An HLA-A2.1-Transgenic Rabbit Model to Study Immunity to Papillomavirus Infection¹

Human papillomaviruses (HPVs)³ represent a group of significant human pathogens that are strongly correlated with various human cancers, most notably, cervical cancer. HPV-associated human cancers account for up to 250,000 deaths per year from cervical cancer alone, as well as an unknown additional number of patients with cancers of the head and neck, skin, and other anogenital sites. Recent advances in HPV vaccine developments have led to promising results in phase II/III clinical trials of a protective virus-like particle (VLP) vaccine (1, 2, 3, 4). The vaccine develops a protective response via the induction of virus capsid neutralizing Abs. One potential caveat of the current VLP vaccine is that the immunity that develops is HPV type-specific such that protection will occur only to those types that are represented in the VLP vaccine (5). As HPV-associated cancers are comprised of at least 15 antigenically unrelated HPV types, a multi-HPV-type VLP vaccine, or a vaccine that develops more broadly protective neutralizing Abs, would be required for more complete protection against HPV-associated cancer. A final consideration of the VLP vaccine is that only protection against new infections will be achieved and thus the vaccine is expected to have little or no impact on existing HPV disease (6). A second form of immunity thus needs to be activated for patients with current HPV infections and this response must target virus-infected cells through Ag-specific adaptive cell-mediated immunity. More specifically, different viral proteins expressed from the early genes E1, E2, E6, and E7 are predicted to be the necessary targets of a successful cell-mediated immune clearance of existing HPV infections (7, 8, 9, 10, 11, 12, 13).

To assess host immune responses to HPV infections, preclinical models are required for initial vaccine and immunotherapeutic testing. The extreme species and tissue specificity of papillomaviruses make these studies particularly challenging (14). Current preclinical models of natural papillomavirus infections include rabbit, dog, and bovine models (15). There are currently no natural papillomavirus infection models for small laboratory rodents for testing cell-mediated immune responses to ongoing infections. Rabbit, dog, and bovine models are difficult models to work with as they must be studied in outbred populations, although several inbred rabbit strains are available for additional immune studies. The species restriction barriers of papillomaviruses prevent infection and growth of HPVs in all immunocompetent laboratory animal models.

The cottontail rabbit papillomavirus (CRPV)/rabbit model offers several advantages as a preclinical model for studying host immunity to papillomavirus infections (16). The model has been used extensively to study protective immunity to VLP-based vaccines as well as to vaccines comprised of fragments of the minor capsid protein, L2. The rabbit model has also been used to test cell-mediated immunity to various viral early proteins including E1, E2, E6, E7, and E8 (7, 17, 18). A powerful advantage of this model is that papillomas can be generated by direct infections of the skin with viral DNA in the absence of encapsidation by the viral coat proteins (19, 20). This latter observation provides opportunities to genetically alter the vial genome by site-directed mutagenesis and to engineer epitopes into the various viral genes for testing specific immunity. As an example of this technology, we have shown that an HPV16 E7 T cell epitope can be engineered into the CRPV E7 gene of the CRPV genome, which subsequently retains the functional ability to develop papillomas. However, despite these advantages, studies on viral immunity to CRPV in the context of rabbit MHC molecules provide little useful information for the design, induction, and testing of Ag-specific T cell responses to HPV epitopes in the context of human MHC molecules.

Given the current inability of preclinical models to assess immunity to HPV epitopes in an infection model for papillomaviruses, we have developed a transgenic rabbit model expressing HLA-A2.1. The rationale for selecting the HLA-A2.1 gene includes: the availability of a large number of reagents and cell lines for functional testing; a large database and computer models for prediction of HLA-A2.1-specific epitopes; and successful testing of HLA-A2.1 function in transgenic mice.

Three lines of HLA-A2.1-transgenic rabbits were developed (J. Hu, X. Peng, L. Budgeon, N. Cladel, and N. Christensen, manuscript in preparation). The transgenic rabbits display high levels of HLA-A2.1 protein on the surface of lymphocytes and in all tissues examined by immunohistology. Transgenic rabbits vaccinated with HLA-A2.1-restricted epitopes showed protection against epitope-modified CRPV genomes demonstrating utility of the model. These data indicate that the transgenic rabbit model will provide opportunities to directly test HLA-A2.1-restricted epitopes of HPV proteins in the context of a human MHC class I (MHCI) gene.

