Broad Cross-Protection Is Induced in Preclinical Models by a Human Papillomavirus Vaccine Composed of L1/L2 Chimeric Virus-Like Particles

ABSTRACT

At least 15 high-risk human papillomaviruses (HPVs) are linked to anogenital preneoplastic lesions and cancer. Currently, there are three licensed prophylactic HPV vaccines based on virus-like particles (VLPs) of the L1 major capsid protein from HPV-2, -4, or -9, including the AS04-adjuvanted HPV-16/18 L1 vaccine. The L2 minor capsid protein contains HPV-neutralizing epitopes that are well conserved across numerous high-risk HPVs. Therefore, the objective of our study was to assess the capacity to broaden vaccine-mediated protection using AS04-adjuvanted vaccines based on VLP chimeras of L1 with one or two L2 epitopes. Several chimeric VLPs were constructed by inserting L2 epitopes within the DE loop and/or C terminus of L1. Based on the shape, yield, size, and immunogenicity, one of seven chimeras was selected for further evaluation in mouse and rabbit challenge models. The chimeric VLP consisted of HPV-18 L1 with insertions of HPV-33 L2 (amino acid residues 17 to 36; L1 DE loop) and HPV-58 L2 (amino acid residues 56 to 75; L1 C terminus). This chimeric L1/L2 VLP vaccine induced persistent immune responses and protected against all of the different HPVs evaluated (HPV-6, -11, -16, -31, -35, -39, -45, -58, and -59 as pseudovirions or quasivirions) in both mouse and rabbit challenge models. The degree and breadth of protection in the rabbit were further enhanced when the chimeric L1/L2 VLP was formulated with the L1 VLPs from the HPV-16/18 L1 vaccine. Therefore, the novel HPV-18 L1/L2 chimeric VLP (alone or in combination with HPV-16 and HPV-18 L1 VLPs) formulated with AS04 has the potential to provide broad protective efficacy in human subjects.

IMPORTANCE From evaluations in human papillomavirus (HPV) protection models in rabbits and mice, our study has identified a prophylactic vaccine with the potential to target a wide range of HPVs linked to anogenital cancer. The three currently licensed vaccines contain virus-like particles (VLPs) of the L1 major capsid protein from two, four, or nine different HPVs. Rather than increasing the diversity of L1 VLPs, this vaccine contains VLPs based on a recombinant chimera of two highly conserved neutralizing epitopes from the L2 capsid protein inserted into L1. Our study demonstrated that the chimeric L1/L2 VLP is an effective vehicle for displaying two different L2 epitopes and can be used in a quantity equivalent to what is used in the licensed vaccines. Hence, using the chimeric L1/L2 VLP may be a more cost-effective approach for vaccine formulation than adding different VLPs for each HPV.

INTRODUCTION

Oncogenic human papillomaviruses (HPVs) cause cervical cancer (1), and it is widely considered that all cervical cancers are linked to at least one HPV (2–5). The oncogenic HPV-16 and HPV-18 are detected in approximately 70% of HPV-positive cervical cancers and represent the two HPVs most frequently detected in these cancers (3, 6–8). Other high-risk oncogenic HPVs also detected in cervical cancer include HPV-31, -33, -35, -45, -52, and -58 (3, 7, 8). In addition to cervical cancer, other HPVs, such as HPV-5, have been associated with nonmelanoma skin cancer (9–11), and HPV-6 and HPV-11 are the cause of the majority of genital warts (12).

Three prophylactic vaccines are licensed for the prevention of cervical cancer: AS04-adjuvanted HPV-16/18 L1 (Cervarix; GSK Vaccines) and HPV-6/11/16/18 and HPV-6/11/16/18/31/33/45/52/58 (9vHPV) (Gardasil and Gardasil 9, respectively; Merck & Co., Inc.) (13). All three vaccines contain recombinant virus-like particles (VLPs) assembled from the L1 major capsid proteins of different HPVs and have demonstrated protective efficacy against HPVs in the vaccine (14–18). Moreover, the HPV-16/18 L1 vaccine and HPV-6/11/16/18 vaccine have also demonstrated different degrees of protective efficacy against nonvaccine HPVs (19, 20), and even though head-to-head comparisons have not been performed, the HPV-16/18 L1 vaccine appears to elicit broader cross-protection than the HPV-6/11/16/18 vaccine (21). The development of the 9vHPV vaccine is aligned with the strategy by which protective efficacy is broadened by increasing the diversity of L1 VLPs in the vaccine (18).

An alternative strategy to broaden the protection of HPV vaccines that is also being considered is to use conserved epitopes from the L2 minor capsid protein incorporated into HPV L1 VLPs (22–29). This strategy is supported by evidence that vaccines based on L2 peptides can confer broad protection in mouse or rabbit HPV challenge models and elicit neutralizing antibodies against a wide range of HPVs (27, 29–33). However, during natural infection, the immunogenicity of L2 epitopes appears subdominant relative to L1 epitopes, probably because the density of L2 exposed on the capsid surface with respect to L1 is low and/or because L2 is buried within the HPV capsid and may be only transiently exposed after receptor engagement on the cell surface (34). Therefore, one way to render L2 epitopes immunogenic for vaccination is to insert L2 peptides into immunogenic HPV L1 VLPs (22–24, 27, 35, 36). One L2 epitope (consisting of amino acid residues 17 to 36; L2_17–36), recognized as a linear neutralizing epitope (RG1 epitope) (30–32), has been incorporated into the capsid surface DE loop of HPV-16 L1 or HPV-18 L1 to create chimeric L1/L2 VLP vaccines (24, 25, 27). Both L1/L2 vaccines (based on either HPV-16 or HPV-18 L1) provided protection in a mouse challenge model against a broad range of high-risk HPVs (25, 27) and generated L2-specific neutralizing antibodies together with L1-specific neutralizing antibodies (24, 25, 27). However, the insertion of epitopes into the DE loop can also disrupt the epitope recognized by the V5 and J4 monoclonal neutralizing antibodies, thus disrupting epitopes important for HPV-16/18 neutralization (22). Other potential insertion sites for epitopes include the C terminus of L1 because atomic modeling and epitope mapping suggest that its C-terminal arm is exposed at the surface of the VLP (35–38). Another L2 epitope has also been evaluated that includes L2 amino acid residues 56 to 75 (L2_56–75) (39, 40).

