A Single-Cycle Vaccine Vector Based on Vesicular S
http://www.100md.com
病菌学杂志 2005年第21期
Section of Microbial Pathogenesis
Department of Pathology, Yale University School of Medicine, New Haven, Connecticut 06510
ABSTRACT
Live attenuated vaccine vectors based on recombinant vesicular stomatitis virus (VSV) are effective in several viral disease models. In this study, we asked if a VSV vector capable of only a single cycle of replication might be an effective alternative to replication-competent VSV vectors. We compared the cellular immune responses to human immunodeficiency virus (HIV) envelope protein (Env) expressed by replication-competent and single-cycle VSV vectors and also examined the antibody response to Env. The single-cycle vector was grown by complementation with VSV G protein and then tested initially for immunogenicity when given by four different routes. When given by the intramuscular route in mice, we found that the single-cycle vector was equivalent to the replication-competent VSV vector in generating high-level primary and memory CD8 T-cell responses as well as antibody responses to Env. Cellular responses were analyzed using major histocompatibility complex class I tetramers and direct measurement of cytotoxic T-lymphocyte activity in vivo. We also found that the recall responses after boosting were equivalent in animals vaccinated with replication-competent or single-cycle vectors. Additionally, we observed recall and heightened memory responses after boosting animals with a single-cycle vector complemented with G protein from a different vesiculovirus. Because expression of HIV Env by G-deleted VSV might allow replication in human cells expressing CD4, we generated a single-cycle VSV recombinant expressing a secreted form of the HIV Env protein. This virus was just as effective as the recombinant expressing the membrane-anchored Env protein at producing CD8 T cells and antibody responses.
INTRODUCTION
Live attenuated vesicular stomatitis virus (VSV) vectors expressing a variety of foreign viral proteins are effective at producing strong immune responses and solid protection against viral challenge in many animal models (11, 18, 19, 21-24, 26). In a monkey AIDS model, VSV-based vaccines have prevented vaccinated monkeys from progressing to AIDS for at least 5 years (24). Because replication-competent vaccine vectors require extensive safety testing and approval for such vectors is very slow, we wanted to examine the immune responses generated by single-cycle VSV vectors.
The VSV glycoprotein (G) is the only viral protein present in the virus envelope. It is responsible for binding to the cellular receptor and for the pH-dependent fusion of the endosomal membrane after receptor-mediated endocytosis (6, 20). VSV particles lacking the G protein are unable to infect cells. However, VSVG viruses (with a deletion of the G gene) can be grown in cells expressing G transiently (28). These G-complemented viruses are able to infect cells and undergo one round of replication, but any virus particles produced by these cells lack G protein and cannot infect other cells.
Previously, our laboratory reported that a VSVG virus expressing influenza virus hemagglutinin (HA) protein was able to induce neutralizing antibody to influenza virus and protection from influenza virus challenge. This VSVG-HA caused no pathogenesis (weight loss) in mice postintranasal vaccination, but it was compromised for immunogenicity and induced approximately 60-fold less influenza virus neutralizing antibody than a replication-competent virus expressing HA after a single inoculation (21). Our previous studies showed that highly attenuated but replication-competent VSV vectors were also not very effective at inducing cellular immune responses when given intranasally but were equivalent to recombinant wild-type VSV (rVSV) vectors when given intraperitoneally (17) or intramuscularly (unpublished data). These studies led us to the understanding that a vector must be examined by more than one route of vaccination before its potential can be judged.
In the study described here, we initially examined the immune response to the single-cycle VSVG vector given via four routes. Because T-cell responses are especially important for protection from progression to AIDS, we focused initially on the ability of the single-cycle vector to generate CD8 T-cell responses to human immunodeficiency virus (HIV) Env. We found that the intramuscular route elicited one of the strongest Env-specific CD8 T-cell responses following vaccination with a single-cycle vector expressing HIV Env. We further demonstrate that a response equivalent to or greater than that generated by a replication-competent vector was generated after intramuscular vaccination with the single-cycle vector. Similar results were obtained when we examined antibody responses to HIV Env.
MATERIALS AND METHODS
Plasmids and viruses. The DNA sequence encoding HIV Env gp140 was amplified from pVSVEnvG (10) with the forward primer 5'-GGACGCG TCT CGAGA TTATGAGAGTG AAGGAG-3' and the reverse primer 5'-CGA TCCCCCC GGGCTAGCTC AACTTGCCCA TTTATCTAATTCC-3'. The amplified DNA product lacked the sequences encoding the transmembrane (TM) and cytoplasmic domains of HIV Env and had a stop codon terminating translation after the DNA sequence encoding amino acids DKWAS, just before the TM domain. The primers introduced the underlined XhoI and NheI restriction enzyme sites upstream and downstream of the coding sequence. The PCR product was digested with XhoI and NheI, purified, and ligated into a pVSVXN2 vector (27) that had been digested with the same enzymes. pVSVgp140 was recovered as previously described (17), and the insert sequence was verified (Yale Keck Sequencing Facility). pVSVGgp140 was constructed by removing the DNA sequence encoding VSV G from the plasmid pVSVgp140 by digestion with MluI and XhoI. The vector was then treated with T4 DNA polymerase to fill in both 3' ends and religated. Recovery of virus from this plasmid was as described previously (28). The recovered virus supernatants were filtered through a 0.2-μm-pore-size filter (Acrodisc) to remove residual vaccinia virus and passaged onto BHK-21 cells transfected with pCAGGS-G, a plasmid that expresses VSV G protein. Recombinant wild-type VSV, VSV-EnvG, and VSVG-EnvG have been described previously (1, 10, 13). VSVG-EnvG and VSVG-gp140 were grown on Vero cells transfected with pCAGGS (14) encoding either the VSV Indiana (I) serotype G protein (15) or the Chandipura G(Ch) protein. pCAGGS-G(Ch) was made by digesting pBS-G(Ch) (25) and pCAGGS-MGFP (K. Dalton and J. K. Rose, unpublished data) with XhoI and NheI. The G(Ch) insert and pCAGGS vector were ligated to generate the plasmid pCAGGS- G(Ch).
Complementation of G viruses was performed as follows. One 15-cm-diameter dish of confluent Vero cells was collected and electroporated with 50 μg pCAGGS-G at 130 V for 70 ms given in four pulses (Electro Square porator; BTX, San Diego, CA). Cells were replated on a 10-cm dish and placed at 37°C in a 5% CO2 incubator for 3 h. The electroporated cells were then heat shocked at 43°C in 5% CO2 for 3 h. At this time the medium was changed to remove cell debris and returned to 37°C. Twenty-four hours after transfection, cells were infected with VSVG-EnvG or VSVG-gp140 viruses. Supernatant was collected after 24 to 48 h upon observation of cytopathic effect in all cells. The supernatant was collected and passed through a 0.2-μm-pore-size filter. Supernatants were aliquoted and frozen at –80°C, and the titer for a frozen aliquot was determined on Vero cells transfected with pCAGGS-G. Approximately 5 x 106 cells were electroporated for each six-well titer plate.
Metabolic labeling and SDS-PAGE. To confirm expression of proteins of the appropriate size, cells were metabolically labeled as described previously (17). Briefly, BHK cells were infected with VSV recombinants for 4 h and then labeled for 1 hour with 100 μCi of [35S]methionine. Cell lysates were fractionated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE; 10% acrylamide), and proteins were visualized using a PhosphorImager (Molecular Dynamics).
Inoculation of mice. Eight-week-old female BALB/c mice were obtained from Charles River Laboratories and kept for at least 1 week before experiments were initiated. Mice were housed in microisolator cages in a biosafety level 2-equipped animal facility. Viral stocks were diluted to appropriate titers in serum-free Dulbecco's modified Eagle's medium (DMEM). For intraperitoneal and subcutaneous vaccinations, mice were injected with 6.25 x 105 PFU of virus in 200 μl. For intranasal vaccination, mice were lightly anesthetized with Metafane (methoxyflurane; Medical Developments Australia Pty. Ltd.) and administered 6.25 x 105 PFU virus in a 25-μl volume using a 200-μl pipette. For intramuscular vaccinations, mice were injected with 6.25 x 105 to 1.25 x 106 PFU in a 50-μl volume in the back hind leg muscle. The Institutional Animal Care and Use Committee of Yale University approved of all animal experiments done in this study.
Tetramer assay. The tetramer assay was performed on fresh splenocytes as previously described (17). Cells that were Env tetramer+, activated (CD62Llo), and CD8+ were identified on peak days using flow cytometry. Memory cells were identified as Env tetramer+, activated or resting (CD62Llo and CD62Lhi), and CD8+ cells. In parallel, rVSV-vaccinated animals were used to determine background levels of tetramer binding.
