Revisited: Borrelia burgdorferi sensu lato infections in hard ticks (Ixodes ricinus) in the city of Hanover (Germany)

Background The present study investigated the prevalence of Borrelia burgdorferi sensu lato (s.l.) genospecies in Ixodes ricinus ticks collected in Hanover, Northern Germany, in 2010. At the same time the study served as fifth-year-follow-up study for data comparison with 2005. Methods A total of 2100 questing ticks were collected and analysed by quantitative real-time PCR (qPCR) with subsequent species differentiation via Reverse Line Blot and Sanger sequencing. Simultaneously, results obtained in 2010 were compared to infection rates from 2005 to evaluate the development of B. burgdorferi s.l. infection rates in Hanoverian ticks. Results Overall, 22.7% (476/2,100) of collected ticks were tested positive for B. burgdorferi s.l. infections. Adult ticks showed an infection rate of 33.3% (124/372), subdivided into 29.6% (58/196) positive males and 37.5% (66/176) positive females. Nymph and larvae infection rates were found to be 20.3% (344/1,697) and 25.8% (8/31), respectively. Species identification was successful for 59.2% (282/476) of positive ticks with B. afzelii as the most frequently detected genospecies, followed by B. garinii (including B. bavariensis) and B. spielmanii. B. burgdorferi sensu stricto (s.s.), B. bissettii, B. valaisiana and B. lusitaniae were also identified. Significant differences concerning seasonal fluctuations as well as local differences were observed. Comparing infection rates of Hanoverian ticks between years, a significant increase (P = 0.002) could be observed for larvae with 1.7% positives (2/60) in 2005 and 25.8% positives (8/31) in 2010. In the latter year, coinfections with Borrelia and Rickettsiales were detected in a total of 7.8% (163/2,100) of collected ticks. Of these, 7.3% (153/2,100) were coinfected with Rickettsia spp., 0.3% (7/2,100) with A. phagocytophilum and 0.1% (3/2,100) were coinfected with all three pathogens. Between years 2005 and 2010, no statistically significant differences in coinfection rates were found. Conclusions Comparing B. burgdorferi s.l. infections in Hanoverian I. ricinus ticks in 2010 with data from 2005, a statistically significant increase of infected larvae was noted, whereas the other stages revealed no statistically significant differences. Whether the increased larvae infection rate is an isolated event or results from factual circumstances, e.g. increasing effectiveness of transovarial transmission due to unknown factors, has to be evaluated in further studies.

In Europe, several reservoir hosts for Lyme diseaseassociated Borrelia spp. are known as, for example, small rodents, hedgehogs, squirrels, lizards and various bird species [16,[19][20][21][22][23]. Transmission of B. burgdorferi s.l. from ticks to hosts may occur through different developmental stages of I. ricinus, as transstadial transmission is very efficient. Even though larvae are occasionally affected, transovarial transmission in ticks seems to be inefficient [24]. Prevalence rates of B. burgdorferi s.l. in I. ricinus range from 6.1% in France to 22.5% in Western Switzerland [14,[25][26][27][28]. In Germany, prevalences of Borrelia-positive questing I. ricinus ticks range from 3.1% in Northern Germany to 27.0% in Thuringia or locally up to 37% in Bavaria [13,[29][30][31][32][33]. The present study was conducted to determine the Borrelia infection rate in Hanoverian ticks in 2010 and at the same time to serve as fifth-year-follow-up study to monitor changes or stagnation of tick infection rates when compared to data from 2005 [9]. For this, 2,100 ticks were collected in different recreational areas in 2010 in the northern German city of Hanover and subsequently analysed for B. burgdorferi s.l. infections.