In the present study, the objective was to broaden the protective efficacy conferred by HPV L1-based VLP vaccines by constructing chimeric L1/L2 VLP antigens. The antigens were constructed by inserting the well-conserved L2_17–36 and L2_56–75 neutralizing epitopes within the DE loop and/or C-terminal arm of HPV-16 or HPV-18 L1 proteins. The immunogenicity and/or efficacy of the different formulations was compared using mouse and rabbit models. From these analyses, a chimeric VLP was selected that contained two different L2 epitopes in the DE loop and C terminus of the HPV-18 L1 protein. The vaccine containing this chimeric VLP conferred broader protective efficacy than the currently licensed HPV-16/18 L1 vaccine in the rabbit quasivirion (QsV) challenge and mouse pseudovirion (PsV) challenge models.

MATERIALS AND METHODS

Animal husbandry and vaccinations.

Animal husbandry and experiments were ethically reviewed and carried out in accordance with European Directive 2010/63/EU and the GlaxoSmithKline Biologicals' policy on the care, welfare and treatment of animals. BALB/c mice were purchased from Harlan, Horst, The Netherlands, and rabbits were purchased from CER Groupe, Aye, Belgium. For the challenge experiments, rabbits were purchased from Robinson, Inc., North Carolina, USA, and the protocols were approved by the Penn State University's Institutional Animal Care and Use Committee. A vaccine was administered as an intramuscular injection in the hindlimb in 6- to 8-week-old female mice or 9- to 12-week-old female rabbits. The injection volume was 50 μl for mice and 500 μl (or 50 μl or 10 μl in some experiments as indicated in the figure legends) for rabbits. All vaccines contained VLPs (clinical grade) and AS04 (produced at GSK Vaccines, Rixensart, Belgium). A 500-μl dose of vaccine contained 20 μg of a given VLP (or 100 μg in some experiments as indicated in the figure legends) and AS04 [which contains 50 μg of monophosphoryl lipid A (MPL) adsorbed on 500 μg Al(OH)₃]. A 500-μl dose of HPV-16/18 L1 vaccine, which represented the equivalent of one human dose, contained 20 μg of HPV-16 L1 VLPs, 20 μg of HPV-18 L1 VLPs, and AS04. Dosing regimens are indicated in the figure legends. PsV (mice) or QsV (rabbits) challenge was performed 1 month after the last vaccine dose unless otherwise stated.

Recombinant protein design.

Chimeric L1/L2 coding sequences were prepared by GeneArt (Regensburg, Germany) and cloned into the pAcSG2 baculovirus transfer vector (pAcSG2-L1L2). The DE-loop insertion points for L2 sequences in HPV-16 and HPV-18 L1 were between amino acid residues 137 and 138. The C-terminal arm insertion points for L2 sequences in HPV-16 L1 were between amino acid residues 431 and 432 in HPV-16 L1 and between amino acid residues 431 and 432 in HPV-18 L1. The L2_17–36 peptide fragment corresponded to residues 15 to 34 in HPV-33 L2, and the L2_56–75 peptide fragment corresponded to residues 55 to 74 in HPV-58 L2 (the peptide fragments aligned with HPV-16 L2 amino acid residues 17 to 36 and 56 to 75, respectively). Baculoviruses were cloned by plaque purification by cotransfection of the Hi-5 cell line with a pAcSG2-L1L2 vector and baculovirus genomic DNA (BacVector-3000). A small-scale expression analysis was performed for each plaque and for each construct, and the best producer was selected for larger-scale production based on a protocol described previously (41). Purity was assessed by SDS-PAGE.

Electron microscopy.

VLP samples (25 μg/ml) were prepared for electron microscopy negative-staining analysis in accordance with a standard two-step negative-staining method. Briefly, nickel grids (400-mesh) were preexposed to a glow discharge and carbon coated with a Formvar film before adsorption of the VLPs for 10 min at room temperature. After VLP adsorption, the grid was briefly washed with distilled water before a 30-s incubation in 2% (wt/vol) phosphotungstic acid in water supplemented with 1% (wt/vol) trehalose. The grid was blotted to remove excess solution and dried before examination by transmission electron microscopy under a Zeiss LEO 912 Omega instrument at 100 kV.

HPV-16/18 L1 VLP and L2 peptide ELISA.

Microtiter plates (96-well Maxisorp Immuno-plate; Nunc, Denmark) were coated with HPV-16 and HPV-18 L1 VLPs (1 μg/ml in phosphate-buffered saline [PBS]) or L2 peptides (2 μg/ml in PBS) overnight at 4°C, followed by saturation buffer (PBS, 0.1% Tween 20, 1% bovine serum albumin [BSA]) for 1 h at 37°C. For the enzyme-linked immunosorbent assay (ELISA), plates were incubated for 1.5 h at 37°C with serum samples serially diluted in saturation buffer. After routine washing steps, standard incubations with biotin-conjugated anti-rabbit or anti-mouse Ig (Amersham), streptavidin-horseradish peroxidase (Dako, United Kingdom), 0.04% o-phenylenediamine (Sigma), and 0.03% H₂O₂ in 0.1% Tween 20, and 0.05 M citrate buffer (pH 4.5) were used to detect bound primary antibodies by measuring absorbance at 492/620 nm. ELISA titers were calculated from a reference by SoftMax Pro, version 6 (using a four-parameter equation; Molecular Devices, LLC, CA, USA) and expressed in ELISA units (EU) per milliliter.

Pseudovirion (PsV)-based neutralization assay.

The PsV neutralization assay was performed as described previously (42). Briefly, purified PsVs composed of L1 and L2 capsid proteins and the reporter plasmid p2CMVSEAP (where CMV is cytomegalovirus and SEAP is secreted alkaline phosphatase) were generated in 293TT cells (human embryonic kidney cell line expressing simian virus 40 [SV40] T antigen). Pseudovirion stocks were combined with serial dilutions of serum on ice for 1 h and then incubated at 37°C for 72 h with 293TT target cells (preplated for 3 to 4 h). The SEAP content in the clarified supernatant of the 293TT cell cultures was determined using a Great Escape SEAP chemiluminescence kit (BD Bioscience, Belgium) and LMAX II luminometer (Molecular Devices). The neutralization titer was defined as the reciprocal of the dilution that caused a 50% reduction in SEAP activity compared with that of the control without serum.

Mouse pseudovirion (PsV) challenge model.

