Transovarial transmission of Borrelia spp., Rickettsia spp. and Anaplasma phagocytophilum in Ixodes ricinus under field conditions extrapolated from DNA detection in questing larvae

Background Ixodes ricinus constitutes the main European vector tick for the Lyme borreliosis pathogen Borrelia burgdorferi (sensu lato), the relapsing fever borrelia Borrelia miyamotoi, as well as Anaplasma phagocytophilum and several Rickettsia species. Under laboratory conditions, a transovarial transmission to the next tick generation is described for Rickettsia spp. and Borrelia spp., especially regarding B. miyamotoi, whereas the efficiency of transovarial transfer under field conditions is largely unstudied. Methods In order to better estimate the potential infection risk by tick larvae for humans and animals, 1500 I. ricinus larvae from 50 collected “nests” (larvae adhering to the flag in a clumped manner) were individually examined for Borrelia, Rickettsia and A. phagocytophilum DNA using quantitative real-time PCR (qPCR). Results Thirty-nine of 50 nests each (78.0%, 95% CI: 64.0–88.5%) were positive for Borrelia spp. and Rickettsia spp. DNA, and in three nests (6.0%, 95% CI: 1.3–16.5%) A. phagocytophilum DNA was detected. Overall, DNA from at least one pathogen could be detected in 90.0% (45/50, 95% CI: 78.2–96.7%) of the nests. Of the 1500 larvae, 137 were positive for Borrelia spp. DNA (9.1%, 95% CI: 7.7–10.7%), 341 for Rickettsia spp. DNA (22.7%, 95% CI: 20.6–24.9%) and three for A. phagocytophilum DNA (0.2%, 95% CI: 0–0.6%). Quantity of Borrelia spp. and Anaplasma spp. DNA in positive larvae was low, with 2.7 × 100Borrelia 5S-23S gene copies and 2.4 × 101A. phagocytophilum msp2/p44 gene copies detected on average, while Rickettsia-positive samples contained on average 5.4 × 102gltA gene copies. Coinfections were found in 66.0% (33/50, 95% CI: 51.2–78.8%) of the nests and 8.6% (38/443, 95% CI: 6.1–11.6%) of positive larvae. In fact, larvae had a significantly higher probability of being infected with Borrelia spp. or Rickettsia spp. when both pathogens were present in the nest. Conclusions This study provides evidence for transovarial transmission of Rickettsia spp. and Borrelia spp. in I. ricinus under field conditions, possibly facilitating pathogen persistence in the ecosystem and reducing the dependence on the presence of suitable reservoir hosts. Further studies are needed to prove transovarial transmission and to explain the surprisingly high proportion of nests containing Rickettsia and/or Borrelia DNA-positive larvae compared to infection rates in adult ticks commonly reported in other studies.


Background
Ixodes ricinus is the most widespread tick species in Europe and acts as a vector for a range of bacterial and viral tick-borne pathogens with relevance for human and animal health. In Germany, different genospecies of the B. burgdorferi (sensu lato) complex, B. miyamotoi, A. phagocytophilum and Rickettsia spp. are among the most important bacterial pathogens transmitted by I. ricinus. These pathogens cause various diseases in humans and animals. Borrelia burgdorferi (s.l.) is the causative agent of Lyme borreliosis (LB), while B. miyamotoi causes febrile illness and has been associated with meningoencephalitis in immunocompromised patients in Europe [1,2]. Furthermore, several Rickettsia spp. may cause spotted fever or lymphadenopathy in humans [3]. In I. ricinus, R. helvetica, R. monacensis, R. massiliae and R. felis have been detected, with R. helvetica being the most frequent [3][4][5][6][7]. Another member of the order Rickettsiales, A. phagocytophilum, may cause granulocytic anaplasmosis in humans, dogs, horses, goats, sheep and cattle [8][9][10].
Transmission of these pathogens between or in ticks may occur in various ways, including co-feeding, transstadial, sexual as well as transovarial transmission. Cofeeding transmission, whereby infection spreads from one tick to another feeding in close proximity on the host, seems to be very rare [11][12][13]. Transstadial transmission, i.e. the transmission of pathogens from one developmental stage to the next, occurs in Borrelia spp. as well as Rickettsia spp. and A. phagocytophilum [14][15][16][17]. In contrast, sexual transmission, i.e. transmission from a male to a female tick during copulation, has only been described for Rickettsia species as well as some relapsing-fever borreliae [18,19]. During transovarial transmission, the offspring of an infected female is affected. An efficient transovarial transmission rate of up to 100% has been shown for Rickettsia spp. under laboratory conditions [17,20], whereas efficiency is generally believed to be much lower for Borrelia burgdorferi (s.l.), which are mainly transmitted to larvae via a blood meal from an infected host or rarely by co-feeding [21][22][23]. Borrelia miyamotoi, on the other hand, can be transmitted transovarially from the female tick to more than 90.0% of its larvae [22,24]. In contrast, a low to inefficient transovarial transmission rate has been described for A. phagocytophilum [25][26][27]. Transovarial transmission facilitates the persistence of pathogens in the ecosystem and thus reduces dependence on suitable reservoir hosts.
The main hosts of I. ricinus larvae are rodents [28], but humans may also serve as accidental hosts [29]. Many prevalence studies have examined pathogen prevalence in nymphs and adult ticks, but only a few have examined tick larvae. Therefore, the aim of this study was to determine Borrelia spp., Rickettsia spp. and A. phagocytophilum DNA in questing I. ricinus larvae to extrapolate potential transovarial transmission rates of these pathogens under natural conditions and assess the possible human or animal infection risk due to I. ricinus larvae.

