Evaluation of Viral Migration of Different Variants of Equid Alphaherpesvirus 1 in the Central Nervous System of Hamsters by Immunohistochemistry

The intranasal inoculation of equid alphaherpesvirus1 (EHV-1) Brazilian variants A4/72, A9/92, A3/97, Iso72/10 and the Argentine variant AR4 in a Syrian hamster model Mesocricetus auratus induced severe encephalitis. Clinical signs included weight loss, lethargy, somnolence, anorexia, and intense salivation two days post-inoculation (dpi), followed by neurological signs such as loss of proprioception, walking in circles, spastic paralysis, seizures, recumbency and death at 3 rd dpi (A9/92 and A4/72 variants) and 4 th dpi. Respiratory signs such as dyspnea and serosanguinous nasal discharge were also observed. Histopathological changes in brain included mixed inflammatory infiltrate with predominance of mononuclear cells, neuronal degeneration, liquefactive necrosis, hemorrhagic foci, leptomeningitis, perivascular edema, mononuclear infiltration, and perivascular cuffing. Immunohistochemical examination showed viral replication in neurons restrict predominantly to olfactory bulb and frontal cortex (variants AR4 and A3/97) and in groups of cells from distant regions, such as the caudal diencephalon and rostral mesencephalon (variants Iso72/10) and absence of viral antigen labeling of variants A9/92 and A4/72 despite these variants were the most neurovirulent, so new experiments not staggered in days but in hours post inoculation are needed to better understand the viral migration of these variants.

Studies of EHV-1 pathogenicity have usually employed murine and equine models [13][14][15][16][17][18]. However, studies involving adult horses are costly, labor intensive, and limited to a small number of animals [19]. Besides this experimentally inoculated horses may have antibodies to EHV-1 (immune memory) that can neutralize virus replication and, consequently, influence the type and magnitude of lesions produced and serological cross-reaction between EHV-1 and EHV-4 may interfere in the results [17].
Experiments using suckling Syrian hamsters have been conducted for virus isolation and studies of the pathogenesis of EHV-1 infection since the 1950s, when the detection of intranuclear inclusion bodies in the hepatic cells of hamsters inoculated intraperitoneally became possible [20]. Equid alphaherpesvirus 1 was first isolated in Brazil in the 1960s from a sample of aborted equine fetus liver inoculated intraperitoneally into suckling hamsters [1].
The first neurological signs of EHV-1 infection in suckling hamsters in Brazil were reported after intraperitoneal inoculation with a suspension of liver, spleen, and lung from an aborted equine fetus from Campinas, São Paulo [21]. The isolate, designated A4/72, was shown to cause tonic-clonic convulsions and paralysis in the infected hamsters. After the 1980s, isolation and study of EHV-1 in Brazil was accomplished predominantly via VERO cell culture and in equine dermal cells [22][23][24].
In subsequent years, the model of EHV-1 infection in hamster was used for the study of antiviral agents [25] and immune response to the virus [26,27]. Recently, hamster models have been used to study the neuropathogenesis of EHV-9 [28][29][30][31], but there are few studies about neuropathogenicity of EHV-1 in hamster models [32].
With advances in molecular biology, hypotheses of the neuropathogenicity of EHV-1 have been proposed. Gene mutations of the open reading frame (ORF) 64, which encodes infected cell protein 4, could be linked to neuropathogenicity [33].
A mutation exchanging adenine for guanine at position 2254 in ORF 30 (A 2254 -G 2254 ) that leads to an alteration in the aminoacid 752 sequence (from N 752 -D 752 ) is also linked to the neuropathogenicity of EHV-1, but only the amino acid variation D 752 is linked to inflammation of the central nervous system and ataxia, while the N 752 amino acid mutation is not shown to cause neurological signs in the natural host [34,35].
Recent studies have shown that EHV-1 with ORF30 mutation (N 752 -D 752 ) did not lead to altered virus replication in FHK and mouse neuron cultured cells and residue 752 in the essential DNA Polymerase of EHV-1 was not required for virus growth [36]. The ORF37 (UL24) deletion has also been mentioned as a factor of lost of neuropathogenicity in a mouse model [37].
Despite the presented studies there is currently no consensus with respect to alterations that may make an EHV-1 variant neuropathogenic, so the aim of the present study was to characterize tissue damage and viral migration throughout the Central Nervous System of different variants of EHV-1 using a hamster model.
The A9/92 (6 th passage) and A4/72 (21 th passage) viruses were propagated in VERO cells, A3/97 (5 th passage) and Iso72/10 (1 st passage) in ED cells and AR4 (2 nd passage) in RK-13 cells. Cell lines were grown in Eagle's Minimum Essential Medium (EMEM) supplemented with 10% fetal calf serum. After three cycles of freezing and thawing, the cells and infectious supernatant were centrifuged at 8000×g for 20 min to remove cell debris. The final infectious supernatant was fractionated into small volumes and stored at -70°C until use [8].

