INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The overwhelming majority of human viral vaccines used clinically consist of either inactivated whole virus particles or live attenuated viruses. Attenuated viruses have proven to be the most effective vaccines for humans. Additionally, macaques infected with simian immunodeficiency virus (SIV) strains attenuated for in vivo virulence by deletion of the nef gene or other regulatory sequences have been shown previously to be protected from challenge with pathogenic SIV (1, 5, 12, 16-21, 30, 32, 38-40, 46, 49, 50, 55, 56, 59). Indeed, attenuated SIV strains are generally accepted as being among the most effective vaccines evaluated to date in nonhuman primate models (34). However, a major concern with attenuated virus vaccines is safety. This is especially true for viruses such as human immunodeficiency virus (HIV) or SIV that have a high mutation rate and integrate into the host's genome and for which the outcome of a pathogenic infection is potentially lethal. With the advent of recent techniques that allow in vivo expression of antigens from a DNA construct (15, 61) and information regarding the expression of nucleocapsid (NC) mutant virions from proviral constructs (23, 29), it has become possible, in principle, to duplicate the steps and immunological exposure of infection with an attenuated virus but without the associated risks of a replicating virus (42, 54).

We have previously shown that cells transfected with retroviral NC mutant proviral DNA expressed viral proteins and assembled as budded, morphologically authentic viral particles that had the full complement of properly processed viral proteins but had RNA levels reduced by as much as 97% compared to those of wild-type virus. These mutant virus particles are replication defective (at least 10⁵-fold less infectious than a comparable level of wild-type virus [23-25, 29]) and bind to target cells and induce CD4-gp120^SU-dependent "fusion from without" (2, 53). Virion particles produced from these constructs effectively incorporate many of the immunologically relevant steps of the viral life cycle including particle assembly, budding from the cell (24, 25, 28, 29), attachment to receptors, and conformational changes induced upon receptor binding leading to membrane fusion. This approach may be particularly important in view of recent results suggesting that exposure of transition epitopes induced by conformational changes upon interaction of HIV envelope glycoproteins with receptors may facilitate the development of broadly neutralizing antibody responses (42).

In this report, we describe the results of vaccine experiments in which we tested the premise that virions from such a DNA construct might represent a useful vaccine immunogen. Macaques were immunized with DNA constructs encoding an NC mutant virus, strain SIV(Mne). In view of the precedent for superior responses seen to DNA-priming-soluble-protein-boosting immunization regimens, an SIV protein boost was also administered in a subset of animals to examine any additional immunological benefits. Animals were subsequently challenged intravenously with pathogenic homologous virus. Virologic, immunologic, and clinical parameters were monitored in immunized and control (nonimmunized) macaques to assess vaccine efficacy.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Plasmids and mutagenesis. SIV(Mne), originally isolated from a pig-tailed macaque (Macaca nemestrina) with lymphoma, was cultured in HuT 78 cells (6). A single cell clone of infected HuT 78 cells, E11S, was used as the source of DNA to obtain the full-length proviral clone, SIV(Mne) clone 8 (7, 33). Clone 8 is flanked on both the 3' and 5' ends by approximately 1 kbp of uncharacterized cellular DNA from the HuT 78 cells. Each flanking region contains a SalI restriction endonuclease site; the provirus with the flanking DNA was cloned into the SalI site of pSVori/neo, described previously (23). Site-directed mutagenesis was performed to create the NCZF2 mutant (previously identified as Cys 33-Cys 36; pRB130) as described by Gorelick and coworkers (23). The mutation in NCZF2 consists of a 12-nucleotide deletion in the gene coding for the NC protein [nucleotide positions 1772 to 1783 of the SIV(Mne) sequence (GenBank accession no. M32741) were deleted]. The nucleotide region removed corresponds to the amino acids Cys-Trp-Lys-Cys; these are the first four amino acids of the second (carboxy-terminal) NC Zn²⁺ finger. The wild-type SIV(Mne) plasmid clone pRB86 (23) was used for comparative purposes.

View larger version (14K): [in this window] [in a new window]   FIG. 1.   Comparison of the NCZF2 and S8-NCZF2 sequences. The NCZF2 is the proviral DNA clone used in both study 1 and study 2. This clone contains the NC deletion mutation, and the LTR regions are shown. Restriction endonuclease sites used for cloning and construction of the S8-NCZF2 are noted at the top of the figure. The S8-NCZF2 used in study 2 has a portion of the 5' U3 region replaced with the cytomegalovirus enhancer from pCEP4 (Invitrogen Corp.). Additionally, a portion of the 3' U3 region and all of the 3' R and U5 regions have been replaced with the SV40 poly(A) signal.

Cell lines. Human 293 cells that express the SV40 large T antigen (293T) were obtained and cultured as described previously (24) in Dulbecco's modified Eagle medium containing 2 mM L-glutamine and 10% heat-inactivated fetal bovine serum. The cell line AA2-clone 5 is derived from the AA2 cell line (13) and is very sensitive to HIV type 1 and SIV infection. AA2-clone 5 cells were cultured in RPMI 1640 medium containing 2 mM L-glutamine, 10% (vol/vol) heat-inactivated bovine calf serum, and 2 µg of hexadimethrine bromide (Polybrene; Sigma Chemical Co., St. Louis, Mo.) per ml. All cell lines were maintained in an atmosphere of 5 to 7% CO₂ at 37°C.

Transfections. For viral particle analyses, mutant and wild-type viral clones were transfected using the calcium phosphate mammalian cell transfection kit from Eppendorf Scientific, Inc. (Westbury, N.Y.). Log-phase 293T cells, grown in 150-cm² flasks, were transfected, and virus was harvested as described previously (24) by centrifuging cell-free culture fluids at 120,000 × g for 1 h at 4°C in a Beckman SW28.1 rotor (Beckman-Coulter, Inc., Fullerton, Calif.).

