DNA Immunization with Plasmids Encoding Fusion and Nucleocapsid Proteins of Bovine Respiratory Syncytial Virus Induces a Strong Cell-Mediated Immunity and Protects Calves against Challenge

Abstract

Respiratory syncytial viruses (RSV) are one of the most important respiratory pathogens of humans and cattle, and there is currently no safe and effective vaccine prophylaxis. In this study, we designed two codon-optimized plasmids encoding the bovine RSV fusion (F) and nucleocapsid (N) proteins and assessed their immunogenicity in young calves. Two administrations of both plasmids elicited low antibody levels but primed a strong cell-mediated immunity characterized by lymphoproliferative response and gamma interferon production in vitro and in vivo. Interestingly, this strong cellular response drastically reduced viral replication, clinical signs, and pulmonary lesions after a highly virulent challenge. Moreover, calves that were further vaccinated with a killed-virus vaccine developed high levels of neutralizing antibody and were fully protected following challenge. These results indicate that DNA vaccination could be a promising alternative to the classical vaccines against RSV in cattle and could therefore open perspectives for vaccinating young infants.

Bovine respiratory syncytial virus (BRSV) and human respiratory syncytial virus (HRSV) belong to the Pneumovirus genus of the Paramyxoviridae family (52). These two negative single-stranded RNA viruses share common genomic, antigenic, epidemiological, and pathological characteristics (62).

BRSV and HRSV are major causative agents of severe respiratory tract diseases in cattle and infants worldwide, respectively (20, 31, 62). Both BRSV infection and HRSV infection can remain asymptomatic or cause severe respiratory tract diseases leading sometimes to death (62). Seventy percent of calves exhibit a positive serological response against BRSV at the age of 12 months, and mortality can reach up to 20% in some outbreaks (31, 61). From figures available in industrialized countries, the number of annual HRSV infections worldwide can be estimated around 64 million and mortality could be as high as 160,000 (20).

For these reasons, efficient vaccines against HRSV and BRSV are needed. However, their development has been hampered since the dramatic vaccine failure in the 1960s. In fact, vaccination with formalin-inactivated, alum-adjuvanted virus predisposed children to a far more serious, and sometimes fatal, form of pathology in the case of natural infection (29). Subsequently, it was found that a similarly inactivated BRSV vaccine could induce strikingly similar immunopathology (47). Further studies in mice and cattle suggested that exacerbation of disease resulted from a polarized type 2 T-helper cell response characterized by increased production of interleukin-4 (IL-4) and IL-5 cytokines, high levels of immunoglobulin G1 (IgG1) and IgE, and a lack of BRSV-specific CD8⁺ T cells, resulting in enhanced pulmonary eosinophilia (10, 13, 18, 25, 27, 63, 67).

Recently, DNA vaccines have emerged as a promising alternative to the modified live and killed-virus (KV) vaccines. Direct immunization with naked DNA results in the production of immunogenic antigens in the host cell which can readily go through processing and presentation via both class II and class I pathways and engender long-lasting humoral and cell-mediated immunity. Furthermore, DNA vaccines mimic live attenuated virus in their ability to induce both humoral and cellular responses but are considered to be safer and to offer several technical advantages (21, 22). Finally, since the immunizing protein is not present in the vaccine preparation, plasmid DNA is not susceptible to direct inactivation by maternal antibodies (44).

So far, DNA vaccination against HRSV has been mainly investigated in mice or cotton rats (6, 8, 32, 33, 58). These studies demonstrated that plasmids encoding the HRSV fusion (F) or attachment (G) proteins primed both humoral and cell-mediated immunity and protected against HRSV infection without significantly enhancing pulmonary pathology following challenge. Despite these promising results, very few studies confirmed the ability of DNA vaccines to protect against RSV infection in a natural host. DNA immunization with plasmid encoding BRSV F or G protein primed the humoral response of young calves, reduced virus excretion, and partially protected them after experimental infection (48, 53). Similarly, DNA immunization against BRSV F and nucleocapsid (N) proteins was shown to be safe, immunogenic, and partially protective in infant rhesus monkeys (64). Even if these reports highlight the potential of DNA vaccination, it seems that the efficacy of this strategy has to be improved in terms of the quality and intensity of the response induced.

Codon optimization and protein boost following DNA vaccination are two commonly used methods that improve the efficacy of DNA immunization (21, 66). In this report, we designed codon-optimized plasmids encoding BRSV F and N proteins and assessed their immunogenicity in young calves.

MATERIALS AND METHODS

Plasmids.

Full-length nonoptimized F and N genes of BRSV were amplified by reverse transcription-PCR (RT-PCR) from viral mRNA extracted from cell culture supernatant infected with the BRSV strain RB94 as previously described (7). Synthetic constructs carrying BRSV F (Fsyn) and N (Nsyn) genes were manually designed by reverse translation of the amino acid sequence of the RB94 strain in order to avoid negatively cis-acting sequences and restore bovine codon usage. The prediction of splicing sites was performed at the NetGene2 server (http://www.cbs.dtu.dk/services/NetGene2/) (9). The presence of other cis-acting elements like the polyadenylation sequence (AAUAAA) (65) or the Shaw-Kamen instability motif (AUUUA) (49) was detected manually. The CUSP software available at EMBOSS (http://emboss.sourceforge.net/) was used to restore bovine codon frequency. The codon adaptive index (CAI), which is a measurement of the relative adaptativeness of the codon usage of a gene compared with the codon usage of highly expressed genes, was calculated using the CAI software available at EMBOSS with a user-defined set of codon relative adaptiveness values based on codon usage in a compendium of bovine genes (Ebovsp.cut) (ideal value of 1.0). The resulting synthetic sequences for F and N genes were submitted to GenBank under accession no. AJ971803 and AJ971804, respectively, and were provided by Geneart GmbH (Germany). Native and optimized genes encoding F (pF and pFsyn, respectively) and N (pN and pNsyn, respectively) proteins were cloned into the HindIII-BamHI window of the pCDNA3.1 plasmid (Invitrogen) with an ATG initiation codon presented in an optimized Kozak sequence context (30). All plasmids were restriction verified, sequenced, and purified with Endofree plasmid Giga kits (QIAGEN) according to the manufacturer's instructions.