CTL assay in vivo. The cytotoxic T-lymphocyte (CTL) assay was performed as described previously (2) using Env peptide p18-I10 (N-RGPGRAFVTI-C; Research Genetics). On days 6 and 30 postvaccination, donor cells were prepared and transferred into recipient (vaccinated) mice as previously described (17). After 2 h, the recipient mice were euthanized and spleens were obtained and prepared as previously described (17). CFSEhi and CFSElo populations were identified by flow cytometry. Percent specific lysis was calculated by using the following formula: percent specific lysis = [1 – (ratio vaccinated/ratio control)] x 100, where "ratio" = (percent CFSElo/percent CFSEhi).
ELISA. Blood was obtained from mice at 56 days after vaccination. Serum was collected and heat inactivated as previously described (25). Preparation of the antigen, gp140, expressed by recombinant vaccinia virus vBD1, was previously described, along with the enzyme-linked immunosorbent assay (ELISA) procedure (25). Twenty μl of gp140 was used in each well, and the serum was diluted in phosphate-buffered saline (PBS) from 1:25 to 1:1,600. Plates were read in a Bio-Rad ELISA plate reader at an absorbance of 415 nm. Backgrounds obtained from serum of preimmune mice were subtracted.
Determination of VSV neutralizing antibody titers. Blood was obtained from mice at 28 days after vaccination. Serum was collected and heat inactivated as previously described (25). The serum was diluted in PBS so that the final dilution in the first well of a 96-well flat-bottom plate was 1:10, with subsequent twofold dilutions to 1:10,240. Samples were assayed in duplicate. One hundred PFU of rVSV diluted in serum-free DMEM was added to each well. The plates were incubated at 37°C-5%CO2 for 1 h, after which 4,000 BHK-21 cells diluted in 100 μl of 10% FBS-DMEM were added to each well. Plates were incubated at 37°C-5%CO2 for 2 days, and cytopathic effect was observed. The titer is reported as the highest dilution of serum that gave 100% virus neutralization.
RESULTS
Construction of single-cycle VSV vector expressing HIV EnvG. We constructed the single-cycle VSVG-EnvG virus by deleting the VSV G gene from the plasmid pVSV-EnvG (1) and performing a recovery in the presence of complementing VSV G protein (28). The EnvG gene encodes the extracellular and transmembrane domains of the HIV IIIB Env protein but has the VSV G cytoplasmic domain replacing the Env cytoplasmic domain. This modification of Env enhances its incorporation into VSV virions (10, 16). The virus obtained was then grown on Vero cells transfected with pCAGGS-G (15).
To confirm the sizes of the proteins expressed from the single-cycle vectors, we infected BHK cells with rVSV, VSV-EnvG, VSVG-EnvG, or VSVG-gp140 (which expresses an HIV Env lacking the transmembrane and cytoplasmic domains of HIV Env). Cells were labeled with [35S]methionine between 4 and 5 h postinfection, and lysates were analyzed by SDS-PAGE (Fig. 1B). The indicated five VSV structural proteins were synthesized in cells infected with rVSV or VSV-EnvG. VSV-EnvG also expressed the indicated (145-kDa) EnvG protein. VSVG-EnvG expressed the EnvG protein but did not express VSV G protein. VSVG-gp140 also lacked G expression and expressed a protein slightly smaller than EnvG.
Analysis of CD8 T-cell responses to the VSVG-EnvG vector given by four routes. To examine the immune responses to Env encoded by the single-cycle G vector and to determine if the route of infection affected the magnitude of the responses, we tested four different vaccination routes: intranasal (i.n.), intramuscular (i.m.), subcutaneous (s.c.), and intraperitoneal (i.p.). Initially we measured the immune response by looking at the HIV Env-specific, activated CD8 T cells using an major histocompatibility complex (MHC) I tetramer assay and fluorescence-activated cell sorter (FACS) analysis of splenocytes from vaccinated mice. For all four routes, the single-cycle vector elicited a primary immune response to Env by day 6 postvaccination, although the magnitude of the response depended on the route of vaccination. Mice vaccinated with VSVG-EnvG by the s.c or i.n. route gave modest primary CD8 T-cell responses (4.2% ± 1.9% and 0.82% ± 0.05% tetramer+, activated [CD62Llo] CD8 T cells, respectively). In contrast, mice vaccinated via the i.p. or i.m. route gave higher responses (14.4% ± 0.4% and 12.7% ± 0.3%, respectively) (Fig. 2A and C).
Populations of tetramer+ CD8 T cells (CD62Lhi + CD62Llo) recognizing Env were retained at 30 days after vaccination with VSVG-EnvG. We designated these as "memory" T cells because their numbers remain constant after 30 days (8). After s.c., i.n., or i.p. vaccination, the percentages of tetramer+, CD8 memory T cells were 1.2 ± 0.04, 1.1 ± 0.43, and 1.7 ± 0.24, respectively. Vaccination by the i.m. route yielded the largest population of Env-specific memory CD8 T cells at day 30 (3.2% ± 0.0%) (Fig. 2B and C). Because i.m. vaccination yielded an excellent primary response and the best memory response, our further studies focused only on this route of vaccination. This is also a convenient and practical route for vaccination.
Analysis of CD8 T-cell responses elicited by single-cycle versus replication-competent VSV vectors using MHC I tetramers and in vivo CTL assays. Live replication-competent VSV vectors expressing HIV Env elicit high levels of Env-specific CD8 T-cell responses and long-lasting Env-specific CD8 memory responses (7, 17). Because we observed a strong Env-specific immune response after immunization with the VSVG-EnvG virus, we next compared this vector directly with the replication-competent vector. We vaccinated mice intramuscularly with the same titer of each virus and looked at the primary (day 6 postvaccination) and the memory (day 30 postvaccination) response. We then measured Env-specific T cells by tetramer staining and by an in vivo cytotoxic T-cell assay.
The primary and memory responses to Env following vaccination with VSVG-EnvG were equal to or greater than those seen following vaccination with VSV-EnvG. The percentage of Env tetramer+, activated (CD62Llo) CD8 T cells on day 6 following i.m. vaccination with VSVG-EnvG or VSV-EnvG was about 13 to 14%. By day 30 postvaccination the populations of activated and resting (CD62Llo and CD62Lhi) Env-specific T cells had decreased to 2.2% ± 0.8% for mice vaccinated with VSV-EnvG and 3.8% ± 0.1% for mice vaccinated with VSVG-EnvG (Fig. 3; Table 1).
To determine if the antigen-specific T cells generated by the VSVG-EnvG virus were cytotoxic, we performed CTL assays in vivo using an assay described previously (2, 17). The assay measures cytotoxicity by quantitating loss of target cells loaded with Env peptide in vaccinated or control animals. Splenocytes from a na?ve mouse were used to create two populations of cells, one labeled with a high concentration of carboxyfluorescein diacetate succinimidyl ester (CFSE) and loaded with p18-I10 Env peptide and the other labeled with a low concentration of CFSE and lacking peptide. Vaccinated mice were then injected i.v. with equal numbers of these cells, and killing was measured by flow cytometry. At day 6 post-i.m. vaccination, mice that were vaccinated with VSVG-EnvG specifically killed 82% ± 1.5% of peptide-loaded splenocytes and mice vaccinated with VSV-EnvG specifically killed 73% ± 1.8% of peptide-loaded splenocytes (Fig. 3B and C).
We also compared the cytotoxic capabilities of the T cells present at day 30 post-VSV-EnvG and -VSVG-EnvG vaccination. Upon transfer of peptide-loaded and unloaded cells on day 30 postvaccination, mice that received VSVG-EnvG produced 12.9% ± 4.6% specific killing and mice that received VSV-EnvG produced 11.8% ± 2.4% specific killing (Fig. 3B and C). The primary and memory Env-specific CTL response elicited by the single-cycle VSVG-EnvG was clearly not attenuated compared to those generated by the replication-competent vector. In fact, they were usually higher.
Recall of CD8 T cells following a vaccinia virus-Env boost. After quantitating the primary and memory CD8 T-cell responses to the replicating and single-cycle vectors, we wanted to examine the recall of Env-specific CD8 T cells following a boost. Previously, we found that boosting VSV-Env-primed animals with a vaccinia virus recombinant expressing HIV Env protein (vPE16) (4) generated a strong recall response (8). We therefore boosted mice that had received either VSV-EnvG or VSVG-EnvG on day 0 with vPE16 on day 30 post-primary vaccination.
Six days after the vPE16 boost (day 36), the levels of tetramer+ CD8 T cells in both the replicating and single-cycle VSV-primed mice were boosted from about 2 to 3% to 24%. These responses clearly represented a boost, because na?ve mice that received vPE16 produced only about 6% Env-specific CD8 T cells at day 6 postvaccination (Fig. 4; Table 1).
The recall response elicited following vPE16 boosting decreased only slightly by day 60, to 16.7% ± 3.3% following VSV-EnvG prime and vPE16 boost and 19.5% ± 1.5% following VSVG-EnvG prime and vPE16 boost, leaving a very high-level memory population (Fig. 4; Table 1). The memory population following vPE16 alone was 3.0% ± 0.19%.
