Establishment of a longitudinal pre-clinical model of lyssavirus infection

Abstract

Traditional mouse models of lyssavirus pathogenesis rely on euthanizing large groups of animals at various time points post-infection, processing infected tissues, and performing histological and molecular analyses to determine anatomical sites of infection. While powerful by some measures, this approach is limited by the inability to monitor disease progression in the same mice over time. In this study, we established a novel non-invasive mouse model of lyssavirus pathogenesis, which consists of longitudinal imaging of a luciferase-expressing Australian bat lyssavirus (ABLV) reporter virus. In vivo bioluminescence imaging (BLI) in mice revealed viral spread from a peripheral site of inoculation into the central nervous system (CNS), with kinetically and spatially distinct foci of replication in the footpad, spinal cord, and hindbrain. Detection of virus within the CNS was associated with onset of clinical disease. Quantification of virus-derived luminescent signal in the brain was found to be a reliable measure of viral replication, when compared to traditional molecular methods. Furthermore, we demonstrate that in vivo imaging of ABLV infection is not restricted to the use of albino strains of mice, but rather strong BLI signal output can be achieved by shaving the hair from the heads and spines of pigmented strains, such as C57BL/6. Overall, our data show that in vivo BLI can be used to rapidly and non-invasively identify sites of lyssavirus replication and to semi-quantitatively determine viral load without the need to sacrifice mice at multiple time points.

Keywords: Bioluminescence imaging; CNS; Luciferase; Lyssavirus; Preclinical model.

Fig 1.. In vitro demonstration that pCMV-ABLV-luc produces infectious virus.

(A) Schematic representation of pCMV-ABLV-luc, showing the location of the firefly luciferase gene between the ABLV G and L genes. (B) HEK293T cells transfected with the indicated plasmids were lysed, incubated with luciferin and quantified for luminescence. Values are mean luminescence generated from triplicate samples (***p=0.0001, One-way ANOVA of log-transformed data, Dunnett’s Multiple Comparisons Test, error bars, SEM). (C) HEK293T cells transfected with the indicated plasmids or infected with WT ABLV were stained with anti-rabies G to detect virions. Bar, 20 μm.

Fig 2.. Bioluminescence imaging of mice infected with increasing doses of ABLV-luc.

(A) Mice were infected with 2×10³, 2×10⁴, or 2×10⁵ FFU of ABLV-luc on day 0 (n=3 mice/group) and bioluminescence imaging was used to detect virus. (B) Bioluminescence imaging of mice infected with 2×10³, 2×10⁴, or 2×10⁵ FFU of ABLV-luc. Values for pseudocolor intensity scale are mean luminescent intensity. Viral burden was quantified as mean luminescence intensity in the footpad (C), spines (D), and brains (E) of infected mice. Inset images in (C) show same data with a reduced y-axis scale to better enable visualization of quantitative differences from controls.

Fig 3.. Loss of body weight and mortality in response to increasing doses of ABLV-luc.

(A) Percent starting body weight as an indicator of disease. (B) Kaplan-Meier survival plot of mice infected with increasing doses of ABLV-luc (*p=0.0224, Log-rank Mantel Cox test).

Fig 4.. 360° in vivo imaging of ABLV-luc infection to better determine anatomical location of zones of intense viral replication.

(A) Bioluminescence imaging of ABLV-luc infection in a mouse infected with 2×10⁵ FFU ABLV-luc at 10 days post-infection. Images were taken in 45° increments to more precisely localize the anatomical origin of the luminescence signals. Bioluminescence imaging of ABLV-luc infection in a mouse infected with 2×10⁵ FFU ABLV-luc at day 7 (B) and day 10 (C) post-infection. Images were taken at an angle of 270°.

Fig 5.. ABLV-luc and WT ABLV show indistinguishable virulence in vivo.

(A) Mice were infected with 2×10⁴ FFU of ABLV-luc or WT ABLV on day 0 (n=6 mice/group) and monitored for signs of disease thereafter. (B-D) Cumulative disease score, percent starting body weight, and Kaplan-Meier survival plot of mice infected with ABLV-luc or WT ABLV (n.s., not significant, Log-rank Mantel Cox test).

Fig 6.. Correlation between ABLV-luc bioluminescent imaging intensity and ABLV-N transcript number in the brains of infected animals.

B6 albino mice were infected with 2×10⁵ FFU of ABLV-luc and luminescence intensity in the brain was quantified using a Bruker Xtreme II instrument. Immediately after imaging, animals were euthanized, and total RNA was harvested from the brain. ABLV-N transcripts were measured by qRT-PCR, as described in 2.6. Each dot represents a single animal. Line represents linear regression analysis performed using a least squares fit.

Fig 7.. In vivo BLI comparing luminescence signals between B6 albino, shaved C57BL/6 mice, unshaved C57BL/6 mice, and uninfected C57BL/6 mice.

(A) C57BL/6 or B6 albino mice were uninfected or infected with 2×10⁵ FFU of ABLV-luc on day 0 (n=3 mice/group) and bioluminescence imaging was used to detect virus. (B) Hair was shaved from the backs and heads of three infected C57BL/6 mice and BLI was performed to detect virus. Signal was compared to B6 albino mice, one unshaved, infected C57BL/6 mouse, and one shaved, uninfected C57BL/6 mouse. Viral burden was quantified as mean luminescence intensity in the spines (C) and brains (D) of infected mice (n.s., not significant, Unpaired t test). For numerical data from inoculated foot, spine and brain of each animal, see Table 1.