Spikes in Hendra spillover: early warning through the bat urinary metabolome

Transcription

1

2 Spikes in Hendra spillover: early warning through the bat urinary metabolome by Michelle Baker, Gary Crameri, Ina Smith July 2015 RIRDC Publication No 15/064 RIRDC Project No PRJ This research was funded by the Commonwealth of Australia, the State of New South Wales and the State of Queensland under the National Hendra Virus Research Program

3 2015 Rural Industries Research and Development Corporation. All rights reserved. ISBN ISSN Spikes in Hendra spillover: early warning through the bat urinary metabolome Publication No. 15/064 Project No. PRJ The information contained in this publication is intended for general use to assist public knowledge and discussion and to help improve the development of sustainable regions. You must not rely on any information contained in this publication without taking specialist advice relevant to your particular circumstances. While reasonable care has been taken in preparing this publication to ensure that information is true and correct, the Commonwealth of Australia gives no assurance as to the accuracy of any information in this publication. The Commonwealth of Australia, the Rural Industries Research and Development Corporation (RIRDC), the authors or contributors expressly disclaim, to the maximum extent permitted by law, all responsibility and liability to any person, arising directly or indirectly from any act or omission, or for any consequences of any such act or omission, made in reliance on the contents of this publication, whether or not caused by any negligence on the part of the Commonwealth of Australia, RIRDC, the authors or contributors. The Commonwealth of Australia does not necessarily endorse the views in this publication. This publication is copyright. Apart from any use as permitted under the Copyright Act 1968, all other rights are reserved. However, wide dissemination is encouraged. Requests and inquiries concerning reproduction and rights should be addressed to RIRDC Communications on phone Researcher Contact Details Name: Michelle Baker Address: CSIRO, Australian Animal Health Laboratory, Biosecurity Flagship, Geelong, VIC, 3220 Phone: Fax: Michelle In submitting this report, the researcher has agreed to RIRDC publishing this material in its edited form. RIRDC Contact Details Rural Industries Research and Development Corporation Level 2, 15 National Circuit BARTON ACT 2600 PO Box 4776 KINGSTON ACT 2604 Phone: Fax: Web: Electronically published by RIRDC in July 2015 Print-on-demand by Union Offset Printing, Canberra at or phone ii

4 Foreword To date, over 80 viruses have been identified in various species of bats, many of which are highly pathogenic in humans and other species. Hendra virus which is carried by flying foxes was discovered in 1994 following the spillover of Hendra virus from bats into horses and subsequently into humans. Hendra virus has a mortality rate of 60% in humans and 75% in horses and annual spillover events from bats into horses have occurred since its discovery. However, during 2011 a significant increase in Hendra virus prevalence occurred with 18 spillover events from bats into horses. The triggers associated with spillover events remain unknown and there are currently no biomarkers available for the early detection of increased Hendra virus prevalence. This project investigated the presence of viral co-infections in bats associated with differences in Hendra virus prevalence to assess the extent of viral co-infections present in bats, particularly associated with peaks in Hendra virus prevalence. In addition, a proof of principle analysis was performed to assess the use of metabolomics to identify biomarkers excreted in bat urine associated with viral infection in bats. The research demonstrated that bat colonies with a high Hendra virus prevalence contain a correspondingly high prevalence of other viruses, including Adenoviruses and paramyxoviruses belonging to the genera henipavirus and rubulavirus. Metabolomics analysis of experimentally infected bats shows a clear difference in the metabolites present in infected and uninfected bats and demonstrated that urine is suitable for biomarker detection in bats. The identification of viral co-infections associated with Hendra virus in bats provides information on viruses that may have the potential to impact human and animal health in the future. Viral coinfections and metabolic markers can be used to predict viral spillover events and this information can be used by decision makers in the assessment of the risk of spillover of Hendra virus and other viruses from bats. This report is an addition to RIRDC s diverse range of over 2000 research publications and it forms part of the National Hendra Virus Research Program, which aims to minimise the impact of Hendra virus. Most of RIRDC s publications are available for viewing, free downloading or purchasing online at Purchases can also be made by phoning Craig Burns Managing Director Rural Industries Research and Development Corporation iii

5 Acknowledgements Linfa Wang, Jenn Barr, Shawn Todd, Vicky Boyd, Glenn Marsh and Mary Tachedjian from the CSIRO, Australian Animal Health Laboratory (AAHL) provided intellectual input, contributed to sequencing, virus isolation, data analysis and sample collection. Craig Smith, Dan Edson, Alice Broos, Lee McMichael and Deb Melville from Biosecurity Queensland and Amy Burroughs and Bronwyn Clayton from AAHL contributed to sample collection. Malcolm McConville, David De Souza and Dedreia Tull from The University of Melbourne performed mass spectrometry on the bat serum and urine samples and statistical analysis on the resulting metabolomics profiles. iv

