The prevalence of Anaplasma platys and a potential novel Anaplasma species exceed that of Ehrlichia canis in asymptomatic dogs and Rhipicephalus sanguineus in Taiwan.

Canine anaplasmosis is regarded as an infection by Anaplasma platys rather than zoonotic Anaplasma phagocytophilum in subtropical areas based on the assumption that the common dog tick species is Rhipicephalus sanguineus, which transmits E. canis and presumably A. platys. We investigated asymptomatic dogs and dog ticks from 16 communities in Nantou County, Taiwan to identify common dog tick species and to determine the prevalence of Anaplasma and Ehrlichia spp. Of total 175 canine blood samples and 315 ticks, including 306 R. sanguineus and 9 Haemaphysalis hystricis, 15 dogs and 3 R. sanguineus ticks were positive for E. canis, while 47 dogs and 71 R. sanguineus ticks were positive for A. platys, via nested PCR for 16S rDNA and DNA sequencing of selected positive amplicons. However, among the dogs and ticks that were positive to A. platys 16S rDNA, only 20 dogs and 11 ticks were positive to nested PCR for A. platys groEL gene. These results revealed the importance of searching for novel Anaplasma spp. closely related to A. platys in dogs and ticks. Seropositivity to a commercial immunochromatographic test SNAP 4Dx Anaplasma sp. was not significantly associated with PCR positivity for A. platys but with infestation by ticks carrying A. platys (P<0.05). Accordingly, R. sanguineus may be involved in transmission of A. platys but may not act as a reservoir of E. canis and PCR results for 16S rDNA could be a problematic diagnostic index for A. platys infection.

Bacteria in the genus Anaplasma is a tick-borne rickettsial organism characterized by a distinct cellular tropism and host range and causes various diseases among animals [6]. The Anaplasma species known to infect dogs are A. platys and A. phagocytophilum, which are transmitted by different tick species. A. phagocytophilum, the causative agent of human granulocytic anaplasmosis, is a zoonotic pathogen that causes pathological conditions, represented by polyarthritis, cytoplasmic inclusions in neutrophils and various hematological abnormalities as well as fever, anorexia, lethargy, and lameness, in a wide range of mammals, including dogs, horses, ruminants, and humans. In contrast, A. platys is a rare organism that exclusively infects platelets and periodically causes profound thrombocytopenia only in dogs [1]. The clinical signs of A. platys infection are mild or often unrecognizable, despite the severity of thrombocytopenia; however, more severe morbidity has been reported in some countries [4, 14]. In addition, because R. sanguineus also transmits Ehrlichia canis, A. platys infections often occur concomitantly with monocytic ehrlichiosis caused by E. canis, resulting in severe clinical manifestations in dogs. Compared with A. phagocytophilum and infections involving this organism, which have been vigorously studied worldwide as an important emerging disease, A. platys and the infections it causes have been studied only in endemic regions. Despite the difference between these two Anaplasma spp., the serological diagnosis of A. platys infections in dogs has relied on a commercial ELISA kit utilizing a cross-reaction by antibodies against A. phagocytophilum and A. platys [10, 35], partly due to the unavailability of isolated A. platys culture strains [20]. However, the use of this kit against these two species may limit the ability to reveal the true presence of zoonotic anaplasmosis in areas where R. sanguineus is common [29].

A common vector for A. phagocytophilum is the Ixodes ricinus species complex, whereas the vector for A. platys is still unclear, although R. sanguineus has been suspected [1]. The geographic distribution of these tick species coincides with the distribution of either granulocytic or thrombocytic anaplasmosis in the region. Ixodes ricinus is distributed widely in Europe, causing tick-borne fever and related complications in ruminants and horses by transmitting A. phagocytophilum. I. scapularis plays a pivotal role in the spread of A. phagocytophium infections in domestic animals and humans from the upper Midwest to the northeastern states of the U.S.A. and Canada [19]. While the Ixodes ricinus species complex prefers cold, moist weather, R. sanguineus was originally a tropical species. Accordingly, the prevalence of A. platys infections has been reported more frequently in regions at lower latitudes, such as the state of Arizona in the U.S.A., Venezuela, Brazil, and Thailand [10, 16, 28, 31]. However, as R. sanguineus is highly adaptable to the environment, it is distributed worldwide [7], and the distribution of A. platys has reached northern temperate regions [17]. Moreover, the distribution of I. scapularis has shifted southward because of recent climate change [11]. To predict the risk of transmission of Anaplasma spp., such as A. phagocytophilum or A. platys, understanding the regional climate and the most likely tick species involved is of major importance. Information regarding how ticks and infectious organisms are distributed in a region aids in ascertaining the likelihood of exposure to these agents and will improve the diagnostic accuracy of these tick-borne diseases in domestic animals and humans.

