Investigation of neuropathogenesis and control of an equine-pathogenic Australian West Nile virus isolate in an established CD1 Swiss white mouse and a novel rabbit model

Abstract

West Nile virus (WNV) is a globally significant mosquito-borne neurotropic virus. In 2011, alarge-scale arboviral encephalitis outbreak occurred in Australia, involving close to 1000 horses. Avirulent WNV strain, WNVNSW2011, was isolated from the brain of an encephalitic horse during thisoutbreak. Preliminary assessment showed that this strain is intermediate in virulence and belongs inthe same lineage as the highly virulent North American strains. However, the pathogenesis ofWNV in general, let alone this novel Australian strain, remains largely unclear.The first part of the project aimed to investigate the pathogenesis of lethal WNVNSW2011 infection inthe established young adult Swiss white (CD1) mouse model, since previous survival challengeexperiments demonstrated high lethality (~70%) of WNVNSW2011 in this animal model. I conducteda pilot time-course experiment, which demonstrated only benign pathological and virologicaloutcomes in sacrificed mice on day 3 and 7 post-infection (pi), despite the relatively high lethalityseen in previous survival trials. In a subsequent experiment, I characterised the disease phenotypein moribund/dead and surviving mice after WNVNSW2011 infection. Notably, I detected a highdegree of intra-group variability in the neuropathology, neural infection and glial cell activation.Similar variability was also observed in CD1 mice challenged with another Australian WNV strainof intermediate virulence, WNVNSW2012. These inconsistencies in the end-point disease suggest thatrandom sacrifice of infected mice at any given time-point may not give reliable information of thepathogenic processes towards lethal encephalitis. Therefore, conventional sacrifice experimentsmay not be the best approach for investigating pathogenesis of virus strains of intermediatevirulence. This previously unacknowledged challenge to investigating pathogenesis of virus strainsof intermediate virulence should inform the planning and execution of future studies with such virusstrains.The second part of the project addressed the need for a resistant small animal model for studyingthe in vivo control of virulent WNV infections. Since rodent models typically produce lethaldisease after infection with virulent WNV strains, they may not reflect what happens in the natural“dead-end” hosts, humans and horses, in whom most WNV infections (> 90%) are non-lethal.Even cases with neuroinvasion are typically not fatal, unless underlying co-morbidities orimmunosuppressive disease exist.I chose rabbits as a candidate for a novel animal model, due to the ease of performing intra-vitalsampling and monitoring, as well as the possibility for tissue sharing in different assays, given thelarger size of rabbits than mice. Rabbits are also hind-gut fermenters, and thus may be a betterrepresentation of a horse than a mouse. New Zealand White (NZWRs; Oryctolagus cuniculus) and North American cottontail rabbits (CTRs; Sylvilagus sp.) were challenged with WNV strains ofdifferent virulence (WNVNSW2011 and WNVTX8667), and the prototype Murray Valley encephalitisvirus strain (MVE1-51), by the footpad route. The rabbits were consistently resistant to developingsevere disease after virulent WNV and MVEV infection, regardless of age, gender and species,despite productive virus replication in the draining popliteal lymph node (PLN). The resultantviraemia for all rabbits was also of low magnitude and transient. Furthermore, establishment ofvirus infection in the brain was not a feature of the disease, despite mild to moderateneuropathology in a subset of rabbits. This capacity to control flavivirus infection may beexplained by the rapid antiviral innate immune response detectable by day 3 pi, as well as a fastanti-WNV neutralising antibody response detectable from day 7 pi.Comparisons of the tissue-specific transcriptional profile of important cytokine and chemokinegenes suggested that virus control in rabbits was achieved differently depending on the virus, aswell as the rabbit species. However, common transcriptional upregulation of IFNγ in the drainingPLN, TNFα in the non-draining PLN and IL6 in the spleen was detected in all rabbits on day 3 pi,suggesting a common tissue-specific innate immune response signature against flavivirus infection.Involvement of anti-viral, pro- and anti-inflammatory signalling molecules, such as NF-kB2,STAT1, and STAT3, was also detected mostly in infiltrating mononuclear leucocytes in brainsections with elevated levels of IFN-I/II, TNFα, IL6 and IL10 transcripts. Thus, immunocompetentrabbits are able to resolve virulent flavivirus infections via a well-coordinated antiviral, pro- andanti-inflammatory immune response. Given the similarities to the non-lethal disease profile of mosthuman and equine infections, rabbits are a valuable model for studying the mechanisms of viruscontrol and pathogenesis of non-lethal disease.This project has therefore advanced the field of WNV research by describing a previouslyunacknowledged complication for studies of the pathogenesis of infection involving virus strains ofintermediate virulence, as well as having established an alternative animal model for studying viruscontrol and induction of non-lethal disease.