Long-term trends of tick-borne pathogens in regard to small mammal and tick populations from Saxony, Germany

Background Rodents are important in the life-cycle of ticks as hosts for immature developmental stages. Both rodents and ticks are of public health interest as they are reservoirs and vectors for different tick-borne pathogens (TBP). The aim of this study was to reassess the prevalence of TBP in previously studied areas of the city of Leipzig (Saxony, Germany). Methods In the years 2015–2017 rodents and ticks were collected in parks and forest areas in Saxony. DNA was extracted from the rodents, attached and questing ticks. Samples were screened for the presence of Anaplasma phagocytophilum, Babesia spp., Borrelia burgdorferi (s.l.), “Candidatus Neoehrlichia mikurensis” (CNM), Bartonella spp., Hepatozoon spp. and Rickettsia spp. using PCR methods. Rodent, attached nymph and questing tick (nymph and adult) samples were tested individually, while attached larvae were further processed in pools. Results A total of 165 rodents (Apodemus agrarius, n = 1; A. flavicollis, n = 59; Arvicola terrestris, n = 1; Myodes glareolus, n = 104), 1256 attached ticks (Ixodes ricinus, n = 1164; Dermacentor reticulatus, n = 92) and 577 questing ticks (I. ricinus, n = 547; D. reticulatus, n = 30) were collected. The prevalence levels in rodents were 78.2% for Bartonella spp., 58.2% for CNM, 49.1% for B. burgdorferi (s.l.) 29.1% for Rickettsia spp. and 24.2% for Hepatozoon spp. The minimal infection rates (MIR) in attached larvae ticks were 39.8% for Rickettsia spp., 32.7% for Bartonella spp., 7.1% for CNM and 8.8% for B. burgdorferi (s.l.) and the prevalence rates in attached nymphs were 33.7% for Bartonella spp., 52.9% for Rickettsia spp., 13.5% for CNM and 11.3% for B. burgdorferi (s.l.) Both rodents and attached ticks were negative for Babesia spp. The prevalence in questing ticks was 18.2% for Rickettsia spp., 7.3% for CNM, 6.4% for B. burgdorferi (s.l.) and 1.4% for Babesia spp. All tested samples were Anaplasma-negative. Sequencing revealed the occurrence of 14 identified species. Conclusions This research is the first evaluation of the prevalence for Hepatozoon spp. in rodents from Germany. In comparison to earlier studies, detected pathogens species remained the same; however, the prevalence for particular pathogens differed.