The protocol of the mouse PsV challenge was based on that described by Tumban et al. (43). Briefly, mice were injected subcutaneously with 3 mg of medroxyprogesterone acetate (Depo-Provera; Pfizer, Belgium) for 4 days before challenge with HPV PsVs. The day before challenge, rabbit serum was passively transferred into mice by the intraperitoneal route. Mice were pretreated intravaginally with 50 μl of a nonoxynol-9-based spermicidal gel (Conceptrol; Revive Personal Products Company, NJ, USA) to partially disrupt the epithelium (44) and were challenged 6 h later with an intravaginal instillation of 30 μl of HPV pseudovirions diluted in medium viscosity (2%, wt/vol) carboxymethyl cellulose. HPV infection was monitored by measuring luciferase expression in the genital tract 2 days after challenge. Anesthetized mice were instilled intravaginally with 20 μl of luciferin and imaged 5 to 10 min later during a 1-min exposure using a Xenogen IVIS in vivo imager (PerkinElmer, Belgium). Luminescence data were expressed as a ratio of the radiance signal (in photons [p] per second per square centimeter) in the vaginal region to that in the thoracic region. In a given treatment group, full protection was defined as a geometric mean luminescence signal that is statistically significantly lower than that in the control group. Because of the large variability in the luminescence signals and the small group sizes, partial protection was defined as a geometric mean luminescence signal that is ≤17% of that in the control group and a relative difference that is not statistically significant. The cutoff of 17% was used to define partial protection because it represented the highest relative geometric mean luminescence signal in which full protection was observed among all of the evaluations conducted.

Rabbit quasivirion (QsV) challenge model.

The rabbit QsV challenge protocol was based on that described by Mejia et al. (45). Briefly, rabbits were anesthetized with a mixture of ketamine hydrochloride (40 mg/kg) and xylazine (5 mg/kg). After the animals' backs were shaved, 0.5-cm-diameter sites on the left and right flanks were scarified with a scalpel. Three days after scarification, the rabbits were again anesthetized, and a 50-μl suspension of cottontail rabbit papillomavirus (CRPV) virion stock (control) or HPV capsid/CRPV genome quasivirus stock in PBS (standardized by genome content for all capsid types) was applied to the prescarified site using the edge of a syringe needle. Transdermal fentanyl (ear) patches (Duragesic; 12.5 μg/h) or subcutaneous injections of buprenorphine (0.02 to 0.05 mg/kg twice daily for 3 days) were administered as analgesics postscarification and postinfection as needed. The sizes and frequencies of lesions were evaluated at 2 to 9 weeks postchallenge. In the absence of a lesion, a diameter of 0.1 mm was assigned. In a given treatment group, full protection was defined as the absence of papillomas in all animals, and partial protection was defined as the presence of papillomas with a geometric mean papilloma size that is statistically significantly smaller than that in the control group.

Statistics.

A Shapiro-Wilk test was used to evaluate normality. Where the normality assumption was satisfied, parametric statistical analyses were performed on log-transformed data. An analysis of variance (ANOVA) with one or two factors (as indicated in the figure legends) was applied using a heterogeneous variance model followed by the Tukey's method or Dunnett's method (with adjustments for multiple comparisons) to identify differences between treatment groups (see figure legends). Where normality assumptions were not satisfied, a Wilcoxon rank sum test was applied (see figure legends). Statistical significance was assigned at a P value of ≤0.05, and for the parametric analysis of antibody concentrations and neutralizing titers, the difference between the groups was also >2-fold. All analyses were performed using SAS software (version 9.2; SAS Institute, Inc., Cary NC, USA).

RESULTS

Characterization of L1/L2 chimeric VLPs.

The well-conserved L2_17–36 and L2_56–75 neutralizing epitopes were inserted within the DE loop and/or C terminus [L1/L2_DE(17–36) and/or L1/L2_CT(56–75), respectively] of the HPV-16 or -18 L1 protein. The epitopes were derived from high-risk HPV-33 and HPV-58 L2 sequences, respectively. The HPV-33 peptide sequence is 100% identical to the respective sequence in the low-risk HPV-11, and the HPV-58 peptide sequence is 100% identical to the respective peptide in the low-risk HPV-6 oncogenic virus (Fig. 1A).

FIG 1.

Synthesis and characterization of virus-like particle (VLP) chimeras. Alignment of L2 peptide sequences from different human papillomaviruses (HPVs) with respect to HPV-33 L2_17–26 and HPV-58 L2_56–65, as indicated. HPVs include the 11 most prevalent HPVs in cervical cancer (i.e., those HPVs that have been detected in ≥1% of cervical cancers [3]) and HPV-5 that is associated with nonmelanoma skin cancer (11). Amino acid residues that differ from the respective principal comparators are shaded in gray. (B) Description of the chimeras' designs, yields, and VLP sizes. (C) Representative transmission electron micrographs of seven different successfully purified VLPs (from which VLP sizes were estimated). Scale bar, 100 nm.

A series of 10 chimeric L1/L2 proteins was produced by inserting L2_17–36 and/or L2_56–75 epitopes in the DE loop and/or C-terminal region of the HPV-16 and HPV-18 L1 backbones. From these constructs, seven chimeric proteins were selected based on their ability to form VLPs and on yield and immunogenicity (Fig. 1B). Both HPV-16 and HPV-18 L1/L2 chimeras formed VLPs with DE-loop and/or C-terminal insertion of the L2_17–36 and/or L2_56–75 epitope. However, insertion of L2_56–75 in the DE loop yielded overly large VLPs (data not shown) that were not selected for further analysis. The remaining seven HPV-16 or -18 L1/L2 chimeric proteins were then compared in terms of yield and VLP size and shape (Fig. 1B and C). Analyses indicated that the C-terminal insertions resulted in smaller-sized VLPs and that the highest yields were obtained with the chimeras based on the HPV-18 L1 backbone.

Anti-L1 and anti-L2 antibody reactivity induced by different L1/L2 chimeras.

The immunogenicities of the seven L1/L2 VLPs (adjuvanted with AS04) were evaluated in rabbits. The analysis of antibody concentrations elicited after a four-dose schedule revealed that the HPV-18 L1 chimera containing L2_17–36 within the DE loop [HPV-18 L1/L2_DE(17–36) vaccine] induced an anti-HPV-33 L2_17–36 antibody geometric mean titer (GMT) that was similar to that induced by the equivalent HPV-16 L1 chimera [HPV-16 L1/L2_DE(17–36) vaccine; 0.8-fold, P = 0.826] (Fig. 2A). The HPV-18 L1/L2_DE(17–36) chimera induced an anti-HPV-33 L2_17–36 antibody GMT that was 3.7-fold higher (P < 0.05) than that induced by the HPV-18 L1 chimera containing L2_17–36 at the C terminus [HPV-18 L1/L2_CT(17–36) vaccine]. Given these immunogenicity results and that the constructs based on the HPV-18 L1 backbone provided better yields than those based on the HPV-16 L1 backbone (Fig. 1B), three L1/L2 chimeras based on the HPV-18 L1 backbone (Fig. 1B, numbers 4, 5, and 7) were selected for further immunogenicity analyses.

