HIV-1 p24 vaccine protects cats against feline immunodeficiency virus infection

Background: Based on previous analysis of feline immunodeficiency virus (FIV)-specific cross-reactive antibodies to HIV-1 p24, cats vaccinated with HIV-1 p24 were evaluated for cross-reactive immunity to FIV.

Objective: To determine the level of cross-reactivity that exists between HIV-1 and FIV p24 and its implications for vaccine prophylaxis.

Methods: Specific-pathogen-free cats were immunized three times with HIV-1 p24 in Ribi adjuvant, with (n = 18) or without cytokine (n = 6). Control cats were immunized three times with adjuvant (n = 10) or phosphate-buffered saline (PBS; n = 5). All immunized cats were challenged with either subtypes B or A/B FIV, and monitored by virus isolation, proviral PCR, FIV-specific antibodies, and feline interferon-γ ELISpot for T-cell activities.

Results: Of 18 cats vaccinated with subtype B HIV-1 (HIV-1_LAI/LAV, HIV-1_UCD1) p24 in Ribi/cytokine adjuvant 14 (78%) were protected against FIV challenges (subtype A_gag and B_gag) that infected all 15 adjuvant- or PBS-immunized cats. Furthermore, only three of six (50%) cats vaccinated with FIV p24 in Ribi/cytokine adjuvant were protected against similar FIV challenge. HIV-1 p24 vaccination induced weak cross-reactive antibodies to FIV p24, which did not correlate with vaccine efficacy. However, the peripheral blood mononuclear cells from HIV-1 p24-vaccinated/protected cats at 33-34 weeks post-FIV challenge responded to three T-cell responsive peptides at the carboxyl-terminus of the FIV p24, whereas those cells from the infected control cats had minimal to no responses to the same peptides.

Conclusions: These results suggest the importance of including lentiviral p24 as vaccine immunogen for human AIDS vaccine. Moreover, these results suggest the potential importance of evolutionarily conserved, cross-protective epitopes in vaccine protection.

Introduction

Although major strides have been made with antiviral drug therapy, the development of an effective vaccine against HIV-1 is still considered central to the control of the AIDS epidemic ^[1]. Multiple HIV vaccine designs using different HIV-1 strains are currently in various phases of clinical trials ^[1-3]. Even in light of these clinical trials, it is still unclear what HIV-1 epitopes and immune mechanisms are essential for vaccine protection ^[1,4]. Similar issues were faced during the development of an feline immunodeficiency virus (FIV) vaccine for domestic cats ^[5]. As a means to broaden FIV vaccine efficacy, a dual-subtype vaccine was developed using FIV strains from long-term nonprogressor cats ^[5,6]. This vaccine demonstrated moderate to significant protection of cats against both homologous and heterologous FIV challenges ^[5-8]. Furthermore, this vaccine induced not only broad virus neutralizing (VN) antibodies ^[7] but antibodies cross-reactive to HIV-1 proteins, especially to HIV-1 core protein (p24) and group-specific antigens (Gag) ^[9]. The FIV epitopes responsible for providing the dual-subtype vaccine protection have yet to be determined.

Gag and other antigens conserved among viruses from the same subfamily frequently induce antibodies that cross-react with other subfamily members ^[9-13]. Some cross-reactive antigens have been successfully used as immunogens for vaccine against viruses from the same subfamily, such as the vaccinia virus vaccines for smallpox in humans and human measles vaccines for canine distemper in puppies ^[14,15]. Therefore, protective vaccines based on cross-reactive antigens can provide broad immunity, and may be useful against viruses that are currently evolving in a new host, such as HIV infection of humans.

