| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Cell and Tumor Biology |
1 McArdle Laboratory for Cancer Research, University of Wisconsin, Madison, Wisconsin and 2 Section of Cancer Genomics, Genetics Branch, Center for Cancer Research, National Cancer Institute, NIH, Bethesda, Maryland
Requests for reprints: Paul F. Lambert, McArdle Laboratory for Cancer Research, University of Wisconsin, 1400 University Avenue, Madison, WI 53706. Phone: 608-262-8533; Fax: 608-262-2824; E-mail: lambert{at}oncology.wisc.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
146-151 mice, which express a mutant form of E6 that cannot bind a family of cellular proteins known as PDZ domain proteins but retains the ability to inactivate p53. In skin carcinogenesis experiments, the K14E6146-151 transgene failed to contribute to the promotion stage of skin carcinogenesis but retained the ability to contribute to the progression stage. Cytogenetic analysis indicated that, although gains of chromosome 6 are consistently seen in tumors arising on K14E6WT mice, they are infrequently seen in tumors arising on K14E6146-151 mice. This observation supports the premise that the nature of cancer development in these two mouse strains is distinct. Based on these studies, we conclude that E6 contributes to cancer through its disruption of multiple cellular pathways, one of which is mediated through its interaction with PDZ domain partners and the other through E6's inactivation of p53. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
High-risk HPV E6 proteins associate with several cellular partners. Among these is p53, which E6 binds and inactivates by targeting it for proteosome-mediated degradation (6). It is clear, however, that some E6-induced phenotypes in vivo and in tissue culture are independent of p53 inactivation. For example, p53-null mice do not display epithelial hyperplasia or suprabasal DNA synthesis, two phenotypes observed in K14E6WT mice (3). Additionally, some mutant E6 proteins that are unable to inactivate p53 retain the ability to transform cells (7–9) or induce phenotypes in vivo (10). Conversely, some mutant E6 proteins that retain the ability to target p53 for degradation are unable to induce transformation (8, 11). These observations have led investigators to question what other E6-interacting proteins might contribute to these and other phenotypes induced by E6. A particularly intriguing group of proteins is the PDZ domain proteins. High-risk but not low-risk HPV E6 proteins associate with PDZ domain proteins. Two of E6's PDZ domain partners, Dlg and Scribble, have been defined as tumor suppressors in Drosophila. Drosophila carrying mutant alleles for either of these proteins display epithelial hyperplasia and dysplasia, two traits observed in K14E6WT mice (12, 13). To date, E6 has been found to associate with six PDZ domain proteins: Magi1, Magi2, Magi3, MUPP1, Dlg, and Scribble (9, 14–18). PDZ domains are 80– to 90–amino acid protein-protein interaction domains that normally associate with a 5–amino acid sequence that is usually found at the COOH terminus of the binding partner (19). PDZ domain proteins are thought to act as scaffolding proteins, organizing multiprotein cellular signaling complexes (for review, see ref. 20).
To study the role of E6's interactions with PDZ domain proteins in vivo, we previously generated transgenic mice that express a mutant E6 protein that is unable to associate with PDZ domain proteins but retains the ability to inactivate p53. Mice treated with 4 Gy of ionizing irradiation were analyzed for their ability to synthesize DNA. In response to the DNA damage induced by ionizing irradiation, nontransgenic mice stopped synthesizing DNA. In contrast, both K14E6WT and K14E6146-151 mice continued to synthesize DNA, indicating that both wild-type and mutant E6 protein inactivate p53 (21, 22). Studies with these mice indicated that E6's interactions with PDZ domain proteins correlate with E6's induction of epithelial hyperplasia and suprabasal DNA synthesis (21). The focus of this study was to determine whether E6's interactions with PDZ domain proteins contribute to other E6-induced phenotypes, specifically tumorigenesis. E6 contributes to two distinct stages in skin carcinogenesis: promotion, a step involved in the formation of benign tumors, and progression, the malignant conversion of benign tumors to frank cancer (23). Here, we show that distinct biochemical activities of E6 contribute to these two stages in tumor development. Specifically, E6's interaction with PDZ domain proteins correlates with its contribution to the promotion stage of carcinogenesis but does not correlate with E6's contribution to progression.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
