Upregulation of Id1 by Epstein-Barr Virus-encoded LMP1 confers resistance to TGFβ-mediated growth inhibition

LMP1 confers growth and transforming properties to epithelial cells by activating multiple signal cascades. These include the PI3K/Akt, ERK-MAPK and NFκB signalling pathways amongst others. Activation of these three pathways results in suppression of the transcriptional activity of Foxo3a [4]. One consequence of Foxo3a inactivation by LMP1 is inhibition of DNA repair [14]. Here, we examine additional downstream consequences of Foxo3a inactivation by LMP1. In keeping with previously published findings, we demonstrate that transient expression of LMP1 in HEK-293 cells stimulated Akt, Erk1/2 and IκB phosphorylation in a dose dependent manner and was accompanied by Foxo3a phosphorylation and protein degradation (Figure 1A) [2, 14]. In agreement with previously published studies, reduction of p27^kip, a transcriptional target of Foxo3a by LMP1 was also observed [15]. An examination of a nasopharyngeal epithelial cell line NP69 stably expressing LMP1 (NP69-LMP1) revealed increased cytoplasmic levels of Foxo3a, and an overall reduction in total Foxo3a and p27^kip protein (Figure 1B). Stable expression of LMP1 was also accompanied by increased phosphorylation of IκB, Akt and Erk1/2. To examine further the effect of LMP1 on Foxo3a-mediated transcription, luciferase assays were performed using promoter reporter constructs of two established Foxo3a target genes: p27^kip and Bim. As shown in Figure 1C, transient expression of LMP1 in HEK293 cells attenuated the activity of both promoter reporters in a dose-dependent manner. In the reciprocal experiment, exogenous expression of Foxo3a enhanced the activities of both p27^kip and Bim promoter reporters. This induction was antagonised by LMP1 expression (Figure 1D). Taken together, these data confirm that LMP1-induced phosphorylation, nuclear translocation and degradation of Foxo3a ablate Foxo3a transcriptional activity in epithelial cells.

A recent report has shown that Foxo3a downregulates Id1 transcription [6]. This led us to investigate whether inactivation of Foxo3a by LMP1 impacted on Id1 expression. Firstly, immunoblotting confirmed increased levels of Id1 expression in epithelial cells expressing LMP1 (Figure 1A & 1B), findings that are consistent with previously published data [16, 17]. Furthermore, transient overexpression of HA-tagged Foxo3a in NP69-LMP1 cells resulted in a marked reduction in Id1 protein expression (Figure 2A), confirming the reciprocal relationship between Foxo3a and Id1 expression. Using an Id1 (-1695) promoter reporter construct, we found that transient expression of LMP1 augmented Id1 promoter activity in HEK-293 cells in a dose-dependent manner (Figure 2B). While Foxo3a inhibited the transcriptional activity of Id1 promoter, LMP1 counteracted this negative effect (Figure 2C). Foxo3a has been shown to repress Id1 transcription through direct binding to the Id1 promoter at position -134 to -128 bp upstream of the ATG [6]. To evaluate further the interplay between Foxo3a, Id1 and LMP1, a shorter Id1 promoter construct (-353) was transfected into NP69 nasopharyngeal epithelial cells together with increasing amounts of LMP1. As shown in Figure 2D, LMP1 increased the luciferase activity of this shorter Id1 promoter construct. In addition, the suppressive effect of Foxo3a on this shorter Id1 promoter element was antagonised by LMP1 (Figure 2E). Taken together, these data confirm that LMP1 limits the ability of Foxo3a to repress Id1 promoter transcription.