Although cross-protection against HIV-1 with prior HIV-2 infection has been reported in multiple retrospective studies ^[16,17], controversy still exists with multiple studies reporting no protection ^[18,19]. Even though amino acid (aa) sequences of the structural proteins exhibit only limited identity (< 60%) between HIV-1 and HIV-2 ^[20], some of the cross-reactive epitopes between these two major HIV groups have been reported to induce cross-neutralizing antibodies and cross-reactive cytotoxic T lymphocyte (CTL) activities ^[21-24]. Moreover, prime-boost with poxvirus-vectored recombinant HIV-1 vaccine followed by HIV-1 protein, conferred cross-protection against HIV-2 challenge in macaques ^[25]. However, cross-reactive antigen-induced protective immunity has not been reported against distinct heterologous-species viruses (HIV-1 and FIV). The current studies were undertaken to determine the level of cross-reactivity that exists between HIV-1 and FIV.

Material and methods

Vaccine efficacy of studies 1-4

Four separate studies were performed to evaluate the protective efficacy of HIV-1 p24 vaccine against FIV_Bang and FIV_FC1 strains in cats (Table 1, Studies 1-4). FIV_FC1 belongs to subtype B and FIV_Bang is a recombinant of subtype A_{gag,pol,env(V1-V3)} and subtype B_env(V4-V9) ^[8]. The p24 from subtype B HIV-1 (HIV-1_UCD1 or HIV-1_LAI/LAV) that cross-reacted strongly with sera from dual-subtype FIV vaccinated cats was used as vaccine immunogen ^[7]. The HIV-1 p24 vaccine consisted of recombinant p24 (200-250 μg) of HIV-1_UCD1 or HIV-1_LAI/LAV mixed in modified Ribi adjuvant (Corixa Corporation, Hamilton, Montana, USA) (1 ml/dose) containing recombinant human interleukin-12 (rHuIL-12, Genetics Institute, Cambridge, Massachusetts, USA; 5 μg/dose) (Ribi/rHuIL-12), recombinant feline IL-18 (rFeIL-18; 5 μg/dose) (Ribi/rFeIL-18), or without cytokine (Ribi). The modified Ribi adjuvant (modification of Corixa's Ribi R-730) contained cell wall skeleton (25 μg/dose) ^[26] and Escherichia coli M15 lipopolysaccharide (5-50 EU/dose) as our approach to enhance toll-receptor recognition ^[27]. In Study 1 (Table 1), the cats were immunized with either HIV-1_UCD1 p24 in Ribi/rHuIL-12 (Group 1), HIV-1_LAI/LAV p24 in Ribi/rHuIL-12 (Group 2), HIV-1_UCD1 p24 in Ribi (Group 3), or Ribi alone (Group 4). In Study 2, the cats were immunized with either HIV-1_LAI/LAV p24 in Ribi/rHuIL-12 (Group 1), FIV_Bang p24 in Ribi/rHuIL-12 (Group 2), FIV_Pet/Shi p24 in Ribi/rHuIL-12 (Group 3), or Ribi/rHuIL-12 (Group 4). In Study 3, the cats were immunized with HIV-1_UCD1 p24 in Ribi/rFeIL-18 (Group 1) or Ribi (Group 2); and the control cats were immunized with Ribi/rFeIL-18, Ribi, or phosphate-buffered saline (PBS; Group 3). In Study 4, the cats were immunized with HIV-1_UCD1 p24 in Ribi/rHuIL-12 (Group 1) or with PBS (Group 2).

Table 1

All cats were challenged intravenously with 15 median cat infectious dose (CID₅₀) of inoculum consisting of either pooled FIV_Bang-infected plasma (Studies 1-3) or pooled FIV_FC1-infected peripheral blood mononuclear cells (PBMC) (Study 4) derived directly from infected animals ^[7]. Infection status of the cats was monitored by virus isolation based on reverse transcriptase activity and proviral PCR performed on the PBMC collected every 3-4 weeks post-inoculation (wpi) and on the lymph node (LN) and bone marrow (BM) cells at 18 or 24 wpi (Study 2) ^[7] or at 33-34 and 52 wpi (Studies 1, 3, 4). Cats were considered FIV negative by the absence of detectable virus, VN antibodies, and immunoblot antibodies to FIV [nucleoprotein/matrix (p10/p15), polymerase (p65, p50), transmembrane envelope (gp40), and surface envelope (gp95)] at 1: 50 and 1: 250 serum dilutions ^[7].

