Skin Hyperproliferation and Susceptibility to Chem
http://www.100md.com
病菌学杂志 2005年第23期
International Agency for Research on Cancer, World Health Organization, 69372 Lyon Cedex 08, France
Deutsches Krebsforschungszentrum, D-69120 Heidelberg, Germany
Gyn?kologische Molekularbiologie Frauenklinik der FSU Jena, Bachstr. 18, D-07743 Jena, Germany
ABSTRACT
The oncoproteins E6 and E7 of human papillomavirus type 38 (HPV38) display several transforming activities in vitro, including immortalization of primary human keratinocytes. To evaluate the oncogenic activities of the viral proteins in an in vivo model, we generated transgenic mice expressing HPV38 E6 and E7 under the control of the bovine homologue of the human keratin 10 (K10) promoter. Two distinct lines of HPV38 E6/E7-expressing transgenic mice that express the viral genes at different levels were obtained. In both lines, HPV38 E6 and E7 induced cellular proliferation, hyperplasia, and dysplasia in the epidermis. The rate of occurrence of these events was proportional to the levels of HPV38 E6 and E7 expression in the two transgenic lines. Exposure of the epidermis of nontransgenic mice to UV led to p21WAF1 accumulation and cell cycle arrest. In contrast, keratinocytes from transgenic mice continued to proliferate and were not positive for p21WAF1, indicating that cell cycle checkpoints are altered in keratinocytes expressing the viral genes. Although the HPV38 E6/E7-expressing transgenic mice did not develop spontaneous tumors during their life span, two-stage carcinogen treatment led to a high incidence of papillomas, keratoacanthomas, and squamous-cell carcinomas in HPV38 mice compared with nontransgenic animals. Together, these data show that HPV38 E6 and E7 display transforming properties in vivo, providing further support for the role of HPV38 in carcinogenesis.
INTRODUCTION
Nonmelanoma skin cancers (NMSC), basal-cell carcinoma (BCC), and squamous-cell carcinoma (SCC) are the most common malignancies occurring worldwide. Several findings from experimental systems in vitro and in vivo have shown that UV plays a direct role in skin carcinogenesis (3). In addition, a subgroup of the epitheliotropic human papillomaviruses may cooperate with UV in the development of NMSC (28). Individuals with an autosomal recessive disorder termed epidermodysplasia verruciformis (EV) are susceptible to infection by specific HPV types that belong to the genotype beta branch of the HPV phylogenetic tree (7). Individuals with EV develop confluent flat warts that, in approximately 30% of cases, progress to SCC on sun-exposed areas (11, 14, 20, 27). The involvement of an infective agent in skin carcinogenesis is supported by the fact that immunosuppressed organ transplant recipients have a 50- to 100-fold higher risk of developing NMSC than the general population (4, 16, 32, 33). In addition, independent studies have reported that DNA of cutaneous HPV types can be detected in SCC and BCC of immunocompetent individuals, suggesting their role in skin carcinogenesis in the general population (28). The potential carcinogenic role of the EV HPV types is consistent with the fact that other members of the papillomavirus family, i.e., the mucosal high-risk types, are associated with anogenital cancers (36) and a subset of head and neck cancers (17).
The products of two early genes, E6 and E7, are the major oncoproteins of the high-risk mucosal HPV types (e.g., HPV16 and HPV18) (23, 24). Both viral proteins are able to alter the regulation of the cell cycle and apoptosis by binding and inactivating several cellular proteins, including products of tumor-suppressor genes. HPV16 E6 binds p53 and promotes its degradation, mediating the interaction of p53 with the ubiquitin protein ligase E6AP (23). Similarly, HPV16 E7 interacts with the retinoblastoma protein (pRb), a negative cell cycle regulator, inducing its degradation via the ubiquitin pathway (25). In contrast to the E6 and E7 of the mucosal HPV types, very little is known about the oncoproteins of the EV HPV types. Initial studies on HPV5 and HPV8, the types most frequently detected in EV patients, showed lower in vitro transforming activity of their E6 and E7 compared with the oncoproteins of the high-risk mucosal HPV types (13, 35). However, in a recent study, transgenic (Tg) mice expressing the early region of HPV8 under the K14 promoter spontaneously developed multifocal skin tumors and, in 6% of cases, SCC (29).
We have recently identified an EV HPV type (HPV38) that displays in vitro transforming properties (6). HPV38 E7 inactivates pRb and induces loss of G1/S transition control with efficiency similar to HPV16 E7. In addition, HPV38 E6 and E7 can immortalize primary human keratinocytes that are naturally infected by the virus, suggesting a role of HPV38 in skin carcinogenesis (6). To examine this hypothesis, we have generated FVB/N Tg mouse lines expressing the HPV38 E6 and E7 genes under the control of the bovine promoter-enhancer region homologous to the human K10 promoter that is active in the suprabasal layer of the epidermis (2). The validity of this approach has been clearly demonstrated in previous studies on high-risk mucosal HPV types (9, 10). We found that HPV38 E6 and E7 expression in the mouse epidermis led to keratinocyte proliferation, hyperplasia, dysplasia, and loss of UV-induced cell cycle checkpoints. In addition, HPV38 E6/E7-Tg mice are more susceptible than non-Tg mice to the development of skin tumors upon exposure to chemical carcinogens. Together, these data confirm the transforming properties of HPV38 and further support its role in human carcinogenesis.
MATERIALS AND METHODS
Plasmid construction and generation of Tg mice. The E6 and E7 open reading frames of HPV38 were amplified by PCR using as template the entire HPV38 genome and were cloned in a pPolyIII vector in front of a 3.5-kb fragment of the bovine KVI (homolog of the human K10) promoter (2) using SmaI and EcoRI. An EcoRI-BamHI SV40 polyadenylation fragment was amplified and cloned after the E7 open reading frame. The complete insert, containing the K10 promoter, HPV38 E6/E7, and the simian virus 40 poly(A) was isolated as a NotI-NotI (4.5-kb) fragment (see Fig. 1A) and microinjected, at a concentration of 3 ng/μl, into the pronuclei of fertilized eggs to generate Tg mice as described previously (1). We generated two lines (2 and 6) of HPV38 E6/E7-Tg mice in a B6D2F1 genetic background. For the identification of Tg mice, DNA was extracted from the tails of 5-week-old mice and PCR performed using two different pairs of primers: (i) specific primers located in the K10 promoter (5'-CAT GTG GGA TAC ACC CTC-3') and poly(A) sequence (5'-AGA CAT GAT AAG ATA CAT TGA T-3') and (ii) primers located in the 5' (5'-ATG GAA CTA CCA AAA CCT CA-3') and 3'(5'-TTA TCG TCC GCC ATT GCG-3') regions of the E6 and E7 genes, respectively. The K10 HPV16 E6/E7-Tg mice were previously described (1). HPV16 E6/E7-positive animals were identified by PCR analysis using primers located in the 5' (5'-ATG CAC CAA AAG AGA ACT GCA A-3') and 3' (5'-TGG TTT CTG AGA ACA GAT GGG-3') regions of HPV16 E6 and E7, respectively. The quality of DNA extracted from the different mice was assessed by amplification of a region of the GAPDH (glyceraldehyde-3-phosphate dehydrogenase) gene by using the following primers: 5'-CTT CAC CAC CTT CTT GAT GT-3' (forward) and 5'-CTC ACT CAA GAT TGT CAG CA-3' (reverse). After establishment of the two HPV38 E6/E7-Tg lines, B6D2F1 mice were backcrossed for at least six generations into FVB/N genetic background. All experiments described in this study were performed with FVB/N Tg mice and in accordance with the guidelines of the IARC animal care and use committee.
Total RNA isolation and reverse transcription-PCR (RT-PCR) analysis. Total RNA was isolated from skin of dorsal area of six-week-old Tg animals using the EZ1 RNA isolation kit (QIAGEN, Courtaboeuf, France). cDNAs were synthesized from 1 μg of total RNA samples by reverse transcription using the first-strand cDNA synthesis kit (MBI Fermentas; Euromedex, Mundesheim, France), and PCR was performed using specific HPV38 E7 primers (5'-ATG ATT GGG AAA CAA GCT AC-3' and 5'-TGG TTT CTG AGA ACA GAT GGG-3'). PCR products were separated by electrophoresis through 2% agarose gels.
In situ hybridization. Radioactively labeled partially hydrolyzed riboprobes were generated from a clone containing full-length HPV38 E6/E7 cDNA. RNA-RNA in situ hybridization was performed as previously described (8). In brief, serial cryo-sections (5 μm) were mounted on 3-aminopropyl-triethoxy silane-coated slides, fixed in 4% paraformaldehyde in 2 x SSPE (1x SSPE is 0.18 M NaCl, 10 mM NaH2PO4, and 1 mM EDTA [pH 7.7]), digested with proteinase K (0.5 μg/ml), and prehybridized at 42°C for 2 to 4 h. Hybridization was performed overnight at 42°C in 50% formamide, 2 x SSPE, 10% dextran sulfate, 10 mM Tris-HCl, pH 7.5, 1x Denhardt's solution, 500 μg/ml tRNA, 100 μg/ml herring sperm DNA, 0.1% SDS, and 105 cpm/μl of radioactive probe. After hybridization, slides were washed once in 50% formamide, 2 x SSPE, and 0.1% sodium dodecyl sulfate (SDS) for 30 min at 50°C, treated with RNaseA (50 μg/ml in 2 x SSC and 0.1% SDS), and washed again in 50% formamide, 0.5 x SSPE, and 0.1% SDS for 30 min at 37°C. Slides were dehydrated in alcohol, dried, dipped in film emulsion (Kodak NTB 2 solution 1:1 with 600 mM ammonium acetate), and exposed for 14 days at 4°C.
