Nuclear respiratory factor 1 promotes the progression of EBV-associated gastric cancer and maintains EBV latent infection

This study aimed to investigate the association of Epstein-Barr virus (EBV) with nuclear respiratory factor 1 (NRF1) and the biological function of NRF1 in EBV-associated gastric cancer (EBVaGC). Western blot and qRT-PCR were used to assess the effect of latent membrane protein 2A (LMP2A) on NRF1 expression after transfection with LMP2A plasmid or siLMP2A. The effects of NRF1 on the migration and apoptosis ability of GC cells were investigated by transwell assay and flow cytometry apoptosis analysis in vitro, respectively. In addition, we determined the regulatory role of NRF1 in EBV latent infection by western blot and droplet digital PCR (ddPCR). LMP2A upregulated NRF1 expression by activating the NF-κB pathway. Moreover, NRF1 upregulated the expression of N-Cadherin and ZEB1 to promote cell migration. NRF1 promoted the expression of Bcl-2 to increase the anti-apoptotic ability of cells. In addition, NRF1 maintained latent infection of EBV by promoting the expression of the latent protein Epstein-Barr nuclear antigen 1 (EBNA1) and inhibiting the expression of the lytic proteins. Our data indicated the role of NRF1 in EBVaGC progression and the maintenance of EBV latent infection. This provided a new theoretical basis for further NRF1-based anti-cancer therapy.


Introduction
The Epstein-Barr virus (EBV), also known as human herpesvirus 4 (HHV4), is a large double-stranded DNA (dsDNA) virus belonging to the gamma-herpesvirus subfamily. EBV is a well-characterized oncovirus associated with several malignancies, mainly including lymphoma, nasopharyngeal carcinoma (NPC) and EBV-associated gastric cancer (EBVaGC) [1]. Based on the report of The Cancer Genome Atlas (TCGA) in 2014, a new classification was proposed to divide GCs into four subtypes: namely EBVaGC, microsatellite instability tumors, chromosomal instability tumors, and genomic stability tumors [2]. EBVaGC is the most common of EBV-related tumors, with unique genomic aberrations and distinct clinicopathological features [3,4]. In EBVaGC, the latent infection type of EBV is between type I and type II, mainly expressing Epstein-Barr nuclear antigen 1 (EBNA1), EBV-encoded small RNAs (EBERs) and BamHI-A rightward transcripts BARTs [5]. Furthermore, latent membrane protein 2A (LMP2A) expression could be detected in 40% of EBVaGC cases [6].
LMP2A is one of the most important molecules in EBV latency, and it can promote gastric cancer progression in multiple ways. PTEN is a well-known tumor suppressor that could interfere with the migration and growth of GC cells by inhibiting the PI3K/AKT signaling [7]. LMP2A induces phosphorylation of STAT3, which activates DNA methyltransferase 1 (DNMT1), resulting in promoter methylation of PTEN and loss of expression [8]. Moreover, LMP2A could lead to enhanced cell growth and anti-apoptotic effects by activating PI3K/Akt signaling [9,10]. LMP2A could also induce IL-6 production by activating NF-κB and STAT, leading to cell growth and survival [11,12]. LMP2A plays an important role in regulating cell function. LMP2A could Edited by Hartmut Hengel. promote cell migration by mediating Notch signaling and inhibit apoptosis by transforming growth factor β1 (TGFβ1) [13,14].
Nuclear respiratory factor 1 (NRF1) is a transcription factor, usually acting as a transcription activator [15]. NRF1 could activate the expression of various nuclear genes necessary for mitochondrial biogenesis and function. In addition, target genes of NRF1 also play important roles in RNA metabolism, splicing and cell cycle control [16]. Moreover, some studies have shown that NRF1 plays a crucial role in the transcriptional initiation of genes encoded by some viruses. For example, NRF1 is one of the key factors to start transcription at one of the major transcriptional start sites of hepatitis B virus (HBV) X mRNA [17]. With the deepening of research, the role of NRF1 in tumorigenesis has also been studied. In prostate cancer, overexpression of NRF1 inhibits cell proliferation and migration [18]. In breast cancer, the increased expression of NRF1 is a potential biomarker for poor prognosis [19]. Our laboratory found that EBV could upregulate the expression of NRF1, but the specific regulation mode and biological function of NRF1 in gastric cancer were unclear. This paper aims to study the regulatory effect of LMP2A on NRF1 in EBVaGC, and whether NRF1, as a transcription factor, also reflects the transcription of EBVencoded genes.

