Frequency of Anaplasma phagocytophilum, Borrelia spp., and coinfections in Ixodes ricinus ticks collected from dogs and cats in Germany

Background Changing geographical and seasonal activity patterns of ticks may increase the risk of tick infestation and tick-borne pathogen (TBP) transmission for both humans and animals. Methods To estimate TBP exposure of dogs and cats, 3000 female I. ricinus from these hosts were investigated for Anaplasma phagocytophilum and Borrelia species. Results qPCR inhibition, which was observed for ticks of all engorgement stages but not questing ticks, was eliminated at a template volume of 2 µl. In ticks from dogs, A. phagocytophilum and Borrelia spp. prevalence amounted to 19.0% (285/1500) and 28.5% (427/1500), respectively, while ticks from cats showed significantly higher values of 30.9% (464/1500) and 55.1% (827/1500). Accordingly, the coinfection rate with both A. phagocytophilum and Borrelia spp. was significantly higher in ticks from cats (17.5%, 262/1500) than dogs (6.9%, 104/1500). Borrelia prevalence significantly decreased with increasing engorgement duration in ticks from both host species, whereas A. phagocytophilum prevalence decreased only in ticks from dogs. While A. phagocytophilum copy numbers in positive ticks did not change significantly over the time of engorgement, those of Borrelia decreased initially in dog ticks. In ticks from cats, copy numbers of neither A. phagocytophilum nor Borrelia spp. were affected by engorgement. Borrelia species differentiation was successful in 29.1% (365/1254) of qPCR-positive ticks. The most frequently detected species in ticks from dogs were B. afzelii (39.3% of successfully differentiated infections; 70/178), B. miyamotoi (16.3%; 29/178), and B. valaisiana (15.7%; 28/178), while B. afzelii (40.1%; 91/227), B. spielmanii (21.6%; 49/227), and B. miyamotoi (14.1%; 32/227) occurred most frequently in ticks from cats. Conclusions The differences in pathogen prevalence and Borrelia species distribution between ticks collected from dogs and cats may result from differences in habitat overlap with TBP reservoir hosts. The declining prevalence of A. phagocytophilum with increasing engorgement duration, without a decrease in copy numbers, could indicate transmission to dogs over the time of attachment. The fact that this was not observed in ticks from cats may indicate less efficient transmission. In conclusion, the high prevalence of A. phagocytophilum and Borrelia spp. in ticks collected from dogs and cats underlines the need for effective acaricide tick control to protect both animals and humans from associated health risks. Graphical Abstract