Tick larvae collection and molecular species identification
Questing tick larvae were collected by the flagging method at different sampling sites in northern Germany during 2010-2018. The flag was dragged approximately one meter over the ground, and all larvae adhering to the flag in a clumped manner were defined as a "nest" originating from a single female (Fig. 1). After freezing the flags with the collected larvae overnight at − 20 °C, larvae were picked off the flag and stored individually at − 20 °C until genomic DNA isolation.
Tick species identification was performed on genomic DNA of two larvae per nest by amplification and sequencing of a part of the 16S rRNA gene, using primers described by Mangold et al. [30]. The reaction mixture and thermoprofile were set up as described by Hauck et al. [31], except that the amount of DNA template was increased to 6 μl and the number of PCR cycles to 41. Amplicons were separated by electrophoresis on 1.5% agarose gels stained with GelRed ® (Biotium Inc., Fremont, CA, USA) and visualized under UV light. Obtained PCR products were Sanger-sequenced at Microsynth Seqlab Laboratories (Göttingen, Germany) and aligned with 16S rRNA gene sequences published by Estrada-Peña et al. [32]

Tick larvae testing for bacterial pathogen DNA
In total, 1500 larvae originating from 50 larvae nests were examined for Borrelia spp., Rickettsia spp. and A. phagocytophilum. Per nest, 10-40 larvae (average 30 larvae per nest, standard deviation, SD: 11.6) were examined, depending on nest size. In case of a small nest (≤ 40 larvae) all larvae were examined, whereas for larger nests 40 larvae were tested. Larvae were individually homogenized using 0.5 ml polysterene pistils (VWR, Darmstadt, Germany). Genomic DNA was extracted from individual tick larvae using the Nucleo Spin ® 8 Blood Core Kit (Macherey-Nagel, Düren, Germany) according to the manufacturer's specifications, with previously described amendments [5]. Until further use, isolated genomic DNA was stored at − 20 °C.
Testing for Borrelia spp. (B. burgdorferi (s.l.) and B. miyamotoi) and Rickettsia spp. by duplex quantitative real-time PCR (qPCR) was carried out as described previously [33,34]. For Borrelia spp., the 5S-23S rRNA intergenic spacer (IGS) region was targeted based on a primer-TaqMan ™ minor groove binder (MGB) probe combination designed by Strube et al. [35]. For detection of Rickettsia spp., the citrate synthase (gltA) gene was amplified based on a primer-TaqMan ™ probe combination by Stenos et al. [36]. The reaction set-up and thermal cycling were performed as described previously [5]. Regarding A. phagocytophilum, larvae nests from 2010 and 2011 were tested by targeting the 16S rRNA gene using a primer-TaqMan ™ probe combination by Sirigireddy and Ganta [37]. At the same time, successful DNA isolation was confirmed by duplex amplification of the Ixodes ITS2 region as previously described [5]. For larvae nests collected from 2015 to 2018, the msp2/ p44 gene with a primer-TaqMan ™ probe combination by Courtney et al. [38] was targeted. Again, successful DNA isolation was confirmed by simultaneous amplification of the Ixodes ITS2 region. The duplex reaction set-up and thermal cycling were carried out according to Blazejak et al. [7].

