First Record of Hepatozoon spp. in Alpine Wild Rodents: Implications and Perspectives for Transmission Dynamics across the Food Web

Among the Apicomplexa parasites, Hepatozoon spp. have been mainly studied in domestic animals and peri-urban areas. The epidemiology of Hepatozoon spp. is poorly investigated in natural systems and wild hosts because of their scarce veterinary and economic relevance. For most habitats, the occurrence of these parasites is unknown, despite their high ecosystemic role. To fill this gap for alpine small mammals, we applied molecular PCR-based methods and sequencing to determine the Hepatozoon spp. in 830 ear samples from 11 small mammal species (i.e., Apodemus, Myodes, Chionomys, Microtus, Crocidura and Sorex genera) live-trapped during a cross-sectional study along an altitudinal gradient in the North-Eastern Italian Alps. We detected Hepatozoon spp. with an overall prevalence of 35.9%. Two species ranging from 500 m a.s.l. to 2500 m a.s.l. were the most infected: My. glareolus, followed by Apodemus spp. Additionally, we detected the parasite for the first time in another alpine species: C. nivalis at 2000–2500 m a.s.l. Our findings suggest that several rodent species maintain Hepatozoon spp. along the alpine altitudinal gradient of habitats. The transmission pathway of this group of parasites and their role within the alpine mammal community need further investigation, especially in consideration of the rapidly occurring environmental and climatic changes.


Introduction
Protozoa parasites are a highly diverse group of successful organisms globally distributed via a wide range of hosts [1] and whose role is crucial for ecosystem functioning [2]. Among the blood protozoa parasites, the phylum Apicomplexa has received much attention for its zoonotic relevance, for example, the vector borne diseases due to Plasmodium spp. or the Toxoplasma spp. transmitted in urban cycles that include domestic species and humans. Other protozoan parasites, such as the Hepatozoon genus represented by 340 species, have also been commonly recorded in peri-urban cycles sustained by domestic species (e.g., cats and dogs) [3][4][5][6][7][8], although no zoonotic relevance has been reported so far [9,10]. Further, Hepatozoon spp. have been detected in amphibians, reptiles, birds that act as intermediate hosts, and mammals, which can be both paratenic and reservoir hosts [11].
The transmission of Hepatozoon spp. can occur through different modalities that most often involve the ingestion of infected vectors or hosts. In mammals, for example, animals can become infected through grooming behavior [12] by eating infected arthropod ectoparasites attached to another individual, such as ticks [13], fleas [14], and mosquitoes [15]. Further, predator-prey food web transmission routes can occur through the ingestion of prey either containing infective meronts or infested by infected vectors [16]. Additionally, vertical transplacental transmission has also been reported [17].
Hepatozoon spp. have been found in a diversity of habitats across the globe, such as temperate broad-leaved forests [18,19], conifer forests [20], Mediterranean scrubland [21], subtropical forests and savannah [22], and taiga [23], likely indicating complex transmission cycles that are still largely unknown. However, the occurrence and prevalence of Hepatozoon spp. in some crucial ecosystems remain unexplored, for example, across altitudinal gradients where the rapid succession of habitats often corresponds to the high biodiversity of partially sympatric hosts. Under global and climate changes, such mountain habitats are particularly exposed to abiotic and biotic variations that favor upward distributional shifts of host mammal species, with local modifications of the communities [24,25]. This may result in parasite expansion in terms of geographic distribution and diversity of the host species involved [26,27]. The monitoring of the occurrence of a parasite with a very plastic transmission cycle, such as Hepatozoon spp., is therefore of crucial importance in these habitats undergoing intense dynamics. This is the case for Alpine ecosystems, where Hepatozoon spp. were identified in wild carnivores [18,20] but have never been recorded in alpine small mammals [28]. The occurrence of the parasite in the small mammal community is of particular interest, as it prevalently depends on arthropods acting as vectors, the vertical transmission route being secondary. The main arthropods transmitting Hepatozoon spp. to rodents are blood ectoparasites, such as ticks, fleas, mites, and mosquitoes [29], the occurrence of which is shifting upward in the Alps as the temperature and humidity conditions become milder with the climate change [30].
In this work, we conducted a cross-sectional study, i.e., we assessed the parasite prevalence without specifically considering the temporal pattern from 2019 to 2021 in the Italian Alps by intensively live-trapping small mammals from forested habitats up to above the tree line. Our main aim was to evaluate the occurrence of Hepatozoon spp. across the altitudinal gradient of the Alpine habitat also in consideration of the anthropic pressure on such habitat. To the best of our knowledge, this is the first study investigating Hepatozoon spp. occurrence in small mammals, and more generally in rodents in Italy, with a special focus on the Alpine habitat.