FIG 2.

Cross-reactive antibody and neutralization responses induced by L1/L2 chimeric VLP vaccines in rabbits. (A) Serum anti-HPV-33 L2_17–36 Ig titers in rabbits injected with the HPV-18 L1/L2_DE(17–36) vaccine, HPV-16 L1/L2_DE(17–36) vaccine, or HPV-18 L1/L2_CT(17–36) vaccine. (B) Serum anti-HPV-18 L1 Ig, anti-HPV-33 L2_17–36 Ig, or anti-HPV-58 L2_56–75 Ig titers in rabbits injected with the HPV-18 L1/L2_CT(56–75) vaccine, HPV-18 L1/L2_DE(17–36) vaccine, or HPV-18 L1/L2_DE(17–36)-L2_CT(56–75) vaccine. (C) Serum anti-L2_17–36 Ig titers for L2 epitopes derived from HPV-5, -6, -11, -16, -35, -52, and -59 in rabbits injected with the HPV-18 L1/L2_DE(17–36) vaccine or HPV-18 L1/L2_DE(17–36)-L2_CT(56–75) vaccine. (D) Serum anti-L2_56–75 Ig titers for L2 epitopes derived from HPV-5, -11, -33, -45, and -58 in rabbits injected with the HPV-18 L1/L2_CT(56–75) vaccine or HPV-18 L1/L2_DE(17–36)-L2_CT(56–75) vaccine. (E) Serum neutralization titers (using a pseudovirion-based neutralization assay) against pseudovirions derived from HPV-18, -5, -6, -11, -16, -31, -33, -45, -52, and -58 in rabbits injected with the HPV-18 L1/L2_CT(56–75) vaccine, HPV-18 L1/L2_DE(17–36) vaccine, or HPV-18 L1/L2_DE(17–36)-L2_CT(56–75) vaccine. Histogram bars represent geometric mean titers (GMTs), and gray circles represent individual titers. Hatched bars represent GMTs at the lower cutoff limit for the assay. Each vaccine group contained three rabbits. One vaccine dose contained 20 μg of the given VLPs, and each vaccine was administered as four doses at 0, 14, 42, and 70 days. All titers were measured at 14 days postdose 4.

The immunogenicity analysis of the three vaccines found that the addition of the L2_56–75 epitope at the C terminus of the HPV-18 L1/L2_DE(17–36) chimera [HPV-18 L1/L2_DE(17–36)-L2_CT(56–75)) did not lower the anti-HPV-33 L2_17–36 antibody GMTs (1.1-fold; P = 0.993) or anti-HPV-18 L1 GMTs (3.3-fold, P = 0.378) [Fig. 2B, compare HPV-18 L1/L2_DE(17–36)-L2_CT(56–75) with HPV-18 L1/L2_DE(17–36)]. Similarly, the addition of the L2_17–36 epitope in the DE loop of the HPV-18 L1/L2_CT(56–75) chimera did not lower anti-HPV-58 L2_56–75 antibody GMTs (1.2-fold; P = 0.965) or anti-HPV-18 L1 Ig GMTs (1.1-fold; P = 0.981) [Fig. 2B, compare HPV-18 L1/L2_DE(17–36)-L2_CT(56–75) with HPV-18 L1/L2_CT(56–75)]. The analyses therefore indicated that the insertion sites were suitable for carrying the L2 epitopes.

The HPV-18 L1/L2_DE(17–36) (hereafter referred to as L1/L2_DE) vaccine induced anti-L2_17–36 antibody responses to nearly all HPVs evaluated except HPV-59. Antibody GMTs against the L2 peptide from HPV-33 were similar to (i.e., <10-fold different from) those against HPV-5, -6, -16, -35, and -52, whereas this GMT was >100-fold higher than that against the HPV-59 L2 peptide (Fig. 2C). Similar magnitudes of anti-L2_17–36 antibody responses to the different HPVs were induced with the HPV-18 L1/L2_DE(17–36)-L2_CT(56–75) (hereafter referred to as L1/L2_DE-L2_CT) vaccine.

The HPV-18 L1/L2_CT(56–75) (hereafter referred to as L1/L2_CT) vaccine induced anti-L2_56–75 antibody responses to all HPVs evaluated, such that GMTs against L2 peptides from HPV-5, -11, -33,-45, and -58 were less than 10-fold different in magnitude from each other (Fig. 2D). Similar magnitudes of anti-L2_56–75 antibody responses to the different HPVs were induced with the L1/L2_DE-L2_CT vaccine.

The functionality of anti-L2 antibodies was evaluated using a standard PsV-based neutralization assay. The L1/L2_DE-L2_CT vaccine induced neutralizing antibodies against all PsVs evaluated (Fig. 2E). The L1/L2_DE vaccine induced neutralizing antibodies against all PsVs except PsV-31, and the L1/L2_CT vaccine induced neutralizing antibodies against PsV-16, PsV-31, and PsV-45 only. The relatively low reactivity against nonvaccine PsVs in comparison with that against PsV-18 most likely reflected the reduced sensitivity of the assay for detecting L2-mediated neutralization (46). Nevertheless, the results suggested that the L1/L2_DE and L1/L2_DE-L2_CT vaccines elicited broader functional immune responses than the L1/L2_CT vaccine.

The lower reactivity of anti-L2_17–36 antibodies against HPV-59 and HPV-31 may reflect a proline substitution at position 30 in the L2 protein (Fig. 1A) because the proline substitution in HPV-31 was found to reduce the neutralizing activity of anti-HPV-16 L2_28–42 antibodies (47). However, the cross-reactivity against HPV-31 was achieved with anti-L2_56–75 antibodies induced by the doubly chimeric L1/L2_DE-L2_CT vaccine. Therefore, these results demonstrated that HPV-18 L1 protein could harbor two sites for displaying different HPV-neutralizing and cross-reactive L2 epitopes.

HPV-specific antibody GMTs were not further boosted with a fourth dose or by using 100-μg VLP doses (Fig. 3A to C). Similarly, functional antibody levels after the fourth dose were not higher with 100-μg VLP doses than with 20-μg VLP doses (compare Fig. 2E and 3D), suggesting that in the rabbit model a plateau of response is achieved after three injections of doses containing 20 μg of VLPs.