Boosting of responses with G vectors complemented with the same or different G proteins. Infection with VSV produces a strong neutralizing antibody (nAb) response to the G protein. This nAb response prevents boosting of mice primed by VSV with the same VSV vector (25). Because VSVG-EnvG does not express G, we wanted to determine if the complementing G in the particles of VSVG-EnvG was sufficient to induce a neutralizing antibody response to VSV and prevent boosting.
Initially, we determined the VSV neutralizing titer in serum from mice infected with VSV-EnvG and VSVG-EnvG. Twenty-eight days post-i.m. infection, mice infected with both viruses made a neutralizing antibody response to VSV. The titers generated to the G virus were nearly as high as those to the replication-competent virus (Fig. 5A). We next determined if any boosting of CTL responses could be obtained by i.m. vaccination in the presence of the neutralizing antibody. We primed mice on day zero with VSVG-EnvG and then boosted with the same virus on day 30. On day 36, 6 days postboosting of these mice, we observed a twofold increase of Env-specific T cells from day 30 (Fig. 5B; Table 1). However, this recall response was twofold lower than a primary response to VSVG-EnvG on day 6 postvaccination, indicating some effect of the exposure to the virus in the priming vaccination (Fig. 5B; Table 1). However, by day 60, 30 days postboost, the tetramer+ T-cell population had been maintained, indicating that the boost was effective at increasing the memory cell population.
To overcome the interference of nAb in boosting, our group has previously described the use of VSV glycoprotein exchange vectors which employ the same basic vector, but with a G protein from a different VSV serotype or from another vesiculovirus that is not cross-neutralized by antibody to the VSV G (Indiana) serotype (25). To determine if boosting by VSVG-EnvG could be increased using a different G envelope, we complemented the VSVG-EnvG with the G protein from a vesiculovirus called Chandipura.
On day 30 post-i.m. vaccination with VSVG-EnvG (G Indiana complemented), we boosted mice with VSVG-EnvG (complemented with Chandipura G protein). On day 36, 6 days postboost, we saw a strong recall response of Env tetramer+ cells. By day 60 the boost was more evident, with a significant population of memory cells that was over twofold higher than the memory response in these mice on 30 days postvaccination and over threefold higher than the memory response of the boosting virus alone (Fig. 3C; Table 1). While boosting with VSVG-EnvG complemented with Chandipura G produced a slightly better memory response postboost than VSVG-EnvG complemented with Indiana G, both responses were still two- to threefold lower than the memory response produced by a vaccinia boost. However, these experiments still introduce an exclusively nonreplicating viral boosting strategy that is effective.
Cellular immune response to a secreted Env protein expressed from a G virus. Because the VSVG-EnvG virus we are using in this study incorporates the EnvG protein into the viral membrane, it will bind to, fuse with, and replicate in human cells expressing CD4 and appropriate coreceptors (1, 10). A low level of HIV Env (lacking the G tail) is also incorporated into VSV particles. Because the use of viruses expressing membrane-anchored HIV Env proteins in humans would raise new safety issues, we wanted to determine if expression of a non-membrane-anchored Env protein from the G vector would be able to induce equivalent immune responses. This Env protein would not be incorporated into the VSV envelope and thus could not mediate infection of CD4 T cells. We therefore constructed a VSVG virus that expresses the extracellular domains of gp120 and gp41, but lacks the Env transmembrane and cytoplasmic domains (designated gp140). This VSVG-gp140 was recovered and grown in VSV G-complementing cells. As described above, we confirmed expression of the appropriate VSV proteins and the gp140 protein in cells infected with this virus (Fig. 1B).
To examine the CD8 T-cell response to this virus, mice were vaccinated by the i.m. route with either VSVG-EnvG or VSVG-gp140. On day 6 postvaccination, mice that had been vaccinated with VSVG-EnvG or VSVG-gp140 generated comparable responses of 15.1% ± 1.3% and 14.2% ± 1.2%, respectively (Fig. 6).
Antibody responses to EnvG or gp140 expressed from single-cycle vectors. HIV gp140 is a secreted protein lacking the transmembrane and cytoplasmic domain, while the HIV EnvG protein is transported to the plasma membrane and incorporated into VSV particles. Because of these differences in localization, we wanted to determine if there were any significant differences in the antibody responses to these proteins encoded by single-cycle vectors. We therefore infected mice with either VSVG-EnvG or VSVG-gp140 and measured the antibody response to HIVgp140 in the serum 8 weeks postvaccination. We also included a group infected with replication-competent VSV-EnvG to determine if replication had any effect on the induction of antibody. The ELISA titer for antibodies to Env protein was determined as described previously (25). The results in Fig. 7 show that the total antibody titer generated to HIV Env is very similar in the mice inoculated with VSV-EnvG or in VSVG-EnvG-infected mice and slightly greater in the VSV gp140-infected mice. These results are thus parallel to what we have seen for cellular immune responses after i.m. vaccination. Because HIV neutralizing antibodies are not generated in mice without the use of boosting vectors (25), we did not attempt to measure HIV neutralization with these sera.
DISCUSSION
Attenuated VSV recombinants expressing appropriate foreign antigens have been used as effective live viral vaccines in several animal disease models (5, 11, 18, 19, 21-24, 26). However, because these viruses are based on a live recombinant virus, they are subject to strict regulatory standards before testing in humans can begin. VSV recombinants that lack the G glycoprotein gene can be grown in a cell line expressing VSV G (28) and are then capable of only a single round of replication. Viral vectors capable of only a single cycle of replication are subject to much less stringent regulatory standards. For example, the single-cycle RNA virus vectors based on Venezuelan equine encephalitis virus, a neuroinvasive virus, were approved for human trials as an AIDS vaccine (3; http://www.iavireport.org/issues/0803/Vaxbriefs.asp) without any requirement for neurovirulence testing (R. Johnston, personal communication).
In this study we have shown that single-cycle VSV vectors elicit potent CTL and antibody responses to the foreign protein HIV Env encoded in the vector. When given intramuscularly, these immune responses are equivalent to the responses generated by the replication-competent VSV vector expressing HIV Env.
It is interesting that this single-cycle vector generates a much lower cellular immune response than the replication-competent vector when given intranasally. This result is consistent with results obtained using a highly attenuated VSV vector (VSVCT1-EnvG) generated by truncation of the VSV G cytoplasmic domain (17). That vector induced weak immune responses to Env compared to those generated by the recombinant wild-type VSV vector when given intranasally but generated equivalent immune responses when given by the i.p. (17) or i.m. (unpublished observations) route. We suggested previously that systemic spread of rVSV after intranasal inoculation might be required to generate a strong immune response and that the poorly replicating vectors might not be capable of spreading from the lungs into the blood and other organs to induce potent immune responses. Earlier studies have shown that rVSV given intranasally replicates in the lungs, causes a transient viremia, and spreads to other organs (21).
Taken together with the results presented here, it is clear that only a single cycle of replication is required to generate potent immune responses when VSV vectors are given by the intramuscular route. It may be that rVSV undergoes only a single round of replication after intramuscular inoculation because of rapid immune responses or other restrictions on replication in muscle. This rapid control of replication is consistent with the fact that viremia and pathogenesis (weight loss) in mice have not been observed following i.m. vaccination (unpublished results).
We have noted a consistent trend (although not statistically significant) in our data (Fig. 3; Table 1) that the cellular immune responses to HIV Env generated by the single-cycle VSVG-EnvG vector were typically better than those induced by the replicating VSV-EnvG vector. Why would a single-cycle vector be more potent than a replication-competent vector? If we first accept that a single cycle of replication after intramuscular injection is sufficient to induce an optimal immune response, then the slightly better responses to HIV Env may be due simply to the absence of the upstream G gene in the VSVG-EnvG virus. Each upstream gene results in a attenuation of expression the downstream gene (9, 12). Therefore, deletion of the G gene would be expected to increase expression of EnvG relative to the remaining VSV genes during a single round of replication. In fact, EnvG expression was increased more than twofold relative to the upstream N gene in the G vector compared to the replication-competent vector. Furthermore, the lack of G expression would be expected to eliminate some competition for MHC class I presentation and could result in upregulation of CD8 T-cell responses to Env.
In previous studies, where a VSVG vector expressing an influenza virus HA protein was used as an influenza vaccine, it produced much lower neutralizing antibody responses than the replication-competent vector but was able to provide protection from influenza virus challenge in animals given a booster inoculation (21). The poor response in these studies can now be explained by the choice of the intranasal, rather than the intramuscular, route for vaccination. Our studies suggest that repetition of the influenza virus studies using G vectors given by the intramuscular route would likely yield more robust anti-HA responses.
Previous studies from our laboratory have shown that VSV vectors expressing SHIV proteins protect macaques from progression to AIDS following a SHIV challenge (18, 24). Since the rVSV vector in the macaques studies was identical to the replication-competent VSV vector used in this study, our findings may suggest that a VSV vector that undergoes only a single cycle of replication may be equally effective in macaques as the rVSV at producing Env-specific CD8 T cells when given by the appropriate route. Even though the attenuated rVSV has not caused any measurable disease symptoms in more than 100 vaccinated macaques, rare vaccine-associated disease cannot be ruled out. A single-cycle vector is a potent alternative that would be safe for use even in people with severe immunodeficiency.