Vaccine immunogenicity

The FIV_Bang p24 reactivity of the antibodies from vaccine-immunized cats and adjuvant/PBS-immunized controls were determined by immunoblot analysis using FIV_Bang p24 as a substrate. The cross-reactive specificity of these antibodies (1: 200 dilution) was determined by ELISA ^[28] using overlapping 28-30 mer peptides of FIV_Bang p24 as substrates. Only those reactivities that were threefold the pre-immunization/pre-infection sera were considered positive. The FIV/HIV-1 specific cellular immune responses were determined by feline interferon-γ (FeIFNγ) ELISpot of dendritic cell (DC)-primed PBMC from vaccine-immunized and PBS-immunized cats. Briefly, plastic adherent BM cells were cultured in RPMI media containing recombinant feline granulocyte-macrophage colony-stimulating factor (50 ng/ml; R&D Systems, Minneapolis, Minnesota, USA) and rFeIL-4 (25 ng/ml) produced using a previously described method ^[29,30]. On culture day 10, the non-adherent fraction was re-cultured for additional 2 days in RPMI media containing recombinant human tumor necrosis factor-α (50 ng/ml, R&D Systems), to induce maturation ^[31]. The mature DC were then incubated with peptide (2 μg/0.2 ml/well) in 96 wells (2 × 10⁴ cells/well) for 18 h and transferred to FeIFNγ ELISpot plates containing 2 × 10⁵ autologous PBMC. ELISpot plates were processed according to manufacturer's method (R&D Systems) and analyzed with an ELISpot reader (MVS Pacific, LLC, Minneapolis, Minnesota, USA).

Sequencing, expression, and purification of HIV-1 p24, FIV p24, and FeIL-18

The proviral genes of HIV-1_LAI/LAV (kindly provided by F. Barre-Sinoussi, Pasteur Institute, France; NCBI accession #K02013), HIV-1_UCD1 (isolated from an HIV-positive individual from San Diego, California, USA; NCBI #AY679786), FIV_Bang (NCBI #AY684181), FIV_Pet (NCBI #NC001482), and FIV_Shi (NCBI #D37818; #AY679785) were used to derive the viral p24 proteins. The whole gag sequence was obtained by PCR amplification of proviral DNA from HIV-1_UCD1- or HIV-1_LAI/LAV-infected HuT-78 cells, and cloned into pCR2.1 vector (TA cloning kit, Invitrogen, Carlsbad, California, USA). Three to eight clones were used for sequence confirmation. The HIV-1_LAI/LAV p24 sequence was identical to the reported HIV-1_HXB2 p24 sequence (NCBI #K03455) and differed from HIV-1_UCD1 by eight amino acids. The confirmed sequence of p24 gene was subcloned into pQE30 for expression in E. coli M15[pREP4] (QIAexpressionist, Qiagen Inc., Valencia, California, USA). Recombinant p24 product was purified on Ni-NTA resin using the manufacturer's protocol (> 97% purity as determined by silver stain analysis). The specificity of the product was determined by immunoblot analysis using human and feline polyclonal antibodies reactive to either HIV-1 or FIV p24, respectively.

The rFeIL-18 was produced from an E. coli expression system using a previously described method with modifications ^[32,33]. The FeIL-18 sequence was obtained by reverse transcription (RT)-PCR amplification of mRNA from feline splenocytes. Amplified products were cloned into pCR2.1 vector for sequence confirmation based on published FeIL-18 sequence ^[32]. The confirmed sequence was subcloned into pMAL-c2G (New England Biolabs Inc., Beverly, Massachusetts, USA) for expression in E. coli ER2508 strain as a fusion protein (FeIL-18/maltose-binding protein). The fusion protein was cleaved with Genenase I and the cleaved product was purified by amylose resin chromatography (New England Biolabs Inc.) to > 95% purity. Biological activity of the rFeIL-18 was determined by proliferation-based IL-18 bioassay using PBMC instead of spleen cells ^[34].

