Molecular Identification of Selected Tick-Borne Protozoan and Bacterial Pathogens in Thoroughbred Racehorses in Cavite, Philippines

Tick-borne diseases (TBDs) considerably impair equine health and productivity. Moreover, TBDs, particularly equine piroplasmosis, impede international movement and trade of equids, which is a vital component of the global horse racing industry. In the Philippines, horse racing is a lucrative industry generating millions of USD annually. However, information on equine TBDs is scarce. This study intended to describe molecularly the equine tick-borne infections in a racehorse park in Cavite, Philippines and identify the risk factors associated with the infections. One hundred twenty-four (n = 124) thoroughbred racehorses were sampled and screened for selected tick-borne protozoan and bacterial pathogens using polymerase chain reaction (PCR) assays. Racehorses were positive for Babesia caballi (12.10%; 15/124), Theileria equi (0.81%; 1/124), Anaplasma phagocytophilum (10.48%; 13/124), Borrelia burgdorferi sensu lato (38.71%; 48/124), A. marginale (0.81%; 1/124), and Coxiella burnetii (0.81%; 1/124). Rickettsia was not detected in the samples. Gender was determined as a significant risk factor for B. caballi infection. Sequencing analysis revealed that seven partial 18S rRNA B. caballi isolates shared 98.63–100% identity with each other and were classified as genotype A. Meanwhile, the sequence obtained from the lone T. equi-positive sample was 99.77% identical to isolates from Spain, Switzerland, China, Saudi Arabia, and South Korea, and was confirmed as genotype E based on the 18S rRNA gene. Eight Anaplasma 16S rRNA partial sequences were highly identical to A. phagocytophilum and A. ovis. Partial sequences of Borrelia 5–23S rRNA were most closely related to B. japonica and other Borrelia sp. isolates from various countries. This study reports the first molecular detection of Borrelia and Anaplasma and the identification of B. caballi and T. equi genotypes in racehorses in the Philippines. Findings from this study shall be useful in crafting equine tick and TBD control and prevention programs in the country.


Sample Population and Demographics
We sampled a total of 124 (n = 124) horses raised in a race park in Cavite, Philippines ( Figure 1). All horses were thoroughbreds and were specifically raised for racing. The majority of the horses were female (56.45%), while the age of the horses ranged from 1 to 20 years (median = 5). The sample population was composed of 12.90% (16/124) yearlings, 17.74% (22/124) colt/filly, and 69.35% (86/124) stallion/mare (Table 1). During sample collection, ticks and other ectoparasites were not observed in the bodies of the horses. The caretakers reported that ectoparasite control practices included daily bathing of the horses and the use of ducks as biological means of controlling ectoparasites. Moreover, none of the sampled horses presented any clinical signs related to TBDs.

Identifying Risk Factors Associated with TBP Detection
Univariable analysis by Pearson's chi-squared or Fisher's exact test was done to assess the association of PCR-positivity with the different variables (Tables 1 and 3). Samples positive for T. equi, A. marginale, and C. burnetii were excluded in the analysis due to low detection rates. Female horses had significantly higher B. caballi infection (18.57%) compared to those of male horses (3.70%) (p = 0.013). Gender was negligible in A. phagocytophilum (p = 0.77) and Bbsl (p = 0.58) infections. On the other hand, B. caballi (25%) and A. phagocytophilum (18.75%) were more frequently detected in yearlings, while Bbsl detection rates were comparable among various age groups. Nevertheless, the differences among age groups were not significant, as shown by p values of 0.17, 0.44, and 0.96, respectively. Multivariable logistic regression analysis was performed for B. caballi, A. phagocytophilum, and Bbsl (Table 4). Variables that had p values ≤ 0.20 for B. caballi (gender: p = 0.013; age group: p = 0.17; Bbsl positivity: p = 0.09), A. phagocytophilum (Bbsl positivity: p = 0.13), and Bbsl (B. caballi positivity: p = 0.09; A. phagocytophilum positivity: p = 0.13) in the univariable analysis were included in the subsequent methods. Using the backward stepwise method, the B. caballi model, which included gender and Bbsl positivity as predictors, showed the lowest Akaike information criterion (AIC) value (87.54); thus, it was selected as the final model. Analysis indicated that gender was a significant risk factor for B. caballi infection (p = 0.026), with female horses 5.77 times more likely to test positive for B. caballi than male horses (Table 4). Meanwhile, the final models for the other TBPs identified Bbsl positivity, and B. caballi and A.phagocytophilum positivity as non-significant predictors for detecting A. phagocytophilum and Bbsl in horse samples, respectively (Table 4).