Histological and immunohistochemical analysis. Tissue samples from six-week-old mice were fixed in 4% formaldehyde for 24 h at room temperature, transferred to phosphate-buffered saline and embedded in paraffin. Sections of 5-μm thickness were cut and either stained with hematoxylin/eosin or used for immunostaining using the following primary antibodies: anti-BrdU (1:50,000 dilution) (Sigma, Lyon, France) and anti-Ki-67 MM1 (1:200 dilution) (Novocastra, Newcastle, United Kingdom). Staining was performed using biotin-labeled goat anti-mouse immunoglobulin G or goat anti-rabbit immunoglobulin G (Vector Labs-ABCYS, Paris, France) and ABComplex AP amplification (Dako Cytomation SA, Trappes, France). The percentages of positive cells were determined by counting 400 to 500 hematoxylin-stained cells under 40x magnification in five to six different fields of epidermis. The two-sample t test (Student's t test) was used for statistical analysis of significance. A P value of <0.01 was considered to be significant.
UVB treatment and cell cycle analyses. Six-week-old animals were irradiated with 0.45 J/cm2 UVB (Bio-Spectra, Vilber Lourmat, Marne-La-Vallée, France). One hour after irradiation, the treated and control animals were injected intraperiteonally with BrdU (100 μg in 100 μl phosphate-buffered saline). Two hours after BrdU injection, the animals were sacrificed and skin in the ear region was isolated. For immunohistochemical analysis, a section of the ear was fixed in 10% phosphate-buffered saline-buffered formalin for 24 h and embedded in paraffin or snap-frozen in liquid nitrogen for DNA and RNA analysis. Immunohistochemistry was performed using the following antibodies: anti-BrdU (B2531; Sigma, Lyon, France) (dilution 1:50,000), anti-Ki-67 (NCL-Ki-67-MM1; Novocastra, Newcastle, United Kingdom) (dilution 1:250) and anti-p21WAF1 (sc-6246; Santa Cruz, Santa Cruz, CA) (dilution 1:200).
Preparation of protein extracts and immunoblot analysis. For the preparation of total protein extracts, mouse epidermis was homogenized in lysis buffer (20 mM Tris-HCl, pH 8.0, 200 mM NaCl, 0.5% Nonidet P40, 1 mM EDTA, 10 mM NaF, 0.1 mM Na3VO4, 1 mM phenylmethylsulfonyl fluoride [PMSF], 1 μg/ml leupeptin, and 1 μg/ml aprotinin) by using a high-speed homogenizer (Ultra-Turrax T25, Jankel and Kunkel, Illkirch, France). After centrifugation (12,000 x g, 15 min at 4°C), the supernatant was collected and protein concentration determined using BC assay reagent (UP95424A; Uptima Interchim, Montlu?on, France). Extracts were fractionated by electrophoresis on an SDS-polyacrylamide gel. Proteins were transferred onto a Polyscreen polyvinylidene difluoride membrane (NEN Life Sciences, Boston, MA) in a Trans-Blot SD semidry electrophoretic transfer cell (Bio-Rad, Marne le Coquette, France) (130 mA, 1 h 30 min). Immunoblot analyses were performed using the following antibodies: anti-?-tubulin (TUB2.1; Sigma, Lyon, France), anti-p21WAF1/CIP1 (p21-F5; SC-6246; Santa Cruz, Santa Cruz, CA), anti-Bak (SC-7873; Santa Cruz, Santa Cruz, CA).
Two-stage mouse skin carcinogenesis. Induction of tumor development was performed on 8-week-old Tg mice (8 mice HPV38 E6/E7-expressing line 2, 8 mice HPV38 E6/E7-expressing line 6, and 12 mice expressing HPV16 E6/E7) and their non-Tg littermates (13 mice). The dorsal skin of the mice was shaved one week before topical application of a single dose of the initiator 7,12-dimethylbenz[a]anthracene (DMBA) (catalog no. 39570; Fluka, Saint Quentin Fallavier, France) (400 nmol dissolved in 200 μl of acetone). Starting 10 days later, the dorsal skin was treated twice weekly with the tumor promoter 12-O-tetradecanoylphorbol 13-acetate (TPA) (P8139; Sigma, Lyon, France) (20 nmol in 200 μl of acetone) for 13 weeks. After 15 weeks of treatment, four mice of each group were killed for histological analysis of skin lesions, while the remaining mice were kept under observation for development of neoplastic lesions.
The number of papillomas was recorded individually for each mouse twice per week. Papillomas and all other lesions were examined histologically on hematoxylin-eosin-stained paraffin sections for detailed diagnosis. The formation of papillomas in each animal was monitored over time and the percentages of papilloma-bearing animals in each group were plotted as a Kaplan-Meier graph (GraphPad Software Inc., San Diego, CA). We also determined the average number of papillomas per animal in each group by counting all papillomas in non-Tg or Tg mice and dividing them by the number of animals in each group. The size of the tumors was measured at different time points and plotted using GraphPad Prism 4 (GraphPad Software Inc., San Diego, CA). t test or log rank test for group data was used to determine whether the difference in data obtained with non-Tg and Tg mice was statistically significant.
RESULTS
Generation of HPV38 E6/E7-Tg mice. To characterize the biological properties of the HPV38 E6 and E7 oncoproteins in vivo, we generated Tg mouse lines expressing the two viral genes in skin keratinocytes. We used the bovine KVI promoter-enhancer region homologous to the human keratin 10 (K10) promoter that has been used previously for expression of HPV16 E6 and E7 in the suprabasal layers of the epidermis (1). Tg mice with expression of HPV38 E6 and E7 in the skin were made by microinjection of the construct shown in Fig. 1A into the pronuclei of one-cell mouse embryos. Transgene-positive offspring were identified by PCR using two different pairs of primers, as described in Materials and Methods. Two independent Tg lines (2 and 6) of animals were identified. Expression of the HPV38 E6 and E7 genes was determined in each line by semiquantitative RT-PCR. As shown in Fig. 1B, line 6 expressed the HPV38 E6 and E7 genes at higher levels than line 2. Next, we determined the location of the viral transcripts in the epidermis of the Tg animals using in situ hybridization. Sections of ear tissue of non-Tg and Tg mice were hybridized with HPV38 E6/E7 32P-antisense or sense RNA. A strong positive signal in the epidermis of HPV38 E6/E7-Tg mice was observed with antisense RNA, but not with sense probe (Fig. 1C).
HPV38 E6 and E7 induce cellular proliferation in the epidermis of Tg mice. To detect possible morphological alterations in the epidermis of the Tg mice, we performed histological analysis of skin specimens from mice of the two HPV38 E6/E7-Tg lines. For comparison, we included in the analysis HPV16 E6/E7-Tg mice that have a marked and diffuse increase in skin thickness (1). Most of the epidermis in sections from the ears of 6-week-old HPV38 E6/E7-Tg mice (lines 2 and 6) displayed no apparent morphological alteration (Fig. 2A and data not shown). However, hyperproliferative patches of epithelium (at least two for each section) were clearly detected in some areas of the HPV38 E6/E7-Tg mice but never in non-Tg animals (Fig. 2 and data not shown). The formation of these hyperproliferative patches may be due to continuous scratching of the skin that generate microlesions, healing, and irritation. Two representative skin sections, showing hyperplasia and dysplasia, from HPV38 E6/E7-Tg mice are shown in Fig. 2B. Skin epidermal hyperplasia was defined as a lesion containing several layers of basal cells that retain the ability to differentiate (Fig. 2B, middle panel). In addition, dysplasia was described as a lesion containing progressive increase in the number of proliferative basal and suprabasal cells throughout the thickness of the epidermis with nuclear and cytoplasmic abnormalities and occasionally mitosis (Fig. 2B, right panel).
The bromodeoxyuridine (BrdU) incorporation assay and Ki-67 immunohistochemistry showed increased proliferation in the epidermis of HPV38 E6/E7-Tg mice (Fig. 3A). Scoring of BrdU- and Ki-67-positive cells in different randomly selected skin areas showed that the percentage of proliferative keratinocytes was increased approximately twofold in the skin of HPV38 E6/E7-Tg mice compared with non-Tg mice (Fig. 3B). In addition, in situ hybridization using HPV38 E6/E7 32P-antisense RNA probe showed the E6/E7 transcript to be expressed throughout the hyperproliferative patches observed in the epidermis of the HPV38 E6/E7-Tg mice (Fig. 3C). In agreement with this observation, keratinocytes positive for BrdU and Ki-67 were detected in the basal and suprabasal layers of these lesions (Fig. 3D).
Together, these data show that inducing HPV38 E6 and E7 gene expression in mouse epidermis results in keratinocyte hyperproliferation.
