INTRODUCTION

Top Abstract Introduction Materials and methods Results Discussion References

Currently the most effective means of vaccination under development by the human immunodeficiency virus (HIV) scientific research community is attenuated virus. The studies of attenuated lentiviruses have been limited to animal models (2, 9, 13, 27, 31, 38, 47) and a few naturally occurring cases in the human population (10, 22, 26). All of the naturally attenuated viruses isolated from both humans and macaques have been reported to cause little or no pathogenicity in the infected hosts (22, 26), although reversion to pathogenic virus has been reported (46). Attenuated isolates, constructed by molecular biologic techniques (9, 13, 33), have been used to immunize adult nonhuman primates with little observed pathology, although recent studies have found that they can cause disease in newborn macaques (4, 48). The attenuated-virus vaccine strategy has been tested and proven to be very effective at protecting juvenile to adult animals from homologous challenge virus (2, 9, 27, 31, 47). Inactivated isolates tested in vivo have included both natural (2, 3, 31) and constructed deletions in the nef coding region (9), and constructs with deletions in the accessory genes vif, vpr, and vpx and in downstream portions of the nef coding regions found in the 3' long terminal repeat have also been studied (13). As was originally predicted by Daniel et al. (9) during their studies involving the nef gene deletion mutants of simian immunodeficiency virus (SIV), the greater the number of deletions built into the virus accessory genes, the less efficiently the virus grows both in vitro and in vivo, suggesting a lower pathogenicity. This lower growth potential is associated with a reduced immune response and less-effective protection following challenge (13). Evaluation of the immune responses following immunization has revealed that the attenuated virus can induce potent immunity associated with both humoral and cellular responses (18, 47). Upon challenge with homologous and heterologous SIV/HIV-2, animals immunized with nef deletion mutants or mutants with deletions in nef and vpr were usually completely protected from the challenge. Immunized animals infected with the challenge virus were found to limit the replication and survive significantly longer than unimmunized controls; however, this has not always been true for all attenuated-virus isolates (11, 46).

The mechanism of protection induced by attenuated lentiviruses is currently unknown. Immunized animals have been shown to produce strong cellular and humoral responses to the immunizing virus, including cytotoxic T-cell production (18), helper T-cell responses (15), and neutralizing antibodies (18). Unfortunately, no correlates of the protective immunity have been determined. Passive antibody transfer studies recently performed by Almond et al. (3) have not shown any linkage between the antibodies developed following immunization and the observed protection. In addition, recent studies by Langlois et al. (23) have shown that the antiviral antibodies generated by attenuated-SIV-immunized animals fail to neutralize the SIV challenge stocks in vitro. In vivo depletion of lymphocytes with anti-CD57 or anti-CD8 monoclonal antibody treatments did not affect the protective responses of the immunized animals (39). These studies suggest that the mechanism may be viral interference by receptor blocking (45) or the presence of other nonspecific immune mediators, such as cytokines or chemokines (6, 7, 12, 41).

The present studies were designed to focus on the humoral responses generated during immunization and to determine the breadth of the protective response by challenging with a pathogenic simian-human immunodeficiency virus (SHIV) that contains an HIV-1-derived envelope. The challenge virus, SHIV_89.6PD, is highly infectious and causes a rapid pathology in macaques (25, 28, 34, 40). It was constructed by using the SIV₂₃₉ molecular clone as the backbone with the substitution of the HIV-1_89.6 envelope gene for the SIV₂₃₉ envelope (35). The construct has recently been described by Reimann et al. (35), and an in vivo-passaged isolate has been derived and cloned (20). This isolate infects macaques when they are challenged either by intravenous injection or application on mucosal surfaces (25, 40) and causes a rapid disease course with a near-total loss of detectable CD4 cells. For the present studies we chose to use two nef-deleted SIV strains, SIV_239nef (9) and SIV_PBj6.6nef (33). The results of this study show that these attenuated viruses can totally protect some animals from infection with the SHIV_89.6PD and that if infection occurs, the acute-disease course can be altered. However, animals that were unable to control the immunizing viruses were found to have difficulty in controlling the challenge virus and were subsequently found to progress to disease more rapidly than the controls.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and methods Results Discussion References

