Pathogens Manipulating Tick Behavior—Through a Glass, Darkly

Pathogens can manipulate the phenotypic traits of their hosts and vectors, maximizing their own fitness. Among the phenotypic traits that can be modified, manipulating vector behavior represents one of the most fascinating facets. How pathogens infection affects behavioral traits of key insect vectors has been extensively investigated. Major examples include Plasmodium, Leishmania and Trypanosoma spp. manipulating the behavior of mosquitoes, sand flies and kissing bugs, respectively. However, research on how pathogens can modify tick behavior is patchy. This review focuses on current knowledge about the behavioral changes triggered by Anaplasma, Borrelia, Babesia, Bartonella, Rickettsia and tick-borne encephalitis virus (TBEV) infection in tick vectors, analyzing their potential adaptive significance. As a general trend, being infected by Borrelia and TBEV boosts tick mobility (both questing and walking activity). Borrelia and Anaplasma infection magnifies Ixodes desiccation resistance, triggering physiological changes (Borrelia: higher fat reserves; Anaplasma: synthesis of heat shock proteins). Anaplasma infection also improves cold resistance in infected ticks through synthesis of an antifreeze glycoprotein. Being infected by Anaplasma, Borrelia and Babesia leads to increased tick survival. Borrelia, Babesia and Bartonella infection facilitates blood engorgement. In the last section, current challenges for future studies are outlined.


Introduction
Vector-borne diseases (VBDs) are caused by parasites, bacteria and viruses, leading to more than 700,000 deaths yearly [1]. Many VBDs are caused by pathogens vectored by arthropods, among which mosquitoes, sand flies, triatomine bugs and ticks are major players [2][3][4]. Pathogens represent a significant selective pressure on their hosts [5]. The pathogen-host interaction shapes coevolution on both sides. Pathogens can manipulate many phenotypic traits of their hosts, thus maximizing their fitness [6][7][8]. A classic example is represented by Plasmodium infection making host-borne odors more attractive to Anopheles mosquitoes [9][10][11] (but see [12]). Similarly, Leishmania modify the odor of infected hosts, making it more attractive for sand flies [13,14], and Hepatozoon infection can lead to odor changes in snake and frog hosts, which leads to higher feeding rates by Culex pipiens and Culex territans mosquito vectors [15].
Among the phenotypic traits that can be modified, a fascinating category is represented by the behavioral manipulation of vectors to enhance pathogens transmission. According to an earlier classification by Hurd [16], manipulation can be achieved through various mechanisms, such as boosting the chances of contacts among vectors and hosts, reducing vector reproduction to increase nutrients available for the microorganisms, and/or magnifying vector survival.
Behavioral alterations caused by pathogens on their vectors include higher biting rates, reported for Yersinia pestis-infected Xenopsylla cheopis fleas [17], La Crosse virus-infected Aedes triseriatus

Pathogens Infecting Ticks Lead to Major Behavioral Changes
Most of the studies about how pathogen infection can affect tick behavior have focused on Borrelia bacteria, the causative agents of Lyme borreliosis, followed by Anaplasma-related studies. Only a few studies have considered the behavioral changes triggered by Babesia, Bartonella, Rickettsia, and TBEV infection. The following subsections review current knowledge about how infection with different pathogens may affect tick behavior. The first subsection is dedicated to the behavioral changes in Borrelia-infected ticks, followed by subsections focusing on the behavioral changes in ticks infected by other pathogens. Since it is hard to disentangle behavioral results from their whole biological context, this review critically discusses behavioral research along with current information on the impact of pathogen infection on other phenotypic traits. Physiological research shedding light on how pathogens can finely manipulate tick basic metabolism shows how this profoundly impacts tick behavior and ecology.