Tick material
Questing ticks were collected each month from April to October 2010 in ten different recreation areas in the city of Hanover, the capital of the Northern German federal state Lower Saxony [34]. The city of Hanover was elected as the "German Capital of Biodiversity" in 2011 and is nicknamed "The green metropolis" as it houses numerous parks and the largest continuous urban woodland in Europe. At each sampling site, 30 ticks were collected per month resulting in 210 ticks per defined location over the sampling period and a total of 2,100 ticks overall. Tick species and developmental stages were determined microscopically based on morphological parameters [35].
DNA isolation and detection of Borrelia spp.
Genomic DNA isolation was carried out as described previously [34,36]. Genomic DNA was eluted twice with 70 μl and 60 μl double-distilled water, respectively, to obtain a final volume of 100 μl genomic DNA. Detection of B. burgdorferi s.l. was achieved by minor groove binder probe-based qPCR targeting the 5S-23S intergenic spacer (IGS) region described by Strube et al. [37] with the modification that Absolute Blue QPCR low ROX mix (Thermo Fisher; containing Thermo-Start™ DNA Polymerase) was used and tick DNA template was increased to 10 μl. All qPCR reactions were performed as duplicates and each run included a negative template control as well as plasmid standard positive controls containing 10 0 − 10 6 copies of the Borrelia 5S-23S IGS and Ixodes ITS2 region, respectively [37]. Samples were resubjected to qPCR analysis, if only one well was determined positive and species identification was unsuccessful with the Reverse Line Blot (RLB). If no amplification could be detected in the second qPCR run, samples were considered as questionable positives.
Identification of B. burgdorferi s.l. genospecies by Reverse Line Blot (RLB) and Sanger sequencing Species identification of Borrelia-positive tick samples was achieved by RLB. Sanger sequencing was additionally applied to samples that were determined B. garinii-positive in RLB to further differentiate between B. garinii and B. bavariensis.
RLB was preceded by amplification of the B. burgdorferi s.l. 5S-23S IGS region using biotin linked forward primer 5SCB and reverse primer 23SN2 (10 pmol each) as described by Rijpkema et al. [38]. For the 25 μl reaction set up, 12.5 μl Thermo-Start™ PCR Master Mix (Thermo Scientific, Surrey, England), 1 μl of each primer and 5 μl tick DNA template were added to the corresponding amount of H 2 O. Cycling conditions were based on the protocol by Burri et al. [39] with addition of a polymerase activation step (15 min, 94°C). Each run included positive controls using template DNA of the following Borrelia  [41]. Moreover, a modified probe for B. burgdorferi s.l. (SL2; 5'-[AmC6T]-CCATATTTTTATCTTCCATCTCTA-3'; 500 μM) was added as positive control for successful hybridization procedure. Furthermore, a probe for relapsing fever-like spirochetes (RFLNE; 250 μM) [40] was added to the RLB. All reactions were performed as duplicates.
Following sample hybridization as described by Rijpkema et al. [38], the membrane was washed with 2× SSPE-0.5% SDS at 45°C for 15 min, then incubated with streptavidinperoxidase for 30 min at 42°C, and finally washed with 2× SSPE-0.5% SDS for 10 min at 42°C. Chemiluminescent hybridization signals were achieved by use of ECL detection reagent (GE Healthcare) and detected with Bio Imaging System MF-ChemiBIS 3.2 (Biostep, Jahnsdorf, Germany) during 2 to 10 min exposition (depending on signal strength). Reactions were performed as duplicates and all samples were tested at least twice.
To further differentiate between B. garinii and B. bavariensis, GA positive samples were reamplified and products showing visible gel bands were custom sequenced. Identification of those two species was based on four SNPs occurring in the 5S-23S IGS region of B. bavariensis and B. garinii ( Figure 1). Data was statistically analysed by application of chisquare test followed by Yates correction with SigmaStat® software (version 3.11) with subsequent Bonferroni-Holm correction. For analysis, α was defined with 0.05 and H 0 was rejected if P ≤ 0.05. Ticks classified as questionable were included as positives in statistical analysis.