TGFβ is a potent regulator of squamous epithelial homeostasis acting as a tumour suppressor by inducing cell cycle arrest. Id1 has multiple oncogenic functions imparting resistance to TNFα and anti-cancer drug-induced apoptosis [18]. Here, we sought to investigate whether Id1 confers pro-survival properties in NP69 cells, a nasopharyngeal epithelial cell line that is responsive to TGFβ-mediated growth inhibition [19]. Using both cell-cycle and proliferation assays, we found that stable expression of Id1 in NP69 cells enhanced cell proliferation and overcame TGFβ-mediated G1 cell cycle arrest (Figure 3A & 3B). Inhibition of TGFβ-mediated growth arrest by LMP1 in B cells and fibroblasts has been reported previously [10, 11]. Using an epithelial cell model, we set out to explore whether the resistance to TGFβ afforded by LMP1 was associated with increase expression of Id1. Firstly, NP69-pLNSX and NP69-LMP1 cells were transduced with retroviruses containing either individual shRNAs to Id1 (shId1B or shId1C), or both (shId1B+C). After drug selection, the suppressive function of Id1 shRNAs in stably established cell lines was validated [see Additional file 1].

Using cell cycle analysis, NP69-LMP1 pSuper.retro cells maintained in normal KSF medium were found to contain higher percentage of cells in the S and G2 phases of the cell cycle compared to NP69-pLNSX pSuper.retro control cells, demonstrating that LMP1 promotes cell proliferation (Figure 3C). NP69-LMP1 pSuper.retro cells grown in normal medium (31.1%) or in medium supplemented with TGFβ (30.1%) showed similar percentage of cells in the G1 phase of the cell cycle, while the NP69-pLNSX pSuper.retro control cells treated with TGFβ had higher percentage of cells in the G1 phase (50.1%) compared to untreated cells (38.4%). These findings confirm the ability of LMP1 to protect against TGFβ-mediated G1 cell cycle arrest in this nasopharyngeal epithelial cell line (Figure 3C). The role of Id1 in this response was established as NP69-LMP1 cells expressing shRNAs to Id1 exhibited clear-cut cell cycle arrest, with 48.1% of the cell population in the G1 phase compared to NP69-LMP1 Super.retro, where 30.1% of the cell population was in the G1 phase (Figure 3C). This finding was further supported by MTT cell proliferation assays. As shown in Figure 3D, NP69-LMP1 pSuper.retro cells were relatively refractory to TGFβ-mediated growth inhibition. However, silencing Id1 by shRNA reduced the growth of NP69-LMP1 cells in normal medium (-17.92%) and the growth inhibition was increased further in the presence of TGFβ (-45.65%). Taken together, these data confirm that Id1 plays a significant role in LMP1-mediated cell proliferation and resistance to the growth inhibitory effects of TGFβ.

To determine whether Id1 confers resistance to TGFβ-mediated cytostasis by inhibiting TGFβ-mediated SMAD transcription, an Id1 expression vector was co-transfected along with the SMAD-responsive reporter construct, pGL3(CAGA) or p3Tplux into HEK-293 cells. Twenty-four hours post-transfection, cells were subjected to TGFβ treatment for 16 hours prior to harvesting for luciferase reporter analysis. As shown in Figure 4A, increased expression of Id1 resulted in a dose-dependent reduction of TGFβ-induced SMAD transcription. This experiment was also performed in HepG2 hepatocellular liver carcinoma cells and NP69 nasopharyngeal epithelial cells, where similar findings were observed [see Additional file 2]. All these results suggest that Id1 is able to inhibit the TGFβ-SMAD-mediated transcription.

LMP1 has previously been reported to inhibit SMAD transcription [12, 13]. Here, we reveal that induction of Id1 by LMP1 plays a direct role in this inhibition. As shown in Figure 4B, expression of LMP1 in HEK-293 cells suppressed TGFβ-mediated transcriptional activity of pGL3(CAGA) and p3Tlux reporter constructs. However, addition of Id1 shRNA to silence the expression of Id1 antagonised this suppressive effect, while scrambled shRNA treatment showed no such effect. Similar findings were also observed in HepG2 and NP69 cells [see Additional file 3]. The effects of Id1 silencing are similar to the effects of Foxo3a activation, as Foxo3a has been shown to negatively regulate the expression of Id1 [6]. Exogenous expression of Foxo3a also antagonised the suppressive effect of LMP1 on TGFβ-mediated SMAD transcription (Figure 4C and see Additional file 4). In summary, LMP1 induction of Id1 participates in suppressing the TGFβ-SMAD-mediated transcription.

