Derived LMP1 in Balb/3T3 Cells Promotes Stress-Induced Cell Death and Impairs LMP1-Mediated Transformation

Induction of inducible nitric oxide synthase by Epstein-Barr virus B95-8- derived LMP1 in Balb/3T3 cells promotes stress-induced cell death and impairs LMP1-mediated transformation

Jau-Song Yu*,1, Hsing-Chen Tsai1, Chih-Ching Wu1, Li-Ping Weng1, Hsin-Pai Li2, Pei-Jung Chung2 and Yu-Sun Chang2

1Department of Cell and Molecular Biology, Institute of Basic Medical Science, Medical College of Chang Gung University, Tao- Yuan, Taiwan, Republic of China; 2Department of Microbiology and Immunology, Institute of Basic Medical Science, Medical College of Chang Gung University, Tao-Yuan, Taiwan, Republic of China

The latent membrane protein 1 (LMP1) of Epstein-Barr Introduction virus (EBV) causes cellular transformation and activation of several intracellular signaling events. In this Nitric oxide (NO), a short-lived free radical produced report, we show that BLMP1 (encoded by the LMP1 by many cell types, is recognized as an important gene derived from the B95-8 strain of EBV) triggers the biological mediator in diverse physiological and expression of inducible nitric oxide synthase (iNOS) in pathological processes (Moncada et al., 1991; Nathan, Balb/3T3 fibroblasts. Intriguingly, NLMP1, a natural 1992). The compound is synthesized from L-arginine in sequence variant of LMP1 identified in EBV-positive an oxidative reaction catalyzed by nitric oxide synthase nasopharyngeal carcinoma biopsy, does not similarly (NOS) isozymes, which generates NO and L-citrulline induce iNOS expression. BLMP1-induced iNOS in (Moncada et al., 1991). Three NOS isoforms, namely Balb/3T3 cells is active to produce nitric oxide (NO), neuronal NOS (nNOS or NOS1), inducible NOS and NO production can be blocked by several iNOS (iNOS or NOS2), and endothelial NOS (eNOS or inhibitors. When subjected to environmental stress, Balb/ NOS3), have been identified and characterized (Bredt 3T3 cells that produce NO lose viability more rapidly et al., 1990; Marsden et al., 1992; Nathan and Xie, than non NO-producing cells. Blockage of NO genera- 1994). The nNOS and eNOS enzymes are constitutively tion by iNOS inhibitors enhances the viability of NO- expressed, whereas iNOS is induced de novo in many producing cells under stress conditions. The activities of cell types by bacterial cell wall lipopolysaccharides, caspase-3 and c-Jun N-terminal kinase, two important environmental stress and various cytokines, resulting in regulators mediating stress-induced apoptosis, are signif- the production of large amounts of NO. For example, icantly potentiated following heat shock treatment of a well-characterized cytokine, tumor necrosis factor- BLMP1-expressing/NO-producing cells, compared to a(TNF-a), induces iNOS expression in cells, either parental and NLMP1-expressing cells. Furthermore, independently or via synergistic action with other treatment with iNOS inhibitor augmented the cloning inflammatory cytokines such as IL-1b and INFg efficiency (in culture) and tumor growth (in nude mice) of (Kleinert et al., 1996). In mouse macrophages, BLMP1-expressing/NO-producing cells. Collectively, cooperative signaling by TNF receptor 1 (TNFR1) the results demonstrate that BLMP1 induces iNOS and TNF receptor 2 (TNFR2) is required for the expression and NO production in Balb/3T3 cells, which maximal induction of iNOS by TNF-a (Riches et al., leads to the alteration of cell functions, including 1998). sensitivity to environmental stress, capability to colonize NOS expression has been linked to tumorigenesis. independent of anchorage and tumorigenicity in nude Elevated expression of NOS isoforms was observed in mice. Our data additionally implicate that the differ- human gynecological, breast and central nervous ential iNOS induction potential of the two LMP1 forms system tumors, colon adenomas, esophageal adenocar- may represent the basis of a functional difference cinomas, melanomas and prostate cancer (Thomsen et between the two LMP1 proteins. al., 1994, 1995; Cobbs et al., 1995; Ambs et al., 1998a; Oncogene (2002) 21, 8047 – 8061. doi:10.1038/sj.onc. Wilson et al., 1998; Aaltoma et al., 2001; Massi et al., 1205990 2001). In several of these cases, iNOS expression correlated well with tumor grade and could be detected Keywords: LMP1; iNOS; nitric oxide; caspase-3; JNK in tumor endothelial cells, tumor-infiltrating immune cells, and/or tumor cells themselves, all of which are important regulators in the growth of solid tumors. Animal studies showed that a human colon adeno- *Correspondence: Dr J-S Yu; E-mail: [email protected] carcinoma genetically engineered to express iNOS Received 29 October 2001; revised 20 August 2002; accepted 20 constitutively grew faster than the wild-type cells in August 2002 nude mice (Jenkins et al., 1995), and inhibition of iNOS induction by LMP1 of EBV J-S Yu et al 8048 iNOS by a selective inhibitor reduced the growth of transmembrane region and a 200 amino acid C- this xenografted tumor (Thomsen et al., 1997). These terminal cytoplasmic effector domain (Liebowitz et observations implicate an oncogenic role of iNOS in al., 1986). Molecules such as TNFR-associated factors human cancer. On the other hand, although low levels (TRAF) and TNFR-associated death domain protein may prevent apoptosis of malignant cells and enhance (TRADD) interact directly with the C-terminal domain tumor growth (Ambs et al., 1998b), increased produc- of LMP1 (Mosialos et al., 1995; Devergne et al., 1996; tion of endogenous NO by tumor cells transfected with Izumi and Kieff, 1997). These interactions are essential a functional iNOS gene has autotoxic effects, causing for LMP1-mediated activation of the transcription the suppression of tumor growth and metastasis via factor, NF-kB (Devergne et al., 1996; Sylla et al., triggering of apoptosis (Xie and Fidler, 1998). 1998; Kieser et al., 1999). Furthermore, kinase path- Recently, a regulatory role for NO in controlling ways, including NF-kB-inducing kinase (NIK) and c- viral reactivation in EBV-infected cells was reported Jun N-terminal kinase (JNK), responsible for the (Mannick et al., 1994; Kawanishi, 1995). Low levels of activation of transcription factors, NF-kB and AP-1, iNOS were detected in EBV-infected B cells and respectively are activated by LMP1 (Kieser et al., 1997; epithelial cells, and NO production inhibited apoptosis Sylla et al., 1998). and reactivation of EBV virus (Mannick et al., 1994; At least two major variants of LMP1 genes have Gao et al., 1999). However, to date, the identity of the been reported, one derived from the B95-8 strain viral protein responsible for iNOS induction is unclear. (designated BLMP1), and the other from NPC biopsies Epstein-Barr virus (EBV) is a human herpes virus (NLMP1) (Hu et al., 1991; Chen et al., 1992). NLMP1 closely associated with several malignancies, including mainly differs from BLMP1 by a 10 amino acid Burkitt’s lymphoma (Esptein et al., 1964; Henle et al., deletion in the C-terminal region (Chen et al., 1992). 1968), nasopharyngeal carcinoma (NPC) (Klein et al., Despite 95% sequence identity with BLMP1, NLMP1 1974; Pathmanathan et al., 1995), B-cell lymphoproli- is more tumorigenic than its homolog, as assessed in ferative disease in immunosuppressed patients (Hanto the rodent model (Chen et al., 1992; Li et al., 1996) et al., 1985), peripheral T-cell lymphoma (Su and and human keratinocytes (Hu et al., 1993). Further- Hsieh, 1992), and gastric carcinoma (Iwasaki et al., more, the 10-amino acid deletion at the C-terminus 1998). EBV is maintained in the latent stage, which appears to be essential for the higher transforming allows expression of a limited number of viral gene ability of NLMP1, compared to BLMP1 (Li et al., products (Kieff and Liebowitz, 1990). These gene 1996). These results indicate that sequence variation in products include EBV nuclear antigen 1 (EBNA1), the two LMP1 forms may have a profound effect on which is essential for maintenance of the EBV episomal the biological activity of the proteins. genome in EBV-positive cells, and latent membrane Since TNFR1 and 2 are responsible for iNOS protein 1 (LMP1), present in more than 60% NPC and induction in cells by TNF-a, and LMP1 mimics the EBV-associated Hodgkin’s lymphoma (Fahraeus et al., TNFR superfamily, it is possible that LMP1 has a 1988). LMP1 has oncogenic potential, since it trans- similar potential to trigger iNOS expression. In this forms rodent fibroblast cells in vitro (Wang et al., 1985; report, we confirm that LMP1 induces iNOS expres- Baichwal and Sugden, 1988), and induces B cell sion, and consequently the production of large lymphoma in transgenic mice model in vivo (Kulwichit amounts of NO in Balb/3T3 cells. Interestingly, the et al., 1998). LMP1 expression additionally affects the induction potential of BLMP1 is much greater than growth characteristics of B cells, epithelial cells and that of NLMP1. Using biochemical and clongenic hematopoietic stem cells (Dawson et al., 1990; assays, we characterize the effects of iNOS induction/ Fahraeus et al., 1990; Peng and Lundgren, 1992; NO production on Balb/3T3 cells. Furthermore, we Fairbairn et al., 1993). Alteration of cell function by examine the alteration of tumorigenicity of Balb/3T3 LMP1 may be due to its ability to induce the cells in nude mice when iNOS is induced by LMP1. expression of numerous proteins, including B cell markers (CD23 and CD40), vimentin, cell adhesion Results molecule (ICAM1), epithelial growth factor receptor (EGFR), interleukin-10, cyclin D2, a calcium-camodu- LMP1 induces iNOS expression in Balb/3T3 cells lin-dependent protein kinase, and cyclooxygenase-2 (Birkenbach et al., 1989; Wang et al., 1990; Nakagomi Inducible NO synthase (iNOS) is triggered by tumor et al., 1994; Mosialos et al., 1994; Arvanitakis et al., necrosis factor-a (TNF-a) in Balb/3T3 cells (Kleinert et 1995; Murono et al., 2001). Furthermore, up-regulation al., 1996), resulting in the production of NO. Nitric of Bcl-2 and A20 by LMP1 confers anti-apoptosis oxide is toxic at levels above a certain threshold in activity in B cells (Henderson et al., 1991; Laherty et cells. LMP1 is an integral membrane protein with al., 1992). biological functions mimic to the TNFR superfamily. The pleiotropic and profound effects of LMP1 on However, LMP1 is constitutively activated, in contrast diverse biological processes appear to be due to its to the TNF receptor, which requires binding to TNF-a comparable properties to a constitutively active TNFR for activation. In order to examine whether LMP1 superfamily member (Gires et al., 1997; Eliopoulos and induces iNOS similar to TNF-a, unselected mixed Rickinson, 1998). LMP1 is a 62 kDa polypeptide with G418-resistant cells generated by the transfection of a 25 amino acid N-terminus, a 161 amino acid expression vectors containing the two variants of

