTCCAGAAAATGCC
GAPDH	ATGGGGAAGGTGAAGGTCG	GGGGTCATTGATGGCAACAATA
EBNA1	AGTCGTCTCCCCTTTGGAAT	TCCTCACCCTCATCTCCATC
c-JUN	GAGCGGACCTTATGGCTACA	GTGTTCTGGCTGTGCAGTTC
ERK1	CGGCTTCCTGACGGAGTATG	CCCAGGATGCCCAGAATGT
ERK2	TTCGAGCACCAACCATCGAG	TGTTGAGCAGCAGGTTGGAA
CREB	GTTGTTGTTCAAGCTGCCTCT	AGACGGACCTCTCTCTTTCGT
LC3B	TCAGGTTCACAAAACCCGCC	GCGTTTGTGCCAACTGTGAT
PIK3CA	GACCTTCGGCTTTTTCAACCC	GACGTCCCTAAGATCCACAGC

Flow cytometry

AGS-EBV cells were treated with either DMSO or 5 µM C₆-Cer for 24 h and subsequently analyzed. The cells were fixed and permeabilized using the eBioscience Foxp3/Transcription Factor Staining Buffer set according to the manufacturer’s instructions. Following this, the cells were stained with mouse anti-human-BZLF1 (1:50) and rabbit anti-human-LC3B (1:200) for 2 h. Subsequently, Cy3-conjugated goat anti-mouse IgG and Cy5-conjugated goat anti-rabbit IgG were applied for 1 h in the dark. The analysis was performed using FACSAria fusion flow cytometers (BD Biosciences), with BZLF1-Cy3 detected in the PE-A channel and LC3B-Cy5 in the APC-A fluorescence channel.

Virus isolation and measuring viral copy number

AGS-EBV cells (7.5 × 10⁵) were seeded in a 100-mm-diameter dish. After 24 h, the cells were treated with the following drugs: DMSO, C₆-Cer, TPA, ceranib-2, Genz-123346, or U0126. For U0126 co-treatment, cells were pretreated with U0126 2 h before C₆-Cer treatment. The cells were incubated at 37°C in a 5% CO₂ incubator for 72 h, allowing the production of viral particles. Cell supernatants were obtained by centrifuging cultures for 5 min at 500 × g at room temperature, and the supernatants were then passed through a 0.45-μm-pore-size filter (Hyundai Micro, Seoul, Korea). Subsequently, the supernatant was ultracentrifuged using an F50L-8 × 39 rotor (Thermo) at 109,200 × g for 2 h at 4°C. Viral DNA was extracted from the pellet using a QIAamp DNA Mini Kit (Qiagen) according to the manufacturer’s protocol. Real-time PCR amplification of EBNA-1 was carried out using TOPreal Qpcr 2× PreMIX SYBR-Green with low ROX. The primer sequences for EBNA-1 are listed in Table 1. The PCR conditions were 95°C for 30 s, followed by 40 cycles of 95°C for 10 s and 60°C for 30 s. For the dissociation curve, the reaction mixtures were incubated at 95°C for 60 s and then ramped from 55°C to 95°C at a heating rate of 0.1°C/s, with fluorescence being measured continuously. The relative viral copy number was calculated using the Cq values.

MTT assay

The IC₅₀ values of the used drugs were analyzed using the MTT assay. AGS cells (5 × 10³), AGS-EBV cells (7.5 × 10³), SNU-719 cells (1 × 10⁴), YCCEL1 cells (1 × 10⁴), and HFE-145 cells (7.5 × 10³) were seeded in 96-well plates. After 24 h, the drug was administered for a duration of 24 h. Subsequently, 20 µL of MTT solution (5 mg/mL) was added to each well. After a 4-h incubation period, the medium containing MTT solution was aspirated, and 100-µL DMSO was added to each well. The absorbance at 595 nm was then measured with a SoftMax apparatus (Molecular Devices, Sunnyvale, CA, USA) 5 min after the addition of DMSO.