Electron microscopy. For transmission electron microscopy, 293T cells transfected with mutant and wild-type virus expressing DNA clones (pRB130 and pRB86, respectively) were fixed 72 h posttransfection with 1.25% (vol/vol) glutaraldehyde in phosphate-buffered saline (PBS). Cell pellets were embedded and processed as described previously (22). For cell-free-virus examinations, particles from the transfections were pelleted at 100,000 × g for 1 h at 4°C through a 20% (wt/vol) sucrose (in PBS) cushion and fixed in 1.25% (vol/vol) glutaraldehyde (in PBS). Embedded cell and viral pellets were sectioned and examined in an Hitachi H-7000 electron microscope operated at 75 kV.

SIV(Mne) gp120^SU-CD4⁺-mediated cell fusion. To determine if the SIV(Mne) NC mutant particles could mediate gp120^SU-CD4⁺-dependent fusion, CEMx174 cells were incubated with wild-type or mutant SIV(Mne) in a fusion-from-without assay, as previously described (2, 53). The presence of characteristic syncytia was evaluated by inverted phase-contrast microscopy 6 h following the addition of virus. The appearance of syncytia at this time reflects fusion from without, mediated between target cells, as this is insufficient time for expression of adequate levels of envelope glycoproteins as a consequence of newly synthesized virus from infection. The CD4⁺ dependence of fusion was determined by preincubating target cells with Leu 3a antibody (25 µg/ml; Becton Dickinson, Franklin Lakes, N.J.) for 15 min at 4°C prior to addition of virus. Virus was obtained by transfecting NCZF2 (pRB130 [23]), S8-NCZF2, or wild-type (pRB86 [23]) SIV(Mne) DNA into 293T cells and was purified on an Iodixanol gradient according to the manufacturer's recommendations (Nycomed Pharma, Oslo, Norway).

Animals. Pig-tailed macaques (M. nemestrina) were maintained in stainless steel cages in biological safety level 3 facilities. Sedation prior to immunizations, virus inoculations, or venipuncture was performed using Telazol (Elkins-Sinn, Inc., Cherry Hill, N.J.; 0.03 mg/kg of body weight administered intramuscularly [i.m.]). Animals were housed in pairs for study 1 at the Washington Regional Primate Research Center (WRPRC; Seattle, Wash.) and individually housed in study 2 at the National Institutes of Health primate facility (Bethesda, Md.). Both facilities are accredited by the American Association for Accreditation of Laboratory Animal Care. Animal care was provided in accordance with the procedures outlined in the "Guide for the Care and Use of Laboratory Animals" (National Institutes of Health publication no. 86-23, 1985).

Immunizations. For study 1, nine male juvenile M. nemestrina animals (ranging from 1.7 to 2.2 years in age) were randomly assigned to two groups. Prior to each DNA inoculation, each animal was injected with bupivacaine-HCl in the quadriceps muscle as described previously (62). The site of bupivacaine-HCl inoculation was marked, and the DNA was injected at this exact site 24 h later. Five animals (group X) were inoculated with 100-µg or 500-µg injections of NCZF2 DNA (Fig. 1) as summarized in Table 1. Four control animals (group Z) received escalating doses of the plasmid DNA vector, lacking SIV sequences.

View this table: [in this window] [in a new window]   TABLE 1.   SIV(Mne) NCZF2 and pSVori/neo control DNA immunization, boost, and challenge schedule (study 1)

View this table: [in this window] [in a new window]   TABLE 2.   SIV(Mne) NCZF2, S8-NCZF2, and pSVori/neo-pCEP4 control DNA immunization and challenge schedule (study 2)a

SIV protein antigen boosts. SIV(Mne) Gag-Pol particles suspended in PBS and lentil lectin-purified gp160^Env (lot 3392-126, pool III; also in PBS) were produced in BSC-40 (African green monkey) cells and were obtained from Shiu-Lok Hu (WRPRC). The SIV Gag-Pol particles produced by a vaccinia virus vector (48) contained 500 µg of p28^CA per dose, were mixed with incomplete Freund's adjuvant, and were delivered into the right thigh. The SIV(Mne) gp160^Env was lentil lectin purified, mixed with incomplete Freund's adjuvant, and delivered into the left thigh (100 µg of gp160^Env per dose). Both formulations were mixed at a 1:1 (vol/vol) ratio of protein to adjuvant. Five animals (J93138, F93264, F93255, F93253, and F93121) from both the vaccinated and control groups (indicated in Table 1) were boosted i.m. with a mixture containing the Gag-Pol particles and gp160^Env proteins. The remaining four animals were injected at the same sites with PBS and incomplete Freund's adjuvant (1:1 PBS/adjuvant ratio).

