These authors contributed equally.
EGFR overexpression induces activation of telomerase via PI3K/AKT-mediated phosphorylation and transcriptional regulation through Hif1-alpha in a cellular model of oral–esophageal carcinogenesis
Article first published online: 12 DEC 2010
© 2010 Japanese Cancer Association
Volume 102, Issue 2, pages 351–360, February 2011
How to Cite
Heeg, S., Hirt, N., Queisser, A., Schmieg, H., Thaler, M., Kunert, H., Quante, M., Goessel, G., von Werder, A., Harder, J., Beijersbergen, R., Blum, H. E., Nakagawa, H. and Opitz, O. G. (2011), EGFR overexpression induces activation of telomerase via PI3K/AKT-mediated phosphorylation and transcriptional regulation through Hif1-alpha in a cellular model of oral–esophageal carcinogenesis. Cancer Science, 102: 351–360. doi: 10.1111/j.1349-7006.2010.01796.x
- Issue published online: 25 JAN 2011
- Article first published online: 12 DEC 2010
- Accepted manuscript online: 13 NOV 2010 08:43AM EST
- (Received May 24, 2010/Revised October 27, 2010/Accepted November 7, 2010/Accepted manuscript online November 13, 2010/Article first published online December 12, 2010)
Telomerase plays an important role during immortalization and malignant transformation as crucial steps in the development of human cancer. In a cellular model of oral–esophageal carcinogenesis, recapitulating the human disease, immortalization occurred independent of the activation of telomerase but through the recombination-based alternative lengthening of telomeres (ALT). In this stepwise model, additional overexpression of EGFR led to in vitro transformation and activation of telomerase with homogeneous telomere elongation in already immortalized oral squamous epithelial cells (OKF6-D1_dnp53). More interestingly, EGFR overexpression activated the PI3K/AKT pathway. This strongly suggested a role for telomerase in tumor progression in addition to just elongating telomeres and inferring an immortalized state. Therefore, we sought to identify the regulatory mechanisms involved in this activation of telomerase and in vitro transformation induced by EGFR. In the present study we demonstrate that telomerase expression and activity are induced through both direct phosphorylation of hTERT by phospho-AKT as well as PI3K-dependent transcriptional regulation involving Hif1-alpha as a key transcription factor. Furthermore, EGFR overexpression enhanced cell cycle progression and proliferation via phosphorylation and translocation of p21. Whereas immortalization was induced by ALT, in vitro transformation was associated with telomerase activation, supporting an additional role for telomerase in tumor progression besides elongating telomeres. (Cancer Sci 2011; 102: 351–360)
Immortalization and malignant transformation are important steps in the development of human cancer. They involve a complex accumulation of genetic and epigenetic events mainly in the proto-oncogenes and tumor-suppressor genes, as well as the ability to maintain telomeres. In a genetic model of oral–esophageal carcinogenesis we recapitulated the malignant transformation of primary human epithelial cells to invasive squamous cancer cells using a defined set of genetic alterations frequently found in the corresponding human cancer.(1) In this model, overexpression of cyclin D1 in combination with the inactivation of p53 induced immortalization of well-defined primary epithelial cells via an alternative lengthening of telomeres (ALT) mechanism in order to maintain telomere length.(2) Additional overexpression of the epidermal growth factor receptor (EGFR) triggered in vitro transformation and activation of telomerase and activated the PI3K/AKT signaling pathway. Subsequent overexpression of c-myc finally induced the fully malignant phenotype and led to invasive squamous cancer cells. Interestingly, in this stepwise model of human carcinogenesis in vitro transformation was strongly associated with activation of telomerase despite ample telomere length, elongated by ALT. Thus, cells were immortalized even before EGFR overexpression. This strongly suggested an additional role for telomerase in tumor progression besides just elongating telomeres. Such an additional role was also indicated by Stewart et al.,(3) demonstrating that activation of ALT substituted for telomerase in order to immortalize cells but failed to functionally substitute for telomerase in the process of malignant transformation. Consistent with previous in vitro transformation models,(4,5) human fibroblasts coexpressing the SV40 early region and oncogenic Ras as well as exhibiting an ALT phenotype were immortalized and permitted anchorage-independent growth. Nevertheless, these cells failed to form tumors in nude mice.(3) However, the additional expression of hTERT ultimately allowed these ALT-positive cells to form tumors in immunodeficient mice. Furthermore, the introduction of a mutant version of hTERT unable to elongate telomeres also induced malignant transformation. These and our observations together with other recent reports suggested that the activation of telomerase serves other important functions in tumor progression besides just maintaining telomere length.(3,6–9) In our in vitro model of human carcinogenesis, EGFR overexpression induced the activation of telomerase, leading to increased proliferation, an enhanced migratory potential and in vitro transformation. The aim of the present study was to elucidate how EGFR signals the activation of telomerase and how this leads to enhanced proliferation and in vitro transformation to ultimately further shed light on the role and regulation of telomerase in human carcinogenesis.
