First report on molecular characteristics and risk factor analysis of Ehrlichia canis in dogs in Khon Kaen, Thailand

Abstract Background and Aim: Ehrlichia canis is a well-known cause of both anemia and thrombocytopenia in dogs. There are insufficient epidemiological data on this blood parasite in Thailand and the association of infections with hematological abnormalities. This study aimed to analyze the molecular characteristics and to identify E. canis as well as the risk factors associated with E. canis infection in dogs in Khon Kaen, Thailand. Materials and Methods: Blood samples from 126 dogs that visited animal clinics were subjected to molecular detection using nested polymerase chain reaction for E. canis 16S rRNA gene. The risk factors and hematological profiles associated with the infection were analyzed using the logistic regression test in program SPSS version 19. Results: Forty-one dogs were infected, indicating a 32.5% molecular infection rate of E. canis. The factors significantly associated with E. canis infection include animal housing status, low packed cell volume, low red blood cell count, and low platelets (p<0.05). Ten positive samples were amplified, sequenced, and phylogenetically analyzed. Sequence and phylogenetic analysis confirmed the current ten samples as E. canis compared with reference sequences in GenBank, using the BLAST program hosted by NCBI, which showed 99.74-100% similarity. Conclusion: This study provided the first data of infection rate of E. canis using nested PCR and molecular characteristics of E. canis in randomly selected domestic dogs in Khon Kaen, Thailand.

Khon Kaen is a city located northeast of Thailand. Although a couple of reports have shown E. canis infection, a study was conducted in a smallscale sub-district area using the conventional polymerase chain reaction (PCR) technique [14]. Another study used multiplex PCR to detect E. canis in leftover blood samples negative for E. canis microscopically in a clinic [15]. This study covered private animal clinics on a large scale in the city municipality area and the Veterinary Teaching Hospital, Khon Kaen University (KKU), located in the Muang district, Khon Kaen. Nested PCR was conducted rather than multiplex or conventional PCR. Furthermore, this parasite's molecular characteristics and phylogenetic analysis have never been described in the northeast of Thailand.
Consequently, this study aimed to investigate the molecular characteristics of E. canis infection in the city municipality area of Khon Kaen and to assess the association between infection and various factors, including the age of the dogs, rearing status, tick infestation status, packed cell volume (PCV) level, red blood cell (RBC) count, white blood cell (WBC) count, and platelet count. It is hoped that the results will assist in controlling and preventing E. canis infection in dogs and facilitate a decline in disease transmission to humans.

Ethical approval
The study was reviewed and approved by the Institutional Animal Care and Use Committee of KKU, No IACUC-KKU-7/64. The written consent from owners of animals participating in the research was obtained.

Study period, area, and sample collection
From June to December 2020, 126 blood samples were randomly collected from dogs visiting seven private animal clinics and the KKU Veterinary Teaching Hospital in the city municipality area, Muang district, Khon Kaen, Northeast Thailand ( Figure-1). The sample size calculation was based on 36.73% prevalence of E.canis infection from the previous study in Buriram, the province located near Khon Kaen [16], with a 10% error. Dogs with or without any clinical signs were randomly selected in this study. Three milliliters of blood samples were collected from the cephalic or saphenous vein, filled in EDTA-anticoagulant (BD Vacutainer®, New Jersey, USA) tubes, and kept in an icebox (4-8°C) before being sent to the laboratory within eight hours at the Faculty of Veterinary Medicine, KKU, Khon Kaen, for further investigation.

Risk factor and hematological analysis
Some information, including dog age (≤1 year and >1 year), the duration of tick infestation (no ticks, <1 month, 1-6 months, and >6 months), and animal housing status (indoor, confined to grounds, and free-roaming), were collected by interviewing the pet owners.

Molecular detection
DNA was extracted from blood samples using the GF-1 blood DNA kit (Vivantis, Selangor Ehsan, Malaysia), according to the manufacturer's instructions. The nested PCR protocol followed the previous study in amplifying the 16S rRNA gene. The molecular-grade water was used as the negative control. The first reaction using primers ECC (5' AGAACGAACGCTGGCGGCAAGCC 3') and ECB (5' CGTATTACCGCGGCTGCTGGCA 3') and the second reaction using primers CANIS (5' CAATTATTTATAGCCTCTGGCTATAGGA 3') and HE3 (5'-TATAGGTACCGTCATTATCTTCCC TAT 3'), yielded 478 and 389 bp amplicons, respectively [18]. The PCR products were sequenced using primer CANIS and HE3 (BTSeq™ Barcode-Tagged Sequencing; Celemics, Seoul, South Korea). The current sequences were compared with the sequences in GenBank, using the BLAST program hosted by NCBI, National Institutes of Health, USA (http://www.ncbi. nlm.nih.gov) to evaluate the similarity before being aligned using BioEdit version 7 software (developed by Tom Hall, North Carolina State University) [19]. The phylogenetic tree was constructed using the neighbor-joining method with 500 replicates for bootstrap analysis, MEGA version 7 software (Molecular Evolutionary Genetics Analysis software available at https://www.megasoftware.net/) [20].

Statistical analysis
The risk factors, RBC, WBC, platelets, and PCV, were analyzed in relation to the infection. The relationship between risk factors and prevalence was analyzed using the logistic regression test with p<0.05 in SPSS software v 19.0 (IBM Corp., NY, USA).
According to the statistical analysis of risk factors, there was an association between animal housing status and E. canis infection (p<0.01), while age and tick infestation was not associated (p>0.05). The free-roaming dogs had a higher E. canis infection rate of 53.3% (24/45) than the other groups (Table-1).
The phylogenetic tree was constructed using the neighbor-joining method in the MEGA (version7) program. Ehrlichia spp., Ehrlichia ewingii, Ehrlichia chaffeensis, and Anaplasma platys sequences from the GenBank were compared as the out-group. The results revealed that the samples were closely related to E. canis from GenBank accession no. MH620196, EU781689, MG029068, KF972450, and JX861392 ( Figure-4).

