Molecular Diagnosis, Prevalence and Importance of Zoonotic Vector-Borne Pathogens in Cuban Shelter Dogs—A Preliminary Study

The present study aimed to determine the prevalence of zoonotic vector-borne pathogens, including Anaplasma platys, Anaplasma phagocytophilum, Borrelia burgdorferi sensu lato, Ehrlichia canis and Rickettsia spp. in shelter dogs from Cuba. Blood samples were collected from 100 shelter dogs and examined by molecular methods. Overall, 85 (85%; 95% CI: 77.88–92.12) dogs tested positive for at least one vector-borne pathogen using species-specific qPCR assays. Among the positive samples, E. canis was the most prevalent 62% (95% CI: 52.32–71.68), followed by A. platys 40% (95% CI: 30.23–49.77) and Rickettsia felis 27% (95% CI: 18.15–35.85), whereas 36% (95% CI: 26.43–45.57) showed co-infections. All samples were negative for A. phagocytophilum and B. burgdorferi s.l. The presence of 248 Rhipicephalus sanguineus ticks collected from the dogs was not statistically associated with the occurrence of infections. Thrombocytopenia was the most frequent haematological alteration found in PCR-positive dogs; it was statistically associated with the presence of E. canis, as well as co-infections (p < 0.05). The phylogenetic analyses of A. platys and E. canis based on 16S rRNA, groEL and gltA genes showed a low genetic diversity between Cuban strains. The present study demonstrates the high prevalence of vector-borne pathogens with zoonotic potential in shelter dogs from Cuba.


Introduction
Canine vector-borne diseases (CVBDs) consist of a group of infectious diseases caused by a range of pathogens transmitted by arthropod vectors, including ticks, mosquitoes, fleas and lice [1]. Clinical signs commonly associated with these diseases include anorexia, pyrexia, lethargy, weight loss, bleeding disorders and icterus progressing to fatal outcomes in some dogs [2]. In addition, some CVBD-causing pathogens are a cause of major zoonotic concern and constitute a serious human health hazard worldwide [1].