FIG 3.

Immunogenicity with respect to VLP quantity and number of doses. Serum anti-HPV-33 L2_17–36 Ig (A), anti-HPV-58 L2_56–75 Ig (B), or anti-HPV-18 L1 Ig (C) titers in rabbits injected with three doses (at 0, 14, and 42 days) or four doses (at 0, 14, 42, and 70 days) of the HPV-18 L1/L2_DE vaccine (A and C), HPV-18 L1/L2_CT vaccine (B and C), or HPV-18 L1/L2_DE-L2_CT vaccine (A to C), with the three-dose regimen containing 20 μg of VLPs or 100 μg of VLPs per dose. (D) Serum neutralization titers against pseudovirions derived from HPV-18, -5, -6, -11, -16, -31, -33, -45, -52, and -58 in rabbits injected with three doses of the HPV-18 L1/L2_DE vaccine, HPV-18 L1/L2_CT vaccine, or HPV-18 L1/L2_DE-L2_CT vaccine, where each vaccine dose contained 100 μg of VLPs. Histogram bars represent geometric mean titers (GMTs), and gray circles represent individual titers. Hatched bars indicate GMTs at the lower cutoff limit for the assay. Each vaccine group contained three rabbits. All titers were measured at 14 days after the last dose.

Hence, the overall outcome of the immunogenicity evaluations was to select the L1/L2_DE and L1/L2_DE-L2_CT VLPs for efficacy evaluations, in which, for the initial rabbit experiments, the vaccines contained 20 μg of VLPs/dose and were administered in a three-dose schedule. Based on other similar evaluations (data not shown), for the mouse experiments, vaccines contained 2 μg of VLPs/dose and were administered on a two-dose schedule.

Protective efficacy of LI/L2 chimeric VLP vaccines in the rabbit QsV challenge model.

Protective efficacy was evaluated in a rabbit QsV challenge model (45) (Fig. 4). Rabbits were injected with (i) CRPV L1 VLP (negative control), (ii) HPV-16/18 L1 VLP, (iii) L1/L2_DE VLP, or (iv) L1/L2_DE-L2_CT VLP AS04-adjuvanted vaccines. One month later, rabbits were challenged with QsVs formed by HPV-18, -11, or -58 L1/L2 capsids carrying the CRPV genome. Infection and the degree of infection were determined by the presence of papillomas and mean papilloma diameters, respectively, at the two QsV challenge sites for each animal. The protective efficacy was assessed by statistical analysis of the ratio of mean papilloma diameters in the test vaccine group to the mean papilloma diameters in the negative-control vaccine comparison (and termed relative geometric mean diameter, or RMD).

FIG 4.

Protective efficacy of L1/L2 chimeric VLP vaccines in the rabbit quasivirion challenge model. (A) Papilloma diameters in vaccinated rabbits 8 weeks after challenge by quasivirions (QsVs) formed by HPV-18, -11, or -58 capsids carrying the CRPV genome. QsVs were injected at two sites on the back of each rabbit. Rabbits (five per group) were injected with the CRPV L1 vaccine, HPV-16/18 L1 vaccine, HPV-18 L1/L2_DE(17–36) vaccine, or HPV-18 L1/L2_DE(17–36)-L2_CT(56–75) vaccine. Gray circles represent mean papilloma diameters of the two QsV injection sites for each rabbit relative to the geometric mean of the mean papilloma diameters in the CRPV L1 vaccine group, and vertical lines represent geometric means of the mean papilloma diameters in each group relative to the geometric mean of the mean papilloma diameters in the CRPV L1 vaccine group (termed RMDs). In the absence of a papilloma at a QsV injection site, a cutoff diameter of 0.1 mm was assigned. Significant differences (***, P < 0.001) in results relative to those from the CRPV L1 vaccine group were identified using the Wilcoxon rank sum test and Dunnett's method (adjusted for multiple comparisons). (B) Serum anti-HPV-18 L1 Ig, anti-HPV-33 L2_17–36 Ig, or anti-HPV-58 L2_56–75 Ig titers in rabbits (5 per group) injected with the HPV 16/18 L1 vaccine, HPV-18 L1/L2_DE vaccine, or HPV-18 L1/L2_DE-L2_CT vaccine. Gray circles represent individual titers, and vertical lines represent geometric mean titers in each group. ND, not determined. One vaccine dose contained 20 μg of the given VLPs, and each vaccine was administered as three doses at 0, 14, and 28 days.

All animals that received the CRPV L1 VLP control vaccine developed papillomas after challenge with the QsVs evaluated, except one rabbit out of five that were challenged with QsV-11 (Fig. 4A). As expected the HPV-16/18 L1 vaccine and the two HPV-18 L1/L2 chimeric vaccines fully or nearly fully protected against QsV-18 (RMD of ≤3%, P < 0.001), with only one animal injected with the L1/L2_DE-L2_CT vaccine having a papilloma at one site of injection. In contrast to the HPV-16/18 L1 vaccine, the L1/L2_DE and L1/L2_DE-L2_CT vaccines protected against QsV-11 and QsV-58. Moreover, the L1/L2_DE-L2_CT VLP vaccine provided a higher degree of protection than the L1/L2_DE vaccine because it fully protected against both QsV-11 and QsV-58 (no papillomas; RMD of ≤3%, P < 0.001), whereas the L1/L2_DE vaccine fully protected against QsV-11 (no papillomas; RMD of 3%, P < 0.001) and only partially protected against QsV-58 (4/5 animals had papillomas; RMD of 23%, P < 0.01). The two chimeric vaccines induced the expected VLP-specific antibodies on the day of challenge (anti-HPV-18 L1, anti-HPV-33 L2_17–36, or anti-HPV-58 L2_56–75 antibodies) (Fig. 4B).

LI/L2 chimeric VLP vaccines confer protection at 1 and 6 months in the mouse challenge model.

