Investigation of the molecular epidemiology of West Nile virus at the human-animal interface in South Africa

Abstract

West Nile virus (WNV) is an emerging, neurotropic zoonotic arbovirus that belongs to the Flaviviridae family and the Flavivirus genus. Since its discovery in 1937 in Uganda, Africa, WNV has continued to spread outside of its original geographic range and has now been named the most widespread cause of arboviral neurological disease, occurring on every continent except Antarctica. West Nile virus is maintained in nature through an enzootic bird-mosquito transmission cycle, with Culex (Cx.) univittatus as the principal vector in South Africa. Birds are used as sentinels for WNV in many countries, but in Africa birds seldom die or show signs of disease therefore alternative animals- such as horses- are used as sentinels for WNV. In equines, 90.00 % of symptomatically infected horses will develop neurological disease, and 35 % of infections are fatal. Besides horses, additional vertebrate hosts such as domestic animals, cattle, sheep, goats, and wildlife exist. In humans 80.00 % of WNV infections are asymptomatic, and usually between 1-10.00 % of cases will progress to neurological disease with a fatality rate of 10.00 %, although recent outbreaks in Europe have seen an increase in neurological disease manifestations up to 77.00 % and mortality rates of 40.00 %. This type of data is currently not available for South Africa.

This study aimed to investigate the molecular epidemiology of WNV in South Africa using a One-health approach and reports on the active circulation of endemic lineage 2 strains amongst mosquitoes, animals, and humans. The burden of disease and association with neurological disease was identified through syndromic surveillance in animals and humans. Additionally, a next-generation sequencing (NGS) pipeline was developed to sequence WNV full genomes from clinical specimens.

Chapter one provides an in-depth review, first starting broad with the topic of flaviviruses, then taking a closer look at WNV highlighting the importance of this virus as a pathogen of medical and veterinary importance.

Chapter two describes the optimization of a real-time Pan-flavivirus reverse transcription-polymerase chain reaction (RT-PCR) capable of detecting all members of the Flavivirus genus to identify infections in mosquitoes, animals and humans and the comparison thereof to the historically used assay.

In chapter three, flavivirus screening was performed on mosquitoes collected at urban, peri-urban and wildlife sites across the northeastern regions of South Africa. This chapter identified West Nile- and Banzi viruses in 16/1471 (1.08 %, minimum infection rate [MIR] = 0.39) and 2/1471 (0.14 %, MIR = 0.05) pools respectively. Phylogenetic analysis characterised WNV strains as lineage 2. Cytochrome oxidase 1 confirmed morphological identification of potential vector species but also identified new potential vectors for these flaviviruses in South Africa. Minimum infection rates were calculated per site to identify places of increased risk for spillover to humans and animals and identified Kyalami (Gauteng, peri-urban, MIR = 2.53) and the Kruger National Park (wildlife, MIR = 1.22) as ‘hot spots’ for WNV infection in animals and humans.

Chapter 4 describes the use of molecular and serological tools in conjunction with syndromic surveillance to identify WNV as the aetiological agent responsible for acute febrile or neurological disease in animals in South Africa. From January 2017 to December 2020, WNV was detected in 11/835 (1.32 %) horses by real-time RT-PCR and in 69/657 (10.50 %) through IgM serodiagnosis that could be confirmed with a microtiter virus neutralization test (micro-VNT). Therefore, a total of 77/835 (9.22 %) acute WNV infections were diagnosed in equines experiencing clinical disease. A further 7/243 (2.88 %) diseased or dead non-equine animals tested positive for WNV by real-time RT-PCR. Phylogenetic analysis confirmed all PCR positive cases to be lineage 2 infections, except for a single lineage 1 case in a lion. Pairwise distance analysis confirmed circulation of endemic lineage 2 strains amongst mosquitoes and animals in South Africa. A geographical correlation between the animals and areas with high mosquito MIR was observed.

In chapter 5, a retrospective pilot study identified WNV in 3.65 % of cerebrospinal fluid (CSF) specimens sampled from patients suffering acute undiagnosed neurological disease in Gauteng, South Africa through molecular and serodiagnosis. Phylogenetic analysis of two WNV strains identified in the CSF of two children confirms circulation of endemic neuroinvasive strains between mosquitoes, animals and people in Gauteng, South Africa. Prospective syndromic surveillance focusing on acute febrile diseases of unknown cause (AFDUC), with or without neurological signs, identified WNV as the aetiological agent in 8.26 % of cases through IgM serodiagnosis and virus neutralization assay confirmation. Odds ratio analysis identified clinical signs associated with infection and elucidated socioeconomic and demographic risk factors for WNV infection in South Africa. Patients residing close to the rural Mpumalanga site were more likely to contract WNV infection, which correlates with the high mosquito MIR identified in Chapter three.

Finally, chapter 6 describes three next-generation sequencing approaches to obtain full genomes from WNV clinical specimens or cell culture isolates. For the first time in South Africa, full genome sequences originating in mosquitoes and a non-equid animal were obtained. Phylogenetic analysis shows a high nucleotide and amino acid similarity amongst the sequences and to their closest relative, WNV HS101/08, a neuroinvasive strain isolated from the brain of a horse in South Africa. This confirms the active dissemination of neuroinvasive strains between mosquitoes and animals in South Africa, as predicted in the previous chapters. Phylogenetic analysis elucidated that circulating stains were genetically related despite arising from different spatial and temporal parameters.

The success of predicting areas of increased WNV spillover to animals and humans using mosquito MIR highlights the benefit of viewing arboviral infections in the context of One-health. Mosquito surveillance is a valuable tool as it allows timely detection of species diversity, distribution, and infection rates which are vital for estimating the risk and transmission of vector-borne disease. Surveillance in humans with AFDUC and/or neurological signs helped to define the burden of disease and indicated an increased risk for WNV associated hospitalisations in summer and late autumn in South Africa suggesting that the virus is still missed likely due to a lack of clinical awareness of the potential risk for more severe disease. Active surveillance at the animal-human interface may aid in predicting outbreaks and may assist in the prevention, diagnosis, and control of these arboviruses at both the local and international levels.