Rabies Virus-Based Vectors Expressing Human Immunodeficiency Virus Type 1 (HIV-1) Envelope Protein Induce a Strong, Cross-Reactive Cytotoxic T-Lymphocyte Response against Envelope Proteins from Different HIV-1 Isolates

Abstract

Novel viral vectors that are able to induce both strong and long-lasting immune responses may be required as effective vaccines for human immunodeficiency virus type 1 (HIV-1) infection. Our previous experiments with a replication-competent vaccine strain-based rabies virus (RV) expressing HIV-1 envelope protein from a laboratory-adapted HIV-1 strain (NL4–3) and a primary HIV-1 isolate (89.6) showed that RV-based vectors are excellent for B-cell priming. Here we report that cytotoxic T-lymphocyte (CTL) responses against HIV-1 gp160 are induced by recombinant RVs. Our results indicated that a single inoculation of mice with an RV expressing HIV-1 gp160 induced a solid and long-lasting memory CTL response specific for HIV-1 envelope protein. Moreover, CTLs from immunized mice were not restricted to the homologous HIV-1 envelope protein and were able to cross-kill target cells expressing HIV-1 gp160 from heterologous HIV-1 strains. These studies further suggest promise for RV-based vectors to elicit a persistent immune response against HIV-1 and their potential utility as efficacious anti-HIV-1 vaccines.

New antiretroviral strategies against human immunodeficiency virus (HIV) type 1 (HIV-1) have resulted in a dramatic decrease in mortality among infected humans in developed countries, but the development of a successful vaccine to prevent infection is still the major goal to halt the HIV-1 pandemic. A human being is infected with HIV-1 every 10 s on average, and in heavily affected countries in Africa, such as Zambia and Uganda, nearly 40% of young adults are HIV-1 seropositive.

Currently, a variety of HIV vaccine strategies are being investigated, including recombinant proteins (16, 33, 37), peptides (6, 8, 30), naked DNA (3, 5, 10, 25, 32, 34, 38), replication-competent and non-replication-competent (replicon) live viral vectors (7, 13, 19, 27–29, 36), and prime-boost combinations (for a review, see reference 4). A large number of these vaccine strategies have been tested in the simian immunodeficiency virus (SIV) macaque model system, but to date no potent protective immunity has been obtained, although some amelioration of disease course has been seen (5, 13, 29). So far, the only effective method for protecting macaques from SIV infection is the use of live attenuated SIV. Studies showed that a genetically modified, nef deletion SIV strain that does not cause disease in rhesus monkeys induced high anti-SIV titers of antibodies and cytotoxic T-lymphocyte (CTL) activity (12, 22). Subsequent challenge of the immunized animals with infectious doses of a pathogenic SIV strain yielded protection from infection (12). A major drawback for the use of attenuated lentivirus vaccine approaches is the finding that even SIV with a nef deletion can give rise to an AIDS-like disease in both neonatal and adult macaques (1, 2, 14). Additional concerns regarding the use of attenuated lentiviruses arise from the recent finding that in some instances, recombination of live attenuated SIV with challenge virus results in an even more virulent strain (18). Nevertheless, study results have indicated that live viral vectors may be excellent vaccine candidates for an HIV-1 vaccine.

We have recently developed a new potential HIV-1 vaccine based on an attenuated replication-competent rabies virus (RV) expressing HIV-1 gp160 from both a laboratory-adapted strain (NL4–3) and a primary HIV-1 isolate (89.6) (36). The HIV-1 envelope protein was stably and functionally expressed and induced a strong humoral response directed against the HIV-1 envelope protein after a single boost with recombinant gp120 in mice. Moreover, high neutralization titers against HIV-1 could be detected in the mouse sera.

The immune response(s) required to protect against HIV-1 infection is currently unknown, but a protective immune response against HIV-1 might require both major arms of the immune system. Recent reports on vaccine approaches using recombinant HIV-1 envelope protein suggested that an exclusively humoral response is not sufficient to protect against HIV-1 infection, but the passive transfer of three monoclonal antibodies directed against HIV-1 envelope protein resulted in protection of macaques against subsequent challenge with a pathogenic HIV-1/SIV chimeric virus (26). Other studies indicated that a cell-mediated response plays an important role in controlling HIV-1 infection (9, 17). Exposed but uninfected individuals often have HIV-1-specific CTLs but no detectable antibodies against HIV-1 (31, 35).

Little information is available regarding the induction of CTL responses against foreign proteins expressed by rhabdovirus-based vectors. In this study, we analyze the potency of recombinant RVs expressing HIV-1 envelope protein to induce HIV-1-specific CTLs.

Induction of long-lasting HIV-1 gp160-specific CTLs.