Oncogene iNOS induction by LMP1 of EBV J-S Yu et al 8049 LMP1, BLMP1 and NLMP1, together with a neo- exposed for a long time (data not shown). Expression gene-containing vector into Balb/3T3 cells, were of LMP1 proteins in these cells was confirmed by analysed for iNOS expression by Western blotting. immunoblot analyses with the S12 monoclonal anti- As shown in Figure 1a (left panel), BLMP1 induced body (Figure 1a, right panel). high levels of iNOS in cells, following 2 week selection Similar results were obtained when randomly with G418. However, NLMP1 did not induce selected individual cell clones stably expressing BLMP1 detectable levels of iNOS, unless the immunoblot was and NLMP1 were tested. Results displayed in Figure 1b show that iNOS expression was induced in BLMP1 clones (D4, D11 and D13), but not NLMP1 clones (E7, E9 and E14). As a positive control, high-level expression of iNOS was induced by TNF-a in Balb/ 3T3 cells (Figure 1b, lower panel, lane TNF-a). In agreement with the above results, iNOS was additionally detected at the transcription level. The iNOS mRNA was highly expressed in BLMP1- expressing D4 and D11 clones, but not in NLMP1- expressing E9 and E14 cultures (Figure 1c). To exclude the possibility that induction of iNOS by BLMP1 in Balb/3T3 cells occurs fortuitously during the clonal selection process, transient transfection experiments were performed. The RT – PCR data in Figure 1d confirm that transient transfection of LMP1 expression plasmids into Balb/3T3 cells induces iNOS mRNA, specifically, to a significantly higher level when the BLMP1-expressing vector is employed. The results demonstrate that LMP1 proteins induce iNOS expression, and the induction potential of BLMP1 is much greater than that of NLMP1.

NO production is significantly augmented in BLMP1- expressing Balb/3T3 cells Induction of iNOS results in NO production, which can be assayed by measuring the end products, nitrate and nitrite. To investigate whether this reaction occurs in BLMP1-expressing cells, the concentrations of nitrate and nitrite were determined in 3-day cultured Figure 1 Induction of iNOS by LMP1 in Balb/3T3 cells. (a) Ex- medium of D11 and D13 cells. Our results (Table 1) tracts (60 mg) of parental and unselected/mixed clones of BLMP1- show that 4.41 and 0.95 nmol nitrate and nitrite/ml or NLMP1-expressing Balb/3T3 cells were separated on a 10% were released in the medium from D11 and D13 SDS-gel, and electroblotted onto PVDF membrane. The mem- 5 brane was probed with anti-iNOS antibody (left panel) and (3610 cells), respectively. These ﬁndings correlate well monoclonal antibody, S12, against LMP1 (right panel) individu- with the higher levels of iNOS observed in D11 cells ally. Positions of LMP1 and iNOS are indicated. Lane P, parental than D13 cells (Figure 1b). On the other hand, cells; lane B, BLMP1-expressing cells; lane N, NLMP1-expressing parental, 3T3neo and NLMP1-expressing cell clones, cells. Molecular weights (kDa) of markers are shown on the left of both panels. (b) Parental Balb/3T3 cells were treated with or without TNF-a (20 ng/ml) at 378C for 24 h. Extracts (60 mg) of parental cells, TNF-a-pretreated parental cells, BLMP1- and Table 1 Nitric oxide production in Balb/3T3 cells stably expressing NLMP1-expressing cell clones were subjected to immunoblotting LMP1 analyses with a monoclonal antibody against LMP1 (S12) and Nitrate plus nitratea anti-iNOS antibody, respectively. Lane P, parental cells; lanes Cell Clone (nmol/ml medium) D4, D11 and D13, BLMP1-expressing cells; lanes E7, E9 and E14, NLMP1-expressing cells; lane TNF-a, TNF-a-treated paren- Parental cells – NDb tal cells. (c) Poly (A)+mRNA (10 mg) isolated from various cell 3T3neo cells – ND clones was resolved by electrophoresis in a formaldehyde/1% BLMP1-expressing cells D11 4.41 agarose gel and transferred to a nitrocellulose membrane. After D13 0.95 UV cross-linking, the membrane was hybridized with the 32P-la- NLMP1-expressing cells E7 ND beled probe for iNOS or b-actin, and the hybridized signal was E9 ND detected as described in Materials and methods. (d) Balb/3T3 cells (36106) were transiently transfected with 1 mg pCMV vector, Parental, 3T3neo and LMP1-expressing cells (36105) were plated on pCMVBLMP1 or pCMVNLMP1 expression plasmids by electro- 35-mm dishes containing 2 ml DMEM. The medium was replaced by poration (BioRad electroporator) at 0.24 KV. Transfected cells 2 ml DMEM without phenol red, 4 h after plating. Culture medium were allowed to grow for 24 and 32 h, and harvested for RT – (100 ml) was collected 24 h later, and subjected to nitrate/nitrate PCR analysis using the primers for iNOS, as described in Materi- assays. aNitrate plus nitrate accumulated in the cell culture medium, als and methods bND, not detectable

Oncogene iNOS induction by LMP1 of EBV J-S Yu et al 8050 E7 and E9, did not produce detectable quantities of nitrite/ml in medium from D11 cells as compared to NO under the same assay conditions (Table 1). 3.5 nmol nitrate and nitrite/ml in parental cells without The kinetics of NO production in LMP1-expressing TNF-a treatment), and 3.5-fold higher than in TNF-a- cells was further assessed using the D11 clone treated parental cells (28 nmol nitrate and nitrite/ml expressing the highest levels of iNOS protein and compared to 8 nmol nitrate and nitrite/ml). To certify parental cells treated with or without TNF-a.As that NO production was not due to the higher growth presented in Figure 2a, parental cells only secreted rate of D11 cells, a time-course study was performed. small quantities of NO over 3 days in the absence of The amount of NO produced in D11 cells plated at 0.5, TNF-a, whereas a *2 – 3-fold increase was observed in 1and56105 cells/ml was measured on days 1, 2, 3 and the presence of TNF-a. In contrast, NO production at 4, respectively. As illustrated in the left panel of Figure day 3 in D11 cells was about eightfold higher than that 2b, NO production increased continuously over the 4 in untreated parental cells alone (28 nmol nitrate and days, when cells were plated at low concentrations (0.5