Study Area and Animal Sampling
The study was carried out in two sites located in the Province of Trento (Italy): Cembra Valley (CEV; 46.13020 N-11.17843 E, altitude 1000 m a.s.l.) and Calamento Valley (CAV; 46.12092 N-11.48842 E, altitude from 500 to 2500 m a.s.l.) ( Figure 1). CEV comprises peri-urban areas with an alternation of small villages and areas of anthropic employment (quarries and forest harvest) and semi-naturalized forests. CAV is instead characterized by the full vegetation succession, from broad-leaf woodland and mixed forest to conifer and stands forest, shrubs, and alpine prairies. The area is relatively undisturbed, with few anthropic activities (traditional cattle husbandry and seasonal tourism). In CEV, small mammal monitoring was performed from 2019 to 2021 at monthly or bimonthly intervals during the winter (November-March) and summer (April, June, and August), respectively, in a mixed broad-leaved and coniferous forest at 1000 m a.s.l. In CAV, live trapping was performed in 2019 and 2020 at monthly intervals during the summer (May-October) along an altitudinal gradient (from 500 m a.s.l. to 2500 m a.s.l.), corresponding to the succession of alpine habitats (i.e., from broad-leaf woodland down the valley up to grasslands above the tree line). Small mammals were captured using standard Ugglan Multiple Live Traps (model 2, Granhab, Sweden) set in a grid array, following a protocol that aimed at capturing especially rodents to compare assemblages across an altitudinal gradient. Capture-markrecapture (CMR) techniques were adopted, and each animal was individually tagged with a subcutaneous Passive Integrated Transponder (PIT) tag (Trovan ® Ltd., Douglas, UK). At each capture event, we recorded information on individuals, such as species, PIT tag code, sex, body mass, breeding status, and ectoparasites load. Additionally, at first capture, an ear biopsy (using sterile disposable ear punch needles, Ø 3 mm; 20 mg) was collected for each animal and individually placed in tubes stored at −80 • C.
All animal handling procedures and ethical issues were approved by the Provincial Wildlife Management Committee (Prot. n. S044-5/2015/277268/2.4).

DNA Extraction, Amplification, and Sequencing
For DNA extraction, the ear samples (506 from CEV and 326 from CAV) were incubated overnight at 56 • C and isolated using the DNeasy 96 Blood & Tissue Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions.
This study is part of a wider project investigating tick-borne pathogen circulation in small mammals. Therefore, the identification of Hepatozoon spp. was a bycatch while assessing the protocol for Babesia spp. detection on ear tissue samples. We applied a conventional polymerase chain reaction (PCR) method using the Babesia BJ1 5 -GTC TTG TAA TTG GAA TGA TGG-3 and BN2 5 -TAG TTT ATG GTT AGG ACT ACG-3 primers [46]. Indeed, beyond Babesia spp., BJ1 and BN2 primers can amplify the 18S RNA gene of some other protozoans [47], such as the 600-bp fragment of Hepatozoon spp. [48][49][50]. To discriminate the two parasites and detect coinfections, we analyzed all the samples separately (protocol of Babesia spp. not described) by specifically adapting the PCR reactions and temperature cycling profiles for each protozoan [47].
In particular, for Hepatozoon spp. detection, the PCR reactions were composed of 3-mM MgCl 2 , 0.1-mg/mL BSA, 0.2-mM each dNTP, 0.3-µM of each primer, 5× Flexi buffer, and 1.25-U/µL Go Taq Hot Start Polymerase (Promega) in 50 µL of total volume, including 1 µL of extracted DNA. After the initial denaturation at 95 • C for 2 min, 40 cycles of 95 • C for 20 s, 46 • C for 15 s, and 72 • C for 50 s were performed before a final incubation at 72 • C for 5 min. Negative controls, i.e., samples that did not contain DNA but only reagents, were included in each molecular essay to cross-check the quality for DNA extraction and PCR amplification. The success of amplification was confirmed using the Qiagen QIAxcel ® capillary electrophoresis system. Positive PCR products were purified using ExoSAP-IT (USB, Cleveland, OH, USA) according to the manufacturer's instructions and then sequenced using Sanger sequencing. The DNA sequences were analyzed and aligned using Sequencher software version 5.4.6 (Gene Codes Corp., Ann Arbor, MI, USA) and Clustal X software version 2.0 [51] and subsequently compared with the data stored in the GenBank database with the Basic Local Alignment Search Tool (Blast; online version).
To verify the robustness of the PCR method, we also analyzed and sequenced double ear samples on a subset of individuals (N = 20, 2 samples from 10 individuals).