Results
In total, 100 dog blood samples were collected from 11 municipalities in the Havana province ( Figure 1). Using sensitive species-specific PCR assays and sequence confirmation, E. canis, A. platys, and Rickettsia felis were detected in dogs. Neither A. phagocytophilum nor B. burgdorferi s.l. DNA was identified in any of the dog blood samples included in this study. Out of 100 blood samples, 85 (85%; 95% CI: 77.88-92.12) tested positive for at least one vector-borne pathogen using real-time qPCR assays. Among these positive samples, E. canis was the most prevalent 62% (95% CI: 52.32-71.68), followed by A. platys 40% (95% CI: 30. 23-49.77) and Rickettsia spp. 27% (95% CI: 18. 15-35.85). Dogs were most often co-infected with E. canis and A. platys in 28 (28%; 95% CI: 19.05-36.95), followed by E. canis and Rickettsia spp. in 13 (13%; 95% CI: 6.29-19.71), A. platys and Rickettsia spp. in 11 (11%; 95% CI: 4.76-17.24), and triple mixed infections were detected in 8 (8%; 95% CI: 2.59-13.41) dogs. The results from the real-time qPCR testing are summarized in Table 1. Tick infestation was observed in 57 out of 100 sampled dogs, and a total of 248 ticks were submitted for identification to species level. All the ticks collected were identified morphologically as R. sanguineus, and consisted of 111 females, 131 males and 6 nymphs. The presence of ticks was not statistically associated with the occurrence of PCR-positives infections (p = 0.115), i.e., E. canis  Table S1). Most of the collected ticks were visibly engorged with blood. Ticks were collected throughout the year, and adult ticks were seen in every month during the sample collection. A complete blood count (CBC) was available for 90 of the 100 sampled dogs. Table 2 shows the main statistics values obtained from the haematological parameters determined for the tested animals, which were distributed into five groups, including E. canis-, A. platys-, Rickettsia spp.-PCR positive, co-infected and non-infected dogs. The most common haematological abnormalities among tested dogs included thrombocytopenia ( 49) were more frequent in PCR-positive dogs, for red blood cell count (RBC), haemoglobin concentration (Hb), haematocrit (HCT), mean corpuscular volume (MCV), mean corpuscular haemoglobin (MCH), mean corpuscular haemoglobin concentration (MCHC), and total white blood cell counts (WBCs), there were no statistically significant differences between the five groups (p > 0.05). However, the mean values of platelet counts in E. canis-PCR positive (p = 0.018) and co-infected dogs (p = 0.016), as well as MCH (p = 0.042) and MCHC (p = 0.027) values for co-infected dogs were statistically different from non-infected animals (p < 0.05).  The sequence analysis of the nearly full-length 16S rRNA gene sequences obtained from E. canis (1434 bp) and A. platys (1431 bp) Cuban isolates revealed high identities >99% with several sequences of E. canis (e.g., LC269822, EF139459) and A. platys (e.g., EU106856, CP000107) available in GenBank, respectively. The nucleotide sequences obtained in the present study were not 100% identical to each other and may represent local variants that exist within the studied region. In addition, partial E. canis-gltA (507 bp), A. platys-groEL (625 bp) and Rickettsia spp.-htrA (434 bp) gene sequences were obtained from PCR-positive samples. The sequences obtained from ompA and ompB fragment genes were not evaluable. The E. canis-gltA obtained sequences were 100% identical to each other and to reported sequences from China (CP025749), the Philippines (LC428206) and Zambia (LC373038), while, for A. platys-groEL, sequences were 100% identical to each other and to reported sequences from the Democratic Republic of Congo (AF478129), Japan (AY077621) and Venezuela (AF399916). The Rickettsia spp.-htrA (434 pb) were 100% identical to each other and showed high identities >99% with the 17 kDa surface antigen gene sequences of R. felis type reference isolates reported from Mexico (GU447234) and the USA (CP000053). For E. canis-gltA, A. platys-groEL and R. felis-htrA gene sequences no nucleotide variation was observed among the sequenced PCR amplicons.
The phylogenetic analysis based on 16S rRNA gene sequences were grouped into two main clades of Anaplasma spp. and Ehrlichia spp. In addition to A. platys and E. canis strains, closely related species of the tick-borne parasites were included, such as Anaplasma bovis, Anaplasma centrale, Anaplasma capra, Anaplasma marginale, Anaplasma ovis, Anaplasma phagocytophilum, Candidatus Anaplasma camelii, Ehrlichia chaffeensis, Ehrlichia ewingii, Ehrlichia muris, Ehrlichia minasensis and Ehrlichia ruminantium. A biologically divergent member of the family Anaplasmataceae, Rickettsia parkeri was used as an outgroup. As expected, the resultant phylogenetic tree revealed that A. platys and E. canis Cuban isolates were clustered tightly with other A. platys and E. canis strains reported around the world, respectively ( Figure 2). In addition, phylogenetic analysis based on the alignment of the A. platys-groEL partial gene sequences obtained in this study was compared with several A. platys reference sequences, and other Anaplasma spp. found in GenBank. A. platys-groEL Cuban genotype was tightly classified in A. platys cluster grouped with other strains isolated from different host species worldwide, supported with 100% bootstrap value ( Figure 3). Moreover, phylogenetic tree based on the alignment of gltA partial gene sequences of Ehrlichia spp. found in GenBank shows the presence of five clusters represented by E. canis, E. chaffeensis, E. ewingii, E. minasensis, Ehrlichia sp. and Rickettsia monacensis as an outgroup. The E. canis-gltA Cuban strain was grouped within the E. canis clade formed by strains isolated from different host species worldwide, supported with 100% bootstrap value ( Figure 4).