6 Contents Foreword... iii Acknowledgements... iv Contents... v Executive Summary... vii Introduction... 1 Objectives... 3 Methodology... 4 Part A: Virus identification and discovery... 4 Part B: Metabolite discovery... 4 Results... 5 Implications Recommendations References v

7 Tables Table 1. Summary of all paramyxoviruses isolated from this study (from Barr et al., 2015)... 9 Table 2. Isolation detail of the four novel rubulaviruses (from Barr et al., 2015) Table 3. Significant metabolites detected in bat urine and serum samples following Menangle virus infection. Metabolites were considered significant if the BH adjusted p value was < Figures Figure 1. Bat urine samples collected in Boonah, Queensland from July 2011 to November The presence of viral RNA corresponding to multiple bat viruses provided evidence for high levels of viral co-infections during spikes in Hendra virus prevalence in July-August Boxed regions correspond to colder months Figure 2. Viral RNA was detected in bat urine samples collected from locations in Queensland (Cairns and Hervey Bay) and New South Wales (Alstonville, Nambucca Heads and Sydney). The presence of multiple viruses provides evidence for viral co-infections, particularly during spikes in Hendra virus prevalence. * indicate spikes in Hendra virus prevalence Figure 3. Bat urine samples collected at Cedar Grove, Queensland from September 2009 to May * indicates a spike in Hendra virus prevalence Figure 4. Principal components analysis scores plots of (A) 345 metabolites measured using GC-MS in bat serum samples collected from captive bats pre- (blue) and post- (red) Menangle virus infection. (B) 409 metabolites measured using GC-MS in bat urine samples from captive bats pre-(purple) and post (blue) Menangle virus infection and from pooled urine pre-infection (green)... 8 vi

8 Executive Summary What the report is about In 2011 a significant increase in the spillover of Hendra virus from bats into horses occurred. In response to this increase, the New South Wales, Queensland and federal government committed funds to research leading to strategies to minimise the impact of Hendra virus. This report describes a project that was funded to identify viral co-infections and metabolic biomarkers associated with increases in Hendra virus prevalence in bats. These biomarkers have the potential to be used to identify periods of high risk associated with spillover events. Who is the report targeted at? This report is targeted at the National Hendra virus taskforce, horse owners and veterinarians, animal health committees, staff in veterinary diagnostic laboratories, the research community and interested individuals. Where are the relevant industries located in Australia? The most relevant industry for this project is the horse industry, in particular the horse racing industry and horse owners in general in Queensland and New South Wales where Hendra virus cases have been identified. Aims/objectives The main objective of this project was to provide information on virus dynamics in bats by determining the role of viral co-infection in changes in Hendra virus prevalence in bat populations. A secondary objective was to develop urine metabolic profiles for bats, whether of secondary metabolites, hormones, or signalling molecules, that are occurring in response to the pathophysiological event/s that precipitate increased viral replication. These may include such factors as nutritional stress, movement stress, pregnancy, birth or lactation - and having a predictive association with detection of viral infection. Translation of this data to a population landscape context will assist in recognition of periods of increased risk of ramping up of Hendra virus replication in bats that may in turn be applied at the local population level. This information will be used to develop tools and strategies for long term monitoring of bat populations that will inform both local and more wideranging risk management practices involving horses and other animals at risk of Hendra virus infection. This project is relevant to RIRDC s goal of providing knowledge to address rural issues with the potential to identify biomarkers which could be used as tools to detect periods of increased Hendra virus risk. Methods used The methodology for the virus identification and discovery part of the project included standard virus isolation techniques, polymerase chain reaction (PCR) and luminex technologies. Viruses were detected using both real-time and conventional PCR as well as next generation sequencing. Luminex bead based multiplex PCR assays were also used to enable rapid and efficient screening of large numbers of samples for a variety of virus targets. Mass spectrometry techniques were used to provide a broad overview of the serum and urinary metabolome of experimentally infected and uninfected bats. vii