A recent study has shown that rVSV vectors expressing SHIV Env and Gag proteins are more effective at generating cellular immune responses in macaques when given by the i.n. route compared to the i.m. route (5). We have not observed this same dichotomy in mice with replicating vectors. In fact, the intranasal vaccination route generally yields somewhat lower cellular immune responses than the i.m. or i.p. routes (7, 17). We have not yet used our single-cycle vectors in macaques, but our results in mice indicate that comparison of the magnitude of the immune responses generated by different routes of vaccination should be undertaken.
ACKNOWLEDGMENTS
We thank members of the Rose laboratory for helpful discussions and suggestions on the manuscript.
This work was supported by NIH grants R37-AI40357 and RO1-AI45510 and by an HIV-1 Vaccine Design and Development Team contract (NIH NO1-AI-25458).
REFERENCES
Boritz, E., J. Gerlach, J. E. Johnson, and J. K. Rose. 1999. Replication-competent rhabdoviruses with human immunodeficiency virus type 1 coats and green fluorescent protein: entry by a pH-independent pathway. J. Virol. 73:6937-6945.
Coles, R. M., S. N. Mueller, W. R. Heath, F. R. Carbone, and A. G. Brooks. 2002. Progression of armed CTL from draining lymph node to spleen shortly after localized infection with herpes simplex virus 1. J. Immunol. 168: 834-838.
Davis, N. L., A. West, E. Reap, G. MacDonald, M. Collier, S. Dryga, M. Maughan, M. Connell, C. Walker, K. McGrath, C. Cecil, L. H. Ping, J. Frelinger, R. Olmsted, P. Keith, R. Swanstrom, C. Williamson, P. Johnson, D. Montefiori, and R. E. Johnston. 2002. Alphavirus replicon particles as candidate HIV vaccines. IUBMB Life 53:209-211.
Earl, P. L., S. Koenig, and B. Moss. 1991. Biological and immunological properties of human immunodeficiency virus type 1 envelope glycoprotein: analysis of proteins with truncations and deletions expressed by recombinant vaccinia viruses. J. Virol. 65:31-41.
Egan, M. A., S. Y. Chong, N. F. Rose, S. Megati, K. J. Lopez, E. B. Schadeck, J. E. Johnson, A. Masood, P. Piacente, R. E. Druilhet, P. W. Barras, D. L. Hasselschwert, P. Reilly, E. M. Mishkin, D. C. Montefiori, M. G. Lewis, D. K. Clarke, R. M. Hendry, P. A. Marx, J. H. Eldridge, S. A. Udem, Z. R. Israel, and J. K. Rose. 2004. Immunogenicity of attenuated vesicular stomatitis virus vectors expressing HIV type 1 Env and SIV Gag proteins: comparison of intranasal and intramuscular vaccination routes. AIDS Res. Hum. Retrovir. 20:989-1004.
Florkiewicz, R. Z., and J. K. Rose. 1984. A cell line expressing vesicular stomatitis virus glycoprotein fuses at low pH. Science 225:721-723.
Haglund, K., I. Leiner, K. Kerksiek, L. Buonocore, E. Pamer, and J. K. Rose. 2002. High-level primary CD8+ T-cell response to human immunodeficiency virus type 1 gag and env generated by vaccination with recombinant vesicular stomatitis viruses. J. Virol. 76:2730-2738.
Haglund, K., I. Leiner, K. Kerksiek, L. Buonocore, E. Pamer, and J. K. Rose. 2002. Robust recall and long-term memory T-cell responses induced by prime-boost regimens with heterologous live viral vectors expressing human immunodeficiency virus type 1 Gag and Env proteins. J. Virol. 76:7506-7517.
Iverson, L. E., and J. K. Rose. 1981. Localized attenuation and discontinuous synthesis during vesicular stomatitis virus transcription. Cell 23:477-484.
Johnson, J. E., M. J. Schnell, L. Buonocore, and J. K. Rose. 1997. Specific targeting to CD4+ cells of recombinant vesicular stomatitis viruses encoding human immunodeficiency virus envelope proteins. J. Virol. 71:5060-5068.
Kahn, J. S., A. Roberts, C. Weibel, L. Buonocore, and J. K. Rose. 2001. Replication-competent or attenuated, nonpropagating vesicular stomatitis viruses expressing respiratory syncytial virus (RSV) antigens protect mice against RSV challenge. J. Virol. 75:11079-11087.
Kretzschmar, E., L. Buonocore, M. J. Schnell, and J. K. Rose. 1997. High-efficiency incorporation of functional influenza virus glycoproteins into recombinant vesicular stomatitis viruses. J. Virol. 71:5982-5989.
Lawson, N. D., E. A. Stillman, M. A. Whitt, and J. K. Rose. 1995. Recombinant vesicular stomatitis viruses from DNA. Proc. Natl. Acad. Sci. USA 92:4477-4481.
Niwa, H., K. Yamamura, and J. Miyazaki. 1991. Efficient selection for high-expression transfectants with a novel eukaryotic vector. Gene 108:193-199.
Okuma, K., Y. Matsuura, H. Tatsuo, Y. Inagaki, M. Nakamura, N. Yamamoto, and Y. Yanagi. 2001. Analysis of the molecules involved in human T-cell leukaemia virus type 1 entry by a vesicular stomatitis virus pseudotype bearing its envelope glycoproteins. J. Gen. Virol. 82:821-830.
Owens, R. J., and J. K. Rose. 1993. Cytoplasmic domain requirement for incorporation of a foreign envelope protein into vesicular stomatitis virus. J. Virol. 67:360-365.
Publicover, J., E. Ramsburg, and J. K. Rose. 2004. Characterization of nonpathogenic, live, viral vaccine vectors inducing potent cellular immune responses. J. Virol. 78:9317-9324.
Ramsburg, E., N. F. Rose, P. A. Marx, M. Mefford, D. F. Nixon, W. J. Moretto, D. Montefiori, P. Earl, B. Moss, and J. K. Rose. 2004. Highly effective control of an AIDS virus challenge in macaques by using vesicular stomatitis virus and modified vaccinia virus Ankara vaccine vectors in a single-boost protocol. J. Virol. 78:3930-3940.
Reuter, J. D., B. E. Vivas-Gonzalez, D. Gomez, J. H. Wilson, J. L. Brandsma, H. L. Greenstone, J. K. Rose, and A. Roberts. 2002. Intranasal vaccination with a recombinant vesicular stomatitis virus expressing cottontail rabbit papillomavirus L1 protein provides complete protection against papillomavirus-induced disease. J. Virol. 76:8900-8909.
Riedel, H., C. Kondor-Koch, and H. Garoff. 1984. Cell surface expression of fusogenic vesicular stomatitis virus G protein from cloned cDNA. EMBO J. 3:1477-1483.
Roberts, A., L. Buonocore, R. Price, J. Forman, and J. K. Rose. 1999. Attenuated vesicular stomatitis viruses as vaccine vectors. J. Virol. 73: 3723-3732.
Roberts, A., E. Kretzschmar, A. S. Perkins, J. Forman, R. Price, L. Buonocore, Y. Kawaoka, and J. K. Rose. 1998. Vaccination with a recombinant vesicular stomatitis virus expressing an influenza virus hemagglutinin provides complete protection from influenza virus challenge. J. Virol. 72:4704-4711.
Roberts, A., J. D. Reuter, J. H. Wilson, S. Baldwin, and J. K. Rose. 2004. Complete protection from papillomavirus challenge after a single vaccination with a vesicular stomatitis virus vector expressing high levels of L1 protein. J. Virol. 78:3196-3199.
Rose, N. F., P. A. Marx, A. Luckay, D. F. Nixon, W. J. Moretto, S. M. Donahoe, D. Montefiori, A. Roberts, L. Buonocore, and J. K. Rose. 2001. An effective AIDS vaccine based on live attenuated vesicular stomatitis virus recombinants. Cell 106:539-549.
Rose, N. F., A. Roberts, L. Buonocore, and J. K. Rose. 2000. Glycoprotein exchange vectors based on vesicular stomatitis virus allow effective boosting and generation of neutralizing antibodies to a primary isolate of human immunodeficiency virus type 1. J. Virol. 74:10903-10910.
Schlereth, B., J. K. Rose, L. Buonocore, V. ter Meulen, and S. Niewiesk. 2000. Successful vaccine-induced seroconversion by single-dose immunization in the presence of measles virus-specific maternal antibodies. J. Virol. 74:4652-4657.
Schnell, M. J., L. Buonocore, M. A. Whitt, and J. K. Rose. 1996. The minimal conserved transcription stop-start signal promotes stable expression of a foreign gene in vesicular stomatitis virus. J. Virol. 70:2318-2323.