Laboratory Studies
Several studies demonstrate how Borrelia spp. can promote behavioral modifications in tick vectors. A pioneer study by Lefcort and Durden [48] highlighted that nymphs of I. scapularis, infected by Borrelia burgdorferi, showed increased phototaxis and attraction to vertical surfaces compared to non-infected ticks. Both behavioral alterations may be adaptive, boosting Borrelia transmission, since increased phototaxis and attraction to vertical surfaces could increase the likelihood of contact between the tick vector and a potential reservoir host. Additionally, from a physiological point of view, it has been reported that B. burgdorferi-infected I. scapularis ticks can up-regulate the tick histamine release factor (tHRF), allowing a higher blood flow to the bite site and improved vascular permeability. In this way, ticks may engorge more quickly, and the likelihood of becoming infected by Borrelia is also higher [49]. In contrast, B. burgdorferi-infected I. scapularis adults showed lower mobility than uninfected ones, questing at lower heights, being less able to overcome physical obstacles, and avoiding climbing on vertical surfaces [48]. Reduced mobility of adult ticks may represent a by-product of B. burgdorferi infection, possibly non-adaptive for the pathogen or for the tick. However, one may consider that this behavior may be adaptive, since reduced tick movement contribute reducing water and energy loss. Current evidence does not allow any definitive conclusion. While I. scapularis nymphs are crucial for infecting reservoir hosts, thus indirectly newly hatched tick larvae, adults are less prone to feed on reservoir hosts, and so they are not considered mayor players in the spread of Borrelia bacteria [45]. However, these conclusions cannot be generalized for all Ixodes species. Indeed, further research by Romashchenko et al. [50] showed that I. persulcatus adults with a higher prevalence of B. burgdorferi s.l. reached a higher questing height (where the likelihood of intercepting a large host for feeding purposes is higher) than non-infected ticks.
Larvae, nymphs and adults of I. ricinus naturally infected by B. burgdorferi s.l. showed reduced mobility compared to uninfected ones [51]. The same applies to Borrelia-infected adults of the taiga tick, I. persulcatus [51][52][53], suggesting that being infected by B. burgdorferi s.l. boosts mobility of I. persulcatus adults with exoskeleton deformities (i.e., crater-like depressions on the exoskeleton, damaged female scutum and undeveloped palps), collected from a heavy metal polluted area. The reasons at the basis of the observed differences are unclear, and may be partially linked to the short observation time of the mobility experiments (i.e., 3 min). Later, at variance with the results by Alekseev and co-workers, Perret [54] reported that I. ricinus nymphs naturally infected by B. burgdorferi s.l. exposed to desiccating conditions for a prolonged time are more active than uninfected ones. Similarly, Borrelia afzelii-infected I. ricinus nymphs showed higher mobility [i.e., longer walking activity and higher velocity (+10%)] over uninfected ones [55].

Into the Woods-Field Studies
In addition to laboratory research, it is important to consider what is known about the impact of Borrelia infection on ticks in the natural setting. A flagging collection study in the proximity of St. Petersburg (Russia) [56] showed that the appearance of B. burgdorferi s.l.-infected I. persulcatus ticks was related to gradients between the surface and soil temperature and the soil and air relative humidity. More Borrelia-infected nymphs were detected at the temperature interval 10-14 • C compared to 15-20 • C and 21-26 • C (no collected individuals in this latter case) [56]. In dense woodlands nymphs of the western black-legged tick, Ixodes pacificus, infected by B. burgdorferi were found at higher densities on logs and trunks than in the leaf litter [57]. A field study carried out on I. ricinus in western Germany showed that B. burgdorferi s.l.-infected adult females were more frequently present on protective clothes of human volunteers compared to infected females caught by blanket dragging in the same study area [58]. These field results highlight that being infected by Borrelia may contribute to boost host-finding strategies at least for I. pacificus and I. ricinus. Further field data on other Ixodes species are needed. It is also important to determine the effects of other Borrelia genospecies on tick behavioral traits.

Borrelia Manipulation of the Host Odors
Do Borrelia infections manipulate host odor that make hosts more attractive for tick vectors? Similar questions have been addressed about other pathogens, making the host odor more attractive for their insect vectors [13,59]. However, for ticks this question remained unanswered. Recently, van Duijvendijk et al. [60] showed that I. ricinus nymphs were more attracted to bank voles (Myodes glareolus) infected by B. afzelii over uninfected voles, and that infected nymphs feeding on bank voles showed a higher body weight than uninfected ones.
But what happens when the tick vector selects a host not competent for Borrelia transmission? Is the pathogen able to manipulate the tick vector boosting avoidance of the incompetent host? Berret and Voordouw [61] attempted a reply to both questions, showing that being infected by rodent-or bird-specialized Borrelia genospecies did not affect attraction (in terms of questing rates)to mouse odor. Thus, no evidence for the qualitative manipulation hypothesis [6] in the Borrelia-Ixodes interaction has been detected.

