|Home | About | Journals | Submit | Contact Us | Français|
Human papillomavirus (PV) (HPV) types 2, 27, and 57 are closely related and, hence, represent a promising model system to study the correlation of phylogenetic relationship and immunological distinctiveness of PVs. These HPV types cause a large fraction of cutaneous warts occurring in immunocompromised patients. Therefore, they constitute a target for the development of virus-like particle (VLP)-based vaccines. However, the immunogenic structure of HPV type 2, 27, and 57 capsids has not been studied yet. Here we provide, for the first time, a characterization of the B-cell epitopes on VLPs of cutaneous alpha-HPVs using a panel of 94 monoclonal antibodies (MAbs) generated upon immunization with capsids from HPV types 2, 27, and 57. The MAbs generated were characterized regarding their reactivities with glutathione S-transferase-L1 fusion proteins from 18 different PV types, the nature of their recognized epitopes, their isotypes, and their ability to neutralize HPV type 2, 27, 57, or 16. In total, 33 of the 94 MAbs (35%) showed type-specific reactivity. All type-specific MAbs recognize linear epitopes, most of which map to the hypervariable surface loop regions of the L1 amino acid sequence. Four of the generated MAbs neutralized pseudovirions of the inoculated HPV type efficiently. All four MAbs recognized epitopes within the BC loop, which is required and sufficient for their neutralizing activity. Our data highlight the immunological distinctiveness of individual HPV types, even in comparison to their closest relatives, and they provide a basis for the development of VLP-based vaccines against cutaneous alpha-HPVs.
Recently licensed prophylactic vaccines confer efficient protection against infections by human papillomavirus (PV) (HPV) types 16 and 18, thereby aiming to prevent approximately 70% of all cervical cancer cases (17, 39). These vaccines are composed of virus-like particles (VLPs), which spontaneously assemble from the major capsid protein L1 via 72 pentamers (capsomeres) as subunits (2, 23, 26).
In the process of vaccine development, monoclonal antibodies (MAbs) proved to be valuable tools for the immunological analysis of recombinantly produced capsids and capsomeres (51) as well as for serological studies (25, 49, 56). Moreover, the identification and characterization of many neutralizing epitopes of HPV types 11 and 16 have been facilitated by the employment of MAbs (6, 11, 30-32, 41, 42, 55). Such epitopes to neutralizing antibodies are mostly conformation dependent, but a few neutralizing MAbs that recognize linear epitopes have also been generated (16, 18). Most neutralizing MAbs are HPV type specific due to the hypervariable nature of their respective epitopes, which typically reside in the surface-exposed loop regions of the L1 protein (10). In contrast, cross-reactive MAbs targeting rather conserved L1 epitopes are generally nonneutralizing.
HPV types 2, 27, and 57 are the three members of Alphapapillomavirus species 4 (20). They are very closely related, and HPV types 2 and 27 hardly fulfill the requirement of more than 10% nucleotide variation in the L1 open reading frame to be classified as distinct types (8). Therefore, they represent a promising model system to study the immunological distinctiveness of closely related HPV types. Pathologically, HPV types 2, 27, and 57 infect primarily the cutaneous epithelia, thereby causing common skin warts, which often occur ubiquitously and confluently in immunocompromised patients (1, 24, 28). It is our long-term goal to develop a prophylactic L1 VLP-based vaccine to alleviate the burden provoked by HPV-induced skin lesions in these patients. However, to date, neither the structure nor the immunogenicity of HPV type 2, 27, and 57 capsids has been elucidated.
The purpose of the present study was twofold. First, we sought to generate MAbs specific for HPV types 2, 27, and 57 as tools for type-specific diagnostic assays. Second, we aimed to exploit the generated MAbs for an investigation of the B-cell epitopes on capsids of HPV types 2, 27, and 57.
Baculoviruses recombinant for wild-type (WT) or mutant HPV type 2, 27, and 57 L1 genes were generated by using a protocol outlined previously (37). All point mutations in the L1 genes (C172S and C422S for HPV type 2, C173S and C424S for HPV type 27, and C173S and C423S for HPV type 57) were introduced using the QuikChange multisite-directed mutagenesis kit (Stratagene). The L1 genes were introduced into transfer plasmid pVL1392 (Invitrogen) by PCR amplification with primers inserting the restriction sites NotI and EcoRI. All constructs were sequenced.
For the production of recombinant Autographa californica nuclear polyhedrosis viruses (AcNPVs), 2 μg of the respective transfer plasmid and 0.2 μg of linearized DiamondBac baculovirus DNA (Sigma) were cotransfected by calcium phosphate precipitation into 5 × 106 Sf9 cells. All recombinant A. californica viruses were amplified at least three times before they were used for productive infections. The titers of all AcNPVs were determined by a plaque assay as described previously (33).
