Priming-Boosting Vaccination with Recombinant Mycobacterium bovis Bacillus Calmette-Guérin and a Nonreplicating Vaccinia Virus Recombinant Leads to Long-Lasting and Effective Immunity

ABSTRACT Virus-specific T-cell responses can limit immunodeficiency virus type 1 (HIV-1) transmission and prevent disease progression and so could serve as the basis for an affordable, safe, and effective vaccine in humans. To assess their potential for a vaccine, we used Mycobacterium bovis bacillus Calmette-Guérin (BCG)-Tokyo and a replication-deficient vaccinia virus strain (DIs) as vectors to express full-length gag from simian immunodeficiency viruses (SIVs) (rBCG-SIVgag and rDIsSIVgag). Cynomolgus macaques were vaccinated with either rBCG-SIVgag dermally as a single modality or in combination with rDIsSIVgag intravenously. When cynomologus macaques were primed with rBCG-SIVgag and then boosted with rDIsSIVgag, high levels of gamma interferon (IFN-γ) spot-forming cells specific for SIV Gag were induced. This combination regimen elicited effective protective immunity against mucosal challenge with pathogenic simian-human immunodeficiency virus for the 1 year the macaques were under observation. Antigen-specific intracellular IFN-γ activity was similarly induced in each of the macaques with the priming-boosting regimen. Other groups receiving the opposite combination or the single-modality vaccines were not effectively protected. These results suggest that a recombinant M. bovis BCG-based vector may have potential as an HIV/AIDS vaccine when administered in combination with a replication-deficient vaccinia virus DIs vector in a priming-boosting strategy.

As the rate of new infections with human immunodeficiency virus type 1 (HIV-1) continues to increase globally, an effective preventive vaccine is urgently needed to stem further spread of the virus (24). Because long-term survival in humans has been observed when HIV-1 replication is controlled by protective immunity (12,29), targeted experimental immunogens have been designed to closely mimic the long-lasting protective immunity induced in long-term human survivors by the natural infection (8,25). Recently, various vaccine modalities, including live viral vectors and DNA, have been used to elicit protective immunity in nonhuman primate models (9). However, before an HIV-1 vaccine regimen can be considered promising, it must be shown to be not only effective at inducing protective immunity, but also safe, affordable, and compatible with other vaccines (2,32).
When it comes to safety, traditional live vaccines, which have been administered safely to both the healthy and the infected, may be the vectors of choice for HIV-1 vaccines. In order to fully take advantage of the potential benefits of traditional live vectors in HIV-1 vaccine development, we studied the Myco-bacterium bovis bacillus Calmette-Guérin (BCG) substrain Tokyo 172 (6) and the replication-deficient vaccinia virus vaccine strain DIs (22,50), both of which have been shown to be nonpathogenic when inoculated into immunodeficient animals (41,51,53) as live recombinant vaccine vehicles (1,(17)(18)(19)(46)(47)(48). As further evidence of the potential of the live vectors for use in HIV/AIDS vaccines, we noted that a recombinant M. bovis BCG vector candidate vaccine for HIV-1-induced positive immune responses in animals (17,46). Moreover, we found that recombinant vaccinia virus DIs encoding the simian immunodeficiency virus (SIV) gene was effective at eliciting anti-SIV immunity in mice when administered as a booster antigen after priming with SIV DNA (47). In this study, we have developed a new combination regimen, priming with recombinant M. bovis BCG-SIV Gag followed by boosting with rDIsSIVgag. This immunization regimen elicited effective positive immunity against an immune deficiency virus in macaques for the 1 year they were under study. tion of the Japanese Association for Laboratory Animal Science, 1987, under the Japanese Law Concerning the Protection and Management of Animals (46) and were maintained in accordance with the guidelines set forth by the Institutional Animal Care and Use Committee of National Institute of Infectious Diseases, Japan. Once approved by an institutional committee for biosafety level 3 experiments, these studies were conducted at the Tsukuba Primate Center, National Institute of Infectious Diseases, Japan, in accordance with the requirements specifically stated in the laboratory biosafety manual of the World Health Organization. The animals' condition was monitored by analyzing a hemogram parameter as well as absolute CD4 ϩ and CD8 ϩ T-lymphocyte counts with an automated blood analyzer Celltac (Nihon Koden, Tokyo Japan), as described below.
