|Home | About | Journals | Submit | Contact Us | Français|
Epstein-Barr virus (EBV) is a ubiquitous herpesvirus that infects more than 90% of the world's adult population and is linked to multiple malignancies, including Burkitt lymphoma, Hodgkin disease, and nasopharyngeal carcinoma (NPC). The EBV oncoprotein LMP1 induces transcription of the epidermal growth factor receptor (EGFR), which is expressed at high levels in NPC. EGFR transcription is induced by LMP1 through a p50 NFκB1-Bcl-3 complex, and Bcl-3 is induced by LMP1-mediated activation of STAT3. This study reveals that LMP1, through its carboxyl-terminal activation domain 1 (LMP1-CTAR1), activates both STAT3 and EGFR in a serum-independent manner with constitutive serine phosphorylation of STAT3. Upon treatment with EGF, the LMP1-CTAR1-induced EGFR was additionally phosphorylated and STAT3 became phosphorylated on tyrosine, concomitant with upregulation of a subset of STAT3 target genes. The kinase responsible for LMP1-CTAR1-mediated serine phosphorylation of STAT3 was identified to be PKCδ using specific RNAi, a dominant negative PKCδ, and the PKCδ inhibitor rottlerin. Interestingly, inhibition of PKCδ also inhibited constitutive phosphorylation of EGFR and LMP1-CTAR1-induced phosphorylation of ERK. Inhibition of PKCδ blocked LMP1-CTAR1-mediated transformation of Rat-1 cells, likely through the inhibition of ERK activation. These findings indicate that LMP1 activates multiple distinct signaling pathways and suggest that PKCδ functions as a master regulator of EGFR, STAT3, and ERK activation by LMP1-CTAR1.
Infecting more than 90% of the world's population, the Epstein-Barr virus (EBV) is a ubiquitous human gammaherpesvirus. EBV is transmitted through saliva and infects oropharyngeal epithelial cells and B-lymphocytes, resulting in a life-long infection (38). Persistent, latent EBV infection is present in several lymphoid and epithelial malignancies, including Burkitt lymphoma, Hodgkin disease, and nasopharyngeal carcinoma (NPC). The tumors are latently infected with expression of a subset of viral genes in the malignant cells which contain EBV episomes (38).
EBV latent membrane protein 1 (LMP1) is considered the major oncogene as it is essential for B-lymphocyte transformation and can also transform rodent fibroblast cells (19, 47). LMP1 functions as a constitutively active tumor necrosis factor receptor (TNFR) due to aggregation through its six transmembrane domains and interactions with tumor necrosis factor-associated factors (TRAFs) (31, 34). Multiple signaling pathways are activated by LMP1, including mitogen-activated protein kinase (MAPK), c-Jun N-terminal kinase (JNK), phosphatidylinositol 3-kinase (PI3K)/Akt, and NF-κB (7, 28, 36, 39). LMP1 induces transcription of many genes to affect apoptosis, cell cycle progression, cell proliferation, and migration (9, 10, 41). Two major signaling domains in the C-terminal tail of LMP1, carboxyl-terminal activation regions 1 and 2 (CTAR1 and CTAR2), mediate LMP1-associated signaling pathways by recruiting different TRAFs (44). LMP1-CTAR1 recruits TRAF1/2/3/5 through a TRAF-binding motif residing from amino acids (aa) 204 to 208 (PQQAT), whereas CTAR2 recruits TRAF2 and TRAF6 through adapters TRADD and BS69 (44).
In addition to distinctive TRAF binding, CTAR1 and CTAR2 also have different effects on signaling activation and cellular transformation. LMP1-CTAR1 specifically induces MAPK and PI3K/Akt signaling, while CTAR2 activates the JNK pathway (6, 28, 29). LMP1-CTAR2 activates strong canonical NF-κB signaling, whereas LMP1-CTAR1 induces more complex NF-κB signaling, including canonical, noncanonical, and atypical pathways (22, 24, 31). LMP1-CTAR1 also uniquely induces expression of TRAF1, ID1, and the epidermal growth factor receptor (EGFR) (9, 31). Importantly, LMP1-CTAR1 has been shown to be required for LMP1-mediated transformation of B-lymphocytes and rodent fibroblasts, while CTAR2 is dispensable (16, 19, 28). These findings suggest that LMP1-CTAR1 may have more significant effects in the development of malignancy.