Ixodes Behavior Meets Physiology
Borrelia affects tick physiology in several ways. For example, B. burgdorferi infecting I. scapularis salivary glands can usurp a tick salivary protein, salp15, which binds to the spirochetes protecting them from antibody-mediated killing; thus, facilitating infection in mice [62]. The studies summarized above (with few exceptions, see [51][52][53]) highlight that Borrelia infection-as a general trend-impacts tick mobility, with special reference to host-seeking activity. One may consider that a more mobile tick is exposed to a higher risk of desiccation (thus needing to move more often in the ground litter to recover water), and may also need higher energy reserves, because of its increased metabolic activity [63]. The study by Herrmann and Gern [64] shows that the infection by B. burgdorferi s.l. makes I. ricinus ticks more resistant to desiccation when exposed to challenging thermo-hygrometric conditions. The effect was more pronounced on female ticks than on males and nymphs, and highest survival was noted for B. afzelii-infected ticks [64]. A further study examined I. ricinus nymphs collected in the field, showing that Borrelia-infected ones lived longer than uninfected conspecifics when exposed to unfavorable thermo-hygrometric parameters [55].
As concerns the impact of Borrelia infection on tick energy reserves, it has been reported that I. ricinus nymphs questing in the field have a higher fat content when infected by B. burgdorferi s.l. if compared to uninfected ones [65]. The same study clarified that Borrelia exploited only a tiny fraction of these energy reserves, thus the tick can benefit from the increased fat reserves for its own metabolism [65]. As stressed by Herrmann and Gern [63], ticks water recovering occurs through active absorption of external aqueous vapor, and this process is costly from an energetic point of view. Thus, the higher energy reserves of Borrelia-infected ticks can contribute to allow a higher resistance to desiccation.
Borrelia burgdorferi s.l. infection and energy reserves affect horizontal mobility in I. ricinus nymphs. Ticks with high-fat reserves are more prone to explore dryer areas, but-among them-infected individuals stay longer in a fixed position [66]. This partially contrasts with the earlier findings by Lefcort and Durden [48]. Anyway, it has been claimed that Borrelia manipulates I. Ricinus nymphs to remain still, waiting for possible hosts, even under desiccating conditions that usually increase walking activity in non-infected conspecifics [66]. The overall picture is interesting and outlines how Borrelia infection contribute in multiple ways to its own and vector fitness. Remaining still under challenging thermo-hygrometric conditions is highly dangerous for ticks, and only individuals with abundant energy reserves can perform a prolonged recovery of water vapor from the air [67]. The adaptive significance of this behavior deserves further studies, as well as validation of the findings on other Ixodes species.
Lastly, understanding how the behavioral changes characterizing Borrelia-infected ticks may contribute to the spread of Lyme disease is a crucial, but neglected, research topic. A theoretical model developed by Gassner and Hartemink [45] stresses how behavioral changes triggered by Borrelia infection may contribute to an increased transmission risk. Applying a next-generation matrix approach the authors showed that the increased tick vector survival deeply affects the basic reproduction number R 0 for Borrelia pathogens.

Behavioral Changes in Anaplasma-Infected Ticks
Anaplasma infection leads to important physiological changes in Ixodes ticks, impacting tick survival, questing, and feeding activity [68][69][70]. Indeed, as recently highlighted by Cabezas-Cruz et al. [71], Anaplasma phagocytophilum infecting I. scapularis manipulates the expression of key genes, boosting in turn both vector fitness and the pathogen transmission [68,72]. Three major facts related to Anaplasma infection on ticks deserve consideration.
First, ticks exposed to challenging thermo-hygrometric conditions experience a high risk of desiccation. Anaplasma phagocytophilum infecting I. scapularis ticks induce the synthesis of heat shock proteins (i.e., hsp20 and hsp70), which reduce the risk of desiccation, thus enhancing tick survival rates [73]. From a behavioral point of view, gene knock-down of these hsp is linked with reduced questing speed in I. scapularis males, but with relevant changes according to the tested temperature. At 22 • C, only the knock-down of hsp70 gene leads to a significant decrease in tick questing activity (−50%) [73]. Anaplasma phagocytophilum infection does not downregulate the I. scapularis production of the tick protective antigen subolesin, which-if knocked down artificially-leads to reduced questing rates (i.e., −50% and −66% at 4 • C and 37 • C, respectively) [73], among other negative effects on tick biology [71]. The production of another crucial protein for vectors of A. phagocytophilum, such as I. scapularis and I. ricinus, the tudor staphylococcal nuclease, is not affected by the pathogen, thus minimizing damage for tick feeding (assessed in term of tick weight), thus their survival [74].
Second, being exposed to cold temperatures during the winter may lead to significantly higher mortality rates in tick populations. Anaplasma infection can lead to beneficial effects on its tick vectors. Indeed, it has been showed that the infection of I. scapularis nymphs by A. phagocytophilum trigger the synthesis of an antifreeze glycoprotein that protect ticks from cold [75].
Third, A. phagocytophilum infecting the salivary glands of I. scapularis facilitates the infection process through inhibition of the intrinsic apoptosis pathway, while tick cells respond through the extrinsic apoptosis pathway limiting the infection and boosting vector feeding and survival [69,71,76,77].
Overall, the interaction between A. phagocytophilum-I. scapularis is one of the better studied pathogen-tick relationships, with fully clarified mechanisms magnifying tick and pathogen fitness. Applying this successful approach for other important Anaplasma species (e.g., A. marginale, A. centrale, A. ovis and A. platys) and their tick vectors is an important timely challenge.

