Street rabies virus causes dendritic injury and F-actin depolymerization in the hippocampus

Rabies is an acute viral infection of the central nervous system and is typically fatal in humans and animals; however, its pathogenesis remains poorly understood. In this study, the morphological changes of dendrites and dendritic spines in the CA1 region of the hippocampus were investigated in mice that were infected intracerebrally with an MRV strain of the street rabies virus. Haematoxylin and eosin and fluorescence staining analysis of brain sections from the infected mice showed very few morphological changes in the neuronal bodies and neuronal processes. However, we found a significant decrease in the number of dendritic spines. Primary neuronal cultures derived from the hippocampus of mice (embryonic day 16.5) that were infected with the virus also showed an obvious decrease in the number of dendritic spines. Furthermore, the decrease in the number of dendritic spines was related to the depolymerization of actin filaments (F-actin). We propose that the observed structural changes can partially explain the severe clinical disease that was found in experimental models of street rabies virus infections.


INTRODUCTION
Rabies virus (RV) is a neurotropic virus that primarily targets the central nervous system (CNS). The deadly RV produces a variety of nervous system symptoms; however, patients eventually die of circulatory insufficiency (Hemachudha et al., 2002). In contrast to the dramatic and severe clinical manifestations related to neuronal dysfunctions, only mild lesions in the CNS are observed during post-mortem examinations. Previous studies had demonstrated that the fetal rabies caused neuronal dysfunction, including ion channel dysfunction and neurotransmitter abnormalities rather than neuronal damage (Bouzamondo et al., 1993;Iwata et al., 1999;Tsiang, 1993), and that the infection with silver haired bat rabies virus resulted in the downregulation of several proteins that were relevant to synaptic physiology. Furthermore, the downregulation of these proteins can block synaptic vesicles from docking and fusing to the plasma membrane; therefore, the release and uptake of neurotransmitters is reduced (Dhingra et al., 2007).
Dendrites (the branched projections of a neuron) and dendritic spines constitute major post-synaptic sites for excitatory synaptic transmission (Harris, 1999;Hering & Sheng, 2001;Sheng, 2001). Dendritic spines are highly motile and can undergo remodelling; in addition, their structural plasticity is tightly coordinated with synaptic function (Kasai et al., 2010). Spine loss or alteration has been described in patients with neurodegenerative diseases, such as Alzheimer's disease (Fiala et al., 2002;Law et al., 2004). However, few studies have addressed if the neuronal dysfunction observed in the street rabies virus infection was associated with dendritic spine plasticity.
In this study, we used in vivo and in vitro methods to investigate the morphological changes of dendrites and dendritic spines in the hippocampus of mice that were infected by an MRV strain of the street rabies virus (GenBank accession no. DQ875050.1). The results showed that the RV infection decreased the number of dendritic spines. Further analyses confirmed that the dendritic changes were related to the depolymerization of filamentous actin (F-actin), a cytoskeleton protein that helps to regulate the morphogenesis and dynamics of dendritic spines (Fischer et al., 2000;Matus et al., 2000).

Distribution of MRV antigen in the CNS
The MRV antigen has been detected in numerous brain regions, such as the cortex, the thalamus and the cerebellum (data not shown). In the hippocampus, MRV expression began to appear in the CA1 region at day 4 post-infection (p.i.), and viral antigen was seen in the perikarya and the intact processes of hippocampal pyramidal neurons (Fig. 1a, b). By contrast, almost no positive staining could be found in the neuronal processes at day 7 p.i., and the antigen was located mainly in the perikarya (Fig. 1c, d). At day 7 p.i., the mice became moribund, and a greater number of variably sized RVpositive inclusion bodies were found to appear in the infected areas (the inset in Fig. 1d).

