Copyright is owned by the Author of the thesis. Permission is given for a copy to be downloaded by an individual for the purpose of research and private study only. The thesis may not be reproduced elsewhere without the permission of the Author. A thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Veterinary Science at Massey University, Palmerston North, New Zealand. Marie Moinet 2020 © Marie Moinet 2020 Abstract Leptospirosis is an important zoonosis in New Zealand where it has historically been associated with livestock. Formerly negligible in human cases notified, Leptospira borgpetersenii serovar Ballum—associated with rodents and hedgehogs (Erinaceus europaeus)—is now preponderant. The role of wild introduced mammals in the epidemiology of leptospirosis has been overlooked in New Zealand but remains a critical question. In this thesis, we determined the prevalence of Leptospira serovars, renal colonisation and seroprevalence in wild mammals and sympatric livestock. During a cross- sectional and a longitudinal survey, house mice (Mus musculus), ship rats (Rattus rattus) and hedgehogs were trapped in farms with a history of leptospirosis to collect sera and kidneys. Urine and sera from livestock (dairy or beef cattle, sheep) and dogs were also collected on the same farms. Sera were tested by microagglutination test to identify serovars/serogroups that circulate in wildlife for comparison with those circulating in livestock. Urine and kidney samples were used to determine prevalence by qPCR, to isolate circulating leptospires by culture and subject them to whole genome sequencing, in order to determine their phylogenetic relationships and compare them to other sequences locally, nationally and internationally. Capture-mark recapture (CMR) methods were used to investigate the population dynamics of mice naturally infected with Ballum. Finally, the level of lesions and bacterial load in kidneys were assessed visually by histopathology and put in perspective with other results to investigate reservoir dynamics. Direct or indirect presence of Ballum was found in all wild and domestic species investigated. Overall apparent prevalence in mice, rats and hedgehogs was respectively 46%, 95% CI [39, 52%], 44% [26, 62%] and 27% [11, 50%]. It varied greatly between seasons in mice, with a spring peak (83 to 86%) and minimum in autumn (31 to 37%). Mice densities reached up to 56 mice/ha and varied seasonally in the opposite way, resulting in a relatively constant density of infected mice, ranging 3-8 infected mice/ha. An extremely low rate of mutations hindered the investigation of transmission pathways using genomics. However, despite little or no lesions in all species, the bacterial load was markedly higher in mice, suggesting rats and hedgehogs are secondary hosts. Control strategies to mitigate exposure to Leptospira in NZ should include wild mammals, and especially mice. i Résumé La leptospirose est une zoonose grave en Nouvelle Zélande et y a historiquement été associée au bétail. Autrefois négligeable dans les cas humains notifiés, Leptospira borgpetersenii serovar Ballum—associé aux rongeurs et aux hérissons (Erinaceus europaeus)—est désormais prépondérant. Le rôle des mammifères sauvages envahissants dans l'épidémiologie de la leptospirose a été ignoré en Nouvelle-Zélande mais reste une question critique. Dans cette thèse, nous avons déterminé la prévalence des sérotypes de Leptospira, la colonisation rénale et la séroprévalence dans la faune sauvage et le bétail sympatrique. Au cours de deux études transversale et longitudinale, des souris domestiques (Mus musculus), des rats noirs (Rattus rattus) et des hérissons ont été capturés dans des fermes à haut risque de leptospirose pour collecter du sérum et des reins. De l'urine et du sérum de bétail (bovins laitiers ou allaitants, ovins) et de chiens ont également été collectés dans les mêmes fermes. Les sérums ont été testés par test de microagglutination pour identifier les sérovars/sérogroupes qui circulent dans la faune sauvage et les comparer avec ceux qui circulent dans le bétail. Des échantillons d'urine et de rein ont été utilisés pour déterminer la prévalence par qPCR, pour cultiver et isoler les leptospires circulants et séquencer leur génome, afin de déterminer leurs relations phylogénétiques et de les comparer à d'autres séquences à l’échelle locale, nationale et internationale. Des méthodes de capture-recapture ont été utilisées pour étudier la dynamique des populations de souris naturellement infectées par Ballum. Enfin, le niveau de lésions et la charge bactérienne dans les reins ont été évalués visuellement par histopathologie et mis en perspective avec d'autres résultats pour étudier la dynamique de réservoir. La présence directe ou indirecte de Ballum a été trouvée dans toutes les espèces sauvages et domestiques étudiées. La prévalence apparente globale chez la souris, le rat et le hérisson était respectivement de 46%, 95% IC [39 – 52%], 44% [26 – 62%] et 27% [11 – 50%]. Elle variait fortement d'une saison à l'autre chez la