9 Results/key findings Virus family PCRs have demonstrated that bat colonies with a high Hendra virus prevalence contain a correspondingly high prevalence of other viruses including paramyxoviruses from the genera henipavirus and rubulavirus. A seasonal trend in the presence of viruses was also observed indicating that environmental triggers may be associated with spillover events. Virus isolation from bat urine collected during the 2011 Hendra virus spillover events have resulted in the isolation of over 40 viruses including a large number of new viruses which are yet to be classified. Urine and serum collected from captive female flying foxes experimentally infected with the bat paramyxovirus, Menangle virus were used for a proof of principle study to examine the metabolic profile of infected and uninfected bats. Matching serum and urine samples were used for initial validation to determine the utility of using urine to identify changes in the metabolic profile. A clear difference in the metabolic profiles of the infected and uninfected bats was identified in both urine and serum. The metabolic profile of pooled urine samples collected from the cages of uninfected bats were tested to provide information on the profile expected from a mixed bat population consisting of both males and females. The results of this analysis demonstrated that a similar metabolic profiles were detectable in pooled samples to those of individual samples collected from uninfected bats. This analysis supports the feasibility of using pooled urine samples collected in the field from wild bat colonies. To our knowledge this is the first metabolomics analysis performed on bat samples and provides important preliminary data on metabolites that may be appropriate biomarkers for identifying metabolic changes associated with changes in viral replication. Implications for relevant stakeholders for: This project has identified biomarkers for the assessment of the risk of Hendra virus spillover from bats into horses and other susceptible species and provided information on the feasibility of using urine as a non-invasive method for biomarker detection. This is the first study of its kind to identify viral co-infections and changes in the metabolic profile of bats associated with viral infection. The ability to identify biomarkers associated with increased viral replication has the potential to be used by the horse industry and veterinarians to assess the risk posed by bats. Recommendations This report has demonstrated that peak periods of Hendra virus spillover from bats are associated with a peak in other viral infections. Current recommendations associated with reducing the risk of Hendra virus spillover to horses including removal of fruit trees in horse paddocks and avoiding areas associated with bats should remain in place and may assist in reducing the risk of spillover of other viruses. The metabolomics component of the work is now ready to be validated using field samples from wild bat colonies. Separate from this project, a handheld device for use in the field is currently under development and could be applied for the detection of metabolic biomarkers associated with increased viral replication. viii

10 Introduction The prevalence of Hendra virus in pooled urine collected from flying fox colonies varies considerably over time; ambient levels have ranged from 0 to 10% and, until recently, the prevalence in colonies near spillover events have remained below 5%. In 2011, the prevalence of Hendra virus in urine samples collected from colonies spatially linked with Hendra virus outbreaks ranged from 3-67% at one location over two months. By comparison with the previous 15 years, 2011 was clearly unusual for the number of recorded outbreaks, but the contribution of environmental, seasonal or physiological changes (including events such as bat pregnancy, birth and lactation) is unknown. Accordingly, the likelihood of recurrence in years to come cannot be estimated. Urine samples collected from flying fox colonies linked with Hendra virus outbreaks provide a unique window of opportunity to determine the prevalence of other viruses and the metabolic changes associated with increases in Hendra virus prevalence. Current monitoring of bat populations for Hendra virus focuses on detecting changes in Hendra virus viral Ribonucleic acid (RNA) in urine. However, by the time increased Hendra virus viral RNA is detected in bat urine, the risk of exposure of susceptible species is already present. Samples collected in 2011 provided the first opportunity to examine Hendra virus prevalence in urine collected from bats in the vicinity of outbreak events. These samples provided evidence for an increase in the rate and persistence of Hendra virus shedding by bats associated with Hendra virus spillover events. These preliminary studies have also provided evidence for an increase in the levels and number of other viruses coinciding with periods of high Hendra virus prevalence. These results are consistent with the possibility that the increase in Hendra virus in bat populations is not simply due to the introduction of Hendra virus into a naïve population but due to more complex events which lead to increased replication of a large number of viruses. It is also possible that increased viral co-infections predispose bats to Hendra virus infection/replication. The goal of this project was to identify biomarkers that precede or coincide with increases in Hendra virus prevalence and, by association with environmental and physiological events, thus identifying triggers for Hendra virus spillover events. We hypothesised that viruses may be persistently maintained within individuals and physiological changes trigger an active infection of multiple viruses. To date, over 80 viruses have been identified in various species of bats with this number increasing steadily as new viruses are discovered (Calisher et al., 2006). The extent of viral-co-infection in bats is unknown but it is likely that bats harbour a variety of viruses simultaneously. Increases in Hendra virus prevalence have been detected in bats using standard PCR techniques specific for Hendra virus. However, it is possible that Hendra virus is only one of many viruses that undergo a spike in prevalence and bats may be shedding a cocktail of viruses in both urine and faeces, some of which may contribute to Hendra virus spillover events or lead to infections of their own in horses or other exposed species. Identification of viruses that accompany or precede increases in Hendra virus may lead to the identification of agents that can be used as markers to predict Hendra virus spillover events and will increase our understanding of viral dynamics in bats in general. In addition, further identification of bat viruses will provide insight into other viruses that may have the potential to impact human and animal health in the future. As these viruses can be measured in urine and potentially faeces, they may provide early indicators of Hendra virus status that can be detected noninvasively. The factors that influence Hendra virus spillover from flying foxes to horses are unknown. However, it is possible that environmental and/or physiological stressors predispose bats to viral infection or result in viral recrudescence. Evidence for viral recrudescence in bats has recently been reported in a long term study of captive pteropid bats in which evidence of Nipah virus recrudescence was detected (Sohayati et al., 2011). A recent study of fruit bats on an isolated island in the gulf of Guinea (Peel et al., 2112) also provided evidence for the presence of antibodies against henipaviruses, consistent with the possibility that persistent henipaviruses infections are maintained in individual animals with the potential for recrudescence. These results are consistent with the ability of bats to harbour viruses at undetectable levels with the potential for changes in replication to occur in response to triggers such as stress. Flying foxes may experience stress from external factors including climatic conditions, food 1

