Serological relationship between cutaneous human ... - CiteSeerX

Journal of General Virology (2009), 90, 136–143

DOI 10.1099/vir.0.006189-0

Serological relationship between cutaneous human papillomavirus types 5, 8 and 92 Alessandra Handisurya,1 Ratish Gambhira,2 Christina Schellenbacher,1 Saeed Shafti-Keramat,1 Ola Forslund,3 Michel Favre4 and Reinhard Kirnbauer1 Correspondence Reinhard Kirnbauer reinhard.kirnbauer @meduniwien.ac.at

1

Laboratory of Viral Oncology, Division of Immunology, Allergy and Infectious Diseases (DIAID), Department of Dermatology, Medical University Vienna, Vienna, Austria

2

Department of Pathology, The Johns Hopkins School of Medicine, Baltimore, USA

3

Department of Medical Microbiology, Malmoe University Hospital, Lund University, Malmoe, Sweden

4

Unite´ postulante de Ge´ne´tique, Papillomavirus et Cancer Humain, Institut Pasteur, Paris, France

Received 1 August 2008 Accepted 5 September 2008

Evidence of a possible association of cutaneous human papillomavirus (HPV) types, especially members of the genus Betapapillomavirus, and the development of non-melanoma skin cancer (NMSC) is accumulating. Vaccination with virus-like particles (VLPs) consisting of self-assembled L1, the major capsid protein, has been introduced to control anogenital HPV infection. This study examined the serological relationship between betapapillomavirus (b-PV) types 5 and 8 and the new type HPV-92, which has recently been isolated from a basal cell carcinoma containing a high number of viral genomes. Following expression by recombinant baculoviruses, the L1 protein of HPV-92 self-assembled into VLPs that elicited high-titre antibodies after immunization, similar to VLPs from HPV-5 and -8. Haemagglutination inhibition (HAI) assays were used as a surrogate method for the detection of virion-neutralizing antibodies, which correlates with protection from infection. Antisera raised against HPV-5 and -8 VLPs displayed HAI activity not only against the homologous type, but also against heterologous HPV types 5, 8 and 92, whereas HAI activity of antisera against HPV-92 VLP was restricted to the homologous type. The results of neutralization assays using HPV-5 pseudovirions were consistent with those from HAI assays. Cross-neutralizing immune responses by VLP vaccination against heterologous HPV types may provide broader protection against the multiplicity of HPV types detected in NMSC. If a close link to HPV infection can be conclusively established, these results may provide a basis for further evaluation of VLPs of b-PVs as candidates for a prophylactic skin-type HPV vaccine, aimed at reducing the incidence of NMSC.

INTRODUCTION Papillomaviruses (PVs) comprise a large group of small, non-enveloped DNA tumour viruses that infect epithelial cells of the skin and mucosa, causing papillomas and warts in many different vertebrate species including humans. More than 100 human papillomavirus (HPV) types have currently been completely characterized. The aetiological role of highrisk mucosal HPV types in the development of anogenital – especially cervical – neoplasias and a subset of oropharyngeal carcinomas is firmly established (Howley & Lowy, 2007; Kirnbauer et al., 2008). In addition, the possible involvement of cutaneous HPVs in the pathogenesis of non-melanoma skin cancer (NMSC) is under active investigation. NMSCs, in particular basal cell carcinomas (BCCs) and squamous cell carcinomas (SCCs), are the most prevalent 136

malignancies in the Caucasian population and their incidence is increasing (Diepgen & Mahler, 2002). The main risk factor for development of these keratinocyte skin cancers is UV radiation, with immunosuppression and genetic factors appearing to play an adjunct predisposing role. Due to behavioural changes and ageing of the world’s population, these NMSCs represent increasing health and socioeconomic problems. It is estimated that at least half of white-skinned transplant recipients will develop NMSC (Diepgen & Mahler, 2002). Preventative strategies have focused on raising public awareness to use precautions against UV exposure. In recent years, evidence of a possible association of HPV infection and development of NMSC has accumulated. In contrast to cervical cancer, a defined adjunct causal role for HPV infection, in addition to UV radiation-induced DNA damage, has yet to be established. 006189 G 2009 SGM Printed in Great Britain

Serological relationship of HPV-5, -8 and -92

The oncogenic potential of certain HPV types in human skin cancer was originally recognized in studies of patients with the rare genodermatosis epidermodysplasia verruciformis (EV) (Orth, 2006). EV patients display increased susceptibility to cutaneous HPV infection, and mutations in the EVER1 or EVER2 genes have been linked to the disease (Ramoz et al., 2002). There exist 24 fully characterized HPV types belonging to the genus Betapapillomavirus (which historically have been designated EV types), which have been isolated from the characteristic macular and verrucous EV skin lesions. Approximately 50 % of EV patients develop SCC on sunexposed areas. HPV-5 and HPV-8 are the types most often present in EV skin cancers, and viral genes are always transcribed (Orth, 2006). Detection of betapapillomavirus (b-PV) DNA in skin cancer is not restricted to EV patients, as NMSCs of immunocompetent patients frequently harbour b-PV DNA, although generally at very low copy numbers of less than one in 100 cells. Nested PCR-based epidemiological studies have reported b-PV DNA in about 30–50 % of NMSCs of immunocompetent individuals (Harwood et al., 2004; Pfister, 2003), whereas in organ transplant recipients, a group with a 50–100-fold increased risk of developing NMSC, the detection rate increases up to 90 % (de JongTieben et al., 1995). Furthermore, b-PV DNA has been found in pre-malignant skin lesions such as actinic keratoses and extragenital Bowen’s disease, but also in normal or psoriatic skin and plucked hairs (Antonsson et al., 2000; Boxman et al., 1997; Favre et al., 1998; Forslund et al., 2003c). Independently, several epidemiological studies have found an association between the detection of b-PV DNA in normal skin or hair follicles, preferentially on sun-exposed sites, and increased risk of developing premalignant or invasive SCC (Harwood et al., 2004; Termorshuizen et al., 2004).