Morphological changes to cell bodies and dendrites in the CA1 region of the hippocampus
The changes that occurred to cell bodies and dendrites in the hippocampus after MRV infection were studied by both conventional histological staining and immunolabelling using microtubule-associated protein 2 (MAP2), a neuronal cytoskeleton protein. As shown in Fig. 1(e, f), haematoxylin and eosin (H&E) staining of paraffinembedded tissue sections did not demonstrate well-defined cytopathological changes in MRV-infected neurons. Similarly, fluorescence staining for MAP2 suggested that both the cell bodies and the dendrites remained intact in the infected area, even at the late infection stage (Fig. 2).
MRV infection decreased the number of dendritic spines in vivo. Alexa Fluor 488 phalloidin conjugate was used to show the dendritic spines in the CA1 region of the hippocampus. The punctate labelling of phalloidin, which is typical for the location of F-actin in dendritic spines (Capani et al., 2001), was observed in the PBS-injected mice ( Fig. 3a, b). By contrast, a significant decrease in fluorescence intensity and punctate labelling was observed in the MRV-infected mice ( Fig. 3c-e). In addition, many rope-like structures were observed to appear along the dendrites in the brains of the injected mice (arrows in Fig.  3d), indicative of a reorganization of F-actin occurring post-synaptically.
MRV infection decreased the number of dendritic spines in vitro. To confirm our in vivo results, cultures of primary hippocampal neurons were prepared. At 28 days after plating, nearly 95 % of the cells isolated from the hippocampus expressed the neuronal marker MAP2 as assessed by using immunofluorescence staining indicating that they were neurons (data not shown). These neurons also exhibited a well-developed network and a high level of growth, indicative of a healthy and viable culture.
Alexa Fluor 488 phalloidin staining revealed that most pyramidal hippocampal neurons showed mushroom-like structures 28 days after plating, suggesting the presence of well-defined dendritic spines. Infection with MRV resulted in a dramatic decrease in the number and size of spines, as revealed by phalloidin staining. When compared with those at day 4 p.i. (Fig. 4d-f), the infected neurons appeared seriously damaged at day 6 p.i. and continuous staining was not found in their cell bodies or dendrites for either F-actin or MAP2 (Fig. 4g-i). Due to the presence of positive punctuates in the dendritic cytoplasm, spines could not be unequivocally distinguished at this infection stage, therefore only the changes of spines occurring at day 4 p.i. were examined and quantified. A significant decrease in the number of spines was found in the infected cells (4±2.0 mm 21 , n512) compared with those in the control (12±3.0 mm 21 , n510; P,0.01, Student's t-test) (Fig. 4).

Depolymerization of F-actin by MRV
The intensity of F-actin and G-actin in mice infected by MRV was analysed by Western blotting and the ratio of Factin to G-actin was calculated. The intensity (P,0.01, n56-8 sections per mouse, n510 mice) and ratio (P,0.01, n56 mice per group, Student's t-test) were significantly decreased in the CA1 region of the infected mice when compared with that in the PBS controls ( Fig. 5a, b). It has been suggested that the polymerization or depolymerization of F-actin is regulated by the phosphorylation state of cofilin (Allison et al., 1998;Capani et al., 2001;Kuriu et al., 2006;Renner et al., 2009). Thus, we examined if the depolymerization of F-actin (caused by RV infection) was related to a decrease in the phosphorylated form of cofilin (phospho-cofilin, p-cofilin). Western blot analysis revealed that RV infection has no effect on the total levels of actin and cofilin; however, it can increase the levels of p-cofilin in the hippocampi of furious rabid mice (Fig. 5c, d) (P,0.01, n510 mice each group).