11 availability, habitat changes and social interactions. Stressful life history events including mating, pregnancy and lactation may also play a role in viral dynamics. Data collected from Hendra virus outbreaks from 1994 to 2010 has demonstrated a higher risk of Hendra virus spillover from May to October (Plowright et al., 2011). As flying foxes in Southern Queensland mate in March and April and give birth in October, the peak in Hendra virus spillover events corresponds with the period when adult females are pregnant and giving birth and juveniles have undergone a decline in maternal immunity. These are two factors which may increase the prevalence of Hendra virus in bat populations. Similarly, changes during pregnancy have been speculated to be involved in the spillover of Ebola Zaire virus from bats to apes (Leroy et al., 2005). However, the physiological changes that accompany increases in viral prevalence resulting in spillover events are unknown and at present these links remain speculative. Metabolites represent the ultimate product of gene, mrna, and protein activity, providing information as close to the phenotype as possible. Metabolomics is the quantitative analysis of small metabolites (<1500kDa) and metabolic fluxes in biological systems and is the most sensitive method for detecting upstream changes in physiology. Factors such as gender, age, diet, disease, hormone, and stress levels cause changes in the production of metabolites and their products which in turn can be used as indicators or biomarkers of an altered physiological state. As these factors would each be expected to produce a unique metabolic profile, identification of metabolic patterns associated with Hendra virus spillover and correlation with environmental and/or physiological events may reveal the triggers responsible for increases in Hendra virus replication in bats. The use of metabolomics is well established in various aspects of human medicine including prognosis, diagnosis, personalised medicine, human performance assessment and in environmental monitoring (Munger et al., 2008; Nicholson, 2006; Vinayavekhin et al., 2009). The identification of metabolic changes associated with environmental and/or physiological stress and/or increases in viral replication offer the opportunity to predict changes in Hendra virus replication in bats prior to spillover events. A variety of biological fluids including urine, serum, blood, saliva, hair and cerebrospinal fluid have been shown to provide a reliable reflection of the metabolome. As changing metabolic patterns can be detected in urine, bat colonies can be monitored non-invasively and on a routine basis, providing an early warning system, offering the opportunity to prevent susceptible species such as horses from exposure to Hendra virus. Metabolomics approaches also have the potential to identify new pathways linked to the regulation of an infectious disease which may be used as targets for therapeutic intervention. 2

12 Objectives Changes in Hendra virus prevalence in flying foxes have been strongly associated with an increased incidence of spillover of Hendra virus to horses. However, not only is there currently no firm data on the social or environmental triggers for increases in Hendra virus prevalence in bats, there are no tools for timely prediction of heightening shedding by individual bat populations. The main objective of this project was to determine the role viral co-infection in changes in Hendra virus prevalence in bat populations. A secondary objective was to develop urine metabolic profiles for bats, to assess the feasibility of using urinary metabolic markers to identify secondary metabolites, hormones or signalling molecules that are occurring in response to the pathophysiological event that precipitates increased Hendra virus replication such as nutritional stress, movement stress, pregnancy, birth or lactation - and having a predictive association with changes in infection status or viral replication. The study aimed to deliver lab-based analysis of field samples collected in both surveillance and strategic sampling programs to identify changes in urinary metabolites that were temporally associated with subsequent alteration in Hendra virus prevalence in bats. Translation of this data to a population landscape context will assist in recognition of periods of increased risk of ramping up of Hendra virus replication in bats that may in turn be applied at local population level. This information will be used to develop tools and strategies for long term monitoring of bat populations that will inform both local and more wide-ranging risk management practices involving horses and other animals at risk of Hendra virus infection. 3