Materials and Methods
Cell culture and retroviral infection. Normal diploid human oral keratinocytes (OKF6) were established from a biopsy of normal floor of the mouth of clinically and genetically normal tissue. OKF6 cells have been cryopreserved within their first two serial passages in culture and characterized extensively.(1,2) They express keratinocyte markers as K5 and K14 and retain the ability to form a squamous epithelium in organotypic culture (data not shown). OKF6 and the lines generated thereof (OKF6-D1_dnp53, OKF6-D1_dnp53_EGFR) are grown in defined keratinocyte-SFM with defined growth supplement (Life Technologies Inc., Rockville, ML, USA) and final Ca2+-concentration of 0.4 mM. The medium was supplemented with antibiotics.
Determination of replicative life span. Serial cultures of the two different OKF6 lines (OKF6-D1_dnp53, OKF6-D1_dnp53_EGFR) studied were performed in 10-cm dishes by plating 105 cells, refeeding the cells every second day and subculturing every 4–5 days. The doubling number of each passage was calculated using the formula PD = (nf/n0)/log2, where n0 is the initial number of cells and nf is the final number of cells.
Western blot analysis. Lysates from exponentially growing cells were harvested in a buffer (50 mM HEPES [pH 7.4], 0.1% Nonidet P-40, and 250 mM NaCl) with 1 mM protease inhibitor PMSF (Sigma–Aldrich, St Louis, MO, USA) and Phosphatase Inhibitor Cocktail IV (Merck, Darmstadt, Germany) in a dilution of 1:50. A total of 7 μg of total protein was separated on 10% SDS-polyacrylamide gels and transferred to Immobilon membranes (Millipore Corp., Bedford, MA, USA). Blocking was performed in 5% milk, 10 mM Tris-HCl (pH 7.4), 150 mM NaCl and 0.2% Tween-20 overnight, followed by incubation with primary antibodies as indicated for 2 h. The secondary antibody was peroxidase-conjugated anti-mouse, anti-rabbit or anti-goat Ig at 1:5000 (Amersham Biosciences, Pittsburgh, PA, USA). Detection was by chemiluminescence (ECL plus; Amersham Biosciences). The primary antibodies used were polyclonal EGFR antibody 1:800 (sc-03; Santa Cruz Biotechnology, Santa Cruz, CA, USA), monoclonal Hif1-alpha antibody 1:800 (sc-13515; Santa Cruz Biotechnology), polyclonal AKT1/PKBα 1:800 (06-556; Upstate, Lake Placid, NY, USA) and monoclonal phospho-AKT 1:800 (Ser-473 and Thr-308; Cell Signaling Technology, Danvers, MA, USA) antibodies, polyclonal ERK1/2 1:800 (sc-154; Santa Cruz Biotechnology) and monoclonal phospho-ERK1/2 1:1000 (sc-7383; Santa Cruz Biotechnology) antibodies, polyclonal p21 1:800 (sc-756; Santa Cruz Biotechnology) and phospho-p21 1:800 (sc-20220-R; Santa Cruz Biotechnology) antibodies, as well as monoclonal hTERT antibody 1:1000 (NB 100–297; Novus Biologicals, Littleton, CO, USA). Quantification was done with AIDATM 3.20 image software (Raytest, Straubenhardt, Germany).
RT-PCR. mRNA of the different OKF6 cell types (OKF6-D1_dnp53, OKF6-D1_dnp53_ EGFR) was extracted with the RNeasy protocol (Qiagen, Hilden, Germany) as described previously(10) and reverse transcribed using the Superscript_RNaseH_Reverse Transcriptase Systems (Invitrogen, Carlsbad, CA, USA). For RT–PCR the cDNA concentration was adjusted at 100 ng/mL. The oligonucleotide primers hTRTF3 5′-AAG TTC CTG CAC TGG CTG AT-3′ and hTRTR7 5′-CAC GAC GTA GTC CAT GTT CA-3′ were used to amplify the N-terminal region of the telomerase gene. The samples were subjected to 30 cycles of PCR (94°C for 45 s, 60°C for 45 s and 72°C for 30 s). The oligonucleotide primers p21 sense 5′-CAG GGG ACA GCA GAG GAA GA-3′and p21 antisense 5′- TTA GGG CTT CCT CTT GGA GAA-3′ were constructed using Vector NTI software. The samples were subjected to 40 cycles of PCR (94°C for 10 s, 55°C for 10 s and 72°C for 10 s). β-tubulin served as the internal control. The PCR products were separated on an ethidium bromide-stained 2% agarose gel.
