Reassessing West Nile virus action thresholds for improved disease control

Project Summary/Abstract West Nile virus (WNV) is the most consequential arthropod-borne virus (arbovirus) in the US, and due to the lack of effective treatments or vaccines, preventing human exposures to infectious mosquitoes is key to controlling transmission. Mosquito-based surveillance for WNV in the US is robust and provides policy makers information on the human risk of WNV infection in near "real-time" in the form of the Vector Index (VI), a measure of the abundance of infected mosquitoes in a given area. Human cases are nearly always preceded by an increase in the number of mosquitoes testing positive for WNV, and thus an increase in the VI, weeks prior to onset. However, an increase in the VI is not always followed by the detection of human cases. These observations suggest the accuracy of the VI can be improved. The VI is calculated by determining the number of mosquito pools that test positive for WNV, typically through standard RT-qPCR assays. While these assays are inherently quantitative, the outputs are interpreted in a binary fashion, either the presence or absence of WNV RNA. Quantity of WNV RNA from mosquito surveillance pools, as measured by cycle threshold (CT) values, varies dramatically and is not considered when calculating or interpreting the VI. The objective of this proposal is to determine how the variability in WNV CT values from mosquito surveillance pools is associated infectiousness (i.e. the point at which a mosquito can transmit WNV) in mosquitoes and to use these data to refine the VI. The central hypothesis is that WNV CT values are a reliable predictor of infectiousness in mosquitoes, and therefore useful indicators of human risk of infection. This hypothesis will be tested in two independent and complimentary aims. Aim 1: Identify the biological underpinnings of the observed variability in WNV RNA quantities from mosquitoes. This aim will make use of a robust experimental design to expose mosquitoes to i) a range of WNV titers via lab-derived bloodmeals and ii) viremic blood collected through the course of experimental infection in birds to determine how these factors affect WNV growth, time to infectiousness, and titers in saliva. Importantly, the results of these experiments will make the explicit connection between whole body CT values (what is measured in WNV surveillance) and the ability of mosquitoes to transmit WNV. Aim 2: Incorporate WNV CT values generated from mosquito surveillance into measures of human risk. This aim will result in two biostatistical modeling strategies that incorporate WNV CT values produced during routine mosquito-based surveillance into measures of the VI. The accuracy of these models to predict the timing and magnitude of human WNV cases will be determined using a robust retrospective dataset of over a decade of human and mosquito WNV surveillance data from Nebraska and Colorado. Taken together, this R21 is addressing a significant knowledge gap in arbovirus surveillance in the US by using innovative experimental and biostatistical approaches to identify the biological underpinnings of WNV RNA variability in naturally infected mosquitoes and to use these data to better estimate human risk and inform control campaigns.