UVB-induced cell cycle checkpoints are altered in keratinocytes of HPV38 E6/E7-Tg mice. Exposure of cells to UVB radiation leads to DNA damage that in turn induces cell cycle arrest or apoptosis. Since HPV DNA replication is entirely dependent on the proliferative status of an infected keratinocyte, it is likely that cutaneous HPV types have developed a mechanism to circumvent the negative cell cycle regulation induced by UV. To test this hypothesis, we determined whether HPV38 E6 and E7 can prevent cell cycle arrest induced by UVB irradiation. Non-Tg and HPV38 E6/E7-Tg mice were irradiated with UVB, and the proliferative status of skin keratinocytes was evaluated by BrdU incorporation and Ki-67 immunostaining. Figure 4A and B show that UVB irradiation drastically reduced the number of proliferating keratinocytes in the skin of non-Tg mice. In contrast, the effects of UV were clearly less severe in the epidermis of Tg mice, in which a high percentage of cells were still positive for BrdU or Ki-67 after exposure to UVB (Fig. 4). Immunohistochemical staining for the cell cycle inhibitor p21WAF1 showed that in non-Tg mice, UV irradiation leads to accumulation of this protein in skin keratinocytes, but no accumulation was observed in the two lines of HPV38 E6/E7-Tg mice (Fig. 5A and B). Immunoblot analysis of skin protein extracts confirmed the immunohistochemical data, showing no increase in p21WAF1 levels in epidermis of Tg mice after UV irradiation. In addition, we determined the levels of the proapoptotic protein Bak that is highly stabilized in response to UV irradiation and degraded by E6 from EV HPV type 5 (15). As shown in Fig. 5C, no accumulation of Bak occurred in the skin of HPV38 E6 and E7 Tg mice after UV exposure. Together, these data show that HPV38 E6 and E7 interfere with cellular defense processes induced by UV irradiation.
HPV38 E6 and E7 act synergistically with chemical carcinogens in the development of skin tumors. In contrast to mucosal HPV16 E6/E7-Tg or EV HPV8-Tg mice (18, 29, 31), HPV38 E6/E7 animals did not develop spontaneous tumors or any skin lesions during their life span of approximately 2 to 2.5 years. Therefore, we checked whether HPV38 E6 and E7 cooperate with chemical carcinogens to induce skin tumors in our Tg mouse model, as previously shown for other papillomaviruses (12, 30). Non-Tg (FVB/N, 13 animals), HPV16 E6/E7-Tg (12 animals), and HPV38 E6/E7-Tg (8 for each line) mice were exposed to a two-stage carcinogen treatment with DMBA and TPA (Fig. 6A). After initiation of the chemical treatment, animals were monitored every 3 days over a period of 20 weeks. At week 15, four mice of each group were sacrificed for histological analysis of skin lesions, while the remaining mice were kept under observation for development of skin carcinomas.
In the first 15 weeks, HPV38 E6/E7-Tg mice rapidly developed numerous papillomas, in contrast to the non-Tg mice (Fig. 6B). Skin lesions were detected in HPV38 E6/E7-Tg line 6 mice approximately 7 weeks after the beginning of the carcinogenic treatment, and 1 week later in HPV38 E6/E7-expressing line 2 mice and HPV16 E6/E7-expressing mice (Fig. 6C), indicating a dose-dependent effect of HPV38 oncoprotein expression. At week 9, 100% of the Tg animals had developed papillomas, while non-Tg mice did not develop any skin lesion until week 9 (Fig. 6C). The average number of tumors per animal and tumor size were clearly greater in Tg mice than non-Tg mice (Fig. 6D and E). After four mice of each group were sacrificed at week 15, we performed histological analysis on 44 lesions for each HPV38 line, 55 lesions of HPV16 Tg mice, and 21 lesions of non-Tg mice. The frequencies of papillomas and keratoacanthomas were similar in HPV38 E6/E7-Tg mice (lines 2 and 6) and HPV16 E6/E7-Tg mice (data not shown). We observed in the papillomas hyperkeratinized and hyperproliferative areas (Fig. 7, top panels), while keratoacanthomas showed the typical keratin pearls together with hyperproliferation (Fig. 7, middle panels). At weeks 17 and 18, two out of four HPV38 E6/E7-expressing mice of line 2 and one out of four HPV38 E6/E7-expressing mice of line 6 had developed SCC (Fig. 7, bottom panels and data not shown). No SCCs were observed in non-Tg or HPV16 E6/E7-Tg mice after 20 weeks of DMBA/TPA treatment. These results show that HPV38 E6 and E7 are able to cooperate with chemical carcinogens to induce skin tumors.
DISCUSSION
Although the mucosal high-risk HPV types have been extensively characterized and their role in human carcinogenesis has been proven, very little is known about the cutaneous HPV types. In this study, we generated two lines of Tg mice expressing E6 and E7 of the EV HPV38 under the control of the bovine homolog of the human K10 promoter. We have previously shown that both proteins are able to corrupt the cell cycle and senescence programs in primary keratinocytes, leading to immortalization (6; unpublished data). In agreement with our in vitro data, expression of HPV38 E6 and E7 in mouse epidermis resulted in alteration of cell cycle regulation. Although most of the skin did not present any morphological alteration, patches with increased epithelial thickness were detected in both HPV38 E6/E7-Tg lines. Markers of cellular proliferation were detected in the basal and upper layers of the epidermis in these hyperplastic/dysplastic regions. In addition, in epidermis with no morphological alteration, we observed increased basal cell proliferation in the HPV38 E6/E7-Tg mice. A similar phenomenon was also observed in HPV16 E6/E7-Tg mice, in which E6 and E7 expression was under the control of the same promoter as in the HPV38 E6/E7-expressing animals (1). In this previous study, it was shown that production of autocrine factors induced by HPV16 E6 and E7 was responsible for the stimulation of proliferation of basal cells (1). We are currently investigating in HPV38 E6/E7-expressing mice the possible role of autocrine factors in this event. However, it is also possible that the bovine homolog of the human K10 promoter is weakly active in the basal layer. In fact, in situ hybridization staining showed a positive signal for HPV38 E6 and E7 mRNA also in cells attached to the basal membrane.
In contrast to HPV38 E6/E7-expressing animals, HPV16 E6/E7-Tg mice have a diffuse increase of epidermis thickness. Thus, HPV38 E6 and E7 proteins appear to be less efficient than HPV16 E6 and E7 in promoting morphological alterations of the epidermis when expressed by the K10 promoter. Differences between HPV16 and HPV38 in the intrinsic properties of E6 and E7 are most likely responsible for the difference in phenotype between the two Tg mouse lines. We have previously shown that E7 proteins of HPV16 and HPV38 in vitro have similar efficiency in degrading pRb and deregulating the G1/S transition (6). Thus, other E6 and/or E7 activities may be responsible for the difference observed in the skin of HPV16 E6/E7-Tg and HPV38 E6/E7-Tg mice. HPV16 E6, but not HPV38 E6, has a PDZ-binding motif at the C terminus, which mediates interaction with several PDZ partners, including hDLG, hSCRIBBLE, MUPP1, and MAGI (23). Interestingly, Tg mice expressing an HPV16 E6 mutant that lacks the PDZ-binding motif did not show any epithelial hyperplasia, in contrast to mice expressing wild-type HPV16 E6 (26). Although HPV38 E6 and E7 are not as efficient as HPV16 E7 and E6 in inducing morphological alteration, they appear to cooperate better with chemical carcinogens in the induction of skin cancers. In fact, while E6 and E7 from both HPV types showed similar efficiencies in promoting the formation of keratoacanthomas and papillomas, only HPV38 E6/E7 animals (lines 2 and 6) developed SCC after 20 weeks of DMBA treatment.
UV light is a key risk factor for the development of skin cancer, inducing accumulation of DNA damage including C to T transitions at dipyrimidine sequences (19). This event normally activates cellular defense processes leading to p53 activation that in turn induces cell cycle arrest or apoptosis to allow repair or elimination of the damaged cells, respectively. HPV38 E6 and E7 enable cells to overcome the cell cycle arrest induced by UVB irradiation; we detected only a small reduction in numbers of epidermal proliferative cells after UVB exposure of HPV38 E6/E7-Tg mice. In contrast, in non-Tg mice, UVB irradiation results in almost complete cell cycle arrest. Expression of the cell cycle inhibitor p21WAF1 is under the control of p53 and plays a key role in UV-induced cell cycle arrest. No p21WAF1 accumulation was observed in the epidermis of HPV38 E6/E7-Tg mice upon UV irradiation. HPV38 E6 does not promote p53 degradation (5), but we have recently observed in primary human keratinocytes and Tg mice that expression of HPV38 E6 and E7 results in an alteration of p53 transcriptional functions (R. Accardi and M. Tommasino, unpublished data). Thus, p21WAF1 down-regulation in HPV38 E6/E7-Tg mice may be due to p53 inactivation. Alternatively, HPV38 E6 can repress p21WAF1 transcription by altering p53-independent pathways, as shown for HPV16 E6 (21, 22).
Cutaneous HPV types (both EV and non-EV) are able to antagonize UV-induced apoptosis by promoting the ubiquitin-mediated degradation of Bak, a proapoptotic protein that is stabilized in response to UV irradiation (15). We did not detect Bak accumulation in the epidermis of our HPV38 E6/E7-Tg mice after UV irradiation, showing that this property is shared by HPV38. Thus, HPV38 E6 and E7 oncoproteins appear to have several properties that allow infected epidermal keratinocytes to circumvent the adverse effects of UV radiation. Such activity would be essential to guarantee viral DNA replication in sun-exposed cells and, as a side effect, might favor the accumulation of DNA damage and so facilitate the development of skin cancer. Although we do not yet know whether UV and HPV38 act synergistically in our animal model, the cooperation between HPV38 E6 and E7 and DMBA-induced DNA damage in tumor development support this hypothesis.
Studies on human specimens indicated that the presence of the viral genome may not be required in the later stages of carcinogenesis (34). This may imply a need to consider a new scenario for the role of EV HPV in NMSC pathogenesis, such as a "hit-and-run" mechanism. This mode of action is clearly different from that established for mucosal high-risk HPV in cervical cancer, where E6 and E7 expression is constantly required for maintenance of the neoplastic phenotype of the infected cell.