13 Methodology Part A: Virus identification and discovery Virus detection was performed using both real-time and conventional PCR as well as next generation sequencing. Conventional PCR strategies were used to identify unknown viruses within field samples focussing on virus family specific primers against the paramyxovirus family as well as other families of known bat viruses including adenoviruses, reovirus and coronavirus. Genus directed primers were also utilised to identify paramyxoviruses in the rubulavirus and henipavirus genera. Luminex beadbased multiplex PCR assays were developed to enable rapid and efficient screening of large numbers of samples for a variety of specific target viruses including Hendra virus, Menangle virus, Nelson Bay virus and newly identified bat viruses including Cedar virus (henipa), Teviot virus (rubula), Grove virus (rubula) and Yeppoon virus (rubula). To determine whether genetic variation in Hendra virus or other viral co-infections play a role in Hendra virus spillover events, their genetic variation was examined by PCR and sequencing. Virus isolation techniques utilising primary P. alecto cell cultures routinely used at AAHL (Crameri et al., 2009) was used to identify known and unknown viruses in bat urine under the unique BSL4 facilities at AAHL. Part B: Metabolite discovery The second component of the project was the identification of potential metabolomic biomarkers from bat urine. This part of the project represents a proof of principle study using matching serum and urine samples collected from captive animals experimentally infected with Menangle virus, a bat-borne paramyxovirus. The nature of these samples allowed for a controlled analysis of the effects of viral infection on the serum and urinary metabolome of individual bats of known age, sex and viral status. A comparison with pooled urine samples collected from plastic sheets at the bottom of the bat cage was used to obtain information on the complexity expected from samples collected from wild bat colonies. Pooled samples were from adult females and captive born juvenile bats of mixed sex. Metabolomics was used to analyse and quantitate low molecular weight (<1,500 daltons) metabolites in bat urine using mass spectrometry techniques, providing a broad overview of the urinary and serum metabolome for correlation with viral infection. Metabolite extraction was carried out using standard ethanol techniques. Analysis of the serum and urinary metabolome was then performed using three major platforms; liquid chromatography mass spectrometry (LC-MS), gas chromatography-ms (GS- MS) and nuclear magnetic resonance spectroscopy (NMR) to provide a targeted and quantitative analysis. Multivariate statistical methods were used to provide a crude overview of differences between Menangle virus positive and Menangle virus negative samples. Marker metabolite identification, pattern recognition software and database searches were used to identify unique metabolic patterns associated with viral infection. As our captive bats are under constant environmental conditions, this analysis will provide insights into changes in metabolism in the absence of impacts from environmental or lifecycle events. 4