146-151 mice has been described previously (3, 21). At weaning, tail samples were collected and DNA was isolated for genotyping. The mice were genotyped by PCR using the following primers: 709-1 (5'-GGCGGATCCTTTTATGCACCAAAAGAGAACTG-3') and 709-4 (5'-CCCGGATCCTACCTGCAGGATCAGCCATG-3'). The PCR conditions are (a) 96°C for 3 minutes, (b) 94°C for 1 minute, (c) 60°C for 1 minute, (d) 72°C for 1 minute, (e) steps b to d repeated 34 times, and (f) 72°C for 7 minutes. Mice were housed in the Association for Assessment of Laboratory Animal Care–approved McArdle Laboratory Animal Care Unit of the University of Wisconsin Medical School. All procedures were carried out according to an animal protocol approved by the University of Wisconsin Medical School Institutional Animal Care and Use Committee. Immunohistochemical analysis for bromodeoxyuridine, keratin 14, keratin 10, and Scribble. At the time of euthanasia, ear and torso skin were harvested, fixed overnight in 10% buffered formalin, paraffin embedded, and cut into 5 µm histologic sections. Sections were deparaffinized in xylenes and rehydrated through a graded series of ethanol/water washes (twice for 2 minutes in xylene, 1 minute in 100% ethanol, 1 minute in 95% ethanol, 1 minute in 80% ethanol, 1 minute in 70% ethanol, and 1 minute in 50% ethanol). For bromodeoxyuridine (BrdUrd) staining, 9-day-old mice were injected with BrdUrd 1 hour before euthanasia. Following deparaffinization and rehydration, histologic sections were washed in PBS, endogenous peroxidase activity was quenched by a 10-minute incubation in 0.3% H2O2 in methanol, and the nuclear antigen was unmasked by a 20-minute incubation in 2 N HCl. The slides were then boiled in 10 mmol/L citrate buffer (pH 6.0) for 20 minutes to achieve further unmasking. Following a wash in PBS, the slides were blocked in 10% horse serum/PBS for 30 minutes and then incubated with anti-BrdUrd (Calbiochem, Darmstadt, Germany) antibody diluted 1:40 in block solution for 3 hours at room temperature. The biotinylated secondary antibody and streptavidin-peroxidase conjugate were added according to the manufacturer's instructions (Vectastain ABC kit, Vector Laboratories, Burlingame, CA). The antibody-peroxidase complex was detected by incubating the sections with 3,3'-diaminobenzidine solution (Vector Laboratories) for 1 to 2 minutes. The reaction then was quenched in water. The slides were then counterstained with hematoxylin, dehydrated through a series of graded ethanol/water washes and xylenes, and coverslipped. For keratin 14 (K14) and keratin 10 (K10) immunofluorescence, the deparaffinized, rehydrated histologic sections were washed in PBS and incubated in anti-K14 antibody (1:1,000; Covance, Richmond, CA) overnight at room temperature in 10% horse serum/3% bovine serum albumin (BSA)/PBS. After washing in PBS, Alexa 594 anti-rabbit secondary antibody (1:200; Molecular Probes, Eugene, OR) in 3% BSA/PBS was applied for 30 minutes at room temperature. The slides were washed in PBS and incubated in FITC-conjugated anti-K10 antibody (1:100; Covance) for 5 hours at room temperature. The slides were then washed, counterstained with 4',6-diamidino-2-phenylindole (Vector Laboratories), and coverslipped. For Scribble immunofluorescence, the deparaffinized, rehydrated histologic sections were washed in PBS and incubated in trypsin (1 mg/mL) for 20 minutes at room temperature. After washing in PBS, the slides were blocked in 10% chicken serum/PBS for 1 hour at room temperature. The slides were then incubated overnight at 4°C in goat anti-Scribble antibody (1:100; Santa Cruz Biotechnology, Santa Cruz, CA). After washing in PBS, Alexa 488 chicken anti-goat secondary antibody (1:200; Molecular Probes) was applied for 30 minutes at room temperature. The slides were then washed, counterstained with propidium iodide (Vector Laboratories), and coverslipped.
Multistage skin carcinogenesis study. The backs of 4- to 6-week-old mice were shaved and painted once with 1.0 nmol 7,12-dimethylbenz(a)anthracene (DMBA) in acetone. Starting 2 weeks following DMBA treatment, mice were treated twice weekly with 15 nmol 12-O-tetradecanoylphorbol-13-acetate (TPA) in acetone for 20 weeks. The mice were then allowed to age for up to 40 weeks after DMBA treatment dependent on remaining carcinoma-free. Benign (papilloma) and malignant (carcinoma) tumors were counted biweekly.