Animals. Animals used in this study were captive-bred, juvenile rhesus macaques from a breeding colony in the United States or imported pigtailed macaques. All animals were screened and confirmed to be free of antibodies to SIV, simian retrovirus (SRV), and simian T-cell leukemia virus type 1 (STLV-1) and free of isolatable SIV and SRV following culture performed prior to their inclusion in the study. The animals were housed in accordance with American Association for Accreditation of Laboratory Animal Care standards.

Virus stocks. Construction of the SIV_239nef and SIV_PBj6.6nef viruses has been described previously (9, 33). The SIV_239nef virus stock was supplied by Ronald C. Desrosiers. The SIV_PBj6.6nef construct was supplied by Frank Novembre. The attenuated-virus stocks used for this experiment were grown in CEM×174 cells in an acute infection with harvest at day 12 postinoculation (p.i.). The stocks were filtered through a 0.45-µm-pore-size filter and frozen in aliquots at 165°C. The titers of these stocks were 10³ 50% tissue culture infective doses (TCID₅₀) per ml, as measured by using log dilutions on CEM×174 cells. The rhesus macaques were inoculated intravenously with 1.0 ml of the undiluted SIV_239nef stock, and the pigtailed macaques were inoculated intravenously with the SIV_PBj6.6nef stock. The SHIV_89.6 construction has been previously described (35). The pathogenic isolate derivation has been described in detail by Reimann et al. (34). The specific isolate (SHIV_89.6PD) used in these studies was derived in CEM×174 cells from culture with the plasma of the second in vivo-passage of the 89.6 isolate (25). The virus stock was produced from acute infection of macaque peripheral blood lymphocytes for 7 days. The titer of the stock was 2.5 × 10⁴ TCID₅₀/ml in CEM×174 cells and 2.5 × 10⁴ 50% macaque infectious doses. The animals were challenged with 25 50% infective doses of the SHIV_89.6PD construct/ml by intravenous injection.

Virus detection and isolation. SIV p27 was detected in the plasma of the macaques by using an SIV p27-specific enzyme-linked immunosorbent assay (ELISA) (Coulter Diagnostics, Miami, Fla.). For isolating virus, macaque peripheral blood lymphocytes were isolated by using a 95% Ficoll-Hypaque separation solution (Sigma, St. Louis, Mo.). Approximately 2.5 × 10⁶ cells were placed in 2.5 ml of RPMI 1640 medium containing 10% fetal bovine serum, 0.5% gentamicin, 5.0% glutamine, and 5 mg of phytohemagglutinin-p (PHA-p) (Sigma) per ml. Following 3 days of culture, the supernatant was removed and a portion was tested for the presence of SIV p27 (Coulter Diagnostics). The cells were washed free of PHA, adjusted to 10⁶/ml, and placed in media without PHA-p and with 10% recombinant interleukin-2 (Boehringer Mannheim, Indianapolis, Ind.). An equal number of human 3-day PHA-blasted peripheral blood mononuclear cells (PBMC) were then added. Cultures were readjusted to 10⁶ cells/ml every 3 to 4 days with a 100% medium change and assayed twice per week for the presence of SIV p27. Cultures were held for up to 4 weeks. A culture was designated as positive after positive test results were obtained from two successive specimens.

Flow cytometry. Peripheral blood lymphocyte subset analysis was performed on a FACScan flow cytometer (Becton Dickinson, Mountain View, Calif.) with a panel of mouse anti-human monoclonal antibodies to B cells (phycoerythrin-conjugated CD20), T cells (fluorescein isothiocyanate-conjugated CD2), and T-cell subsets (phycoerythrin-conjugated CD4 and fluorescein isothiocyanate-conjugated CD8) (Becton Dickinson). Analysis was performed by a whole-blood-lysis procedure as directed by the manufacturer.