EGFR is a member of the ErbB receptor tyrosine kinase family. Multiple signaling pathways are activated by the EGFR, including Ras/MAPK, Src kinases, JAKs/STATs, and PI3K-Akt (17). EGFR signaling is targeted by proteins of several oncogenic viruses to mediate transformation, including v-ErbB, E5, HBVx, and LMP1 (32). Elevated amounts of EGFR are detected in NPC, and its expression and secretion from the cell in exosomes correlates with the levels of LMP1 (30, 48). Treatment with EGFR tyrosine kinase inhibitors induces cell cycle arrest and inhibits cell proliferation of NPC cell lines (49). These findings suggest that the effect of LMP1 on EGFR is an important factor in EBV-mediated carcinogenesis and is a prime target for development of therapeutic strategies.
Signal transducers and activators of transcription (STATs) consist of a family of cytoplasmic proteins that, upon stimulation by cytokines or growth factors, translocate into the nucleus and transactivate cellular target genes that are involved with cell proliferation, cell cycle regulation, apoptosis, angiogenesis, and differentiation (20). The transcriptional activity of STAT3 is regulated by phosphorylation. Phosphorylation of tyrosine 705 and serine 727 affect translocation and the transcriptional activity of STAT3, respectively (1). Serine-phosphorylated STAT3 induced by LMP1 has been shown to specifically bind to the promoter and enhancers that regulate expression of Bcl-3 leading to the formation of p50/Bcl-3 complexes, the unique form of NF-κB that regulates EGFR expression (21). STAT3 is activated in NPC and B-cell lymphomas that develop in LMP1-transgenic mice (2, 40). Inhibition of STAT3 adversely affects the growth of lymphomas in vitro (40). A positive autoregulatory loop, in which LMP1-activated STAT3 regulates LMP1 expression through effects on the novel LMP1 terminal repeat promoter, has been described (4). LMP1-mediated tyrosine phosphorylation of STAT3 has been linked to JAK1 or JAK3, although whether LMP1 interacts directly with JAKs is controversial (12, 13). In this study, the effects of LMP1 on EGFR and STAT3 activation were further analyzed. The data indicate that LMP1 through effects on PKCδ induces serine phorphorylation of STAT3, ERK activation, and constitutive phosphorylation of EGFR.
Cervical carcinoma cell line C33A cells, Rat-1 cells, and 293T cells were cultured in Dulbecco's modified Eagle's medium (DMEM) (Gibco) supplemented with 10% fetal bovine serum (FBS) (Sigma) and antibiotic/antimycotic (Gibco) at 37°C with 5% CO2. Constructs expressing HA-tagged pBabe LMP1, pBabe-HA-LMP1-CTAR1 (which contains aa 1 to 231 of LMP1) and pBabe-HA-LMP1-CTAR1 (1-220) (which contains aa 1 to 220 of LMP1 and has signaling and transforming abilities that are similar to those of pBabe LMP-1-CTAR1), were generated as described previously (9, 29). To generate double stables, Rat-1 cells stably expressing full-length LMP1 or vector control were transfected with a PKCδ dominant negative (DN) construct (43) by FuGENE6 transfection according to the manufacturer's instructions (Roche). Double stables were selected in DMEM containing 8 μg/ml puromycin and 10 μg/ml G418. Small chemical inhibitors, including rottlerin, Gö6976, U0126, LY294002, and AG1478, were purchased from Calbiochem. To test the effects of inhibitors on LMP1-CTAR1-mediated pathways, the growth medium was replaced with serum-free DMEM supplemented with an antibiotic or antimycotic for 24 h. Inhibitors were added for 5 h before lysates were prepared. Dimethyl sulfoxide (DMSO; Sigma) was added at 1:1,000 as the vehicle control. To test the activity of EGFR, serum-starved cells were treated with 500 ng/ml EGF (Austral Biologicals) and 100 μM AG1478 for 10 or 30 min before cell lysates were harvested.