Statistical analyses

Individual immunization groups in each study (Table 1) and combined groups from different studies (Table 2) were analyzed for statistical significant difference by analysis of variance (ANOVA). Comparisons that demonstrated overall significance by ANOVA were evaluated by two-way paired t test (SAS program, version 8.0) and were considered to have statistical difference when P < 0.05.

Table 2

Results

Challenge efficacy studies

In Study 1 (Table 1), all four cats (100%, Group 1) vaccinated with HIV-1_UCD1 p24 in Ribi/rHuIL-12 were protected against FIV_Bang challenge that infected all three control cats (Group 4). In contrast, only two of four cats (50%, Group 2) vaccinated with HIV-1_LAI/LAV p24 in Ribi/rHuIL-12, and one of three cats (33%, Group 3) vaccinated with HIV-1_UCD1 p24 in Ribi were protected. Statistically significant difference was observed only between the protection rates of Groups 1 and 4 (P < 0.01). All vaccinated/protected cats in Study 1 were FIV negative even at 52 wpi. In Study 2, all three cats (100%, Group 1) vaccinated with HIV-1_LAI/LAV p24 in Ribi/rHuIL-12 were protected, while all three Ribi/rHuIL-12-immunized cats (Group 4) were infected. However, only one of three FIV_Bang p24-vaccinated cats (33%, Group 2) and two of three FIV_Pet/Shi p24-vaccinated cats (67%, Group 3) were protected.

Using a different T-helper 1 cell (Th1)-promoting cytokine adjuvant, in Study 3, two of three cats (67%, Group 1) vaccinated with HIV-1_UCD1 p24 in Ribi/rFeIL-18 were protected as of 52 wpi (Fig. 1, Table 1), while all six control cats (Group 3) were infected (typical results from three cats shown, Fig. 1; Table 1). Again, only one of three cats (33%, Group 2) vaccinated with HIV-1_UCD1 p24 in Ribi was protected, similar to the observation in Study 1 that showed HIV-1 p24 formulated in Ribi/rHuIL-12 to be more effective than those formulated in only Ribi (Table 1). In order to determine whether HIV-1 p24 protection is effective against challenge strains distinctly different from the subtype A gag of FIV_Bang, HIV-1_UCD1 p24-vaccinated cats were challenged with subtype B FIV_FC1 in Study 4 (Fig. 1, Study 4; Table 1). Three of four cats (75%, Group 1) vaccinated with HIV-1_UCD1 p24 in Ribi/rHuIL-12 were protected as of 52 wpi against FIV_FC1 challenge that infected all three control cats (Group 2). These observations demonstrate the protective efficacy of subtype B HIV-1 p24 against FIV strains containing distinct subtype A and B gag domains. Furthermore, analysis of the vaccinated/unprotected cats of Studies 3 and 4 (Fig. 1) indicates that none of these cats had enhancement in FIV infection.

Fig. 1

Cross-reactive antibodies to FIV p24

ELISA- and immunoblot-antibody titers to HIV-1 and FIV p24 proteins in Studies 1-4 increased with each HIV-1 p24 immunization (only immunoblot results shown, Fig. 2a). Thirteen of fifteen (87%; 12 ELISA positive, Fig. 3a; 13 immunoblot positive, data not shown) vaccinated/protected cats tested and six of eight (75%, data not shown) vaccinated/unprotected cats developed either ELISA or immunoblot antibodies to FIV p24 after the third vaccination before challenge. The protected cats were negative for antibodies to FIV transmembrane peptide, TM695-705 (Fig. 3a), while all five FIV-infected cats were positive for antibodies to TM(695-705). These results suggest that the development of cross-reactive antibodies to FIV p24 protein does not correlate with vaccine protection. Furthermore, no VN antibodies to FIV subtype A, B, and D strains were detected in the cats with either HIV-1 or FIV p24 vaccination (data not shown). In contrast, the sera from cats immunized with commercial whole-virus FIV_Pet/Shi vaccine (Fel-O-Vax FIV, Fort Dodge Animal Health, Fort Dodge, Iowa, USA) had moderate titers of VN antibodies to FIV_Pet and FIV_Shi but only weak VN titers to FIV_Bang and FIV_FC1.