Cell culture, and treatments
In this study, we used the EBV-positive gastric cancer cell lines SNU719 and GT38. SNU719 cells were provided by Professor Qian Tao (The Chinese University of Hong Kong). GT38 cells were a kind gift of Takeshi Sairenji (Tottori University, Japan). MGC803 and HGC27, EBV-negative gastric cancer lines, were provided by Lei Jiang (Peking University). All cell lines were cultured in DMEM (Gibco, USA) providing 10% fetal bovine serum (Biological Industries, Israel), and 2% penicillin-streptomycin, at 37 °C with 5% CO 2 .

Western blot analysis
All cells were washed twice by cold phosphate-buffered saline (PBS) and lysed in radioimmunoprecipitation assay (RIPA) buffer mixture (RIPA: phenylmethylsulfonyl fluoride [PMSF]: phosphatase inhibitors, 100: 1:1) on ice for 20 min. The cell lysate was then centrifuged at 12,000 rpm at 4 °C for 20 min. The centrifuged supernatant was mixed with 4 × loading buffer and boiled for 5 min to denature the proteins. The proteins were separated by 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred onto a polyvinylidene difluoride (PVDF) membrane (Millipore, USA) by a transfer device. After transfer, the PVDF membranes were blocked in 5% non-fat milk for 2 h at room temperature. Next, the primary antibody was incubated overnight at 4 °C, and the second antibody was incubated for 2 h at room temperature the next day. The proteins of interest were visualized using an enhanced chemiluminescence detection system. The following specific antibodies were used: anti-β-actin, anti-NF-κB (p65), anti-Phospho-NF-κB (p-p65), anti-vimentin, anti-caspase3, anti-Bcl-2 and HRP-linked anti-mouse/rabbit secondary antibody, all of which were purchased from CST (Mass., USA). Anti-NRF1, anti-E-cadherin, and anti-N-cadherin antibodies were purchased from Abcam (USA); anti-ZEB1 was purchased from Proteintech (China); anti-Bax was purchased from Bioss (China); anti-EBV gp350 antibody was purchased from SinoBiological (China); and anti-EBV VCA antibody was purchased from GeneTex (USA). All these antibodies were used at a dilution of 1:1000. Anti-BZLF1, and anti-EBNA1 antibodies were purchased from Santa Cruz Biotechnology Inc. Antibodies (USA) and used at a dilution of 1:200.

Overexpression of LMP2A and NRF1
Recombinant expression plasmid (Hanbio, China) carrying LMP2A was constructed, and its vector structure was pcDNA3.1-EF1A-MCS-3Flag-CMV-EGFP. NRF1 plasmid was produced by GeneChem Co. (China), and its vector structure was Ubi-MCS-3FLAG-CBh-gcGFP-IRESpuromycin. Plasmid transfection was performed using the Lipofectamine 2000 transfection reagent (Invitrogen, China) according to the manufacturer's protocols. A total of 2.5 μg of plasmid DNA per six-well plate was used for transfection. The cells were harvested after 48 h. Transfection efficiency was measured by qRT-PCR or Western Blot.

Migration assay
In the transwell migration assay, the transwell chamber was topped with a non-coated membrane with 8-μm pores (Corning, USA). The cells were diluted by using serumfree medium, and 200 μl cell suspension (MGC803 and HGC27: 3×10 4 cells, SNU719: 1×10 5 cells) was added to the upper chamber and 750 μl of culture medium containing 20% fetal bovine serum was added to the lower chamber. After incubation for 48 h, the cells were fixed with methanol, stained with hematoxylin, and counted under a microscope at 200×magnification. All experiments were performed in three independent replicates.

