High-risk human papillomaviruses (HPV) cause certain anogenital and head and neck cancers. E6, one of three potent HPV oncogenes that contribute to the development of these malignancies, is a multifunctional protein with many biochemical activities. Among these activities are its ability to bind and inactivate the cellular tumor suppressor p53, induce expression of telomerase, and bind to various other proteins, including Bak, E6BP1, and E6TP1, and proteins that contain PDZ domains, such as hScrib and hDlg. Many of these activities are thought to contribute to the role of E6 in carcinogenesis. The interaction of E6 with many of these cellular proteins, including p53, leads to their destabilization. This property is mediated at least in part through the ability of E6 to recruit the ubiquitin ligase E6-associated protein (E6AP) into complexes with these cellular proteins, resulting in their ubiquitin-mediated degradation by the proteasome. In this study, we address the requirement for E6AP in mediating acute and oncogenic phenotypes of E6, including induction of epithelial hyperplasia, abrogation of DNA damage response, and induction of cervical cancer. Loss of E6AP had no discernible effect on the ability of E6 to induce hyperplasia or abrogate DNA damage responses, akin to what we had earlier observed in the mouse epidermis. Nevertheless, in cervical carcinogenesis studies, there was a complete loss of the oncogenic potential of E6 in mice nulligenic for E6AP. Thus, E6AP is absolutely required for E6 to cause cervical cancer. Cancer Res; 70(12); 5064–73. ©2010 AACR.
Human papillomaviruses (HPV) are etiologically associated with cancers of the head and neck and the anogenital tract, which includes the cervix. Infection with HPVs is attributed with almost all cases of cervical cancer (1, 2). Cervical cancer accounts for ∼5% of the world's total tumor burden (3) and is the second leading cause of cancer-related deaths among women (4, 5). Approximately 70% of all cervical cancers are attributed to infection with the “high-risk” HPV types 16 or 18. In cervical cancers, two HPV genes, E6 and E7, are expressed (6, 7). HPV E6 and E7 have transforming and oncogenic properties in tissue culture (8, 9) and in animal models (10-12). E6 and E7 are primarily known for their abilities to target the tumor suppressors p53 (13, 14) and pRb (15), respectively, for inactivation and ubiquitin-mediated degradation.
High-risk E6 is known to target p53 for ubiquitin-mediated degradation through its interaction with the E3 ligase E6-associated protein (E6AP). Similarly, E6 has been found to induce the degradation of other cellular proteins, and in many cases, it is mediated through interactions with E6AP (16, 17). Although the majority of data support the requirement of E6AP in E6-mediated degradation of p53 (18, 19), our previous studies in the mouse skin indicate that HPV16 E6 is able to target p53 for inactivation and degradation in the absence of E6AP (20). Results from other groups using either in vitro or in vivo systems also suggest the existence of E6AP-independent means for high-risk HPV E6 to target p53 for ubiquitin-mediated degradation (21, 22). Therefore, it remained unclear what role E6AP would have in E6-mediated acute and oncogenic phenotypes in the context of the cervix.
In mice, E6 and E7 cooperate to induce cervical carcinogenesis (23, 24). HPV E6 induces cervical cancer through its ability to inactivate p53 and other cellular proteins, including PDZ-containing proteins that interact directly with E6 (24, 25). Similar requirements were needed for E6 to induce skin cancers in mice (26). Because HPV16 E6 can target p53 for inactivation in the skin through E6AP-independent means (20), it was of interest to determine the importance of E6AP in cervical carcinogenesis.
Further interest fueled our pursuits of E6AP in the cervix. In addition to its function as an E3 ubiquitin ligase, E6AP also functions as a steroid receptor coactivator (27) and may be differentially regulated in cancers of the breast and prostate (28, 29). Recent evidence shows that the nuclear hormone receptor estrogen receptor α (ERα) is a necessary cofactor in HPV-mediated cervical carcinogenesis (30). Given the involvement of E6AP in the regulation of nuclear hormone receptors, its importance in cervical cancer warranted further investigation.