Fig. 2

Fig. 3

Sequence and B-cell epitope analyses

The HIV-1_LAI/LAV p24 sequence was identical to the reported HIV-1_HXB2 p24 sequence (NCBI #K03455) and differed from HIV-1_UCD1 p24 by eight amino acids (Fig. 2b). Sequence comparisons between HIV-1_UCD1 p24 (231 aa) and either FIV_Bang or FIV_FC1 p24 (223 amino acids) showed only 30.9% and 30.3% amino acid identity, respectively (Fig. 2b). Only regions toward the carboxyl-terminus contained relevant homology, consisting of four amino acids between HIV-1_UCD1 and FIV_Bang (HIV-1 residues 210-213: TLEE) and seven amino acids between HIV-1_UCD1 and FIV_FC1 (HIV-1 residues 193-199: NANPDCK). Comparisons between HIV-1_LAI/LAV and either FIV_Bang p24 or FIV_FC1 p24 exhibited 31.3% and 30.8% amino acid identity, respectively. The regions that had the longest identical sequence were the same as those observed between HIV-1_UCD1 and the two FIV strains. Overall, the carboxyl-terminal region contained more conservation between the HIV-1 and FIV p24 sequences. Sequence analysis of p24 from HIV-1_UCD1 and HIV-1_LAI/LAV demonstrates seven of eight amino acid differences found at the amino-terminus (Fig. 2b). Based on antibody reactivity to overlapping 28-30 mer peptides of FIV p24 (Fig. 3a), the antibodies produced by HIV-1_UCD1 p24-vaccinated cats before challenge reacted strongest with peptide FB4/71-100, moderately with peptide FB11/197-223, and minimally with peptides FB3/53-81, FB9/161-188, and FB10/178-207. Therefore, at least two cross-reactive B-cell epitopes are recognized by the antibodies that are induced by HIV-1 p24 vaccination. According to our hydropathy plot analysis, the reactive peptides are in the hydrophilic regions, which generally contain the B-cell antigenic sites (data not shown). The pattern of cross-reactivity to FIV peptides (FB4/71-100 and FB11/197-223) with sera from HIV-1 p24-vaccinated cats is similar to the antibody reactivity of sera from FIV-infected cats (Fig. 3a). The presence of antibodies to peptides FB4/71-100 and FB11/197-223 did not correlate with vaccine protection, since protection was also achieved in HIV-1 p24-vaccinated cats that had no cross-reactive antibodies to these peptides. In fact, two vaccinated/protected cats developed no ELISA and immunoblot antibodies to FIV p24 protein and peptides.

Cellular immune analyses

A selected number of cats from Studies 3 and 4 were tested for cellular immune function (FeIFNγ ELISpot) after challenge and after demonstration of full protection at 33-34 wpi. Since the DC-priming was unavailable in our laboratory until recently, only three vaccinated/protected cats (#JB6 and #IW1, Study 3; #MF3, Study 4) and two infected control cats (#MK4 and #MG5, Study 4) were tested by FeIFNγ ELISpot analysis of the DC-primed PBMC. The PBMC from all three vaccinated/protected cats had strong FeIFNγ ELISpot responses to three FIV p24 peptides (F4.5/122-131, F6/155-164, F7/183-191) (≥ baseline, Fig. 3B1). Two of these cats also responded to peptide F7.8/210-219. Furthermore, the PBMC from all three cats responded to the three counterpart HIV-1 p24 peptides (H4.5/130-140, H6/162-172, H7/191-199) and also to peptide H7.8/217-226 (≥ baseline, Fig. 3B2). In contrast, the PBMC from infected control cat (#MK4) had only minimal FeIFNγ ELISpot responses to three FIV p24 peptides (F2/15-25, F7.8/210-219, F8/214-223) (≥ baseline, Fig. 3b3), while the PBMC from another infected cat (#MG5) had no responses to any of the nine FIV p24 peptides (HIV-1 peptides not tested). Since these targeted peptides corresponded to HIV-1 sequences that are the reported CTL epitopes for HIV-1-positive individuals ^[36], we speculate that they may represent FIV-specific CTL epitopes. Future studies using separated CD4 and CD8 T-cell subpopulations from vaccinated cats will determine whether the responses are CTL, Th, or both.

