Intravenous inoculation of a bat-associated rabies virus causes lethal encephalopathy in mice through invasion of the brain via neurosecretory hypothalamic fibers

Abstract

The majority of rabies virus (RV) infections are caused by bites or scratches from rabid carnivores or bats. Usually, RV utilizes the retrograde transport within the neuronal network to spread from the infection site to the central nervous system (CNS) where it replicates in neuronal somata and infects other neurons via trans-synaptic spread. We speculate that in addition to the neuronal transport of the virus, hematogenous spread from the site of infection directly to the brain after accidental spill over into the vascular system might represent an alternative way for RV to invade the CNS. So far, it is unknown whether hematogenous spread has any relevance in RV pathogenesis. To determine whether certain RV variants might have the capacity to invade the CNS from the periphery via hematogenous spread, we infected mice either intramuscularly (i.m.) or intravenously (i.v.) with the dog-associated RV DOG4 or the silver-haired bat-associated RV SB. In addition to monitoring the progression of clinical signs of rabies we used immunohistochemistry and quantitative reverse transcription polymerase chain reaction (qRT-PCR) to follow the spread of the virus from the infection site to the brain. In contrast to i.m. infection where both variants caused a lethal encephalopathy, only i.v. infection with SB resulted in the development of a lethal infection. While qRT-PCR did not reveal major differences in virus loads in spinal cord or brain at different times after i.m. or i.v. infection of SB, immunohistochemical analysis showed that only i.v. administered SB directly infected the forebrain. The earliest affected regions were those hypothalamic nuclei, which are connected by neurosecretory fibers to the circumventricular organs neurohypophysis and median eminence. Our data suggest that hematogenous spread of SB can lead to a fatal encephalopathy through direct retrograde invasion of the CNS at the neurovascular interface of the hypothalamus-hypophysis system. This alternative mode of virus spread has implications for the post exposure prophylaxis of rabies, particularly with silver-haired bat-associated RV.

Figure 1. Strain and Inoculation Route Dependent Pathogenicity of SB and DOG4 in Mice.

Six- to eight-week-old mice were inoculated i.m. or i.v. with 10⁶ ffu of SB or DOG4. (A) The graph depicts the weight curves of ten mice per group as group averages (mean±standard error). Percentages relate to the body weight at the day of inoculation. (B) Kaplan-Meier plots show the survival probability per day for each experimental group. Data are combined from two independent experiments. Survival curves were statistically different (p<0.0001) between SB i.m. (n = 16) and DOG4 i.m. (n = 17), SB i.v. (n = 18) and DOG4 i.v. (n = 17) as well as between DOG4 i.m. and DOG4 i.v.

Figure 2. Strain and Inoculation Route Dependent Distribution of Viral Antigen in the Brain.

Brains of mice inoculated i.m. or i.v. with 10⁶ ffu of SB or DOG4 were analyzed when the animals were moribund or at the end of the 20-day observation period as indicated in the panels A to D. Sagittal sections were stained immunohistochemically against RV, subsequently visualized by an enzymatic reaction and documented with bright field microscopy. Abbreviations: Cb, cerebellum; cc, corpus callosum; CGMB, central gray substance of midbrain; Cx, cerebral cortex; Hy, hypothalamus; MO, medulla oblongata; RF, reticular formation; SptN, septal nuclei; Tg, midbrain tegmentum; Tec, tectum; Th, thalamus.

Figure 3. Residual DOG4 RNA in the CNS and Serum VNA Titers after i.v. Inoculation.

Ten mice were infected i.v. with 10⁶ ffu of DOG4 and euthanized pair wise at the indicated time points. (A) The number of RV genome equivalents per microgram total RNA isolated from spinal cord and brain was quantified by qRT-PCR. Data are mean RNA copy numbers (+standard error) calculated for each mouse pair. (B) From the same mice, blood was obtained immediately before euthanasia and VNA serum titers were determined in international units (IU).

Figure 4. Viral Burden in the CNS after SB i.m. or i.v. Inoculation.

Twelve mice were infected i.m. or i.v. with 5×10⁶ ffu of SB. Three mice per group were euthanized at each indicated time point and the number of RV genome equivalents per microgram total RNA isolated from spinal cord (A) and brain (B) was quantified by qRT-PCR. Data are mean RNA copy numbers (+standard error) calculated for three mice per time point. Asterisks indicate significant differences in the virus load at the indicated time point in dependence on the inoculation route (*, p<0.05).

Figure 5. SB Progression within the CNS after i.m. or i.v. Inoculation.

CNS tissue from four of twelve mice infected either i.v. (panel A–D) or i.m. (panel E–H) with 10⁷ ffu of SB was analyzed in more detail by immunohistochemical staining against RV. Animal numbers relate to Table 1. The red bar represents the section plane (stereotaxic coordinates: interaural 2.34 mm, bregma −1.46 mm) for the initial screening of all mouse brains (see text). Each vertical blue bar stands for another plane that was analyzed by several sections. The area caudal of the broken blue line represents the findings for the cervical spinal cord. The background color of each plane symbolizes its percentage of SB positive structures (nuclei, fiber tracts, cortical areas). Distinct structures were identified and counted using a stereotaxic mouse brain atlas . Gray colored regions were not analyzed. The total percentage of RV immunoreactive structures for each brain is noted in the left corner of each panel.

Figure 6. Potential Viral Invasion and Migration Pathways after i.v. (left) or i.m. (right) SB Inoculation.

Potential viral migration pathways between all RV antigen positive cortical areas and nuclei in brain and spinal cord, analyzed and identified as described in Figure 5, were composed manually by using the BrainInfo database (http://braininfo.rprc.washington.edu/) for published efferent and afferent connections. Depicted connections were chosen by evaluation of all possibilities with regard to the best match for each animal as well as between the animals in dependence on time and morbidity. Presented is a simplified version, summarized on the level of hierarchically higher-ordered structures.

Figure 7. Infection of Neurosecretory Hypothalamic Neurons by SB after i.v. Inoculation.

The panels show the hypothalamic paraventricular (PVN; panel A, D and F), arcuate (ARC; panel B) and supraoptic (SO; panel C, E and G) nuclei of mice inoculated i.v. with 10⁷ SB four days before euthanasia. The sections were stained immunohistochemically against RV only (panel A–C) or co-stained against RV and antidiuretic hormone (ADH; panel D and E) or oxytocin (Oxy; panel F and G). Abbreviation: 3^rd V, third ventricle. Black or white bars represent 20 (panel D–G), 25 (panel B, C) or 75 µm (panel A). Arrowheads point to neurons and fibers containing both RV and ADH, or Oxy respectively.