Future potential applications of the HLA-A2.1-transgenic rabbit model can be envisaged in the testing of CD8⁺ CTL-inducing vaccines against other human pathogens that are permissive or semipermissive in rabbits including human T cell leukemia virus 1/2, adenovirus, EBV-like viruses, tuberculosis, and syphilis (21, 22, 23, 24).

The procedures for producing transgenic rabbits have been previously described (25). In brief, microinjection of the DNA construct HLA-A2.1 (3 μg/ml) into the pronucleus of fertilized eggs was conducted. The injected eggs were incubated for 1 h following microinjection and then transplanted into the oviducts of pseudopregnant recipient rabbits that had been mated with a sterile (vasectomized) male at the same time that the donors were mated. The F₁-F₅-transgenic rabbits were produced by breeding A2-transgenic rabbits with EIII/JC inbred rabbits maintained in our facility.

Peptides: HPV16E7/82–90 (LLMGTLGIV) (26), CRPVE1/161–169 (LLFRQAHSV); CRPVE1/245–253 (ALLSQLLGV), CRPVE1/42–50 (SLLDDTDQV), CRPVE1/301–311 (MLQEKPFQL), CRPVE1/149–157 (ILNANTARV), SV40LT/281–289 (KCDDVLLLL) (27), and HIVGagP17/77–85 (SLYNTVATL) were synthesized in the core facility of the Pennsylvania State University College of Medicine. The HLA-A2.1/HPV16E7/82–90 tetramer complex was supplied by the Tetramer Core Facility of National Institutes of Health. Primary and secondary Abs used in our flow cytometry, immunohistochemistry, and immunofluorescent microscopy were: HLA-A2.1-specific mAb (BB7.2; American Type Culture Collection), anti-rabbit MHCI Ab (Mab73.2; Spring Valley Laboratories), FITC- or PE-conjugated anti-rabbit CD8 Ab (12.C7; Research Diagnostics), Alexa Fluor 488 and 594 goat or donkey anti-mouse IgG1 or IgG (Molecular Probes/Invitrogen Life Technologies), FITC anti-mouse IFN-γ (R&D Systems), and PE anti-mouse CD90.2 (BD Biosciences), PE and FITC anti-mouse IgG (Jackson ImmunoResearch Laboratories).

PBLs were isolated in blood drawn from the rabbit ear artery as described previously (28).

PBLs collected from transgenic rabbits were cultured in DMEM (complement with 10% FBS, 10 mM HEPES, 2 mM l-glutamine, 1 mM sodium pyruvate, 100 U/ml⁻¹ penicillin, and 100 μg/ml⁻¹ streptomycin). The attached cells were maintained and subcultured up to 2 mo and labeled as RA2 cells. RA2 cells was immortalized with SV40 large T Ag (pPVU0Neo, with a G418 selection marker; provided by Dr. M. J. Tevethia, Hershey Medical Center, Hershey, PA) (29) and were selected and maintained in medium containing 1 mg/ml G418 and labeled as RA2LT.

Rabbit ear biopsies and other tissues were collected and frozen in liquid nitrogen and kept subsequently at −70°C. Regular immunohistochemistry staining procedures were conducted on the frozen sections of these ear biopsies (30). Positive signals showed red under bright field microscopy and photo images were stored digitally using Adobe Photoshop CS.

A total of 10⁶ PBLs from HLA-A2.1-transgenic and nontransgenic rabbits or cultures of RA2, T2, Caski, and normal rabbit cells were labeled with BB7.2 and secondary PE anti-mouse IgG (1:50). The cells were then washed three times and then analyzed with a fluorescent-activated cell scanner (FACScan; BD Biosciences) in the Core Facility of the Pennsylvania State University College of Medicine. Control-labeled cells were stained using the same protocol but without the primary Ab and were used as negative controls for FACScan analysis. The expression levels were determined as mean fluorescence intensity (MFI) units. Data were analyzed with CellQuest software (BD Biosciences).

Rabbit spleen cells, ear biopsy sections, and trypsinized RA2 cells were incubated in primary (BB7.2) and secondary (Alex Fluor 594 anti-mouse-IgG) Ab and subsequently stained with anti-rabbit MHCI and secondary Ab (Alex Fluor 488 anti-mouse IgG1) on ice for 45 min, respectively. After three washes with 2% FBS/PBS, the cells were again incubated in 1 μg/ml Hoechst dye for 5 min. Digital photos were taken using an ACT-1 Nikon microscope image processing system. Colocalization of the two MHCI molecules was accessed with Adobe Photoshop CS.