Discussion

Cats immunized with HIV-1 p24 vaccine were protected against FIV-challenge infection. Furthermore, the HIV-1 p24 vaccine was effective against in vivo-derived inocula of both subtype B FIV_FC1 and recombinant subtype A/B FIV_Bang. An overall protection rate of 78% (14 of 18; Group A1, Table 2) achieved in the cats vaccinated with either HIV-1_UCD1 or HIV-1_LAI/LAV p24 in Ribi/cytokine, was significantly different (P < 0.001) from the 0% protection rate (0 of 15; Group A2) observed in the corresponding combined control group. The 78% protection rate was better than the 50% (3 of 6; Group E1) combined protection rate of the cats vaccinated with either FIV_Bang or FIV_Pet/Shi p24 in Ribi/rHuIL-12 (P = 0.170, Groups E1 versus E2; Table 2). These observations suggest that the HIV-1 p24 vaccine is as effective as or may be better than FIV p24 vaccine in protecting cats against FIV infection.

Although antibody titers to HIV-1 p24 in Studies 1-4 increased with each immunization, HIV-1 p24 vaccine protection did not correlate with the induction of VN antibodies or with the level or the B-cell epitope specificity of the antibodies to either HIV-1 or FIV p24. The cross-reactive antibodies reacted with FIV peptides FB4/71-100 and FB11/197-223, which also reacted strongly with antibodies from FIV-infected cats. None of these peptide sequences corresponded to the major homology region (MHR) previously reported for HIV-1, SIV, and FIV ^[10,11]. These cross-reactive antibodies to FIV p24 were most likely negative for VN activity because no VN antibodies to FIV were detected in the sera from HIV-1_LAI/LAV or HIV-1_UCD1 p24-vaccinated cats. The absence of VN antibody production in p24-vaccinated cats was anticipated because VN antibodies only reactive to envelope glycoproteins have been reported to date ^[37]. Moreover, our preliminary passive immunization results demonstrate 100% passive protection (2/2) with sera from dual-subtype FIV vaccinated cats with high VN antibody titers, but only 25% passive protection (1/4) with sera from HIV-1_LAI/LAV p24-vaccinated cats at a challenge FIV dose that infected all (2/2) control (data not shown).

Based on our FeIFNγ ELISpot studies, strong CTL/Th responses to FIV peptides F4.5/122-131, F6/155-164, and F7/183-191 were observed in the vaccinated/protected cats (Fig. 3). Peptide F6/155-164 overlaps the MHR, which contains CTL, Th, and antibody epitopes recognized by HIV-positive individuals ^[36]. Since these vaccinated/protected cats were tested by FeIFNγ ELISpot analysis at 33-34 wpi, it remains to be determined whether their responses to these FIV p24 peptides were due to the HIV-1 p24 vaccination or to an occult or low-level infection. However, these cats were negative for FIV infection by immunoblot analysis and by virus isolation and proviral PCR of the tissues collected at the same time as the PBMC used for the ELISpot analysis (Fig. 1). Furthermore, both infected control cats tested by ELISpot lacked major responses to these peptides, suggesting that the specific FIV p24-peptide responses of the vaccinated/protected cats, most likely, are not caused by active infection. Nevertheless, a possibility exists that the FIV challenge provided a boosting effect without causing active infection.

