Update on epidemiology of canine babesiosis in Southern France

Background Canine babesiosis is an emerging or re-emerging disease caused by Babesia and Theileria protozoans, also called piroplasms, transmitted by Ixodid ticks. In Europe, four etiological agents have been identified to date, namely Babesia canis, B. vogeli, B. gibsoni and Theileria annae. France has a high prevalence of canine babesiosis and two tick species, Dermacentor reticulatus and Rhipicephalus sanguineus, are supposed to transmit B. canis and B. vogeli respectively. In southern France, where dog infections with B. vogeli were recently confirmed, no comprehensive study was performed to date on piroplasm species infecting dogs. Thus, a large scale survey involving veterinary clinics, kennels and tick collection from the environment was conducted from 2010 to 2012 in this area. Results From 2010 to 2012, 140 dog blood samples and 667 ticks were collected. All blood and a subset of ticks were screened for the presence of piroplasms by PCR amplification of 18S rDNA. B. vogeli, B. canis and T. annae were detected in 13.6, 12.9 and 0.7 % dogs respectively. B. vogeli and B. canis were detected in 10.5 % and in 1.6 % R. sanguineus ticks including 1.3 % co-infections. B. canis was the only species detected in D. reticulatus ticks (9.7 %). B. canis infections were only recorded in the southwest of France whereas B. vogeli was mainly found in the southeast. Finally, a significantly higher prevalence of B. vogeli infection was found in Gard compared to Corsica and Drôme regions, both in dogs (p < 0.002) and R. sanguineus ticks (p < 0.02) although R. sanguineus was the main ticks species removed from dogs in those three areas. Conclusions The survey confirmed the circulation of both B. canis and B. vogeli in dogs in southern France with differences in distribution probably linked to the distribution of their respective vectors. It also showed differences in prevalence of B. vogeli infection in areas similar in terms of risk of dogs infestation with R. sanguineus. Further studies focusing on genetic and microbiota of R. sanguineus ticks should be conducted to explore other biological interactions that may explain the differences observed.


Background
Babesiosis (or piroplasmosis) is an emerging or reemerging tick-borne disease caused by intraerythrocytic protozoa of the genera Babesia and Theileria, also known as piroplasms [1]. In dogs, infection by these parasites usually induces a syndrome characterized by hyperthermia and anaemia that can be fatal when complicated [1]. The severity of the disease depends on various factors including the Babesia/Theileria species involved and the age and immune status of the host [1]. In the last few decades, thanks to molecular biology, new piroplasma species were evidenced bringing to at least seven the number of species able to infect dogs in the world [1][2][3]. In Europe, four of these species have been identified to date, namely B. canis, B. vogeli, B. gibsoni and B. microti-like also known as Theileria annae, Babesia sp. "Spanish dog" or Babesia vulpes sp. nov [3,4]. B. canis is the most widely distributed species coinciding with the distribution of its known vector Dermacentor reticulatus [5,6]. B. vogeli is generally found around the Mediterranean basin where R. sanguineus is the predominant tick species [3]. Infections by B. gibsoni and B. microti-like seem to remain sporadic and their vectors are currently unknown. In addition, DNA of Theileria equi, Babesia caballi and Babesia rossi have been very exceptionally detected in dog blood samples in Europe however the epidemiological significance of these observations is unclear [7,8].
France has a high prevalence of canine babesiosis overall but incidence rates vary considerably according to the seasons and locality [9][10][11]. Although, the disease occurs almost in the whole country, more cases are generally reported in the southwest, including west of the Mediterranean basin, while they are more scarce in Corsica and east of the Mediterranean basin [12][13][14]. Both vectors, D. reticulatus and R. Sanguineus, are supposed to transmit Babesia canis and Babesia vogeli respectively in the country but knowledge on their distribution, pathogens they transmit as well as climatic conditions favorable to transmission is still lacking, especially in southern France where reports are scarce or interested small foci [5,8,9,15]. Moreover, the taxonomic status of R. sanguineus remain controversial and it is now clearly established that several species are gathered into the so called R. sanguineus group [16]. In particular, it was shown that ticks of the R. sanguineus group can be divided into two main lineages: ticks of "temperate" and "tropical" areas [17][18][19]. The impact of those genetic variations of ticks among the R. sanguineus group on B. vogeli transmission has not been evaluated to date. In the continuity of our previous reports [11,15] a large scale survey was conducted in the French Mediterranean basin in order to (i) enhance knowledge on prevalence of Babesia vogeli and other piroplasm infections in dogs (ii) try to provide key elements of response to the inhomogeneous distribution of canine babesiosis cases and (ii) appraise the importance of each vector, R. sanguineus and D. reticulatus, in the transmission of the disease in this region.
From 2010 to 2012, dog blood samples and ticks were collected from veterinary practices, kennels and the environment in southern France and tested for the presence of piroplasms using molecular tools. The study confirmed the circulation of Babesia canis and B. vogeli in dogs with differences in prevalence of infections between regions. It also supports the role of both R. sanguineus and D. reticulatus ticks in the transmission of the disease in this area.