Figure 2 Detection of nitrate/nitrite in the culture supernatant of LMP1-expressing Balb/3T3 cells. (a) Nitrate/nitrite accumulation in the culture supernatant of LMP1-expressing Balb/3T3 cells. Parental Balb/3T3 and LMP1-expressing D11 cells (26105) were seeded in 35-mm culture dishes containing 2 ml DMEM without phenol red. Parental cells were treated with or without TNF-a (20 ng/ml) for 3 days, while D11 cells were left untreated. Culture supernatants (100 ml) were collected every 24 h, and subjected to a nitrate/nitrite assay described in Materials and methods. (b) Time-course of NO production in BLMP1-expressing D11 cells. (Left panel) Various concentrations of BLMP1-expressing D11 cells (1.56106 -&-, 36105 -*-, and 1.56105 cells -*-) were seeded in 35-mm culture dishes containing 3 ml DMEM without phenol red, and cultured for 4 days. Culture supernatants (100 ml) were collected every 24 h, and subjected to a nitrate/nitrite assay. (Right panel) BLMP1- expressing D11 cells (1.56106) were seeded in a 35-mm culture dish containing 3 ml DMEM without phenol red. Culture supernatants (100 ml) were collected every 4 h, and subjected to a nitrate/nitrite assay

Oncogene iNOS induction by LMP1 of EBV J-S Yu et al 8051 and 16105 cells/ml). Higher plating concentrations and did not alter the expression of LMP1 in D11 cells (56105 cells/ml) showed a rapid initial increase, (Figure 3b). Blockage of NO production by iNOS followed by a plateau in NO levels by day 2. The inhibitors in D11 cells may be due to the direct kinetics of the latter was further analysed by measuring inhibition of pre-existing iNOS activity, synthesis of NO production at shorter intervals. NO production in the enzyme, or both. To further evaluate this D11 cells increased linearly for 24 h, and reached possibility, the level of iNOS in D11 cells was maximum levels at *28 to 32 h (Figure 2b, right examined after prolonged treatment with inhibitor panel). compounds. The concentration of iNOS did not change in SMT- and L-NMMA-treated cells (Figure 3b). In contrast, expression of iNOS in Dex-treated Blockage of NO production in LMP1-expressing Balb/ D11 cells was significantly inhibited. These results 3T3 cells by iNOS inhibitors suggest that SMT and L-NMMA block iNOS activity To ensure that augmentation of NO production in in D11 cells, while Dex inhibits iNOS synthesis. Thus, LMP1-expressing Balb/3T3 cells is indeed due to the at least two different mechanisms are potentially enzymatic reaction catalyzed by induced iNOS, the involved in the inhibition of NO production in D11 effects of several known iNOS inhibitors, including cells by these compounds. G N -monomethyl-L-arginine (L-NMMA), S-methylisothiourea (SMT) and dexamethasone (Dex), were LMP1-expressing, NO-producing cells are considerably tested in D11 cells. As shown in Figure 3a, treatment sensitive to environmental stress with any one of these compounds resulted in more than 70% NO reduction within 3 days, indicating that During the maintenance of individual cell clones, it was iNOS induced by LMP1 generates NO. All three observed that BLMP1-expressing and NO-producing compounds had little cytotoxic effect on the growth of cultures always contained a significantly higher D11 cells under our assay conditions (data not shown), proportion of dead cells that detached from the culture dish. This phenomenon was less evident in parental, 3T3neo and NLMP1-expressing 3T3 cells. To investigate whether over-expression of BLMP1 enhances stress-induced cell death, the viability of BLMP1- expressing, NLMP1-expressing, 3T3neo and parental cells was investigated after treatment with various environmental stresses. Following prolonged culturing without replacing culture media, these four groups of cells exhibited similar viability over the initial 2 – 3 days (Figure 4a). However, on the fourth day, only *50% of BLMP1-expressing D11 cells remained viable, and cell death was observed on day 5, in contrast to more than 60% viability of parental, 3T3neo and NLMP1- expressing E9 cells. Similarly, D11 cells lost viability more rapidly than NLMP1-expressing E9, 3T3neo and parental cells after heat shock (Figure 4b) and UV irradiation (Figure 4c). These results collectively indicate that BLMP1-expressing, NO-producing Balb/ 3T3 cells are much more sensitive to environmental stress than NLMP1-expressing, 3T3neo and parental cells.

Potentiation of heat shock-induced activation of caspase-3 in LMP1-expressing NO-producing Balb/3T3 cells Prolonged culturing, heat shock and UV irradiation are typical environmental stress factors that trigger cellular apoptosis. Accordingly, the activity of caspase- 3, an ICE/CED-3 family of cysteine proteases essential Figure 3 Blockage of NO production in BLMP1-expressing for apoptotic execution (Martins and Earnshaw, 1997; Balb/3T3 cells by iNOS inhibitors. (a) BLMP1-expressing D11 cells (26105) seeded in 35-mm culture dishes containing 2 ml Nicholson and Thornberry, 1997), was measured in DMEM without phenol red were treated with L-NMMA BLMP1-expressing, NLMP1-expressing, 3T3neo and (50 mM), S-methylisothiourea (SMT) (25 mM) or dexamethasone parental cells, following stress treatment. The basal (Dex) (25 mM) for 3 days, or left untreated. Culture supernatants activity of caspase-3 was low in all four cell types (100 ml) were collected every 24 h, and assayed for nitrate/nitrite. (b) BLMP1-expressing D11 cells were harvested at day 3 after (Figure 5a). When cells were subjected to heat-shock at drug treatment, and 60 mg cell extracts were prepared for analysis 448C for 30 min, caspase-3 was dramatically activated of LMP1 and iNOS expression by immunoblotting in BLMP1-expressing, NO-producing D11 cells, but

Oncogene iNOS induction by LMP1 of EBV J-S Yu et al 8052

Figure 5 Caspase-3 activity in parental and LMP1-expressing Balb/3T3 cells after heat shock treatment. (a) Parental, 3T3neo, BLMP1-expressing D11 and NLMP1-expressing E9 cells in 100 mm dishes were left untreated, or exposed to heat shock at 448C for 30 min, and incubated at 378C for another 1, 2, 4, and 6 h, respectively. Cells were harvested, and 100 mg cell extracts were analysed for caspase-3 activity, using Z-DEVD-AFC as the fluorogenic substrate, as described in Materials and methods. (b) Parental, 3T3neo and unselected/mixed clones of BLMP1-expressing (Bmix) or NLMP1-expressing (Nmix) Balb/ Figure 4 Differential response of parental and LMP1-expressing 3T3 cells were left untreated, or subjected to heat shock and incu- Balb/3T3 cells to environmental stress. (a) Parental, 3T3neo, bated at 378C for 4 h. Cell extracts (100 mg) were analysed for BLMP1- and NLMP1-expressing Balb/3T3 cells (26105) were caspase-3 activity, as described above. The experiment was re- plated on 35-mm culture dishes containing 3 ml DMEM. Cells peated three times, with reproducible results were cultured for the various time intervals indicated, without replacing media. Viable cells were determined using the trypan blue exclusion method. (b,c) Cells (26105) plated in 35-mm culture dishes for 24 h were exposed to heat shock (448C for 30 min) pivotal for inducing apoptosis (Kyriakis et al., 1994; 2 (b) or UV light (200 J/m )(c), and incubated at 378C for various Derijard et al., 1994; Verheij et al., 1996). JNK was time intervals, as indicated. Cell viability was determined by trypan blue exclusion immunoprecipitated from extracts of parental, 3T3neo, BLMP1-expressing D11 and NLMP1-expressing E9 cells with or without heat shock, followed by a kinase activity assay, using GST-c-Jun as substrate. Substan- not in parental, 3T3neo or NLMP1-expressing E9 cells. tial activation of JNK to comparable levels occurred Similar results were observed in unselected/mixed rapidly within 1 h upon heat-shock treatment of all four clones expressing BLMP1 (Bmix) and NLMP1 (Nmix) groups (Figure 6a). However, interestingly, in parental, (Figure 5b), indicating that the significant caspase-3 3T3neo and E9 cells, levels of JNK activation dropped activation in heat-treated D11 cells is not due to clonal significantly to *10% at 6 h following treatment, variation during the selection process, but is caused by whereas relatively high JNK activity (*40% maximum BLMP1 expression. activation level) persisted in D11 cells at this time-point (Figure 6a). This was additionally observed in unselected/mixed BLMP1-expressing clones, compared to Heat shock induces persistent JNK activation in BLMP1- parental, 3T3neo and NLMP1-expressing cell clones expressing, NO-producing Balb/3T3 cells (Figure 6b). The data indicate that the pattern of JNK In addition to caspase-3 activation, c-Jun N-terminal activation induced by heat shock in D11 cells is distinct kinase (JNK) activated by environmental stress is for BLMP1-expressing cells.