Comparison of B. burgdorferi s.l. tick infection rates in 2010 vs. 2005
To compare results concerning infection rates of different tick stages obtained in 2010 to those obtained in 2005, questionable results were added to positive tested ticks in 2005 [9]. As a result, modified 2005 data was used for analysis as follows: 153 positive adults divided into 75 male and 78 female ticks (75 positive and 3 questionable), 96 nymphs (92 positive and 4 questionable) and 2 larvae (1 positive and 1 questionable). Data of adult and nymphal stages were statistically analysed as described above. Final significances were based on corrected α-values. To compare larval stages collected in 2005 with those collected in 2010, Fisher's exact test was used due to low sample size. Final significances were based on corrected α-values.

Coinfections of ticks with B. burgdorferi s.l. and Rickettsiales
To analyze the tick coinfection rates with Borrelia and Rickettsiales, obtained data on B. burgdorferi s.l. infections were compared with data on infections with Rickettsiales published previously [34]. Statistical analysis of present coinfections with Rickettsia spp. or A. phagocytophilum and comparison with data from 2005 was conducted as described above. Comparison of coinfections with all three pathogens between years was carried out by using Fisher's exact Test (SigmaStat® software version 3.11).

Results
Collection of 2,100 questing ticks resulted in 372 adult ticks (196 males and 176 females), 1,697 nymphs and 31 larvae all identified as I. ricinus. A total of 22.7% (476/ 2,100) of collected ticks was found to be infected with B. burgdorferi s.l.. Male adults showed an infection rate of 29.6% (58/196) whereas 37.5% (66/176) of female adults were determined positive resulting in an overall infection rate of 33.3% (124/372) for adult ticks. Nymphs showed an infection rate of 20.3% (344/1,697) and larval infection rates were determined as 25.8% (8/31). Statistically significant differences concerning developmental stages were observed between adults and nymphs (P < 0.001; α loc = 0.017). On level below, adult males (P < 0.001; α loc = 0.0083) as well as adult females (P = 0.003; α loc = 0.01) were significantly more often infected than nymphs. A detailed overview of infection rates of different stages during the collection period is shown in Table 1. Over the collection period, two main infection peaks of ticks could be observed with a first increase in June (25.3%; 76/300) and July (28%; 83/300) as well as a second peak in October (32.7%; 98/300), the latter is simultaneously the highest seasonal prevalence. The lowest Borrelia-prevalence was determined in May (17.7%; 53/300). Statistically significant differences were observed between October and the months of April, May and August. P-values and adjusted α-values are shown in Figure 2.

