Canine tick-borne pathogens in Cyprus and a unique canine case of multiple co-infections. Ticks and Tick-borne Diseases, 8, 341-346.

Canine tick-borne pathogens such as Ehrlichia canis and Hepatozoon canis are widespread in the Mediter- ranean basin but have never been reported or investigated in Cyprus. We describe herein the presence of canine tick-borne pathogens in three dogs with clinical signs compatible with vector-borne diseases from Paphos area of Cyprus. Molecular and phylogenetic analysis revealed the presence of E. canis , Anaplasma platys , H. canis , Babesia vogeli and Mycoplasma haemocanis in Cyprus. One dog co-infected with E. canis , H. canis , B. vogeli and M. haemocanis is, to the best of our knowledge, the ﬁrst report of this multiple co-infection in dogs. The tick-borne pathogens reported in the current study should be considered in the differential diagnoses in dogs exposed to ticks in Cyprus.


Introduction
Cyprus is a European Union member island state situated in the eastern Mediterranean basin (35 • 10 N and 33 • 22 E) with a temperate climate. Several tick-borne pathogens are endemic in domestic ruminants (goats, sheep and cattle) (Chochlakis et al., 2012Psaroulaki et al., 2009), wild animals (mouflon, fox and hare) (Chochlakis et al., 2012;Ioannou et al., 2011;Psaroulaki et al., 2014) and migratory bird populations of the island . Rhipicephalus sanguineus sensu lato, the main vector for several canine tick-borne pathogens (CTBP), is the most common tick species found in dogs in Cyprus (Chochlakis et al., 2012;Dantas-Torres, 2010;Le Riche et al., 1974). The only vector-borne disease (VBD) previously reported in dogs in Cyprus is leishmaniosis (Mazeris et al., 2010). Up to now, no studies have yet reported the presence of CTBP in Cypriot dogs.
This report describes, for the first time, the molecular detection of Ehrlichia canis, Anaplasma platys, Hepatozoon canis, Babesia vogeli and Mycoplasma haemocanis in three naturally-infected dogs from Cyprus with clinical signs compatible with VBD.

Material and methods
Three dogs from Cyprus, designated as Cases 1, 2 and 3, with no overseas travel history and with clinical signs compatible with VBD, were admitted to Cyvets Veterinary Center, at Paphos, in 2013 and included in this study.
At clinical presentation, as part of diagnostic procedures, blood samples were collected into EDTA and heparin tubes and analysed by an automated haematology impedance analyzer (Vet ABC, Scil, Viernheim, Germany) and by a VetScan VS2 chemistry analyzer (Abaxis, USA), respectively. Blood smears prepared on glass slides at the time of blood collection were air-dried and stained with Giemsa (Merck, Darmstadt, Germany) and examined by light microscopy. Additionally, plasma samples were evaluated for the presence of antibodies against E. canis/E. ewingii, Borrelia burgdorferi sensu lato, A. platys/Anaplasma phagocytophilum and the presence of Dirofilaria immitis antigen using the SNAP 4Dx ® Plus test (IDEXX Laboratories, Westbrook, Maine, USA). Urine samples were collected via cystocentesis and analysed using the Multistix 10SG urine reagent strips (Siemens AG, Munich, Germany). Following diagnostic testing and with written owner consent, surplus stored EDTA-blood was used retrospectively for this study. DNA was extracted from 100 mL of EDTA-blood using a commercial kit (NucleoSpin ® Blood, Machery-Nagel, Germany) according to the manufacturer's instructions. Real-time quantitative polymerase chain reaction (qPCR) assays were used to investigate for the presence of haemotropic mycoplasmas (M. haemocanis and 'Candidatus Mycoplasma haematoparvum'), B. burgdorferi sensu lato and Bartonella henselae, conventional PCRs were used to evaluate for Ehrlichia/Anaplasma spp., Hepatozoon spp., Babesia spp. and Leishmania spp. infection, and a nested E. canis-specific PCR assay was also performed (Table 1) (Barker et al., 2010;Christodoulou et al., 2012;Dawson et al., 1996;Demaerschalck et al., 1995;Inokuma et al., 2002;Matjila et al., 2008;Oosthuizen et al., 2008;Parola et al., 2000;Robinson et al., 2010). To confirm the presence of amplifiable DNA and absence of PCR inhibitors, all qPCR reactions were duplexed with an internal control for glyceraldehyde-3-phosphate dehydrogenase. For each PCR run, positive controls using DNA from known infected dogs and negative controls using water were included. After conventional PCR amplification, the amplicons from any positive Ehrlichia/Anaplasma spp., Hepatozoon spp., Babesia spp. and E. canis PCR assays were purified (ExoSAP-IT, Affymetrix, USB, Cleveland, Ohio, USA) and the DNA sequenced on both strands using the respective forward and reverse PCR primers. For each DNA amplicon forward and reverse sequences were assembled, and a consensus sequence constructed, using ClustalW in Bioedit. DNA sequences were deposited in GenBank and checked for identity against previously deposited sequences, using NCBI BLAST (Altschul et al., 1990). Sequences obtained in this study were aligned to selected sequences of the respective genes, from the same and related species of organism, previously deposited in Gen-Bank using ClustalW. Phylogenetic trees were constructed using the maximum likelihood program, corrected for nucleotide substitutions by the Kimura-2 parameter model, in MEGA version 6 (Tamura et al., 2013). The data set was resampled 1000 times to generate bootstrap percentages.