Statistical analysis. Spontaneous skin tumor development was compared using the two-sided Fisher's exact test. For the multistage skin carcinogenesis study, the average number of papillomas per mouse was compared at 20 weeks after DMBA treatment using the Wilcoxon rank-sum test. To compare carcinoma incidence, a Kaplan-Meier survival curve was plotted and the log-rank test was used to compare the percentage of mice developing carcinomas among the three groups. The log-rank test takes into account data accrued throughout 20 to 40 weeks after DMBA treatment and is designed to determine whether any temporal differences exist. All statistics calculations were carried out using the MSTAT statistics software program.3
Western blot analysis. To isolate protein from epidermis, the dorsal skin of mice was harvested at the time of euthanasia and the underlying dermis was scraped away using a scalpel blade. On ice, the remaining epidermal scrape was sonicated thrice for 5 seconds in radioimmunoprecipitation assay (RIPA) buffer [50 mmol/L Tris (pH 8.0), 150 mmol/L NaCl, 1% NP40, 0.5% sodium deoxycholate, 0.1% SDS] containing protease inhibitors (50 µmol/L tosyl-lysine chloromethyl ketone, 1 µmol/L aminocaproic acid, 10 µmol/L benzamidine, 2 µmol/L leupeptin, 2 µmol/L pepstatin, 0.05 µmol/L aprotinin). Lysates were rotated at 4°C for 30 minutes followed by centrifugation at 14,000 rpm for 30 minutes at 4°C. The supernatant was collected and protein concentrations were determined using the Lowry method (Bio-Rad, Hercules, CA). In tissue culture cells, cellular protein lysates were isolated by scraping cells into RIPA buffer containing protease inhibitors. To separate the soluble from the insoluble fraction of cellular lysates, cells were lysed in 250 mmol/L NaCl, 0.1% NP-40, 50 mmol/L HEPES, pH 7.0 containing protease inhibitors. The lysates were vortexed every 5 minutes for 30 minutes on ice. Following a 30-minute centrifugation at 4°C, the supernatant was collected and the protein concentrations were determined using the Lowry method. For fraction isolation the remaining pellet represented the insoluble fraction. Equal amounts of protein were loaded onto precast SDS-PAGE gels (Bio-Rad ready gels). After transfer to polyvinylidene difluoride membranes (Millipore, Billerica, MA), the blots were blocked overnight at 4°C in 5% milk/PBS-T (PBS containing 0.1% Tween 20). Specific proteins were detected using the following antibodies: goat anti-Scribble, mouse anti-human Dlg (PharMingen, San Diego, CA), a mixture of mouse anti-E7 (1:200; Santa Cruz Biotechnology) and mouse anti-E7 (1:150; Zymed, San Francisco, CA), horseradish peroxidase (HRP) donkey anti-ß-actin (1:5,000; Jackson ImmunoResearch, West Grove, PA), and HRP rabbit anti—glyceraldehyde-3-phosphate dehydrogenase (GAPDH; 1:800,000; Chemicon, Temecula, CA). Enhanced chemiluminescence detection was employed according to the manufacturer's instructions (Amersham Biosciences, Piscataway, NJ).
Isolation of tumor genomic DNA. The tumors were isolated and the skin was removed before the addition of 500 µL digestion buffer [100 mmol/L Tris (pH 8.0), 200 mmol/L NaCl, 5 mmol/L EDTA, 0.2% SDS, 200 µg/mL proteinase K] and incubation at 55°C overnight. DNA was precipitated from the supernatant the following day in 2-propanol and washed with 70% ethanol. The resuspended DNA was then phenol/CHCl3 extracted and reprecipitated with sodium acetate (0.3 mol/L) and 70% ethanol. The air-dried DNA pellet was resuspended in DNase-free water at 37°C and stored at 4°C until use.
