|Home | About | Journals | Submit | Contact Us | Français|
The E6 and E7 oncoproteins of high-risk human papillomaviruses (HPVs) are together sufficient to cause cellular transformation. Nucleophosmin (NPM) was identified as a protein with increased levels in two-dimensional (2-D) gel analysis of human foreskin keratinocytes (HFKs) expressing E7 following methylcellulose-induced differentiation. Analysis of NPM expression in E7-expressing cells and E6- and E7-expressing cells in culture and in organotypic rafts confirmed the increased levels observed in 2-D gel analysis. The elevated expression of NPM was determined to be posttranscriptional and was attributed to increased v-akt murine thymoma viral oncogene (AKT) activity in the E6- and E7-expressing cells. Depletion of NPM caused a reduction in the replicative capacity of E7- and E6/E7-expressing HFKs and an increase in markers of differentiation. Also, the p53 and pRb tumor suppressor levels are increased with the knockdown of NPM in E6/E7-expressing cells, and, interestingly, p14ARF is relocalized from the nucleolus to the nucleoplasm and cytoplasm in these cells. The results show for the first time that NPM is required for the proliferation and inhibition of differentiation observed in HPV E6- and E7-expressing primary cells.
The E6 and E7 oncoproteins of human papillomavirus type 16 (HPV-16) have been shown to cause immortalization of primary human keratinocytes and are expressed in malignant cancers caused by HPV-16 infection (27, 28). E6 is best known for its ability to bind and degrade the tumor suppressor p53, whereas E7 can inactivate the pRb family of tumor suppressors (2, 3, 26). E6 is one of the earliest genes expressed during HPV infection and has been shown to bind sites at both the C terminus and the DNA binding domain of p53. Degradation is mediated by the ubiquitin ligase E6-associated protein (E6-AP/UBC3A), leading to degradation of p53 via the 26S proteasome (14, 34). Another mechanism by which E6 inhibits p53 activity is by binding to p300/CBP and inhibiting the coactivation of p53-dependent gene transcription (30).
E7 can bind to and inactivate the pRb family of tumor suppressors, Rb, p107, and p130 (5). These proteins play a major role in regulating the cell cycle, transcriptional repression, and tumor suppression (7, 11). E7 has the ability to override normal cell cycle activities by binding to the hypophosphorylated form of Rb, prematurely pushing cells into the S phase and resulting in disruption of differentiation. Recent data have indicated the role of E7 in pRb-independent mechanisms that target other cellular proteins and disrupt their normal function (1).
In an attempt to identify other significant targets of E7 we carried out a two-dimensional (2-D) gel analysis of proteins from E7-expressing primary human foreskin keratinocytes (HFKs) during methylcellulose-induced differentiation. Nucleophosmin (NPM) was identified as a protein showing increased levels compared to the vector control cells. NPM is a nucleolar phosphoprotein that is abundant in tumor and proliferating cells (9, 21). Although it is localized in the nucleoli, NPM has the ability to shuttle between the nucleus and cytoplasm and can bind and chaperone proteins to alter their cellular localization (4). Regarded as a proto-oncogene, NPM is overexpressed in a range of cancers and is used as a marker for colon, gastric, and ovarian cancers, with increased levels of NPM correlating with tumor progression (8). It is also the most frequently mutated gene in acute myeloid leukemia (AML), with approximately 35% of patients showing an abnormality in the gene (9). NPM functions through sustaining ribosome biogenesis, inhibiting apoptosis and disrupting differentiation, and upregulation of NPM in cells leads to an increase in proliferation (4). In this report, we provide the first evidence of a role for NPM in HPV-mediated proliferation and inhibition of differentiation. We show that NPM is upregulated by E7 at the protein level through the ability of E7 to deregulate v-akt murine thymoma viral oncogene (AKT) and that this upregulation is required for proliferation of cells and for the inhibition of differentiation.
