Materials and Methods

Nineteen permanently disabled red-tailed hawks were obtained from the Raptor Center at the University of Minnesota and various rehabilitators throughout the United States, in accordance with provisions of permits issued by the US Fish and Wildlife Service. These were acquired over a period of 6 months and held in suitable indoor laboratory animal facilities at the University of Minnesota until the experiment was undertaken. All procedures were conducted in accordance with Institutional Animal Care and Use Committee protocols from the University of Minnesota and the University of Louisiana. Each hawk underwent physical examination, clinicopathologic analyses, and testing for serum antibodies against WNV and St. Louis encephalitis virus (a related cross-reacting flavivirus) by use of PRNTs,^b and only healthy, antibody-negative birds were used in the experiment.

Fourteen birds were selected at random and placed in a treatment group room; the remaining 5 birds were used as sham-vaccinated controls and placed in a separate room. Both groups were maintained with accepted practices for captive raptors.¹⁷ Briefly, hawks were housed in indoor rooms that were controlled for light and ventilation. Rooms were constructed of cinder block walls and concrete floors, all of which were sealed with epoxy resin finishes. Room temperature was maintained constant at 20°C, and photoperiod was regulated (12 hours of light and 12 hours of dark) with standard fluorescent lighting. Hawks were fed coturnix quail ad libitum, which were supplied at the rate of approximately three-quarters of a quail/hawk/d. Water for drinking and bathing was provided in shallow floor pans. Rope-covered perches of suitable size and height for nonflying red-tailed hawks were placed in various locations on the floor. All rooms were cleaned once daily with the hawks in situ.

Administration of the vaccine or adjuvant only—A recombinant WNV DNA-plasmid West Nile vaccine developed at the CDC in Fort Collins, Colo was used in the experiment.¹⁸ The Escherichia coli plasmid contained WNV genes expressing premembrane and envelope proteins of the WNV and was the same plasmid used in some previous studies.^10,11,18 The plasmid was obtained for our use under a material transfer agreement with the CDC and produced in quantity by a commercial company.^c The vaccine was produced in the laboratory of one of the coauthors (GJC). In the final vaccine product, the DNA concentration was 500 μg in 500 μL of PBS solution, and it was mixed 1:1 with a 2% aluminum phosphate adjuvant.^d

Injection of the vaccine or adjuvant only was undertaken in facilities at the University of Minnesota. The 14 hawks in the treatment group received two 1-mL doses of the vaccine via IM injection in the femoral muscle group delivered via a 25-gauge needle. There was a 3-week interval between vaccine doses. The 5 hawks in the control group received two 1-mL doses of the vaccine adjuvant via IM injection in the femoral muscle group delivered via a 25-gauge needle. There was a 3-week interval between adjuvant doses. Three weeks after the second injection, the hawks were loaded in fiberglass animal-holding containers^e and driven in a van from Saint Paul, Minn, to Baton Rouge, La, to undergo experimental challenge with live virus in appropriate biosafety facilities at the Louisiana State University School of Veterinary Medicine. Upon arrival, hawks were transferred to a suitable animal-holding space similar to that at the University of Minnesota and allowed 1 week of acclimation prior to experimental challenge with WNV. During this time, they were observed for overall health and food consumption; however, any further handling or sampling was eschewed in the interest of not subjecting them to additional stress.

Live WNV challenge—Four weeks after the second injection of vaccine or adjuvant only, each hawk received an SC injection of 0.1 mL of a suspension containing 10⁵ pfu of the Louisiana strain of WNV/mL. The virus was isolated from the kidney of a Blue Jay (Cyanocitta cristata) that died in Louisiana during the summer of 2001 and was maintained in storage at −80°C. To create the challenge inoculum, the stored WNV isolate was inoculated onto Vero cells and harvested after 24 hours growth. The day of live virus challenge was designated as day 0.