HPV type 2, 27, and 57 VLPs and capsomeres were produced as described previously (38). Briefly, 2 × 108 Trichoplusia ni High Five cells (Invitrogen) were infected at a multiplicity of infection of 2 with baculovirus recombinant for WT L1 genes or mutant L1 genes for the production of VLPs and capsomeres, respectively (29, 44). After a 3-day incubation, insect cells were harvested and lysed by sonication. The lysate was cleared by centrifugation and layered onto a two-step gradient with 40% sucrose on top of a 57.5% CsCl solution. Subsequently, the gradient was centrifuged at 96,500 × g at 10°C in a Beckman SW32 rotor for 3 h, the sucrose cushion was discarded, and the cesium chloride fraction was transferred into a Quick-Seal tube (Beckman) and centrifuged again for 16 h at 184,000 × g at 20°C in a Sorval TFT 65.13 rotor. The gradient was fractionated into 1-ml aliquots, and the purity and L1 content were determined by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) followed by Coomassie staining. The L1 protein concentration was determined by a Bradford assay (1a) with Rotiquant (Roth).
For further VLP purification, the peak fractions were pooled and dialyzed against 50 mM HEPES (pH 7.4)-0.3 M NaCl and centrifuged for 10 min at 20,000 × g and 4°C. Subsequently, heparin affinity chromatography was carried out using 1-ml HiTrap columns (GE Healthcare). VLPs were eluted with 50 mM HEPES (pH 7.4)-1 M NaCl and analyzed by SDS-PAGE followed by Coomassie staining and Western blot analysis. The concentration of the L1 protein in the eluates was determined by a Bradford assay (1a) and by comparison to a bovine serum albumin standard on a Coomassie-stained SDS-PAGE gel. The structure of the particles was confirmed by electron microscopy.
BALB/c mice were immunized three times in 4-week intervals subcutaneously with 10 μg of VLPs emulsified in Freund's complete adjuvant for the first immunization or Freund's incomplete adjuvant for the booster immunizations. Mice were boosted again intraperitoneally with 10 μg of antigen devoid of any adjuvant 3 days before the immunization end point. In total, five, three, and two mice were immunized with capsids of HPV types 2, 27, and 57, respectively. Mice were sacrificed, and spleen cells were isolated and fused with myeloma SP2/0-Ag14 cells (ratio, 5:1) by using polyethylene glycol (Sigma-Aldrich) as described previously (27). The reactivity of the hybridoma supernatants (from 9,200, 6,600, and 4,400 clones upon immunization with HPV type 2, 27, and 57 VLPs, respectively) with HPV type 2, 27, and 57 glutathione S-transferase (GST)-L1 fusion proteins was screened by enzyme-linked immunosorbent assay (ELISA). The strongest-reacting hybridoma clones were isolated and expanded by subcloning.
All PV L1 open reading frames were expressed as fusion proteins with GST at the N terminus and a peptide consisting of the last 11 amino acids from the simian virus 40 large T antigen (TAg) at the C terminus from pGEX4T3 plasmids (GE Healthcare) in Escherichia coli Rosetta cells (Merck) as described previously (35). Expression constructs for HPV type 6, 16, and 18 L1 were characterized previously by Sehr et al. (46); expression constructs for HPV types 1a, 2a, 3, 4, 5, 10, 27, 38, 41, 57, and 77 were specified previously by Michael et al. (35); the HPV type 7 L1 expression construct was described previously by Casabonne et al. (7); and expression constructs for HPV types 11 and 32 were reported previously by Waterboer et al. (52). A multiplex assay for the simultaneous analysis of reactivity with various GST-L1-TAg constructs was previously described (53, 54). Briefly, the GST-L1-TAg fusion proteins of different HPV types from the cleared bacterial lysates were coupled to distinctly fluorescence-labeled polystyrene beads (Luminex). Subsequently, all antigen-loaded bead sets were pooled and incubated for 1 h at 37°C with hybridoma supernatants, which had been serially diluted in blocking buffer. Bound MAbs were stained with biotinylated goat anti-mouse antibody and streptavidin-R-phycoerythrin (Molecular Probes) as a reporter. By using an xMAP Luminex analyzer, the individual bead sets were identified, and the corresponding reporter fluorescence activities were quantified. Fluorescence intensities were considered positive when exceeding the cutoff, defined as the autofluorescence of the individual bead sets (fluorescence upon incubation with culture supernatant of 293T cells [Invitrogen]) plus twice the determined background for each hybridoma supernatant (fluorescence for GST-TAg as the antigen).
All incubation steps were carried out for 1 h at 37°C unless indicated otherwise. For the blocking and dilution of antibodies, 5% skim milk in phosphate-buffered saline (PBS)-0.05% Tween 20 (PBS-T) was used; after each incubation step, plates were washed four times with PBS-T; assays were developed with horseradish peroxidase-conjugated goat anti-mouse antibody (Dianova) and 2,2′-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) (ABTS; Sigma) substrate; the absorbance at 405 nm was measured by using an ELISA reader, the supernatant of a hybridoma clone raised with HPV type 16 E6 as an antigen was used as negative control in triplicate (43); and the cutoff was defined as the absorbance values of the negative controls plus three times the standard deviation.