Cell apoptosis detection
For apoptosis assays, 5×10 5 differently treated cells were seeded into 12-well plates, cultured for 48 h, washed twice with PBS, and then tested using Annexin V-FITC apoptosis assay kit (BD Biosciences, USA) and APC Annexin V (SIMUBIOTECH, China) according to the manufacturer's instructions. The cells were re-suspended in 500 μl binding buffer containing Annexin V-FITC and propidium iodide (PI). After 15 min of incubation at room temperature in the dark, these cell suspensions were sorted by fluorescence-activated cell sorting (FACS) using CytExpert (Beckman, USA). In addition, cisplatin at a concentration of 20 μM was added to MGC803 cells to improve the apoptosis rate. Three independent replicates were performed for all experiments.

Luciferase reporter gene assay
In this paper, the JASPAR website (http:// jaspar. gener eg. net/#) was used to predict the binding sites of NRF1 and EBNA1 promoter BamHI-Q promoter (Qp), and we found two binding sites and constructed the corresponding Qp wild-type and mutant reporter plasmids with pGL3-Basic as vectors (Tsingke, China). In addition, pRL-TK (Tsingke, China) was used as a control for transferring into the construction plasmid. To test the effect of NRF1 on the expression of EBNA1, wild-type plasmids (pGL3-Basic-Qp-WT) or mutated plasmids (pGL3-Basic-Qp-MUT) were cotransfected with NRF1 plasmid or NC plasmid into HEK-293 T cells. And all constructed plasmids were confirmed by sequencing. Cell lysates were harvested 48 h after transfection, and luciferase activity was detected using the dual luciferase reporter system (Promega, China).

Determination of EBV DNA copy number
DNA was extracted from EBV-positive gastric cancer cell lines transfected with NRF1 and NC plasmids according to the instructions of the DNA extraction kit (TianGen, China). EBV copy number was determined by detecting EBV BamH I-W fragment using Droplet digital PCR (ddPCR). The primers were used for BamH I-W: 5′-CCA GAC AGC AGC CAA TTG TC-3′, and 5′-GGT AGA AGA CCC CCT CTT AC-3′. The ddPCR reaction mix was obtained by mixing 10 μl of 2X QX200 ddPCR EvaGreen Supermix (Bio-Rad Laboratories, #1864033), 0.5 μl of 10 μM Fwd/Rev primer mixture, 8 μl of RNase/DNase free water and 1ul DNA, resulting in a final reaction volume of 20 μl. The resulting reaction system together with Droplet Generation Oil (Bio-Rad Laboratories) yielded 40 μl of droplets reaction mixture through a QX200 droplet generator (Bio-Rad Laboratories). Then, the droplets were transferred to a 96-well plate and amplified in a T100 Thermal Cycler amplification apparatus (Bio-Rad Laboratories) under the following conditions: 95 °C for 10 min, 95 °C for 30 s, for 40 cycles of expansion at 60 °C for 1 min (annealing), the microdroplet was stabilized at 90 °C for 5 min followed by infinite maintenance at 4 °C. After amplification, we read the positive and negative droplets through the QX200 droplet reader (Bio-Rad Laboratories). All experiments were performed in three independent replicates.

Statistical analysis
Data were analyzed with using Student's t-test. Analyses were performed using GraphPad Prism software (Graph-Pad Software, USA). Differences were considered statistically significant if P < 0.05. All data were expressed as the mean ± standard error of the mean (SEM), and all experiments were repeated at least three times.

LMP2A upregulates the expression of NRF1
In order to explore the effect of EBV on the expression of NRF1, we stably transfected LMP2A plasmid in the EBVnegative gastric cancer cell lines MGC803 and HGC27 (MGC803-LMP2A and HGC27-LMP2A). The transfection efficiency was estimated based on the proportion of cells expressing green fluorescent proteins (GFP) from the plasmid and showing green fluorescence (Fig. 1a). In addition, LMP2A and NRF1 expression levels at the transcriptional level were detected by qRT-PCR. LMP2A was effectively overexpressed in both MGC803 and HGC27, but at the transcriptional level, LMP2A did not regulate NRF1 significantly (Fig. 1b, d). However, the protein expression level of NRF1 increased compared with the control group (Fig. 1c,  e). To further verify the regulation of NRF1 by LMP2A, we further interfered with LMP2A in EBV-positive gastric cancer cells SNU719. In Fig. 1f, we detected a decrease in the transcription level of LMP2A by qRT-PCR, but the regulatory effect of LMP2A on NRF1 was not obvious at the transcription level. But after LMP2A in SNU719 was interfered, the protein level of NRF1 decreased (Fig. 1g).
In conclusion, LMP2A upregulates NRF1 expression at the protein level.