Comparative genome hybridization analysis. The protocol was done as reported previously (24).4 Briefly, genomic DNA (2 µg) from each tumor and from normal female mouse genomic DNA was differentially labeled via nick translation. Labeled tumor and reference DNA (500-800 ng each) were combined with 15 µg mouse CotI DNA (Invitrogen, Carlsbad, CA) and 1 µg salmon testes DNA (Sigma, St. Louis, MO) in hybridization buffer, denatured, and allowed to pre-anneal for 1 hour before addition to pretreated and denatured slides containing normal mouse metaphase spreads. Hybridization occurred for 48 hours at 37°C in a moist chamber. Tumor DNA was detected with avidin-FITC (Vector Laboratories) followed by amplification of the signal with a biotinylated sheep anti-avidin antibody (Vector Laboratories) and another layer of avidin-FITC. Reference DNA was detected with mouse anti-digoxigenin (Sigma) followed by amplification with rabbit anti-mouse-TRITC (Sigma) and goat anti-rabbit-TRITC (Sigma). Images were acquired using a Leica DM-RXA microscope (Leica, Wetzlar, Germany), Photometrics CCD camera (Photometrics, Tucson, AZ), and CW-4000 imaging software (Leica, Cambridge, United Kingdom). Images were analyzed and ratios were plotted along the length of the chromosome using CW4000 software (Leica, Cambridge, United Kingdom).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
146-151 mice, we discovered that E6's interactions with PDZ domain partners correlated with its ability to induce epithelial hyperplasia (21). Furthermore, we observed an absence in the induction of suprabasal DNA synthesis in the epidermis of the K14E6146-151 mice (21), a finding reproduced in Fig. 1A to C. To determine if E6's interactions with PDZ domain proteins also contribute to the delay in terminal differentiation observed in K14E6WT mice, we stained sections of ear epidermis with antibodies against the differentiation-specific markers K14 and K10. K14 is normally detected selectively in the basal compartment of the epidermis, whereas K10 is specifically detected uniformly within the terminally differentiating suprabasal compartment. As seen previously with the K14E6WT mice (3), expression of wild-type HPV16 E6 in the epidermis led to an expansion of the K14-positive compartment and a delay in the onset of K10-positive staining within the suprabasal compartment compared with that observed in the nontransgenic mice (Fig. 1D and E). Neither the expansion in K14 staining nor the delay in the onset of K10 staining was observed in the K14E6146-151 mice (Fig. 1F). This comparison and further studies described herein were carried out on lines of K14E6WT and K14E6146-151 mice that were matched for their level of transgene expression based on real-time PCR (21). These data indicate that E6's ability to interact with PDZ domain proteins contributes to both the induction of suprabasal DNA synthesis and the delay in differentiation observed in the epidermis, two properties that are thought to contribute to the epithelial hyperplasia observed in K14E6WT mice.
|
14% of K14E6WT mice develop spontaneous skin tumors, whereas no nontransgenic mice were found to develop spontaneous skin tumors in the same period (Table 1; ref. 3). To determine whether the domain of E6 that mediates its interaction with PDZ domain proteins is necessary for predisposing mice to spontaneous skin tumors, we monitored a cohort of K14E6146-151 mice for the same 15-month monitoring period. We found a single skin tumor to arise in 1 of 59 mice or an incidence of 1.7% (Table 1). This frequency was significantly lower than that observed in the K14E6WT mice (P = 0.004), indicating that E6's ability to bind PDZ domain partners correlates with its ability to cause tumors.
|
146-151 mice a single dose of DMBA at 4 to 6 weeks of age. Thereafter, the mice were treated twice weekly for 20 weeks with the promoting agent TPA and monitored for papilloma development. After the TPA treatment, the mice were allowed to age for 20 more weeks and were monitored for carcinoma development. During the first 20 weeks of the study, we found that K14E6WT mice developed a significantly greater average number of papillomas per mouse (2.82) than the nontransgenic mice (0.71; P = 0.03; Fig. 2A), in agreement with our previous studies (23). This increase in papilloma formation reflects E6's ability to contribute to promotion. In contrast, the K14E6146-151 mice did not develop a significantly different average number of papillomas per mouse (0.9) than the nontransgenic mice (P = 0.88). Furthermore, like the nontransgenic mice, the incidence of papillomas in the K14E6146-151 mice differed significantly from that observed in the K14E6WT mice (P = 0.006). Thus, we conclude that the ability of E6 to interact with PDZ domain partners correlates with its ability to contribute to the promotion stage of carcinogenesis.
|
146-151 (P = 0.002) groups developed carcinomas (Fig. 2B). This result indicates that, although E6's interactions with PDZ domain proteins correlate with E6's ability to act in the promotion stage of carcinogenesis, these PDZ interactions are not required for E6's progression activity in carcinogenesis. This is the first evidence showing that the contribution of E6 to the multiple stages of carcinogenesis can be mechanistically separated.
Comparison of genomic alterations in tumors derived from K14E6WT and K14E6146-151 mice. Characterization of the genomic alterations in carcinomas formed in K14E6WT mice revealed the consistent gain of mouse chromosome 6 in all of the cell lines established from these tumors and often the additional gain of chromosome 10 (25). This was strikingly different from the tumors that arose in mice transgenic for K14E7 in which no consistent genomic alterations were observed. In the original study, cell cultures were established from tumors and the cells were cytogenetically analyzed. To ensure that cytogenetic changes seen in that earlier study were not the result of genomic instability in tissue culture, early and late passages of cells from the K14E6WT tumors were reanalyzed and found to have the same cytogenetic abnormalities (data not shown). For this study, we did not culture the tumors but simply prepared genomic DNA from the dissected K14E6146-151 tumor mass, as the tumor masses were well defined and surrounding normal tissue could be removed by microdissection. Eleven tumors were analyzed by comparative genome hybridization (CGH). Eight of the tumors showed copy number alterations in at least one chromosome, indicating that the tumor cells were well represented in the DNA samples extracted from the tumors. Gains in regions of chromosomes 7 and X, which were observed in approximately half of the tumors, were the most common changes seen (Fig. 3). The minimum region of overlap for the five tumors containing a regional gain on chromosome 7 were bands B5 to C. This corresponds approximately to 43.3 to 6.5 Mb. Bands E1 to E3 constituted the minimal overlap region on the X chromosome. Of greatest significance, gains of chromosome 6, which had been observed previously in 100% of cells derived from K14E6WT tumors, were only observed in 2 of 11 K14E6146-151 tumors. Thus, it seems that tumor formation in the absence of the PDZ domain of E6 is not dependent on aneuploidy for mouse chromosome 6.