Background
Recent changes in geographical distribution and seasonal activity patterns of ticks, such as the castor bean tick Ixodes ricinus, may increase the risk of tick infestation and tick-borne pathogen (TBP) transmission for both humans and animals.While in humans Lyme borreliosis caused by species of the Borrelia burgdorferi sensu lato (s.l.) complex is the most frequent tick-borne disease (TBD) in Europe, borreliosis in dogs is less common but may be associated with febrile illness, polyarthritis, and glomerulonephritis [1].Next to species of the B. burgdorferi s.l.complex, B. miyamotoi causing relapsing-fever-like illness is frequently detected in hard ticks from Germany [2,3].Moreover, granulocytic anaplasmosis caused by Anaplasma phagocytophilum is common in dogs [4,5] with increasing incidence and clinical relevance in Germany [6].In severe cases, the Gramnegative and obligate intracellular bacterium can lead to severe febrile, life-threatening illness, sometimes accompanied by central nervous system disorders or gastrointestinal symptoms [7,8].Cats may also be affected, and clinical cases of A. phagocytophilum-positive felines with lethargy, loss of appetite, fever, severe inflammatory processes, and thrombocytopenia are increasingly reported [5,9,10].
Several European studies have examined the prevalence of A. phagocytophilum in field-collected questing ticks, resulting in values between 1.0% in a study from Germany [11] and 40.5% in a study from Denmark [12].Prevalence is typically higher in adult than nymphal ticks, as shown for example in a study from Denmark with infection rates of 14.5% in nymphs compared with 40.5% in adult I. ricinus ticks [12].Local investigations in the northern German cities of Hamburg and Hannover revealed A. phagocytophilum prevalences of 2.1% and 4.5%, respectively [13,14], in field-collected adult I. ricinus, compared with only 1.0% in a study from southern Germany [11].In ticks collected from dogs and cats in Europe, between 1.0% and 22.3% were qPCR-positive for A. phagocytophilum, with most of the studies concentrating mainly on dogs [15][16][17][18][19][20].
Similarly, Borrelia prevalence in questing ticks varies considerably between different regions and countries in Europe, with a mean prevalence of 3.6% in the British Isles to 19.3% in Central Europe [21].Similar to A. phagocytophilum, the average infection rate is higher in I. ricinus adults (14.9%) than in nymphs (11.8%) [21].A study from the City of Hanover, northern Germany, monitored the Borrelia spp.prevalence at the same locations over several years, showing a rather stable prevalence between 22.7% and 25.5%, with significantly higher infection rates in adult ticks than in nymphs, e.g., 32.6% versus 18.4% in 2020 [22].In another study from northern Germany, a prevalence of 31.7% in adults and 28.6% in nymphs was detected [3].Although the clinical relevance of Lyme borreliosis in dogs may be rather low [1] and questionable in cats [23], assessing the prevalence of Borrelia spp. in ticks from dogs and cats does not only shed light on their own infection risk but may also serve as an indicator of human risk.Dogs usually accompany their owners, and in addition, both dogs and cats may introduce ticks into the home, which could represent a human health risk.In ticks collected from dogs, the overall Borrelia prevalence varied between 2.1% in the UK and 44.0% in the Ukraine [24,25].For ticks from cats, one study reported a Borrelia prevalence of 18.0% [26], while another study indicated a prevalence of 10.2% for ticks from both host species [27].
The use of ticks removed from hosts for prevalence studies is not without controversy, as ticks detached from a host may have acquired the infection during the actual bloodmeal or by cofeeding transmission [28].Nevertheless, the main aim of the recent study was to assess the TBP contact risk for dogs and cats and to unravel potential host-related differences.Moreover, taking tick engorgement status into account may help to clarify the direction of pathogen transmission during the bloodmeal ("tick to host" or "host to tick") on the basis of different patterns of prevalence between nonengorged and engorged ticks.A study from northeastern Poland showed differences in pathogen prevalence between the stages of engorgement of ticks collected from dogs [17].While 34.4% of 436 engorged ticks were positive for Borrelia spp., only 4.7% of 21 nonengorged ticks were positive.This increased prevalence in engorged ticks is in contrast to a publication from southern Germany, which reported a prevalence of 33.3% in 26 non-engorged and 13.6% in 110 engorged female I. ricinus [29].
The wide range in reported prevalence values indicates the need for a comprehensive study on ticks collected from dogs and cats to assess their actual infection risk.Therefore, the present study investigated 3000 female I. ricinus specimens removed from dogs and cats during a Germany-wide tick collection study [30].Ticks were selected to equally represent the two host species and different stages of engorgement to allow a meaningful comparison of Borrelia spp.and A. phagocytophilum prevalence and infection intensity in terms of qPCR copy numbers.To exclude interference of engorgement or bloodmeal, respectively, with qPCR sensitivity, a spiking experiment was performed and the PCR setup was adjusted accordingly.

Tick collection study and sample selection
In the frame of a tick collection study, 197 veterinary practices distributed across Germany submitted ticks between March 2020 and October 2021 [30].Information on the clinical presentation of sampled hosts was not available.The submitted specimens were morphologically identified, and female I. ricinus specimens were measured using the OLYMPUS cellSens Entry (v.3.2) software paired with an OLYMPUS SC50 camera adapter to determine the coxal index as an estimate of attachment duration [30,31].On this basis, female I. ricinus were assigned to one of four engorgement categories (non-engorged [0 h to ≤ 24 h], partially engorged stage one [> 24 h to ≤ 72 h], partially engorged stage two [> 72 h to ≤ 144 h], fully engorged [> 144 h]), which were adapted from Franta et al. [32] by dividing the period of "slow feeding" into two stages for a better differentiation within this long time span (> 24 h to ≤ 144 h).Ticks with implausible calculated engorgement times > 245 h were excluded.The ticks were then stored individually at −20 °C until further use.
From the total number of 14,383 received I. ricinus specimens, a representative sample of 3000 females, 1500 from dogs and 1500 from cats, was selected for pathogen analyses.Per host species, 375 specimens of each engorgement category (non-engorged, partially engorged stage one, partially engorged stage two, and fully engorged) were included.Within these categories, further selection according to the coxal index was performed to represent the stages of engorgement as evenly as possible.Additionally, ticks from dogs and cats infested with a single or multiple ticks (hereafter single and multiple infestations) were selected approximately equally as well, with 1-5 ticks analyzed per multiple infested animal.Within these categories, the samples were further chosen via randomized numbers.The origin of the selected samples was plotted on a map of Germany using R version 4.1.0[33] with geographic data distributed by OpenStreetMap under the Open Database License (www.opens treet map.org/ copyr ight) to exclude a geographic bias (Fig. 1).