Anti-SIV antibody ELISA. Circulating SIV antibodies were detected by an HIV-2 ELISA (Genetic Systems, Seattle, Wash.) which is cross-reactive for SIV antibody. The ELISA used a fixed plasma dilution of 1:200. All samples from an individual animal were run at the same time. The assay was run as directed except that the plates were read on a Vmax kinetic ELISA reader (Molecular Devices, Menlo Park, Calif.) at 650 nm. Values of 10³ optical density (OD) units per min minus background were used as ELISA units.

ELISA measurement of HIV-1 gp160-specific serum IgG binding. Affinity-purified oligomeric HIV-1 gp160₄₅₁ (Advanced Bioscience Laboratories, Kensington, Md.) (0.65 µg/ml in phosphate-buffered saline) (PBS) (pH 7.4; 0.01% thimerosal) was coated onto Immulon 2 microtiter plates (Dynatech, Chantilly, Va.), left overnight at 4°C, and assayed as described previously (44). Oligomeric gp160 was affinity purified from cell cultures infected with HIV-1₄₅₁ as described previously (19). Briefly, plates were washed twice with wash buffer (PBS with 0.1% Tween 20 [pH 7.4]) prior to the incubation with twofold dilutions of macaque serum, diluted in diluent (wash buffer with 5% skim milk [pH 7.4]), for 1 h at 37°C. Plates were washed three times with wash buffer and incubated with horseradish peroxidase-conjugated goat anti-monkey immunoglobulin G (IgG) (Nordick Laboratories, San Clemente, Calif.) (diluted 1:4,000 in serum diluent). After a 1-h incubation at 37°C, plates were washed three times and substrate [2,2'-azinobis(3-ethylbenzthiazolinesulfonic acid); Kirkegaard and Perry, Gaithersburg, Md.] was then added. The reaction was stopped with addition of 0.5% sodium dodecyl sulfate after 30 min at 37°C.

Quantitative PCR of viral RNA. Circulating levels of viral RNA were determined with an externally controlled PCR assay. Serum samples were rapidly thawed at 37°C, and to each sample an equal volume of 1× PBS containing 5 mg of heat-inactivated bovine serum albumin (Sigma) per ml was added. Samples were centrifuged at 12,000 × g for 30 min at 4°C. The viral pellets were resuspended and lysed in 800 µl of Tri-reagent (Molecular Research Center, Woodland, Tex.) and purified following the manufacturer's instructions. Quantitative liquid hybridization was carried out by a modification of the method described by Vahey et al. (43) by using SIV gag sequences. The assay employs external standards consisting of cloned templates of a cognate region of SIV_smh4 gag DNA from a plasmid clone, p5'4i (16), obtained from Vannesa Hirsch. SIV gag RNA was amplified by reverse transcriptase-linked PCR. Reverse transcription was carried out in 40 µl of solution containing 7 µl of 5× first-strand buffer (Gibco-BRL), 20 U of RNasin (Promega, Madison, Wis.), 0.4 µl of 25 mM deoxynucleoside triphosphates and 200 U of Superscript RT (Gibco-BRL, Gaithersburg, Md.). Incubation was for 10 min at 45°C. SIV gag DNA was amplified in 100 µl of a solution containing 10 µl of 10× PCR buffer (Perkin Elmer, Norwalk, Conn.), 0.8 µl of 25 mM deoxynucleoside triphosphates (Promega), and 2.5 U of Taq polymerase (Perkin Elmer) with 1.0 mM 5' primer, 5'-CCCGCAGTAAAGAATTGGATGAC-3', and 1.0 mM 3' primer, 5'-ACTTCCAGCAGCCCTGTCTTCT-3'. Amplification was for 24 cycles of 95°C for 30 s, 50°C for 30 s, and 72°C for 3 min.