The pBabe (puro), pBabe-E6stopE7 (E7), and pBabe E6/E7 retroviral constructs used were described previously (10). pSuper-retro constructs expressing short-hairpin RNAs (shRNA) against no known annotated gene (shScr) were cloned as previously described (31), as were the pSuper-retro constructs expressing shRNAs targeting Rb and p53(15). The following sequences were used for shRNAs targeting NPM: forward, 5′-CCA GTG GTC TTA AGG TTG AAG TGT GG-3′; reverse, 5′-TCC AGA TAT ACT TAA GAG TTT CAC ATC CTC CTC C-3′. Before transfection into ΦNYX-GP packaging cells, all retroviral plasmid constructs were sequenced. Small interfering RNAs (siRNAs) targeting AKT (SignalSilence 6211, 6510, and 6511) were purchased from Cell Signaling. siRNAs targeting NPM (sense, UGA UGA AAA UGA GCA CCA G) and a Scrambled control (ACG GUA ACA GUC ACU GAG C) were designed and purchased from Darmacon.
Primary human foreskin keratinocytes (HFKs) were isolated from neonatal foreskin and transduced with retrovirus produced using the ΦNYX-GP packaging line as previously described (10).
pBabe shScr, pBabe shRNA targeting NPM (shNPM), E6/E7 shScr, and E6/E7 shNPM HFK cell lines were induced to differentiate by using organotypic rafts as previously described (10). Raft cultures were harvested, fixed in 4% paraformaldehyde, embedded in paraffin, sectioned, and stained with hematoxylin and eosin (H&E). Bromodeoxyuridine (BrDU) (20 μmol/liter) was added to the raft culture 12 h before harvest to label DNA-synthesizing cells. Cell lines were also induced to differentiate by suspension in 1.6% methylcellulose (29).
Proteins were extracted from gel samples and digested into tryptic peptides as previously described (35). The tandem mass spectrometry (MS/MS) spectra were searched using SEQUEST7 and the Bioworks browser (both from Thermo Corporation, San Jose, CA) and the publicly available European Bioinformatics Institute nonredundant human.fasta sequence database (http://www.ebi.ac.uk/IPI/IPIhuman.html) to determine possible sequence correlations of known proteins.
Protein lysate concentrations were either 30 or 50 μg for all blots, as described previously (29). In this study, the following primary antibodies were used: mouse monoclonal anti-Rb and mouse monoclonal anti-p53 (BD PharMingen) (1:1.000); mouse monoclonal anti-B23 and rabbit polyclonal anti-C23 (1:1,000) and mouse monoclonal antiactin (1:20,000) (Santa Cruz Biotechnology); mouse monoclonal anti-p14ARF (Neomarkers) (1:500); and rabbit polyclonal anti-K1 (Covance) (1:5,000). Secondary antibodies used in this study were goat anti-mouse horseradish peroxidase (HRP) and goat anti-rabbit HRP (Santa Cruz Biotechnology) (1:2,000). Luminescence was detected by either Perkin-Elmer or Pierce enhanced chemiluminescence (ECL), and the signal was detected using an Alpha Innotech FluorChem SP imaging system.
RNA was extracted with a High Pure RNA isolation kit (Roche), according to the manufacturer's instructions. FastStart SYBR green Master (Roche) was used according to the manufacturer's instructions to amplify PCR products, and fluorescence was monitored using a DNA engine Peltier thermal cycler (Bio-Rad) equipped with a Chromo4 real-time PCR detection system (Bio-Rad). cDNA samples were diluted 1:10 and quantified by amplification using a series of dilutions of control cDNA. The following cycling conditions were used: initial denaturation at 95°C for 10 min, followed by 40 cycles of 95°C for 15 s, 58°C for 15 s, and 60°C for 60 s. Expression levels were assessed in triplicate and normalized to ribosomal large protein P0 (RPLP0) control levels. Graphs produced represent the combined results of three independent replicate experiments.
pBabe- and E6/E7-expressing HFKs were pulse-labeled with 110 μCi/ml of EasyTag EXPRESS35S protein-labeling mix-[35S]methionine-cysteine-2 mCi (74 MBq) stabilized aqueous solution (catalog no. NEG772002MC; Perkin Elmer). After 3 h, the cells were washed and labeled and were media chased with fresh unlabeled media. Cells were harvested at indicated time points thereafter, and immunoprecipitation (IP) and Western blotting were performed. Samples were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) followed by staining in colloidal Coomassie blue (Invitrogen), and gels were dried under vacuum conditions (Bio-Rad Geldryer). Dried gels were exposed to a Phosphoimager screen and read using a Fuji FLA-7000 analyzer. Bands were normalized to protein levels in the immunoprecipitate.