Antibodies.

Monoclonal antibodies specific for BRSV F (AK13A2 and AK8D1) were produced by immunizing BALB/c mice with the human Long strain and have been previously described (35). Mab19 was kindly provided by G. Taylor (Institute for Animal Health, Compton, United Kingdom) and has been described by Arbiza et al. (4). Monoclonal antibodies specific for BRSV N (AL5G1 and MoAbα-N) were produced by immunizing BALB/c mice with the RB94 strain by a protocol similar to that described by Matheise et al. (35). The specificity of the AL5G1 and MoAbα-N antibodies was tested by immunoblotting and immunoprecipitation procedures (data not shown). The K71 bovine polyclonal anti-BRSV antibody was produced in gnotobiotic calves after repeated inoculation with the RB94 BRSV strain and was provided by G. Wellemans (Veterinary and Agrochemical Research Centre, Brussels, Belgium).

Cell culture.

Vero and COS-7 cells were grown in minimal Eagle's medium (Gibco) supplemented with 10% fetal calf serum, 2.5 mM glutamine, and antibiotics.

Viruses.

BRSV RB94 was grown on Vero cell monolayers. The challenge inoculum consisted of lung lavage fluid collected 6 days after intratracheal inoculation of a calf with the BRSV strain VRS 3761 (kindly provided by Gilles Meyer, Ecole Nationale Vétérinaire de Toulouse, France). This strain was successively passaged on three young calves by a protocol similar to that described by Tjørnehøj et al. (56). This inoculum contained 10^3.5 50% tissue culture infective doses (TCID₅₀)/ml and was free of bovine viral diarrhea virus, bovine herpesvirus 1, bovine parainfluenza virus 3, bovine coronavirus, bovine adenovirus 5, endotoxins, bacteria, and mycoplasmas.

In vitro expression of plasmids in COS-7 cells.

COS-7 cells were grown in 24-well plates at 8 × 10⁴ cells per well and were transiently transfected with 400 ng of recombinant or empty pcDNA3.1 plasmids using Lipofectamine as described by the manufacturer (Invitrogen). Twenty-four hours after transfection, monolayers were fixed in 4% paraformaldehyde and the expression of the F and N proteins was checked by an indirect immunostaining procedure using bovine polyclonal bovine anti-BRSV serum (K71) and monoclonal antibodies specific for the F protein (AK13A2, AK8D1, and mAb19) or for the N protein (AL5G1 and MoAbα-N).

DNA immunization of mice.

Four groups of five 6- to 8-week-old, BALB/c mice were housed in animal facilities. Mice were immunized twice by intramuscular injections with 50 μg of plasmid construct (pCDNA3, pF, pN, pFsyn, or pNsyn). Three weeks after the second immunization, sera were collected in each group and pooled, and BRSV-specific IgG levels were measured by enzyme-linked immunosorbent assay (ELISA) performed as previously described (47). Briefly, RB94 cell cultures from classic culture were sedimented by an overnight centrifugation at 90,000 × g. The pellet was suspended in phosphate-buffered saline (PBS) containing 0.1% N-octyl β-d-glucopyranoside (Sigma) for permeabilization of virus envelopes. A predetermined optimal dilution of this preparation was adsorbed on polyvinyl microplates (Maxisorp; Nunc) overnight at 4°C. Plates were washed with PBS-0.1% Tween 20 and blocked with 100 μl/well of hydrolyzed casein in PBS-0.1% Tween 20 for 2 h at room temperature. Serial fivefold dilutions of murine sera (50 μl/well) were added to wells, and the plates were incubated for 1 h. Goat anti-mouse IgG antibody conjugated to horseradish peroxidase (Dako) was used as secondary antibody at a dilution of 1:2,000 in PBS containing hydrolyzed casein. Antibody titer was defined as the logarithm of the inverse of the dilution giving an optical density (OD) value two times over that of the corresponding preimmune serum.

Immunization of calves.

Seventeen 1-month-old calves, deprived of colostrum, were housed in separate isolation rooms. The calves were free of bovine viral diarrhea virus, bovine herpesvirus 1, BRSV, and bovine parainfluenza virus 3. As shown in Table 1, three groups of four calves were immunized intramuscularly and intradermally twice at a 3-week intervals (day 0 and day 21) with either 250 μg of each of the pFsyn and pNsyn plasmids (F+N/KV and F+N) or 500 μg of the empty pCDNA3.1 vector (pCDNA3/KV). The other calves received saline. Three weeks after the second DNA vaccination (day 42), the calves of the F+N/KV and pCDNA3/KV groups were vaccinated once by intramuscular immunization with 5 ml of a commercially available, Quil-A-adjuvanted vaccine, containing 10^5.5 TCID₅₀ of inactivated BRSV (Bovipast; Intervet). The other groups received saline solution. Three weeks after the third vaccination (day 63), all of the calves were challenged with BRSV strain VRS 3761 by intratracheal injection of 15 ml of inoculum. One calf (control) was not infected.

TABLE 1.

