|Home | About | Journals | Submit | Contact Us | Français|
Epidermal growth factor receptor (EGFR) tyrosine kinase is commonly overexpressed in human cancers; however, the cellular mechanisms regulating EGFR expression remain unclear. p53, p63 and p73 are transcription factors regulating many cellular targets involved in controlling the cell cycle and apoptosis. p53 activates EGFR expression, whereas TAp63 represses EGFR transcription. The involvement of p73 in the regulation of EGFR has not been reported. Here, a strong correlation between EGFR overexpression and increased levels of the oncogenic ΔNp73 isoform in head and neck squamous cell carcinoma (HNSCC) cell lines was observed. Ectopic expression of TAp73, particularly TAp73β, resulted in suppression of the EGFR promoter, significant downregulation of EGFR protein and efficient induction of cell death in all six EGFR-overexpressing HNSCC cell lines. EGFR overexpression from a heterologous LTR promoter protected lung cancer cells from TAp73β-induced EGFR suppression and apoptosis. Expression of TAp73β efficiently induced promyelocytic leukaemia (PML) protein expression and PML knockdown by shRNA attenuated the downregulation of EGFR and induction of apoptosis by p73 in HNSCC cells. Furthermore, PML was found to be important for E1A-induced suppression of EGFR and subsequent killing of HNSCC cells. Our data therefore suggest a novel pathway involving PML and p73 in the regulation of EGFR expression.
The promyelocytic leukaemia (PML) gene was initially identified as a tumour suppressor gene inactivated in patients with acute promyelocytic leukaemia as a result of its fusion with the retinoic acid receptor-α gene (Goddard et al., 1991; Kakizuka et al., 1991; Kastner et al., 1991; de The et al., 1991). PML localizes to distinct speckled subnuclear structures commonly known as PML oncogenic domains (PODs) (Salomoni and Pandolfi, 2002). PML is a tumour suppressor gene with pro-apoptotic activity (Wang et al., 1998). It modulates p53 function in response to oncogene-induced stress and DNA damage by functioning as a transcriptional co-activator of p53 by regulating p53 acetylation by CBP/p300 (Fogal et al., 2000; Guo et al., 2000; Pearson et al., 2000). Furthermore, PML-induced premature senescence has been shown to involve stabilization and activation of p53 (Bischof et al., 2002).
The p53 family member, p73, has several isoforms as a consequence of alternative promoter usage and C-terminal splicing (Melino et al., 2002). The transactivating isoforms of p73 (TAp73) have a diverse efficiency in the transactivation of downstream cellular targets whereas the anti-apoptotic, N-terminally truncated, ΔNp73 isoforms lack transactivation ability (Zhu et al., 1998; Lee and La Thangue, 1999; Ueda et al., 1999; Melino et al., 2004; Wu et al., 2004; Klanrit et al., 2008). Unlike p53, p73 is not mutated but its isoforms, particularly the ΔNp73 isoforms, are frequently overexpressed in many types of cancers (Zaika et al., 2002). The p53 family members, p53, p63 and p73 have been demonstrated to physically interact with PML and colocalize in PODs (Guo et al., 2000; Bernassola et al., 2004, 2005). Overexpression of PML results in the increased transactivation potential of p53, TAp63 and TAp73 and leads to the accumulation of p63 and p73 by inhibiting their ubiquitin- and proteasome-dependent degradation (Guo et al., 2000; Bernardi and Pandolfi, 2003; Bernassola et al., 2004, 2005). Furthermore, PML has been shown to be a transcription target of both p53 and TAp73 (de Stanchina et al., 2004; Lapi et al., 2008).
Epidermal growth factor receptor (EGFR) is an important regulator of proliferation, angiogenesis, migration, tumourigenesis and metastasis. EGFR dysregulation has been associated with many cancers especially carcinomas of lung, head and neck, and colon. These studies further validate EGFR as an important molecular target in cancer treatment (reviewed in Prenzel et al., 2001 and Kalyankrishna and Grandis, 2006). A better understanding of EGFR biology particularly pathways involved in the regulation of EGFR expression should allow the development of more effective therapeutic approaches targeting EGFR-induced signalling.
Several transcription factors including Sp1, interferon-regulated factor-1, EGFR transcription factor, activator protein-1, activator protein-2 and p53 have been shown to activate EGFR expression (reviewed in Nishi et al., 2001). Other factors including p63 and PML have been shown to suppress EGFR transcription (Vallian et al., 1998; Nishi et al., 2001). PML appears to suppress EGFR transcription through interacting with Sp1 and hence blocking Sp1-mediated transactivation of the EGFR (Vallian et al., 1998). However, the precise nature of EGFR regulation by PML and its importance in EGFR signalling, cellular proliferation and development of cancer remain unclear.
