Pathogen communities of songbird-derived ticks in Europe’s low countries

Background Birds play a major role in the maintenance of enzootic cycles of pathogens transmitted by ticks. Due to their mobility, they affect the spatial distribution and abundance of both ticks and pathogens. In the present study, we aim to identify members of a pathogen community [Borrelia burgdorferi (s.l.), B. miyamotoi, ‘Ca. Neoehrlichia mikurensis’, Anaplasma phagocytophilum and Rickettsia helvetica] in songbird-derived ticks from 11 locations in the Netherlands and Belgium (2012–2014). Results Overall, 375 infested songbird individuals were captured, belonging to 35 species. Thrushes (Turdus iliacus, T. merula and T. philomelos) were trapped most often and had the highest mean infestation intensity for both Ixodes ricinus and I. frontalis. Of the 671 bird-derived ticks, 51% contained DNA of at least one pathogenic agent and 13% showed co-infections with two or more pathogens. Borrelia burgdorferi (s.l.) DNA was found in 34% of the ticks of which majority belong to so-called avian Borrelia species (distribution in Borrelia-infected ticks: 47% B. garinii, 34% B. valaisiana, 3% B. turdi), but also the mammal-associated B. afzelii (16%) was detected. The occurrence of B. miyamotoi was low (1%). Prevalence of R. helvetica in ticks was high (22%), while A. phagocytophilum and ‘Ca. N. mikurensis’ prevalences were 5% and 4%, respectively. The occurrence of B. burgdorferi (s.l.) was positively correlated with the occurrence of ‘Ca. N. mikurensis’, reflecting variation in susceptibility among birds and/or suggesting transmission facilitation due to interactions between pathogens. Conclusions Our findings highlight the contribution of European songbirds to co-infections in tick individuals and consequently to the exposure of humans to multiple pathogens during a tick bite. Although poorly studied, exposure to and possibly also infection with multiple tick-borne pathogens in humans seems to be the rule rather than the exception.


Background
Songbirds are swift transporters of ticks and tick-borne pathogens, spreading them over long distances on bird migration and dispersal routes. They are important pathogen reservoirs and carriers of infected ticks in areas that are less accessible to mammals, but still frequently visited by humans, such as islands, green space and gardens in urbanized areas [1][2][3][4][5]. Not only their contribution in the terrestrial cycles of pathogens has become clear during the past decades, but also their importance in maintaining tick populations is now generally recognized [6][7][8][9].
In Europe, bird-associated Borrelia burgdorferi (s.l.) species such as B. garinii and B. valaisiana [10][11][12][13][14] have been associated with human Lyme borreliosis [15,16]. However, limited information is available on the birds' contribution to the cycles of other human tick-borne pathogenic agents, as well as the mechanisms of cooccurrence of more than one pathogenic agent in individual birds and bird-derived ticks ("co-infection") [17]. Understanding the mechanisms underlying co-infections in ticks is important, as co-infections in hosts in which tick bites are relatively low (e.g. humans) can result from the attachment of a single co-infected tick rather than sequential bites of multiple singly-infected ticks [18].
Simultaneous infections of multiple pathogen species can lead to increased pathogenicity, can affect pathogen proliferation dynamics in the hosts, can influence the host's immune responses, can affect the distribution of pathogens in the host body and can complicate the diagnosis and treatment of disease [19][20][21][22][23].
The scope of our study is to identify the members of the pathogen community in tick species derived from songbird species in Europe's Low Countries (Belgium and the Netherlands), to define their infection prevalence, and to investigate whether the occurrences of different pathogenic agents are independent of each other. Songbirds could be classified in nine categories following their foraging habitats (see Table 1) based on the information provided in a reference work [41] and the expert knowledge of two experienced bird-watchers of the University of Antwerp (J. Elst and D. Heylen). Immediately after collection, the ticks of an individual bird were immersed in a single tube filled with ethanol (80%), which was subsequently stored at -20°C until species identification and DNA extraction. Ticks were identified to species and life stage by trained and experienced technicians who used various taxonomic keys [42][43][44][45].

Bird trapping and collection of ticks
DNA extraction, qPCR assays and sequencing procedures DNA was extracted from ticks using a Qiagen DNA extraction procedure. For the detection of B. burgdorferi (s.l.), a duplex qPCR was used, based on the detection of fragments of ospA and flagellin genes [5]. A conventional PCR assay, targeting the 5S-23S intergenic region (IGS), was performed for B. burgdorferi (s.l.) species identification [46]. Conventional PCR assays were carried out in a Px2 thermal cycler (Thermo Electron Corporation, Breda, the Netherlands) and visualized on a 2% agarose gel. Both strands of PCR products were sequenced by BaseClear (Leiden, the Netherlands), according to the company's protocol and using the same forward and reverse primers as in conventional PCR. BLAST analyses and in-house molecular epidemiological databases (Bionumerics 7.1. -Applied Math, Belgium) were used to identify Borrelia burgdorferi (s.l.) species. These databases contain all our DNA sequences from (field) isolates, together with (reference) sequences from GenBank [47,48].
For detection of B. miyamotoi, a qPCR assay was used that targets a region of the flagellin gene, specific for B. miyamotoi [49]. For detection of A. phagocytophilum and 'Ca. N. mikurensis' DNA, a single duplex qPCR assay was used, which is described elsewhere [50,51]. This qPCR assay targets specific regions of genes msp2 (major surface protein 2) for A. phagocytophilum and groEL (heat shock protein) for 'Ca. N. mikurensis'. For detection of R. helvetica, we used a multiplex qPCR assay, targeting the gltA gene, as described earlier [52].