LMP2A regulates NRF1 expression through the NF-κB pathway
The data above suggested that LMP2A could upregulate NRF1 expression. There are several motifs at the N-terminus of LMP2A that are docking sites for the tyrosine kinases Lyn, Syk and the ubiquitin ligase Nedd4/Itchy [20], which activate PI3-K/Akt [10], NF-κB [21] and Wnt/β-catenin [22] signaling pathway, promote cell growth, inhibit apoptosis and differentiation, and contribute to cell transformation. S. Miranda et al. found that NRF1 could be upregulated and exert oxidative stress when NF-κB pathway was activated [23]. Therefore, to further explore how LMP2A regulates NRF1, we examined the expression levels of the NF-κB pathway-related proteins p65 and phosphorylated p65 (p-p65) in HGC27-LMP2A and MGC803-LMP2A. p65 is one of the key molecules of the NF-κB pathway, and its phosphorylation represents the activation of the NF-κB classical pathway. We found that the expression level of p-p65 was significantly increased compared with the control group ( Fig. 2a-b). In addition, after treatment with siLMP2A in SNU719, the expression level of p-p65 was also significantly reduced (Fig. 2c). Furthermore, we treated SNU719 with the NF-κB pathway inhibitor BAY11-7085 and cultured cells at concentrations of 2 and 4 μM, respectively, for 24 h. Western blot showed that NRF1 expression was inhibited in a dosedependent manner (Fig. 2d). To explore the key molecules of the NF-κB pathway that affect NRF1 expression, we treated SNU719 with sip65. The results showed that the expression of NRF1 was decreased when p65 was interfered (Fig. 2e). These two partial results show that LMP2A upregulates the expression of NRF1 by activating the NF-κB pathway.

NRF1 affects the ability of cell migration in EBVaGC
We transfected NRF1 plasmid in MGC803 and HGC27 (MGC803-NRF1 and HGC27-NRF1). The transfection efficiency was estimated based on the proportion of cells expressing GFP from the plasmid and showing green fluorescence (Fig. 3a). Transwell assay showed that NRF1 expression significantly promoted the migration of MGC803-NRF1 and HGC27-NRF1 (Fig. 3b-c). To further validate our above results, SNU719 was treated with siNRF1. Consistently, the cell migration ability of SNU719-siNRF1 was also reduced compared with the control group (Fig. 3d). In addition, we also detected changes in proteins associated with Epithelial-mesenchymal transition (EMT). Western blot results showed that N-cadherin and ZEB1 expression was increased in MGC803-NRF1 and HGC27-NRF1, while E-cadherin expression was decreased, compared with the control group (Fig. 3e-f). After interference with the expression of NRF1 in SNU719, the expression of N-cadherin and ZEB1 decreased, while the expression of E-cadherin increased (Fig. 3g). In conclusion, NRF1 enhances the ability of cell migration in GC.

NRF1 affects the anti-apoptotic ability of cells in EBVaGC
To explore the effect of NRF1 on anti-apoptotic ability of cells, we used flow cytometry to detect the level of cell apoptosis. As shown in Fig. 4a, when MGC803 was induced by cisplatin, the level of apoptosis was significantly increased to more than 20%. However, after overexpression of NRF1, the level of cell apoptosis was decreased. When HGC27 was induced by cisplatin, the apoptosis rate was around 15%. Consistent with the results in MGC803, apoptosis level of HGC27-NRF1 was also significantly downregulated compared to the control group (Fig. 4b). Whereas the number of apoptotic cells increased after treatment with siNRF1 in SNU719 (Fig. 4c). To investigate which protein changes affect the anti-apoptotic ability of cells, we examined the expression of apoptosis-related proteins by Western Blot.
Compared with the control group, the expression of caspase3 was decreased and the expression of Bcl-2 was increased in MGC803-NRF1 and HGC27-NRF1 (Fig. 4d-e). After treatment of SNU719 with siNRF1, western blot showed that caspase3 expression was increased while Bcl-2 expression was decreased (Fig. 4f). These results were consistent with the results of flow cytometry apoptosis described above. As well, accordingly, these results indicate that NRF1 enhances the anti-apoptotic ability of cells in GC.