patients with SCC or actinic keratosis, a precursor lesion of SCC (Feltkamp et al., 2003). A relatively new HPV type, designated (candidate) HPV92, was isolated from a BCC of an 89-year-old immunocompetent Australian male (Forslund et al., 2003a). Intriguingly, a high viral load comprising approximately 100 DNA copies per tumour cell was observed, compared with the low amount of HPV DNA usually found in keratinocyte skin cancers. Additionally, HPV-92 has been detected in a fraction of BCCs, raising the possibility of a causal association (Forslund et al., 2007). HPV-92 DNA has also been detected in a fraction of SCCs, solar keratoses and keratoacanthomas of immunocompetent as well as immunocompromised patients (Forslund et al., 2003a, b, c). HPV-92 belongs to species 4 b-PVs (de Villiers et al., 2004) and is phylogenetically distantly related to the oncogenic types HPV-5 and -8, with an amino acid identity of 71 % in the L1 protein (Fig. 1). HPV-5 and -8, both species 1 b-PVs, share 87 % amino acid sequence identity in L1. The major capsid protein L1 of PVs self-assembles into virus-like particles (VLPs) that are highly immunogenic (Kirnbauer et al., 1992, 1993). Generally, systemic immunization with VLPs induces mainly type-restricted neutralizing antiserum, so a polyvalent vaccine would be required to protect against the multiplicity of b-PV types that may play a role in the development of SCCs of the skin. However, low levels of cross-neutralization have been detected after VLP immunizations, but only between the closely related genital types HPV-6 and -11, HPV-16 and -33, HPV-31 and -33, and HPV-18 and -45 (Giroglou et al.,

Gene transcription of b-PV types has been demonstrated in a proportion of SCCs from both immunocompetent and immunocompromised individuals. Although the level of transcriptional activity has been low, these results imply a possible biological role in tumour development (Pfister, 2003; Purdie et al., 2005). The oncogenes E6 and E7 of bPVs have weak transforming potential in vitro compared with E6 and E7 of high-risk mucosal types such as HPV-16. However, b-PV E6 targets the pro-apoptotic protein Bak for degradation (Jackson et al., 2000) and thus may protect cells from UV-induced apoptosis. By favouring accumulation of UV-induced DNA damage, b-PVs may contribute at an early stage of skin carcinogenesis. In addition, E6 of some b-PVs (HPV-8 and -38) can activate telomerase, and a fraction of HPV-8 E6/E7 transgenic mice spontaneously develop SCCs (Bedard et al., 2008; Schaper et al., 2005). Seroepidemiological data have also suggested an association between NMSC and skin-type HPV infection. Increased seroprevalence for antibodies (Abs) to b-PV types, including HPV-5 and -8, has been observed in http://vir.sgmjournals.org

Fig. 1. Phylogenetic tree of members of the genus Betapapillomavirus (species 1 to 5) indicating the genetic relationships of HPV-5, -8 and -92. 137

A. Handisurya and others

2001; Harper et al., 2006; Roden et al., 1996; White et al., 1998), showing amino acid sequence identity in the L1 protein of 92.4, 80.9, 78.4 and 87.4 %, respectively. If a preventative b-PV VLP vaccine were developed, information on cross-neutralization would be clinically relevant, as vaccine formulations are likely to include a limited number of HPV VLP types. The aim of this study was to determine the serological relationship between the cutaneous types HPV-5, -8 and -92, which have been associated with the pathogenesis of NMSC.

METHODS Construction of recombinant baculovirus stocks and generation of HPV VLPs. To construct recombinant baculoviruses

expressing the major capsid protein L1 of the new HPV type 92, the L1 open reading frame was amplified from HPV-92 DNA by inverse-touchdown PCR (Slupetzky et al., 2001) using the primers 59-GCGCCAGATCTAATATGTCCTATTGGCTTCCAG-39 and 59GCGCCAGATCTCTCATAACACATGAGTTCC-39, which correspond to HPV-92 sequences between nt 5638–5656 and 7212–7230, respectively, and contain recognition sites for BglII (underlined), GC clamps (italic) and a polyhedrin leader sequence (AAT) preceding the ATG start codon. The amplified L1 gene was cloned into the baculovirus transfer vector pSynwtVI2 via BglII sites and final clones were verified by restriction enzyme digest and nucleotide sequencing. By co-transfection of Spodoptera frugiperda (Sf9) insect cells with transfer vector and linearized baculovirus DNA (BaculoGold; BD Biosciences Pharmingen), recombinant baculoviruses were generated using standard methods (Handisurya et al., 2007). Recombinant baculoviruses expressing L1 of HPV-5 and -8 have been described previously (Favre et al., 2000; Stark et al., 1998). Sf9 insect cells were subsequently infected at a high m.o.i. and high-molecular-mass structures were separated by density gradient ultracentrifugation. VLP-containing bands were collected and dialysed against PBS containing 0.5 M NaCl and 0.05 % NaN3. Purified VLPs were analysed by Western blotting or Coomassie Brilliant Blue staining. Briefly, samples were denatured in SDS sample buffer containing 2 % b-mercaptoethanol and analysed by 10 % SDSPAGE. Following electrophoresis, proteins were stained with Coomassie Brilliant Blue or, for Western blotting, transferred to membrane and probed overnight at 4 uC with rabbit polyclonal immune serum raised against HPV VLPs. Finally, blots were incubated with horseradish peroxidase-labelled goat anti-rabbit Ab (diluted 1 : 20 000; Kirkegaard & Perry Laboratories) and developed using an enhanced chemiluminescence system (Amersham Biosciences AB). To analyse the self-assembly of L1 and the integrity of VLPs, purified VLP preparations were negatively stained with 1 % uranyl acetate and analysed by transmission electron microscopy using a JEOL 1010 electron microscope at 80 kV and 630 000 magnification. Generation of anti-HPV VLP immune serum. Purified HPV-5, -8