Lysed SIV ELISA. Sucrose density gradient-purified SIV(Mne) clone E11S grown in HuT 78 cells was lysed with 1% (vol/vol) Triton X-100 and 0.5% (wt/vol) sodium desoxycholate followed by sonication. Following extraction with ethyl ether, the aqueous phase was evaporated to less than one-half of its original volume under a stream of N₂ gas and diluted to approximately 1 mg of protein per ml using sterile deionized water. Solubilized SIV(Mne) was diluted to 5 µg/ml in 0.1 M sodium carbonate buffer (pH 9.6) just prior to use. To each well of a 96-well plate, 100 µl of this viral lysate was added and incubated for 18 h at 4°C. Plates were washed with 0.05% (vol/vol) Tween 20 in PBS (T-PBS buffer) and then blocked by adding 350 µl of sterile 0.25% (wt/vol) gelatin in PBS per well. After washing, 100 µl of diluted plasma samples from the macaques was added to the appropriate wells; dilutions and pipetting of serum samples into the enzyme-linked immunosorbent assay (ELISA) plates were performed using a Packard MultiPROBE model 100 robotic liquid handling system (Packard Instrument Company, Meriden, Conn.). After incubation for 2 h at 37°C, the plates were washed five times with T-PBS buffer. Bound antibody was detected by adding 100 µl of goat anti-human immunoglobulin G (IgG) labeled with alkaline phosphatase (Sigma Chemical Co.) diluted 1:5,000 in T-PBS buffer containing 0.5% wt/vol bovine serum albumin. The assay was developed by incubating 100 µl of substrate (1 mg of p-nitrophenyl phosphate per ml in 0.1 M glycine HCl [pH 10.4], 1 mM MgCl₂, and 1 mM ZnCl₂) per well for 30 min at room temperature in the dark. Colored product was detected at 405 nm using a Vmax plate reader (Molecular Devices Corporation, Sunnyvale, Calif.).

Immunoblotting for the detection of SIV(Mne) antibodies in M. nemestrina plasma. For immunoblot analysis, plasma samples obtained from animals after challenge were diluted as indicated in the figure legends. Plasma samples were reacted with Immobilon-P (Millipore Corp., Bedford, Mass.) filter strips containing transferred, sodium dodecyl sulfate-polyacrylamide gel electrophoresis-separated proteins from disrupted, sucrose gradient-purified SIV(Mne) clone E11S, as previously described (9).

Plasma viral load analysis. Levels of virion-associated SIV RNA in plasma samples were measured using a quantitative real-time reverse transcriptase PCR (RT-PCR) technique, as described previously (58). The analytical sensitivity of this assay is typically ~300 copy eq of SIV gag sequence per reaction, with interassay variation of <25% (coefficient of variation). The clinical sensitivity of the assay depends on the volume of source specimen analyzed.

Virus isolation assays. For virus isolation, 4 × 10⁶ peripheral blood mononuclear cells (PBMC) or lymph node cells isolated from the pig-tailed macaques were incubated with 5 × 10⁶ AA2-clone 5 cells as described previously (8). RT assays were performed weekly to monitor cultures for the presence of virus (9).

Detection of proviral DNA in PBMC and lymph node samples. Nested PCR was performed as described previously using 1 µg of total DNA obtained from Histopaque-1077 (Sigma Chemical Co.)-isolated PBMC or lymph node cells (37). The limit of detection of this particular assay is 3 to 30 DNA copies (37).

Neutralizing antibody assays. Serum neutralizing antibody titers were assayed as described previously (37) except that SIV(Mne) E11S virus was incubated with plasma or serum samples at 37°C for 2 h instead of at 4°C for 16 h.

Lymphocyte proliferation assays. Proliferation assays of study 1 samples were performed essentially as described by Clerici and coworkers (14) using, in addition to the Env peptides described previously (15), SIV(Mne) gp120^SU, p28^CA, p16^MA, and p8^NC proteins (AIDS Vaccine Program, National Cancer Institute [NCI]-Frederick Cancer Research and Development Center, Frederick, Md. [10, 35]). For study 2, assays were performed as described by Lifson and coworkers (45).

Virus challenge. The challenge virus stock, produced in HuT 78 cells, was SIV(Mne) clone E11S, lot no. B1818. The virus was diluted into RPMI 1640 medium containing 10% (vol/vol) fetal bovine serum, and 3 ml of a 10⁴ dilution of the stock (representing 20 animal infectious doses [AID] of virus) was administered by intravenous infusion.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Characterization of SIV(Mne) NC mutant and wild-type particles. We have previously described the characteristics of a number of SIV(Mne) NC mutant viruses (23), including NCZF2 (identified originally as Cys 33-Cys 36), one of the clones used in the present study. The mutant plasmid constructs produce virions that are replication defective. Transmission electron micrographs of the particles resulting from transfection of 293T cells with the NCZF2, S8-NCZF2, and wild-type SIV(Mne) DNA constructs are shown in Fig. 2. Wild-type virus preparations contain uniformly sized particles, approximately 100 nm in diameter, and both mature and immature forms are evident. The SIV NC mutants also contain mature and immature forms, although the ratio of immature to mature particles was greater in the mutant than in wild-type preparations. Well-defined spikes were easily visualized on the surface of both the mutant and wild-type particles (Fig. 2).

View larger version (112K): [in this window] [in a new window]   FIG. 2.   Electron micrograph analysis of mutant and wild-type SIV(Mne). For virus particle analysis, 293T cells were transfected with NCZF2, S8-NCZF2, or SIV(Mne) wild-type DNA as described in Materials and Methods. Cell or virus pellets were fixed in 1.25% (vol/vol) glutaraldehyde, and samples were examined. Arrows indicate particles with a mature morphology. Particles were visualized at ×90,000. A 100-nm size marker is indicated in the upper left-hand panel.

CD4⁺-dependent cell fusion. To assess the ability of NC mutant viruses to mediate binding and fusion of CD4⁺ cells, NCZF2 and S8-NCZF2 mutant particles were compared to wild type in a virion-induced fusion-from-without assay that has been described previously (2, 53). Similar syncytium formation was evident within hours of addition of virus to target cells when NCZF2, S8-NCZF2, or wild-type SIV(Mne) particles were incubated with CEMx174 cells (Fig. 3). Fusion events were readily blocked by Leu 3a (anti-CD4) antibody.

