Human papillomaviruses (HPV) have been implicated in the development of nonmelanoma skin cancer (NMSC). The molecular mechanisms by which these viruses contribute towards NMSC are poorly understood. We have used an in vitro skin-equivalent model generated by transducing primary adult human epidermal keratinocytes with retroviruses expressing HPV genes to investigate the mechanisms of viral transformation. In this model, keratinocytes expressing HPV genes are seeded onto a mesenchyme composed of deepidermalized human dermis that had been repopulated with primary dermal fibroblasts. Expression of the HPV8 E7 gene caused both an enhancement of terminal differentiation and hyperproliferation, but most strikingly, the acquisition of the ability to migrate and invade through the underlying dermis. The basement membrane integrity was disrupted in a time-dependent manner in areas of invading keratinocytes, as evidenced by immunostaining of its protein components collagen types VII, IV, and laminin 5. This was accompanied by the overexpression of extracellular matrix metalloproteinases MMP-1, MMP-8, and MT-1-MMP. These results suggest that the cutaneous HPV type 8 that is frequently found in NMSC of epidermodysplasia verruciformis patients may actively promote an invasive keratinocyte phenotype. These findings also highlight the importance of epithelial-extracellular matrix-mesenchymal interactions that are required to support cell invasion.

Nonmelanoma skin cancer (NMSC) is the most common human cancer in the Caucasian population, with about one million cases in the United States and >40,000 cases occurring annually in the United Kingdom (1, 2). Many epidemiologic studies have shown that the incidence of NMSC has been increasing rapidly over the last decades, and at present, it represents about 30% of all cancer incidence (3). Basal cell carcinoma is the most common NMSC, whereas squamous cell carcinoma (SCC) accounts for 20% of all cutaneous malignancies.

Human papillomaviruses (HPV) are small DNA viruses that induce epithelial hyperproliferation that extends from benign to premalignant and malignant lesions of the cutaneous and mucosal epithelia (4). The association between HPV and skin cancer was first identified in patients with the rare inherited disorder epidermodysplasia verruciformis (EV; ref. 5). These individuals have a predisposition to infection with a specific group of HPV types (EV types) and cancers frequently harbor the oncogenic HPV types 5 or 8 (6). Whereas EV patients are very rare, recent epidemiologic studies have shown that about 30% of NMSC in immunocompetent individuals contain HPV DNA that invariably seems to be EV HPV types. These findings suggest a previously uncharacterized role for EV HPV types in NMSC development in immunocompetent individuals (7, 8).

The two main viral oncoproteins of anogenital HPVs, E6 and E7, display different functions towards the development of cancer, such as cellular transformation, cell death inhibition and immortalization. The major target of the E6 protein on high-risk mucosal HPVs is p53, whose functions are disrupted after its E6-promoted proteolysis (9). The E7 protein functions in cellular transformation by interaction principally with pRb (10) and other cellular targets (for a review, see ref. 11). In vivo, HPV E7 can uncouple cellular differentiation and proliferation and hence, retain differentiating keratinocytes in a DNA replication competent stage. E7 from cervical HPVs can immortalize human foreskin keratinocytes in vitro, alone or in cooperation with E6. However, little is known about the abilities of the cutaneous HPV oncogenes in cellular transformation or immortalization. The E7 genes of cutaneous skin cancer-associated HPV types 8 and 47 failed to induce any detectable transformation of rodent cells (11, 12). In collaboration with the activated Ha-ras gene, however, HPV5 and HPV8 gave rise to transformed cell lines (13). The E7 of the plantar wart-specific HPV1 (a low-risk papillomavirus) fully transformed the mouse fibroblast cell line C127 (14). In the case of cutaneous HPVs, the degree of morphologic transformation by E7 genes therefore seems not correlated to the risk of malignant conversion of the lesions induced by the corresponding HPVs. Little work has been done on cutaneous HPVs with the natural target cell of the virus, the adult epidermal keratinocyte. This may have limited the findings of previous studies investigating the functional activities of early genes. In this sense, it has been recently shown that the E6 and E7 genes of HPV38, a cutaneous EV type, can transform and induce a long-lasting proliferation of primary adult human keratinocytes (15).

The basement membrane (BM) is a specialized form of extracellular matrix (ECM) that is mainly composed of collagen, nidogen, laminins, and perlecan, that separates epithelial cells from the underlying supporting stroma. In tumor development, epithelial cells disrupt the BM, proliferate, and migrate within the connective tissues. The invasive character is facilitated by the expression of specific extracellular matrix metalloproteinases (MMP) that, besides their main activity in connective tissue–remodeling functions, are involved in promoting aspects of tumor growth, such as cellular proliferation, adhesion and dispersion, or migration. Using an organotypic culture of skin, we show that the early E7 gene from high-risk HPV type 8 can modify the normal differentiation and proliferation programmes of primary adult human cutaneous keratinocytes and, more importantly, confer on them the ability to invade the underlying dermis. Furthermore, this phenotype is accompanied by the overexpression of MMP-1, MMP-8, and MT1-MMP in localized areas of the skin culture.