Experimental assessments—A blood sample (1 mL) was collected from each hawk on days 0 (before administration of the live WNV), 1, 2, 4, 6, 8, 10, 12, and 14 to assess serum anti-WNV antibody titers; the sample was also used for assessment of the degree of viremia on days 1, 2, 4, 6, 8, 12, and 14. Blood was transferred immediately into evacuated tubes^f containing K₂EDTA and held in an ice-water bath. Tubes were centrifuged for 10 minutes at 0.1 × g in a refrigerated centrifuge^g at 4°C. Buffy coats were harvested and stored for a maximum of 15 days at −80°C until quantitatively cultured for determination of viremia severity. Plasma was removed, decanted into individual cold storage tubes, and stored at −80°C for 4 months prior to antibody testing via PRNT. To assess shedding of virus, a single swab^h was inserted first into the choana and then in the cloaca of each bird at the same time points as blood sample collection. Swabs were placed in viral transport media in a 2-mL cryovial, placed in an ice bath during collections, and later stored at −80°C until further analysis. After the 14th day, all birds were euthanized (ie, anesthetized via inhalation of isoflurane [mask induction] followed by administration of an IV overdose of pentobarbital euthanasia solution).

Assessment of viremia—Quantification of virus in blood samples was accomplished by use of a plaque formation assay¹⁹ with Vero cells.ⁱ Briefly buffy coats were thawed and dilutions from 1 × 10⁻¹ through 1 × 10⁻⁷ were made in Hank's balanced salt solution. Two hundred microliters of each dilution was added to triplicate wells of Vero cell monolayers contained in a 24-well cell culture plate.^j Vero cell plates were incubated for 60 minutes at 37°C in 5% CO₂. After incubation, each well was overlaid with approximately 1.5 mL of a mixture of 3% methylcellulose prepared in MEM supplemented with 10% fetal bovine serum and then returned to the incubator. After 60 hours of incubation, 500 μL of 4% formalin was added to each well and plates were kept for 4 hours at room temperature (approx 20°C). The overlay solution and formalin were removed, and plates were washed under a gentle stream of warm tap water. Each dilution well was stained with approximately 1 mL of Gram-crystal violet stain^k for a minimum of 20 minutes followed by a rinse with tap water. For each sample, the number of plaques in 3 wells selected from among the dilution range containing 20 to 200 plaques were counted, and the mean number of plaques for the 3 wells was calculated; the level of viremia was recorded as pfu per milliliter of blood.

Antibody titer determination by use of a PRNT—Antibody titers were determined by use of a PRNT following an established method.²⁰ Briefly, serum samples were diluted 1:5 in serum neutralization medium (1 × MEM; 2.2 g of NaHCO₃/L; 3% heat-inactivated fetal bovine serum; and 2× antimicrobial-antimycotic solution [200 U of penicillin/mL, 200 μg of streptomycin/mL, and 500 ng of amphotericin B/mL]). Positive and negative control sera were also run on each batch. Negative control serum was commercially purchased specific-pathogen-free chicken serum.^l Positive control serum was antiserum obtained from mallards that had been experimentally infected with WNV at the National Wildlife Health Center, Madison, Wisconsin. Samples and control sera were heat-inactivated at 56°C for 30 minutes prior to being further diluted (doubling dilutions to 1:320) in an equal volume of serum neutralization medium containing 100 pfu of WNV Dilutions were then incubated at 37°C for 1 hour. Following the neutralization step, each virus-serum mixture was inoculated onto 4-day-old Vero cells and allowed to adsorb at 37°C in a humidified atmosphere containing 5% CO₂ for 1 hour with periodic rocking. Infected cell cultures were then overlaid with 1% gum tragacanth and 1× MEM supplemented with 3% heat-inactivated fetal bovine serum and 2× antimicrobial-antimycotic solution, and incubation was resumed. On day 4 after adsorption, cultures were inactivated with neutral-buffered 10% formalin and subsequently stained with 0.25% Gram-crystal violet stain^k for plaque detection. Test serum samples that had a 90% reduction in plaques, relative to the negative control serum sample, were considered positive for WNV-neutralizing antibodies.