NRF1 maintains EBV latent infection
We further investigated the effect of NRF1 on EBV latency. After overexpression of NRF1 in SNU719 and GT38, Western blot results showed that the latent protein EBNA1 expression was increased in SNU719-NRF1 and GT38-NRF1, whereas the lytic proteins BZLF1, BRLF1, gp350 and VCA expression were decreased compared with the control group ( Fig. 5a-b). In addition, we also examined the effect of NRF1 on EBV-encoded genes at  (d, e) The protein level of Bcl-2, Bax, cas-pase3 and NRF1 in MGC803-NRF1 and HGC27-NRF1 cells detected by Western blot (***, p < 0.001; **, p < 0.01; ns: not significant). (f) The protein level of Bcl-2, Bax, caspase3 and NRF1 in SNU719-siNRF1 cells detected by Western blot (***, p < 0.001; **, p < 0.01; *, p < 0.05; ns: not significant). Data were representative of three independent experiments the transcriptional level. qRT-PCR results showed that overexpression of NRF1 upregulated the transcription of EBNA1, but not the expression of BZLF1 and BRLF1 (Fig. 5c-d). To make our results more comprehensive and accurate, we measured the DNA copy number of EBV by ddPCR. As shown in Fig. 5e-f, the viral copy numbers of SNU719-NRF1 and GT38-NRF1 was significantly lower than the control groups. To further explore whether NRF1 could regulate EBNA1's promoter Qp, we predicted two binding sites based on the JASPAR website (Fig. 5g). In Fig. 5h, dual luciferase assays showed NRF1 significantly upregulated the activity of Qp-WT compared to the control group and NRF1 was combined with the second binding site. After the first binding site mutation, the effect of NRF1 on Qp activity was not significantly different from that before the mutation. In summary, NRF1 maintains latent infection of EBV by increasing transcription levels of EBNA1 and downregulating the expression levels of some lytic proteins.