The protective efficacies of various vaccines were then evaluated in a mouse PsV challenge model (43, 44) in comparison with a saline negative control (Fig. 5). In addition to HPV-16/18 L1, L1/L2_DE, and L1/L2_DE-L2_CT vaccines, other combinations were also evaluated in which HPV-16 and HPV-18 L1 VLPs were combined with L1/L2_DE VLPs or L1/L2_DE-L2_CT VLPs. Vaccinated mice were challenged by intravaginal inoculation with PsVs, including a luciferase reporter gene construct encapsulated by L1 and L2 proteins. The degree of infection was determined 2 days later by in vivo measurements of luciferase-generated luminescence a few minutes after intravaginal instillation of luciferin. In two experiments, mice were challenged with PsVs at either 1 month or 6 months after the second vaccine dose. Both experiments used PsVs representing HPV-11, -16, -35, or -58. The four viruses were chosen based on a range of anticipated outcomes (PsV-16, for harboring homology with HPV-16 L1, PsV-11 for harboring homology with L2_DE(17–36), PsV-58 for harboring homology with L2_CT(56–75), and PsV-35 for harboring partial [90%] homology with either of the two L2 peptides). The first experiment also used PsVs representing HPV-45 and -59 [two viruses which have relatively low, i.e., ≤75%, similarity with the L2_DE(17–36) peptide].

FIG 5.

LI/L2 chimeric VLP vaccines confer protection at 1 and 6 months in the mouse pseudovirion challenge model. Luciferase-generated luminescence was measured a few minutes after intravaginal instillation of luciferin in mice 1 month (A) or 6 months (B) after vaccination and 2 days after challenge by intravaginal inoculation with pseudovirions (PsVs) formed from a luciferase reporter gene construct encapsulated by L1 and L2 proteins from HPV-11, -16, -35, -58, -45, or -58 (A) or from HPV-11, -16, -35, or 58 (B). Mice (five per group) were injected with saline, HPV-16/18 L1 vaccine, HPV-18 L1/L2_DE(17–36) vaccine, or HPV-18 L1/L2_DE(17–36)-L2_CT(56–75) vaccine or with vaccines containing combinations of HPV-16/18 L1 VLPs with HPV-18 L1/L2_DE(17–36) VLPs or with HPV-18 L1/L2_DE(17–36)-L2_CT(56–75) VLPs. All luminescence measurements are normalized to the geometric mean luminescence measurement in the saline control group (100%). Gray circles represent individual relative luminescence measurements, and vertical lines represent geometric means of the relative luminescence measurements in each group. Significant differences (*, P < 0.05; **, P < 0.01; ***, P < 0.001) in results relative to those from the saline group were identified using ANOVA (with time [month] and treatment as factors) and Dunnett's method (adjusted for multiple comparisons). Note that there was no significant effect of time in the ANOVA model. One vaccine dose contained 2 μg of the given VLPs, and each vaccine was administered as two doses, 1 month apart.

One month after vaccination, the HPV-16/18 L1 vaccine fully protected against PsV-16 (relative luminescence to that of the saline control [RLSC] of 1%; P < 0.05) but did not protect against PsV-11, -35, -45, -58, or -59 (RLSC of ≥47%; P ≥ 0.681) (Fig. 5A). In contrast, the L1/L2_DE-L2_CT vaccine fully protected against PsV-11, -35, -45, -58, and -59 (RLSC of ≤6%; P < 0.05) and partially protected against PsV-16 (RLSC of 2%; P = 0.061), and the L1/L2_DE vaccine fully protected against PsV-11, -45, and -58 (RLSC of ≤8%; P < 0.05), partially protected against PsV-16 (RLSC of 6%; P = 0.227) and PsV-59 (RLSC of 5%; P = 0.061), but did not protect against PsV-35 (RLSC of 26%; P = 0.674). The L1/L2_DE-L2_CT combination with HPV-16/18 L1 fully protected against all six PsVs evaluated (RLSC of ≤10%; P < 0.05) (Fig. 5A). As with L1/L2_DE alone, the combination of L1/L2_DE and HPV-16/18 L1 fully protected against PsV-11 and PsV-45 (RLSC of ≤9%; P < 0.05) and did not protect against PsV-35 (RLSC of >100%). In contrast to L1/L2_DE alone, the combination of L1/L2_DE and HPV-16/18 L1 fully protected against HPV-16 (RLSC of 1%; P < 0.05), only partially protected against PsV-58 (RLSC of 14%; P = 0.404), and did not protect against PsV-59 (RLSC of 66%; P = 0.992).

Six months after vaccination, the extent of protection conferred by the vaccines was similar to that observed 1 month after vaccination (there was no significant interaction with time in the two-factor ANOVA) (Fig. 5B), except that the L1/L2_DE vaccine fully protected against PsV-16 (RLSC of 3%; P < 0.001) and did not protect against PsV-58 (RLSC of 67%; P = 0.935). The L1/L2_DE-L2_CT VLP alone or in combination with HPV-16 and HPV-18 L1 VLPs fully protected against all four PsVs evaluated (PsV-11, -16, -35, and -58; RLSC of ≤17%, P < 0.05) (Fig. 5B).

The immunogenicities of the two L1/L2 chimeric vaccines were as expected 1 month after vaccination, with VLP-specific antibodies being detected in accordance with the presence (or absence) of the L2 epitopes and HPV-18 L1 backbone in the vaccines (Fig. 6A). At 6 months after the receipt of the L1/L2_DE-L2_CT vaccine, anti-HPV-18-L1-, anti-HPV-33-L2_17–36-, and anti-HPV-58-L2_56–75-specific antibodies were detected, although the GMTs had declined by 2.0-, 3.7-, and 6.0-fold, respectively, compared with the levels at 1 month postdose 2 (P < 0.01) (Fig. 6A). The anti-HPV-18-, anti-L2_17–36-, and anti-L2_56–75-specific antibody responses to the L1/L2_DE-L2_CT vaccine appeared to be both independently boosted by keyhole limpet hemocyanin (KLH)-conjugated L2_17–36 and L2_56–75 peptides, respectively, such that, although not significant, the antibody GMTs appeared to be higher than the GMT of the saline boost control by 2.4-fold (P = 0.156) and 2.7-fold (P = 0.110), respectively, at 1 week postboost (Fig. 6B).

FIG 6.

Memory response to the L1/L2_DE-L2_CT vaccine in mice. (A) Serum anti-HPV-18 L1 Ig, anti-HPV-33 L2_17–36 Ig, or anti-HPV-58 L2_56–75 Ig titers in mice injected with the L1/L2_DE vaccine or L1/L2_DE-L2_CT vaccine at 1 month (1 M) postdose 2 (24 to 25 mice per group) and in mice injected with the L1/L2_DE-L2_CT vaccine at 6 months (6 M) postdose 2 (14 to 25 mice per group). Significant differences (**, P < 0.01; ***, P < 0.001) between 1-month and 6-month titers in the L1/L2_DE-L2_CT vaccine group were identified using ANOVA and Tukey's method (adjusted for multiple comparisons). (B) Serum anti-HPV-33 L2_17–36 Ig or anti-HPV-58 L2_56–75 Ig titers at 6 months in mice (18 to 23 mice per group) after priming (saline or L1/L2_DE-L2_CT vaccine) and 1 week after boosting with saline or KLH-conjugated L2_17–36 peptides and/or KLH-conjugated L2_56–75 peptides. Gray circles represent individual titers, and vertical lines represent geometric mean titers in each group. Relevant significant differences (***, P < 0.001) in results relative to those from the no-prime saline boost group were identified using ANOVA and Tukey's method (adjusted for multiple comparisons). One vaccine dose contained 2 μg of the given VLPs, and the vaccine was administered as two doses, 1 month apart.