Recombinant retrovirus production and retroviral transduction were performed as previously described (29). Subconfluent 293T cells in 100-mm plates were transfected by FuGENE 6 transfection reagent (Roche) with 5 μg pBabe (vector), pBabe-HA-LMP1-CTAR1 or pBabe-HA-LMP1-CTAR1 (1–220), and 5 μg pVSV-G and 5 μg pGag/Pol expressing plasmids. After incubation at 37°C for 24 h, the medium was replaced with fresh medium (DMEM supplemented with 10% FBS and antibiotic/antimycotic) and cells were incubated at 33°C for 24 h. The cell supernatant was clarified, and the virus-containing supernatant was collected. Stable cell lines were produced by transduction with clarified virus supernatant with 4 μg/ml Polybrene for 24 h at 37°C, followed by selection with 1 μg/ml of puromycin (Sigma).
Cells were harvested at 90% confluence, washed with cold phosphate-buffered saline (PBS) (Gibco), scrape harvested in cold PBS, centrifuged at 1,000 × g for 5 min, and lysed with radioimmunoprecipitation assay (RIPA) buffer (20 mM Tris-HCl [pH 7.5], 150 mM NaCl, 1 mM EDTA, 1% NP-40, 0.1% sodium dodecyl sulfate [SDS], 0.1% deoxycholic acid) supplemented with 0.5 mM phenylmethylsulfonyl fluoride (PMSF), 1 mM sodium orthovanadate (Na3VO4), protease, and phosphatase inhibitor cocktail (Sigma). Cell lysates were clarified, and the protein concentration was determined using the Bio-Rad DC protein assay system. Equal amounts of protein were used for SDS-polyacrylamide gel electrophoresis (PAGE), and the samples were transferred to Optitran (Schleicher and Schuell) for Western blot analysis. Membranes were blocked for 1 h at room temperature in Tris-buffered saline containing 0.1% Tween 20 and 5% nonfat dry milk. Primary antibodies include anti-GAPDH, anti-STAT3, anti-ERK, anti-phospho-ERK (Tyr 204) (Santa Cruz), anti-phospho-Akt (Ser 473), anti-phospho-STAT3 (Ser 727 and Tyr 705), anti-phospho-EGFR (Tyr 992/1045/1068; Cell Signaling), anti-activated EGFR (BD Biosciences), anti-Bcl-3 (Millipore and Santa Cruz), and anti-HA tag (Covance). A rabbit antiserum generated against the carboxyl-terminal 100 amino acids of the EGFR fused to glutathione S-transferase was kindly provided by H. Shelton Earp (University of North Carolina at Chapel Hill) and used to detect expression of total EGFR. Secondary antibodies used to detect Abs-bound proteins include horseradish peroxidase (HRP)-conjugated anti-mouse, anti-rabbit (Amersham Pharmacia), and anti-goat (Dako) antibodies. After incubation with secondary antibodies, blots were developed using Pierce Supersignal West Pico or Femto chemiluminescence system according to the manufacturer's instructions, followed by exposure to X-ray film (ISCBioexpress).
Plates of C33A cells were transfected at 80% confluence in Optimem (Gibco) media by Lipofectamine 2000 transfection according to the manufacturer's instructions (Invitrogen) with PKCδ DN or pSUPER.PKCdelta.RNAi (43, 45). Following a 4-h incubation with the DNA-Lipofectamine complexes, the medium was replaced with serum-free DMEM. At 24 h posttransfection the cells were washed with PBS and harvested as described above for immunoblot analysis.
Total RNA of cells was isolated using RNeasy kit (Qiagen). Quantitative real-time PCR (QRT-PCR) was performed using the Quantifast SYBR green RT-PCR kit (Qiagen). PCR products were detected using the ABI 7900HT sequence detection system (Applied Biosystems) and analyzed using SDS 2.0 software (Applied Biosystems). The cycle threshold (CT) was determined as the number of PCR cycles required for reactions to reach an arbitrary fluorescence value within the linear amplification range. The change in CT (ΔCT) was determined between the same target gene primer sets and different samples, and the change in ΔCT (ΔΔCT) was determined by adjusting for the difference in the number of cycles required for GAPDH to reach the CT. The fold change was determined as 2ΔΔCT, since each PCR cycle results in a 2-fold amplification of PCR products. Primers used in this study are listed in Table 1.
