KSHV Hijacks FoxO1 to Promote Cell Proliferation and Cellular Transformation by Antagonizing Oxidative Stress

Abstract

Reactive oxygen species (ROS) are a group of a highly short-lived molecules that control diverse behaviors of cells. Normal cells maintain ROS balance to ensure their functions. Because of oncogenic stress, cancer cells often have excessive ROS, also known as oxidative stress, which are often counteracted by enhanced antioxidant systems in order to maintain redox homeostasis. Kaposi’s sarcoma-associated herpesvirus (KSHV) is an oncogenic virus associated with Kaposi’s sarcoma (KS), which manifests hyper inflammation and oxidative stress as the hallmarks. We have previously shown that excessive ROS can disrupt KSHV latency by inducing viral lytic replication, leading to cell death. Paradoxically, most KS tumor cells are latently infected by KSHV in a highly inflammatory and oxidative stress tumor microenvironment, which is in part due to the activation of alternative complement and TLR4 pathways, indicating the existence of an enhanced antioxidant defense system in KS tumor cells. In this study, we show that KSHV upregulates antioxidant genes, including SOD2 and CAT by hijacking the forkhead box protein O1 (FoxO1), to maintain intracellular ROS level. Moreover, the fine-tuned balance of ROS level in KSHV-transformed cells is essential for cell survival. Consequently, KSHV-transformed cells are extremely sensitive to exogenous ROS insult such as treatment with a low level of hydrogen peroxide (H₂O₂). Either chemical inhibition or knockdown of FoxO1 by siRNAs decreases the expression of antioxidant genes and subsequently increases the intracellular ROS level in KSHV-transformed cells, resulting in the inhibition of cell proliferation and colony formation in soft agar. Mechanistically, KSHV-encoded microRNAs and vFLIP upregulate FoxO1 by activating the NF-κB pathway. These results reveal a novel mechanism by which an oncogenic virus counteracts oxidative stress by upregulating FoxO1, which is essential for KSHV-induced cell proliferation and cellular transformation. Therefore, FoxO1 might be a potential therapeutic target for KSHV-related malignancies.

1. Introduction

The oncogenic signals in cancer cells reprogram metabolic pathways to generate biosynthetic precursors and energy required to support uncontrolled anabolic proliferation (1). This aberrant energy production and biomacromolecule synthesis often leads to excessive reactive oxygen species (ROS) accumulation, also known as oxidative stress, which is present in almost all cancer cells (2). ROS is a double-edged sword with pleiotropic physiological functions (3). As a second messenger, a low level of ROS is essential to maintain cellular metabolic activities. In contrast, excessive or even modest concentrations of ROS can be toxic to the cells, resulting in cell death (2,4). To overcome this Achilles heel, cancer cells have evolved to hijack the cellular detoxification functions, which is crucial for maintaining the cellular homeostasis and cell survival (5,6).

Kaposi’s sarcoma-associated herpesvirus (KSHV) is one of the seven human oncogenic viruses, which is etiologically associated Kaposi’s sarcoma (KS), primary effusion lymphoma (PEL), a subset of multicentric Castleman’s disease (MCD), and KSHV-associated inflammatory cytokine syndrome (KICS) (7,8). Like other herpesviruses, the KSHV life cycle has two phases: the latent and lytic phases (9,10). In KS tumors, most of tumor cells are latently infected by KSHV. During latency, only limited viral products are expressed including vFLIP (ORF71), vCyclin (ORF72), LANA (ORF73) and a cluster of 12 precursor microRNAs (pre-miRNAs), indicating their essential roles in latency and tumorigenesis (9,11-13). Despite extensive studies, the mechanism underlying KSHV-induced tumorigenesis remains unclear, which is in part due to the lack of a relevant model of KSHV-induced cellular transformation and tumorigenesis. We have previously shown that KSHV can efficiently infect and transform rat primary embryonic metanephric mesenchymal precursor (MM) cells (14). Compared to MM cells, KSHV-transformed MM (KMM) cells grow faster, and are immortalized and transformed. Upon subcutaneous injection into nude mice, KMM cells can induce tumors with pathological and virological features resembling human KS tumors (14). This model enables the investigation of the underlying mechanism of KSHV-induced oncogenesis.

Like other cancer cells, cells transformed by oncogenic viruses have dysregulated metabolic pathways, which are required for their anabolic proliferation (11-13,15-17). KSHV enhances the aerobic glucose metabolism and fatty acid synthesis to promote cell survival during early infection of primary human dermal microvascular endothelial cells (18,19). In KSHV-transformed cells, KSHV enhances glutaminolysis and the urea cycle to promote cellular transformation and cell survival (12,20). The dysregulation of metabolic pathways likely leads to oxidative stress. Importantly, high levels of oxidative stress and inflammation are observed in all four clinical forms of KS (21-23). Thus, in order to maintain the cellular homeostasis, KSHV has likely evolved specific mechanisms to counter cellular oxidative stress.

