A Targeted Mutation within the Feline Leukemia Virus (FeLV) Envelope Protein Immunosuppressive Domain To Improve a Canarypox Virus-Vectored FeLV Vaccine

Abstract

We previously delineated a highly conserved immunosuppressive (IS) domain within murine and primate retroviral envelope proteins that is critical for virus propagation in vivo. The envelope-mediated immunosuppression was assessed by the ability of the proteins, when expressed by allogeneic tumor cells normally rejected by engrafted mice, to allow these cells to escape, at least transiently, immune rejection. Using this approach, we identified key residues whose mutation (i) specifically abolishes immunosuppressive activity without affecting the “mechanical” function of the envelope protein and (ii) significantly enhances humoral and cellular immune responses elicited against the virus. The objective of this work was to study the immunosuppressive activity of the envelope protein (p15E) of feline leukemia virus (FeLV) and evaluate the effect of its abolition on the efficacy of a vaccine against FeLV. Here we demonstrate that the FeLV envelope protein is immunosuppressive in vivo and that this immunosuppressive activity can be “switched off” by targeted mutation of a specific amino acid. As a result of the introduction of the mutated envelope sequence into a previously well characterized canarypox virus-vectored vaccine (ALVAC-FeLV), the frequency of vaccine-induced FeLV-specific gamma interferon (IFN-γ)-producing cells was increased, whereas conversely, the frequency of vaccine-induced FeLV-specific interleukin-10 (IL-10)-producing cells was reduced. This shift in the IFN-γ/IL-10 response was associated with a higher efficacy of ALVAC-FeLV against FeLV infection. This study demonstrates that FeLV p15E is immunosuppressive in vivo, that the immunosuppressive domain of p15E can modulate the FeLV-specific immune response, and that the efficacy of FeLV vaccines can be enhanced by inhibiting the immunosuppressive activity of the IS domain through an appropriate mutation.

INTRODUCTION

Feline leukemia virus (FeLV) is a gammaretrovirus responsible for fatal diseases in cats. FeLV infection has a major impact on cat life expectancy (1), and the virus remains one of the major feline pathogens. FeLV is also a threat to wild felids, including endangered species (2–4). The prevalence of FeLV infection in cats has been reduced by the management of infected animals and vaccination but is still high in some populations (5–7).

New sensitive molecular techniques have been developed to detect FeLV infection and have changed our understanding of FeLV pathogenesis (8–11). All surveys in the field are based on the detection of p27 antigen in the blood and measure mainly the prevalence of persistently antigenemic cats. It is likely that the prevalence of infection is underestimated, since latent or regressive infections are not detected by the classical methods. More studies will be needed to further characterize the different outcomes of FeLV infection and their long-term impact (12–14).

Prevention of FeLV-related diseases relies on vaccination. Classical inactivated or subunit vaccines (15) as well as a canarypox virus-vectored vaccine (16) have been widely used and have contributed to the reduction in the prevalence of productive infection. Vaccine performances were recently reviewed in the light of the new molecular assays (17). Under the conditions of an experimental challenge, vaccines protect against persistent antigenemia but do not provide sterilizing immunity. Prevention of persistent antigenemia is the primary goal of vaccination, because persistently antigenemic cats are those that will develop FeLV-related diseases (18).

Under natural conditions, protection against FeLV infection is mediated primarily by an early cytotoxic T cell response followed by the production of neutralizing antibodies (19, 20). Innate immunity probably plays an important role, as suggested by abortive infections in naive cats and some in vitro studies (21). In vaccinated cats, several studies have shown that protection against persistent antigenemia can be obtained in the absence of neutralizing antibodies (22–24), suggesting a key role for vaccine-induced cell-mediated immune responses (25). However, like other retroviruses, FeLV has developed mechanisms to interfere with the host immune response and establish persistent infection. In particular, the transmembrane (TM) moiety of the envelope glycoprotein (p15E) contains an immunosuppressive domain (ISD) which is conserved among gammaretroviruses (26). The immunosuppressive activity of FeLV p15E has been documented in vitro (27–30) and to some extent in vivo (31).

We had previously shown that the Env proteins of a murine leukemia virus (MLV) (32), of the simian Mason-Pfizer monkey virus (MPMV) (33), and of human endogenous retroviruses (HERV) of the HERV-H and HERV-FRD families (34, 35) can antagonize the immune system-dependent elimination of tumor cells injected into immunocompetent mice following transduction of these cells by an Env expression vector. We have also shown that this immunosuppressive (IS) activity is absolutely required for MLV propagation in immunocompetent mice, with a mutant virus—carrying an introduced specific mutation abolishing IS activity without altering the mechanical fusogenic activity of the envelope protein—being rejected by the mouse immune system (36). Importantly, we established via selective cell subset depletions that this rejection was dependent on natural killer (NK) cells and cytotoxic T lymphocytes (CTL), demonstrating that envelope IS activity inhibits cytotoxic effectors of both innate and adaptive immunity. We also demonstrated an enhanced immunogenicity of the envelope proteins inactivated for their IS activity by specific point mutations (35, 36), suggesting that such mutations may be useful for optimizing vaccine antigens.

Here we measured the IS activity of the FeLV envelope protein and found that this protein is indeed immunosuppressive. Furthermore, based on sequence comparison between the MLV and FeLV Env proteins, we identified a key amino acid responsible for IS activity of FeLV and a substitution resulting in loss of this activity without significant alteration of the “mechanical” properties of the mutant protein. We further substituted the mutant “optimized” FeLV Env antigen for the wild-type (wt) antigen present in a canarypox virus-vectored vaccine (ALVAC-FeLV), and studied the impact of this substitution on the immunogenicity and efficacy of the vaccine.