View larger version (96K): [in this window] [in a new window]   FIG. 3.   Fusion of CEMx174 cells with mutant and wild-type SIV(Mne). CEMx174 cells were pretreated with either PBS or Leu 3a (anti-CD4) antibody (Becton Dickinson; 25 µg/ml) and then exposed to a 100× concentration of wild-type or no virus (A to D). For panels E and F, CEMx174 cells were pretreated with PBS and then incubated with a 100× concentration of NCZF2 and S8-NCZF2, respectively. All virus concentrates had similar levels of RT activity. Cells were incubated for 6 h at 37°C with virus and observed at an ×100 magnification. Syncytia are indicated by the arrows.

Immunization study design. The overall experimental design involved immunization of M. nemestrina with the DNA constructs and proteins described and monitoring of immune responses, followed by intravenous challenge with a pathogenic strain of SIV. Following challenge, the animals were monitored virologically, immunologically, and clinically to assess the possible protective efficacy of vaccination. For study 1, follow-up was for more than 2 years postchallenge and is continuing informally for surviving animals. For study 2, the postchallenge follow-up period is 46 weeks. Data from studies 1 and 2 are presented and compared.

Prechallenge immune responses. Humoral immune responses were examined by ELISA, immunoblotting, viral neutralization assays, and other analyses. In study 1, at week 45 after the initial DNA immunization (prior to the SIV[Mne] gp160^Env protein and Gag-Pol particle boost), two animals (F93255 and F93271) had weak but detectable antibodies to SIV based on the lysed SIV ELISA (Fig. 4). In study 2, only one animal (92P005 immunized with S8-NCZF2) showed measurable antibodies to SIV in the lysed SIV ELISA at week 30 after the first DNA immunization (Fig. 4). These antibody responses provide indirect evidence for in vivo expression of viral proteins from the immunizing DNA constructs NCZF2 and S8-NCZF2 in these animals. Antibody responses were not detected by ELISA in the other eight SIV DNA-exposed animals or in any of the six control animals from either study (data not shown).

View larger version (13K): [in this window] [in a new window]   FIG. 4.   Lysed SIV ELISA. Serial dilutions of plasma samples taken prior to and after the last DNA injection were analyzed. Plasma samples were added to 96-well plates that were coated with detergent-disrupted SIV(Mne) clone E11S, grown in HuT 78 cells as described in Materials and Methods. Duplicate samples were measured, and error bars, depicting 1 standard deviation from the mean, are shown for each dilution. After washing, bound antibodies were detected colorimetrically following the addition of goat anti-human IgG labeled with alkaline phosphatase and p-nitrophenyl phosphate substrate. Absorbance of colored products at 450 to 650 nm was monitored using a Vmax plate reader (Molecular Devices Corporation). Only animals that were positive for antibodies are presented; all other immunized animals had no significant increase in antibody responses after the DNA immunizations compared to preimmunization responses. Antibody responses were clearly detected in two of the macaques inoculated with NCZF2 DNA (F93255 and F93271, study 1) and one animal injected with the S8-NCZF2 DNA (92P005, study 2). All of the vector control DNA-immunized monkeys were antibody negative as expected. , plasma from animals prior to NCZF2 or S8-NCZF2 DNA immunization (preimmunized samples); , plasma from NCZF2/S8-NCZF2 DNA-immunized animals (week 45 and 26 after the initial DNA immunization for studies 1 and 2, respectively). The antibody-positive animal (J93255) that was subsequently boosted with Gag-Pol particles and gp160Env protein prior to challenge is indicated with an asterisk.

View larger version (32K): [in this window] [in a new window]   FIG. 5.   Immunoblot analyses of pre- and post-protein boost serum from vaccinated and control animals in study 1. Serum samples from DNA-immunized and control animals were tested for SIV-specific humoral responses on the day of the Gag-Pol particle and Env protein boosts (week 72 post-DNA immunization), and at weeks 1 and 2 (weeks 73 and 74 post-DNA immunization, respectively) following the protein boost. The serum was tested at a 1:30 dilution. Antibodies to Env proteins (gp120SU and gp32TM) and to Gag proteins (p28CA and p16MA) are evident. The animals that were not boosted with the proteins did not show any change in their antibody profiles over the 2-week postboost period. The numerals at the top of the Western blot strips indicate weeks post-DNA immunization. Note that animal F93121 had antibodies that reacted with p28CA prior to boosting. This reactivity was observed even prior to the first DNA immunization and has been observed occasionally in animals from other studies (3).

Virologic outcome of challenge with pathogenic SIV(Mne). All macaques in both studies were challenged by intravenous administration of 20 AID of SIV(Mne) E11S (11, 41, 60). Infection was monitored by quantitation of plasma viral RNA levels, seroconversion (for SIV control animals that did not receive any SIV DNA or antigens), virus isolations, and PCR for viral DNA from PBMC and lymph node mononuclear cells.

