Trichomonas Infection in a Community of Free-Ranging Domestic and Wild Columbiformes and Bonelli's Eagle (Aquila fasciata)

Trichomonas gallinae is a pathogen of conservation relevance, whose main maintenance hosts are Columbiformes, but spillover to avian predators has been described. The goal of this study was to characterize the epidemiology of Trichomonas spp. in a community of free-ranging domestic and wild Columbiformes and an endangered predator, Bonelli's eagle Aquila fasciata. We surveyed 253 live-captured Rock doves, 16 nestling Bonelli's eagles and 41 hunted Columbiformes. Oro-esophageal swabs were incubated in culture media and Trichomonas spp. isolated from Bonelli's eagle (6.3%, CI95 1.1-28.3), Turtle dove Streptopelia turtur (56.3%, CI95 39.3–71.8), Wood pigeon Columba palumbus (83.3%, CI95 43.7–97.0) and Rock dove Columba livia (68.4%, CI95 62.4–73.8). Infected Rock doves showed significantly poorer body condition than uninfected ones (p = 0.022). From a subset of 32 isolates, 18S and ITS1/5.8S/ITS2 rRNA genes were sequenced and Maximum-Likelihood trees inferred. Four ribotypes of Trichomonas spp. were identified. In this study area Trichomonas spp. seem to persist in a multi-host system involving several species of Columbiformes. Conservation actions aimed at increasing the availability of trophic resources for Bonelli's eagles through Rock dove restocking should consider the risk of pathogen transmission and of introduction of alien strains.


INTRODUCTION
Trichomonosis is an avian disease of conservation relevance, caused by infection with Trichomonas gallinae (1)(2)(3). Typical lesions in the upper digestive tract range from mild inflammation to caseous masses that can partially obstruct the digestive or respiratory tracts and progress to systemic disease (3). Trichomonosis can be present as subclinical to subacute, depending on the host species, immune status and the pathogenicity of the strain (1,2,(4)(5)(6)(7)(8). Transmission can be direct, such as through "crop milk" or consumption of infected prey or indirect such as through contaminated water (3,(9)(10)(11).
Trichomonas gallinae is an important pathogen of Columbiformes, Falconiformes, and Passeriformes (3). It has been implicated in the decline of Band-tailed pigeons (Patagioenas fasciata) in western USA (2) and a highly pathogenic strain has recently emerged in North-Western Europe, causing population declines of finch species (3,7,12). Columbiformes are maintenance hosts for T. gallinae and the Rock dove (Columba livia) is considered the natural host of this pathogen (3). All species of Columbiformes occurring in Europe have been shown to harbor endemic Trichomonas gallinae infection (8,13,14). Particularly the Turtle dove (Streptopelia turtur) was shown to be heavily infected by Trichomonas gallinae (8,15). European populations of Turtle doves heavily declined in the last decades and their migration routes pass through the Iberian Peninsula (16), with the potential of disseminating new strains. On the other hand, raptors are considered spillover hosts, acquiring infection from infected prey (3,17).
Bonelli's eagle (Aquila fasciata) is a conservation priority species in Europe because of widespread declines in the last decades of the 20th century (18,19). Most of the European population inhabits the Iberian Peninsula where its diet is based on Columbiformes, Red-legged partridge (Alectoris rufa) and Wild rabbit (Oryctolagus cuniculus) (20)(21)(22). Marginal populations of this species were shown to have low genetic variability (23), which could make them more vulnerable to emerging pathogens (24). Trichomonosis was shown to kill 2-14% of nestling Bonelli's eagles in several Iberian populations (25,26). A marginal population of Bonelli's eagle comprising 22-24 pairs breeds in the Douro Internacional/Arribes del Duero natural parks and adjoining areas in northwestern Iberian Peninsula (19,27,28). This population is characterized by a low productivity, which has been linked to food shortage (28). Conservation management actions include supplementary feeding, either direct or through restocking of traditional pigeon lofts with Rock doves (27,29).
Traditional pigeon lofts (TPL) are human constructions typical of Northwestern Iberian Peninsula (Figure 1), scattered across the landscape and originally aimed at producing Rock doves for human consumption and organic fertilizer for crops (30). These buildings provide nocturnal shelter and breeding sites for Rock doves, which are otherwise free-ranging (29,30). Nowadays most TPL are abandoned but some have been recovered and restocked to increase prey availability for Bonelli's eagles (29).
The aim of this study was to investigate the epidemiology of Trichomonas spp. in the community of free-ranging domestic and wild Columbiformes and their endangered avian predator, Bonelli's eagle, in an area of Northwestern Iberian Peninsula. Additionally, we aimed to contribute to elucidating if restocking TPL with Rock doves could increase the exposure of Bonelli's eagles to infection by Trichomonas spp.