14 Results Part A. Virus identification Primers specific for 11 bat-borne viruses covering the paramyxovirus, henipavirus and rubulavirus families have been used in a luminex multiplex x-tag assay to screen urine samples collected from bat colonies during both peak and non-peak Hendra periods. A total of 500 urine samples across seven locations in Queensland (Cedar Grove, Boonah, Hervey Bay, Cairns) and New South Wales (Sydney, Alstonville, Nambucca Heads) were screened using this approach. There was a high prevalence of viral co-infections in urine samples that were collected in July and August 2011 which corresponded to the Hendra virus spike period. Both Hendra virus positive and negative samples were positive by PCR for up to seven of the eleven different viruses included in the x-tag assay. For example, 20 urine samples collected on the July 26, 2011 at Boonah in Queensland were tested by the x-tag assay, of these, 19 contained viral RNA from at least one of seven different paramyxoviruses. This indicates that within the animals responsible for the 20 pooled urine samples there were potentially 51 separate ongoing or convalescent virus infections. Data obtained from Boonah over 2011 and 2012 also shows a seasonal trend in the presence of viral RNA in bat urine which may be associated with environmental conditions such as temperature (Figure 1). Similar spikes in viral RNA during periods of high Hendra virus prevalence were observed at other locations in Queensland and New South Wales (Figures 2 and 3). Urine samples collected during the non-spike period rarely contained viral RNA for Hendra or other viruses. These results provide strong support for the presence of viral co-infections in bats and for the excretion of multiple viruses during periods of high Hendra virus prevalence. Virus isolation was attempted on 32 pooled urine samples from Cedar Grove across two time points which resulted in 10 isolations of 5 different viruses; 15 pooled urine samples from Boonah across three time points resulted in 37 isolations of 7 different viruses; 7 pooled urine samples from Nambucca Heads from one time point resulted in 5 isolations of 2 different viruses and 16 pooled urine samples from Hervey Bay across three time points resulted in 8 isolations of 5 different viruses. Virus sequencing resulted in the identification of four novel paramyxoviruses, all in the genus Rubulavirus. The isolate with closest sequence relatedness to the known bat rubulavirus Tioman paramyxovirus (TioPV) was designated Teviot paramyxovirus (TevPV). Menangle paramyxovirus (MenPV) was isolated from the 2009 Cedar Grove collection and previously reported (Barr et al., 2012). A novel rubulavirus, designated Yeppoon paramyxovirus (YepPV), was isolated initially from the Yeppoon 2009 collection. A virus most related to YepPV was subsequently isolated from Cedar Grove and designated Grove paramyxovirus (GroPV). The fourth novel rubulavirus was isolated in P. alecto kidney (PaKi) cells and designated Hervey paramyxovirus (HerPV). These results are described in a recently published manuscript and shown in Tables 1 and 2 (Barr et al., 2015). In many cases, multiple viruses were identified from a single pooled urine sample and in some cases in only one cell type. There was a correlation between the high success rate in virus isolation and Hendra virus prevalence consistent with the possibility that multiple viruses are excreted from bats during peaks in Hendra virus prevalence. The potential of the newly discovered viruses to cause disease in humans and other species awaits further investigation. Further work will also include an analysis of the prevalence of viral co-infections between different bat colonies across time to provide information on the spatial and temporal aspects of viral excretion. Four Hendra virus isolates from bat urine (collected from three different locations) and 13 isolates from horses (collected from 10 different locations) from were sequenced by 454 sequencing. Analysis of partial genome sequences revealed no significant changes in the bat or horse derived Hendra virus variants. This information confirms the stability of Hendra virus indicating that it is unlikely that changes in the Hendra virus genome were responsible for the spike in spillover events that occurred in

15 Figure 1. Bat urine samples collected in Boonah, Queensland from July 2011 to November The presence of viral RNA corresponding to multiple bat viruses provided evidence for high levels of viral co-infections during spikes in Hendra virus prevalence in July-August Boxed regions correspond to colder months. 6

16 Figure 2. Viral RNA was detected in bat urine samples collected from locations in Queensland (Cairns and Hervey Bay) and New South Wales (Alstonville, Nambucca Heads and Sydney). The presence of multiple viruses provides evidence for viral co-infections, particularly during spikes in Hendra virus prevalence. * indicate spikes in Hendra virus prevalence. Figure 3. Bat urine samples collected at Cedar Grove, Queensland from September 2009 to May * indicates a spike in Hendra virus prevalence. 7

17 Part B. Metabolomics Paired urine and serum samples collected from captive female flying foxes that were experimentally infected with the bat paramyxovirus, Menangle virus were used for a proof of principle study to examine the metabolic profile of bats in response to viral infection. These samples provided the opportunity to examine the metabolic profile of samples in individuals of known age, sex and infection status under controlled conditions. Individual urine and serum samples were collected pre- and postinfection with Menangle virus and subjected to mass spectrometry analysis. Matching urine and serum samples were collected from the 14 adult female bats. Pooled urine samples were collected from the cages pre-infection and included adult females and their captive born juvenile pups which were of both sexes. A total of 345 metabolites were detected from serum samples collected from captive bats pre- and post- infection with Menangle virus using an untargeted GS-MS approach. To obtain an overview of the effect of viral infection on the metabolite profile, a multivariate analysis using principal component analysis (PCA) was performed on the MS data matrix. The PCA showed that there was a separation of the uninfected and Menangle virus infected groups (Figure 1A). A similar analysis was performed on individual urine samples from the same group of captive bats pre- and post-menangle virus infection and from pooled urine samples collected from the cages pre-infection. As the pooled samples were from cages consisting of the adult females and their captive born juvenile pups, this provided information on the complexity expected from pooled samples from wild bat colonies. A total of 409 metabolites were detected in the bat urine samples. Similar to the pattern observed for the serum samples, there was a clear distinction between the profiles of infected and uninfected bats when they were subjected to PCA analysis (Figure 1B). A. B. Figure 4. Principal components analysis scores plots of (A) 345 metabolites measured using GC-MS in bat serum samples collected from captive bats pre- (blue) and post- (red) Menangle virus infection. (B) 409 metabolites measured using GC-MS in bat urine samples from captive bats pre-(purple) and post (blue) Menangle virus infection and from pooled urine pre-infection (green) In the serum samples sixteen metabolites differed between uninfected and infected bats, including five that were significantly different and are shown in Table 3. Although not significant using stringent statistical testing, isoleucine was identified as significantly different in pre- and post-infected bats using standard t-tests between the pre and post-infection samples. In the serum samples, eight metabolites that differed significantly between uninfected and infected bats were identified in the bat urine dataset (Table 3). Isoleucine was a commonly changing metabolite between the serum and urine samples. The metabolic profile obtained for the pooled urine samples was more similar to the preinfection samples providing evidence that this technique can be applied for the analysis of pooled urine samples from wild bat colonies. The information provided from this study will form the basis for 8