Case reports
3.1. Case 1 A 1.5-year-old, entire male, mixed-breed dog presented with a 6-day history of depression and anorexia. The dog lived outdoors in rural Paphos-area. Physical examination revealed pale mucous membranes with widespread petechial haemorrhages. Body condition score was 4/9. Haematology showed a mild normochromic normocytic anaemia (haematocrit [Hct] 34.5%; reference interval [RI] 44-57%), monocytosis (1.4 × 10 9 /L; RI 0.0-0.5 × 10 9 /L), lymphocytosis (4.0 × 10 9 /L; RI 1.0-3.6 × 10 9 /L) and a moderate thrombocytopenia (65 × 10 9 /L; RI 200-460 × 10 9 /L), while no abnormalities were detected on biochemistry and urine analysis. Giemsa stained peripheral blood smear examination revealed the presence of gamonts of Hepatozoon spp. intracellularly in 0.2% of neutrophils. Seropositivity for E. canis/E. ewingii was also detected. Based on the physical examination, haematological abnormalities and serological results, a diagnosis of Hepatozoon spp. infection was established, with presumptive co-infection with Ehrlichia spp. Treatment with doxycycline (Ronaxan, Merial, Lyon, France; 5 mg/kg orally twice daily for 28 days) was given. Abnormalities were not detected on repeat physical examinations, haematological analysis and blood smear examination on days 28 and 134 following treatment initiation. Subsequent PCR on pre-treatment stored blood samples and sequencing analysis confirmed both E. canis and H. canis infection and revealed concurrent infection with B. vogeli and M. haemocanis.

Negative Negative Negative
Borrelia burgdorferi (ospA gene) (Demaerschalck et al., 1995) SL Lyon, France; 5 mg/kg orally twice daily for 28 days). On days 28 and 134 following treatment, the dog was described as being healthy by the owner via telephone communication, but repeat examination was declined. Subsequent PCR on stored pre-treatment blood samples and sequencing analysis confirmed A. platys infection.

Sequence and phylogenetic analysis
The sequencing results of products derived from positive Ehrlichia/Anaplasma spp., Hepatozoon spp., Babesia spp. and E. canis PCR assays, together with the lengths of the sequence data obtained, are shown in Table 1, alongside the GenBank accession numbers for the generated sequences.
On phylogenetic analysis, all of the H. canis sequences compared clustered together, including those from Cases 1 and 2, in a single clade, separate from H. felis and H. ursi (Fig. 1A). All B. vogeli sequences compared, including that from Case 1, also clustered together in a single clade, distinct to other Babesia and Theileria spp. (Fig. 1B). All E. canis sequences compared, including those from Cases 1 and 2, clustered together in a single clade supported, distinct to other Ehrlichia and Anaplasma spp. (Fig. 1C). Most of the A. platys sequences compared, including that from Case 3, clustered closely together in a single clade (Fig. 1D), although one A. platys sequence from China branched separately, albeit with a relatively low bootstrap value.