Statistical Analyses
The prevalence of Hepatozoon spp. was calculated for each rodent species with a 95% confidence interval (CI) using the EpiR package [52] in the R program [53]. We applied the Two Proportion Z-test to analyze the differences of the prevalence rates among rodent species. The differences were considered statistically significant if the p-values were <0.05.
Moreover, we investigated the role of the host species and environmental context on the probability of rodent infection with Hepatozoon spp., controlling for the life history traits. Specifically, we applied two analytical designs: (i) we evaluated the effect of anthropogenic pressure on rodent infection by comparing the study areas (CEV: high pressure; CAV: low pressure), only for a homogeneous altitude; (ii) we assessed how the probability of infection would vary across species along the alpine altitudinal gradient in a particularly wild context (CAV: 5 belts of 500 m, from 500 m a.s.l. to 2500 m a.s.l.). To this end, we fitted Generalized Linear Mixed Models (GLMMs; [54]) with binomial distribution of errors to the probability of infection of all samples with anthropogenic pressure and altitudinal belts, respectively, for the two designs, as the main covariates together with the species. We controlled for the sex and breeding status of the captured rodents (juveniles, subadults, and adult classes for both males and females). Finally, to account for spatial autocorrelation between captures, we included a grid as the most parsimonious random effect (Appendix A). Model selection was performed on the basis of the AICc score [55] (see Appendix A for details). We fitted the models using R software 4.1.2 [53] and packages tidyverse [56], ppsr [57], MuMIn [58], and glmmTMB [59].
In both proportion tests and in GLMMs, we discarded those sporadic species that were captured only a few times (<20 capture events).
All accidentally caught shrews tested negative for Hepatozoon spp. and were therefore not included in the following analysis (see Table 1). The results from the 10 double samples were identical and confirmed the robustness of the PCR method.  Table 1 for details), with an overall prevalence of 35.9% (see Appendix B for sequences).
In Calamento Valley (CAV; 294 samples), four rodent species tested positive (Table 1). My. glareolus showed a statistically significant higher prevalence compared to A. flavicollis and C. nivalis (Z-test, p-value = 6.075 × 10 −12 and p-value = 0.027, respectively) ( Figure 2b). Among the species with too few captures to compute prevalence, Hepatozoon spp. was recorded in M. arvalis, in one individual out of the two captures and A. sylvaticus, in three out of eight captures.
The results of the GLMMs showed that rodent species have a crucial role in driving the infection of Hepatozoon spp. Conversely, both human pressure and the altitudinal gradient were not selected in the best models, indicating their scarce relevance in influencing the probability of becoming infected by Hepatozoon spp. (Appendix A). In particular, in the first analysis between study areas at 1000 m a.s.l., the infection probability depended on the species, with My. glareolus more likely to be infected than A. flavicollis (β = 1.07 ± 0.21; p-value = 3.27 × 10 −7 ; reference category: A. flavicollis). In the second analysis, the infection of Hepatozoon spp. depended on the additive effect of the species and breeding status, confirming that My. glareolus was more prone to becoming infected if compared to A. flavicollis (β = 1.97 ± 0.30; p-value = 5.98 × 10 −11 ; reference category: A. flavicollis), while no significant difference emerged between A. flavicollis and C. nivalis (Appendix A). In addition, a minor effect of the breeding status emerged, in which juveniles seemed to be marginally more infected compared to adults (β = 1.29 ± 0.70; p-value = 0.06; reference category: Adults), while we did not detect any significant difference between adults and subadults (Appendix A).
It is relevant to observe that, after sequencing, two samples (one from S. araneus and one from My. glareolus) previously assigned to Hepatozoon spp. were instead confirmed as Babesia microti (see Appendix B).