A recent study on EV HPV type 8 has shown that Tg mice expressing the entire early region under control of the K14 promoter developed spontaneous benign tumors and, in a small percentage, SCC (29). We did not observe any formation of spontaneous tumors in our HPV38 E6/E7-Tg mice during their life span. The difference in phenotype between our two Tg models may have various explanations: (i) E6 and E7 of the two HPV types may display different transforming activities, (ii) E2 that is expressed in HPV8 Tg mice may cooperate with E6 and E7 in tumor development, or (iii) the expression of viral proteins by different promoters in different layers of the epidermis may influence tumor development. Further studies are required to address this issue.
In summary, we present here data that confirm the transforming activity of HPV38 in an in vivo model. The fact that skin keratinocytes of HPV38 E6/E7-Tg mice are unable to respond to cellular stresses strongly suggests that viral infection and UV irradiation may act synergistically in the induction of NMSC.
ACKNOWLEDGMENTS
We are grateful to all members of our laboratory for their cooperation, Martyn Plummer for statistical analysis, Dominique Galendo and Marie-Pierre Cros for animal colony maintenance, Nicole Lyandrat and Christine Carreira for histological service, and John Cheney for critical reading of the manuscript.
REFERENCES
Auewarakul, P., L. Gissmann, and A. Cidarregui. 1994. Targeted expression of the E6 and E7 oncogenes of human papillomavirus type 16 in the epidermis of transgenic mice elicits generalized epidermal hyperplasia involving autocrine factors. Mol. Cell. Biol. 14:8250-8258.
Blessing, M., J. L. Jorcano, and W. W. Franke. 1989. Enhancer elements directing cell-type-specific expression of cytokeratin genes and changes of the epithelial cytoskeleton by transfections of hybrid cytokeratin genes. EMBO J. 8:117-126.
Bowden, G. T. 2004. Prevention of non-melanoma skin cancer by targeting ultraviolet-B-light signalling. Nat. Rev. Cancer 4:23-35.
Boyle, J., R. M. MacKie, J. D. Briggs, B. J. Junor, and T. C. Aitchison. 1984. Cancer warts, and sunshine in renal transplant patients. A case-control study. Lancet i:702-705.
Caldeira, S., R. Filotico, R. Accardi, I. Zehbe, S. Franceschi, and M. Tommasino. 2004. p53 mutations are common in human papillomavirus type 38-positive non-melanoma skin cancers. Cancer Lett. 209:119-124.
Caldeira, S., I. Zehbe, R. Accardi, I. Malanchi, W. Dong, M. Giarre, E.-M. deVilliers, R. Filotico, P. Boukamp, and M. Tommasino. 2003. The E6 and E7 proteins of cutaneous human papillomavirus type 38 display transforming properties. J. Virol. 77:2195-2206.
de Villiers, E. M., C. Fauquet, T. R. Broker, H. U. Bernard, and H. zur Hausen. 2004. Classification of papillomaviruses. Virology 324:17-27.
Durst, M., D. Glitz, A. Schneider, and H. zur Hausen. 1992. Human papillomavirus type 16 (HPV 16) gene expression and DNA replication in cervical neoplasia: analysis by in situ hybridization. Virology 189:132-140.
Eckert, R. L., J. F. Crish, S. Balasubramanian, and E. A. Rorke. 2000. Transgenic animal models of human papillomavirus-dependent disease. Int. J. Oncol. 16:853-870.
Griep, A. E., and P. F. Lambert. 1994. Role of papillomavirus oncogenes in human cervical cancer: transgenic animal studies. Proc. Soc. Exp. Biol. Med. 206:24-34.
Hartevelt, M. M., J. N. Bavinck, A. M. Kootte, B. J. Vermeer, and J. P. Vandenbroucke. 1990. Incidence of skin cancer after renal transplantation in The Netherlands. Transplantation 49:506-509.
Helfrich, I., M. Chen, R. Schmidt, G. Furstenberger, A. Kopp-Schneider, D. Trick, H. J. Grone, H. Zur Hausen, and F. Rosl. 2004. Increased incidence of squamous cell carcinomas in Mastomys natalensis papillomavirus E6 transgenic mice during two-stage skin carcinogenesis. J. Virol. 78:4797-4805.
Iftner, T., S. Bierfelder, Z. Csapo, and H. Pfister. 1988. Involvement of human papillomavirus type 8 genes E6 and E7 in transformation and replication. J. Virol. 62:3655-3661.
Jablonska, S., J. Dabrowski, and K. Jakubowicz. 1972. Epidermodysplasia verruciformis as a model in studies on the role of papovaviruses in oncogenesis. Cancer Res. 32:583-589.
Jackson, S., C. Harwood, M. Thomas, L. Banks, and A. Storey. 2000. Role of Bak in UV-induced apoptosis in skin cancer and abrogation by HPV E6 proteins. Genes Dev. 14:3065-3073.
Kiviat, N. B. 1999. Papillomaviruses in non-melanoma skin cancer: epidemiological aspects. Semin. Cancer Biol. 9:397-403.
Kreimer, A. R., G. M. Clifford, P. Boyle, and S. Franceschi. 2005. Human papillomavirus types in head and neck squamous cell carcinomas worldwide: a systematic review. Cancer Epidemiol. Biomarkers Prev. 14:467-475.
Lambert, P. F., H. Pan, H. C. Pitot, A. Liem, M. Jackson, and A. E. Griep. 1993. Epidermal cancer associated with expression of human papillomavirus type 16 E6 and E7 oncogenes in the skin of transgenic mice. Proc. Natl. Acad. Sci. USA 90:5583-5587.
Lee, D. H., and G. P. Pfeifer. 2003. Deamination of 5-methylcytosines within cyclobutane pyrimidine dimers is an important component of UVB mutagenesis. J. Biol. Chem. 278:10314-10321.
Majewski, S., S. Jablonska, and G. Orth. 1997. Epidermodysplasia verruciformis. Immunological and nonimmunological surveillance mechanisms: role in tumor progression. Clin. Dermatol. 15:321-334.
Malanchi, I., R. Accardi, F. Diehl, A. Smet, E. Androphy, J. Hoheisel, and M. Tommasino. 2004. Human papillomavirus type 16 E6 promotes retinoblastoma protein phosphorylation and cell cycle progression. J. Virol. 78:13769-13778.
Malanchi, I., S. Caldeira, M. Krutzfeldt, M. Giarre, M. Alunni-Fabbroni, and M. Tommasino. 2002. Identification of a novel activity of human papillomavirus type 16 E6 protein in deregulating the G1/S transition. Oncogene 21:5665-5672.
Mantovani, F., and L. Banks. 2001. The human papillomavirus E6 protein and its contribution to malignant progression. Oncogene 20:7874-7887.
Munger, K., A. Baldwin, K. M. Edwards, H. Hayakawa, C. L. Nguyen, M. Owens, M. Grace, and K. Huh. 2004. Mechanisms of human papillomavirus-induced oncogenesis. J. Virol. 78:11451-11460.
Munger, K., J. R. Basile, S. Duensing, A. Eichten, S. L. Gonzalez, M. Grace, and V. L. Zacny. 2001. Biological activities and molecular targets of the human papillomavirus E7 oncoprotein. Oncogene 20:7888-7898.
Nguyen, M. L., M. M. Nguyen, D. Lee, A. E. Griep, and P. F. Lambert. 2003. The PDZ ligand domain of the human papillomavirus type 16 E6 protein is required for E6's induction of epithelial hyperplasia in vivo. J. Virol. 77:6957-6964.
Orth, G., S. Jablonska, M. Jarzabek-Chorzelska, S. Obalek, G. Rzesa, M. Favre, and O. Croissant. 1979. Characteristics of the lesions and risk of malignant conversion associated with the type of human papillomavirus involved in epidermodysplasia verruciformis. Cancer Res. 39:1074-1082.
Pfister, H. 2003. Chapter 8: human papillomavirus and skin cancer. J. Natl. Cancer Inst. Monogr. 31:52-56.
Schaper, I. D., G. P. Marcuzzi, S. J. Weissenborn, H. U. Kasper, V. Dries, N. Smyth, P. Fuchs, and H. Pfister. 2005. Development of skin tumors in mice transgenic for early genes of human papillomavirus type 8. Cancer Res. 65:1394-1400.
Song, S., A. Liem, J. A. Miller, and P. F. Lambert. 2000. Human papillomavirus types 16 E6 and E7 contribute differently to carcinogenesis. Virology 267:141-150.
Song, S., H. C. Pitot, and P. F. Lambert. 1999. The human papillomavirus type 16 E6 gene alone is sufficient to induce carcinomas in transgenic animals. J. Virol. 73:5887-5893.
Walder, B. K., D. Jeremy, J. A. Charlesworth, G. J. Macdonald, B. A. Pussell, and M. R. Robertson. 1976. The skin and immunosuppression. Australas. J. Dermatol. 17:94-97.
Walder, B. K., M. R. Robertson, and D. Jeremy. 1971. Skin cancer and immunosuppression. Lancet ii:1282-1283.
Weissenborn, S. J., I. Nindl, K. Purdie, C. Harwood, C. Proby, J. Breuer, S. Majewski, H. Pfister, and U. Wieland. 2005. Human papillomavirus-DNA loads in actinic keratoses exceed those in non-melanoma skin cancers. J. Investig. Dermatol. 125:93-97.