Behavioral Changes in Babesia-Infected Ticks
At variance with the plethora of studies on the effect of Borrelia infection on tick behavior, the impact of Babesia infection on tick vectors has been considered only in a couple of studies. First, Randolph [78] showed that Babesia microti infection magnified the feeding success (in terms of mean engorged weight) and survival (in terms of the percentage molt of larvae to nymphs) of the shrew tick Ixodes trianguliceps. Both changes are not linked to the level of infection by B. microti [78]. Later, Hu et al. [79] focused on the potential effect of B. microti infection on I. scapularis feeding time, engorged body weight, and molting rate. Being infected delays the tick engorgement, but the body weight of nymphs engorged on infected hosts is higher than that of the conspecifics that fed on uninfected ones. Larvae, but not nymphs, fed on infected hosts show higher molting rates than those that fed on non-infected hamsters. These differences are not detected testing various tick instars on a different host, i.e., Peromyscus leucopus mice [79]. The mechanisms that promote feeding success, development and survival of Babesia-infected ticks are unknown.

Behavioral Changes in Bartonella-Infected Ticks
Studies on the behavioral effects of Bartonella infection in ticks are limited. Recently, the salivary glands transcriptome of I. ricinus females infected or not by Bartonella henselae was investigated. Bartonella henselae infection of salivary glands up-regulate I. ricinus serine protease inhibitor (IrSPI), a member of the BPTI/Kunitz family of serine protease inhibitors, and IrSPI silencing strongly reduces weight of I. ricinus feeding adults [80].

Behavioral Changes in Rickettsia-Infected Ticks
Despite the public health importance of rickettsiosis vectored by ticks, our knowledge about how the Rickettsia infection may affect the behavior of tick vectors is limited. Frątczak et al. [81] showed that I. ricinus adult males and females are attracted toward an area irradiated by a 900 MHz electromagnetic field in a radiation-shielded tube. Ticks infected by Rickettsia spp. as well as those co-infected by Rickettsia spp. and Borrelia s.l. are more attracted by the electromagnetic field compared to the uninfected ones [81]. Understanding the effects of anthropogenic environmental modifications on the pathogen-tick interaction and the related changes in vector behavior represents a highly valuable area of research, which is of practical interest for public health.

Behavioral Changes in TBEV-Infected Ticks
Ixodid ticks feeding on humans or animals show a higher TBEV prevalence compared to field-collected ticks from the same area. This is clear with I. ricinus from Germany [82] as well I. persulcatus and Ixodes pavlovskyi from Russia [83,84]. Two main hypotheses have been formulated to explain these findings. First, the TBEV infection magnifies tick mobility and host-seeking activity. Second, feeding boosts the titer of TBEV already present in unfed questing ticks to a detectable level. The study by Belova et al. [85] provided data in support of both theories. Ixodes ricinus adults infected by TBEV trying to reach a bait are more active and tolerant than non-infected ones when exposed to growing concentrations of the repellent N,N-diethyl-meta-toluamide (DEET). Of note, about 6% of the infected adults can go over an area where DEET was formulated at 1%, while none of the uninfected ticks could do so [85].
Furthermore, it should be noted that TBEV-infected I. persulcatus females are more active, showing a higher mobility both in terms of walking speed and length of the trajectory when trying to reach a bait [86]. Moreover, I. persulcatus infected with the TBEV reach a higher questing height compared to uninfected conspecifics [50,53].