Cell Culture. Cell cultures were incubated at 37°C in a humidified 10% CO2 atmosphere. The NIH 3T3-mouse fibroblast line PT67 (Clontech, Heidelberg, Germany), which was used to replicate amphotropic retroviruses, and NIH 3T3 cells were maintained in DMEM, supplemented with 10% FCS and antibiotics. Human dermal fibroblasts and epidermal keratinocytes were isolated from discarded abdominal skin obtained from plastic surgery. Briefly, thin sheets of skin were removed by using a dermatome and digested with trypsin. Primary human keratinocytes were isolated, propagated on lethally irradiated NIH 3T3 feeder cells, and grown in keratinocyte culture media composed of three parts DMEM and one part Ham's F12 with 10% FCS and supplements as described (16). For the liberation of fibroblasts, skin was enzymatically digested in collagenase D (Roche, Lewes, United Kingdom) and cells were passaged in DMEM supplemented with 10% FCS and antibiotics. Before reaching confluency, cells were trypsinized, resuspended in FCS with 10% DMSO and stored in liquid nitrogen.

HPV Expression Vectors and Production of High Titer Retroviruses. The Moloney murine leukemia retrovirus vector pLXSN (17) was used to generate recombinant retroviruses containing HPV8 genes. This vector contains a gene conferring resistance to neomycin, which is transcribed from a SV40 promoter. E6 and E7 open reading frames were amplified together by PCR by using the primers 5′-TTACAATGCTGTGACTTGTGCAAT-3′ and 5′-CACTACATTCAGCTTCCAAAATACA-3′, and cloned in pCR-Blunt (Invitrogen, Karlsruhe, Germany). The plasmid obtained was digested with BamHI and XhoI, and the DNA fragment containing the E6 and E7 open reading frames was inserted into pLSXN treated with BamHI and XhoI, to obtain pLXSN-8E6E7. The E7 open reading frame alone was amplified by PCR by using the primers 5′-AAGCTTGAATTCGAGTTTGCAGGCTTTGTAAGC-3′ and 5′-GAAGCTTGGATCCCCTTCTTTAGATGTACTACC-3′, which contain EcoRI and BamHI restriction endonuclease sites at their 5′ ends, respectively. After PCR, the amplimers were digested and subsequently cloned into BamHI/EcoRI digested-pLXSN, thus obtaining pLXSN-8E7. The HPV8 genes were cloned downstream of the Moloney murine leukemia virus 5′ long terminal repeat sequence. Recombinant retroviruses were produced by transfecting pLXSN-derived DNA into PT67 cells with Superfect reagent (Qiagen, Hilden, Germany), following the manufacturer's indications. Two days after transfection, cells were plated into selection media containing 500 μg/mL G418 for 3 days. Resistant cells were grown to confluence at which time retrovirus-containing cellular supernatants were collected.

Infection of Human Cutaneous Keratinocytes with Recombinant Retroviruses. Keratinocytes at passage 1 were seeded out in defined keratinocyte serum-free medium (Invitrogen) at a cell density of 9 × 104 cells/cm2 in 6 cm dishes. Retroviral supernatants were mixed with an equal volume of DMEM in the presence of 5 μg/mL of hexadimethrine bromide (polybrene, Sigma, Poole, United Kingdom) and added to the keratinocytes. Spin infection was made by centrifugation for 1 hours at 300 × g, and the cells washed with PBS and the cultured in defined keratinocyte-serum-free medium. After 2 days, cells were selected with G418 (500 μg/mL) for 3 days after which time only infected keratinocytes survived. The cultures were trypsinized before reaching confluency and were used immediately in organotypic cultures as described below. The use of pooled stable cell populations minimizes possible variations due to the apparent randomness of the viral integration site in the cellular chromosomes.

Preparation of Organotypic Cultures with Deepidermalized Dermis. The organotypic cultures were done as described (18). Glycerol-preserved skin (Euro Skin Bank, Beverwijk, Holland) was washed thrice in PBS and incubated in PBS containing antibiotic mix (600 units/mL penicillin-G, 600 μg/mL streptomycin sulfate, 250 μg/mL gentamicin sulfate, and 2.5 μg/mL fungizone) at 37°C for up to 10 days. Epidermis was then mechanically removed using forceps, and deepidermalized dermis was cut into 2 × 2 cm squares and placed in culture plates with the papillary dermal surface on the underside. Stainless steel rings were placed on top of the dermis, and normal human dermal fibroblasts (5 × 105 cells) were inoculated into the rings on the reticular dermal surface. Following an overnight incubation, the deepidermalized dermis was inverted to orient the papillary dermal surface on top before the rings were replaced. Retroviral-infected human cutaneous keratinocytes (3 × 105 cells) were seeded inside the rings onto the dermis. After 2 days, the dermis was raised to the air-liquid interface in the same orientation, by placing the composites on stainless steel grids for 14 days. The medium was refreshed every 3 days. After 14 days, the composites were removed from the grids, fixed in 10% formalin and embedded in paraffin. Cultures were also incubated for 21 days in the air-liquid interface to investigate the progressive destruction of the BM. Deparaffinized sections were stained with H&E for histologic examination.

