The fundamentals of zoonotic disease ecology are often poorly understood despite the fact that they can have serious public health consequences and are emerging with alarming frequency 
. In addition, threatened species, as well as overall biodiversity, can be negatively impacted by disease 
. This study incorporated data collected over a ten year period on 791 pumas, bobcats, and domestic cats, sampled across 5 study areas that varied in both ecosystem characteristics and degree of urbanization (). Data provide new and unanticipated findings about the distribution of three pathogens capable of infecting and being transmitted among three felid species whose ranges overlap, particularly along urban edges. This study revealed specific associations between variables of interest and exposure to pathogens, and found that transmission route was consistently associated with the variables driving exposure. Results
have implications for the routes in which emerging or invading pathogens could move within and between these species.
spp. exposure, as predicted, was generally higher in California locations when compared to Colorado locations. The initial prediction was based on the known association between C. felis
spp. known to infect cats, coupled with data on the relationship between flea distributions and climate 
. Potential arthropod vectors occur in higher numbers in regions with warmer temperatures and higher humidity
and these climate differences may drive the higher exposure levels seen in California felids. Domestic cats in California had substantially higher Bartonella
spp. exposure than non-domestic cats which could be related to high domestic cat densities in urban areas, leading to locally amplified flea populations and increased transmission opportunities. Published densities of domestic cats adjacent to our Ventura County site are much greater (2–3 orders of magnitude) than non-domestic cats 
, a situation that is potentially similar in other urban sites as well. Previous research has demonstrated that Bartonella
seroprevalence in domestic cats in the Los Angeles region is even higher than predicted based on flea prevalence estimates 
, and while flea and host species richness is higher in non-urban areas, flea infestation was higher in urban/disturbed sites with low vector richness 
. In support of these previous findings, we observed higher Bartonella
spp. seroprevalence among domestic cat populations nearer large urban centers in California (i.e.
Orange County, CA and Ventura County, CA). Despite vector research that supports the felid seroprevalence patterns reported here, flea data was not collected during the course of this study. Future flea collections, along with sequencing of bacterial isolates from fleas and feline blood, are needed to definitively relate host seroprevalence to differences in vector abundance and distribution. Additional vectors, such as ticks, should be examined as well based on evidence that other vectors could be involved in Bartonella
spp. transmission 
. These more detailed data might also help reveal why bobcats from the rural Colorado Western Slope had substantially higher Bartonella
spp. exposure when compared to sympatric pumas and domestic cats.
has received substantial attention because it is a ubiquitous pathogen with significant zoonotic potential 
. Studies have primarily focused on human infection and disease, but links to marine mammal population declines have also been documented 
. The primary factors associated with T. gondii
seroprevalence in this study were the same factors associated with Bartonella
spp. exposure: age, location, and species. The similarity between the models for the two pathogens may be related to the fact that exposure for both is typically indirect (i.e.
, vector transmission for Bartonella
and ingestion of infected prey or environmental contact for T. gondii
) rather than direct contact between conspecifics. Despite both pathogens having similar predictors in the final models, the overall seroprevalence patterns differed. Pumas had higher T. gondii
seroprevalence across all regions compared to bobcats, suggesting pumas have increased exposure to the pathogen. This could be reflective of factors including: (1) larger home range and spatial scales of pumas, resulting in increased exposures; (2) a larger dietary intake, resulting in a greater opportunity for ingestion of prey species with encysted toxoplasmosis intermediate forms; (3) a diet that consists of prey species with higher T. gondii
exposure; or, (4) increased susceptibility of pumas to T. gondii
infection. Dual exposure to both T. gondii
and FIV was substantially higher in pumas compared to both bobcats and domestic cats. This is likely simply related to both T. gondii
and FIV being independently common and so they are more likely to co-occur, but is also possible that there may be an interaction between the two pathogens that increases opportunities for infection, or that some aspects of puma behavior are “risky,” leading to increased exposure to both pathogens.