Schnell, M. J., J. E. Johnson, L. Buonocore, and J. K. Rose. 1997. Construction of a novel virus that targets HIV-1-infected cells and controls HIV-1 infection. Cell 90:849-857.(Jean Publicover, Elizabet)
Department of Pathology, Yale University School of Medicine, New Haven, Connecticut 06510
ABSTRACT
Live attenuated vaccine vectors based on recombinant vesicular stomatitis virus (VSV) are effective in several viral disease models. In this study, we asked if a VSV vector capable of only a single cycle of replication might be an effective alternative to replication-competent VSV vectors. We compared the cellular immune responses to human immunodeficiency virus (HIV) envelope protein (Env) expressed by replication-competent and single-cycle VSV vectors and also examined the antibody response to Env. The single-cycle vector was grown by complementation with VSV G protein and then tested initially for immunogenicity when given by four different routes. When given by the intramuscular route in mice, we found that the single-cycle vector was equivalent to the replication-competent VSV vector in generating high-level primary and memory CD8 T-cell responses as well as antibody responses to Env. Cellular responses were analyzed using major histocompatibility complex class I tetramers and direct measurement of cytotoxic T-lymphocyte activity in vivo. We also found that the recall responses after boosting were equivalent in animals vaccinated with replication-competent or single-cycle vectors. Additionally, we observed recall and heightened memory responses after boosting animals with a single-cycle vector complemented with G protein from a different vesiculovirus. Because expression of HIV Env by G-deleted VSV might allow replication in human cells expressing CD4, we generated a single-cycle VSV recombinant expressing a secreted form of the HIV Env protein. This virus was just as effective as the recombinant expressing the membrane-anchored Env protein at producing CD8 T cells and antibody responses.
INTRODUCTION
Live attenuated vesicular stomatitis virus (VSV) vectors expressing a variety of foreign viral proteins are effective at producing strong immune responses and solid protection against viral challenge in many animal models (11, 18, 19, 21-24, 26). In a monkey AIDS model, VSV-based vaccines have prevented vaccinated monkeys from progressing to AIDS for at least 5 years (24). Because replication-competent vaccine vectors require extensive safety testing and approval for such vectors is very slow, we wanted to examine the immune responses generated by single-cycle VSV vectors.
The VSV glycoprotein (G) is the only viral protein present in the virus envelope. It is responsible for binding to the cellular receptor and for the pH-dependent fusion of the endosomal membrane after receptor-mediated endocytosis (6, 20). VSV particles lacking the G protein are unable to infect cells. However, VSVG viruses (with a deletion of the G gene) can be grown in cells expressing G transiently (28). These G-complemented viruses are able to infect cells and undergo one round of replication, but any virus particles produced by these cells lack G protein and cannot infect other cells.
Previously, our laboratory reported that a VSVG virus expressing influenza virus hemagglutinin (HA) protein was able to induce neutralizing antibody to influenza virus and protection from influenza virus challenge. This VSVG-HA caused no pathogenesis (weight loss) in mice postintranasal vaccination, but it was compromised for immunogenicity and induced approximately 60-fold less influenza virus neutralizing antibody than a replication-competent virus expressing HA after a single inoculation (21). Our previous studies showed that highly attenuated but replication-competent VSV vectors were also not very effective at inducing cellular immune responses when given intranasally but were equivalent to recombinant wild-type VSV (rVSV) vectors when given intraperitoneally (17) or intramuscularly (unpublished data). These studies led us to the understanding that a vector must be examined by more than one route of vaccination before its potential can be judged.
In the study described here, we initially examined the immune response to the single-cycle VSVG vector given via four routes. Because T-cell responses are especially important for protection from progression to AIDS, we focused initially on the ability of the single-cycle vector to generate CD8 T-cell responses to human immunodeficiency virus (HIV) Env. We found that the intramuscular route elicited one of the strongest Env-specific CD8 T-cell responses following vaccination with a single-cycle vector expressing HIV Env. We further demonstrate that a response equivalent to or greater than that generated by a replication-competent vector was generated after intramuscular vaccination with the single-cycle vector. Similar results were obtained when we examined antibody responses to HIV Env.
MATERIALS AND METHODS
Plasmids and viruses. The DNA sequence encoding HIV Env gp140 was amplified from pVSVEnvG (10) with the forward primer 5'-GGACGCG TCT CGAGA TTATGAGAGTG AAGGAG-3' and the reverse primer 5'-CGA TCCCCCC GGGCTAGCTC AACTTGCCCA TTTATCTAATTCC-3'. The amplified DNA product lacked the sequences encoding the transmembrane (TM) and cytoplasmic domains of HIV Env and had a stop codon terminating translation after the DNA sequence encoding amino acids DKWAS, just before the TM domain. The primers introduced the underlined XhoI and NheI restriction enzyme sites upstream and downstream of the coding sequence. The PCR product was digested with XhoI and NheI, purified, and ligated into a pVSVXN2 vector (27) that had been digested with the same enzymes. pVSVgp140 was recovered as previously described (17), and the insert sequence was verified (Yale Keck Sequencing Facility). pVSVGgp140 was constructed by removing the DNA sequence encoding VSV G from the plasmid pVSVgp140 by digestion with MluI and XhoI. The vector was then treated with T4 DNA polymerase to fill in both 3' ends and religated. Recovery of virus from this plasmid was as described previously (28). The recovered virus supernatants were filtered through a 0.2-μm-pore-size filter (Acrodisc) to remove residual vaccinia virus and passaged onto BHK-21 cells transfected with pCAGGS-G, a plasmid that expresses VSV G protein. Recombinant wild-type VSV, VSV-EnvG, and VSVG-EnvG have been described previously (1, 10, 13). VSVG-EnvG and VSVG-gp140 were grown on Vero cells transfected with pCAGGS (14) encoding either the VSV Indiana (I) serotype G protein (15) or the Chandipura G(Ch) protein. pCAGGS-G(Ch) was made by digesting pBS-G(Ch) (25) and pCAGGS-MGFP (K. Dalton and J. K. Rose, unpublished data) with XhoI and NheI. The G(Ch) insert and pCAGGS vector were ligated to generate the plasmid pCAGGS- G(Ch).
Complementation of G viruses was performed as follows. One 15-cm-diameter dish of confluent Vero cells was collected and electroporated with 50 μg pCAGGS-G at 130 V for 70 ms given in four pulses (Electro Square porator; BTX, San Diego, CA). Cells were replated on a 10-cm dish and placed at 37°C in a 5% CO2 incubator for 3 h. The electroporated cells were then heat shocked at 43°C in 5% CO2 for 3 h. At this time the medium was changed to remove cell debris and returned to 37°C. Twenty-four hours after transfection, cells were infected with VSVG-EnvG or VSVG-gp140 viruses. Supernatant was collected after 24 to 48 h upon observation of cytopathic effect in all cells. The supernatant was collected and passed through a 0.2-μm-pore-size filter. Supernatants were aliquoted and frozen at –80°C, and the titer for a frozen aliquot was determined on Vero cells transfected with pCAGGS-G. Approximately 5 x 106 cells were electroporated for each six-well titer plate.
Metabolic labeling and SDS-PAGE. To confirm expression of proteins of the appropriate size, cells were metabolically labeled as described previously (17). Briefly, BHK cells were infected with VSV recombinants for 4 h and then labeled for 1 hour with 100 μCi of [35S]methionine. Cell lysates were fractionated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE; 10% acrylamide), and proteins were visualized using a PhosphorImager (Molecular Dynamics).
Inoculation of mice. Eight-week-old female BALB/c mice were obtained from Charles River Laboratories and kept for at least 1 week before experiments were initiated. Mice were housed in microisolator cages in a biosafety level 2-equipped animal facility. Viral stocks were diluted to appropriate titers in serum-free Dulbecco's modified Eagle's medium (DMEM). For intraperitoneal and subcutaneous vaccinations, mice were injected with 6.25 x 105 PFU of virus in 200 μl. For intranasal vaccination, mice were lightly anesthetized with Metafane (methoxyflurane; Medical Developments Australia Pty. Ltd.) and administered 6.25 x 105 PFU virus in a 25-μl volume using a 200-μl pipette. For intramuscular vaccinations, mice were injected with 6.25 x 105 to 1.25 x 106 PFU in a 50-μl volume in the back hind leg muscle. The Institutional Animal Care and Use Committee of Yale University approved of all animal experiments done in this study.
Tetramer assay. The tetramer assay was performed on fresh splenocytes as previously described (17). Cells that were Env tetramer+, activated (CD62Llo), and CD8+ were identified on peak days using flow cytometry. Memory cells were identified as Env tetramer+, activated or resting (CD62Llo and CD62Lhi), and CD8+ cells. In parallel, rVSV-vaccinated animals were used to determine background levels of tetramer binding.
CTL assay in vivo. The cytotoxic T-lymphocyte (CTL) assay was performed as described previously (2) using Env peptide p18-I10 (N-RGPGRAFVTI-C; Research Genetics). On days 6 and 30 postvaccination, donor cells were prepared and transferred into recipient (vaccinated) mice as previously described (17). After 2 h, the recipient mice were euthanized and spleens were obtained and prepared as previously described (17). CFSEhi and CFSElo populations were identified by flow cytometry. Percent specific lysis was calculated by using the following formula: percent specific lysis = [1 – (ratio vaccinated/ratio control)] x 100, where "ratio" = (percent CFSElo/percent CFSEhi).