Study design

Group (no. of calves)	Vaccination(s)			Challenge (day 63)	No. of calves euthanized
	Day 0	Day 21	Day 42		Day 67	Day 69	Day 82
Mock (4)	Saline	Saline	Saline	BRSV	1	1	2
F+N/KV (4)	pFsyn+pNsyn	pFsyn+pNsyn	KV	BRSV	1	1	2
F+N (4)	pFsyn+pNsyn	pFsyn+pNsyn	Saline	BRSV	1	1	2
pCDNA3/KV (4)	pCDNA3.1	pCDNA3.1	KV	BRSV	1	1	2
Control (1)	Saline	Saline	Saline	Saline	0	1	0

All calves were bled weekly for examination of humoral and cellular responses. Bronchoalveolar lavage (BAL) fluids were collected from days 63 to 77 by instillation and aspiration of 50 ml of PBS directly into the lungs using a 120-cm-length PVC dual-flow tube (Vygon, Belgium). On days 68, 70, and 82, one or two calves of each group were euthanized. Immediately after euthanasia, lungs were excised and macroscopic lesions were photographed and recorded on a standard lung diagram to score the extent of pneumonic consolidation as previously described (56). A score of 0 was given to a lung without macroscopic lesion. A score of between 1 and 5 was given according to the extent of the lung lesions: 1 represents 1 to 5%, 2 represents 5 to 15%, 3 represents 15 to 30%, 4 represents 30 to 50%, and 5 represents >50%. Finally, six pieces of tissue from the right cranial, middle, and caudal lobes (two pieces in each lobe) were excised.

Evaluation of clinical signs.

Clinical signs were recorded daily after challenge. Respiratory rate was measured via a stethoscope for a full minute. Rectal temperature, the presence of nasal discharge, spontaneous or induced cough, lung sounds, depression, and anorexia were also noted. Examination of calves was done by qualified people who were not aware of group assignments. Clinical scores were then assigned to each calf as previously described (23).

Humoral immune response of calves.

The ELISA anti-BRSV kit (CER, Belgium) was used to determine BRSV-specific antibody titers. IgG, IgG1, or IgG2 was detected by the following horseradish peroxidase-conjugated antibodies: a rabbit anti-bovine IgG antibody (Sigma), an IgG1-specific monoclonal antibody 1C8 (kindly provided by J.-J. Letesson, University of Namur, Belgium), or a sheep anti-bovine IgG2 (Serotec), respectively. All sera were tested at a 100-fold dilution, and the corrected OD (COD) was calculated for each serum by subtraction of the OD of the serum on control antigen from the OD of the same serum on BRSV antigen. The COD was then expressed as a percentage of the COD of a positive reference serum (K71).

Neutralizing antibody titers were determined in a virus seroneutralization assay as previously described (47). Briefly, twofold-diluted sera were incubated with 100 TCID₅₀ of RB94 virus for 1 h at 37°C and added on Vero cells grown in 96-well culture microplates. Each dilution was tested in four wells. After 4 days of incubation, monolayers were fixed in 4% paraformaldehyde and viral expression on the monolayer was revealed by an immunoperoxidase assay with a bovine anti-BRSV serum (K71). Virus neutralization titers were expressed as the reciprocal of the highest dilution of serum that completely prevented viral expression.

Lymphocyte proliferation assay.

Heparinized blood samples were 10-fold diluted in RPMI 1640 medium (Gibco) supplemented with 2 mM glutamine, 5 × 10⁻⁵ M β-2-mercaptoethanol, 100 U/ml of gentamicin, and 2.5 μg/ml amphotericin B (Fungizone). Two hundred microliters of diluted blood samples was added to wells of U-bottom microplates (Nunc). After 6 days of incubation with 20 μl of BRSV antigen (sonicated and heat-inactivated culture supernatant of Vero cells infected with RB94 virus, containing 10⁵ TCID₅₀/ml) or 20 μl of control antigen (sonicated and heat-inactivated culture supernatant of mock-infected Vero cells), 0.8 μCi of methyl [³H]thymidine in 25 μl RPMI 1640 medium was added to each well. After 18 h of incubation, cells were harvested and radioactivity incorporated into the DNA was measured by liquid scintillation counting (Betaplate; Pharmacia). The results were expressed as stimulation indexes (SI) corresponding to the number of cpm obtained with the BRSV antigen divided by the number of cpm obtained with the control preparation. The lymphoproliferation was considered significant when the SI was equal or superior to 2.2. All tests were set up in sextuplicate cultures.

IFN-γ assay.

Gamma interferon (IFN-γ) production was measured after a 24-h in vitro stimulation of heparinized blood with 20 μl of the above-described BRSV or control antigens. The amounts of IFN-γ in the plasma were quantified by ELISA with the bovine IFN-γ EASIA kit (Biosource Europe). The results were expressed as the SI corresponding to the OD obtained with the BRSV antigen divided by the OD obtained with the control preparation. IFN-γ production was considered significant when the SI was equal or superior to 1.8.

Hpt assay.

The haptoglobin (Hpt) concentration was determined by measuring hemoglobin binding capacity in serum (50). Briefly, 10 μl of serum, blank, or standard was incubated at 25°C with 90 μl of methemoglobin for 10 min. Guaiacol (1.5 ml; Sigma) was added immediately followed by 0.5 ml of hydrogen peroxide. After 8 min incubation at 25°C, the OD at 470 nm was read. The Hpt concentration in the sample was calculated as follows: [(sample absorbance − blank absorbance)/(standard absorbance − blank absorbance)] × standard concentration (g/liter).

IF.