Genomic DNA isolation
Ticks were homogenized in 20 µl phosphate buffered saline with polystyrene pistils (Carl Roth GmbH, Karlsruhe, Germany) in individual tubes, before addition of 25 µl proteinase K and 200 µl BQ1 buffer from the Nucleospin ® 96 Blood Core Kit (Macherey-Nagel GmbH & Co KG, Düren, Germany) and overnight incubation at 56 °C.Further, the DNA isolation followed the instructions of the Nucleospin ® 96 Blood Core Kit, with minor modifications as described [14,34,35].The 100 µl eluate of genomic DNA was stored at −20 °C until further use.

Detection of A. phagocytophilum and Borrelia spp. by quantitative real-time PCR evaluated by spiking experiments
The success of the DNA isolation was checked randomly for 100 ticks each from dogs and cats by targeting the I. ricinus internal transcribed spacer (ITS) 2 region in probe-based qPCR [36], including a serial plasmid standard (10 0 -10 6 copies per reaction, ligated into TOPO ™ TA vectors [Invitrogen ™ , Thermo Fisher Scientific, Schwerte, Germany]) and a negative control.Because of the high concentration of tick DNA from engorged specimens, which may inhibit exponential PCR amplification [37], 1 µl tick DNA template was used in the 25 µl reaction volume for ITS2 detection.The Mx3005 multiplex quantitative polymerase chain reaction (qPCR) system (Agilent Technologies Inc., Santa Clara, USA) was used for all qPCR reactions.The reaction setup has been described in detail by Strube et al. [36] with previously published adaptations by May and Strube [13].The thermoprofile for both reactions included initial denaturation at 95 °C for 15 min followed by 45 cycles of 95 °C for 15 s and 60 °C for 60 s.
For pathogen detection, a duplex probe-based qPCR was performed targeting the msp2/p44 multi copy gene of A. phagocytophilum [38] and the 5S-23S rRNA intergenic spacer (IGS) region [36], which occurs in duplicate in the B. burgdorferi s.l.complex and as a single copy in B. miyamotoi [39].The qPCR setup included a negative control and serial plasmid standards (10 0 -10 6 copies per reaction, ligated into TOPO ™ TA vectors [Invitrogen ™ , Thermo Fisher Scientific, Schwerte, Germany]) for quantification.All samples were tested in duplicate.Initially, 10 µl template was used in the 25 µl reaction volume, while negative samples were later retested using only 2 µl on the basis of the result of a spiking experiment performed to exclude possible false-negative results due to PCR inhibition in engorged ticks.The spiking experiment was performed with six samples from each engorgement category and each host species (n = 48), which showed a negative qPCR result based on 10 µl template.Of these samples, 10, 5, and 2 µl template were tested as described above (if enough template was left), using a mastermix spiked with the Borrelia spp.and A. phagocytophilum plasmid standard, amounting to 10 4 plasmid copies each per qPCR reaction.Further, a positive control containing only the plasmid standards but no template was included to generate reference cycle threshold (Ct) values.A second spiking experiment was conducted using 10 µl and 2 µl DNA of 22 non-engorged laboratory-bred I. ricinus females as template.

Borrelia species identification by reverse line blot (RLB)
For all Borrelia qPCR-positive ticks, a conventional PCR targeting the 5S-23S ribosomal IGS region with primers B5S-Bor, BMiya-For, and 23S-Bor was performed [40,41], followed by RLB differentiation with 11 different Borrelia species-specific probes as described before [40,42].Extrapolated from the spiking experiments (see section above), the amount of template was reduced from 10 µl to 2 µl for samples that were qPCR-positive with 2 µl but not 10 µl template volume, to avoid potential interference and inhibition in the RLB as well.