Assessment of viral shedding—Upon collection, each choanal-cloacal swab specimen was placed in 500 μL of BA-1 medium^m (containing 1× medium 199, 0.05M Tris-HCl, 1% bovine serum albumin, 0.0375% NaHCO₃, penicillin G [200 μg/mL], streptomycin [200 μg/mL], and amphotericin B [0.5 μg/mL]) and stored at −80°C. Ribonucleic acid extraction was conducted by use of an extraction kitⁿ according to the manufacturer's protocol. Briefly, 250 μL of sample was mixed with 250 μL of the cell lysis buffer containing mercaptoethanol. Then, 250 μL of cold 70% ethanol was added and samples were vortexed and centrifuged for 30 seconds. The supernatant was pulled through the spin column by use of a vacuum manifold.ⁿ Seven hundred microliters of the wash buffer was added to each column followed by 500 μL of the second wash buffer and a repeat wash with 500 μL of that second wash buffer. Each spin column was placed in collection tubes and centrifuged for 2 minutes. The collection tubes were discarded, and the spin columns were placed in RNAse-free tubes. Thirty microliters of diethylpyrocarbonate-treated water was added to each column, and these columns were centrifuged for 2 minutes. The eluted RNA obtained was stored at −20°C until the reverse transcriptase PCR assay was performed.

Each extracted sample was tested for the presence of WNV RNA by use of a real-time reverse transcriptase PCR assay. The samples underwent 45 cycles of amplification in a sequence detection system.^o Five microliters of extracted RNA was added to 20 μL of master mixⁿ and 50pM of WNV-specific forward primer (5′-TCAGCGATCTCTCCACCAAAG-3′; GenBank accession No. AF196835), 50pM of WNV-specific reverse primer (5′-GGGTCAGCACGTTTGTCATTG-3′; Gen-Bank accession No. FJ527738.1), and 10pM of WNV-specific probe (FAM-TGCCCGACCATGGGAGAAGCTC-BHQ-1; GenBank accession No. AF196835.2). The extracted RNA samples were tested together with samples prepared for a standard curve containing known amounts of WNV The resulting cycle threshold values of the samples were then compared with the curve and reported as number of pfu/5 μL of sample.

Data analysis—Geometric mean values of the viremia levels (log₁₀ pfu/mL of blood) and viral shedding activity (pfu) for the vaccinated and nonvaccinated control birds at each postchallenge time point were compared by use of a 1-tailed unpaired Student t test. The serum antibody GMTs for the vaccinated and non-vaccinated control birds at the prechallenge (day 0) and each postchallenge time point were also compared by use of a 1-tailed unpaired Student t test. Values of P < 0.05 were considered significant.

Results

None of hawks had a visible reaction to administrations of the vaccine or adjuvant alone. Three vaccinated birds developed weak antibody titers (1:10) following vaccination. All remaining birds seroconverted following viral challenge. For the treated and control birds, geometric mean values of the levels of viremia and viral shedding at the postchallenge time points were calculated (Tables 1 and 2). Serum anti-WNV antibody GMTs for treated and control birds before and after live virus challenge were also determined (Table 3). Following live virus challenge, significant differences in severity of viremia and antibody GMTs were evident between vaccinated and nonvaccinated birds. Differences in seroconversion rate and duration and magnitude of viral shedding were also observed but were not significantly different.

Viremia—In 1 vaccinated bird, viremia did not develop after live virus challenge. Mean values of the level of viremia (pfu/mL of blood) in vaccinated and control birds differed significantly (P < 0.05) on postchallenge days 2, 4, and 6 (Table 1). In addition, the proportion of birds that had detectable viremia on day 6 (2/5 control birds and 1/14 vaccinated birds) was further evidence of differences between the 2 groups.

Viral shedding patterns—Oropharyngeal-cloacal shedding of WNV was evident in both vaccinated and control birds on postchallenge days 2, 4, and 6 but was greatly diminished on day 8 with some continued low level shedding through day 14 (Table 2). The 1 bird in which viremia was not evident had intermittent, low-level shedding on days 4, 8, and 12. Viral shedding was detected in 8 of the 14 vaccinated hawks and 5 of the 5 control hawks on day 8 and in 7 of the 14 vaccinated birds and all control birds on day 10. On day 12, only 4 vaccinates and 2 control birds continued to shed small numbers of virus, and on day 14, 2 of each shed small numbers of virus.