The expression of the E1A gene of human adenovirus 5 was previously shown to induce PML protein levels and cause the re-organization of PODs in p53-mutated human cancer cell lines (Flinterman et al., 2003, 2007). In addition, E1A expression induced a significant increase in the TAp73 mRNA and protein levels in head and neck squamous cell carcinoma (HNSCC) and neuroblastoma cell lines (Flinterman et al., 2005). Besides, E1A has been shown to downregulate EGFR expression in several HNSCC cell lines (Flinterman et al., 2003, 2007).
As outlined above, p53 has been shown to transactivate EGFR (Ludes-Meyers et al., 1996; Sheikh et al., 1997) whereas p53 family member TAp63γ suppresses EGFR transcription (Nishi et al., 2001). However, an association between p73 and EGFR has not yet been demonstrated. In this study we investigated whether p73 isoforms affect EGFR regulation and sought to determine the importance of PML in the regulation of EGFR expression by E1A or p73.
Expression of endogenous p73 and EGFR protein levels in a panel of cell lines was analysed by western blotting using an anti-EGFR monoclonal antibody and a rabbit anti-p73α (p73SAM) antibody that reacts only with TAp73α and ΔNp73α but not with other p73 and p63 isoforms (Sayan et al., 2005). All the HNSCC cell lines had mutations in p53 gene except HN30 cell line that has a wild-type p53 gene, Figure 1a, consequently they have stabilized and/or truncated p53 protein (Gusterson et al., 1991; Sakai and Tsuchida, 1992; Yeudall et al., 1995, 1997). Figure 1a demonstrates that all HNSCC cell lines expressed high levels of EGFR compared to other cancer cell lines including p53-null non-small-cell lung carcinoma H1299 cells, p53-null oesteosarcoma Saos-2 cells and neuroblastoma SH-SY5Y cells (known to have wild-type but non-functional p53) (Goldschneider et al., 2004). The normal human fibroblasts 1BR3 and its matched SV40 transformed, 1BR3-LT and normal human embryonal lung fibroblasts 6689, showed undetectable or very low level of EGFR. All HNSCC cell lines, except HN30, also exhibited high levels of ΔNp73α, in particular HN5 and metastatic HSC-3 M3 cells. Low levels of the endogenous TAp73α were present in all HNSCC cell lines and this was significantly lower than ΔNp73α with the exemption of HN30 cells that expressed higher TAp73α protein than ΔNp73α. Interestingly, low-passage normal human keratinocytes Km3 expressed high levels of EGFR and ΔNp73α whereas TAp73α was barely detectable, indicating the importance of EGFR and ΔNp73α in the survival and renewal of early passage normal human keratinocytes. Other cancer cell lines demonstrated low (Saos-2 and H1299 cells) to moderate levels (SH-SY5Y cells) of TAp73α whereas ΔNp73α was barely detectable in these cells. The observed expression pattern of p73 isoforms in SH-SY5Y cells is in agreement with a previous report by Goldschneider et al. (2004) demonstrating higher level of endogenous TAp73α than ΔNp73α in SH-SY5Y. Both the TAp73α and ΔNp73α isoforms were undetectable in normal 1BR3, 6689 and transformed 1BR3-LT fibroblasts. The expression of p73β isoform was analysed by reverse transcription (RT)–PCR (currently no reliable p73 isoform-specific antibodies are available) and compared with endogenous p73α. HNSCC cell lines examined were found to express low levels of p73β isoform. The expression ratio of p73α and p73β of each cell line varied from 0.9 to 1.67 (Supplementary Figure 1).
TAp73β has been shown to have the strongest apoptosis-inducing effect in Saos-2 cells (Klanrit et al., 2008). Therefore, the ability of TAp73β in killing a range of HNSCC cell lines was examined. HNSCC cells were left either uninfected or infected with the control Ad-GFP, Ad-GFP-p53 or Ad-GFP-TAp73β at a multiplicity of infection (MOI) of 10 and cell survival was analysed by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (Figure 1b). Expression of TAp73β and p53 for 48 h efficiently induced cell death in all HNSCC cell lines tested. Importantly, HNSCC cells were found to be more sensitive to TAp73β than p53 whereas lung cancer, H1299 cells were more sensitive to p53- than to TAp73β-induced apoptosis.