Concluding Remarks and Outstanding Challenges for a Research Agenda
This analysis of current knowledge about how, when and why pathogens trigger behavioral changes in their tick vectors shows that most studies focused on Ixodes species infected by Borrelia bacteria, followed by research on Anaplasma-Ixodes interactions. Only a limited number of studies investigated tick behavioral changes associated to infection by Babesia, Bartonella, Rickettsia, and TBEV.
However, even if limited, the knowledge available on the topic contributes to depict a fascinating scenario. As a general trend, being infected by Borrelia and TBEV substantially increase the overall tick mobility, with special reference to exploiting higher questing heights (this is likely connected with increased phototaxis), often coupled with a higher walking speed and longer walking activity, if compared to non-infected ticks. Noteworthy analogies have been found examining changes triggered post-Borrelia and Anaplasma infection. Both pathogens magnify desiccation resistance of Ixodes ticks (Figure 1). This has been linked to fine physiological changes, i.e., storing higher fat reserves in Borrelia-infected ticks, and synthesizing heat shock proteins in Anaplasma-infected ticks. Anaplasma also improves cold resistance in infected ticks. Additionally, being infected by both pathogens, as well as by B. microti, allowed ticks to achieve a longer survival. Babesia infection also led to higher feeding success (i.e., higher engorged weight), a successful engorgement through upregulation of tHRF and IrSPI characterizes Borreliaand Bartonella-infected ticks, respectively (Figure 1). Overall, behavioral and physiological changes triggered by pathogens substantially contribute to the tick fitness [63,71], fostering a win-win strategy in the pathogen-tick interaction [69].
Following the criteria expressed for parasite/pathogen-host interactions [7], the term "manipulation" can be used only when the research shows direct fitness benefits (thus an adaptive value of the observed change) for the infecting pathogen, mainly improving its transmission and dispersal rates. This is not disconnected from benefits for the tick vector, often including a longer survival (Figure 1). In other cases, reducing the infection (as in the A. phagocytophilum-Ixodes interaction) or lowering its detrimental effects, can also lead to significant changes, which can be beneficial for the vector. Even the pathological state triggered by the infection can lead to behavioral changes [7,87], which may even reduce the vectorial ability of the involved species. These seem the case of earlier studies showing reduced feeding rates in Cx. pipiens infected by the Rift Valley fever virus [39], as well as the decreased flight activity in Cx. tarsalis infected by the Western equine encephalomyelitis virus [88]. Other pathogen-vector interactions need to be studied further, e.g., the vesicular stomatitis virus-Culicoides sonorensis one, which led to a decrease in the midge biting behavior 2 days' post-inoculation (DPI), followed by a sharp increase 3 DPI, but not 4 DPI [89]. Besides these examples on insects, a possible case fitting this category can be that of B. burgdorferi-infecting I. scapularis adults, which results in lower tick mobility, with detrimental effects also on questing [48]. Overall, caution remains the first and most useful rule when using the term "adaptive". In conclusion, based on the analysis of the literature considered in this review, there are some points that can be useful for developing a perspective research agenda.
Despite the wide diversity of tick species worldwide [90], only Ixodes species have been considered in studies analyzing pathogen-related behavioral changes. How pathogens affect behavioral traits of other tick vectors of huge medical importance, e.g., those belonging to the genera Amblyomma, Dermacentor, and Rhipicephalus, is still overlooked.
Most of the studies have been carried out on Anaplasmaand Borrelia-Ixodes interactions; more research efforts on other key tick-borne pathogens (with special reference to Babesia, Coxiella, Ehrlichia, Rickettsia, and Theileria species) are urgently needed, elucidating the physiological pathways promoting vector manipulation, as recently carried out for Anaplasma and Bartonella research [69,71,80].
Standardizing methods for behavioral research is crucial to allow data comparisons. For example, high reproducibility of questing-related data can be obtained using mechatronic arenas, which also reduce animal testing in laboratory [91], and allows the dissection of host-borne cues [92]. The use of mechatronic arenas also allows repeated testing, thus assessing both intra-and inter-individual variability (e.g., questing success and responses to repellents) of the affected traits in highly standardized conditions over time.
While most studies focused on the effect of pathogen infection on mobility, questing and feeding activities, one may consider, they affect the timing of mating (as showed for Ae. triseriatus infected by La Crosse virus, [29]), mate recognition and mating success of ticks? To the best of my knowledge, this issue has never been investigated.
Moving the focus from studying single traits to a more systematic analysis of all behavioral traits potentially altered by the pathogen infection, and how they correlate each other, represents a major challenge for future research [7].
Both for Borrelia and TBEV-infected ticks, field data have been found extremely useful, fitting the behavioral changes observed in laboratory, which may be partially responsible of enhanced host-seeking activities. Further field research focusing on the impact of other pathogen infections on tick behavioral and ecological traits is needed.
Lastly, the role of manipulation is still not considered in most epidemiological models [10], and the behavioral differences between infected and uninfected vectors need to be carefully evaluated for the correct development and implementation of tick control tools within the Integrated Vector Management and One Heath approaches [2,93,94].
Funding: This study received no external funding.