Co-immunoprecipitation. Protein of exponentially growing cells was extracted using 1× CHAPS lysis buffer (10 mM Tris, 1 mM MgCl2, 1 mM EGTA, 0.1 mM Benzamidin [Sigma, Deisenhofen, Germany] 5 mM β-Mercaptoethanol 0.5% CHAPS [Roche Diagnostics, Mannheim, Germany], 10% Glycerol). Total protein (300 μg) was immunoprecipitated with 1 μg of hTERT (sc-7214) or Hif1-alpha (sc-13515; both Santa Cruz Biotechnology) primary antibody, respectively, at 4°C overnight, followed by incubation with Protein A/Protein G Plus Agarose (Santa Cruz Biotechnology) for 2 h. Normal anti-goat (sc-2028) or anti-mouse IgG (sc-2025; both Santa Cruz Biotechnology) served as a negative control for precipitation. After intensive washing and centrifugation steps, the proteins were separated on 10% SDS-polyacrylamide gels and transferred to Immobilon membranes (Millipore, Billerica, MA, USA). Incubation with primary antibodies was performed as indicated (1:800). The secondary antibody was peroxidase-conjugated anti-rabbit immunoglobulin (1:5000; Amersham Biosciences). Detection was by chemiluminescence (ECL plus; Amersham Biosciences).
Telomeric repeat amplification protocol (TRAP) assays/telomeric length assays. Generated oral keratinocytes were assayed for telomerase activity by using the PCR-based TRAP assay(11) (TRAPeze; Millipore). Cellular extracts (50 ng) along with a heat-inactivated control were used for the TRAP assays. Telomere length was measured by hybridizing an α-32P-labeled telomeric 5′-(CCCTAA)3-3′ probe to 10 μg of HinfI- and RsaI-digested genomic DNA as previously described.(11)
Quantitative fluorescence in situ hybridization (Q-FISH). The different OKF6 cell types were harvested after 24 h of treatment with 0.1 μg/mL Colcemid (Invitrogen). After hypotonic swelling in 75 mM KCl for 30 min at 37°C, the cells were fixed and stored in 3:1 methanol/acetic acid. Before hybridization the cells were dropped on cover slides and dried for 5 min at 80°C, followed by overnight drying at room temperature. After washing with PBS the slides were fixed in 3.7% formamide, extensively washed with PBS and treated with 550 U/mL pepsin (Sigma-Aldrich) in pH2 water (15 min, 37°C). The formaldehyde fixation and washing steps were repeated and the slides were dehydrated in ethanol and finally air dried. A hybridization mixture (2 × 10 μL), containing 70% ultra pure deionized formamide, 0.3 μg/mL Cy-3-conjugated (C3TA2)3 peptide nucleic acid (PNA) probe (PBIO/Biosearch Technologies, Novato, CA, USA), 0.25% (wt/vol) NEN blocking reagent (DuPont, Wilmington, DW, USA) in 10 mM Tris (pH 7) was added to the slide. A cover slip was added followed by DNA denaturation (3 min at 80°C). After hybridization overnight at room temperature the slides were washed with 70% formamide, 10 mM Tris, 0.1% BSA (pH 7.0–7.5) and with 0.1 M Tris, 0.15 M NaCl (pH 7.0) containing 0.08% Tween-20. The slides were dehydrated with ethanol, air dried and covered by 2 × 10 μL antifade solution VECTASHIELD (Vector Laboratories Inc., Burlingame, CA, USA) containing 0.2 mg/mL of 49-6-diamidino-2-phenylindole (DAPI). Analysis was done using a camera system (Spot RT3; Diagnostic Instruments, Inc., Sterling Heights, MI, USA) on a fluorescence microscope (Leica, Solms, Germany).
Wild-type and mutant hTERT-promoter constructs. The hTERT-promoter reporter gene constructs were generated using a 1242 bp hTERT-promoter plasmid designated as full length, as previously described.(10) Sense primers were designed with an NheI restriction site and antisense primer with an NcoI restriction site. Deletion constructs were generated by PCR, accounting for potential transcription factor binding sites.(12–15) The NheI- and NcoI-digested PCR products were agarose gel purified and ligated into the promoterless luciferase reporter gene construct, pGL3 (Promega, Madison, WI, USA), positioning the translation start site (ATG) at +58, which is identical to the luciferase translation start. The pGL3-promoter plasmid was used to control for transfection efficiency. All plasmids were transformed in DH5α-cells, purified by a modified alkaline lysis method (Qiagen) and verified by DNA sequencing.