Statistical analyses
Statistical analyses were performed using R version 4.1.1717[33].The A. phagocytophilum and Borrelia spp.prevalence, the coinfection rate, and the number of mixed infections with different Borrelia species was compared between ticks from dogs and cats via χ 2 test or Fisher's exact test in case of small values.
Further, generalized linear mixed models (GLMM) with binomial error structure were constructed separately for ticks from dogs and cats to analyze the influence of different predictors on the A. phagocytophilum and Borrelia spp.prevalence.As fixed factors, the tick's stage of engorgement, whether the tick was from a single or a multiple infestation, and its coinfection status with A. phagocytophilum or Borrelia spp., respectively, were included in the model, while host animal ID was entered as a random factor.Multiple comparisons between the different stages of engorgement were performed using the function "glht" from the package "multcomp" [43] with Tukey contrasts.The full models were compared with null models based only on the random factor using likelihood ratio tests (R function "anova").
The detected copy numbers for the msp2/p44 gene of A. phagocytophilum and the 5S-23S rRNA IGS region of Borrelia spp.(in the following referred to as A. phagocytophilum and Borrelia spp.copy numbers) were compared among the different engorgement categories of ticks from dogs and cats, respectively, via Kruskal-Wallis rank-sum tests followed by Dunn's all-pairs rank-comparison test in case of a significant result.For all samples, the obtained copy numbers were extrapolated to 100 µl to quantify the copy number per examined tick.
Finally, the overall Borrelia species distribution was compared between the different engorgement categories of ticks from dogs and cats, respectively, by Fisher's exact tests with simulated P-values.In case of a significant result, the prevalence of each species was compared between the engorgement categories by individual Fisher's exact tests, followed by Bonferroni-Holm correction of P-values.

Spiking experiments to evaluate pathogen qPCR detection
Using 10 µl template, 35.4% (17/48) and 29.2% (14/48) of the spiked tick DNA samples from dogs and cats tested negative for A. phagocytophilum and Borrelia spp., respectively, with no differences in the detection rate between ticks from dogs and cats.False negatives occurred independently of the engorgement duration.Further, Ct values were unaffected by the engorgement state.With decreasing amount of template, the number of false-negative results decreased, while Ct values of positive samples stayed mostly consistent, indicating that qPCR inhibition was not gradual but occurred in a "yes or no" manner.With 5 µl template, detection of A. phagocytophilum was still not successful in 2.1% (1/48) of all tick samples, while 18.8% (9/48) of the tick samples still yielded a false-negative Borrelia result.With 2 µl template, in all tick samples from which template was left (40/48; 83.3%), no PCR inhibition was noted for A. phagocytophilum nor Borrelia species.
When using DNA from unfed laboratory-bred ticks, no difference between 10 µl or 2 µl template was noted, as in both spiking experiments the detection rate was 100% (22/22) with Ct values consistent with the positive control.

Pathogen prevalence in ticks collected from dogs and cats
Detection of the I. ricinus ITS2 region indicated successful DNA isolation of all randomly selected samples.However, as the spiking experiment revealed inhibition of the qPCR when using 10 and 5 µl template of ticks detached from dogs and cats independent of engorgement status, all tick samples initially negative using 10 µl were retested with 2 µl template only.
Further, the GLMM revealed a significant positive association of A. phagocytophilum and Borrelia spp. in ticks from dogs, but not in those from cats (Table 2).There was no distinct geographical pattern in prevalence, with infected ticks originating from all over the country (Fig. 1).

Prevalence and copy numbers of A. phagocytophilum and Borrelia spp. in relation to tick engorgement status
In ticks collected from dogs, the highest prevalence of A. phagocytophilum as well as Borrelia spp. was measured in non-engorged specimens, decreasing significantly with increasing time of engorgement (Fig. 2).Regarding A. phagocytophilum, the prevalence amounted to 25.1% (94/375) in non-engorged ticks but only to 14.9% (56/375) in fully engorged ticks, with a significant difference between non-engorged ticks and partially engorged   ticks from stage 2 or fully engorged specimens, as well as between partially engorged ticks from stage 1 and stage 2 (Table 2).The median A. phagocytophilum copy numbers varied between 6.96 × 10 1 in non-engorged and 1.80 × 10 2 in fully engorged ticks, but the increase with increasing engorgement time was not statistically significant after P-value correction (Kruskal-Wallis test, χ 2 = 8.52, df = 3, P 0.036; Fig. 2, Dunn's post hoc tests, P = 0.077 between non-engorged and fully engorged ticks).Concerning Borrelia spp., with 48.5% (182/375) of the non-engorged ticks from dogs being qPCR positive but only 12.0% (45/375) of the fully engorged ticks, all stages of engorgement showed significantly different prevalence values compared with non-engorged ticks and the other engorgement stages (Table 2).Seven different Borrelia species were detected in non-engorged ticks, six each in partially engorged ticks of phase 1 and phase 2, and two, precisely B. afzelii and B. miyamotoi, in fully engorged specimens (Fig. 3).A significant difference in prevalence between the engorgement categories was only detected with regard to B. miyamotoi (Fisher's exact test, Bonferroni-Holm-corrected P = 0.007).Further, median Borrelia spp.copy numbers varied between 5.68 × 10 2 copies in non-engorged and 3.75 × 10 2 copies in fully engorged ticks (Fig. 2).Additionally, a significant influence of the time of engorgement on the Borrelia copy numbers was detected (Kruskal-Wallis test, χ 2 = 13.54,df = 3, P = 0.004), with significantly higher copy numbers in nonengorged specimens compared with partially engorged ticks of stage 1 (Dunn's post hoc test, P = 0.003).In ticks from cats, no significant difference in A. phagocytophilum prevalence between the stages of engorgement was noted, with a prevalence of 33.6% (126/375) in fully engorged ticks and 31.7%(119/375) in non-engorged specimens (Table 2).Furthermore, no significant influence of the time of engorgement on the Anaplasma copy numbers was detected (Kruskal-Wallis test, χ 2 = 3.44, df = 3, P = 0.328) with a median of 8.85 × 10 1 copies in non-engorged and 1.52 × 10 2 copies in fully engorged ticks (Fig. 2).