Sampling
The survey was conducted at 13 locations in southern France ( Fig. 1; Table 1) and involved veterinary clinics, kennels and tick collection from the environment. Study sites were selected in order to cover regions with different ecological characteristics. To enhance knowledge on prevalence of infection with Babesia vogeli and potential risk of babesiosis transmission to dog by R. sanguineus ticks, six of the sampling locations were situated in the Gard department, an area where were recently confirmed several cases of infections of dogs by this pathogen [15]. For this specific area, meteorological data (mean monthly maximum and minimum temperatures and rainfall) were obtained from the French national meteorological service (http://france.meteofrance.com/, Nîmes-Courbessac station).
Dogs were included in the survey in case of clinical suspicion of canine babesiosis and/or infestation with ticks. In kennels, dogs housed for more than 3 years were also selected. For all dogs included, a questionnaire was completed in order to obtain background information on pets in particular the usual place of residence, the history of any recent travel, the vaccination status regarding babesiosis and informed consent of owners. The analyses were realized in a context of diagnosis in dogs infested by ticks or suspect of canine babesiosis, thus no approval by an ethics committee was required. Finally, only dogs living around the sampling area without known travel history were kept for subsequent analyses. For each dog, blood was sampled from the cephalic or the jugular veins and stored in collection tubes with EDTA (anticoagulant) at 5°C ± 1°C until DNA extraction within 5 days of collection. A blood smear was also made extemporaneously with fresh or whole blood in EDTA, air-dried, stained using the May Grünwald Giemsa method and stored at room temperature before observation under light microscopy to search for intra-erythocytic forms of piroplasms.
Ticks were collected from infested dogs or from the environment by visual picking or flagging. All ticks were stored in 70 % ethanol until identification.

Ticks identification
Ticks were all identified under light microscopy using morphological identification keys [20,21]. For ticks of the R. sanguineus group, morphological identification was stopped at the group level because of the actual confused taxonomic status of several species inside the group. To unequivocally identify some R. sanguineus tick specimens, PCR amplification of the mitochondrial 16S rRNA and 12S rRNA genes ( Table 2) and BLAST analysis of the sequences was done as previously described [15]. In particular, all ticks of the R. sanguineus group retrieved from the environment were identified by the molecular method because of the greater variety of species possibly collected when using the flagging method (without a selective host trap).
DNA extraction from blood and tick samples DNA was extracted from dog blood samples and a subset of ticks for diagnostic PCR analysis to detect piroplasms. Since R. sanguineus and D. reticulatus ticks are the only known vectors of piroplasms in Europe to date and the possibility of piroplasm transmission to dogs by other ticks species has never been demonstrated, only ticks belonging to these two species were kept for subsequent analysis. All dog blood samples and all R. sanguineus and D. reticulatus ticks retrieved from the environment were analyzed. However, for ticks collected from dogs, DNA was extracted from all specimens if the tick infestation level was low and from a subset of specimens selected at random, in numbers proportional to the stage and sex (e.g. larva, nymph, adult male and female) of the ticks found on the host, in cases of heavy infestation. DNA was extracted as previously described [15] and the quality of extracted DNA was assessed by PCR amplification of mitosin or mitochondrial 16S rRNA genes specific for dogs and ticks respectively ( Table 2). Quantification of total DNA was systematically performed after each DNA extraction using a spectrophotometer (Nanodrop ® ).