|
146-151 mice. Surprisingly, we observed no differences in the amount of Scribble or Dlg protein in the epidermal scrapes from the K14E6WT mice compared with the nontransgenic or K14E6146-151 mice (Fig. 4A). The absence of any reduction in these PDZ domain partners could reflect differences in how the E6 protein affects mouse versus human homologues or differences in the levels of E6 protein expressed in the mouse model versus in the natural host cells expressing the E6 gene from its homologous viral promoter. We therefore did Western analysis on a HPV16-positive human keratinocyte cell line, NIKS, which harbors the intact viral genome as a low-copy nuclear plasmid and which on terminal differentiation supports the entire viral life cycle. The presence of the HPV16 replicon in this human keratinocyte cell line led to no change in the steady-state levels of Dlg or Scribble (Fig. 4A). Considering that the degradation of the proteins by E6 could be an event that occurs only during tumor development, we looked at Scribble and Dlg protein levels in cell lines isolated from skin tumors from K14E7WT and K14E6WT mice that arose spontaneously or arose as a consequence of topical treatment with chemical carcinogens. The levels of Dlg and Scribble present in these tumors were equivalent to each other and to that seen in epidermal scrapes from 9-day-old nontransgenic mice (Fig. 4B). We extended this analysis to human cell lines derived from a CIN lesion and frank cervical cancer. Clonal cell populations from a human CIN lesion harboring the HPV16 genome either extrachromosomally (W12E) or integrated into the host genome (W12I) and from a HPV16-positive cervical cancer lesion (CaSki) were compared to determine if steady-state Dlg or Scribble protein levels are altered with the increased E6/E7 RNA expression that occurs on viral integration (Fig. 4C). These cells were also compared with C33a cells, a HPV-negative cervical cancer cell line, to determine whether changes in protein levels were dependent on the expression of E6 versus the development of cancer. E7 expression was detected in W12E, W12I, and CaSki cells but not in C33a cells, consistent with the HPV status of each of these cell lines. E7 expression was increased in W12I cells compared with W12E cells; however, no apparent increase was observed in the CaSki cells compared with the W12E cells (Fig. 4C). Dlg protein levels were inversely correlated with those of E7 in the different cell lines; that is, a decrease was observed in W12I cells but not in CaSki cells compared with W12E cells. However, Dlg protein was difficult to detect in C33a cells, suggesting that loss of Dlg may correlate with cancer development in some cases and may not be dependent on expression of E6 (Fig. 4C). In contrast, Scribble protein seemed to be slightly decreased in W12I and CaSki cells compared with W12E cells. This decrease was not observed in C33a cells, suggesting that decreased Scribble levels are dependent on expression of E6 (Fig. 4C). Studies by Massimi et al. have suggested that E6 expression may alter the levels of Dlg and Scribble specifically in the nuclear and cystoplasmic fraction (26). In order to determine if this was true in NIKS cells we carried out Western blot analysis comparing the levels of Dlg and Scribble in the soluble and insoluble fractions of NIKS and NIKS harboring the HPV16 genome (Fig. 4D). In contrast to previous studies we were unable to observe any differences in the levels of Dlg between NIKS and NIKS harboring HPV16 in both the soluble and insoluble fractions. Scribble analysis indicated that, if anything, Scribble levels might be upregulated in the presence of HPV16. Although E6 as expressed in our E6 transgenic mice or from the viral genome in NIKS does not alter the steady-state level of Scribble or Dlg, it may alter its function by changing its cellular localization. We therefore carried out immunofluorescence studies on Scribble, for which a suitable antibody is available. As expected, immunofluorescence analysis revealed that Scribble was primarily localized to the cell membrane in normal mouse epidermis (Fig. 5A). We were unable to detect any differences in Scribble localization among 9-day-old, K14E6WT, nontransgenic, and K14E6146-151 mice. Likewise, we observed no difference in the Scribble staining pattern in organotypic raft cultures of the human keratinocyte cell line NIKS that did or did not harbor the HPV16 genome (Fig. 5B). Interestingly, we observed a difference in Scribble localization in a raft of CaSki cells. Here, Scribble appeared more punctate and cytoplasmic compared with both the NIKS and the NIKS harboring the HPV16 genome rafts (Fig. 5B). This is consistent with studies that monitored the localization of Scribble in normal and progressively dysplastic HPV-positive cervical epithelium (27). To determine if a similar pattern was observed in our mouse model, we isolated tumors from K14E6WT and nontransgenic mice treated with DMBA and TPA and stained them for Scribble. Similar to what we observed in the skin of 9-day-old mice, Scribble was detected at the membrane of normal tissue adjacent to tumors from both K14E6WT and nontransgenic mice (Fig. 6A and D). In contrast to the normal tissue, Scribble appeared more punctate and cytoplasmic in tumor sections from both K14E6WT and nontransgenic mice (Fig. 6B, C, E, and F).
