Molecular and biological determinants of pathogenicity of Africa lyssaviruses

Abstract

The Lyssavirus genus currently consists of eleven species of which four has been isolated in Africa [(Rabies virus (RABV), Lagos bat virus (LBV), Mokola virus (MOKV) and Duvenhage virus (DUVV)]. RABV occurs worldwide and is also widespread throughout Africa. In southern Africa, two distinct variants of RABV are adapted to canines (canid variant) and to herpestids (mongoose variant). LBV, MOKV and DUVV appear to be exclusive to Africa. Studies on the pathogenesis of African lyssaviruses are limited, but one study reported that LBV and MOKV are not lethal via the intramuscular (i.m.) route of inoculation in murine models. A more recent study contradicts this in that several LBV isolates and one isolate of MOKV were found to be lethal to mice when administered i.m. It should be noted that such pathogenesis studies are of limited value, due to some differences in models and the ranges of isolates studied. Since limited sequence information has been available for the African lyssaviruses, very few studies included a comparison of pathogenic domains. This study aimed at comparing a number of pathogenic domains on representative isolates of African lyssavirus genomes and link that to differences in pathogenicity profiles observed when these lyssaviruses were inoculated in a mice.

Mice were inoculated i.m. with RABV (canid and mongoose variant), LBV, MOKV and DUVV to compare their relative pathogenicity. In these experiments, all the viruses from all four viral species were found to be lethal to mice, with a single exception, viz. the isolate LBVNig1956. Generally, LBV, MOKV and DUVV isolates had high percentage mortality (50 to 75%) compared to RABV (canid variant) (25%) when inoculated at the same dose (titre). Different isolates of LBV, in particular, exhibited different pathogenicity profiles and we concluded that the pathogenicity of different isolates from all of the viral species can be very variable. Therefore, the pathogenicity of an isolate in a mouse model is not necessarily a trait of the species, but of the specific isolate.

There was detectable virus neutralizing antibodies (VNA) as early as day 7 in all mice. These titers increased on day 14 (above 300) and remained high on days 50 and 60 (above 200). No significant difference was observed in the level of VNA between mice that died of rabies and those that survived infection. This indicated that the presence of VNA in serum may not be the only requirement for clearance of lyssavirus infections, contrary to some reports. Lyssaviruses have been reported to spread to the salivary glands after replication in the central nervous system. In this study, no viral RNA was detected in the salivary glands of mice that succumbed to rabies. This may indicate that the pathogenesis of lyssaviruses in mice is not the true reflection of their pathogenesis in natural or reservoir hosts.

A number of pathogenic domains on the phosphoprotein, matrix and the glycoprotein were compared between representative isolates of the African lyssaviruses. Substitution of domains such as Arg 333 on the glycoprotein was previously shown to be important in pathogenicity of some lyssaviruses. Generally our findings suggested that domains important in pathogenicity of a particular lyssavirus isolate or strain may not necessarily be of the same importance in the pathogenicity of another isolate. More pathogenesis studies on African lyssaviruses are crucial for a better understanding of the factors that determine their pathogenicity.