Abstract
Ten crows (Corvus brachyrhynchos) and three blue jays (Cyanocitta cristata), species indigenous to North America, were intravenously inoculated with 103 PFU of West Nile virus (WNV) strain NY99 for production of positive tissues for Canadian surveillance. Both species developed clinical signs 4 days postinoculation (dpi). Virus was detected in blood, cloacal and tracheal swabs, and in a number of organs by reverse transcriptase-polymerase chain reaction (RT-PCR) and virus isolation (titers reaching over 107 PFU/0.1 g). Virus appeared as early as 1 dpi in blood (102-103 PFU/ml) and spleen (103-104 PFU/0.1 g of tissue), whereas kidney, liver, intestine, gonads, heart, skeletal muscle, and lung tested positive for WNV in a later stage of the infection. Immunostaining (IHC) using heterologous rabbit anti-WNV polyclonal antiserum detected viral antigen in a wide range of organs, starting at 2 dpi. Detection of WNV antigen in the brain of blue jays and crows by IHC was laborious as only few cells, not present in all sections, would stain positive. Mononuclear cells appeared to be an important target for virus replication, contributing to virus spread throughout tissues during the infection. This conclusion was based on the positive IHC staining of these cells in organs before virus antigen detection in parenchymal cells and supported by virus isolation and RT-PCR-positive results in white blood cells. The inability of blue jays and crows to perch and fly may reflect weakness due to generalized infection and marked skeletal muscle involvement, although involvement of the central nervous system cannot be excluded.
West Nile virus (WNV) is an arthropod-borne virus, enveloped, with a single-stranded, positive-sense RNA genome. The virus belongs to the Japanese encephalitis virus serocomplex within the genus Flavivirus, family Flaviviridae. WNV was considered to be an Old-World arbovirus, widely spread throughout Africa, the Middle East, Europe, and Western Asia, 18 until it was isolated on the American continent in 1999 during a human encephalitis outbreak with accompanying epizootic in a number of species in New York. 1,2,14 It became established in North America the next year, and appears to be spreading throughout the continent. 5,8,16
WNV is transmitted mainly by the bites of infected mosquitoes. Wild birds and mosquitoes are considered to be primary reservoirs of the virus, although a number of species including humans can be infected with the virus and some will develop a potentially fatal disease. 10,11,19
It appears that in the natural infection of corvids with WNV, North American (1999) and related Israel (1998) isolates are more virulent than previously circulating virus strains. Dead birds surveillance in the USA in the year 2000 found 63 avian species positive for WNV, with crows representing 89% of the total positive birds. 13,17 During the initial WNV surveillance program in the USA, WNV was isolated from brain, kidney, and heart, and the viral genome was detected by Taqman reverse transcriptase–polymerase chain reaction (RT-PCR) from brain, kidney, liver, and lung. 13,20 A serosurvey in 1999 in New York detected only 1% of crows with anti-WNV antibodies, similar to 12% in a 1999–2000 survey in Israel, 15,17 likely reflecting the low number of survivors. 2,9,14,21 In contrast, surveys conducted in 1955 and 1956 in Egypt detected WNV antibodies in 88% and 65% of crows, respectively, with no major mortality observed. 22,23 Interestingly, during outbreaks (1964–1968) in the delta of the Volga river, crows and other Corvidae (with 20% seroprevalence) were considered to have an important role in the transmission cycle of the WNV only in the drier, agricultural areas of the upper delta, whereas aquatic birds were found to be important in the ecology of WNV in the lower marshy parts of the Volga delta. 3,4
In contrast to field infections, crows of both the New and Old World appear to be highly sensitive to experimental infections irrespective of the WNV isolate. Experimental infections in crows (Corvus corone sardonius) were used during the ecologic studies of WNV in Egypt in the mid-1950s to obtain information on vertebrate hosts of WNV. 22 Taylor and colleagues infected 13 crows (C. corone sardonius) with the Egypt 101 strain of WNV through Culex mosquito bites. The mortality in the inoculated birds was 100% within 3–7 days postinoculation (dpi), with a high level of viremia. The virus was isolated from brain and spleen. In a second experiment, virus titers in crow blood reached up to 108 LD50 in suckling mice per 0.05 ml of serum, and the virus circulated for up to 6 days in the surviving birds. 23 A similar mortality rate was observed by McLean et al. in the North American crow, Corvus brachyrhynchos, infected with the strain NY99. The birds became viremic with 100% mortality after inoculation. In addition, transmission of the virus was observed between inoculated and control birds, indicating the possibility of direct transmission under the experimental conditions. 17,18 This finding was supported by the experimental infection of North American birds with WNV conducted by Komar and coauthors. 12
Canada established a WNV surveillance program in 2000 based on the US experience in 1999. Subsequently, positive crow and blue jay control tissues required for surveillance testing were prepared. This article describes the experimental infection, viral distribution, pathology in the organs of blue jays and crows, and temporal distribution of the WNV in tissues of infected crows with accompanying pathologic changes, using virus isolation, RT-PCR, histopathology, and immunohistochemistry.