Experimental Animals
Twenty-four Syrian hamsters Mesocricetus auratus were obtained from the animal facility of the Department of Pathology, School of Veterinary Medicine and Animal Science, University of SãoPaulo, SãoPaulo, Brazil. Hamsters were separated into six groups, five experimental and an untreated control, with groups housed separately in polycarbonate micro-isolator cages, each connected to an individual ventilation system with HEPA filters on both the in-and out-flow. Filtered and autoclaved water and commercial pellets were provided ad libitum. All procedures were approved by the Ethics Committee on Animal Experiments of the Biological Institute (CETEA-IB) on registration protocol number CETEA-IB 106/10.
After inoculation, the infected hamsters were weighed and evaluated twice daily to register of clinical signs that were recorded as absent, slight, moderate or severe by the same operator of the experiment. At the onset of severe clinical signs they were euthanized along with the control group by an overdose of Sevoflurane and necropsied to collect samples of brain, lung, spleen, liver, thymus, heart, and kidney [18].

Histopathology
The collected organs were fixed for 48h in 10% buffered formalin then transferred to 70% ethanol, dehydrated in ethanol series, cleared in xylene, and embedded in paraffin. Sections were cut at 3µm, deparaffinized, rehydrated, and stained with H&E [11]. The histological score was classified in absent, slight, moderate or severe according to the intensity of microscopical lesions by the same pathologist of the experiment.

Immunohistochemistry
Sections of organs were dewaxed in xylol at 37°C for 30min, and endogenous peroxidase was blocked in a solution of 20ml of 30% hydrogen peroxide (H 2 O 2 ) diluted in 80ml of methanol for 30min. Antigen retrieval was performed with citrate buffer (pH 6.0) by heating in a microwave oven (1000W for 15min) prior to incubation with the primary antibody in a humid chamber at 37°C for 30min with a 1:1000 dilution of goat antiserum specific for ERV/EHV-1 (VMRD ® USA Inc. catalog 210-70-ERV, Pullman, Washington, USA). Finally, they were incubated with streptavidin-biotin complex (a combination of anti-mouse, -rabbit, and -goat immunoglobulins; LSAB+System-HRP, Dako ® ref. K0690-Dako ® Cytomation, Carpinteria, California, USA) for 45min in a humid chamber at room temperature. The reaction was revealed upon exposure to 3,3' diaminobenzidine chromogen solution (DAB, Dako ® code ref K3468). The slides were counterstained with Mayer's hematoxylin and mounted with synthetic resin (Entellan-Merck ® ) [11]. The presence of viral antigen was classified in: -absent when no antigen was stained; + few when found less than 5 stained cells in 40 x field; ++ moderate when found 5-30 stained cells in 40 x field or +++ numerous when found more than 30 stained cells in 40 x field according to the evaluation of the same pathologist of the experiment.

Clinical Signs
The hamsters inoculated with the A/92 and A4/72 variants presented high weight loss at 2 days post-inoculation (dpi), with severe disease at 3 dpi characterized by intense salivation and neurological signs such as loss of proprioception, walking in circles, spastic paralysis, seizures, and recumbency, followed by death. Those inoculated with the Iso72/10 variant showed similar neurological signs, but the acute phase started at 4 dpi and included respiratory signs such as dyspnea and serosanguinous nasal discharge.
Animals inoculated with A3/97 and AR4 variants showed weight loss, lethargy, somnolence, and respiratory signs, predominately dyspnea and serosanguinous nasal discharge. At 4 dpi, the hamsters exhibited neurological signs such as hyperexcitability in response to external stimuli, tremors, and seizures, but effects were less severe than observed with the other studied variants. The control group showed normal weight gain and did not exhibit behavior, respiratory, or neurological abnormalities ( Figure 1; Table 1).   The lungs were edematous and did not collapse upon opening of the thoracic cavity. They presented diffuse reddish coloration in the pulmonary lobes corresponding to hemorrhagic areas. The histological examination showed intense and diffuse mixed inflammatory infiltrate with predominance of mononuclear cells, areas of moderate to intense congestion, hemorrhagic foci, thickened alveolar walls, moderate to intense foamy macrophage infiltration in the alveolar lumen, presence of free red blood cells and alveolar edema, intact bronchiolar epithelium, and slight to moderate mucous secretion. The pleura showed discrete swelling and moderate emphysema.
Cellular debris and free red blood cells in the bronchiolar lumen were observed in hamsters inoculated with the A9/92 and AR4 variants. In hamsters inoculated with Iso72/10 and AR4 variants, syncytia and intranuclear inclusion bodies were observed in the parenchyma. No alterations were found in spleen, liver, thymus, kidney, or heart of inoculated hamsters. No macroscopic or histological changes were observed in any organ of the control group.