Differential nef RNA PCR assay. A differential PCR assay was developed to determine the presence of a full nef sequence. The assay was designed to include a pair of primers that bind outside of the 5' and 3' ends of the deletions described by Kestler et al. (21). Sequences were obtained from the GenBank database. The reaction conditions were identical to those described above. The paired primers were 5'-GATGGATATCTGCAATCCCCAGGAGGATTA-3' for the 5' end and 5'-TAAATCCCTTCCAGTCCCCCCTTTTC-3' for the 3' end. The amplified product corresponds to a 299-bp length for the wild-type nef gene and an approximately 166-bp product for the nef deletion mutants. The amplified products were visualized by the liquid hybridization method described above with an antisense probe, 5'-GGGCTTGAGCTCACTCTCTTGTGAGGG-3'.

Virus neutralization assays. Virus-neutralizing titers were determined with cytopathic effect reduction assays. Cytopathic effect reduction titers were determined as described by Montefiori et al. (29, 30) for SIV and SHIV isolates.

Immunoblots. SIV_Mne/E11s (5) was grown in AA2 cells, clarified, and harvested by ultracentrifugation. This virus has significant immunologic cross-reactivity to both SIV₂₃₉ and SIV_PBj antigens. Pelleted virus was subjected to sodium dodecyl sulfate-gel electrophoresis and then transferred to nitrocellulose by the method of Towbin et al. (42). HIV-1 strips were obtained from a commercial source (Cambridge Biotech, Rockville, Md.). Immunoblots were probed with macaque plasma diluted to 1:100. Immune complexes were detected by a goat anti-human IgG (Sigma) by the enhanced chemiluminescence system (Amersham, Arlington Heights, Ill.) with Kodak XAR-5 X-ray film (Kodak, Rochester, N.Y.).

	RESULTS

Top Abstract Introduction Materials and methods Results Discussion References

Virus isolation and immune response following SIV_nef immunization. Two nef-deleted viruses, SIV_239nef and SIV_PBj6.6nef, were used to immunize two different macaque species, rhesus and pigtailed, respectively. In each virus, the nef gene was altered by deletion of a portion coding only for the nef protein. Within 1 week following an intravenous inoculation, virus was isolated from all the inoculated animals (Table 1). The rhesus macaques receiving the SIV_239nef stock had moderate to high levels of virus as assessed by numbers of positive cultures (Table 1) and circulating virus RNA levels (Fig. 1). Three of five animals (348, 354, and 356) remained culture positive throughout the period of immunization (more than 80 weeks), while the remaining two animals (343 and 344) were culture positive intermittently during the immunization period. Of the three animals that remained culture positive throughout the immunization period, all were positive for circulating SIV gag RNA at the time of the challenge, and the number of copies ranged from 10² to 10³ per ml (Table 2). These results were in contrast to those for the pigtailed macaques immunized with the SIV_PBj6.6nef, of which only one animal had any positive isolations after 20 weeks, three were isolation positive only during the first 4 weeks of the immunization period, and none had detectable SIV RNA in their plasma at any time during the immunization period.

View this table: [in this window] [in a new window]   TABLE 1.   Virus isolation following immunization with SIVnef

View larger version (18K): [in this window] [in a new window]   FIG. 1.   Plasma RNA levels following immunization with SIV239nef. The lower limit of detection for the assay was 200 copies/ml. The values on the y axis are expressed as log10. Rhesus macaques infected with the pathogenic molecular clone SIV239 produce 105 to 107 RNA copies/ml within 3 weeks of intravenous challenge.

PBj6.6nef

239nef

239nef

PBj6.6nef

View larger version (29K): [in this window] [in a new window]   FIG. 2.   Antibody responses following immunization with attenuated virus and after challenge with SHIV89.6PD. Levels of antibody are expressed as kinetic units (103 OD units/min) on the y axis for each individual animal. Solid symbols are data for animals immunized with attenuated virus. A, responses of animals immunized with the SIVPBj6.6nef isolate; B, responses of animals immunized with SIV239nef. Open symbols are data for control animals. The break in the x axis demarcates the period prior to challenge (to the left) and that following challenge (to the right), at week 92 p.i. for panel A and at week 86 for panel B.