MATERIALS AND METHODS

Mice and cell lines.

Six-week-old C57BL/6 and BALB/c mice were obtained from Harlan (Gannat, France). The 293T (ATCC CRL-11268), G355-5 (ATCC CRL-2033), and MCA205 (a kind gift from L. Zitvogel) cell lines were cultured in Dulbecco modified Eagle medium (DMEM) supplemented with 10% fetal calf serum, streptomycin (100 μg/ml), and penicillin (100 units/ml).

Plasmids.

The sequence coding for FeLV envelope was retrieved by PCR from the pFGA5 plasmid containing an infectious molecular clone of FeLV subgroup A (37) digested with BspHI, using the high-fidelity Pfx Platinum polymerase (Invitrogen) and primers 1 and 2 (Table 1). The amplification product was digested with XhoI and MluI and ligated in the phCMV-envT vector (38) opened with the same enzymes, to yield the phCMV-FeLVenv vector.

TABLE 1.

Primer list

Primer no.	Primer sequence (5′→3′)	Primer name
1	atacatctcgagaccggtccaactagaaccatggaaagtccaacgcaccc	Env FeLV XhoAge-S
2	atacatacgcgttcatggtcggtccggatcg	Env FeLV Mlu-AS
3	gtgagggaaaatgcaaccccc	FeLV SU-S
4	aatgctgcgcagagccctccccgttgtagg	FeLV ER FspI-AS
5	gggctctgcgcagcattaaaagaagaatgt	FeLV FspI-S
6	acatggcccagccggccctacaaatggccatgc	FeLV TM64-Sfi-S
7	acatacgcgtttagagtccggtgtgatccgc	FeLV TM64-Stop Mlu-AS

The phCMV-FeLVenvER vector with mutant env was obtained by three fragment ligations. Briefly, PCR fragments were generated using phCMV-FeLVenv as a template and the primers pairs 3-4 and 5-2. These PCR fragments were then digested with BfuAI and FspI or FspI and MluI and ligated in the phCMV-FeLVenv vector digested with BfuAI and MluI.

To generate the pDFG expression vectors for the 64 amino acids corresponding to the FeLV ectodomain, PCR fragments were retrieved using the phCMV-FeLVenv vectors coding for the wt or mutant FeLV Env proteins and the primer pair 6-7. The PCR fragments were then digested with SfiI and MluI and inserted into the pDFG vector digested with the same enzymes, yielding the pDFG-FeLVenvecto and pDFG-FeLVenvectoER vectors. All the plasmids were checked by sequencing.

Detection of wt and mutant FeLV Env expression by flow cytometry.

To assess wild-type and mutant FeLV-Env expression, 7.5 ×10⁵ 293T cells were transfected by calcium phosphate precipitation with 3 μg of phCMV plasmids encoding the wt or mutant FeLV envelope proteins or an irrelevant protein as a negative control. FeLV Env proteins were detected 36 h later using the FeLV 046IIAH anti-gp70 monoclonal antibody, which was detected by goat anti-mouse–fluorescein isothiocyanate (FITC) antibodies (Amersham), on a FACSCalibur flow cytometer (Becton, Dickinson).

Detection of wt and mutant FeLV Env expression by a functional infectivity assay.

The viral pseudotyping assay was performed as previously described (38), using the wild-type phCMV-Env protein expression vector and its mutant derivatives. Briefly, pseudotype viruses were produced by cotransfecting 7.5 ×10⁵ 293T cells with 0.55 μg of the wild-type phCMV-Env protein expression vector, 1.75 μg of an MLV gag and pol vector, and 1.75 μg of a defective retroviral vector marked with a β-galactosidase reporter gene (pMFGsnlslacZ) by calcium phosphate precipitation. At 36 h posttransfection, viral supernatants were collected and filtered through 0.45-μm-pore-size membranes. Target cells (feline G355-5) were seeded in 24-well plates at a density of 10⁴ cells per well and incubated overnight at 37°C. Five to 500 microliters of virus samples containing 4 μg of Polybrene per ml was added to the cells and centrifuged for spinoculation at 1,200 × g for 2 h 30 min at 25°C. After removal of the supernatants, the cells were incubated in regular medium for 60 h at 37°C. Viral titers were then measured by X-Gal (5-bromo-4-chloro-3-indolylphosphate) staining of the cells and expressed as lacZ focus-forming units (FFU)/ml of viral supernatant.

Establishment of FeLV Env-expressing tumor cells and MCA205 tumor rejection assay.

The IS indexes of wild-type FeLV Env and the mutant were measured following the general procedure described previously (32, 34), by first generating stably transfected MCA 205 cells with retroviral expression vectors for wild-type FeLV Env and the mutant. 293T cells (7.5 × 10⁵) were cotransfected with the FeLV Env-expressing pDFG retroviral vector to be tested (1.75 μg) and expression vectors for the MLV proteins (0.55 μg for the amphotropic MLV env vector and 1.75 μg for the MLV gag and pol vector [38]). At 36 h posttransfection, viral supernatants were harvested for infection of MCA205 tumor cells (2.5 ml of supernatant per 5 × 10⁵ cells, supplemented with 8 μg/ml Polybrene). Cells were maintained in selective medium (400 units/ml hygromycin) for 3 weeks. For in vivo assays, tumor cells were washed with phosphate-buffered saline (PBS) and trypsinized, and 10⁶ cells were inoculated subcutaneously in a shaved area of each BALB/c mouse right flank. Tumor growth was monitored by palpation twice or thrice weekly and tumor area (mm²) determined by measuring perpendicular tumor diameters. The IS index was calculated as (A_env − A_none)/A_none, where A_env and A_none are the mean areas, at the peak of growth, of tumors from BALB/c mice injected with cells expressing the env gene to be tested or with control cells transfected with a vector carrying an empty cassette, respectively. The extent of “immunosuppression” can be quantified by this IS index, with a positive index indicating that env expression facilitates tumor growth (as a consequence of its IS activity) and a null or negative index pointing to no effect or even enhanced rejection, respectively (the latter may be explained by a stimulation of the immune response of the host against the new foreign antigen, represented by a nonimmunosuppressive Env, expressed at the surface of the tumor cells).