or -92 VLPs were administered to New Zealand White rabbits (Charles River Laboratories) to generate anti-HPV VLP immune sera (Zamora et al., 2006). Fifty micrograms of the respective HPV VLP preparations were injected four times at 2–4-week intervals using complete Freund’s adjuvant for the prime and incomplete Freund’s adjuvant for the boost inoculations. Serum samples were collected prior to and 10–14 days after the second, third and final immunizations. Induction of Abs was verified by native VLP ELISA using the 138

respective homologous VLP preparation as the antigen, as described previously (Kirnbauer et al., 1994; Zamora et al., 2006). Briefly, intact VLPs were diluted in PBS to a final concentration of 0.01 mg ml21 and 0.1 ml aliquots were coated overnight at 4 uC onto 96-well microtitre plates (Maxisorp; Nunc). After three washes with PBS, wells were blocked with 1 % non-fat dried milk in PBS. Rabbit immune serum raised against HPV-5, -8 or -92 VLPs was used as the primary Ab. Serial dilutions of sera ranging from 1022 to 1027 were added to triplicate wells and incubated. Detection was performed using horseradish peroxidase-labelled goat anti-rabbit Ab (Kirkegaard & Perry Laboratories) and the peroxidase substrate diammonium 2,29azino-bis(3-ethylbenz-thiazoline-6-sulfonate) (Roche Diagnostics). Absorbance values were determined at 405 nm. Specific A405 values were calculated by subtracting mean values obtained in wells without antigen (PBS only) from values obtained in antigen-coated wells. Replication variation in the assays was less than 5 %. Total IgG was purified from rabbit pre-immune and immune sera using an ImmunoPure IgG Purification kit A Plus (Pierce Biotechnologies). Haemagglutination inhibition (HAI) assay. Total IgG purified

from rabbit pre-immune or immune serum, a murine monoclonal Ab (mAb) specifically recognizing HPV-5 (M. Favre, unpublished data) and an antiserum raised against bovine papillomavirus 1 (BPV-1) VLPs were cleared of non-specific haemagglutination (HA) and complement activity by incubation at 56 uC for 30 min followed by centrifugation for 5 min at 10 000 g and 4 uC. VLPs of HPV-5, -8 or -92 were serially diluted in PBS containing 0.1 % BSA in 96-well plates. BPV-1 and BK virus VLPs (a kind gift of J. Schiller, National Cancer Institute, MD, USA) were used as positive controls. A mock VLP preparation derived from Sf9 cells infected with mutant (prototype) HPV-16 L1 that fails to assemble into intact particles was used as a negative control for HA. VLPs were incubated with rabbit IgG (final dilution 1 : 100) for 1 h at room temperature in a final volume of 75 ml, to allow proper adhesion of the Abs to the VLPs. Twenty-five microlitres of a 1 % (v/ v) suspension of erythrocytes from fresh citrate-treated blood of BALB/c mice was added to the samples and the plates were incubated for 3 h at 4 uC and photographed. To determine the HAI titre, thresholds were determined at which specific HA by addition of VLPs was observed. HPV VLPs were dispersed at identical concentrations and serial dilutions (ranging from 1 : 100 to 1 : 819200) of rabbit IgG were added to the wells. HAI assays were performed as described above. HPV-5 pseudovirion (PsV) neutralization assay. HPV-5 PsVs were generated by co-transfection of 293 TT cells and neutralization assays were performed as described previously (Buck et al., 2005). Briefly, 293 TT cells [30 000 cells in 100 ml neutralization buffer (10 % FCS in Dulbecco’s modified Eagle’s medium without phenol red) per well] were plated in non-treated sterile 96-well plates (Nalge Nunc) 3– 4 h in advance to allow proper adherence. HPV-5 PsVs were incubated for 1 h at 37 uC in 5 % CO2 with serial dilutions of total IgG purified from rabbit immune serum raised against HPV-5, -8 or -92. Incubation with IgG from pre-immune serum of the same animals served as appropriate controls. After the incubation period, the PsV/IgG mixtures were placed on 293 TT cells and further cultivated for 72 h. Finally, supernatants were harvested and clarified by centrifugation at 1500 g for 5 min. The secreted alkaline phosphatase (SEAP) content in the clarified supernatants was determined using p-nitrophenyl phosphate (Sigma) dissolved in diethanolamine (Acros Organics). Absorbance values were measured at 405 nm, corresponding to the amount of SEAP in the wells. Immune IgG dilutions showing at least 50 % reduction in SEAP

Journal of General Virology 90

Serological relationship of HPV-5, -8 and -92 activity compared with pre-immune IgG were considered to be neutralizing. Data shown are the means±SD of triplicate wells of a representative experiment.