Paraffin-embedded organotypic raft sections were deparaffinized with 2 washes in xylene and rehydrated with 4 washes in step-down concentrations of ethanol. Antigen retrieval with citrate buffer (Dako) was performed for 20 min in a steamer. The sections were then washed three times in phosphate-buffered saline (PBS), and the primary antibody was applied. The following primary antibodies (diluted in 10% fetal bovine serum) were used for this study: rabbit polyclonal anti-keratin 1 (Covance) (1:4,000; no retrieval); mouse monoclonal anti-B23 and rabbit polyclonal anti-C23 (Santa Cruz Biotechnology) (1:200); mouse monoclonal anti-BrDU (BD PharMingen) (for antigen retrieval; 1:200); rabbit polyclonal anti-Ki67 (LabVision) (1:200); mouse monoclonal anti p14ARF (Neomarkers) (1:100); mouse monoclonal anti-p14ARF (Sigma Aldrich) (1:100); and goat anti-mouse and anti-rabbit secondaries conjugated to Alexafluor (488 nm or 594 nm) (Molecular Probes) (1:400). Cervical intraepithelial neoplasia grade III (CIN3) lesions were described and processed as previously described (25). Images were taken using an Olympus BH-2 microscope and an Olympus D25 camera and Cell B software (Olympus). For raft sections, Ki67-positive cells were counted for a minimum of 15 fields of view (1,000 μM). For coverslips, a minimum of 15 fields of view and 400 cells were counted for each slide and expressed as percentages of control numbers. Immunofluorescence images were captured using a Leica AF6000 inverted microscope and Leica AF imaging software. Exposure times and deconvolution settings were kept constant within each experiment.
In an attempt to elucidate data for novel proteins that are upregulated in the presence of the HPV-16 E7 oncogene during differentiation, a proteomics approach was taken. 2-D gel analysis of HFK lysates during 12 h of methylcellulose-induced differentiation revealed an increase in the levels of various proteins in cells expressing the E7 oncogene compared to pBabe control cell results (Fig. 1a and b). Mass spectrometric analysis of the proteins revealed NPM to be one of the proteins whose level appears to be elevated in differentiating E7 HFKs. To confirm these findings, HFKs were transduced with a retrovirus expressing either pBabe control vector or HPV-16 E7 and cells were resuspended in methylcellulose for 12 h to induce differentiation. Western blotting confirmed an increase in NPM in E7-expressing differentiating cells (Fig. (Fig.1c)1c) and proliferating cells (Fig. (Fig.1d)1d) compared to pBabe control cells. Next, we wanted to determine whether NPM levels are also elevated in E7 cells in a more biologically relevant model system for keratinocyte differentiation, namely, an organotypic raft culture system. Sections of organotypic rafts generated from E7-expressing HFKs showed that NPM levels were increased compared to the pBabe control raft levels (Fig. (Fig.1e).1e). We also examined NPM levels in E6/E7-expressing cells, as these proteins are expressed in the context of HPV-16 infection. Western blotting of proteins extracted from organotypic rafts and immunofluorescence staining of sections from organotypic rafts revealed a substantial increase in NPM levels in cells expressing E7 alone and in E6/E7-expressing cells compared to pBabe control cells (Fig. 1d and e). To determine whether HPV-16-positive cervical lesions had increased NPM levels, we used immunohistochemical staining of sections from matched normal and cervical intraepithelial neoplasia grade III (CIN3) lesions. Investigation of the CIN3 lesions, all of which were positive for HPV-16, revealed increased levels of NPM compared to control epithelia within the same biopsy specimen (Fig. (Fig.1f1f).