Mutant promoter constructs containing mutant nucleotides spanning region −185 to −182 bp (HIF1-alpha binding site) were generated by site-directed mutagenesis using overlap extension PCR as described by Ho et al.(16) and named “DC-700del.HIF1-alpha”. Complementary oligonucleotide primers were used to generate two DNA fragments having overlapping ends. These fragments were combined in a subsequent “fusion” reaction, in which the overlapping ends anneal, allowing the 3′ overlap of each strand to serve as a primer for the 3′ extension of the complementary strand. The resulting fusion product was amplified further by PCR. Specific alterations in the nucleotide sequence were introduced by incorporating nucleotide changes into overlapping oligonucleotide primers. The PCR primers used in site-directed mutagenesis of the HIF1α binding site contain a restriction site for NheI (primer a) and a restriction site for NcoI (primer d). hTERT-HIF1α primer a (sense) 5′-AGG CTA GCG CAA TGC GTC CTC GGG TTC GTC-3, hTERT-HIF1α primer b (antisense) 5′-CGG GTC CCC AGT CCC TCC GCT TAA TGG GAA-3′, hTERT-HIF1α primer c (sense) 5′-CAG GAC CGC GCT TCC CAT TAA GCG GAG GGA-3′, and hTERT-HIF1α primer d (antisense) 5′-GTC CAT GGC AGG ACG CAG CGC TGC CTG AAA-3′.
Transient transfections. Transient transfections of all hTERT-promoter constructs and the HIF1-alpha expression vectors (pcDNA-wtHif1-alpha and pcDNA-dnHif1-alpha; see Chen et al.(17)) were carried out using an improved lipofectamin method (Effectene; Qiagen). OKF6-D1_dnp53 and OKF6-D1_dnp53_EGFR were plated at a density of 1.35 × 105 cells/six-well and transfected 24 h later with 0.6 μg of the respective luciferase reporter plasmid. Defined keratinocyte-serum free medium (SFM) was replaced by keratinocyte-SFM (KSFM; Invitrogen) supplemented with 40 μg/mL bovine pituitary extract (Invitrogen), 1.0 ng/mL EGF (Invitrogen), 100 units/mL penicillin and 100 μg/mL streptomycin (Sigma–Aldrich) 3 h before transfection. The pGL3 plasmid containing the SV40 promoter and the empty pGL3 plasmid served as positive and negative control, respectively. Thereby, the SV40 promoter activity in a given transfection experiment can be used to control transfection efficiencies and uniformity. Cells were harvested 48 h after transfection. One hundred microliters of lysate with uniform protein concentrations was measured for 30 s with a Monolight luminometer (Analytical Luminescence Laboratory, San Diego, CA, USA). Each experiment was performed in triplicate and at least three sets of independent transfection experiments were performed. Values were then expressed as x-fold increase or decrease compared with the full-length promoter. Activities were expressed as the mean of at least three independent transfection experiments.
Immunofluorescence. OKF6-D1_dnp53 and OKF6-D1_dnp53_EGFR cells were grown in six-well chambers on cover slides for 24 h. The cells were fixed with 3.7% formaldehyde/PBS-solution for 10 min, permeabelized with 0.5% NP-40 in PBS for 10 min and blocked with PBS containing 0.2% cold water fish gelatine (Sigma–Aldrich) and 5 g/L BSA (PBG) for 1 h. The cellular localization of p21Cip1/WAF1 and phosphorylated p21Cip1/WAF1 was determined using primary monoclonal antibodies against p21Cip1/WAF1 (Becton Dickinson, Franklin Lakes, NJ, USA) and Phospho-p21Cip1/WAF1 (Santa Cruz Biotechnology). Cells were incubated with these antibodies diluted 1:100, 1:500 or 1:1000 in PBG overnight at 4°C. After extensive washing in PBS the samples were further incubated with Fluorescein (FITC)-conjugated AffiniPure donkey anti-rabbit IgG (JacksonImmunoResearch, West Grove, PA, USA) and Alexa Fluor 594 donkey anti-mouse IgG (Invitrogen) as secondary antibodies diluted 1:500 for 1 h at room temperature. VECTASHIELD Mounting Medium with DAPI (Vector Laboratories) was used for DNA-counterstaining to label cell nuclei. The fluorescent cells were visualized in a Leica confocal fluorescence microscope (Leica TCS SP2 AOBS).