Polyclonal rabbit anti-HPV type 57 VLP serum was blocked with WT AcNPV-infected insect cell lysate, protein G purified (GE Healthcare), and used to coat immunosorbent 96-well plates (Nunc) overnight at 4°C. Subsequently, plates were blocked, and 100 ng VLPs per well was added. Hybridoma supernatants were applied in 1:10 dilutions. The assay was developed, and the absorbance was measured.
Capsomeres were dialyzed against PBS for 1 h at room temperature (RT). Immunosorbent 96-well plates (Nunc) were coated with 100 ng capsomeres per well overnight at 4°C. Plates were blocked, and hybridoma supernatants were applied (1:10).
GST-L1 fusion proteins were produced recombinantly in E. coli Rosetta cells (Stratagene) and column purified as previously described (9). Fusion proteins were eluted from 1-ml GSTrap columns (GE Healthcare) with 20 mM l-glutathione (Sigma) in buffer L (50 mM Tris [pH 8.2], 0.2 M NaCl, 1 mM EDTA, 2 mM dithiothreitol). The concentration of L1 protein in the eluates was determined by a Bradford assay (1a) and by comparison to a bovine serum albumin standard on a Coomassie-stained SDS-PAGE gel. SDS was added to a final concentration of 0.5%, and the proteins were denatured at 90°C for 10 min. Immunosorbent 96-well plates (Nunc) were coated with 500 ng of denatured L1 protein per well. Following blocking, hybridoma supernatants were applied in 1:10 dilutions.
GST-L1 ELISAs were performed as previously reported (47). Briefly, immunosorbent 96-well plates (Nunc) were coated overnight at 4°C with 200 ng per well of glutathione-casein (Sigma) in 50 mM carbonate buffer (pH 9.6). After blocking, wells were loaded with 12.5 μg of lysates from E. coli Rosetta cells (Stratagene) overexpressing the GST-L1 fusion protein. Hybridoma supernatants were used either undiluted or in 1:10 dilutions.
The class and subclass of the MAbs in the hybridoma supernatants were determined using a mouse MAb isotyping kit (Thermo Fisher Scientific). Plates were coated with 90 ng/well of sheep anti-mouse immunoglobulin G (IgG) (Dianova) or with goat anti-mouse IgM (Invitrogen) and blocked. Hybridoma supernatants were applied undiluted and in 1:20 and 1:100 dilutions. The culture supernatant of 293T cells (Invitrogen) and fetal calf serum (Invitrogen) were used as negative controls. The assay was developed, and the absorbances of serially diluted IgM, IgG1, IgG2a, and IgG2b MAbs of known concentrations (Invitrogen) were employed to generate a standard.
Lysates of recombinant bacteria expressing HPV type 2, 3, 10, 16, 27, 57, and 77 GST-L1 fusion proteins were separated by SDS-PAGE and transferred onto polyvinylidene difluoride (Millipore) membranes according to the manufacturer's instructions. After blocking for 1 h with 5% skim milk (Roth) in PBS, the membranes were incubated for 1 h with 1:10 dilutions of the hybridoma supernatants in the blocking solution. Membranes were washed three times with PBS-0.05% Tween 20 (Sigma) and developed with horseradish peroxidase-conjugated goat anti-mouse antibody (Dianova) and enhanced chemiluminescence substrate for horseradish peroxidase detection (AppliChem).
For the production of HPV type 2, 16, 27, and 57 pseudovirions, a protocol outlined previously (4) was applied. Codon-modified (according to the “as-different-as-possible” method ) nucleotide sequences for the L1 and L2 open reading frames from HPV types 2, 27, and 57 were synthesized (HPV type 2 open reading frames from Blue Heron GenBank accession numbers FJ976666 to FJ976667] and HPV type 27 and HPV type 57 open reading frames from GenScript [GenBank accession numbers FJ976662 to FJ976665]) and cloned into a vector described as pRalw for HPV type 2 (http://home.ccr.cancer.gov/Lco/plasmids.asp). The codon-modified HPV type 16 L1 and L2 genes were amplified by PCR using bicistronic L1/L2 expression plasmid p16sheLL as a template (5) and introduced into the same vector. Mutant codon-modified L1 genes for the production of chimeric pseudovirions were generated by overlapping PCR.
The human cell line expressing high levels of simian virus 40 TAg, 293TT (3), was transfected with the L1 and L2 expression constructs along with a reporter plasmid encoding Gaussia luciferase (50) by using PolyFect transfection reagent (Qiagen). After a 72-h incubation, cells were harvested and lysed with Brij 58 (0.35% final concentration; Sigma). Benzonase was added (250 U; Merck), and capsids were matured overnight at 37°C. Subsequently, NaCl was added to a final concentration of 0.85 mol/liter.