Immunohistochemistry
Positive immunohistochemical labeling was found in a group of neurons from a necrotic foci in the frontal pole region of hamsters inoculated with A4/72. The hamsters inoculated with Iso72/10 presented immunolabeling in neurons of the olfactory bulb and in groups of cells of the caudal diencephalon and rostral mesencephalon (Figure 3a, b).
The hamsters inoculated with A3/97 and AR4 variants showed positive immunolabeling only in the olfactory bulb region (internal granule cells) and in groups of neurons of the frontal pole cortex (Figure 3 c, d). The meninges and other areas of the CNS such as striated septum, pons, and medulla oblongata showed no positive labeling with any variant (Table  3). No immunoreactivity was observed in the organs of hamsters inoculated with A9/92 variant and from negative control group.

Discussion
The neurological signs observed in hamsters inoculated with the EHV-1 variants tested were similar to those described after natural infection in the host equine species: lethargy, anorexia, and ataxia reaching peak intensity at 2-3 dpi [41][42][43]. Meningeal congestion was the only macroscopic alteration in brain of the inoculated hamsters, this alteration has also been observed in mice and horses as a result of EHM, associated with congestion and hemorrhagic foci in spinal cord [7,8,18].
The microscopic lesions were similar to those reported in hamsters infected with EHV-9 and in mice and hamsters inoculated with EHV-1, with the olfactory bulb and frontal cortex being the primary affected regions after intranasal inoculation [8,16,28,29,31,32]. In hamster and murine models, the intranasally inoculated virus migrates from the nasal mucosa by neural dissemination via the olfactory neuroepithelium throughout the olfactory bulb, olfactory nerve, and ventricular surface, leading to neuronal degeneration, mainly in cortical areas and the hippocampus, with associated generalized ventriculitis [16,18,28,30].
As reported by other authors, we also found variation in severity of clinical signs, according the pathogenicity of the virus variant [15,30]. Variants A9/92 and A4/72 showed high neurovirulence, compared to the other tested variants, with results similar to those reported in mice after intranasal inoculation [18]; the Iso72/10 variant showed lighter neurological signs with acute phase starting only at 4 dpi and presented also respiratory signs while A3/97 and AR4 variants caused predominantly respiratory signs [32]. In a study using several line ages of mice, the A3/97 variant caused no weight loss or apparent clinical signs of disease, although it was possible to recover the virus from lung of all infected mice and from brain of a single BALBc nude mouse [18].
In adult horses, EHM induces microvascular damage to the central nervous system via initiation of an inflammatory cascade followed by degenerative thrombotic vasculitis of small blood vessels in the brain or spinal cord with resultant ischemic neuronal necrosis, perivascular mononuclear cuffing, congestion, hemorrhage, diffuse gliosis, perineuronal and perivascular edema [7,12,44]. In hamsters is suggested that these lesions may be the result of the viral replication within neurons, as opposed to vascular changes [32].
Effects of the less pathogenic EHV-1 variants have been reported to remain confined to the olfactory bulb region, while the more virulent variants migrates via neurons throughout the brain by anterograde axonal transport and slow axonal flow, infecting neurons and causing severe tissue damage throughout the CNS [16]. In contrast, in the present study, one less neuropathogenic variant (Iso72/10) were immunohistochemically detected even in distant regions from the olfactory bulb such as caudal diencephalon and rostral mesencephalon, while A4/72 one of the more virulent variant were detected only in a group of cells of necrotic area from the frontal pole region, so further studies are needed to better understand EHV-1 migration.
The route of inoculation seems to be an important factor in the development and progress of experimental herpes infection and migration. In a study of EHV-9, clinical signs of encephalitis including salivation, tremors, uncoordinated movement, convulsions, and severe moribund states were present in 100% of hamsters intranasally inoculated but in only 25% of those inoculated by oral and peritoneal routes [30]. On the other hand, noneurological signs were observed in hamsters inoculated via intravenous or ocular routs [30].