Gene identifiers (GeneID) were as follows: EGFR, 1956; SOCS3, 9021; Bcl-xL, 598; Fos, 2353; CEBPD, 1052; GAPDH, 2597; Cyclin D1, 595; Bcl-3, 602; EGF, 1950; amphiregulin, 374; TGFα, 7039; HB-EGF, 1839; epiregulin, 2069.
Focus formation assays were performed as described previously (28). Subconfluent Rat-1 fibroblast cells plated in 6-well plates were transduced with recombinant retrovirus for 24 h. Fresh medium (DMEM supplemented with 10% fetal bovine serum and antibiotic/antimycotic) were then changed every other day for 10 days. The effect on focus formation by inhibition of PKCδ using rottlerin was determined using 1 μM rottlerin or vehicle control DMSO added daily with freshly changed medium. After foci could be clearly identified, cells were stained with 1% crystal violet (dissolved in 50% methanol) and images were taken under a stereomicroscope.
To test the effect of the PKCδ DN on rodent fibroblast transformation, cell lines stably expressing pBABE vector control, LMP1, pBABE with PKCδ DN, and LMP1 with PKCδ DN were established and grown for 10 to 14 days to reach confluence and assessed for loss of contact inhibition. Cells were stained with crystal violet and observed for overgrowth and loss of contact inhibition at ×10 magnification with phase contrast.
Data from three or more independent experiments were used to compute mean averages and standard errors. Statistical significance was evaluated using a computerized, paired two-tailed Student t test. Differences were considered significant at P < 0.05.
We have previously shown that LMP1, through LMP1-CTAR1, induces EGFR expression through activation of NF-κB p50 and transcriptional upregulation of Bcl-3 by activated STAT3 (21, 46). To investigate the effects of LMP1 on activation of EGFR and STAT3, C33A cells stably transduced with LMP1-CTAR1 were cultured in 10% serum or serum starved for 24 h and whole-cell lysates were analyzed by immunoblotting (Fig. 1A). In serum-supplemented medium, LMP1-CTAR1 induced expression and activation of EGFR, determined using antibody specific for activated EGFR (lane 2). LMP1-CTAR1 also induced both tyrosine and serine phosphorylation of the transcriptional activator STAT3 but did not affect the total level of STAT3. The levels of Bcl-3 were increased approximately 2-fold. Interestingly, EGFR and STAT3 were still phosphorylated in the absence of serum and expression of Bcl-3 was induced to levels similar to those for cells grown in 10% FBS (lane 3). These data indicate that the activation of EGFR and STAT3 by LMP1 are independent of the addition of growth factors in the medium.
It is known that EGFR interacts with STAT3 and functions as a tyrosine kinase to activate STAT3 (37). To determine the effects of EGF treatment on LMP1-CTAR1-induced EGFR and STAT3 phosphorylation, serum-starved LMP1-CTAR1-expressing C33A cells were treated with 500 ng/ml of EGF with and without AG1478, a selective inhibitor of the EGFR tyrosine kinase, and analyzed by Western blotting (Fig. 1A). Treatment with EGF for 10 min greatly increased the tyrosine phosphorylation of STAT3 but did not affect the serine phosphorylation of STAT3 or expression of Bcl-3 (lane 4). The EGF-induced tyrosine phosphorylation of STAT3 was inhibited by AG1478, indicating that EGF-induced STAT3 tyrosine phosphorylation is dependent on EGFR kinase activity (lane 5). Treatment with EGF for 30 min resulted in slightly lower tyrosine phosphorylation of STAT3 than was found with 10 min of EGF treatment (lane 6). Importantly, expression of Bcl-3 was not reduced by inhibition of EGFR kinase, indicating that the induction of Bcl-3 by LMP1 was not mediated through EGFR activation of STAT3.
The major band signal detected with antibody specific for activated EGFR was not changed by EGF treatment; however, a subtly shifted band was detected in both 10-min and 30-min EGF-treated samples (Fig. 1A, lanes 4 and 6, indicated by the dashed arrow) that was not detected with simultaneous AG1478 treatment (lanes 5 and 7). The shifted band was more apparent in a second experiment with less protein (exp. 2). These data suggest that in addition to effects of LMP1-CTAR1 in the absence of serum, EGF treatment stimulated additional phosphorylation of EGFR and the tyrosine phosphorylation of STAT3.