EGFR overexpression leads to enhanced expression and activation of telomerase in ALT-positive immortalized oral keratinocytes. Human oral keratinocytes (OKF6) become senescent when grown for up to eight passages in vitro, but could be immortalized by exogenous telomerase overexpression. Interestingly, OKF6 could also be immortalized by overexpression of cyclin D1 and inactivation of wild-type p53, as shown previously(1,2) and now confirmed in an independent set of experiments. Of note, these immortalized oral epithelial cells (OKF6-D1_dnp53) did not activate telomerase but utilized an ALT mechanism to elongate their telomeres. Ectopically, overexpressed EGFR in these OKF6-D1_dnp53 cells (retrovirus-mediated gene transfer of the pFBEGFR-hyg vector) resulted in a significantly higher population-doubling (PD) rate in OKF6-D1_dnp53_EGFR cells compared with already immortalized OKF6-D1_dnp53 cells, whereby EGFR was transduced at different time points after cells were immortalized, leading to the same results. This PD rate could be further increased by the addition of exogenous EGF, as shown in Fig. 1, with representative cells transduced at passage 62, correlating with day 250. As we described previously(1) and now confirmed, EGFR overexpression alone induced a robust activation of telomerase in ALT-positive OKF6-D1_dnp53 cells using a TRAP assay (Fig. 2A). Finally, telomere length measurement by telomeric restriction fragment analysis (TRF) revealed a shift from heterogeneous telomere ends including very long telomeres of up to 40 kb pairs, consistent with ALT in OKF6-D1_dnp53 cells to a very homogeneous telomere length of around 20 kb in OKF6-D1_dnp53_EGFR cells (Fig. 2B). To further evaluate telomere dynamics on a chromosomal level in a single cell, both cell types were additionally analyzed by Q-FISH analysis. Corresponding to the results from the TRF analysis, telomere signals were highly heterogeneous varying from a very strong signal to undetectable telomere signals within the same metaphase spread in ALT-positive OKF6-D1_dnp53 cells correlating with telomere length. In contrast, OKF6-D1_dnp53_EGFR cells revealed homogenous telomere signals as a characteristic feature of telomerase activity (Fig. 2C).
The role of EGFR overexpression in this reactivation of telomerase became more obvious when exponentially growing OKF6-D1_dnp53_EGFR cells were treated with 50 μM tyrphostin AG 1478 (Calbiochem, Gibbstown, NJ, USA), an inhibitor of the EGF receptor tyrosine kinase (Fig. 2A). After treatment, OKF6-D1_dnp53_EGFR cells displayed decreased telomerase activity in a time-dependent fashion leading to a total loss of telomerase activity after 48 h. These findings correlated with the hTERT mRNA levels, which revealed a basal expression of hTERT in OKF6-D1_dnp53-cells that was increased in OKF6-D1_dnp53_EGFR cells, and abrogated when the cells were exposed to tyrphostin AG 1478 (Fig. 2D).
EGFR overexpression activates the PI3K/AKT signaling pathway and leads to phosphorylation of hTERT. Next we sought to identify the pathways induced by EGFR overexpression and involved in the strong activation of telomerase in OKF6-D1_dnp53_EGFR cells. The two major growth factor-stimulated signal transduction pathways used by EGFR are the MAPK/ERK and the PI3K/AKT pathway. Several lines of evidence already suggested that PI3K and its downstream target AKT are activated in oral–esophageal cancer.(18,19)
Western blot analysis revealed a similar level of AKT expression as well as ERK1/2 expression in OKF6-D1_dnp53 and OKF6-D1_dnp53_EGFR cells. Nevertheless, increased phosphorylation of AKT on Ser-473 was demonstrated in OKF6-D1_dnp53_EGFR cells only, confirming our previous results (Fig. 3A).(1) Tyrphostin AG 1478 abolished the activation, that is, phosphorylation of the PI3K /AKT pathway, supporting the role of the EGF receptor in AKT phosphorylation in this setting (Fig. 3A). In addition, AKT was also phosphorylated at Thr-308 (data not shown). In contrast, ERK1/2 levels as well as phosphorylation status remained unchanged in OKF6-D1_dnp53_EGFR cells (Fig. 3B).
Previous studies reported that the region surrounding Ser-824 of hTERT corresponds to a consensus sequence for phosphorylation, which could eventually result in enhanced telomerase activity.(20,21) Indeed, EGFR overexpression in oral keratinocytes led to a direct interaction between hTERT and phospho-AKT, suggesting AKT-dependent phosphorylation of hTERT, as we confirmed in immunoprecipitated (anti-hTERT) protein of OKF6-D1_dnp53_EGFR cells by IP western blotting with anti-phospho-AKT antibody (Fig. 3C). This effect can be specifically abrogated by the PI3K inhibitor LY294002 (Calbiochem). In contrast, no direct interaction between phospho-AKT and hTERT could be detected in OKF6-D1_dnp53 cells, indicating that EGFR overexpression specifically led to phosphorylation of hTERT through the PI3K /AKT pathway in oral keratinocytes.