View larger version (39K): [in this window] [in a new window]   FIG. 6.   Postchallenge plasma RNA levels in vector control, NCZF2, and S8-NCZF2 DNA-vaccinated M. nemestrina macaques. SIV plasma RNA levels are reported as SIV gag RNA equivalents per milliliter of plasma determined by real-time RT-PCR (58). The shaded areas indicate the first 6 weeks postchallenge. After the x-axis break, the remaining follow-up periods are shown. The follow-up periods for studies 1 and 2 are 106 and 46 weeks postchallenge, respectively. Closed symbols and open symbols indicate plasma viral RNA levels that are above and below the limits of detection, respectively. The threshold sensitivity of the analyses varies due to the specimen volume analyzed. (A) Compilation of data for the control DNA-immunized animals from study 1 as follows: F93253* (diamonds), F93121* (squares), F93254 (triangles), and J93042 (circles). Animals F93121*, F93254, and F93253* died from AIDS at weeks 61, 102, and 111 postchallenge, respectively. (B) Compilation of data for the two control DNA-immunized animals from study 2, 92P002 (diamonds) and 92P011 (squares). (C) Compilation of data for the NCZF2 DNA-immunized animals from study 1, designated as follows: J93138* (diamonds), F93264* (squares), F93255* (triangles), J93044 (circles), and F93271 (inverted triangles). Animal F93271 died from AIDS at 63 weeks postchallenge. (D) Compilation of plasma viral RNA loads of NCZF2 and S8-NCZF2 DNA-immunized animals from study 2, designated as follows: NCZF2 DNA-immunized animals, 92P010 (diamonds), 92P013 (squares), and 92P014 (triangles); S8-NCZF2 DNA-immunized animals, 92P004 (circles), 92P005 (inverted triangles), and 93P004 (stars). The asterisks next to the animal identifiers indicate those that were boosted with the SIV(Mne) gp160Env protein and Gag-Pol particles (see Materials and Methods).

View this table: [in this window] [in a new window]   TABLE 6.   Detection of SIV-infected cells via cocultivation with AA2-clone 5 cells and nested PCR, from NCZF2-vaccinated, SIV (Mne) E11S-challenged M. nemestrina macaques (study 1)a

View this table: [in this window] [in a new window]   TABLE 7.   Detection of SIV-infected cells via cocultivation with AA2-clone 5 cells and nested PCR, from NCZF2-vaccinated, SIV(Mne) E11S-challenged M. nemestrina macaques (study 2)a

View larger version (37K): [in this window] [in a new window]   FIG. 7.   Postchallenge CD4+ levels in vector control, NCZF2, and S8-NCZF2 DNA-immunized M. nemestrina macaques. Levels are reported as CD4+ cells per microliter of blood. The follow-up periods for studies 1 and 2 are 106 and 44 weeks postchallenge, respectively. The shaded area indicates CD4+ levels less than 200 cells per µl of blood, and AIDS in this work is defined as a condition where an animal has less than 200 cells per µl of blood over two or more consecutive samplings. Closed symbols indicate CD4+ levels greater than 200 cells per µl of blood, and open symbols indicate levels less than 200 cells per µl of blood. (A) Control DNA-immunized animals from study 1 are as follows: F93253* (diamonds), F93121* (squares), F93254 (triangles), and J93042 (circles). Animals F93121*, F93254, and F93253* died from AIDS at weeks 61, 102, and 111 postchallenge, respectively. (B) Compilation of data for the two control DNA-immunized animals from study 2, 92P002 (diamonds) and 92P011 (squares). (C) Compilation of data for the NCZF2 DNA-immunized animals from study 1, designated as follows: J93138* (diamonds), F93264* (squares), F93255* (triangles), J93044 (circles), and F93271 (inverted triangles). Animal F93271 died from AIDS at 63 weeks postchallenge. (D) Compilation of CD4+ levels from NCZF2 and S8-NCZF2 DNA-immunized animals in study 2, designated as follows: NCZF2 DNA-immunized animals, 92P010 (diamonds), 92P013 (squares), and 92P014 (triangles); S8-NCZF2 DNA-immunized animals, 92P004 (circles), 92P005 (inverted triangles), and 93P004 (stars). The asterisks next to the animal identifiers indicate those that were boosted with the SIV(Mne) gp160Env protein and Gag-Pol particles (see Materials and Methods).

Postchallenge antibody responses. Postchallenge humoral immune responses were examined by the immunoblot analyses as shown in Fig. 8. Four of six animals immunized with control DNA (Fig. 8A and C) exhibited a typical antibody response to SIV(Mne) infection with progressively greater titers to Gag and Env proteins as infection progressed (the exceptions were F93042 and 92P011). Macaques that were inoculated with control DNA and boosted with SIV(Mne) Env protein and Gag-Pol particles (Fig. 5) had positive antibody responses to p28^CA and showed minor (if any) responses to gp120^SU on the day of challenge as expected (Fig. 8A). Animals F93271 (Fig. 8B) and 92P005 (Fig. 8C), which showed evidence of the most extensive SIV infection postchallenge, were the only NCZF2 DNA-vaccinated animals that had immunoblot patterns typical of an SIV infection. The immunoblot of plasma from the other nine NCZF2 DNA-immunized animals showed a variable pattern of SIV antibody responses. For example, the antibody patterns shown by animals F93255, J93044, 92P014, and 93P004 were distinctly different from the pattern seen for the SIV-infected control DNA macaques; plasma from J93044 exhibited an immune response only to gp120^SU through 106 weeks postchallenge, and there was no detectable antibody response at all in animal 93P004 throughout the 46-week postchallenge period.

View larger version (68K): [in this window] [in a new window]   FIG. 8.   Immunoblot analyses of postchallenge serum from vaccinated and control animals. Immunoblot analyses of SIV-specific antibody responses in macaques challenged with SIV(Mne) clone E11S are shown. (A and B) For study 1 (106 week postchallenge follow-up period), data for the control plasmid (pSVori/neo)-immunized macaques (A) and antibodies present in NC mutant (NCZF2) DNA-immunized animals after challenge (B). (C) Data for the macaques in study 2 (46 week postchallenge follow-up period). The various groups immunized with either NCZF2, S8-NCZF2, or control () DNA are indicated at the top of the panel. Pig-tailed macaque plasma (diluted 100-fold) was reacted with sodium dodecyl sulfate-polyacrylamide gel electrophoresis-separated proteins from disrupted, sucrose gradient-purified SIV(Mne) clone E11S as previously described (9). Antibodies to the Env proteins gp120SU and gp32TM; to the Gag proteins p28CA, p16MA, p8NC, and p6Gag; and to the Vpx protein p14 were detected in infected animals at various times after SIV challenge. Antibodies to p28CA and to gp120SU were also evident on the day of challenge (week 0) in some of those macaques. Macaques boosted with Gag-Pol particles and gp160Env protein prior to challenge are indicated with an asterisk. Numerals at the top of the Western blot strips indicate weeks postchallenge.