Material and Methods
Virus
WNV variant (strain) NY99 kindly provided by Dr. Michael Drebot, NML, Health Canada, Winnipeg, was used for bird inoculation and virus neutralization tests. The virus stocks were prepared and titrated on Vero V-76 cells.
Cells
Preparation of virus stocks, virus titration, and isolation was done using low-passage Vero V-76 cells (up to passage 50), maintained in M-199 medium, supplemented with 25 mM N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid and 3 mM
Crows and blue jays
Four blue jays were provided by a registered bird bander under permit from the Province of Manitoba (Department of Conservation). Twelve crows were supplied by a private crow-removing company, Hawkeye (Ontario). Possession permits for both species were obtained from the Canadian Wildlife Services, Winnipeg. The animal work was approved by the animal care committee of the Canadian Science Centre for Human and Animal Health and met Canadian Council on Animal Care guidelines. The wild-captured birds were all considered to be of adult age and at least 1 year old. The blue jays were numbered 1–4, the crows 5–12.
Both species were intravenously inoculated with 0.1 ml of 103 PFU/bird of WNV strain NY99. Blue jay No. 2 died on day 4 (postmortem was performed 3 days later), the remaining two inoculated blue jays were euthanatized at 5 dpi (Nos. 3 and 4). One control blue jay (No. 1) was injected with phosphate-buffered saline (PBS) and euthanatized on day 7. Ten crows were inoculated on day 0. Two birds per day were euthanatized on days 1 (Nos. 16 and 11), 2 (Nos. 13 and 7), 3 (Nos. 14 and 10), 4 (Nos. 12 and 15), and 5 postinoculation (Nos. 8 and 9). Two control crows (Nos. 5 and 6), inoculated with PBS, were euthanatized on day 8.
Sample collection
Blood was collected on day 0 (crows and blue jays), 1, 2, 3, 4 (crows only) and on day 5 (crows and blue jays) postinoculation for detection of virus and serum antibodies. Tissues (brain, trachea, lung, spleen, heart, liver, kidney, pancreas, gonads) for virus isolation and RT-PCR were collected during postmortem examination on days 1, 2, 3, 4, and 5 and kept frozen until they were processed. Cloacal, pharyngeal, and nares swabs were collected before the inoculation and at necropsy using Dacron Fiber Tipped wood applicator swabs (Fisher Brand), placed into 2 ml of Eagle's Minimum Essential Medium-alpha supplemented with gentamycin sulfate (50 µg/ml) and
Tissues for pathology (brain, sinuses, trachea, lung, heart, liver, spleen, kidney, pancreas, gonads, sciatic nerve, skeletal muscle, proventriculus, gizzard, cecum, cecal tonsil, ileum, duodenum, tongue, tonsil, skin) were collected at necropsy and fixed in 10% neutral buffered formalin.