To determine if the STAT3 phosphorylation induced by EGF treatment affected the expression of STAT3 target genes, the RNA of EGF-treated LMP1-CTAR1 cells was isolated and analyzed for expression of several previously identified STAT3 targets, including SOCS3, Bcl-xL, Fos, CEBPD, and Cyclin D1, using real-time quantitative RT-PCR (Fig. 1B) (42). The RNA levels of SOCS3 and Bcl-xL were not significantly changed after 10 or 30 min of treatment with EGF. Expression of CEBPD went up by 2.2-fold with 10 min of EGF treatment but decreased down to 1.4-fold after 30 min of treatment with EGF. The expression of Cyclin D1 increased by 2.7-fold after 10 min of EGF treatment and stayed at 3-fold after 30 min of treatment with EGF. The mRNA of Fos changed most dramatically with a 2-fold induction after 10 min of EGF treatment and an increase of approximately 15.8-fold following 30 min of treatment with EGF. Statistical analysis revealed that the changes of Cyclin D1 and Fos were significant (P < 0.05). These data suggest that in C33 cells in response to EGF treatment a subset of previously identified targets of STAT3 are induced.
To determine if LMP1-CTAR1-mediated serum-independent activation of EGFR reflects effects on EGFR ligands, real-time quantitative RT-PCR was performed (Fig. 2A). Expression of the EGFR ligands, EGF and amphiregulin, was not detected using QRT-PCR (data not shown). Transcription of TGFα, heparin-bound EGF (HBEGF), and epiregulin (EREG) was also assessed in serum-starved vector control and LMP1-CTAR1-expressing C33A cells. EGFR expression was analyzed as a positive control, and expression levels were normalized to GAPDH. The mRNA levels for TGFα and HBEGF were not affected in LMP1-CTAR1-expressing cells compared to those in vector control cells, while the mRNA expression of epiregulin was induced by 2.4-fold in LMP1-CTAR1-expressing cells in the absence of serum compared to that in vector control cells. This result indicates that in C33 cells, LMP1-CTAR1 increases epiregulin transcription that could contribute to LMP1-CTAR1-mediated EGFR activation.
To determine if LMP1-CTAR1 activates EGFR by mimicking ligand-triggered autophosphorylation of EGFR, EGFR phosphorylation was evaluated by Western blot analysis using antibodies specific for phosphorylated tyrosine residues, Tyr 992, Tyr 1045, and Tyr 1068, that are sites of autophosphorylation (Fig. 2B) (33). Compared to LMP1-CTAR1-expressing C33A cells, vector control cells express a very low level of EGFR and the phosphorylation of EGFR was not detectable after EGF treatment. In contrast, phosphorylation of Tyr 1045, Tyr 1068, and Tyr 992 (to a lesser extent) were all increased after EGF treatment in serum-starved LMP1-CTAR1-expressing cells concomitant with the size shift detected by activated EGFR antibody (indicated by the dashed arrow). These specific EGF-induced sites of phosphorylation were blocked by the EGFR kinase inhibitor AG1478, confirming that they are dependent on EGFR autophosphorylation. However, these three sites of phosphorylation were not detected in LMP1-CTAR1-expressing cells without EGF treatment, suggesting that the constitutive phosphorylation of EGFR was not due to autophosphorylation of these sites. This finding also suggests that the basal levels of EGFR activation are likely not due to epiregulin binding.
