|Home | About | Journals | Submit | Contact Us | Français|
The envelope (Env) glycoproteins of HIV and other lentiviruses possess neutralization and other protective epitopes, yet all attempts to induce protective immunity using Env as the only immunogen have either failed or afforded minimal levels of protection. In a novel prime-boost approach, specific-pathogen-free cats were primed with a plasmid expressing Env of feline immunodeficiency virus (FIV) and feline granulocyte-macrophage colony-stimulating factor and then boosted with their own T lymphocytes transduced ex vivo to produce the same Env and interleukin 15 (3 × 106 to 10 × 106 viable cells/cat). After the boost, the vaccinees developed elevated immune responses, including virus-neutralizing antibodies (NA). Challenge with an ex vivo preparation of FIV readily infected all eight control cats (four mock vaccinated and four naïve) and produced a marked decline in the proportion of peripheral CD4 T cells. In contrast, five of seven vaccinees showed little or no traces of infection, and the remaining two had reduced viral loads and underwent no changes in proportions of CD4 T cells. Interestingly, the viral loads of the vaccinees were inversely correlated to the titers of NA. The findings support the concept that Env is a valuable immunogen but needs to be administered in a way that permits the expression of its full protective potential.
Despite years of intense research, a truly protective AIDS vaccine is far away. Suboptimal immunogenicity, inadequate antigen presentation, and inappropriate immune system activation are believed to have contributed to these disappointing results. However, several lines of evidence suggest that the control or prevention of infection is possible. For example, despite repeated exposures, some individuals escape infection or delay disease progression after being infected (1, 14, 15). Furthermore, passively infused neutralizing antibodies (NA) (28, 42, 51) or endogenously expressed NA derivatives (29) have been shown to provide protection against intravenous simian immunodeficiency virus challenge. On the other hand, data from several vaccine experiments suggest that cellular immunity is an important factor for protection (6, 32). Therefore, while immune protection against human immunodeficiency virus (HIV) and other lentiviruses appears feasible, the strategies for eliciting it remain elusive.
Because of its crucial role in viral replication and infectivity, the HIV envelope (Env) is an attractive immunogen and has been included in nearly all vaccine formulations tested so far (28, 30, 31). Env surface (SU) and transmembrane glycoproteins (gp) are actively targeted by the immune system (9, 10, 47), and Env-specific antibodies and cytotoxic T lymphocytes (CTLs) are produced early in infection. The appearance of these effectors also coincides with the decline of viremia during the acute phase of infection (30, 32). Individuals who control HIV infection in the absence of antiretroviral therapy have Env-specific NA and CTL responses that are effective against a wide spectrum of viral strains (14, 23, 35, 52, 60). At least some of the potentially protective epitopes in Env appear to interact with the cellular receptors during viral entry and are therefore highly conserved among isolates (31, 33, 39, 63). However, these epitopes have complex secondary and tertiary structures and are only transiently exposed by the structural changes that occur during the interaction between Env and its receptors (10, 11, 28). As a consequence, these epitopes are usually concealed from the immune system, and this may explain, at least in part, why Env-based vaccines have failed to show protective efficacy. Indeed, data from previous studies suggested that protection may be most effectively triggered by nascent viral proteins (22, 28, 30, 48, 62).
We have conducted a proof-of-concept study to evaluate whether presenting Env to the immune system in a manner as close as possible to what occurs in the context of a natural infection may confer some protective advantage. The study was carried out with feline immunodeficiency virus (FIV), a lentivirus similar to HIV that establishes persistent infections and causes an AIDS-like disease in domestic cats. As far as it is understood, FIV evades immune surveillance through mechanisms similar to those exploited by HIV, and attempts to develop an effective FIV vaccine have met with difficulties similar to those encountered with AIDS vaccines (25, 37, 66). In particular, attempts to use FIV Env as a protective immunogen have repeatedly failed (13, 38, 58). Here we report the result of one experiment in which specific-pathogen-free (SPF) cats primed with a DNA immunogen encoding FIV Env and feline granulocyte-macrophage colony-stimulating factor (GM-CSF) and boosted with viable, autologous T lymphocytes ex vivo that were transduced to express Env and feline interleukin 15 (IL-15) showed a remarkable level of protection against challenge with ex vivo FIV. Consistent with recent findings indicating the importance of NA in controlling lentiviral infections (1, 59, 63), among the immunological parameters investigated, only the titers of NA correlated inversely with protection. Collectively, the findings support the notion that Env is a valuable vaccine immunogen but needs to be administered in a way that permits the expression of its full protective potential.