For immunofluorescence (IF), frozen lung samples from the right cranial, middle, and caudal lobes were sectioned at 5 μm, put on slides, and fixed in acetone for 15 min at 4°C. Slides were incubated with a biotinylated bovine anti-BRSV serum (K71) for 1 h at 37°C. After washing, fluorescein isothiocyanate (FITC)-conjugated streptavidin (Sigma) was added, the mixture was incubated for 1 h at 37°C, and the slides were observed on a fluorescence microscope. Slides were then examined without any information on the group assignments.

Histopathology.

Samples of the macroscopic lung lesions were taken, formalinized, embedded in paraffin wax, sectioned at 5 μm, and stained with hematoxylin and eosin or using May-Grünwald-Giemsa stains. Slides were then examined without any information on the group assignments.

RNA extraction and cDNA synthesis.

RNA extraction and cDNA synthesis were performed as previously described (7). RNA was extracted from 200 μl of lung homogenate or BAL fluids with an RNeasy RNA purification kit (QIAGEN) as recommended by the manufacturer. Purified RNA was dissolved in 20 μl of RNase-free water. Five microliters of RNA solution was used as a template for cDNA synthesis in the presence of 1 μl of 200 U of Moloney murine leukemia virus reverse transcriptase (200 U) (Invitrogen), 4 μl of 5× First Strand buffer (Invitrogen), 2 μl of 0.1 M dithiothreitol, 2 μl of 1× hexamer (Roche), 1 μl of 10 mM deoxynucleoside triphosphates (dNTPs; Roche), and 5 μl of RNase-free water. After a 50-min incubation at 37°C, the reverse transcriptase was inactivated at 70°C for 10 min.

Real-time PCR.

BRSV and bovine β-actin sequences were amplified by real-time PCR as previously described (7). Each PCR was performed on 1 μl of cDNA mixed with 12.5 μl of Master Mix Platinum Taq polymerase (Invitrogen), 0.5 μl of ROX dye (6-carboxyl-X-rhodamine; Invitrogen), 100 nM of each primer, 200 nM of probe, and RNase-free water for a final volume of 25 μl. Amplifications were performed as follows: 2 min at 50°C, 10 min at 95°C, and 45 cycles at 95°C for 15 s and 59°C for 1 min. The detection limit of the assay was 10² RNA copies. The standard curves demonstrated a linear range from 10³ to 10⁸ copies.

For IFN-γ and IL-4, real-time PCRs were performed with Taqman Master Mix kit (Applied Biosystem) in a final volume of 25 μl containing 5 μl of cDNA, 300 nM of each primer, and 900 nM of the probes described in Table 2. Amplifications were performed as follows: 10 min at 95°C and 40 cycles at 95°C for 15 s and 60°C for 1 min. In each assay, control RNAs were used to construct standard curves. The detection limits for IFN-γ and IL-4 were 10³ RNA copies, and standard curves demonstrated a linear range up to 10⁸ copies (J.-F. Toussaint, unpublished data).

TABLE 2.

Sequences of primers and probes used to detect IFN-γ and IL-4 by real-time RT-PCR

Target and primer or probe	Sequencea
Bovine IFN-γ
Forward primer	ATGGATATCATCAAGCAAGACATGTT
Reverse primer	TCATCCACCGGAATTTGAATC
Probe	FAM-TCCTCCAGTTTCTCAGAGCTGCCATTCA-TAMRA
Bovine IL-4
Forward primer	TTCTGCAGGGTTGGAATTGAG
Reverse primer	TCATTCACAGAACAGGTCTTGCTT
Probe	FAM-GTTCAAGCACGTGTGGCTCCTGTAGATAC-TAMRA

^aFAM, 6-carboxyfluorescein; TAMRA, 6-carboxytetramethylrhodamine.

Statistics.

The data collected were statistically analyzed using Graphpad Prism software. Two-way analysis of variance was used to compare the evolution of IgG titers, lymphoproliferation, and IFN-γ SI. Peak clinical signs and viral load values were compared by one-way analysis of variance and Tukey's test with a family error rate of 0.05. Logarithmic transformation was applied to fulfill the conditions of variances in homogeneity and normality when necessary.

RESULTS

Codon optimization of the BRSV F and N genes.

Codon optimization of the BRSV F and N genes was performed to restore bovine codon usage and remove potential negatively cis-acting elements such as splicing sites, polyadenylation signals, and Shaw-Kamen instability motifs. As a result of codon optimization, the GC contents of the F and N genes were increased from 35% and 39% to 47% and 51%, respectively, and the CAI was improved from 0.551 to 0.710 for the fusion gene and from 0.559 to 0.736 for the nucleocapsid gene (Table 3).

TABLE 3.

Effects of codon optimization

Gene	No. of instability sites native/optimizeda	CAI for native/optimized	No. (%) of native nucleotides altered/total	% GC content native/optimized
Fusion	6 P, 3 SK, 4 S/0	0.551/0.710	240/1,725 (14)	35/47
Nucleocapsid	1 SK, 18 S/0	0.559/0.736	150/1,176 (13)	39/51

^aP, polyadenylation signal; SK, Shaw-Kamen instability motif; S, splicing site.

In vitro protein expression of the pFsyn and pNsyn plasmids and immunogenicity in mice.

The impact of codon optimization on protein expression was investigated by transfection of COS-7 cells with equal amounts of native or optimized plasmids encoding BRSV F and N proteins. Twenty-four hours after transfection, protein expression was revealed by a polyclonal bovine anti-BRSV serum (K71) (Fig. 1). Expression of F or N proteins was detected in cells that received native or optimized plasmids, but no signal was observed in cells transfected with the empty pcDNA3.1 plasmid. Importantly, the results showed that expression levels were greater in cells transfected with codon-optimized plasmids. Similar results were obtained when monoclonal antibodies specific for different epitopes of the fusion protein (AK13A2, AK8D1, and mAb19) or the nucleocapsid protein (AL5G1 and MoAbα-N) were used instead of the polyclonal K71 serum (data not shown).

