Patogenia de la infección experimental con
dos cepas mediterráneas del virus West Nile
en perdiz roja
Gamino V, Pérez-Ramírez E, Gutiérrez-Guzmán AV, Sotelo E, Jiménez-
Clavero MA, Höfle U. Pathogenesis of two Mediterranean West Nile
virus strains in experimentally infected red-legged partridges. En
107
RESUMEN
La re-emergencia del virus West Nile (WNV) en Europa y la cuenca mediterránea en las últimas dos décadas se atribuye a cambios aminoacídicos que se han producido en las cepas del virus con el consecuente incremento en su virulencia para humanos, caballos y aves. En el presente trabajo, estudiamos las diferencias en la patogenia de la infección experimental de perdices rojas (Alectoris rufa), de siete semanas de edad, con dos cepas mediterráneas de WNV, Morocco/2003 (MO03) y Spain/2007 (SP07). El objetivo fue determinar las causas del diferente curso de la infección y la diferente mortalidad encontrados en un trabajo previo. Además, estudiamos la dinámica de activación y llegada de células inflamatorias al cerebro y cerebelo de las perdices infectadas. En ambas infecciones, el día 6 post inoculación (dpi) fue cuando la carga viral fue mayor, el antígeno viral estuvo más ampliamente distribuido y fue más abundante, y las lesiones microscópicas fueron más severas. Los órganos más afectados fueron el corazón, hígado y bazo. En el cerebro y cerebelo, la microglía fue la población inflamatoria más prevalente y las células T CD3+ infiltraron este tejido el 6 dpi. Comparando ambas infecciones, se observaron relativamente pocas diferencias en la carga viral, distribución del virus y naturaleza y severidad de las lesiones; sin embargo, en las perdices infectadas con MO03 las lesiones microscópicas en el cerebro y cerebelo aparecieron de forma más aguda y fueron más severas, lo que dio lugar a una encefalitis más marcada. Considerando nuestros resultados, el mayor impacto de la infección por MO03 para las perdices rojas está probablemente relacionado con marcadores específicos que hacen que esta cepa sea más neurovirulenta.
108
ABSTRACT
West Nile virus (WNV) re-emergence in Europe and the Mediterranean basin in the last two decades is thought to be related to amino acid substitutions in the virus strains with a subsequent increase in their virulence for humans, horses and birds. In the present work, we studied differences in the pathogenesis of the experimental infection of 7-week-old red-legged partridges (Alectoris rufa) with two different Mediterranean WNV strains, Morocco/2003 (MO03) and Spain/2007 (SP07). The objective was to elucidate the reasons of the different infection course and mortality observed in a previous work. Additionally, we studied the dynamics of inflammatory cell activation and recruitment into the brain of the infected partridges. In both infections, day 6 post-inoculation (dpi) was the day when viral load was higher, virus antigen was more widespread and abundant, and microscopic lesions were more severe. The most affected organs were the heart, liver and spleen. In the brain, the more prevalent inflammatory cell population was microglia and CD3+ T cells infiltrated this tissue on 6 dpi. Comparing both infections, few differences were observed in viral load, virus distribution and lesion nature and severity; however, in MO03 infected partridges microscopic lesions in the brain were more acute and severe, which led to a more marked encephalitis. Considering our results, the higher impact of MO03 infection on red-legged partridges is probably related to specific markers of this WNV strain that make it more neurovirulent.
109
INTRODUCTION
West Nile virus (WNV) is an arthropod-borne flavivirus whose natural cycle involves bird hosts and mosquito vectors, with horses and humans as accidental or dead- end hosts (Kramer et al., 2008). Nowadays, it is considered one of the most widely distributed arboviruses in the world, causing thousands of human, equine and bird encephalitis cases and mortalities, both in the Old and New World (CDC; ECDC). In the Mediterranean basin, WNV activity has increased in the last two decades, and it has been associated to several outbreaks affecting mainly humans and horses (Murgue et al., 2001; Sotelo et al., 2011a). In this area, with the exception of the outbreak reported in migrating storks (Ciconia ciconia) and geese (Anser anser domesticus) in Israel between 1998-1999 (Malkinson et al., 2002), there have only been sporadic cases of avian mortality (Jourdain et al., 2007; Jiménez-Clavero et al., 2008; Monaco et al., 2010; Savini et al., 2012, 2013). WNV re-emergence is thought to be associated mainly to genetic factors of the virus and/or environmental factors (Sotelo et al., 2011a). The variability of amino acid sequence of Mediterranean WNV isolates between 1998 and 2003 was low. However, recent isolates show higher frequency of amino acid substitutions that can lead to phenotypic changes (Sotelo et al., 2009; 2011a), as demonstrated in field and in several experimental studies (Beasley et al., 2005; Davis et al., 2005; Wicker et al., 2006). In birds, genetic differences in WNV strains have been shown to modify the degree of peripheral virus replication and its virulence (Brault et al., 2004; Langevin et al., 2005; Totani et al., 2011).