Vaccines.

Cats were immunized with a canarypox virus (ALVAC vector) expressing the env and gag/pro genes of FeLV type A. The ALVAC-FeLV has been described elsewhere (22). Two ALVAC-FeLVs were designed: one with the native env gene (vCP2295) and the other with the mutated env gene (vCP2296). In both constructs, the env and gag/pro genes were inserted at the C5 and C3 loci, respectively, under the control of the same H6 promoter. The only difference between vCP2295 and vCP2296 was the mutation in the FeLV env gene in vCP2296: vCP2295 expressed the native env gene, and vCP2296 expressed the optimized env gene. Both viruses were grown and titrated on primary chicken embryo fibroblasts (CEF). The vaccines were produced according to the same process and consisted of a clarified viral suspension of vCP2295 or vCP2296. The proper expression of env was checked by immunofluorescence on CEF-infected cells and immunoblotting (quantitative dot blotting).

Vaccination and challenge in cats.

Two studies were performed to compare the efficacies of vCP2295 and vCP2296 at different doses. Both experimentations were approved by the Merial Ethical Committee. The first one aimed at studying the immune response and protection after vaccination with a low dose of ALVAC-FeLV, corresponding to approximately one dose protecting 50% of the cats (PD₅₀) of vCP2295 (16). The second study was done at a higher dose to assess the dose response of each vaccine. In each study, 8- to 12-week-old specific-pathogen-free (SPF) kittens (Charles River, France) were randomly assigned to three groups of 10 kittens according to sex, litter, and age. In the first study, cats were vaccinated via the subcutaneous route on days 0 and 28 with 10^5.7 doses infecting 50% of the cell cultures (CCID₅₀) of vCP2295 or vCP2296. In the second study, cats were vaccinated via the subcutaneous route on days 0 and 28 with 10^6.8 CCID₅₀ of vCP2295 or 10^6.4 CCID₅₀ of vCP2296 (vaccine titers were checked by titration on CEF after each vaccination).

Two weeks after vaccination, cats were challenged via the oro-nasal route with 10^4.2 CCID₅₀ of a virulent strain of feline leukemia virus (FeLV-A-Glasgow). Blood samples were collected from week 3 to week 15 postchallenge for weekly detection of FeLV p27 antigen in the serum and quantification of FeLV proviral load in peripheral blood mononuclear cells (PBMC) every 3 weeks.

In the first study, the vaccine-induced FeLV-specific immune response was measured by gamma interferon (IFN-γ) and interleukin-10 (IL-10) enzyme-linked immunosorbent spot assay (ELISpot assay) at 1 week after the second vaccine injection (day 35).

Determination of antigenemia and proviremia.

On a weekly basis from week 3 until week 15 postchallenge, sera were tested for FeLV p27 antigenemia by using the Witness FeLV kit or the Viracheck FeLV kit (Synbiotics Corporation, MO, USA) according to the manufacturer's instructions. If a cat was found to be positive on at least 5 occasions and/or was positive at the end of the follow-up period, it was considered persistently antigenemic.

Blood samples were collected in EDTA tubes on day 44 before challenge and every 3 weeks after challenge for white blood cell (WBC) count and FeLV proviremia monitoring on PBMC using quantitative PCR as previously described (10). Proviremia was expressed as provirus copy number/50,000 WBCs.

DC derivation.

Dendritic cells (DCs) were used for T cell stimulation in the ELISpot assay. One week before running the ELISpot assay, blood was collected in heparin-treated tubes. PBMC were isolated by Pancoll density gradient centrifugation (600 × g for 30 min), washed twice in sterile PBS (with centrifugation at 400 × g for 10 min), and subsequently counted with an ABX Pentra 120 cell counter (Horiba, France). Cells were cultivated in 6-well plates in complete RPMI (RPMI 1640, penicillin-streptomycin, 2 mM β-mercaptoethanol, 10% fetal calf serum) supplemented with feline granulocyte-macrophage colony-stimulating factor (GM-CSF) and feline IL-4 as previously described (25). Cells were incubated overnight at 37°C ± 2°C in 5% CO₂ for monocyte adherence. The following day, nonadherent cells were eliminated and fresh complete RPMI with cytokines was added to each well. Adherent cells were cultured for 7 days at 37°C ± 2°C in 5% CO₂, and fresh medium supplemented with cytokines was added every 2 to 3 days.

Measurement of FeLV-specific IFN-γ immune responses by ELISpot assay.