Immunohistochemistry. For immunohistochemistry, 4-μm-thick wax sections were deparaffinized through xylene treatment and rehydrated through descending grades of ethanol to distilled water. Sections were not subjected to antigen retrieval. The sections were rinsed in PBS and nonspecific binding of antibodies was blocked with serum for 5 minutes at room temperature, before application of the primary antibody. Sections were incubated with the following primary mouse monoclonal antibodies against: collagen VII (LH7.2, supernatant, ref. 19, vimentin [V9], 1:100 dilution, Dako, Ely, United Kingdom); proliferating cell nuclear antigen (PC10, 1:100 dilution, Dako); laminin V (C20, 1:100 dilution, Santa Cruz Biotechnology, Heidelberg, Germany); collagen IV (1:100 dilution, ICN, Asse-Relegem, Belgium); MT1-MMP (Ab 4, 1:100 dilution, Merck Biosciences, Nottingham, United Kingdom), MMP1 (Ab-1, 1:100 dilution, Merck Biosciences), and MMP8 (C20, 1:100 dilution, Santa Cruz Biotechnology). For the controls, the primary antibody was omitted to verify expression patterns (results not shown). Sections were incubated over night at 4°C. The following day, the sections were washed in PBS and then incubated for 30 minutes with a biotinylated secondary antibody and processed with a streptavidin-biotin-peroxidase detection system (Vectastain ABC Kit, Vector Laboratories, Peterborough, United Kingdom) as recommended by the manufacturer. The sections were developed using TSA Fluorescein System (Perkin-Elmer Life Sciences, Inc., Boston, MA). Original magnifications are at 400×.

The organotypic culture of skin is a useful system for the in vitro analysis of skin biology because it can mimic keratinocyte differentiation more effectively than normal monolayer cultures. Cellular functions that require epithelial differentiation, cell-ECM interactions, or keratinocyte-fibroblast paracrine communications can be observed and analyzed. To study the effect of cutaneous HPV gene expression on epithelial cells, we used the organotypic system of skin with retrovirally transduced normal human epidermal keratinocytes.

HPV8 Early Genes Disrupt the Normal Differentiation and Proliferation Programs of Keratinocytes in Regenerated Epithelium. Human keratinocytes were infected with HPV8 E6E7– and HPV8 E7–containing retrovirus and control empty retrovirus pLXSN and seeded onto deepidermalized dermis repopulated with human fibroblast as described in Materials and Methods. After 14 days in the air-liquid interface, conditions that induce epidermal differentiation, the organotypic cultures were paraffin-embedded and 4-μm sections were H&E-stained for histologic examination. Cells infected with the pLXSN-retrovirus generate a normal epithelium with the distinct strata of keratinocyte differentiation over the dermis (Fig. 1). The result shows that the events driving the generation of a normal epithelium are occurring. By contrast, distinct features are observed on E6E7 or E7 cells. These cultures retain the capacity to undergo terminal differentiation. However, when compared with control cultures generated by transducing keratinocytes with empty retrovirus, the organotypic cultures of HPV8 E7–transduced cells show features of hyperkeratinization across the epithelium. The increased cornification was observed not only in the outer epithelial cells, but also within the epithelium itself, leading to the formation of horn pearls (Fig. 1A, features that were commonly detected in the E7-transduced cultures that were not apparent in controls. These structures are similar to concentrically keratinized structures present in well-differentiated spontaneous human skin SCC (20), in severe combined immunodeficient mice transplanted with EV-associated SCC (21), and in skin cancer induced in K14-HPV16 transgenic mice (22).

Figure 1.

Expression of early HPV8 genes causes invasion of keratinocytes into the dermis. A, normal human epidermal keratinocytes were infected with pLXSN (Control), pLXSN-8E6E7, or pLXSN-8E7 containing retroviruses and used in organotypic cultures as described in Materials and Methods. After 14 days in the air-liquid interface, cultures were paraffin-embedded and sections were H&E stained for histologic examination. H, horn pearl-like structure of highly cornified keratinocytes; D, areas of dermis surrounded by keratinocytes; I, infiltrating keratinocytes; F, fibroblasts. B, vimentin staining indicating that the dermal cells in the HPV8 E7 or E6/E7 cultures are fibroblasts (arrow).

Figure 1.

Expression of early HPV8 genes causes invasion of keratinocytes into the dermis. A, normal human epidermal keratinocytes were infected with pLXSN (Control), pLXSN-8E6E7, or pLXSN-8E7 containing retroviruses and used in organotypic cultures as described in Materials and Methods. After 14 days in the air-liquid interface, cultures were paraffin-embedded and sections were H&E stained for histologic examination. H, horn pearl-like structure of highly cornified keratinocytes; D, areas of dermis surrounded by keratinocytes; I, infiltrating keratinocytes; F, fibroblasts. B, vimentin staining indicating that the dermal cells in the HPV8 E7 or E6/E7 cultures are fibroblasts (arrow).

