Materials and Methods

Animals—Thirty-five mixed-breed dogs and Beagle dogs (males and females) that ranged from 26.3 to 42.6 weeks of age were used in the studies. None of the dogs had detectable antibodies to CIV, as determined by use of HI and VN tests at the commencement of the study. Dogs were housed in groups in environmentally controlled rooms, fed a commercially formulated diet, and maintained in accordance with the animal use and care guidelines of the Merial Institutional Animal Care and Use Committee.

Preparation of vaccines—Two canarypox-vectored EIV vaccines encoding HA proteins of A/Equine/OH/03 or A/Equine/KY/94 (designated vCP2242 and vCP1529, respectively) were used in the study. Optimized synthetic HA gene sequences of A/Equine/OH/03 were used for the generation of vCP2242. Synthetic HA gene was obtained by use of commercially availablesoftware^b for chemical synthesis of an array of oligonucleotides that encompass the gene. The oligonucleotides were assembled by use of a PCR-based strategy to generate the complete, full-length synthetic gene. For vCP1529, HA coding sequences of A/Equine/KY/94 with tailored 5′ and 3′ ends, which were obtained by standard, coupled reverse transcriptase–PCR techniques, were used. The HA genes (synthetic and generated by use of the reverse transcriptase–PCR technique) were subcloned into the canarypox virus vector to create the canarypox virus–recombinant viruses by use of essentially the same procedures described elsewhere⁴ for the creation of recombinant canarypox virus–vectored West Nile virus vaccine.

Briefly, genes encoding HA of A/Equine/OH/03 and A/Equine/KY/94 were subcloned into a canarypox C5 insertion vector (plasmid containing a vaccinia virus H6 promoter and the flanking arms of the canarypox C5 locus) to generate an expression cassette consisting of the HA gene under the control of the H6 promoter. To generate the canarypox virus recombinant, the insertion plasmid was transfected into primary chicken embryo fibroblast cells that were subsequently inoculated with canarypox virus. After incubation for 24 hours, the transfected-infected cells were harvested, sonicated, and used to screen for recombinant virus. The recombinant plaques were screened by use of an in situ plaque-lift hybridization method that used an HA-specific probe. After 4 sequential rounds of plaque purification, the recombinant confirmed by hybridization as 100% positive for the HA insert was amplified and used to prepare vaccine stocks (identified as vCP1529 and vCP2242), which were stored at −80°C. Expression of HA in cells infected with recombinant virus was detected by use of immunofluorescence staining and western immunoblotting that used a cocktail of HAspecific monoclonal antibodies. Both vaccines were diluted in PBS solution before being used to achieve the indicated target titers.

Experimental design—Dogs were randomly allocated into 1 of 4 groups. Dogs of group 1 (n = 8) were inoculated SC with vCP1529 (canarypox-vectored EIV [A/Equine/Kentucky/94]). The 9 dogs of group 2 were inoculated SC with vCP2242 (canarypox-vectored EIV [A/Equine/Ohio/03]). The 9 dogs of group 3 were vaccinated with vCP2242 via the transdermal route by use of a needle-free injection system^c that used a 13mm vial adapter with a nozzle size of 0.0024 cm. The 9 dogs of group 4 were not inoculated and served as negative control dogs. Dogs of groups 1 and 2 were inoculated SC in the shoulder area with 1 mL of vaccine, whereas dogs of group 3 were inoculated with 0.25 mL of vaccine that was administered transdermally in the proximal half of the hind limb at a location overlying the semimembranosus-semitendinosus muscle groups. Dogs were inoculated twice (days 0 and 28) with an estimated dose of 10^5.7 TCID₅₀ of the respective inoculum, regardless of the volume injected.

All dogs were examined for local reactions at the site of inoculation and for general clinical signs. Rectal temperature was recorded on days −1 through 7 and days 28 through 35. Blood samples were collected from all dogs on days −1, 7, 13, 21, 28, (before the second inoculation), 35, and 42. Serum was separated and stored frozen at −20°C until analysis.

Antibody determination—All sera were tested by use of HI and VN tests, and the laboratory investigators were not aware of the source of the serum samples with respect to group assignments at time of testing.

All sera were tested for antibodies against CIV (A/Canine/FL/43/04-PR) by use of an HI test at a reference laboratory.^d Briefly, sera were heat inactivated (56°C for 30 minutes). Sera were then treated with potassium periodate followed by incubation with turkey RBCs to remove nonspecific HAs and inhibitors. Four hemagglutination units of antigen were added to serial 2-fold dilutions of treated sera and incubated at 18° to 22°C for 30 minutes. Turkey RBCs were added, and plates were incubated for an additional 30 minutes. Wells were then examined for evidence of hemagglutination. The lowest dilution of serum tested was 1:8, and antibody titers were expressed as the end point dilutions.