In a recent experimental study we demonstrated that the red-legged partridge (Alectoris rufa), an endemic Mediterranean gallinaceous bird species (Blanco-Aguiar et al., 2003), is susceptible to the experimental infection with two Western Mediterranean WNV strains, Morocco/2003 (MO03) and Spain/2007 (SP07) (Sotelo et al., 2011b). However, virulence of both strains was markedly different; with 70%
110 mortality in MO03 infected partridges as compared to 30% mortality in SP07 infected birds. In that study, it was indicated that one or a combination of the 13 amino acid differences detected between the two WNV strains were probably determining differences in their virulence. Nevertheless, in the aforementioned study, detailed pathogenesis of the infection by each strain that led to differences in the onset of clinical signs and mortality was not addressed.
For this reason, in the present study, we compare the dynamics of virus appearance and distribution as well as of microscopic lesion distribution and severity in different tissues during the course of the experimental infection of red- legged partridges with these two Mediterranean WNV strains. Additionally, and considering the importance of the central nervous system (CNS) infection in the pathogenesis of WNV disease, we study the dynamics of immune cell activation and recruitment into the brain of the partridges. To our knowledge, this has not been previously documented in WNV infected birds.
MATERIAL AND METHODS
Animals
Recently hatched red-legged partridge chicks were obtained from a commercial breeder and raised in the experimental farm of the Instituto de Investigación en Recursos Cinegéticos (IREC) until they were six weeks of age. Prior to the experimental infection, they were tested by ELISA for the presence of WNV antibodies (ID Screen© West Nile Competition, IdVet, Montpellier, France) and by real time RT-PCR for detection of WNV genome in cloacal and oral swabs, in order to ensure they were not previously infected by the virus (Sotelo et al., 2011b).
111
Viruses
In this study we used two different strains of WNV that had been isolated in the Mediterranean basin; Morocco/2003 (MO03), isolated from a horse and cultivated in Vero cells, and Spain/2007 (SP07), isolated from a golden eagle (Aquila chrysaetos) and cultivated in BSR cells (clone of BHK-21 cell line). Details on the inoculates are given in Sotelo et al. (2011b).
Experimental infection
Partridges were transported to the biosafety level 3 (BSL-3) facilities in the Centro de Investigación en Sanidad Animal (CISA). After five days for acclimatization, two groups of ten 7-week-old red-legged partridges were subcutaneously inoculated in the cervical region with 104 PFU/individual of either WNV SP07 or MO03 diluted in
up to 0.1 mL Dulbecco's Minimum Essential Medium (DMEM) (supplemented with 2 mM L-glutamine, 100 U/mL penicillin and 100 μg/mL streptomycin). A control group was sham-inoculated with an equivalent volume of DMEM and placed in a separate cage. Inoculated partridges were observed daily for clinical signs or death and two birds of each group were euthanized by intravenous injection of embutramide (T61 ®, Intervet – Schering-Plough, Madrid, Spain) on days 3 (Nos. 1- 4), 6 (Nos. 5-8) and 14 (Nos. 9-11) post-inoculation (dpi).
The experiment was performed following biosafety animal welfare and ethical rules applicable in the EU. Food and water were provided ad libitum throughout the experiment.