In summary, as with the rabbit model, the L1/L2_DE-L2_CT vaccine conferred a greater breadth of protection than either the L1/L2_DE VLP or HPV-16/18 L1 vaccine in the mouse model. The combination with HPV-16 and HPV-18 L1 VLPs in the vaccine appeared to affect the degree of protection conferred by L1/L2_DE VLPs against certain PsVs (PsV-58 or -59), whereas the degree and, hence, breadth of protection conferred by L1/L2_DE-L2_CT VLPs appeared unaffected. Moreover, the degree of protection conferred by HPV-16 and HPV-18 L1 VLPs against PsV-16 appeared potentially enhanced by the addition of the L1/L2 chimera in the vaccine.

The capacity of the L1/L2_DE-L2_CT vaccine to confer such a breadth of protection appeared to persist to at least 6 months after vaccination and was related to the persistence and boostability of anti-L2 antibodies. Hence, the L1/L2_DE-L2_CT vaccine may function by inducing L2-specific long-lived plasma cell and B memory cell populations, a propensity that may also be due to the vaccine including AS04 (42).

Further evaluation of the protective efficacy in the rabbit of L1/L2_DE-L2_CT VLPs alone or in combination with HPV-16 and HPV-18 L1 VLPs.

The rabbit challenge model was then used to evaluate the breadth of protection conferred by a dose range of L1/L2_DE-L2_CT vaccine formulations so as to further evaluate under more sensitive experimental conditions the L1/L2_DE-L2_CT VLP combination with HPV-16 and HPV-18 L1 VLPs (referred to as HPV-16/18 L1/L2).

First, three regimens were compared: three doses of 2 μg of VLPs (and 1/10 of the HPV-18 L1 VLP dose in the human vaccine), two doses of 2 μg of VLPs, and two doses of 0.4 μg of VLPs (Fig. 7A and B). Anti-HPV-18 L1, anti-L2_17–36, and anti-L2_56–75 Ig GMTs were 4.1-, 8.3-, and 5.5-fold lower, respectively, with the two-dose, 0.4-μg VLP regimen than with the three-dose, 2-μg VLP regimen (Fig. 7A). The three-dose, 2-μg VLP regimen fully protected the rabbits against QsV challenge with QsV-11, -16, -31, -35, -39, and -45 (no papillomas; RMD of ≤2%, P < 0.05) and partially protected against QsV-6, -18, and -59 (≥2/5 animals per group without papillomas; RMD of ≤4%, P < 0.05) (Fig. 7B). The degree and breadth of protection were lowered slightly with the two-dose, 2-μg VLP regimen and more extensively with the 0.4-μg VLP regimen than with the three-dose, 2-μg VLP regimen, such that the 0.4-μg VLP regimen partially protected against QsV-16, -18, -35, -39, and -45 (≥2/5 animals per group without papillomas; RMD of ≤7%, P < 0.05). The lower protection against QsV-6, -31, and -59 using the 0.4-μg VLP regimen than with the other regimens may have reflected the lower anti-L2_56–75 GMT in this group, given that the L2_56–75 peptide shared a high degree of homology with peptides from the respective QsVs used in the challenge (Fig. 1 and 7A, anti-L2_CT Ig). However, even with this low-dose regimen, the L1/L2_DE-L2_CT vaccine conferred a substantial degree of cross-protection in this model.

FIG 7.

Protective efficacy in rabbits with different L1/L2_DE-L2_CT VLP regimens alone or in combination with HPV-16/18 L1 vaccine. (A) Serum anti-HPV-18 L1 Ig, anti-HPV-33 L2_17–36 Ig, or anti-HPV-58 L2_56–75 Ig titers in rabbits (five per group) vaccinated with different regimens with the L1/L2_DE-L2_CT vaccine (3d-1/10, three doses at 0, 1, and 6 months and 2 μg of VLPs per dose [50 μl], which is equivalent to 1/10 the quantity of VLPs in a human dose; 2d-1/10, two doses at 0 and 6 months and 2 μg of VLPs per dose; 2d-1/50, two doses at 0 and 6 months and 0.4 μg of VLPs per dose [10 μl]) or with CRPV L1 VLPs (three doses at 0, 1, and 6 months and 2 μg of VLPs per dose). All titers were measured 1 month after the last dose. Gray circles represent individual titers, and vertical lines represent geometric mean titers in each group. Significant differences (*, P < 0.05) in results relative to those from the three-dose group were identified using ANOVA and Dunnett's method (adjusted for multiple comparisons). (B and C) Papilloma diameters in vaccinated rabbits at 8 weeks after challenge by quasivirions (QsVs) formed by HPV-18, -6, -11, -16, -31, -35, -39, -45, or -59 capsids carrying the CRPV genome. QsVs were injected at two sites on the back of each rabbit, and the papilloma diameter was calculated as the mean diameter of the two sites. For the data shown in panel B, vaccinated rabbits from the experiments shown in panel A or C (five per group) were injected with the HPV-16/18 L1/L2 vaccine (incorporating L1/L2_DE-L2_CT VLPs), HPV-16/18 L1 vaccine, or the CRPV L1 vaccine, using the two-dose regimen (0 and 6 months) and 0.4 μg of VLPs per dose (10 μl). The histogram bars represent geometric means of the mean papilloma diameters in each group, normalized to the respective geometric mean papilloma diameters in the CRPV L1 VLP vaccine control group (termed RMDs). In the absence of a papilloma at an inoculation site, a cutoff diameter of 0.1 mm assigned. Significant differences (*, P < 0.05; **, P < 0.01; ***, P < 0.001) in results relative to those from the CRPV L1 vaccine group were identified using ANOVA and Dunnett's method (adjusted for multiple comparisons).