Virus isolation
Virus isolation from the tissues was performed as a plaque assay on Vero V-76 cells in 12-well plates (Costar, Corning Inc., Corning, NY) seeded at density 105 cells/cm2, and incubated for 24 hours before the plaque assay. The tissues were prepared for the virus isolation in the following way: the section of each tissue was weighed, placed into a volume of Dulbecco's PBS without Ca++ and Mg++ (Sigma) representing 90% of the weight of the tissue, and than transferred into a homogenization bag. The tissue was homogenized for 30 seconds using the BagMixer 100 blender (Topac, Hingham, MA), with grinding speed set at 9 strokes per second. Clarified (1,550 × g, 20 minutes) 10% w/v tissue emulsion in Dulbecco's PBS without Ca++ and Mg++ (Sigma) was 10−1 to 10−6 serially diluted in M-199, and duplicates of 400 µl/well incubated on cells for 1 hour at 37 C 5% CO2. The inoculum was removed and the wells overlayed with 2 ml of 1.5% carboxymethyl-cellulose sodium salt, medium viscosity (Sigma) in M-199 with 2% FBS. Plates were then incubated for 4 days in 5% CO2 incubator at 37 C, fixed with 4% formalin after the incubation, and stained with 0.5% of crystal violet (80% methanol in Dulbecco's PBS). Before virus isolation from swabs, more gentamycin (500 µg/ml final) was added to the diluent after removal of the swab. The sample was then incubated at room temperature for 1 hour and clarified at 1,550 × g for 10 minutes. Virus isolation from swabs and blood was carried out using the same protocol as for tissues. The isolation for 1 and 2 dpi was also done from blood fractions prepared as described in the RT-PCR section.
The sensitivity of virus isolation was determined by spiking negative control heart tissues from one crow and one blue jay with virus control before homogenization, titrating out the samples on Vero V-76 cells and comparing the titers with the virus control titer.
Reverse transcriptase–polymerase chain reaction
The 10% homogenized fresh frozen tissue suspensions prepared for virus isolation, as described above, were used simultaneously for RT-PCR. RNA was extracted from 100 µl of the tissue suspension using TriPure Reagent (Roche Diagnostic Corporation, Indianapolis, IN) according to the manufacturer's instructions. For blood samples, 100 µl of whole blood or erythrocyte, leukocyte, and plasma fractions were used for RNA extraction. The blood fractions were prepared using the Ficoll-Paque PLUS (Amersham Pharmacia Biotech AB, Uppsala, Sweden) centrifugation method, as recommended by the manufacturer. Harvested cells were resuspended in PBS to correspond with the original blood volume. The RT-PCR reaction was carried out using the Qiagen OneStep RT-PCR kit (Qiagen Sciences, Germantown, MD) with primers: 9483 (5′ CAC CTA CGC CCT AAA CAC TTT CAC C) and 9794 C (5′ GGA ACC TGC TGC CAA TCA TAC CAT C), yielding an amplicon size of 311 bp. The RT reaction was carried out for 60 minutes at 50 C, followed by 45 cycles of PCR (1 minute at 95 C, 30 seconds at 60 C, 20 seconds at 72 C, with final extension for 7 minutes at 72 C and hold at 4 C).
Antibody detection
Preinoculation bleed (day 0) and final bleed (days 1, 2, 3, 4, 5) sera from both avian species were tested for antibody presence by the microtiter plaque reduction neutralization test against WNV, NY99, based on De Madrid and Porterfield. 7 The assay was carried out with modifications in the following manner: 100 µl of serial twofold dilutions of sera (starting with 1:10) was incubated with 100 µl of 200 PFU of Egypt 101 strain for 1 hour at 37 C, 5% CO2 in cell-free 96-well microtiter plates (Costar, Corning Inc.). The serum-virus mixture (100 µl/well) was transferred onto the Vero V-76 cells grown in 96-well plates, seeded 24 hours before the test at density 105 cells/cm2. The medium was removed from the wells just before the transfer of the serum-virus mixture. The plates were incubated for 1 hour at 37 C, 5% CO2. The carboxymethyl-cellulose overlay (100 µl/well, 1.5% in M-199, 2% fetal calf serum) was added to the inoculum and incubated for 3 days at 37 C, 5% CO2. The cells were fixed with 4% formaldehyde at the end of the incubation period and stained with 0.5% crystal violet. The number of plaques per well was verified by virus back titration. Serum dilutions causing at least 70% plaque reduction were considered positive for presence of anti-WNV antibodies.
Histopathology
Standard histologic procedures were used to prepare the formalin-fixed tissues for microscopic examination. Tissue sections were stained with hematoxylin and eosin.