Sequence Identities and Phylogenetic Analysis
Randomly selected amplicons were sequenced to obtain representative sequences for B. caballi (n = 7), T. equi (n = 1), A. phagocytophilum (n = 6), and Bbsl (n = 17). Babesia caballi 18S rRNA isolates from the current study (MW714970-MW714976) shared 98.63-100% identity with each other and were 99.15-100% identical to a B. caballi isolate from a horse in China. Phylogenetic analysis indicated that the predominant genotype of the B. caballi population in racehorses in Cavite, Philippines was genotype A and these sequences were most closely related to a B. caballi isolate from China ( Figure 2). On the other hand, the sole T. equi sequence (MW714977) had the highest shared percent identity (99.77%) with T. equi 18S rRNA isolates from China, Kazakhstan, Russia, Mongolia, Saudi Arabia, South Korea, and Spain. The phylogeny indicates that MW714977 is classified as genotype E and formed a sister clade with dog and horse T. equi isolates from Saudi Arabia and South Korea, respectively ( Figure 3).   Anaplasma 16S rRNA sequences from eight distinct genotypes were obtained from six PCR-positive samples (MZ150516-MZ150518; MZ150520-MZ150524). MZ150516 was most closely related (99.03%) to a novel Anaplasma sp. (A. phagocytophilum-like) isolate from a sika deer in Japan (JN055357). Meanwhile, the remaining sequences (n = 7) shared 98.81-99.68% identity with each other and had >99% identities with A. ovis isolates from China, Mongolia, Russia, and Kenya. Phylogeny inference revealed the grouping of the Anaplasma sp. (A. ovis-like) lineages generated in the present study, deviating from the A. ovis clade. Additionally, MZ150523 and MZ150524 formed a well-supported subclade within the aforementioned group ( Figure 4). The alignment analysis indicated that the A.ovis-like variant MZ150523 had nine single nucleotide polymorphisms (SNPs) compared to A. ovis sequences from Russia (MW600403), China (KJ639880), and a horse isolate (MZ150522) from this study ( Figure S1). On the other hand, MZ150516 clustered with other A. phagocytophilum-like isolates from East Asian countries and clearly diverged from the main A. phagocytophilum-like clade, forming a phylogenetically distinct subclade ( Figure 4). Moreover, 21 SNPs and two deletions were found by the alignment analysis of MZ150516 with U02521, the Webster strain of A. phagocytophilum ( Figure S2). The lone A. marginale groEL sequence (MZ408312) was 100% identical to previous isolates from the Philippines, while several reamplification attempts of the C. burnetii-positive sample failed; hence, no sequence for C. burnetii was generated.