Sample collection
Detailed necropsies were performed on euthanized individuals. Samples of brain, heart, lung, liver, spleen, kidney, thymus, bursa of Fabricius and feather pulp were collected into sterile polypropylene tubes filled with 1 mL of Hanks' balanced solution (10% glycerol, 200 U/mL penicillin, 200 μg/mL strepctomycin, 100 U/mL
112 polymixin B sulphate, 250 μg/mL gentamicin and 50 U/mL nystatin) and stored at -70 ºC until analysis. In addition, samples of brain, oral mucosa, thymus, heart, trachea, lung, liver, spleen, kidney, small and large intestine, pancreas, cecal tonsils, bursa of Fabricius, pectoral muscle and skin with feather follicles were fixed in 10% neutral buffered formalin.
Virus genome detection
RNA was extracted from tissue samples after homogenization and tested by real time RT-PCR for the presence of WNV genome as described in Sotelo et al. (2011b).
Histopathology
Formalin-fixed tissue samples were trimmed, embedded in paraffin and processed to obtain hematoxylin and eosin stained sections. These were independently examined by two different investigators (UH and VG) to determine the presence of WNV associated lesions. When lesions were present, these were graded according to their distribution (focal, multifocal or diffuse) and severity (mild, moderate or marked).
Immunohistochemistry
Tissue sections were also mounted on Vectabond™ reagent (Vector Laboratories, Inc., Burlingame, CA) pretreated slides. Immunohistochemical detection of WNV antigen was performed using a rabbit polyclonal antibody (BioReliance, Product 81- 015, Rockville, MD) following the protocol described previously (Gamino et al., 2012). We also characterized the inflammatory cell population in the cerebrum and cerebellum. For that purpose, we used primary antibodies, reagents and protocols detailed in table 1. In all cases endogenous peroxidase activity was inhibited with a peroxidase blocking reagent (Dako EnVision®+System-HRP (AEC), DakoCytomation, Carpinteria, CA) (CD3, CD79, GFAP) or with 3% H202 diluted in methanol (RCA),
113 rinses were performed using 0.1% Tris-buffered saline/Tween20 (TBS 0.05 M, pH 7.5), non-specific primary antibody labeling was blocked with 2% albumin from bovine serum (BSA) (Sigma-Aldrich Chemie, Steinheim, Germany) diluted in 0.1% TBS/Tween20, and sections were counterstained with Mayer's hematoxylin.
In the immunohistochemistry (IHC) for WNV antigen detection, tissue sections of experimentally infected red-legged partridges, in which presence of WNV had been confirmed by real time RT-PCR, served as positive controls. Negative controls included several sections with substitution of the primary antibody by 2% BSA-0.1% TBS/Tween20 and negative rabbit antibody (BioReliance, Product 81– 015), as well as tissue sections of a red-legged partridge negative for WNV infection.
In the IHC for T and B cells, positive controls included sections of spleen and bursa of Fabricius of red-legged partridges. Negative controls included substitution of the primary antibody by 2% BSA-0.1% TBS/Tween20 and a brain section of a red-legged partridge that tested negative for WNV presence by real time RT-PCR. A brain section of a non-infected partridge of the same age served as reference for RCA and GFAP.
We graded virus antigen staining according to its distribution and abundance. Moreover, we characterized the distribution of inflammatory cells in the brain and analyzed changes in their relative abundance during the course of infection.
114
*Primary antibody products: RCA-1 product No. B-1085 (Vector Laboratories); GFAP product No. Z0334 (DakoCytomation, Glostrup, Denmark); CD3 product No. A0452 (DakoCytomation); CD79a product No. RM-9118 (Thermo Fisher Scientific, Runcorn, UK).
**Proteinase K (DakoCytomation); RT: room temperature (22-25 ºC). †Antibodies were diluted in 2% BSA-0.1% TBS/Tween20.
♦ON: overnight.
♠Goat anti-rabbit IgG (Vector Laboratories) was diluted 1:200 in 0.1% TBS/Tween20 and applied for 1 hr at RT; Labelled polymer-HRP anti-rabbit (Dako EnVision®+System-HRP (AEC), DakoCytomation) was applied according to manufacturer's recommendation.