Rabbit polyclonal anti-WNV serum
The animal work was approved by the animal care committee of the Canadian Science Centre for Human and Animal Health and met the Canadian Council on Animal Care guidelines. Two New Zealand White rabbits (specific pathogens free) obtained from Charles River, Canada, were injected intramuscularly with 0.2 ml of 107 PFU of live WNV, topotype strain Egypt 101, and boosted with the same amount of virus on day 21. The rabbits were bled before virus inoculation and on days 7, 14, and 21 postinoculation. The rabbits were exsanguinated under anesthesia at 28 dpi and euthanatized at the end of bleeding. The neutralizing serum antibody titers (2,560 for rabbit No. 1 and 1,280 for rabbit No. 2) were the same against both viruses, WNV Egypt 101 and WNV NY99.
Immunohistochemistry
Formalin-fixed, paraffin-embedded sections of individual organs from blue jays and crows, treated with proteinase K (DAKO®, Carpinteria, CA), were stained with polyclonal rabbit serum (see above) using the DAKO EnVision™ + System Peroxidase (3,3′-diaminobenzidine [DAB]) anti-rabbit kit. Rabbit anti–bovine serum albumin serum (ICN/Cappel Pharmaceuticals, Aurora, OH) was used as irrelevant primary antibody in the negative-staining control slides. The optimum dilution of the primary antibodies was determined to be 1:400 in 1 M Tris–HCl, pH 7.6, on positive control tissues (kindly provided by Dr. Louis Sileo, USGS, Madison, WI). The tissues were incubated with primary antibodies for 30 minutes at 37 C, followed after washing by incubation with DAKO® Labeled Polymer, horseradish peroxidase Anti-Rabbit according to the manufacturer's instructions, and observed using DAB in chromogen solution. The stained tissue sections were counterstained with Gill's hematoxylin. Staining of uninfected bird tissues (birds No. 5 and 6) was performed along with the immunostaining (IHC) staining of tissues from infected birds.
The performance of the polyclonal rabbit serum was compared with the monoclonal antibody NIAID V-554-701-562 (WNV strain B956; ATCC) because we were unable to obtain antibodies against homologous virus at the time of the experiment.
Results
Clinical signs
The control birds (blue jays and crows) appeared clinically normal throughout the study. All inoculated blue jays appeared normal until 4 dpi, when one blue jay (bird No. 2) became mildly lethargic and was reluctant to fly. The bird was found dead on day 5. The remaining two birds were euthanatized on 5 dpi when they became lethargic and reluctant to fly similar to bird No. 2. All crows were clinically normal until 4 dpi when the birds became depressed, ate and drank less, produced less feces, and were reluctant to fly. On 5 dpi, crow No. 8 was only able to walk, stumbled and staggered while walking, and was perched on the floor waterer.
Reverse transcriptase–polymerase chain reaction
The RT-PCR was done simultaneously with the virus isolation using the same sample preparation (whole blood, blood fractions, plasma, pharyngeal rectal and nares swabs) or the same tissue homogenates (brain, trachea, lung, spleen, heart, liver, kidney, pancreas, gonads). The 311-bp amplicon from the 3′ end of the NS5 gene region of the genome was detected in all tissue samples that yielded virus on isolation (Table 1). The RT-PCR detected virus RNA 1 day earlier than virus isolation only in swabs from trachea at 1 dpi from crows, and from cloaca only in crow No. 16. RT-PCR was performed on whole blood from blue jays and blood fractions (plasma, red blood cells, white blood cells) from crows at 1 and 2 dpi. The RT-PCR worked well on the whole blood of blue jays, but no positive amplicons were obtained from the whole blood of crows, despite the positive results from individual fractions (Fig. 1).
Virus isolation, blue jays and crows. Virus titer is given in log10 PFU/0.1 g of a tissue or 0.1 ml of blood.∗.
∗ ND = not done; tox = toxicity.