HA ELISpot plates (Millipore, France) were coated overnight at +4°C with purified anti-IFN-γ monoclonal antibody AF764 (R&D Systems, Minneapolis, MN, USA) diluted in carbonate-bicarbonate buffer (0.2 M, pH 9.6). Coated plates were washed three times in sterile PBS and unoccupied sites blocked with sterile complete RPMI. DCs (10⁵) were loaded with 15-mer overlapping peptide pools spanning the FeLV Env and Gag/Pro proteins at 1 μg/ml for 1 h. Loaded DCs were transferred into ELISpot plates, and 5 × 10⁵ PBMC were added into each well. DCs loaded with an irrelevant peptide pool served as a negative control. Cells were stimulated for 20 to 24 h at 37°C + 5% CO₂. ELISpot plates were washed three times in PBS–0.05% Tween and incubated overnight at +4°C with biotinylated anti-IFN-γ monoclonal antibody BAF764 (R&D Systems), and then plates were washed three times in PBS–0.05% Tween and incubated with horseradish-peroxidase-streptavidin solution (R&D Systems) for 1 h at 37°C. Plates were finally washed three times in PBS–0.05% Tween and incubated for 15 min in the dark with the AEC (3-amino-9-ethylcarbazole) substrate solution. The plates were extensively washed with tap water and dried. The spots were counted with a charge-coupled device (CCD) camera system (MicroVision, Redmond, WA, USA). The frequency of peptide-specific IFN-γ spot-forming cells (SFC) was calculated as follows: number of peptide-specific IFN-γ SFC = number of IFN-γ SFC upon individual FeLV peptide pool restimulation − number of IFN-γ SFC upon irrelevant peptide pool restimulation.

Measurement of FeLV-specific IL-10 immune responses by ELISpot assay.

The IL-10 ELISpot assay was performed with precoated plates according to the manufacturer's instructions (R&D Systems). Purified PBMC (5 × 10⁵) were directly restimulated using overlapping peptide pools at 1 μg/ml in completed RPMI and set down in ELISpot IL-10-coated plates for 20 to 24 h at 37°C with 5% CO₂. The frequency of peptide-specific IL-10 SFC was calculated as follows: number of peptide pool-specific IL-10 SFC = number of IL-10 SFC upon individual FeLV peptide pool restimulation − number of IL-10 SFC upon irrelevant peptide pool restimulation.

Statistical analysis.

The 3 groups were compared for the frequency of antigenemia by using Fisher's exact test with the following scoring system: no antigenemia, 0; transient antigenemia, 1; persistent antigenemia, 2. If the global group effect was significant, pairwise comparisons were done using the Bonferroni adjustment method.

Proviral loads were compared between groups by a mixed model with repeated measures in the period from day 63 to day 147. Fixed effects used in the mixed model were the main factors “group” and “day” and the corresponding interaction. The random effect was the “animal” effect. With repeated measures, observations of an animal over time are not independent. The model used takes into account these correlations by modeling the covariance structure within subjects. The Toeplitz type was chosen for the covariance matrix structure. Pairwise comparisons were done using Tukey's test.

The dose responses of the vCP2295 and vCP2296 vaccines were evaluated by using a logistic regression model. The model was used to evaluate the dose which would provide 80% protection (PD₈₀) against the same experimental challenge.

The preventable fraction (see Table 3) is equal to [(percentage of antigenemic cats among controls) − (percentage of antigenemic cats among vaccinates)]/(percentage of antigenemic cats among controls).

TABLE 3.

Results of the vaccination/challenge studies

Vaccine	Vaccine dose of 105.7 CCID50		Vaccine dose of 106.4 or 106.8 CCID50a		Dose providing 80% protectionb
	Protection	Preventable fraction (%)	Protection	Preventable fraction (%)
vCP2295 (native env gene)	4/10	40	8/10	78	106.8
vCP2296 (mutated env gene)	6/10	60	9/10	89	106.0
Controls	0/10	0	1/10	0

^avCP2295 titer, 10^6.8 CCID₅₀/dose; vCP2296 titer, 10^6.4 CCID₅₀/dose.

^bBased on logistical regression analysis.

The frequency of FeLV-specific IFN-γ- or IL-10-positive cells was expressed in log units (in the absence of spot, a value of 1 was attributed) and compared between groups by analysis of the variance (ANOVA) or the Kruskal-Wallis (KW) test depending on the normality of the distribution (Shapiro-Wilk test) and the homogeneity of variances (Cochran's C test).

Statistical analyses were carried out with SAS software (release 9.1) or STATGRAPHICS Plus (release 5.1). The significance level was set at 5% for all tests.

RESULTS

Immunosuppressive activity of the FeLV envelope protein.

We have previously demonstrated that the Env proteins of murine leukemia viruses (MLVs), MPMV retrovirus, and a series of murine or human ERVs are immunosuppressive by using an in vivo tumor rejection assay (32, 33, 35, 36). The IS function is carried by the transmembrane (TM) subunit (Fig. 1A), more precisely by a highly conserved domain of the extracellular moiety of the TM subunit (the “ectodomain”), which is large enough to adopt a definite conformation (see the crystallographic structures of the syncytin-2 ectodomain and of other retroviral envelopes in references 39, 40, and 41). In previous studies, we have delineated the immunosuppressive domain (ISD) responsible for the IS function and identified the key residue governing the IS activity (35, 36).

FIG 1.

Characterization of the fusogenic and IS activities of FeLV Env and generation of a fusion-positive, immunosuppression-negative specific mutant. (A) Schematic representation of the FeLV Env. The left panel shows the three-dimensional structure of the Env trimer, with the surface (SU) and transmembrane (TM) subunits and the immunosuppressive domain (ISD). On the right, a more detailed representation shows the linear organization of the FeLV Env, with the furin cleavage site, the hydrophobic fusion peptide and the transmembrane anchor. The ISD amino acid sequence is shown for the MLV and FeLV Env proteins, with the E→R substitution indicated. (B) IS activities of FeLV and MLV Env wild-type and mutant proteins in the in vivo MCA205 tumor rejection assay. Tumor growth was monitored twice or thrice weekly and tumor areas (mm²) determined by measuring perpendicular tumor diameters. Results are expressed as described previously (35, 36) with an IS index calculated as (A_env − A_none)/A_none at days 7 and 8 postinjection, where A_env and A_none are the mean areas of tumors from BALB/c mice injected with cells expressing the FeLV or MLV env gene to be tested or with control cells transfected with an empty vector (none). (C) Infectivities of FeLV Env and its mutant derivative as expressed on the surface of MLV viral pseudotypes, using G355 cells as a target. Titers (LacZ-positive focus-forming units/ml of viral supernatant) were measured as described previously (38) (mean values ± standard deviations [SD]).