Background
Small mammals are important hosts for the developmental immature stages of ticks in their natural life-cycle. Moreover, small mammals serve also as reservoirs [1] for various zoonotic agents. Ixodes ricinus is the most prevalent tick species in Europe and is responsible for the transmission of most zoonotic tick-borne pathogens (TBP) [2]; however, Dermacentor reticulatus is a rising concern as a potential vector of TBP.
Anaplasma phagocytophilum and "Candidatus Neoehrlichia mikurensis" (CNM) are Gram-negative, obligate intracellular bacteria which are tick-borne and mainly transmitted by I. ricinus [3]. However, D. reticulatus has also been described to harbour both [4,5]. There are four ecotypes of A. phagocytophilum and only two are vectored by I. ricinus [6]. While A. phagocytophilum is known to cause mild to severe symptoms in humans, dogs and other mammals, CNM is rather an opportunistic agent mostly affecting immunosuppressed humans and dogs [7,8]. CNM is considered to be harboured by rodents such as Myodes glareolus and Apodemus flavicollis [1]. Whereas roe deer, wild boars and hedgehogs are regarded as reservoirs for A. phagocytophilum, the reservoir function of small mammals is disputable, as there are supportive as well as refutable studies [1,[9][10][11][12].
Rickettsia spp. are likewise zoonotic Gram-negative, obligate intracellular bacteria which may be subdivided in four groups: (i) the spotted fever group (SFG); (ii) the typhus group; (iii) the Rickettsia bellii group; and (iv) the Rickettsia canadensis group [13]. Most rickettsiae belonging to the SFG are tick-borne and zoonotic. While I. ricinus is thought to be a vector in particular for Rickettsia monacensis and R. helvetica, D. reticulatus seems to be the main vector for R. raoultii in Europe [1,13,14]. While R. helvetica and R. slovaca are considered to be harboured by sika deer and dogs, and by wild boars and domestic ruminants, respectively, the reservoir host for R. raoultii is still not clear [15][16][17]. Nonetheless, small mammals have previously been found positive for all three aforementioned Rickettsia species [18,19].
Species of the Borrelia burgdorferi (sensu lato) complex are the causative agents of Lyme disease which is the most prevalent tick-borne disease in Europe [20]. Ixodes ricinus is known to be the main vector and small mammals are expected to be key reservoirs for B. afzelii which is a species of the B. burgdorferi (s.l.) complex [21].
Bartonella spp. are zoonotic, Gram-negative, vectorborne bacteria. Rodents are known to be reservoirs for most Bartonella species [22], whilst a variety of arthropods such as fleas, lice, keds and ticks are considered to transmit these pathogens. In Germany, human cases of bartonellosis, mainly caused by B. henselae, have been previously reported [23].
Babesia spp. and Hepatozoon spp. are small intracellular parasites which are harboured by many different vertebrate hosts including birds and mammals in Europe [24,25]. Babesia microti is mostly found in voles of the genus Microtus, in particular M. agrestis in Europe. However, there are also reports of B. microti in other rodent species such as M. glareolus and A. flavicollis [26]. Ixodes ricinus is believed to be the main vector of several Babesia spp. [27]. However, I. trianguliceps, a rodent-associated tick species, seems to be the key vector of B. microti in Europe. Human babesiosis caused by B. microti was previously reported in a human from Germany [28].
In the past, Hepatozoon spp. in rodents were not directly examined in Germany; however, there was an accidental finding of Hepatozoon sp. in one rodent previously tested by our study group [29] and other findings in M. glareolus and M. oeconomus previously from Poland, but neither in A. flavicollis nor insectivores [30]. Thus far the Hepatozoon species obtained from small mammals in Europe are either non-pathogenic or of unknown pathogenicity to humans [31]. Hepatozoon canis, which is highly pathogenic to dogs, was previously found in I. ricinus and D. reticulatus collected from foxes in Germany [32]. Most previous examinations on TBP in hosts and vectors from nature were carried out in a time frame of few years only and did not reassess the same areas again. Thus, long term studies on ticks, small mammals and TBP are scarce. However, it may be of importance to survey the dynamics of TBP in hosts and vectors in the purpose of predicting the distribution and maintenance of TBP in the future. Previous research showed a quite high prevalence of the previously mentioned TBPs in small mammals and ticks from Saxony, Germany [4,18,29,[33][34][35][36].
The present study reassessed TBPs in small mammal and tick populations from sites in Saxony which were previously examined by our group for TBP over the last 9 years [4,18,29,[33][34][35][36]. Thus, the aims of this study were: (i) collection of rodents, their attached ticks and questing ticks in Saxony, Germany; (ii) assessment of the prevalence of the mentioned pathogens in collected rodents and ticks; (iii) comparison of the current results with our previous studies from the last 9 years [4,18,29,[33][34][35][36]. , n = 83) were more prevalent than larvae (9.8%, CI: 5.0-17.8%, n = 9). The maximum infestation rate on rodents was 135 ticks per host (M. glareolus) with a mean value of 7.6 (SD= 16.43).

Comparing the present results with previous studies
The results from this study were compared with results obtained in 2009-2014 from the same sites [4,18,29,[33][34][35]. Regarding the numbers and diversity of captured small mammals, there is a visible decreasing trend. In the past, a total of 10 small mammal species were captured, while in the present study only 4 rodent species were found. Furthermore, the species of attached ticks were more diverse in the previous investigations, since I. trianguliceps and unidentified Dermacentor and Ixodes ticks were also found. In the present study, A. phagocytophilum was absent in each type of tested sample, while previously it had been detected in small mammals, questing and attached ticks [4,29]. Rodents and attached ticks were also Babesia-negative, whereas before they had been positive [29,34]. Regarding questing ticks, the prevalence for Babesia spp. in I. ricinus slightly decreased from 4.1% in 2009 to 1% in the present study (P = 0.0359) [29]. However, in this investigation DNA of Babesia was additionally found in questing D. reticulatus. In the present study B. burgdorferi (s.l.) was detected in questing ticks (also only in I. ricinus) with no statistical differences compared to the former research [33]; however, the present prevalence in small mammals (49.1%) was much higher than in the past (31.2%) (P < 0.0001). Borrelia burgdorferi (s.l.) in attached ticks had not been tested in the previous investigations. The prevalence for Rickettsia spp. in questing, attached ticks and small mammals seems to be stable over the years as it had been similar in the past [18,33]. The infection levels of CNM seem to be increasing. The prevalence from this research was significantly higher than in the last study [4] in small mammals (41.2 vs 58.2%, P = 0.0003) and the prevalence for attached ticks in the past was fluctuating from 1.9 to 9.8% while now the average MIR for larvae was 7.1% and the average prevalence for nymphs was 13.5%. Bartonella spp. remained the most often detected pathogen in small mammals [35].