Yamashita, T., K. Segawa, Y. Fujinaga, T. Nishikawa, and K. Fujinaga. 1993. Biological and biochemical activity of E7 genes of the cutaneous human papillomavirus type 5 and 8. Oncogene 8:2433-2441.
zur Hausen, H. 2002. Papillomaviruses and cancer: from basic studies to clinical application. Nat. Rev. Cancer 2:342-350.(Wen Dong, Ulrich Kloz, Ro)
Deutsches Krebsforschungszentrum, D-69120 Heidelberg, Germany
Gyn?kologische Molekularbiologie Frauenklinik der FSU Jena, Bachstr. 18, D-07743 Jena, Germany
ABSTRACT
The oncoproteins E6 and E7 of human papillomavirus type 38 (HPV38) display several transforming activities in vitro, including immortalization of primary human keratinocytes. To evaluate the oncogenic activities of the viral proteins in an in vivo model, we generated transgenic mice expressing HPV38 E6 and E7 under the control of the bovine homologue of the human keratin 10 (K10) promoter. Two distinct lines of HPV38 E6/E7-expressing transgenic mice that express the viral genes at different levels were obtained. In both lines, HPV38 E6 and E7 induced cellular proliferation, hyperplasia, and dysplasia in the epidermis. The rate of occurrence of these events was proportional to the levels of HPV38 E6 and E7 expression in the two transgenic lines. Exposure of the epidermis of nontransgenic mice to UV led to p21WAF1 accumulation and cell cycle arrest. In contrast, keratinocytes from transgenic mice continued to proliferate and were not positive for p21WAF1, indicating that cell cycle checkpoints are altered in keratinocytes expressing the viral genes. Although the HPV38 E6/E7-expressing transgenic mice did not develop spontaneous tumors during their life span, two-stage carcinogen treatment led to a high incidence of papillomas, keratoacanthomas, and squamous-cell carcinomas in HPV38 mice compared with nontransgenic animals. Together, these data show that HPV38 E6 and E7 display transforming properties in vivo, providing further support for the role of HPV38 in carcinogenesis.
INTRODUCTION
Nonmelanoma skin cancers (NMSC), basal-cell carcinoma (BCC), and squamous-cell carcinoma (SCC) are the most common malignancies occurring worldwide. Several findings from experimental systems in vitro and in vivo have shown that UV plays a direct role in skin carcinogenesis (3). In addition, a subgroup of the epitheliotropic human papillomaviruses may cooperate with UV in the development of NMSC (28). Individuals with an autosomal recessive disorder termed epidermodysplasia verruciformis (EV) are susceptible to infection by specific HPV types that belong to the genotype beta branch of the HPV phylogenetic tree (7). Individuals with EV develop confluent flat warts that, in approximately 30% of cases, progress to SCC on sun-exposed areas (11, 14, 20, 27). The involvement of an infective agent in skin carcinogenesis is supported by the fact that immunosuppressed organ transplant recipients have a 50- to 100-fold higher risk of developing NMSC than the general population (4, 16, 32, 33). In addition, independent studies have reported that DNA of cutaneous HPV types can be detected in SCC and BCC of immunocompetent individuals, suggesting their role in skin carcinogenesis in the general population (28). The potential carcinogenic role of the EV HPV types is consistent with the fact that other members of the papillomavirus family, i.e., the mucosal high-risk types, are associated with anogenital cancers (36) and a subset of head and neck cancers (17).
The products of two early genes, E6 and E7, are the major oncoproteins of the high-risk mucosal HPV types (e.g., HPV16 and HPV18) (23, 24). Both viral proteins are able to alter the regulation of the cell cycle and apoptosis by binding and inactivating several cellular proteins, including products of tumor-suppressor genes. HPV16 E6 binds p53 and promotes its degradation, mediating the interaction of p53 with the ubiquitin protein ligase E6AP (23). Similarly, HPV16 E7 interacts with the retinoblastoma protein (pRb), a negative cell cycle regulator, inducing its degradation via the ubiquitin pathway (25). In contrast to the E6 and E7 of the mucosal HPV types, very little is known about the oncoproteins of the EV HPV types. Initial studies on HPV5 and HPV8, the types most frequently detected in EV patients, showed lower in vitro transforming activity of their E6 and E7 compared with the oncoproteins of the high-risk mucosal HPV types (13, 35). However, in a recent study, transgenic (Tg) mice expressing the early region of HPV8 under the K14 promoter spontaneously developed multifocal skin tumors and, in 6% of cases, SCC (29).
We have recently identified an EV HPV type (HPV38) that displays in vitro transforming properties (6). HPV38 E7 inactivates pRb and induces loss of G1/S transition control with efficiency similar to HPV16 E7. In addition, HPV38 E6 and E7 can immortalize primary human keratinocytes that are naturally infected by the virus, suggesting a role of HPV38 in skin carcinogenesis (6). To examine this hypothesis, we have generated FVB/N Tg mouse lines expressing the HPV38 E6 and E7 genes under the control of the bovine promoter-enhancer region homologous to the human K10 promoter that is active in the suprabasal layer of the epidermis (2). The validity of this approach has been clearly demonstrated in previous studies on high-risk mucosal HPV types (9, 10). We found that HPV38 E6 and E7 expression in the mouse epidermis led to keratinocyte proliferation, hyperplasia, dysplasia, and loss of UV-induced cell cycle checkpoints. In addition, HPV38 E6/E7-Tg mice are more susceptible than non-Tg mice to the development of skin tumors upon exposure to chemical carcinogens. Together, these data confirm the transforming properties of HPV38 and further support its role in human carcinogenesis.
MATERIALS AND METHODS
Plasmid construction and generation of Tg mice. The E6 and E7 open reading frames of HPV38 were amplified by PCR using as template the entire HPV38 genome and were cloned in a pPolyIII vector in front of a 3.5-kb fragment of the bovine KVI (homolog of the human K10) promoter (2) using SmaI and EcoRI. An EcoRI-BamHI SV40 polyadenylation fragment was amplified and cloned after the E7 open reading frame. The complete insert, containing the K10 promoter, HPV38 E6/E7, and the simian virus 40 poly(A) was isolated as a NotI-NotI (4.5-kb) fragment (see Fig. 1A) and microinjected, at a concentration of 3 ng/μl, into the pronuclei of fertilized eggs to generate Tg mice as described previously (1). We generated two lines (2 and 6) of HPV38 E6/E7-Tg mice in a B6D2F1 genetic background. For the identification of Tg mice, DNA was extracted from the tails of 5-week-old mice and PCR performed using two different pairs of primers: (i) specific primers located in the K10 promoter (5'-CAT GTG GGA TAC ACC CTC-3') and poly(A) sequence (5'-AGA CAT GAT AAG ATA CAT TGA T-3') and (ii) primers located in the 5' (5'-ATG GAA CTA CCA AAA CCT CA-3') and 3'(5'-TTA TCG TCC GCC ATT GCG-3') regions of the E6 and E7 genes, respectively. The K10 HPV16 E6/E7-Tg mice were previously described (1). HPV16 E6/E7-positive animals were identified by PCR analysis using primers located in the 5' (5'-ATG CAC CAA AAG AGA ACT GCA A-3') and 3' (5'-TGG TTT CTG AGA ACA GAT GGG-3') regions of HPV16 E6 and E7, respectively. The quality of DNA extracted from the different mice was assessed by amplification of a region of the GAPDH (glyceraldehyde-3-phosphate dehydrogenase) gene by using the following primers: 5'-CTT CAC CAC CTT CTT GAT GT-3' (forward) and 5'-CTC ACT CAA GAT TGT CAG CA-3' (reverse). After establishment of the two HPV38 E6/E7-Tg lines, B6D2F1 mice were backcrossed for at least six generations into FVB/N genetic background. All experiments described in this study were performed with FVB/N Tg mice and in accordance with the guidelines of the IARC animal care and use committee.
Total RNA isolation and reverse transcription-PCR (RT-PCR) analysis. Total RNA was isolated from skin of dorsal area of six-week-old Tg animals using the EZ1 RNA isolation kit (QIAGEN, Courtaboeuf, France). cDNAs were synthesized from 1 μg of total RNA samples by reverse transcription using the first-strand cDNA synthesis kit (MBI Fermentas; Euromedex, Mundesheim, France), and PCR was performed using specific HPV38 E7 primers (5'-ATG ATT GGG AAA CAA GCT AC-3' and 5'-TGG TTT CTG AGA ACA GAT GGG-3'). PCR products were separated by electrophoresis through 2% agarose gels.
In situ hybridization. Radioactively labeled partially hydrolyzed riboprobes were generated from a clone containing full-length HPV38 E6/E7 cDNA. RNA-RNA in situ hybridization was performed as previously described (8). In brief, serial cryo-sections (5 μm) were mounted on 3-aminopropyl-triethoxy silane-coated slides, fixed in 4% paraformaldehyde in 2 x SSPE (1x SSPE is 0.18 M NaCl, 10 mM NaH2PO4, and 1 mM EDTA [pH 7.7]), digested with proteinase K (0.5 μg/ml), and prehybridized at 42°C for 2 to 4 h. Hybridization was performed overnight at 42°C in 50% formamide, 2 x SSPE, 10% dextran sulfate, 10 mM Tris-HCl, pH 7.5, 1x Denhardt's solution, 500 μg/ml tRNA, 100 μg/ml herring sperm DNA, 0.1% SDS, and 105 cpm/μl of radioactive probe. After hybridization, slides were washed once in 50% formamide, 2 x SSPE, and 0.1% sodium dodecyl sulfate (SDS) for 30 min at 50°C, treated with RNaseA (50 μg/ml in 2 x SSC and 0.1% SDS), and washed again in 50% formamide, 0.5 x SSPE, and 0.1% SDS for 30 min at 37°C. Slides were dehydrated in alcohol, dried, dipped in film emulsion (Kodak NTB 2 solution 1:1 with 600 mM ammonium acetate), and exposed for 14 days at 4°C.