Since the alignment of the amino acid sequences of the MLV and FeLV ectodomains discloses high conservation between the two proteins (Fig. 1A), it was likely that they could share common functions, and we therefore tested the immunosuppressive activity of the FeLV Env protein, using an in vivo tumor rejection assay. The rationale of this assay can be summarized as follows: while injection of MCA205 tumor cells (H-2^b) into allogeneic BALB/c mice (H-2^d) leads to the formation of no tumor or transient tumors that are rapidly rejected, injection of the same cells but stably expressing an immunosuppressive retroviral Env protein leads to the growth of larger tumors that persist for a longer time, in spite of the expression of the new exogenous antigen, resulting in a positive immunosuppression index (see Materials and Methods). This difference is not associated with a difference in intrinsic cell growth rate, since it is not observed in syngeneic C57BL/6 mice, and it is immune system dependent. MCA205 cells were transduced with expression vectors for the FeLV Env ectodomain or control proteins and assayed for tumor cell growth upon both syngeneic and allogeneic engraftment. Using this test and as expected from amino acid sequence similarities, we now demonstrate that the FeLV Env ectodomain is immunosuppressive (Fig. 1B) to the same extent as the MLV Env used as a positive control, as evidenced by the positive immunosuppression index.

Selective modulation of the immunosuppressive activity of the FeLV Env protein.

Based on the sequence comparison between homologous domains of the MLV and FeLV Env proteins and on previous results (35, 36), the amino acid at position 14 of the ISD, namely, a glutamic acid as for MLV Env, was identified as the putative key residue responsible for IS activity (Fig. 1A). To test this hypothesis, we mutated this amino acid and substituted an arginine for the glutamic acid at position 14 of the FeLV Env ISD (Fig. 1A), as described for MLV Env (35, 36). We first checked that this substitution had not significantly affected the structural properties of the protein via an infectivity assay using Env-defective MLV retroviral particles pseudotyped with either the wt or the mutant FeLV Env. As shown in Fig. 1C, mutation of the E14 residue only slightly altered FeLV Env infectivity, a result consistent with detection by flow cytometry of equivalent levels of wild-type and mutant FeLV Env proteins in transfected 293T cells (mean fluorescence intensity values after background subtraction were 91 and 94 for the FeLV Env wt and mutant, respectively). We then tested the E14R mutant protein for its IS activity, using the in vivo rejection assay described above. As illustrated in Fig. 1B, the tumors expressing the mutant FeLV Env displayed a negative immunosuppression index, demonstrating that the E→R mutation results in inhibition of the IS activity of the FeLV Env ectodomain, as similarly observed for the MLV control.

Vaccine-induced immune response.

Our previous work demonstrated that the mutation of the IS activity of retroviral envelopes results in increased immune responses toward vaccine antigens (36). We therefore set out to test the impact of the E→R mutation in the FeLV envelope on the immunogenicity and protective efficacy of a canarypox virus-vectored vaccine (ALVAC-FeLV). Two ALVAC-FeLV vaccines were generated (see Materials and Methods), both expressing the FeLV env and gag/pro genes, with vCP2295 carrying the wild-type env and vCP2296 carrying the E→R mutated env. Upon vaccination of the cats, the vCP2296 vaccine induced a higher frequency of FeLV-specific IFN-γ-secreting cells than the vCP2295 control vaccine (Table 2; Fig. 2A), as measured by ELISpot assay using Gag/Pro/Env overlapping peptides (see the description of the vaccination protocol and assay in Materials and Methods). This difference was observed for both the Gag/Pro-specific and Env-specific responses (Fig. 2A). The number of cats with FeLV gag/pro/env-specific IFN-γ producing cells was also significantly higher in the vCP2296-vaccinated group than in the vCP2295-vaccinated group (Fisher's exact test, P = 0.04): only two out of the 10 cats vaccinated with vCP2295 had responses above 10 spots, whereas seven out of the 10 cats vaccinated with vCP2296 did. Conversely, the median frequency of FeLV gag/pro/env-specific IL-10-producing cells tended to be lower in vCP2296-vaccinated cats than in vCP2295-vaccinated cats (Table 2; Fig. 2B). The number of cats with FeLV gag/pro/env-specific IL-10-producing cells was significantly higher in the vCP2295-vaccinated group than in the vCP2296-vaccinated group (Fisher's exact test, P = 0.02): all vCP2295-vaccinated cats had FeLV-specific IL-10-producing cells, whereas half of the vCP2296-vaccinated cats had no detectable IL-10-positive cells. As a result, the ratio of FeLV gag/pro/env-specific IFN-γ- to IL-10-positive cells was higher in vCP2296-vaccinated cats than in vCP2295-vaccinated cats (Table 2; Fig. 2C). This difference in the IFN-γ/IL-10 ratio concerned both FeLV gag/pro (stimulation with Gag/Pro peptides only)- and FeLV env (stimulation with Env peptides only)-specific responses (Fig. 2).

TABLE 2.