Seroconversion—Three of 14 vaccinated hawks had neutralizing antibody titers (1:10) in response to the vaccination protocol as assessed by use of the PRNT on day 0 (ie, 4 weeks after the second vaccination) and on days 1 and 2 after live virus challenge. The time required for antibody titers to increase and also the magnitude of the antibody response after live virus challenge differed between vaccinated and nonvaccinated birds (Table 3). Eight of 14 vaccinated birds had titers > 1:10 on postchallenge day 4, whereas none of the control birds had measurable antibody titers. On day 6, all birds in both groups had seroconverted; however, the GMT of the vaccinated birds (GMT ± SD, 250 ± 1.7) was greater than that of the control birds (61 ± 3.6). On day 10 and thereafter, the difference in antibody titers between the 2 groups was negligible.

Discussion

The adjuvanted WNV DNA-plasmid vaccine used in the present study expressed and secreted premembrane and envelope proteins containing virus-like particles in transformed cells and was effective in attenuating viremia and viral shedding in hawks challenged with WNV. This plasmid was the same as that used in previous studies.^10,11,18 In an early study,¹⁸ mice and horses were vaccinated with plasmid DNA without adjuvant. However, in the follow-up clinical trials in horses, a plasmid vaccine formulated with a proprietary lipid-based adjuvant^a was more immunogenic than vaccine without adjuvant (unpublished data). Additionally, the potency of DNA vaccines in laboratory rodents and nonhuman primates increases substantially following formulation with conventional aluminum adjuvants.²¹ Thus, the vaccine used in the present study was formulated in aluminum phosphate adjuvant. In the experience of one of the authors, the use of aluminum phosphate adjuvant rather than a lipid adjuvant eliminated the possibility of an emulsion-induced necrotic reaction at the injection site.

In the present study, hawks were administered live WNV 4 weeks after the second injection of vaccine or second injection of adjuvant only. At the time of viral challenge, 3 of 14 vaccinated hawks had detectable serum antibody titers at a 1:10 dilution. In a study by Turrell et al¹⁰ involving DNA-plasmid-vaccinated fish crows (Corvus ossifragus), 6 of 9 vaccinated birds seroconverted. In another study,¹¹ seroconversion occurred in all condors (Gymnogyps californianus) that received the same vaccine; at day 21 after vaccination, titers ranged from 4 to 128. In another investigation¹⁴ involving similar DNA-plasmid constructs but different vaccine component types (ie, subunit as opposed to plasmid), seroconversion occurred in all domestic geese (Anser anser domesticus Var Embden) following vaccination; in those birds, postvaccination titers were between 20 and 40, as determined via PRNTs. All birds (including the nonviremic bird) in the study reported here were seropositive by day 6 following live virus challenge (GMT, 250 for vaccinates and 61 for nonvaccinates; titer ranges, 80 to > 320 and 10 to 320, respectively). This finding suggested an anamnestic response by the vaccinated birds. However, by day 10, there was no difference in serum antibody titers between the 2 groups. The reasons for the observed low rate of seroconversion following DNA-plasmid vaccination in the present study, compared with results of other studies, are not apparent but may reflect interspecies differences in the response to this type of vaccine. Interspecific variation in response to WNV vaccines has been reported.^9,12,13,15