DISCUSSION
Human and animal rabies cases that are caused by a street rabies virus infection show neuronal dysfunction, such as abnormal neurotransmitter release and ion channel dysfunction, without any evidence of a significant degree of neuronal death. Stein et al. (2010) described the viral antigen distribution in the brainstems, cerebella, hippocampi and cerebra in 13 different species, and the hippocampus was suggested as the optimal site for RV detection in dogs and cats. Our study demonstrated that MRV was distributed in the cortices, the thalamuses, the cerebella, and particularly the hippocampi in mice at 7 days p.i.; in agreement with previous reports, no obvious damage to the neuronal soma was detected at this time point. Therefore, the neuronal processes were also examined. Scott et al. (2008) found relatively few changes in the perikarya and neuronal processes in the hippocampi of challenge virus (CVS)-infected transgenic mice expressing yellow fluorescent protein (YFP). Li et al. (2005) reported that the apical dendrites in the hippocampi became disorganized in the mice intracerebrally infected with the pathogenic CVS-N2C, a fixed RV strain. In contrast, Scott et al. (2008) found relatively few changes in the perikarya and neuronal processes in the hippocampi of CVS-infected YFP mice, and they attributed the discrepancy in these findings to the route of inoculation and the pathogenicity of the strain.
In the present study, no obvious morphological changes were seen in the processes or cell bodies of pyramidal neurons at day 7 p.i., supporting the idea that the fetal rabies virus causes neuronal dysfunction rather than neuronal damage. This is also consistent with the strain-dependent pathogenicity differences that were suggested by Scott et al. (2008), since we used a street rabies virus.
Dendritic spines are small membranous protrusions that typically receive inputs from the excitatory synapses of axons. Dendritic spines contain neurotransmitter receptors, organelles and signalling systems that are essential for synaptic function and plasticity, and numerous brain disorders are associated with abnormal dendritic spines (Hutsler & Zhang, 2010;Glantz & Lewis, 2000). Despite the presence of severe neuronal dysfunction in rabies cases, few studies have focused on the dendritic spine alterations caused by rabies. Studies have showed dendritic morphological alterations of cortical pyramidal neurons (including loss of dendritic spines) with RV infection (Torres-Fernández et al., 2007). Our results showed a significant decrease in the number of dendritic spines in certain brain regions of rabid mice in vivo and in vitro. This decrease may partially explain why RV infection causes neuronal dysfunction rather than neuronal damage.
The cytoskeleton plays an important role in intracellular transport, cellular division and cell shape. The cytoskeleton of a dendritic spine is primarily made of F-actin. Dynamic changes in F-actin have been found in herpes simplex virus 1-infected human neuroblastoma cells (Xiang et al., 2012). In the case of RV, alterations in the actin-based cytoskeleton have been described in the neuroblastoma cells by Ceccaldi et al. (1997), who found that the nucleocapsid of the virus had no direct action on the kinetics of actin polymerization, but could inhibit the actin-binding effect induced by dephosphorylated synapsin I (Bloom et al., 2003;Ceccaldi et al., 1997;Valtorta et al., 1992). Synapsin I is known to interact with the actin-based cytoskeleton to induce the release of neurotransmitters and the recycling of synaptic vesicles. Thus, we sought to determine if the decrease in the number of dendritic spines was associated with changes in F-actin. We analysed the levels of F-actin in the stratum radiatum of the CA1 region and found that an RV infection causes the depolymerization of F-actin. A number of proteins that regulate the actin cytoskeleton have been identified, including cofilin (also named actin depolymerization factor) (Hotulainen et al., 2005;Kiuchi et al., 2007), profilin, gelsolin, drebrin and CaMKII (Ackermann & Matus, 2003;Bamburg, 1999;Ivanov et al., 2009;Lisman et al., 2002;Nag et al., 2009). Cofilin can depolymerize F-actin through binding at the interface between the actin monomers. Cofilin is inactivated by phosphorylation; therefore, the dephosphorylation of cofilin at the Ser3 residue leads to cofilin activation and the subsequent depolymerization of F-actin. However, our study showed an increase in phosphorylated cofilin, which may be attributed to the compensatory mechanisms associated with F-actin dynamics (Shi & Ethell, 2006).

In our study, we have determined that an RV infection can decrease the number of dendritic spines in vivo and in vitro.
We also showed that this decrease may be partially caused by F-actin depolymerization. Future work will analyse the functional significance of these cytoskeletal and synaptic changes.  (Laothamatas et al., 2008), we referred to mice exhibiting furious rabies as furious rabid mice.  Co-culture of primary hippocampal and cortical neurons. As previously described, the primary neuronal cultures were derived from the cerebral cortices and the hippocampi of mice on embryonic day 16.5 (E16.5) (Fath et al., 2009). The cortices and hippocampi were incubated with 0.25 % trypsin (Invitrogen), and 50 ml of 10 mg ml 21 DNase I stock (Invitrogen) was added for 30 s to break down the DNA and avoid tissue clumping. The cells were then titrated in glass pipettes. The cell suspension was adjusted to the appropriate concentration to obtain 1610 3 living cells from the hippocampus and 1610 5 living cells from the cortex per 75 ml Dulbecco's modified Eagle's medium (DMEM; Invitrogen) supplemented with 10 % FBS (Invitrogen). The dissociated neurons were initially plated in 10 % FBS/DMEM on poly-D-lysine-treated (0.1 mg ml 21 ; Sigma) coverslips with wells. A total of 75 ml of the hippocampal cell suspension was directly plated onto a coverslip, and 75 ml of the cortical cell suspension was slowly plated in a ring surrounding the hippocampal cells. After 2 h, the plating medium was carefully replaced with equilibrated neurobasal media containing B27 supplement (Gibco) and 2 mM L-glutamine (Invitrogen). The cells grew for up to 4 weeks without any further change of medium. The cells were infected with MRV for 120 h at 10 TCID 50 per well and then fixed with 4 % paraformaldehyde (Invitrogen) and mounted for subsequent fluorescence staining.