Discussion
This study reassessed the prevalence of TBP over 9 years in ticks and rodents from sites previously examined by our group in the surroundings of Leipzig, Saxony, Germany [4,18,29,[33][34][35]. Although such long-term investigations are scarce, they may be of importance from a public health point of view to survey dynamics of TBP in hosts and vectors as this may help to predict the distribution and maintenance of TBP in future. The number of captured rodents and questing ticks as well as their species diversity has been decreasing through the years. In contrast, the average tick infestation on rodents has been rising in recent years. A reason for this phenomenon may be the so-called dilution effect. This effect describes that the higher the number of individuals in a host population the lower the tick burden per host individual [37]. In line with a former study, D. reticulatus was exclusively found on M. glareolus while I. ricinus did not have such a host association [18]. CNM is widespread in rodents across Eurasia with a prevalence ranging between 10.8-52.7% in Germany and other European countries, such as the Netherlands and Slovakia [36,38,39]. Earlier, it was described that male rodents were more often infected with CNM than females [4]. The present research confirms a sex-biased difference in the prevalence for CNM in M. glareolus. Previous studies explained this bias by a higher activity rate of males and due to immunosuppressive effects and higher aggression levels resulting in a higher chance of encountering the pathogen through fights [40]. Through wounds, scratches and/or bites pathogens may be transmitted directly to the bloodstream. Previous studies from Austria, France and the Netherlands showed a moderate prevalence (1.7-22%) in the known CNM vector, I. ricinus [41][42][43]. The prevalence in the present study was statistically lower in questing ticks than in previous studies [36]. CNM has been rarely investigated in D. reticulatus ticks. In this study it could be found only in attached D. reticulatus and not questing individuals, suggesting that it was probably a temporary uptake through the blood meal. Previously B. burgdorferi (s.l.) was described in rodents in other European countries with a prevalence of up to 77% in Austria [44]. In the present study, the prevalence of B. burgdorferi (s.l.) in rodents has significantly increased in the years 2015-2017 compared to 2012-2014 (from 31 to 49%) [33]. A previous investigation showed B. burgdorferi (s.l.) has many mechanisms to elude the hosts' immune system, thus persisting in their rodent host [45]. One proven effect is described by a T-dependent B cell response which is subverted during infection in reservoir hosts. This could be a reason for the rise of the prevalence over the years. However, a dilution effect may not be ruled out as the population size of rodents decreased over the years, while the tick density increased per rodent. As described earlier for CNM, male M. glareolus were likewise more often infected than females. Sequencing from rodent samples confirmed the presence of pathogenic B. afzelii, the main rodent-associated Borrelia species [46]. Although the prevalence in small mammals increased, it did not vary in ticks over the years in this study. The prevalence in questing and attached I. ricinus ticks from the present study was in line with other European countries, e.g. Estonia, Belarus, Slovakia and Austria (8.2-13.5%) [14,47,48]. Rickettsia spp. were found in almost 24% of the rodents from this study which was higher compared to the prevalence detected in other parts of Germany, e.g. Mecklenburg-Western Pomerania, Thuringia and Baden-Wuerttemberg (6.8-9.4%) [49], and similar to a study from Lithuania (27.6%) [50]. Previous investigations in Europe revealed the occurrence of R. helvetica in A. agrarius, A. flavicollis and M. glareolus [51]. An earlier study by our group also showed the presence of R. raoultii in small mammals [18]. DNA of Rickettsia spp. was found in larvae attached to positive as well as to negative rodents, which supports the hypothesis of transovarial transmission of Rickettsia in ticks [52]. The current prevalence of 10.3% in I. ricinus is relatively low compared to the prevalence from previous studies in Germany (18-25%) and other European countries, e.g. France (16%) [18,33,53]. The infection level in attached (20-78.4%) and questing (76.9%) D. reticulatus ticks from the present study was much higher than in Dermacentor ticks from Poland and the Czech Republic (18-41%) [54,55]. The previous prevalence from the same sites showed a likewise high prevalence in questing D. reticulatus (70.5%) [33]. Rickettsia raoultii was detected only in questing D. reticulatus ticks with a very high prevalence and in M. glareolus with a low infection rate, which is in accordance with studies suggesting the transovarial transmission of R. raoultii in D. reticulatus is more significant than feeding on reservoir hosts in order to maintain in the natural life-cycle [18]. Bartonella spp. in rodents are highly prevalent in Europe with prevalence rates ranging between 16-56% in France, Denmark and Poland [56][57][58]. In the present study the prevalence was 78% in rodents and thus the highest in comparison to all other examined TBPs. A previous examination at the same