ELISA. Blood was obtained from mice at 56 days after vaccination. Serum was collected and heat inactivated as previously described (25). Preparation of the antigen, gp140, expressed by recombinant vaccinia virus vBD1, was previously described, along with the enzyme-linked immunosorbent assay (ELISA) procedure (25). Twenty μl of gp140 was used in each well, and the serum was diluted in phosphate-buffered saline (PBS) from 1:25 to 1:1,600. Plates were read in a Bio-Rad ELISA plate reader at an absorbance of 415 nm. Backgrounds obtained from serum of preimmune mice were subtracted.
Determination of VSV neutralizing antibody titers. Blood was obtained from mice at 28 days after vaccination. Serum was collected and heat inactivated as previously described (25). The serum was diluted in PBS so that the final dilution in the first well of a 96-well flat-bottom plate was 1:10, with subsequent twofold dilutions to 1:10,240. Samples were assayed in duplicate. One hundred PFU of rVSV diluted in serum-free DMEM was added to each well. The plates were incubated at 37°C-5%CO2 for 1 h, after which 4,000 BHK-21 cells diluted in 100 μl of 10% FBS-DMEM were added to each well. Plates were incubated at 37°C-5%CO2 for 2 days, and cytopathic effect was observed. The titer is reported as the highest dilution of serum that gave 100% virus neutralization.
RESULTS
Construction of single-cycle VSV vector expressing HIV EnvG. We constructed the single-cycle VSVG-EnvG virus by deleting the VSV G gene from the plasmid pVSV-EnvG (1) and performing a recovery in the presence of complementing VSV G protein (28). The EnvG gene encodes the extracellular and transmembrane domains of the HIV IIIB Env protein but has the VSV G cytoplasmic domain replacing the Env cytoplasmic domain. This modification of Env enhances its incorporation into VSV virions (10, 16). The virus obtained was then grown on Vero cells transfected with pCAGGS-G (15).
To confirm the sizes of the proteins expressed from the single-cycle vectors, we infected BHK cells with rVSV, VSV-EnvG, VSVG-EnvG, or VSVG-gp140 (which expresses an HIV Env lacking the transmembrane and cytoplasmic domains of HIV Env). Cells were labeled with [35S]methionine between 4 and 5 h postinfection, and lysates were analyzed by SDS-PAGE (Fig. 1B). The indicated five VSV structural proteins were synthesized in cells infected with rVSV or VSV-EnvG. VSV-EnvG also expressed the indicated (145-kDa) EnvG protein. VSVG-EnvG expressed the EnvG protein but did not express VSV G protein. VSVG-gp140 also lacked G expression and expressed a protein slightly smaller than EnvG.
Analysis of CD8 T-cell responses to the VSVG-EnvG vector given by four routes. To examine the immune responses to Env encoded by the single-cycle G vector and to determine if the route of infection affected the magnitude of the responses, we tested four different vaccination routes: intranasal (i.n.), intramuscular (i.m.), subcutaneous (s.c.), and intraperitoneal (i.p.). Initially we measured the immune response by looking at the HIV Env-specific, activated CD8 T cells using an major histocompatibility complex (MHC) I tetramer assay and fluorescence-activated cell sorter (FACS) analysis of splenocytes from vaccinated mice. For all four routes, the single-cycle vector elicited a primary immune response to Env by day 6 postvaccination, although the magnitude of the response depended on the route of vaccination. Mice vaccinated with VSVG-EnvG by the s.c or i.n. route gave modest primary CD8 T-cell responses (4.2% ± 1.9% and 0.82% ± 0.05% tetramer+, activated [CD62Llo] CD8 T cells, respectively). In contrast, mice vaccinated via the i.p. or i.m. route gave higher responses (14.4% ± 0.4% and 12.7% ± 0.3%, respectively) (Fig. 2A and C).
Populations of tetramer+ CD8 T cells (CD62Lhi + CD62Llo) recognizing Env were retained at 30 days after vaccination with VSVG-EnvG. We designated these as "memory" T cells because their numbers remain constant after 30 days (8). After s.c., i.n., or i.p. vaccination, the percentages of tetramer+, CD8 memory T cells were 1.2 ± 0.04, 1.1 ± 0.43, and 1.7 ± 0.24, respectively. Vaccination by the i.m. route yielded the largest population of Env-specific memory CD8 T cells at day 30 (3.2% ± 0.0%) (Fig. 2B and C). Because i.m. vaccination yielded an excellent primary response and the best memory response, our further studies focused only on this route of vaccination. This is also a convenient and practical route for vaccination.
Analysis of CD8 T-cell responses elicited by single-cycle versus replication-competent VSV vectors using MHC I tetramers and in vivo CTL assays. Live replication-competent VSV vectors expressing HIV Env elicit high levels of Env-specific CD8 T-cell responses and long-lasting Env-specific CD8 memory responses (7, 17). Because we observed a strong Env-specific immune response after immunization with the VSVG-EnvG virus, we next compared this vector directly with the replication-competent vector. We vaccinated mice intramuscularly with the same titer of each virus and looked at the primary (day 6 postvaccination) and the memory (day 30 postvaccination) response. We then measured Env-specific T cells by tetramer staining and by an in vivo cytotoxic T-cell assay.
The primary and memory responses to Env following vaccination with VSVG-EnvG were equal to or greater than those seen following vaccination with VSV-EnvG. The percentage of Env tetramer+, activated (CD62Llo) CD8 T cells on day 6 following i.m. vaccination with VSVG-EnvG or VSV-EnvG was about 13 to 14%. By day 30 postvaccination the populations of activated and resting (CD62Llo and CD62Lhi) Env-specific T cells had decreased to 2.2% ± 0.8% for mice vaccinated with VSV-EnvG and 3.8% ± 0.1% for mice vaccinated with VSVG-EnvG (Fig. 3; Table 1).
To determine if the antigen-specific T cells generated by the VSVG-EnvG virus were cytotoxic, we performed CTL assays in vivo using an assay described previously (2, 17). The assay measures cytotoxicity by quantitating loss of target cells loaded with Env peptide in vaccinated or control animals. Splenocytes from a na?ve mouse were used to create two populations of cells, one labeled with a high concentration of carboxyfluorescein diacetate succinimidyl ester (CFSE) and loaded with p18-I10 Env peptide and the other labeled with a low concentration of CFSE and lacking peptide. Vaccinated mice were then injected i.v. with equal numbers of these cells, and killing was measured by flow cytometry. At day 6 post-i.m. vaccination, mice that were vaccinated with VSVG-EnvG specifically killed 82% ± 1.5% of peptide-loaded splenocytes and mice vaccinated with VSV-EnvG specifically killed 73% ± 1.8% of peptide-loaded splenocytes (Fig. 3B and C).
We also compared the cytotoxic capabilities of the T cells present at day 30 post-VSV-EnvG and -VSVG-EnvG vaccination. Upon transfer of peptide-loaded and unloaded cells on day 30 postvaccination, mice that received VSVG-EnvG produced 12.9% ± 4.6% specific killing and mice that received VSV-EnvG produced 11.8% ± 2.4% specific killing (Fig. 3B and C). The primary and memory Env-specific CTL response elicited by the single-cycle VSVG-EnvG was clearly not attenuated compared to those generated by the replication-competent vector. In fact, they were usually higher.
Recall of CD8 T cells following a vaccinia virus-Env boost. After quantitating the primary and memory CD8 T-cell responses to the replicating and single-cycle vectors, we wanted to examine the recall of Env-specific CD8 T cells following a boost. Previously, we found that boosting VSV-Env-primed animals with a vaccinia virus recombinant expressing HIV Env protein (vPE16) (4) generated a strong recall response (8). We therefore boosted mice that had received either VSV-EnvG or VSVG-EnvG on day 0 with vPE16 on day 30 post-primary vaccination.
Six days after the vPE16 boost (day 36), the levels of tetramer+ CD8 T cells in both the replicating and single-cycle VSV-primed mice were boosted from about 2 to 3% to 24%. These responses clearly represented a boost, because na?ve mice that received vPE16 produced only about 6% Env-specific CD8 T cells at day 6 postvaccination (Fig. 4; Table 1).
The recall response elicited following vPE16 boosting decreased only slightly by day 60, to 16.7% ± 3.3% following VSV-EnvG prime and vPE16 boost and 19.5% ± 1.5% following VSVG-EnvG prime and vPE16 boost, leaving a very high-level memory population (Fig. 4; Table 1). The memory population following vPE16 alone was 3.0% ± 0.19%.
Boosting of responses with G vectors complemented with the same or different G proteins. Infection with VSV produces a strong neutralizing antibody (nAb) response to the G protein. This nAb response prevents boosting of mice primed by VSV with the same VSV vector (25). Because VSVG-EnvG does not express G, we wanted to determine if the complementing G in the particles of VSVG-EnvG was sufficient to induce a neutralizing antibody response to VSV and prevent boosting.