For epitope presentation by transgenic rabbit cells, a CTL clone (A2.1-58.3) that specifically recognizes an SV40 large T Ag (SV40LT) epitope 281–289 was used as effector cells and the assay was conduced as described previously (27). RA2 cells pulsed with SV40LT281–289 or HIV GagP17/77–85 were used as target cells for exogenous epitope-presenting assay. RA2 and normal rabbit cells stably transfected with SV40LT were used as target cells in the endogenous peptide-presenting assay. A mouse A2 cell line stably transfected with SV40LT (A2.1-364-A) (27) was used as positive control. The IFN-γ-producing CD8⁺ T cells shown as the bars represented the effector cell population (percentage).

The HPV16E7/82–90 DNA vaccine and the CRPVE1 multivalent epitope DNA vaccine were designed and synthesized commercially (Genscript). Five repeats of the HPV16E7/82–90 epitope or five CRPV E1 epitopes screened by MHCI epitope prediction software (31) were separated by AAY linkers (32).

DNA-coated gold particles were attached to the inside of small tubing as previously described (18). HLA-A2.1-transgenic and nontransgenic rabbit inner ear skin was bombarded with DNA-coated gold particles by a helium-driven gene gun delivery system (33). HLA-A2.1-transgenic and nontransgenic outbred rabbits were divided into two groups. One group was vaccinated with HPV16E7/82–90 DNA vaccine (n = 4) or CRPVE1 epitope vaccine (n = 4) and the other group received the control vaccine (ubiquitin). The rabbits were boosted twice with the vaccines (20 μg DNA/time/rabbit) at 3-wk intervals.

CD8⁺ T cells were positively selected from PBLs and spleen cell suspensions from vaccinated rabbits with an EasySep PE-Magnetic Selection kit (StemCell Technologies). The cells were cultured with gamma-irradiated and HPV16E7/82–90 peptide-pulsed syngeneic PBLs or spleen cells complemented with 10 U of recombinant human IL-2 (Cys¹⁴⁵Ser; R&D Systems). The CTLs were then stimulated with RA2 cells stably expressing the E7 epitope. After 2 wk of stimulation, cells were labeled with FITC anti-rabbit CD8 and HLA-A2.1 tetramers loaded with the E7 peptide (LLMGTLGIV) or an antigenically unrelated control peptide. Two-color flow cytometry was performed with CellQuest software (BD Biosciences). Two-color-labeled cells represented the effector cells that had the potential to eliminate infected cells.

A colorimetric method was used to quantitatively measure lactate dehydrogenase release upon cell lysis (cytoTox96 Nonradioactive Cytotoxicity Assay kit; Promega). Briefly, in vitro-stimulated rabbit CTLs were harvested and divided into quadruplicate wells using E:T ratios of 20:1, 6.7:1, 2.2:1, and 0.7:1 and cocultured with target cells at 37°C for 4 h. Supernatant collected from each well was added to assay buffer and quantitated by an ELISA plate reader at 490 nm. Specific release was calculated for the percent cytotoxicity using the following formula: test and effector spontaneous and target minimum/target maximum and target minimum × 100.

Viral infection and data analysis were conduced as described previously (18). The protection rates were calculated as the number of sites without papillomas/total challenged sites and statistical significance was determined by Fisher’s exact test (p < 0.05 was considered significant).

To determine whether the transgene was properly expressed on the cell surface, we collected PBLs from both nontransgenic and transgenic rabbits and labeled them with an HLA-A2.1-specific mAb (BB7.2) and subsequently a PE-conjugated secondary Ab. The prepared cells from nontransgenic and transgenic rabbits were analyzed by flow cytometry. The transgenic rabbit PBLs labeled with the secondary PE-conjugated goat-anti-mouse IgG were used as negative control for FACScan analysis (open lines in Fig. 1, A and B). HLA-A2.1-transgenic rabbit PBLs showed homogeneous expression (filled graph in Fig. 1,B) whereas no transgene protein was detected on cells from nontransgenic siblings (filled graph in Fig. 1,A). The transgenic rabbits have been backbred with an inbred rabbit line (EIII/JC) in our animal facility for five generations (F-F₅) and the offspring have retained the transgene and maintained expression of HLA-A2.1 on cell surface of splenocytes. Representative data from rabbits of different generations are shown in Fig. 1 C.

