Comparative pathogenesis of different phylogroup I bat lyssaviruses in a standardized mouse model

Abstract

A plethora of bat-associated lyssaviruses potentially capable of causing the fatal disease rabies are known today. Transmitted via infectious saliva, occasionally-reported spillover infections from bats to other mammals demonstrate the permeability of the species-barrier and highlight the zoonotic potential of bat-related lyssaviruses. However, it is still unknown whether and, if so, to what extent, viruses from different lyssavirus species vary in their pathogenic potential. In order to characterize and systematically compare a broader group of lyssavirus isolates for their viral replication kinetics, pathogenicity, and virus release through saliva-associated virus shedding, we used a mouse infection model comprising a low (102 TCID50) and a high (105 TCID50) inoculation dose as well as three different inoculation routes (intramuscular, intranasal, intracranial). Clinical signs, incubation periods, and survival were investigated. Based on the latter two parameters, a novel pathogenicity matrix was introduced to classify lyssavirus isolates. Using a total of 13 isolates from ten different virus species, this pathogenicity index varied within and between virus species. Interestingly, Irkut virus (IRKV) and Bokeloh bat lyssavirus (BBLV) obtained higher pathogenicity scores (1.14 for IRKV and 1.06 for BBLV) compared to rabies virus (RABV) isolates ranging between 0.19 and 0.85. Also, clinical signs differed significantly between RABV and other bat lyssaviruses. Altogether, our findings suggest a high diversity among lyssavirus isolates concerning survival, incubation period, and clinical signs. Virus shedding significantly differed between RABVs and other lyssaviruses. Our results demonstrated that active shedding of infectious virus was exclusively associated with two RABV isolates (92% for RABV-DogA and 67% for RABV-Insectbat), thus providing a potential explanation as to why sustained spillovers are solely attributed to RABVs. Interestingly, 3D imaging of a selected panel of brain samples from bat-associated lyssaviruses demonstrated a significantly increased percentage of infected astrocytes in mice inoculated with IRKV (10.03%; SD±7.39) compared to RABV-Vampbat (2.23%; SD±2.4), and BBLV (0.78%; SD±1.51), while only individual infected cells were identified in mice infected with Duvenhage virus (DUVV). These results corroborate previous studies on RABV that suggest a role of astrocyte infection in the pathogenicity of lyssaviruses.

Fig 1. In vitro replication kinetics of lyssavirus isolates in Na 42/13 cells infected with a MOI of 0.001.

The mean and standard errors of three replicates are indicated.

Fig 2. Experimental setup. Outline of the animal experiment and sample collection.

Fig 3. Kaplan-Meyer survival plots of the individual isolates following i.m. infection with low (A) and high doses (B).

Six BALB/c mice were inoculated per group. Mock-infected control mice did not develop any clinical signs and, hence, were omitted for better visualization.

Fig 4. Incubation periods after low dose (A) and high dose (B) i.m. infection.

Mean values are provided as horizontal lines. Statistical differences between the mean of RABV-DogA as a reference challenge strain and the means of other lyssaviruses are indicated (* p ≤ .05; ** p ≤ .01; *** p ≤. 001; ordinary one-way ANOVA with Tukey’s multiple comparison test).

Fig 5. Clinical signs in individual mice inoculated with the indicated lyssaviruses by the i.m. (n = 156; 78 high dose and 78 low dose), i.c. (n = 39) or i.n. (n = 78) route of infection.

The color code per cell represents the predominant clinical sign for each individual mouse before euthanasia or death. For clarity, high and low dose i.m. infections were combined.

Fig 6. Intramuscular pathogenicity index (IMPI) of the different lyssavirus isolates obtained in the mouse model.

Depicted are mean pathogenicity indices (median bar) of combined datasets of i.m. low (lower values) and high dose (upper values) infected animals. A maximum index value of 2 would be reached if all mice died at day 1 post-infection.

Fig 7. Comparison of the astrocyte tropism of different bat lyssaviruses with a high and low IMPI.

A) Indirect immunofluorescence of solvent-cleared brain sections demonstrates the presence of lyssavirus phosphoprotein P (red), neurons (blue, marker: NeuN) and astrocytes (green, marker: glial fibrillary acidic protein, GFAP). Insets show RABV P accumulation (red) at GFAP-positive cells (green). x, y = 387.5 μm, 387.5 μm; z = 65.5 μm (IRKV), 36.5 μm (RABV-Vampbat), 64.5 μm (BBLV), 65.5 μm (DUVV). Scale bar = 100 μm (overview), 15 μm (inset). B) Percentage of infected astrocytes (black dots) and neurons (gray squares). Per virus, 3 to 11 × 10³ astrocytes and 5 to 17 × 10³ neurons were counted in independent confocal z-stacks in two animals per isolate (one animal for DUVV). Each dot represents the ratio of infected astrocytes or neurons in an analyzed z-stack. Mean values are provided as horizontal lines. Statistical comparisons of the means between the different groups are indicated for those with a statistically significant difference (* p ≤ .05; ** p ≤ .01; *** p ≤. 001; ordinary one-way ANOVA with Tukey’s multiple comparison test). C) Corresponding data table for the measurements.

Fig 8. Comparison of virus shedding in lyssavirus-infected mice.

Percentage of animals positive/negative for viral RNA (A, B) and viable virus (C, D) in either salivary glands or oral swabs or both according to individual lyssaviruses (A, C) or grouped according to RABVs and non-RABV bat lyssaviruses (B, D). Correlation between ct-values as obtained in RT-qPCR and results of virus isolation in salivary glands (E) and oral swabs (F). Here, only animals were considered were active shedding (positive salivary gland and positive corresponding oral swab) was observed. Individual ct-values are shown and the mean is indicated. Successful virus isolations in cell culture are highlighted.