Multiple kinases have been shown to mediate serine phosphorylation of STAT3, including several pathways that are activated by LMP1 (5). To assess the effects of specific kinases on LMP1-induced STAT3 phosphorylation, LMP1-CTAR1-expressing C33A cells were treated with either DMSO or inhibitors of previously identified serine kinases of STAT3, including rottlerin (PKCδ inhibitor), Gö6976 (PKCα/β inhibitor), U0126 (MEK/ERK inhibitor), and LY294002 (PI3K inhibitor) (Fig. 3A) (5). As previously shown, LMP1-CTAR1 (lane 2) induced EGFR expression, Bcl-3 expression, and phosphorylation of STAT3 compared to parental C33A cells (lane 1). The PKCδ inhibitor rottlerin inhibited LMP1-CTAR1-induced serine phosphorylation but not tyrosine phosphorylation of STAT3 at 5 μM and more significantly at 40 μM (lane 3 and 4). At 40 μM, rottlerin also reduced LMP1-CTAR1-induced Bcl-3 protein expression, confirming previous findings that Bcl-3 transcription was regulated by serine-phosphorylated STAT3. The effects of the inhibitors on the levels of STAT3 serine phosphorylation and Bcl-3 expression from three independent experiments were quantitated and graphically represented (Fig. 3). The levels of tyrosine-phosphorylated STAT3 could not be quantitated due to the high background with the anti-Tyr antibody. Treatment with 5 μM rottlerin decreased serine-phosphorylated STAT3 by approximately 50%, while treatment with 40 μM rottlerin decreased STAT3 serine phosphorylation by more than 80%. Similarly, Bcl-3 was decreased 40 to 60% with increasing rottlerin treatment. Treatment with the PKCα/β inhibitor, Gö6976, did not decrease serine phosphorylation of STAT3, suggesting that LMP1-CTAR1-induced serine phosphorylation of STAT3 was PKCδ specific and independent of PKCα/β (Fig. 3A, lane 5, Fig. 3B). MAPK has been shown to phosphorylate STAT3, and inhibition of MEK/ERK with U0126 effectively decreased LMP1-CTAR1-induced ERK phosphorylation but did not block LMP1-CTAR1-induced serine phosphorylation of STAT3 (lane 6). Similarly, inhibition of PI3K with LY294002 blocked LMP1-CTAR1 activation of Akt but did not affect LMP1-induced serine STAT3 phosphorylation or ERK activation (lane 7). These results suggest that PKCδ, but not PI3K, MAPK, or PKCα/β, is responsible for LMP1-CTAR1-induced serine phosphorylation of STAT3.
LMP1-CTAR1 has been previously shown to significantly activate Akt and ERK in rodent fibroblasts (29). However, in the highly transformed, malignant parental C33A cell line, ERK and Akt were highly activated and only slightly elevated levels of Akt and ERK phosphorylation were detected in LMP1-CTAR1-expressing C33A cells. Importantly, inhibition of PKCδ also blocked phosphorylation of ERK (Fig. 3A, lane 3 and 4), suggesting that LMP1 through PKCδ activates MAPK, leading to phosphorylation of ERK. The ability to specifically inhibit activation of PI3K or ERK indicates that these pathways are independently activated by LMP1.
To evaluate the effects of rottlerin on the activation of the endogenous kinases in the absence of LMP1-CTAR1, the pBabe vector control cells were treated with rottlerin and compared to the effects in the LMP1-CTAR1 cells (Fig. 3B). Elevated levels of serine-phosphorylated STAT3 were readily detected in the LMP1-CTAR1-expressing cells, and treatment with rottlerin effectively inhibited serine phosphorylation in a dose-dependent manner in the LMP1-CTAR1 expressing cells and also eliminated the low levels of serine phosphorylated STAT3 in the vector control cells. Bcl-3 was not detected in the vector control cells; however, rottlerin clearly dose-dependently decreased Bcl-3 in the LMP1-CTAR1 cells. An increase in activated ERK was not detected in the LMP1-CTAR1 C33 cells in comparison to the level in the pBabe vector control; however, 5 μM rottlerin treatment blocked the activation of ERK in the LMP1-CTAR1 cells and at 40 μM decreased activated endogenous ERK (Fig. 3B). These data suggest that in addition to serine phosphorylation of STAT3, PKCδ is also responsible for the LMP1-CTAR1-induced and endogenous ERK phosphorylation in C33A cells. Interestingly, treatment with rottlerin also inhibited the constitutive phosphorylation of EGFR. PKCδ has been previously shown to be an upstream mediator of EGFR activation, and in some systems, this activation has been suggested to be mediated by c-src (15).
It has been shown that at higher doses rottlerin can inhibit other kinases; therefore, to demonstrate the specificity of PKCδ for these effects, vector control and LMP1-CTAR1 cells were transfected with either a PKCδ RNA interference (RNAi) or a dominant negative (DN) PKCδ (Fig. 4) (43, 45). The RNAi had slight effects on serine phosphorylation of STAT3, ERK activation, or EGFR phosphorylation. However, the PKCδ DN reduced STAT3 serine phosphorylation, ERK phosphorylation, and levels of activated EGFR. Bcl-3 and total EGFR were slightly decreased, while total levels of STAT3 and activation of Akt were not affected. These data support the findings with rottlerin and suggest that PKCδ is responsible for LMP1-CTAR1 mediated effects on serine-phosphorylated STAT3, constitutive activation of EGFR, and activation of ERK.