Discussion
To the authors' knowledge, this is the first study addressed to investigate canine arthropod-borne pathogens in stray dogs housed in animal shelters from Cuba, and demonstrated that rescued dogs housed in shelters from the investigated areas showed high prevalence rates for several arthropod-borne pathogens. Eighty five percent of the dogs tested PCR-positive to pathogenic organisms. This finding can be easily explained since stray dogs usually are neither protected by preventive measures against ectoparasites nor receive any proper veterinary care, and therefore are at increased risk of infections by arthropod-borne pathogens. Importantly, all infections detected here have a relevant zoonotic potential since E. canis, A. platys and Rickettsia spp. human infections have been reported from Venezuela [4], Grenada [23], and Spain [24], respectively.
The sample size in the current study was rather small. Nonetheless, it is of importance to note that the overall prevalence infection rate with at least one zoonotic pathogen (i.e., E. canis, A. platys and Rickettsia spp.) recorded during this study was higher than that reported in previous molecular studies conducted in dogs from Italy (44/145; 30.3%) [25], Thailand (78/181; 43.1%) [26], Brazil (118/181; 65.2%) [27], and Haiti (111/207; 53.6%) [28]. This worldwide variation in infection rates are likely attributed to several factors, including the demography of dog populations, the type and number of samples analysed, the extent of tick infestations, and the sensitivities of diagnostic methods employed [29]. E. canis and A. platys were the most prevalent pathogens (62% and 40%, respectively), while Rickettsia spp. was less frequently detected (27%). Moreover, E. canis and A. platys were the most frequently detected either as single infections (29% and 9%, respectively) or as coinfections (28%). Coinfections with multiple tick-borne pathogens in dogs are quite frequently reported [30] and often occur due to concomitant transmission by the same tick vector, R. sanguineus [31], which was the only tick species found in this study. The presence of co-infections is clinically important because it may pose a diagnostic and therapeutic problem as infected dogs frequently show unspecific symptoms, such as fever, weight loss, inappetence, lethargy or apathy, which give no indication of the possible causative agent [2]. Although in this study a high prevalence of mixed infections was observed, the sampled dogs showed no clinical signs consistent with E. canis, A. platys and Rickettsia spp. infections more than pale mucosa, anorexia, apathy, dehydration and poor body condition. Most of the dogs were asymptomatic, which may reflect the existence of chronic, subclinical or mild infections that makes the clinical diagnosis of infected dogs in the studied region difficult [32].
Prior to this study, Navarrete et al. [21] described the presence of E. canis as a canine tick-borne pathogen in Cuba; however, to date, the thrombocytopenia in dogs was a problem of unknown aetiology. In this study, the presence of E. canis infections was associated with significantly lower platelet count values compared to non-infected dogs (p = 0.018), this fact was even more significant when co-infection was considered (p = 0.016). In general, this result is consistent with several studies carried out under both natural and experimental conditions, which concluded that thrombocytopenia is the major haematological abnormality associated with E. canis infections [33]. The presence of thrombocytopenia is commonly used alone as a useful haematological marker of E. canis infection for the diagnosis of CME. However, a previous study conducted by Santos et al. [34] demonstrated that the diagnosis of E. canis infection in dogs just based on the occurrence of thrombocytopenia is not sufficient, and screening for other tick-borne pathogens such as A. platys and Babesia spp. is recommended to reach a definite diagnosis.
The resultant phylogenetic analysis based on the nearly full length 16S rDNA sequences revealed that A. platys and E. canis strains from Havana, Cuba, were tightly grouped with other A. platys and E. canis isolates from dogs around the world (Figure 2). These results are consistent with a previous report described by de la Fuente et al. [35], which support the hypothesis that A. platys strains are neither geographically nor host segregated. In addition, a highly conserved genetic profile was observed for the A. platys and E. canis strains based on the groEL and gltA partial gene sequences analysis, respectively. Sequence analysis of the E. canis-gltA and A. platys-groEL genes performed on three Cuban strains revealed 100% identity, even though the analysed samples were obtained from different areas. The sequence alignments and phylogenetic trees suggested little genetic diversity and homogeneous evolution within A. platys and E. canis strains, based on the close similarity amongst their 16S rDNA, groEL and gltA sequences from geographically diverse areas studied in this report. The results obtained were in concordance with previous reports of slight genetic variation between sequenced genes from different A. platys and E. canis strains [36,37]. The choice of molecular markers with an appropriate mutation rate is an essential step in phylogenetic analysis. Consistent with previous investigations conducted by Marsilio et al. [38] and Ben Said et al. [20], nucleotide variability of both groEL and gltA genes have proved be useful as markers to clarify evolutionary relationship and correct identification among Anaplasma ssp. and Ehrlichia spp., respectively. These conclusions are consistent with other reports, in which gltA and groEL genes indicated higher interspecies nucleotide variability than that observed for the 16S rRNA gene [38,39]. However, further studies are needed in Cuba to investigate the genetic variability among different A. platys and E. canis strains.
The presence of A. phagocytophilum and B. burgdorferi s.l. DNA was not identified in any of the blood samples tested. The negative results for A. phagocytophilum and B. burgdorferi s.l. is in accordance with the absence of the main vector of these pathogens, which are hard ticks other than R. sanguineus that have never been reported in Cuba [40]. Regarding B. burgdorferi s.l., there is a previous report in Cuba by Rodríguez et al. [41] that described the detection of antiborrelial antibodies and clinical signs resembling Lyme disease in humans, but, according to our research, the existence of B. burgdorferi s.l. still remains unproven.
Twenty-seven (27%) of 100 samples were positive in the Rickettsia species DNA screening qPCR, and the subsequent sequence analysis identified R. felis presence. The qPCR assay used in the present study was developed by Stenos et al. [42], and can detect most Rickettsia species in the spotted fever and typhus groups with a high specificity and sensitivity, capable of detecting one target copy gene per PCR reaction. However, despite a high prevalence of Rickettsia species found in Cuban dogs by gltA gene real-time qPCR, we were only capable of obtaining a partial Rickettsia-htrA gene sequence by conventional PCR. Unfortunately, rickettsial DNA could not be amplified in any of the samples when tested by PCR based on ompA and ompB genes, limiting additional phylogenetic inferences. The variable successful amplification of different genes is likely explained by the fact that the molecular detection of rickettsial DNA from blood samples based on conventional PCR shows low sensitivity. This point of fact is due to the pathogenic mechanisms of Rickettsia spp. that once infect endothelial cells, the bacterial load in blood is decreased until too low numbers, which makes them incapable of being detected by molecular analysis [43]. This is the first detection of R. felis, a member of the spotted fever group Rickettsia (SFGR), infections in dogs from Cuba. A study conducted by Ng-Nguyen et al. [44] demonstrates based on molecular evidence the role of the domestic dog (Canis lupus familiaris) as a mammalian reservoir for R. felis and as a potential source of human rickettsial infection. A previous study conducted by Noda et al. [45] described the detection of "Candidatus Rickettsia amblyommii" in Amblyomma mixtum tick species by PCR, which constitutes the first report of an SFGR member in Cuba. The high prevalence of R. felis found in the tested Cuban dogs highlights the substantial importance of this pathogen on human health, since Rickettsiosis has become a re-emerging problem worldwide and suggests it may be causing unreported or unstudied SFGR in Cuba. The SFGR infection among dogs in Cuba was interesting given the findings that include descriptions of the occurrence and clinical significance of R. sanguineus-associated Rickettsia infections in dogs and humans [46]. These findings indicate the need for further studies regarding the presence of Rickettsia spp. in R. sanguineus from Cuba.