To determine whether the increase in NPM protein levels observed in E6/E7-expressing cells and rafts was due to increased transcription, RNA from three independent sets of pBabe- and E6/E7-expressing HFKs was extracted for real-time PCR. NPM levels were normalized to an RPLP0 control, and values from the three individual experiments were averaged. Results showed that there was no significant increase in NPM RNA levels in E6/E7-expressing cells compared to pBabe control cells, suggesting that the increase in NPM levels was not due to increased transcription (Fig. (Fig.2a2a).
Having determined that the increase in NPM observed in E6/E7-expressing cells was not transcriptional, we next wanted to identify a possible mechanism to explain the increased protein levels of NPM in these cells. Previous data from our laboratory show that E7 upregulates AKT activity through deregulation of pRb protein (25). A recent report (20) has shown that AKT interacts with NPM and protects it from degradation. Therefore, either AKT was depleted or the activity was inhibited in pBabe- and E6/E7-expressing HFKs by small interfering RNAs (siRNA), an AKT inhibitor (AIV), and a PI3-kinase inhibitor (PI103). Reduced levels of AKT protein correlated with reduced levels of NPM in both pBabe- and E6/E7-expressing HFKs (Fig. (Fig.2b).2b). In addition, the treatment of cells with AKT or PI-3 inhibitors resulted in a reduction of AKT activity, as measured by a reduction in glycogen synthase kinase-3β phosphorylation (pGSK-3β), and a corresponding decrease in NPM protein levels (Fig. 2c and d). The half-life of NPM in control and E6/E7-expressing cells was measured in 35S-labeled cells and was found to be 24 h in control cells, but over the same period the levels in E6/E7-expressing cells did not change (Fig. (Fig.3a).3a). In line with a role for AKT in NPM stability, when AKT was depleted and stability determined at 16 h after the 35S labeled amino acids were washed out, the level of NPM in depleted E6/E7-expressing cells dropped to control cell levels. However, no change in NPM localization was observed when AKT was depleted using siRNA (Fig. (Fig.3b).3b). Indeed, NPM was localized to the nucleolus in all cell types examined, including normal HFKs and pBabe- and E6/E7-expressing cells (data not shown).
Cells expressing E6/E7 have the ability to override the normal process of cell cycle control and differentiation. To determine the effect of knockdown of NPM on proliferation and differentiation, E6/E7-expressing HFKs were infected with retrovirus expressing scrambled short-hairpin RNAs (shRNA) expressing either a scrambled shRNA (shScr) or an shRNA targeting NPM (shNPM) and stable cell lines were generated. Alternatively, NPM was depleted by RNA interference (RNAi) molecules (siScr and siNPM) and proliferation was investigated in short-term assays. The cell lines were tested for the level of proliferation as monolayers and in organotypic rafts by the use of BrDU incorporation and Ki67 staining, respectively, while the RNAi-treated cells were studied only as monolayer cultures. BrDU- and Ki67-positive cells were counted and quantified against DAPI (4′,6′diamidino-2-phenylindole)-positive cells, in approximately 15 different fields of view. A 40% decrease in proliferation of cells in a monolayer (Fig. 4a and b) and a 50% reduction in proliferation of E6/E7-expressing HFKs in raft cultures compared to scrambled E6/E7-expressing cell results (Fig. 4c and d) were observed. Similar reductions in proliferation were also observed in pBabe cells when NPM was depleted (Fig. (Fig.4e).4e). Taken together, these results suggest that E6/E7-expressing cells with reduced NPM levels are less proliferative than pBabe control cells.
To determine the effects of knockdown of NPM on differentiation, control or NPM-depleted HFKs were differentiated by two different methods. First, the cells were grown to confluence and then treated with 1.5 mM CaCl2 to induce differentiation and harvested at various times thereafter. Alternatively, organotypic rafts were produced with the depleted cells. As expected, both Western blotting of proteins from CaCl2-treated cells and immunofluorescence analysis of sections from organotypic rafts showed low levels of K1 expression in the E6/E7-expressing scrambled control cells (Fig. 5a and b). Interestingly, E6/E7-expressing cells with NPM knockdown showed increased levels of K1 in both Western blot analysis of CaCl2-treated cells and immunoflorescence staining of raft sections (Fig. 5a and b). These data show that when the elevated levels of NPM in E6/E7-expressing cells are reduced, there is a decrease in proliferation and cells start to differentiate. Therefore, NPM contributes to the inhibition of differentiation observed in E6/E7-expressing cells.