Study Area
The study area consists of Douro Internacional/Arribes del Duero natural parks (comprising 193,000 hectares) and adjoining areas in Northwestern Iberian Peninsula (centroid 41 • 11 ′ 36 ′ ′ N, 6 • 45 ′ 49 ′′ W), encompassing both Portuguese and Spanish territory (Figure 2A). The area is an extensive plateau (600-800 m asl) where Douro river and several main tributaries (Águeda, Sabor, Côa, Tormes, and Huebra rivers) carved steep canyons in granite or shale rock. The landscape is highly fragmented, with oak (Quercus pyrenaica, Q. faginea, Q. rotundifolia, and Q. suber) and juniper (Juniperus oxycedrus) woods interspersed with shrub (mainly Cytisus sp. and Cistus sp.) and agricultural areas (mainly wheat Triticum sp., olive Olea europaea, vine Vitis vinifera, and pasture). The area is included in the Mesomediterranean and Supramediterranean bioclimatic zones showing Mediterranean-Subcontinental climate with large thermal amplitude, hot and dry summers and relatively cool and humid winters (31,32).  (Figure 2A) by closing the exits during the night and manually capturing them inside. To prevent repeated sampling of individuals, each TPL was sampled only once and Rock doves were kept in cages and released when all were sampled (<2 h of restraint). Samples from wild Columbiformes were obtained from animals hunted for recreational purposes, according to Portuguese and European legislation. No animals were killed for the purpose of this study. This study was carried out in accordance with national and European legislation and the recommendations of the "Guidelines on the care and use of wildlife" (33). An oro-esophageal swab was obtained from each bird, immediately inoculated in InPouch TM media (Biomed Diagnostics Inc., White City, OR, USA) and kept at room temperature and away from sunlight until incubation (7-50 h post collection). From individual Rock doves, we collected data on sex, age class, weight, wing length, and plumage (wild-type or other). The residuals of a linear regression between log weight and log wing length of each Rock dove were used as indicators of body condition. The presence of macroscopic lesions compatible with trichomonosis was systematically investigated in the Bonelli's eagle nestlings but not in the Columbiformes.

Sample and Data Collection
Data were collected on each TPL during the sampling procedure and included geographical location, number of Rock doves, presence of other avian species, provision of water, location within occupied Bonelli's eagle territories (<3 km straight line from the center of a territory occupied by a pair during the previous breeding season) and the interval between collection and incubation of the biological samples.

Laboratory Analysis
Culture media were incubated at 37 • C for 6 days and isolation of Trichomonas spp. was assessed by daily bright field microscopy at 100x magnification and considered positive when motile, flagellated protozoans were present in the culture media. Media containing Trichomonas were centrifuged at 1,430 × g for 10 min, the supernatant discarded, and the pellet re-suspended in PBS and stored at −20 • C. DNA was extracted from a subset of 32 Trichomonas isolates, 1-2 per host species and TPL studied, with DNeasy Blood & Tissue Kit (Qiagen) and purified with Exo-AP Clean-up (Thermo Scientific), according to the manufacturer's instructions.