Discussion
This study reports for the first time the occurrence and prevalence of Hepatozoon spp. in wild rodents in the Italian Alps. In particular, we found a high prevalence in the most common woodland rodent hosts, i.e., My. glareolus and A. flavicollis, as well as in two other alpine species: the sympatric A. sylvaticus and C. nivalis at high altitudes.
This study is part of a wider project investigating tick-borne pathogen circulation in small mammals in the Italian Alps. In this context, the detection of rodent infection by Hepatozoon spp. represented a bycatch of the assessment of Babesia spp. For this reason, Babesia-specific primers were used for the screening, and the positive samples were further identified through sequencing. The identification of two positive samples of Babesia microti underlines the risk of mismatch when Babesia-specific primers are used for detecting Hepatozoon spp., as already underlined by other studies [47]. This finding suggests that follow-up sequencing is a compulsory step to discriminate this parasite by other protozoa when their detection is due to nonspecific protocols. For these reasons, we were not able to further identify Hepatozoon spp. at the species level due to the low quality of the obtained sequences. The identification could be refined by using Hepatozoon-specific primers (HEPF/HEPR [47]).
European studies on Hepatozoon spp. in wild rodents have been generally based on blood [32,47,61,62], organs [14,22,23], and, more rarely, on skin [13,63] samples, which are often collected through invasive sampling methods that, in many cases, require the suppression of the animals. In live rodents, Hepatozoon spp. can be identified only through blood sampling, to the best of our knowledge [31,32]. This comes with the drawback that traditional microscopy based on a blood smear can underestimate Hepatozoon spp. prevalence, especially with low intensity of infection [47]. On the contrary, PCR-based assays, such as those that we implemented using ear biopsy samples collected from livetrapped rodents, are considered more sensitive and robust. Coupling molecular screening approaches with less invasive sampling methods may accomplish conservation issues in cases of endangered and vulnerable species, e.g., C. nivalis [64], limiting the impact on the studied species. Within this framework, we believe that our contribution may provide a useful methodological approach. Although the comparison of past studies based on different protocols or matrices might not be feasible, the comparison between different procedures to assess Hepatozoon spp., e.g., by comparing ear tissue samples with skin ones as those collected in studies [13,63], could be an interesting direction of research.
The number of species of the small mammal community in which Hepatozoon spp. was detected, its first identification in C. nivalis (33.3%), and its occurrence at 2000 and 2500 m a.s.l., the highest elevation at which these parasites have ever been recorded, represent key findings to speculate on the transmission pathways of Hepatozoon spp. in the Alpine range. Additionally, this is the first record of Hepatozoon spp. in A. sylvaticus (30.0%) using molecular approaches [34,65]. Since small mammals can become infected from Hepatozoon spp. mainly though the ingestion of infected vectors, it is reasonable to consider Hepatozoon spp. as a vector-borne infectious agent for small mammal hosts and specifically in our case, although we did not assess the infective status and transmission dynamics of arthropod vectors. Our results suggest that Hepatozoon spp. are common in alpine environments, showing a high prevalence in a broad spectrum of alpine small mammal species. In particular, we did not detect any relevant difference in Hepatozoon spp. prevalence between anthropic and wild systems, implying a stable and ubiquitous presence of Hepatozoon spp. in the analyzed small mammal species. This is interesting, as Hepatozoon spp. have been previously studied mainly in domestic species [3][4][5][6][7][8]. The low host specificity displayed by Hepatozoon spp. [23,29] may indicate that these parasites persist in the environment sustained by competent and reservoir hosts, such as small mammals [66], both in wild and more anthropic settings. These conditions may increase the risk of spillover events among wildlife and domestic animals [67,68]. For example, A. sylvaticus generally shares food resources, part of the habitats and vectors, with other rodent competitor species, especially in case of high rodent and vector density [69], as it may occur in permissive environmental conditions (low elevations, in our case). The high