To determine whether p73 has an effect on EGFR transcription, Saos-2 cells, which are p53 deleted, express very low level of endogenous TAp73 and EGFR and have no detectable ΔNp73 (Figure 1a), were co-transfected with a luciferase reporter construct under the control of the EGFR promoter (Figure 2a) together with plasmids encoding TAp73α, TAp73β, TAp73γ, TAp73δ, ΔNp73α, p53 or control empty vector. Luciferase activity was measured 24 h after transfection. As shown in Figure 2b, there was a 36, 15 and 19% reduction of the EGFR promoter activity by TAp73β, TAp73γ and TAp73δ, respectively. Conversely, TAp73α did not suppress the EGFR promoter activity. In agreement with previous studies (Ludes-Meyers et al., 1996; Sheikh et al., 1997), p53 significantly transactivated EGFR by 7.4-fold compared with the control empty vector (Figure 2b). TAp73β and TAp73δ were also shown to suppress EGFR transcription in HNSCC cell line HN30 that has low levels of endogenous p73 (Figure 2c). Interestingly, p53 had no significant effect on the EGFR promoter in this cell type (Figure 2c), which could be due to the presence of wild-type p53 in HN30 cells (Yeudall et al., 1997).
Furthermore, using the luciferase reporter assay, we demonstrated that PML expression resulted in the suppression of the EGFR promoter in a dose-dependent manner (data not shown).
Next, the effect of exogenous TAp73β on EGFR protein levels was investigated. The EGFR-overexpressing HNSCC cell lines HN5 and H357 were left either uninfected or infected with Ad-GFP-TAp73β, Ad-GFP-p53 or control Ad-GFP at an MOI of 10. Expression of both p53 and TAp73β resulted in a dramatic downregulation of EGFR protein levels in H357 cells (Figure 2d, left). Although, in HN5 cells, a significant downregulation of EGFR was shown only with TAp73β and no reduction was observed in the Ad-GFP-p53-infected HN5 cells (Figure 2e, left). Furthermore, in H357 cells, apoptosis, detected by poly-(ADP-ribose)-polymerase (PARP) 85 kDa fragment, was observed 24 h after treatment with both Ad-GFP-p53 and Ad-GFP-TAp73β. In H357, TAp73β was significantly more efficient in inducing apoptosis than p53 as detected by a higher level of cleaved PARP (Figure 2d). Apoptosis was detected in HN5 cells 24 h after infection but only in cells expressing TAp73β and not p53 (Figure 2e). These data indicate that p53-deficient HNSCC cells are less sensitive to apoptosis induced by wild-type p53 than by TAp73β, confirming the MTT results shown in Figure 1b.
Promyelocytic leukaemia expression was analysed using the rabbit anti-PML antibody clone 3573 that detects various isoforms of PML proteins. Western blot analysis showed that expression of p53 or TAp73β in H357 cells highly induced the expression of PML (Figure 2d). Conversely, HN5 cells showed the induction of PML only after infection with Ad-GFP-TAp73β (Figure 2e). In addition, H357 cells infected with Ad-GFP-p53 and Ad-GFP-TAp73β for 24 h showed disaggregated PODs compared to the control cells infected with Ad-GFP (Figure 2d). Interestingly, altered PODs were only observed in HN5 cells infected with Ad-GFP-TAp73β whereas HN5 cells expressing green fluorescent protein (GFP) or GFP-p53 showed normal PODs (small, round speckles), though with less intensity than controls (Figure 2e). These results demonstrate a clear correlation between the increased levels of PML, downregulation of EGFR, and apoptosis induced by p53 and p73.
To further determine whether TAp73β is involved in the suppression of EGFR and the induction of apoptosis, non-small-cell lung cancer H1299 cells that express low level of EGFR were infected with pBABEpuro-EGFR expression vector and stable single or mixed clones of cells expressing EGFR were selected (Flinterman et al., 2003). Parental H1299, control vector-transduced cells (H1299-VE), the mixed population of exogenous EGFR-expressing cells (H1299-MC) and high-level exogenous EGFR-expressing clones (H1299-C2 and H1299-C3) were infected with Ad-GFP-TAp73β or Ad-GFP at an MOI of 10. Cell viability was measured by the MTT assay at 48 h after infection. As shown in Figure 3a, the viability of H1299-MC, H1299-C2 and H1299-C3 infected with Ad-GFP-TAp73β was significantly higher than that observed in H1299 and H1299-VE cells (P<0.001).
To determine whether the resistance of exogenous EGFR-expressing H1299 cells to apoptosis was specific to TAp73β expression, the parental H1299 cells and H1299-C3 were treated with different concentrations of DNA-damaging drugs, carboplatin and 5-fluorouracil (5-FU). No difference in cell viability between the different cell lines was detected after 3-day drug treatment (Figure 3b), indicating that overexpression of EGFR did not protect cells against apoptosis induced by DNA-damaging drugs whereas it significantly protected them against TAp73β-induced cell death.