18 further work to examine the metabolic profile of wild bats. Further work will examine changes in the metabolic profiles of bats at different levels of Hendra virus prevalence and in response to environmental conditions including climate and food availability and lifestyle events such as mating, pregnancy and weaning. This work has the potential to identify triggers associated with increases in viral shedding and reveal metabolites with the potential to be used as biomarkers to predict the risk of spillover events. Table 1. Summary of all paramyxoviruses isolated from this study (from Barr et al., 2015) Genus (Proposed) Species Total No. Location (No. of times isolated at that location) of Isolations Henipavirus Cedar (CedPV) 1 Cedar Grove (1) Hendra (HeV) 8 Cedar Grove (1), Yeppoon (2), Tolga Scrub (1), Boonah (1), Hervey Bay (3) Rubulavirus Menangle (MenPV) 11 Cedar Grove (1), Boonah (9), Hervey Bay (1) Hervey (HerPV) 9 Hervey Bay (1), Boonah (5), Nambucca Heads (3) Grove (GroPV) 10 Cedar Grove (4), Boonah (6) Teviot (TevPV) 17 Cedar Grove (3), Boonah (12), Nambucca Heads (2) Yeppoon (YepPV) 8 Yeppoon (1), Hervey Bay (3), Boonah (4) 9

19 Table 2. Isolation detail of the four novel rubulaviruses (from Barr et al., 2015) (Proposed) Isolate # Location of CPE first observed Species Isolation Cell type Passage No. Hervey (HerPV) 1 Hervey Bay PaKi 1 2 Boonah PaKi 1 3 Boonah PaKi 1 4 Boonah Vero 1 5 Boonah Vero 1 6 Boonah PaKi 1 7 Nambucca Heads Vero 2 8 Nambucca Heads Vero 2 9 Nambucca Heads PaKi 2 Grove (GroPV) 1 Cedar Grove Vero 2 2 Cedar Grove PaKi 1 3 Cedar Grove Vero 1 4 Cedar Grove PaKi 3 5 Boonah Vero 1 6 Boonah PaKi 1 7 Boonah PaKi 1 8 Boonah Vero 1 9 Boonah Vero 1 10 Boonah Vero 1 Teviot (TevPV) 1 Cedar Grove PaKi 1 2 Cedar Grove Vero 1 3 Cedar Grove Vero 1 4 Boonah Vero 1 5 Boonah Vero 1 6 Boonah Vero 1 7 Boonah PaKi 1 8 Boonah Vero 1 9 Boonah PaKi 1 10 Boonah Vero 1 11 Boonah Vero 1 12 Boonah PaKi 1 13 Boonah Vero 1 14 Boonah Vero 1 15 Boonah Vero 1 16 Nambucca Heads Vero 2 17 Nambucca Heads Vero 2 Yeppoon (YepPV) 1 Yeppoon PaKi 2 2 Hervey Bay Vero 1 3 Hervey Bay PaKi 1 4 Hervey Bay PaKi 2 5 Boonah Vero 1 6 Boonah PaKi 1 7 Boonah Vero 1 8 Boonah PaKi 1 10

20 Table 3. Significant metabolites detected in bat urine and serum samples following Menangle virus infection. Metabolites were considered significant if the BH adjusted p value was <0.05. Serum Leucine Cholesterol Serine Lactic acid b-alanine Isoleucine* *statistically significant by t-testing Urine 5, oxoproline Asparagine Glycine Threonine Isoleucine Galactonic acid Mannitol 3,5-dihydroxybutanoic acid 11