Two thousand 50% tissue culture infectious doses (TCID 50 ) of SHIV KS661c, a pathogenic molecular clone, were intrarectally administered as a challenge virus (39). The parent virus, SHIV-C2/1, is an SHIV-89.6 variant isolated by in vivo passage in cynomolgus macaques (40,42) and the original SHIV-89.6 strain was kindly provided by Y. Lu at the Harvard AIDS Institute (Boston, MA) (26,37). SHIV-C2/1 and SHIV KS661c were shown to infect cynomologus macaques by both the intravenous and intrarectal routes (39). Both viruses induced high levels of viremia and marked CD4 ϩ T-cell depletion within 2 and 3 weeks after inoculation, respectively (39,40,42). Virus stocks were stored at Ϫ125°C and thawed just prior to use.
Production and preparation of recombinant M. bovis BCG (rBCG) and vaccinia virus DIs expressing full-length SIV Gag. Detailed methods for plasmid construction were described previously (7,17,18,21). Briefly, a DNA fragment encoding the full-length gag sequence of SIVmac239 was cloned downstream of the hsp60 promoter (52) and then inserted into the multicloning site of the plasmid pSO246 (28). Recombinant Mycobacterium bovis BCG substrain Tokyo 172 that stably expressed the inserted DNA fragment (designated rBCG-SIVgag) was then selected and used for all rBCG inoculations. For the Western blot analysis, the transformant of rBCG was grown in 7H9-ADC broth for 2 weeks and a portion of the culture medium was periodically collected, sonicated and blotted using the monoclonal antibody IB6, as described previously (47). Since the recombinant DIs virus (rDIs) encoding the SIVmac239 gag-pol open reading frame elicited remarkably high SIV Gag-specific T-cell responses but low polymerase responses in mice (47), confirming the findings of a previous report (20), we named it rDIsSIVgag. The rDIsSIVgag and rDIs encoding ␤-galactosidase (rDIsLacZ) were prepared with chicken embryo fibroblast (CEF) cells (18,47). Virus preparations were purified by sucrose density gradient ultracentrifugation and were adjusted to 10 7 PFU/ml. P27 antigen generation in cells was measured by antigen-specific enzyme-linked immunosorbent assay (42).
Virus-specific IFN-␥ ELISPOT assays. ELISPOT assays were performed using the method developed by and following the direct instructions of Mothe and Watkins, Wisconsin University Primate Center (31,46). In brief, 96-well flatbottomed plates (U-CyTech-BV, Utrecht, Netherlands) were coated with antigamma interferon (IFN-␥) monoclonal antibody MD-1 (U-CyTech-BV). Freshly isolated peripheral blood mononuclear cells (PBMC) were added with either concanavalin A or pooled Gag peptides (AIDS Research and Reference Reagent Program, National Institutes of Health, Rockville, MD). The cells were then incubated in anti-IFN-␥-coated plates before lysing with ice-cold deionized water. After plates had been washed, rabbit anti-IFN-␥ polyclonal biotinylated detector antibody (1 g per well; U-CyTech-BV) was added. The plates were reacted with gold-labeled antibiotin immunoglobulin G solution by adding 30 l of the activator mix (U-CyTech-BV) to each of the wells and allowing them to develop for 15 min.
Wells were imaged and spot-forming cells (SFC) were counted using the KS ELISPOT compact system (Carl Zeiss, Germany) (31,46). An SFC was defined as a large black spot with a fuzzy border. To determine significance levels, we established a baseline for each peptide using the average and standard deviation of the number of SFC for each peptide. A threshold significance value corresponding to this average and two standard deviations were then determined. A response was considered positive if the number of SFC exceeded the threshold significance level of the sample with no added peptide.