Clinical outcome of challenge with pathogenic SIV(Mne). M. nemestrina was chosen as the macaque species for these vaccine experiments because these animals develop AIDS (pathogenic SIV disease with CD4⁺ depletion to <200 cells/µl of blood) more rapidly than do other macaque species, after infection with SIV(Mne) E11S. Although comparatively few vaccine protocols have been performed using this species with E11S challenge, 83% (20 out of 24 animals) of M. nemestrina macaques infected with SIV(Mne) E11S develop AIDS within 2 years of intravenous or intrarectal inoculation (R. E. Benveniste, unpublished data). This natural history background allowed prevention of progressive disease to also be used as a criterion for assessment of vaccine efficacy. As expected, three of the four animals inoculated with control DNA (study 1) showed progressive disease within the approximately 2-year follow-up period (Fig. 7A). Animals that developed progressive disease and ultimately succumbed to AIDS had a higher number of virus-infected PBMC (Tables 4 to 7) and higher levels of plasma viral RNA (Fig. 6) than those that remained healthy. In the NCZF2 DNA-vaccinated group, only one animal (F93271) had progressed to AIDS within the 106-week follow-up period. The remaining NCZF2 DNA-vaccinated animals have shown substantial to apparently complete control of plasma virus levels, without depletion of CD4⁺ T cells or clinical evidence of progressive SIV disease or AIDS, although plasma SIV RNA levels appeared to be increasing in animal F93264 toward the end of the follow-up period (Fig. 6C). There has been insufficient time in study 2 to observe progression to disease (i.e., CD4⁺ T-cell depletion).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The experiments described here are based on the use of DNA immunization to produce an engineered mutant virus that is replication incompetent but is nevertheless capable of authentically completing most steps of a single viral life cycle. The underlying premise was that such an immunization would allow the opportunity for exposure of the immune system to a full range of authentic viral determinants, in a manner that mimics the highly effective immune sensitization achieved with live attenuated viruses, but without the risks inherent in using a replication-competent virus. The SIV(Mne) mutant DNA (NCZF2) used in this study contains the entire SIV(Mne) genome except for a deletion of 12 nucleotides that encode the first 4 amino acids in the second (C-terminal) Zn²⁺ finger of the NC protein. Transfection of cells in vitro with this mutant SIV(Mne) DNA construct (Fig. 1) resulted in budding of complete virus particles containing the full complement of viral proteins, including the viral enzymes RT and protease, but which contained reduced levels of genomic RNA and were also replication defective (23, 66). Electron microscopic analyses of particles obtained after transfections showed particles budding from transfected cells that were ultrastructurally indistinguishable from wild-type SIV(Mne) virions (Fig. 2). In addition, these NC mutant SIV(Mne) particles were shown to bind and fuse cells in a CD4-gp120^SU-dependent manner, in effect completing the steps of viral replication from budding to entry into susceptible cells (Fig. 3).

Although DNA immunizations with NCZF2 clones were intended to induce an immune response in macaques similar to the responses provided by attenuated viral vaccines, the measured immune responses resulting from the NCZF2 immunizations in these studies were low to undetectable, suggesting that the injected NCZF2 DNA may have been poorly expressed. Only 3 of 11 immunized animals had an antibody response detectable prior to challenge or protein boosting using a lysed SIV ELISA (Fig. 4). While the high background reactivity in preimmune serum samples complicated the interpretation of results for four animals, even with this assay there was minimal evidence of humoral responses to the immunization. A weak anamnestic response in the NCZF2 DNA-immunized animals boosted with the SIV(Mne) proteins was suggested, based upon evidence presented in Fig. 5 and corroborated by an additional lysed SIV ELISA detecting the presence of IgG and IgM antibodies (data not shown). Please note that the differences in the Western blot intensities obtained for the DNA-immunized and protein-boosted animals from week 74 (Fig. 5) and week 75 (day of challenge, Fig. 8A and B) are due to differences in the serum dilutions between Fig. 5 (diluted 1:30 to obtain high sensitivity) and Fig. 8A and B (tested at a dilution of 1:100) and also due to different lots of immunoblot strips used for Fig. 5 and 8A and B. There was no clear evidence of serum neutralizing antibody responses or lymphocyte proliferation in response to incubation with viral antigens in either study (data not shown). Cellular cytotoxicity (cytotoxic T-lymphocyte [CTL]) assays were not performed because there was no CTL assay for M. nemestrina available to us when these experiments were initiated.

By design, we performed these studies in a model involving challenge of immunized macaques with a pathogenic SIV strain so that we would be able not only to evaluate vaccine-induced protection from infection but also to evaluate possible vaccine-induced modulation of infection and clinical course, in the event that infection was not prevented (43). Consistent with the cumulative experience of infecting M. nemestrina with SIV(Mne) E11S, three of the four control animals in the long-term study (study 1) developed progressive SIV infection, CD4⁺ depletion, and AIDS during the follow-up period. The one control animal that did not develop AIDS (J93042) has maintained low to undetectable plasma SIV RNA levels and CD4⁺ cell numbers usually greater than 1,000 CD4⁺ cells/µl of blood for over 3 years postchallenge (Fig. 7A). Among the animals immunized with NCZF2 DNA, only one developed SIV disease and AIDS within the formal 106-week follow-up interval (Fig. 7C; F93271). The other NCZF2-immunized animals showed low to undetectable plasma viral RNA levels, without evidence of sustained depletion of CD4⁺ T cells, although SIV RNA levels were increasing in animal F93264 (Fig. 6C) toward the end of the formal follow-up period.