FIG. 1.

In vitro expression of pFsyn and pNsyn plasmids in COS-7 cells. COS-7 cells in 24-well plates were transfected with 400 ng of the plasmids described. Twenty-four hours after transfection, expression of the BRSV F and N proteins was revealed in an immunoperoxidase assay with a polyclonal anti-BRSV antibody (K71).

The immunogenicity of the plasmids was then investigated in mice. Five groups of four mice twice were given 50 μg of plasmid at a 3-week interval. Three weeks after the second immunization, sera were pooled and BRSV-specific antibodies were measured by ELISA. No antibody response was detected after injection of mice with either the empty pCDNA3 plasmid or nonoptimized vectors, but pFsyn and pNsyn immunizations elicited IgG titers of 3.01 and 1.86, respectively.

Overall, these results showed that the codon optimization of the BRSV F and N genes increased both their in vitro expression as well as their immunogenicity in mice.

BRSV-specific immune response of calves following vaccinations and challenge.

The global immune response against BRSV induced by coadministration of pFsyn and pNsyn plasmids was evaluated weekly by measuring levels of BRSV-specific IgG and lymphoproliferative response after antigenic stimulation of whole blood cells. Immunization with pFysn and pNsyn plasmids resulted in seroconversion 5 weeks after the second vaccination (day 56), as shown by stable antibody levels in F+N calves (Fig. 2A). KV induced the production of BRSV-specific antibodies 2 weeks after injection in the pCDNA3/KV groups and raised antibody levels in the F+N/KV group compared to the F+N calves (P < 0.01). Ten days after challenge (day 73), seroconversion of mock-vaccinated calves was observed as well as enhanced amounts of IgG in the other groups, especially in the pCDNA3/KV calves.

FIG. 2.

BRSV-specific immune response following vaccination and challenge. Calves were vaccinated twice with DNA (days 0 and 21) followed by protein boost (day 42) and challenge (day 63) at 3-week intervals. (A) IgG levels were measured by ELISA, and COD values were transformed to percentages of a standard positive serum. (B) In vitro lymphocyte proliferation was evaluated by measuring [³H]thymidine incorporation after antigenic stimulation of whole blood cells with control or BRSV antigen, and values were expressed as an SI. Bars represent means (+standard deviations) in each group on days 0, 21, 42, 56, 63, 69, and 73.

DNA priming induced a lymphoproliferative response 5 weeks after the second vaccination (day 56) in F+N and F+N/KV calves, as shown by higher SI compared to the mock-vaccinated and pCDNA3/KV groups (P < 0.05) (Fig. 2B). The KV boost was not able to stimulate lymphoproliferation in the pCDNA3/KV group even if it induced slightly higher SI values in F+N/KV calves compared to the F+N calves (P > 0.05). The challenge did not enhance the SI values measured in F+N and F+N/KV groups but it induced a lymphoproliferative response in pCDNA3/KV and mock-vaccinated calves on day 73.

Ig isotype profile and neutralizing antibody titers.

BRSV-specific humoral immunity primed by DNA immunization and KV boost was further investigated by analyzing the IgG isotype profile and neutralizing antibody levels on days 63 and 69. On these days, F+N calves had low but significant BRSV-specific IgG1 (Fig. 3A) and IgG2 (Fig. 3B) compared to mock-vaccinated calves (P < 0.05). Nevertheless, these antibodies did not possess a neutralizing activity (Fig. 3C). KV boost resulted in significantly higher levels of IgG1 (P < 0.01), IgG2 (P < 0.01), and neutralizing antibodies (P < 0.05) in F+N/KV and pCDNA3/KV animals compared to F+N and mock-vaccinated calves. Interestingly, on days 63 and 69, F+N/KV calves had significantly lower IgG1 levels (P < 0.05) and similar amounts of IgG2 but larger neutralizing antibody titers than pCDNA3/KV calves (P < 0.05). SN/IgG_total and IgG2/IgG1 ratios were then calculated for each calf in the F+N/KV and pCDNA3/KV groups. These ratios could not be determined in the two other groups because the titers before and after challenge were too weak. As shown in Fig. 3D, animals that received the F+N priming had higher IgG₂/IgG₁ and SN/IgG_total ratios than pCDNA3/KV calves (P < 0.05).

FIG. 3.

Immunoglobulin isotype profile and neutralizing antibody titers at challenge (day 63) and 6 days after challenge (day 69). IgG1 (A) and IgG2 (B) levels were measured by isotype-discriminating ELISA, and COD values were expressed as percentages of a standard positive serum. Neutralizing antibody titers (C) were determined by neutralization assay and expressed as the base 2 log of the maximum positive dilution. Bars represent means + standard deviations measured in each group on days 63 and 69. Individual IgG2/IgG1 and SN/IgG_total ratios (D) were calculated in the F+N/KV and pCDNA/KV groups on days 63 and 69.

Cytokine production ex vivo and in vivo.

The effect of DNA priming on the type of immune response induced was examined by quantifying cytokine production in vitro and in vivo. Secretion of IFN-γ was measured after antigenic stimulation of whole blood cells at challenge (day 63) and 6 days after challenge (day 69). As shown in Fig. 4A, DNA-primed calves (F+N and F+N/KV groups) displayed significant IFN-γ production on the day of the challenge and this response was slightly improved on day 69. Similarly to what we observed in lymphoproliferation assay, KV boost did not enhance SI measured in F+N/KV calves nor did it stimulate IFN-γ production in the pCDNA3/KV group.