Sample preparation for lipid analysis

For the lipid analyses, independently cultured AGS and AGS-EBV cells were counted, and 3 × 10⁶ cells were dispensed into a 2-mL safe-lock Eppendorf tube. The dispensed cells were centrifuged at 18,000 × g for 1 min at 4°C. The supernatant was removed, and the cells were resuspended in sterile distilled water. The cells were then centrifuged at 18,000 × g for 1 min, and the supernatant was removed. The cell pellets were kept at −80°C until use. To each cell pellet, 300 µL of methanol and 250 µL of water were added, and the sample mixture was vortexed. The tubes were frozen in liquid nitrogen for 5 min and thawed on ice three times. After 1,000 µL of MTBE was added to the tubes, all tubes were placed into universal microtube holders. Following vortexing with a vortex mixer at 1,500 rpm at 4°C for 1 h, the samples were centrifuged at 16,000 × g for 10 min at 4°C. Then, 1 mL of the hydrophobic layer was transferred to a new 2-mL safe-lock Eppendorf tube, and the solvent was completely dried with a nitrogen purge at room temperature. The samples were then kept at −80°C until analysis.

Lipid analysis by thin-layer chromatography

For TLC, lipid extract was dissolved in 40 µL of chloroform:methanol (9:1, vol/vol) mixture and stored in an amber glass vial with a Teflon-coated cap. The TLC plate was pre-washed with methanol and dried overnight in a 50°C dry oven before use. Samples (5 µL) were spotted on a completely dried TLC plate using a 5-µL 22-gauge micro syringe (Hamilton). A chloroform:methanol:water (80:10:1, vol/vol/vol) solvent system was used to separate ceramide and glycosyl ceramide. A solvent system was poured into the TLC chamber, and vapor equilibration was reached after 1 h with the lid sealed at room temperature. The sample on the TLC plate was placed in a chamber for development and then dried. Staining with primuline was used for the visualization of lipid spots (14). One hundred milligrams of primuline were completely dissolved in 10-mL acetone, and deionized water was added up to 100 mL to prepare a 1 mg/mL primuline stock solution, which was stored at 4°C in the dark. Immediately before staining, 1 mL of the primuline stock solution was mixed with 100 mL of PBS. The TLC plate was pre-washed by soaking in PBS for 10 minutes, then soaked in primuline diluted in PBS, and incubated in the dark for 10 min. The TLC plate was washed by soaking in PBS for 5 min and then dried at 50°C for at least 1 h. Lipid spots were then visualized under 365-nm UV.

Lipid extraction for lipidomics

For lipidomic analysis, samples were extracted by a liquid-liquid extraction method. In brief, frozen cell samples were thawed on ice and homogenized in 250 µL of −80°C methanol with a Precellys bead beater twice at 5,500 × g for 30 s. In between homogenization, the samples were cooled for 2 min on dry ice to prevent lipid deformation. An extraction solvent (850 µL of −20°C MTBE) was added, and after being vortexed for 1 min, the samples were extracted on a ThermoShaker for 1 h at 1,500 rpm and 4°C. For the next step, 4°C 210 µL of water was added followed by 1-min vortexing and 15-min shaking at 1,500 rpm. Finally, the samples were centrifuged for 5 min at 16,000 × g, and 800 µL of the upper layer was transferred into separate tubes. After being completely dried under a gentle stream of nitrogen gas, the samples were stored at −80°C until analysis. For resuspension, dried samples were reconstituted with 70 µL of methanol:toluene (9:1, vol/vol), and a quality control (QC) sample was prepared by pooling the same volume of each sample together.