study sites [35] detected a lower prevalence of 65.8% and the following species: B. grahamii, B. taylorii, Bartonella sp. N40; and a variety of uncultured Bartonella species. In the present study, only B. taylorii and uncultured Bartonella strains were detected. Bartonella taylorii is known to be non-pathogenic to humans and the uncultured Bartonella spp. are currently of unknown pathogenicity [59]. Previously it had been shown that the prevalence for Bartonella spp. is significantly higher in Apodemus than in Myodes due to a deficiency in resolving the infection in Apodemus [60]. However, it was also shown that the prevalence of Bartonella spp. in M. glareolus, studied over 11 years, was subjected to great fluctuations and may even double over the years before declining again, as the prevalence is dependent on changes in the rodent population such as density and mean age [61]. Bartonella spp. could not be detected in questing ticks from the present study, supporting the hypothesis that ticks play a subordinate role in the transmission of rodent-associated Bartonella. Earlier studies from our group, however, support the hypothesis that ticks play a role in the life-cycle of Bartonella spp. since B. chomelii was detected in ticks attached to rodents. This Bartonella species is, however, associated with domesticated ruminants [62]. In the present study, seven attached larvae pools/nymphs were positive for Bartonella spp. even though the host was negative. Previously our group suggested that D. reticulatus plays a subordinate role in the transmission cycle compared to I. ricinus. However, the present study found almost equally high prevalence rates in attached D. reticulatus and I. ricinus. To our knowledge, there are thus far no studies focussed on the presence of Hepatozoon spp. small mammals in Germany. Studies from Spain, Slovakia and Poland reported a prevalence range of 4.5-41.6% in different rodent species, including A. flavicollis and M.glareolus [30,63,64]. In the present study, the prevalence of Hepatozoon spp. in rodents was 31.1%. In accordance with a study from Slovakia, M. glareolus showed a significantly higher prevalence than A. flavicollis [64]. This was also observed in rodents from Finland and Poland [30,65]. Hepatozoon strains detected in small mammals from this study are known to have a wide host range and were previously detected in small mammals and reptiles [66]. It is not surprising that attached ticks as well as questing ticks were negative for Hepatozoon spp. in the present study, as rodent-associated Hepatozoon spp. are mainly transmitted by rodent-associated fleas [67]. Babesia DNA in this research was barely detected in questing ticks (1.4%) and not at all in rodents or in attached ticks. However, previous investigations from the same study sites revealed a similar prevalence in questing ticks (1.6%) and very low prevalence in attached ticks (0.3-0.5%) and rodents (0.6-2.5%) [29,34]. The prevalence for Babesia in rodents from other European studies showed similarly low levels in rodents; however, the study from the UK reported a much higher prevalence (27.2%) [68]. The prevalence in questing ticks in former studies from Sweden and Poland varied but also in a lower range (up to 4.6%; B. venatorum, B. microti and B. divergens) [69,70]. In the present study, B. venatorum and B. microti were detected in I. ricinus, and B. capreoli in D. reticulatus. Babesia venatorum and B. microti are zoonotic agents and have been previously detected in I. ricinus from other European countries [69][70][71]. Thus far only the "Jena" strain of B. microti is thought to be pathogenic to humans in Europe [72]. However, the B. microti strain detected in this study showed 99% identity to a non-pathogenic Ukrainian B. microti strain. Babesia capreoli, which is thought to be non-pathogenic, has previously been described in I. ricinus, with reindeer serving as main hosts in Europe [71,73]. Interestingly, exclusively these three Babesia species described here were also previously detected at the same study sites [29].
In other studies from Germany, the prevalence for A. phagocytophilum in ticks varied between 1.9-8.9% [74][75][76]. In this investigation A. phagocytophilum DNA was neither detected in rodents nor in ticks. However, earlier results from our group showed a low prevalence in both rodents (1.1%; [4]) and questing ticks (5.3%; [29]). The explanation for the observed decline may be the effect of the resistance against A. phagocytophilum developed by rodents which may persist from 12 weeks up to a year, protecting them from re-infection and preventing uninfected ticks from infection, thus interrupting the infection cycle [77].
In comparison to the overall prevalence for TBPs in attached and questing ticks from this study, the level for rodents was generally higher, also resulting in a high co-infection rate. Even though the co-infection levels in questing ticks were very low as well as the prevalence for Babesia spp. (only 3 out of 220 ticks), most Babesia-positive ticks were co-infected leading to the assumption that infections with Babesia favour co-infections with other pathogens. The current co-infection level in rodents (over 70%) is much higher compared to a study from Austria where only 8.1% of the rodents were infected with more than one pathogen [78].