Crandell feline kidney fibroblast (CrFK) and human epithelial 293T cell lines were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% heat-inactivated fetal calf serum (FCS), penicillin (100 U/ml), streptomycin (10 μg/ml), and l-glutamine (2 mM) (Sigma-Aldrich, Milan, Italy). The human erythroblastoid TF-1 cell line (ATCC CRL-2003), a human erythroleukemic cell line dependent on several cytokines for growth, including GM-CSF, was cultured in RPMI 1640 medium (Sigma-Aldrich) supplemented as described above plus 4.5 g/liter d-glucose, 1.5 g/liter sodium-bicarbonate, 10 mM HEPES, 1 mM sodium pyruvate, 1% nonessential amino acids (Sigma-Aldrich) (complete RPMI 1640 medium), and 2 ng/ml recombinant human GM-CSF (Schering-Plough, Milan, Italy). MBM cells, a feline T-lymphoid line expressing the FIV CD134 receptor and the CXCR4 coreceptor (46, 65), were cultured in complete RPMI 1640 medium supplemented with 5 μg/ml concanavalin A (ConA) (Sigma-Aldrich) and 20 U/ml recombinant human IL-2 (Roche, Milan, Italy). The murine T-lymphocyte CTLL-2 cell line (ATCC TIB-214, kindly provided by Gregg Dean, North Carolina State University, Raleigh, NC) was maintained in complete RPMI 1640 medium supplemented with 20 U/ml IL-2. FL-4 cells, a line of feline T lymphocytes chronically infected by FIV Petaluma (FIV-Pet) (68), were grown in complete RPMI 1640 medium.
FIV Env was derived from pKKS (Fig. (Fig.1A),1A), a replication-competent molecular clone obtained by the cloning of env of in vivo-readapted FIV-Pet (4, 53) into the p34TF10 backbone (GenBank accession no. NC_001482). For the DNA vaccine, the region encoding Env, Rev, and the Rev-responsive element (RRE) was amplified from pKKS with L-env-S (5′-CGCGGATCCGCCACCATGGCAGAAGGATTTGCAGCCAATAGAC-3′) and stop-Rev-AS (5′-CGGAATTCCAGTCCCTAGTCCATAAGCATTC-3′), encoding BamHI and EcoRI sites, respectively, and by using Pfu Ultra proofreading DNA polymerase (Agilent Technologies, Cedar Creek, TX). The 2,600-bp amplicon was then cloned into plasmid pVIVO-2 (Invitrogen, Milan, Italy) downstream of the human ferritin heavy-chain promoter (pEnv). The cloned gene was checked by complete cycle sequencing using an automated DNA sequencer (GE Healthcare, Milan, Italy). The vector used to produce the boost had unmodified pKKS Env.
The feline GM-CSF coding sequence was obtained from the total RNA of alveolar macrophages of a euthanized FIV-infected cat. Total RNA was extracted with an RNeasy minikit (Qiagen, Milan, Italy) according to the manufacturer's instructions and reverse transcribed (400 ng) with random hexamers (GE Healthcare), and cDNA was amplified with primers GM-OS (5′-ATGTGGCTGCAGAACCTGC-3′) and GM-OAS (5′-CCAGCAGTCAAAGGGGATTG-3′), designed from the GenBank feline GM-CSF coding sequence (accession no. NM_001009846). The amplified product (435 bp) was cloned into pCR2.1-TOPO (Invitrogen) and reamplified with primers GM-BspHI-S (5′-GCTCATGAGCCACCATGTGGCTGCAGAACCTGC-3′) and GM-XbaI-AS (5′-GCTCTAGATTACTTCTGGTCTGGTCCCCAGC-3′), bearing BspHI and XbaI restriction sites, respectively. Finally, GM-CSF (98% homologous to the sequence in the GenBank database) was inserted downstream of the ferritin light-chain promoter of pEnv (pEnv/GM-CSF). The sequence for feline IL-15 was derived from pND14-Lc-IL-15 (pIL15) (kindly provided by Gregg Dean). Briefly, the IL-15 coding sequence was amplified with primers IL15Nru-S (5′-ACGCTCGCGAATGGATGCAATGAAGAGAGGGCTC-3′) and IL15Sac-AS (5′-TCCCCGCGGTCAAGAAGTGTTGATGAACATTTG-3′) and Pfu Ultra proofreading DNA polymerase. The amplicon was enzyme digested and cloned into vector v-B as described below.
Vector v-A was produced from pKKS by deleting the gag-pol coding sequence downstream packaging site (ψ) and the 120-nucleotide (nt) gag sequence known to be important for RNA encapsidation (7, 8) and upstream of the 3′-end pol sequence overlapping vif (nt 749 to 5154). The deleted segment was replaced with a polycloning site to insert the cytomegalovirus promoter (CMVp) and the green fluorescent protein (GFP) reporter gene. Vector v-B was derived from v-A by replacing the 5′ long terminal repeat (5′-LTR) U3 region (nt 1 to 203) with CMVp, and v-C was obtained from vFIV-B by replacing the vif-ORF-A region (nt 5230 to 6238) with the phosphoglycerate kinase promoter (PGKp). Vector vIL15/Env was obtained from v-B by replacing GFP with the IL-15 amplicon obtained as described above. Final and intermediate constructs were checked by cycle sequencing.