FIG. 4.

Cytokine production in vitro and in vivo before and after challenge. (A) IFN-γ production was measured by ELISA after antigenic stimulation of whole blood cells with control or BRSV antigen, and values were expressed as SI. Bars represent means + standard deviations measured in each group on days 63 and 69. (B) Levels of IFN-γ (left panel) and IL-4 (middle panel) transcripts were quantified by real-time RT-PCR in lung homogenates of calves euthanized on days 67 and 69, using β-actin (β-act) as an internal reference, and ratios between IFN-γ and IL-4 mRNAs were calculated for each calf (right panel).

The amounts of IFN-γ and IL-4 mRNAs produced in the right caudal lung lobes of calves euthanized four (day 67) and six days (day 69) following challenge were measured by real-time RT-PCR. Levels of IFN-γ mRNA measured in the lungs of calves that received pFsyn and pNsyn priming (F+N and F+N/KV) were up to 100-fold higher than those of the other calves (Fig. 4B, left panel), confirming the results observed after in vitro stimulation of whole blood cells. Analysis of IL-4 mRNA levels did not reveal such differences between groups as there was no obvious IL-4 induction in the lung (Fig. 4B, middle panel). Finally, ratios between IFN-γ and IL-4 were calculated as a measure of Th bias for each calf (Fig. 4B, right panel). These ratios were higher in calves that received the pFsyn and pNsyn priming compared to mock-vaccinated and pCDNA3/KV calves, reflecting an increase in Th1 response.

Clinical signs.

After experimental infection on day 63, a clinical score was calculated for each calf as previously described (23). Whereas the mock-vaccinated calves had an important increase of rectal temperature and respiratory rate two days after infection (day 65), calves from either F+N, F+N/KV or pCDNA3/KV groups didn't display any clinical signs earlier than 4 to 5 days after infection (Fig. 5). In all groups, peak clinical scores were recorded between days 68 and 70. Mock-vaccinated calves had the most severe respiratory disease compared to pCDNA3/KV (P < 0.01), F+N (P < 0.001) and F+N/KV (P < 0.001) animals. DNA-primed calves displayed reduced clinical signs compared to pCDNA3/KV group (P < 0.05), and there was no significant difference between F+N and F+N/KV groups.

FIG. 5.

Clinical signs following challenge. Rectal temperature, respiratory rhythm, presence of nasal discharge, lung sounds, and anorexia were recorded daily after challenge, and clinical scores were calculated as previously described (56). Data represent means (+standard deviations) in each group from day 63 (challenge) to day 77.

Postmortem examination.

On days 67, 69, and 82, gross pneumonic lesions were recorded and scored as previously described (56). On day 67, only limited lesions were observed in cranioventral lung parts of F+N, pCDNA3/KV, and mock-vaccinated calves (Table 4). On days 69 and 82, extended consolidated areas were seen, covering up to 50% of the lung surface (score of 5) in the mock-vaccinated group. pCDNA3/KV and F+N calves displayed reduced lesions compared to mock-vaccinated calves, but consolidation was observed in cranioventral and caudal regions of the lung. Interestingly, only one F+N/KV calf had one spot of consolidated lung tissue in the cranial lobe. The assessment of macroscopical lesions was confirmed by measuring the concentration of an acute-phase-response protein, Hpt. Serum Hpt was detected in all groups from day 68 and peaked around days 70 to 72. The highest Hpt concentrations were measured in pCDNA3/KV and mock-vaccinated calves, and there was a good correlation between the magnitude of the peak Hpt concentration and the extent of lung lesions. Overall, the observed clinical signs were severe. Finally, viral load and expression of the virus in cranial, middle, and caudal lung lobes were assessed by real-time RT-PCR and immunofluorescence (IF), respectively. As shown in Table 4, real time RT-PCR revealed the presence of BRSV in all infected calves but the viral load measured in calves that received the F+N priming was lower than those in to mock-vaccinated and pCDNA3/KV calves. Viral expression was detected by IF in all four lobes of mock-vaccinated calves, in two lobes (cranial and caudal) of pCDNA3/KV calves, and on the cranial lobe of the F+N calf killed at day 69. Samples collected from the control calf that was mock infected did not display gross lesions and were free of BRSV.

TABLE 4.

Lesion score, peak Hpt concentration, and BRSV detection following challenge

Group and day euthanized	Lesion scorea	Peak Hpt concn (mg/ml)	Viral loadb,c	IF detectionc,d
Mock
67	1	0	4.4 ± 0.1	cr, ca, m
69	3	0	4.2 ± 0.2	cr, ca, m
82	5-3	1.80-0.68	−	ND
F+N/KV
67	0	0	1.6 ± 0.2	−
69	1	0.05	0.5 ± 0.3	−
82	0-0	0.28-0	−	ND
F+N
67	1	0	2.2 ± 0.2	cr
69	2	0.01	1.8 ± 0.3	−
82	1-2	0.14-0.79	−	ND
pCDNA3/KV
67	1	0	3.0 ± 0.6	cr, m
69	2	0	3.3 ± 0.2	cr, m
82	4-2	1.81-0.9	−	ND
Control
69	0	0	−	−

^aAs described by Tjornehoj et al. (56).

^bMean (±standard deviation) viral load expressed as the log of BRSV mRNAs per 10⁶ b-actin mRNAs detected in the right cranial, middle, and caudal lung lobes.

^c−, not detected.

^dLung lobes in which viral expression was detected by IF. cr, cranial; m, middle; ca, caudal; ND, not done.