Lymphocyte proliferative responses. SIV-specific proliferative responses were measured in freshly isolated PBMC as described by Gauduin et al. and Hel et al. (14,15). PBMC were cultured in flat-bottomed 96-well plates with either concanavalin A or purified SIVmac251 p27 protein (Advanced BioScience Laboratories, Rockville, MD) (15) for three days before the addition of [ 3 H]thymidine. Cells were harvested 16 h later to determine uptake.
Plasma viral RNA copy numbers. Plasma viral RNA copy numbers were measured using a real-time quantification assay based on the TaqMan system (Applied Biosystems, Foster City, CA) and the Prism 7700 sequence detection system (Applied Biosystems), as reported previously (30,46). Briefly, viral RNA was extracted and purified from macaque plasma samples using a QIAamp viral RNA mini kit (QIAGEN, Valencia, CA). The RNA was subjected to reversetranscription and amplification using a TaqMan EZ RT-PCR Kit (Applied Biosystems) with SIV Gag consensus primers SIVmac239-1224F and SIVmac239-1326R, and the SIV Gag consensus Taqman probe FAM-SIV-1272T. To obtain control RNA for quantification, SIVmac239 gag RNA was synthesized using T7 RNA polymerase and pKS460, a template plasmid that contains SIVmac239 gag under control of the T7 promoter.
To measure the RNA recovery rate, 10 5 copies of SHIV KS661c, in which the viral RNA copy number was previously determined by branched DNA assay (Bayer), were extracted and purified using the same kit as for the sample. Plasma viral load was calculated based on the standard curve of control RNA and the RNA recovery rate. All assays were carried out in duplicate.
Statistical analysis. Data analysis was carried out using the Stat View program (SAS Institute, Cary, NC) and data are expressed as the mean Ϯ standard deviation. A P value of Ͻ0.05 was considered significant.

Construction and preparation of recombinant M. bovis BCG Tokyo 172 and vaccinia virus DIs expressing whole SIV Gag.
In initial studies, we cloned DNA encoding SIVmac239 gag downstream of the hsp60 promoter and the expression unit was inserted into the KpnI restriction site of plasmid pSO246. We also constructed a recombinant M. bovis BCG vaccine based on the Tokyo 172 strain expressing the full-length gag gene of SIVmac239 (rBCG-SIVgag) (Fig. 1A). The presence of SIV Gag-specific DNA was confirmed in recombinant bacteria by DNA-PCR (42). To determine the in vitro expression of the SIV Gag protein in the cells, we analyzed cell extracts of rBCG-SIVgag bacteria after 2 weeks of culture by Western blot using anti-SIV Gag monoclonal antibody IB6. The rBCG clone produced an SIV Gag recombinant protein that strongly reacted as a single band with the specific monoclonal antibody (Fig. 1B). The concentration of SIV Gag p27 protein in transformed bacteria was 28.56 Ϯ 8.30 ng/10 8 CFU of bacilli. In contrast, neither SIV Gag protein nor gag DNA was detected in bacteria transformed with rBCG-pSO246, a control construct lacking the SIV gag insert and used as a vector control (Fig. 1B).
rDIsSIVgag and a control vaccinia virus, rDIsLacZ, were propagated in CEF and adjusted to 10 7 PFU/ml. Using Western blot, we confirmed the expression of each foreign gene in the cell extract, and purified virions used as immunogens in this study.
Immune induction after single-modality or combined immunization regimens with vaccine candidates. We examined whether rBCG expressing the full-length gag gene of SIVmac239 would be suitable for use in combined prime-boost protocols with the replication-deficient vaccinia virus strain DIs recombinant. The rBCG was intradermally delivered to the inner region of the thigh and rDIs was intravenously administered into the small saphenous vein on the back of the leg. Of the 15 macaques registered in this study, 13 were divided into five groups and immunized using either a single-modality regimen plus vector controls or with a priming-boosting regimen consisting of M. bovis BCG and vaccinia virus DIs recombinants ( Table 1). The remaining two macaques were inoculated with phosphate-buffered saline and served as naïve controls throughout the experiment.