In subtropical Taiwan, R. sanguineus has been recognized as the most common dog tick species [33]; therefore, A. platys and E. canis infections are of veterinary importance [5, 15]. However, the epidemiology of A. platys infections in dogs and ticks has not been studied. Moreover, one study identified rodents carrying A. phagocytophilum in Taichung city, which is located adjacent to Nantou County [23]. A previous study by our group revealed that the seroprevalence of canine anaplasmosis and ehrlichiosis differs across communities, ranging from virtually absent to more than 60%, despite the fact that local dogs are equally heavily- infested by ticks [38]. We hypothesized that this difference in local seroprevalence may be due to the differences in common dog tick species and the frequency of carriers present in each community. Hence, a cross-sectional, community-based study was conducted in Nantou County, Taiwan. The objectives of this study were to confirm the identity of the common tick species in local dogs and determine the prevalence of Anaplasma/Ehrlichia spp. known to infect dogs and ticks, in addition to comparing results between communities.

MATERIALS AND METHODS

Study area and population

Whole blood samples and infesting ticks from apparently healthy dogs were collected at the time of the annual rabies vaccination campaign in urban and rural communities of Nantou County (23°58’N 120°58’E) in central Taiwan in the spring and fall of 2010 and 2011 [38] (Fig. 1). Prior to sample collection, the health condition of each dog was evaluated by a veterinarian and sample collection was conducted to the dogs whose owners gave consent to test their dogs. More than 80% of the county is covered with high mountains, and most of the studied communities are scattered in wooded mountainous areas. In Fig. 1, the studied communities are indicated by a circle simultaneously indicating the degree of tick infestation and the local seroprevalence of E. canis and Anaplasma sp. infection, as determined via the SNAP 4Dx test (IDEXX Laboratories, Westbrook, ME, U.S.A.). The seroprevalence of each community was compared with that of the entire study population, and each community was assigned to one of two groups, based on showing either a higher seroprevalence (high seroprevalence) or lower seroprevalence (low seroprevalence) than the study population: 21.1% for Anaplasma sp. and 11.4% for E. canis (Table 1). A total of 315 ticks and 175 dog blood samples were collected from 16 communities for this study. Among the 315 ticks, 291 were removed from 54 dogs that had provided blood samples (1 to 35 ticks per dog), and 24 ticks were collected without a blood sample from the host dog. After removal from the dogs, the ticks were preserved in tubes containing 70% ethanol and stored at room temperature prior to further analysis. Using a 23G sterile needle syringe, a maximum of 4 ml of peripheral blood was drawn from each dog into a sterile EDTA-containing vacutainer and maintained at 4°C during shipping to the laboratory for DNA extraction.

Fig. 1.

Location of the study area, Nantou County, Taiwan. The 16 studied communities are represented by a circle simultaneously indicating both the degree of tick infestation (the bigger, the higher) and the local seroprevalence of E. canis and Anaplasma sp. infection determined via SNAP 4Dx (White: low seroprevalence. Black: high seroprevalence*). *low: the local seroprevalence of E. canis and/or Anaplasma sp. was lower than the average for the entire population. high: the local seroprevalence of E. canis and/or Anaplasma sp. was higher than the average for the entire population

Table 1.