Borrelia (geno-)species and Rickettsia species identification
To determine the (geno-)species of Borrelia-positive tick larvae, the Reverse Line Blot (RLB) technique was performed for larvae nests from 2010/2011 by amplifying a fragment of the B. burgdorferi (s.l.) 5S-23S rRNA IGS region using biotin-linked forward primer 5SCB and reverse primer 23SN2 as published by Tappe et al. [39]. For larvae collected during 2015-2018, the biotin-linked forward primer B5S was used instead of 5SCB, and a second, hydrolase-23S rRNA region specific biotin-linked forward primer was added for specific amplification of B. miyamotoi as described by Blazejak et al. [40]. Moreover, the RLB setup was modified to include the BisNE1 probe as described by Springer et al. [29].
For Rickettsia species identification, a subset of the Rickettsia-positive samples was subjected to real-time pyrosequencing of a sequence stretch of the rickettsial citrate synthase (gltA) gene as previously described [41].

Statistical analyses
Statistical analyses were conducted in R v. 3.3.1 [42]. To assess factors influencing larval prevalence of Borrelia spp. and Rickettsia spp., generalized linear mixed effect models (GLMMs) with binomial error structure were constructed. Anaplasma phagocytophilum prevalence was not statistically analysed because of the low number of positive larvae. The fixed factors were "location/ year" (Hanover 2010, 2015, 2017, 2018, Mellendorf 2018 and Hamburg 2011) and "coinfection of the nest" (yes/ no). Due to the small number of nests originating from some locations, only nests from Hanover, Mellendorf and Hamburg were included (n = 1220 larvae from 43 nests). Furthermore, "nest ID" was included as a random effects term. Each model was compared to a null model including only the random effects term in a likelihood ratio test (R function 'anova' , test = 'chisq').
Finally, the predictive variable "location/year" was subjected to post-hoc analysis, computing all pairwise differences between factor levels in a Tukey's test based on the parameters of the fitted GLMM.

