A total of 12 horses of different breeds and ages were infected with West Nile virus (WNV) via the bites of infected Aedes albopictus mosquitoes. Half the horses were infected with a viral isolate from the brain of a horse (BC787), and half were infected with an isolate from crow brain (NY99-6625); both were NY99 isolates. Postinfection, uninfected female Ae. albopictus fed on eight of the infected horses. In the first trial, Nt antibody titers reached >1:320, 1:20, 1:160, and 1:80 for horses 1 to 4, respectively. In the second trial, the seven horses with subclinical infections developed Nt antibody titers >1:10 between days 7 and 11 post infection. The highest viremia level in horses fed upon by the recipient mosquitoes was approximately 460 Vero cell PFU/mL. All mosquitoes that fed upon viremic horses were negative for the virus. Horses infected with the NY99 strain of WNV develop low viremia levels of short duration; therefore, infected horses are unlikely to serve as important amplifying hosts for WNV in nature.
West Nile virus (WNV), a flavivirus related to Japanese encephalitis, St. Louis encephalitis, and Murray Valley encephalitis viruses, was responsible for outbreaks of encephalomyelitis in humans and horses in New York (NY) during 1999 (
Previous WNV experimental infection studies in equids are few, and attempts to induce and study clinical disease caused by WNV in equids have yielded equivocal results (
Previous studies did not resolve the question of whether WNV-infected equids produce viremia levels of sufficient magnitude and duration to infect vector mosquitoes. Schmidt and Mansoury (
Because of the lack of definitive studies on the potential for equines to serve as amplifying hosts for WNV following vector-borne transmission and the implications for public health and veterinary concerns, studies were needed of horses infected by mosquito bite with virus strains isolated during the 1999 outbreak in New York. Accordingly, we investigated the course of clinical disease (
Mares and geldings of varying ages and breads were used (
The horses were examined for signs of disease twice a day. Their body temperatures were recorded twice a day from day 0 (day of infection) to day 10, then daily through day 28, unless they were euthanized earlier. Pulse and respiration rates were recorded daily for the first 2 weeks after infection. Blood was collected for serum twice a day from day 0 through day 14, then daily through day 28, and twice a week for the duration of the project. Oral and rectal swabs were obtained daily from days 0 through 10, and the samples were placed into 1 mL of BA-1 medium (M-199 salts, 1% bovine serum albumin, 350 μ/L sodium bicarbonate, 100 units/mL penicillin, 100 μ/L streptomycin, and 1 μ/L amphotericin in 0.05 M Tris, pH 7.6) and stored frozen for virus isolation. Blood samples were analyzed daily from day 0 through day 10 by using a QBC-V hematology analyzer (Clay Adams, Becton, Dicksinson and Company, Franklin Lakes, NJ). The methods used for histopathology and immunocytochemistry have been described (R. Bowen, submitted for pub.).
Horses 5 through 9 were used to test a vaccine for WNV and were not part of this research project.
Mosquitoes used for transmitting WNV to susceptible horses by bite were infected with either of two strains of WNV isolated from New York animals infected during the 1999 outbreak. Strain BC787 (1 passage in suckling mouse brain) was isolated from the brain of a horse, and NY99-6625 (1 passage in Vero cell culture) was isolated from the brain of a crow.
The Aedes albopictus mosquitoes used in these experiments were from a colony strain from Lake Charles, LA. Mosquitoes were reared in an insectary maintained at 26.7°C (± 0.5°) in approximately 80% relative humidity and with a long photophase (light:dark 16:8). Larvae were fed liver powder and rabbit chow as desired. Cohorts of 3- to 5-day-old adult female Ae. albopictus were inoculated intracoelomically with either of the two virus strains, placed in separate cages, given 5% sucrose for maintenance, and incubated under the insectary conditions described above. The virus dose per mosquito, based on appropriate dilutions of stock virus of known titer, was estimated to be approximately 170 Vero PFUs. Sucrose was withheld prior to feeding on equines.
Infected mosquitoes were allowed to feed on horses in two separate trials. A pilot experiment was conducted with four horses to determine probable periods of peak viremia before a more comprehensive experiment was conducted, which involved the feeding of large numbers of uninfected (recipient) mosquitoes. On day 8 postinoculation (May 10, 2000), caged infected mosquitoes were placed in a securely taped styrofoam container and transported to the equine holding facility. Feeding was accomplished by holding a cage of mosquitoes, with a gloved hand, against a shaved area immediately behind the left shoulder of each horse. Based on viremia profiles from the initial trial, a second trial was begun on July 10, 2000, in which infected mosquitoes that had been incubated 12 days postinoculation were allowed to feed on eight horses. Each horse was exposed to the mosquitoes for 5 minutes in each trial.