We have found that LMP1 suppresses TGFβ-mediated SMAD transcription without affecting SMAD phosphorylation. Over a time-course of 48 hours, TGFβ treatment stimulated phosphorylation of the SMAD2 and SMAD3 proteins in both NP69-pLNSX and NP69-LMP1 cells as early as 2 hours after the addition of TGFβ and high levels of phosphorylated SMAD proteins persisted thereafter although gradually declining by 48 hours post-stimulation (Figure 5). Expression of p21 protein, a downstream target of TGFβ who expression is required for TGFβ-mediated cytostasis, gradually increased in NP69-pLNSX control cells 2 hours after TGFβ treatment and reached its peak, with almost a three fold induction, at 12 hours. Its expression then declined to basal levels by 48 hours post-stimulation. In NP69-LMP1 cells, a relatively modest induction of p21 protein was observed upon TGFβ treatment; however, the overall p21 protein level in NP69-LMP1 was significantly lower compared to that in NP69-pLNSX control cells. These findings demonstrate that LMP1's suppressive effect on TGFβ-mediated induction of p21 is independent of SMAD phosphorylation, suggesting that the suppressive effect of LMP1 on SMAD transcriptional activity does not involve formation of activated SMAD complex.

We found that expression of the Id1 protein increased in both NP69-pLNSX and NP69-LMP1 cells 2 hrs after TGFβ addition. Thereafter, high levels of Id1 persisted in NP69-LMP1 cells, while in NP69-pLNSX cells, the levels of Id1 protein gradually decreased reaching basal levels 48 hours post-stimulation. During the time course following TGFβ treatment, the levels of Foxo3a did not change significantly in either NP69-pLNSX or NP69-LMP1 cells although the overall levels of Foxo3a protein were lower in NP69-LMP1 compared to NP69-pLNSX cells. These data show that Id1 is induced in LMP1-expressing cells in response to TGFβ stimulation and that this induction is not likely associated with the expression and/or activity of Foxo3a.

Massagué and colleagues have demonstrated that Id1 is transiently induced by TGFβ-activated SMAD3 but long-term TGFβ stimulation results in Id1 transcriptional repression, which is dependent on induction of the ATF3 transcriptional repressor [20]. Here, we found that the basal levels of ATF3 were low in NP69-LMP1 cells relative to NP69-pLNSX cells. After addition of TGFβ, the expression of ATF3 increased in NP69-pLNSX cells at 4 hours and peaked at 12 hours, while in NP69-LMP1 cells, ATF3 protein was slightly increased at 4 hours but was reduced thereafter. This finding suggests that LMP1 inhibition of ATF3 may prolong TGFβ-mediated induction of Id1. The effect of LMP1 on ATF3 suppression was further confirmed in NP69 cells, where transfection of LMP1 suppressed ATF3 protein expression in a dose dependent manner (Figure 6).

In an examination of primary NPC tumours (T1, T2 and T3), which displayed strong, moderate, and weak expression of LMP1 respectively, we observed a positive correlation between expression of LMP1 and that of Id1, whereas expression of Foxo3a was inversely correlated with LMP1 expression. For example, tumour T1 shows strong staining for both LMP1 and Id1, but weak Foxo3a nuclear staining. In contrast, tumour T3 showed strong nuclear staining of Foxo3a but weak detection of LMP1 and Id1 proteins. While in the normal nasopharyngeal epithelium (NP) which is LMP1 negative, we found weak Id1 expression but strong nuclear Foxo3a staining (Figure 7). These data suggest that LMP1 is involved in suppressing Foxo3a activity and increasing Id1 expression during NPC progression.