Our previous experiments with a recombinant RV expressing HIV-1 envelope protein from a laboratory-adapted HIV-1 strain (NL4–3) and a primary HIV-1 isolate (89.6) showed that RV-based vectors are excellent for B-cell priming (36). In the present study, we analyze the memory CTL response against HIV-1 envelope protein expressed by attenuated RV-based vectors. As noted, increasing evidence suggests that the induction of a vigorous, long-lasting CTL response will be an important feature for a successful HIV-1 vaccine.

To analyze the potency of RV-based vectors to induce a cytotoxic response against HIV-1, we immunized six mice with 2 × 10⁷ focus-forming units (FFU) of the previously described (36) recombinant RV expressing HIV-1_NL4–3 envelope protein (SBN-NL4–3). Three mice were sacrificed 105 or 135 days after infection, and the spleens were removed. One-third of the splenocyte culture was infected at a multiplicity of infection (MOI) of 1 with a recombinant vaccinia virus expressing HIV-1_NL4–3 gp160 for 16 h, deactivated using psoralen (Sigma) and UV treatment, and added back to the culture as presenter cells. Stimulated effector cells were analyzed 7 days after activation for their ability to kill P815 target cells infected with wild-type vaccinia virus, a recombinant vaccinia virus expressing HIV-1_NL4–3 gp160, or a recombinant vaccinia virus expressing HIV-1 Gag.

As shown in Fig. 1, a strong cytotoxic response was detected only against P815 target cells infected with the recombinant vaccinia virus expressing HIV-1 envelope protein. Only a low percentage of lysis was observed for P815 cells infected with the other two vaccinia viruses. Of note, these responses were achieved after a single inoculation with a recombinant RV expressing HIV-1 envelope protein, indicating that RV-based vectors are able to induce long-lasting CTLs after a single vaccination.

FIG. 1.

CTLs from HIV-1 gp160-immunized mice induce long-lasting HIV-1 gp160-specific CTLs. Groups of three 6- to 8-week-old female BALB/c mice (Harlan Sprague-Dawley) were inoculated i.p. with 2 × 10⁷ FFU of recombinant RV expressing HIV-1_NL4–3 envelope protein. At 105 (A) to 135 (B) days after the single inoculation, spleens from three mice were aseptically removed and combined, and single-cell suspensions were prepared. Red blood cells were lysed with ACK lysing buffer (BioWhittaker) and washed twice in RPMI-10 medium containing 10% fetal bovine serum. Splenocytes were divided into effector and stimulator cells. Stimulator cells were prepared by infection with vaccinia virus expressing the envelope protein from HIV-1_NL4–3 (vCB41) (▪) at an MOI of 1 for 2 h. Cells were washed once to remove excess virus and incubated for 16 h. After incubation, the vaccinia virus was inactivated using psoralen at a final concentration of 5 μg/ml for 10 min, treated with long-wave UV for 4 min, and washed twice. Stimulator cells were added back to the effector cell population at a ratio of 3:1, and 10% T-STIM (Collaborative Biomedical Products) was added as a source of interleukin-2. The cytolytic activity of cultured CTLs was measured by a 4-h assay with ⁵¹Cr-labeled P815 target cells 7 days after in vitro stimulation. Target cells were prepared by infection with vaccinia virus expressing HIV-1_NL4–3 gp160 for 1 h at an MOI of 10. Cells were washed to remove excess virus and incubated for 16 h. To measure background, target cells were infected with vaccinia virus expressing HIV-1 Gag (vP1287) (○) or wild-type vaccinia virus (vP1170) (▵). Target cells were washed once, labeled with 100 μCi of ⁵¹Cr for 1 h, washed twice, and added to effector cells at various E/T ratios for 4 h. The percent specific ⁵¹Cr release was calculated as 100 × [(experimental release − spontaneous release)/(maximum release − spontaneous release)]. Maximum release was determined from supernatants of cells that were lysed by the addition of 5% Triton X-100. Spontaneous release was determined from target cells incubated without added effector cells.

CTLs from HIV-1 gp160-immunized mice cross-kill target cells expressing heterologous HIV-1 envelope proteins.

There is a significant difference in HIV-1 envelope amino acid sequences, but cross-protection between divergent viruses will be a likely requirement for a protective HIV-1 vaccine. To analyze the potency of our vaccine candidate to induce cross-reactive CTLs against gp160 from different HIV-1 strains, we screened splenocytes from mice immunized with a recombinant RV expressing HIV-1 gp160 against P815 target cells expressing homologous and heterologous HIV-1 envelope proteins. This approach seemed to be most promising because the sequences of the HIV-1 envelope proteins used are quite different (Table 1) and this method does not require knowledge of a certain epitope that is conserved among different HIV-1 envelope proteins. For HIV-1 gp160, no H-2^d-restricted CTL epitopes are known for primary isolates (HIV Molecular Immunology Database, Los Alamos National Laboratory, Theoretical Biology and Biophysics, Los Alamos, N.Mex.).