EGFR overexpression differentially regulates telomerase promoter activity through a Hif1-alpha binding site. It is assumed that hTERT is mainly regulated on a transcriptional level and the transcriptional regulation of hTERT expression has been extensively analyzed in cancer cells(12–14,22) as well as normal and premalignant keratinocytes.(10) To further assess the role of the transcriptional regulation of hTERT with respect to activation of telomerase and the PI3K/AKT pathway by EGFR, we used the 1242 bp hTERT promoter and a set of 11 deletion constructs, as described previously.(10) The 1242 bp hTERT promoter designated as full length, subcloned in the promoterless pGL3 luciferase reporter plasmid, was transiently transfected into OKF6-D1_dnp53 and OKF6-D1_dnp53_EGFR cells, respectively. Since the full-length hTERT promoter proved to be sufficient to achieve gene expression in both cell types, this region was subjected to functional analysis through deletion constructs containing 1011, 905, 700, 357, 290, 247, 187, 161, 101, 48 and 16 bp of the flanking DNA sequence 5′ to the putative transcription start site in the respective cell types.
The analysis revealed no significant difference in promoter activity between both cell types with the deletion constructs representing the hTERT-promoter region upstream of 700 bp. With the −357 construct, hTERT promoter activity increased in OKF6-D1_dnp53 cells, consistent with the role of the known silencer region located between 635 and 438 bp. With the −290, −247 and the −187 constructs, as well as downstream of 101 bp, both cell types displayed roughly similar transcriptional activities. In contrast, with the construct −161, hTERT promoter activity was almost doubled in OKF6 D1_dnp53 cells compared with OKF6-D1_dnp53_EGFR cells, suggesting a positive cis-regulatory element between 187 and 161 bp in the EGFR-overexpressing cells (Fig. 4A). Indeed, computer analysis of the hTERT promoter revealed a Hif1-alpha binding site between 187 and 161 bp (5′-CCACGTGGCGGG-3′).
Hif1-alpha has been identified as a positive regulator of telomerase expression in different normal and tumor cell lines.(23–25) To elucidate the role of a potential Hif1-alpha binding site as a cis-acting element in OKF6-D1_dnp53_EGFR cells, we additionally performed site-directed mutagenesis(16) of the site between 185 and 182 bp in the hTERT core promoter. Luciferase reporter analysis revealed a more than 50% decrease of transcriptional activity with the mutated construct compared with the not-mutated counterpart (Fig. 4B). Additionally, when we co-transfected vectors encoding wild-type Hif1-alpha in OKF6-D1_dnp53_EGFR cells together with the promoter deletion construct −187 containing the intact Hif1-alpha binding site, we observed significantly higher luciferase activity compared with the pcDNA empty vector, as well as compared with the vector encoding dominant-negative Hif1-alpha (Fig. 4C). Previous studies demonstrated that PI3K/AKT signaling enhanced transcription of telomerase in normal and tumor cells.(26–28) To further prove the key role of the PI3K/AKT pathway in the transcriptional regulation of hTERT via Hif1-alpha in OKF6-D1_dnp53_EGFR cells we performed transfection experiments followed by inhibition of the PI3K/AKT pathway. OKF6-D1_dnp53_EGFR cells transfected with deletion construct −187 showed a >50% decrease of luciferase activity when LY294002 was added to the medium 12 h after transfection. In contrast, no effect on luciferase activity by LY294002 could be observed with deletion construct −161, which lacks the Hif1-alpha binding site (Fig. 4D). These findings correlated with hTERT mRNA levels, which decreased considerably when OKF6-D1_dnp53_EGFR cells were exposed to LY294002 (Fig. 4E).
Taken together, these results indicate an important role of the PI3K/AKT pathway and Hif1-alpha as a downstream target in the transcriptional regulation of hTERT activated through EGFR overexpression.