View larger version (44K): [in this window] [in a new window]   FIG. 3.   Immunoblot analysis of serum collected from attenuated-virus-immunized animals on the day of challenge. Reactivities to SIVmne/e11s-derived antigens (A) and HIV-1IIIB-derived antigens (B) are shown.

Virus isolation and immune response after SIV_89.6PD challenge. SHIV_89.6PD grows to high titers in vivo in macaques following intravenous challenge, causing a rapid decline in detectable number of circulating CD4⁺ cells in naive macaques (34). We challenged the immunized macaques with 25 TCID₅₀ of the SHIV_89.6PD stock via the intravenous route. These challenges were performed in the immunized rhesus and pigtailed macaques at 86 and 92 weeks, respectively. Two control animals per immunization group were also inoculated in a similar fashion. All animals were monitored for virus isolation (Table 4) and circulating CD4⁺ cell loss (Fig. 4). As expected, the controls showed a rapid decline in circulating CD4⁺ cells by 2 weeks p.i. (Fig. 4). This rapid decline was absent in all of the vaccinated animals during the early follow-up period, even in those confirmed to be infected with the SHIV_89.6PD. An inguinal lymph node biopsy specimen was collected at 2 weeks postchallenge from each animal, and following the preparation of a single-cell suspension of the tissues, virus isolation and lymphocyte subpopulation determinations were performed (Table 5). The results confirmed the findings for the peripheral blood, except that one animal, animal 2581, had SHIV isolated from the inguinal lymph node biopsy specimen collected at 2 weeks but was never found to show detectable circulating infected PBMC. The biopsy specimens from immunized animals showed moderate or no change in % CD4 or CD4:CD8 ratio from prechallenge levels in lymph nodes. The controls had a near-total loss of detectable CD4⁺ cells in the lymph nodes by 2 weeks p.i. During the follow-up period, the infected control rhesus macaques had a slow rebound in CD4⁺ cell numbers (Fig. 4) and each developed moderate to high levels of SIV antibody (Fig. 2). This was in contrast to the control pigtailed macaques that had no recovery of CD4⁺ cells and developed few or no SIV-specific antibodies.

View this table: [in this window] [in a new window]   TABLE 4.   Isolation of virus from SIVnef-immunized animals following SHIV89.6PD challenge

View larger version (32K): [in this window] [in a new window]   FIG. 4.   Levels of circulating CD4+ lymphocytes in peripheral blood during the first 40 weeks following SHIV challenge. Solid symbols are data for vaccinated animals, and open symbols are data for control animals. A, levels for animals immunized with the SIVPBj6.6nef isolate; B, levels for animals immunized with SIV239nef.

239nef

PBj6.6nef

SIV RNA in plasma. Circulating SIV gag RNA was detected in three of the five SIV_239nef-immunized animals and none of the five SIV_PBj6.6nef-immunized animals at the time of challenge (Table 2 and Fig. 5). Following challenge with the SHIV_89.6PD, all three SIV_239nef RNA-positive, immunized rhesus macaques and one SIV_PBj6.6nef-immunized pigtailed macaque had circulating viral RNA (Fig. 5) that contained an undeleted nef gene. The assay results for the remaining animals were all below the level of detection (<200 copies/ml). SIV RNA was easily detected in all of the control animals by 1 week postchallenge, and these levels peaked by 2 to 4 weeks. The controls for both groups had high levels of circulating RNA (>10⁶ RNA copies/ml) (Fig. 5). The overall levels were observed to be higher in the naive pigtailed macaques than in the rhesus macaques. In addition, the pigtailed macaques developed a higher sustained level of RNA, whereas by 16 weeks postchallenge the levels for both control rhesus macaques were at or near the lower limit of detection of the assay.