°Avidin-biotinylated enzyme complex (ABC system, Vector Laboratories) was applied for 30 min according to manufacturer's recommendation and 3,3´-diaminobenzidine tetrahydrochloride (DAB, Vector Laboratories) was applied for 30 sec according to manufacturer's recommendations. AEC+substrate chromogen (Dako EnVision®+System-HRP (AEC), DakoCytomation) was applied for 15 min (CD3, CD79) and 3 min (GFAP) according to manufacturer's recommendation.
RESULTS
Macroscopic findings
Macroscopic lesions were observed in the necropsy of all euthanized animals from 3 dpi on. The most affected organs were the heart, spleen, liver and kidneys, with no remarkable difference between SP07 and MO03 inoculated partridges. Main macroscopic lesions were pale myocardium and pale hepatic, splenic and renal parenchyma, which, in some cases, also showed diffuse petechiae. On 14 dpi, one SP07 inoculated bird (No. 10) had congestion in the kidney and spleen that also was observed in the heart and the lung of one MO03 infected partridge (No. 11). Partridge No. 11 showed cerebral vessel injection that was also observed in one SP07 infected bird (No. 9).
Cell population/ marker Antibody reference* Pretreatment** Primary antibody dilution† Primary antibody incubation♦ Secondary
antibody♠ Detection system°
Microglia-
macrophages Lectin RCA-1 biotinylated
Citrate buffer- microwave heat (22 min)
1:600 45 min RT Goat anti-rabbit IgG ABC-DAB
Astrocytes Polyclonal rabbit anti-GFAP Proteinase K (7 min RT) 1:500 4ºC ON
Labelled polymer-HRP anti-rabbit AEC+ substrate chromogen
T cells Polyclonal rabbit anti-human CD3
Citrate buffer- microwave heat (22 min) 1:500 4ºC ON B cells Monoclonal rabbit anti-human CD79a 1:10 45 min RT
Table 1. Detail of reagents and protocols used in the IHC for inflammatory cell characterization in the CNS of experimentally WNV infected red-legged partridges.
115
Virus genome detection
In the WNV real time RT-PCR, all tissue samples collected in the necropsy of inoculated animals tested positive from 3 dpi on, with lower Ct values (i.e. higher viral loads) on 6dpi (Table 2). Ct values were similar between MO03 and SP07 infected partridges or slightly lower in some tissues of the MO03 infected group (Table 2). On 3 dpi, the spleen followed by the kidney and heart were the organs with highest viral load (Table 2). On 6 dpi, the highest viral load was detected in the heart but also in the spleen and kidney (Table 2). On this dpi, the feather pulp was markedly positive, especially in bird No. 8, and the brain showed the lowest Ct values (Table 2). On 14 dpi, viral loads were low in most cases, and several tissues were negative in the SP07 infected partridge (No. 10) (Table 2). Control animals tested negative for WNV RNA throughout the experiment.
*NA: tissue sample no analyzed.
◦No Ct: tissue sample with Ct ≥40 (considered negative).
Histopathology
Histologic lesions appeared as early as 3 dpi, but were more severe and widespread on 6 dpi (Table 3). The most affected tissues in both groups were the heart, spleen, liver and skin with feather follicles (Figures 1-4). The brain, lung and kidney were
3 dpi 6 dpi 14 dpi
SP07 MO03 SP07 MO03 SP07 MO03
TISSUE No.1 No. 2 No. 3 No. 4 No. 5 No. 6 No. 7 No. 8 No. 9 No. 10 No. 11
Brain 30.3 32.8 31.4 33.9 25.2 26.5 25.2 24.8 NA No Ct 37.6 Heart 28.7 28.5 25.4 24.8 19.6 20.3 23.7 18.8 NA 38.0 33.8 Lung NA 30.9 31.0 30.1 26.3 25.7 29.2 24.2 NA No Ct 38.8 Liver NA 30.2 29.6 30.9 28.7 25.7 30.8 28.2 NA No Ct No Ct Spleen 25.3 23.7 26.5 24.4 29.7 22.7 24.6 21.8 NA 36.1 34.0 Kidney 26.3 27.5 24.5 26.0 23.0 22.4 23.7 20.9 NA 37.5 36.1 Thymus 27.6 29.3 31.9 23.5 30.2 23.4 33.0 NA NA No Ct 29.0
Bursa of Fabricius NA* 32.0 32.8 34.2 27.9 24.8 31.5 29.8 NA No Ct 31.3
Feather pulp No Ct◦ 31.4 27.4 31.3 18.2 20.8 22.8 16.9 NA 34.6 33.7
Table 2. Detection of WNV genome by real time RT-PCR (Ct values) in tissues of experimentally infected red-legged partridges.