The time course of infection in hamsters seems also to be important in the migration of the herpes virus throughout the CNS. EHV-9 viral antigens were detected in the olfactory epithelium, nerve and bulb in the first 48 hours post inoculation, while in the cerebral cortex and areas such as those connecting the trigeminal sensory nerve root to the brain stem, pons and medulla oblongata, positive detection was found at 60h PI [29]. Recent studies in mice have shown the EHV-1 and 9 migration from the olfactory epithelium to the olfactory bulb on day 3 post inoculation, but without viral migration to other areas of the CNS, probably due to the higher susceptibility to infection of the hamster in comparison to mice [16,45].
The dose of the inoculums and the type of anesthetic used must also be considered in pathogenicity and virulence studies. Intranasal inoculation of hamsters with EHV-9 at10 1 PFUs was reported to not cause death, whereas with 10 4 PFUs, hamsters became severely ill at 3 dpi and died at 4 dpi [28]. In a study with mice intranasally inoculated with EHV-1 (AR4 variant) it was observed that clinical signs were more severe in animals anesthetized with ether; ketamine/xylazineanesthetized animals presented the highest cell death rates, while those anesthetized with isoflurane showed the highest proliferation rates, despite this isoflurane was considered a good anesthetic for experimental assays, since it produced few known side effects, animal suffer and showed a better post anesthesia recovery [46].
Studies have suggested that genotype alterations may be responsible for EHV-1 neuropathogenicity. EHV-1 variants with the exchange of adenine for guanine at position 2254 in ORF30 (A 2254 -G 2254 ), leading to alteration in the aminoacid 752 sequence (from N 752 -D 752 ), have been associated with the potential to cause neurological disease. This genotype alteration was associated with 30 of 32 outbreaks of neurological disease for EHV-1 in 2001-2006 in United Kingdom and United States [34,47].
Prevalence of EHV-1 variants considered neuropathogenic (G 2254 /D 752 ) increased from 3.6% in the 1960s to 13.3% in the 1990s in central Kentucky's thoroughbred broodmare population [19]. This number increased to 19.4% in 2000-2006, suggesting that viruses of the neuropathogenic genotype were increasing in prevalence in the latent reservoir, leading to greater risk of outbreak of equine herpesvirus neurological disease [19].
The amount of EHV-1 nasal shedding significantly increased in animals infected with the D 752 variant compared with N 752 variant (p=0.001); the D 752 variant led, on average, to a four-fold higher amount of nasal EHV-1 shedding compared to infections with the N 752 variant, suggesting that neuropathogenic variants could have a selective advantage and are systematically increasing in prevalence in domestic horse populations [48].
A experiment comparing the partial nucleotide sequences of ORF72 (glycoprotein D-gD), ORF64 (ICP4), and ORF30 (DNA polymerase) genes to corresponding sequences of EHV-1, found that the reference variant showed no molecular variation in the ICP4, gD, or viral DNA polymerase gene regions from the evaluated variants, in the experimental the EHV-1 Brazilian variants analyzed, including A3/97, A4/72, and A9/92, were classified as non-neuropathogenic variants (N 752 ) based on the ORF30 analysis, suggesting that other factors, such as the immuneresponse of the host species, could be involved in EHV-1 neuropathogenicity [49].
Recent studies have shown neuropathogenic variants 01c1 to exhibit growth kinetics similar to those of non-neuropathogenic variants in fetal horse kidney and cultured neurons and suggest that the D/N 752 dissimilarity in ORF30 may not be related to replication ability in fetal horse kidney and neural cell lines [36].
In a study to evaluate the ability of neuropathogenic (EHV-1) and non-neuropathogenic (Jan-E and Rac-H) EHV-1 variants to infect neuronal cells, real-time PCR analysis demonstrated that both neuropathogenic and non-neuropathogenic EHV-1 variants replicated in primary murine neurons and ED cells at a similar level [50].

Conclusions
The five EHV-1 variants tested induced severe encephalitis in the hamster model; viral replication in neurons restrict predominantly to olfactory bulb and frontal cortex were observed in variants AR4 and A3/97, more related with respiratory clinical signs than neurological; while the variant Iso72/10, related to both neurological and respiratory signs presented migration in groups of cells of caudal diencephalon and rostral mesencephalon and the most neurovirulent strains A9/92 and A4/72 presented no labeling, suggesting that further studies, not staggered in days but in hours post inoculation, are needed to better understand the viral migration of these neurovirulent variants.