Tick larvae collection and molecular species identification
Total nest size of the 50 collected nests varied between 10 and 1643 larvae. Most nests were collected in Hanover (20/50), Hamburg (12/50) and Mellendorf (11/50). Detailed numbers of larvae per nest including collection sites and dates are presented in Table 1. Molecular species identification revealed that all collected nests belonged to I. ricinus.
Within nests, larval pathogen prevalence differed considerably, ranging between 2.5% and 97.5% for Rickettsia spp., 2.5% and 82.5% for Borrelia spp., as well as 2.5% and 8.3% for A. phagocytophilum. Except for Rickettsiainfections in Hanover (2010, 2015, 2018) vs Mellendorf (2018), no statistically significant regional differences in larval pathogen prevalence were detected. Additionally, to this influence of region, the differences in Rickettsia prevalence within nests might be explained by the degree of rickettsial development in the ovarian tissues of the mother tick at the time of oviposition. Burgdorfer and Brinton [20] described that experimentally infected female ticks with generalised massive infections transmitted Rickettsia spp. to 100% of their offspring, and this was also observed regarding R. helvetica [46]. However, females with a mild rickettsial infection or in the initial phase, had a lower percentage of infected larvae [20].
Overall, Rickettsia prevalence in the collected tick larvae was 22.7%. This is comparable to previous studies, as prevalence of Rickettsia spp. in field-collected Ixodes larvae in Hanover 2005 and 2010 amounted to 27.3% (24/88) and 16.1% (5/31) [4,5]. In Germany, R. helvetica is the most frequent Rickettsia spp. in I. ricinus [3][4][5][6][7], and was the only Rickettsia spp. that could be defined in larvae in the present study. The proportion of R. helvetica-infected female ticks producing at least one positive egg or larva was described as 100% [17]. The high Rickettsia DNA detection rates in the present study and the identification of R. helvetica in the entire subset of sequenced samples provide evidence for a high transovarial transmission rate for R. helvetica in I. ricinus under field conditions. This Table 2 Results of the GLMM testing the influence of location/ year and coinfection in the nest on Borrelia spp. prevalence in larvae collected in Hanover, Hamburg and Mellendorf (n = 1220 larvae from 43 nests) Notes: The full model was significantly different from a null model containing only the random factor "nest ID" (χ 2 = 14.0, df = 6, P = 0.029). Multiple comparisons between levels of the factor "location/year" were performed using Tukeyʼs contrasts with single-step P-value adjustment.  Table 3 Results of the GLMM testing the influence of location/ year and coinfection in the nest on Rickettsia spp. prevalence in larvae collected in Hanover, Hamburg and Mellendorf (n = 1220 larvae from 43 nests) Notes: The full model was significantly different from a null model containing only the random factor "nest ID" (χ 2 = 28.9, df = 6, P < 0.001). Multiple comparisons between levels of the factor "location/year" were performed using Tukeyʼs contrasts with single-step P-value adjustment. Significant P-values (P < 0.05) are shown in bold  ). Furthermore, the calculated copy numbers may be underestimated, because the tick DNA samples were tested against a plasmid standard. The plasmid standard represents an ideal template, while the tick DNA template is more complex; therefore, the copy numbers of standard versus tick DNA are not fully comparable. However, it should be kept in mind that DNA detection does not necessarily imply that viable bacteria were present in the larvae. The pathogens could have died in the egg, during development from egg to larva or during larval hatching. Furthermore, it is also possible that DNA from dead and lysed pathogens in the mother tick may have entered the eggs and was detected.
In contrast to Rickettsia spp., a rather inefficient transovarial transmission is assumed for B. burgdorferi (s.l.). Nevertheless, van Duijvendijk et al. [47] showed that flagged larvae can transmit B. afzelii and B. miyamotoi to rodents. Thus, larvae also pose a potential infection risk. Borrelia-prevalence in individual, field-collected unfed I. ricinus larvae varied from 0% to 25.8% in previous studies [33,[47][48][49]. This is in line with the overall larval Borrelia spp. prevalence of 9.1% determined in the present study. Within positive nests, 12.3% of larvae were positive on average, but rates of up to 82.5% were detected. Likewise, Burgdorfer et al. [50] described a transovarial transmission rate of 60.0% and 100.0% in two I. ricinus females infected with Borrelia spirochetes. However, in another study, only one of sixteen B. burgdorferi (s.l.)-infected I. persulcatus females could transmit the pathogen via eggs to the larvae. The infection rate of these larvae was 7.0% [51]. In contrast, an efficient transovarial transmission from female ticks to more than 90.0% of their larvae has been described for the relapsing fever borrelia B. miyamotoi [22,24]. In individual field-collected larvae, B. miyamotoi showed a prevalence of 0-4.2% [22,47,52]. Among Borrelia-positive I. ricinus ticks collected in 2015 in Hanover, B. miyamotoi showed a prevalence of 18.2% in successfully differentiated Borrelia-positive ticks, subdivided into 20.0% in nymphs, 14.0% in females and 18.2% in males [31]. Thus, larvae infected with B. miyamotoi were expected. However, species differentiation using RLB was only successful in one case due to the low number of 5S-23S IGS copies in Borrelia-positive larvae. In this larva originating from a nest with a transovarial transmission rate of 82.5%, B. spielmanii was identified. For the remaining nests, it was not possible to determine whether the DNA originated from B. burgdorferi (s.l.) or B. miyamotoi, and thus not possible to investigate whether there was a correlation between rather low or high transmission rates and the two Borrelia species. Furthermore, the nests may have been infected with several Borrelia (geno-)species, including B. miyamotoi. Additionally, transovarial transmission rates of B. burgdorferi (s.l.) might vary between genospecies. Further laboratory studies examining transovarial transmission in ticks infected with known Borrelia (geno-)species are needed to achieve a clearer picture and to assess the potential Borrelia infection risk posed by larvae.
Due to the low A. phagocytophilum prevalence, there were mainly coinfections with Borrelia spp. and Rickettsia species. In total, 66.0% of all nests and 2.5% of all larvae were coinfected with Rickettsia spp. and Borrelia spp., while for nymphs and adult ticks, coinfection rates between 7.3% and 22.9% have been described in northern Germany [4,37,40]. Furthermore, larvae had a significantly higher probability of being infected with Borrelia spp. or Rickettsia spp. when the nest was coinfected, leading to the hypothesis that coinfection of the mother tick promotes transovarial transmission efficiency.
Transovarial transmission of A. phagocytophilum is described as absent or inefficient in the literature [16,25]. A study by Baldridge et al. [53] showed transovarial transmission rates of 10.0-40.0% in Dermacentor albipictus, while Jahfari et al. [27] and Hagedorn [54] detected A. phagocytophilum in 1.3% of field-collected I. ricinus larvae. In Hanover, positive A. phagocytophilum larvae have not been detected in previous studies [5,44]. The detection of A. phagocytophilum DNA in three unfed larvae in three different nests in the present study shows that transovarial transmission may be possible, but that it is inefficient. Therefore, a potential risk of A. phagocytophilum infection by I. ricinus larvae appears to be low.

Conclusions
Detection of pathogen DNA in questing I. ricinus larvae provides evidence for transovarial transmission of Rickettsia spp. and Borrelia spp. under field conditions. In consequence, I. ricinus larvae might serve as a source of human or animal infection with these pathogens. However, further studies investigating the percentage of viable transovarially transmitted Rickettsia and Borrelia species are necessary. As expected, transovarial transmission of A. phagocytophilum was rare.