We next compared the effect of TAp73β on the expression of PML, EGFR and cleaved PARP between H1299-VE and H1299-C3. As shown in Figure 3c, the level of endogenous EGFR expression was significantly reduced in H1299-VE at 24 h post-infection with both Ad-GFP-p53 and Ad-GFP-TAp73β but not with Ad-GFP, used at an MOI of 10. In contrast, the exogenous EGFR level was reduced only by treatment with Ad-GFP-p53 but not with Ad-GFP-TAp73β showing a clear transcriptional regulation of endogenous but not exogenous EGFR by TAp73. Induction of PML and apoptosis detected by cleaved PARP were observed in both H1299-VE and H1299-C3 infected with Ad-GFP-p53 and in H1299-VE infected with Ad-GFP-TAp73β but not in H1299-C3 infected with Ad-GFP-TAp73β. Importantly, the downregulation of EGFR was concurrently observed with the induction of PML and the presence of 85 kDa cleaved PARP protein.
We have previously demonstrated that E1A expression alters the level and pattern of PML expression in EGFR-overexpressing HNSCC cell lines (Flinterman et al., 2003). Furthermore, the sensitivity of cells to killing by p53 or p73 correlated with EGFR downregulation and with the ability to induce PML expression. We therefore speculated that PML is an important factor in TAp73β- and E1A-mediated suppression of EGFR and induction of apoptosis in p53-deficient HNSCC cells. To test this hypothesis, PML protein was silenced using expression vector encoding shRNA targeting PML. H357 and HN5 cells were electroporated with PML shRNA and selected cell populations were established. Downregulation of PML protein was confirmed by western blot analysis and immunofluorescence using both polyclonal (clone 3573) and monoclonal (PG-M3; Santa Cruz, Middlesex, UK) antibodies (Figures 4a and b). The mixed populations of cells expressing empty vector, H357pRSempty and HN5pRSempty were used as controls for H357 and HN5, respectively. As shown in Figure 4, the basal level of PML in untreated cells is very low, nonetheless a clear reduction in PML level was observed in shPML clones using both antibodies (Figures 4a and b).
MTT assay showed a significant reduction in E1A-induced cell death in PML knockdown clones derived from both H357 and HN5 cells as compared with controls (Figures 4c and d, left). There was no significant cell death in all cell types infected with the control Ad-Del virus (Figures 4c and d, right). Similar to the effect observed with Ad-E1A, PML knockdown H357 and HN5 cells showed increased resistance to Ad-GFP-TAp73β-induced killing as compared with controls (Figures 4c and d, middle). However, knockdown of PML did not confer resistance to H357 cells in response to cell death induced by treating the cells for 48 and 72 h with 100 μg/ml carboplatin (Supplementary Figure 2).
Next, the effect of PML knockdown on the downregulation of EGFR and apoptosis by E1A was investigated. As shown in Figures 5a and c, a significant reduction in EGFR level was observed in Ad-E1A-infected H357pRSempty cells (36% EGFR remaining), whereas only a small decrease was observed in Ad-E1A-infected H357shPML-C1 (86% EGFR remaining). Cleaved PARP was only detected in H357pRSempty cells expressing E1A. These data provide strong evidence that the function of PML is important for the suppression of EGFR and induction of cell death by E1A in HNSCC cells.
Next the effect of TAp73β on the level of endogenous EGFR and apoptosis in H357shPML-C1 and H357pRSempty cells was investigated. Western blot analysis (Figure 5b) showed that the exogenous expression of TAp73β resulted in more efficient suppression of EGFR protein levels in H357pRSempty cells than in H357shPML-C1 (27.8% in H357pRSempty versus 48.1% in H357shPML-C1 of EGFR remaining) (Figures 5b and c). In addition, TAp73β expression strongly associated with apoptosis detected by cleaved PARP antibody in H357pRSempty cells. Expression of TAp73β induced PARP cleavage in H357shPML-C1 but to a lesser extent than control H357pRSempty cells. p53 expression also suppressed EGFR level in H357pRSempty cells but by a significantly lesser amount than that observed by TAp73β. Furthermore, suppression of EGFR by p53 was similar in H357pRSempty cells and H357shPML-C1 and therefore irrespective of PML knockdown.
To determine whether the suppression of EGFR protein levels was a cause or a consequence of apoptosis-induced protein degradation, the same blot was probed with an antibody against another high molecular weight protein, c-Cbl, an E3 ubiquitin ligase known to be important in the degradation of the EGFR (de Melker et al., 2001). Similar level of c-Cbl was detected in all samples irrespective of TAp73β or p53 expression (data not shown).
