Many avian species are common residents of cities and suburban environments in Europe. Domestic chicken, which are predominantly kept at rural farms, can also be encountered in suburban backyards of cities worldwide. This may be due to an existing tradition of backyard farming in some countries, while more recently, the trend of keeping chickens by private individuals was inspired by the popularity of organic/free-range eggs and meat (1–3). Ornamental duck species, such as the mandarin and wood duck (Aix galericulata and Aix sponsa), have been purposely introduced in European parks and estates with suitable freshwater habitats. While the mandarin duck was not endemic when it was first introduced, it has managed to establish breeding populations in the Netherlands and has been sighted in the wild as far south as Algeria recently (4, 5). The close proximity of non-endemic to endemic avians, both wild and domestic, facilitates the possibility of physical interaction and provides opportunity for pathogen transmission. Birds raised extensively or in backyards are especially vulnerable, as confinement and isolation are usually not practiced. One of the greatest global health challenges in the 21^st century thus far has been the emergence of ‘new’ diseases, often zoonoses originating from wildlife, along with increased geographical spread of diseases (2, 6). Birds can transmit a variety of pathogenic bacteria, viruses and parasites to other birds, but also to mammals, including humans, thus playing a role in various transmission chains. Some of the most notorious zoonotic pathogens directly transmitted by birds to humans via ingestion or inhalation are the bacteria Salmonella typhimurium, Campylobacter jejuni, the Influenza A viruses, as well as Cryptococcus neoformans, a pathogenic fungus and the Apicomplexan parasites Cryptosporidium meleagridis and Toxoplasma gondii. Of these, only cryptococcosis may be considered an emerging disease, associated with immunodeficiencies (HIV), transplantation and immunosuppressive therapy (7, 8), while the other pathogens are responsible for ‘old’ diseases. Systematic active surveillance of bird reservoirs is in place only for the bacterial pathogens and avian influenza, while detection of C. neoformans, C. meleagridis and T. gondii occurs primarily through scientific investigations. Of the three, infection with T. gondii is associated with the highest disability adjusted life-years (DALY) (9). T. gondii infections occur globally with variable prevalence in humans and animals. The mode of infection is always ingestion, and as humans, animals and even the environment are reservoirs of the parasite, epidemiological research on T. gondii requires a One Health approach. The parasite’s life cycle is complex with any warm-blooded species acting as intermediate hosts and members of the family Felidae as definitive hosts. The role of cold-blooded hosts in the life cycle of the parasite is still unclear, despite the apparent success of short-term replication of T. gondii tachyzoites in some cold-blooded species after experimental infection and detection of parasite DNA in others (10, 11). There are three infectious life forms of T. gondii, the free living tachyzoite and two encysted forms: the bradyzoite (in tissue cysts) and the sporozoite (in oocysts) (12). The parasite can switch between tachyzoite and bradyzoite interchangeably in all hosts, but only in the definitive hosts can tachyzoites differentiate into intermediates required for sexual reproduction, which results in the formation of oocysts (13). Oocysts are expelled into the environment from the digestive tract of the definitive host where they undergo a process of maturation, under permissive environmental conditions, to yield sporocysts with sporozoites. Sporocysts are characterized by high tolerance of environmental conditions, with the capacity to remain infectious for extended periods of time in soil of various chemical compositions and also in fresh and marine water (14–16). T. gondii infection can occur after ingestion of tissue cysts in undercooked or raw meat, oocysts in mollusks, or on vegetables and fruits, as well as oocysts from untreated water, commonly from surface sources. Host to host transmission occurs dominantly via predation and on occasion vertically, while the frequency of environmental transmission is likely variable, depending on a multitude of factors (17). Avian species, similar to mammals, are exposed to multiple sources of infection, due to a highly variable diet and feeding habits. As with mammals, susceptibility to toxoplasmosis varies between avian species, thus, while domestic chicken handle the infection quite well (18), for others it can be lethal, even leading to near extinction in extreme cases, as with the ‘alalā (Corvus hawaiiensis) (19). High susceptibility to infection can reduce the host’s efficiency to transmit the parasite and vice versa, thus the capacity to maintain the parasite in the ecosystem differs by host biodiversity and population. In terms of spreading the parasite, unlike terrestrial species, the ability to fly allows birds to overcome a number of natural and anthropogenic environmental barriers, while migratory species are an exceptional category due to the additional ability to traverse vast distances over land and water. Thus, birds may be instrumental in disseminating and introducing T. gondii to remote, inaccessible and isolated ecosystems. A recent review and meta-analysis calculated the pooled global prevalence of T. gondii infection in birds to be 25% (23–28%), and 9% (4–17%), using serological and molecular detection methods, respectively (20). The lowest pooled prevalence (by all methods combined) of 16% (12–20%) was determined in Europe. Another study, by Wilson et al. (21), which examined seroprevalence in 24,344 wild birds from all continents except Australia, determined that omnivorous terrestrial species had the highest seroprevalence, while strictly carnivorous species came in second. In addition, the seroprevalence was greater in terrestrial as compared to aquatic species, while dietary preferences of aquatic species had a negligible impact on seroprevalence (21). The aim of this study was to assess the occurrence of T. gondii infection in different species of birds living in habitats with varying degrees of urbanization and human population density. Additionally, possible environmental (water) and animal reservoirs of infection relevant for birds were examined for the presence of parasite gDNA.