Oncogene iNOS induction by LMP1 of EBV J-S Yu et al 8053

Figure 6 JNK activity in parental and LMP1-expressing Balb/3T3 cells after heat shock treatment. (a) Parental, 3T3neo, BLMP1- expressing D11 and NLMP1-expressing E9 cells cultured in 100-mm dishes were left untreated or exposed to heat shock at 448C for 30 min, and incubated at 378C for another 1, 2, 4, and 6 h, respectively. Cells were harvested, and the activity of JNK1 immunoprecipitated from 500 mg cell extracts was measured, using GST-c-Jun (1-79) as substrate. The kinase reaction products were analysed by electrophoresis on a 12.5% SDS-gel, and phosphorylated GST-c-Jun (1 – 79) was identiﬁed by autoradiography (left panel). Relative JNK1 activity in each sample was quantiﬁed with a densitometer (right panel). (b) Parental, 3T3neo, unselected/mixed clones of BLMP1-expressing (Bmix) and NLMP1-expressing (Nmix) Balb/3T3 cells were left untreated, or subjected to heat shock and incubated at 378C for 6 h. JNK1 activity was measured from 500 mg cell extracts as described above. The experiment was re- peated three times with reproducible results

Stress-induced death of BLMP1-expressing, NO- percentage of viable cells increased by 15 – 25% (Figure producing cells is attenuated by iNOS inhibitors 7c). The results clearly demonstrate that NO produced To investigate whether NO production in BLMP1- by BLMP1-expressing clones is deleterious to the cells, expressing cells is responsible for enhanced sensitivity and this effect may be attenuated by iNOS inhibitors to stress, the effects of two iNOS inhibitors, amino- that prevent NO synthesis in cells. guanidine (AG) and L-NMMA, were examined in D11 cells. Examination of NO levels in cells treated with Effect of iNOS inhibitor on anchorage-independent iNOS inhibitors revealed that by day 2, the production growth of LMP1-expressing 3T3 cells in methylcellulose of NO in D11 cells was effectively blocked (Figure 7a), and *70 – 80% cells retained viability, compared to It is known that LMP1 can transform Balb/3T3 cells 560% control cells. This difference was more (Wang et al., 1985; Baichwal and Sugden, 1988). pronounced on the third day, whereby *15 to 25% Moreover, cells expressing LMP1 acquire the ability cells retained viability in drug-treated cells, while viable to grow in an anchorage-independent manner, which is cells were barely detected in the control cell group a classical method for the assessment of the oncogenic (Figure 7b). Similarly, when D11 cells were grown in potential of a gene product. Since BLMP1 specifically media containing iNOS inhibitors for 2 days and induces iNOS expression and NO production, which in subsequently subjected to heat shock treatment, the turn damages the cells, we attempted to determine

Oncogene iNOS induction by LMP1 of EBV J-S Yu et al 8054

Figure 7 Effect of iNOS inhibitors on the viability of D11 cells under stress. (a,b) D11 cells (66105 cells) were cultured in 35-mm dishes containing 3 ml DMEM in the absence and presence of 1 mM AG or L-NMMA for 1 – 3 days without refreshing the medium. The amount of nitrate/nitrite in culture supernatants (a) and cell viability (b) was analysed every 24 h. (c) D11 cells 5 (6610 cells) were cultured in the absence and presence of 1 mM AG or L-NMMA for 2 days. Non-viable cells were removed by replacing the culture media, and viable cells were treated with heat shock (448C for 30 min) and incubated at 378C for another 4 h. Cell viability was subsequently determined as described above

whether iNOS expression/NO production aﬀects the BLMP1-mediated iNOS induction in Balb/3T3 cells anchorage-independent growth of cells containing attenuates the ability of the cell to grow in an BLMP1. BLMP1-expressing D11 and D13 cells were anchorage-independent manner. cultured in medium containing methylcellulose, in the presence and absence of 1 mM AG for 2 weeks, and the Effect of iNOS inhibitor on the growth of numbers of the resulting colonies were scored. D11 and LMP1-expressing 3T3 cells in nude mice D13 cells formed colonies in methylcellulose, both in the presence and absence of AG (result not illustrated). We further examined the possible role of iNOS However, more colonies (14 – 25%) of BLMP1-expres- induction in the tumorigenicity of LMP1-expressing sing D11 and D13 cells were observed in the presence 3T3 cells in vivo. Cell clones including 3T3neo, E9, D11 of AG, compared to cells cultured in the absence of the and D13 were inoculated subcutaneously into nude inhibitor (Table 2). This was more obvious in the D11 mice, which were administrated intraperitoneally with clone, which expresses higher levels of iNOS. In or without AG (200 mg/kg) every other day until day contrast, approximately 30% reduction in colony 45. All mice tolerated well this regime. No tumors were formation was observed when E9 cells were tested in formed in mice injected with 3T3neo cells, either in the the presence of AG (Table 2). As the controls, both presence or absence of AG. Without AG treatment, parent and 3T3neo cells could only grow very few subcutaneous tumors arose in all the mice inoculated spontaneous colonies without or with AG under the with E9 or D11 cells, with the E9 clone-derived tumors same condition (Table 2). These results signify that growing much faster than those from D11 cells (Figure

Oncogene iNOS induction by LMP1 of EBV J-S Yu et al 8055 Table 2 Efficiency of colony formation of LMP1-expressing Balb/ al., 1994). Furthermore, although both BLMP1 and 3T3 cells in the absence and presence of iNOS inhibitor in NPC-derived LMP1 activate transcription factors, NF- methylcellulose kB and AP-1, it appears that NLMP1 functions more Colony number Ratio efficiently in Jurkat cells (Fielding et al., 2001), but not Cell clone +AG 7AG (+AG/7AG) C33A cells (Yeh et al., 1997). Additional reports D11 (BLMP1) 660 474 1.39 demonstrate that the expression levels of several cell 558 507 1.10 surface molecules (CD40, CD44, and CD54), inter- 975 768 1.27 leukins-6 and -8, and epidermal growth factor receptor (Average) 731 583 1.25 D13 (BLMP1) 477 375 1.27 differ between BLMP1- and NLMP1-expressing cells 519 501 1.04 (Johnson et al., 1998; Miller et al., 1998; Dawson et al., 540 489 1.10 2000). Using Balb/3T3 cells as a model in this study, (Average) 512 455 1.14 we show that BLMP1 induces considerably higher E9 (NLMP1) 471 573 0.83 426 720 0.59 iNOS expression, which may be the basis of another 447 594 0.75 functional difference between the two LMP1 proteins. (Average) 448 629 0.71 The constitutive iNOS in BLMP1-expressing Balb/ 3T3 cells is catalytically active in NO synthesis, since a D11, D13 and E9 cells (4000 cells each) were seeded in 1.1% significant amount of the metabolic end products are methylcellulose prepared in culture media without and with 1 mM AG. Colonies with a diameter of 40.1 mm were scored under the detected in the culture supernatant (Table 1). In the microscope after 2 weeks. The results obtained from three BLMP1-expressing D11 clone, which exhibits the independent experiments are displayed. As experimental controls, highest iNOS levels, NO production was 2 – 3-fold 25+3 colonies were obtained with parental cells and 3T3neo control higher than that released from TNF-a-treated parental cells, in the presence or absence of AG cells (Figure 2a). NO was either toxic or protective, depending on its chemistry in a given biological milieu (Moncada et al., 1991). NO has cytotoxic effects at 8a). Only one of the four D13 clone-injected mice grew high concentrations, whereas at lower concentrations, palpable tumor by day 45. AG treatment had little the compound may have the opposite effect and protect effect on the kinetics of E9 clone-derived tumor against apoptotic cell death from diverse stimuli formation, but significantly enhanced the growth rate (Dimmeler and Zeiher, 1997; Muschel et al., 1998). of D11 and D13 clone-derived tumors (Figure 8a). To understand the possible biological functions of Interestingly, all four mice who received D13 cells constitutively expressed iNOS, we evaluated the effects developed tumors in the presence of AG. AG of NO on LMP1-expressing cells. Several lines of treatment systematically caused a reduction of the evidence indicate that iNOS induction and NO plasma nitrate/nitrite levels in all the three groups of production due to BLMP1 expression has deleterious mice (Figure 8b), indicating the efficacy of AG to effects on cells. Firstly, BLMP1-expressing, NO-produ- reduce NO production in mice. Immunoblot analysis cing cells lost viability more rapidly than parental, confirmed the expression of LMP1 and iNOS of the 3T3neo and NLMP1-expressing Balb/3T3 cells under cells cultured from E9, D11 or D13 clone-derived various environmental stress conditions (Figure 4). tumor mass, and the patterns were almost identical to Secondly, enhanced activation of caspase-3 was those observed from these cells prior to being injected observed in BLMP1-expressing, NO-producing cells into mice (Figure 8c). In another experiment, we during heat shock treatment (Figure 5). Thirdly, further examine the dose effect of AG on the growth blockage of NO production by iNOS inhibitors (AG of D11 clone-derived tumors in nude mice. Again, and L-NMMA) in these cells partially prevented cell tumors grew faster in the AG-treated mice than in the death during environmental stress (Figure 7). Overall, control group. At day 29, the average tumor size was these results suggest that Balb/3T3 cells harboring more than twofold larger in the presence of 200 mg/kg BLMP1 are considerably more sensitive to environ- AG and was almost fourfold in the presence of mental stress than those expressing NLMP1. This 400 mg/kg AG (Figure 8d). Thus, results collectively finding is consistent with previous reports that over- support that inhibition of iNOS by iNOS inhibitor expression of BLMP1 is toxic to many cell lines, enhances the tumor formation activity of BLMP1- including six human B lymphocyte lines, Balb/3T3 expressing cell clones. fibroblasts, 143/EBNA 1 cells (human osteosarcoma cell line expressing EBV EBNA-1 gene), HEp-2 cells (human carcinoma cell line) (Hammerschmidt et al., Discussion 1989) and human epithelial cell line RHEK-1 (Lu et al., 1996). Furthermore, the cytotoxic effect of BLMP1 Following the identification of an LMP1 variant, expression on Balb/3T3 cells is mediated, at least in NLMP1, from NPC biopsies (Chen et al., 1992; Hu part, through induction of iNOS and the subsequent et al., 1993), efforts have been made to distinguish the production of nitric oxide. functional role of this protein in relation to BLMP-1 Stable or transient expression of BLMP1 and from B95-8 cells. It is known that NLMP1 is more NLMP1 activates the JNK pathway in epithelial or B oncogenic than BLMP1 (Chen et al., 1992; Hu et al., cells (Kieser et al., 1997; Eliopoulos and Young, 1998; 1993; Li et al., 1996) and less immunogenic (Trivedi et Dawson et al., 2000). Stable expression of both