View larger version (22K): [in this window] [in a new window]   FIG. 5.   Early plasma virus RNA levels following challenge with SHIV89.6PD. The lower limit of detection for the assay was 200 copies/ml. Solid symbols are data for vaccinated animals, and open symbols are data for control animals. A, levels for animals immunized with the SIVPBj6.6nef isolate; B, levels for animals immunized with SIV239nef. The values on the y axis are expressed as log10.

	DISCUSSION

Top Abstract Introduction Materials and methods Results Discussion References

This study was initially designed to determine the breadth and limits of protection associated with the immunity developed following immunization with attenuated lentiviruses. To achieve this we chose to challenge with a pathogenic SHIV that contained an HIV-1 envelope and the SIV₂₃₉ gag-pol. This SHIV isolate contains the envelope gene from the HIV-1_89.6 isolate, which is classified as a primary isolate with dual tropism for both lymphocytes and monocytes as defined by recent studies by Rucker et al. (36) showing that it can bind the CCR2b, CCR3, CCR5, and CXCR4 coreceptors. Following in vivo passage a highly pathogenic SHIV, SHIV_89.6PD, was isolated (34, 35). The choice of a SHIV as a challenge virus for our study was made due to the recent finding that attenuated-virus immunization could very effectively protect animals from infection with homologous strains of SIV and nonpathogenic SHIV without any obvious correlate of protection (2, 9, 13, 27, 31, 37, 38, 47). Additional studies involving passive transfer of antibodies generated following immunization and the removal of CD8⁺ lymphocytes by monoclonal antibodies have also been inconclusive as to their protective values (3, 39). Currently no definitive correlation has been made as to the means of protection, leading some investigators to speculate that the protection may be due to nonspecific immunity or to viral interference (31).

Our goal for this study was to determine if attenuated virus could protect against a pathogenic virus containing a highly diverse envelope and, if protection was observed, whether it could be attributed to specific immunity. In addition, we were interested in determining if animals that were not protected from infection had an altered disease course. We chose to use two attenuated isolates, SIV_239nef and SIV_PBj6.6nef, to compare differences in growth and immunizing potential. The SHIV challenge virus contains the gag-pol portion of the SIV₂₃₉ isolate, so that the SIV_239nef is totally homologous in sequence to the gag-pol portion of the challenge virus, while the SIV_PBj isolate varies up to 10% in the gag-pol region. The findings of this study demonstrate that the immunizing attenuated viruses stimulated effective broadly protective responses in some animals. These responses do not appear to induce a sterilizing immunity, except in the case of animal 2580, but appear to be very effective at blocking the acute pathogenicity of the SHIV_89.6PD. The mechanism of protection is unlikely to be associated with specific envelope antibodies since the SHIV envelope is highly diverse and neutralizing antibodies were absent at the time of challenge. This corroborates the findings of Almond et al. (3), which showed that passive antibody treatments with immune serum from attenuated-virus-immunized animals did not protect from homologous pathogenic challenge, and of Langlois et al. (23), which showed that sera from attenuated-virus-immunized macaques failed to neutralize the challenge viruses. In addition, the protection does not appear to be associated with a viral receptor interference mechanism (45), since the animals that were replicating virus at a high level at the time of challenge were more susceptible to SHIV infection and disease than others that had little or no documented virus replication. This is emphasized by our finding that the immunizing virus SIV_PBj6.6nef, which had limited growth potential in the immunized macaques, was more protective than the more-efficiently growing SIV_239nef. The protective responses observed may be associated with gag-specific responses or possibly with nonspecific immune responses, such as the presence of -chemokines or interferons. These responses are not sufficient to inhibit infection with the challenge virus but may act to decrease and possibly block its spread.