116 moderately affected and the oral mucosa, pancreas, cecal tonsils, thymus and intestines showed, in most cases, mild lesions (Table 3). There were no microscopic lesions in the trachea and pectoral muscle.
Main microscopic findings were the presence of lymphoplasmacytic, histiocytic and/or granulocytic inflammatory infiltrates, cellular necrosis and/or degeneration, and hemosiderosis in the spleen and liver (Table 3 and figures 1-4). In the brain, the most consistently observed lesion was endothelial cell swelling (Figure 5), both in the cerebrum and cerebellum, but also mild neuronal necrosis (affecting also Purkinje cells) diffuse gliosis or glial nodules, and rare perivascular cuffing (Table 3 and figures 5 and 6). All these lesions were observed in both groups but were slightly more severe in some tissues of MO03 infected partridges, especially in the brain (Table 3).
On 3 dpi, the main microscopic finding was the presence of widespread moderate to marked inflammatory infiltrates, although mild cardiac myofiber degeneration was also observed (Table 3). In the brain, endothelial cell swelling was the only detected lesion, mainly in MO03 infected partridges (Table 3). On 6 dpi, microscopic lesions increased in severity, with the appearance of cellular necrosis in many tissues, which was especially severe in the liver in both groups (Table 3 and figures 1 and 3). One MO03 infected partridge (No. 8) showed multifocal macrophages with foamy and distended cytoplasm in the spleen (Figure 2). On this dpi, mild to moderate neuronal necrosis and gliosis appeared in the brain of MO03 infected partridges (Table 3 and figures 5 and 6). On 14 dpi, microscopic lesions were still widespread, with moderate to marked severity in the heart and spleen that was milder in the liver (Table 3). On this dpi, renal tubular epithelial cell necrosis appeared and both SP07 and M003 infected birds showed mild to moderate lesions in the brain (Table 3).
117 Incidental findings included the presence of granulomas in the liver and pancreas of birds euthanized on 3 and 6 dpi and coccidian oocysts in the enterocytes of the large intestine. These parasites were especially numerous in individuals euthanized on 6 dpi and in partridges infected with MO03. Thus, one MO03 infected partridge (No. 8) showed an associated severe necrosis of the intestinal mucosa.
Table 3. Microscopic lesions and IHC staining gradation on different days post-inoculation (dpi) in tissues of experimentally WNV infected red-legged partridges.
3 dpi 6 dpi 14 dpi
SP07 MO03 SP07 MO03 SP07 MO03
No. 1 No. 2 No. 3 No. 4 No. 5 No. 6 No. 7 No. 8 No. 9 No. 10 No. 11
Cerebrum†
Neuronal necrosis − − − − − − + + + − +
Gliosis − − − − − + + ++ + + −
Perivascular cuffing − − − − − − − − − + + Endothelial cell swelling + + ++ ++ + + + ++ ++ + + Immunohistochemical staining♦ − − − − − − − − − − −
Cerebellum
Purkinje cell necrosis − − − − − − ++ + + NA* +
Gliosis − − − − − − ++ + − NA ++
Perivascular cuffing − − − − − − − − + NA − Endothelial cell swelling − − ++ ++ + + +++ ++ ++ NA ++ Immunohistochemical staining − − − − − − − − − NA − Heart Myofiber necrosis-degeneration + + − + +++ ++ +++ +++ ++ + ++ Inflammatory infiltrate − ++ − ++ +++ +++ +++ +++ + ++ +++ Immunohistochemical staining ± − + + ++ ++ + +++ − − − Lung BALT necrosis − − − − − − − − + − − Inflammatory infiltrate − +++ − ++ +++ +++ +++ +++ + − +++ Immunohistochemical staining − − − − − ± − + − − − Liver Hepatocyte necrosis − − − − − +++ − +++ − − − Inflammatory infiltrate − ++ ++ ++ ++ ++ ++ +++ − ++ + Hemosiderosis − − − − + + ++ ++ ++ + + Immunohistochemical staining − − − − − − − + − − − Spleen
Lymphoid cell necrosis − − − − − − − − ++ − − Lymphoid cell depletion − − ++ + + + + + − − − Granulocytic infiltrate +++ +++ ++ ++ + +++ − − +++ ++ +++ Eosinophilic material deposits − − − − +++ +++ ++ ++ ++ − ++ Hemosiderosis − − − − +++ +++ + +++ ++ − + Immunohistochemical staining + + + ± − + + + − − −
118
†Tissue lesions were graded according to their distribution and severity: −, no lesion; +, focal and mild or moderate / multifocal and mild; ++, focal and marked / multifocal and moderate / diffuse and mild; +++, multifocal and marked / diffuse and moderate or marked.