RESULTS The major capsid protein L1 of HPV-92 selfassembles into highly immunogenic VLPs To determine whether the major capsid protein L1 of the new HPV type 92 had the intrinsic capacity to self-assemble into VLPs, similar to other HPV types, HPV-92 L1 was expressed by recombinant baculoviruses in Sf9 insect cells and particles were purified by density-gradient centrifugation. Denatured HPV-92 L1 VLPs migrated on 10 % SDSPAGE as an approximately 55–60 kDa protein (Fig. 2a), corresponding to the expected molecular mass for PV L1 proteins. L1 of HPV-5 and -8, although similar in predicted size and molecular mass, migrated slightly differently, possibly due to different post-translational modifications. Analysis by transmission electron microscopy revealed spherical structures of approximately 50–55 nm in diameter composed of capsomers (the pentamer subunit of VLP, consisting of five L1 molecules) (Fig. 2b). The spherical particles were morphologically similar to the

VLPs of other HPV types (Kirnbauer et al., 1992). In addition, assembly of capsomers into long rod-like structures, as well as incomplete particles, was observed. The major capsid proteins of HPV-5 and -8 also selfassembled into VLPs, as described previously (data not shown) (Favre et al., 2000; Stark et al., 1998). To determine the immunogenicity of HPV-92 VLPs, a New Zealand White rabbit was immunized. Serum samples were taken prior to and after immunizations and tested in an ELISA using intact HPV-92 VLPs as the antigen. HPV-92 VLPs readily induced specific Abs with high titres of .600 000, even after administration of the first dose (data not shown). Similarly, HPV-5 and -8 VLPs were also highly immunogenic when administered by a similar protocol. The rabbit immune sera showed high Ab titres (.600 000) in the ELISA assay against HPV-5 and -8 VLPs. All preimmune sera lacked immunoreactivity. Detection of linear cross-reactive epitopes present in HPV-5, -8 and -92 VLPs To determine whether the VLPs of HPV-5, -8 and -92 shared linear cross-reactive epitopes, purified VLPs were analysed by Western blotting. Polyclonal rabbit immune sera obtained by immunization with HPV VLPs were used as the primary Ab. Immune serum to each of the three HPV VLP types was able to recognize the homologous VLP preparations by Western blotting, showing a prominent band with the expected molecular mass of approximately 55–60 kDa. In addition, each immune serum reacted with the L1 protein of the other two HPV types. Fig. 2(c) shows a representative Western blot using rabbit immune serum (at a dilution of 1 : 2000) raised against HPV-5 VLPs, which recognized the L1 of purified HPV-5, -8 and -92 VLP preparations at the expected molecular masses. Abs raised against HPV-5 and -8 display HAI activity against HPV-5, -8 and -92

Fig. 2. (a) Following expression of the major capsid protein, L1, in Sf9 insect cells, self-assembled VLPs of HPV-5, -8 and -92 were purified by density-gradient centrifugation. Preparations were analysed by SDS-PAGE and Coomassie Brilliant Blue staining. Bands corresponding to the expected size of L1 and molecular mass markers (M) are indicated. (b) Transmission electron microscopy of a negatively stained HPV-92 VLP preparation, demonstrating spherical particles with a diameter of 50 nm. In addition, rod-like assemblies from capsomers, as well as smaller and incomplete particles, are visible. Bar, 200 nm. (c) Western blot analysis of purified HPV-5, -8 and -92 L1 VLP preparations, using rabbit immune serum raised against HPV-5 VLPs (diluted 1 : 2000). Faster-migrating bands are probably degradation products of the full-length L1 protein. Molecular mass markers (M) are indicated. http://vir.sgmjournals.org

Incubation of PV virions or VLPs with mouse erythrocytes leads to HA due to cellular receptor binding (Favre et al., 1974; Roden et al., 1996). The HAI assay detects Abs that bind to correctly assembled capsids and appears to correlate with inhibition of virion binding to extracellular matrix (Day et al., 2007). This assay does not measure nonneutralizing Abs to internal or denatured epitopes, but represents a stringent surrogate assay to detect neutralizing Abs and thus has been useful for the examination of serological relatedness of PVs and for measuring protective Ab titres after VLP vaccination (Roden et al., 1996). IgG purified from rabbit sera raised against the VLPs of HPV-5, -8 or -92 were next tested for their HAI activity to homologous and heterologous HPV VLP types (Table 1). IgG from the respective pre-immune sera served as negative controls. 139