Second, using the two-dose regimen with vaccines containing 0.4 μg of VLPs, the HPV-16/18 L1/L2 vaccine was compared with the HPV-16/18 L1 vaccine (Fig. 7C). With this dose regimen, HPV-16/18 L1 vaccine fully protected against QsV-16 and QsV-18 (no papillomas; RMD of 1%, P < 0.001) and, in contrast to the earlier experiment, partially protected against QsV-11 (3/5 animals per group without papillomas; RMD of 3%, P < 0.05). Although results were statistically insignificant, the HPV-16/18 L1 vaccine provided some protection against QsV-31 and QsV-45 (3/5 and 2/5 animals per group without papillomas, respectively; RMDs of 6% [P = 0.125] and 11% [P = 0.236], respectively).

In contrast and using the same low-dose regimen, the HPV-16/18 L1/L2 vaccine fully protected against QsV-16 and QsV-45 (no papillomas; RMD of 1%, P < 0.001) and partially protected against QsV-6, -11, -18, -31, -35, and -39 (≥3/5 animals per group without papillomas; RMD of ≤4%, P < 0.05). Although results were statistically insignificant, the HPV-16/18 L1/L2 vaccine provided some protection against QsV-59 (1/5 animals without papillomas, RMD of 27%, P = 0.417). The HPV-16/18 L1/L2 vaccine also appeared to confer broader protection than the L1/L2_DE-L2_CT vaccine (e.g., with QsV-6, -11, -31) (compare Fig. 7B and C), indicating that the HPV-16 L1 VLP also contributed to protection.

DISCUSSION

The development of recombinant chimeric L1/L2 VLP antigens based on different HPVs represents one strategy to broaden the protective efficacy of HPV vaccines. In the present study, using the mouse and rabbit challenge models, the AS04-adjuvanted L1/L2_DE-L2_CT vaccine conferred broader protection than the licensed HPV-16/18 L1 vaccine to prevalent high-risk HPVs associated with cervical cancer (3) and two low-risk HPVs (HPV-6 and -11) associated with benign genital warts (12). The degree and breadth of protection were further enhanced when the L1/L2_DE-L2_CT VLP vaccine was formulated with HPV-16 and HPV-18 L1 VLPs. The L1/L2_DE-L2_CT vaccine also induced neutralizing antibodies against HPVs associated with cervical cancer and benign genital warts and against the HPV (HPV-5) associated with nonmelanoma skin cancer (9–11).

Both the rabbit and mouse challenge models satisfied the aim of discriminating vaccines based on the breath of protection. Where comparisons could be made, the responses to challenge were consistent between models; for example, the L1/L2_DE-L2_CT vaccine and L1/L2_DE vaccine, in contrast to the HPV-16/18 L1 vaccine, conferred protection against the challenge viruses based on HPV-11 and HPV-58. However, the number of animals per group limited the power of the statistical analyses in identifying responses to challenge that were partially protective and may explain the equivocal responses to PsV-16 challenge in mice vaccinated with L1/L2_DE-L2_CT and to QsV-11 challenge in rabbits vaccinated with HPV-16/18 L1. Nevertheless, the relevance of the rabbit challenge model was supported by the observation that with the relatively low-dose regimen, the HPV-16/18 L1 vaccine conferred a degree of cross-protection against HPV-31 and HPV-45, two HPVs to which cross-protection has been observed with the vaccine in humans (19). Although the neutralization assay was not optimized for detecting L2-mediated neutralization (46, 48), neutralization antibodies were observed against all HPVs tested. Furthermore, five HPVs (HPV-6, -11, -16, -31, and -58) were evaluated both in the neutralization assay and in the challenge model for rabbits vaccinated with L1/L2_DE-L2_CT. For each of these five HPVs, vaccination generated neutralization antibodies and conferred protection against challenge, thus supporting the relevance of the neutralization assay.

The cross-neutralization titers and capacity for cross-protection conferred by the L1/L2_DE-L2_CT vaccine reflected the conserved nature of the two different L2 functional epitopes inserted into the DE loop and C-terminal arm of the HPV-18 L1 protein. There was no evidence of immune interference at the level of immunogenicity or protection between the two inserted L2 epitopes. In contrast, there was evidence that both epitopes could independently contribute to the generation of neutralizing antibodies to a given HPV, which was advantageous for targeting those HPVs (e.g., HPV-31 and -59) in which L2_17–36-specific antibodies were less effective due to proline-to-serine substitution in the targeted L2_17–36 epitope (47). In addition, the L2_17–36-specific antibodies did not appear to confer protection against HPV-35. Therefore, the L2_17–36 epitope alone may be insufficient to target some of the high-risk HPVs.

Based on purification yields and physical properties, the HPV-18 VLP was selected as a backbone in preference to HPV-16 L1 for the insertion of L2 peptides even though both HPV-16 L1 and HPV-18 L1 proteins have been used to develop immunogenic chimeric L1/L2 VLPs previously (24, 25, 27). The HPV-18-specific immunogenicity and protective capacity of HPV-18 L1 appeared generally unaffected by the insertion of L2 epitopes in the L1/L2_DE-L2_CT vaccine. Similarly, as with the HPV-16/18 L1 vaccine (42), the L1/L2_DE-L2_CT vaccine induced HPV-18 L1-specific antibodies, as well as L2-specific protective immune responses, that were long lasting. Hence, the present study has made an advance on earlier studies (22–24, 27) by demonstrating that the HPV-18 L1 VLP can be used as an effective vehicle for two different L2 epitopes, one located at an internal DE loop and the other at the C-terminal arm. Moreover, the L1/L2_DE-L2_CT VLP conferred a significant level of protection even with a low-dose regimen.

At least one of the two L2 peptide sequences in the L1/L2_DE-L2_CT VLP was ≥90% identical to the equivalent peptide sequences in the 11 most prevalent oncogenic HPVs (3), including those HPVs represented in the 9vHPV L1 vaccine (18). Hence, on the basis of protection against QsV-35 and QsV-39 in the rabbit model, the L1/L2_DE-L2_CT vaccine may have the potential to confer protection against HPVs with L2 sequences that are at least 90% identical to those in the vaccine. On the basis of protection against PsV-59 conferred by the L1/L2_DE vaccine in the mouse model, this potential may extend to HPVs with L2 sequences that are at least 65% identical to those in the vaccine. Therefore, an AS04-adjuvanted vaccine containing HPV-18 L1/L2_DE-L2_CT VLPs alone or in combination with HPV-16 and HPV-18 L1 VLPs has the potential to provide broad cross-protective efficacy in human subjects beyond the high-risk HPVs tested in this study. In addition, the use of a chimeric protein in the VLP may be more cost-effective for a vaccine formulation than adding different VLPs representing each HPV.