UCLA

Many pathogens, and all zoonotic pathogens, are capable of infecting multiple hosts. Understanding the transmission dynamics of multi-host, generalist pathogens is a major frontier in disease ecology, with far-reaching implications for both animal and human health. A central challenge is that multiple lines of evidence are required to identify maintenance hosts and assess the relative transmission contributions of multiple species. This evidence can unfortunately be difficult to obtain, particularly in wildlife systems, due to resource limitations and low sample accessibility. To address these challenges, I utilize computational tools to maximize the insights gained from limited wildlife data, using the globally significant zoonotic pathogen Leptospira interrogans in California’s coastal wildlife as a case study. Leptospira interrogans serovar Pomona presents a unique long-term case study of multi-host, generalist pathogen dynamics in California wildlife. This pathogen has affected the California sea lion (Zalophus californianus) population for decades, causing low levels of infection year-round and recurrent cyclical outbreaks of disease every few years. More recently, Leptospira has been found to be circulating among terrestrial wildlife in the California Channel Islands as well. Despite multi-year surveillance of the bacteria in California sea lions (Zalophus californianus) and Channel Island wildlife, the mechanisms governing transmission and persistence in this system are still unclear, as are potential connections to circulation of Leptospira among mainland wildlife host species. In the following chapters, I investigate the transmission potential and prevalence of Leptospira interrogans in California’s coastal wildlife. Within ecological systems, infection prevalence is critical to understanding pathogen dynamics, as it reflects transmission risk to others. However, uncertainty in the accuracy of diagnostic assays makes prevalence estimation difficult, particularly in wildlife where test methods are often not validated and sample sizes may be low. Bayesian latent class analysis (BLCA) offers a statistical solution to this problem, but research detailing its limitations and usefulness in biological systems is lacking. In my first chapter, I estimate disease prevalence and diagnostic test accuracy using simulations to assess the ability of BLCA to produce accurate estimates across a range of biological conditions. I demonstrate that this method is effective, but has the potential to bias estimates depending on underlying biological system traits (e.g., sample size, test accuracy, and true prevalence). I use the California sea lion system as a case study to assess infection prevalence and test accuracy, describing situations in which this method would be preferable to results from a single high quality diagnostic test. Our findings directly benefit scientists and veterinary professionals working on the California sea lion system, and, more generally, they validate a statistical tool and show ecologists when this technique may be of use. In chapter two, I develop models to predict Leptospira shedding, indicative of transmission potential, in California sea lions. Shedding can be detected via polymerase chain reaction (PCR) of urine or kidney samples to identify Leptospira DNA, but obtaining these samples is difficult, and historical data are limited. Antibody titers were previously identified as predictive of shedding in this species, but antibody results take time and are not always available. I utilized LASSO regression to assess if shedding predictions from antibody titers improve in the presence of additional environmental, clinical, and demographic data. I then exclude antibody results to identify more accessible data that are predictive of shedding in their absence, and show that these predictions are robust to differences in the underlying sample population. Extrapolations to out-of-sample data provide accurate shedding estimates in the broader sea lion population, providing key information for understanding Leptospira transmission and persistence in California sea lions. Understanding multi-host pathogen dynamics requires identification of possible hosts and the assessment of pathogen prevalence and transmission links in relevant host species. A closely related strain of Leptospira interrogans serovar Pomona has been identified in California sea lions, Channel Island foxes (Urocyon littoralis), and island spotted skunks (Spilogale gracilis amphiala), but it is unknown if mainland coastal wildlife play a transmission role in this multi-host pathogen system. In my final chapter, I conduct the first extensive survey of Leptospira in Southern California wildlife, using serology to investigate possible links to Leptospira in sea lions. Sampling primarily focused on five core species in the greater Los Angeles region: coyotes (Canis latrans), raccoons (Procyon lotor), Virginia opossums (Didelphis virginiana), striped skunks (Mephitis mephitis) and fox squirrels (Sciuris niger). Infections were detected in all core species except fox squirrels, and all five species exhibited Leptospira exposure and were reactive to serovar Pomona. This evidence of widespread Leptospira circulation demonstrates a potential risk to both animal and human health across the Los Angeles region, and animals with primary reactivity to serovar Pomona represent future sampling targets to assess possible transmission links in the broader multi-host system. Using statistical techniques to analyze multi-year surveillance data from California sea lions and mainland terrestrial mammals, I address critical knowledge gaps in the ecology of Leptospira in the coastal California ecosystem. Maximizing the information gained from limited data allows us to better understand the local prevalence and transmission ecology of this globally significant zoonosis, with direct applications for public health and wildlife management. Extending these methods to other systems will empower future multi-host pathogen studies, addressing key challenges in wildlife disease surveillance and ecology.