Domestic cats, despite the attention they generate as a source of T. gondii
infections in households, had extremely low seroprevalence overall to this parasite. This is supported by data from several countries that show human exposure to T. gondii
is often related to the meat supply and meat consumption, rather than exposure to domestic cats 
, and is likely reflective of the limited ingestion of intermediate host species around urban areas by domestic cats versus their free-ranging relatives. It is possible that stray and/or feral domestic cat populations consisted of disproportionately younger animals, and that this contributed to the lower than expected T. gondii
seroprevalence reported here 
. The higher seroprevalence in non-domestic felids suggests that T. gondii
could represent a scenario where disease exposure is higher in undeveloped areas and along an urban edge versus developed, urban areas. A recent study modeling T. gondii
transmission via environmental contact or intermediate host suggested that urban transmission was more dependent upon environmental exposure whereas suburban/rural spread was more dependent upon ingestion of intermediate hosts 
. A similar model of Baylisascaris procyonis
transmission in raccoons (Procyon lotor
) also predicted differences in intermediate host exposures in urban versus rural landscapes 
. The highest T. gondii
seroprevalence levels seen in this study were in non-domestic felids from rural study areas, although substantial exposure was also seen in urban Orange County, CA pumas and bobcats. Our data would therefore support a robust means of sylvatic infection (resulting in high infection rates of pumas and bobcats outside of urban centers) with a less efficient exposure in urban settings (as reflected by low exposure in feral domestic cats). The broad spatial scales examined confound a detailed analysis of pathogen exposure in relation to urbanization and warrant further study on finer spatial scales.
In contrast to bartonellosis and toxoplasmosis, FIV is transmitted horizontally through direct contact, especially aggressive interactions or mating. Location was associated with Bartonella spp. and T. gondii seroprevalence, but was not a predictor of FIV exposure in any species. This does not necessarily rule out spatial heterogeneity in FIV exposure or focal regions where FIV is more prevalent, but such associations were not detected here. The FIV model explained less variation than the models for the other two pathogens and it is possible that this unexplained variation could be related to localized clusters that were not accounted for in the analysis. Future analysis on finer spatial scales may help to pinpoint additional sources of variation.
While not a strong predictor, males were more likely to be infected with FIV than females in all three species. This pattern is consistent with previous research 
, which has related the frequency of aggressive interactions among male felids to increased FIV transmission opportunities. Species-specificity of FIVs has been well-documented 
and we noted substantial differences in FIV seroprevalence among bobcats, pumas, and domestic cats. Greater genetic diversity among non-domestic FIV strains suggests that domestic cat FIV emerged relatively recently whereas FIV of wild felids has been established for a longer period of time 
. In this study, FIV exposure was consistently lower in domestic cats compared to non-domestics, which is consistent with other serosurveys of domestic cats in the US 
. While domestic cat sample collection focused on feral and stray animals, it is likely that some domestic cats were “owned” at some point and those conditions, which could include neutering and some degree of isolation from other cats, resulted in decreased FIV exposure. Interestingly, the most common pathogen combination in domestic cats was exposure to both FIV and Bartonella
spp. While previous research has found an association between Bartonella
spp. in felids and another feline pathogen, Feline Leukemia Virus 
, it is also possible that the association found here simply reflects similar risk factors.
In summary, this analysis suggests that environment, species, and individual behaviors are important factors in disease occurrence, and that anthropogenic influences may alter pathogen structure in wild populations. Conversely, wild populations of felids appear to be important reservoirs for some highly prevalent human diseases. Diseases which spread within species – such as FIV – are less likely to be influenced by geographic location and are more likely dependent upon individual behaviors and intra-specific encounters. Pathogens that are spread by vectors, like Bartonella spp., are more likely to occur in regions supporting vector success and vector exposure to the pathogen, though domesticated animals may serve as focal bioaccumulators that could impact prevalence among susceptible wildlife. Pathogens transmitted by environmental contamination or ingestion of infected prey (i.e. toxoplasmosis) can have broad regional associations, and our studies provide evidence that T. gondii exposure is substantial in rural settings, and therefore wildlife may serve as reservoirs for domestic animals and/or human toxoplasmosis when at the urban-wildland interface.
Domestic cat densities are higher in urban areas 
, while puma and bobcat populations can decrease as a consequence of urbanization and habitat isolation 
. Wild felids isolated by habitat fragmentation exhibit “home-range pile-up” 
and the potential for increased contact rates with conspecifics. Contact between domestic cats and non-domestic felids can lead to cross-species FIV transmission events, such as has apparently occurred with feline leukemia virus transmission between domestic cats and pumas in Florida 
. Additionally, recent spatial analyses have suggested that landscape features, and in particular roads, could impact FIV infections in pumas 
. Additional studies will focus on use of chronic pathogen genetic signatures to trace wildlife movement in fragmented landscapes and predictive modeling for fine scale analysis of diseases in carnivores impacted by human development.