Molecular methods for piroplasms detection and characterization
All tick and blood DNA samples were screened using two PCR methods : (i) a Babesia/Theileria genera specific nested PCR method using primer sets BTF1/BTR1 and BTF2/BTR2 that was shown to be very sensitive for the detection of piroplasms on blood [22] and (ii) a specific PCR that was recently validated for detection of Babesia vogeli in ticks (Table 2) [15]. The amplification reactions for Babesia/Theileria genera specific nested PCR were carried out in a thermocycler (Biometra ® T gradient, Goettingen, Germany) in 25 μl of reaction mixture containing 30 ng of extracted DNA, 200 μM (each) deoxyribonucleotides, 0.625 U of Hotstartaq DNA polymerase (Qiagen, Hilden, Germany) and 12.5 pmol of each primer in the reaction buffer provided by the manufacturer (Qiagen, Hilden, Germany) [22]. B. vogeli specific PCR was performed as explained before [15]. Départements where ticks and/or blood sampling where performed are indicated in grey. The survey was conducted at 13 locations indicated by black dots corresponding to veterinary practices, kennels or the environment as detailed in Table 1. Sites codes are as follows: CO, Corsica; DR, Drôme; HE, Hérault; HG, Haute-Garonne; HP, Hautes-Pyrénées; GA, Gard. We thank J.F. Bradu for providing the basemap of French départements Positive (B. rossi and B. vogeli DNA) and negative (reaction mix without DNA) controls were used in each PCR assay. Amplified DNA fragments were separated by electrophoresis through 1.5 % agarose gels stained with ethidium bromide and visualized under ultraviolet light. All PCR reactions were strictly performed according to the protocols described in the original papers.
To discriminate between species within the Babesia/ Theileria positive samples detected by nested PCR, a Restriction Fragment Length Polymorphism method (RFLP) was tested on PCR products from the second round of amplification. Amplified fragments of 800-bp 18S rDNA (10 μL) were digested using TaqI (65°C) and HinfI (37°C) enzymes (Promega, Madison, WI, USA) for 3 h according to a protocol adapted from Carret et al. [23]. Restricted fragments were examined by electrophoresis on 2 % agarose gels and the profiles were compared to those of the 7 species that were detected to date in dog blood samples in Europe B. caballi, B. canis, B. gibsoni, B. rossi, B. vogeli, T. equi, and T. annae obtained from infected animals or parasite cultures provided by different laboratories. The identities of these positive controls had been previously confirmed by PCR amplification, sequencing and BLASTN analysis of 18S rRNA gene.
Finally, most Babesia/Theileria positive PCR products from dogs and ticks were sequenced at BIOFIDAL-DTAMB (FR BioEnvironment and Health, Lyon, France) to assess the possible genetic variability of species and ensure

Statistical analysis
Statistical analyses were performed using R software [24]. Prevalences of infection in each area were estimated by calculating the ratio of positive samples divided by the total number of samples analyzed. Prevalences of B. vogeli infections in dogs and R. sanguineus ticks from Gard, Corsica and Drôme were compared using a Fisher's exact test with a significance threshold defined as p < 0.05.

Nucleotide sequence accession numbers
Sequences of piroplasms obtained from bloods and/or ticks were deposited in GenBank with the following accession numbers: JX304662 to JX871892 for B. vogeli; KC902833 and KC593877 to KC593879 for B. canis; and JX454779 for T. annae. Mitochondrial 16S rDNA sequences from ticks of the Rhipicephalus group affiliated to R. sanguineus "temperate" species after BLAST analyses were deposited in GenBank with the following accession numbers: JQ362399 to JQ362409 and JX304684 to JX304708. Mitochondrial 12S rDNA sequences from ticks of the Rhipicephalus group affiliated to R. sanguineus "temperate" species were deposited in GenBank with accession numbers JX304709 to JX304744. Mitochondrial 16S rDNA sequences from ticks of the Rhipicephalus group affiliated to R. pusillus after BLAST analyses were deposited in GenBank with accession numbers KC593861 to KC593876. Only sequences of excellent quality were deposited in GenBank.

Samples collection and ticks identification
From 2010 to 2012, 155 dogs were selected from which 140 bloods and 635 ticks were retrieved (Tables 1 and 3

Specificity of nested PCR-RFLP for piroplasms detection
A nested PCR-RFLP method was tested with the seven piroplasm species detected to date in dog blood samples in Europe (Fig. 2). This method confirmed its ability to discriminate the seven Babesia/Theileria species tested. Its accuracy for detection of Babesia/Theileria species in dog blood samples as well as in tick samples was established since no inappropriate bands were observed. Consequently, this method was used to screen for piroplasms in dog blood samples and ticks.