Histological and immunohistochemical analysis. Tissue samples from six-week-old mice were fixed in 4% formaldehyde for 24 h at room temperature, transferred to phosphate-buffered saline and embedded in paraffin. Sections of 5-μm thickness were cut and either stained with hematoxylin/eosin or used for immunostaining using the following primary antibodies: anti-BrdU (1:50,000 dilution) (Sigma, Lyon, France) and anti-Ki-67 MM1 (1:200 dilution) (Novocastra, Newcastle, United Kingdom). Staining was performed using biotin-labeled goat anti-mouse immunoglobulin G or goat anti-rabbit immunoglobulin G (Vector Labs-ABCYS, Paris, France) and ABComplex AP amplification (Dako Cytomation SA, Trappes, France). The percentages of positive cells were determined by counting 400 to 500 hematoxylin-stained cells under 40x magnification in five to six different fields of epidermis. The two-sample t test (Student's t test) was used for statistical analysis of significance. A P value of <0.01 was considered to be significant.
UVB treatment and cell cycle analyses. Six-week-old animals were irradiated with 0.45 J/cm2 UVB (Bio-Spectra, Vilber Lourmat, Marne-La-Vallée, France). One hour after irradiation, the treated and control animals were injected intraperiteonally with BrdU (100 μg in 100 μl phosphate-buffered saline). Two hours after BrdU injection, the animals were sacrificed and skin in the ear region was isolated. For immunohistochemical analysis, a section of the ear was fixed in 10% phosphate-buffered saline-buffered formalin for 24 h and embedded in paraffin or snap-frozen in liquid nitrogen for DNA and RNA analysis. Immunohistochemistry was performed using the following antibodies: anti-BrdU (B2531; Sigma, Lyon, France) (dilution 1:50,000), anti-Ki-67 (NCL-Ki-67-MM1; Novocastra, Newcastle, United Kingdom) (dilution 1:250) and anti-p21WAF1 (sc-6246; Santa Cruz, Santa Cruz, CA) (dilution 1:200).
Preparation of protein extracts and immunoblot analysis. For the preparation of total protein extracts, mouse epidermis was homogenized in lysis buffer (20 mM Tris-HCl, pH 8.0, 200 mM NaCl, 0.5% Nonidet P40, 1 mM EDTA, 10 mM NaF, 0.1 mM Na3VO4, 1 mM phenylmethylsulfonyl fluoride [PMSF], 1 μg/ml leupeptin, and 1 μg/ml aprotinin) by using a high-speed homogenizer (Ultra-Turrax T25, Jankel and Kunkel, Illkirch, France). After centrifugation (12,000 x g, 15 min at 4°C), the supernatant was collected and protein concentration determined using BC assay reagent (UP95424A; Uptima Interchim, Montlu?on, France). Extracts were fractionated by electrophoresis on an SDS-polyacrylamide gel. Proteins were transferred onto a Polyscreen polyvinylidene difluoride membrane (NEN Life Sciences, Boston, MA) in a Trans-Blot SD semidry electrophoretic transfer cell (Bio-Rad, Marne le Coquette, France) (130 mA, 1 h 30 min). Immunoblot analyses were performed using the following antibodies: anti-?-tubulin (TUB2.1; Sigma, Lyon, France), anti-p21WAF1/CIP1 (p21-F5; SC-6246; Santa Cruz, Santa Cruz, CA), anti-Bak (SC-7873; Santa Cruz, Santa Cruz, CA).
Two-stage mouse skin carcinogenesis. Induction of tumor development was performed on 8-week-old Tg mice (8 mice HPV38 E6/E7-expressing line 2, 8 mice HPV38 E6/E7-expressing line 6, and 12 mice expressing HPV16 E6/E7) and their non-Tg littermates (13 mice). The dorsal skin of the mice was shaved one week before topical application of a single dose of the initiator 7,12-dimethylbenz[a]anthracene (DMBA) (catalog no. 39570; Fluka, Saint Quentin Fallavier, France) (400 nmol dissolved in 200 μl of acetone). Starting 10 days later, the dorsal skin was treated twice weekly with the tumor promoter 12-O-tetradecanoylphorbol 13-acetate (TPA) (P8139; Sigma, Lyon, France) (20 nmol in 200 μl of acetone) for 13 weeks. After 15 weeks of treatment, four mice of each group were killed for histological analysis of skin lesions, while the remaining mice were kept under observation for development of neoplastic lesions.
The number of papillomas was recorded individually for each mouse twice per week. Papillomas and all other lesions were examined histologically on hematoxylin-eosin-stained paraffin sections for detailed diagnosis. The formation of papillomas in each animal was monitored over time and the percentages of papilloma-bearing animals in each group were plotted as a Kaplan-Meier graph (GraphPad Software Inc., San Diego, CA). We also determined the average number of papillomas per animal in each group by counting all papillomas in non-Tg or Tg mice and dividing them by the number of animals in each group. The size of the tumors was measured at different time points and plotted using GraphPad Prism 4 (GraphPad Software Inc., San Diego, CA). t test or log rank test for group data was used to determine whether the difference in data obtained with non-Tg and Tg mice was statistically significant.
RESULTS
Generation of HPV38 E6/E7-Tg mice. To characterize the biological properties of the HPV38 E6 and E7 oncoproteins in vivo, we generated Tg mouse lines expressing the two viral genes in skin keratinocytes. We used the bovine KVI promoter-enhancer region homologous to the human keratin 10 (K10) promoter that has been used previously for expression of HPV16 E6 and E7 in the suprabasal layers of the epidermis (1). Tg mice with expression of HPV38 E6 and E7 in the skin were made by microinjection of the construct shown in Fig. 1A into the pronuclei of one-cell mouse embryos. Transgene-positive offspring were identified by PCR using two different pairs of primers, as described in Materials and Methods. Two independent Tg lines (2 and 6) of animals were identified. Expression of the HPV38 E6 and E7 genes was determined in each line by semiquantitative RT-PCR. As shown in Fig. 1B, line 6 expressed the HPV38 E6 and E7 genes at higher levels than line 2. Next, we determined the location of the viral transcripts in the epidermis of the Tg animals using in situ hybridization. Sections of ear tissue of non-Tg and Tg mice were hybridized with HPV38 E6/E7 32P-antisense or sense RNA. A strong positive signal in the epidermis of HPV38 E6/E7-Tg mice was observed with antisense RNA, but not with sense probe (Fig. 1C).
HPV38 E6 and E7 induce cellular proliferation in the epidermis of Tg mice. To detect possible morphological alterations in the epidermis of the Tg mice, we performed histological analysis of skin specimens from mice of the two HPV38 E6/E7-Tg lines. For comparison, we included in the analysis HPV16 E6/E7-Tg mice that have a marked and diffuse increase in skin thickness (1). Most of the epidermis in sections from the ears of 6-week-old HPV38 E6/E7-Tg mice (lines 2 and 6) displayed no apparent morphological alteration (Fig. 2A and data not shown). However, hyperproliferative patches of epithelium (at least two for each section) were clearly detected in some areas of the HPV38 E6/E7-Tg mice but never in non-Tg animals (Fig. 2 and data not shown). The formation of these hyperproliferative patches may be due to continuous scratching of the skin that generate microlesions, healing, and irritation. Two representative skin sections, showing hyperplasia and dysplasia, from HPV38 E6/E7-Tg mice are shown in Fig. 2B. Skin epidermal hyperplasia was defined as a lesion containing several layers of basal cells that retain the ability to differentiate (Fig. 2B, middle panel). In addition, dysplasia was described as a lesion containing progressive increase in the number of proliferative basal and suprabasal cells throughout the thickness of the epidermis with nuclear and cytoplasmic abnormalities and occasionally mitosis (Fig. 2B, right panel).
The bromodeoxyuridine (BrdU) incorporation assay and Ki-67 immunohistochemistry showed increased proliferation in the epidermis of HPV38 E6/E7-Tg mice (Fig. 3A). Scoring of BrdU- and Ki-67-positive cells in different randomly selected skin areas showed that the percentage of proliferative keratinocytes was increased approximately twofold in the skin of HPV38 E6/E7-Tg mice compared with non-Tg mice (Fig. 3B). In addition, in situ hybridization using HPV38 E6/E7 32P-antisense RNA probe showed the E6/E7 transcript to be expressed throughout the hyperproliferative patches observed in the epidermis of the HPV38 E6/E7-Tg mice (Fig. 3C). In agreement with this observation, keratinocytes positive for BrdU and Ki-67 were detected in the basal and suprabasal layers of these lesions (Fig. 3D).
Together, these data show that inducing HPV38 E6 and E7 gene expression in mouse epidermis results in keratinocyte hyperproliferation.