To determine the role of E6AP in the oncogenic effects of E6, we generated HPV transgenic mice on an E6AP-null (E6AP−/−) background. Our results show that in the cervix, loss of E6AP ablates the oncogenic activity of E6. We conclude that E6AP is a critical mediator of the oncogenic activities of HPV16 E6 in the cervix.
Materials and Methods
The mouse lines K14E6WT (31), K14E7WT (10), and E6AP−/− (20, 32) were maintained on the inbred FVB/n background. All mice were bred and maintained in the American Association for Accreditation of Laboratory Animal Care–approved McArdle Laboratory Animal Care Facility in accordance with an institutionally approved animal protocol.
Estrogen treatment and analysis of reproductive tracts
For acute studies, 5-week-old virgin female mice were treated with 17β-estradiol for 6 weeks to achieve a state of persistent estrus (24, 25). Mice acutely treated with estrogen were either exposed to 10 Gy of ionizing radiation from a 137Cs source or left untreated and sacrificed 24 hours later. One hour before sacrifice, the nucleotide analogue 5-bromo-2′-deoxyuridine (BrdUrd; 12.5 mg/mL in PBS) was injected i.p. at 10 μL/1 g body weight to label any newly synthesized DNA.
For acute studies that require ovariectomy, 8-week-old virgin female mice were ovariectomized and allowed to finish the remaining cycle and recover for 10 days. The mice were then injected with either estrogen (30 μg/kg weight) or ethanol vehicle and sacrificed 24 hours later. Animals were perfused with 4% paraformaldehyde, and reproductive tracts were harvested for analysis.
For long-term cancer studies, 5-week-old virgin female mice were either untreated or treated with 17β-estradiol for a period of 6 to 9 months, after which the reproductive tracts were harvested and fixed according to previously described methods (23) and sectioned. Every tenth, 5-μm section was stained with H&E and evaluated histopathologically, with the worst lesion scored as the final diagnosis. The severity of disease was evaluated by taking the spectrum of lesions within a given group and given a score of 0 to 5. Specifically, a diagnosis of hyperplasia is given a value of 0; cervical intraepithelial neoplasia 1 (CIN1) is given a value of 1 and increases progressively with each higher lesion grade to a score of 5 for large invasive cancer (LIC).
Quantification of DNA synthesis and DNA damage response
To quantify the levels of DNA synthesis in the epithelium of the cervix, basal and suprabasal BrdUrd-positive cells were counted and divided by the total number of cells and multiplied by 100 to determine the percentage. To quantify the DNA damage response, the total number of BrdUrd-positive cells was counted and divided by the total number of cells in the epithelium and multiplied by 100. BrdUrd was counted in at least eight, 40× microscopic fields per mouse, with a total of at least three or more mice per genotype group.
A two-sided Wilcoxon rank sum test was used for determining statistical significance in the quantification of DNA synthesis, tumor multiplicity, tumor burden, and severity of disease. A two-sided Fisher's exact test was used in determining significance for cancer incidence.
E6-induced cervical epithelial hyperplasia is not dependent on E6AP
Previous results indicate that the ubiquitin ligase E6AP is not required for E6-induced epithelial hyperplasia in the skin (20). To determine whether E6AP is required for E6-induced cervical epithelial hyperplasia, we crossed K14E6WT transgenic mice onto an E6AP−/− background and monitored for proliferation in young female mice treated acutely with estrogen. Acute treatment with estrogen is necessary to avoid variability in the thickness of the cervical epithelium and synchronizes all mice. Animals were injected with the nucleotide analogue BrdUrd 1 hour before sacrifice. Proliferation was subsequently measured by quantifying the incorporation of BrdUrd into nascently synthesized DNA via BrdUrd-specific immunohistochemistry. Similar to previous results, K14E6WT transgenic mice were able to induce hyperplasia in the cervical epithelium as indicated by higher levels of suprabasal DNA synthesis relative to nontransgenic (NTG) mice (0.9% versus 0.4%; P = 0.03; Fig. 1A, inset). Loss of E6AP in K14E6WT transgenic mice (K14E6WTE6AP−/−) did not result in a loss of epithelial hyperplasia (0.9%) and was not different from that of K14E6WT mice (P = 1). Surprisingly, the loss of E6AP in the absence of HPV E6 (E6AP−/−) also resulted in significantly higher levels of suprabasal DNA synthesis relative to NTG mice (1.0% versus 0.4%; P = 0.02). Thus, E6AP is not the mediator of HPV E6–induced hyperplasia.