Close modal

More importantly, H&E-stained sections of cells transduced either with both E6 and E7 genes, or only with the E7 gene, revealed a striking phenotype, in that the cells seem to have lost their normal polarity and invaded into the dermal matrix (Fig. 1A). We noted that, especially on E7-transduced keratinocytes, instead of keratinocytes migrating upwards as in a normal epithelium, irregular masses of epidermal cells proliferate down into the dermis. Islands of dermis surrounded by keratinocytes and keratinocytes deep within the dermis were also observed. In contrast to the epithelium derived from control cells where the rete ridges are of equal length, in this study we observed that those of the epithelium generated by HPV8-tranduced cells were more heterogeneous in length. Together these observations suggest that the HPV8-transduced cells, when compared with controls, have acquired the ability to invade surrounding tissues.

Another feature of the regenerated skin produced from HPV8-transduced keratinocytes in this system, when compared with cultures generated by control retrovirally infected keratinocytes, was the presence of fibroblasts in the dermis (Fig. 1A). The identity of these cells was confirmed by positive staining for vimentin (Fig. 1B). As these fibroblasts were initially seeded onto the reticular side of the deepidermalized dermis, their presence in the main body of the dermis indicates an increased migratory capacity, possibly as a result of soluble factors produced by the HPV8-transduced keratinocytes that in turn may also influence keratinocyte proliferation and migration.

In normal epithelium proliferating cells are restricted to basal cell layers. To analyze whether E7 expression resulted in a hyperproliferative phenotype we investigated the expression pattern of proliferating cell nuclear antigen (PCNA) in the organotypic cultures. In control cultures, PCNA expression was, as expected, restricted to basal cells. In marked contrast, increased expression of PCNA was evident in both the HPV8 E7 or HPV8 E6/E7 epithelium. Here PCNA expression was found not only in basal cells but also in suprabasal layers and was most evident in the areas of invading keratinocytes. In addition, HPV-transduced cells also showed some cytoplasmic staining of PCNA suggesting that the expression of the viral genes may alter the cellular localization of the protein (Fig. 2).

Figure 2.

PCNA expression in raft cultures. Sections of organotypic cultures of pLXSN (Control), pLXSN-8E7, or pLXSN-8E6/E7-transduced keratinocytes were stained using an antibody to PCNA (green) as described in Materials and Methods. Note the abundant PCNA staining in the HPV8 cultures in multiple cell layers, PCNA expressing in control cells is limited to basal cells. Propidium iodide staining of nuclei (red). Merged images, nuclei of cells expressing PCNA (yellow).

Figure 2.

PCNA expression in raft cultures. Sections of organotypic cultures of pLXSN (Control), pLXSN-8E7, or pLXSN-8E6/E7-transduced keratinocytes were stained using an antibody to PCNA (green) as described in Materials and Methods. Note the abundant PCNA staining in the HPV8 cultures in multiple cell layers, PCNA expressing in control cells is limited to basal cells. Propidium iodide staining of nuclei (red). Merged images, nuclei of cells expressing PCNA (yellow).

Close modal

The Basement Membrane Integrity Is Disrupted by HPV8 E7–Transduced Keratinocytes in Organotypic Cultures. The BM separates the epithelial and dermal compartments and, as such, represents the initial barrier to tumor cell invasion. Mechanistically, disruption of the BM is a necessary event to allow the migrating cells to invade the dermal compartment. To investigate whether the BM was disrupted in E7-transduced cells, the presence of collagen VII was analyzed by immunohistochemistry. Collagen VII is a component of BM that is still present, structurally intact and forms a continuous layer in the acellular dermis used in the organotypic cultures (23). Degradation of the BM is a marker for early tumor growth (24). In the system used here, the preparation of the dermal substrate leaves the integrity of the BM intact. In contrast to the continuous distribution pattern of the protein along the BM in control cultures (Fig. 3A), collagen VII staining is lost in areas in which E7-transduced keratinocytes are growing downward into the dermis. Interestingly, prolonging the incubation time at the air-liquid interface of the E7-transduced cultures to 21 days, revealed a progressive loss of collagen VII staining of the BM, such that collagen VII–specific staining at the BM was then completely absent. Because collagen VII can normally be synthesized by the keratinocytes under these culture conditions (25), the lack of expression in the HPV-transduced cultures may be due to inhibition of protein synthesis or the active degradation of the protein. We further probed whether the BM function was compromised by investigating the distribution patterns of two other important BM components, collagen IV and laminin V. Both these markers were normally expressed in control cells at both 14 and 21 days. However, as with collagen VII, the intensity of both collagen IV and laminin V staining of the BM zone at the epithelial/dermal juncture was found to be lower at 14 days, and was further reduced by 21 days, in the HPV8 cultures (Fig. 3B and C). In contrast, collagen IV staining of the remnants BM that surrounded dermal blood vessels was evident at both 14 and 21 days (Fig. 3B). Taken together, these findings indicate that expression of HPV8 E7 leads to the progressive loss of BM integrity.