|
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
146-151 mice retain an increased incidence of progression indicates that some property of E6 other than its interaction with PDZ domain proteins mediates E6's contribution to malignant conversion. This provides the first direct evidence that distinct activities of E6 contribute to distinct stages in carcinogenesis. We suspect that this other property is E6's inactivation of p53, as mice deficient for p53 specifically show an enhanced rate and incidence of progression but no increase in initiation or promotion (28).
CGH analysis indicated that the chromosomal aberrations in tumors from K14E6146-151 mice were not similar to those observed previously in K14E6WT tumors. In this study, CGH analyses were carried out on DNA extracted directly from carcinomas isolated from the K14E6146-151 mice, whereas the prior CGH analysis of tumors from K14E6WT mice was carried out on DNA isolated from cell cultures derived from those tumors. For several reasons, we do not believe that culturing the cells is responsible for the striking difference in frequency of gains in chromosome 6. First, the consistent gain of chromosome 6 seen in cells from the K14E6WT tumors was not seen in cells from the K14E7WT tumors, indicating that culturing of squamous carcinomas per se is not responsible for gains in chromosome 6. Second, prior studies have shown that karyotypes remain quite stable with passaging of cells derived from thymic lymphomas (29). Third, and perhaps most importantly, we reanalyzed early and late passages of the cell populations derived from the K14E6WT tumors that were used in the prior study analysis (25) and found that early and late passages show the same cytogenetic changes as described previously (data not shown). Although CGH analysis of DNA extracted directly from the tumor mass avoids the potential concern for tissue culture artifacts, it raises the concern that cytogenetic abnormalities might be missed if the fractional content of the tumor cell type over normal and/or stromal cells is insufficient. In our analysis, 8 of 11 DNA samples isolated from the K14E6146-151 tumors showed one or more chromosomal gain or loss; therefore, we are confident at least for those 8 samples there was sufficient representation of tumor cells.
Although a consistent gain of chromosome 6 was present in all of the tumors from K14E6WT mice (25), only 2 of 11 carcinomas from K14E6146-151 mice displayed a gain of chromosome 6. These results indicate that E6's interactions with PDZ domain proteins correlate with a gain of chromosome 6 in tumors. This finding may indicate a functional link between E6's interactions with PDZ domain proteins and one or more genes on chromosome 6. However, it is unclear whether this link would be direct or indirect. Several genes on chromosome 6 present interesting potential targets for a direct link. These include K-ras, ephrin receptors B6 and A1, and transforming growth factor-
(TGF-). K-ras, a commonly mutated protein in cancer, is an interesting target; however, in our previous studies, no up-regulation of this gene was observed in the K14E6WT carcinomas (25). The ephrin receptors B6 and A1 are part of the receptor tyrosine kinase family of ephrin receptors. Alterations in expression of ephrin receptors and ligands have been found in a variety of cancers and in some cases have been used for diagnostic purposes (30, 31). Interestingly, both ephrin receptors and their ligands contain PDZ-binding motifs at their COOH terminus; however, it is unclear what functional role the motif plays (ref. 32; for review, see ref. 33). TGF- is a ligand for the epidermal growth factor receptor, whose expression has been implicated in E6-induced transformation and tumorigenesis (34, 35). Interestingly, TGF- membrane trafficking has been linked to a Golgi-associated PDZ domain protein, p59/GRASP55 (36).