Statistical analysis
Generalized linear mixed effects models (GLMM) were fitted to test whether co-occurrences of different pathogen species in individual ticks were independent of each other (logit-link, binomial-distributed residuals), taking into account the correlation structure of cofeeding ticks that were obtained from the same individual [53]. In these models, individual birds were nested within bird species; both were modelled as random effects. For the inference of the maximum likelihood estimates of the fixed effects, Kenward Roger approximation was used to estimate the denominator degrees of freedom of the F-distributed test statistics, which takes into account the correlation of observations within the same cluster [53,54]. For those bird species of which at least 10 individuals were caught, mean tick infestation intensity (i.e. the average number of ticks in infested individuals) was calculated. For bird species with at least 10 infected individuals per tick stage, estimates of proportions of infected ticks are given by their arithmetic mean ± standard error (i.e. the square root of the estimated variance/ the square root of the number of bird individuals) in the main text. Data management and statistical analyses were performed using SAS v9.2 (SAS Institute, Cary, North Carolina, USA).
The occurrence of B. miyamotoi in ticks was very low [prevalence in ixodid ticks: 0.6% (4/670)]. It was only found in one I. ricinus larva from E. rubecula and one I. ricinus nymphs from T. merula, Phylloscopus collybita, T. philomelos each.
For A. phagocytophilum, we detected DNA in 5.1% (34/670) of all ixodid ticks. The nymphal infection prevalence in the four bird species with more than 10 infested birds varied between 1.4 ± 0.9% (P. modularis) and 8 ± 4.9% (T. iliacus). Furthermore, A. phagocytophilum DNA was found in two I. frontalis females from one individual blackbird as well.