Vaccine-induced immune response at day 35 (FeLV-specific IFN-γ- or IL-10-positive cells) and FeLV status after challenge

Vaccine	Cat	FeLV-specific SFC/106 PBMC		IFN-γ/IL-10 ratioa	Persistent FeLV p27 antigenemia postchallenge
		IFN-γ positive	IL-10 positive
vCP2295 (native env gene)	1	58	26	2.2	−
	2	6	56	0.1	−
	3	0	90	0	+
	4	0	352	0	+
	5	0	134	0	−
	6	100	16	6.3	−
	7	6	666	0	+
	8	4	548	0	+
	9	2	346	0	+
	10	0	1,300	0	+
vCP2296 (mutated env gene)	11	700	0	700.0	−
	12	40	522	0.1	+
	13	324	182	1.8	+
	14	40	60	0.7	+
	15	100	0	100.0	−
	16	4	0	4.0	−b
	17	72	0	72.0	−
	18	44	174	0.3	−
	19	0	0	0	−
	20	0	536	0	+

^aIn the absence of a spot in the IL-10 ELISpot assay, a value of 1 was used for the calculation of the ratio.

^bCat 16 was transiently antigenemic (positive at weeks 3 and 6 postchallenge).

FIG 2.

Vaccine-induced specific responses. (A) Vaccine-induced FeLV gag/pro/env (left)-, gag/pro (middle)- and env (right)-specific responses measured by IFN-γ ELISpot assay at 1 week after vaccination (day 35). The frequency of FeLV-specific IFN-γ-producing cells (expressed in log₁₀) was higher in vCP2296-vaccinated animals (KW test, P = 0.07 for env and P = 0.05 for gag/pro; ANOVA, P = 0.05 for env/gag/pro-specific responses). (B) Vaccine-induced FeLV gag/pro/env (left)-, gag/pro (middle)- and env (right)-specific responses measured by IL-10 ELISpot assay at 1 week after vaccination (day 35). The median frequency of FeLV Env-specific IL-10-producing cells (expressed in log₁₀) was higher in vCP2295-vaccinated cats (KW test, P = 0.02 for env-, P = 0.23 for gag/pro-, and P = 0.08 for env/gag/pro-specific responses). (C) Ratio of vaccine-induced FeLV gag/pro/env (left)-, gag/pro (middle)- and env (right)-specific IFN-γ-producing cells to vaccine-induced FeLV-specific IL-10-producing cells, expressed in log₁₀ (day 35). The median ratio was higher in vCP2296-vaccinated cats (KW test, P = 0.01 for env-, P = 0.04 for gag/pro, and P = 0.03 for env/gag/pro-specific responses). The box-and-whisker plots show the median, lower, and upper quartiles as well as extreme values.

Protection against FeLV challenge. (i) First study.

In the first study, cats were vaccinated with a low dose of close to one PD₅₀ of ALVAC-FeLV expressing a nonmodified envelope glycoprotein (10^5.7 CCID₅₀/dose). At 2 weeks after vaccination, all cats were inoculated with 10^4.2 CCID₅₀ of a highly virulent FeLV strain via the oro-nasal route. In the control group (group C), 100% of the cats became persistently antigenemic and were therefore considered not protected (Table 3). In the vCP2295-vaccinated group (group A), 40% of cats were protected against p27 persistent antigenemia: 4 cats were never found positive, and 6 cats presented a persistent antigenemia. In the vCP2296-vaccinated group (group B), 60% of cats were protected against persistent antigenemia. Five cats were never found positive, one cat had a transient antigenemia, and four cats had a persistent antigenemia. In this assay, the incidence of antigenemia was significantly different between groups (Fisher's exact test, P = 0.03), but in pairwise comparisons, a trend toward a statistical significance was observed only between the vCP2296-vaccinated group and controls (adjusted P values with Bonferroni's method were as follows: B versus C, P = 0.06; A versus C, P = 0.260; and A versus B, P = 1).

In group C and in those cats from groups A and B which became persistently antigenemic after challenge, the mean proviral load reached 10^4.9 to 10^5.1 copies/50 × 10³ PBMC at 6 weeks after challenge and remained fairly stable on average until the end of the study at 15 weeks postchallenge. Interestingly, the proviral loads of some persistently antigenemic cats in the vaccinated groups decreased at the end of the follow-up period (Fig. 3A). Proviral loads in the controls were consistent with those reported in other studies (8, 9, 10, 17).

FIG 3.

FeLV proviral loads after challenge in the first (A) and the second (B) studies. Proviremia was expressed in logarithm (log) provirus copy number/50,000 WBCs. Green lines represent cats which were not persistently antigenemic. Red lines represent cats considered persistently antigenemic (cats found positive for circulating p27 on at least 5 occasions and/or positive at the end of the follow-up period).

In vCP2295-vaccinated cats which did not become persistently antigenemic, the average proviral load reached 10^2.6 copies/50 × 10³ PBMC at 3 weeks after challenge and decreased until 15 weeks postchallenge (10^1.2 copies on average). Similarly, in vCP2296-vaccinated cats which resisted persistent antigenemia, the average proviral load peaked at 10^2.0 copies/50 × 10³ PBMC at 3 weeks after challenge and decreased to 10^0.8 copies on average. In 3 vCP2296-vaccinated cats, the proviral load remained below 100 DNA copies per 50 × 10³ PBMC. The mixed-model analysis with repeated measures showed a significant difference between the vaccinated groups and the control group (A versus C, P = 0.04; B versus C, P = 0.01; and A versus B, P = 0.56).

(ii) Second study.