Sample Collection and DNA Extraction
The sample collection was performed between September 2016 and August 2017 in animal shelters housing dogs from ten municipalities of Havana City, Cuba (Figure 1). The climate of this region is tropical and humid with two marked climatic periods, a dry season from November to April with temperatures varying from 15 to 26 • C, and a wet season from May to October with temperatures typically range between 22 and 32 • C. The annual average temperature varies between 22 and 28 • C, and relative humidity of 80%. As is typical of most animal shelters in Cuba, the population included stray dogs and dogs abandoned by their owners for various reasons. Whole blood samples were collected from 100 randomly selected dogs of different breeds, sex and age. Samples were drawn from the jugular vein using sterile Vacutainer needles and K2EDTA-coated tubes (Becton-Dickinson Vacutainer Systems, Franklin Lakes, NJ, USA), and maintained at 4 • C until DNA extraction within 24 h of blood collection, which was performed using the Wizard Genomic DNA Purification Kit (Promega, Madison, WI, USA) according to the manufacturer's instructions. The DNA samples were eluted in 100 µL of DNA Rehydration Solution and stored at −20 • C until used as template for polymerase chain reaction (PCR) assays. An extraction control (DNA-free distilled water) was included for every 20 samples extracted. Sampled dogs were subjected to a thorough external physical exam looking for the presence of ticks, including their ears, heads, necks, chests, bellies, and paws. A representative sample of up to ten ticks was manually removed per infested dog using forceps and ensuring that the mouth parts remained intact. All collected specimens were deposited in labelled plastic tubes, covered by a piece of cloth, secured by rubber band, and transported alive to the laboratory for identification using a stereomicroscope (Carl Zeiss AG, Oberkochen, Germany) according to the standard taxonomic key described by Estrada-Peña et al. [47]. Once identified, the ticks were preserved in 70% ethanol (Merck®, Kenilworth, NJ, USA) using 1.5 mL plastic sterile.