Discussion
The long-established horse racing industry is a huge income generator in the Philippines. Despite this, the equine TBD research in the Philippines remains scarce and has progressed only modestly in recent decades. In this study, we molecularly identified major agents of TBDs in the blood DNA samples of Philippine racehorses.
EP is one of the most critical diseases causing devastating economic impacts to the equine industry [8,9]. The current study demonstrates the prevalence of EP in Philippine racehorses. The detection rates for B. caballi (12.10%) and T. equi (0.81%) in this study reflect the results of the previous EP investigation in slaughter and racehorses from Batangas and Manila, wherein B. caballi was the predominant EP agent identified [30]. It should be noted that the sampling site of the study conducted in 2010 in Manila transferred the horse racing operations to the present sampling location in Cavite. Conversely, T. equi was more frequently detected by PCR in horses in Cebu and Bohol provinces [31]. An in-country variance of the frequency of EP infections has been documented in several studies and was attributed to the differences in extrinsic factors such as geographic features [32,33], microclimate [34], and management practices of the animals [35]. The horses positive for EP did not exhibit any clinical sign associated with EP, a hallmark of persistent infections [7]. This type of infection is common in countries where EP is prevalent, such as the Philippines [6]. Inapparent carriers of EP pose a transmission risk as ticks can efficiently transmit the parasites acquired from persistently infected horses to naïve ones [36]. Moreover, EP carrier horses used for sports, such as racing, may have suboptimal performance compared with their healthy counterparts despite the lack of evident clinical signs [9]. Nevertheless, EP-carrying horses should be treated to eliminate parasites as developing overt clinical EP can still occur, especially in fatigued and immunosuppressed horses [37].
Identifying EP risk factors is vital to better tailor fit the formulation of control measures for EP in a particular horse population. Gender was identified as an important risk factor for B. caballi infection in this study. A higher positivity (18.37%) and higher odds (5.77) for B. caballi infection was recorded for female horses compared to that of male horses, which contradicts the findings of Qablan et al. [38]. However, previous investigations which evaluated gender as a risk factor for EP have inconsistent and contrasting results [1,39].
Although not significantly different, we likewise observed an indirect relationship between B. caballi infection rates and age. This can be attributed to the more robust immunity in older animals that leads to the elimination of B. caballi parasite [34,40], a feature of B. caballi persistent infections.
Equine piroplasma genotyping based on the 18S rRNA gene has been extensively used to characterize parasite populations in the field [6]. This study's identification of T. equi genotype E provides invaluable information on the T. equi population circulating in Philippine horses. This genotype was implicated in prior fatal cases of equine theileriosis [41] and from apparently healthy horses from China [42] and South Korea [43]. As we only sequenced one isolate, further study on T. equi-infected horses should be conducted in various provinces to fully elucidate the different types of existing T. equi populations in the Philippines. On the other hand, B. caballi genotyping is relatively limited, with isolates only classified into three genotypes [6]. We confirmed genotype A as the predominant B. caballi in the tested horses in the present survey. Unlike T. equi, B. caballi 18S rRNA genotypes reveal inadequate insight to the parasite's geographical distribution. Thus, characterization using different gene markers shall be beneficial to fully characterize B. caballi populations in equines.
We herein report the first molecular identification of Anaplasma in horses in the Philippines. While Anaplasma is relatively more frequently reported in cattle and water buffaloes in the country [44], information about it in other Philippine livestock is lacking. For this survey, we initially aimed to detect A. phagocytophilum, the species responsible for EGA. Interestingly, sequencing the 16S rRNA gene confirmed the presence of novel variants of Anaplasma instead. The Anaplasma sequences generated in this study exhibited the same values for sequence similarity and percent identities (98.81-99.68% for A. ovis-like and 99.03% for A. phagocytophilum-like). The sequence similarity values fall within the cutoff values of 98.65% for species delineation for prokaryotes based on the 16S rRNA gene [45], indicating that these sequences are possibly new variants of A. ovis and A. phagocytophilum. This result reinforces the previously described high diversity of Anaplasma in the Philippines [46] and may indicate the expansion of Anaplasma species in various hosts.
The detection of A. ovis-like variants in horses is particularly noteworthy, as horses and goats are raised together inside the race park. The two animals were observed to interact very closely during the sampling. Cross-species transmission, with goats implicated as pathogen source, may be a plausible explanation for detecting A. ovis-like variants in horses in this study. Presently, the competency and host specificity of equine Anaplasma vectors in the Philippines has not been reported, and the status of Anaplasma infections in goats is not confirmed. Hence, verifying this premise will be helpful to confirm Anaplasma inter-host transmission and the clinical impact of A. ovis-like variant infection to its incidental hosts. On the other hand, an A. phagocytophilum-like variant was detected in this study. In a prior survey in Cebu province, the authors identified several A. phagocytophilum-like variants in cattle [47]. As A. phagocytophilum has a diverse host range [48], numerous strains and variants have evolved which considerably differ in their ability to cause diseases [49]. Notably, the phylogenetic placing of the identified A. phagocytophilum-like variant in a subclade with isolates from East Asia suggests a geographic-related grouping. Nonetheless, using a multi-locus characterization approach may provide a more comprehensive strain differentiation of A. phagocytophilum in horses [50].
The first molecular evidence of Borrelia in the Philippines is documented in the present study. The presence of Borrelia has never been confirmed in the Philippines before, and the detection of DNA fragments in horse samples may indicate possible non-clinical Bbsl infection. In particular, we detected the Bbsl genospecies B. japonica, which is vectored by I. ovatus ticks and thought to be a Japan-specific, non-pathogenic human Bbsl genospecies [19], and two Borrelia sp. genotypes closely related to a Borrelia isolate highly similar to B. garinii. Cases of horses acting as potential reservoir for Bbsl causing human disease have been documented in South Korea [51] and Belgium [52] before. Given its zoonotic nature, the detection of Bbsl in horses demands a more extensive inquiry into the disease transmission risk to humans who have close contact with the infected animals.
We detected one horse sample positive for C. burnetii. Recently, C. burnetii was molecularly detected in cattle, water buffaloes, and Rhipicephalus microplus ticks from nearby provinces of Rizal and Quezon [53] while sera from human and ruminants collected from Laguna, Northern Samar, and General Santos City were seropositive for C. burnetii [54]. On the contrary, Rickettsia was not detected in horse samples in the present study. Molecular studies of Rickettsia in the Philippines are limited. Rickettsia was recently identified in dogs in Luzon [55], but was not detected in ruminants and ticks [53]. However, non-detection of Rickettsia could not rule out probable infections as detecting this bacterium using PCR assays in whole blood DNA can be less sensitive than detection in DNA extracted from cutaneous and serum samples [56].
Despite the information on equine TBPs uncovered in the current study, the competent tick vectors of the detected TBPs remain unidentified. The tick species transmitting B. caballi, T. equi, and A. ovis-like variant have not been reported in Cavite province, but potential tick vectors, such as Rhipicephalus microplus and Rhipicephalus sanguineus, were previously identified from various animal hosts in nearby provinces [55,57,58]. In addition, the detection of an A. phagocytophilum-like variant and Bbsl in the current study should compel further investigation on the competent tick vectors of these potentially zoonotic pathogens. Ticks belonging to the Ixodes genus are the only confirmed competent vectors of A. phagocytophilum. Still, previous studies suggest an alternative epidemiological cycle in Rhipicephalus ticks [59], which may explain our results. Likewise, Ixodes ticks are the only confirmed competent tick vectors of Borrelia [15]. Detecting these Ixodes-transmitted pathogens is not unlikely, and the presence of Ixodes ticks in Cavite is highly probable, as a tick survey conducted 50 years ago identified I. granulatus ticks in wild rodents [60].
In addition, several matters are beyond the scope of our study. We recommend employing direct methods such as immunofluorescence microscopy and isolation of the organism, and serological confirmation of infection by antibody detection to address the various sensitivity issues of PCR assays for future studies. In particular, the relatively high detection rate for Borrelia in the current study may be complemented by serological confirmation, the more common and accepted diagnostic test for Borrelia infection [3,18].
Altogether, the information gathered from this molecular survey revealed the TBPs harbored by thoroughbred racehorses and the horses' potential role as reservoirs of TBPs. Moreover, this study describes data on EP genotypes, novel variants of Anaplasma, and molecular evidence of Borrelia in Philippine racehorses for the first time.