Pseudotyped FIV particles were prepared by cotransfecting 293T cells with vector, packaging, and Env plasmids. Packaging pΔenv1 was produced from pΔ00, a replication-competent molecular clone of FIV-Pet derived from p34TF10 (56). A detailed description of pΔenv1 was described elsewhere previously (55). The Env plasmid was pEnv/GM-CSF; pEE14-Env, also encoding KKS Env (Lonza, Milan, Italy); or pVSV-G, providing gp G of vesicular stomatitis virus (VSV-G). Briefly, 3 × 106 293T cells were seeded into 10-cm tissue culture dishes (BD Biosciences, Milan, Italy) the day before and cotransfected with 10 μg pΔenv1, 8 μg of plasmid vector, and 2 μg of pEnv/GM-CSF, pEE14-Env, or pVSV-G by using a standard calcium phosphate precipitation protocol. Vector supernatants were harvested at 48 h posttransfection, filtered (0.45 μm), tested for FIV p25 capsid content as a measure of the viral particles produced (54), aliquoted, titrated as described previously (55), and stored at −80°C until use. Transfection efficiency was evaluated by counting GFP-positive cells by flow cytometry with a FACScan apparatus using CELLQuest, version 2, software (BD Biosciences).
CrFK, 293T, and MBM cells, seeded at 0.2 × 106 to 0.5 × 106 cells/well in 24-well plates the day before, had their culture medium replaced with vector supernatant (1 ml/well) containing approximately 10 transduction units/cell. After 4 h in 5% CO2 at 37°C, the supernatants were replaced with fresh culture medium and further incubated for 2 to 3 days. Unless otherwise stated, the efficiency of transduction was evaluated by counting GFP-positive cells by flow cytometry at day 2. The shortest incubation period permitting a satisfactory transduction efficiency of freshly cultured T lymphocytes was determined by using Ficoll-purified peripheral blood mononuclear cells (PBMCs) from three individual donor cats and several conditions of ConA (5 μg/ml) and IL-2 (20 U/ml) treatment, namely, 4 h of ConA followed by 4 h of IL-2, 4 h of ConA followed by 12 h of IL-2, and 16 h or 24 h of ConA-IL-2 costimulation. Parallel PBMCs incubated for 24 h with ConA followed by 24 h with IL-2 served as a reference control. All cultures were then transduced with VSV-G-pseudotyped FIV particles delivering v-B and examined for GFP expression by flow cytometry 1 and 2 days later. Based on the results of these preliminary tests, the procedure selected for transducing the autologous T lymphocytes used for boosting was as follows: the PBMCs purified from 10 ml of blood were costimulated with ConA-IL-2 for 16 h and then incubated for 24 h with VSV-G-pseudotyped particles delivering vIL15/Env (10 transduction units/cell). The cells were then harvested and resuspended in 1 ml of apyrogenic saline, the number of viable cells was counted, and the cells were reinoculated into the donor animals.
Env expression was examined by Western blotting (WB) and/or immunofluorescence analyses. WB was carried out by using anti-FIV SU monoclonal antibody vpg-2 (kind gift of Brian Willett, University of Glasgow, Glasgow, United Kingdom), followed by a horseradish peroxidase-conjugated rabbit anti-mouse IgG antibody (Sigma-Aldrich). Immunofluorescence was performed with Env monoclonal antibody 5F7, kindly provided by Philip Andersen (Idexx Laboratories, Westbrook, MN), followed by a fluorescein isothiocyanate-conjugated goat anti-mouse IgG antibody (Sigma-Aldrich). GM-CSF expression was evaluated with a lymphoproliferation assay using GM-CSF-responsive TF-1 cells. CrFK cells (106 cells) were transfected with 10 μg pEnv/GM-CSF or pVIVO2, and supernatants were collected 48 h later, passed through 0.45-μm filters (Corning Life Sciences, Amsterdam, Netherlands), diluted as indicated, and added to TF-1 cells seeded at 2 × 105 cells/ml in 96-well plates. Parallel cultures incubated with 2 ng/ml recombinant human GM-CSF or medium alone served as positive and negative controls, respectively. After 48 h of incubation, the cells were pulsed with 0.75 μCi/well [3H]thymidine for 5 h and counted with a 1450 Microbeta Trilux β counter (Perkin-Elmer, Monza, Italy). Results were expressed as the ratio between the average counts per min (cpm) measured in the stimulated cells versus those of negative-control cells (stimulation index [SI]). SI values of ≥2 were taken to indicate specific proliferation. IL-15 expression was tested with a lymphoproliferation assay using CTLL-2 cells. Supernatants from 106 CrFK cells transfected 2 days earlier with 10 μg of vEnv/IL-15, v-B (negative control), or pIL15 (positive control) were filtered (0.45 μm), diluted, and added to CTLL-2 cells (2 × 105 cells/well). These supernatants were then incubated for 16 h at 37°C, pulsed for 5 h with 0.75 μCi/well [3H]thymidine, and analyzed as described above. An SI of ≥2 (calculated relative to cells incubated with medium alone) was considered to indicate specific proliferation.
