Abstract
Although conventional immunohistochemistry for neurotropic rabies virus (RABV) usually shows high preference for neurons, non-neuronal cells are also potential targets, and abortive astrocyte infection is considered a main trigger of innate immunity in the CNS. While in vitro studies indicated differences between field and less virulent lab-adapted RABVs, a systematic, quantitative comparison of astrocyte tropism in vivo is lacking. Here, solvent-based tissue clearing was used to measure RABV cell tropism in infected brains. Immunofluorescence analysis of 1 mm-thick tissue slices enabled 3D-segmentation and quantification of astrocyte and neuron infection frequencies. Comparison of three highly virulent field virus clones from fox, dog, and raccoon with three lab-adapted strains revealed remarkable differences in the ability to infect astrocytes in vivo. While all viruses and infection routes led to neuron infection frequencies between 7-19%, striking differences appeared for astrocytes. Whereas astrocyte infection by field viruses was detected independent of the inoculation route (8-27%), only one lab-adapted strain infected astrocytes route-dependently [0% after intramuscular (i.m.) and 13% after intracerebral (i.c.) inoculation]. Two lab-adapted vaccine viruses lacked astrocyte infection altogether (0%, i.c. and i.m.). This suggests a model in which the ability to establish productive astrocyte infection in vivo functionally distinguishes field and attenuated lab RABV strains.