This is the first report on the molecular occurrence of T. gondii infection in wild birds in Serbia which was determined to be 24.1%, nearly threefold higher than the average pooled global molecular prevalence in birds in the last two decades (20). Of the 16 species analyzed (Figure 1), the only domestic species were backyard chickens, which also had the highest occurrence of infection, 41.4% for a single species, while in wild species it was in rooks, 32.1% (Table 1). Previously, we determined seroprevalence of infection by modified agglutination assay (MAT) in pheasants (66.6%) and city dwelling feral pigeons (7.1%) (unpublished data). While it is not possible to directly compare sero- and molecular prevalence, due to fundamentally different approaches to detection—indirect (antibodies in sera or meat juice by various tests) and direct (PCR based detection of the parasite gDNA), the results suggest that infection is more frequent in pheasants as compared to feral pigeons, which was recapitulated by this study. Unlike the findings of Wilson et al. (21), based on a large global set of avian species and individuals, there were no statistical differences in the occurrence of infection based on diet or between terrestrial versus aquatic species (Table 4). However, this study found a higher molecular occurrence of infection in birds that are omnivorous (Corvidae, both species taken together, n = 47) as compared to strict carnivores (owls and raptors, combined, n = 20) with 31.9 and 20%, respectively. In recent years, several studies were published on the prevalence of T. gondii in wild bird species in Europe, mostly from southern Europe. The greatest number of individuals analyzed were of Falco tinnuculus (Eurasian kestrel) in Italy (n = 167 in 2016, n = 91 in 2017), with seroprevalences of 33.3 and 14.3% in 2016 and 2017, respectively (25), followed by Columba livia (feral pigeons) from Spanish zoos (n = 142), with a seroprevalence of 9.2% (26). Next, Buteo buteo (common buzzard) from France with a seroprevalence of 51% (27), Ciconia ciconia (white stork) from Portugal with a seroprevalence of 31.4% (28), Pica pica (Eurasian magpie) with 15.2% from Spain and Anas indicus (bar-headed goose) from Spanish zoos, with a seroprevalence of 7.1% (26, 29). In terms of molecular prevalence, in common buzzards from Turkey it was 92% (30), 18.75% in Spatula clypeata (northern shoveler) and in 6.25% Anas crecca (Eurasian teal) from Italy (31) and finally, merely 1.85% in white storks from Poland (32). Taken together, the published data shows that there is a significant divergence in infection prevalence, making any firm conclusions regarding the true epidemiological significance of birds in terms of toxoplasmosis difficult to reach. It is equally difficult to ascertain whether the frequency of infection correlates fully with diet alone. Perhaps this is due to relatively few individuals analyzed per species, as compared to the large set used by Wilson et al. (21). Indeed, the number of individuals of many wild species in a number of studies rarely exceeds 10, with 1–3 being most common. An inevitable limitation in terms of sampling wild birds is ‘convenience’ sampling of blood and tissues of game and birds in rehabilitation or rescue centers (such as in this study), and particularly in zoos, which may bias the results by over- and underestimating infection occurrence. Thus, even a relatively low occurrence of infection, 7.7% in owls as determined here, or the white storks from Poland (32), may confer a high capacity to transmit T. gondii to other species in environments in which there are large numbers of these birds. When considering human infection, wild game, such as pheasants and mallards, with a combined frequency of infection of 26% are relevant as reservoirs, however, it is unclear how probable or frequent infections actually are, as opposed to infections resulting from domestic species, like chickens, which are far more commonly consumed. Chickens are also efficient bioindicators for the contamination of the environment with oocysts, due to their feeding habits (pecking off the ground) (33). The occurrence of infection in chickens in Europe varies, with the greatest overall detected in backyard chickens 47% (34) in Germany and 41.2% in Greece (35), followed by free-range chickens, 3.2% in Greece (35), 5.6% in Portugal (36), and 18% in Ireland (37), and the lowest in caged chickens, 1.7% in Greece (35), 3.7% in Germany (34), and 0.4% in the Czech Republic, as expected (38). Interestingly, these studies show that even though free-range chickens are also exposed to the environment, and possibly oocysts, their seroprevalence is lower as compared to backyard chickens. The results of this study showed a statistically significant difference between the occurrences of infection in those from rural backyards, 80%, as opposed to 21% from peri-urban backyards (Table 4). However, given the relatively low number of samples analyzed here, it is unclear whether this variability in occurrence of infection is a general trend and how a country-wide pooled occurrence may compare to the published data. However, if variability of infection occurrence is inherent, a pooled prevalence in backyard chickens on a country-wide scale may not provide relevant information with respect to the true infection risk for humans. In terms of the significance of avians as reservoirs of infection for other animals, the key host to consider are cats (owned, feral, wild) which facilitate environmental transmission of T. gondii. It has been suggested that chickens are an efficient source of infection for cats (18), but it has also been estimated that cats kill anywhere from 136 million (Poland), 64.8 million (UK), and 1.73 million (Finland) wild birds annually and that ‘death by cat’ is the leading cause of mortality for garden birds in France and Belgium (39). With respect to domestic cats, which are most numerous worldwide, the above findings suggest that more relevant reservoirs of T. gondii may be wild birds living in cities and/or close proximity to urban areas. The molecular occurrence of infection between wild species residing in the urban, peri-urban and rural areas was not statistically different, 22.4, 27.3, and 22.7%, respectively (Table 4), suggesting equal opportunities for infection of other animals in these areas. The most numerous birds in the city (n = 79), all from the residential neighborhood of Surčin, were Corvidae (rooks and hooded crows) and Columbidae (wood pigeons), with a combined molecular occurrence of infection of 25.3%, indicating a fairly high probability for transmission. However, infection was detected in 15.6% of the wood pigeons, as opposed to 32.1% of rooks and 31.6% of hooded crows. As both Corvidae and Columbidae are suitable as cat prey, the results point to the conclusion that despite a high combined occurrence of infection, cats are far more likely to become infected when preying on Corvidae. To analyze potential animal and environmental T. gondii reservoirs in the residential neighborhood of Surčin, we collected rodents from local exterminators and Sava River water. Based on available information, the mice were collected from rodent traps which were set around small farms in close proximity to each other. The high molecular occurrence of infection (56.2%), may therefore be explained in part by the fact that mice live in groups on small territories, thus providing the opportunity for a relatively large number of them to become infected by a locally occurring source of T. gondii oocysts or tissue cysts. It stands to reason therefore that carnivorous or scavenging birds with habitats/feeding grounds within the mouse territory as well as cats living in the area have a fairly high probability of infection. T. gondii DNA was detected several years ago in the Danube River in water sampled in another residential neighborhood of Belgrade, Zemun (24). The locations on the Sava sampled here are on both banks of the river, several hundred meters apart. T. gondii DNA was detected at three locations of the four sampled. Although systematic sampling of the Sava and Danube in the Belgrade metropolitan area has not been done to date, the combined results point to the fact that oocyst contamination may be fairly high in some locations. The findings here point to a local source of oocysts, likely to be environmental rather than animal, as the sampling was done on different days. As much of the area of Surčin is within the river’s floodplain, the oocysts could have originated from the surrounding soil. Along with natural reservoirs, anthropogenic sources (waste) may also contribute to infection in omnivorous species and may in part explain the relatively high molecular occurrence of infection in hooded crows, rooks and mice. Based on the currently published data, including this study, the significance of avians in the epidemiology of toxoplasmosis is still unclear. It appears that species living in close proximity to humans in urban areas are more likely to be preyed upon by domestic cats (owned and feral), thus even a low prevalence of infection is sufficient to maintain the parasite in these environments. As far as parasite dissemination is concerned, it is not entirely clear whether migratory or sedentary birds have a greater capacity to spread T. gondii. Migratory birds may cover great distances, yet migrations occur twice annually, while some sedentary species, like rooks and wood pigeons, visit different feeding grounds on a daily basis. Due to the frequency of localized travel, sedentary birds may be more likely to facilitate spread on a smaller geographical scale. The city dwelling sedentary species analyzed here share their habitats/feeding grounds not only with cats but also other carnivorous wildlife, which increases the likelihood for transmission and spread of T. gondii. Our results further suggest that perhaps Corvids are important reservoirs of the parasite in urban environments due to a relatively high molecular occurrence of infection. Corvids may even be useful as bioindicators for the presence of T. gondii in urban environments, due to their population size and omnivorous feeding habits. Our findings did not support a strong dependency of the occurrence of infection on diet, when analyzed as a sole risk factor, possibly due to the low number of strictly carnivorous avian species (owls and raptors). Our analysis of the natural sources of infection however suggests that on a local scale, diet must be considered in context of the species ecology and habitat characteristics and availability of anthropogenic food sources in a particular area. Approximately half of the 11 species of city dwelling birds analyzed here originated from Surčin, which features a mosaic of natural wetlands, fishponds and forests along with agricultural plots, gardens and orchards which provide shelter, nesting and breeding habitats as well as a variety of food sources. Altogether, this resilient matrix of habitat and resources supports the increase in population size and biodiversity of avian species in close proximity to a highly urbanized area. With respect to T. gondii, this ecological setup facilitates environmental and host-to-host transmission and spread to other avian species, terrestrial wildlife and humans, thus feeding parasite strains through both the ‘wild’ and ‘domestic’ cycles. The importance of avian hosts in terms of the local population genetics of T. gondii remains to be investigated.