Female SPF cats free from FIV and feline leukemia virus antibodies were purchased from IFFA Credo (Lyon, France). The cats were 8 months old at the start of immunization, were housed individually in our climatized facility in accordance with European Community guidelines, had ad libitum access to water and a proprietary brand of food, and were assigned to the experimental groups at random. The cats were monitored by routine hematochemical analyses, circulating lymphocyte subsets, and immunological responses 2 weeks before the initiation of the experiment, during immunization, and after challenge. Prior to any procedure, the animals were sedated with ketamine-diazepam intramuscularly. The challenge virus was FIV-Pet that had been consecutively passed in vivo several times and, as a result, had reacquired virulence and neutralization characteristics typical of those of wild-type FIV (53). The preparation used was pooled plasma from acutely infected SPF cats titrated in vivo, frozen in aliquots, and diluted when used to contain 10 50% cat infectious doses (CID50) in 1 ml.
Env-specific CTL activity in PBMCs was measured as previously described (5). Briefly, the day before the assay, frozen target autologous fibroblasts expressing Env and GFP or GFP alone (negative control) were seeded into 48-well plates (1 × 104 cells/well) and left to adhere overnight. Fresh Ficoll-purified PBMCs were added onto autologous and heterologous fibroblasts at an effector-to-target cell ratio of 50:1. After overnight incubation at 37°C in 5% CO2, the PBMCs were extensively washed out, and the target cells were trypsinized and fixed for flow cytometry. Samples were analyzed in triplicate by using CELLQuest, version 2, software, and the percent specific anti-Env CTL activity was calculated by using the following formula: (% fluorescent target cells incubated alone − % fluorescent target cells incubated with autologous effector cells) − (% fluorescent target cells incubated alone − % fluorescent target cells incubated with heterologous effector cells)/% fluorescent target cells incubated alone × 100. Values of cell lysis above 6% (i.e., % fluorescent target cells incubated alone − % fluorescent target cells incubated with heterologous effector cells + 3× the standard deviation) were considered positive.
Gamma interferon (IFN-γ)-secreting T lymphocytes were counted by an enzyme-linked immunospot (ELISpot) assay using pooled 15-mer peptides overlapping the entire Env of FIV-Pet as an antigen and the feline IFN-γ development module (R&D Systems, Minneapolis, MN) (54). Briefly, 100 μl of IFN-γ capture antibody was poured into 96-well polyvinylidene difluoride (PVDF) microplates (Millipore, Milan, Italy) and kept overnight at 4°C in the dark. The wells were then washed three times with cold phosphate-buffered saline (PBS), and any unreacted site was blocked with 200 μl/well of blocking buffer (1% bovine serum albumin [BSA] and 5% sucrose in PBS) at room temperature for 2 h. The buffer was replaced with 100 μl complete RPMI 1640 medium containing 2 × 105 Ficoll-purified PBMCs stimulated with Env peptides. PBMCs stimulated with phorbol myristate acetate or medium alone served as positive and negative controls, respectively. After 24 h at 37°C in 5% CO2, the plates were washed four times with PBS-0.05% Tween 20 and kept overnight at 4°C with an affinity-purified biotinylated polyclonal sheep anti-feline IFN-γ antibody. Unbound antibody was removed by three washes with the above-described Tween solution, and the assay mixture was developed with the ELISpot Blue Color module according to the manufacturer's protocol. The plates were then washed in distilled water and air dried, and the frequency of Env-specific IFN-γ-secreting T lymphocytes was determined by using a binocular dissecting microscope. Numbers of IFN-γ-secreting T lymphocytes above 20 (i.e., numbers of spots observed in the absence of Env peptides + 3× the standard deviation) were considered positive.
Whole FIV binding antibodies were tested with an enzyme-linked immunosorbent assay (ELISA) against disrupted FIV-Pet as described previously (43) and are expressed as the reciprocal of the highest dilution giving a positive reading. Reactivities of ≥10 were considered positive. Env binding antibodies were measured as described previously (26). Briefly, High Binding microtiter plates (Greiner, Lumezzane, Italy) were coated overnight with 5 μg/well Galanthus nivalis lectin (Vector Laboratories, Burlingame, CA) in 1 mM NaHCO3 buffer. The wells were then washed three times with TBST buffer (100 mM NaCl, 50 mM Tris, 0.05% Tween 20 [pH 7.6]), and unreacted sites were blocked with TBST supplemented with 3% BSA at room temperature for 1 h. Each well then received 50 μl of FIV-Pet pretreated overnight with 0.5% Triton at 4°C, ultracentrifuged at 20,000 × g for 90 min, and resuspended in PBS. After 1 night at room temperature, the plates were washed twice with TBST, postcoated with PBS-5% skim milk for 1 h, and incubated with 50 μl of plasma diluted 1:50 to 1:400 in PBS containing 4% Tween 20, 5% FCS, and 0.5% BSA. After 1 h, the wells were washed four times with TBST and further incubated with primary mouse anti-feline IgG (AbD Serotec, Raleigh, NC) diluted 1:500 in PBS buffer, washed four times in TBST, incubated with tetramethylbenzidine-H2O2 (TMB substrate kit; Euroclone, Pero, Italy) for 10 min in the dark, and finally read at 450 nm/650 nm. Optical densities (OD) above 0.085 (i.e., average values for three naïve SPF cats + 3× the standard deviation) were considered positive. NA were measured against 10 50% tissue culture infectious doses of FIV-Pet grown in MBM cells, with 50% inhibition of reverse transcriptase (RT) production as a readout (21). The mixtures of virus and diluted plasma were incubated at 4°C for 1 h, inoculated into 1 × 105 MBM cells, washed out after 5 h at 37°C, and finally replaced with fresh complete RPMI 1640 medium. RT activity in the supernatants was evaluated at day 7. Neutralizing titers were defined as the reciprocal of the highest plasma dilution that reduced the RT levels produced by FIV exposed to the same dilution of pooled serum from 10 normal cats by ≥50%. Titers of ≥8 were considered positive.