Previous studies have shown that inhibition of LMP1-CTAR1-mediated activation of ERK blocked rodent fibroblast transformation (29, 39). To test the requirement for PKCδ in LMP1-mediated transformation and confirm PKCδ-mediated activation of ERK, the effect of its inhibition was assessed using a focus formation assay in Rat-1 fibroblast cells (Fig. 5A). Subconfluent Rat-1 cells were transduced with LMP1-CTAR1 (1–220)-containing retrovirus at three different dilutions (10−1, 10−2, and 10−3). Many foci were induced at the 10−1 dilution of LMP1-CTAR1, and focus formation decreased proportionally with the 10−2 and 10−3 retrovirus dilutions, indicating that focus formation was LMP1-CTAR1 specific. Treatment with 1 μM rottlerin blocked focus formation at all three dilutions. Western blot analysis confirmed that ERK activation correlated with the levels of LMP1 retrovirus and ERK phosphorylation was blocked by rottlerin treatment (Fig. 5B). Immunoblotting with total-ERK antibody indicated that the total level of ERK was not affected and GAPDH expression was assessed as an internal loading control. Comparison of the pBabe control Rat-1 cells with LMP1 cells confirmed the robust activation of ERK in Rat-1 cells in contrast to the slight effects in C33 cells, and this activation was decreased by treatment with 5 μM rottlerin and totally blocked by 40 μM rottlerin (Fig. 5C). These findings suggest that PKCδ is required for LMP1-mediated transformation of Rat-1 cells and that PKCδ mediates the activation of ERK by LMP1.
To test the effects on transformation of specific inhibition of PKCδ, the PKCδ DN was stably expressed in the pBABE control and CTAR1 Rat-1 cells and the loss of contact inhibition was evaluated (Fig. 6). As previously described, the vector control Rat-1 cells formed monolayers that were contact inhibited whereas the LMP1 cells continued to grow and form mounds of cells. The PKCδ DN effectively blocked this process, and both the vector control and the LMP1 cells stopped growing after formation of a confluent monolayer (Fig. 6A). Evaluation of ERK activation indicated that the PKCδ DN inhibited LMP1-mediated ERK activation but did not affect basal levels of activated ERK (Fig. 6B).
This study shows that LMP1-CTAR1 induces EGFR expression and activation of EGFR, STAT3, and ERK through its effects on PKCδ (Fig. 7). In the presence of LMP1 and in the absence of serum, EGFR was constitutively phosphorylated but not at the canonical autophosphorylation sites at tyrosine 992, 1045, or 1068. These tyrosines are also known to be critical for recruitment of STAT3 to EGFR (18, 33). In LMP1-expressing cells, STAT3 was constitutively phosphorylated on serine and tyrosine and treatment with EGF induced abundant tyrosine phosphorylation of STAT3 and expression of a subset of known STAT3 target genes (42). LMP1-CTAR1 also induced serum-independent serine phosphorylation of STAT3, which was not changed by treatment with EGF or its inhibitor, AG1478. These data indicate that the LMP1-CTAR1-induced serine phosphorylation of STAT3 is not dependent on EGFR signaling. Specific inhibition of multiple serine kinases revealed that only the PKCδ-inhibitor rottlerin reduced serine phosphorylation of STAT3, suggesting that PKCδ, but not PKCα/β, ERK, or PI3K, is responsible for LMP1-CTAR1-induced serine phosphorylation of STAT3. These findings were confirmed using a PKCδ DN. LMP1 has not been previously shown to activate PKCδ-dependent signaling pathways, although it does regulate Annexin 2 and Ezrin through PKCα/β (8, 27). In addition to blocking serine phosphorylation of STAT3, rottlerin and PKCδ also decreased expression of Bcl-3. We have previously shown that LMP1 induced binding of serine-phosphorylated STAT3 to the Bcl-3 promoter and enhancer to regulate Bcl-3 (21). These findings suggest that distinct cellular genes are regulated by serine- or tyrosine-phosphorylated STAT3. PKCδ-mediated serine phosphorylation of STAT3 has been shown to be important for keratinocyte proliferation, and LMP1 activation of STAT3 is required for the enhanced growth properties of LMP1 transgenic lymphocytes (11, 40).