ROS has been shown to regulate KSHV life cycle. Exogenous ROS, specifically hydrogen peroxide (H₂O₂), can disrupt KSHV latency and induces viral reactivation by activating the mitogen-activated protein kinase (MAPK) pathways (24,25) , while inhibition of forkhead box O protein 1 (FoxO1) induces oxidative stress resulting in KSHV reactivation (26). Of note, ROS induction by silver nanoparticles is shown to inhibit PEL tumor growth by inducing viral lytic reactivation (27). Hence, ROS regulates not only KSHV life cycle but also the homeostasis of KSHV-transformed cells. It is inevitable for KSHV to evolve a mechanism to counter oxidative stress and maintain viral latency, particularly in tumor cells.

FoxOs transcription factor family includes four proteins, FoxO1, FoxO3, FoxO4 and FoxO6, which are central regulators of cellular homeostasis (28). FoxOs proteins, especially FoxO1, play a key role in the regulation of energy metabolism, such as glucose and lipid metabolism (29). Traditionally, FoxOs negatively regulate mammalian cell survival by inducing the expression of apoptotic genes, and therefore are widely regarded as tumor suppressor genes (28). FoxOs are downregulated in several types of cancer such as ovarian, breast and prostate cancers (30). However, in other studies, FoxOs have been shown to promote tumor development by maintaining tumor stem cell properties, enhancing cancer cell drug resistance, or activating the PI3K-AKT pathway, implying a tumor-promoting role of FoxOs in cancer (31,32). The facts that FoxOs are highly conserved ROS sensors and key proteins against intracellular ROS, and they are induced by oxidative stress, further complicate their roles in cancer (33). Specifically, FoxOs proteins can directly bind to the promoters of antioxidant genes such as superoxide dismutase 2 (SOD2) and catalase (CAT) to induce their expressions, resulting in the clearance of intracellular ROS (33).

In this study, we have shown that KSHV-transformed cells are sensitive to exogenous H₂O₂, and identified FoxO1 as a critical antioxidant protein in KSHV-transformed cells. Mechanistically, KSHV-encoded vFLIP and microRNAs concomitantly upregulate FoxO1 to antagonize ROS in a NF-κB-dependent manner, which is critical for KSHV-induced cell proliferation and cellular transformation. These results suggest that FoxO1 might be a potential therapeutic target for KSHV-induced cancers.

2. Methods and Materials

2.1. Cell culture and reagents

Human embryonic kidney 293T cells (HEK293T) and rat primary embryonic metanephric mesenchymal precursor cells (MM) were cultured in DMEM containing 10% fetal bovine serum (FBS; Sigma), and antibiotics (100 μg/ml penicillin and 100 μg/mL streptomycin). KSHV-transformed MM cells (KMM) were maintained in DMEM supplemented with 10% FBS, and antibiotics containing 200 μg/mL hygromycin, 100 μg/ml penicillin and 100 μg/mL streptomycin. BJAB cells were cultured in RPMI medium containing 20% FBS. KSHV-infected BJAB cells (BJAB-KSHV) were cultured as BJAB cells in the presence of 100 μg/ml hygromycin. The FoxO inhibitor AS1842856 from Sigma-Aldrich (Cat. 344355) was dissolved in DMSO.

2.2. siRNA knockdown

The control (NC) and siRNAs for RelA were purchased from Sigma-Aldrich. A pool of scrambled oligonucleotides containing four random sequences used as a control (NC) and a pool of siRNAs containing four pair of siRNAs for FoxO1 and FoxO3 were purchased from Dharmacon. Five pmol of each siRNA oligoribonucleotide duplex was transfected into MM and KMM cells that were seeded in 24 well plate using LipofectamineRNAi Max (Thermo Fisher, 13778150). Cells were examined for knockdown efficiency by RT-qPCR and Western-blotting analysis after 2 and 3 days post-transfection, respectively.

2.3. Lentiviral shRNA knockdown

The linearized pLKO1 plasmid was ligated with shRNA sequences for FoxO1 and FoxO3. The sequences for FoxO1 shRNAs were shRNA1: GCTGCCCAGGCCGGAGTTTAA; shRNA2: GCTGTCAGCACCGACTTTATG; shRNA3: GGAGAAGAGCTGCATCCATGG. The sequences for FoxO3 shRNAs were shRNA1: GCCGAGGCCGGGCAGCCAAGA; shRNA2: GCCCTTCCCAACTCTCCAAGT. The sequence for a non-targeting shRNA was: TTGTACTACACAAAAGTACTG. To ontain lentiviral virion progeny, HEK293T cells was transfected with the targeted shRNA plasmid, pMD2.G and psPAX2 packaging plasmids using the polyethylenimine reagent (Polyscien, 23966-2). The cells were harvested following 48 h and 72 h transfection, respectively. Cells were firstly treated with 8 μg/ml polybrene, and then a defined volume of supernatants containing lentivirus was added to infect the cells by centrifuging at 1,500 rpm for 1 h. At 2 and 3 days post-transduction, cells were checked for knockdown efficiency by RT-qPCR and Western blotting analysis, respectively.