Ethical Statements
The methodologies and procedures performed in this study were approved by Obihiro University of Agriculture and Veterinary Medicine, Obihiro, Hokkaido, Japan (animal experiment permit no. 20-128; DNA experiment permit no. 1723-4). Animal handling practices adhered to the Philippine Animal Welfare Act (Republic Act 8485 as amended by R.A. 10631) and were approved by the Institutional Animal Care and Use Committee (IACUC) of Cavite State University, Indang, Cavite, Philippines (approval no. 2019-001). The caretakers of the horses and the race park manager were informed about the background and aims of the study and gave their consent before sampling was conducted under the supervision of the race park veterinarian.

Location of Sampling and Animals
Cavite is a province situated southwest of Manila and located in the southern part of Luzon island, Philippines. The province is characterized by various topographies ranging from coastal plains to upland mountainous areas. Most areas in Cavite is suited for agricultural production, which is supported by its climate type composed of two pronounced seasons: the dry season from November to April and the rainy season from May to October [61]. Two of the three major racehorse parks in the Philippines are located in Cavite province. For this study, a total of 124 racehorses were randomly selected as samples from a race park in Naic, Cavite. Name, gender, age, breed, clinical signs, and history of tick infestation of each animal were recorded through the information provided by the caretakers. Briefly, about 3 ml of whole blood was drawn from the jugular vein and collected in Vacutainer ® EDTA tubes (BD, Franklin Lakes, NJ, USA). The blood samples were kept cool in an ice box during transportation to the laboratory. All samples were processed within the same day of the sampling.