In the second study, cats were vaccinated with higher doses, namely, 10^6.8 CCID₅₀/dose for vCP2295 (group A) and 10^6.4 CCID₅₀/dose for vCP2296 (group B). At 2 weeks after vaccination, all cats were challenged as in the first study. In the control group (group C), 90% of cats became persistently infected after challenge. In the vCP2295-vaccinated group, 80% of cats were protected against FeLV infection (Table 3). One cat was found to be positive at the end of the study (i.e., on day 147) and was considered persistently antigenemic. In the vCP2296-vaccinated group, despite the lower vaccine dose than that of the vCP2295 vaccine, 90% of cats were protected against p27 persistent antigenemia and were never found positive. Both “vaccinates-versus-controls” pairwise comparisons were significant (A versus C, P = 0.02, and B versus C, P = 0.003). No difference between vaccinated groups was evidenced (A versus B, P = 1).

Cats from groups A and B presented a proviremia ranging from 0 to 10^4.8 copies/50 × 10³ PBMC at 3 weeks after challenge (Fig. 3B). Proviremia then decreased in most cats until 15 weeks postchallenge (10^1.5 and 10^0.8 copies/50 × 10³ PBMC on average in groups A and B, respectively). In the cat that became antigenemic on day 147, proviremia remained low over the whole follow-up period but increased sharply at the end of the study (10^5.6 copies/50 × 10³ PBMC on day 147). Proviremia in all cats but one in control group (group C) remained high until the end of the study. Mean proviremia ranged from 10^4.5 to 10^4.9 copies/50 × 10³ PBMC, consistent with the first study. According to the mixed-model analysis on repeated measures, the average provirus loads in groups A and B were significantly lower than that in group C over the whole period of observation (A versus C and B versus C, P < 0.0001; A versus B, P = 0.99).

(iii) Vaccine dose response.

Since the two studies were carried out under the same standardized conditions and resulted in comparable infection rates in the controls, they could be combined to determine the dose response by logistical regression (16) and to estimate the dose which would provide 80% protection (PD₈₀) against a validated challenge (≥80% of persistent antigenemia in controls). Those criteria are derived from the monograph on feline leukemia vaccines of the European Pharmacopeia. Accordingly, based on the results from the whole set of vaccinated animals (i.e., 60 cats), the estimated PD₈₀s are 10^6.8 CCID₅₀/dose for vCP2295 and 10^6.0 CCID₅₀/dose for vCP2296 (Table 3).

Relationship between immune response and protection against FeLV challenge.

A cat was considered protected if it was not persistently antigenemic. For both vCP2295- and vCP2296-vaccinated cats, the antigenemic status was consistent with proviral loads: cats that did not develop persistent antigenemia were those that were able to control their proviremia, as illustrated by proviral loads usually lower at peak (3 weeks after challenge) and progressively declining to values below 100 copies/50 × 10³ PBMC (Fig. 3). To analyze further the correlation between the protection status and the immune response induced upon vaccination, we pooled the data in Table 2 from all the vaccinated cats. The IFN-γ/IL-10 ratios for FeLV-specific responses on day 35 after vaccination were compared between vaccinated cats which were protected against FeLV persistent antigenemia and vaccinated cats which became persistently antigenemic (Fig. 4). A close to 2-log-higher median IFN-γ/IL-10 ratio was observed in protected cats than in persistently antigenemic ones, and this was statistically significant (Table 2; Fig. 4). This higher IFN-γ/IL-10 ratio in protected cats than antigenemic ones was observed for each of the env-, gag/pro-, and gag/pro/env-specific responses (Fig. 4).

FIG 4.

Ratio of vaccine-induced FeLV-specific IFN-γ-producing cells to FeLV-specific IL-10-producing cells in cats protected against persistent antigenemia versus antigenemic cats, for the whole set of vaccinated animals (both vCP2295 and vCP2296). The data shown are for the env (A)-, gag/pro (B)-, and gag/pro/env (C)-specific responses at day 35. The median ratio was higher in cats that did not develop persistent antigenemia (KW test, P = 0.003 for env-, P = 0.02 for gag/pro-, and P = 0.004 for env/gag/pro-specific responses). The box-and-whisker plots show the median, lower, and upper quartiles as well as extreme values.

DISCUSSION

A canarypox virus expressing the env and gag/pro genes of FeLV (ALVAC-FeLV) was shown to fully protect cats against persistent antigenemia after FeLV challenge (16). Previous attempts to further increase the efficacy of the ALVAC-FeLV and thereby reduce its protective dose by deleting a short domain of p15E (22), which was reported to be immunosuppressive, were unsuccessful. Deletion of this domain did not increase the efficacy of the vaccine, probably because the envelope glycoprotein was not properly expressed, impairing the FeLV Env-specific immune response. Here, we compared the immunogenicities of two ALVAC-FeLVs, vCP2295 and vCP2296, which differ only by the replacement of a glutamic acid (vCP2295) by an arginine at position 14 of the IS domain (vCP2296). We showed that this mutation reduced the immunosuppressive activity of the envelope-associated IS domain, preserved the infectivity of the envelope protein, and improved the efficacy of the vaccine.