Viral RNA was extracted from plasma by using the QIAamp viral RNA kit (Qiagen, Milan, Italy), reverse transcribed, and amplified by RT-TaqMan PCR with the IQ5 multicolor real-time PCR detection system (Bio-Rad, Milan, Italy). Sensitivity was 200 FIV RNA copies/ml plasma. Genomic DNA was extracted from PBMCs by using the QIAamp DNA blood minikit (Qiagen), and proviral DNA was quantified from 0.4 μg of genomic DNA by TaqMan PCR. The sensitivity was 100 FIV DNA copies/μg genomic DNA. RT and amplification conditions were described elsewhere previously (54).
Peripheral blood CD4 and CD8 T lymphocytes were counted as previously described (18). Briefly, 30 μl of heparinized blood was incubated with 3 μl of anti-feline CD4- or CD8-phycoerythrin (clones vpg34 and vpg9, respectively; AbD Serotec, Raleigh, NC) on ice for 30 min. Red cells were lysed, and leukocytes were fixed with 30 μl Opti-Lyse (Beckman Coulter, Milan, Italy) for 10 min at room temperature, followed by 300 μl of distilled water for 10 min. FlowCount microspheres (30 μl, 1.014 microspheres/μl; Beckman Coulter) were added, and samples were analyzed by flow cytometry.
Statistical analyses were performed by using SPSS 15.0 software for Windows (SPSS Inc., Chicago, IL). Differences between vaccine and control groups were investigated by arbitrarily assigning to the negative results a value equal to the lower limit of sensitivity of the detection method used.
The DNA immunogen was produced by cloning FIV Env and feline GM-CSF into plasmid pVIVO2. The resulting construct, pEnv/GM-CSF, was tested for Env and GM-CSF expression in transfected CrFK and 293T cells. Env expression was analyzed by WB of cell lysates collected 2 days posttransfection. As shown in Fig. Fig.1B,1B, precursor 135-kDa gp and SU gp95 showed similar patterns regardless of whether they were produced by pEnv/GM-CSF-transfected CrFK or chronically FIV-infected FL-4 cells, indicating that the transfected CrFK cells processed Env correctly. Transfected 293T cells yielded similar findings (data not shown). As described below, when used to pseudotype FIV particles, the Env expressed by pEnv/GM-CSF permitted their efficient replication in MBM cells, indicating that it had a functional conformation. The release and biological activity of GM-CSF were checked with a proliferation assay using TF-1 cells. Cell-free supernatants from pEnv/GM-CSF-transfected CrFK cells 2, 7, or 10 days earlier produced growth stimulation comparable to that of human GM-CSF, while the supernatant of empty-pVIVO2-transfected CRFK cells did not (Fig. (Fig.1C).1C). Thus, pEnv/GM-CSF encoded functional Env and GM-CSF and did so for a prolonged period of time.
Preparation of the Env- and IL-15-expressing autologous T lymphocytes used for boosting required the development of appropriate vectors delivering Env and IL-15, the definition of favorable conditions for generating efficient transducing particles, the evaluation of the particles generated for transducing efficiency in fresh feline T lymphocytes, and the definition of minimum stress conditions for obtaining transduction-competent T lymphocytes.
The Env vector was derived from pKKS (Fig. (Fig.1A)1A) and delivered the same env sequence cloned into pEnv/GM-CSF and feline IL-15 (or GFP, used for evaluating the vector in vitro). It was selected by comparing three appositely developed constructs for GFP and Env expression in transfected CrFK and 293T cells: v-A, which had env under the control of the 5′-LTR as in wild-type FIV and IL-15 expression driven by a CMVp; v-B, which was derived from v-A by replacing the U3 region of the 5′-LTR with the CMVp; and v-C, which differed from v-B for having env under the control of the PGKp (Fig. (Fig.2A).2A). GFP expression was much greater in 293T than in CrFK cells (approximately 30% versus 85%), basically reflecting the different susceptibilities of these cells to transfection, but did not differ between vectors. In contrast, Env expression was highest with v-B in both CrFK and 293T cells, intermediate for both cell lines with v-C, and, as expected due to the low or null activity of FIV LTRs in nonfeline cells (49), minimal with v-A in 293T cells (Fig. (Fig.2B).2B). The low level of activity of v-C, which was tentatively attributed to promoter interference but not further investigated, was consistent in three experiments, thus leading to the selection of v-B.