A. Handisurya and others

As expected rabbit immune IgG to HPV-5, -8 or -92 showed HAI of homologous VLPs. Unexpectedly, antiHPV-5 IgG also revealed HAI of heterologous HPV-8 and -92 VLPs. Similarly, IgG to HPV-8 VLP displayed HAI activity to VLPs of heterologous HPV-5 and -92. In contrast, anti-HPV-92 IgG displayed HAI activity restricted to HPV-92 VLPs, and not to HPV-5 or -8 VLPs. As a control, a non-neutralizing mAb against HPV-5 VLPs did not reveal activity in the three HAI assays (data not shown). These data are the first to demonstrate that b-PV VLPs bind to mouse erythrocytes similarly to the genital HPV types, suggesting a common receptor. To verify the specificity of the results, IgGs were further tested for HAI using VLPs of BPV-1. No HAI activity was observed for IgG raised against HPV-5, -8 or -92 (Table 1), whereas neutralizing antiserum raised against BPV-1 VLPs inhibited HA induced by BPV-1 VLPs. The haemagglutinating VLPs of BK virus were used to exclude non-specific serum HA of the IgG purified sera (data not shown). Finally, to rule out the possibility that insect cell proteins contaminating VLP preparations non-specifically mediated HA, a mock VLP preparation derived from Sf9 insect cells infected with mutant (prototype) HPV-16 L1, known to fail to assemble into intact particles, was included. As expected, this preparation did not induce HA of mouse erythrocytes, thus ruling out false-positive HA by insect cell proteins (data not shown) (Roden et al., 1996). To determine HAI titres of immune IgG, VLPs of HPV-5, -8 and -92 were dispersed at identical concentrations, and serially diluted purified IgG was added. IgG to HPV-5 VLPs displayed HAI activity against homologous HPV-5 VLPs with a titre of 3200, but lower titres of 200–400 against VLPs of HPV-8 and -92 (Table 1). Remarkably, anti-HPV-8 IgG showed equal HAI titres of 800 to HPV-8 and heterologous HPV-5 VLPs and HPV-92 VLPs. In contrast, anti-HPV-92 VLP IgG showed HAI restricted to the homologous HPV-92 VLP preparation, with an HAI titre of 1600 (Table 1). The observed differences in HAI titres to homologous types might be attributed, at least in Table 1. HAI activity (titre) Purified IgG*

Anti-HPV-5 Anti-HPV-8 Anti-HPV-92 Anti-BPV-1D

VLP type HPV-5

HPV-8

HPV-92

BPV-1D

3200 800 2 2

200 800 2 2

200–400 800 1600 2

2 2 2 ND

*IgG purified from immune serum raised against HPV-5, -8 or -92 VLPs was tested for its ability to inhibit VLP-induced HA of mouse erythrocytes. DBPV-1 VLPs and IgG purified from anti-BPV-1 rabbit immune serum served as controls. ND, Titre not determined; 2, negative result. 140

part, to variable efficiency of assembly for VLPs used in the assays. HPV-5 PsV neutralization assay Propagation of PVs is difficult to achieve in keratinocyte raft cultures and is limited to a few types that do not include HPV-5, -8 and -92. Furthermore, infection with native HPV virions can only be quantified by RT-PCR, making neutralization assays complex. These limitations have been partially overcome by the recent generation of PV-based gene-transfer vectors, also known as PsVs, that can readily be utilized in quantitative neutralization assays (Buck et al., 2005). Currently, few HPV PsV types are available, and include HPV-5 only for the genus Betapapillomavirus (kindly provided by Christopher Buck, National Cancer Institute, MD, USA). Thus, we used an HPV-5 PsV-based assay to detect (cross-)neutralizing Abs in sera raised against HPV-5, -8 or -92 VLPs. Serial dilutions of IgG purified from rabbit sera immunized with HPV-5, -8 or -92 VLPs, ranging up to a dilution of 1 : 819 200, were tested for their ability to neutralize HPV-5 PsVs. IgG purified from rabbit serum immunized with HPV-5 VLPs effectively neutralized HPV-5 PsVs with a titre of 1 : 51 200 (Fig. 3a). In accordance with the results obtained by HAI assays, anti-HPV-8 IgG also neutralized heterologous HPV-5 PsVs, with a mean neutralization titre of 1 : 12 800, indicating effective cross-neutralization (Fig. 3b). IgG to HPV-92 did not neutralize HPV-5 PsV (Fig. 3c), corroborating the results of the HAI assays.

DISCUSSION This study showed that VLPs of skin HPV types 5, 8 and 92 are highly immunogenic and elicit high-titre antisera similar to the VLPs of genital-mucosal types. Using HAI as stringent surrogate assay, induced antisera scored positively with their respective (homologous) types, suggesting the induction of neutralizing Abs. Unexpectedly, the HAI assay detected activity to the respective heterologous type in sera induced by HPV-5 and -8 VLPs, and also to the more distantly related HPV92, indicating the presence of cross-reactive epitopes on the surface of all three VLP types tested. HAI activity represents a well-established surrogate marker for detection of neutralizing antisera that correlate with protection from HPV infection and associated disease in vivo. Thus, whether HPV VLPs induce HA cross-inhibiting antisera should have a predictive value for cross-protection following vaccination. This raises the possibility of b-PV VLPs as candidates for a prophylactic skin-type HPV vaccine, aimed at reducing the incidence of NMSC, if a close link to HPV infection can be established conclusively. Importantly, and in contrast to anogenital cancer, in NMSC of immunocompetent patients no prevalent highrisk HPV types have emerged so far. Thus, induction of cross-neutralizing immune responses to several heterologJournal of General Virology 90

Serological relationship of HPV-5, -8 and -92

neutralization, as antiserum to HPV-31 VLPs crossneutralizes HPV-33 pseudovirions (78.4 % L1 identity) more efficiently than those of the more closely related HPV-16 (83.5 % L1 identity) (Giroglou et al., 2001). Abs reactive to VLPs of HPV-5 variants have been detected in dermatological proliferative disorders, although their neutralizing activity has not been evaluated due to the lack of appropriate assays (Favre et al., 2000). In addition, divergent HPV-16 variants have been found to be serologically cross-reactive (Cheng et al., 1995). Although immunity induced by VLP vaccination is predominantly type-restricted, recent results from human vaccine trials and the high degree of amino acid homology between L1 proteins of HPV-5 and -8 (87 %) have indicated that vaccination with HPV-5 VLPs might also induce crossneutralization to HPV-8 and vice versa, thus providing biologically relevant cross-protection. In contrast, given the more distant relationship between the L1s of HPV-92 and -5 (71 %), and HPV-92 and -8 (71 %), our finding that immunization with HPV-5 or -8 VLPs generated Abs that inhibited HA by HPV-92 VLPs was unexpected. b-PV types reveal 81–90 % sequence identity for L1