Antigen-specific T-cell responses in all 15 animals were monitored by SIV Gag peptide-specific IFN-␥-ELISPOT assays (Fig. 2). Fifty weeks postinfection, the rBCG/rDIs group showed the highest SIV Gag-specific IFN-␥-ELISPOT responses; that group's responses peaked at 1,020 Ϯ 360 SFC/10 6 PBMC at 56 weeks postinfection or 2 weeks after the second booster inoculation ( Fig  responses in other rBCG and rDIs groups, both at 56 weeks postinfection and before mucosal challenge with pathogenic SHIV (P Ͻ 0.05). Furthermore, the number of SFC in the control and in the two naïve macaque groups did not exceed twenty during the 64-wk immunization period. Thus, the two booster inoculations of rDIs in rBCG-immunized animals effectively induced Gag peptide-specific IFN-␥-ELISPOT responses in peripheral blood, with the booster effect of DIs somewhat resembling that observed in our previous report on DNA/DIs prime-boost immunization in mice (47). We further studied the induction of SIV Gag-specific IFN-␥ ELISPOT by stimulating PBMC with SIV Gag p27 protein 56 weeks postinfection (Fig. 3A). The rBCG/rDIs group expressed whole-protein-specific IFN-␥ responses of 615 Ϯ 49 cells per million PBMC and the highest peptide-specific ELISPOT responses at 56 weeks postinfection (Fig. 2) of all five groups, with the peptide-specific responses being higher than the protein-specific responses ( Fig. 2 and 3A). Other groups exhibited fewer than 200 cells per million PBMC.
To characterize the cellular immune responses in the rBCG/ rDIs group, PBMC from the rBCG/rDIs-immunized macaques were compared with those of the rBCG and control groups by staining the surface for CD8 and intracellular SIV Gag-specific IFN-␥ expression (CD8 ϩ IFN-␥ ϩ cells) and then performing flow cytometric analysis (Fig. 3B). In vitro stimulation of PBMC with SIV Gag peptides in macaques 08, 10, and 46 of the rBCG/rDIs group generated a higher percentage of CD8 ϩ IFN-␥ ϩ T cells (1.34, 0.60, and 0.75%, respectively) than it did in animals of the rBCG group. Furthermore, non-CD8 ϩ T cells in PBMC from each animal of the rBCG/rDIs group expressed higher levels of SIV Gag-specific IFN-␥ activities (macaque 8: 0.42%; macaque 10: 0.29%; macaque 46: 0.55%) than did those of the other two animal groups. The vector control animals had fewer than 0.03% of both CD8 ϩ IFN-␥ ϩ and non-CD8 ϩ IFN-␥ ϩ double-positive cells in PBMC. These findings show that the rBCG/rDIs prime-boost immunization augmented numbers of both IFN-␥-specific intracellular stainingpositive cells and ELISPOT in the immunized animals, and that antigen-specific IFN-␥ activities were highly induced in CD8 ϩ as well as in non-CD8 ϩ T cells, the latter most likely being CD4 ϩ T cells.
Mucosal challenge study with virulent SHIV KS661c for vaccine efficacy. Ten weeks after the second booster immunization or 64 weeks postinfection, the macaques were chal- lenged by intrarectal inoculation with 2 ϫ 10 3 TCID 50 or 50 50% monkey infectious doses (MID 50 ) of SHIV KS661c, a molecular clone derived from an SHIV-89.6 variant. As shown in Fig. 4, only those macaques in the rBCG/rDIs group first primed with rBCG-SIVgag and then boosted with two inoculations of rDIsSIVgag showed evidence of protective immune responses (rBCG/rDIs). For two animals in this group (macaques 10 and 46), plasma viremia levels remained undetectable (Ͻ500 RNA copies/ml, shadow in left panel of rBCG/rDIs in Fig. 4) and CD4 ϩ T-cell counts stayed above 500 cells/l (shadow in right panel of rBCG/rDIs in Fig. 4) for the entire year of testing. The third animal in this group (macaque 08) had fluctuating levels of viremia that were still significantly lower than those of animals in the other immunization groups.