The 82% protection rate (9/11; Group B1, Table 2) of the cats vaccinated with HIV-1_UCD1 p24 in Ribi/Th1-promoting cytokine (rHuIL-12 or rFeIL-18) was statistically different from the 33% protection rate (2/6; Group D1) of the cats vaccinated with HIV-1_UCD1 p24 in Ribi (P = 0.049, Table 2). Such comparison demonstrates the importance of using adjuvants that induce strong cellular immunity for optimizing the efficacy of HIV-1 p24 subunit vaccines. Based on the ELISpot analysis, the HIV-1 p24 vaccination alone or in combination with FIV challenge induced strong IFNγ-specific cellular immune responses to HIV/FIV p24 peptides. These observations taken together with our result showing no correlation between the levels of protection and cross-reactive antibodies, suggest that the cross-protection observed with HIV-1 p24 vaccination is more likely to be mediated by cellular immunity.

Recent trials in HIV-1 vaccine development are utilizing gag-expressing recombinant vaccines with or without p24 CTL epitopes ^[3,38-40]. In one study, macaques immunized with SIV gag plasmid DNA priming followed by adenovirus-vectored SIV gag boost had significantly lower challenge virus (SHIV 89.6P) load when compared to the virus load in unvaccinated macaques ^[38]. In another study, immunization three times with adenovirus-vectored SIV gag vaccine also conferred partial protection against SHIV challenge ^[39]. These gag-based vaccines elicited CD4 T-cell response in addition to the dominant anti-Gag CD8 CTL response ^[39,40], suggesting that T-cell immunity may be essential for the partial protection. Our current findings demonstrate the importance of viral p24 immunogen in the efficacy of gag-based vaccines and further support the role of vaccine-induced T-cell responses in conferring protection.

The FIV epitopes important for FIV vaccine protection have yet to be defined, but findings from our studies suggest that at least some of the protective vaccine epitopes for cellular immunity reside on FIV_Pet/Shi and HIV-1 p24. Studies by others demonstrated no protection with the FIV_UK8 p24 vaccine and FIV_AM19K Gag protein-containing vaccine against homologous FIV_UK8 and FIV_AM19 challenges, respectively ^[5,41,42]. In fact, FIV_UK8 p24 vaccination enhanced homologous challenge infection ^[41]. Although statistical difference was not achieved, the efficacy of the HIV_UCD1 p24/Ribi/rHuIL-12 vaccine (4/4, Table 1) against FIV_Bang challenge was better than the efficacy of the FIV_Bang p24/Ribi/rHuIL-12 vaccine (1/3). The minimum-to-no sterilizing protection with FIV_Bang p24 vaccine in Study 2 and with FIV_FC1 p24/Ribi/rHuIL-12 vaccine (0/4 protected) in preliminary study (data not shown) are similar to the findings of SIV gag vaccines against homologous backbone SHIV challenge, where only a decrease in virus load was observed and no sterilizing immunity ^[38,39]. These results suggest what may be important is the selection of the strain(s) to be used for the vaccine in addition to the adjuvant used.

This is the first study to demonstrate protection against a feline AIDS lentivirus with a vaccine consisting of only HIV-1 protein, more specifically subtype B HIV-1 p24. In summary, three main points can be derived from the above findings: (i) HIV-1 proteins from subtype B strains can protect cats against FIV infection, suggesting that the epitopes shared by members of the lentivirus family may have protective properties; (ii) such protection did not correlate with antibody immunity; and (iii) Th1-promoting cytokines in the adjuvant greatly enhanced the HIV-1 p24 vaccine efficacy. These findings taken together suggest that the cross-protection observed in HIV-1 p24-vaccinated cats is most likely mediated by the cellular immunity. Future studies should determine whether FIV-cat model using HIV-1 vaccine is a useful small animal AIDS model to characterize protective epitopes for HIV-1 vaccine design.

Acknowledgements

This work was supported by NIH R01-AI30904 and JKY Miscellaneous Donors Fund.

Conflict of interest statement required by NIH to be placed in acknowledgment: J.K. Yamamoto is the inventor of record on a University of Florida patent and may be entitled to royalties from companies that are developing commercial products that are related to the research described in this paper.