RT-PCR results for 1 and 2 days postinnoculation (dpi) in crow blood and swabs. Lanes 1–5: crow No. 11, 1 dpi (cloacal swab, tracheal swab, white blood cells, red blood cells, plasma). Lanes 6–10: crow No. 16, 1 dpi (cloacal swab, tracheal swab, white blood cells, red blood cells, plasma). Lanes 11–16: crow No. 13, 2 dpi (cloacal swab, tracheal swab, white blood cells, red blood cells, plasma, whole blood). Lanes 17–22: crow No. 7, 2 dpi (cloacal swab, tracheal swab, white blood cells, red blood cells, plasma, whole blood). Lane 23: negative control. Lane 24: positive control. Lane 25: 100-bp DNA ladder. The bright band indicates the 500-bp marker.
Virus isolation
The processing of samples for virus isolation did not significantly influence the detection of infectious virus. The titer between the virus control and tissues spiked with virus before processing differed within 1 log: spiked crow heart yielded 107. 2 PFU/ml, blue jay heart 107. 1, and the virus control 106. 3 PFU/ml.
Virus isolation was attempted from blood, spleen, liver, kidney, gonads, brain, heart, lung, trachea, and pancreas (Table 1). At 1 dpi, virus was isolated from the spleen of both crows (Nos. 11 and 16) and from the blood of crow No. 11, likely detecting the virus present in the blood in both cases. Beside the whole-blood isolation, virus isolation from white blood cells, red blood cells, and plasma was attempted at 1 and 2 dpi. On 1 dpi, virus was detected only in the plasma of bird No. 11 (167 PFU/ml of original blood volume). On 2 dpi (birds Nos. 13 and 7), the virus was found in white blood cells (350 PFU/ml), plasma (300 PFU/ml), and red blood cells (125 PFU/ml). Viremia was detected until euthanasia or death (days 4 and 5), with the peak at 3 dpi. On day 4 postinoculation, neutralizing antibodies in the serum became detectable, possibly affecting the virus titers (Table 1). Virus isolation from the organs listed above corresponded with the immunohistochemistry staining, except for pancreas. Although we isolated virus from the pancreas specimens, IHC was negative, and we conclude that the detected virus originated from blood or duodenum accidentally collected together with pancreas. Overall highest titers per tissue weight were detected in the lungs (more than 108 PFU/0.1 g of tissue) on day 4 postinoculation. The titers obtained from trachea would start at 101. 8 PFU/ml at 2 dpi and reach 105. 23 PFU/ml on 5 dpi. The blue jay No. 2, dead at 4 dpi and processed 3 days later, still had relatively high virus titer in the tracheal swab (103. 7 PFU/ml) but no detectable virus in the cloacal swab. Cloacal swabs from birds immediately after euthanasia ranged from 101.67 PFU/ml at 2 dpi to 105.48 PFU/ml on day 5 postinoculation.
Gross pathology
All control birds (blue jays and crows) had no significant lesions. Blue jay No. 2 had mild congestion of the cerebral cortex, proventriculus, testicles, spleen, and left lung, whereas blue jay No. 3 had severe splenomegaly, with congestion and hemorrhage. The crows had mild congestion of the cerebral cortex and spleen from 1 dpi through 4 dpi. On dpi 5, crow No. 9 had pinpoint hemorrhages in the proventriculus, skeletal muscle, and liver. Crow No. 8 on 5 dpi was more severely affected with moderate to severe congestion of the cerebral vasculature, lung, heart, skeletal muscle, kidneys, and proventriculus, with distension of the gall bladder. The air sacs were opaque and adhered to the gizzard.
Histopathology
No lesions were observed in tissues collected from the control birds. In the lungs of blue jays inoculated with WNV, there was vacuolation of cells in the lamina propria of respiratory capillaries as well as in the connective tissues surrounding smooth muscle of the tertiary bronchi. Scattered throughout the lamina propria were phagocytic cells filled with debris.
Randomly distributed necrosis of individual hepatocytes became noticeable at 3 dpi in crows. By 4 and 5 dpi, there were a low to moderate number of hepatocytes in both species with cytoplasmic vacuolation, karyolysis, and karyorrhexis. Mononuclear phagocytic cells were prominent in periportal and sinusoidal areas, together with a mixed inflammatory reaction.