The Id1 (-1695) and (-353) promoter luciferase reporters were generated by PCR and cloned into the pGL3 basic vector (Promega). The Id1 shRNA B & C expression vectors were generated by inserting a fragment of synthesised Oligo (Sigma) with the sequence of Id1 coding region into pSUPER.retro.puro vector (oligoengine). The sequence of Id1 shRNA B is GATCCCCGCGCGCTGAAGGCCGGCAATTCAAGAGATTGCCGGCCTTCAGCGCGCTTTTTA and Id1 shRNA C is GATCCCCGGTGCGCTGTCTGTCTGAGTTCAAGAGACTCAGACAGACAGCGCACCTTTTTA. pECE-HA-Foxo3a was kindly provided by M. Deckert [31], pGL2-Bim vector is gift of P.J. Coffer [32], GFP-Foxo3a is a gift from MC Hung [33]. The pGL2-p27kip promoter construct was provided by T Sakai [34]. P3Tplux was provided by J Massagué Laboratory, Memorial Sloan-Kettering Cancer Center, New York, USA and pGL3(CAGA) was kind gift from CS Hill (The Cancer Research UK London Research Institute, London, UK).

The detailed procedures of Western blotting have been described previously [35]. Briefly, cells were lysed in RIPA buffer. Total cell lysates (5-25 μg of protein) were separated by 10% SDS-PAGE and then electrophoretically transferred to nitrocellulose membrane prior to immunoblotting. Antibodies specific for Phospho-IκBα (Ser32), Phospho-Erk1/2 (Ser217/221), phospho-Akt (Ser473), phospho-Foxo3a (Ser253), Foxo3a and p27 were purchased from Cell Signalling, USA. Antibody specific for ATF3 was from Abcam. Antibodies to SP-1 and α-tubulin were purchased from Santa Cruz, USA. Antibodies to LMP1 were purchased from Dako and β-Actin from Sigma, UK.

The expression of LMP1 and Foxo3a in paraffin-embedded NPC specimens was examined by immunohistochemistry as described previously [36]. Primary antibodies used in this study were anti-LMP1 mouse monoclonal antibody (DAKO) and anti-Foxo3a rabbit polyclonal antibody (Cell Signalling).

1×10⁵ cells grown in 24-well plates were co-transfected with 40 ng of luciferase reporter constructs together with different amounts of expression vectors as indicated in the text. RSV-β-Gal vector (50 ng) was transfected as an internal control to normalise for transfection efficiency. Two days post-transfection, cells were lysed in reporter lysis buffer (Promega) and then assayed for luciferase and β-gal activities. For detection of Id promoter activity, transfected cells were cultured in serum free medium for 6 hrs before harvesting. For detection of TGFβ responsive promoter activity of pGL3(CAGA) and p3TLux constructs, cells were cultured in medium containing 0.2% FBS and 5 ng/ml TGFβ for 16 hrs prior to harvesting.

Cells (5 × 10⁵) were fixed in ice cold 70% ethanol for 1 hr. Prior to analysis, fixed cells were washed with PBS, treated with RNase (1 μg/ml) and stained with propidium iodide (50 μg/ml) for 30 min at 37°C. Cell cycle analysis was carried out on a XL-MCL flow cytometer (Beckman-Coulter) and data analyzed using the MultiCycle AV DNA Analysis software (Phoenix Flow Systems).

For MTT assay, cells (5×10³ per well) were seeded into 96-well plates. One day after cell seeding. TGFβ1 (10ng/ml) was added. MTT assay was analyzed each 24 hrs by adding MTT solution (5 mg/ml; 10 μl/well) and cells were incubated at 37°C for 5 hrs. The culture media were aspirated and DMSO (200 μl/well) was added to dissolve the formazan crystals. The absorbance was measured at a wavelength of 570 nm. Each time point was performed in triplicate. Results are presented as relative growth rate by dividing the absorbance value of the cells at indicated time points by the absorbance value of the cells one day after cell plating. Each data point is represented by the mean and SD.