Initially, we determined the VSV neutralizing titer in serum from mice infected with VSV-EnvG and VSVG-EnvG. Twenty-eight days post-i.m. infection, mice infected with both viruses made a neutralizing antibody response to VSV. The titers generated to the G virus were nearly as high as those to the replication-competent virus (Fig. 5A). We next determined if any boosting of CTL responses could be obtained by i.m. vaccination in the presence of the neutralizing antibody. We primed mice on day zero with VSVG-EnvG and then boosted with the same virus on day 30. On day 36, 6 days postboosting of these mice, we observed a twofold increase of Env-specific T cells from day 30 (Fig. 5B; Table 1). However, this recall response was twofold lower than a primary response to VSVG-EnvG on day 6 postvaccination, indicating some effect of the exposure to the virus in the priming vaccination (Fig. 5B; Table 1). However, by day 60, 30 days postboost, the tetramer+ T-cell population had been maintained, indicating that the boost was effective at increasing the memory cell population.
To overcome the interference of nAb in boosting, our group has previously described the use of VSV glycoprotein exchange vectors which employ the same basic vector, but with a G protein from a different VSV serotype or from another vesiculovirus that is not cross-neutralized by antibody to the VSV G (Indiana) serotype (25). To determine if boosting by VSVG-EnvG could be increased using a different G envelope, we complemented the VSVG-EnvG with the G protein from a vesiculovirus called Chandipura.
On day 30 post-i.m. vaccination with VSVG-EnvG (G Indiana complemented), we boosted mice with VSVG-EnvG (complemented with Chandipura G protein). On day 36, 6 days postboost, we saw a strong recall response of Env tetramer+ cells. By day 60 the boost was more evident, with a significant population of memory cells that was over twofold higher than the memory response in these mice on 30 days postvaccination and over threefold higher than the memory response of the boosting virus alone (Fig. 3C; Table 1). While boosting with VSVG-EnvG complemented with Chandipura G produced a slightly better memory response postboost than VSVG-EnvG complemented with Indiana G, both responses were still two- to threefold lower than the memory response produced by a vaccinia boost. However, these experiments still introduce an exclusively nonreplicating viral boosting strategy that is effective.
Cellular immune response to a secreted Env protein expressed from a G virus. Because the VSVG-EnvG virus we are using in this study incorporates the EnvG protein into the viral membrane, it will bind to, fuse with, and replicate in human cells expressing CD4 and appropriate coreceptors (1, 10). A low level of HIV Env (lacking the G tail) is also incorporated into VSV particles. Because the use of viruses expressing membrane-anchored HIV Env proteins in humans would raise new safety issues, we wanted to determine if expression of a non-membrane-anchored Env protein from the G vector would be able to induce equivalent immune responses. This Env protein would not be incorporated into the VSV envelope and thus could not mediate infection of CD4 T cells. We therefore constructed a VSVG virus that expresses the extracellular domains of gp120 and gp41, but lacks the Env transmembrane and cytoplasmic domains (designated gp140). This VSVG-gp140 was recovered and grown in VSV G-complementing cells. As described above, we confirmed expression of the appropriate VSV proteins and the gp140 protein in cells infected with this virus (Fig. 1B).
To examine the CD8 T-cell response to this virus, mice were vaccinated by the i.m. route with either VSVG-EnvG or VSVG-gp140. On day 6 postvaccination, mice that had been vaccinated with VSVG-EnvG or VSVG-gp140 generated comparable responses of 15.1% ± 1.3% and 14.2% ± 1.2%, respectively (Fig. 6).
Antibody responses to EnvG or gp140 expressed from single-cycle vectors. HIV gp140 is a secreted protein lacking the transmembrane and cytoplasmic domain, while the HIV EnvG protein is transported to the plasma membrane and incorporated into VSV particles. Because of these differences in localization, we wanted to determine if there were any significant differences in the antibody responses to these proteins encoded by single-cycle vectors. We therefore infected mice with either VSVG-EnvG or VSVG-gp140 and measured the antibody response to HIVgp140 in the serum 8 weeks postvaccination. We also included a group infected with replication-competent VSV-EnvG to determine if replication had any effect on the induction of antibody. The ELISA titer for antibodies to Env protein was determined as described previously (25). The results in Fig. 7 show that the total antibody titer generated to HIV Env is very similar in the mice inoculated with VSV-EnvG or in VSVG-EnvG-infected mice and slightly greater in the VSV gp140-infected mice. These results are thus parallel to what we have seen for cellular immune responses after i.m. vaccination. Because HIV neutralizing antibodies are not generated in mice without the use of boosting vectors (25), we did not attempt to measure HIV neutralization with these sera.
DISCUSSION
Attenuated VSV recombinants expressing appropriate foreign antigens have been used as effective live viral vaccines in several animal disease models (5, 11, 18, 19, 21-24, 26). However, because these viruses are based on a live recombinant virus, they are subject to strict regulatory standards before testing in humans can begin. VSV recombinants that lack the G glycoprotein gene can be grown in a cell line expressing VSV G (28) and are then capable of only a single round of replication. Viral vectors capable of only a single cycle of replication are subject to much less stringent regulatory standards. For example, the single-cycle RNA virus vectors based on Venezuelan equine encephalitis virus, a neuroinvasive virus, were approved for human trials as an AIDS vaccine (3; http://www.iavireport.org/issues/0803/Vaxbriefs.asp) without any requirement for neurovirulence testing (R. Johnston, personal communication).
In this study we have shown that single-cycle VSV vectors elicit potent CTL and antibody responses to the foreign protein HIV Env encoded in the vector. When given intramuscularly, these immune responses are equivalent to the responses generated by the replication-competent VSV vector expressing HIV Env.
It is interesting that this single-cycle vector generates a much lower cellular immune response than the replication-competent vector when given intranasally. This result is consistent with results obtained using a highly attenuated VSV vector (VSVCT1-EnvG) generated by truncation of the VSV G cytoplasmic domain (17). That vector induced weak immune responses to Env compared to those generated by the recombinant wild-type VSV vector when given intranasally but generated equivalent immune responses when given by the i.p. (17) or i.m. (unpublished observations) route. We suggested previously that systemic spread of rVSV after intranasal inoculation might be required to generate a strong immune response and that the poorly replicating vectors might not be capable of spreading from the lungs into the blood and other organs to induce potent immune responses. Earlier studies have shown that rVSV given intranasally replicates in the lungs, causes a transient viremia, and spreads to other organs (21).
Taken together with the results presented here, it is clear that only a single cycle of replication is required to generate potent immune responses when VSV vectors are given by the intramuscular route. It may be that rVSV undergoes only a single round of replication after intramuscular inoculation because of rapid immune responses or other restrictions on replication in muscle. This rapid control of replication is consistent with the fact that viremia and pathogenesis (weight loss) in mice have not been observed following i.m. vaccination (unpublished results).
We have noted a consistent trend (although not statistically significant) in our data (Fig. 3; Table 1) that the cellular immune responses to HIV Env generated by the single-cycle VSVG-EnvG vector were typically better than those induced by the replicating VSV-EnvG vector. Why would a single-cycle vector be more potent than a replication-competent vector? If we first accept that a single cycle of replication after intramuscular injection is sufficient to induce an optimal immune response, then the slightly better responses to HIV Env may be due simply to the absence of the upstream G gene in the VSVG-EnvG virus. Each upstream gene results in a attenuation of expression the downstream gene (9, 12). Therefore, deletion of the G gene would be expected to increase expression of EnvG relative to the remaining VSV genes during a single round of replication. In fact, EnvG expression was increased more than twofold relative to the upstream N gene in the G vector compared to the replication-competent vector. Furthermore, the lack of G expression would be expected to eliminate some competition for MHC class I presentation and could result in upregulation of CD8 T-cell responses to Env.
In previous studies, where a VSVG vector expressing an influenza virus HA protein was used as an influenza vaccine, it produced much lower neutralizing antibody responses than the replication-competent vector but was able to provide protection from influenza virus challenge in animals given a booster inoculation (21). The poor response in these studies can now be explained by the choice of the intranasal, rather than the intramuscular, route for vaccination. Our studies suggest that repetition of the influenza virus studies using G vectors given by the intramuscular route would likely yield more robust anti-HA responses.
Previous studies from our laboratory have shown that VSV vectors expressing SHIV proteins protect macaques from progression to AIDS following a SHIV challenge (18, 24). Since the rVSV vector in the macaques studies was identical to the replication-competent VSV vector used in this study, our findings may suggest that a VSV vector that undergoes only a single cycle of replication may be equally effective in macaques as the rVSV at producing Env-specific CD8 T cells when given by the appropriate route. Even though the attenuated rVSV has not caused any measurable disease symptoms in more than 100 vaccinated macaques, rare vaccine-associated disease cannot be ruled out. A single-cycle vector is a potent alternative that would be safe for use even in people with severe immunodeficiency.