The conditions that permitted the preparation of transducing particles of sufficiently high titers were determined by transfecting 293T and CrFK cells with v-B, pΔEnv1, and one of the following Env constructs: pVSV-G, pEE14-Env, or, with the additional purpose of evaluating the functional properties of the Env encoded, pEnv/GM-CSF. Regardless of the plasmid combination used, the transducing particles released into supernatants had titers of approximately 108 and 109 RNA copies per ml in CrFK and 293T cell supernatants, respectively (data not shown), and exhibited profiles that closely resembled the intact virus by WB (Fig. (Fig.2C2C).
The transduction efficiencies of the particles were first assessed in the MBM T-cell line. As shown in Fig. Fig.3A,3A, ,22 days after transduction, GFP expression levels were around 40% for the particles pseudotyped with VSV-G or the Env provided by pEnv/GM-CSF and over 50% with the particles pseudotyped with pEE14-Env; also, transduced cells expressed SU 95gp and the precursor 135-kDa gp at ratios similar to those for FL-4 cells regardless of whether transduction had been performed 2 days earlier (Fig. (Fig.3B)3B) or 7 days earlier (data not shown). When tested with PBMCs stimulated with ConA for 24 h followed by IL-2 for another day, the VSV-G-pseudotyped particles transduced three times more cells than did the Env-pseudotyped ones (Fig. (Fig.3A).3A). Neither particle affected MBM cell and PBMC viability and growth at 1 week (data not shown). Thus, VSV-G pseudotyping was slightly superior for PBMC transduction. As mentioned above, the study also showed that the Env encoded by pEnv/GM-CSF was functional.
After preliminary tests outlined in Materials and Methods, the in vitro manipulation of primary T lymphocytes was kept to a minimum by costimulating freshly harvested PBMCs with ConA and IL-2 for 16 h and incubating them with the transducing particles for 24 h. This procedure consistently permitted the transduction of 30 to 50% of cells.
Before the preparation of the boosting inocula, the GFP gene in v-B was replaced with the feline IL-15 gene, and the construct thus obtained, vIL15/Env, was tested for the ability to generate functional IL-15. The supernatants of vIL15/Env-transfected CrFK cells (Fig. (Fig.3C)3C) and MBM cells (data not shown) efficiently stimulated the growth of CTLL-2 cells, thus showing that vIL15/Env generated biologically active IL-15. Finally, when examined by immunofluorescence, vIL15/Env-transduced CrFK cells were clearly positive for FIV Env, which in nonpermeabilized cells was concentrated mostly at the cell surface (Fig. (Fig.3D3D).
Seven SPF cats (vaccinees) were primed at weeks 0, 3, and 12 with 300 μl of 1 μg/μl pEnv/GM-CSF in apyrogenic water injected intramuscularly into two sites of the hind thigh. Twenty minutes earlier, the injection sites had been shaved and rubbed with Aldara (Graceway Pharmaceuticals, Bristol, TN), a topical imiquimod-containing cream known to exert an adjuvant effect. At week 16, the same animals were boosted intraperitoneally with vIL15/Env-transduced autologous T lymphocytes generated as described above. The number of viable cells reinfused ranged between 3 × 106 and 10 × 106 cells (Table (Table1).1). A second group of four cats (mock vaccinees) were mock primed with empty pVIVO2 DNA and mock boosted with their own T lymphocytes processed as described above for the preparation of the boost given to the vaccinees except for the use of empty viral particles. No local reactions or adverse clinical signs were observed during immunization. A final group of four animals was left untreated and served as naïve controls.
At selected times during the course of immunization, vaccinees and mock vaccinees were evaluated for anti-Env cell-mediated immune responses (Fig. (Fig.4).4). For the latter group, Env-specific CTL activity and IFN-γ-secreting T lymphocytes were essentially nil at all times tested. On the contrary, the vaccinees possessed weak but clearly detectable levels of both effectors starting from 2 weeks after the second DNA inoculation. Furthermore, in these animals, levels of both CTL activity and IFN-γ-secreting T lymphocytes underwent a marked increase following the boost. Thus, by week 26, 5 weeks postboost, the levels of these effectors were 2-fold (cat GF) to 8-fold (cat GG) higher than those at preboost. Although at challenge (9 weeks postboost), these levels had generally declined, at this time they were also still markedly higher than those at boost (Table (Table1).1). Humoral immunity was assessed at the start of immunization, when all animals were uniformly negative (data not shown), at boost, and at challenge. In line with data from previously reported FIV DNA vaccination studies (24, 25, 37, 54), at boost, serology was still essentially negative except for two vaccinees (cats GD and GE) who, somewhat surprisingly, showed measurable anti-Env binding antibodies and NA. On the other hand, at challenge, all vaccinees possessed moderate to high levels of anti-whole-FIV antibodies and anti-Env binding antibodies, and, most importantly, five of seven vaccinees had NA at titers of ≥128 (Table (Table1).1). Collectively, these data demonstrated that the inoculation of autologous T lymphocytes transduced ex vivo to express Env and IL-15 had proven a valuable boost. Incidentally, the efficacy of boosting with the engineered autologous T lymphocytes used was also underlined by the development of moderate levels of whole FIV binding antibodies in the absence of Env antibodies in the mock vaccinees (Table (Table1).1). By WB these antibodies reacted solely with FIV capsid protein p25 (data not shown), thus indicating that the traces of FIV capsid antigen present in the mock boost as a result of the fact that the autologous T lymphocytes had been exposed to empty viral particles had sufficed to generate some antibody response.