Oncogene iNOS induction by LMP1 of EBV J-S Yu et al 8056 BLMP1 and CAO-LMP1 (resembling NLMP1 in this study) in BL41 Burkitt lymphoma and SCC12F human epithelial cells resulted in low-level (2 – 3-fold), but constitutive activation of JNK (Eliopoulos and Young, 1998; Dawson et al., 2000). However, in this study, we observed little or no JNK activation in Balb/3T3 cells stably expressing BLMP1 or NLMP1, compared to parental or 3T3neo cells (Figure 6). Interestingly, the kinetics of JNK activation in Balb/3T3 cells stably expressing BLMP1 or NLMP1 was distinct in response to heat shock treatment. In contrast to the rapid and transient JNK activation in NLMP1-expressing E9, 3T3neo and parental cells, BLMP1-expressing, NO- producing D11 cells exhibited prolonged JNK activation after heat shock (Figure 6a). Similarly, at later stages, i.e., 6 h after heat shock, relatively higher levels of activated JNK were detected in unselected/mixed clones expressing BLMP1 than in parental, 3T3neo and NLMP1-expressing cells (Figure 6b). These findings imply that the upstream mechanism(s) regulating JNK activity in Balb/3T3 cells may be altered significantly by BLMP1, but not NLMP1. As activation of JNK was not impaired in either BLMP1- or NLMP1- expressing Balb/3T3 cells, the mechanism(s) responsible for the down-regulation of stress-induced JNK activity in Balb/3T3 cells may be affected by BLMP1, but not NLMP1. The hypothesis that the duration of JNK activation may be a determining factor in cell death or proliferation (Chen et al., 1996) is further supported by several recent studies suggesting that JNK activation is required for caspase activity during apoptosis triggered by variety stimuli (Seimiya et al., 1997; Chan et al., 2000; MacFarlane et al., 2000). In the present study, we found that during the later stages of heat- shock treatment, enhanced caspase-3 activation in BLMP1-expressing, NO-producing D11 cells correlated well with the sustained high levels of JNK activity (see Figures 5a and 6a for comparisons). Therefore, it is feasible that the enhanced sensitivity of these cells to environmental stress is attributed to the alteration of JNK-mediated caspase activation. At present, the mechanism by which specific expression of BLMP1 in Balb/3T3 cells affects the regulation of JNK and/or caspase activities is unknown. One possibility is that BLMP1 induces large quantities of NO, which is deleterious to cells. To examine this theory, we investigated the effects of iNOS inhibitors (AG and L-NMMA) on heat shock- Figure 8 Effects of iNOS inhibitor on the tumorigenicity of induced caspase-3 activation in BLMP1-expressing, LMP1-expressing Balb/3T3 cells in nude mice. (a – c) Tumor NO-producing D11 cells. Although more than 70% growth in response to AG treatment. Nude mice were inoculated NO production was inhibited by pretreatment with with BLMP1-expressing cells (D11 and D13), NLMP1-expressing 1mM AG or L-NMMA, levels of heat shock-induced cells (E9) or 3T3neo control cells respectively, followed by treatment of AG (200 mg/kg) every other day. (a) The kinetics of tumor development. Data represent the mean tumor volume+s.e.m. (b) The plasma nitrate/nitrite concentrations of mice. Equal amounts of plasma (50 ml) from the mice in each group were tumor mass were resolved in SDS – PAGE, followed by immuno- mixed together and subjected for nitrate/nitrite level determina- blot analysis with anti-LMP1 or anti-iNOS antibody. (d) Dose ef- tion. The values are the mean level of nitrate/nitrite in each fect of AG on the growth of D11 clone-derived tumors in nude group. Ctrl, mice without any treatment. (c) Expression of mice. Nude mice were inoculated with D11 cells, followed by LMP1 and iNOS in cells cultured from tumor mass. Extracts treatment with or without AG (200 or 400 mg/kg) every other (100 mg) of cells cultured from the E9, D11 or D13 clone-derived day. Data represent the mean tumor volume+s.e.m. at day 29

Oncogene iNOS induction by LMP1 of EBV J-S Yu et al 8057 caspase-3 in D11 cells were not affected (data not for iNOS. It can also inhibit diamine oxidase, the rate- illustrated). A simple explanation is that NO may not limiting enzyme in terminal catabolism of polyamines, be the sole cause for BLMP-expressing cell suscept- and the formation of highly reactive advanced ibility to stress-induced apoptosis, although the glycosylation end products that are linked to tissue possible significance of NO in this process cannot be damage in diabetes and aging (for review, see Nilsson, excluded. It is known that long-term exposure to high 1999). Therefore, the side effects of AG resulted from doses of NO leads to S-nitrosylation involving the the inhibition of the two pathways can not be ruled out transfer of a nitric oxide group to cysteine sulfhydryls in the present study. However, the efficacy of AG to onto cellular targets, including metabolic, structural inhibit NO production in vitro and in vivo demon- and signaling proteins (Stamler, 1994; Jaffrey et al., strated in this study and in mice infected with 2001). This may in turn lead to irreversible cell lymphocytic choriomeningitis virus (LCMV) (Camp- damage, irrespective of the further addition of iNOS bell, 1996), as well as the profound effects of iNOS inhibitors. In this context, NO-producing cells may be expression in the growth or apoptosis of tumors particularly susceptible to damage, despite temporary (Jenkins et al., 1995; Thomsen et al., 1997; Ambs et treatment with iNOS inhibitors. al., 1998b; Xie and Fidler, 1998) strongly implicate the Both BLMP1 and NLMP1 induce transformation of role of iNOS expression/NO production in attenuating Balb/3T3 cells when the LMP1 gene is driven by an the LMP1-mediated transformation. appropriate promoter. LMP1-transformed cells grow In conclusion, we show that BLMP1, but not as colonies in an anchorage-independent manner NLMP1, induces iNOS expression and NO production (Baichwal and Sugden, 1988; Chen et al., 1992; Li et in Balb/3T3 cells. This biological activity may be al., 1996). In concurrence with these reports, we significant in distinguishing between the functional observed that all three clones employed in the present roles of the BLMP1 and NLMP1 proteins. study (BLMP1-expressing D11 and D13, and NLMP1- expressing E9 clones) formed colonies in methylcellulose, and the number of colonies significantly increased Materials and methods (14 – 25%) in NO-producing BLMP1 clones, D11 and D13 (Table 2). Moreover, in D11 cells, the percentage Materials increase in colonies following AG treatment corre- [g-32P]ATP was purchased from Amersham (Buckingham- sponded to enhanced cell viability (*20%) under stress shire, UK). Polyvinylidene fluoride (PVDF) membrane was (see Table 2 and Figure 7 for comparison). This obtained from Millipore (Bedford, MA, USA). Dulbecco’s specific effect of iNOS inhibition by AG on the modified Eagle’s medium (DMEM), dexamethasone, methyl- anchorage-independent growth of BLMP1-expressing cellulose, goat anti-rabbit and anti-mouse IgG antibodies cells was not observed in NLMP1-expressing E9 cells, conjugated with alkaline phosphatase were from Sigma (St. which exhibit undetectable levels of iNOS (Table 2). In Louis, MO, USA). DMEM without phenol red was fact, AG treatment resulted in fewer E9 colonies. The purchased from GIBCOBRL. Anti-iNOS (N-20) and anti- reason for the differential response of BLMP1- and JNK1 (C17) antibodies were procured from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Anti-LMP1 mono- NLMP1-expressing cells to iNOS inhibitors is currently clonal antibody, S12, was generated by Dr Thorley-Lawsen. unknown. However, these results imply that the Nitrate/nitrite assay kits were from Cayman Chemical Co. G cellular levels of NO play an important role in the (Ann Arbor, MI, USA). N -monomethyl-L-arginine (L- anchorage-independent growth of BLMP1-expressing NMMA), S-methylisothiourea (SMT), aminoguanidine (AG) Balb/3T3 cells. Therefore, in addition to inducing and Z-DEVD-AFC were from Calbiochem (La Jolla, CA, considerable susceptibility to stress-triggered cell death, USA). CDP-StarTM (a chemiluminescent substrate for alka- iNOS expression/NO production may affect cell growth line phosphatase) was obtained from Boehringer Mannheim in methylcellulose. This finding is not unexpected, since (Mannheim, Germany). Molecular weight marker proteins growth of an adherent cell in a semi-solid background (Mark12) were from Novel Experimental Technology (San or in suspension acts as a stress inducing factor leading Diego, CA, USA). BCA protein assay reagent and Protein A- Sepharose CL-4B were supplied by Pierce (Rockford, IL) and to anoikis, a special type of apoptosis (Frisch and Amersham Pharmacia Biotech AB (Uppsala, Sweden), Ruoslahti, 1997; Valentinis et al., 1999). respectively. GST-c-Jun (1-79) was expressed in E. coli and The possible role of iNOS induction in the LMP1- purified with a glutathione affinity column for JNK activity mediated transformation was further evaluated by assays (Chan et al., 2000). nude mice model in vivo. Albeit the effects of iNOS inhibition with respect to the in vitro functional tests Cell culture and drug treatment were small (Table 2 and Figure 7), administration of AG caused a quite significant enhancement of the Mouse Balb/3T3 fibroblasts were cultured at 378C in a 95% tumor growth in mice inoculated with the NO- air/5% CO2 and water-saturated atmosphere in DMEM producing, BLMP1-expressing D11 and D13 cells supplemented with 10% heat-inactivated fetal bovine serum, 100 U/ml penicillin and 100 mg/ml streptomycin. Stable Balb/ (Figure 8). In contrast, this treatment had no obvious 3T3 cell lines expressing BLMP1 and NLMP1 were generated effect on the growth of tumor derived from NLMP1- by co-transfection of BLMP1-expressing vector (pEcoD4) expressing E9 cells. The results clearly show that these and NLMP1-expressing vector (pT7E) with pSV2neo cells respond to AG in a similar trend both in vitro and (containing neomycin-resistance gene) (Chen et al., 1992; Li in vivo. It is known that AG is not a specific inhibitor et al., 1996), followed by Geneticin sulfate (G418) selection