The results presented indicate that although the immunizing viruses are attenuated in phenotype, they are not totally avirulent in all hosts. Previous studies by Whatmore et al. (46) have shown that nef deletion does not eliminate pathogenicity in macaques. More recently, even a triple-deletion mutant has been shown to cause some pathology in neonates (4, 48). Our results support this finding, in that some immunized animals failed to control the attenuated virus. This is highlighted by the very effective protective responses observed in most SIV_PBj6.6nef-immunized animals as contrasted to the SIV_239nef-immunized group, which appear to be related to the overall growth of the immunizing virus. As was observed, the SIV_PBj6.6nef as the immunizing virus was effectively controlled in all but one animal following the early period of immune development. In addition, the circulating RNA from the SIV_PBj6.6nef was not detected in any animals during the immunization period. These animals very effectively blocked the SHIV infection and/or growth; one in five was completely protected and three in five controlled the SHIV after challenge. The SIV_PBj6.6nef isolate has been used to infect an additional 15 animals, and circulating RNA was detected in only 1 animal, at a level of 10³ RNA copies/ml, during the early infection (data not shown). This indicates that the isolate is significantly more attenuated than its wild-type parent strain, which grows to high levels in pigtailed macaques (24, 32). This is in contrast to the observations that all SIV_239nef-immunized animals had circulating viral RNA following immunization and that three of these five animals were unable to control virus growth during the >80-week period of immunization. Following challenge, all three of these SIV_239nef-immunized animals failed to control the SHIV infection and all died prior to the controls, suggesting an enhanced disease course due to undetected immune system damage caused by the immunizing virus.

The variation of the responses in the two types of macaques, rhesus and pigtailed, towards the challenge virus is of interest. These macaques are commonly used in SIV and/or SHIV research, and the pigtailed macaque has been shown to be more susceptible to most SIV and some HIV-1 isolates than the rhesus macaque (1, 14, 17, 24). These studies indicate that this is also the case for the SHIV_89.6PD isolate. The virus infected both macaque types, and the patterns of early growth and pathogenicity are similar, with rapid virus growth and a rapid loss of circulating and tissue CD4⁺ cells. Overall differences can be observed beginning at the early time points, when higher virus loads are observed in the pigtailed macaques. By a few months postchallenge the differences become quite apparent, with CD4⁺ cells returning, antibodies developing, and virus RNA levels dropping in the control rhesus macaques, while the pigtailed macaques had no apparent CD4⁺ cell rebound or humoral response toward the virus. The control pigtailed macaques regained few, if any, circulating CD4⁺ cells and generated only minimal levels of SHIV antibodies, and both controls died during the study period. This indicates that the SHIV_89.6PD isolate is more pathogenic in pigtailed macaques and makes the observed protective responses induced by the SIV_PBj6.6nef even more significant.

Use of attenuated HIV constructs has recently been proposed for inclusion in human clinical trials (8). The specific constructs would have deletions additional to those in the viruses reported here. Each additional mutation has been shown to cause the virus to grow less efficiently and to induce a diminished immune response (13). Our findings suggest that a window of opportunity exists for a safe and effective attenuated-lentivirus vaccine to work. This window is limited first by the potential of the immunizing virus to cause damage to the immune system and second by the level of the immunity generated following vaccination. The width of this opening is currently unknown. Our findings indicate that single-deletion mutants should not be used, due to the potential for the immunizing virus to cause immune system damage and the potential for enhanced disease upon contact with a pathogenic virus. The findings of Baba et al. (4) and Wyand et al. (47) indicate that the deletions of nef, nre, and vpr do not eliminate the potential safety problems in neonatal macaques, and this phenomenon could also apply to immunosuppressed or immunocompromized individuals. Desrosiers et al. (13) have developed mutants with four and five deletions, each having the expected loss in growth potential in vitro. When a four-deletion mutant was used in vivo, it generated low-level infections with low or nondetectable immune responses (13). Protection studies involving the four-deletion mutant have not been reported in macaques to date, but the limited immune response developed suggests that it will not generate sufficiently high levels of protection. The five-deletion mutant failed to attain measurable growth in vitro. These studies indicate that additional methods of attenuation or combinations of deletions should be tested and that their potential use in developing vaccine candidates should be scrutinized for both safety and generated immunity in animal models prior to their use in humans.