Fig. 3. Neutralization of HPV-5 PsVs by IgG raised against HPV5, -8 and -92 VLPs. HPV-5 PsVs were incubated in the presence of the indicated dilutions of IgG purified from immune sera (shaded bars) obtained from rabbits immunized with HPV-5 (a), HPV-8 (b) or HPV-92 (c) VLPs. IgG from pre-immune sera (open bars) from the same animals served as appropriate controls.

ous HPV types may be of outstanding significance for vaccine strategies, to provide broad protection against the multiple types found in skin tumours. Two recently introduced HPV vaccines, Gardasil and Cervarix, comprising VLPs of low-risk (HPV-6 and -11) and/or high-risk (HPV-16 and -18) genital types have shown predominantly type-restricted efficacy in preventing persistent genital HPV infection and associated disease caused by the targeted types (Garland et al., 2007; Paavonen et al., 2007). Protection correlated with the induction of high-titre neutralizing antisera. In addition, limited cross-neutralization in vitro and cross-protection in vivo have been observed for closely related types, e.g. HPV6 and -11, HPV-16 and -33, HPV-31 and -33, and HPV-18 and -45 (Giroglou et al., 2001; Harper et al., 2006; Roden et al., 1996; White et al., 1998). These genotypes showed 92.4, 80.9, 78.4 and 87.4 % amino acid sequence identity in L1, respectively. However, the degree of sequence identity is not necessarily a good predictor of the efficiency of crosshttp://vir.sgmjournals.org

proteins of types 12, 19, 25, 36 and 47 when compared with L1 of HPV-5 or -8. Furthermore, 71–80 % sequence identity is shared between the L1s of HPV-5 and -8 and those of HPV-20, -24, -49, -75, -76, -93 and -96. Another cutaneous HPV type putatively involved in NMSC development, HPV-38, shares 69 and 70 % identity to L1 of HPV-5 and -8, respectively. Considering the high degree of similarity in the L1 major capsid proteins, which are the major target for neutralizing Abs, it is tempting to speculate that vaccination with, for example, HPV-5 or -8 VLPs might induce cross-neutralization to additional closely related cutaneous types. This hypothesis can be tested once the VLPs or PsVs of other skin HPV types become available. Further determination of the serological relationship between members of the b-PVs may offer interesting information about possible vaccine candidates targeting the multiplicity of skin HPV types. HAI titres induced by immunization with HPV-5 VLPs were 8–16-fold lower to heterologous HPV-8 or -92 VLPs compared with homologous HPV-5, suggesting that these VLPs share a fraction of the cross-reactive epitopes. Of note, IgG to HPV-8 VLPs displayed comparable HAI titres to HPV-8 VLPs, and also to HPV-5 and -92 VLPs. In contrast, HAI activity of the antiserum raised against HPV92 VLPs was type-restricted to HPV-92, and did not crossinhibit HA by HPV-5 and -8 VLPs. These results may indicate that HPV-92 VLPs display cross-reactive epitope(s) in an immuno-subdominant fashion. It is noteworthy that the HAI assay is stringent in detecting only a fraction of neutralizing Abs that prevent VLP binding to extracellular matrix, but not those Abs that neutralize by a mechanism after cell attachment, which might lead to an underestimate of the neutralization capacity of the sera (Day et al., 2007; Roden et al., 1996). 141

A. Handisurya and others

Native HPV virions are notoriously difficult to propagate in vitro; thus, neutralization assays using PsV technology have been used to detect neutralizing Abs (Buck et al., 2005). However, only a limited number of HPV PsV types are available, among these HPV-5 but not HPV-8 or -92 PsVs. Neutralization assays using HPV-5 PsVs, which were produced in mammalian cells independently by another laboratory (Buck et al., 2005), corroborated the findings obtained by HAI assays. IgG raised against HPV-5 and -8 VLPs did (cross)-neutralize HPV-5 PsVs at a comparable titre, whereas antisera to HPV-92 VLPs were nonneutralizing in this assay. Our study demonstrates that b-PVs HPV-5, -8 and -92 are related serotypes, although representing distinct genotypes. If a close link to HPV infection can be established conclusively, these findings may be relevant for further studies on the development of a broad-spectrum vaccine to prevent b-PV infection of the skin, with the ultimate goal of decreasing the incidence of NMSC in both immunosuppressed and immunocompetent individuals.

ACKNOWLEDGEMENTS We thank John Schiller for providing BK virus VLPs, Christopher Buck for the HPV-5 PsV vectors, Subhashini Jagu for help with the HPV-5 PsV neutralization assays, Petra Lu¨hrs and Maria Sibilia for supply of mouse erythrocytes, and Doug Lowy and Richard Roden for critical reading of the manuscript. This study was supported by a grant to R. K. by the Austrian Science Foundation (FWF) P-18990P13.

High frequency of detection of epidermodysplasia verruciformisassociated human papillomavirus DNA in biopsies from malignant and premalignant skin lesions from renal transplant recipients. J Invest Dermatol 105, 367–371. de Villiers, E. M., Fauquet, C., Broker, T. R., Bernard, H. U. & zur Hausen, H. (2004). Classification of papillomaviruses. Virology 324,

17–27. Diepgen, T. L. & Mahler, V. (2002). The epidemiology of skin cancer.

Br J Dermatol 146 (Suppl 61), 1–6. Favre, M., Breitburd, F., Croissant, O. & Orth, G. (1974).