Pathogen prevalence in single and multiple infested dogs and cats
In total, ticks from 535 single infested dogs and 371 cats were selected for pathogen detection, as well as a total of 965 ticks from 474 multiple infested dogs, and 1129 ticks from 527 multiple infested cats.
For Borrelia spp., 24.3% (130/535) of the single infested dogs carried infected ticks and 49.6% (184/371) of the single infested cats.Concerning  5).There was no significant difference in prevalence of A. phagocytophilum nor Borrelia spp. between ticks from single and multiple infestations of dogs, but both occurred significantly more often in ticks from multiple than single infested cats (Table 2).
Considering only cases of multiple infested animals, the number of actually examined ticks varied from one to five.From 450 dogs, more than one tick was examined, of which 20.0% of the examined ticks (188/941) from 148 dogs were A. phagocytophilum positive.Concerning cats, 492 animals contributed more than one tick to pathogen detection, of which 364/1.096 (33.2%) specimens from 262 cats were A. phagocytophilum positive.The distribution of dogs infested with qPCRpositive ticks was as follows: In 73.7% (109/148) of cases, only a single tick was A. phagocytophilum positive, while two ticks were positive in 25.7% (38/148) of cases and three ticks in only one case (0.7%).Regarding cats, 66.0% (173/262) contributed a single, 30.5% (80/262) contributed two, and 3.4% (9/262) contributed three or more A. phagocytophilum-positive ticks (Table 6).