To address the mechanism of decreased proliferation and increased K1 expression in E6/E7-expressing shNPM cells, we analyzed the expression of p53 and pRb in these cells. In E6/E7-expressing HFKs, there were normally low levels of both p53 and pRb; this was confirmed by Western blotting (Fig. (Fig.5c).5c). However, there was a modest but consistent increase of both pRb (approximately 28% increase) and p53 (approximately 24% increase) levels in NPM-depleted E6/E7-expressing cells (Fig. (Fig.5c5c).
NPM is a shuttling protein that binds to and influences the localization of a number of proteins, including p14ARF (ARF). Since ARF is regulated in part by Rb through E2F and since it stabilizes p53, we investigated the levels of ARF in E6/E7-expressing cells compared to control cells. ARF levels were significantly higher in E6E7 cells (Fig. (Fig.6a).6a). Localization of ARF is important for biology, so both E6/E7-expressing cells with or without NPM depletion and control pBabe cells were analyzed for subcellular localization of ARF. It should be noted that in control and E6/E7-expressing cells, NPM in HFKs was localized to the nucleolus and colocalized with the nucleolar marker C23 both in cycling (Fig. (Fig.6b)6b) and during differentiation (data not shown). In pBabe-expressing cells, ARF was localized to the cytoplasm-nucleoplasm (Fig. (Fig.6b),6b), whereas in E6/E7-expressing shRNA control cells, ARF colocalized with NPM in the nucleolus (Fig. (Fig.6b).6b). In E6/E7-expressing cells with depleted NPM, ARF was relocalized to the nucleoplasm-cytoplasm (Fig. (Fig.6b).6b). This suggests that in E6/E7-expressing shRNA control cells, NPM sequesters ARF to the nucleolus and consequently reduces the stabilizing effect of ARF on p53 in these cells. However, upon depletion of NPM in E6/E7-expressing cells, ARF relocalizes from the nucleolus to nucleoplasm, which may result in the stabilization of p53 observed in E6/E7-expressing HFKs with NPM knockdown (Fig. (Fig.5b).5b). This relocalization of ARF from the nucleolus to nucleoplasm was also observed in E6/E7-expressing cells when AKT activity was inhibited using the AKT inhibitor AIV (Fig. (Fig.7a),7a), suggesting that AKT activity and the ability to stabilize NPM are important for E6/E7 to relocalize ARF to the nucleolus.
To determine whether ARF localization to nucleoli in E6/E7-expressing cells was due to disruption of the p53 and pRb functions, we next generated HFK cells with knockdown of p53 or of pRb or of both p53 and pRb. NPM levels increased only when both p53 and pRb were depleted (Fig. (Fig.7b),7b), while in cells with only p53 or pRb depleted, NPM levels are similar to control cell levels (data not shown). Interestingly, when both pRb and p53 were depleted, ARF was localized to the nucleoli, as was the case in E6/E7-expressing cells (Fig. (Fig.7c).7c). NPM remains in the nucleoli in these cells (data not shown). These results suggest that abrogation of the activity of both p53 and pRb is required for the nucleolar localization of ARF observed in E6/E7-expressing cells.
The intricate balance between cell proliferation and differentiation is crucial for maintenance of homeostasis and normal development within the cell, and disruption of either of these processes may result in oncogenesis. We and others have previously reported on the ability of the E6 and E7 oncoproteins to disrupt the normal process of differentiation of HFKs by targeting key tumor suppressors such as p53 (28) and pRb (17, 37), resulting in increased levels of cell survival proteins such as AKT (25) and disruption of the cell cycle (24, 31). In this study, we investigated the possible role of the nucleolar phosphoprotein nucleophosmin (NPM) in the differentiation process of HFKs. Frequently overexpressed in tumors and highly proliferating cells, NPM has previously been shown to reduce the susceptibility of cells to the onset of differentiation and apoptosis (13, 33). NPM has also been reported to play a crucial role in sustaining ribosome biogenesis in cancer cells (12) and is now regarded as important in the development of various cancers (9, 13, 22), with the first reported NPM inhibitor developed as an anticancer agent in 2008 (32). NPM was identified as one of the proteins that were also upregulated in a proteomic screening undertaken to characterize proteins altered in expression in differentiating cells expressing E7 and subsequently shown to be upregulated when both E6 and E7 are present. The function of E6 targeting p53 for degradation, combined with the ability of E7 to deregulate pRb and family members as well as AKT, has been extensively characterized and is known to be required for inhibition of differentiation and promotion of proliferation. This report describes another pathway involving NPM that is utilized by cells with E6/E7 to maintain proliferative capacity and to potentially circumvent the activity of p53.
