Are We Getting Closer to the Treatment of Rabies?

Rodney E Willoughby Jr


Future Virology. 2009;4(6):563-570. 

In This Article

Unique Immunology

Rabies evades host immunity by restricting its tropism to neuronal cells and a few other specialized cell types. Most neurons reside behind the blood–brain barrier, while peripheral nerves are at least partially isolated by perineurium and myelin sheaths. The virus replicates at low levels, such that surface expression of the rabies glycoprotein is very low.[3] The rabies virus phosphoprotein downmodulates the type I interferon signaling pathway by interfering with STAT1 phosphorylation and localization.[15]

Neurons do not usually express MHC class I molecules. This means that the classical surveillance mechanism for intracellular pathogens is evaded.[16] Underexpression of MHC class I molecules is teleologically plausible for terminally differentiated cells such as neurons. For the optimal evolutionary survival of the host after encephalitis, the clearance of virus infection from neurons must be noncytolytic, without apoptosis or necrosis. Conventional clearance through antibody-dependent cell-mediated cytotoxicity or cytolytic T cells would result in permanent brain damage or death. The mechanism for noncytolytic clearance has not been well elucidated, but appears similar in rabies virus infections and the better-studied Sindbis virus infection in mice.[16,17] Antibody production within the brain, requiring local B cells and CD4+ T cells, is essential for protection against and clearance of the rabies virus; cells must cross the blood–brain barrier.[18] A subset of neutralizing monoclonal antibodies (MAbs), reactive with surface-expressed viral antigens, are internalized into the cytoplasm and inhibit virus transcription and translation.[17] This noncytolytic mechanism for clearing viruses is probably more common than generally appreciated, because most infectious encephalitis is resolved without permanent sequelae to the host.

A second clearance mechanism in the Sindbis model requires cytokines (notably IFN-γ) to clear viruses from brain astrocytes. While astrocytes are not generally infected by rabies virus, recent metabolomic studies of cerebrospinal fluid (CSF) from humans infected by rabies show very high levels of quinolinic acid (QA), a tryptophan degradation pathway under the control of IFN-γ.[19] IFN-γ has been largely neglected in rabies research, but is critical to the humoral immune response that clears rabies virus from the mouse brain.[18] The inferred 'upstream' presence of high levels of IFN-γ in human rabies suggests that both components (antibodies and type II interferons) required for noncytolytic clearance of neurotropic Sindbis alphavirus from the brain are also present in human rabies encephalitis.

The low immunological profile of wild-type rabies viruses contrasts with many fixed rabies viruses that dominate the virological literature. Fixed viruses have been passaged since the late 19th century and have very short consistent, incubation periods in animal models. In contrast to the wild-type viruses, many fixed rabies viruses are highly inflammatory and proapoptotic. The rapid immune response to these strains often limits extension of peripherally inoculated rabies virus to the level of the spinal cord – causing permanent poliomyelitis-like paresis – while generally sparing the brain.[4] 'Attenuation', whether low virulence, proinflammatory or both, maps genetically to the surface membrane-expressed rabies G glycoprotein. The G glycoprotein is the major attachment protein for rabies virus, containing the four major antigenic sites, including conformation-dependent and the linear epitopes. Two epitopes suffice for universal protection by MAbs.[20] Fixed strains of rabies virus replicate at much higher levels and bud promiscuously from infected cells. The overexpression of the G glycoprotein increases the induction of apoptosis in cell lines and immunogenicity in animals.[4] These laboratory-adapted strains are ideal for engineering recombinant viruses that are capable of providing protection against the challenge virus even after single vaccination schedules.[21] Laboratory-adapted strains are not useful for understanding neural dysfunction or the immunology of wild-type rabies viruses.

The inflammatory properties of fixed strains may explain the notable phenomenon of early death in rabies. Early death is induced experimentally by vaccination or transfer of effector cells or antibodies during symptomatic disease, and leads to accentuated symptoms and earlier demise. Based on case reports and very limited numbers from the Milwaukee Protocol (MP) registry, this phenomenon may occur in bat-associated rabies in humans.[22] The pathophysiology of early death is unknown.[23] It is thought that by intensifying the immunity of the rabies virus at time of active disease – by systemic immunization with killed, genetically inflammatory, fixed vaccine strains – the immune response is diverted away from noncytopathic, antibody-mediated clearance and toward broader recruitment of effector cells and antibody-dependent cytotoxicity. Enhancement of wild-type disease after formalin-fixed measles and respiratory syncytial virus vaccines in children is well known.[24] Administration of passive antibody, raised against such phenotypically altered strains, also appears deleterious.[23,25] The systemic immune response, extended to the brain, may result in neuronal loss or demyelination.