A trivial explanation for our observation that E6AP is not required for E6-induced cell proliferation is that E6 is itself causing lower steady-state levels of E6AP. To determine if this is the case, we performed E6AP-specific Western blot analysis. We observed no obvious differences in the levels of E6AP in NTG versus K14E6WT, K14E7WT, or K14E6WTE7WT transgenic tissues (Supplementary Fig. S1). Thus, in our mouse model, E6 does not alter levels of E6AP.
Inhibition of the DNA damage response by HPV E6 in the absence of E6AP
Previous results show that HPV E6 transgenic mice retain the ability to inhibit DNA damage responses in response to ionizing radiation in the epidermis of the mouse skin despite the absence of E6AP (20). Given these results, we wanted to determine whether HPV E6 maintains the ability to inhibit the DNA damage response in the absence of E6AP in the cervical epithelium. Mice acutely treated with estrogen were either given 10 Gy of ionizing radiation or left untreated and harvested 24 hours later. One hour before sacrifice, all of the mice were injected with BrdUrd to label nascently synthesized DNA. We assessed two different responses to DNA damage: (a) the ability to cease DNA synthesis and (b) the induction and subsequent accumulation of p53. As expected, DNA synthesis was significantly reduced in both NTG (2.0 versus 0.3%; P = 0.01) and E6AP−/− (3.4 versus 0.8%; P = 0.03) mice on radiation (Fig. 1B and C). Similar to previous results observed in the skin, DNA synthesis in K14E6WT transgenic mice was not significantly changed upon radiation (2.6% versus 2.1%; P = 0.09). Interestingly, K14E6WTE6AP−/− transgenic mice did not show a reduction in DNA synthesis on DNA damage (2.3% versus 3.5%; P = 0.16), suggesting that, in the cervical epithelium, E6AP is not required for the targeted inactivation of E6 of p53 in inhibiting the DNA damage response. This result is in kind to that observed previously in the skin (20).
To address whether E6 can prevent the accumulation of p53 after DNA damage in the absence of E6AP, immunohistochemistry for p53 was performed on sections from both unirradiated and irradiated mice. In the absence of any DNA damage, p53 was undetectable in all genotypes (data not shown). As expected, p53 was highly induced in the cervical epithelium of NTG and E6AP−/− mice upon radiation (Fig. 1C). In contrast to skin where E6 is able to largely prevent induction of detectable p53 following irradiation, p53 levels in the irradiated K14E6WT cervix varied from nil to modest levels, with positive cells found predominantly in the basal layer. Thus, in the cervix, E6 partially prevents the accumulation of p53 in response to irradiation, although it efficiently prevents DNA damage–associated growth arrest (Fig. 1B). The cervical epithelia of irradiated K14E6WTE6AP−/− transgenic mice also displayed low to modest levels of p53 protein, similar to K14E6WT mice, indicating that E6AP is not required for the ability of E6 to prevent accumulation of p53 following irradiation. This is similar to what was observed in the skin of E6AP-null mice (20).