Scattered skeletal muscle fibers of blue jays were swollen, finely vacuolated, and had loss of striation, with occasional fragmentation of sarcomeres. Myocardial fibers were occasionally swollen, with pale staining and clumping of sarcoplasm.
In the spleen of both species, there was moderate to marked congestion. Necrosis of lymphocytes in splenic lymphoid follicles was prominent at 2 dpi. By 5 dpi, lymphoid necrosis was widespread. Inflammatory cell aggregates were prominent in the walls and periphery of sheathed arterioles. Fibrin deposition was present in these areas of inflammatory reaction. Macrophages in perivascular locations were few at 2 dpi, but became increasingly prominent in later stages of infection, eventually spreading throughout the splenic parenchyma.
Immunohistochemistry
A number of organs were tested for presence of viral antigen by immunohistochemistry using rabbit polyclonal antiserum prepared against WNV. The overview of the organs and the time course of the antigen detection are summarized in Table 2, with the exception of IHC-negative organs.
Immunohistochemistry in positively stained organs.∗.
∗ dpi = days postinnoculation; ND = not done; o = ovary; m ϕ = resident monocytes/macrophages stained positive; T = tubular glands only; L = lamina propria.
† Only few parenchymal cells stained positive for the WNV antigen.
‡ Few tubules stained positive for the WNV antigen.
No virus antigen was detected in pancreas, sciatic nerve, skin, or feather follicles of both blue jays and crows. From thirty tested sections per brain, only very few neuronal cells were found positive in brains of blue jays and crows in one or two slides at 4 and 5 dpi. Antigen-positive monocytes/macrophages were observed on some sections at 3–5 dpi within the blood vessels. Staining was prominent in a large number of myocardial cells in the heart of blue jays, whereas only a few myocardial cells from crows were antigen positive, starting at 4 dpi. There were pronounced differences in staining patterns of skeletal muscles between crows and blue jays, with almost no staining in crow tissues compared with relatively prominent staining in tissues of blue jays (Figs. 2–4).
Heart; crow, bird No. 12, 4 days postinnoculation (dpi). WNV antigen is present only in occasional myocardial cells. Immunoperoxidase labeling, hematoxylin counterstain. Bar = 45 µm.
Heart; blue jay, bird No. 3, 5 dpi. WNV antigen is present in a high number of myocardial cells. Immunoperoxidase labeling, hematoxylin counterstain. Bar = 45 µm.
Skeletal muscle; blue jay, bird No. 3, 5 dpi. WNV antigen is present in the striated muscle cells. Immunoperoxidase labeling, hematoxylin counterstain. Bar = 31 µm.
There was staining of macrophages located in the tracheal submucosa. Numerous antigen-laden cells appeared to be moving through the tracheal mucosa, with an occasional cell located on the lumenal surface. Positively stained macrophages were very prominent in submucosal lymphoid tissue sites. Tracheal epithelium was vacuolated and attenuated, but brush borders were usually intact and staining in epithelial cells was not a feature.
Positive staining was detected in liver at 2 dpi in hepatocytes and Kupffer cells. Gonads were affected in blue jays as well as in crows starting at 3 dpi. Antigen was detected in both ovaries and testis. It appears that at a later stage in the infection, the infected granulosa cells of ovaries started to produce large amounts of viral antigen (Figs. 5–7).
Testis; blue jay, bird No. 4, 5 dpi. Immunostaining indicates infection of scattered individual interstitial cells. Immunoperoxidase labeling, hematoxylin counterstain. Bar = 45 µm.
Ovary; crow, bird No. 10, 3 dpi. WNV antigen staining is present in the theca cells of the follicle. Immunoperoxidase labeling, hematoxylin counterstain. Bar = 31 µm.
Ovary; crow, bird No. 12, 4 dpi. Immunolabeling indicates infection of granulosa cells that appear to convey viral antigen into the yolk as they slough. Immunoperoxidase labeling, hematoxylin counterstain. Bar = 31 µm.
Antigen-positive monocytes/macrophages were detectable in the sinusoids of the red pulp of the spleen at 1 dpi in crow No. 11. By day 3 postinfection, antigen-stained macrophages were in lymphoid follicles, and after 4 dpi they were widespread throughout the entire splenic parenchyma (Figs. 8, 9).