After feeding, the caged mosquitoes were sealed in a styrofoam container that was rinsed externally with diluted bleach and then transported to a laboratory. Mosquitoes were anesthetized with CO2 and sorted on wet ice. Most blood-fed specimens were placed in stoppered vials and frozen at -70°C until they were triturated and tested for virus. Mosquitoes were ground individually in 1 mL each of BA-1 diluent in TenBroeck (Wheaton Science Products, Millville, NJ) tissue grinders. Samples were centrifuged for 4 minutes at 20,000 x g in a refrigerated microcentrifuge. Supernatants were poured into screw-cap vials, which were kept on wet ice until serial dilutions were made, and the samples were tested for virus by plaque assay.
During the second experiment, on days 3, 4, and 5 after infected mosquitoes had fed on the horses, uninfected 3- to 7-day-old Ae. albopictus females were permitted to feed on the 8 horses in lots of approximately 40 mosquitoes per cage. In addition, 3 lots of recipient mosquitoes were fed on horse 11 on days 8 and 9 postinfection, after the horse showed signs of clinical illness. Recipient mosquitoes were allowed to feed in an area behind the right shoulder of the animals on the side opposite from where the infected mosquitoes had fed. Fed mosquitoes were transported to a laboratory and sorted with a mechanical aspirator, and each cohort was placed in a separate 0.2-L cage. Blood-engorged mosquitoes were given 5% sucrose for maintenance, and the 0.2-L cages were placed inside larger cages and incubated as described. Recipient mosquitoes that fed on horses on days 3 and 4 postinfection were incubated for 10 days. The 8 cohorts of recipient mosquitoes fed on day 5 postinfection were incubated only 7 days because a large number had died. Following incubation, mosquitoes were immobilized by being chilled briefly and were placed in tubes while still alive. They were then frozen at -70°C until processed and tested for virus by plaque assay. All mosquitoes fed on days 3 to 5 postinfection were individually disrupted by sonic energy (
Some of the WNV-infected mosquitoes used for horse challenge were chilled at 4°C for 5 minutes and held on ice. Mosquitoes were immobilized by removing their legs and wings. Individual mosquito proboscises were placed in a capillary tube containing Type “B” immersion oil. Saliva was collected in the oil-filled tube (
The WNV equine immunoglobulin (Ig)M capture enzyme-linked immunosorbent assay (ELISA) is a microtiter plate format assay designed to detect IgM antibodies to WNV. Unused 96-well plates (Immunolon II; Dynatech Laboratories, Chantilly, VA) were coated overnight at 4°C with 75 μL of goat anti-horse IgM antibody (Kirkegaard & Perry Laboratories, Gaithersburg, MD) in carbonate buffer (0.015 M NaCO3, 0.035 M NaHCO3, pH 9.6) at a 1:1,000 dilution. For quality control, 36 outside wells of the plate were not used in the test. After the coating buffer was decanted, 200 μL of blocking buffer (phosphated-buffered saline [PBS], 0.05% Tween 20, 5% dry milk) was added to each well, and the wells were incubated at room temperature for at least 30 minutes. After the mixture was washed five times with wash buffer (PBS, 0.05% Tween 20), 50-μL samples of horse serum, diluted 1:400 in wash buffer, were applied; each sample was added to 6 wells. Positive and negative control equine sera were included on each plate, yielding a final capacity of eight test samples per microtiter plate. Diluted serum was incubated for 1 hour at 42°C. After the serum was washed five times, 50 μL of WNV tissue culture antigen (CDC Cat. No. VA2395), diluted 1:300 in wash buffer, was added to half the wells to which diluted serum was added. And 50 μL normal control tissue culture antigen (CDC Cat. No. VB2396), diluted 1:300 in wash buffer, was added to the remaining wells so that each test serum sample had three viral antigen wells and three normal antigen wells. The antigen was incubated overnight at 4°C. After the samples were washed five times, 50 μL of horseradish peroxidase, conjugated with a monoclonal antibody to WNV (6B6C-1 CDC; ABD [Arboviral Branch] reference reagent) and diluted 1:2000 in blocking buffer, was added to each well and incubated for 1 hour at 42°C. After the wells’ contents were washed 10 times with wash buffer, 75 μL tetramethylbencidine (TMB) substrate (GibcoBRL, Life Technologies Inc, Gaithersburg, MD) was added to each well. After incubation with the substrate for exactly 10 minutes, the reaction was stopped by adding 50 μL 1N H2SO4. The OD of each well was measured at 450 nm, and results were calculated as follows:
The Final A450nm = [A450 nm of test sera on West Nile antigen] – [A450nm of normal sera on West Nile antigen]
Serum samples were tested for virus-neutralizing (Nt) antibodies to WNV (NY99-4132 strain, 1 Vero passage) by the plaque-reduction neutralization test in Vero cell culture. Earlier serum samples were also tested for Nt antibodies to SLEV (TBH-28 strain, passage history unknown). Briefly, diluted serum was heat inactivated at 56°C for 30 minutes and mixed with an equal volume of a virus preparation BA-1 containing 8% normal human serum, so that the number of infectious virus particles in the final dilution was approximately 100/0.1 mL. A volume of 0.1 mL was then injected onto a Vero cell monolayer and processed as for the plaque assay. Samples were screened by testing once at a final dilution (after mixture with challenge viruses) of 1:10. Any sample that neutralized the challenge virus dose at a level of >70% was confirmed by testing in duplicate and titrated by serial twofold dilutions. Neutralization at a level of >90% was considered positive for each dilution. Horses having preexposure sera that neutralized SLEV at >70% were excluded from the study. Sera from horses 1 to 4 were challenged with approximately 300 PFU of WNV (Horse Brain WN Isolate # BC 787), and sera from horses 9 to 16 had an average of 86 PFU of WNV added.