Using standard immunohistochemistry methods, we were able to detect HLA-A2.1 protein expressed extensively in transgenic rabbit ear tissue (Fig. 1,E) but not in tissues from nontransgenic rabbits (Fig. 1 D). We also examined HLA-A2.1 expression in other organ tissues (e.g., liver, lung, tongue, etc.), and all showed strong levels of HLA-A2.1 protein, especially in spleen tissue (data not shown).

HLA-A2.1-transgenic rabbits were generated by traditional methods of gene insertion so that endogenous rabbit MHC genes are retained. Having demonstrated expression of HLA-A2.1 on transgenic rabbit cell surfaces, we wished to determine whether HLA-A2.1 influenced the expression and localization of rabbit MHCI on the cell surface. Using immunofluorescent microscopy, we showed that HLA-A2.1 colocalized with rabbit MHCI on the cell surface of rabbit spleen cells (Fig. 2,A), cells of ear tissue (Fig. 2,B), and spontaneously immortalized RA2 cells (Fig. 2 C).

Next, we wanted to determine whether the HLA-A2.1 protein on transgenic rabbit cell surfaces was functional. For functional testing, we designed experiments to answer two questions: 1) whether HLA-A2.1 could correctly bind specific epitopes added exogenously; 2) whether transgenic rabbit cells could correctly process and present HLA-A2.1-specific epitopes.

A rabbit cell line (RA2) was initially established from cultures of an HLA-A2.1 rabbit PBLs. RA2 cells expressed HLA-A2.1 on the cell surface at levels comparable to that of two human HLA-A2.1-positive cell lines (CaSki and T2, Fig. 3,A) while no positive signals were found on normal rabbit cell surfaces of control cells (Fig. 3 A).

We used the RA2 cell line to test for correct binding and presentation to an HLA-A2.1 peptide-specific CTL (SV40LT/281–289) using the intracellular cytokine assay for IFN-γ production. If RA2 cells could properly bind and present the specific HLA-A2.1-restricted epitope (SV40LT/281–289) to this epitope-specific CTL clone (A2.1-58.3), CTL would be activated and secret cytokines, particularly, IFN-γ. Therefore, more CTLs should be stained positively for intercellular IFN-γ following brefeldin A treatment. In this experiment, we incubated RA2 cells with SV40LT/281–289 (the positive epitope) or HIVGag P17/77–85 (a negative control epitope) and subsequently cultured these peptide-pulsed cells with an epitope-specific CTL clone (A2.1-58.3). Our results showed significantly more IFN-γ-producing CD8 T cells in the presence of RA2 cells pulsed with SV40LT/281–289 when incubated with RA2 cells pulsed with an irrelevant HLA-A2.1-restricted epitope (HIVGag P17/77–85) (p < 0.01, unpaired t test, Fig. 3 B). We conclude that HLA-A2.1-transgenic rabbit cells could effectively and specifically present exogenous epitopes to CTLs.

We next wanted to determine whether RA2 cells could properly process and present this epitope endogenously to the CTL clone. RA2 cells were transfected with SV40LT and the selected line was designated as RA2LT. RA2LT cells were then used as stimulator cells in cocultures of the SV40LT/281–289 epitope-specific CTL clone A2.1-58.3. A mouse HLA-A2.1 cell line expressing SV40LT (A2.1-364-A) and a nontransgenic rabbit cell line (normal rabbit cell line) transfected with SV40LT were designated as mA2LT and RLT and used as positive and negative control cells, respectively. The same protocol described previously was used for this experiment. We detected significantly more IFN-γ-secreting CD8 T cells in cocultures of mouse LT and RA2LT cells compared with those containing control rabbit LT cells (p < 0.01, unpaired t test, Fig. 3 C).

Collectively, these data show that HLA-A2.1 molecules expressed on transgenic rabbit cell surfaces could present both exogenous and endogenous epitopes effectively.

Transgenic rabbits were next tested for CTL responses to HLA-A2.1-restricted epitopes. To test for in vivo function, we constructed a DNA vaccine expressing an HLA-A2.1-specific epitope from HPV16E7 (aa 82–90, Fig 4,A) based on the alignment between CRPVE7 and HPV16E7 and had been shown a biological epitope in other studies (Fig. 4,B). This vaccine had five repeats of this epitope and was used for immunizing both normal and transgenic rabbits. Rabbit spleen cells were harvested and examined for tetramer binding of HLA-A2.1-peptide complexes. Ex vivo staining showed that a small population of CD8⁺ T cells from splenocytes was tetramer positive (Fig. 4,C). We positively selected CD8 T cells from spleen cells using an Easysep magnetic positive selection system and expanded them in culture by stimulating with peptide-pulsed syngeneic cells. The frequency of tetramer-positive T cells increased substantially following in vitro stimulation of cells from the HLA-A2.1-transgenic rabbit but not from cultured cells from the control group rabbit (Fig. 4,C). No tetramer-binding CD8 T cells could be found with an irrelevant tetramer as a control stain (Fig 4,C). Therefore, these tetramer-binding CD8 T cells were specific for HPV16E7/82–90 epitope. These tetramer-positive CD8 T cells also showed epitope-specific killing in vitro (Fig. 4 D).