In addition, a mouse anti-Fas agonistic monoclonal antibody (clone CH-11) was used to induce apoptosis through the extrinsic, death receptor pathway. CH-11 efficiently induced apoptosis, detected by cleaved PARP, in H357 cells at 8 h after treatment. However, induction of PML and downregulation of EGFR protein levels were not observed even though a high level of cleaved PARP was detected (Figure 5d; data not shown). This result suggests that apoptosis induced in HNSCC cells through the Fas pathway does not cause EGFR suppression.
Furthermore, western blot analysis showed that the levels of EGFR detected in H357shPML-C1 were in general slightly higher (11–30% increase) than the levels detected in H357pRSempty cells, further providing evidence for the involvement of PML in the control of EGFR expression (Figure 5e).
To examine whether PML knockdown has any effect on the p73-mediated regulation of EGFR transcription, H357pRSempty cells and H357shPML-C1 were infected with Ad-GFP-TAp73β and EGFR mRNA levels were analysed by quantitative real-time RT–PCR. As shown in Figure 5f, expression of TAp73β resulted in a reduction in EGFR mRNA level in both H357pRSempty cells and H357shPML-C1. However, TAp73β-induced suppression of EGFR mRNA was significantly higher in H357pRSempty cells than in H357shPML-C1 (47.7% in H357pRSempty cells versus 74.7% in H357shPML-C1 of EGFR mRNA level, P<0.001). Collectively, these data suggest that TAp73β suppresses EGFR at both mRNA and protein levels and that PML is important for TAp73β-mediated suppression of EGFR transcription as depletion of PML hindered E1A- and TAp73β-induced EGFR suppression.
In this study, a clear correlation between EGFR overexpression and increased levels of the dominant-negative p73 isoform ΔNp73α was demonstrated in a panel of HNSCC cell lines, suggesting a link between p73 and EGFR dysregulation in these cancers. In addition, the EGFR-overexpressing HNSCC cell lines harbouring a defective p53 pathway were extremely sensitive to killing by TAp73. This data therefore identify TAp73 as an important anti-cancer therapeutic target for EGFR-overexpressing cancers with p53 mutation. These findings also support our previous observation showing the importance of an intact TAp73 function in the sensitivity of HNSCCs to chemotherapy and to the induction of apoptosis by viral proteins E1A and apoptin (Bergamaschi et al., 2003; Klanrit et al., 2008).
We have previously shown that the expression of human adenovirus gene, E1A, results in a significant suppression of EGFR mRNA and protein and the induction of cell death in several HNSCC cell lines (Flinterman et al., 2003, 2007). In addition, E1A strongly transactivates TAp73 expression resulting in the p53-independent induction of pro-apoptotic targets (Flinterman et al., 2005; Klanrit et al., 2008). Function of p73 was shown to be crucial for E1A-induced cell death as inhibition of p73 function by overexpression of ΔNp73 or using p73-functional deficient c-Abl knockout mouse embryonic fibroblasts blocked this activity (Klanrit et al., 2008). Importantly, E1A expression was shown to induce PML expression (Flinterman et al., 2003). PML consecutively regulates the expression of a number of important genes including p53, p63 and p73 (Guo et al., 2000; Bernassola et al., 2004, 2005), and has been shown to suppress EGFR expression (Vallian et al., 1998). Mice and cells lacking PML are resistant to a vast variety of apoptotic stimuli (reviewed in Bernardi et al., 2008). PML is important for the stabilization and hence increased activity of p73 (Bernassola et al., 2004). Furthermore, PML is the direct transcriptional target of p73/YAP and PML transcriptional activation by p73/YAP is under the negative control of Akt/PKB kinase (Lapi et al., 2008). These independent but complementary findings led us to speculate a link between E1A, TAp73 and PML in the regulation of EGFR expression in head and neck cancers. The data obtained here clearly demonstrated the efficient suppression of EGFR by TAp73 in head and neck cancers. Furthermore, the induction of PML in HNSCC cells was shown to be an important indicator of the sensitivity of these cells to killing by TAp73. The luciferase reporter assay showed that TAp73β and TAp73δ were the most efficient isoforms in suppressing the EGFR promoter. Previously, another p53 family member, TAp63γ, was shown to repress the activity of the EGFR promoter resulting in the downregulation of endogenous EGFR expression (Nishi et al., 2001). This effect is believed to be through the interaction of TAp63γ with Sp1 (Nishi et al., 2001). Interestingly, TAp73 isoforms, in particular TAp73β, have been shown to suppress the human telomerase reverse transcriptase promoter activity, through interaction of TAp73β with Sp1 (Racek et al., 2005). Therefore, the observed suppression of EGFR by TAp73β might be partly mediated through its interaction with Sp1. However, the precise nature of TAp73β-mediated EGFR suppression remains unclear and needs further investigation.