souris, avec un pic printanier (83 à 86%) et un minimum en automne (31 à 37%). Les densités de souris ont atteint jusqu'à 56 souris/ha et ont varié de façon saisonnière dans le sens opposé, résultant en une densité relativement constante de souris infectées, allant de 3 à 8 souris infectées/ha. Un taux de mutations extrêmement faible a entravé l’étude des voies de transmission basée sur la génomique. Cependant, malgré peu ou pas de lésions chez toutes les espèces, la charge bactérienne était nettement plus élevée chez la souris, ce qui suggère que les rats et les hérissons sont des hôtes secondaires. Les stratégies de contrôle pour atténuer l'exposition à Leptospira en Nouvelle-Zélande devraient inclure la faune sauvage, et en particulier la souris. i Dedication To Yves, Jean-Baptiste, and Liam, my wise, kind and mischievous angels Acknowledgments This quote from Gertrude Stein seems appropriate to begin this section: “Silent gratitude isn’t very much use to anyone”. Science nowadays is seldom a one-(wo)man job, and it is particularly true for this project that linked together a diversity of disciplines and skills. Thank you, Jackie Benschop, for inviting me to carry out this research work, and for your continuous and kind help and guidance. Your leadership style has been very instructive. I would also like to express my gratitude to all my co-supervisors and advisors. To Peter Wilson, who was involved in the early stages of this project and encouraged me to take a scientific approach in building the outline of Chapter 2. To David Wilkinson, who literally and figuratively ‘wet his shirt’ to help me throughout this project, and without whom the Chapter 5 would still be in a development stage. I learnt tremendously by your side and appreciated your great skills in bioinformatics, vital support in ℝ coding and Linux debugging and kind guidance in best genomics methods to apply. You are a terrific pedagogue. To Emilie Vallée, who was always available for comments and help and gave support and guidance on statistical methods in Chapter 3 and participated in the development of the Bayesian Latent Class Models. To James C. Russell, my reference in ecology, who lent me Longworth traps to catch mice, helped with the conceptualisation of the field work and welcomed me at the University of Auckland to help with the development of the spatially explicit capture recapture models in Chapter 4. To Cord Heuer, who challenged my understanding of epidemiological methods and greatly helped me in building the structure of Chapter 6. To Julie Collins-Emerson, who was my reference for all work related to laboratory methods and Leptospira throughout this project. Thank you for your great availability and diligent comments. To Danielle Aberdein, who trained me in the recognition of renal histopathological lesions and provided guidance at all stages of development of the Chapter 6. The research project described in this thesis was conducted thanks to the financial support of various funders and I would like to acknowledge their support. The main source of funding for this project was a large portion of the Massey University research medal team award that my colleagues from the mEpiLab and Epicentre jointly won in 2013. Without the excellence in their work being acknowledge through this award, the present work would not have been possible. The costs of field operations were covered thanks to donations from the Southern Rangitikei Veterinary Services and from v the Wairarapa Veterinary Association. I would also like to acknowledge the financial support of Hawkes Bay Medical Research Foundation and Massey University School of Veterinary Science Postgraduate Funds that covered a part of the costs of laboratory analyses and whole genome sequencing. I am very thankful to the Kathleen Spragg Agricultural Research Trust for their generous travel award that allowed me to present my results at the European Leptospirosis Society Meeting at Alghero, attend the Mathematical Biology Modelling days of Besançon in 2018, and visit the Leptospirosis Reference Centre in Amsterdam University Medical Centre and the Unit of Biology of Spirochetes at the Pasteur Institute in Paris. The knowledge I acquired during that travel in Europe was invaluable for the rest of my thesis project. Thanks to Marga Goris, Ahmed Ahmed, Mathieu Picardeau and their teams for welcoming me so warmly. My thanks expand to Professor Angeli Kodjo and his team at the Leptospires Laboratory in VetAgro Sup Lyon who hosted me for a few days at the very beginning of this project in 2016. Thanks to the Massey University Institute of Veterinary, Animal and Biomedical Science Postgraduate travel fund, I was able to attend the Wildlife Disease Association – Wildlife Society of NZVA joint conference in Christchurch in 2016 and a workshop on Spatially Explicit Capture Recapture Methods in Melbourne in 2017. Both were very useful in understanding the context of wildlife diseases surveillance in New Zealand and having a better understanding of statistical methods used in Chapter 4.