PCR Amplification and Sequencing
To verify the presence of amplifiable DNA in the samples, a real-time qPCR assay for the canine housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was performed as previously described by Sieber-Ruckstuhl et al. [48]. All DNA samples were analysed by real-time qPCR using the primers and probe set previously described for A. phagocytophilum [49], A. platys [50], B. burgdorferi s.l. [51], E. canis [52] and Rickettsia spp. [42]. The PCR reactions included 500 nM of each primer, 250 nM probe, 0.2 µL of Uracil-DNA Glycosylase (UNG, Eurogentec S.A., Seraing, Belgium), 10 µL of the qPCR Mastermix (Eurogentec S.A., Seraing, Belgium) and 5 µL of DNA in a final volume of 20 µL. All real-time qPCR assays in this study were run on an ABI 7500 FAST Real-Time PCR System (Applied Biosystems; Thermo Fisher Scientific, Reinach, Switzerland) with an initial step of 2 min at 50 • C and a denaturation for 10 min at 95 • C, followed by 45 cycles of 15 s at 95 • C and 1 min at 60 • C. All primers and probes are listed in the Table 3. In addition, DNA sequencing was performed for molecular characterization on samples randomly selected among the DNA PCR-positive samples. Selected DNA samples were used as a template in conventional PCR assays with genus-and species-specific primers for Ehrlichia/Anaplasma spp. (16S rRNA gene) [53], A. platys (groEL gene) [39], E. canis (gltA gene) [38], and Rickettsia spp. (ompA, ompB and htrA genes) [54][55][56]. Each PCR reaction consisted of 10 µL of 5× Phusion HF buffer (Finnzymes, Espoo, Finland), 400 nM each primer, 200 nM each deoxynucleotide triphosphate (dNTP) (Sigma-Aldrich, Buchs, Switzerland), 1 U Phusion DNA Polymerase (Finnzymes, Espoo, Finland), 5 µL of DNA template, and nuclease-free water (Thermo Fisher, Darmstadt, Germany) in a final volume of 50 µL. The conventional PCR assays were run on a Biometra T-Personal 48 Thermocycler (Biometra, Gottingen, Germany). All PCR reactions were performed including negative, positive and extraction controls in each run. The cycling conditions and primers for sequence analysis are listed in Table 4. Amplified PCR products were electrophoresed in 1.5% agarose gels (100 V, 45 min), pre-stained with GelRed™ DNA Stain (Biotium, Hayward, CA, USA) and visualized under UV light. The molecular weight of the obtained products was determined using the GeneRuler™ 100 bp Plus DNA Ladder (Thermo Fisher Scientific, Darmstadt, Germany) as a molecular weight marker.

Sequence Analysis
PCR products were purified with the QIAquick Gel Extraction Kit (Qiagen, Hilden, Germany) following the manufacturer's instructions. The purified PCR products were cloned using a pCR 2.1 Invitrogen TOPO TA cloning kit (Thermo Fisher Scientific, Dreieich, Germany) followed by transformation into Escherichia coli Top 10F´competent cells according to the manufacturer's protocol. Plasmid DNA was extracted from the recombinant clones using a QIAprep Spin Miniprep kit (Qiagen, Hilden, Germany) and sent for sequencing in both directions with universal primers of M13 gene (M13f: 5' -GTA AAA CGA CGG CCAG-3'; M13r: 5' -CAG GAA ACA GCT ATG AC-3') to a commercial laboratory (Microsynth, Balgach, Switzerland). Obtained sequences were analysed using BLAST: Basic Local Alignment Search Tool (http://blast.ncbi.nlm.nih.gov/Blast.cgi) to determine the closest similarities to corresponding sequences of the reference strains reported in the GenBank database [57]. Theoretical translation of nucleotide sequences into amino acid sequences using the ExPASy translate tool, available on the ExPASy molecular biology server (http://www.expasy.org) [58], and the protein sequences were aligned using the ClustalW, included in the package BioEdit v.7.0.0 (Ibis Biosciences, Carlsbad, CA, USA).