The acute response to estrogen is affected in mice without E6AP
In addition to its function as an ubiquitin ligase, E6AP also functions as a steroid receptor coactivator (27). E6AP has been reported to target various nuclear hormone receptors (29), including ERα, for ubiquitin-mediated degradation (33). Because E6AP is reported to play a role in estrogen signaling, we investigated whether loss of E6AP has any effect in the cervix, where estrogen signaling plays an important role. Effects from estrogen signaling can be observed quickly, and thus, the acute response to estrogen was assessed in groups of NTG and E6AP−/− mice. Mice were ovariectomized to eliminate endogenous levels of estrogen and treated 10 days later with either vehicle or estrogen for 24 hours, and after which, the reproductive tracts were harvested. Ovariectomy results in a thin epithelium throughout the reproductive tract due to a lack of proliferation signals from ovarian hormones. The normal response to estrogen is an increase in uterine wet weight (34) and the induction of epithelial hyperplasia. The epithelium from NTG and E6AP−/− groups treated with ethanol vehicle were hypoplastic. On estrogen treatment, the majority of reproductive tracts from NTG mice were overtly (data not shown) and histologically thickened (Fig. 2). Acute treatment of E6AP−/− mice with estrogen, on the contrary, did not result in either an overt or a histologically thickened epithelium. It is not clear why this is the case because, at the time of sacrifice, the proliferative index in the cervical epithelium of these mice was similar to NTG mice as assessed by BrdUrd incorporation over a 1-hour period before sacrifice, and there was no detectable programmed cell death (data not shown). One possible explanation is that there is a slight delay in the response of the E6AP-null tissue to estrogen, given that E6AP-null tissue does undergo hyperplasia when treated long term with estrogen (see below).
E6-induced cervical carcinogenesis is dependent on E6AP
Given that ERα signaling is affected in E6AP−/− mice, we investigated whether loss of E6AP would result in a reduced ability of HPV E6 to induce cervical cancer. We therefore compared mice nulligenic for E6AP and bred them to K14E6WT mice. K14E6WT mice develop cervical cancer after treatment with physiologic levels of estrogen for 9 months (24). Cervical lesions that arise in these mice are histopathologically similar to that found in HPV16-infected women. Specifically, we observe a progressive disease that includes the development of CIN grades 1 to 3, followed by the onset of frank cancer of the cervix. We therefore compared the severity of cervical disease by comparing the complete spectrum of lesions observed within any given group (see Materials and Methods for details). Similar to previous findings, K14E6WT mice have an increase in the severity of disease, which was significantly different from NTG mice (P < <0.001). K14E6WT transgenic mice have an increase in tumor incidence, multiplicity, and tumor burden (defined as the average area of tumor mass) relative to NTG mice (Supplementary Fig. S2; Tables 1 and 2). In K14E6WTE6AP−/− transgenic mice, however, the severity of disease was significantly reduced (P < 0.001) when compared with K14E6WT. Specifically, K14E6WTE6AP−/− mice have reductions in the incidence of cervical cancer (6.3%; P = 0.04), tumor multiplicity (0.13; P = 0.06), and tumor burden (P = 0.04). Indeed, the incidence of cervical disease in the K14E6WTE6AP−/− mice was not significantly different from that observed in NTG mice. These results indicate that the capacity of E6 to induce cervical cancer was completely abrogated on an E6AP-null background.
In human cervical cancers, E6 is always expressed with E7, a second HPV oncogene. HPV16 E7 is the more potent oncogene in the mouse cervix, as K14E7WT mice develop cervical cancers after only treatment with 6 months of 17β-estradiol (23). In comparison, K14E6WT mice only develop cervical cancers when treated for 9 months (24). Double-transgenic mice (K14E6WTE7WT) expressing both HPV16 E6 and E7 develop larger cancers than K14E7WT mice when treated for 6 months with 17β-estradiol (23). To assess the importance of E6AP in cervical carcinogenesis in the presence of both E6 and E7, we assessed the severity and incidence of cervical disease in K14E6WTE7WT mice on an E6AP-sufficent or nulligenic background after treatment with estrogen for 6 months (Supplementary Fig. S2; Tables 2 and 3). Histopathologic diagnosis revealed that in the absence of E6AP (K14E6WTE7WTE6AP−/−), the severity of disease is significantly reduced in the cervix of HPV mice (P = 0.03). Moreover, K14E6WTE7WTE6AP−/− transgenic mice did not show any incidence of large invasive cancers (LIC) as opposed to an incidence of 41% in K14E6WTE7WT mice on the E6AP-sufficient background. To determine if E6AP status had any influence on carcinogenesis when only E7 is expressed, we compared the severity and incidence of cervical disease in K14E7WT mice on an E6AP-sufficent or nulligenic background (Table 3). No significant differences in cervical disease were observed. Based on these results, we conclude that E6AP is specifically a critical cofactor for HPV16 E6 in cervical carcinogenesis.