Oncogene iNOS induction by LMP1 of EBV J-S Yu et al 8058 for 2 weeks. Individual cell clones stably expressing BLMP1 Northern blot analysis, 10 mg poly(A)+mRNA isolated from and NLMP1 were further selected from mixed clones. various cell clones was resolved by electrophoresis in a Similarly, 3T3neo cells were established by transfection of formaldehyde/1% agarose gel, and transferred to a nitro- pSV2neo into Balb/3T3 cells, followed by selection with cellulose membrane. After UV cross-linking, the membrane G418. was hybridized to 32P-labeled probes at 688C, and washed in For drug treatment, cells (*5–66105) were plated onto high-stringency conditions (0.26standard saline citrate/0.1% 35-mm culture dishes containing 2 ml DMEM. Four hours SDS at 558C). Autoradiography was performed by exposure after plating, the medium was replaced by 2 ml DMEM of the washed membrane to BioMax film (Kodak). without phenol red, and cells were incubated with or without L-NMMA (50 mM), S-methylisothiourea (25 mM) or dexa- RT – PCR analysis for iNOS expression methasone (25 mM) for 1, 2 and 3 days, respectively. High doses of iNOS inhibitor, aminoguanidine or L-NMMA Total RNA was prepared from freshly harvested, trans- (1 mM), were used when required. Small aliquots (0.2 ml) of fected Balb/3T3 cells, using TRIzoltm, according to the medium were removed at various time-points and stored manufacturers’ instructions (Gibco). Oligonucleotide primers at 7208C for the determination of nitrate/nitrite content. specific for iNOS, IN-1 and IN-2 were prepared. Primer IN- Cells were washed twice with ice-cold PBS and lysed in 250 ml 2 was selected for the reverse transcription-polymerase chain solution A (20 mM Tris-HCl, pH 7.4, 1 mM EDTA, 1 mM reaction (RT – PCR). PCR amplification of iNOS performed EGTA, 1% Triton X-100, 1 mM benzamidine, 1 mM as described previously (Asano et al., 1994) yielded a 380- phenylmethylsulfonyl fluoride, 50 mM NaF, 20 mM sodium bp fragment. Denaturation, annealing, and elongation pyrophosphate, 1 mM sodium orthovanadate) on ice for temperatures for PCR were 94, 55, and 728C, respectively, 10 min. Cell lysates were collected and sonicated on ice at 1 min each for 32 cycles, followed by a final 10 min (model W-380, Heat Systems-Ultrasonics) for 3610 s at 50% extension at 728C. PCR products were separated by agarose power output, followed by centrifugation at 15 000 g for gel electrophoresis, transferred to nylon membranes 20 min at 48C. The resulting supernatants were used as cell (Hybond-N, Amersham, UK), and the iNOS gene was extracts. detected by Southern blot hybridization to a specific radiolabeled probe. Environmental stress Nitrate/nitrite assays Cells (*5–66106) were plated on 100-mm culture dishes 24 h before the experiment, and subsequently subjected to The amount of NO produced by cells was indirectly - heat shock or UV irradiation, as previously described (Chan measured by the formation of nitrate (NO3 ) and nitrite - et al., 1998; Tang et al., 1998). Briefly, heat shock treatment (NO2 ) (two stable end products of NO) in culture super- was performed by placing culture dishes in a 448C water bath natants. The total amounts of nitrate and nitrite were for 30 min, and incubating at 378CinaCO2 incubator for determined with the nitrate/nitrite assay kit (Cayman indicated intervals. Ultraviolet irradiation (254 nm, UVC) Chemical Co.), using manufacturer’s protocols. Briefly, was achieved with a UV gene linker (UV Stratalinker, model supernatant (80 ml) was incubated with nitrate reductase 1800, Stratagene, La Jolla, CA, USA) equipped with energy (20 ml) at room temperature for 1 h, followed by incubation output control. Prior to irradiation with UV light at 200 J/ with 100 ml Griess reagent (1% sulfanilic acid, 0.1% m2, medium was removed from the culture dish. The medium naphtylethylenediamine dihydrochloride, 2.5% phosphoric was immediately added back into the dish after UV acid) for 10 min. The optical density of each sample was irradiation, and incubated at 378CinaCO2 incubator for analysed at 540 nm in a microplate reader with sodium nitrite the various time intervals indicated. as a standard. For measurement of the plasma nitrate/nitrite levels, the plasma derived from blood samples were first clarified by centrifugal ultrafiltration (Microcon YM-30 Filter Cell viability assay Devices, Millipore, Bedford, MA, USA). The filtered plasma Cell viability was determined using the trypan blue exclusion was then processed as described above. test. Briefly, cultured cells after treatment were trypsinized from the culture dish, and an equal volume of trypan blue Immunoblots solution was mixed with cell suspensions for staining. Stained non-viable and unstained viable cells were counted over Immunoblotting was carried out essentially as described in several days. Cell viability was represented as a percentage of previous reports (Yu and Yang, 1994). Briefly, cell extracts the ratio of viable to total (viable and non-viable) cell were resolved on SDS gels and electroblotted onto PVDF number. membranes. Membranes were probed with anti-iNOS antibody (0.2 mg/ml) or anti-LMP1 monoclonal antibody, S12 (0.25 mg/ml) (Mann et al., 1985). Proteins of interest were Northern blot analysis detected with goat anti-rabbit or anti-mouse IgG antibody Total RNA was isolated from cultured cells with the RNeasy conjugated with alkaline phosphatase, and CDP-StarTM (a Midi Kit (Qiagen), and Poly (A)+mRNA was purified by chemiluminescent substrate for alkaline phosphatase), using Oligotex Direct mRNA Kit (Qiagen), according to manu- manufacturers’ protocols. For re-probing the same facturer’s instructions. Probes were prepared by reverse membrane with another type of antibody, the blotted transcription-polymerase chain reaction (RT – PCR) amplifi- membrane was initially incubated with 200 ml stripping cation, and labeled, using a random priming kit (Amersham). buffer (62.5 mM Tris-HCl, pH 6.7, 100 mM 2-mercaptoetha- For iNOS, the oligonucleotide primers, IN-1 (5’-GTGAGG- nol, 2% SDS) at 508C for 30 min, with occasional agitation ATCAAAAACTGGGG-3’) and IN-2 (5’-ACCTGCAGGTT- to strip off bound antibodies. After washing three times in GGACCAC-3’), were synthesized as described by Asano et TTBS buffer (20 mM Tris-HCl, pH 7.4, 0.5 M NaCl, 0.05% al. (1994). b-actin mRNA detected by a probe corresponding Tween-20), the stripped membrane was probed with another to part of b-actin cDNA was used as an internal control. For antibody.