Coincidentally, this animal also had significantly decreased CD4 ϩ T-cell counts. All macaques in the rBCG/rDIs group remained clinically healthy during the one-year observation period. Those in the rDIs/rBCG group maintained antigen-specific immune responses (Fig. 6), but showed no protective immunity against viral challenge, except for macaque 36 who showed fluctuation in the number of CD4 ϩ T cells, with numbers dipping at times below 500 cells/l (rDIs/rBCG in Fig. 4). macaques in the other three groups all showed high levels of plasma viremia and a loss of CD4 ϩ T cells, suggesting that vaccination with rBCG and rDIs, either alone or as a priming agent, may not be suitable to induce effective, long-term positive immunity against mucosal challenge by virulent virus. By day 170 after FIG. 4. Plasma viral loads and CD4 ϩ T-cell counts after viral challenge. Postchallenge plasma viral RNA copies and absolute CD4 ϩ T-cell counts in peripheral blood were detected in macaques in each of five groups immunized with a consecutive prime-boost regimen consisting of rBCG-SIVgag and rDIsSIVgag. In the study, 13 macaques were divided into five groups following the experimental designs described in Table 1. VOL. 79, 2005 PRIME-BOOST RECOMBINANT BCG/DIs ELICITS IMMUNITY challenge, six of the 13 macaques had died with symptoms consistent with simian AIDS: four with interstitial pneumonia, one with neurological disturbances, and one with acquired hemorrhagic diathesis. Analysis of the cumulative survival rate using the Kaplan-Meier plot showed that the rBCG/rDIs group vaccinated with the priming-boosting regimen had a superior survival rate (P ϭ 0.012) to the other groups receiving vaccine protocols (P ϭ 0.548) and to the control group (Fig. 5). These findings demonstrate that a prime-boost immunization with rBCG-SIVgag/rDIsSIVgag controlled virulent immunodeficiency virus infection in macaques for at least 1 year and more significantly improved survival rates than did other vaccine protocols. Immune correlates of protection after viral challenge. In order to study virus-specific immune enhancement by SHIV challenge, we followed the postchallenge expansion of the virus-specific IFN-␥-positive cells in each animal by comparing the virus-specific IFN-␥-positive cell numbers pre-and postchallenge (Fig. 6). In all of the challenged animals of the rBCG/rDIs group, the mean number of IFN-␥-positive cells expanded from 369 Ϯ 73 at the time of viral challenge to 629 Ϯ 41 cells per 10 6 PBMC at 7 days after viral challenge, the sharpest increase noted with any of the animal groups. The animals of the rDIs/rBCG group showed much less enhancement, from a mean of 108 Ϯ 46 cells per 10 6 PBMC before challenge to 224 Ϯ 64 postchallenge, demonstrating that cellular immune responses are enhanced by viral challenge in the initial viral infection period in animals. Although in the rBCG/ rDIs group high levels of IFN-␥ production were observed in both CD8 ϩ and non-CD8 ϩ T cells in all three monkeys, macaque 10 and macaque 46 maintained undetectable setpoint levels of plasma viral load and normal numbers of CD4 ϩ lymphocytes, while macaque 08 did not. The macaques showed no clinical sign of weight loss, lymphoadenopathy, splenomegaly, anemia, or thrombocytopenia in the 1-year observation period. Furthermore, macaques in the rDIs group survived under low immune induction but exhibited CD4 ϩ T-cell loss and plasma viremia. Notably, all macaques in the control group exhibited very low levels of immune induction by viral challenge and showed no viral control.