UVB-induced cell cycle checkpoints are altered in keratinocytes of HPV38 E6/E7-Tg mice. Exposure of cells to UVB radiation leads to DNA damage that in turn induces cell cycle arrest or apoptosis. Since HPV DNA replication is entirely dependent on the proliferative status of an infected keratinocyte, it is likely that cutaneous HPV types have developed a mechanism to circumvent the negative cell cycle regulation induced by UV. To test this hypothesis, we determined whether HPV38 E6 and E7 can prevent cell cycle arrest induced by UVB irradiation. Non-Tg and HPV38 E6/E7-Tg mice were irradiated with UVB, and the proliferative status of skin keratinocytes was evaluated by BrdU incorporation and Ki-67 immunostaining. Figure 4A and B show that UVB irradiation drastically reduced the number of proliferating keratinocytes in the skin of non-Tg mice. In contrast, the effects of UV were clearly less severe in the epidermis of Tg mice, in which a high percentage of cells were still positive for BrdU or Ki-67 after exposure to UVB (Fig. 4). Immunohistochemical staining for the cell cycle inhibitor p21WAF1 showed that in non-Tg mice, UV irradiation leads to accumulation of this protein in skin keratinocytes, but no accumulation was observed in the two lines of HPV38 E6/E7-Tg mice (Fig. 5A and B). Immunoblot analysis of skin protein extracts confirmed the immunohistochemical data, showing no increase in p21WAF1 levels in epidermis of Tg mice after UV irradiation. In addition, we determined the levels of the proapoptotic protein Bak that is highly stabilized in response to UV irradiation and degraded by E6 from EV HPV type 5 (15). As shown in Fig. 5C, no accumulation of Bak occurred in the skin of HPV38 E6 and E7 Tg mice after UV exposure. Together, these data show that HPV38 E6 and E7 interfere with cellular defense processes induced by UV irradiation.
HPV38 E6 and E7 act synergistically with chemical carcinogens in the development of skin tumors. In contrast to mucosal HPV16 E6/E7-Tg or EV HPV8-Tg mice (18, 29, 31), HPV38 E6/E7 animals did not develop spontaneous tumors or any skin lesions during their life span of approximately 2 to 2.5 years. Therefore, we checked whether HPV38 E6 and E7 cooperate with chemical carcinogens to induce skin tumors in our Tg mouse model, as previously shown for other papillomaviruses (12, 30). Non-Tg (FVB/N, 13 animals), HPV16 E6/E7-Tg (12 animals), and HPV38 E6/E7-Tg (8 for each line) mice were exposed to a two-stage carcinogen treatment with DMBA and TPA (Fig. 6A). After initiation of the chemical treatment, animals were monitored every 3 days over a period of 20 weeks. At week 15, four mice of each group were sacrificed for histological analysis of skin lesions, while the remaining mice were kept under observation for development of skin carcinomas.
In the first 15 weeks, HPV38 E6/E7-Tg mice rapidly developed numerous papillomas, in contrast to the non-Tg mice (Fig. 6B). Skin lesions were detected in HPV38 E6/E7-Tg line 6 mice approximately 7 weeks after the beginning of the carcinogenic treatment, and 1 week later in HPV38 E6/E7-expressing line 2 mice and HPV16 E6/E7-expressing mice (Fig. 6C), indicating a dose-dependent effect of HPV38 oncoprotein expression. At week 9, 100% of the Tg animals had developed papillomas, while non-Tg mice did not develop any skin lesion until week 9 (Fig. 6C). The average number of tumors per animal and tumor size were clearly greater in Tg mice than non-Tg mice (Fig. 6D and E). After four mice of each group were sacrificed at week 15, we performed histological analysis on 44 lesions for each HPV38 line, 55 lesions of HPV16 Tg mice, and 21 lesions of non-Tg mice. The frequencies of papillomas and keratoacanthomas were similar in HPV38 E6/E7-Tg mice (lines 2 and 6) and HPV16 E6/E7-Tg mice (data not shown). We observed in the papillomas hyperkeratinized and hyperproliferative areas (Fig. 7, top panels), while keratoacanthomas showed the typical keratin pearls together with hyperproliferation (Fig. 7, middle panels). At weeks 17 and 18, two out of four HPV38 E6/E7-expressing mice of line 2 and one out of four HPV38 E6/E7-expressing mice of line 6 had developed SCC (Fig. 7, bottom panels and data not shown). No SCCs were observed in non-Tg or HPV16 E6/E7-Tg mice after 20 weeks of DMBA/TPA treatment. These results show that HPV38 E6 and E7 are able to cooperate with chemical carcinogens to induce skin tumors.
DISCUSSION
Although the mucosal high-risk HPV types have been extensively characterized and their role in human carcinogenesis has been proven, very little is known about the cutaneous HPV types. In this study, we generated two lines of Tg mice expressing E6 and E7 of the EV HPV38 under the control of the bovine homolog of the human K10 promoter. We have previously shown that both proteins are able to corrupt the cell cycle and senescence programs in primary keratinocytes, leading to immortalization (6; unpublished data). In agreement with our in vitro data, expression of HPV38 E6 and E7 in mouse epidermis resulted in alteration of cell cycle regulation. Although most of the skin did not present any morphological alteration, patches with increased epithelial thickness were detected in both HPV38 E6/E7-Tg lines. Markers of cellular proliferation were detected in the basal and upper layers of the epidermis in these hyperplastic/dysplastic regions. In addition, in epidermis with no morphological alteration, we observed increased basal cell proliferation in the HPV38 E6/E7-Tg mice. A similar phenomenon was also observed in HPV16 E6/E7-Tg mice, in which E6 and E7 expression was under the control of the same promoter as in the HPV38 E6/E7-expressing animals (1). In this previous study, it was shown that production of autocrine factors induced by HPV16 E6 and E7 was responsible for the stimulation of proliferation of basal cells (1). We are currently investigating in HPV38 E6/E7-expressing mice the possible role of autocrine factors in this event. However, it is also possible that the bovine homolog of the human K10 promoter is weakly active in the basal layer. In fact, in situ hybridization staining showed a positive signal for HPV38 E6 and E7 mRNA also in cells attached to the basal membrane.
In contrast to HPV38 E6/E7-expressing animals, HPV16 E6/E7-Tg mice have a diffuse increase of epidermis thickness. Thus, HPV38 E6 and E7 proteins appear to be less efficient than HPV16 E6 and E7 in promoting morphological alterations of the epidermis when expressed by the K10 promoter. Differences between HPV16 and HPV38 in the intrinsic properties of E6 and E7 are most likely responsible for the difference in phenotype between the two Tg mouse lines. We have previously shown that E7 proteins of HPV16 and HPV38 in vitro have similar efficiency in degrading pRb and deregulating the G1/S transition (6). Thus, other E6 and/or E7 activities may be responsible for the difference observed in the skin of HPV16 E6/E7-Tg and HPV38 E6/E7-Tg mice. HPV16 E6, but not HPV38 E6, has a PDZ-binding motif at the C terminus, which mediates interaction with several PDZ partners, including hDLG, hSCRIBBLE, MUPP1, and MAGI (23). Interestingly, Tg mice expressing an HPV16 E6 mutant that lacks the PDZ-binding motif did not show any epithelial hyperplasia, in contrast to mice expressing wild-type HPV16 E6 (26). Although HPV38 E6 and E7 are not as efficient as HPV16 E7 and E6 in inducing morphological alteration, they appear to cooperate better with chemical carcinogens in the induction of skin cancers. In fact, while E6 and E7 from both HPV types showed similar efficiencies in promoting the formation of keratoacanthomas and papillomas, only HPV38 E6/E7 animals (lines 2 and 6) developed SCC after 20 weeks of DMBA treatment.
UV light is a key risk factor for the development of skin cancer, inducing accumulation of DNA damage including C to T transitions at dipyrimidine sequences (19). This event normally activates cellular defense processes leading to p53 activation that in turn induces cell cycle arrest or apoptosis to allow repair or elimination of the damaged cells, respectively. HPV38 E6 and E7 enable cells to overcome the cell cycle arrest induced by UVB irradiation; we detected only a small reduction in numbers of epidermal proliferative cells after UVB exposure of HPV38 E6/E7-Tg mice. In contrast, in non-Tg mice, UVB irradiation results in almost complete cell cycle arrest. Expression of the cell cycle inhibitor p21WAF1 is under the control of p53 and plays a key role in UV-induced cell cycle arrest. No p21WAF1 accumulation was observed in the epidermis of HPV38 E6/E7-Tg mice upon UV irradiation. HPV38 E6 does not promote p53 degradation (5), but we have recently observed in primary human keratinocytes and Tg mice that expression of HPV38 E6 and E7 results in an alteration of p53 transcriptional functions (R. Accardi and M. Tommasino, unpublished data). Thus, p21WAF1 down-regulation in HPV38 E6/E7-Tg mice may be due to p53 inactivation. Alternatively, HPV38 E6 can repress p21WAF1 transcription by altering p53-independent pathways, as shown for HPV16 E6 (21, 22).
Cutaneous HPV types (both EV and non-EV) are able to antagonize UV-induced apoptosis by promoting the ubiquitin-mediated degradation of Bak, a proapoptotic protein that is stabilized in response to UV irradiation (15). We did not detect Bak accumulation in the epidermis of our HPV38 E6/E7-Tg mice after UV irradiation, showing that this property is shared by HPV38. Thus, HPV38 E6 and E7 oncoproteins appear to have several properties that allow infected epidermal keratinocytes to circumvent the adverse effects of UV radiation. Such activity would be essential to guarantee viral DNA replication in sun-exposed cells and, as a side effect, might favor the accumulation of DNA damage and so facilitate the development of skin cancer. Although we do not yet know whether UV and HPV38 act synergistically in our animal model, the cooperation between HPV38 E6 and E7 and DMBA-induced DNA damage in tumor development support this hypothesis.
Studies on human specimens indicated that the presence of the viral genome may not be required in the later stages of carcinogenesis (34). This may imply a need to consider a new scenario for the role of EV HPV in NMSC pathogenesis, such as a "hit-and-run" mechanism. This mode of action is clearly different from that established for mucosal high-risk HPV in cervical cancer, where E6 and E7 expression is constantly required for maintenance of the neoplastic phenotype of the infected cell.