Discussion
The hard tick I. ricinus serves as main vector for spirochetes of the B. burgdorferi s.l. complex in Central Europe. Parts of this complex are responsible for Lyme disease, a sickness that has increased rapidly during the past 20 years in the northern hemisphere [5,44]. In several studies, infection rates of ticks with B. burgdorferi s.l. in Germany were investigated to assess the potential infection risk for humans resulting in a broad range of infection rates ranging from 3.1% in Northern Germany to 27.0% in Thuringia and 36.2% in Bavaria [29][30][31][32][33]. Besides data collection in different geographical regions, it is of importance to monitor tick infection rates over time to assess whether human infection risk increases or decreases. Thus, the present study served not only as a status survey for the Northern German state capital Hanover but also as fifth-year-follow-up survey of Borrelia infections in I. ricinus ticks. The total infection rate of B. burgdorferi s.l.-infected ticks in 2010 was 22.7% with a significant difference between adults and nymphs (33.3% and 20.3% infected individuals, respectively), but not between adults and larvae (25.8% infected individuals). The expected highest burden in adult ticks is in accordance with different previous studies [6,9,31] and is most likely related to the combination of transstadial Borrelia transmission and the number of blood meals for development, which is connected to a higher probability to acquire bacteria from infected hosts. The rather high prevalence of Borrelia-infected larvae in 2010 might be explained by an interrupted blood meal with failed further development as well as by transovarial transmission. Even though transovarial transmission of B. burgdorferi s.l. was found to be rather inefficient [24,45], it is the most plausible explanation as it seems unlikely that a quarter of collected larvae were removed during feeding from a Borrelia-infected host. However, solely in one of eight Borrelia-positive tested larvae, B. burgdorferi s.l. was confirmed by RLB. Overall, RLB confirmed 312 (65.5%) out of 476 Borrelia qPCR-positive ticks and identified 282 (59.2%) successfully concerning their genospecies. Reason for the generally lower RLBpositive rate is most likely a higher sensitivity of qPCR. The 30 tick samples in which RLB resulted in detection of B. burgdorferi s.l., but genospecies identification failed, might be explained by the higher sensitivity of the SL2 probe compared to genospecies-specific probes. Furthermore, not all in Europe occurring B. burgdorferi s.l. genospecies were detected by RLBprobes are not available for B. kurtenbachii, B. carolinensis and B. finlandensis. Moreover, ticks might have been infected with Borrelia miyamotoi, a Borrelia species associated with the relapsing fever group. B. miyamotoi, originally detected in Japan in I. persulcatus [46], was also previously found in I. ricinus ticks in Europe [47,48]. In contrast to B. burgdorferi s.l., transovarial transmission could be demonstrated [49,50]. However, the primerprobe-combination [9] used in the present study does not detect B. miyamotoi, but it is specific for the B. burgdorferi s.l. complex.
Regarding seasonal distribution of B. burgdorferi s.l. infection in Hanoverian I. ricinus ticks, the two peak course of infected ticks in June/July and October may depend on biological and climatic conditions like host disposability, temperature and humidity as these factors can affect pathogen transmission from hosts to ticks [51][52][53][54]. Concerning different sampling sites in the city of Hanover, tick infection rates range from 16.2% at the location "Maschpark" up to 32.9% at "Misburger Wald". These differences in numbers of Borrelia-infected ticks might be explained by different incidences of Borreliainfected hosts like small rodents or birds at single sampling locations.
B. burgdorferi s.l. genospecies identification resulted in findings comparable to results from previous studies,  where B. afzelii was the predominant detected species in ticks [33,55,56], but also B. garinii including B. bavariensis and B. spielmanii were frequent findings in Borreliainfected I. ricinus ticks [6,12,55]. Comparing mentioned genospecies distribution in 2010 with data from Hanoverian ticks collected in 2005, B. afzelii, B. garinii including B. bavariensis and B. spielmanii were the most frequently detected species in both studies. B. afzelii was the most mono-infecting species, B. spielmanii was found more often in multiple-infections than as mono-infecting species and frequent findings of double-infections contained a combination of B. afzelii and B. spielmanii [9]. Doubleor multiple-infections (cf. Table 3) may result from blood meals on different hosts each infected with one genospecies, from one host having a multiple infection causing ticks to take up a number of genospecies, or from cofeeding of infected ticks. Several combinations including B. afzelii or B. garinii as detected in the present study were also observed by Pichon et al. [57]. The detected combination of B. afzelii and B. burgdorferi s.s. was also described amongst others in red squirrels [23].
Concerning coinfections of Borrelia-positive ticks with Rickettsiales in 2010, statistically significant differences were observed between different tick stages. This is in accordance with data from 2005, where adult ticks also showed higher infection rates than nymphs [42,43]. No statistically significant differences in coinfection rates were found between years.

Conclusion
In summary, comparing B. burgdorferi s.l. infections in Hanoverian I. ricinus ticks in 2010 with data from 2005, a significant increase of infected larvae was determined whereas nymphs and adult ticks did not show significantly changed infection rates. Whether this development of infected larvae is an actual fact through rather effective transovarial transmission under field conditions due to unknown factors, or an isolated event should be evaluated by further studies. The Hanoverian tick infection rate and distribution of B. afzelii, B. garinii and B. spielmanii as most abundant genospecies will be further monitored in the next fifth-year-follow-up study starting in April 2015.