Discussion
We have shown that half of the songbird-derived I. ricinus ticks, that readily feed on humans, contained DNA of one or more bacteria that are pathogenic to humans: B. burgdorferi (s.l.), R. helvetica, A. phagocytophylum, B. miyamotoi, 'Ca. N. mikurensis'. The presence of the DNA in the ticks shows that songbirds carry infected ticks and may facilitate bacterium transmission. Transmission facilitation via birds may occur either via the infection of bird tissue on which ticks feed or via cofeeding of ticks in close spatial and temporal proximity to each other. The latter transmission pathway can occur in the absence of a systemic infection, allowing some pathogens (e.g. Borrelia species) to evade the hostile immune system of otherwise incompetent hosts [55]. Ground-dwelling birds, especially the members of the family Turdidae, had the highest infestation intensities and also yielded the highest numbers of infected ticks overall. They are known to run a greater risk of I. ricinus exposures, as they forage inside the habitat of this abundant tick species (i.e. ground and lower vegetation strata) [4,56]. Particularly the blackbird (T. merula) and the song thrush (T. philomelos), two very common birds of European woodlands and gardens, contributed strongly to the number of infected ticks. In line with other European studies, birds were predominantly infested by immature I. ricinus stages and rarely by adult females. Adult I. ricinus are typically found on mediumsized and larger mammals (e.g. roe deer) on which they copulate [57]. In contrast but not surprising, we found substantial numbers of semi-engorged adult female I. frontalis on the birds; all developmental stages I. frontalis feed on birds [42].
We found a strong association of B. garinii, B. valaisiana and B. turdi with avian reservoir hosts, which has previously been shown by numerous European field studies concluding that birds act both as competent reservoirs and transmitters for these species [10-12, 29, 58-60]. Given that vertical transmission of B. burgdorferi (s.l.) spirochetes in I. ricinus ticks only rarely occurs [61,62], their presence in larvae (Tables 2, 3) indicates that they were acquired either via (local) infection in the host or via co-feeding with an infected nymph. Borrelia turdi, recently discovered in Europe and strongly associated with I. frontalis, was also found in I. ricinus nymphs. Transmission experiments have shown that I. ricinus can transmit B. turdi to naïve avian hosts and, seen the extreme host range of this tick species, I. ricinus could potentially act as a bridging vector towards mammals, including human [60]. Recent experimental and observational studies based on larval and nymphal infections show non-homogeneous distributions of the avian Borrelia species in bird-derived ticks, indicating differential transmission and amplification of these species depending on the avian reservoir hosts and tick species [12,60,63].
An interesting outcome of our study and previous field studies is that several of the ground-dwelling birds (T. merula, T. philomelos, E. rubecula and P. modularis) were frequently infested with ticks that carried B. afzelii, a species that is associated with rodent reservoir hosts [64,65]. The fact that all B. afzelii-infected ticks were nymphs suggests that these individuals had fed as a larva on a B. afzelii-infected mammal, moulted and maintained their infections when feeding on the birds. Findings of bird-derived infected larvae in other studies have led to speculations that particular strains of B. afzelii can also use bird hosts for transmission [14]. PCR-based screening outcomes, like ours and others, should, however, be interpreted with caution as they are based on the detection of specific DNA sequences and do not necessarily mean that viable, infectious microorganisms are present. A recent experimental study investigating transmission of B. afzelii in blackbirds and great tits showed that nymph-to-adult transstadial transmission of B. afzelii DNA could occur. However, the positive signal in the adult ticks turned out not to be viable and infectious spirochetes, as shown by the BSK culture test [66]. It is, therefore, necessary to identify the B. afzelii strains found in bird-derived ticks from the wild and test via culture-based infection methods and tick transmission experiments whether they are still infectious and transmittable after being exposed to bird blood during tick feeding.
Also for the more sporadically observed mammalassociated pathogens (B. miyamotoi and 'Ca. N. mikurensis') that we detected in bird-derived ticks, including larvae (Table 2), experiments are needed to investigate whether viable bacteria survive the birds immune system and/or are transmitted during tick feeding. Studies in the USA and Europe have implicated small rodent species as the reservoir hosts for B. miyamotoi [34,67], but a limited number of studies reported B. miyamotoi infections from bird-derived ticks as well [17,29,68]. For the rodent-associated 'Ca. N. mikurensis' [50] observed in bird-derived I. ricinus larvae and nymphs of our study and others [28,29], a role for songbirds as transmission facilitators could be expected.
Compared to the other A. phagocytophilum and 'Ca. N. mikurensis' , the number of ticks with R. helvetica DNA was high. A good comprehension of the transmission dynamics of rickettsial bacteria in songbirds is still lacking. Within infected ticks, a proportion of the bacteria could have a maternal origin, as R. helvetica, in contrast to Borrelia burgdorferi (s.l.), can be transmitted transovarially [69,70]. However, the experimental study of Heylen et al. [17] using great tits (Parus major) exposed to a community of pathogens, clearly shows rapid R. helvetica transmission via co-feeding (cf. mammals [71,72]) and/or fast systemic infection (as found in mammals experimentally injected with different rickettsial species [73]). Our and other's finding of R. helvetica in bird-derived ticks, including larvae [3, 29-31, 51, 74, 75] reinforces the presumption that songbirds can become bacteraemic and effectively facilitate the transmission of this pathogen via host tissue [30].
Further, our study provides evidence that grounddwelling birds, especially thrushes, are important hosts in the transmission cycles of A. phagocytophilum. Bacteraemia of this pathogen has been shown to develop in songbirds [30], which is likely the reason for the reports of infected bird-derived ticks here and other locations [25,26,29,30,76]. Probably not all birds are equally competent in the transmission; in a great tit (Parus major) experiment by Heylen et al. [17] no transmission facilitation occurred despite the presence of A. phagocytophilum in challenge nymphs. Our finding of infections in a bird-specialized tick (I. frontalis) that is never found on other vertebrate hosts (two infected adult females cofeeding with infected I. ricinus nymphs on the same blackbird individuals) gives further indication that birds facilitate A. phagocytophilum transmission, either via cofeeding transmission or systemic infections. Although the host-specific strains of A. phagocytophilum were not identified, the avian ecotype IV that has been isolated by Jahfari et al. [51] from blackbird tissues and blackbirdderived ticks is to be expected.
The co-infections found in individual ticks and birds strongly suggest that simultaneous transmission of different bacterium species can occur and that birds are permissive for multiple pathogens, as experimentally shown in Heylen et al. [28]. Especially, the fact that coinfections were found in (sets of ) larvae provides the strongest indication. However, larvae could also acquire pathogens via the maternal line from other hosts than the individual songbird from which they were collected, through vertical transmission (e.g. in B. miyomatoi and R. helvetica) [69,70,77]. In larvae, the most frequent observed co-infection was B. burgdorferi (s.l.) with R. helvetica (E. rubecula, T. troglodytes, T. philomelos), but also A. phagocytophilum with R. helvetica and 'Ca. N. mikurensis' with R. helvetica (both from T. iliacus) were observed. Only for one bacterial combination, 'Ca. N. mikurensis' + B. burgdorferi (s.l.), we found that the occurrence of the one pathogen is more likely when another pathogenic agent is present. Interestingly, also in mammals, this combination of pathogens was much higher than expected from the prevalence of each pathogen [39]. This positive association could be the result of variation in general susceptibility among birds, but could also indicate transmission facilitation, as has been suggested in other studies on tick-borne co-infections [18,34,78]. The pathways that lead to facilitation can only be elucidated with experimental studies in which pathogen-driven physiological, cellular and biochemical interactions are disentangled.

Conclusions
Our findings highlight the contribution of songbirds to co-infections in individual ticks. In addition, not only avian but also mammalian bacterium species are transported via bird-derived ticks, highlighting the need to experimentally test whether latter pathogens are viable and infectious in birds. Furthermore, future studies should focus on the reservoir competence of members in the bird community and how the different vector-bird-niches contribute to the pathogen transmission dynamics.