To analyze the neutralizing capacities of the hybridoma supernatants or mouse sera, 293TT cells were seeded at a concentration of 4.5 × 104 cells per well onto 96-well plates in Dulbecco's modified Eagle's medium (Sigma) supplemented with 10% fetal calf serum (Gibco). The following day, the pseudovirions were diluted (HPV type 2, 27, and 57 pseudovirions at a 1:20,000 dilution for data shown in Fig. Fig.22 and and44 and at a 1:2,000 dilution for data presented in Fig. Fig.55 and HPV type 16 pseudovirions at a 1:500,000 dilution for data depicted in Fig. Fig.22 and and44 and at a 1:50,000 for data shown in Fig. Fig.5)5) in Dulbecco's modified Eagle's medium and incubated with serially diluted hybridoma supernatants or mouse sera for 15 min at RT before they were applied onto the 293TT cells in duplicates. After a 5-day incubation, Gaussia luciferase activity in the culture supernatants was quantified using a Gaussia luciferase substrate (New England Biolabs), black 96-well OptiPlates (Perkin-Elmer), and a luminescence reader (Victor3 1420 multilabel plate reader; Perkin-Elmer). The percent neutralization at a given hybridoma supernatant dilution was calculated using the following equation: neutralization at dilution X (%) = 100 − 100 × mean luminescence upon infection with preincubated pseudovirions at dilution X/mean luminescence upon infection with untreated pseudovirions.
For the mapping of linear epitopes, a method described previously (21) was employed. Briefly, immunosorbent 96-well plates (Nunc) were coated overnight at 37°C with 125 ng of streptavidin (Sigma) per well to let the liquid evaporate. Plates were blocked with 5% skim milk in PBS-T for 1 h at RT and washed three times with PBS-T. Five micrograms of biotin-labeled, synthetic 20-mer peptides (mimotopes) with 10-amino-acid overlaps spanning the HPV type 2, 16, 27, and 57 L1 sequences was added to the individual wells, and the plates were incubated at RT for 1 h. Subsequently, the plates were washed three times with PBS-T and incubated with the hybridoma supernatants (1:10 in blocking solution) for 1 h at RT. After three washes with PBS-T, assays were developed with horseradish peroxidase-conjugated goat anti-mouse antibody and ABTS. The absorbance at 405 nm was measured by using an ELISA reader. As negative control, the supernatant of a hybridoma clone raised with HPV type 16 E6 as an antigen was used in triplicate (43). The cutoff value was defined as absorbance values of the negative controls plus five times the standard deviation.
All MAbs were raised, screened, and expanded using HPV type 2, 27, and 57 capsids as antigens as described in Materials and Methods. In total, 58, 28, and 8 MAbs were generated upon immunization with VLPs from HPV types 2, 27, and 57, respectively. They were designated TS2, TS27, and TS57, respectively. We employed a fluorescence-labeled bead-based multiplex assay (53, 54) to characterize the MAbs for reactivity with L1 proteins of 18 different PV types representative of the genera alpha (HPV types 2, 3, 6, 7, 10, 11, 16, 18, 27, 32, 57, and 77), beta (HPV types 5 and 38), gamma (HPV type 4), delta (bovine PV type 1), mu (HPV type 1), and nu (HPV type 41). Eighty-five of the 94 tested MAbs (90%) reacted with GST-L1 fusion proteins of the HPV type used for immunization (see Table S1 in the supplemental material).
Twenty-seven (47%) of the MAbs originating from immunizations with HPV type 2 capsids reacted only with the cognate GST-L1. Six MAbs (TS2-11, TS2-12, TS2-14, TS2-15, TS2-16, and TS2-18; 6%) did not exhibit reactivity with the GST-L1 fusion proteins of HPV type 2 but reacted with the closely related type 27. Of the 24 cross-reactive MAbs, 5 (21%) reacted exclusively with GST-L1 fusion proteins of Alphapapillomavirus species 4 (HPV types 2, 27, and 57), whereas 5 MAbs (21%) were reactive with GST-L1 proteins from Alphapapillomavirus species 2 (HPV types 3, 10, and 77) and species 4. Interspecies but intragenus cross-reactivity as well as intergenus cross-reactivity were each observed for seven MAbs (29%).
Four (14%) of the MAbs generated with HPV type 27 VLPs show type-specific reactivity with the homologue GST-L1 fusion protein. One MAb reacted only with GST-L1 of HPV type 11, HPV type 16, and bovine PV type 1 (TS27-14). The 24 cross-reactive MAbs obtained after immunization with HPV type 27 VLPs are subdivided into sets showing exclusively intraspecies cross-reactivity (12 MAbs; 50%), intragenus reactivity (two MAbs; 8%), or intergenus reactivity (ten MAbs; 42%).