Hemagglutinating activity of bovine papilloma virus. Virology 60, 572–578. Favre, M., Orth, G., Majewski, S., Baloul, S., Pura, A. & Jablonska, S. (1998). Psoriasis: a possible reservoir for human papillomavirus type

5, the virus associated with skin carcinomas of epidermodysplasia verruciformis. J Invest Dermatol 110, 311–317. Favre, M., Majewski, S., Noszczyk, B., Maienfisch, F., Pura, A., Orth, G. & Jablonska, S. (2000). Antibodies to human papillomavirus

type 5 are generated in epidermal repair processes. J Invest Dermatol 114, 403–407. Feltkamp, M. C., Broer, R., di Summa, F. M., Struijk, L., van der Meijden, E., Verlaan, B. P., Westendorp, R. G., ter Schegget, J., Spaan, W. J. & Bouwes Bavinck, J. N. (2003). Seroreactivity to

epidermodysplasia verruciformis-related human papillomavirus types is associated with nonmelanoma skin cancer. Cancer Res 63, 2695– 2700. Forslund, O., Antonsson, A., Higgins, G., Ly, H., Delius, H., Hunziker, A. & de Villiers, E. M. (2003a). Nucleotide sequence and phylogenetic

classification of candidate human papilloma virus type 92. Virology 312, 255–260. Forslund, O., DeAngelis, P. M., Beigi, M., Schjolberg, A. R. & Clausen, O. P. (2003b). Identification of human papillomavirus in

keratoacanthomas. J Cutan Pathol 30, 423–429.

REFERENCES Antonsson, A., Forslund, O., Ekberg, H., Sterner, G. & Hansson, B. G. (2000). The ubiquity and impressive genomic diversity of

human skin papillomaviruses suggest a commensalic nature of these viruses. J Virol 74, 11636–11641.

Forslund, O., Ly, H., Reid, C. & Higgins, G. (2003c). A broad spectrum

of human papillomavirus types is present in the skin of Australian patients with non-melanoma skin cancers and solar keratosis. Br J Dermatol 149, 64–73. Forslund, O., Iftner, T., Andersson, K., Lindelof, B., Hradil, E., Nordin, P., Stenquist, B., Kirnbauer, R., Dillner, J. & de Villiers, E. M. (2007).

Bedard, K. M., Underbrink, M. P., Howie, H. L. & Galloway, D. A. (2008). The E6 oncoproteins from human betapapillomaviruses

Cutaneous human papillomaviruses found in sun-exposed skin: betapapillomavirus species 2 predominates in squamous cell carcinoma. J Infect Dis 196, 876–883.

differentially activate telomerase through an E6AP-dependent mechanism and prolong the lifespan of primary keratinocytes. J Virol 82, 3894–3902.

Garland, S. M., Hernandez-Avila, M., Wheeler, C. M., Perez, G., Harper, D. M., Leodolter, S., Tang, G. W., Ferris, D. G., Steben, M. & other authors (2007). Quadrivalent vaccine against human papillo-

Boxman, I. L., Berkhout, R. J., Mulder, L. H., Wolkers, M. C., Bouwes Bavinck, J. N., Vermeer, B. J. & ter Schegget, J. (1997). Detection of

human papillomavirus DNA in plucked hairs from renal transplant recipients and healthy volunteers. J Invest Dermatol 108, 712–715. Buck, C. B., Pastrana, D. V., Lowy, D. R. & Schiller, J. T. (2005).

Generation of HPV pseudovirions using transfection and their use in neutralization assays. Methods Mol Med 119, 445–462. Cheng, G., Icenogle, J. P., Kirnbauer, R., Hubbert, N. L., St Louis, M. E., Han, C., Svare, E. I., Kruger Kjaer, S., Lowy, D. R. & Schiller, J. T. (1995). Divergent human papillomavirus type 16 variants are

serologically cross-reactive. J Infect Dis 172, 1584–1587. Day, P. M., Thompson, C. D., Buck, C. B., Pang, Y. Y., Lowy, D. R. & Schiller, J. T. (2007). Neutralization of human papillomavirus with

mavirus to prevent anogenital diseases. N Engl J Med 356, 1928–1943. Giroglou, T., Sapp, M., Lane, C., Fligge, C., Christensen, N. D., Streeck, R. E. & Rose, R. C. (2001). Immunological analyses of

human papillomavirus capsids. Vaccine 19, 1783–1793. Handisurya, A., Gilch, S., Winter, D., Shafti-Keramat, S., Maurer, D., Schatzl, H. M. & Kirnbauer, R. (2007). Vaccination with prion

peptide-displaying papillomavirus-like particles induces autoantibodies to normal prion protein that interfere with pathologic prion protein production in infected cells. FEBS J 274, 1747–1758. Harper, D. M., Franco, E. L., Wheeler, C. M., Moscicki, A. B., Romanowski, B., Roteli-Martins, C. M., Jenkins, D., Schuind, A., Costa Clemens, S. A. & Dubin, G. (2006). Sustained efficacy up to 4.5

monoclonal antibodies reveals different mechanisms of inhibition. J Virol 81, 8784–8792.

years of a bivalent L1 virus-like particle vaccine against human papillomavirus types 16 and 18: follow-up from a randomised control trial. Lancet 367, 1247–1255.

de Jong-Tieben, L. M., Berkhout, R. J., Smits, H. L., Bouwes Bavinck, J. N., Vermeer, B. J., van der Woude, F. J. & ter Schegget, J. (1995).