Piroplasms detection in dogs
All blood samples were screened for the presence of piroplasms using Babesia/Theileria specific nested PCR-RFLP as well as a B. vogeli specific PCR method. Results of piroplasm detections in blood samples are shown in Table 4. Among

Piroplasms detection in R. sanguineus and D. reticulatus ticks
Results of piroplasms detection in ticks are presented in Table 5.
Among the 588 R. sanguineus ticks collected from dogs (N = 574) and the environment (N = 14 after removal of the ticks finally affiliated to the species R. pusillus) 57.3 % were females, 35.2 % were males and 7.5 % were nymphs. Among them, a representative set of 248 R. sanguineus ticks were selected for piroplasm screening as follows: 121 from Gard (63. 6

Seasonal distributions of infections
As sampling from dogs was done regularly in Gard, it was possible to follow R. sanguineus tick activity,   Table 1 and Fig. 1 (Fig. 3). In 2010 and 2011, R. sanguineus ticks were mainly collected from dogs from March to June with a peak in infestation in April. Canine babesiosis cases, mainly caused by B. vogeli in this area, were recorded from March to July with a similar peak in observations in April. Interestingly, during the 29 months of the survey, first dog infestations with R. sanguineus were detected when the mean monthly temperature went above a threshold value of 10°C in spring.
All B. canis infection cases reported in other areas in the study were recorded in winter or very early spring and when associated with tick infestations, only D. reticulatus was found.