Vaccinees, mock vaccinees, and naïve cats were challenged intraperitoneally with 10 CID50 of ex vivo FIV-Pet 9 weeks after the boost and then monitored for infection for 10 months.
At postchallenge (p.c.), all the naïve and mock-vaccinated controls became readily infected (FIV RNA viremia and proviral loads were consistently evident by 2 weeks p.c.), and all the animals that could be monitored for a sufficient time (naïve cat GP died 1 month p.c. as a consequence of an unexplained renal failure) exhibited the kinetics of molecular markers typical of a florid FIV infection (Fig. (Fig.5).5). These findings were corroborated by quantitative FIV isolation performed, as previously described (43), at 3 months p.c., revealing numbers of infectious units ranging between 1 and 100 infectious units per 106 PBMCs, and by the prompt immune responses to FIV that developed (Fig. (Fig.66 and data not shown). In particular, NA became evident at different times p.c., peaked at titers of 1:16 or 1:32, and then either remained stable or decreased slightly over time (Fig. (Fig.6).6). Only cat GQ, i.e., the animal with the highest levels of plasma viremia and proviral load seen in the study (Fig. (Fig.5),5), remained NA negative throughout the follow-up period. Finally, all the controls underwent the expected progressive decline in levels of circulating CD4 T cells (Fig. (Fig.77).
For the vaccinees, the outcome of challenge was quite different. Two cats (cats GD and GF) proved totally protected in that they constantly tested negative for FIV viremia and proviral load (Fig. (Fig.5)5) and had stable levels of CD4 T cells throughout the follow-up period (Fig. (Fig.7).7). Moreover, repeated attempts to culture FIV from their PBMCs were constantly negative (data not shown). Three cats (cats GE, GH, and GI) underwent a markedly attenuated infection as revealed by (i) levels of plasma viremia and proviral load in the PBMCs that remained near or below the limit of detection except for one or two small blips that occurred during the first 2 months p.c. (Fig. (Fig.5),5), (ii) virus cultures that were positive only occasionally and at high PBMC inputs (data not shown), and (iii) circulating CD4 T-cell values that remained stable throughout the observation period (cats GE and GH) or exhibited only a transient moderate drop that occurred soon after an early blip in viremia (cat GI). Finally, the remaining two vaccinees (cats GC and GG) were scored as overtly infected, but their viral RNA and DNA levels were frequently lower than those of the controls (Fig. (Fig.5),5), and their CD4 T-cell values were modestly impacted throughout the 10 months of observation (Fig. (Fig.7).7). On the other hand, the challenge had little or no consequences for the FIV immunological status observed for the vaccinees prechallenge. The only changes observed were some brisk fluctuations in CTL activity and IFN-γ-secreting T-lymphocyte numbers (data not shown) that were clearly independent of challenge outcome and a marginal increase in the NA titers of the two vaccinees that were overtly infected (Fig. (Fig.66).
For statistical analysis of the outcome of challenge, any negative result was conservatively assigned a value equal to the lower limit of sensitivity of the detection assay used, and mock-vaccinated and naïve animals were considered together since they had shown no appreciable differences in the parameters examined. The geometric mean of plasma viremia of the vaccinees was significantly lower than that of the controls at months 0.5 (P = 0.011), 1 (P = 0.004), 2 (P = 0.004), 3 (P = 0.0006), 6 (P < 0.0001), and 10 (P < 0.0001), and the same was true for the proviral loads at months 1 (P = 0.01), 2 (P < 0.0001), 3 (P < 0.0001), 6 (P < 0.0001), and 10 (P < 0.0001) (Mann-Whitney test for independent samples). The number of subjects who at any time p.c. underwent a >10% reduction in levels of circulating CD4 T cells relative to the time of challenge was also significantly lower in the vaccinee group than in the control group (1/7 versus 7/7; P = 0.0007 by Fisher's exact test). Finally, the levels of plasma viremia observed at any time p.c. were found to correlate negatively with the titers of NA at the corresponding times (r, −0.388; P = 0.011 by Pearson test).