Protection against FeLV infection is mediated by an early FeLV env- or gag-specific CTL response (19, 20). Here, the immunogenicities of vCP2295 and vCP2296 ALVAC-FeLV were assessed by measuring the frequencies of IFN-γ- and IL-10-producing cells following stimulation with FeLV Env or Gag/Pro peptides. Consistent with the relatively weak immunogenicity of FeLV antigens reported in previous works (25), a low IFN-γ response was detected in vCP2295-vaccinated cats (Table 2). Remarkably, vCP2296 induced a higher frequency of both FeLV env- and FeLV gag/pro-specific IFN-γ-positive cells than vCP2295 (Table 2; Fig. 2). Similarly, a UV-inactivated Friend murine leukemia virus in which the IS domain of the envelope glycoprotein was mutated induced a stronger production of IFN-γ by CD4⁺ and CD8⁺ T cells in vaccinated mice than its wild-type counterpart (36). Since the mutation that we introduced in the ALVAC-FeLV minimally alters antigenicity, with only one amino acid being mutated and with the structure and fusion function of the mutated Env being unaltered, it is likely that the shift in the quality of the immune response observed for both Env and Gag/Pro antigens with the vCP2296 vaccine results from the knocking down of the Env-dependent immunosuppressive activity. Conversely, in cats vaccinated with vCP2295 expressing the native env gene, the frequency of FeLV-specific IL-10-positive cells was higher than in vCP2296-vaccinated cats. IL-10 is known as a cytokine with broad immunosuppressive effects. It is one of the most common subversion effectors used by viruses to escape the host immune response (42–44), and it plays a pivotal role in viral persistence (45, 46). IL-10 can suppress CD4⁺ and CD8⁺ T cell responses and may therefore be detrimental for vaccine-induced responses (47). Indeed, we show that a high frequency of IL-10-producing cells is associated with a lack of protection against persistent antigenemia (Table 2) and that all the cats that had no detectable IL-10-producing cells were protected against persistent antigenemia. The nature of the IL-10-producing cells has not been assessed, but cells being induced by FeLV-derived peptides could be FeLV-specific T helper cells, adaptive regulatory T cells, or CD8⁺ T cells (48). However, we cannot rule out that the IS domain of FeLV p15E could also induce IL-10 production by other cells such as macrophages, DCs, or natural regulatory T cells; a difference in the way that vCP2295 and vCP2296 modulate IL-10 production by antigen-presenting cells would then contribute to the orientation of the FeLV-specific immune response, with a stronger shift toward a Th1 type of response for vCP2296.

Our work clearly establishes that the cats protected against an FeLV challenge have close to a 2-log-higher ratio of FeLV-specific IFN-γ- to IL-10-producing cells than the nonprotected cats (see “protected” versus “antigenemic” cats in Fig. 4), whatever the nature of the ALVAC-FeLV vaccine. This result is in accordance with the known importance of cell-mediated immune responses for protection against FeLV (25). Since the IFN-γ/IL-10 ratio for FeLV-specific responses is significantly higher for the cats vaccinated with vCP2296 than for those vaccinated with vCP2295 (see above) (Table 2 and Fig. 2), this characteristic feature of the vCP2296-induced immune response is predictive of and consistent with a better protection efficacy.

In terms of vaccine dose response, vCP2295 and vCP2296 were compared in two independent FeLV challenge studies, using low doses of ALVAC-FeLV close to the PD₅₀ and PD₈₀ of vCP2295, respectively. In both studies, protection was higher in the vCP2296 group than in the vCP2295 group (Table 3). Although the difference in the protection rates between vCP2295 and vCP2296 within each study was of limited extent (for ethical reasons, the number of animals was limited), calculation by logistical regression of the PD₈₀ values as derived from the whole set of data (the two series of assays, 60 cats) provided a PD₈₀ for vCP2296 that was 0.8 log unit lower than that for vCP2295 (i.e., the vCP2295 dose required to achieve 80% protection is 6 times higher than that for vCP2296). This dose difference adds to the above-mentioned difference in the extent and nature of the immune responses elicited by the two vaccines before challenge and strengthens the difference between the two vaccines. It is noteworthy that in the context of further work on the sole optimized vCP2296 construct (10 cats vaccinated with 10^5.6 CCID₅₀ vCP2296 and 10 controls, twice), similar protection rates were obtained (data not shown), supporting the validity and reproducibility of our well-controlled FeLV challenge. The values were higher than those for vCP2295 in the present assay, as well as for the ALVAC-FeLV vaccine with a native env gene previously published using the same challenge model (16).

The unraveled differences between the effects of the two ALVAC-FeLV vaccines can be solely assigned to the one amino acid difference in the env gene carried by the vaccine—all the more as the canarypox virus in cats is a nonreplicative vector. The canarypox virus per se has been shown to stimulate innate immunity, in particular the production of type I interferons, the maturation of dendritic cells (DCs), the secretion of tumor necrosis factor alpha (TNF-α), and the production of IFN-γ by natural killer (NK) cells (49–51). This natural adjuvant-like property of the vector may counteract to some extent the detrimental IS activity of FeLV Env expressed by the “wild-type” ALVAC-FeLV vaccine and explain why a high dose of ALVAC-FeLV expressing the native env gene is protective despite its IS activity. This makes the difference between vCP2295 and vCP2296 even more remarkable.

Conclusion.

This study clearly illustrates the immunosuppressive activity of the FeLV envelope protein carried by its IS domain and extends the conclusion previously derived with the murine leukemia virus model where inhibition of IS activity was demonstrated to improve the immunogenicity of recombinant Env and inactivated virions (36). IS activity most probably represents one “strategy” to downregulate the host immune response and avoid viral clearance. When introduced in a vaccine, the Env-associated IS domain has a direct negative impact on its immunogenicity. Here, we establish for a bona fide veterinary vaccine that vaccine efficacy can be improved through a targeted mutation abrogating IS activity of the retroviral envelope protein. Thus, it could be a general strategy to identify immunosuppressive domains within viral antigens and to knock down this function via mutations minimally altering the structure of the antigen to generate more efficient vaccines. As such, the present study provides new promising issues for other vaccine approaches, including those against human viruses for which the same principles should apply. It might also improve the safety of vaccines without impacting their efficacy, by reducing the antigen dose or the amount of adjuvant required for protection.