Oncogene iNOS induction by LMP1 of EBV J-S Yu et al 8059 agar bottom layer (Amresco). Methylcellulose (2 ml of 1.1%) Caspase-3 activity assay prepared in culture medium was added to each dish after 1 Caspase-3 activity was measured using Z-DEVD-AFC week to feed cells. Colonies with a diameter of 40.1 mm (Calbiochem) as the fluorogenic substrate. Cell lysate (approximately 300 cells) were scored under the microscope (100 mg) and Z-DEVD-AFC (0.1 mM) were incubated in after 2 weeks. 250 ml caspase assay buffer (25 mM HEPES, pH 7.5, 0.1% CHAPS, 10 mM dithiothreitol, 100 U/ml aprotinin) for 3 h at Animal tumor model 378C. Ice-cold caspase assay buffer (1.25 ml) was added to the mixture, and the relative caspase-3 activity was BALB/cAnN-nude mice, 5 – 6 weeks old, were obtained from determined, using a fluorescence spectrophotometer (Hitachi, the National Laboratory Animal Breeding and Research F-2000; excitation 400 nm, emission 505 nm). Center (Nangang, Taipei, Taiwan), kept under specific pathogen-free conditions in Animal Center of Chang Gung University, and used for experiments at age of 7 – 8 weeks. Immunoprecipitation and JNK activity assay One day before inoculation of cells, mice were injected Immunoprecipitated JNK activity was measured as intraperitoneally with 0.2 ml of PBS or PBS containing AG previously described (Chan et al., 2000). Before immunopre- (200 mg/kg). Cells (46106) suspended in 0.2 ml Earle’s cipitation, protein concentrations of cell extracts were diluted balanced salts solution (EBSS) were inoculated subcuta- to equal amounts with solution A. For immunoprecipitation, neously into nude mice (four mice per group), and the mice 0.5 ml cell extracts (1.0 mg/ml) were incubated with 10 ml were injected with 0.2 ml of PBS or PBS containing AG JNK1(C17) antibody (200 mg/ml) at 48C for 1.5 h, followed (200 mg/kg) every other day. For dose effect experiment, one by 40 ml protein A-Sepharose CL-4B (30%, v/v) for another group of mice were injected with 400 mg/kg AG. Tumor 1.5 h with shaking. Immunoprecipitates were collected by development was evaluated by measuring three perpendicular centrifugation, washed three times with 1 ml buffer C (20 mM diameters by calipers on alternative days from day 15 after Tris-HCl, pH 7.0, 0.5 mM dithiothreitol) containing 0.5 M injection. A tumor with a minimal size of 5-mm diameter was NaCl, and resuspended in 40 ml buffer C. For the JNK scored as positive. The mice were sacrificed at day 45, and activity assay, immunoprecipitates were incubated in a 50 ml blood samples (100 ml/mouse) from each group were collected mixture containing 20 mM Tris-HCl (pH 7.0), 0.5 mM and pooled together for measurement of the total nitrate/ 32 dithiothreitol, 0.2 mM [g- P] ATP, 20 mM MgCl2, and nitrite concentration. Tumors from each group were excised, 0.1 mg/ml GST-c-Jun (1-79) at room temperature for cut into small pieces, and cultured in DMEM containing 10 min with shaking. The reaction was stopped by the 10% heat-inactivated fetal bovine serum, 100 U/ml penicillin, addition of 26Laemmli sample buffer, and samples were 100 mg/ml streptomycin and 0.8 mg/ml G418 for 2 weeks. analysed by electrophoresis on 12.5% SDS-polyacrylamide gels. Phosphorylated GST-c-Jun (1-79) (*37 kDa) was identified by autoradiography. Densitometrical analyses were accomplished with a Computing Densitometer (Molecular Acknowledgements Dynamics). This work was supported by grants (NSC89-2318-B-182- 005-M51, NSC90-2316-B-182-003) from the National Science Council of Taiwan, a Striving for Excellency grant Anchorage-independent growth in methylcellulose (89-B-FA04-1-4) from the Ministry of Education, Taiwan Anchorage-independent growth was assayed in 1.1% methyl- and grants (CMRP-724) from Chang Gung University, cellulose with a 0.9% agar bottom layer. Briefly, 4000 cells Tao-Yuan, Taiwan, ROC. PJ Chang is supported by a were seeded in 3 ml culture medium containing 1.1% fellowship from National Health Research Institute grant methylcellulose in 60-mm dishes pre-coated with a 0.9% (NHRI-EX90-8703SL).

References

Aaltoma SH, Lipponen PK and Kosma VM. (2001). Chan WH, Yu JS and Yang SD. (1998). J. Protein Chem., 17, Anticancer Res., 21, 3101 – 3106. 485 – 494. Ambs S, Merriam WG, Bennett WP, Felley-Bosco E, Chan WH, Yu JS and Yang SD. (2000). Biochem. J., 351, Ogunfusika MO, Oser SM, Klein S, Shields PG, Billiar 221 – 232. TR and Harris CC. (1998a). Cancer Res., 58, 334 – 341. ChenML,TsaiCN,LiangCL,ShuCH,HuangCR, Ambs S, Merriam WG, Ogunfusika MO, Bennett WP, Ishibe Sulizeanu D, Liu ST and Chang YS. (1992). Oncogene, 7, N, Hussain SP, Tzeng E, Geller DA, Billiar TR and Harris 2131 – 2140. CC. (1998b). Nat. Med., 4, 1371 – 1376. Chen YR, Wang X, Templeton D, Davis RJ and Tan TH. Arvanitakis L, Yaseen N and Sharma S. (1995). J. Immunol., (1996). J. Biol. Chem., 271, 31929 – 31936. 155, 1047 – 1056. Cobbs CS, Brenman JE, Aldape KD, Bredt DS and Isreal Asano K, Chee CB, Gaston B, Lilly CM, Gerard C, Drazen MA. (1995). Cancer Res., 55, 727 – 730. JM and Stamler JS. (1994). Proc. Natl. Acad. Sci. USA, 91, Dawson CW, Eliopoulos AG, Blake SM, Barker R and 10089 – 10093. Young LS. (2000). Virology, 272, 204 – 217. Baichwal VR and Sugden B. (1988). Oncogene, 2, 461 – 467. Dawson CW, Rickinson AB and Young LS. (1990). Nature, Birkenbach M, Liebowitz D, Wang F, Sample J and Kieff E. 344, 777 – 780. (1989). J. Virol., 63, 4079 – 4084. Derijard B, Hibi M, Wu IH, Barrett T, Su B, Deng T, Karin Bredt DS, Hwang PM, Glatt CE, Lowenstein C, Reed RR M and Davis RJ. (1994). Cell, 76, 1025 – 1037. and Snyder SH. (1990). Nature, 347, 768 – 770. Campbell IL. (1996). J. Neuroimmunol., 71, 31 – 36.