Hif1-alpha is phosphorylated and stabilized in EGFR-overexpressing cells. Phosphorylation has been identified as an important mechanism of translational modification and stabilization of Hif-1alpha.(29,30) Furthermore, Hif1-alpha and beta have been characterized as targets of PI3K/AKT signaling in EGFR and HER2-overexpressing cells.(31–33) We observed a direct interaction between phospho-AKT and Hif1-alpha performing co-immunoprecipitation. Immunoprecipitation of protein from OKF6-D1_dnp53_EGFR cells with anti-Hif1-alpha antibody and subsequent IP western blotting with anti-phospho-AKT antibody revealed a protein complex containing phosphorylated-AKT as well as Hif1-alpha. This effect was specifically abolished by the PI3K inhibitor LY294002 (Fig. 5A). Moreover, protein levels of Hif1-alpha were elevated in OKF6-D1_dnp53_EGFR cells. This overexpression could be reversed by inhibition of the EGF receptor (Fig. 5B), suggesting the EGFR-dependent phosphorylation and stabilization of Hif1-alpha by phospho-Akt.
Increased levels of phosporylated p21 and cytoplasmic translocation in OKF6-D1_dnp53_EGFR cells. A recent report demonstrated that members of the HER family of receptor tyrosine kinases, such as EGFR, can induce cytoplasmic localization of p21 through PI3K /AKT-dependent phosphorylation at threonine 145 to promote cell growth.(34) Since telomerase regulation seems to be mediated through the PI3K /AKT pathway and heavily depends on the cell cycle, we aimed to analyze the cell cycle regulator p21 as a potential player involved in the accelerated proliferation observed in OKF6-D1_dnp53_EGFR cells. RT-PCR specific for p21 revealed equal levels of p21 mRNA expression in both cell types (Fig. 6A). In western blot analysis, levels of unphosphorylated p21 were considerably higher in OKF6-D1_dnp53 cells (Fig. 6B). However, compared with OKF6-D1_dnp53 cells, EGFR-overexpressing cells revealed a higher level of phosphorylated p21 (Fig. 6C). Furthermore, in OKF6-D1_dnp53 cells p21 was predominantly located in the nucleus. In contrast, almost no nuclear signal of p21 could be detected in OKF6-D1_dnp53_EGFR cells. These cells showed an increased level of phosphorylated p21 mainly restricted to the cytosol (Fig. 6D). Taken together, these results suggest that p21 is post-transcriptionally modified in OKF6-D1_dnp53_EGFR cells, leading to cytoplasmic translocation of p21.
Immortalization and malignant transformation are important steps in the development of human cancer involving a complex accumulation and interplay of genetic and epigenetic events. Immortalization is thereby closely linked to the maintenance of telomeres. In the majority of human tumors telomere maintenance is assured by the activation of telomerase. However, accumulating evidence suggests that telomerase has activities beyond telomere maintenance. Previously we found that overexpression of cyclin D1 and loss of p53 in primary oral keratinocytes induced immortalization via an ALT mechanism.(2) Interestingly, additional overexpression of EGFR led to activation of telomerase and, of note, in vitro transformation. Moreover the PI3K/AKT pathway was activated. Subsequent c-myc overexpression finally induced the cancerous phenotype.(1) Consequently, generated cancer cells were telomerase positive as human esophageal cancer, while ALT is recruited in an early stage of human carcinogenesis.(1) In the present study we show that EGFR overexpression uses the PI3K/AKT pathway to activate telomerase through direct phosphorylation of hTERT by phospho-AKT, as well as transcriptional regulation by Hif1-alpha as the key transcription factor. Furthermore, EGFR-induced cell cycle progression and proliferation correlated with the phosphorylation and translocation of p21.