Spleen; crow, bird No. 7, 2 dpi. WNV antigen is present in macrophages surrounding blood vessel. Immunoperoxidase labeling, hematoxylin counterstain. Bar = 31 µm.
Spleen; crow, bird No. 8, 5 dpi. Immunolabeled macrophages are scattered throughout the parenchyma. Lymphocyte depletion is marked. Immunoperoxidase labeling, hematoxylin counterstain. Bar = 31 µm.
Viral antigen was first detected in kidney at 3 dpi in macrophages located interstitially near blood vessels. Staining in tubules (distal convoluted and collecting) was detectable at 4 dpi, and in blue jays antigen was detected at 5 dpi in glomeruli.
Positive staining was present in the intestine of some birds, primarily in crypt epithelial cells, with heaviest staining in latter stages of disease progression. In proventriculus, staining first appeared within macrophages of the intraglandular and submucosal lamina propria, was then present in the glandular epithelial cells of the compound tubular alveolar glands, and later was present in the columnar epithelium in the centers of individual glands as well as on the lumenal surface.
Discussion
Based on our results, tracheal swabs and spleen from dead crows are recommended for sample collection. Virus could be detected in these tissues, both by virus isolation and RT-PCR, not only in the early stages of infection but also 2 days after death. With suspensions from liver and kidney collected several days postmortem, a cytotoxic effect in the cell culture was observed; however, these two types of samples yield consistent results on RT-PCR. Trachea posed a problem during homogenization, but all the other tested organs provided good specimens for both RT-PCR and virus isolation. The use of polyclonal rabbit sera raised against heterologous strain did not affect results of the virus neutralization assays. It was not expected to be a concern also because of reported high antigenic cross-reactivity among Flaviviruses (the original WNV outbreak was initially considered, based on serology, to be an outbreak of St. Louis encephalitis) and even higher among the WNV strains/isolates. 6
Immunohistochemical detection of virus was found to be suitable in birds, which died later in the infection (4 to 5 dpi in our experiment). Interestingly, we were not able to readily detect viral antigen in the crow and blue jay brain sections. A number of sections had to be examined to detect antigen-positive cells, although the staining was quite prominent.
IHC was used for better understanding of pathogenesis and to confirm the sites of virus replication because virus detection/RT-PCR in individually harvested organs may not always indicate involvement of the organ in virus replication. For example, although we isolated virus from the pancreas samples, it is likely that the detected virus originated from blood or duodenum accidentally collected together with pancreas because no viral antigen was detected in pancreatic cells by immunohistochemistry. Whereas Komar and others reported virus isolation from skin of crows, we did not detect virus antigen by IHC in skin, perhaps for a similar reason. 12
We hypothesize that monocytes/macrophages are important target cells, contributing to the virus spread throughout tissues during the infection. The virus was detected very early in the infection by RT-PCR and virus isolation in the white blood cell fractions. In addition, macrophages were the first cells to stain by IHC in all the organs, followed later by parenchymal cells, and lastly, epithelial cells of some organ systems. In a later stage of the infection, the intensively stained macrophages became enlarged, most likely because of phagocytosis of other infected cells and debris containing viral antigen in addition to the virus replication. Virus replication in epithelial cells was not consistent in all organ systems and will require further study to elucidate the reason for this.
Almost all the tested organs in both species were involved in virus replication: lymphoid tissue and cells of the lung, spleen, liver, intestine, gonads, and kidney. In blue jays, the striated muscle tissues (heart, skeletal muscle) showed surprisingly prominent IHC staining, not seen in the crows. This apparent replication in muscle may be related to clinical signs seen in the experimental birds and could conceivably be linked to their inability to perch or fly in the later stages of the disease process.
Our results correspond well with the pathologic findings presented in field cases by Steele and colleagues 21 and Anderson and colleagues. 1 They illustrate the hematogenous spread of the virus within the host, and indirectly confirm suggested bird to bird transmission, through shedding from the respiratory or digestive tract.
Footnotes
Acknowledgements
We would like to thank Kevin Hills and Jason Gren for technical support and Stacey Halayko and Julie Kubay for the animal care aspects of the work.