2.4. Soft agar assays

Soft agar assay was conducted by preparing two layers of agarose (Biosharp, BS081). For the base layer, agarose at the final concentration of 0.5% was evenly mixed with 20% FBS in DMEM. For the top layer, 2x10⁴ cells were suspended in 1 ml of DMEM containing 20% FBS and 0.3% agarose. The top layer was plated onto the base layer in 6 well-plates. After 2-3 weeks, colonies were photographed at 40x magnification and colonies with diameter >50 μm were subjected to statistical analysis.

2.5. Reverse transcription quantitative real-time polymerase-chain reaction (RT-qPCR)

Total RNA was extracted using TRI reagent (Sigma, T9424,) based on the instructions of the manufacturer. Reverse transcription of total RNA into cDNA was done using Maxima H Minus First Strand cDNA Synthesis Kit (Thermo Fisher, K1652). qPCR analysis was conducted by using the SYBR Green method (Bio-Rad, 172-5272). All reactions were run in triplicates. The relative expression for each target gene was calculated by the 2^−ΔΔCt after normalized to β-actin. The primers for rat FoxO1 were: 5'-GTGAACACCATGCCTCACAC-3' (FoxO1-F) and 5'-CACAGTCCAAGCGCTCAATA-3' (FoxO1-R); rat FoxO3 were: 5'-CTCAGCCAGTGGACAGTGAA-3' (FoxO3-F ) and 5'-GCTCTGGAGTAGGGATGCTG-3' (FoxO3-R); rat SOD2: 5’-CACATTAACGCGCAGATCA-3’ (SOD2-F) and 5’-AGGCTGAAGAGCAACCTGAG-3’ (SOD2-R); rat CAT: 5’-GTGGTTTTCACCGACGAGAT-3’ (CAT-F) and 5’- CAAACACCTTTGCCTTGGAG-3’ (CAT-R); and rat β-actin: 5’-CCATGTACCCAGGCATTGCT-3’ (β-actin-F) and 5’-AGCCACCAATCCACACAGAG-3’ (β-actin-R).

2.6. Western blot analysis

Cell were lysed in 1x laemmli buffer. The obtained cell lysates were separated in SDS polyacrylamide gels, and the proteins were electrophoretically transferred onto nitrocellulose membranes (GE Healthcare). The membranes were then blocked with 5% skim milk at room temperature for 1 h, which was sequentially incubated with primary and secondary antibodies. The signal was probed by using Luminiata Crescendo Western HRP substrate (EMD Millipore, WBLUR0500,). The images were recorded with a ChemiDoc MP Imaging System (Bio-Rad, 17001402) at Chemi channels. The primary antibodies used included rabbit monoclonal antibodies (mAbs) for FoxO1 (CST, 2880), FoxO3 (CST, 2497) and β-tubulin ((Sigma, 7B9), FoxO3 (CST, 9139), p-FOXO1(T24)/FOXO3a (T32) (CST, 9464), GAPDH (CST, 5174); and p65 (CST, 8242).

2.7. Cell cycle analysis and apoptosis assay

Cell cycle was conducted as previously described (60). Briefly, cells were pelleted and then fixed with 70% ethanol, which was then stained by propidium iodide (PI) and subsequently analyzed by flow cytometry. Apoptotic cells were examined by double staining of Fixable Viability Dye eFluor 660 (eBioscience, 650864) and PE-Cy7 Annexin V (eBioscience, 88810374). FACS Canto System (BD Biosciences) was used to perform flow cytometry, and analysis was conducted by FlowJo V10.

2.8. ROS detection

Cells were pelleted and then stained with a cell-permeable fluorogenic probe of ROS, the CellROX Deep Red Reagent (Thermo Fisher, C10422). This probe is nonfluorescent when in a reduced status but become fluorescent upon oxidation exhibiting maximal excitation/emission at 640/665 nm. Live cells were incubated with 0.5 μm CellROX Deep Red Reagent for 30 min at 37 °C, which was then subjected to flow cytometry analysis.

2.9. H₂O₂ detection

The H₂O₂ level was detected by Hydrogen Peroxide Assay Kit (Abcam, ab102500) following the manufacturer’s instructions. The trypsinzed cells were pelleted by centrifuge at 1,500 rpm for 5 min, and subsequently homogenized by using a Dounce homogenizer (10-50 passes) on ice. The cell lysates were then centrifuged at about 12,000g for 5 minutes at 4°C, and the supernatans were collected and subjected to deproteinization process. The processed supernatant samples were then mixed with the reaction mixture containing the Assay Buffer, OxiRed Probe and HRP at room temperature for 30 minutes protected from light. The H₂O₂ level was immediately measured by colorimetric readings using the microplate reader at OD570 nm. The concentration of H₂O₂ in the samples was calculated according to the standard curve.