A recent study has shown that rVSV vectors expressing SHIV Env and Gag proteins are more effective at generating cellular immune responses in macaques when given by the i.n. route compared to the i.m. route (5). We have not observed this same dichotomy in mice with replicating vectors. In fact, the intranasal vaccination route generally yields somewhat lower cellular immune responses than the i.m. or i.p. routes (7, 17). We have not yet used our single-cycle vectors in macaques, but our results in mice indicate that comparison of the magnitude of the immune responses generated by different routes of vaccination should be undertaken.
ACKNOWLEDGMENTS
We thank members of the Rose laboratory for helpful discussions and suggestions on the manuscript.
This work was supported by NIH grants R37-AI40357 and RO1-AI45510 and by an HIV-1 Vaccine Design and Development Team contract (NIH NO1-AI-25458).
REFERENCES
Boritz, E., J. Gerlach, J. E. Johnson, and J. K. Rose. 1999. Replication-competent rhabdoviruses with human immunodeficiency virus type 1 coats and green fluorescent protein: entry by a pH-independent pathway. J. Virol. 73:6937-6945.
Coles, R. M., S. N. Mueller, W. R. Heath, F. R. Carbone, and A. G. Brooks. 2002. Progression of armed CTL from draining lymph node to spleen shortly after localized infection with herpes simplex virus 1. J. Immunol. 168: 834-838.
Davis, N. L., A. West, E. Reap, G. MacDonald, M. Collier, S. Dryga, M. Maughan, M. Connell, C. Walker, K. McGrath, C. Cecil, L. H. Ping, J. Frelinger, R. Olmsted, P. Keith, R. Swanstrom, C. Williamson, P. Johnson, D. Montefiori, and R. E. Johnston. 2002. Alphavirus replicon particles as candidate HIV vaccines. IUBMB Life 53:209-211.
Earl, P. L., S. Koenig, and B. Moss. 1991. Biological and immunological properties of human immunodeficiency virus type 1 envelope glycoprotein: analysis of proteins with truncations and deletions expressed by recombinant vaccinia viruses. J. Virol. 65:31-41.
Egan, M. A., S. Y. Chong, N. F. Rose, S. Megati, K. J. Lopez, E. B. Schadeck, J. E. Johnson, A. Masood, P. Piacente, R. E. Druilhet, P. W. Barras, D. L. Hasselschwert, P. Reilly, E. M. Mishkin, D. C. Montefiori, M. G. Lewis, D. K. Clarke, R. M. Hendry, P. A. Marx, J. H. Eldridge, S. A. Udem, Z. R. Israel, and J. K. Rose. 2004. Immunogenicity of attenuated vesicular stomatitis virus vectors expressing HIV type 1 Env and SIV Gag proteins: comparison of intranasal and intramuscular vaccination routes. AIDS Res. Hum. Retrovir. 20:989-1004.
Florkiewicz, R. Z., and J. K. Rose. 1984. A cell line expressing vesicular stomatitis virus glycoprotein fuses at low pH. Science 225:721-723.
Haglund, K., I. Leiner, K. Kerksiek, L. Buonocore, E. Pamer, and J. K. Rose. 2002. High-level primary CD8+ T-cell response to human immunodeficiency virus type 1 gag and env generated by vaccination with recombinant vesicular stomatitis viruses. J. Virol. 76:2730-2738.
Haglund, K., I. Leiner, K. Kerksiek, L. Buonocore, E. Pamer, and J. K. Rose. 2002. Robust recall and long-term memory T-cell responses induced by prime-boost regimens with heterologous live viral vectors expressing human immunodeficiency virus type 1 Gag and Env proteins. J. Virol. 76:7506-7517.
Iverson, L. E., and J. K. Rose. 1981. Localized attenuation and discontinuous synthesis during vesicular stomatitis virus transcription. Cell 23:477-484.
Johnson, J. E., M. J. Schnell, L. Buonocore, and J. K. Rose. 1997. Specific targeting to CD4+ cells of recombinant vesicular stomatitis viruses encoding human immunodeficiency virus envelope proteins. J. Virol. 71:5060-5068.
Kahn, J. S., A. Roberts, C. Weibel, L. Buonocore, and J. K. Rose. 2001. Replication-competent or attenuated, nonpropagating vesicular stomatitis viruses expressing respiratory syncytial virus (RSV) antigens protect mice against RSV challenge. J. Virol. 75:11079-11087.
Kretzschmar, E., L. Buonocore, M. J. Schnell, and J. K. Rose. 1997. High-efficiency incorporation of functional influenza virus glycoproteins into recombinant vesicular stomatitis viruses. J. Virol. 71:5982-5989.
Lawson, N. D., E. A. Stillman, M. A. Whitt, and J. K. Rose. 1995. Recombinant vesicular stomatitis viruses from DNA. Proc. Natl. Acad. Sci. USA 92:4477-4481.
Niwa, H., K. Yamamura, and J. Miyazaki. 1991. Efficient selection for high-expression transfectants with a novel eukaryotic vector. Gene 108:193-199.
Okuma, K., Y. Matsuura, H. Tatsuo, Y. Inagaki, M. Nakamura, N. Yamamoto, and Y. Yanagi. 2001. Analysis of the molecules involved in human T-cell leukaemia virus type 1 entry by a vesicular stomatitis virus pseudotype bearing its envelope glycoproteins. J. Gen. Virol. 82:821-830.
Owens, R. J., and J. K. Rose. 1993. Cytoplasmic domain requirement for incorporation of a foreign envelope protein into vesicular stomatitis virus. J. Virol. 67:360-365.
Publicover, J., E. Ramsburg, and J. K. Rose. 2004. Characterization of nonpathogenic, live, viral vaccine vectors inducing potent cellular immune responses. J. Virol. 78:9317-9324.
Ramsburg, E., N. F. Rose, P. A. Marx, M. Mefford, D. F. Nixon, W. J. Moretto, D. Montefiori, P. Earl, B. Moss, and J. K. Rose. 2004. Highly effective control of an AIDS virus challenge in macaques by using vesicular stomatitis virus and modified vaccinia virus Ankara vaccine vectors in a single-boost protocol. J. Virol. 78:3930-3940.
Reuter, J. D., B. E. Vivas-Gonzalez, D. Gomez, J. H. Wilson, J. L. Brandsma, H. L. Greenstone, J. K. Rose, and A. Roberts. 2002. Intranasal vaccination with a recombinant vesicular stomatitis virus expressing cottontail rabbit papillomavirus L1 protein provides complete protection against papillomavirus-induced disease. J. Virol. 76:8900-8909.
Riedel, H., C. Kondor-Koch, and H. Garoff. 1984. Cell surface expression of fusogenic vesicular stomatitis virus G protein from cloned cDNA. EMBO J. 3:1477-1483.
Roberts, A., L. Buonocore, R. Price, J. Forman, and J. K. Rose. 1999. Attenuated vesicular stomatitis viruses as vaccine vectors. J. Virol. 73: 3723-3732.
Roberts, A., E. Kretzschmar, A. S. Perkins, J. Forman, R. Price, L. Buonocore, Y. Kawaoka, and J. K. Rose. 1998. Vaccination with a recombinant vesicular stomatitis virus expressing an influenza virus hemagglutinin provides complete protection from influenza virus challenge. J. Virol. 72:4704-4711.
Roberts, A., J. D. Reuter, J. H. Wilson, S. Baldwin, and J. K. Rose. 2004. Complete protection from papillomavirus challenge after a single vaccination with a vesicular stomatitis virus vector expressing high levels of L1 protein. J. Virol. 78:3196-3199.
Rose, N. F., P. A. Marx, A. Luckay, D. F. Nixon, W. J. Moretto, S. M. Donahoe, D. Montefiori, A. Roberts, L. Buonocore, and J. K. Rose. 2001. An effective AIDS vaccine based on live attenuated vesicular stomatitis virus recombinants. Cell 106:539-549.
Rose, N. F., A. Roberts, L. Buonocore, and J. K. Rose. 2000. Glycoprotein exchange vectors based on vesicular stomatitis virus allow effective boosting and generation of neutralizing antibodies to a primary isolate of human immunodeficiency virus type 1. J. Virol. 74:10903-10910.
Schlereth, B., J. K. Rose, L. Buonocore, V. ter Meulen, and S. Niewiesk. 2000. Successful vaccine-induced seroconversion by single-dose immunization in the presence of measles virus-specific maternal antibodies. J. Virol. 74:4652-4657.
Schnell, M. J., L. Buonocore, M. A. Whitt, and J. K. Rose. 1996. The minimal conserved transcription stop-start signal promotes stable expression of a foreign gene in vesicular stomatitis virus. J. Virol. 70:2318-2323.
Schnell, M. J., J. E. Johnson, L. Buonocore, and J. K. Rose. 1997. Construction of a novel virus that targets HIV-1-infected cells and controls HIV-1 infection. Cell 90:849-857.(Jean Publicover, Elizabet)