METHODS
Immunohistochemistry for RVBV antigen. As previously described , the MRV-infected mice (n515) and the control mice (n510) were anaesthetized with 50 mg kg 21 pentobarbital (Invitrogen) and then intracardially perfused with 50 ml PBS followed by 50 ml of freshly prepared 4 % paraformaldehyde in 0.1 M phosphate buffer (pH 7.4; Invitrogen) at day 4 p.i. and day 7 p.i. The brain was dissected and post-fixed in the same fixative for 2-6 h at 4 uC. The brains were then washed in a series of cold sucrose solutions of increasing concentration. The samples were embedded in an OCT (optimal cutting temperature) compound (Tissue-Tek; Sakura Finetek Japan), frozen on dry ice and cut into transverse sections at a thickness of 40 mm using a freezing microtome (Leica).
Immunostaining was subsequently performed on these free-floating sections. The cryosections were incubated with 0.3 % H 2 O 2 and blocked with 3 % normal goat serum (Invitrogen) in 0.1 M phosphate buffer for 30 min at room temperature; the cryosections were then incubated with a rabbit polyclonal anti-RV antibody (1 : 200 dilution; made and stored in our laboratory) overnight at 4 uC. Next, the sections were incubated with HRP-labelled anti-rabbit IgG (Sigma) for 1 h at room temperature and washed and developed with 3,39diaminobenzidine (DAB; Sigma). The stained sections were examined using an Olympus IX 51 microscope.
Histopathology. The perfused mice brains were dehydrated through a graded series of ethanol and embedded in paraffin. Paraffin sections of 4 mm thickness were prepared and stained with haematoxylin and eosin (H&E; Sigma).
Immunofluorescence. The animals and cultured cells were fixed or sectioned, as described above. After blocking with a normal 3 % goat serum in 0. Western blot analysis. To analyse the cofilin and actin changes that were caused by the MRV infection, brains were collected from the rabid and the control mice. The hippocampi were then individually dissected and homogenized in an SDS-PAGE sample buffer containing 3 % SDS, 2 % b-mercaptoethanol and 5 % glycerol in 60 mM Tris buffer (pH 6.7), as previously described (Ouyang et al., 2005(Ouyang et al., , 2007. The homogenized samples were boiled for 5 min and stored at 220 uC. The protein concentration was determined by the Lowry method. Thirty micrograms of protein were separated by 15 % SDS-PAGE and transferred to polyvinylidene difluoride membranes. After incubation with a primary antibody (1 : 1000; Cell Signalling) that recognized phosphorylated cofilin (p-cofilin) at Ser3, the membranes were incubated with a peroxidase-conjugated secondary antibody and visualized with an ECL detection kit (Pierce). The blots were reprobed for total cofilin (1 : 1000; Cytoskeleton) and actin (1 : 500; Sigma). The signals were scanned for quantitative analysis with ImageJ.
The ratio of F-actin to G-actin was measured by previously published Western blotting techniques (Gu et al., 2006). Briefly, the hippocampi from the control and rabid mice were isolated and homogenized in a cold lysis buffer (10 mM K 2 HPO 4 , 100 mM NaF, 50 mM KCl, 2 mM MgCl 2 , 1 mM EGTA, 0.2 mM dithiothreitol, 0.5 % Triton X-100, 1 M sucrose, pH 7.0) and then centrifuged at 15 000 g for 30 min. The supernatants were used for measuring soluble actin (G-actin). For Factin, the pellets were resuspended in a lysis buffer and an equal volume of 1.5 M guanidine hydrochloride, 1 M sodium acetate, 1 mM CaCl 2 , 1 mM ATP and 20 mM Tris/HCl, pH 7.5. The pellets were then incubated on ice for 1 h with gentle mixing every 15 min to depolymerize the F-actin. The samples were centrifuged at 15 000 g for 30 min, and the supernatants were used to measure the insoluble Factin. Samples from the supernatant (G-actin) and pellet (F-actin) fractions were proportionally loaded and analysed by Western blotting.
Quantitative analysis. To analyse the F-actin levels in the control and MRV-infected groups, all of the sections were scanned with an Olympus FV1000 laser scanning confocal microscope under the same parameters. The fluorescence intensity of F-actin in cryosections was calculated by an experimenter who was blinded to the treatment conditions. For the in vitro experiment, 10-15 neurons were randomly selected, and the mean number of dendritic spines per 10 mm of dendrite was calculated. A Student's t-test was used to determine the statistically significant differences in the dendritic spine number, the F-actin intensity, and the quantified protein expression between the control and the infected groups. All of the data were expressed as the means±SD. Statistical significance was defined as P,0.05.