2.10. Statistical analysis

Results were presented as the mean ± standard error of the mean (SEM). The differences between two groups were analyzed using Student’s t-test, and one-way ANOVA was performed when more than two groups were compared. Statistical tests were two-sided. A P < 0.05 was considered statistically significant. Statistical symbols “*”, “**” and “***” represent p-values < 0.05, < 0.01 and < 0.001, respectively, while “NS” indicates “not significant”.

3. Results

3.1. KSHV-transformed cells are sensitive to exogenous ROS

We examined the intracellular level of ROS in KSHV-transformed KMM cells (14). The flow cytometry analysis showed that the relative ROS level in KMM cells was significantly lower than in MM cells (Fig 1A-B). H₂O₂ is one of the most common ROS species in non-immune cells. Quantification of intracellular H₂O₂ level revealed around 0.3 μM in MM cells compared to 0.1 μM in KMM cells (Fig 1C). These results suggest that KSHV might have evolved a specific mechanism to counter oxidative stress in order to maintain the homeostasis. Thus, we determined whether KSHV-transformed cells were sensitive to exogenous H₂O₂. Although both MM and KMM cells exhibited reduced cell proliferation rates upon treatment with exogenous H₂O₂, KMM cells more sensitive than MM cells (Fig 1D). At the lowest 0.05 mM concentration tested, H₂O₂ had no effect on MM cells but significantly inhibited the proliferation of KMM cells by 50%. Exposure with 0.4 mM H₂O₂ for 4 hours reduced the cell numbers of MM and KMM cells by 79.5% and 48.4% at day 3 post-treatment, respectively (Fig 1E), suggesting that KMM cells might have a stronger anti-oxidant defense system than that of MM cells. Furthermore, treatment with H₂O₂ significantly decreased the efficiency of colony formation in soft agar of KSHV-transformed cells (Fig 1F-G). In agreement with the results of cell proliferation, H₂O₂ induced cell cycle arrest in both MM and KMM cells at concentration as low as 0.1 mM but only significantly increased the numbers of apoptotic cells of both types of cells at concentration of 0.4 mM (Fig 1H-I). Hence, H₂O₂ primarily affects cell cycle progression rather than cell survival of these cells. These results indicate that maintaining the redox balance is essential for KSHV-transformed cells, which likely possess an enhanced antioxidant system.

FIGURE 1.

KSHV-transformed cells are sensitive to ROS. (A-B) The intracellular ROS level is decreased in KSHV-transformed cells. Representative detection of ROS in MM and KMM cells by flow cytometry (A) and quantification of results from 3 repeats (B). (C) The intracellular H₂O₂ level is decreased in KSHV-transformed cells. Detection of intracellular ROS concentration with a colorimetric method. Results are from 3 independent repeats. (D) H₂O₂ inhibits the cell proliferation of MM and KMM cells in a concentration-dependent manner. Cells at 1x10⁴ were seeded in 24 well plate and then treated with indicated concentrations of H₂O_2. The live cells were counted at day 1, 2, and 3 post-treatment. (E) KSHV-transformed KMM cells are more resistant to H₂O₂ than uninfected MM cells. Cells at 1x10⁴ were seeded in 24 well plate and then pulsed with 0.4 mM H₂O₂ for 4 hours. The live cells were counted at day 1, 2, and 3 post-treatment. (F-G) H₂O₂ inhibits colony formation of KMM cells in soft agar. The colony formation in soft agar of KMM cells were examined following treatment with different concentration of H₂O₂. Representative pictures at 40x magnification were shown (F). Colonies with diameter >50μm were quantified in each field and the results were graphed (G). (H-I) H₂O₂ induced G1 cell cycle arrest but not apoptosis. MM and KMM cells were treated with indicated concentrations of H₂O₂ for 24 h and then analyzed for cell cycle progression (H) or apoptosis (I) by flow cytometry.

3.2. KSHV upregulates FoxO1 to antagonize ROS in transformed cells

FoxO1, FoxO3 and FoxO4, the main members of the FoxO protein family, are the major oxidative stress defenders (28-30). To determine whether KSHV hijacks FoxOs to antagonize ROS, we examined their relative expression levels in KSHV-transformed cells. We found that FoxO1 and FoxO3 but not FoxO4 were significantly upregulated in KMM cells compared to the uninfected MM cells at both mRNA and protein levels (Fig 2A-B). The upregulated expressions of FoxO1 and FoxO3 were not specific to our model as they were confirmed in KSHV-infected BJAB cells compared to the uninfected BJAB cells (Fig 2C).

FIGURE 2.