Virus concentration in serum samples, mosquito homogenates, and tissue homogenates was measured by titration in a plaque assay. Briefly, 0.1 mL of sample was added to a monolayer of Vero cells in a six-well cell culture plate (Costar, Corning Incorporated Life Science, Acton, MA) and incubated 1 hour at 37°C in 5% CO2. Cells were overlaid with 3 mL per well of 0.5% agarose in M-199 medium, supplemented with 350 mg/L of sodium bicarbonate, 29.2 mg/L of L-glutamine, and antibiotics. After 48 hours of additional incubation, a second 3-mL 0.5% agarose overlay, containing 0.004% neutral red dye, was added for plaque visualization. The plaques were scored on days 3, 4, and 5 of incubation.
In the first infection trial, 12 to 17 infected mosquitoes fed on each of four horses. Virus titrations were done on 10 blood-engorged mosquitoes from each cohort that fed. All of the 40 mosquitoes tested were infected. Virus titers ranged from 106.6 to 107.9 Vero cell PFU per mosquito and averaged 106.8, 107.2, 107.3, and 107.4 PFU per mosquito in the cohorts that fed on horses 1 to 4, respectively. Mosquitoes infected with the WNV strain isolated from a horse were fed on horses 1 and 2, and mosquitoes infected with a strain isolated from a crow were fed on horses 3 and 4 (
On the basis of these viremia profiles, we decided to feed uninfected mosquitoes on infected horses in the next trial on days 3 to 5 postinfection to determine whether viremia levels were sufficient to infect susceptible mosquitoes. None of these horses developed fever, hematologic abnormalities, or signs of clinical disease. Horses 1, 2, 3, and 4 were euthanized, and necropsies were performed on days 14 and 16 postinfection. Histopathologic lesions indicative of WNV infection were not observed. Virus was not isolated from any tissues from these animals, and viral antigens were not detected in brain, spinal cord, or other tissues by immunocytochemical tests.
In the second trial, 7 to 14 infected mosquitoes fed on each of the eight horses. Virus titrations were done on five mosquitoes from each cohort, and 100% of the mosquitoes tested were infected. Virus titers ranged from 106.5 to 108.0 Vero cell PFU per mosquito and averaged 10 7.5,107.3, 107.3, 107.3, 107.6,107.3, 107.7, and 107.4 PFU per mosquito in the cohorts that fed on horses 9 numbered to 16, respectively. Mosquito saliva was collected from 10 infected mosquitoes used in the second trial, and virus titers ranged from 101.3 to 102.5 PFU per sample of saliva.
Mosquitoes infected with a WNV strain isolated from a horse were fed on horses 9 numbered to 12 and mosquitoes infected with a strain isolated from a crow were fed on horses numbered 13 to 16 (
Twenty-four lots of uninfected recipient mosquitoes were fed on seven infected horses in the mornings of days 3 to 5 postinfection. Virus titers in the seven viremic horses ranged from 101.0 PFU/mL to 102.7 PFU/mL during the times mosquitoes were fed. The feeding success among uninfected mosquitoes was 92%, 81%, and 71% on days 3 to 5, respectively. Following incubation, only live mosquitoes were saved for testing. Survival rates for the mosquitoes that fed on days 3, 4, and 5 were 256(87%) out of 293, 246(92%) out of 266, and 150(64%) out of 236, respectively. All 652 mosquitoes that fed during the 3 days and survived the incubation periods were tested individually for the presence of virus and found to be negative. In addition, after horse number 11 developed clinical symptoms compatible with encephalomyelitis three lots of mosquitoes were fed on animal number 11 on day 8 (a.m. and p.m.) and day 9 (a.m.). The mosquitoes were incubated for 10 days and tested for virus infection in three pools of 30, 41, and 44 specimens. All were negative for the presence of virus.