In this study, we have further confirmed that PML suppresses EGFR promoter activity. This is in agreement with a previous report showing that PML is a transcriptional repressor of EGFR through its association with Sp1, thus inhibiting Sp1-mediated transactivation of EGFR (Vallian et al., 1998). Using the GAL4-responsive promoter, another study has suggested the trans-repressing function of PML to be mediated through its interaction with histone deacetylases (Wu et al., 2001). PML3, a specific PML isoform, has recently been shown to interact with and recruit histone acetyl transferase, Tip60 to PODs. The physical interaction between PML3 and Tip60 protects Tip60 from Mdm2-mediated degradation, suggesting that PML3 competes with MDM2 for binding to Tip60 resulting in altered distribution, dynamics and function of Tip60 (Wu et al., 2009). Tip60 belongs to a multi-molecular complex involved in the cellular response to DNA damage. Tip60 interacts with Tip60 complex protein, p400 (EP400), that was discovered as an E1A-associated protein, to regulate the expression of both pro- and anti-apoptotic genes (Tyteca et al., 2006). We have recently shown that the p400 function is important for E1A-induced suppression of EGFR as p400 knockdown blocked this activity (Flinterman et al., 2007). These studies suggest a possible link between PML and p400/Tip60 in transcriptional modulation of EGFR induced by E1A and p73, which needs to be further investigated.
TAp73β expression in H357 and HN5 cells resulted in a strong induction of PML protein and changes in PML expression pattern from a typical pattern of several, small, round, discrete dots to a dense, patch-like pattern. These changes were accompanied by EGFR downregulation and PARP cleavage, providing evidence that TAp73β induces apoptosis in H357 and HN5 cells possibly by inducing PML and PML-mediated downregulation of EGFR and ultimately apoptosis. Although the changes in PML expression pattern were observed in H357 cells expressing either p53 or TAp73β and HN5 cells expressing TAp73β, we cannot specify which isoforms of PML were affected as the antibody used does not distinguish specific PML isoforms. Previously, p53 has been described as a transactivator of EGFR (Ludes-Meyers et al., 1996; Sheikh et al., 1997). In the present study the expression of p53 in H357 cells resulted in the induction of apoptosis as a consequence of PML induction and EGFR downregulation. This effect could be due to the direct induction of PML by p53 and subsequent repression of EGFR through PML. In support of this notion, the HN5 cells in which p53 failed to induce PML showed no evidence of EGFR downregulation and apoptosis. Alternatively, suppression of EGFR and induction of PML by p53 in H357 cells could be indirect and due to the ability of p53 to activate TAp73 (Chen et al., 2001).
In addition to our previous study demonstrating that exogenous expression of EGFR protected H1299 cells from E1A-induced suppression of EGFR and induction of apoptosis (Flinterman et al., 2003), the present study demonstrated that exogenous expression of EGFR from a heterologous LTR promoter protected H1299 cells from the effects of TAp73β expression including PARP cleavage, induction of PML and downregulation of EGFR, although these cells remained sensitive to killing by Ad-GFP-p53. Besides, p53-induced PML upregulation and EGFR downregulation were comparable between both H1299 cells transduced with the control vector and H1299 cells expressing exogenous EGFR.
Finally a direct function for PML in the regulation of EGFR by p73 and E1A was demonstrated using shRNA-mediated inhibition of PML in HNSCC cell lines, HN5 and H357. However, despite the incomplete knockdown of PML, these cells demonstrated clear resistance to downregulation of EGFR and apoptosis in response to both the expression of E1A or TAp73β. We have previously shown that E1A transcriptionally suppresses EGFR (Flinterman et al., 2007). In this study, TAp73β was also shown to inhibit EGFR promoter activity using a luciferase reporter assay and inhibition of PML by shRNA partially blocked transcriptional suppression of EGFR by TAp73β in H357 cells.
The data from this study show that TAp73β suppresses the expression of EGFR at both the transcriptional and the post-translational levels. However, it appears that the post-translational regulation, possibly protein degradation, is the major mechanism for p73 regulation of endogenous EGFR protein; only 50% suppression of EGFR mRNA by TAp73β in H357pRSempty was detected (Figure 5f). This is in agreement with the luciferase reporter assay that showed a moderate suppression of the EGFR promoter by TAp73β (Figures 2b and c) in contrast to a substantial reduction in the level of EGFR protein (Figures 2d and e). Further investigation of the role of TAp73β as well as E1A on EGFR protein stability is necessary to understand the actual mechanisms involved.