This study observed the infection rate and molecular characteristics of E. canis in dogs in Khon
Kaen. The association between E. canis infection and risk factors, including the levels of PCV, WBC, RBC, and platelet count was evaluated. The infection rate (32.5%) was similar to the previous reports in some other countries. For example, E. canis prevalence was found to be 21.8% in Cambodia [21], 28% in Pakistan [22], and 47.1% in Malaysia [23]. In Colombia, the prevalence was 40.6% [24], whereas E. canis infection prevalence was as low as 6.3% in Northeast Algeria [25]. Different temperatures in different countries influence the tick population, which may cause several E. canis infection prevalence.
In Thailand, several studies have reported the prevalence of E. canis infection from different provinces, including 21.5% in Mahasarakham [26], 25% in Kalasin [17], 36.73% in Buriram [16], and 65.12% in Bangkok and Samutraprakarn [27]. In Khon Kaen, where this research was conducted, other studies were previously reported using different methodologies  [14,15]. For example, one study used multiplex PCR to detect infection in left-over blood samples from animal hospitals, which were primarily negative following microscopic examination and found 33.75% prevalence of E. canis infection [15]. However, this study collected random blood samples from hospitalized dogs. Contrastly, another study that collected samples from Phra Lap Sub-district, Muang District, Khon Kaen, used conventional PCR to examine E. canis and showed a low infection prevalence of 3% [14]. CME occurrence in dogs in Khon Kaen Province was shown to be associated with animal housing status (p=0.001). Dogs living in indoor areas were at less risk of E. canis infection, probably related to low tick infestation [28]. Another report suggested that dogs living in outdoor areas were at higher risk of E. canis infection (Odds Ratio 2.3) than dogs living indoors [29]. In this study, tick infestation was not associated with E. canis infection (p=0.196). The result agreed with a previous study showing that ticks were not considered a risk factor for E. canis positivity [20]. However, a previous study in Mexico showed 72% prevalence of E. canis infection in dogs infested with tick and 54% in dogs without tick infestation [30].
Furthermore, E. canis infection occurs mainly during the warm season compared to the rainy and winter [31]. The differences in climatic conditions have been found to be important factors influencing the population dynamics of ticks in a particular region, resulting in a variable prevalence pattern of CME [32]. The dog's age was not associated with E. canis infection (p=0.488), and this was also supported by the study of E. canis infected dogs in three districts in Punjab of Pakistan [22]. However, one previous report found that dogs younger than 6 months had more E. canis infection than older dogs [31]. However, the gender of animals was not analyzed in the current research since several studies have revealed that it was not associated with the infection [22,33,34].
PCV values showed a significant association with E. canis infection (p=0.001). This result agrees with the previous reports [9,17,18]. The current results show that PCV values of <20% were associated with more E. canis infection (93.8%). This study had similar findings to a previous report [35]. In addition, platelet count was associated with E. canis infection (p=0.002), as supported by the previous reports [36,37]. This study showed that dogs with platelet counts <100,000 platelets/µL had a higher infection rate (69.2%) than other groups (Table-2). Moreover, RBC count was associated with E. canis infection (p=0.001) [9]. Similar to this previous study, the present study showed that anemic dogs had an increased E. canis infection rate (81.2%) [38]. In addition, WBC count was not associated with E. canis infection (p=0.757), and this is in agreement with a previous report [36].
According to the BLAST investigation, the partial sequences of the 16S rRNA gene of E. canis in dogs in Khon Kaen Province were more than 99% identical to genotypes in GenBank isolated from Texas, Tunisia, Mexico, India, and Brazil. The phylogenetic tree and sequence alignments showed low diversity in E. canis strains, similar to the previous reports [39,40]. However, it would be interesting to examine longer 16S rRNA sequences of E. canis strains to detect genetic diversity in Thai dogs. Furthermore, no heterogeneity was observed among the E. canis groups using the 16S rRNA gene. However, the distinct sequence was found to have a 99% similarity to the Ehrlichia sequence reported from China [41]. A previous study from Bangkok showed that the E. canis strains were linked with multiple connected branches, and little genetic diversity was observed, suggesting slow and homogeneous evolution [42]. This study supports this concept as E. canis strains were also joined with multiple connecting branches. The phylogenetic tree and sequence alignments showed low diversity in E. canis strains [43]. The extremely low polymorphism exhibited among the E. canis strains [40], suggests a common origin with no distant divergence in the dendrogram showing the genetic relationship between the 16S rRNA sequences of different Ehrlichia spp. isolates.

Conclusion
This study provided the first data on the infection rate of E. canis using nested PCR and molecular characteristics of E. canis in domestic dogs in Muang district, Khon Kaen Province, Thailand. The phylogenetic trees display a high similarity of the current E. canis and other sequences in GenBank. Conclusively, 32.5% of dogs were infected with E. canis. Dogs living outdoor (p<0.05) were associated with infection, whereas age and tick infestation was not related. Furthermore, low PCV levels, low platelets, and low RBC were associated with E. canis infection. This survey provided preliminary information on the prevalence and factors associated with CME, which can be helpful in developing control and prevention program. Animal housing indoors or confined areas should be concerned, and pet owners should consider controlling tick infestation.

Authors' Contributions
WT: Conceptualization, formal analysis, project administration, and drafted and revised the manuscript. TM and WT: Methodology, data curation, visualization. TM, BRS, AM, and PB: Investigation. TM: Software and the first draft of the manuscript. SS: Supervision. WT and SS: Validation. All authors read and approved the final manuscript.