A microneutralization (VN) test was performed in accordance with the method described in another study.⁵ A CIV isolate (A/Canine/NY/115809/05) was used for the test, and all serum and virus dilutions were made in minimal essential medium containing 2 μg of trypsin/mL. Briefly, heat-inactivated sera were serially diluted and mixed with an equal volume of CIV containing 100 to 300 TCID₅₀ CIV. After incubation at 37°C for 60 minutes, 100 ML of virus-antibody mixtures was transferred into wells containing 18- to 24-hourold monolayers of Madin-Darby canine kidney cells.

Plates were incubated at 37°C for 7 days and examined for cytopathic effects. Each serum sample was tested in triplicate, and antibody titers were determined by use of the Spearman-Karber method. When an end point titer could not be determined, the last titer tested was used for calculations (eg,≥3,548 was recorded as 3,548). A serum titer of 8 (HI tests) or 15 (1:15 was the lowest dilution used in VN tests) or higher was considered to be a positive result for CIV antibodies.

The protein identities between HA genes of A/Canine/FL/43/04, A/Canine/NY/115809/05 isolates, and HA genes expressed by vCP1529 and vCP2242 were determined by use of multiple-sequence alignment of amino acid sequences.⁶ The procedure was performed by use of a commercial software.^e

Statistical analysis—Statistical analyses were performed by use of a commercially available statistical program.^f A repeated-measures ANOVA was used to detect significant differences in antibody titers among groups. Values of P < 0.05 were considered significant.

Results

Animals—All dogs tolerated the inoculations well. None of the dogs had a systemic or local adverse reaction, regardless of the inoculum or route of administration.

Antibody responses—Antibody responses of dogs to 2 doses of vCP1529 or vCP2242 administered on days 0 and 28 were determined by use of the HI and VN tests (Table 1). On day −1, all dogs were seronegative by use of HI (< 8) and VN (< 15) tests. Control dogs remained seronegative throughout the duration of the study.

On day 7 (the first sampling point after the initial inoculation), 5 of 8, 6 of 9, and 9 of 9 dogs of groups 1, 2, and 3, respectively, had detectable antibodies against CIV, as determined by results of HI tests. With the exception of 1 dog of group 1, all inoculated dogs had detectable antibodies (HI tests) against CIV on days 14, 21, and 28 after the initial inoculation. Antibodies against CIV were also detectable by the VN test 7 days after the initial inoculation in 1, 3, and 5 dogs of groups 1, 2, and 3, respectively. Fourteen, 21, and 28 days after the initial inoculation, 4, 4, and 6 dogs, respectively, of group 1 had detectable antibodies (VN tests). For the same time points, 8, 6, and 6 dogs, respectively, of group 2 and all dogs of group 3 had detectable antibodies (VN tests).

After the booster inoculation on day 28, all dogs of groups 1, 2, and 3 developed antibodies (HI and VN tests). Booster inoculation induced an apparent anamnestic response as evidenced by a 4-fold or greater increase in GMT detected 7 and 14 days after the second inoculation.

Although there was no significant difference in antibody titers between SC inoculated groups (group 1 and group 2), vCP2242-induced antibody titers were approximately twice as high (GMT, 139 for vCP1529 and 276 for vCP2242; HI tests) and 3 times as high (GMT, 335 for vCP1529 and 937 for vCP2242; VN tests) as those that were induced by vCP1529. Antibody titers of group 3 (transdermally inoculated group) were significantly higher than those of groups 1 and 2 after the first and second inoculations.

Amino acid sequences—We detected a high degree of homology for the amino acid sequences among HA1 proteins expressed by vaccine viruses and antigens used in HI and VN tests. The HA1 proteins expressed by vCP1529 and vCP2242 had 94.16% and 97.37% identity, respectively, with A/Canine/NY/115809/05 and 96.20% and 98.54% identity, respectively, with A/Canine/FL/43/04 isolates. Comparisons between amino acid sequences of HA1 proteins expressed by vCP1529 and vCP2242 revealed 97.66% identity.

