Recipient Organization
MONTANA STATE UNIVERSITY
(N/A)
BOZEMAN,MT 59717
Performing Department
Microbiology & Immunology
Non Technical Summary
Increases in the frequency of human-wildlife and wildlife-livestock interaction have led to emergence of new pathogens that threaren biosecurity, human health , and livestock health. Bats are hosts of some of the highest-profile emerging diseases, including Ebola, Marburg, Nipah, and Hendra viruses and severe acute respiratory syndrome coronavirus2 (SARS-Co-V-2 the cause of COVID-19). Cross-species transmission often occurs when bats abandon natural habitats to exploit resources associated with human settlements, and when bats are stressed and excreting high levels of pathogen.Spillover risk to livestock and humans is determined by the probability of an infective dose of pathogen passing through a series of potential barriers (Fig. 1): the reservoir host has to be present and infected; pathogen must be released from the reservoir host; pathogen must survive outside of the reservoir host with access to the recipient host (e.g. horse or human); recipient hosts must be exposed to the pathogen in sufficient quantity for an infection to establish; and recipient hosts must be susceptible to the virus (and not vaccinated). Certain conditions widen or close gaps in these barriers. The proposed study will examine the cause of coronavirus excretion from bats, using existing samples from a large study of bat viruses in Australia and Bangladesh.
Animal Health Component
50%
Research Effort Categories
Basic
50%
Applied
50%
Developmental
(N/A)
Goals / Objectives
We propose to assemble the first spatiotemporal dataset of coronavirus shedding from bat populations with existing metadata on bat demography, immune status, nutrition status, physiological status and coinfection status.We will use novel statistical tools to investigate how periods of coronavirus shedding coincides with i) stressful environmental conditions, such as food shortages, ii) shedding pulses of other viruses, including Hendra or Nipah virus.We will use modeling approaches to summarize viral shedding intensity into a single metric (area under the epidemic curve) that can be incorporated into mechanistic and machine learning models designed to predict spillover.These objectives will allow us to identify potential drivers of coronavirus shedding and spillover from bats to inform surveillance of bat viruses with zoonotic potential and help to prevent future pandemics.
Project Methods
Study DesignSample analysis. Fecal samples from both catching and under-roost sampling sessions will be screened with a beta-coronavirus pan-S RT-PCR. Positive samples will be sequenced using conventional Sanger sequencing and by full genome analysis with next generation sequencing, allowing us to select viruses or genes for further genetic characterization. For a selection of viruses, rapid VirCapseq will be used to obtain spike and full-length genome sequences to allow genotypic characterization and set up phenotypic screening.Data analysis The spatiotemporal dynamics of coronavirus shedding and exposure will be modeled using individual bat characteristics and information from under-roost data. Individual-level bat covariates such as species, sex, age, body condition, immunological status and diet, all collected from the existing project, will be integrated with the roost-level data to jointly predict high risk periods and locations for coronavirus shedding. We will develop a surveillance framework to detect early warning signs for periods of viral shedding. Co-occurrence among multiple viruses, and the nature of these interactions (antagonistic or amplifying) will be analyzed using association algorithms (machine-learning techniques).Sample collection and sampling frameworkAll samples for this project have already been collected. No new samples will be collected for this renewal project. All samples were collected under approved biosafety protocols from MSU and local universities and after risk assessments by local institutions. These approved biosafety protocols are routinely used to work with biosecurity level 4 pathogens (BSL4) such as Ebola and Nipah virus and were developed in collaboration with our BSL4 laboratory collaborators, the CDC, and local governments.Coronaviruses are excreted in bat feces. Bat fecal samples were collected as a part of the original project with a longitudinal and spatial sampling effort across the east coast of Australia and throughout Bangladesh. All fecal samples are currently stored in -80 freezers in-country. In addition, we have stored serum and blood in RNA protect, as well as multiple other samples and metadata associated with each bat.In Australia, we longitudinally sampled five permanent, urban bat populations within northeast New South Wales and southeast Queensland., where the majority of Hendra virus spillover events have occurred. Nomadic populations were opportunistically sampled when nomadic bats entered the study area. Most sampling was focused on Pteropus alecto, the species most often associated with virus excretion, spillover events, and urban habituation29. Pteropus poliocephalus, another