Interestingly, inhibition of PKCδ, which is known to activate the MAPK pathway through several mechanisms, also blocked both LMP1-CTAR1-induced and endogenous ERK phosphorylation (17). The activation of ERK by LMP1 was blocked by inhibition of MAPK, which did not affect serine phosphorylation of STAT3. Additionally, the activation of PI3K and Akt was not affected by either inhibition of MAPK or PKCδ, indicating that these pathways are independently activated. These findings suggest that PKCδ functions as a master regulator to activate STAT3 and ERK as distinct pathways (Fig. 7). The inhibition of focus formation and ERK phosphorylation in Rat-1 cells through inhibition of PKCδ suggests that PKCδ is required for LMP1-mediated transformation through activation of ERK.
The serum-independent tyrosine phosphorylation of EGFR may reflect the expression of potential ligands, as low levels of transcription of TGFα and HBEGF and slightly increased epiregulin were detected (Fig. 7). It is possible that different EGFR ligands may be affected by LMP1 expression in distinct cell lines. Interestingly, mice with LMP1 transgenes develop epidermal hyperplasia, and this is consistent with increased expression of activated EGFR and its ligands, TGFα, HBEGF, and epiregulin (3). Upon ligand binding, oligomerization and autophosphorylation of EGFR activate multiple signaling pathways (18). The LMP1-CTAR1 effects on potential EGFR ligands may be responsible for the low levels of tyrosine phosphorylation of STAT3 observed in serum-starved LMP1-CTAR1 cells. However, phosphorylation of the specific tyrosines known to be phosphorylated after ligand binding that enable STAT3 activation was not detected.
Intriguingly, inhibition of PKCδ using rottlerin or the PKCδ DN blocked the intrinsic phosphorylation of EGFR induced by LMP1-CTAR1. Other tyrosine kinases, such as Src, have been shown to phosphorylate EGFR and regulate its activity, and this phosphorylation has been linked to EGFR inhibitor-resistant cancers (35). Activation of c-src by PKCδ leading to the phosphorylation of EGFR has been described (15). The phosphorylation and properties of EGFR are complex, and its phosphorylation profile can differ remarkably in different cell lines. At least nine tyrosines can potentially be phosphorylated by c-src (33). It will be important to characterize the effects of LMP1 on specific sites of EGFR phosphorylation and determine the potential effects of LMP1 on the activity of c-src.
EGFR expression and STAT3 activation are elevated in NPC, and STAT3 activation is important for the invasiveness of EBV-associated NPC, which can be blocked by STAT3 inhibitor treatment (25, 26). The effects of LMP1 on EGFR, STAT3 activation, and regulation of STAT3 target genes may be important factors in the development of NPC. In breast carcinoma, the EGFR signaling pathway induces epithelial-mesenchymal transition (EMT) of cancer cells through STAT3-mediated TWIST gene expression and subsequently E-cadherin downregulation (23). LMP1 has also been shown to induce TWIST expression and EMT in NPC cells (14). These findings suggest that LMP1 mediates the EGFR-STAT3 signaling pathway to regulate cellular transcriptional and biologic properties.
In summary, this study reveals that LMP1-CTAR1 induces the intrinsic activation of EGFR and STAT3 signaling pathways. The combined effects of LMP1 on EGFR and STAT3 likely regulate the transcription of many cellular genes. Importantly, this study has identified PKCδ as the likely serine kinase of STAT3, as well as the upstream effector of ERK and EGFR activation mediated by LMP1-CTAR1 (Fig. 7). The ability of PKCδ inhibitor rottlerin and PKCδ DN to inhibit LMP1-CTAR1-induced transformation points to the importance of this kinase in LMP1-mediated oncogenesis. The further study of the effects of LMP1 on EGFR, PKCδ, and STAT3 will likely increase our understanding of the mechanisms through which EBV contributes to oncogenesis.
We thank H. Shelton Earp for providing anti-EGFR rabbit antiserum.
This work was supported by NIH grants CA32979 and CA19014 to N.R.-T., and D.G.M. was supported by training grant T32CA009156.
Published ahead of print on 9 February 2011.