Discussion
Owing to climatic and environmental changes over the last decades, the geographical and seasonal expansion of ticks toward year-round activity translates to an increased overall risk of infestation and thus infection with tick-associated pathogens such as A. phagocytophilum and Borrelia spp., endangering human and animal health.A tick submission study from Germany revealed a long average time of engorgement of female I. ricinus ticks on dogs (78.8 h) and cats (82.7 h), illustrating the high risk of pathogen transmission [30].Especially regarding cats, the current risk of infection with TBDs is probably underestimated, as a rising number of feline TBD case reports illustrates [9,44].
In previous studies on questing female I. ricinus collected from vegetation in Germany, A. phagocytophilum prevalences between 1.0% [11] and 11.6% [45] were detected, whereas a comparably high prevalence of 40.5% was reported from Denmark [12].In previous studies examining ticks from dogs and cats, between 0.1% female I. ricinus from the UK [15], 3.5% from Finland [20], and 22.3% from Poland [16] were found to be A. phagocytophilum positive.In the present study examining exclusively female I. ricinus ticks, the prevalence in ticks from dogs was quite high compared with these values, with every fifth tick positive, and even higher in ticks from cats with almost every third tick positive.The quite high A. phagocytophilum prevalence might additionally be influenced by the significant positive association with Borrelia spp.infection status, and vice versa.Enhanced acquisition of both organisms by larval ticks from coinfected mice has been shown, related to higher bacterial burdens [46].Moreover, coinfections with B. burgdorferi s.l. and A. phagocytophilum were frequently detected in I. ricinus collected from European hedgehogs [47].
Concerning Borrelia spp., most studies from Germany and neighboring countries have determined a prevalence around 35.0% in questing, non-engorged female I. ricinus  [3,17,29,40,48,49], while the prevalence reached from 1.0% in a study examining adult I. ricinus ticks from dogs sampled in four European countries [26], over 10.5% in a recently published study from Finland [20], to 44.0% in adult engorged I. ricinus ticks from wild and domestic animals in the Ukraine [24].Therefore, the prevalence of 28.5% in female I. ricinus ticks collected from dogs in the present study is similar to that in questing ticks, but considerably lower than the detected infection rate in ticks from cats, with more than every second tick positive for Borrelia species.The fact that the seroprevalence of anti-A.phagocytophilum and anti-Borrelia spp.antibodies in dogs varies enormously in Europe between 1.1% and 56.5% as reviewed by [50], or 1.1% and 17.9% as reviewed by [51], additionally shows that regional investigations can hardly be extrapolated, underlining the importance of nationwide surveys.In the present study, however, the investigated pathogens were detected more or less evenly across Germany.
To evaluate whether the detected prevalence might be influenced by the ticks' engorgement status, spiking trials were performed to exclude possible PCR inhibition due to the bloodmeal, although a DNA isolation kit specifically designed to inactivate PCR inhibitors from blood and other body fluids had already been used.The spiking trials showed no influence of the stage of engorgement, but inhibitors actually did interfere with the qPCR process as false-negative results were obtained in nonengorged as well as engorged ticks when using 10 µl and 5 µl template volume.Interestingly, copy numbers were mainly not affected, being comparable in almost all samples with positive results.Thus, the inhibition does not seem to skew qPCR quantification, but suppresses the detection in general in certain cases similar to a "yes or no" answer.The Borrelia qPCR seemed to be more susceptible to inhibition than the A. phagocytophilum qPCR.Differential susceptibility of different PCR assays to inhibitors has been reported previously [52].Of note, inhibition was only observed when using isolated DNA from host-detached ticks, but not when using DNA from questing ticks as template, suggesting that the initiation of feeding activates or introduces an inhibitor in the tick.If this inhibitor has a biological function, e.g., in the context of an immunomodulation at the location of the tick bite, needs to be investigated further.Therefore, studies on host-detached ticks may be prone to false-negative results at certain template quantities, which should be carefully evaluated in future studies.In the present study, the qPCR was repeated with reduced template volume (2 µl) for every negative tick afterwards, thereby detecting more A. phagocytophilum-and Borrelia-infected ticks in every engorgement category, but the general pattern over the stages of engorgement nevertheless remained as initially detected.
Concerning the overall prevalence of the pathogens detected, the significant differences that were determined between ticks from dogs and cats may be related to the different outdoor behavior and the resulting degree of habitat overlap with or contact to TBP reservoir hosts.Contact of cats with hedgehogs and other small mammals, acting as reservoir hosts for A. phagocytophilum and a variety of Borrelia species [53][54][55], may be provoked by their unrestricted hunting drive.This is further supported by a veterinary submission study showing differences in the distribution of tick species infesting dogs and cats, with 5.0% of the collected ticks identified as the hedgehog tick Ixodes hexagonus representing the second most frequently detected tick species on cats, while in dogs only 1.6% of the collected ticks were identified as I. hexagonus [30].Further, infected ticks might show a stronger affinity to cats than dogs, however, this speculation requires further investigations.
Differences between ticks from dogs and cats were also apparent regarding the Borrelia species distribution, further underlining the hypothesis that both host species have a different probability of acquiring ticks that previously fed on certain reservoir hosts.In ticks from both host species, B. afzelii was detected most frequently, in line with former studies on questing and/or host-associated ticks [3,27,40,48,56].However, B. miyamotoi and B. valaisiana were more frequently detected in ticks from dogs than B. garinii, which was the second most frequently detected species in the above-mentioned previous studies, whereas in ticks from cats, B. spielmanii and B. miyamotoi were more frequent than B. garinii.Further, B. valaisiana, which is associated with birds or lizards as reservoir hosts [57], was detected significantly more frequently in ticks from dogs than cats, while cats harbored significantly more ticks infected with B. spielmanii, whose main reservoir hosts are rodents but which is also found in hedgehogs [54,58].
While in ticks from dogs, the Borrelia species was successfully determined in more than every third tick by RLB, the differentiation rate was lower for ticks from cats, with only every fourth tick successfully examined.This is most likely related to the fact that 430/827 Borrelia-positive ticks from cats contained only very low copy numbers (≤ 10 1 copies), hampering detection by RLB, which is less sensitive than qPCR.For both host species, the overall differentiation success rate was comparable to those from previous studies (31.6%-58.0%)examining questing [3,40,49] and host-associated ticks [17].