Histopathological lesions.

Examination of pulmonary histopathology of the control calves revealed the classical signs of BRSV infection. Four days after infection (day 67), all sections showed local alveolar emphysema, local lobular thickening of the alveolar septa with syncytium formation, and localized peribronchiolar cuffing (Fig. 6A). On day 69, pCDNA3/KV and mock-vaccinated calves had severe bronchopneumonia (Fig. 6B) characterized by the presence of eosinophils in the peribronchiolar infiltrate (Fig. 6C). Eosinophils were observed between the hyperplastic epithelium, mucosa, and submucosa. Their presence was focal: sometimes solitary (e.g., epithelium) and sometimes in small groups. Sections of F+N/KV calf displayed moderate alveolar emphysema, scattered thickening of the alveolar septa, and syncytial formation (Fig. 6D). In the F+N calf there was prominent inflammatory reaction with infiltration of lymphocytes and neutrophils in the septa, near bronchi and bronchioli, and scattered sites of consolidation (Fig. 6E). On day 82, a prominent fibrovascular reaction, probably due to repair, was present on lung sections of pCDNA3/KV and mock-vaccinated calves, and to a lesser extent on sections of F+N calves (Fig. 6F). Thus, the only calves that did not display histological lesions were those from the F+N/KV group.

FIG. 6.

Histopathological features on lung section after challenge. (A, B, and F) Lung sections of mock-vaccinated calves on days 67 (A), 69 (B), and 82 (F). (C) Eosinophil infiltration (arrow) in pneumonic lung of a pCDNA3/KV calf on day 69. (D and E) General aspects of the F+N/KV (D) and F+N (E) calf lung sections on day 69.

Viral replication.

The real-time RT-PCR was used to determine the viral loads in BAL fluids. As shown in Fig. 7, the presence of BRSV was not detected before infection. Virus load peaked around day 67 in all calves and decreased around day 70, except in mock-vaccinated calves in which the decrease was observed at day 72. As expected, peak viral load was significantly higher in pCDNA3/KV and mock-vaccinated groups compared to F+N and F+N/KV groups (P < 0.001). Similar observations were made from the analysis of viral load in lungs (Table 4). At day 73, F+N and F+N/KV groups were free of BRSV, whereas viral RNA was detected up to day 75 and day 77 in pCDNA3/KV and mock-vaccinated groups, respectively.

FIG. 7.

Evolution of viral load in BAL fluids. Virus load was determined by real-time RT-PCR in BAL fluid samples collected on days 63, 66, 68, 70, 73, 75, and 77. Bars represent mean (+standard deviation) standardized amounts (BRSV mRNAs/10⁶ β-actin mRNAs) in each group.

DISCUSSION

DNA vaccines are promising alternatives to the classical modified live virus or KV vaccines. In fact, DNA immunization often primes a Th1 immune response and provides large numbers of memory B and T cells. Furthermore, DNA immunization might avoid the inhibitory effect of maternal antibodies (for reviews, see references 21, 22, and 44). Efficacy of DNA immunization against RSV F and G proteins has well been described in rodents (6, 8, 32, 33, 58). However, immunization of large animals like cattle or monkeys offered only partial protection and require more innovative approaches, like codon optimization and protein boost following DNA vaccination (48, 53, 64).

In this report, we designed and constructed codon-optimized plasmids encoding BRSV F and N proteins and evaluated the immunogenicity of these plasmids in calves. Construction of genes was performed in order to increase the overall GC content, avoid negatively cis-acting sequence motifs, and restore codon adaptation to frequently used bovine codons. We have shown that codon optimization of the BRSV F and N genes led to an increase in antigen levels expressed in mammalian cells and improved antibody response in mice. These results are consistent with similar codon optimization studies reported for genes encoded by different pathogenic organisms (1, 37-39, 43).

Vaccination of young calves with a DNA plasmid containing the BRSV F gene or recombinant vaccinia virus expressing the F or N proteins was shown to stimulate both humoral and cell-mediated immunity (16, 53, 54). More precisely, in these studies, the F gene construct rather primed humoral immunity while immunization against N rather promoted cellular immunity. Consistent with these reports, two immunizations with pNsyn elicited lower IgG levels than pFsyn but resulted in in vitro lymphoproliferation and IFN-γ production, whereas pFsyn did not (M. Tignon, unpublished results). Thus, coimmunization against F and N proteins should stimulate both humoral and cell mediated immunity and could be an interesting vaccine approach. To answer this question, young calves were vaccinated with both codon-optimized pFsyn and pNsyn plasmids. Moreover, in an attempt to increase the immune response against the virus, we boosted the DNA priming with a heterologous boost consisting of a BRSV KV vaccine. Recently, a similar strategy was shown to be efficient in the context of the bovine herpesvirus 1 (57). Here, two administrations of both pFsyn and pNsyn plasmids stimulated the two arms of the immune system, as shown by antibody production, in vitro lymphoproliferation, and IFN-γ production. Furthermore, the use of the KV vaccine considerably increased the levels of BRSV-specific antibody. Overall, these observations confirm the potential of coimmunization against F and N proteins as well as the efficacy of an alternative protein boost following DNA immunization in large animals.

DNA priming with pFsyn and pNsyn elicited lower levels of antibody than the KV vaccine and failed to induce the presence of a detectable neutralizing humoral response. Nevertheless, we have shown that pFsyn/pNsyn priming influenced the antibody response induced by KV vaccination. Indeed, pFsyn/pNsyn-primed calves had lower levels of IgG, especially IgG1, but higher neutralizing antibody titers after KV vaccination than nonprimed calves. As the presence of RSV-neutralizing antibodies is much more important than the absolute antibody levels in the serum (59, 60), our results suggest that pFsyn/pNsyn priming, even if it elicits low BRSV-specific IgG titers, improves the humoral immune response induced by the KV boost.