Figure 3.

BM disruption by invading keratinocytes. The integrity of BM was analyzed by immunostaining of key components collagen VII, collagen IV, and laminin V. A, control or HPV8-transduced keratinocytes were used in organotypic cultures and incubated in the air-liquid interface for either 14 or 21 days. After 14 days, collagen VII staining is interrupted in the HPV8E7 cultures compared with controls and is completely absent by 21 days in culture. B, progressive loss of collagen IV immunoreactivity at the BM zone. Note that collagen IV expression is retained on the remnants of blood vessels in the dermis. C, laminin V expression is reduced at 14 days and absent in 21-day organotypic cultures of HPV8-transduced keratinocytes.

Figure 3.

BM disruption by invading keratinocytes. The integrity of BM was analyzed by immunostaining of key components collagen VII, collagen IV, and laminin V. A, control or HPV8-transduced keratinocytes were used in organotypic cultures and incubated in the air-liquid interface for either 14 or 21 days. After 14 days, collagen VII staining is interrupted in the HPV8E7 cultures compared with controls and is completely absent by 21 days in culture. B, progressive loss of collagen IV immunoreactivity at the BM zone. Note that collagen IV expression is retained on the remnants of blood vessels in the dermis. C, laminin V expression is reduced at 14 days and absent in 21-day organotypic cultures of HPV8-transduced keratinocytes.

Close modal

Increased Expression of MMP1, MMP8, and MT1-MMP in HPV8 E7–Transduced Keratinocytes in Skin Cultures. The migration of keratinocytes through the disruption of the BM and ECM are processes that depend on the activity of MMPs and other degrading proteases. Having observed that the E7-transduced keratinocytes lacked collagen VII, collagen IV, and laminin V expression at the BM, we then went to investigate whether the E7-transduced cells had increased expression of specific MMPs that could account for the disrupted BM staining pattern. We were especially interested in the expression of MMP-1, MMP-8, and MT1-MMP, as they are known to be involved in skin diseases (26). Immunostaining of the organotypic cultures with MMP antibodies shows increased expression of specific MMPs in cultures of E7-expressing keratinocytes when compared with controls. Interestingly, MMP-1 (collagenase-1) was present in the dermis and seemed to be concentrated at the dermal-epidermal interphase (Fig. 4). Whether this represents the site of synthesis or results from binding or retention of the protein by other ECM components is at present unknown. The MMP-1 expression was notably increased in dermis and mainly located along a line between the migrating cells and the dermis. MMP-1 staining is also observed in the epithelium of both control and the E7 cultures. In contrast, overexpression of MMP-8 (collagenase-2) is observed also in the dermis (Fig. 4). Finally, membrane-associated MMP (MT1-MMP) was up-regulated on E7 cultures in the epithelial compartment when compared with control cultures (Fig. 4). No expression of MT1-MMP was observed in the dermis of either control or E7-transduced cells. These findings indicate that the E7 protein of HPV8 promotes the expression of different MMPs in separate skin compartments that can co-operate in the E7-induced keratinocyte phenotype and invasion through the dermis. This induction of MMP expression, rather than an inhibition of gene expression, most likely accounts for the lack of BM components of the HPV-transduced keratinocyte cultures.

Figure 4.

HPV8 E7 protein upregulates the expression of MMP-1, MMP-8, and MT1-MMP. Sections of control or HPV8-derived organotypic cultures were immunostained with specific antibodies to MMP-1, MMP-8, and MT1-MMP, as described in Materials and Methods (green).

Figure 4.

HPV8 E7 protein upregulates the expression of MMP-1, MMP-8, and MT1-MMP. Sections of control or HPV8-derived organotypic cultures were immunostained with specific antibodies to MMP-1, MMP-8, and MT1-MMP, as described in Materials and Methods (green).

Close modal

The key molecular changes responsible for the acquisition of an invasive phenotype by tumor cells are not well understood. The investigation and characterization of the cellular changes that result in the ability of the cells to invade surrounding tissues and migrate to distant body sites will yield new insights into tumor metastasis. Our findings that the HPV8 E7 gene alone is capable of generating an invasive phenotype when expressed in epidermal keratinocytes not only provides the first evidence supporting a direct role for the virus in NMSC development but also serves as a general model to study epidermal invasion.