Statistical Analysis
Differences in prevalence between TPL were analyzed by non-parametric methods (Kruskal-Wallis test). We performed a binomial GLM with Trichomonas spp. infection status in individual Rock doves as dependent variable and individual traits (age class, sex, body condition, and plumage type) as independent variables ( Table 1). We performed another GLM with Trichomonas spp. prevalence in TPL as dependent variable and environmental and sampling features as independent variables (number of Rock doves, presence of other avian species and water availability inside the TPL, location within an occupied Bonelli's eagle territory, and interval from collection of samples to start of incubation) ( Table 1). We considered all TPL <3 km straight distance from an occupied nest as being located within an occupied territory, based on the home ranges of breeding Bonelli's eagles in the Iberian Peninsula (38,39) (Figure 3A). Prevalence of Trichomonas spp. in TPL ranged 14.3-100% (Figure 3B), showing statistically significant differences between TPL (p < 0.001, Kruskal-Wallis test).

Determinants of Infection
In the binomial GLM with Trichomonas spp. isolation in individual Rock doves as dependent variable (Hosmer-Lemeshow χ 2 = 6.445, p = 0.598), the only significant determinant of infection was the body condition, which was significantly lower (p = 0.021) in infected Rock doves ( Table 3 and Figure 3C).

Molecular Epidemiology
Three ribotypes of Trichomonas gallinae and one ribotype of Trichomonas sp. were obtained from the 32 isolates that were sequenced ( Table 4 and  Table 4 and Figure 2B).