Out of the eight MAbs obtained upon immunization with HPV type 57 VLPs, two (25%) reacted only with the corresponding HPV type 57 GST-L1 protein. Of the six cross-reactive MAbs, five (83%) reacted with GST-L1 fusion proteins from multiple-genus alpha-HPV types. One MAb was reactive with the HPV type 57 antigen, and it weakly reacted with HPV type 27 GST-L1.
Interestingly, of the 94 MAbs generated, 48 (51%) were reactive with GST-L1 fusion proteins from HPV types 2 and 27, whereas only 15 (16%) or 14 (15%) reacted with antigens from types 2 and 57 or types 27 and 57, respectively. All 14 MAbs reacting with GST-L1 fusion proteins from HPV types 27 and 57 were also reactive with GST-L1 from HPV type 2. In summary, these data reflect the exceptionally close relatedness between HPV types 2 and 27.
All MAbs were used to further characterize the nature of the epitopes that they recognize. Reactivity with epitopes displayed on VLPs was identified by VLP capture ELISA and by immunofluorescence using insect cells infected with baculoviruses recombinant for HPV type 2, 16, 27, or 57 L1. The recognition of epitopes on capsomeres was analyzed by capsomere ELISA, and reactivity with linear epitopes was assayed by Western blotting as well as ELISA by using denatured GST-L1 as an antigen. The detailed results of this analysis are shown in Table S2 in the supplemental material.
In total, 43 (46%) and 61 (65%) MAbs recognized conformational epitopes on VLPs and capsomeres, respectively, whereas 86 (91%) MAbs reacted with linear epitopes (Fig. (Fig.1).1). All type-specific MAbs reacted with denatured L1 protein, either exclusively (three for HPV type 2 and none for HPV type 27 or 57) or in addition to capsomeres (18 for HPV type 2 and none for HPV type 27 or 57) or capsomeres and VLPs (one for HPV type 2, two for HPV type 27, and two for HPV type 57). This pattern was observed independently of the HPV type against which the MAbs were raised. However, the recognition of denatured protein was not sufficient for type specificity, as numerous MAbs originating from immunizations with HPV type 2 or 27 VLP recognized denatured L1 only or additionally to conformational capsomeres or VLPs but showed broad cross-reactivity. Surprisingly, only a few of the generated MAbs were reactive solely with conformation-dependent epitopes on the surface of HPV type 2, 16, 27, or 57 VLPs. However, these MAbs (TS2-34, TS2-44, TS2-53, TS57-2, TS57-3, TS57-4, TS57-7, and TS57-8) were cross-reactive with GST-L1 fusion proteins of several HPV types from several genera.
For many MAbs that show intraspecies cross-reactivity either alone or in addition to reactivity with GST-L1 from other HPV types, binding to linearized L1 and to native capsomeres or VLPs was observed.
To investigate the neutralizing capacities of the MAbs, we established pseudovirion-based reporter assays for HPV types 2, 27, and 57 as previously described for HPV type 16 (4). Employing Gaussia luciferase as a reporter, we analyzed the ability of the MAbs to prevent the pseudovirus-dependent delivery of the reporter gene by incubating the hybridoma supernatants with the pseudovirions prior to the inoculation of the cells. We found that four MAbs (4%) (TS27-30, TS57-1, TS57-5, and TS57-6) type-specifically neutralized the HPV type used for immunization (see Table S2 in the supplemental material). To quantify their neutralizing potential, we determined the IgG concentration required to achieve half-maximal neutralizing activity (50% inhibitory concentrations [IC50s]) of MAbs TS27-30, TS57-1, TS57-5, and TS57-6 in comparison to sera from mice that had been immunized with VLPs emulsified in Freund's adjuvant. Whereas for the mouse sera, IC50s of about 20 μM were observed, TS27-30 neutralized HPV type 27 with a value of 0.8 μM, and TS57-1, TS57-5, and TS57-6 exhibited values for HPV type 57 neutralization ranging from 0.9 to 1.3 μM (Fig. (Fig.2).2). None of the isolated MAbs was identified to neutralize HPV type 2. IC50s for MAbs TS27-30, TS57-1, TS57-5, and TS57-6 were >300 μM for the neutralization of HPV types other than the respective immunogen HPVs. These data show that TS27-30, TS57-1, TS57-5, and TS57-6 neutralize their corresponding HPV types with high efficiencies.
It was previously proposed that HPV neutralization in vivo is associated with particular antibody isotypes (57). Therefore, we determined the isotypes of the generated MAbs and analyzed their relative reactivities by considering the maximum titers reached in the multiplex assay and the immunoglobulin concentration in the respective hybridoma supernatants. Forty (43%), 6 (6%), 3 (3%), and 45 (48%) MAbs were found to be IgG1, IgG2a, IgG2b, and IgM, respectively (see Table S3 in the supplemental material). Except for one (TS27-30), all MAbs raised upon immunization with HPV type 27 VLPs were of the IgM isotype. In contrast, for the MAbs generated after immunization with VLPs from HPV types 2 and 57, only 18 and none, respectively, were found to be IgM. No evident difference in type specificity was found between IgG and IgM MAbs. However, reactivities exhibited by IgM MAbs were generally lower even though, on average, higher immunoglobulin concentrations were detected in the hybridoma supernatants (see Fig. S1 in the supplemental material). The neutralizing MAbs were isotyped as IgG2b (TS27-30, TS57-1, and TS57-6) or as IgG1 (TS57-5). Neither of these MAbs showed particularly high reactivities compared to all other IgG MAbs (see Fig. S1 in the supplemental material).