DISCUSSION
In the current study, we initially produced rBCG expressing SIV whole Gag. Second, by introducing a priming-boosting regimen combining rBCG-SIVgag with a nonreplicating rDIsSIVgag, we found that the rBCG/rDIs vaccination induced a longlasting and effective immunity that was able to control a highly pathogenic virus after mucosal challenge in macaques. Third, elicitation of virus-specific immunity was observed to be important in exerting viral control in the animals immunized with the prime-boost vaccine regimen. Further investigation using larger groups of animals will be needed to determine whether high levels of immune induction correlate with increased efficacy. In this study, the macaques in the rBCG/rDIs group developed high levels of cellular immunity and were protected against the loss of CD4 ϩ T lymphocytes and the increase of viral RNA levels induced by viral challenge. Furthermore, the rBCG/rDIs group showed no evidence of clinical diseases or mortality after viral challenge during the 1-year period of observation.
The rBCG/rDIs prime-boost vaccine controlled the infection efficiently for the duration of the one-year observation period, reducing viral loads to below the threshold level for RNA copies in peripheral blood and maintaining the CD4 ϩ cell numbers above 500 cells per microliter of peripheral blood in two of the three animals in group 3. The remaining animal in the group showed fluctuations in the two parameters. Viral loads and CD4 ϩ cell numbers were not significantly affected in animal groups following the other vaccine regimens. The level of vaccine efficacy for the rBCG/rDIs group seems to be comparable to that observed in previous studies with DNA/fusion protein of interleukin-2 and immunoglobulin G (5), DNA/ MVA (3), DNA/recombinant adenovirus type 5 (Ad5) (45), and MVA/recombinant vesicular stomatitis virus (36); that is, effective control of pathogenic SHIV 89.6P infection was achieved in macaques for 6 to 8 months.
SHIV KS661c, which was used as a mucosal challenge virus in this study, is a highly pathogenic molecular clone of a variant of SHIV-89.6 possessing a tropism of CXCR-4. In our preliminary study, the SHIV virus infected GHOST-X4 cells in vitro and the virus challenge eliminated the naïve CD4 ϩ T-cell population in the peripheral blood in macaques, findings which confirmed those by Nishimura et al. (33,34). In conjunction with CCR5-tropic pathogenic SIVsmE660, Ourmanov et al. obtained similar results with the partial control of homologous viremia by the recombinant MVA vaccination (35). Furthermore, the potential of the DNA vaccination to induce a broad spectrum of mucosal protection against heterologous SIV/ DeltaB670 has been demonstrated (13).
Although the virus-specific immune elicitation by DNA/Ad5 vaccination was extremely high in immunized animals (45), the efficacy results for a DNA/Ad5 study with an SIVmac239 were not comparable to those for SHIV 89.6 (43). These discrepancies in vaccine efficacy by challenge viruses suggest that SIVmac239 might be a difficult virus to control by the active immunization of various vaccine candidates. Since DNA/Ad5 is expected to elicit higher levels of immunity than either MVA or DNA alone (43)(44)(45), it might be possible to obtain vaccine efficacy in conjunction with different CCR5-tropic SIV or SHIV from SIVmac239. Alternatively, a multicomponent DNA/Ad5 might elicit broad-spectrum immunity as well as protection against SIV or CCR5-SHIV. Recently, a DNA/ Sendai virus vaccination (27) proved to be as effective at controlling SIVmac239 as an attenuated live SIV vaccine (10,11), opening the possibility for studies comparing the protective immunity elicited by ordinary vaccination to that induced by attenuated live SIV vaccination.
Because the lack of an exact HIV-1 macaque model significantly limits our ability to study and calibrate vaccine efficacy, we may need to rely on parameters such as the control of viremia, the loss of CD4 ϩ cells, and the absence of mortality to establish the efficacy of a tested vaccine against an immunodeficiency virus. Certainly, such parameters would represent a more realistic goal for the development of a preventive vaccine in the macaque model. They may also play a key role in the evaluation of vaccine efficacy in human trials I/II using the vaccine modalities developed in the macaque model.