Chromosomal aberrations are generally considered to be a hallmark of the progression stage of carcinogenesis. Therefore, it is interesting that, although E6's interactions with PDZ domain proteins contribute specifically to promotion and not progression (Fig. 2), carcinomas from K14E6WT and K14E6146-151 do not display similar chromosomal aberrations. This result could be explained by invoking a model in which expression of E6 and the gain of chromosome 6 are indirectly related. In such a model, E6's interactions with PDZ domain proteins would lead to a signaling pathway that contributes to promotion and when coupled with a gain of chromosome 6 gives a selective advantage to those cells that then progress on to frank cancer. In contrast, when this signaling pathway is not active (i.e., E6 is unable to interact with PDZ domain proteins), there is no selective advantage for a gain of chromosome 6 in these cells; instead, other genetic changes provide selective advantage for progression. Interestingly, this hypothesis would suggest that tumors in K14E6WT and K14E6146-151 mice are progressing through different mechanisms, an outcome we would not have predicted simply based on our results in the multistage carcinogenesis study.
To date, it is unclear what PDZ domain protein(s) is contributing to these E6-induced phenotypes. Of the six described partners, Scribble and Dlg perhaps seem to be the most likely candidates considering their growth control activities in Drosophila. Studies in vitro and in tissue culture have suggested that both of these proteins are targeted for degradation by E6, possibly explaining E6's PDZ-dependent properties. Using HPV16- and HPV18-positive cervical carcinoma cell lines, Massimi et al. showed that Scribble and Dlg proteins found in the soluble fraction were specifically targeted for degradation (26). Immunofluorescence coupled with small interfering RNA revealed that E6 specifically decreases the levels of Dlg in the nucleus and cytoplasm (26). In our hands, we failed to observe any change in the steady-state level of these proteins in total lysates from both K14E6 and nontransgenic mouse epidermis, in tumors that arose from the E6 transgenic mice, and from human keratinocytes harboring HPV16 (Fig. 4A and B). We did observe changes in Scribble and Dlg expression when comparing their expression in cell lysates from human CIN-derived cell lines, a human HPV-positive cervical cancer–derived cell line (CaSki), and a human HPV-negative cervical cancer cell line (C33a; Fig. 4C). Dlg levels were inversely correlated with E7 levels; however, the low Dlg expression in C33a cells indicates that E6 expression is not required for decreased Dlg expression. On the other hand, Scribble protein appeared to decrease modestly, specifically when the HPV genome had integrated into the genome of human cervical cancer cell lines, suggesting that it may be a more relevant mediator of E6-induced phenotypes (Fig. 4C). Our analysis of soluble and insoluble fractions indicates that Dlg levels are not altered in either fraction in NIKS harboring the HPV16 genome (Fig. 4D). This result contrasts with that of Massimi et al. (26). A possible explanation may be that higher levels of E6 than those in NIKS harboring the HPV16 genome are required for E6 to induce the degradation of soluble Dlg. These effects may also be tissue specific, as decreased levels of Scribble protein are detected in the presence of E6 in the lens of K14E6WT mice (22). E6 could use other mechanisms than destabilizing proteins to alter the function of its PDZ protein partners. These may include changing the localization of PDZ domain protein(s) and/or altering the activity of PDZ domain proteins. Immunofluorescence studies in mouse epidermis suggest that on a gross level E6 does not alter the localization of Scribble (Fig. 5). However, subtle changes in subcellular localization may not have been evident. We noted changes in Scribble localization in the raft composed of the HPV16-positive cervical cancer cell line CaSki (Fig. 5), consistent with a study of human cervical lesions (27). That study indicated that, as normal cervical epithelium becomes progressively more dysplastic, the localization of Scribble becomes more cytoplasmic and is actually undetectable in frank cancers. Interestingly, we observed a similar staining pattern in tumors induced with DMBA and TPA in our mice. However, the change in localization of Scribble did not differ between tumors isolated from nontransgenic and K14E6WT mice (Fig. 6). E6 is up-regulated in its expression in the context of human cervical cancer, a consequence of the viral genome becoming integrated into the host chromosomes. It has been hypothesized that this up-regulation of E6 may cause the changes in localization and/or levels of Scribble in the higher-grade cervical lesions in women. We are not aware of any such up-regulation of E6 expression in the tumors that arise in our K14E6 mouse model.