Discussion
Canine tick-borne pathogens such as E. canis, H. canis, B. vogeli and M. haemocanis have been documented in several European countries especially in the Mediterranean region where the temperate climate sustains the presence and activity of R. sanguineus sensu lato which is considered the major vector for the transmission of these pathogens (Baneth, 2011;Novacco et al., 2010;Sainz et al., 2015; Solano-Gallego and Baneth, 2011). Two previous stud-ies have described the canine tick flora found on dogs in Cyprus, with R. sanguineus sensu lato comprising 87.5-89.4% of ticks collected (Chochlakis et al., 2012;Le Riche et al., 1974). No previous studies have investigated CTBP in Cypriot dogs.
Case 1 was diagnosed with likely acute monocytic ehrlichiosis, and even though it was co-infected with three additional CTBP (H. canis, B. vogeli and M. haemocanis) the clinical and haematological outcome did not appear to be worsened. Co-infection has previously been demonstrated in Greek dogs; one study demonstrated that 65% of dogs in Greece, with naturally-occurring monocytic ehrlichiosis, were also seropositive to H. canis (Mylonakis et al., 2005). In the same study, the prevalence of H. canis parasitaemia and the number of circulating gamonts were low, and H. canis infection did not appear to exacerbate the clinical manifestations or the haematological abnormalities of monocytic ehrlichiosis. In Case 1, there was a low degree of H. canis parasitaemia (<5%), which might suggest limited involvement in the clinical signs of the dog (Baneth, 2011). The fact that there was no cytological evidence of H. canis parasitaemia after follow-up cytological review of the peripheral blood smear, coupled with the complete clinical and haematological recovery following doxycycline treatment, suggests that this infection was not a critical determinant in the clinical and clinicopathologic outcome. The clinical relevance of the B. vogeli and M. haemocanis infections in Case 1 is uncertain, as both pathogens tend to affect splenectomized dogs more severely, and their clinical and haematological abnormalities may overlap with those of monocytic ehrlichiosis (Chalker, 2005;Solano-Gallego and Baneth, 2011). Additionally, doxycycline treatment should be effective against M. haemocanis (Chalker, 2005). To the authors' knowledge this is the first report of a dog being co-infected with E. canis, H. canis, B. vogeli and M. haemocanis.
Dual E. canis and Leishmania spp. infections, as seen in Case 2, are commonly reported and tend to have prominent clinical signs and slower recoveries compared to dogs with single infections of either pathogen (Di Loria et al., 2009;Mekuzas et al., 2009). Since these diseases overlap substantially in their pathology (Mylonakis et al., 2008), the clinical and clinicopathologic abnormalities seen in Ehrli chi a c anis Case 1 KU2 551 18

Ehrli chi a c anis Spain AY3944 65
Ehrli chi a c anis Italy EU439944

Ehrli chi a c anis Case 2 KX756 971
Ehrli chi a c anis Is rael U26740

Ehrli chi a m uris AB013008
Ehrli chi a c haffee nsis AF147752  Table 1) and selected gene fragments available from GenBank, and those of related organisms. Phylogenetic trees were constructed by the maximum likelihood method. The data sets were resampled 1000 times to generate bootstrap percentage values, and values greater than 50% are given at the nodes of the tree. this case could have been attributed to either E. canis or Leishmania spp. infection. In a previous study, dogs that presented with epistaxis, as in Case 2, and with mucosal pallor and pancytopenia, were more likely to be affected by monocytic ehrlichiosis, as opposed to the presence of peripheral lymphadenomegaly, which was more indicative of leishmaniosis (Mylonakis et al., 2008). Similar to Case 1, the H. canis infection in Case 2 was likely to be an incidental finding and most likely of limited clinical significance.

Ehrli chi a c anis
Case 3 was diagnosed with thrombocytotropic anaplasmosis, with thrombocytopenia being the typical clinicopathologic finding of A. platys infection. Although a definitive mode of transmission has yet to be demonstrated for A. platys, R. sanguineus sensu lato is thought to be the most likely vector (Sainz et al., 2015). Unlike the well-recognised A. phagocytophilum-induced immune-mediated polyarthritis (Sainz et al., 2015), A. platys has not previously been associated with canine arthritis. Therefore the neutrophilic joint inflammation in this case cannot be attributed to A. platys infection, especially in view of the lack of joint fluid culture or PCR to rule out common pathogens or synovial fluid PCR to amplify A. platys DNA. Further studies may be warranted to address any potential association of A. platys and arthritis.
The phylogenetic trees of the sequences derived in this study, together with existing sequences available in Genbank, showed grouping of species derived from a range of countries in different continents, with no differentiation according to geographical origin. For example, the sequences derived from the Cypriot samples did not cluster more closely with those obtained from other countries in Europe than those reported from countries of other continents. In our study short sequences from a single gene were analysed only, and although no differences were revealed, further investigation with additional phylogeny and amplification of multiple and longer genes is needed, in order to prove that these pathogens are not specifically adapted to any particular region.

Conclusions
The current study provides the first molecular documentation for the presence of tick borne pathogens E. canis, A. platys, H. canis, B. vogeli and M. haemocanis in dogs in Cyprus. One dog was co-infected with E. canis, H. canis, B. vogeli and M. haemocanis and to the best of our knowledge this is the first report of this multiple co-infection in dogs. It further illustrates the widespread distribution of CTBP in Mediterranean countries where R. sanguineus sensu lato resides. Veterinarians should consider these pathogens for dogs living in or having travelled to Cyprus. Furthermore, future studies should be carried out to determine the epidemiology of CTBP in Cyprus.