The E6 and E7 oncoproteins of mucosal HPV (such as HPV16 and HPV18) display cellular transformation and immortalization activities in vitro. Molecular interactions between HPV16 E6/E7 and cellular proteins have been described that could explain some of their functions towards cancer development. However, there is a paucity of information about the potential role of cutaneous HPV proteins in skin cancer development. It is very probable that the simple model systems employed to characterize the transforming potential of anogenital viruses are inappropriate for the study of cutaneous HPVs. Different HPV types are associated with the development of lesions at particular body sites. This tropism implies that the viral life cycle has a requirement for specific cellular factors present only in keratinocytes derived from that particular body site, and in addition may require additional cues from mesenchymal cells. The dependence of the viral life cycle on cellular differentiation together with the difficulty of generating a fully differentiated stratified epithelium in vitro has hampered the investigation of cutaneous HPV gene function. Considerable progress has been made in functional analysis of HPVs in organotypic cultures using collagen as matrix (27). Previously, such culture analysis has been made using foreskin keratinocytes expressing the E6 and E7 genes of several EV HPV types and of the high-risk mucosal type HPV16 (28). However, no invasive phenotype of the epithelial cells was described in any of the HPV types analyzed. It should be noted however that organotypic cultures using collagen as matrix do not closely mimic an in vivo environment, as they do not maintain the structural integrity of normal skin together and lack ECM proteins, including functionally important molecules such as glycosaminoglycans. These characteristics are fundamental for in vitro models investigating epithelial-mesenchymal interactions (18). The system used in this paper emulates epidermal regeneration more closely through the in vitro culture of keratinocytes onto an acellular dermal substrate repopulated with dermal fibroblasts, with subsequent differentiation of the culture taking place at the air-liquid interface (18, 29). We have used this culture system to investigate HPV8 early gene function in primary keratinocytes. Most significantly, we show that HPV8 E7 promotes a tumorigenic phenotype as evidenced by the invasive behavior of the HPV-transduced keratinocytes. Migration of the keratinocytes downward into dermis was facilitated by the degradation of components of the BM and ECM (collagen VII, collagen IV, and laminin V) through the induction of expression of the extracellular proteinases MMP-1, MMP-8, and MT1-MMP. Most interestingly, these in vitro findings are in accord with similar observations in skin cancer, thereby supporting a direct role of the E7 protein of cutaneous HPV8 in NMSC development. We would hypothesize that other viral types closely linked to cancer development may also share the phenotype produced by expression of HPV8 genes. Our studies suggest that both the target keratinocyte, in addition to the cellular and acellular mesenchymal components of the organotypic system itself, are critical in eliciting the invasive behavior. Increased proliferation was shown by increased PCNA expression in suprabasal cell layers, which is in agreement with Boxman et al. (28), who also described PCNA staining in suprabasal keratinocytes in cultures containing E6 and E7 genes of EV HPVs. Additionally, HPV8 E7 alters the normal differentiation program of the cells resulting in hyperkeratosis and horn pearl formation. These are features also described in SCC of EV and immunocompromised patients and further suggest that HPV8 early genes E6 and E7 can alter the normal homeostasis of the keratinocyte.

Invasion of malignantly transformed cells is a complex, sequential multistage process that involves the controlled degradation of structural barriers such as basement membrane and collagenous ECM and migration of cells through the degraded matrix. Collagen VII degradation is an early event observed in NMSC progression and is used as marker for early invasion. Degradation of collagen VII is observed in E7 cultures (Fig. 3A), most strikingly after prolonging the air-liquid interface incubation time from 14 to 21 days. The decreased levels of collagen IV and laminin V in the HPV8 cultures provided further evidence that BM integrity was compromised.

Immunostaining of the organotypic cultures with MMP antibodies showed increased expression of MMP-1 and MMP-8 in the dermis and MT1-MMP in the epidermis, suggesting that the activity of different MMPs in separated compartments can cooperate in the movement of transduced keratinocytes through the dermis. Because the organotypic culture is a bicellular system in which normal human dermal fibroblasts seeded on the reticular side of the dermis are cocultivated with the keratinocytes, both cell types can contribute to the expression of MMP-1 and MMP-8 located in the dermis. The apparent increase in the number of fibroblasts in the dermis of the HPV8 cultures may facilitate this process. The observations in distribution of expression of MMP-1, MMP-8, and MT1-MMP are consistent with previous findings in tissue samples of basal cell carcinoma and SCC (ref. 30 and references therein; ref. 31). In addition, up-regulation of these MMPs has also been described in SCC lines (31–33), and their role in keratinocyte invasiveness showed. Interestingly, microarray analyses of other HPV associated cancers, such as cervical cancer (34) that is typically associated with HPV types 16 and 18 and cell lines that express specific HPV16 genes (35), have also shown increased expression of MMPs and ECM remodeling proteins. Enhanced MMP-1 production also occurs in other skin conditions where homeostasis is perturbed, such as psoriasis where EV HPV types have been detected (36, 37), or exposure to UV light (38), a well-known NMSC cocarcinogen. The subset of MMPs induced by HPV gene expression supports a model in which native dermal collagen is degraded by the collagenases MMP-1 and MMP-8, especially collagen VII, the main component of the BM of the skin. Subsequently MT1-MMP can degrade gelatin (product of the enzymatic digestion of collagen), as well as other components of the dermis (39), allowing the epithelial cells to migrate in the dermal stroma (40). Interestingly, MT1-MMP also has a direct role in the activation of MMP-2 (gelatinase-A) in the outer side of the cellular membrane. MMP-2 regulates cell migration and proliferation during cancer cell invasion and is also capable of degrading BM and ECM components (41). It cannot be discounted at present that the possible activation of MMP-2 in the external surface of HPV8 E7-expressing keratinocytes may contribute to the migration of the invading cells.