KSHV latent infection upregulates FoxO1 to antagonize ROS. (A-B) FoxO1 is upregulated in KSHV-transformed cells. The levels of FoxO1 mRNA (A) and protein (B) in MM and KMM cells were examined by RT-qPCR and Western-blotting analysis, respectively. (C) KSHV infection upregulates the expression of FoxO1 protein in BJAB cells. FoxO1 protein levels in uninfected and KSHV-infected BJAB cells were examined by Western-blotting analysis. (D) AS1842856 specifically decreases FoxO1 protein level. Western-blotting analysis of FoxO1 protein in MM and KMM cells after treating with the indicated concentrations of FoxO1 inhibitor AS1842856 for 48 h. (E-F) AS1842856 increases the intracellular ROS level in MM and KMM cells. Cells were treated with the indicated concentrations of AS1842856 for 24 h and analyzed by flow cytometry to detect the intracellular ROS levels (E) and the results of 3 independent repeats were presented (F). (G) Western-blotting analysis confirmed the knockdown efficiencies of FoxO1 and FoxO3 by siRNAs. (H) Knockdown of either FoxO1 or FoxO3 increases the intracellular ROS levels in MM and KMM cells. MM and KMM cells were transfected with FoxO1 and FoxO3 siRNA for 2 days, and analyzed by flow cytometry to detect the intracellular ROS levels. (I-J) AS1842856 decreases the mRNA levels of SOD2 (I) and CAT (J) in KSHV-transformed cells. (K-L) Either FoxO1 or FoxO3 knockdown decreased the mRNA levels of SOD2 (K) and CAT (L). MM and KMM cells were transfected with FoxO1 and FoxO3 siRNA for 2 days followed by RT-qPCR analysis of SOD2 and CAT mRNA expression.

Since KSHV upregulated fold change was higher for FoxO1 than FoxO3 at mRNA level (7 vs. 1.3 in Fig 2A) and at protein level (16.14 vs. 3.77 in Fig 2B), we determined whether KSHV upregulation of FoxO1 was responsible for controlling excess ROS in KSHV-transformed cells by treating the cells with a specific FoxO1 inhibitor AS1842856. While AS1842856 is reported to mainly inhibit the transactivation activity of FoxO1, we and others have previously found that it also decreases FoxO1 but not FoxO3 and FoxO4 protein level in iSLK-RGB cells (26,34). Consistently, treatment with AS1842856 decreased the level FoxO1 but not FoxO3 protein level in both MM and KMM cells (Fig 2D). Importantly, FoxO1 inhibition by AS1842856 was sufficient to increase the ROS level in KSHV-transformed cells (Fig 2E-F). Additionally, AS1842856 also increased the ROS level in MM cells, confirming the essential role of FoxO1 in regulating the homeostasis of redox status in normal cells (Fig 2E-F). To confirm these results, we performed siRNA knockdown of FoxO1 (Fig 2G). FoxO1 knockdown increased the intracellular levels of ROS in both MM and KMM cells (Fig 2H). FoxO3 knockdown also increased the intracellular levels of ROS in both MM and KMM cells (Fig 2H). Interestingly, the intracellular ROS level was increased more by FoxO1 than FoxO3 knockdown in KMM cells; however, this trend was reversed in MM cells (Fig 2H), suggesting a more important role of FoxO1 than Fox3 in KMM cells, which was in agreement with their different folds of induction by KSHV.

In response to oxidative stress, FoxO1 and FoxO3 are known to bind to the promoters of anti-oxidant genes including SOD2 and CAT to induce their expression (35). Indeed, siRNA knockdown or pharmacological inhibition of FoxO1 and FoxO3 significantly suppressed the expression of both SOD2 and CAT in KSHV-transformed cells (Fig 2I-L). Altogether, these results indicate that FoxO1 and FoxO3 control the ROS level in KSHV-transformed cells.

3.3. vFLIP and miRNAs mediate KSHV upregulation of FoxO1

KSHV is in tight latency in KMM cells expressing numerous latent genes including vFLIP, vCyclin, LANA and a cluster of 12 pre-microRNA (14). To identify the viral gene(s) that mediates FoxO1 upregulation, we examined MM cells infected by KSHV mutant with a deletion of vFLIP (ΔvFLIP), vCyclin (ΔvCyclin) or a cluster of 10 of the 12 pre-miRNAs (miR-K1-9 and 11, ΔmiR) (36,37). Because of the essential role of LANA in viral genome persistence, it is impossible to obtain cells stably infected by a KSHV mutant with LANA deleted (38). Deletion of vFLIP or the miRNA cluster but not vCyclin restored FoxO1 protein and mRNA levels to those of the MM cells (Fig 3A-B). Consistently, overexpression of vFLIP but not vCyclin in MM cells is sufficient to increase FoxO1’s protein level (Fig 3C). In ΔmiR cells, overexpression of the miRNA cluster restored FoxO1’s protein level (Fig 3D). These results confirmed the essential roles of the miRNA cluster and vFLIP in KSHV upregulation of FoxO1.