Discussion

Analysis of results of the study reported here indicated that in dogs, canarypox-vectored vaccines expressing EIV genes from A/Equine/KY/94 and A/Equine/OH/03 induced substantial antibody titers, as determined by use of HI tests (all dogs had antibody titers≥ 32 after the booster inoculation), and antibodies against CIV A/Canine/FL/43/04-PR and A/Canine/NY/115809/05 isolates, respectively, as determined by use of VN tests. Although vaccinationchallenge studies will be required to substantiate or refute protective efficacy of the vaccines used in this study, antibody titers reported here may be considered predictive of a reasonable expectation of efficacy for use of these vaccines to prevent or ameliorate CIV infection. Serum antibody titers, mainly those determined by HI and VN assays, are generally accepted to be good predictors of immune protection against influenza in humans and other animals because of their ability to block attachment of viral HA to sialic acid–containing receptors.^7–10 Although there has been variation between required antibody titers and the degree of protection in various species, it is generally accepted that a titer of≥ 32 on an HI test correlates with protection and that higher antibody titers are associated with a higher degree of protection.⁷

It is widely reported that infection with influenza viruses can induce long-lasting and subtype cross-reactive immune responses, compared with the response for inactivated vaccines.^7,8 Indeed, use of an inactivated EIV vaccine containing A/Equine/KY/93 antigen induced poor antibody responses against homologous antigen (only 3/6 dogs had detectable antibodies) and no detectable antibody responses against heterologous EIV or CIV antigens in dogs.^a One of the vaccines (vCP1529) used in the study reported here expresses HA antigen of EIV isolate A/Equine/KY/94 (which is considered to be antigenically similar to the antigen used in inactivated EIV vaccine used in the study of CIV), and it induced substantial amounts of antibodies (HI and VN tests) against CIV antigens in all of the inoculated dogs.

Similar observations have been made in horses. For example, there was an outbreak of EIV infection with extensive morbidity in horses vaccinated with an inactivated vaccine that contained EIV antigens from European and American lineages,¹¹ whereas a vaccination-challenge study^12,13 revealed that canarypoxvectored EIV vaccines (vCP1529 and vCP1533) expressing similar antigens used in the aforementioned inactivated vaccines (the EIV HA antigens from European and American lineages) induced substantial protection against the EIV isolated from that outbreak, which was used for challenge exposure. Analysis of these results suggests that the antigen content as well as the antigen presentation may play an important role in induction of cross-reactive immune responses and that vaccines used in the study reported here may mimic natural infection because there was de novo synthesis of HA protein within the host cells, which may have led to production of cross-reactive antibody responses measurable by use of HI and VN tests.

Inoculation with vCP2242 induced approximately 2-fold higher antibody titers (GMT, 139 for vCP1529 and 276 for vCP2242; HI tests) and 3-fold higher antibody titers (GMT, 335 for vCP1529 and 937 for vCP2242; VN tests) than vCP1529, although the antigen content, volume, and route of administration for vCP1529 and vCP2242 vaccines were the same (groups 1 and 2). This result was not surprising because it is generally accepted that one of the factors that determines cross-reactivity is the degree of antigenic similarity in the HA proteins among vaccines and antigens used in the serologic assay. Indeed, comparison between amino acid sequences of HA proteins expressed by canarypox-vectored vaccines and sequences of CIV antigens used in HI and VN tests revealed that the HA protein expressed by vCP2242 is more closely related to the CIV antigens used in HI and VN tests than it is to vCP1529. Although this result may be a good indication of the differences in antibody titers detected by use of HI and VN tests in our study, the role of amino acid positions in terms of predicting antibody cross-reactivity has not been clearly defined for EIV HA type 3.

Although jet injection devices have been used extensively in humans for mass immunization campaigns against smallpox, poliomyelitis, and measles, its use in veterinary medicine has been extremely limited until recently. In addition to its advantages with regard to safety, administration of vaccines by the transdermal route has resulted in improved vaccine immunogenicity.^14,15 Results of the study reported here confirm these findings because administration via the transdermal route of inoculation induced significantly higher antibody titers (HI and VN tests) against CIV than were induced after SC administration. In contrast to our results, investigators in another study¹⁶ reported that IM administration of an adenovirus-vectored swine influenza virus vaccine induced consistently higher antibody responses (HI tests) than after vaccination via needle-free injection. The discrepancy between our results and results of that study may be attributable to several factors, including differences in the vaccines, routes of administration (SC vs IM), animals (dogs vs pigs), and needle-free devices. Although additional studies will be required, use of the transdermal route of administration for other vaccines may provide additional opportunities for improvements, such as dose reduction (antigen content and volume) and increased efficacy of veterinary vaccines.

The canarypox-vectored vaccines used in the study reported here did not cause any adverse reactions in dogs, regardless of the HA type or route of administration. This result further confirms the safety of an attenuated canarypox virus strain^g that can replicate only in avian species and that has been used as the vector for several registered recombinant vaccines for dogs and ferrets (distemper), cats (rabies and FeLV), and horses (EIV and West Nile virus). In addition to their safety features, vaccines used in our study express de novo–synthesized antigens that are expected to stimulate both humoral and cell-mediated immune responses. Thus, these vaccines may have a substantial impact on CIV transmission within the community and may be of great value in protecting dogs against CIV-induced disease.