Boosting animals one time with the Gag-Pol particles and Env protein did in fact induce a weak humoral response in study 1 (Fig. 5). However, although the number of animals that were boosted was small, there was no apparent protective advantage afforded by the protein boost. An NCZF2 DNA-immunized animal (F93264) which was boosted with the protein mixture had increased viral loads later in the follow-up period and has since developed AIDS, whereas animal J93044, which did not receive the protein boost, has remained disease free as well as virus free (also Table 3). The Gag-Pol particle and Env protein boost did not appear to increase the protection of the control DNA-immunized animals either since the two animals that were boosted developed high virus loads (Fig. 6A), rapid CD4⁺ cell decline (Fig. 7A), and progressive SIV disease leading to AIDS.

As described previously for various other SIV-infected macaque systems, viral replication patterns, reflected in plasma viral RNA levels, were predictive of progressive disease (43, 44, 57, 64). Of the control animals in study 1, only J93042 showed sustained apparent suppression of viral replication through the postacute phase of infection, with undetectable plasma virus during the second year postchallenge (Fig. 6A) and no depletion of CD4⁺ T cells (Fig. 7A) or clinical evidence of progressive disease (3 years postchallenge), a pattern that is seen in a small subset of SIV-naïve animals infected with SIV(Mne) E11S (R. Benveniste, personal communication). In contrast, the three other control animals (F93121, F93254, and F93253) showed a more typical pattern of high postacute levels of plasma viral RNA in the range of 10⁵ to 10⁸ copy eq/ml (Fig. 6A). This was associated with progressive depletion of CD4⁺ T cells (Fig. 7A) and death from AIDS at 61, 102, and 111 weeks postchallenge, respectively.

A strong relationship between plasma viral load patterns and progressive disease was also seen for the SIV-immunized animals. The one SIV-immunized animal that progressed to AIDS had postacute viral RNA plasma levels of 10⁵ to 10⁶ copy eq/ml (Fig. 6C), developed progressive depletion of CD4⁺ T cells (Fig. 7C), and was euthanatized shortly after the onset of AIDS at 63 weeks postchallenge. Three of the remaining four SIV-immunized animals exhibited low to undetectable plasma viral RNA levels, without evidence of depletion of CD4⁺ T cells, throughout the 106-week follow-up period. The final SIV-immunized animal, F93264, showed modest but readily measurable levels of plasma viral RNA throughout the follow-up period. Since the conclusion of the scheduled follow-up period for data collection, additional informal measurements indicate that this animal has entered a phase of progressive SIV disease, with increasing plasma RNA levels and circulating CD4⁺ T-cell levels of <200 cells/µl of blood (~3 years after challenge). The remaining three immunized animals (J93138, F93255, and J93044) still show no signs of progressive SIV disease at 3 years postchallenge.

Plasma viral RNA load measurements from study 2 were also compiled and compared with the postchallenge results from study 1. Study 2 used a second construct (S8-NCZF2) in addition to NCZF2 that was designed for enhanced safety, based on removal of portions of the LTR regions involved in strand transfer events during reverse transcription (Fig. 1). The S8-NCZF2 construct also made use of cytomegalovirus enhancer elements in an attempt to increase mutant virus expression. As demonstrated above, there were no significant differences between the two constructs and no significant differences in initial peak plasma viral RNA levels between the two studies (Fig. 6 and Table 3).

Study 2 is still in the early stages of follow-up, with observations continuing. However, the robust relationship between plasma viral RNA values and clinical course in study 1 and in numerous other observations (43) suggests that available viral load measurements should be strongly predictive of eventual clinical course and outcome. The lower plasma virus levels in the SIV-vaccinated animals in study 2 should be associated with a prolonged duration of survival without evidence of progressive disease. It is still too early in the follow-up period to observe declines in CD4⁺ cell populations. Importantly, there were significant differences in the initial peak plasma viral RNA levels between the vaccinated and control groups in both studies (Table 3).

In general, the patterns observed between the plasma viral load measurements (Fig. 6) paralleled the virus isolation (cocultivation experiments) and nested PCR results (Tables 4 to 7). There were a few instances where plasma viral RNA was detected in the absence of detectable provirus (e.g., animal J93138 at week 28). Additionally, there were a few cases where plasma viral genomes were present but virus could not be isolated from PBMC. Such occasional apparent nonconcordance of different viral load assay measurements has been observed by other groups and may reflect both biological and analytical variation (40, 47).

Despite this clear evidence of reduced viral loads in immunized animals from both studies and impressive protection from progressive disease in study 1 (Table 3), no measured immunological response to vaccination appears to correlate with protection. This observation of protection without an apparent immunological correlate is hardly unique. Hosie et al. (36) reported that immunization of cats with feline immunodeficiency virus (FIV) DNA containing an in-frame Pol deletion mutation also failed to elicit a detectable humoral response. Immunization with this FIV mutant DNA elicited CTL responses to Gag and Env in the absence of a detectable humoral response. In spite of the lack of a humoral response, roughly one-half of the immunized animals were protected from challenge with infectious FIV, with protection defined as lack of infection and/or reduced virus levels in the vaccinated animals. Although a CTL response was detected in the immunized cats, there was no clear correlation between elicitation of Env- or Gag-specific CTL responses and the protection observed following challenge. The FIV DNA inoculated into the cats represented a full-length viral genome with a 33-codon deletion in Pol. Transfection of this same DNA into susceptible cells results in viral antigen release and Env-mediated cell fusion, suggesting formation of viruslike particles. It is therefore possible that administration of this FIV mutant DNA into cats could result in budding of virion-like particles. Unfortunately, the challenge FIV does not induce disease readily in infected cats, and so protection from disease could not be assessed.