To examine whether dogs and cats might act as reservoir hosts themselves, infecting the infesting ticks, or whether tick-to-tick transmission by cofeeding might take place during the bloodmeal, pathogen prevalence between ticks from single and multiple infestations as well as between different stages of engorgement was compared.Concerning cats, a multiple infestation type significantly raised the odds of a tick to be pathogenpositive compared with single infestations, while in dogs multiple infestations had no significant influence on the detected pathogen prevalence.Further, in those cases where more than one tick was examined per host, most dogs carried just a single infected tick, while only a quarter harbored more than two to three infected ticks at the same time.In cats that contributed multiple ticks to the analysis, the proportion of animals with more than one infected tick was approximately one-third for A. phagocytophilum and one-half for Borrelia spp., however, this distribution remained the same when excluding nonengorged specimens, which have had little or no chance yet to become infected by the current bloodmeal.Thus, a reservoir function of the cat itself is unlikely, but the data rather suggest a higher pathogen prevalence in ticks from the cats' environment than in the natural habitat of dogs.Further, the attachment sites of the individual ticks argue against a potential influence of cofeeding, as in 55.9% of cases with two or more positive ticks per animal, those ticks were detected on separate body parts (data not shown).
Moreover, a decreasing rather than an increasing frequency of infected ticks was noted with increasing engorgement duration, except for A. phagocytophilum in ticks from cats, where no significant differences were visible at all.One of the main hypotheses for an increasing Borrelia prevalence over the time of engorgement is feeding-induced multiplication of spirochetes [59].According to that, a study by Michalski et al. [17] showed a higher prevalence of Borrelia spp.and A. phagocytophilum in engorged females compared with non-engorged specimens collected from dogs, although the sample sizes of examined non-engorged ticks were very small with 21 specimens for Borrelia spp.and 11 for A. phagocytophilum compared with a higher sample size of 436 and 402 engorged specimens, respectively, in that study.In contrast, a study from the Netherlands showed a decreasing Borrelia prevalence from non-engorged (11.8%; 10/85) to partially engorged ticks (1.9%; 1/53) collected from dogs [60], comparable to the findings of the present study.Other studies also described a decreasing prevalence with increasing time of engorgement [29,61].This might be related to an early transmission of Borrelia spirochetes, an elimination by defecation or a destruction during the time of transfer from midgut to salivary glands [60,62].Further reasons might be a dilution effect due to the engorged blood or complement-mediated Borrelia clearance, as has been described for ruminant blood [63].However, complement-mediated clearance seems less likely here, as most members of the B. burgdorferi s.l.complex are resistant to serum complement of dogs and cats, except for an intermediate susceptibility of B. garinii and B. lusitaniae as reviewed by [64].Further, there are strain-specific differences in serum susceptibility, and a German B. garinii strain (Pbi) showed a high resistance to dog serum in another study [65].Moreover, in adult female ticks removed from humans, both B. garinii and B. afzelii bacterial load decreased with increasing engorgement duration, with a stronger decrease of B. afzelii [66], despite the fact that B. garinii but not B. afzelii is considered susceptible to human complement [64], indicating that immune-mediated clearance probably only played a minor role for the decrease in bacterial load.The Borrelia copy numbers in fully engorged ticks in the present study did not differ from those in non-engorged ticks, but a significant decrease of copy numbers was observed between non-engorged ticks from dogs and stage 1 partially engorged specimens.This may indicate efficient transmission of Borrelia spirochetes to dogs within the early phase of the bloodmeal, supported by the biggest drop of prevalence between these two engorgement stages.The result is comparable to the decrease of B. afzelii and B. garinii load in ticks having fed for more than 36 h on humans, which the authors linked to possible pathogen transmission [66].Further, neither multiplication of the Borrelia spp. in the midgut nor a dilution effect are evident from the present data, as no significant increase nor decrease of the detected copy numbers occurred after this initial decrease.As a cautionary note, it should be kept in mind that the investigated target is a multicopy gene in B. burgdorferi s.l.but a single-copy gene in B. miyamotoi, indicating that the copy numbers are not necessarily linearly related to bacterial load.
Similar to the picture in ticks from dogs, the frequency of Borrelia-infected ticks decreased significantly, but not as sharply, in ticks from cats.Further, in ticks from cats, no significant decrease of the detected copy numbers was evident, possibly indicating a less efficient transmission of Borrelia spirochetes compared with the findings in ticks from dogs, which might go hand in hand with minor clinical symptoms in cats.
With regard to A. phagocytophilum, only Michalski et al. [17] compared non-engorged and engorged ticks from dogs.Further studies have been performed on ticks from ruminants, such as roe deer, showing a distinct A. phagocytophilum prevalence increase with increasing time of engorgement [37], which, however, might be related to the reservoir function of wild ruminants as reviewed in [67].In the present study, a distinct decrease of the A. phagocytophilum prevalence between stage 1 (24-72 h) and stage 2 (72-144 h) partially engorged ticks from dogs was observed, which could possibly indicate the time window when most transmission takes place.Even if the transmission of A. phagocytophilum can already occur after 6 h of feeding, although seldomly resulting in a sufficient infection [68], it is commonly agreed that the onset of transmission occurs around 36 h after the beginning of the feeding process [69].Nevertheless, the A. phagocytophilum copy numbers were not significantly affected by increasing time of engorgement in ticks from dogs nor in ticks from cats.Only in ticks from dogs, a slight increase might indicate that the pathogen is replicating.
Cats also did not show a decline in A. phagocytophilum prevalence, which suggests different transmission kinetics related to the different host species.In general, dogs are regarded as more suitable A. phagocytophilum hosts compared with cats as reviewed in [67].Further studies are necessary to establish the cause of the decreasing prevalence in ticks of dogs with certainty as well as why this decrease was not or not as apparent in ticks from cats and to explain the observed differences between dogs and cats.