Oncogene iNOS induction by LMP1 of EBV J-S Yu et al 8060 Devergne O, Hatzivassiliou E, Izumi KM, Kaye KM, Kulwichit W, Edwards RH, Davenport EM, Baskar JF, Kleijnen MF, Kieff E and Mosialos G. (1996). Mol. Cell. Godfery V and Rabb-Traub N. (1998). Proc. Natl. Acad. Biol., 16, 7098 – 7108. Sci. USA, 95, 11963 – 11968. Dimmeler S and Zeiher AM. (1997). Nitric Oxide, 1, 275 – Kyriakis JM, Banerjee P, Nikolakaki E, Dai T, Rubie EA, 281. Ahmad MF, Avruch J and Woodgett JR. (1994). Nature, Eliopoulos AG and Rickinson AB. (1998). Curr. Biol., 8, 369, 156 – 160. R196 – R198. Laherty CD, Hu HM, Opipari AW, Wang F and Dixit VM. Eliopoulos AG and Young LS. (1998). Oncogene, 16, 1731 – (1992). J. Biol. Chem., 267, 24157 – 24160. 1742. Liebowitz D, Wang D and Kieff E. (1986). J. Virol., 58, 233 – Esptein MA, Achong BG and Barr YM. (1964). Lancet, 1, 237. 702 – 703. Li SN, Chang YS and Liu ST. (1996). Oncogene, 12, 2129 – Fahraeus R, Fu HL, Ernberg I, Finke J, Rowe M, Klein G, 2135. Falk K, Nillson E, Yadav M, Busson P, Tursz T and Kallin LuJJY,ChenJY,HsuTY,YuWCY,SuIJandYangCS. B. (1988). Int. J. Cancer, 42, 329 – 338. (1996). J. Gen. Virol., 77, 1883 – 1892. Fahraeus R, Rymo L, Rhim JS and Klein G. (1990). Nature, MacFarlane M, Cohen GM and Dickens M. (2000). 345, 447 – 449. Biochem. J., 348, 93 – 101. Fairbairn LJ, Stewart JP, Hampson IN, Arrand JR and Mann KP, Staunton D and Thorley-Lawson DA. (1985). J. Dexter TM. (1993). J. Gen. Virol., 74, 247 – 254. Virol., 55, 710 – 720. Fielding CA, Sandvej K, Mehl A, Brennan P, Jones M and MannickJB,AsanoK,IzumiK,KieffEandStamlerJS. Rowe M. (2001). J. Virol., 75, 9129 – 9141. (1994). Cell, 79, 1137 – 1146. Frisch SM and Ruoslahti E. (1997). Curr. Opin. Cell Biol., 9, MarsdenAP,SchappertTKandMichelT.(1992).FEBS 701 – 706. Lett., 307, 287 – 293. Gao X, Tajima M and Sairenji T. (1999). Virology, 258, 375 – Martins LM and Earnshaw WC. (1997). Trends Cell Biol., 7, 381. 111 – 114. Gires O, Zimber-Strobl U, Gonnella R, Ueffing M, Massi D, Franchi A, Sardi I, Magnelli L, Paglierani M, Marschall G, Zeidler R, Pich D and Hammerschmidt W. Borgognoni L, Maria Reali U and Santucci M. (2001). J. (1997). EMBO J., 16, 6131 – 6140. Pathol., 194, 194 – 200. Hammerschmidt W, Sugden B and Baichwal VR. (1989). J. Miller WE, Cheshire JL, Baldwin Jr AS and Raab-Traub N. Virol., 63, 2469 – 2475. (1998). Oncogene, 16, 1869 – 1877. Hanto DW, Frizzera G, Gajl-Peczalska KJ, Balfour HH, Moncada S, Palmer RM and Higgs EA. (1991). Pharmacol. Simmons RL and Najarian JS. (1985). Transplant Proc., Rev., 43, 109 – 142. 17, 89 – 92. Mosialos G, Birkenbach M, Yalamanchili R, VanArsdale T, Henderson S, Rowe M, Gregory C, Croom-Carter D, Wang Ware C and Kieff E. (1995). Cell, 80, 389 – 399. F, Longnecker R, Kieff E and Rickinson A. (1991). Cell, Mosialos G, Hanissian SH, Jawahar S, Vara L, Kieff E and 65, 1107 – 1115. Chatila TA. (1994). J. Virol., 68, 1697 – 1705. Henle G, Henle W and Diehl V. (1968). Proc. Natl. Acad. Sci. Murono S, Inoue H, Tanabe T, Joab I, Yoshizaki T, USA, 59, 94 – 101. Furukawa M and Pagano JS. (2001). Proc. Natl. Acad. Hu LF, Zabarovsky ER, Chen F, Cao SL, Ernberg I, Klein G Sci. USA, 98, 6905 – 6910. and Winberg G. (1991). J. Gen. Virol., 72, 2399 – 2409. Muschel RJ, Soto DE, McKenna WG and Bernhard EJ. HuLF,ChenF,ZhengX,ErnbergI,CaoSL,Christensson (1998). Oncogene, 17, 3359 – 3363. B, Klein G and Winberg G. (1993). Oncogene, 8, 1575 – Nakagomi H, Dolcetti R, Bejarano MT, Pisa P, Kiessling R 1583. and Masucci MG. (1994). Int. J. Cancer, 57, 240 – 244. IwasakiY,ChongJ,HayshiY,IkenoR,AraiK,Kitamura Nathan C. (1992). FASEB J., 6, 3051 – 3064. M, Koike M, Hirai K and Fukayama M. (1998). J. Virol., Nathan C and Xie QW. (1994). Cell, 78, 915 – 918. 72, 8321 – 8326. Nicholson DW and Thornberry NA. (1997). Trends Biochem. Izumi KM and Kieff ED. (1997). Proc. Natl. Acad. Sci. USA, Sci., 22, 299 – 306. 94, 12592 – 12597. Nilsson BO. (1999). Inflamm. Res., 48, 509 – 515. Jaffrey SR, Erdjument-Bromage H, Ferris CD, Tempst P Pathmanathan R, Prasad U, Sadler R, Flynn K and Raab- and Snyder SH. (2001). Nat. Cell Biol., 3, E46 – E49. Traub N. (1995). N.Engl.J.Med.,333, 693 – 698. Jenkins DC, Charles IG, Thomsen LL, Moss DW, Holmes Peng M and Lundgren E. (1992). Oncogene, 7, 1775 – 1782. LS,BaylisSA,RhodesP,WestmoreK,EmsonPCand Riches DWH, Chan ED, Zahradka EA, Winston BW, Moncada S. (1995). Proc. Natl. Acad. Sci. USA, 92, 4392 – Remigio LK and Lake FR. (1998). J. Biol. Chem., 273, 4396. 22800 – 22806. Johnson RJ, Stack M, Hazlewood SA, Jones M, Blackmore Seimiya H, Mashima T, Toho M and Tsuruo T. (1997). J. CG, Hu LF and Rowe M. (1998). J. Virol., 72, 4038 – 4048. Biol. Chem., 272, 4631 – 4636. Kawanishi M. (1995). Intervirology, 38, 206 – 213. Stamler JS. (1994). Cell, 78, 931 – 936. Kieff E and Liebowitz D. (1990). Virology Fields BN and Su IJ and Hsieh HC. (1992). Leukemia Lymphoma, 7, 47 – 53. Knipe DM (ed.). New York: Raven Press, pp 1889 – 1920. Sylla BS, Hung SC, Davidson DM, Hatzivassiliou E, Kieser A, Kaiser C and Hammerschmidt W. (1999). EMBO Malinin NL, Wallach D, Gilmore TD, Kieff E and J., 18, 2511 – 2521. Mosialos G. (1998). Proc. Natl. Acad. Sci. USA, 95, Kieser A, Kilger E, Gires O, Ueffing M, Kolch W and 10106 – 10111. Hammerschmidt W. (1997). EMBO J., 16, 6478 – 6485. Thomsen LL, Lawton FG, Knowles RG, Beesley JE, Klein G, Giovanella BC, Lindahl T, Fialkow PJ, Singh S and Riveros-Moreno V and Moncada S. (1994). Cancer Res., Stehlin JS. (1974). Proc.Natl.Acad.Sci.USA,71, 4737 – 54, 1352 – 1354. 4741. Thomsen LL, Miles DW, Happerfield L, Bobrow LG, Kleinert H, Euchenhofer C, Ihrig-Biedert I and Forstermann Knowles RG and Moncada S. (1995). Br.J.Cancer,72, U. (1996). J. Biol. Chem., 271, 6039 – 6044. 41 – 44.

Oncogene iNOS induction by LMP1 of EBV J-S Yu et al 8061 Thomsen LL, Scott JM, Topley P, Knowles RG, Keerie AJ Wang D, Liebowitz D and Kieff E. (1985). Cell, 43, 831 – and Frend AJ. (1997). Cancer Res., 57, 3300 – 3304. 840. Tang TK, Chang WC, Chan WH, Yang SD, Ni MH and Yu Wang F, Gregory C, Sample C, Rowe M, Liebowitz D, JS. (1998). J. Cell Biochem., 70, 442 – 454. Murray R, Rickinson A and Kieff E. (1990). J. Virol., 64, Trivedi P, Hu LF, Chen F, Christensson B, Masucci MG, 2309 – 2318. Klein G and Winberg G. (1994). Eur. J. Cancer, 30A, 84 – Wilson KT, Fu S, Ramanujam KS and Meltzer SJ. (1998). 88. Cancer Res., 58, 2929 – 2934. Valentinis B, Morrione A, Peruzzi F, Prisco M, Reiss K and Xie K and Fidler IJ. (1998). Cancer Metastasis Rev., 17, 55 – Baserga R. (1999). Oncogene, 18, 1827 – 1836. 75. VerheijM,BoseR,LinXH,YaoB,JarvisWD,GrantS, YehTS,LiSN,WuCJ,LiuST,MengCLandChangYS. Birrer MJ, Szabo E, Zon LI, Kyriakis JM, Haimovitz- (1997). DNA Cell Biol., 16, 1311 – 1319. Friedman A, Fuks Z and Kolesnick RN. (1996). Nature, Yu J-S and Yang S-D. (1994). J. Cell. Biochem., 56, 131 – 380, 75 – 79. 141.

Oncogene