TABLE 1.

Amino acid identity and divergence for HIV-1 strain NL4–3, 89.6, Ba-L, and JR-FL envelope proteins

HIV-1 strain	% Amino acid identity or divergencea
	NL4–3	89.6	Ba-L	JR-FL
NL4–3		83.3	86.9	86.9
89.6	19.0		82.7	84.2
Ba-L	14.4	19.7		88.8
JR-FL	14.4	17.8	12.1

^aValues in italics are percent identities; values in boldface are percent divergences. 

Two groups of six mice were immunized intraperitoneally (i.p.) with 2 × 10⁷ FFU of recombinant RV expressing HIV-1 gp160 from a laboratory-adapted, CXCR4-tropic strain (NL4–3) or a dualtropic (CXCR4 and CCR5) isolate (89.6). Three and five weeks after immunization, three mice from each group were sacrificed, the spleens were removed, and the pooled splenocytes were stimulated with a recombinant vaccinia virus expressing the homologous HIV-1 envelope protein (NL4–3 or 89.6). Seven days after stimulation, effector cells were analyzed for their ability to lyse P815 cells infected with recombinant vaccinia viruses expressing HIV-1 envelope proteins from the laboratory-adapted, CXCR4-tropic HIV-1 strain (NL4–3), the dualtropic strain (89.6), and two primary, CCR5-tropic HIV-1 strains (Ba-L and JR-FL).

The results from two different, independent experiments are shown in Fig. 2A for mice immunized with an RV expressing HIV-1_NL4–3 Env and in Fig. 2B for mice immunized with an RV expressing HIV-1_89.6 Env. As expected, strong, specific lysis of P815 cells expressing the homologous antigen was observed for both groups. More striking, the effector cells were able to cross-kill P815 target cells expressing heterologous HIV-1 envelope proteins. Activated splenocytes from SBN-NL4–3-immunized mice achieved specific lysis of P815 cells expressing gp160 from JR-FL or 89.6 in the 40% range at an effector/target (E/T) ratio of 50:1 and were also able to cross-kill target cells expressing HIV-1_Ba-L gp160. Cross-killing was also observed with effector cells from SBN-89.6-primed mice. P815 target cells expressing heterologous HIV-1 envelope protein were lysed in the same range as that observed for activated splenocytes from mice immunized with SBN-NL4–3, except that only about 20% of P815 cells expressing HIV-1_NL4–3 were lysed. These data indicate that CTLs against HIV-1 gp160 induced by RV-based vectors may be directed against different epitopes within the HIV-1 envelope protein.

FIG. 2.

CTLs from HIV-1 gp160-immunized mice cross-kill target cells expressing heterologous HIV-1 envelope proteins. Groups of six 6- to 8-week-old female BALB/c mice were inoculated i.p. with 2 × 10⁷ FFU of recombinant RV expressing HIV-1 envelope protein from strain NL4–3 (A) or 89.6 (B). At 3 and 4 weeks after the single inoculation, spleens were aseptically removed, and splenocytes were stimulated in vitro with vaccinia virus expressing the homologous HIV-1 envelope protein as described in the legend to Fig. 1. Target cells were prepared by infection with vaccinia virus expressing HIV-1 envelope proteins from strain NL4–3 (vCB41) (▪), 89.6 (vBD3) (●), JR-FL (vCB28) (▴), or Ba-L (vCB43) (♦). To measure background, target cells were infected with vaccinia virus expressing HIV-1 Gag (vP1287) (○) or wild-type vaccinia virus (vP1170) (▵). Chromium release assays were completed as described in the legend to Fig. 1. The results shown are from two different, independent experiments. Error bars show standard deviations.

HIV-1-specific CTL activity is mediated by CD8⁺ T cells.

The phenotype of the T-cell subpopulation mediating cytolytic activity was assessed by selective depletion. Three mice were immunized with 2 × 10⁷ FFU of a recombinant RV expressing HIV-1_NL4–3 envelope protein, and the spleens were removed 18 weeks later. Splenocytes were restimulated with a recombinant vaccinia virus expressing the homologous HIV-1 envelope protein for 7 days. Immunomagnetic bead cell separation was completed to both deplete and positively isolate CD8⁺ T cells from the activated splenocyte culture. Chromium release assays were completed using cultures depleted of CD8⁺ T cells (CD8⁻), cultures of isolated CD8⁺ cells (CD8⁺), or unprocessed cultures (CD8⁺ CD8⁻). P815 target cells were infected with a vaccinia virus expressing HIV-1_NL4–3 gp160 or HIV-1 Gag.