Another recent study used an immunization regimen that included a nearly full-length SIV proviral construct similar to the S8 construct employed in study 2 (31). In that study, immunization induced cellular and humoral responses and was associated with some degree of apparent modulation of viral replication. However, in contrast to the present studies, the challenge virus used was not pathogenic in the macaque species employed (Macaca fascicularis), and so it was not possible to evaluate the effect of immunization on disease course. In another recently reported DNA immunization study, the SIV DNA construct used was similar to the S8 construct utilized in study 2. Wang and coworkers (63) employed SIV(mac239) proviral sequences in their construct and immunized Macaca mulatta by the mucosal route using liposome-mediated DNA delivery methods. Again, results were similar in that vaccination did not prevent infection with SIV(mac239) but did modulate the levels of circulating virus. There was also no apparent immunological correlate of protection from disease in their study (63).

The natural history of SIV(Mne) E11S infection in M. nemestrina is such that plasma viral loads typically are high within the first few weeks after challenge. Viral loads often then diminish to below detectable levels (<300 to 1,000 copies/ml of plasma). Animals unable to control their infections then start to show subsequent increases in virus loads. This viral replication pattern is similar to infection of M. fascicularis with SIV(Mne) E11S, although in this system viral loads usually do not decrease to undetectable levels (31, 48). The transient clearance of measurable virus from the plasma, following peak levels achieved during primary infection, despite the maintenance of cultivable virus in the PBMC compartment throughout, is a unique feature of SIV(Mne) E11S infection in macaques that complicates the interpretation of vaccine studies in which solid protection from challenge is not observed. However, comparison of both initial peak plasma SIV RNA loads and the late follow-up RNA levels in study 1 clearly shows that vaccination modulated the levels of infection. It is interesting to note that the control DNA-immunized animals in study 2 showed high viral loads initially, but they diminished at ~10 weeks postchallenge (Fig. 6B), in contrast to the control animals from study 1 whose virus loads rebounded after week 8 (Fig. 6A). Study 2 is the first study, to our knowledge, that employed adult pig-tailed macaques; thus, differences in the control of virus loads between these two studies may be due to the difference in the ages of the animals among the two groups (1.7 to 2.2 years for study 1 versus 5.6 to 6.6 years for study 2). Similar observations were reported in a previous study that showed increased susceptibility to SIV infection in neonates compared to adult macaques (4).

DNA immunizations offer several potential benefits over conventional vaccines. These range from economical production of the immunogen to stability of the vaccine during transport and delivery at ambient temperature. DNA immunization also appears to be capable of inducing desirable patterns of immune responses (51, 52, 65). The theoretical and practical advantages of DNA vaccines have not escaped the attention of AIDS researchers. There are numerous reports of SIV and HIV DNA vaccines, although most of these studies have used constructs designed to express one or more viral gene products (51, 52, 65), without producing authentically assembled virions, in contrast to the present studies. The prechallenge immunological results suggest that we achieved minimal in vivo expression of viral proteins. Thus, the current results, which demonstrate good protection from progressive disease, likely represent the minimum potential efficacy of this approach. One of the potential advantages of DNA immunization is the ability to achieve sustained expression of the encoded antigen, in theory as long as the episomal immunizing DNA persists. This may be significantly longer than the persistence of a conventional protein immunogen or vaccine antigen encoded by a nonpersisting live vector, such as recombinant poxviruses. If such persistence can be achieved in the context of the strategy proposed in this work, of using DNA immunization with an essentially full-length proviral construct to mimic attenuated retroviral vaccines, then this feature may allow realization of some of the advantages of attenuated viral vaccines.

Using an infectious attenuated retrovirus as a vaccine raises safety concerns because of the high mutation rate of retroviruses and the fact that these viruses can integrate into the host's genome. In addition, attenuated viruses (with deleted accessory genes) have shown the capacity for pathogenesis in neonatal (54) and some juvenile (16) macaques. However, immunization with DNA encoding a replication-defective provirus should be significantly safer than infection using a replication-competent attenuated virus. The available data support this position. Even though the number of animals that we have inoculated with NC mutant proviral DNA is limited, there has been no evidence that the NCZF2 or S8-NCZF2 DNAs resulted in an infection. Supporting the safety of these NC mutant DNA vaccines are the numerous in vitro transfections with similar murine leukemia virus, HIV, and SIV NC mutants without detection of infectious virus. In fact, NC mutants have been generated that package wild-type levels of genome (23, 24, 27) and are still not detectably infectious. The defect for infectivity in NC mutant viruses that package wild-type levels of their genome appears to be postentry but prior to integration. Therefore, mutants in which the Zn²⁺-coordinating residues are deleted not only lack the ability to efficiently package the retroviral genome but also are replication defective due to a defect in an early infection event (26, 27).

The results presented in this work may be considered as an initial proof of the concept that an effective vaccine can be developed by utilizing DNA immunization to mimic the viral replication processes. In view of the observed prechallenge immune responses, it is likely that the construct was poorly expressed in vivo. To enhance expression, we currently have experiments in progress using more efficient promoters and alternative DNA delivery methods. In addition, these constructs may be particularly well suited for use as the priming immunogen in DNA prime-protein boost vaccination regimens, perhaps using conformationally intact whole inactivated virions (2, 53) as the immunogen for the boost phase of immunization. Future studies with such constructs and immunization regimens should allow assessment of the full potential of this approach.