We conducted additional experiments to determine whether specific immunity could be stimulated in the epitope-vaccinated transgenic rabbits described above. Transgenic and nontransgenic rabbits were divided into two groups each and immunized with either the epitope vaccine (groups 1 and 2) or a control vaccine (groups 3 and 4) (Table I). After two booster immunizations at 3-wk intervals, all rabbits were challenged with either wild-type CRPV or CRPV containing the HPV16E7/82–90 DNA sequence genetically engineered into the CRPV E7 gene at four left and right dorsal skin sites, respectively. The papillomas were monitored weekly until week 12. HLA-A2.1-transgenic rabbits vaccinated with the epitope DNA vaccine had more sites free of tumors induced by the epitope-modified CRPV DNA challenge compared with transgenic or nontransgenic rabbits vaccinated with control vaccine or control rabbits with epitope vaccine (Table I). No difference in frequency was observed between all groups of rabbits that were challenged with wild-type CRPV DNA (Table II).

Although the experiment above confirmed the immunogenicity of a well-known A2-restricted HPV16E7 epitope, we wished to determine the versatility of the model by testing HLA-A2.1-predicted epitopes using epitope-prediction software. To answer this question, we designed a multiple epitope DNA vaccine targeting the CRPVE1 gene based on online MHCI epitope predication software (Materials and Methods, Fig. 5,A) (31). These five peptides showed strong affinity binding to HLA-A2.1 molecules and the binding was stable (data not shown). After immunizing transgenic rabbits three times with this epitope vaccine, we challenged rabbits with wild-type CRPV DNA. All HLA-A2.1-transgenic rabbits vaccinated with the epitope vaccine were protected against viral DNA infection while those vaccinated with control vaccine were not (Fig. 5 B). These data further confirmed that the transgenic rabbit model could be used as an infection model to screen effective HLA-A2.1-restricted epitopes for induction of specific immunity to viral infections.

In this report, we describe the development and functional testing of a novel HLA-A2.1-transgenic rabbit model to study immunity to papillomavirus infections. A summary of our findings demonstrated that: 1) HLA-A2.1 protein was expressed and colocalized with rabbit MHCI on the cell surface; 2) HLA-A2.1-transgenic rabbit cells could present exogenous and endogenous epitopes to an HLA-A2.1 epitope-specific CTL clone; 3) HLA-A2.1-restricted HPV16E7/82–90 epitope DNA vaccine stimulated HLA-A2.1 epitope-specific CTLs in rabbits; and 4) enhanced specific protective immunity against CRPV genome containing HPV16E7/82–90 epitope was generated in these HLA-A2.1-transgenic rabbits vaccinated with the HPV16E7/82–90 epitope DNA vaccine. An additional observation was demonstrated in which complete protective immunity to viral infection was induced in transgenic rabbits vaccinated with a multivalent DNA epitope vaccine based on epitope prediction software. Collectively, these data indicate that the HLA-A2.1-transgenic rabbits can be used as a model system to screen HLA-A2.1-restricted epitopes and to test their immunogenicity and induction of protective immunity in vivo.

The CRPV/rabbit model has been used for many years as an important preclinical model for natural papillomavirus infections. Both prophylactic and therapeutic papillomavirus vaccines have been tested in the rabbit model (16). A secondary reason for the increased importance of the rabbit (and canine and bovine) model is that there are currently no laboratory rodent models of natural papillomavirus infections (15). One key observation for expanding the utility of the CRPV/rabbit model is that papillomas can be induced following direct application of viral DNA to cutaneous sites (20, 34). We and others have conducted numerous mutagenesis studies and found considerable plasticity of the CRPV genome following a variety of mutations (35, 36). The ability to alter the CRPV genome has enabled us to genetically relocate an HLA-A2.1-restricted epitope from the HPV16 E7 gene into the CRPV E7 gene without significant loss of viral function. We do predict, however, that not all such epitope relocations will be successful with respect to retention of viral function, but the current study demonstrates proof of principle for such an approach. This option is critical to the analysis of HLA-A2.1-restricted responses to epitopes of an important human pathogen. Without this option, the transgenic rabbit model would be limited to an analysis of HLA-A2.1-restricted epitopes that are present in the native rabbit papillomavirus proteins only.