As shown in Fig. 3, the CD8⁺ T-cell-depleted cultures showed no activity, while the CD8⁺ T-cell-enriched and unprocessed cultures showed high specific lysis at E/T ratios of 25:1 and 12.5:1. Indeed, the CD8⁺ T-cell-enriched population was also enriched in lytic units, as the CTL activity was still on a plateau at 12.5:1, in contrast to the results for the unselected population. These data indicate that cytolytic activity is mediated by the CD8⁺ T-cell subpopulation. Furthermore, these results suggest that in addition to antibodies, recombinant RV-based vectors also generate long-lived anti-HIV-1 CD8⁺ T-cell responses.

FIG. 3.

Cytolytic activity is mediated by CD8⁺ T cells. Groups of three 6- to 8-week-old female BALB/c mice were inoculated i.p. with 2 × 10⁷ FFU of recombinant RV expressing HIV-1 envelope protein from the NL4–3 strain. At 18 weeks after the single inoculation, spleens were aseptically removed, and splenocytes were stimulated in vitro with vaccinia virus expressing HIV-1_NL4–3 envelope protein as described in the legend to Fig. 1. At 7 days after in vitro stimulation, CD8⁺ T cells were depleted from the cell culture (CD8⁻) or enriched (CD8⁺) using Dynabeads mouse CD8 (Lyt2) as described by the manufacturer. Chromium release assays were completed as described in the legend to Fig. 1 with cultures depleted of (CD8⁻) or enriched for (CD8⁺) CD8 T cells or (CD8⁺ CD8⁻) unprocessed cultures. Target cells were prepared by infection with vaccinia virus expressing HIV-1 envelope protein from NL4–3 (vCB41). To measure background, target cells were infected with vaccinia virus expressing HIV-1 Gag (vP1287). Background levels were equal to or below 6% specific lysis. E/T ratios were 25:1 (gray bars) and 12.5:1 (black bars).

Summary.

We previously demonstrated that RV-based vectors expressing HIV-1 envelope proteins are able to induce a humoral response against HIV-1 gp160 after a single immunization followed by a booster injection with recombinant HIV-1 gp120 (36). Expanding evidence suggests that CTL responses play a major role in the immune response against HIV-1 (9). The development of an effective prophylactic HIV-1 vaccine therefore probably requires the selection of an HIV-1 antigen(s) capable of inducing long-lasting and broadly reactive CTL responses. The results presented here indicate that RV-based vectors are excellent vectors for inducing such responses. In contrast to the observed humoral response, a single inoculation of mice with a recombinant RV expressing HIV-1 envelope protein resulted in a vigorous CTL response against HIV-1 Env. In addition, this response was stable for at least 135 days after immunization. One explanation for the strong response may be that RV grows in various cell lines without killing the cells, a characteristic which probably results in longer-lasting expression of HIV-1 genes than that seen with a cytopathogenic viral vector. In addition, the expression of the RV nucleoprotein, which was previously shown to be an exogenous superantigen (23, 24), might help to enhance a general immune response against the HIV-1 envelope protein after a single immunization.

Our recombinant RVs were able to induce cross-reactive CTLs against a variety of different HIV-1 envelope proteins. Previous studies showed that single amino acid exchanges can abrogate CTL cross-reactivity, but other examinations indicated that single or even double amino acid substitutions frequently do not abrogate cross-killing (11, 20, 21). Therefore, the question remains as to whether CTLs induced by recombinant RVs are directed against different epitopes. However, our results are encouraging because several studies have indicated that CTLs from HIV-1-infected individuals show cross-reactivity even with different clades of HIV-1; thus, broad cross-reactivity is an important requirement for an HIV-1 vaccine (11, 35). To our knowledge, only one study has shown cross-clade CTL reactivities induced with a canarypox virus-based HIV-1 vaccine in uninfected volunteers (15). We are currently analyzing whether CTLs induced against HIV-1 gp160 by recombinant RVs are also cross-reactive against HIV-1 envelope proteins from clades other than B.

In summary, we have shown that a single vaccination with a recombinant RV expressing HIV-1 envelope protein elicits a strong, long-lasting CTL response against envelope proteins of different HIV-1 strains. These results further emphasize the use of RV as a potential HIV-1 vaccine. In contrast to the situation for most other viral vectors, only negligible seropositivity for RV exists in the human population, and immunization with an RV-based vector against HIV-1 would not interfere with immunity against the vector itself. Because oral immunization against RV with an RV vaccine strain was successful and apathogenic in chimpanzees (World Health Organization, unpublished document W. H. O./Rab.Res./93.42), an RV-based vector may also be promising in inducing mucosal immunity against HIV-1. Further exploration of this vaccine strategy in macaques will indicate the usefulness of such vectors in inducing protective immunity against HIV-1.