To further characterize the type-specific epitopes recognized by the MAbs generated in this study, we carried out epitope mapping as previously described (21) by using biotinylated 20-mer peptides spanning the L1 sequences of HPV types 2, 16, 27, and 57 with 10-amino-acid overlaps. For cross-reactive MAbs, epitope identification required binding to the corresponding peptides of all the recognized HPV types (only HPV types 2, 16, 27, and 57 were tested). To verify that binding to the employed peptides is specific for MAbs that recognize linear epitopes, we additionally analyzed supernatants of hybridoma clones H16.V5 and H16.E7 (a generous gift by Neil D. Christensen). Both MAbs, which were previously shown to neutralize HPV type 16 by binding to conformation-dependent epitopes (11, 42, 55), did not react with any of the peptides. Of the 86 MAbs binding to denatured L1 proteins, the linear epitopes of 44 (51%) could be identified unambiguously. The remaining epitopes could not be mapped, most likely due to stringent assay conditions, low immunoglobulin concentrations in the applied hybridoma supernatants, and incomplete coverage of all potential epitopes longer than 11 amino acids. The identified epitopes are distributed across most of the L1 protein, with clusters occurring in the hypervariable regions corresponding to the surface loops defined for L1 from HPV type 16 (10). Seven, seven, seven, and five epitopes were found to reside in the BC, DE, FG, and HI loop regions, respectively (Fig. (Fig.3).3). Twelve of the mapped epitopes are bound by type-specific MAbs. Except for TS2-16, which recognizes an amino acid sequence between the FG and the HI loop regions, all these MAbs bind epitopes localized within either of the loop regions. An analysis of the L1 amino acid sequence identity between HPV types 2, 27, and 57 reveals significant sequence variation in the BC, DE, FG, and HI loop regions coinciding with the binding regions of all except one of the type-specific MAbs (Fig. (Fig.3).3). Interestingly, all four neutralizing MAbs were found to bind the same amino acid stretch located in the BC loop of the respectively recognized HPV type, suggesting the major importance of this surface loop as an antigen for immunizations.
To determine the extent to which the BC loop region is involved in neutralization by MAbs TS27-30, TS57-1, TS57-5, and TS57-6, we analyzed their potential to inhibit the in vitro infection of chimeric pseudovirions comprising the BC loop region of HPV type 16 in the L1 backbone of immunogen types 27 and 57 and vice versa. As expected, MAb TS27-30 efficiently neutralized HPV type 27 but not HPV type 16. However, when the HPV type 27 BC surface loop region was replaced by the corresponding amino acid sequence from HPV type 16 L1, the neutralization of HPV type 27 by TS27-30 was abrogated (Fig. (Fig.4A).4A). Conversely, HPV type 16 pseudovirions carrying the BC loop from HPV type 27 L1 were neutralized as strongly as HPV type 27 pseudovirions. Corresponding results were obtained for the analysis of neutralization by TS57-1, TS57-5, and TS57-6: infections by HPV type 57 pseudovirions and by the chimeric pseudovirions comprising the HPV type 57 BC loop in HPV type 16 particles were efficiently inhibited to a similar extent, whereas HPV type 16 pseudovirions and the chimeric pseudovirions with the HPV type 16 BC loop in an HPV type 57 particle context were not neutralized (Fig. (Fig.4B).4B). To verify that neutralization by the MAbs depends on a linear epitope within the BC loop, MAbs were preincubated with 30-mer peptides containing the respective HPV type 27 and 57 amino acid sequences before they were used for neutralization assays. The peptide amino acid sequences are presented in Fig. S2 in the supplemental material. As expected, the neutralizing capacities of the MAbs could be inhibited by the peptides in a dose-dependent manner (data not shown). In summary, these data demonstrate not only that MAbs TS27-30, TS57-1, TS57-5, and TS57-6 bind to the BC loop region but also that a linear epitope within the BC loop is necessary and sufficient for their neutralizing activity.