Studies in mice carrying nulligenic alleles of PDZ domain proteins will be critical to allow for a better understanding of the role these proteins play in mammalian cell signal regulation and in cancer. Mice carrying a Dlg gene trap mutation indicate that Dlg plays a critical role in development, as these mice die around embryonic day 18.5 (e18.5; ref. 37). Our own studies indicate that at e18.5 the Dlg gene trap mutation does not induce suprabasal DNA synthesis or delayed terminal differentiation (data not shown), possibly suggesting that the inactivation of Dlg is not sufficient to account for the these phenotypes in K14E6WT mice. However, K14E6WT mice also do not display these phenotypes at e18.5 likely because expression of the transgene cannot be detected until 6 to 9 days after birth. Additionally, the truncated Dlg protein produced from the gene trap retains all three PDZ domains and therefore may still retain some function. Finally, a combination of these activities may be required for E6 to induce PDZ domain protein-dependent phenotypes. To elucidate this, mice carrying mutant alleles for combinations of PDZ domain proteins will be required. These studies await the availability of mice carrying conditional null alleles of these essential genes.
|
Acknowledgments |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank Amy Liem and the University of Wisconsin Comprehensive Cancer Center Histology Facility for technical assistance, Dr. Norman Drinkwater for critical reading of the article, and Lance Rodenkirch and the W.M. Keck Laboratory for biological imaging.
|
Footnotes |
|---|
4 http://www.riedlab.nci.nih.gov/protocols.asp.
Received 5/13/05. Revised 6/22/05. Accepted 7/ 1/05.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-helix partners displays reduced oncogenic potential in vivo. J Virol 2002;76:13039–48. tethers to the PDZ domain-containing, Golgi membrane-associated protein p59/GRASP55. EMBO J 2000;19:6427–39.[CrossRef][Medline]
This article has been cited by other articles:
|
|
 |
|
  R. M. Vos, J. Altreuter, E. A. White, and P. M. Howley The Ubiquitin-Specific Peptidase USP15 Regulates Human Papillomavirus Type 16 E6 Protein Stability J. Virol., September 1, 2009; 83(17): 8885 - 8892. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S.-H. Chung, K. Wiedmeyer, A. Shai, K. S. Korach, and P. F. Lambert Requirement for Estrogen Receptor {alpha} in a Mouse Model for Human Papillomavirus-Associated Cervical Cancer Cancer Res., December 1, 2008; 68(23): 9928 - 9934. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. C. Spanos, A. Hoover, G. F. Harris, S. Wu, G. L. Strand, M. E. Anderson, A. J. Klingelhutz, W. Hendriks, A. D. Bossler, and J. H. Lee The PDZ Binding Motif of Human Papillomavirus Type 16 E6 Induces PTPN13 Loss, Which Allows Anchorage-Independent Growth and Synergizes with Ras for Invasive Growth J. Virol., March 1, 2008; 82(5): 2493 - 2500. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Topffer, A. Muller-Schiffmann, K. Matentzoglu, M. Scheffner, and G. Steger Protein tyrosine phosphatase H1 is a target of the E6 oncoprotein of high-risk genital human papillomaviruses J. Gen. Virol., November 1, 2007; 88(11): 2956 - 2965. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. P. Maufort, S. M. Genther Williams, H. C. Pitot, and P. F. Lambert Human Papillomavirus 16 E5 Oncogene Contributes to Two Stages of Skin Carcinogenesis Cancer Res., July 1, 2007; 67(13): 6106 - 6112. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. C. Hoover, W. C. Spanos, G. F. Harris, M. E. Anderson, A. J. Klingelhutz, and J. H. Lee The Role of Human Papillomavirus 16 E6 in Anchorage-Independent and Invasive Growth of Mouse Tonsil Epithelium Arch Otolaryngol Head Neck Surg, May 1, 2007; 133(5): 495 - 502. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. H. Storrs and S. J. Silverstein PATJ, a Tight Junction-Associated PDZ Protein, Is a Novel Degradation Target of High-Risk Human Papillomavirus E6 and the Alternatively Spliced Isoform 18 E6 J. Virol., April 15, 2007; 81(8): 4080 - 4090. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Shai, T. Brake, C. Somoza, and P. F. Lambert The Human Papillomavirus E6 Oncogene Dysregulates the Cell Cycle and Contributes to Cervical Carcinogenesis through Two Independent Activities Cancer Res., February 15, 2007; 67(4): 1626 - 1635. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Kuballa, K. Matentzoglu, and M. Scheffner The Role of the Ubiquitin Ligase E6-AP in Human Papillomavirus E6-mediated Degradation of PDZ Domain-containing Proteins J. Biol. Chem., January 5, 2007; 282(1): 65 - 71. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
; n = 26), K14E6WT (
; n = 30), and K14E6
; n = 35) mice. Tumors were scored by visual inspection and palpation on a weekly basis. A, average number of papillomas (i.e., benign tumors) per mouse arising as a function of time following initial treatment with DMBA. B, Kaplan-Meier survival curve shows the number of animals remaining free of carcinomas (i.e., malignant tumors) following the termination of TPA treatment. Significance values for these data are provided in Results.