Discussion
Knowledge on epidemiology of canine babesiosis in southern France is scarce, in particular regarding Fig. 2 PCR-RFLP profiles of 18S rRNA gene fragments from selected piroplasm species. This method was used to discriminate between Babesia/Theileria species from ticks or blood samples known to contain piroplasms following nested PCR targeting the 18S rRNA gene. Lanes 2 to 8 and lanes 10 to 16 show PCR-RFLP products for seven piroplasm species known or supposed to infect dogs in France and in Europe digested with endonucleases TaqI or HinfI, respectively. Lanes 1, 9 and 17 show 100-bp molecular weight markers. Product sizes expected for each piroplasm species after digestion with TaqI or HinfI enzymes are given in the table   Table 1 and Fig. 1 for details) b Number of positive sample/Number analyzed (percentage) per location c RFLP and sequencing confirmed the affiliation of the protozoa to Theileria annae piroplasm species circulating in dogs and tick species involved in their transmission. Moreover, an heterogenic distribution of canine babesiosis cases was previously reported in southern France [12][13][14] and remain unexplained. Recently we confirmed infection of 4 dogs and 8 R. sanguineus ticks with B. vogeli in Gard [15], a department from this region, which prompted us to continue and extend the survey to several locations. Thus, a total of 140 dog blood samples and 667 ticks were collected and screened by PCR for piroplasm detection and characterization. The original sampling method used in the study allowed making the link between pathogens detected in dogs and in ticks retrieved from dogs which is essential for accurate results interpretation. From the 667 ticks collected either on dogs or from the environment in this area, 605 belonged to the R. sanguineus group from which 574 were retrieved from dogs confirming the high propensity of this tick species to parasitize dogs in this area. For ticks of the R. sanguineus group, because of the controversial taxonomic status of tick species in this group, it was decided to analyze selected specimens using molecular tools. Among the 64 ticks analyzed by sequencing, 47 were finally affiliated to "temperate" species of R. sanguineus (= southern lineage) [16][17][18][19], in particular all ticks retrieved from dogs, whereas 17 were affiliated to the species R. pusillus (all collected from the environment). R. sanguineus and R. pusillus ticks species obtained by flagging from the environment were all collected along watercourses susceptible to be frequented by dogs. New flagging campaigns should be performed in the future to confirm the possibility of exophilic development of R. sanguineus ticks in this area. On the other hand, collection of R. pusillus ticks, a species generally known to infest rabbits, may be attributed to the possible presence of such hosts in sampling locations. Different and complementary methods can be used to evaluate the prevalence of piroplasm infection in dogs. Serology is useless as it lacks species specificity compared to molecular tools that are generally more sensitive and specific [2]. Several PCR-RFLP methods, using TaqI and HinfI or other restriction enzymes were previously developed to discriminate between species capable of infecting dogs, notably B. canis, B. gibsoni, B. rossi, B. vogeli and T. annae [23,25,26]. Molecular methods for piroplasm detection in ticks are less common. Here we modified a very sensitive nested-PCR strategy, previously developed for the detection of Babesia and Theileria in blood [22] to adapt it to the analysis of tick samples and to allow a specific detection of piroplasms that may infect dogs in Europe. Results clearly proved the accuracy of the modified PCR-RFLP method to detect and differentiate the seven piroplasm species tested both in dog blood and tick samples. The method was then used to screen dog blood and tick samples in combination with a B. vogeli specific PCR method that had previously proved its ability to detect the pathogen in ticks [15]. The higher number of B. vogeli samples detected by B. vogeli specific PCR compared to Babesia/Theileria nested-PCR showed the higher sensitivity of B. vogeli specific method for B. vogeli detection in bloods and ticks and support the interest of associating both PCR methods in the study. However, as the sensitivity of the methods was only tested on blood samples the possible impact of the samples (ticks vs blood) on the sensitivity of the methods remains unknown.
Overall, three piroplasm species, B. vogeli, B. canis, and T. annae were detected in dogs allowing confirmation of the circulation of both B. vogeli and B. canis in dogs in southern France. For Theileria annae, further studies will be needed to assess the prevalence of infection in dogs in   Table 1 and Fig. 1 [5,27] and it is assumed that each Babesia species is associated with a single tick vector in a delimited geographic area [28]. DNA of B. canis was previously found in R. sanguineus ticks in Italy and vertical transmission was suggested [29,30]. The data presented here, in particular, the strong relation between geographical and temporal occurrence in dogs of B. vogeli and R. sanguineus on one hand and B.
canis and D.reticulatus on the other hand, are highly suggestive of tick-pathogen specificity as generally accepted. However further study should be conducted to assess if B. canis infected R. sanguineus ticks are capable of transmitting the pathogen to dogs. An heterogenic distribution of canine babesiosis cases was previously reported in southern France [12][13][14]. In the study, B. canis was the more frequently detected piroplasm species in dogs in Hérault, Haute-Garonne and Hautes-Pyrénées, 3 departments situated in the southwestern part of France. In Gard, Drôme and Corsica, all situated in the south-eastern part of France, B. vogeli was the main piroplasm species detected in dogs except in Corsica where none of the 36 dog blood samples analyzed was found to be infected with Babesia/Theileria species. Although the number of blood samples obtained from western Mediterranean basin was limited, these results  [15]. a Number of R. sanguineus, D. reticulatus and Ixodes ricinus collected each month from dogs and number of canine babesiosis cases caused by B. vogeli. b Number of R. sanguineus collected each month and mean monthly temperature and rainfall. Only ticks retrieved from dogs in veterinary clinics were included in the analysis to avoid potential bias due to the high number of ticks obtained during sampling campaigns in kennels or the environment suggest that B.canis and B. vogeli could be the main etiological agent of canine babesiosis in south-western and south-eastern France respectively. Moreover, a significantly higher prevalence of B. vogeli infection was found in Gard compared to Corsica and Drôme regions, both in dogs (p < 0.002) and R. sanguineus ticks (p < 0.02). Thus, the study provide new data on epidemiology of canine babesiosis in southern France that can partly explain heterogenic distribution of canine babesiosis cases previously reported in this area. However, analysis of a higher number of cases should be conducted in the future to confirm the distribution of pathogens suggested by the study and confront them to biological and ecological characteristics of vectors. Further studies focusing on genetic and microbiota of R. sanguineus ticks should be conducted as well to explore other biological interactions that may explain the differences observed.
Six of the sampling locations were situated in the Gard department to enhance knowledge on risk of dogs infection by the protozoan B. vogeli that was recently evidenced in this area [15]. In Gard, B. vogeli was detected in 12/49 (24,5 %) of dog and in 20/121 (16.5 %) R. sanguineus ticks. Among the 20 B. vogeli positive ticks, two were non engorged ticks collected by flagging from the environment. These results are consistent with the hypothesized endemicity of B. vogeli infections in this area as previously suggested [15]. Interestingly, evidence of a relationship was found between the first infestations of dogs with R. sanguineus in spring and an increase in the mean monthly temperature above a threshold value of 10°C during the 29 months of the study. Such study associating epidemiological and ecological observations should be continued in the future to assess if such a factor could be a useful indicator of R. sanguineus infestation and B. vogeli infection risk.

Conclusions
The survey conducted from 2010 to 2012 confirmed that both B. canis and B. vogeli piroplasms and their respective vectors D. reticulatus and R. sanguineus parasitize dogs in southern France with differences in distribution probably linked with vectors biotopes. In particular, B. vogeli seems to be the most prevalent piroplasm species in the south-eastern part of the country with a higher prevalence of infection found in the Gard area than in Drôme and Corsica. Better understanding of the spatial and temporal distribution of piroplasm species involved in canine babesiosis cases is needed to improve preventive measures and assess the efficacy of treatments and vaccines currently available in France. This implies a better understanding of ecology, genetic and microbiota of vectors to explore other biological interactions that may explain the differences observed.