A progressive disease leading to cancer is also observed in the vaginal cavity or lower reproductive tract (LRT) in our HPV transgenic mouse model for cervical carcinogenesis, consistent with the role of high-risk HPVs in vaginal cancers (35, 36). We therefore assessed the influence of E6AP on vaginal carcinogenesis. Similar findings were observed as in the cervix. In K14E6WT mice treated with estrogen for 9 months, the absence of E6AP led to a complete abrogation of E6-dependent carcinogenesis in the vagina (Supplementary Fig. S2; Table 1). Likewise, the contribution of E6 was lost in K14E6WTE7WT mice treated with estrogen for 6 months in K14E6WTE7WTE6AP−/− mice (Supplementary Fig. S2; Table 3).
It is interesting to note that in our analysis of the influence of E6AP status on the acute response to estrogen in the cervical epithelia, we observed a defect in hyperplasia induced by estrogen in the E6AP-deficient cervix of ovariectomized mice. Nevertheless, in our 9-month estrogen-treated E6AP−/− mice, we observed hyperplasia, dysplasia, and even cancer at a low frequency, akin to what we observed in E6AP-sufficent mice (Table 1). These results indicate that any effect of E6AP deficiency on the function of estrogen in this tissue is not manifested in the long term and could simply reflect a slight delay in the acute response.
Reduced incidence of cervical cancer is not due to a lack of ERα expression
Given the acute and long-term results observed in mice on the E6AP−/− background, we wanted to determine whether the reduced incidence of cancer in K14E6WTE6AP−/− mice is due to a lack of ERα expression. Previous findings show that ERα is a necessary cofactor in cervical carcinogenesis in our estrogen-dependent HPV model (30). Thus, immunohistochemistry for ERα was performed on long-term–treated mice. There were varying levels of ERα in reproductive tract epithelium across all genotypes (data not shown). No differences were observed in the expression of ERα between NTG and E6AP−/− cancers. In the presence of either HPV E6 or E7, the loss of E6AP resulted in minor reductions in the expression of ERα; however, the cancers still expressed appreciable levels of ERα (Fig. 3A).
Reduced incidence of cancer is not a result of reduced proliferation or expression of the HPV biomarker MCM7
To confirm that the reduction in the severity of disease and cancer incidence in HPV transgenic mice on the E6AP−/− background is not due to an inability to proliferate, immunohistochemistry for BrdUrd was performed on NTG, E6AP−/−, K14E6WT, K14E6WTE6AP−/−, K14E7WT, K14E7WTE6AP−/−, K14E6WTE7WT, and K14E6WTE7WTE6AP−/− mice treated with estrogen for either 6 or 9 months (data not shown). As expected, proliferation was low in NTG mice. Similar to acute studies reported above, E6AP−/− mice had higher levels of proliferation. No differences in the levels of BrdUrd were observed in comparing HPV transgenic (K14E6WT or K14E6WTE7WT) mice either on an E6AP-sufficient or E6AP-insufficient background.
MCM7 is a biomarker for HPV-positive lesions in both clinical human and murine samples (37). Cancers from the above genotypes were evaluated for the expression of MCM7. Cancers from NTG and E6AP−/− mice expressed low but detectable levels of MCM7. Mice expressing either HPV E6 or E7 generally had increased levels of MCM7 expression, driven mostly by the inactivation of pRb by E7 (38) or the dysregulation of the p16/pRb pathway by E6 (24, 39), both of which can lead to subsequent activation of E2F target genes. Indeed, there were no significant differences in the expression of MCM7 between HPV transgenic mice on an E6AP-sufficient versus E6AP-insufficient background (Fig. 3B).