Conclusions
This study reports very high prevalence levels for TBP, especially in rodents. This is the first study focussing on the presence of Hepatozoon spp. in rodents from Germany. Furthermore, over a 9-year long trend, it should be taken into account that the number and the species diversity of rodents and questing ticks collected have been declining, while the average infestation rate for ticks attached to rodents has been increasing. While prevalence for A. phagocytophilum and Babesia spp. in general decreased or/and were not detected at all in the present study, the prevalence for CNM, Bartonella spp. and B. burgdorferi (s.l.), particularly in rodents, seems to be rising. Rickettsia spp. are the only pathogens in which the prevalence in rodents, attached and questing ticks has remained at the same level over the years. Although the prevalence rates for certain pathogens differed between the years, the detected pathogen species did not change with time.

Small mammal trapping
The trapping of small mammals took place in April to October 2015, May to November 2016 and March to October 2017. Twenty-five Sherman© live animal traps (H. B. Sherman Traps Inc., Tallahassee, FL, USA) were set for two successive nights each month at each site at the same time. Apple slices were used as bait and hay as isolation material. The traps were controlled twice a day; captured rodents were anesthetized on the spot with CO 2 and euthanized by cervical dislocation. The rodents were morphologically identified using taxonomic key [79] and dissected in the laboratory. Attached ticks, skin and spleen samples were taken from each rodent and stored at -80 °C until further processing.

Attached and questing ticks
Questing ticks were collected simultaneously with each rodent trapping action by the flagging method. Questing and attached ticks were stored at −80 °C until morphological identification [80] and further analysis. A total of 455 ticks were selected for further PCR analysis examining tick-borne pathogens, including 231 I. ricinus (207 larvae and 24 nymphs) obtained from 64 M. glareolus, 164 I. ricinus (135 larvae and 29 nymphs) from 41 A. flavicollis and 60 D. reticulatus (9 larvae, 51 nymphs) from 15 M. glareolus (Table 1). Altogether 351 larvae were tested in 113 pools: 342 I. ricinus larvae in 108 pools and 9 D. reticulatus larvae in 5 pools. Concerning questing ticks, a total of 194 I. ricinus and 26 D. reticulatus were selected for further molecular examination.

DNA extraction from rodents and ticks
For DNA extraction, 0.6 g of sterile ceramic beads (sized 1.4 mm, Peqlab Biotechnologie, Erlangen, Germany) and 500 μl of PBS were added to each rodent sample. For ticks, 1 g of steel beads (sized 2.8 mm) was used instead of ceramic beads. The samples were then homogenized at 5500× rpm for 3 × 15 s with 10 s break intervals in a Precellys ® 24 tissue homogenizer (Bertin Technologies, Montigny Le Bretonneux, France). Due to financial restrictions, not all ticks were selected for further analysis. Up to five questing ticks per tick species, collection site, per month and year were randomly chosen. Attached ticks were likewise selected with the addition of up to five attached specimens per small mammal species (up to 30 host individuals per rodent species per month and collection site). Attached larvae were further tested in pools of up to 5 individuals according to the selection criteria. DNA was extracted with a QIAamp DNA Mini Kit (Qiagen, Hilden, Germany) as per the manufacturer's recommended protocol, followed by quantitative and qualitative measures with a spectrophotometer (Nan-oDrop ® 2000c, Thermo Fisher Scientific, Waltham, Ma, USA).

PCR methods
All DNA samples were screened for the presence of A. phagocytophilum, Babesia spp., B. burgdorferi (s.l.), CNM and Rickettsia spp. by real-time and/or conventional PCRs. Samples positive for B. burgdorferi (s.l.) were additionally processed via multi locus sequence typing (MLST). All samples were moreover examined for Bartonella spp. and Hepatozoon spp. Details on used PCR protocols are presented in Table 7. For the detection of Hepatozoon spp., the initial annealing was changed to 52 °C. All

Statistical analysis
Confidence intervals (95% CI) for the prevalence of pathogens were determined by the modified Wald method using GraphPad Prism v.4 (Graph Pad Software, San Diego, CA, USA). Chi-square and Fisher's tests were used to test the prevalence levels for significant independence. The significance threshold was set at P = 0.05. The prevalence levels for attached larvae are given as MIR (minimal infection rate) as these were pooled.