To evaluate the relevance of the BC loop for an in vivo immune response to VLPs, we analyzed the sera obtained from the mice immunized with HPV type 2, 27, 57, or 16 VLPs emulsified in Freund's adjuvant. Serial dilutions of the sera were used to preincubate pseudovirions chimeric for the BC loop region before neutralization assays were carried out and the corresponding neutralization IC50s were determined. The mouse sera neutralized the respective immunogen HPV types efficiently but not the heterologous types (Fig. (Fig.5).5). Compared to the neutralization of HPV type 2, 27, or 57 pseudovirions, the replacement of the BC loop by the corresponding amino acid sequence of HPV type 16 led to only a marginal reduction of neutralizing capacities by sera from mice that had been immunized with HPV type 2, 27, or 57 VLPs. A five- to sixfold decrease in the neutralizing capacity by sera of HPV type 16 VLP-immunized mice was observed if HPV type 16 pseudovirions contained the BC loop regions of type 2, 27, or 57. The presentation of the BC loop of HPV type 2, 27, or 57 in an HPV type 16 background was sufficient to achieve a similar neutralization by sera of HPV type 2, 27, or 57 VLP-immunized mice, respectively, compared to the corresponding nonchimeric pseudovirions. HPV type 16 VLP-immunized mouse sera neutralized HPV type 2, 27, or 57 pseudovirions carrying the HPV type 16 BC loop with up to sixfold-decreased efficiencies. These results indicate that the subset of antibodies induced upon immunization with VLPs, which is directed against the BC loop region, is not necessary but sufficient to confer efficient neutralization.
Despite their clinical relevance, HPV types 2, 27, and 57 have not been studied intensively so far. To characterize B-cell epitopes on capsids from these HPVs, we have generated a panel of 94 MAbs. This study represents the first comprehensive analysis of the immunogenicity of VLPs from cutaneous alpha-HPVs. It shows that both cross-reactive and type-specific antibody responses are triggered upon VLP immunization. Type-specific as well as cross-reactive MAbs were found to bind to hypervariable regions in the surface-exposed loop regions. In addition, this study identifies a linear epitope for type-specific and neutralizing MAbs. Interestingly, different MAbs targeting VLPs from two distinct HPV types recognize this epitope in the BC loop region, which is necessary and sufficient for the MAb neutralizing activity and sufficient for neutralization by sera of VLP-immunized mice.
Previously, the generation of MAbs upon immunization with VLPs of HPV types 6, 11, 16, and 18 was reported (12-16, 32). Many of these MAbs were shown to have neutralizing capacities. Characterization of the neutralizing MAbs revealed that most of them bind to conformation-dependent epitopes formed by at least one hypervariable surface loop (11, 30, 34, 42, 55). In this context, the reactivity pattern exhibited by the MAbs presented in this study was unexpected. Eighty-six out of 94 MAbs (91%), including all 27 type-specific MAbs, recognized denatured L1 proteins. Furthermore, with only 4 out of 94 MAbs (4%), fewer MAbs than expected are neutralizing.
For the initial screening of the MAbs generated against HPV types 6, 11, 16, and 18, VLPs or native virions were used as antigens. In contrast, for the generation of the MAbs against HPV types 2, 27, and 57, we employed GST-L1 capture ELISAs to screen MAb reactivity. This screening strategy might have favored the selection of hybridoma clones producing MAbs that preferentially recognize linear epitopes. However, the GST-L1 capture ELISA was previously demonstrated to be equivalent to VLP ELISA (19, 46). Moreover, Rizk et al. previously showed that MAbs raised against VLPs and that recognize conformation-dependent epitopes also bind to GST-L1 fusion proteins, suggesting that they display most conformational and neutralizing epitopes (41).
The large number of MAbs recognizing linear epitopes may also be explained by potentially improperly folded immunogen capsids. It was previously suggested that Freund's adjuvant causes a certain degree of denaturation of the administered antigen (45). However, the same adjuvant was employed for the generation of the MAbs directed against conformational epitopes on VLPs of HPV types 6, 11, 16, and 18 (12-16, 32), so a major impact by the adjuvant used is unlikely. Regarding their ability to properly assemble, the coexpression of the employed L1 constructs with the respective L2 genes in 293TT cells entails the assembly of infectious HPV type 2, 27, and 57 pseudovirions, underscoring their structural integrity. In addition, many of the generated MAbs recognized not only denatured L1 protein but also conformational VLPs or capsomeres. This implies either that the VLP and capsomere preparations used for the respective ELISAs contained large amounts of denatured protein or that the linear epitopes recognized by MAbs are displayed on the surface of conformational capsids. Given the recognition of linear epitopes in the BC loop by the four generated neutralizing MAbs, we strongly favor the latter argument. Moreover, it was previously reported that several MAbs targeting conformational VLPs from HPV types 6, 11, 16, and 18 indeed recognize surface-exposed linear epitopes (11, 12, 18, 31, 32).
We cannot exclude the possibility that the difference in the reactivity patterns of the MAbs generated in this study from previously reported ones may be owing to methodological differences. For instance, differing immunization schedules and VLP dosages were used, which might have entailed a biased generation of hybridoma clones (12). Moreover, we used only splenocytes for the fusion with myeloma cells, compared to the usage of both splenocytes and lymphocytes described previously (36). We do, however, allude to previous reports on the successful generation of MAbs against HPV type 11 using splenocytes as a fusion partner of myeloma cells (15).