MHCI-restricted epitope DNA vaccines have been used successfully in mouse models (32, 37, 38, 39). We exploited the same epitope DNA design in our transgenic rabbit model, and found this method to successfully induce rabbit CD8 T cell responses to defined HLA-A2.1 epitopes. In conjunction with the gene-gun delivery system that preferentially induces cell-mediated immune responses in our immunized rabbits; we achieved complete protection in transgenic rabbits with a predicted five-epitope DNA vaccination. These findings indicate that we can test predicted HLA-A2.1-restricted epitopes from a variety of theoretical and experimental candidate sequences and test their immunogenicity and, more importantly, their protective immunity in vivo using the transgenic rabbit model.

HLA-A2.1-transgenic mice have been used to evaluate antitumor immunotherapeutic strategies preclinically (14). We have immunized mice with recombinant vaccinia virus expressing HPV16E7/82–90 and similar results were found when compared with those in HLA-A2.1-transgenic rabbits. However, it is not possible to use mice to investigate protective immunity to certain human pathogens because mice are not natural hosts for these infectious agents. Rabbits, in contrast, have been demonstrated as permissive or semipermissive to important pathogenic viruses such as human T cell leukemia virus 1 (21), EBV-like viruses (23), HIV type 1 (HIV-1) (40, 41), SV40 (42), and adenovirus type 5 (22), and nonviral pathogens such as syphilis and tuberculosis (24). Therefore, the HLA-A2.1-transgenic rabbits represent an important model to complement the HLA-A2.1-transgenic mouse models to study pathogen-host interactions and immune protection against this subset of pathogens. An additional complementary role for these two transgenic models is envisaged in which the HLA-A2.1 mouse model can be used to more quickly develop HLA-A2.1-restricted CD8⁺ T cell lines using in vitro culture techniques and reagents that are well-defined in current and past protocols.

There are several limitations of the CRPV/rabbit model to study papillomavirus infections. For example, CRPV is a cutaneous papillomavirus, and thus is not a model for studying genital papillomavirus infections that occur with HPV16 and HPV18. However, we do have access to a second rabbit papillomavirus, rabbit oral papillomavirus which targets mucosal tissues. Our HLA-A2.1-transgenic rabbit model can be easily adapted to studies with rabbit oral papillomavirus infection in rabbits. A second deficiency of the CRPV model is that the infection requires a long interval of time for data collection especially for investigations on malignancy. We recently developed several mutant CRPV genomes which show significantly shortened time for cancer development and this latter finding would facilitate studies on immunity to papillomavirus-induced malignancy (data not shown). A final concern is that there are only limited reagents developed for immunological studies using rabbits. Recent studies have screened and identified several human cytokines that are effective in cultures of rabbit cells. For example, recombinant human IL-2 supported rabbit CD8 T cell growth. Other investigators and our preliminary data have shown that human IL-4 and GM-CSF can support rabbit dendritic cell growth in vitro (43). Several rabbit cytokine and chemokine genes have been sequenced so that there is an opportunity to develop Ab probes for molecular and cellular analyses of rabbit immune function. In addition, there is increasing interest in the rabbit as a research model with the development of transgenic rabbit models of atherosclerosis, heart function, and immune function (44, 45).

In this study, we describe the utility of an HLA-A2.1-transgenic rabbit model to analyze HLA-A2.1-restricted epitopes of HPV16 E7 following the embedding of an HPV16 E7 epitope into the E7 gene of CRPVE7 and subsequent infection of the transgenic rabbit skin. We have also demonstrated that the model can be used to test software-predicted HLA-A2.1 epitopes of a papillomavirus gene in vivo after specific epitope vaccination. These initial studies indicate that the HLA-A2.1-transgenic rabbits will provide important additional opportunities to test protective and therapeutic vaccines to several human pathogens that show permissiveness or semipermissiveness in rabbits.

We thank Martin Pickel and Karla Balogh for excellent help with the animals and Dr. Victor Engelhard for the HLA-A2.1 genomic sequence.

The authors have no financial conflict of interest.