Extraction of DNA
DNA was extracted using the QIAmp DNA Blood Mini Kit (Qiagen, Hilden, Germany), following the manufacturer's protocol. About 200 µL of whole blood was used and DNA was eluted in 100 µL elution buffer. DNA concentration was checked using NanoDrop ™ 2000 Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). Samples were stored at −20 • C until the molecular screening.

PCR Assays for the Identification of Tick-Borne Pathogens
Samples were screened using PCR assays specific to selected TBPs. A summarized list of information on the assays used in this study is shown in Table 5. A multiplex nested PCR assay which amplifies the partial 18S rRNA gene of B. caballi (540 bp) and T. equi/T. haneyi (392 bp) was used for detecting equine piroplasma [62]. Meanwhile, PCR assays for the detection of tick-borne bacteria A. phagocytophilum [63], Rickettsia spp. [64], and C. burnetii [65] amplified the partial 16S rRNA gene with final amplicon sizes of 928, 426, and 1450 bp, respectively. The rrf-rrl (5-23S rRNA intergenic spacer)-target primer set was used for the detection of Bbsl which amplified bands of 226-266 bp [66]. For the nested PCR assays, the first assay reactions were set to a volume of 10 µL with a final concentration of 2 mM of each dNTP (New England Biolabs, Ipswich, MA, USA), 1× ThermoPol ® buffer (New England Biolabs), 2 µM of forward and reverse primers, 0.25 U of Taq DNA polymerase (New England Biolabs). The second or single assays were performed to a final volume of 25 µL with the same final concentration of reagents except for the final concentrations of primers in the multiplex EP assay: forward primer (4 µM) and reverse primers (2 µM for B. caballi and 2 µM for T. equi/T. haneyi). About 2 µL of genomic DNA was used for direct amplification from the samples, while 1.5 µL of the first reaction product was used as a template for the multiplex or nested reactions. Thermocycling conditions were performed for each pathogen as before, except for the extension and final extension temperatures, which were set at 68 • C. In all assays, double-distilled water was used as negative controls and confirmed positive DNA samples from previous studies conducted in our laboratory were used as positive controls (Table 5). PCR products were resolved in either 1% or 1.5% agarose gel by electrophoresis, stained with ethidium bromide, and visualized under UV light.

Sequencing and Phylogenetic Analyses
Amplicons were excised from the gel and purified using NucleoSpin ® Gel and PCR Clean-up (Macherey Nagel, Düren, Germany). Samples with high concentration were directly sequenced; otherwise, samples with low concentration were ligated to pGEM ® -T Easy Vector (Promega Corporation, Madison, WI, USA) and inserted in DH5α Escherichia coli calcium competent cells. Plasmids were purified using NucleoSpin ® Plasmid Quick-Pure Kit (Macherey Nagel). Purified products were sequenced using BigDye ™ Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA) and ABI Prism 3100 Genetic Analyzer (Applied Biosystems).

Statistical Analysis
The data were comprised of animal parameters (age and gender) as independent variables and positivity for B. caballi, A. phagocytophilum, or Bbsl as dependent variables. For coinfections, positivity for the first and second TBP was considered independent and dependent variables. Meanwhile, data for other detected TBPs were excluded in the case of low detection rates. Age was classified into three categories: yearling (≤2 years), colt/filly (<4 years), or stallion/mare (≥4 years), while gender was categorized as either male or female. Univariable analysis by Pearson's chi-squared or Fisher's exact test was performed to assess the association of PCR-positivity with the parameters (age group and gender) and positivity to another pathogen. Consequently, multivariable logistic regression was performed for parameters with p value of ≤0.20. Initially, variable selection using the backward stepwise elimination was performed. Then, the models with the lowest AIC values were chosen. A p value of <0.05 was considered significant. All analyses were done using the Stats [69] and Epitools [70] packages in R.

Conclusions
In this survey, B. caballi, T. equi, A. ovis-like, A. phagocytophilum-like, A. marginale, Borrelia burgdorferi sensu lato genospecies, and C. burnetii were molecularly detected in thoroughbred racehorses in Cavite, Philippines. In addition, gender was determined as a significant risk factor for B. caballi infection. Furthermore, we performed the first genotyping of EP agents in Philippine horses. These findings provide crucial information on the TBD agents and call for establishing and implementing strategic treatment and control programs for the neglected tick-borne infections in Philippine horses.