Phylogenetic Analysis
The phylogenetic analysis was performed on the Molecular Evolutionary Genetics Analysis software package version 7.0 (MEGA7) [59], using the neighbor-joining method. Sequences were aligned using MAFFT configured for the highest accuracy and conserved regions identified [60]. After alignment, ambiguous regions (i.e., containing gaps and/or poorly aligned) were removed with Gblocks version 0.91b [61]. For the phylogenetic trees' construction, the best-fit model of the sequence evolution was selected based on Corrected Akaike Information Criterion (cAIC) and Bayesian Information Criterion (BIC) implemented in MEGA7. For E. canis-gltA and A. platys-groEL nucleotide sequences, the Tamura 3-parameter method, as well as for E. canis and A. platys 16S rRNA nucleotide sequences the Kimura 2-parameter, showed the lowest values of cAIC and BIC, and thus were chosen for corresponding tree reconstruction. Rates' variation across sites was fixed to "invariant and gamma distributed". A bootstrap analysis was performed to test the stability of the trees with 1000 replicates. GenBank accession numbers for the sequences used in the analyses are given in Figures 2-4.

Data Analysis
The obtained data were compiled and analysed with Excel 2016 software (Microsoft Corporation, WA, USA), and statistical analysis was performed using the R software (R_Development_Core_Team, 2018). The A. platys, E. canis, Rickettsia spp. and co-infections prevalence rates with 95% confidence intervals (CI) were calculated using a Bayesian approach based on Beta distribution, beta (s + 1; n-s + 1), where s = positives; n = tested animals. The following variables (i.e., HCT, RBC, MCV, Hb, MCH, MCHC, WBC, PLT, segmented neutrophil, lymphocyte monocyte, and eosinophil counts) were tested for statistical association with A. platys, E. canis, Rickettsia spp. and co-infections PCR detection using the PCR-negatives dogs as a control group. The evaluated variables were found not to be normally distributed by Shapiro-Wilks' W test and were analysed by the non-parametric Mann-Whitney U-test. Differences were regarded significant when p < 0.05.

Ethical Approval
Ethical approval of the present study was obtained from the Ethics Committee and Animal Welfare of Centro Nacional de Sanidad Agropecuaria (CENSA), Mayabeque, Cuba. The blood and tick sampling, as well as animal handling, was carried out by registered veterinarians. For the purposes of the study, no animal was sacrificed and the field study did not involve endangered or protected species, harm or cruelty to animals.

Conclusions
In conclusion, the results of this study highlight the high prevalence of vector-borne pathogens with zoonotic potential in apparently healthy shelter dogs in Cuba. The studied region predominantly comprised urban areas, which makes the zoonotic potential a particular concern for human health. The present study also represents the first report of R. felis in dogs from Cuba. The high canine vector-borne pathogens (CVBPs) infection prevalence observed indicates that, in the canine population of the studied region, tick-borne pathogens, such as E. canis, A. platys, R. felis and possibly other members of the SFGR are circulating, which are considered to be both zoonotic and pathogenic bacteria in dogs. In addition, the detection of CVBP infection was correlated with the occurrence of haematological changes and thus our findings suggest a possible long-term health impact of arthropod-borne pathogens on infected shelter dogs. The present study is important to raise a common awareness that stray dogs can serve as immediate proximal sentinels of CVBD-causing pathogens, representing a health threat that requires consideration by Cuban veterinarians and physicians. Based on our results and clinical observations, we encourage a surveillance campaign of CVBDs for monitoring and control, with special emphasis on the investigation in humans, animals and vectors, to obtain a wider epidemiological perspective focused on the One Health approach.
Supplementary Materials: The following are available online at http://www.mdpi.com/2076-0817/9/11/901/s1. Table S1: The correlation about tick presence and related hosts with real-time qPCR diagnostic results of vector-borne pathogens for sampled shelter dogs (n = 100) from Havana City, Cuba. Table S2: Complete blood counts (CBC) obtained for blood samples collected from shelter dogs (n = 90) living in Havana City, Cuba.