FIGURE 3.

vFLIP and the miR cluster mediates KSHV upregulation of FoxO1. (A-B) KSHV vFLIP and the miR cluster are required for the upregulation of FoxO1. Western-blot (A) and RT-qPCR (B) analysis of FoxO1 expression in Mock (MM), WT (KMM), ΔvFLIP, ΔvCyclin and ΔmiR cells. (C) KSHV vFLIP alone increases FoxO1 expression. Detection of FoxO1 expression in MM cells overexpressed with KSHV vFLIP or vCyclin by Western-blot analysis. Cells trancfected with the empty vector was used as a control (Vec). (D-E) Multiple KSHV miRs mediate KSHV induction of FoxO1 expression. Examination of FoxO1 expression in ΔmiRs cells overexpressed with individual miRs by Western-blotting analysis (D) and RT-qPCR (E).

To identify the miRNA(s) mediates KSHV upregulation of FoxO1, we overexpressed individual pre-miRNAs in ΔmiR cells. Western blotting results showed that overexpression of all the pre-miRNAs upregulated the FoxO1 at both protein and transcripy levels, although the levels did not reach that of KMM cells (Fig 3D-E), indicating that multiple pre-miRNAs were required to induce maximum FoxO1 expression. Together, these results indicate that KSHV vFLIP and multiple miRNAs mediate FoxO1 upregulation.

3.4. vFLIP and the miR cluster upregulate FoxO1 in a NF-κB-dependent manner contributing to KSHV suppression of ROS

Since vFLIP and the miRNA cluster upregulate the expression of FoxO1, and FoxO1 mediates KSHV inhibition of ROS, we examined the roles of vFLIP and the miRNA cluster in regulating ROS in KSHV-transformed cells. Deletion of either vFLIP or the miRNA cluster but not vCyclin indeed increased the intracellular ROS levels (Fig 4A-B). Both vFLIP and the miRNA cluster are required for maximal activation of the NF-κB pathway, which is essential for KSHV-induced cell survival and cellular transformation (37,39-42). RelA is one of the key subunits of the NF-κB complex (43) and knockdown of RelA inhibited the NF-κB pathway in KMM cells (37). Knockdown of RelA decreased the protein level of FoxO1 (Fig 4C). Consistently, two specific pharmacological inhibitors of the NF-κB pathway, JSH-23 and BAY11-7082, also dramatically reduced FoxO1 protein expression (Fig 4D). These results indicated that the NF-κB pathway mediates KSHV upregulation of FoxO1 protein. In agreement with these results, inhibition of the NF-κB pathway by either RelA silencing or specific inhibitors of the NF-κB pathway signaigicantly increased the intracellular ROS levels in both MM and KMM cells (Fig 4E-H). It is worth noting that suppression of NF-κB by either by RelA knockdown or using the specific inhibitors failed to completely restore the level of FoxO1 protein in KMM cells to that of MM cells (Fig 4C-D), suggesting possible involvement of other pathways in the upregulation of FoxO1 protein in KMM cells.

FIGURE 4.

vFLIP and miR cluster mediate KSHV suppression of ROS by upregulating FoxO1 in a NF-κB-dependent manner. (A-B) The mRNA cluster and vFLIP mediate KSHV suppression of ROS. Flow cytometry analysis of ROS in uninfected (MM), KSHV-transformed (WT), ΔvFLIP, ΔvCyclin and ΔmiR cells (A) and quantification of the results from 3 independent repeats (B). (C) RelA Knockdown reduces the levels of FoxO1 and FoxO3 proteins. Western-blot detection of RelA, FoxO1 and FoxO3 proteins in MM and KMM cells following transfection with siRNAs to RelA (siRelA) or a non-targeting control (NC) for 3 days. GAPDH was used as an internal control for loading. (D) Inhibition of the NF-κB pathway reduces the levels of FoxO1 and FoxO3 proteins. MM and KMM cells were treated with NF-κB inhibitors JSH23 (30 μM) or BAY 11-7082 (2 μM), and examined for the expression of FoxO1 and FoxO3 proteins by Western-blotting analysis at 24 h post-treatment. (E-F) RelA knockdown increases the intracellular ROS level. Flow cytometry detection of ROS levels in MM and KMM cells following transfection with siRNAs to RelA (siRelA) or a non-targeting control (NC) for 3 days, and results from 3 independent repeats were presented (F). (G-H) Inhibition of the NF-κB pathway increases the intracellular ROS level. Flow cytometry detection of ROS levels in MM and KMM cells following treatment with NF-κB inhibitor JSH23 (30 μM) or BAY 11-7082 (2 μM) for 48 h, and results from 3 independent repeats were presented (H).

