It is intriguing that rabies in humans and dogs can manifest in either furious or paralytic forms. The observation that the same rabid dog transmitted paralytic rabies to one human and furious to another human  suggests that host response may play a role in determining the clinical subtype. Previous neuropathological studies in rabies-infected humans have not been able to demonstrate differences between these two clinical subtypes . Immunohistochemistry has been employed in limited studies to detect rabies nucleocapsid antigen, and assessment done in a semi-quantitative fashion. In terms of RABV antigen abundance, the thalamus, basal ganglia, and brainstem have been shown to be preferentially involved over the cerebral hemispheres regardless of the clinical forms . A study in rabid dogs also showed greater RABV antigen burden in the brainstem compared to supratentorial structures . However, in this animal study, only the furious form was included and no spinal cord was available for analysis.
In the present study, semi-quantitative and quantitative assessment of RABV antigen was performed in both furious and paralytic forms at the early stage of the disease, and the spinal cord was studied for the first time in canine rabies. The viral antigen burden in both clinical forms was greatest in the spinal cord, followed by brainstem, and then the brain. This caudal to rostral trend of decreasing RABV antigen abundance was shown by semi-quantitative assessment, and quantitative assessment of all three parameters (% positive neurons, % antigen area in positive neuron, and % antigen area per neuron). Greater viral antigen amount was found in the furious form in most CNS regions including the spinal cord. At first, this latter finding might be viewed as paradoxical since one might expect the viral load in the spinal cord would be greater in the paralytic form. However, it is known that the clinical weakness in the paralytic form is due to peripheral nerve dysfunction, rather than the presence of RABV in the spinal cord . Of diagnostic interest, in one dog with the paralytic form, RABV antigen was found only in the spinal cord. Therefore, for making a diagnosis of rabies in a dog using an antigen detection method, a spinal cord specimen should be included, especially when the suspected animal is sacrificed shortly after the onset of illness or when not yet unconscious from the disease.
RABV RNA was distributed almost uniformly in all CNS regions by the early stage of the disease, regardless of the clinical subtype, reflecting the rapidly aggressive nature of rabies infection. The larger viral RNA load in all CNS sites in the furious form compared to paralytic form may explain the shorter survival period in the furious form and perhaps the more dominant cerebral symptoms seen in both human and canine rabies. Although it has been shown that RABV RNA is synthesized within a cage of nucleocapsid protein in the Negri body-like structures , the distribution of viral RNA in our study did not mirror that of rabies nucleocapsid protein. Whereas there were significant differences in the amount of RABV antigen at different levels of the CNS, this was generally not the case for viral RNA. While the brainstem, thalamus, caudate and hippocampus contained higher copy numbers of RABV RNA compared to other regions, this was not statistically significant. The cerebellum was the only site to have significantly less viral RNA compared to the brainstem and only in the furious form. Furthermore, the spinal cord, where RABV antigen was most abundantly found, contained relatively low amounts of viral RNA.
The discrepancy between RABV antigen and RNA amounts reflects the dynamic nature of viral transcription and replication and could be related to differences in production and degradation of RABV mRNA as distinct from viral antigen, and/or host factors controlling of transcription such as miRNAs. Different CNS regions may vary in resistance to viral propagation and cytolytic effects. In support of this hypothesis, rat spinal cord motor neurons are more resistant in vitro to RABV-induced cytolysis in a fixed virus model . Other host factors may also play important roles in determining the clinical subtype of rabies infection. A greater immune response occurs with paralytic rabies, as evidenced by cytokine mRNA transcripts in the brain of dogs .
A noteworthy finding in our study is that the greatest degree of inflammation was found in the brainstem of dogs with the paralytic form. Similar to other viral infections, T-cells predominated over B-cells in the inflammatory reaction. These findings correlated with the imaging findings showing more abnormal signals by MRI, indicating a greater degree of inflammation, in dogs with the paralytic form compared to dogs with the furious form, particularly at the brainstem level . Moreover, interleukin-1β and interferon-γ mRNAs were found exclusively in the paralytic form . Our findings also correlate with a mouse model study in which minocycline-treated mice showed reduced inflammation and increased disease severity when infected with SAD-D29, an attenuated strain of rabies . The heightened inflammatory response in the canine brainstem in paralytic rabies is in line with the impairment of neural tract integrity recently demonstrated by increased FLAIR (fluid attenuated inversion recovery) signal on MRI and decreased fractional anisotropy values on diffusion tensor imaging . Disruption of axoplasmic flow by inflammation in the brainstem could potentially retard viral propagation, particularly towards the cerebral hemispheres in the situation of paralytic rabies. This, in turn, could result in a diminished viral load in the brainstem and cerebrum. Why the maximum degree of inflammation was seen in the brainstem is not known. This may reflect the site of initial entry into the CNS by the RABV, since most dogs would be bitten in the neck region, and the virus would then migrate along the cranial nerves to enter the CNS at the level of the brainstem. However, we cannot say with certainty at which sites the dogs in our study were bitten, and thus this remains speculation. We also recognize the small number of animals available for study limits the strength of any conclusions we might reach. Carrying out larger studies in dogs in a controlled laboratory setting would overcome some of these limitations, and would also allow a comparison of clinical and pathologic findings between immune-competent and immune-deficient dogs. However, experimental studies for rabies are generally done in rodents rather than dogs, in part for ethical reasons. There also arise concerns over which virus to use in the laboratory setting. Wild-type virus is difficult to control with respect to the amount of virus to inoculate in order to create a reproducible incubation period, clinical picture, and mortality rate. Laboratory strains of RABV do not duplicate the natural infection. For example, elimination of RABV virulent strains is impeded by apoptosis of migratory inflammatory cells, whereas laboratory RABV strains produce neuronal apoptosis that is not seen with the wild-type virus [22, 26–28]. As well, other studies have shown laboratory strains are associated with activation of the host immune response, whereas there is very little response to wild-type virus . Therefore, there are distinct advantages to studying rabies in the natural setting.
Host factors may not be the only explanation for the two clinical subtypes of rabies. We were previously unable to detect any specific differences in the glycoprotein, phosphoprotein and nucleocapsid genes of RABV isolated from the furious and paralytic forms in humans and animals . The glycoprotein gene was found to have minor sequence variations, and cloned sequences of the viral population derived from a single rabies-infected dog showed minor substitutions at both nucleotide and amino acid levels . However, both reports analyzed viruses isolated only from one part of the brain. Hence, it remains possible that viral genetic polymorphisms might play a role in determining the clinical manifestations and influencing the host inflammatory and immune responses.