In summary, our findings suggest that specific HPV types may play a direct role in skin carcinogenesis. The development of this in vitro model of HPV-associated epithelial tumorigenesis will greatly facilitate the investigation and better understanding of the molecular mechanisms that underpin tumor invasion.

Note: B. Akgül and R. García-Escudero contributed equally to this study.

Grant support: Center of Molecular Medicine Cologne, Germany grant 01KS9502, Cancer Research UK, Marie-Curie fellowship (European Commission, R. García-Escudero), and fellowship of the Dr. Mildred Scheel Stiftung für Krebsforschung/Deutsche Krebshilfe (B. Akgül).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank N. Ojeh for expert technical advice on keratinocyte purification and propagation, D. Roth for pLXSN-8E7, I. Schaper for pLXSN-8E6E7, and I.M. Leigh for helpful discussions.

1
Miller DL, Weinstock MA. Nonmelanoma skin cancer in the United States: incidence.
J Am Acad Dermatol
1994
;
30
:
774
–8.
2
Stern RS. The mysteries of geographic variability in nonmelanoma skin cancer incidence.
Arch Dermatol
1999
;
135
:
843
–4.
3
DePinho RA. The age of cancer.
Nature
2000
;
408
:
248
–54.
4
zur Hausen H. Papillomaviruses in human cancers.
Proc Assoc Am Physicians
1999
;
111
:
581
–7.
5
Jablonska S, Dabrowski J, Jakubowicz K. Epidermodysplasia verruciformis as a model in studies on the role of papovaviruses in oncogenesis.
Cancer Res
1972
;
32
:
583
–9.
6
Majewski S, Jablonska S. Epidermodysplasia verruciformis as a model of human papillomavirus-induced genetic cancer of the skin.
Arch Dermatol
1995
;
131
:
1312
–8.
7
Harwood CA, Proby CM. Human papillomaviruses and non-melanoma skin cancer.
Curr Opin Infect Dis
2002
;
15
:
101
–14.
8
Pfister H. Chapter 8: Human papillomavirus and skin cancer. J Natl Cancer Inst Monogr 2003;52–6.
9
Werness BA, Levine AJ, Howley PM. Association of human papillomavirus types 16 and 18 E6 proteins with p53.
Science
1990
;
248
:
76
–9.
10
Münger K, Phelps WC, Bubb V, Howley PM, Schlegel R. The E6 and E7 genes of the human papillomavirus type 16 together are necessary and sufficient for transformation of primary human keratinocytes.
J Virol
1989
;
63
:
4417
–21.
11
Iftner T, Bierfelder S, Csapo Z, Pfister H. Involvement of human papillomavirus type 8 genes E6 and E7 in transformation and replication.
J Virol
1988
;
62
:
3655
–61.
12
Kiyono T, Nagashima K, Ishibashi M. The primary structure of major viral RNA in a rat cell line transfected with type 47 human papillomavirus DNA and the transforming activity of its cDNA and E6 gene.
Virology
1989
;
173
:
551
–65.
13
Yamashita T, Segawa K, Fujinaga Y, Nishikawa T, Fujinaga K. Biological and biochemical activity of E7 genes of the cutaneous human papillomavirus type 5 and 8.
Oncogene
1993
;
8
:
2433
–41.
14
Schmitt A, Harry JB, Rapp B, Wettstein FO, Iftner T. Comparison of the properties of the E6 and E7 genes of low- and high-risk cutaneous papillomaviruses reveals strongly transforming and high Rb-binding activity for the E7 protein of the low-risk human papillomavirus type 1.
J Virol
1994
;
68
:
7051
–9.
15
Caldeira S, Zehbe I, Accardi R, et al. The E6 and E7 proteins of the cutaneous human papillomavirus type 38 display transforming properties.
J Virol
2003
;
77
:
2195
–206.
16
Rheinwald JG. Methods for clonal growth and serial cultivation of normal human epidermal keratinocytes and mesothelial cells. Oxford: IRL Press; 1989. p. 81–94.
17
Miller AD, Rosman GJ. Improved retroviral vectors for gene transfer and expression.
Biotechniques
1989
;
7
:
980
–2, 984–6, 989–90.
18
Ojeh NO, Frame JD, Navsaria HA. In vitro characterization of an artificial dermal scaffold.
Tissue Eng
2001
;
7
:
457
–72.
19
Leigh IM, Purkis PE, Bruckner-Truderman B. LH7.2 monoclonal antibody detects type VII collagen in the sublamina densa zone of ectodermally-derived epithelia, including skin.
Epithelia
1987
;
1
:
17
–29.
20
Schwartz RA, Stoll HL. Fitzpatrick's Dermatology in General Medicine. 5th ed. The McGraw-Hill Companies inc.; 1999. Vol. 1. p. 840–56.