The role of cell-mediated immunity is more complicated and is a critical parameter in the outcome of RSV infection (for a review, see reference 40). In rodents and calves, T cells are essential for virus clearance but also appear to mediate immunopathology. The CD8 cells seem to play an important role in protection against the disease, possibly through RSV-specific cytotoxic activity, but a strong cytotoxic T lymphocyte response has been also implicated in the pathological outcome of vaccination (11, 12, 17, 25, 45, 51, 55). The role of CD4 Th cells is more uncertain and complicated, but the imbalance of Th1 and Th2 cytokines is likely to play a role in the pathogenesis of RSV disease.

DNA vaccination often promotes a Th1-biased immune response that is characterized by high levels of antigen-specific lymphoproliferative response and the production of large amounts of IFN-γ (5). After two inoculations of the pFsyn and pNsyn constructs, both IFN-γ production and lymphoproliferative response were detected. In contrast to what we observed for the humoral response, KV boost did not enhance the IFN-γ production in calves that were DNA primed and showed poor efficacy for the induction of both IFN-γ production and lymphoproliferative response in nonprimed calves before and after challenge. In a previous report, we demonstrated that this KV vaccine did not stimulate detectable or strong cell-mediated immunity even after two inoculations (28). To approach the problem of IL-4 production, we performed real-time RT-PCR to quantify the amounts of IFN-γ and IL-4 mRNAs in the lungs of calves during the first week of the infection. IFN-γ mRNA levels were increased in calves primed with pFsyn and pNsyn plasmids, whereas there was no clear difference in IL-4 transcripts between groups, confirming what we observed after whole-blood-cell sensitization. Thus, DNA priming using pFsyn and pNsyn plasmids seems to promote a Th1-biased immune response characterized by the production of IFN-γ both in vitro and in vivo. IFN-γ has been shown to promote the production of IgG2, resulting in a higher IgG2/IgG1 ratio (15). Here, in DNA-primed calves, we did not measure higher levels of IgG2. However, higher IgG2/IgG1 ratios were observed after challenge, further confirming that pFsyn/pNsyn priming elicits a Th1 immune response.

The protection conferred by these vaccination protocols was evaluated after experimental infection of calves using a BRSV field strain (VRS 3761) successively passaged on young calves. This method has previously been shown to induce an exceptionally severe pathology in mock-vaccinated calves compared to challenge procedures based on the usage of cell culture-passaged strains (56). In the present study, the first clinical signs of BRSV infection were recorded 3 days following challenge and they peaked 1 week after infection. Postmortem, consolidated areas covered up to 50% of the lung in mock-vaccinated animals. Histopathologically, syncytia, emphysema, alveolitis, and bronchiolitis with infiltration of numerous neutrophils and eosinophils appeared first, rapidly evolving in bronchopneumonia and a prominent fibrovascular reaction indicative of repair. As shown by others, the serum concentration of Hpt, an acute-phase-response protein, peaked 7 to 9 days after infection and the magnitude of the response was well correlated with the severity of the clinical signs and the extent of lung lesions (24). Thus, BRSV challenge resulted in severe pathology similar to that seen in naturally infected calves, allowing reliable evaluation of the protection conferred by each vaccine strategy.

Even if it was not sufficient to fully prevent histopathological changes and gross pneumonic lesions after challenge, the immune response generated by pFsyn and pNsyn plasmids clearly reduced both clinical signs of the infection and virus replication while KV failed to do so. In cattle and mice, secretion of IFN-γ was shown to contribute to NK activation and CD8 cytotoxic activity, resulting in viral clearance and resolution of the infection in the airways (14, 26, 41, 55, 68). Here, we have also shown that the production of IFN-γ was correlated with viral clearance as pFsyn/pNsyn-primed calves efficiently eliminated the virus after challenge. In contrast, KV vaccination, which elicited high levels of neutralizing antibody but did not induce IFN-γ production, failed to reduce viral replication and pneumonic lesions. Thus, it seems that the induction of a strong cell-mediated immunity is a critical parameter for the protection against BRSV infection as the pFsyn/pNsyn vaccine-induced cell-mediated immunity allowed viral clearance even in absence of a neutralizing humoral response. However, prevention of the disease was only achieved when calves were further vaccinated with the KV boost.

In contrast to previous reports, KV vaccination did not induce a strong lymphoproliferative response and did not result in enhanced pathology following challenge (2, 3, 19, 27, 46). However, a single inoculation of the vaccine poorly reduced viral load, clinical signs, histological changes, and gross pneumonic lesions compared to mock-vaccinated calves. Furthermore, it did not prevent eosinophilia which had previously been linked to immunopathology following challenge (25). These results are consistent with a previous report exploring the efficacy of the same KV vaccine (36). This supports the interest in DNA priming before KV vaccination. Similarly, KV vaccines formulated with CpG oligodeoxynucleotides have also been shown to result in a better Th1-type response and lack of immunopathology (34, 42).

In summary, we have demonstrated for the first time that DNA priming against BRSV F and N protein primed a Th1-biased immune response that reduces viral replication and partially protected calves against a highly virulent challenge. Furthermore, KV boost potentiated the immune response primed by DNA vaccination and this heterologous DNA-priming KV vaccine boost offered complete protection against BRSV. In a future work, immune response induced by specific heterologous prime-boost protocols against F or N protein as well as strategies for improving the efficacy of DNA vaccination in cattle like microparticle encapsulation will be evaluated.