Loss of E6AP results in an increased frequency of detectable p53 in the reproductive tract
Due to the reduction in the incidence of cancer and the severity of disease in K14E6WTE6AP−/− mice, we wanted to determine whether the inability of E6 to degrade p53 accounted for these observations. To investigate this question, immunohistochemistry for p53 was performed on mice treated with estrogen. Cancers from NTG, E6AP−/−, K14E6WT, and K14E6WTE6AP−/− mice have little to no detectable p53 (Fig. 3B). However, focal expression of p53 is detectable in the reproductive epithelium of E6AP−/− and K14E6WT mice, with an appreciable increase in K14E6WTE6AP−/− mice. Expression of p53 was more likely to be detected in the epithelium of the LRT and thus may explain the rare occurrence of cancers in this region in mice that only express E6 and not the E7 transgene (24, 36).1 Mice expressing the E7 transgene have detectable levels of p53 in both the dysplastic epithelium and cancers, likely due to the ability of E7 to cause increases in the levels of p53 (40, 41). Accordingly, K14E6WTE7WTE6AP−/− transgenic mice have an appreciable increase in the expression of p53 over that of K14E6WTE6AP−/− mice, which is evident both in the epithelium and in the cancers. In sum, the loss of E6AP results in an increased frequency of detectable p53 in the reproductive tract that correlated with decreases in cancer incidence and the severity of disease in E6 transgenic mice.
In this study, we found that the loss of E6AP had little effect on acute activities of E6 in the cervix, yet had dramatic effects on the ability of E6 to synergize with estrogen to form cancers in the reproductive tract, in the presence or absence of E7. The loss of E6AP resulted in reductions in both cancer incidence and the severity of disease. In sum, this study shows that E6AP plays a critical role in mediating the contributions of HPV16 E6 to cervical cancer.
E6AP is not required for HPV E6–induced acute effects
Similar to previous observations in the skin of HPV E6 transgenic mice (20), E6AP was found to be required neither for the induction of epithelial hyperplasia nor for the inactivation of the DNA damage response in the reproductive tract. E6AP is thought to potentially mediate E6-induced hyperplasia due to its predicted ability to target the PDZ partners of E6, Scribble, and Dlg for degradation. Loss of either of these tumor suppressor genes in Drosophila results in the development of epithelial hyperplasia, loss of cell-cell contacts, and cancer development (42-44). In human cervical cancers, hScrib and hDlg become mislocalized and their levels progressively reduced (45, 46). Observations from the skin and the reproductive tract suggest that E6AP is not the mediator of E6-induced epithelial hyperplasia and/or that in its absence there is alternative ubiquitin ligase that can target these proteins for degradation. Evidence from others suggests that the latter hypothesis may be the case. Work from Massimi and colleagues (22) show that E6 maintains the ability to target both Scribble and Dlg for ubiquitin-mediated degradation in the absence of E6AP, in contrast to reports that argue for a dependence of E6AP (16). Moreover, the ability of E6 to inhibit the DNA damage response in both the skin and reproductive tract (this study) in the absence of E6AP similarly suggests that there are alternative means to inactivating p53. Alternatively, proposed mechanisms by which E6 inactivates p53 include repression of transactivation of p53 target genes either by acetylation (47) or by interaction of E6 with CBP/p300 (48), which may be relevant here.
Implications of our findings
Our studies provide clear evidence for an essential role of E6AP in mediating the oncogenic activity of HPV16 E6 but leave unanswered the mechanism underlying its requirement. Specifically, those acute phenotypes of E6 that were measured were retained on the E6AP-deficient background despite the complete loss of neoplastic disease. Our results suggest that the role of E6AP in mediating the activities of E6 has yet to be fully appreciated.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
We thank Denis Lee for assistance with immunohistochemistry and Dr. Elaine Alarid and members of the Lambert laboratory for insightful discussions.
Grant Support: National Cancer Institute grants CA098428, CA022443, and CA009135.
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.
- Received September 4, 2009.
- Revision received April 16, 2010.
- Accepted April 21, 2010.
- ©2010 American Association for Cancer Research.