Conclusions
The high prevalence of A. phagocytophilum and Borrelia spp.detected in ticks from dogs and cats in this study underlines the high risk of pathogen transmission and thus the need for efficient tick protection via licensed acaricides, as e.g.recommended by the European Scientific Counsel Companion Animal Parasites (ESCCAP), to protect the health of dogs and cats with regard to clinical cases of anaplasmosis or borreliosis.Even if the relevance of Borrelia spp. in veterinary medicine is lower than in human medicine, ticks that are introduced into the human environment with companion animals and then wiped off may represent a risk for humans.Thus, tick protection of dogs and cats is also of public health relevance within a One Health context.The significant differences between dogs and cats concerning the overall pathogen prevalence, coinfections, and Borrelia species distribution are indicative of behavioral differences resulting in varying degrees of contact or habitat overlap with different TBP reservoir hosts.Decreasing prevalence of A. phagocytophilum as well as Borrelia spp. in ticks from dogs with increasing time of engorgement may indicate successful transmission, while cat blood might negatively affect the efficiency of A. phagocytophilum transmission.This could be a further explanation for the low number of clinical cases and/or less pronounced clinical symptoms in cats.However, further studies are needed to unravel the differences between dogs and cats in this regard.

Fig. 1
Fig. 1 Nationwide distribution pattern of ticks positive for A. phagocytophilum (shades of red) and Borrelia spp.(shades of blue) from dogs (A/C) and cats (B/D).The shading indicates the proportion of positive ticks per sampled zip code.Zip code areas that were not sampled are shown in gray

Fig. 2
Fig. 2 Differences in the frequency of ticks infected with A. phagocytophilum (A) and Borrelia spp.(C) as well as A. phagocytophilum msp2/p44 gene (B) and Borrelia 5S-23S IGS copies (D) per stage of engorgement in ticks from dogs (red/royal blue) and cats (light red/light blue).Boxes extend from the 25th to the 75th percentile, with a line at the median and whiskers extending to 1.5 the interquartile range.Significant differences (P < 0.05) are indicated by an asterisk

Fig. 3
Fig. 3 Borrelia species distribution among successfully differentiated ticks from dogs (A) and cats (B) per stage of engorgement.Note the different sample size of positive ticks for each engorgement category indicated above the bars

Table 1
Anaplasma phagocytophilum and Borrelia spp.copy number distribution among positive female I. ricinus collected from dogs and cats

Table 2
Results of the GLMM testing the influence of several predictor variables on the prevalence of A. phagocytophilum (model A and B) and Borrelia spp.(model C and D) infections in female I. ricinus collected from dogs and cats.The models were significantly different from a null model containing only the animal identification number as a random factor (A: chi-square = 26.36,df = 5, P < 0.001; B: chi-square = 15.14, df = 5, P = 0.01; C: chi-square = 145.64,df = 5, P < 0.001; D: chi-square = 30.63;df = 5, P < 0.001).Significant P-values are shown in bold [42]tandard error, GLMM generalized linear mixed model a Multiple comparisons were conducted using the function "glht" from the R package "multcomp"[42]with Tukey contrasts

Table 3
Borrelia species distribution among qPCR-positive female I. ricinus from dogs and cats.Significant P-values (Bonferronicorrected) are shown in bold n.a.not applicable for Fisher's exact test

Table 5
Prevalence of A. phagocytophilum and Borrelia spp. in female I. ricinus from single and multiple infested dogs and cats

Table 6
Distribution of animals from which multiple female I. ricinus were examined according to the number of positive ticks per host.The upper part of the table includes only engorged ticks, while all ticks are included in the lower part