The only horse from the entire study to show clinical signs of disease was horse 11, which became febrile and showed neurologic signs beginning 8 days after infection. This mare progressed to severe clinical disease within 24 hours and was euthanized on day 9. She had a severe encephalomyelitis and relatively high titers of virus (104.0 to 106.8 PFU/tissue) in several areas of the brain and spinal cord. Horse 16 was euthanized on day 30 because of a preexisting respiratory condition unrelated to WNV infection. Horses 9, 10, 12, 13, 14, and 15 were monitored for 91 days after infection. None showed signs of disease at any time during that period, nor was WNV isolated from any of the serum samples collected biweekly through day 91. Oral and rectal swabs collected daily from days 0 through 10 were negative for virus.
All 12 horses became infected with WNV after being bitten by infected mosquitoes as evidenced by clinical encephalomyelitis in horse 11 and Nt antibody titers in the others. In the first trial, Nt antibody titers >1:10 were first detected in horses 1 to 4 on days 8 to 10 postinfection. Titers in the samples drawn on days 12 and 13 were >1:320, 1:20, 1:160 and 1:80 for horses 1 to 4, respectively. These horses were euthanized and underwent necropsy between days 14 and 16. In the second trial, the seven horses that remained well developed Nt antibody titers >1:10 between days 7 and 11 post infection. Peak titers in seven horses that did not exhibit clinical illness were reached between days 9 through 13. No samples were tested between days 13 and 27, by which time the titer of horse 16 was >1:1,280. Titers for horses 9, 10, and 12 to 15 on day 31 postinfection were 1:320, 1:160, 1:160, 1:160, 1:40, and 1:80, respectively.
The IgM capture ELISA results are summarized in
Absorbance Attributed to specific West Nile virus (WNV) Antigen immuglobulin (Ig) M Interaction in the West Nile Equine IgM Capture enzyme-linked immunosorbent assay (ELISA).
The open circle (o) is the a.m. reading, the plus (+) is the p.m. reading, and the filled-in circle is the last reading (all a.m., hence the circle) before the horse was euthanized. Each number corresponds to the horse number.
Ae. albopictus mosquitoes were used in these experiments because this species was known to be susceptible to WNV infection by mouth and by intracoelomic inoculation and to be capable of transmitting the virus by bite (
The key question addressed in these studies was whether horses develop viremia of sufficient magnitude to serve as amplifying hosts of WNV. The low titers we observed are similar to those seen by Schmidt and Mansoury in Egypt (
The highest viremia titer in horses fed on by the recipient mosquitoes was approximately 460 Vero cell PFU/mL in horse 11 during the morning of day 5 postinfection (
Akhter et al. (
The results of the equine IgM-capture ELISA for WNV antibody indicate that this is a simple and efficient method for detecting antibody at about the same time that neutralizing antibody can be detected (
Our limited work, and that of others cited above, support the conclusion that horses infected with WNV develop viremias of low magnitude and short duration and that infected horses are unlikely to serve as important amplifying hosts for WNV in nature.
Suggested citation: Bunning ML, Bowen RA, Cropp BC, Sullivan KG, Davis BS, Komar N, et al. Experimental Infection of Horses with West Nile virus. Emerg Infect Dis. [serial on the Internet]. 2002 Apr [date cited].
We thank the United States Equestrian Team and Bill Barnes for financial support; Geoff Letchworth (United States Department of Agriculture) for helpful comments during planning meetings and assisting in several necropsies; Roger Nasci for providing mosquito eggs for our subcolony; Nick Panella for technical assistance; Grant Campbell, Juan Lubroth, Dan O’Leary, Leann Thomas, Kevin Grayson, and Suzanna Ryan for critical review of study proposal and design and editing; Sherif Zaki for immunohistochemistry support; Kimlea Medlin, Jennifer Lehman, Amy Delorimier, and Andrew Hickey for participation in the clinical study; and Rebecca Deavours and Ellen Peterson for administration support.
Dr. Bunning is a veterinary epidemiologist in the Division of Vector-borne Infectious Diseases Division, Centers for Disease Control and Prevention. He currently serves as a liaison between the Department of Defense and the Centers for Disease Control and Prevention. His interests include epidemiology and research on the West Nile virus.