During the last decade, immortalization and malignant transformation in different cell types could be experimentally dissected by the serial introduction of a limited set of key genetic alterations. The most prominent combination used was the SV 40 early region, H-rasV12 and hTERT.(5,35–37) In this context, we were able to recapitulate a respective human cancer modeling oral–esophageal carcinogenesis without using the viral oncogenes other transformation models depended on. All of these studies provided compelling evidence that dysregulation of a limited set of pathways observed in human cancer should be sufficient to recapitulate human carcinogenesis and that further dissection of the signaling pathways, altered by the genes introduced, can identify key steps in a specific human cancer.(38)
In the present study we sought to identify how EGFR signaling is involved in these important steps of in vitro transformation, especially the strong activation of telomerase, that obviously is not needed for telomere maintenance in this setting. We were able to demonstrate that EGFR overexpression leads to phosporylation of hTERT through the PI3K /AKT signaling pathway, resulting in the activation of telomerase in OKF6-D1_dnp53_EGFR cells. This is consistent with previous studies suggesting that PI3K and its downstream target AKT are involved in esophageal cancer.(18,19) Regarding the key genetic alterations in human transformation models it was recently demonstrated that PI3K/AKT is also a critical target of SV 40 st antigen. Activation of PI3K signaling functionally mimics the expression of SV40 st and conferred anchorage-independent growth and tumorigenicity.(39) The first evidence that hTERT might be a target of phosphorylation came from experiments in breast cancer cell lines treated with protein-phosphatase 2A leading to the reduction of telomerase activity.(40) Studies in recent years have identified two phosphorylation sites of AKT within the hTERT protein and AKT-dependent phosphorylation, and subsequent activation of hTERT was demonstrated in the meantime.(20,21,41) Moreover, phosphorylation of hTERT by PKC isoenzymes has been indentified as an important mechanism of telomerase regulation in head and neck as well as nasopharyngeal cancer,(42,43) both of epithelial origin. A potential mechanism leading to enhanced telomerase activity through phosphorylation could be observed in stimulated CD4 cells, where phosphorylated hTERT is shuttled from the cytosol to the nucleus.(44)
Nevertheless, telomerase activity is mainly controlled on a transcriptional level leading to hTERT mRNA expression and consequently to telomerase activity.(45,46) The transcriptional regulation of hTERT expression has been extensively analyzed in cancer cells.(12–14,22) In addition, we dissected the hTERT promoter in genetically defined premalignant and malignant cells and characterized the differential transcriptional regulation of hTERT in the process of malignant transformation.(10) In the present study, we focused on the transcriptional regulation of hTERT activation following EGFR overexpression. Comparing promoter analysis data of untransformed EGFR negative cells to their EGFR-overexpressing in vitro-transformed counterparts, we could demonstrate that a Hif1-alpha binding site within the hTERT core promoter seems to play an important role in the transcriptional regulation of hTERT. These findings were supported by transfection experiments using Hif1-alpha mutant constructs as well as expression vectors. Furthermore, we were able to link PI3K/AKT signaling and transcriptional regulation of hTERT via Hif1-alpha by using specific inhibitors of PI3K. The role of PI3K/AKT signaling in the transcriptional regulation of telomerase has been demonstrated in several studies.(26–28) Hif1-alpha has been identified as a positive regulator of telomerase expression in different normal and tumor cell lines.(23–25) Moreover, Hif1-alpha has been characterized as a downstream target of EGFR and HER2 signaling via the PI3K/AKT pathway.(31–33) Linking activation of the PI3K/AKT pathway to a Hif1-alpha cis-acting element on the telomerase promoter, we additionally demonstrated a direct interaction of Hif1-alpha and phosphorylated AKT in EGFR-overexpressing cells. Hif1-alpha was previously described as a prominent phosphorylation target,(30,47,48) resulting in the stabilization of Hif-1alpha protein levels. This effect can be abrogated by inhibition of EGFR or its downstream signaling pathway PI3K/AKT, respectively, confirming an important role of EGFR overexpression and downstream PI3K/AKT signaling in the activation and stabilization of Hif1-alpha.
Interestingly, activation of telomerase led to in vitro transformation of previously immortalized cells, which already had maintained telomeres. We observed elevated levels of phosphorylated p21 in these EGFR-overexpressing cells corresponding with its cytoplasmic translocation. Translocation to the cytoplasm removes p21 from its nuclear function as a cell cycle inhibitor leading to increased proliferation. Furthermore, cytosolic p21 is exposed to a different range of potential binding partners.(49) One of the best characterized roles of cytoplasmic p21 is the inhibition of Fas-mediated apoptosis.(50) Asada et al.(51) proposed cytoplasmic translocation as a potential mechanism of p21 regulation. Following studies confirmed these findings, indicating the AKT-dependent phosphorylation of p21 and subsequent cytoplasmic translocation in cells overexpressing HER-2/neu, a sibling of EGFR in the family of Erb-B receptor tyrosine kinases.(34,52) Phosphorylation of p21 through the PI3K/AKT signaling pathway has been described and characterized in several studies.(53–55) Whether p21-shuttling is directly dependent on telomerase now needs to be further elucidated. Taken together, our findings indicate a substantial role of p21 together with telomerase in the enhanced proliferative and migratory capacity of OKF6-D1_dnp53_EGFR cells.
In summary, we propose that EGFR overexpression in oral squamous epithelial cells, immortalized via an ALT mechanism, leads to the robust activation of telomerase via PI3K/AKT signaling through both direct phosphorylation and transcriptional regulation of hTERT through Hif1-alpha. In addition, enhanced proliferation of these in vitro-transformed OKF6-D1_dnp53_EGFR cells is induced by phosphorylation of p21, resulting in its translocation from the nucleus. Overall, this suggests that telomerase induced by EGFR has an immortalization- and telomere elongation-independent role in in vitro transformation and oral–esophageal squamous carcinogenesis.
The authors have no conflict of interest.