The antibody isotypes produced upon an immunological stimulus are governed by the balance between helper T-cell type 1 (Th1) and Th2. Whereas a Th1-mediated response favors IgG2a production, Th2 cells support the generation of IgG1 but inhibit the production of the other IgG subclasses (IgG2a, IgG2b, and IgG3) and of IgM (48). Cytokine and chemokine profiling of human peripheral blood mononuclear cells following immunization with HPV type 16 VLPs suggests that both the Th1- and Th2-mediated arms of the adaptive immune system are activated (22, 40). These observations are consistent with the isotype distribution of the MAbs presented in this study. Furthermore, our results reflect the isotype pattern of previously reported MAbs against HPV types 6 and 11 (15, 16, 34). However, 45 (48%) MAbs were found to be of the IgM isotype, most of which were raised following immunization with VLPs from HPV type 27. Isotype switching may not have occurred to a sufficient degree in the respective mice before the splenocytes were isolated for fusion with the myeloma cells. However, whether this phenomenon reflects the quality of the administered antigen or whether it is due to individual abnormalities of the immunized mice cannot be evaluated.
It was previously suggested that the neutralizing efficiency of MAbs recognizing linear epitopes is generally lower than that of MAbs targeting conformation-dependent epitopes (16, 18). We have determined IC50s with respect to neutralization exhibited by TS27-30, TS57-1, TS57-5, and TS57-6. The respective IC50s (0.8 to 1.3 μM) are significantly lower than the IC50s observed for neutralization by sera obtained from VLP-immunized mice (13.6 to 18.7 μM) (Fig. (Fig.2),2), indicating that neutralization by the MAbs is rather efficient. However, the mouse sera are likely to contain considerable amounts of nonneutralizing antibodies. Therefore, for a stringent comparison of neutralization efficiencies, the MAbs specific for linear epitopes would have to be assayed in parallel with neutralizing MAbs targeting conformation-dependent epitopes on VLPs from HPV type 27 or 57. Unfortunately, no such MAbs are available to date, so we could not undertake such a comparative analysis.
The four generated neutralizing MAbs recognize a linear epitope located in the hypervariable surface BC loop region (10). This hypervariable region was previously described to be an epitope to at least one HPV type 16-neutralizing MAb (11). However, the epitope to this MAb (H16.E7) is strictly conformation dependent.
Together with the DE, FG, and HI loops as well as the C terminus, the BC loop region represents one of the 5-amino-acid stretches that we found to contain significant variations between HPV types 2, 27, and 57 (Fig. (Fig.3).3). Therefore, the type specificity of neutralizing MAbs TS27-30, TS57-1, TS57-5, and TS57-6 is consistent with the type-specific sequence variation of their recognized epitopes. This association also holds true for the type-specific MAbs whose epitopes were mapped to any of the hypervariable regions. However, many of the MAbs to epitopes located in these surface-exposed loop regions are not type specific but show intraspecies cross-reactivity (e.g., TS2-4, TS2-5, TS2-57, and TS2-58), probably reflecting the close relatedness between HPV types 2, 27, and 57, which entails a certain degree of sequence conservation. Supporting this notion, no such exclusively intraspecies cross-reactivity was observed for the characterization of 92 MAbs raised against VLPs from several mucosal species alpha types that are more distantly related (41). Moreover, in contrast to results from L1 sequence analyses of less related HPV types, L1 proteins from HPV types 2, 27, and 57 do not show significant amino acid sequence difference in their EF loop regions.
Sera of VLP-immunized mice were found to neutralize HPV type specifically (Fig. (Fig.22 and and5)5) (T. Senger and L. Gissmann, unpublished data). However, a similarly efficient neutralization of chimeric pseudovirions containing only the BC loop of the immunogen type could be achieved (Fig. (Fig.5).5). This observation suggests that the BC loop region contributes considerably to the epitope pool targeted by the neutralizing antibodies triggered upon immunization with VLPs. We anticipate the remaining neutralizing antibodies in the mouse sera to be directed mostly against epitopes constituted by one or several of the other surface loop regions because they contain the highest sequence variability inducing type-restricted antibodies (Fig. (Fig.3).3). It would be in correspondence with the type specificity of the humoral immune response generated by VLP-immunized mice.
The MAbs generated in this study are valuable tools allowing the type-specific detection and characterization of HPV type 2, 27, and 57 L1 proteins. Moreover, our data highlight the immunological distinctiveness of individual HPV types, even in comparison to their closest relatives. They provide an explanation of why VLP-based vaccines targeting alpha-PVs induce a mostly type-restricted humoral immune response.
We thank Neil D. Christensen for providing us with the hybridoma supernatants. We thank Corinna Klein, Birgit Aengeneyndt, and Monika Oppenländer for their expert assistance.
The work was in part supported by DKFZ-Canceropole Grand-Est (grant to L.G.).
Published ahead of print on 30 September 2009.
†Supplemental material for this article may be found at http://jvi.asm.org/.