The major novelty of the FIV vaccination strategy explored here, relative to the DNA prime-protein boost approaches tested in the past, has been the use of autologous T lymphocytes engineered to express Env as a boosting immunogen. SPF cats were primed with a plasmid expressing FIV Env DNA and feline GM-CSF and then boosted with 3 to 10 million viable autologous T lymphocytes that had been transduced ex vivo to produce FIV Env and feline IL-15 in a sustained fashion. Nine weeks after the boost, when the animals were challenged intraperitoneally against virulent ex vivo FIV, they possessed robust Env-specific immune effectors, including generally high titers of NA, and showed degrees of protection that ranged from complete to substantial with regard to both viral replication and effects of the infection on circulating CD4 T lymphocytes. In marked contrast, all of the mock-vaccinated and naïve control animals challenged in parallel became readily infected and, although none developed AIDS, as typically observed for experimentally infected cats housed in climatized facilities (25), exhibited clear hallmarks of a florid FIV infection.
In the light of the poor results obtained in the past whenever the Env proteins of FIV and other lentiviruses were tested as a vaccine antigen (13, 25-27, 30, 38, 58, 67), the outcome of the present proof-of-concept experiment may represent an important step forward and encourages further studies with this and other animal systems to compare the approach used here to the prime-boost ones previously proposed. On the other hand, the present study was conducted in an entirely homologous virus system and with a challenge virus that, despite having been readapted in vivo, might be more sensitive to neutralization than truly primary isolates (25, 69). Further studies will therefore also need to investigate the breadth of the NA response generated, the protection afforded against genuinely wild-type isolates, and whether this breadth can be extended by administering additional boosts. The detection of broadly acting and high-titer NA in HIV-infected individuals is uncommon and usually occurs after protracted periods of infection (9, 39, 40). However, these findings may provide important insights into the requirement for making efficacious Env-based antilentiviral vaccines.
Previous attempts to exploit cell-based immunogens for vaccinating against FIV gave some promising results (16, 61). In particular, the repeated inoculation of paraformaldehyde-fixed FIV-infected MBM cells afforded marked protection (43-45, 69). However, the use of similarly fixed FIV-infected autologous infected T lymphocytes or of viable dendritic cells loaded with inactivated FIV failed to protect or even aggravated homologous virus challenge (18, 21), thus excluding the possibility that the key to the success of the approach tested here was merely the use of a cellular boost. The reasons for its efficacy may be several that are not mutually exclusive. First, during both priming and boosting, the Env immunogen was generated within the body itself. This may have represented a key advantage since it is generally accepted that most neutralizing epitopes present in lentiviral Env proteins are conformational in nature, concealed in resting virions, and accessible very briefly solely in the very early stages of the virion interaction with susceptible cells (10, 11, 28). Furthermore, the absence in the immunogen of other viral proteins that are known to accelerate Env folding (70) and are also likely to produce undesirable effects, such as immune competition and inappropriate lymphoid activation, may have allowed a better immune recognition of usually elusive Env epitopes. Second, since T cells are the primary target of FIV (64), boosting with viable Env-expressing T lymphocytes may have improved the quality of Env expression and processing and/or may have amplified the Env interaction with its receptors and coreceptors, permitting immune recognition in a manner very similar, if not identical, to what happens during actual FIV infection. Third, immune responses to Env may have been potentiated by the combination of adjuvants that we employed. Both GM-CSF and IL-15 have been shown to be valuable adjuvants in several animal species (2, 12, 34, 36, 57), including the cat (41), while imiquimod, which was applied at the sites of DNA injections, is a Toll-like receptor agonist known to induce the local release of several cytokines (19) as well as dendritic cell maturation and homing to draining lymph nodes (17, 50). To our knowledge, these adjuvants had not been used in concert before, and the actual contribution of each of them will need to be defined in further studies. Last, but not least, the viable Env-expressing T cells used for boosting may have persisted in the body for an extended time, providing protracted antigen stimulation that may have been especially important for eliciting the high titers of NA observed at challenge.
A corollary finding of the present study was that in the vaccinees, the extent of control of FIV replication, as deduced from the reduced viral loads in plasma, correlated with the titer of NA present in the animals. This adds to the substantial body of evidence collected in recent years indicating that although it most likely acts in concert with other immune effectors, NA exerts a crucial role in immune protection against lentiviruses (20, 25, 28, 39, 40, 59, 63). Indeed, the development of immunogens capable of eliciting NA remains an important focus of current AIDS vaccine research (3, 11, 28-30, 59).
We thank Gregg Dean (North Carolina State University, Raleigh, NC) for providing the IL-15-encoding plasmid and cells for testing IL-15 bioactivity and Brian Willett (University of Glasgow, Glasgow, United Kingdom) and Philip Andersen (IDEXX Laboratories, Westbrook, MN) for providing the anti-FIV Env monoclonal antibodies. We also thank Antonio Merico for in vivo experiments and excellent technical assistance.
This study was supported by Ministero dell'Istruzione, dell'Università e della Ricerca grant PRIN 2007 and Ministero della Salute—Istituto Superiore di Sanità, VI° Programma Nazionale di Ricerca sull'AIDS 2006, Progetto Nazionale AIDS—ICAV.
Published ahead of print on 3 February 2010.