DISCUSSION
This study reports the epidemiology of Trichomonas spp. infection in the community of maintenance hosts for this pathogen and apparent spillover to an endangered avian predator. These results suggest that at least the Trichomonas gallinae ribotype 18S-II, ITS-227 I/D (n = 4), and ribotype 18S-VIII, ITS-III of Trichomonas sp. are maintained in the study area by a multi-host-pathogen system (41), where infection is likely transmitted between several species of free-ranging domestic and wild Columbiformes, as also documented in the United Kingdom (13). The inter-species transmission of Trichomonas spp. is supported by the high prevalence of infection in all studied Columbiformes species and particularly by the isolation of two genogroups in multiple species of Columbiformes. Four distinct ribotypes of Trichomonas were detected in Rock doves, two of which are presumed to naturally occur in the study area (T. gallinae ribotypes 18S-VI, ITS-IV/B and 18S-II, ITS-I/D), as they were previously reported in this host species elsewhere in Europe (42). On the other hand, T. gallinae ribotype 18S-IV, ITS-I/D was previously reported only in Eurasian collared doves (36,37), this being to our knowledge the first report of its isolation from Rock doves. Trichomonas gallinae ribotype 18S-IV, ITS-I/D was detected in the present study in two TPL recently restocked with Rock doves acquired in a mixed species collection. This observation highlights the risk of introducing new pathogen strains in local host communities by means of translocations for conservation purposes (43,44). Here we report the first isolation in European wild Columbiformes of Trichomonas sp. ribotype 18S-VIII, ITS-III. This genogroup was previously detected in Europe only in 2007, in a domestic Rock dove in Austria (37).
Frontiers in Veterinary Science | www.frontiersin.org Interestingly the most common and widespread T. gallinae ribotype 18S-VI, ITS-IV/B was only detected in Rock doves, while the less common T. gallinae ribotype 18S-II, ITS-I/D was also detected in Wood pigeon and Bonelli's eagle. Whether this corresponds to a greater ability of the latter for spillover to other host species is still to be determined. Differential inter-species  transmission rates have been suggested for other T. gallinae genogroups, as well as differential pathogenicity (7,15). As we only sequenced the 18S and ITS1/5.8S/ITS2 rRNA genes, more genetic diversity could have gone unnoticed in our sample of Trichomonas isolates. Furthermore, the opportunistic sampling performed in Bonelli's eagle, due to constraints in accessing nests of this endangered species, means there is no temporal overlap between samples collected from this species and those from Columbiformes.
Trichomonosis can be an important cause of mortality for nestling Bonelli's eagles; nevertheless, subclinical infections are more common than clinical ones (25,26). In the present investigation only one nestling was shown to harbor Trichomonas gallinae, but no lesions were detected during handling in the nest and it survived to fledging. We could not isolate Trichomonas from two other nestlings with compatible lesions (oropharyngeal caseous masses), which could be due to a different etiology or a long interval between collection of samples and incubation.  Comparable studies in other populations of Bonelli's eagle report higher prevalence in nestlings, which could be due to the low sample size or to the relatively long interval between collection of the samples and incubation in our study. Trichomonas gallinae was isolated from 36 to 68.8% of nestlings in 3 populations from Southern Iberian Peninsula (5,25,26), where Columbiformes are 28.5-39.2% of the prey items of Bonelli's eagles (20,45). In our study population, Columbiformes are 20.6% of the diet of this species, with a preponderance of Rock doves over Wood pigeons (27). Given the high prevalence of infection by Trichomonas gallinae in all species of domestic and wild Columbiformes and the importance of these prey on the diet of Bonelli's eagles across its Iberian range, most if not all nestlings should be exposed to Trichomonas sp. (26). The variable prevalence of infection in nestlings from different populations could be explained by individual, regional and temporal heterogeneities in the pathogenicity of Trichomonas gallinae strains present in the local population of Columbiformes or the immune status of Bonelli's eagle nestlings. Marginal Iberian populations of Bonelli's eagle were shown to have low genetic diversity (23), which could impair their immune response (24). Further large-scale studies on this predatorparasite-prey system, addressing the link between diet and Trichomonas gallinae prevalence, are needed to elucidate the reasons for this heterogeneity in prevalence between populations of Bonelli's eagles.
One of the drivers of this study was the concern that restocking TPL with Rock doves would increase the exposure of Bonelli's eagles to Trichomonas gallinae (46). While this could occur if the consumption of Rock doves increases in response to this conservation action, we show that Trichomonas spp. infection is widespread in the wild Columbiformes community. The single case of infection in a Bonelli's eagle nestling took place in a territory where no TPL restocking was carried out. Nevertheless, emerging Trichomonas spp. strains might be amplified in TPL, providing opportunity for spillover to Bonelli's eagles. Systematic disease surveillance in TPL should be implemented to detect the potential introduction of new strains and contingency plans prepared to prevent spillover to endangered avian predators.
Turtle doves in our sample showed lower prevalence than those reported elsewhere in Europe [67-100% - (8,(13)(14)(15)]. European populations of this species have undergone a sustained decline in the last decades, and trichomonosis has been suggested as contributing for this decline (8,15,16). The prevalence we determined in other species are comparable to those reported in the literature: 34-70% in Wood pigeon throughout Europe (8,13,14,47) and 45-79% in Rock doves in Spain (5,14).
Interestingly we observed that Rock doves from which Trichomonas spp. were isolated showed poorer body condition. It was shown in other host-pathogen systems that the physiological costs of immune system activation to fight infection trade-off with other physiological needs, such as reproduction, foraging and growth (48,49). In our sample this translated into a lower body condition of infected Rock doves, as previously shown to occur in other free-ranging Columbiformes infected with T. gallinae (15,47). The lower body condition of infected Rock doves could heighten the risk of predation, potentiating interspecies transmission, but being located within a Bonelli's eagle territory was not found to be a risk factor for prevalence in TPL. Studies including the whole community of avian predators of Rock doves are needed to address this question.
Trichomonas spp. seem to persist in the study area in a multihost ecosystem, where intra-and inter-species transmission likely contribute to maintenance, making it demanding to individualize the role of each host species in the epidemiology of infection (41). More detailed molecular epidemiology tools have the potential to further contribute to unveiling the dynamics of pathogen transmission in this host community. Long-term studies will allow uncovering the origin of infection for avian predators such as the Bonelli's eagle, potentially highlighting ways to control spillover to this endangered species. Meanwhile, conservation actions aimed at increasing the availability of trophic resources for Bonelli's eagles should consider the potential risk of increased disease transmission and minimize the risk of introducing alien Trichomonas spp. strains.

ETHICS STATEMENT
This study was carried out in accordance with the recommendations of guidelines on the care and use of wildlife (33), Portuguese and European legislations. The study was conducted under permits 412/2014 and 316/2016 (Junta de Castilla y León, Spain) and 415/2014 (Instituto de Conservação da Natureza e Florestas, Portugal).