Liquid chromatography-mass spectrometry (LC-MS) conditions for lipidomics

Extracted samples were injected into a Thermo Scientific Vanquish ultra-performance liquid chromatography (UPLC) system equipped with a Waters Acquity UPLC charged-surface hybrid (CSH) C18 column (100 mm × 2.1 mm; 1.7 µm) connected to an Acquity UPLC CSH C18 VanGuard pre-column (5 mm × 2.1 mm; 1.7 µm). The injection volume was 2 µL in a positive mode. In LC mobile phases, the flow rate was 0.6 mL/min, and mobile phases with composition were used in positive mode analysis. In positive mode, (A) acetonitrile:water (60:40, vol/vol) with ammonium formate (10 mM) and formic acid (0.1%) and (B) 2-propanol:acetonitrile (90:10, vol/vol) with ammonium formate (10 mM) and formic acid (0.1%) were used. The gradient elution was conducted as follows: 0 min, 15% of phase B; 0–2 min, 30% of phase B; 2–2.5 min, 48% of phase B; 2.5–11 min, 82% of phase B; 11–11.5 min, 99% of phase B; 11.5–12 min, 99% of phase B; 12–12.1 min, 15% of phase B; 12.1–15 min, 15% of phase B; 4 min of post-run. In the MS acquisition, all the samples were run in the full MS mode for quantification, and the data-dependent acquisition (DDA) mode was applied to the QC samples for lipid identification.

Data processing and statistical analysis for lipidomics

The lipidomic data were processed and annotated by MS-data independent analysis (DIAL) software. Lipid identification was conducted by matching m/z and MS/MS information acquired from the DDA of QC samples. All the data were first filtered with the relative standard deviation of QC samples less than 20% and analyzed by using MetaboAnalyst 5.0. There were no missing values to be processed. For statistical analysis, normalization by median, logarithmic transformation (base 10), and Pareto scaling were conducted. PCA and a heatmap were generated for the visualization of the processed data. For multi-group univariate comparison, analysis of variance was applied followed by post hoc Fisher’s least significant difference analysis, and P-values <0.05 were considered statistically significant.

Transfection of siRNA)

All the siRNAs were purchased from GenePharma (Shanghai, China). Transfection experiments were performed using Lipofectamine 3000 (Invitrogen) according to the manufacturer’s protocol. AGS-EBV cells (7.5 × 10⁵ cells/dish) were transfected with 20 nM siRNA using Lipofectamine 3000 in 100-mm-diameter dishes. Cells were harvested to analyze RNA and protein expression 24 h after transfection. The siRNA sequences are presented in Table 2.

TABLE 2.

siRNAs used in this study

siRNA	siRNA sequence (5′→3′)
	Sense	Antisense
Negative control siRNA	UUCUCCGAACGUGUCACGUTT	ACGUGACACGUUCGGAGAATT
siJUN#1	CCAAGAACGUGACAGAUGATT	UCAUCUGUCACGUUCUUGGTT
siJUN#2	GGCACAGCUUAAACAGAAATT	UUUCUGUUUAAGCUGUGCCTT
siERK1#1	GACCGGAUGUUAACCUUUATT	UAAAGGUUAACAUCCGGUCTT
siERK1#2	GCUACUUCCUCUACCAGAUTT	AUCUGGUAGAGGAAGUAGCTT
siERK2#1	CCAGGAUACAGAUCUUAAATT	UUUAAGAUCUGUAUCCUGGTT
siERK2#2	GCUGCAUUCUGGCAGAAAUTT	AUUUCUGCCAGAAUGCAGCTT
siCREB#1	GUCUCCACAAGUCCAAACATT	UGUUUGGACUUGUGGAGACTT
siCREB#2	CUGCCACAAAUCAGAUUAATT	UUAAUCUGAUUUGUGGCAGTT
siLC3B#1	GUAUGAGAGUGAGAAAGAUTT	AUCUUUCUCACUCUCAUACTT
siLC3B#2	GUUCGGGAUGAAAUUGUCATT	UGACAAUUUCAUCCCGAACTT
siPIK3CA#1	CCGUGAGGCUACAUUAAUATT	UAUUAAUGUAGCCUCACGGTT
siPIK3CA#2	GCCAGUGUGUGAAUUUGAUTT	AUCAAAUUCACACACUGGCTT

Statistical analysis

The qRT-PCR and qPCR data were analyzed using the Student’s t-test. Curve fitting and analysis were performed using GraphPad Prism software (GraphPad Software, San Diego, CA, USA). P-values less than 0.05 were considered statistically significant. All results were expressed as means ± standard deviations.

DATA AVAILABILITY

LC-MS raw data have been deposited and are available at MetaboLights: MTBLS7649.