21 Implications This project has implications for the development of assays to identify biomarkers associated with increases in viral replication in bats. The identification of viral co-infections including a number of previously uncharacterised viruses has implications for spillover of other viruses during peak periods of viral replication. Further monitoring will be necessary to determine the impacts of these viruses on other species. Unique metabolic profiles were associated with viral infection and can be measured in urine samples collected non-invasively from bat colonies. This study represents the first analysis of the urinary and serum metabolome of virus infected bats with the identification of a distinct profile associated with viral infection. This information will provide new opportunities for developing assays for use in the field to identify changes associated with viral infection that can be used to predict periods of high risk for a spillover event to occur. The ability to monitor bat colonies in the field has implications for veterinarians and horse owners. 12

22 Recommendations Multiple viruses were associated with spikes in Hendra virus prevalence in bats. As many of these viruses are currently uncharacterised and of unknown disease causing potential in other species, current recommendations around preventing Hendra virus spillover from bats should continue to be adhered to in order to avoid spillover of other viruses. Unknown viral infections of horses or livestock animals in areas associated with bats should be treated with caution and undergo further investigation. Our proof of principle analysis of serum and urine samples from captive, experimentally challenged bats supports the use of metabolomics for biomarker identification. Further validation of samples collected in the field will now be required to take this work to the stage where it can be used to identify markers associated with increased viral replication in bats. 13

23 References Barr, J., Smith, C., Smith, I., de Jong, C., Todd, S., Melville, D., Broos, A., Crameri, S., Haining, J., Marsh, G., Crameri, G., Field, H. & Wang, L.-F. (2015). Isolation of multiple novel paramyxoviruses from pteropid bat urine. Journal of General Virology 96, Barr, J. A., Smith, C., Marsh, G. A., Field, H. & Wang, L.-F. (2012). Evidence of bat origin for Menangle virus, a zoonotic paramyxovirus first isolated from diseased pigs. Journal of General Virology 93, Calisher, C. H., Childs, J., Field, H. E., Holmes, K. & Schountz, T. (2006). Bats: Important Reservoir Hosts of Emerging Viruses. Clinical microbiology reviews 19, Crameri, G., Todd, S., Grimley, S., McEachern, J. A., Marsh, G. A., Smith, C., Tachedjian, M., De Jong, C., Virtue, E. R., Yu, M., Bulach, D., Liu, J.-P., Michalski, W. P., Middleton, D., Field, H. E. & Wang, L.-F. (2009). Establishment, Immortalisation and Characterisation of Pteropid Bat Cell Lines. PLoS ONE 4, e8266. Leroy, E. M., Kumulungui, B., Pourrut, X., Rouquet, P., Hassanin, A., Yaba, P., Delicat, A., Paweska, J. T., Gonzalez, J.-P. & Swanepoel, R. (2005). Fruit bats as reservoirs of Ebola virus. Nature 438, Munger, J., Bennett, B. D., Parikh, A., Feng, X.-J., McArdle, J., Rabitz, H. A., Shenk, T. & Rabinowitz, J. D. (2008). Systems-level metabolic flux profiling identifies fatty acid synthesis as a target for antiviral therapy. Nat Biotech 26, Nicholson, J. K. (2006). Global systems biology, personalized medicine and molecular epidemiology. Mol Syst Biol 2. Peel, A. J., Baker, K. S., Crameri, G., Barr, J. A., Hayman, D. T. S., Wright, E., Broder, C. C., Fernandez-Loras, A., Fooks, A. R., Wang, L.-F., Cunningham, A. A. & Wood, J. L. N. (2112). Henipavirus neutralising antibodies in an isolated island population of african fruit bats. PLoS ONE 7, e Plowright, R. K., Foley, P., Field, H. E., Dobson, A. P., Foley, J. E., Eby, P. & Daszak, P. (2011). Urban habituation, ecological connectivity and epidemic dampening: the emergence of Hendra virus from flying foxes (Pteropus spp.). Proceedings of the Royal Society B: Biological Sciences 278, Sohayati, A., Hassan, L., Sharifah, S., Lazarus, K., Zaini, C., Epstein, J., Shamsyul Naim, N., Field, H., Arshad, S., Abdul Aziz, J., Daszak, P. & Alliance., E. (2011). Evidence for Nipah virus recrudescence and serological patterns of captive Pteropus vampyrus. Epidemiology and Infection 139, Vinayavekhin, N., Homan, E. A. & Saghatelian, A. (2009). Exploring Disease through Metabolomics. ACS Chemical Biology 5,

24