In another study,¹⁰ fish crows were vaccinated with a microencapsulated DNA-plasmid vaccine that contained the same WNV genes as the product used in our experiment. In that previous study,¹⁰ 5 of 10 nonvaccinated birds died, whereas none of the 9 vaccinated birds were affected, following challenge with an SC injection of WNV (10⁵ pfu). In all of the domestic geese administered an experimental subunit vaccine preparation in another study,¹⁴ seroconversion was detected (mean ELISA titer, > 1:528), viremia and cloacal shedding were completely suppressed, signs of morbidity were not evident, and death did not occur as a result of administration of the vaccine or the WNV challenge. Olsen et al¹⁵ challenged Florida sandhill cranes (Grus canadensis pratensis) that had been vaccinated with a killed WNV vaccine. Following challenge, vaccinated birds had lower levels of viremia, shorter periods of cloacal shedding of virus, and higher anti-WNV antibody titers than had nonvaccinated birds. No morbidity or deaths were observed.¹⁵ In another study,²² WNV viremia was induced in 25 species of birds via direct mosquito bite, oral inoculation, or contact with infected birds. In general, it appeared that the inoculating dose of WNV received by these birds was < 10⁷ pfu. The duration of the resultant viremic states were < 7 days in all species and, in most, had waned considerably by day 6 after viral challenge. Twenty-eight birds of 8 passerine species died following viral challenge (range of mean times to death, 4.5 to 9 days); most of those birds died on or before day 6. Curiously, no morbidity or deaths occurred in either the falconiforms (American kestrels [Falco sparverius]) or strigiforms (great horned owls [Bubo virginianus]) used in that experiment, species that are known clinically and pathologically to be susceptible to WNV infection.²³ Nemeth et al²⁴ inoculated juvenile screech owls (Megascops asio) with approximately 10³ pfu of the NY99 strain of WNV, after which 2 of the 5 birds died. On the basis of the apparent susceptibility of red-tailed hawks to WNV infection, we expected deaths to occur among the nonvaccinated control birds in the present study; however, neither morbidity nor deaths were encountered. Clearly, WNV-induced morbidity and death varies with host and the dose and strain of virus used in viral challenges.

In the present study, viremia was detected in vaccinated and control hawks on day 1 following live virus challenge; the severity of viremia peaked on day 4 and then waned rapidly. By day 6, only 2 of the 5 nonvaccinated control birds and 1 of the 14 vaccinated birds had detectable virus in their blood. The attenuation of viremia in vaccinated birds differed from that in control birds on postchallenge days 2, 4, and 6. Experimentally, attenuation of viremia has been used as a surrogate measure of immunity in animals in which clinical disease is not induced.²⁵ In the study of Turrell et al,¹⁰ 6 of 9 vaccinated fish crows developed viremia; the peak level of viremia ranged from log₁₀ 2.9 pfu/mL of blood to log₁₀ 3.8 pfu/mL of blood in vaccinated and nonvaccinated birds, respectively. Additionally, only 5 of 9 vaccinated crows developed neutralizing anti-WNV antibody by day 14 following viral challenge but all 9 survived exposure to the virus. In the present study, WNV shedding was detected on day 2 and the shedding pattern followed the general pattern of changes in viremia, although a few birds in both groups continued to shed after the end of the viremic period. The finding that vaccinated birds shed the virus for a shorter period of time and with apparent less intensity, compared with the nonvaccinated control birds, is consistent with findings of other investigators.¹⁵ Shedding has been reported to extend beyond the viremic period by other investigators as well.^6,21,24

The results of the present study have highlighted the importance of live virus challenge in addition to serologic evaluation in experiments to determine both safety and efficacy of a DNA-plasmid vaccine. Although seroconversion was incomplete and was characterized by low serum antibody titers, WNV DNA-plasmid-vaccinated red-tailed hawks had a notable attenuation in viremia following inoculation with live WNV. In addition, vaccinated birds, regardless of their state of seroconversion, developed detectable and greater magnitude postchallenge antibody titers 2 days sooner than did nonvaccinated control birds; furthermore, vaccinated birds had shorter and less-intense shedding profiles. This attenuation, along with the observed lack of adverse reactions and modest evidence of seroconversion, indicated that the WNV DNA-plasmid construct and adjuvant used in the present study is an effective means of vaccinating red-tailed hawks and possibly other avian species against WNV-induced disease. These findings complement data regarding the use of this vaccine construct in California condors and geese. Because various avian species have wide variation in their response to WNV vaccines whether composed of either conventional killed virus or the newer engineered plasmid types, further research that incorporates live virus challenge is required to determine the true efficacy of any WNV vaccines.