3.5. FoxO1 knockdown suppresses KSHV-induced cell proliferation and transformation

To determine the role of FoxO1 in cell proliferation and cellular transformation of KSHV-transformed cells, we performed siRNA knockdown of FoxO1. siRNA knockdown of either FoxO1 in KMM cells reduced cell proliferation by 50% (Fig 5A). Knockdown of FoxO3 also had similar effect (Fig 5A). Simultaneous knockdown of both FoxO1 and FoxOs reduced cell proliferation by 80% suggesting that both genes mediated cell proliferation of KSHV-transformed cells (Fig 5A). siRNA knockdown of either FoxO1 or FoxO3 in MM cells also reduced cell proliferation but with less effect (Fig 5A). FoxO1 or FoxO3 knockdown induced cell cycle arrest in both MM and KMM cells by increasing the percentage of G1 phase cells from 45% and 32% to 62% and 55%, respectively, while reducing the percentage of S phase cells from 32% and 40% to 25% and 26% in MM and KMM cells, respectively (Fig 5B). In contrast, knockdown of either FoxO1 or FoxO3 didn’t significantly affect the numbers of apoptotic cells in MM and KMM cells (Fig 5C). On the other hand, overexpression of either FoxO1 or FoxO3 enhanced cell proliferation of KMM but not MM cells (Fig 5D-E). Significantly, knockdown of either FoxO1 or FoxO3, or both by lentivirus-mediated expression of specific shRNAs almost completely abolished colony formation in soft agar of KMM cells (Fig 5F-G). Taken together, these results demonstrated that knockdown of FoxO1 or FoxO3 had effect on KSHV-transformed cells similar to those of H₂O₂ and AS1842856 treatments, indicating that FoxO proteins might function through antagonizing oxidative stress to promote KSHV-induced cell proliferation and cellular transformation.

FIGURE 5.

FoxO1 knockdown suppresses KSHV-induced cell proliferation and cellular transformation. (A) Knockdown of either FoxO1 or FoxO3 impairs the proliferation of KMM cells but only with marginal effect on MM cells. MM and KMM cells transfected with a non-targeting control (NC) or a siRNA to FoxO1 (siFoxO1) or FoxO3 (siFoxO3), or both were examined for cell proliferation. The live cells were counted at 24, 48, and 72 h post-transfection. (B-C) Knockdown of either FoxO1 or FoxO3 induces G1 cell cycle arrest (D) but not apoptosis (E) in KMM cells. MM and KMM cells were transfected with a non-targeting control (NC) or a siRNA to FoxO1 (siFoxO1) or FoxO3 (siFoxO3) for three days, and subjected to cell cycle progression (B) or apoptosis (C) analysis. (D) Ectopic expression of FoxO1 and FoxO3 in MM and KMM cells. Western-blotting analysis of FoxO1 and FoxO3 in MM and KMM cells transduced with lentiviruses expressing FoxO1, FoxO3 or a Vector control (Vec). (E) Overexpression of either FoxO1 or FoxO3 enhances the proliferation of KMM but not MM cells. MM and KMM cells transduced with lentiviruses expressing FoxO1, FoxO3 or a Vector control (Vec) were examined for cell proliferation. (F-G) Knockdown of either FoxO1 or FoxO3 inhibits the colony formation of KMM cells in soft agar. KMM cells (2x10⁴) transduced with lentiviruses expressing shRNAs targeting either FoxO1 or FoxO3 were examined for colony formation in soft agar in 6 well-plates for three weeks. Representative pictures at 40x magnification were shown (F), and colonies with diameter >50μm from 3 independent repeats were quantified (G).

3.6. FoxO1 inhibitor AS1842856 suppresses KSHV-induced cell proliferation and cellular transformation

Given the essential role of FoxO1 in KSHV-transformed cells, we investigated the effect of FoxO1 specific inhibitor AS1842856 on KSHV-induced cell proliferation and cellular transformation. Treatment with AS1842856 inhibited the proliferation of KMM cells in a concentration-dependent fashion but with a less effect on MM cells (Fig 6A). Specifically, AS1842856 reduced cell proliferation of KMM cells by 74% and 79%, and MM cells by 45% and 61% at 1 μM and 5 μM at day 3 post-treatment, respectively (Fig 6A). Furthermore, AS1842856 at 1 μM was sufficient to completely abolish colony formation of KMM cells in soft agar (Fig 6B-C). In agreement with these results, AS1842856 significantly induced cell cycle arrest in KMM cells by increasing the numbers of G1-phase cells and decreased S-phase cells at both 1 and at 5 μM but only had marginal effect on MM cells (Fig 6D). In contrast, AS1842856 did not increase the numbers of apoptotic or dead cells of MM and KMM cells (Fig 6E). Taken together, these results showed that AS1842856 strongly reduced cell proliferation and cellular transformation of KMM cells by inhibiting cell cycle progression but not cell survival, which phenocopied the effect of H₂O₂ (Fig 1D-I).

FIGURE 6.

FoxO1 inhibitor AS1842856 suppresses KSHV-induced cell proliferation and cellular transformation. (A) AS1842856 inhibits the proliferation of MM and KMM cells. Cells seeded at 2x10⁴ cells/well in 24-well plates were treated with the indicated concentrations of AS1842856, and the live cells were counted at 24, 48, and 72 h post-treatment. (B-C) AS1842856 inhibits the colony formation of KMM cells in soft agar. MM and KMM cells at 2x10⁴ cells/well were examined for colony formation in soft agar in the presence of the indicated concentrations of AS1842856 in 6 well-plates for three weeks. Representative pictures captured at 40x magnification are shown (B) and the results from 3 independent experiments are presented (C). (D-E) AS1842856 induces G1 cell cycle arrest but not apoptosis in KMM cells. MM and KMM cells were treated with the indicated concentrations of FoxO1 inhibitor AS1842856 for 24 h, and analyzed for cell cycle progression (D) or apoptosis (E).