density of reservoir hosts and of potentially infected vectors may amplify parasite circulation and spreading within the ecosystem, facilitating also the infection of less abundant host species [70]. This mechanism may have also favored the transition of the parasite to high altitudes, especially under increasingly milder climatic conditions. The less-limiting abiotic conditions along the altitudinal gradient occurring under climate change may promote an upward distributional shift of opportunistic small mammal host species (e.g., My. glareolus and Apodemus spp.) and, in turn, the survival and development of arthropod vectors (e.g., mites, fleas, and ticks) [27,71,72]. The combination of abiotic (i.e., climate) and biotic (i.e., interspecific interactions) conditions that permit host and vector persistence, together with the ability of a species to colonize favorable habitats, may therefore alter the local alpine communities, leading to novel host-vector species pairings and, thus, to the emergence of vector-borne diseases in new environments [73][74][75][76]. The occurrence of Hepatozoon spp. in C. nivalis and high altitudes that we recorded is compatible with these dynamics, although there is no track record of previous investigations at the same altitudes in small mammals (see Reference [28]).
The high rate of infection that we observed in different rodent species across a diversity of Alpine habitats may induce relevant cascading effects across the food web. In particular, being carnivore-competent hosts for Hepatozoon spp., a broad spectrum of both domestic (e.g., shepherd dogs [77]) and wild animals (e.g., snakes [78], meso-and large carnivores [20,79], and birds of prey [80]) may become infected through the ingestion of parasitized rodent preys. Specifically, small mammals are paratenic hosts for Hepatozoon spp., meaning that, despite not being necessary for parasite development, they help in maintaining its life cycle in the environment by supporting the cystozoite stages [40] infective for predators via ingestion [81]. The parasite can affect the immune response of predator secondary hosts, leading to population declines [39]. However, at the same time, since vector competence in becoming infected, replicating and transmitting the parasite varies across hosts [81], predators may represent dead end hosts for Hepatozoon spp. so indirectly reducing parasites transmission across ecosystem (i.e., dilution effect). This aspect of the transmission pathway of Hepatozoon spp. needs further investigation.
The prevalence of Hepatozoon spp. that we recorded in the most common rodent hosts, i.e., My. glareolus and A. flavicollis, is in partial accordance with other studies in Europe. In particular, we found that My. glareolus showed the highest prevalence of Hepatozoon spp. among all captured rodent species (58.8%), as was also observed in other studies [13,32,47,63], although with very variable levels (from 3.7% to 87.5% [13,23,47,48,62,82]). Conversely, the prevalence we found for A. flavicollis (28.1%) is higher than all previously published estimates that were below 10% [13,47,48,63] and may support a high prevalence of Hepatozoon spp. infection at lower elevations. The small sample size of Microtus spp. voles does not allow conclusive indications on the rate of infection of this genus (one positive M. arvalis out of five in total), as was the case also for other studies [23,31,32,47].
In conclusion, Apicomplexa remains one of the most poorly investigated groups among protozoa, so that accurate identification (via PCR-based methods coupled with sequencing) of Hepatozoon spp. in new wild hosts and habitats importantly contributes to the understanding of the ecological role of these parasites, especially under the current global change. Collectively, by detecting Hepatozoon spp. in a broad range of rodent hosts throughout the altitudinal gradient of Alpine habitats, we indicated the widespread occurrence of this parasite as likely supported by an expanding availability of arthropod vectors, due to variations in the temperature and moisture linked to climate change [75]. Therefore, this study represents a crucial starting point for future research combining ecological, epidemiological, and molecular analysis to evaluate the trophic transmission route of Hepatozoon spp. across the food chain in habitats particularly exposed to climate change, such as the Alpine range. In particular, studies investigating vector competence and distribution, as well as host ecology and their role in maintaining and transmitting Hepatozoon spp., are needed to clarify the function of this group of parasites in such changing ecosystems.