Harwood, C. A., Surentheran, T., Sasieni, P., Proby, C. M., Bordea, C., Leigh, I. M., Wojnarowska, F., Breuer, J. & McGregor, J. M. (2004).

142

Journal of General Virology 90

Serological relationship of HPV-5, -8 and -92 Increased risk of skin cancer associated with the presence of epidermodysplasia verruciformis human papillomavirus types in normal skin. Br J Dermatol 150, 949–957.

papillomavirus gene expression in cutaneous squamous cell carcinomas from immunosuppressed and immunocompetent individuals. J Invest Dermatol 125, 98–107.

Howley, P. M. & Lowy, D. R. (2007). Papillomavirus. In Fields Virology, 5th edn, pp. 2299–2354. Edited by D. M. Knipe & P. M. Howley. Philadelphia: Lippincott Williams and Wilkins.

Ramoz, N., Rueda, L. A., Bouadjar, B., Montoya, L. S., Orth, G. & Favre, M. (2002). Mutations in two adjacent novel genes are

Jackson, S., Harwood, C., Thomas, M., Banks, L. & Storey, A. (2000).

associated with epidermodysplasia verruciformis. Nat Genet 32, 579–581.

Role of Bak in UV-induced apoptosis in skin cancer and abrogation by HPV E6 proteins. Genes Dev 14, 3065–3073.

Roden, R. B., Hubbert, N., Kirnbauer, R., Christensen, N., Lowy, D. & Schiller, J. (1996). Assessment of the serological relatedness of genital

Kirnbauer, R., Booy, F., Cheng, N., Lowy, D. R. & Schiller, J. T. (1992).

human papillomaviruses by hemagglutination inhibition. J Virol 70, 3298–3301.

Papillomavirus L1 major capsid protein self-assembles into virus-like particles that are highly immunogenic. Proc Natl Acad Sci U S A 89, 12180–12184.

Schaper, I. D., Marcuzzi, G. P., Weissenborn, S. J., Kasper, H. U., Dries, V., Smyth, N., Fuchs, P. & Pfister, H. (2005). Development of

Kirnbauer, R., Taub, J., Greenstone, H., Roden, R., Du¨rst, M., Gissmann, L., Lowy, D. R. & Schiller, J. T. (1993). Efficient self-

skin tumors in mice transgenic for early genes of human papillomavirus type 8. Cancer Res 65, 1394–1400.

assembly of human papillomavirus type 16 L1 and L1–L2 into viruslike particles. J Virol 67, 6929–6936.

Slupetzky, K., Shafti-Keramat, S., Lenz, P., Brandt, S., Grassauer, A., Sara, M. & Kirnbauer, R. (2001). Chimeric papillomavirus-like

Kirnbauer, R., Hubbert, N. L., Wheeler, C. M., Becker, T. M., Lowy, D. R. & Schiller, J. T. (1994). A virus-like particle enzyme-linked

particles expressing a foreign epitope on capsid surface loops. J Gen Virol 82, 2799–2804.

immunosorbent assay detects serum antibodies in a majority of women infected with human papillomavirus type 16. J Natl Cancer Inst 86, 494–499.

Stark, S., Petridis, A. K., Ghim, S. J., Jenson, A. B., Bouwes Bavinck, J. N., Gross, G., Stockfleth, E., Fuchs, P. G. & Pfister, H. (1998).

Kirnbauer, R., Lenz, P. & Okun, M. M. (2008). Human papillomavirus.

In Dermatology, pp. 1183–1198. Edited by J. Bolognia, J. Jorizzo & R. Rapini. London: Mosby. Orth, G. (2006). Genetics of epidermodysplasia verruciformis: insights

into host defense against papillomaviruses. Semin Immunol 18, 362–374. Paavonen, J., Jenkins, D., Bosch, F. X., Naud, P., Salmeron, J., Wheeler, C. M., Chow, S. N., Apter, D. L., Kitchener, H. C. & other authors (2007). Efficacy of a prophylactic adjuvanted bivalent L1

virus-like-particle vaccine against infection with human papillomavirus types 16 and 18 in young women: an interim analysis of a phase III double-blind, randomised controlled trial. Lancet 369, 2161–2170. Pfister, H. (2003). Chapter 8: human papillomavirus and skin cancer.

J Natl Cancer Inst Monogr 31, 52–56. Purdie, K. J., Surentheran, T., Sterling, J. C., Bell, L., McGregor, J. M., Proby, C. M., Harwood, C. A. & Breuer, J. (2005). Human

http://vir.sgmjournals.org

Prevalence of antibodies against virus-like particles of Epidermodysplasia verruciformis-associated HPV8 in patients at risk of skin cancer. J Invest Dermatol 111, 696–701. Termorshuizen, F., Feltkamp, M. C., Struijk, L., de Gruijl, F. R., Bavinck, J. N. & van Loveren, H. (2004). Sunlight exposure and

(sero)prevalence of epidermodysplasia verruciformis-associated human papillomavirus. J Invest Dermatol 122, 1456–1462. White, W. I., Wilson, S. D., Bonnez, W., Rose, R. C., Koenig, S. & Suzich, J. A. (1998). In vitro infection and type-restricted antibody-

mediated neutralization of authentic human papillomavirus type 16. J Virol 72, 959–964. Zamora, E., Handisurya, A., Shafti-Keramat, S., Borchelt, D., Rudow, G., Conant, K., Cox, C., Troncoso, J. C. & Kirnbauer, R. (2006). Papillomavirus-like particles are an effective platform for amyloid-b immunization in rabbits and transgenic mice. J Immunol

177, 2662–2670.

143