host for Hendra virus, was also sampled when it co-occurred with P. alecto.In Bangladesh, we sampled three urban bat populations monthly, and we sampled populations associated with Nipha virus spillover events for two months post spillover.We collected samples from up to 60 individual bats each sample period (described in 30), which will gave us sufficient power to estimate prevalence and detect pulses of Hendra and Nipah virus. Samples were collected over 30 months (three winters).To sample under roost urine and feces, we adapted standardized techniques31, with the exception that sheet placement was based on modeling the bias of pooled versus individual urine and feces. Bat abundance, species composition, and reproductive success were estimated concurrently.Individual bats were captured in mist nets and anaesthetized according to published methods30. Body mass and forearm length was recorded. We collected fur, urine, and blood. A subset of bats were aged by residual premolar analysis30. Bats were released at the roost site after recovery from anesthesia.Metadata that will be associated with each sample will include: total immunoglobulin G (IgG) and IgA (Novus Biologicals) in serum and saliva33,34, gene expression of conserved antiviral and proinflammatory proteins (C-reactive protein, IL-6, TNF-α35), conserved biomarkers of cell (lactate dehydrogenase) and tissue (albumin) damage36,37 from infection and inflammation, fur and fecal cortisol, neutrophil/lymphocyte ratios and differential counts, and blood-borne bacterial infections including Bartonella, hemoplasmas, and Borrelia.Sample selection for coronavirus analysisTo select samples for coronavirus screening from our banked sample collection, we will use a data integration technique with two-phased testing of pooled samples. In the first phase, we will use pooled samples to estimate the population prevalence of coronavirus and inform efficient strategies for the second phase. To combine information from both phases, we will use a Bayesian data integration procedure that combines pooled samples with individual samples for joint inferences about the population prevalence. Data integration procedures result in more efficient estimation of prevalence than traditional procedures that only use individual samples or a single phase of pooled sampling.The first phase pooled sampling will include the following samples:Fecal samples: Bangladesh. 30 fecal samples from three longitudinal catching sites monthly for a year (1,080 samples). Australia. 30 fecal samples from five longitudinal under-roost sites every two months for two years (1800 samples) and 30 fecal samples from two longitudinal catching sites every two months for two years (500 samples).Serological samples. Bangladesh. 40 serum samples from each bat per three longitudinally sampled roosts for two years (1440 samples). Australia. 40 serum samples from each bat per two longitudinally sampled roosts for two years (960 samples).Coronavirus screeningFecal samples from both catching and under-roost sampling sessions will be screened with a beta-coronavirus pan-S RT-PCR developed by Vincent Munster. Positive samples will be sequenced using conventional Sanger sequencing and by full genome analysis with next generation sequencing, allowing us to select viruses or genes for further genetic characterization. For a selection of viruses, rapid VirCapseq will be used to obtain spike and full-length genome sequences to allow genotypic characterization and set up phenotypic screening.Serology will be used to examine multi-pathogen dynamics. We will use the Luminex platform to screen for antibodies to filoviruses, rubulaviruses, paramyxoviruses, henipaviruses, and coronaviruses.. To source the beads and antigens for the Luminex serological analyses, we have established a collaboration with Dr. Eric D. Laing and Dr. Chris Broder in the Department of Microbiology and Immunology, Uniformed Services University.Statistical analysisHealth metrics will be analyzed with a principal components analysis to derive a single axis of bat immune variation38. We will statistically model the longitudinal data, incorporating existing data39,40, to detect shedding pulses (using time-series methods we developed40), and compare timing, frequency, and magnitude of coronavirus shedding pulses to henipavirus shedding pulses. We will use general linear mixed models (GLMMs) to compare serological, physiological, immunological, and nutritional parameters in resident and migratory populations within and outside of Hendra virus and coronavirus pulses.The spatiotemporal dynamics of coronavirus shedding and exposure will be modeled using individual bat characteristics and information from under-roost data. Individual-level bat covariates will be integrated with the roost-level data to jointly predict high risk periods and locations for coronavirus shedding. We will develop a surveillance framework to detect early warning signs for periods of viral shedding. Co-occurrence among multiple viruses, and the nature of these interactions (antagonistic or amplifying) will be analyzed using association algorithms (machine-learning techniques).