21
Adachi A, Kiyono T, Taguchi O, Ohashi M, Ishibashi M. Serial transplantation in SCID mice of an epidermodysplasia verruciformis-associated squamous cell carcinoma without alteration of its histological and virological features.
Virology
1996
;
217
:
380
–3.
22
Coussens LM, Hanahan D, Arbeit JM. Genetic predisposition and parameters of malignant progression in K14-HPV16 transgenic mice.
Am J Pathol
1996
;
149
:
1899
–917.
23
Jackson S. PhD Thesis, University of London, UK.; 1999.
24
Hagedorn HG, Tubel J, Wiest I, Schleicher ED, Nerlich AG. Prognostic aspects of the loss of epithelial basement membrane components in preinvasive and invasive laryngeal carcinomas.
Anticancer Res
1998
;
18
:
201
–7.
25
McKay I, Woodward B, Wood K, Navsaria HA, Hoekstra H, Green C. Reconstruction of human skin from glycerol-preserved allodermis and cultured keratinocyte sheets.
Burns
1994
;
20
Suppl 1:
S19
–22.
26
Kahari VM, Saarialho-Kere U. Matrix metalloproteinases and their inhibitors in tumour growth and invasion.
Ann Med
1999
;
31
:
34
–45.
27
Chow LT, Broker TR. In vitro experimental systems for HPV: epithelial raft cultures for investigations of viral reproduction and pathogenesis and for genetic analyses of viral proteins and regulatory sequences.
Clin Dermatol
1997
;
15
:
217
–27.
28
Boxman IL, Mulder LH, Noya F, et al. Transduction of the E6 and E7 genes of epidermodysplasia-verruciformis-associated human papillomaviruses alters human keratinocyte growth and differentiation in organotypic cultures.
J Invest Dermatol
2001
;
117
:
1397
–404.
29
Prunieras M, Regnier M, Woodley D. Methods for cultivation of keratinocytes with an air-liquid interface.
J Invest Dermatol
1983
;
81
:
28
–33s.
30
Kahari VM, Saarialho-Kere U. Matrix metalloproteinases in skin.
Exp Dermatol
1997
;
6
:
199
–213.
31
Moilanen M, Pirila E, Grenman R, Sorsa T, Salo T. Expression and regulation of collagenase-2 (MMP-8) in head and neck squamous cell carcinomas.
J Pathol
2002
;
197
:
72
–81.
32
Kerkela E, Ala-aho R, Lohi J, Grenman RVMK, Saarialho-Kere U. Differential patterns of stromelysin-2 (MMP-10) and MT1-MMP (MMP-14) expression in epithelial skin cancers.
Br J Cancer
2001
;
84
:
659
–69.
33
Shindoh M, Higashino F, Kaya M, et al. Correlated expression of matrix metalloproteinases and ets family transcription factor E1A-F in invasive oral squamous-cell-carcinoma-derived cell lines.
Am J Pathol
1996
;
148
:
693
–700.
34
Chen Y, Miller C, Mosher R, et al. Identification of cervical cancer markers by cDNA and tissue microarrays.
Cancer Res
2003
;
63
:
1927
–35.
35
Duffy CL, Phillips SL, Klingelhutz AJ. Microarray analysis identifies differentiation-associated genes regulated by human papillomavirus type 16 E6.
Virology
2003
;
314
:
196
–205.
36
Favre M, Orth G, Majewski S, Baloul S, Pura A, Jablonska S. Psoriasis: A possible reservoir for human papillomavirus type 5, the virus associated with skin carcinomas of epidermodysplasia verruciformis.
J Invest Dermatol
1998
;
110
:
311
–7.
37
Weissenborn SJ, Hopfl R, Weber F, Smola H, Pfister HJ, Fuchs PG. High prevalence of a variety of epidermodysplasia verruciformis-associated human papillomaviruses in psoriatic skin of patients treated or not treated with PUVA.
J Invest Dermatol
1999
;
113
:
122
–6.
38
Fisher GJ, Datta SC, Talwar HS, et al. Molecular basis of sun-induced premature skin ageing and retinoid antagonism.
Nature
1996
;
379
:
335
–9.
39
Ohuchi E, Imai K, Fujii Y, Sato H, Seiki M, Okada Y. Membrane type 1 matrix metalloproteinase digests interstitial collagens and other extracellular matrix macromolecules.
J Biol Chem
1997
;
272
:
2446
–51.
40
Nagavarapu U, Relloma K, Herron GS. Membrane type 1 matrix metalloproteinase regulates cellular invasiveness and survival in cutaneous epidermal cells.
J Invest Dermatol
2002
;
118
:
573
–81.
41
Corcoran ML, Hewitt RE, Kleiner DE Jr, Stetler-Stevenson WG. MMP-2: expression, activation and inhibition.
Enzyme Protein
1996
;
49
:
7
–19.