In summary, this study identifies PML as a pivotal mediator of EGFR downregulation and apoptosis in HNSCC cells induced by E1A and TAp73β. We propose a novel p53-independent regulatory pathway by which E1A, through the activation of TAp73 and subsequent induction of PML, represses EGFR expression and consequently induces cell death in the EGFR-overexpressing HNSCC cells (Figure 6). However, different components of the proposed pathway are likely to have multiple important functions in the normal regulation of cell cycle, signalling and apoptosis. These components are already identified as important regulators in different types of cancer. Therefore, the specific and complex inter-relationships shown in this study may well apply to some cell types and not others. Nevertheless, these findings provide new insight into how disruptions in the pathways regulated by the p53 family members could contribute to dysregulation of EGFR and malignant transformation of epithelial cells. Furthermore, this study identifies new targets that may be important to improve response to EGFR-targeted therapies by selecting the most suitable patients. Future studies should include retrospective screening of primary tumours for aberrations in EGFR, p73 and PML, and compare this to response to EGFR-targeted therapies.
Human HNSCC cell lines HN5, HN30, H103, H357, HSC-3 and its metastatic match HSC-3 M3, oesteosarcoma cell line Saos-2, neuroblastoma SH-SY5Y, non-small-cell lung cancer cell line H1299 and its clones, normal human fibroblasts 1BR3, the SV40 large T antigen transformed human fibroblasts (1BR3-LT), normal human embryonal lung fibroblasts (6689), normal human keratinocytes Km3 and human primary embryonal kidney 293A adenoviral packaging cells were cultured as described previously (Flinterman et al., 2003, 2005, 2007; Guelen et al., 2004). Cultures were incubated at 37 °C and 5% CO2.
To establish stably PML knockdown HN5 and H357 cells, 5 × 106 cells were transfected with 5 μg of either pRetroSuper expression plasmid (OligoEngine, Seattle, WA, USA) or pRetroSuper expression plasmid containing the shRNA against PMLexon2 in 100 μl OptiMEM serum-free medium (Invitrogen, Paisley, UK) by electroporation using a Gene Pulser II electroporator (Bio-Rad Laboratories Ltd., Hertfordshire, UK). The suspension of cells containing plasmid was transferred into 2 mm electroporation cuvettes (Bio-Rad Laboratories Ltd.). Cells were electroporated at 160 V using an exponentially decaying pulse. The cells were seeded into a 10 cm dish and incubated for 48 h at 37 °C and 5% CO2 after that the medium was changed and 2.5 μg/ml puromycin was added to select puromycin-resistant cells for 2 weeks. For pRetroSuper expression plasmid containing the shRNA against the exon 2 of PML mRNA, the targeting sequences are indicated in bold:
The pcDNA3-TAp73α, β, γ, δ and ΔNp73α, all are HA tagged, have been described previously (Sayan et al., 2005). Plasmid expressing the full-length PML complementary DNA (cDNA) has been described elsewhere (Vallian et al., 1998). The EGFR promoter luciferase pER1-Luc construct containing full-length EGFR promoter (pER-1) was obtained from Prof Alfred Johnson (Laboratory of Molecular Biology, NCI, NIH, Maryland, MD, USA). The following replication-incompetent adenoviruses were used: Ad-E1A (dl324) containing functional E1A but with complete deletion of E1B and E3, and Ad-Del (dl312) as a control containing a complete deletion of E1A and E3 (Yan et al., 1991). Ad-GFP-TAp73β expressing TAp73β fused to GFP was obtained from Prof Karen Vousden (Beatson Institute for Cancer Research, Glasgow, UK). Ad-GFP-p53 expressing p53 fused to GFP and Ad-GFP was obtained from Prof Bert Vogelstein (The Howard Hughes Medical Institute and Sidney Kimmel Comprehensive Cancer Center, Johns Hopkins University, Baltimore, MD, USA). Adenovirus amplification and purification was essentially performed as described previously (Flinterman et al., 2003). The mouse anti-Fas (CD95) monoclonal antibody clone CH-11 (MBL, Woods Hole, MA, USA) was used as a Fas agonist.
Cell survival was measured by the MTT assay as described previously (Flinterman et al., 2003).
Cells were seeded in Falcon eight-well culture slides (Becton Dickinson, Oxford, UK) and infected with adenoviruses the next day. PML expression was detected by rabbit anti-PML clone 3573 (gift from Dr Kun-Sang Chang, Department of Molecular Pathology, MD Anderson Cancer Center, Huston, TX, USA) and monoclonal anti-PML (PG-M3; Santa Cruz) antibodies as described previously (Flinterman et al., 2003).