Our data show that the upregulation of NPM observed in E6/E7 was not due to increased transcription. Recent work (20) has shown that AKT binds to NPM and prevents its degradation. Previous work had shown that E7 can upregulate AKT through disruption of the pRb and family member functions (25). Therefore, it was logical to test whether this upregulation of AKT was responsible in part for the elevated levels of NPM detected in E6/E7-expressing cells. Indeed, when AKT was depleted in E6/E7-expressing HFKs by siRNA or AKT activity was inhibited by either PI-3 kinase or AKT inhibitors, there was a marked decrease in NPM levels. Depletion of pRb alone did not cause an increase in NPM levels, and so it would appear that either another function of E7 is involved or the other pRb family members, p130 and p107, have a role to play.
The involvement of NPM in the increased proliferation and disruption of differentiation in different types of cells has been previously reported (9, 33). E6/E7-expressing HFKs can override the normal process of differentiation, and in this present report we provide the first evidence that increased levels of NPM in HFKs are important for inhibition of differentiation and proliferation. Stable or transitory knockdown of NPM by shRNA or siRNA molecules, respectively, in E6/E7-expressing HFKs led to a decrease in proliferation and an increase in the levels of the differentiation marker K1. The data demonstrate a clear role for NPM in suppression of differentiation. NPM is a nucleolar phosphoprotein that shuttles between the nucleus and the cytoplasm and takes part in various cellular processes. It has several interacting partners, some of which can be sequestered to the nucleolus and rendered inactive (22). One such target is the tumor suppressor ARF, which has multiple functions that are dependent on localization within the cell (16). ARF has been shown to bind to NPM through the same domain that mediates Mdm2 binding and nucleolar localization and as a consequence inhibits the ability of ARF to modulate p53- and pRb-associated growth arrest functions (18, 38). Also, NPM has been shown to directly impact levels of p53 (19) as well as interact with pRb (23, 36). Here we show that knockdown of NPM in E6/E7-expressing HFKs results in increased levels of p53 and pRb. In support of our results, a recently developed inhibitor of NPM induced apoptosis, upregulated p53, and caused reduced proliferation rates in a number of cell types (32).
Interestingly, ARF localization is altered in HFKs, where, upon knockdown of NPM in E6/E7-expressing cells, ARF is localized in the nucleoplasm and cytoplasm as opposed to being sequestered in the nucleolus. The underlying mechanism that allows ARF to be sequestered to the nucleolus by NPM in E6/E7-expressing cells is the disruption of both p53 and pRb functions, since knockdown of both of these proteins is required for the localization of ARF to the nucleoplasm-cytoplasm. However, the increased levels of NPM in E7 or E6/E7-expressing cells do not solely account for the ability of NPM to sequester ARF to the nucleolus, since E7-expressing cells with increased NPM levels do not sequester ARF to the nucleolus. The fact that E7 does not sequester ARF to the nucleolus may account for the observation of an increase in p53 levels in E7-expressing cells (6). Therefore, while E7 may upregulate NPM levels, E6 must carry out another activity that helps sequester ARF to the nucleolus. In conclusion, our results provide the first evidence of a role for NPM in HPV-mediated carcinogenesis and highlight the potential therapeutic use of an NPM inhibitor in cervical cancer.
This work was supported by grants from the NIH (NIDCR DE15935) and Wellcome Trust (WT082840AIA).
Published ahead of print on 17 March 2010.