♦WNV antigen immunostaining was graded according to its distribution and percentage of stained cells: −, negative staining; ±, focal single cells; +, focal or multifocal and < 20% cells stained; ++, multifocal or diffuse and 20-80% cells stained; +++, multifocal or diffuse and > 80% cells stained.
*NA: tissue sample no analyzed.
Table 3. Microscopic lesions and IHC staining gradation on different days post-inoculation (dpi) in tissues of experimentally WNV infected red-legged partridges (cont).
3 dpi 6 dpi 14 dpi
SP07 MO03 SP07 MO03 SP07 MO03
No. 1 No. 2 No. 3 No. 4 No. 5 No. 6 No. 7 No. 8 No. 9 No. 10 No. 11
Kidney
Tubular epithelial cell necrosis − − − − − − − − +++ +++ +++ Inflammatory infiltrate − ++ − ++ − − − + + − ++ Immunohistochemical staining − − + + + + + ++ − − − Duodenum Inflammatory infiltrate + + − − − + − − NA − + Immunohistochemical staining − − − − − − − + NA − − Large intestine Inflammatory infiltrate − + − ++ − − − − NA − ++ Immunohistochemical staining − − − − + + + + NA − − Pancreas
Acinar cell necrosis − − − − − − − + NA − − Immunohistochemical staining − − ± ± + + + + NA − −
Cecal tonsils
Lymphoid cell necrosis NA NA NA + NA ++ NA NA NA − NA Immunohistochemical staining NA NA NA + NA − NA NA NA − NA
Bursa of Fabricius
Lymphoid cell necrosis − − − − NA − − ++ NA NA − Lymphoid cell depletion − + − − NA ++ ++ ++ NA NA − Immunohistochemical staining − − − − NA − − − NA NA −
Skin + feather follicle
Inflammatory infiltrate skin − NA − +++ − NA NA − NA ++ − Inflammatory infiltrate feather
pulp +++ NA +++ NA NA NA NA +++ NA − +++ Immunohistochemical staining − NA − − − NA − − NA − −
119
Figure 1. Heart; partridge inoculated with SP07 No. 6, 6 dpi. Diffuse and marked necrosis and degeneration of cardiac myofibers, and infiltration of lymphoplasmacytic and histiocytic cells. HE, x100. Inset: detail of necrotic cardiac myofibers which show degeneration, fragmentation and accumulation of hyaline material in the cytoplasm. Infiltration of mononuclear inflammatory cells is also observed. HE, x400.
Figure 2. Spleen; partridge inoculated with MO03 No. 8, 6 dpi. Distension of the cytoplasm of splenic macrophages by a foamy material (black arrows) or a red-brown pigment (white arrows) that displace the nuclei. Mild infiltration of granulocytes (arrowheads). HE, x400.
Figure 3. Liver; partridge inoculated with SP07 No. 6, 6 dpi. Multifocal to coalescing marked necrosis of hepatocytes near the central vein. HE, x40. Inset: detail of the necrotic hepatocytes which
120
show picnosis and lysis of the nuclei and degeneration of the cytoplasm. Mild infiltration of lymphoplasmacytic and histiocytic cells is also observed. HE, x400.
Figure 4. Feather follicle; partridge inoculated with MO03 No. 11, 14 dpi. Diffuse and moderate infiltration of lymphoplasmacytic, histiocytic and granulocytic cells in the feather pulp. HE, x200.