It was recently reported that the AIDS vaccine failed in rhesus macaques approximately six months post-virus challenge, with viral avoidance of cytotoxic T-lymphocyte recognition posing a major limitation to cytotoxic T-lymphocyte-based AIDS vaccines (4). In contrast, the rBCG/rDIs prime-boost vaccine was shown in this study to control viral load throughout the 1-year observation period, suggesting that it may improve the prospects for a vaccine regimen capable of providing longterm protection against HIV-1 replication and disease progression (38,49). Work is under way to determine whether this rBCG/rDIs vaccine will fail to control the plasma viral load in the macaque model, a failure associated with the viral escape of antigen-specific cytotoxic T lymphocytes.
The route of recombinant DIs administration will be key to effectively inducing immunity in humans. In the preliminary study to determine cellular immune induction, hundred times more rDIs was needed to achieve SIV Gag antigen-specific immunity in macaques by the intradermal (10 8 PFU/ml) than by the intravenous (10 6 PFU/ml) route (K. Someya et al., unpublished data). These findings may suggest that replication-defective vaccinia virus DIs is effective at eliciting antigenspecific immunity by intravenous administration. In addition, they suggest that the intravenous inoculation of rDIs may more effectively induce specific immunity than intradermal inoculation, although intravenous inoculation is not practical for use in human.
This study did not show a clear correlation between levels of virus-specific cellular immunity induced by booster inoculations with rDIs to rBCG-primed animals and protection against a highly virulent immunodeficiency virus after mucosal challenge. The levels of both virus-specific IFN-␥ ELISPOT and gamma interferon cytokine staining responses in peripheral blood from animals in the rBCG/rDIs group were the highest of the five groups studied. Why did the prime-boost vaccination of animals of the rBCG/rDIs group prove more effective than the vaccine protocols used with the other groups? We speculate that rBCG priming, which occurs at the skin region of the thigh near the inguinal and iliac lymph nodes VOL. 79, 2005 PRIME-BOOST RECOMBINANT BCG/DIs ELICITS IMMUNITY 12877 draining the genitorectal mucosa, may elicit mucosal immunity in the region (23). Furthermore, we showed that the two booster intravenous inoculations with rDIs help induce a level of protective immunity sufficient to control a mucosal viral challenge in the immunized animals. Although the two intravenous inoculations with rDIsSIVgag alone proved capable of inducing some virus-specific immunity in peripheral blood after the homologous booster immunization in the immunized animals (DIs group), they appeared to provide no protection against the mucosal viral challenge. The M. bovis BCG/DIs prime-boost vaccination might thus provide the opportunity to study the relationship between protection against mucosal viral challenge and elicitation of systemic or mucosal immunity. Our findings regarding the efficacy of the M. bovis BCG/DIs prime-boost vaccine regimen confirm those by Lehner et al. (23) and they further demonstrated a significant association between protection from mucosal rectal infection with SIV and an increase in the levels of CD8 suppressor factor and beta-chemokine. Although we cannot fully explain the differences in vaccine efficacy at this moment, it is likely that the routes of immunization and of challenge, the character of the vaccine vectors and the immunization schedule all play profound roles in eliciting vaccine efficacy in macaques.
Recently, considerable progress has been made in understanding M. bovis BCG as a HIV vaccine vector. Our own group demonstrated that recombinant M. bovis BCG vectors have the potential to deliver an HIV immunogen for desirable immune elicitation in macaques (46). Furthermore, M. bovis BCG vaccine substrain Tokyo 172 was revealed to be avirulent in HIV-infected children (16). The insertion of a full-length SIVmac239 gag into the M. bovis BCG substrain Tokyo 172 does not affect its toxicity, stability, or efficacy against Mycobacterium tuberculosis (54). Furthermore, rBCG has been shown to be nonvirulent in immunodeficient mice (54). These findings highlight the utility of rBCG as a vector for HIV-1 vaccine development.
In summary, our results demonstrate that a prime-boost vaccine regimen using rBCG as the prime and vaccinia virus rDIs as the boost can induce effective immunity against a mucosal infection with a highly virulent immunodeficiency virus for at least a year. Both of the vectors are safe for humans, making them attractive candidates for use in a preventive prime-boost vaccine against HIV-1.