4. Discussion

Inflammation and oxidative stress have been associated with cancer development and progression (2,6). H₂O₂ is one of the most common ROS species in non-immune cells. Therefore, H₂O₂ is thought to present in cancer cells. Likewise, inflammation and oxidative stress are the hallmarks of all clinical forms of KS (21,23,44,45). We and others have previously reported that H₂O₂ is sufficient and essential for KSHV reactivation from latency (24-26). Paradoxically, most KS cells are latently infected by KSHV, indicating that KSHV has evolved to hijack cellular defense mechanisms to detoxify ROS or H₂O₂ in order to maintain viral latency (9). However, the critical factors involved in antioxidant defense in KS tumors remain unknown.

Numerous proteins, such as SOD2, CAT and glutathione peroxidase (GPx) are shown to prevent cells from oxidative stress (46). SOD2 directly catalyzes superoxide (O₂-) into hydrogen peroxide (H₂O₂), which is further broken into H₂O and O₂ by CAT (46). Since H₂O₂ is a metabolic byproduct of mitochondrial respiration, it is expected to be a common and important ROS species in cancer (47). Therefore, SOD2 and CAT that are responsible for H₂O₂ decomposition are likely essential in cancer.

FoxO1 was originally described as a tumor suppressor because it was shown to directly transactivate pro-apoptotic genes such as Bim and Bax (48). Nevertheless, FoxO proteins are now known to be involved in antioxidative protection of cells via directly transactivating SOD2 and CAT (49). Because the tumor microenvironment often contain a high level of ROS, upregulation of FoxO1 might be critical for the survival and homeostasis of cancer cells (48,50). In this study, we observed that exogenous H₂O₂ inhibited the proliferation and colony formation in soft agar of KSHV-transformed cells by inducing G1 cell cycle arrest (Fig 1D-I). These results suggest that the balance of redox status is critical for KSHV-transformed cells. Furthermore, we showed that FoxO1 directly antagonized oxidative stress by upregulating SOD2 and CAT, ensuring KSHV-induced cell proliferation and cellular transformation (Fig 2). Thus, we have established a crucial role of FoxO1 in antagonizing ROS in KSHV-transformed cells, which is essential for KSHV-induced cell proliferation and cellular transformation.

During KSHV latency, only limited viral products are expressed (10). Among them are KSHV vFLIP and the miRNA cluster, indicating their important roles in KS development. Indeed, we have shown that KSHV vFLIP and miRNA cluster are essential for KSHV-induced cell proliferation and cellular transformation by regulaing multiple growth and survival related pathways and cellular metabolism (11-13,20,37,51). Our results show that both KSHV-encoded miRNA cluster and vFLIP are essential for FoxO1 upregulation, which is mediated by the NF-κB signaling pathway (Fig 3 and 4C-D). Although either KSHV vFLIP or miRNA cluster could activate NF-κB pathway, their simultaneous expression is required to ensure the maximal activation of NF-κB in KSHV-transformed cells. The underlying mechanism remains to be further clarified.

Excessive NF-κB activation might be detrimental to the cancer cells (52,53). Numerous studies have shown that FoxO1 can negatively counteract the NF-κB pathway (54-56). In contrast, the NF-κB pathway can activate FoxO1 at hepatocytes by inducing its phosphorylation in response to ferrous ions (57). However, the underlying mechanism of NF-κB regulation of FoxO1 remains unclear. We have shown that RelA silencing or pharmacological inhibition of the NF-κB pathway leads to decreased expression of FoxO1 and increased intracellular ROS level in KSHV-transformed cells (Fig 5). Thus, NF-κB is a central regulator of redox homeostasis in addition to its pro-survival function, which is consistent with previous findings (58,59). These results further confirm the essential role of NF-κB activation by either vFLIP or the miRNA cluster in KSHV-induced tumorigenesis (37,39-41). Thus, both vFLIP and the miRNA cluster play fundamental roles in KSHV-induced tumorigenesis at least in part by regulating redox homeostasis through activating the NF-κB pathway.

In summary, we have found that KSHV encodes vFLIP and the miRNAs to suppress oxidative stress by upregulating FoxO1 in KSHV-transformed cells (Fig 7). This mechanism is likely critical for the proliferation and survival of KSHV-induced tumor cells in a hyper-inflammatory and oxidative tumor microenvironment. Our findings demonstrate the importance of maintaining the balance of redox status in cancer, which could be potentially explored for therapy.

FIGURE 7.

A model of KSHV maintenance of redox homeostasis by regulating FoxO1 in cellular transformation.

Data Availability Statement DA

The data that support the findings of this study are available from the corresponding author upon reasonable request.