Cells were seeded at 4 × 104 cells per well in 96-well plates. Cells were transfected on the next day with 40 ng per well of the EGFR luciferase reporter plasmid and 200 ng of inducer expression plasmid pre-incubated with 0.6 μl LipofectAMINE 2000 (Invitrogen) in 80 μl Dulbecco’s modified Eagle’s medium without fetal calf serum and antibiotics. For normalizing transfection efficiency, 13.2 ng of the pRL-null Renilla luciferase reporter (Promega, Southampton, UK) was transfected into each well. At 24 h after transfection, the medium was removed and the firefly and Renilla luciferase activities were sequentially assayed using Dual-Glo reagents (Promega) in a Wallac Trilux 1450 luminometer (PerkinElmer Life Sciences, Boston, MA, USA).
Western blotting was performed as described previously (Flinterman et al., 2003). The antibodies used were: rabbit anti-p73SAM (Sayan et al., 2005), mouse anti-EGFR clone F4 (gift from Prof William Gullick, Department of Biosciences, University of Kent at Canterbury, UK), rabbit anti-PML clone 3573, mouse anti-PML (PG-M3), mouse anti-p53 clone DO-7 (Novocastra Laboratories, Newcastle Upon Tyne, UK), mouse anti-E1A clone M58 (Pharmingen, BD Biosciences, Oxford, UK), mouse anti-tubulin (Sigma, Gillingham, UK), mouse anti-β-actin (Sigma), rabbit anti-PARP p85 fragment clone G734A (Promega), mouse anti-TAp73 and anti-p53 clone X77 (gift form Prof Thierry Soussi, Institut Curie and Université P & M Curie, Laboratoire de Génotoxicologie des Tumeurs, Paris, France), secondary anti-mouse (Sigma) and anti-rabbit antibodies linked to horseradish peroxidase (GE Healthcare Life Sciences, Little Chalfont, Buckinghamshire, UK).
Total RNA was isolated using the illustra RNAspin Mini Isolation kit (GE Healthcare Life Sciences, Buckinghamshire, UK). RNA (1 μg) was used for reverse transcription with the SuperScript II system (Invitrogen) using random hexamer primers to synthesize the first-strand cDNA. The primers used were as follows:
PCR reactions contained 12.5 μl of Platinum SYBR Green qPCR Supermix-UDG with Rox (Invitrogen), 10.5 μl DNase/RNase-free water, 0.5 μl of each forward and reverse primer (10 μm) and 1 μl of diluted cDNA template. Amplification was performed on an ABI Prism 7900HT Fast Real-Time PCR system (Applied Biosystems, Warrington, Cheshire, UK). All reactions were run in duplicate with a no-template control and -RT control was amplified to control for remaining DNA contamination.
To examine p73α (209 bp) and p73β (135 bp) isoforms in HNSCC panel, primers 5′-CATGGTCTCGGGGTCCCACT-3′ and 5′-CTGCTTCAGGTCCTGCAG-3′ were used and semi-quantitative RT–PCR was performed as described previously (Yokomizo et al., 1999). The products were run using gel electrophoresis and band intensities were quantified using ImageJ software (NIH, Bethesda, MD, USA).
For statistical analyses, the one-way analysis of variance was carried out. Statistically significant difference was defined as P<0.05.
We thank Prof Bert Vogelstein for providing Ad-GFP-p53 and Ad-GFP, Prof Karen Vousden for providing Ad-GFP-TAp73β, Prof Alfred Johnson for providing pER1-luc and Associate Prof Kun-Sang Chang for anti-PML clone 3573 antibody. We also thank Prof Alan Lehmann for 1BR3 and 1BR3-LT, Prof Hugh Paterson for 6689, Prof Barry Gusterson for HN5, Prof Kazuya Tominaga for HSC-3 and HSC-3 M3, Prof Fiona Watts for Km3, Prof Andrew Yeudall for HN30 and Prof Stephen Prime for H103 and H357 cell lines. We especially thank Prof Christoph Borner for the critical and helpful reading of this paper. This work was supported by grants from Cancer Research UK, the Rosetrees Trust and Biomedical Research Centre at Guy’s and St Thomas’ NHS Foundation Trust and King’s College London. Marcella Flinterman was supported by a grant from Cancer Research UK (C1116). Poramaporn Klanrit and Patrayu Taebunpakul were supported by the Royal Thai Government Scholarship.
Conflict of interest The authors declare no conflict of interest.
Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)