Supplementary Materials:
The following supporting information can be downloaded at https: //www.mdpi.com/article/10.3390/microorganisms10040712/s1: Table S1: List of animals captured per site at Cembra and Calamento Valleys (Trento, Italy) and PCR results.

Data Availability Statement:
The data presented in this study are available in the Supplementary Materials section and from the EuroSmallMammals database (https://eurosmallmammals.fmach.it (accessed on 13 February 2022)).

Acknowledgments:
We are grateful to three anonymous reviewers for their comments that greatly improved the clarity and outcomes of this article. We are also grateful to all the people involved in the rodent trapping and data collection, and especially Arnoldi D., Dal Farra S., Inama E. and Savazza S., as well as Casari S. and Rosso F., for their suggestions in molecular analysis. We thank all the members of the Applied Ecology research unit, Animal Ecology research unit, and Conservation Genomics research unit of Fondazione Edmund Mach.

Conflicts of Interest:
The authors declare no conflict of interest.

Appendix A
To investigate the role of host species and environmental context (explanatory variables) on the probability of rodent infection with Hepatozoon spp. (response variable), a set of potentially biological meaningful covariates was considered, also controlling for life history traits, and specifically: Specifically, we applied two analytical designs: (i), we evaluated the effect of anthropogenic pressure on rodent infection by comparing the study areas (Cembra: high pressure; Calamento: low pressure), only for altitudes homogeneous across the sites; (ii), we assessed how the probability of infection would vary across species along the alpine altitudinal gradient in a particularly wild context (Calamento Valley: 5 belts of 500 m from 500 m a.s.l. to 2500 m a.s.l.). To this end, we fitted Generalized Linear Mixed Models (GLMMs; [54]) with binomial distribution of errors to the probability of infection. To account for spatial autocorrelation of captures, we performed a first model selection on random effects based on second-order Akaike Information Criterion (AICc), by maximizing the fixed effects (i.e., considering the full model). In particular, we fitted model with random effect on trap nested within grid ('Plot/Trap'), only on trapping grid ('Plot') and only on trap ('Trap') (Table A1,  Table A4). After the selection of the most appropriate random effect [83], we performed a model selection on fixed effect starting from the full model, by ranking all possible model combinations derived from the full model on the basis of the AICc (Table A2, Table A5). The explanatory variables included in the full model were: small mammal species ('Species'), breeding status ('Status'), sex ('Sex'), anthropogenic pressure characterizing the area ('Area'; used only for first analysis) and altitudinal belts ('Altitude'; used only for second analysis). The model with the lower AICc was selected as the best model [55] (Table A3, Table A6).

Appendix A.1 Models across Study Areas with Different Anthropic Pressure
The most parsimonious model included 'Species' as fixed effect, and trapping grid as random effects (Tables A1 and A2). The model coefficients revealed a strong positive effect of species and specifically, Myodes glareolus was more prone to becoming infected if compared with Apodemus flavicollis (Table A3; Figure A1). Conversely, the area and therefore the degree of anthropic pressure was not retained in the best model. Table A1. Model selection according to AICc to identify the random effect. In bold the model selected. Legend: 'Area': anthropic pressure depending on study area; 'Sex': sex; 'Status': breeding status; 'Species': rodent species, 'Plot': trapping grid; 'Trap': trap station; Fixed effect' = fixed effect modelled; 'Random effect' = random effects modelled; 'AICc' = AIC with a correction for small sample sizes; '∆AICc' = relative differences between the fitted model and the Akaike 'best-ranked' model with the smallest AICc value; 'weight' = relative likelihood of a model.   Table A2). strong positive effect of species and a slightly positive effect of breeding status. Specifically, we confirmed that Myodes glareolus positively drove the infection of Hepatozoon spp. rather than Apodemus flavicollis and Chionomys nivalis (Table A6; Figure A2). Moreover, juvenile animals tended to be more infected compared to adults and sub-adults (Table A6; Figure A3). Conversely, elevation was not selected in the best model.   Table A5).  Table A5).