Discussion
NRF1 was initially recognized as being involved in the transcription of mitochondrial-associated complex genes [15]. Recent evidence also suggested that the NRF1 protein interacted with broad-spectrum transcription factors; Its unique DNA binding recognition site was one of the seven transcription factor binding sites most commonly found in the proximal promoter of ubiquitous genes [24]. NRF1 motifs were present on gene promoters that regulate the cell cycle, chromatin structure, apoptosis, cell adhesion/invasion, DNA repair, DNA methylation and transcriptional inhibition signaling, and epithelial adhesion junctions [25]. These findings suggested that NRF1 was a versatile protein that plays a role in various cellular functions. With the deepening of research, NRF1 has been found to play a role in the progression of some tumors. In human melanoma, NRF1 interacts with the metastasis suppressor KISS1 to promote mitochondrial biogenesis, which provides more energy to promote cancer cell survival [26]. In addition, the translocation of NRF1 into the nucleus could upregulate the expression of some cell cycle genes in 17β-estradiol (E2)-induced breast cancer cell line MCF-7 [27]. In breast cancer, NRF1 could also promote its transcription by interacting with the promoter of CXCR4, and NRF1 could also control CXCR4 expression by generating ROS signals. These further promote the development of breast cancer [28]. The results of our laboratory preliminary experiments found that LMP2A upregulated NRF1 expression via PI3K/Akt signaling pathway in EBVaGC [29]. In this study, NRF1 could also be upregulated by LMP2A, and whether this affects the progression of EBVaGC is worth exploring.
Hagir et al. demonstrated that NRF1 could be transcribed by NF-κB activation in vivo [30]. In GC, transfection of the LMP2A gene into EBV-negative GC cells resulted in an increase in phosphorylated IκBα (p-IκBα) and p65, accompanied by a decrease in total IκBα expression, leading to constitutive activation of NF-κB [31]. In addition, Ryan et al. used DNA binding assays for the first time to demonstrate that LMP2A increases the nuclear localization of the NF-κB p65 subunit [32]. Therefore, in this article we verified that the LMP2A/NF-κB/NRF1 axis hold in EBVaGC. Constitutive activation of NF-κB is known to significantly promote malignant progression of EBVaGC by upregulating genes involved in proliferation and anti-apoptosis. As a transcription factor, NRF1 is highly active in a variety of cancers, especially in breast cancer, where its aberrant expression contributes to brain infiltration by breast cancer stem cells (BCSC). It has been reported that NRF1 could enhance the adhesion and migration ability of BCSC, and help the cancer cells to cross the blood-brain barrier (BBB) for brain metastasis [33]. In addition, the high mitochondrial biological energy caused by high expression of NRF1 contributed to the invasion of esophageal squamous cell carcinoma [34]. Apoptosis is an important cellular program that determines the fate of cells not only during embryonic development, but also in the process from normal to transformed phenotype [35]. In H9C2 cardiomyocytes, NRF1 protected them from hypoxia-induced apoptosis via the mitochondrial pathway and death receptor pathway [36]. In addition, NRF1 protected against CoCl 2 -induced cardiomyocyte apoptosis, possibly by enhancing mitochondrial function against CoCl 2 -induced cell damage [37]. In this paper, we also verified the biological properties of NRF1 in EBVaGC by knocking down or overexpressing NRF1. In GC, NRF1 promotes cell migration and inhibits apoptosis. NRF1 is beneficial to the development of EBVaGC, and targeting NRF1 seems to be a potential strategy for the treatment of EBVaGC.
More importantly, the activation of the NF-κB pathway in some EBV-related cancers could upregulate the expression of EBV latent genes to promote latent viral infection and inhibit viral lysis and reactivation [38]. EBV has a biphasic infection mode: latent infection and lytic infection. In the latent state, the genomic DNA exists in the form of episomes in the nucleus, in which the closed circular plasmid binds to histone proteins, so that only a limited number of viral latent genes are expressed [39]. EBV establishes latent infection in EBV-associated tumors in the presence of latent proteins. In lytic phase, EBV mainly expressed immediate early protein BZLF1 and BRLF1, as well as early protein and late protein VCA, gp350, etc. BZLF1 expression controls the transition from latent to lytic infection. NRF1 could stably regulate ZEB1, and ZEB1 has been shown to be a negative regulator of BZLF1 [40]. Therefore, we further explored whether NRF1 could affect the latent and lytic infection of EBV. In this paper, NRF1 could maintain latent infection of EBV. In particularly, it could upregulate EBNA1 expression at both the transcriptional level and the protein level. It has been shown that EBNA1 could bind to recognition sites in oriP, leading to the emergence of local regions of DNA bending and helical overwinding and underwinding, which played an important role in DNA replication of EBV [41,42]. Therefore, in addition to the above experiments, we further examined the effect of upregulation of EBNA1 by NRF1 on the DNA copy number of EBV. We found that NRF1 reduced EBV replication compared to controls. Given that NRF1 was a powerful transcription factor. NRF1 could actively promote the transcription of the X gene in HBV [17]. In addition, NRF1 could bind to the core promoter of the CD155 gene [43]. Therefore, we tried to verify whether NRF1 could regulate the Qp of EBV. In this paper, we predicted the binding sites of NRF1 to EBNA1 promoter Qp. Qp is a TATA-cassette-free housekeeper-type promoter driven by the EBNA1 latency that maintains latent infection of EBV in EBV-associated tumors [44,45]. As one of the most important promoters of EBV, Qp is also widely studied. Verhoeven et al. found that both p65 and p50 could be combined with Qp in NPC [46]. In addition, heat shock factor 1 (HSF1) binding to heat shock elements in Qp could also induce EBNA1 expression [47]. We verified the binding of NRF1 to Qp by double luciferase reporter gene experiments and found that NRF1 could promote transcription of Qp. In conclusion, NRF1 promotes the progression of EBVaGC and maintains EBV latent infection, so downregulation of NRF1 expression seems to be a potential strategy for the treatment of EBVaGC.

Conclusions
In conclusion, LMP2A upregulated the expression of NRF1 through NF-κB signaling. In turn, NRF1 promoted the expression of EBNA1 and facilitated the latent infection of EBV. In addition, NRF1 promoted the development of EBVaGC. Therefore, inhibiting the activation of NF-κB pathway and downregulating the expression of NRF1 seem to be potential strategies for the treatment of EBVaGC.