A recent study on EV HPV type 8 has shown that Tg mice expressing the entire early region under control of the K14 promoter developed spontaneous benign tumors and, in a small percentage, SCC (29). We did not observe any formation of spontaneous tumors in our HPV38 E6/E7-Tg mice during their life span. The difference in phenotype between our two Tg models may have various explanations: (i) E6 and E7 of the two HPV types may display different transforming activities, (ii) E2 that is expressed in HPV8 Tg mice may cooperate with E6 and E7 in tumor development, or (iii) the expression of viral proteins by different promoters in different layers of the epidermis may influence tumor development. Further studies are required to address this issue.
In summary, we present here data that confirm the transforming activity of HPV38 in an in vivo model. The fact that skin keratinocytes of HPV38 E6/E7-Tg mice are unable to respond to cellular stresses strongly suggests that viral infection and UV irradiation may act synergistically in the induction of NMSC.
ACKNOWLEDGMENTS
We are grateful to all members of our laboratory for their cooperation, Martyn Plummer for statistical analysis, Dominique Galendo and Marie-Pierre Cros for animal colony maintenance, Nicole Lyandrat and Christine Carreira for histological service, and John Cheney for critical reading of the manuscript.
REFERENCES
Auewarakul, P., L. Gissmann, and A. Cidarregui. 1994. Targeted expression of the E6 and E7 oncogenes of human papillomavirus type 16 in the epidermis of transgenic mice elicits generalized epidermal hyperplasia involving autocrine factors. Mol. Cell. Biol. 14:8250-8258.
Blessing, M., J. L. Jorcano, and W. W. Franke. 1989. Enhancer elements directing cell-type-specific expression of cytokeratin genes and changes of the epithelial cytoskeleton by transfections of hybrid cytokeratin genes. EMBO J. 8:117-126.
Bowden, G. T. 2004. Prevention of non-melanoma skin cancer by targeting ultraviolet-B-light signalling. Nat. Rev. Cancer 4:23-35.
Boyle, J., R. M. MacKie, J. D. Briggs, B. J. Junor, and T. C. Aitchison. 1984. Cancer warts, and sunshine in renal transplant patients. A case-control study. Lancet i:702-705.
Caldeira, S., R. Filotico, R. Accardi, I. Zehbe, S. Franceschi, and M. Tommasino. 2004. p53 mutations are common in human papillomavirus type 38-positive non-melanoma skin cancers. Cancer Lett. 209:119-124.
Caldeira, S., I. Zehbe, R. Accardi, I. Malanchi, W. Dong, M. Giarre, E.-M. deVilliers, R. Filotico, P. Boukamp, and M. Tommasino. 2003. The E6 and E7 proteins of cutaneous human papillomavirus type 38 display transforming properties. J. Virol. 77:2195-2206.
de Villiers, E. M., C. Fauquet, T. R. Broker, H. U. Bernard, and H. zur Hausen. 2004. Classification of papillomaviruses. Virology 324:17-27.
Durst, M., D. Glitz, A. Schneider, and H. zur Hausen. 1992. Human papillomavirus type 16 (HPV 16) gene expression and DNA replication in cervical neoplasia: analysis by in situ hybridization. Virology 189:132-140.
Eckert, R. L., J. F. Crish, S. Balasubramanian, and E. A. Rorke. 2000. Transgenic animal models of human papillomavirus-dependent disease. Int. J. Oncol. 16:853-870.
Griep, A. E., and P. F. Lambert. 1994. Role of papillomavirus oncogenes in human cervical cancer: transgenic animal studies. Proc. Soc. Exp. Biol. Med. 206:24-34.
Hartevelt, M. M., J. N. Bavinck, A. M. Kootte, B. J. Vermeer, and J. P. Vandenbroucke. 1990. Incidence of skin cancer after renal transplantation in The Netherlands. Transplantation 49:506-509.
Helfrich, I., M. Chen, R. Schmidt, G. Furstenberger, A. Kopp-Schneider, D. Trick, H. J. Grone, H. Zur Hausen, and F. Rosl. 2004. Increased incidence of squamous cell carcinomas in Mastomys natalensis papillomavirus E6 transgenic mice during two-stage skin carcinogenesis. J. Virol. 78:4797-4805.
Iftner, T., S. Bierfelder, Z. Csapo, and H. Pfister. 1988. Involvement of human papillomavirus type 8 genes E6 and E7 in transformation and replication. J. Virol. 62:3655-3661.
Jablonska, S., J. Dabrowski, and K. Jakubowicz. 1972. Epidermodysplasia verruciformis as a model in studies on the role of papovaviruses in oncogenesis. Cancer Res. 32:583-589.
Jackson, S., C. Harwood, M. Thomas, L. Banks, and A. Storey. 2000. Role of Bak in UV-induced apoptosis in skin cancer and abrogation by HPV E6 proteins. Genes Dev. 14:3065-3073.
Kiviat, N. B. 1999. Papillomaviruses in non-melanoma skin cancer: epidemiological aspects. Semin. Cancer Biol. 9:397-403.
Kreimer, A. R., G. M. Clifford, P. Boyle, and S. Franceschi. 2005. Human papillomavirus types in head and neck squamous cell carcinomas worldwide: a systematic review. Cancer Epidemiol. Biomarkers Prev. 14:467-475.
Lambert, P. F., H. Pan, H. C. Pitot, A. Liem, M. Jackson, and A. E. Griep. 1993. Epidermal cancer associated with expression of human papillomavirus type 16 E6 and E7 oncogenes in the skin of transgenic mice. Proc. Natl. Acad. Sci. USA 90:5583-5587.
Lee, D. H., and G. P. Pfeifer. 2003. Deamination of 5-methylcytosines within cyclobutane pyrimidine dimers is an important component of UVB mutagenesis. J. Biol. Chem. 278:10314-10321.
Majewski, S., S. Jablonska, and G. Orth. 1997. Epidermodysplasia verruciformis. Immunological and nonimmunological surveillance mechanisms: role in tumor progression. Clin. Dermatol. 15:321-334.
Malanchi, I., R. Accardi, F. Diehl, A. Smet, E. Androphy, J. Hoheisel, and M. Tommasino. 2004. Human papillomavirus type 16 E6 promotes retinoblastoma protein phosphorylation and cell cycle progression. J. Virol. 78:13769-13778.
Malanchi, I., S. Caldeira, M. Krutzfeldt, M. Giarre, M. Alunni-Fabbroni, and M. Tommasino. 2002. Identification of a novel activity of human papillomavirus type 16 E6 protein in deregulating the G1/S transition. Oncogene 21:5665-5672.
Mantovani, F., and L. Banks. 2001. The human papillomavirus E6 protein and its contribution to malignant progression. Oncogene 20:7874-7887.
Munger, K., A. Baldwin, K. M. Edwards, H. Hayakawa, C. L. Nguyen, M. Owens, M. Grace, and K. Huh. 2004. Mechanisms of human papillomavirus-induced oncogenesis. J. Virol. 78:11451-11460.
Munger, K., J. R. Basile, S. Duensing, A. Eichten, S. L. Gonzalez, M. Grace, and V. L. Zacny. 2001. Biological activities and molecular targets of the human papillomavirus E7 oncoprotein. Oncogene 20:7888-7898.
Nguyen, M. L., M. M. Nguyen, D. Lee, A. E. Griep, and P. F. Lambert. 2003. The PDZ ligand domain of the human papillomavirus type 16 E6 protein is required for E6's induction of epithelial hyperplasia in vivo. J. Virol. 77:6957-6964.
Orth, G., S. Jablonska, M. Jarzabek-Chorzelska, S. Obalek, G. Rzesa, M. Favre, and O. Croissant. 1979. Characteristics of the lesions and risk of malignant conversion associated with the type of human papillomavirus involved in epidermodysplasia verruciformis. Cancer Res. 39:1074-1082.
Pfister, H. 2003. Chapter 8: human papillomavirus and skin cancer. J. Natl. Cancer Inst. Monogr. 31:52-56.
Schaper, I. D., G. P. Marcuzzi, S. J. Weissenborn, H. U. Kasper, V. Dries, N. Smyth, P. Fuchs, and H. Pfister. 2005. Development of skin tumors in mice transgenic for early genes of human papillomavirus type 8. Cancer Res. 65:1394-1400.
Song, S., A. Liem, J. A. Miller, and P. F. Lambert. 2000. Human papillomavirus types 16 E6 and E7 contribute differently to carcinogenesis. Virology 267:141-150.
Song, S., H. C. Pitot, and P. F. Lambert. 1999. The human papillomavirus type 16 E6 gene alone is sufficient to induce carcinomas in transgenic animals. J. Virol. 73:5887-5893.
Walder, B. K., D. Jeremy, J. A. Charlesworth, G. J. Macdonald, B. A. Pussell, and M. R. Robertson. 1976. The skin and immunosuppression. Australas. J. Dermatol. 17:94-97.
Walder, B. K., M. R. Robertson, and D. Jeremy. 1971. Skin cancer and immunosuppression. Lancet ii:1282-1283.
Weissenborn, S. J., I. Nindl, K. Purdie, C. Harwood, C. Proby, J. Breuer, S. Majewski, H. Pfister, and U. Wieland. 2005. Human papillomavirus-DNA loads in actinic keratoses exceed those in non-melanoma skin cancers. J. Investig. Dermatol. 125:93-97.
Yamashita, T., K. Segawa, Y. Fujinaga, T. Nishikawa, and K. Fujinaga. 1993. Biological and biochemical activity of E7 genes of the cutaneous human papillomavirus type 5 and 8. Oncogene 8:2433-2441.
zur Hausen, H. 2002. Papillomaviruses and cancer: from basic studies to clinical application. Nat. Rev. Cancer 2:342-350.(Wen Dong, Ulrich Kloz, Ro)