PCR results for dogs and ticks

Community ID	Number of sampled dogs and ticks		Number of dogs positive to SNAP 4Dx		Number of samples positive to PCR
	Dogs	R. sanguineus	Anaplasma sp.	Ehrlichia canis	Anaplasma platys (16SrDNA and groEL)		Ehrlichia canis		Anaplasma platys 16SrDNA only
					Dog	Tick	Dog	Tick	Dog	Tick
1	2	2	1	1	0	0	1	0	0	0
2	10	14	0	0	0	0	0	0	6	0
3	19	8	1	1	0	0	2	0	0	3 a)
4	8	4	0	0	0	0	0	0	2	0
5	0	6	NA	NA	NA	0	NA	0	NA	0
6	11	17	4	2	2	0	1	0	0	6
7	17	19	5	6	1	0	2	0	3	1
8	9	6	3	0	1	0	0	0	1 a)	0
9	11	26	0	0	6 a)	1	0	0	0	17 a)
10	7	1	2	1	0	0	0	0	0	0
11	26	56	1	0	1 a)	2 a)	1	0	7	4
12	14	29	9	4	3	1	2 a)	0	0	3
13	19	35	7	3	4 a)	2	4	2 a)	0	6
14	5	71	1	1	2	4 a)	1	1	0	18
15	14	9	2	1	0	0	1	0	8	2
16	3	3	1	0	0	1 a)	0	0	0	0
Total number	175	306 (282 b))	37 (21.1%)	20 (11.4%)	20 (11.4%)	11 (3.6%)	15 (8.6%)	3 (1%)	27 (15.4%)	60 (19.6%)

NA: not available. a) The PCR products were sequenced. b) The number of ticks whose host dogs also provided blood samples and demographic data.

Hematological tests and DNA extraction

After arrival at the laboratory, a complete blood count and serological testing using SNAP 4Dx were performed. A 200 µl aliquot of whole blood that had been separated into a sterile tube and stored at −70°C prior to DNA extraction was analyzed with the DNeasy Blood and Tissue kit (Qiagen, Valencia, CA, U.S.A.) following the manufacturer’s instructions.

Identification and DNA extraction from ticks

The ticks were morphologically identified under a light microscope using taxonomic keys from Walker et al. [34] and Teng and Jiang [32]. Each tick was treated as a single specimen. The tick was cut along the dorsomedial plane into bilaterally symmetrical right and left halves using a new sterile surgical blade. The left half of the body was employed for DNA extraction with the DNeasy Blood and Tissue kit (Qiagen) according to the manufacturer’s recommendation, and the remaining half was stored in 70% alcohol. The body halves of fully engorged females were further cut into 2 to 3 pieces, and DNA was extracted from each piece. Nymphs were cut into halves, and both halves were used for DNA extraction.

Screening via polymerase chain reaction (PCR) and gene sequencing

The DNA samples from dog blood and ticks were screened via PCR for Anaplasma and Ehrlichia species known to infect dogs. For the detection of E. canis, E. chaffeensis, E. ewingii, and A. platys, nested PCR for 16S rRNA genes was performed, whereas for A. phagocytophilum, the msp2/p44 genes were amplified. Briefly, in a final volume of 25 µl, 5 µl of template DNA was first amplified using the outer primers ECC and ECB (universal primers for Anaplasma and Ehrlichia species) [9], in a reaction involving 40 cycles of denaturation at 94°C for 1 min, annealing at 60°C for 1 min, and extension at 72°C for 1 min. For the nested amplification, we employed the specific primer pairs EPLAT 5 and EPLAT 3 [24] and CANIS and HE3 [2, 13] for A. platys and E. canis, respectively, and 2 µl of the primary PCR products was used as a template in a total volume of 25 µl (0.2 mM dNTPs, 2.5 U Taq DNA polymerase, 2.5X reaction buffer, and 10 pmol of each primer) employing an automated thermal cycler (Bio-Rad Laboratories, Hercules, CA, U.S.A.) for 35 cycles. For the detection of Ehrlichia chaffeensis and Ehrlichia ewingii, 2 µl of the same primary PCR products was used as a template to perform nested amplification of 16S rRNA genes with specific primer pairs for each species (HE1 and HE3 [9] or EE72-ewingii and HE3 [8], respectively). To screen for A. phagocytophilum, PCR targeting the p44/msp2 gene was performed using MSP3F and MSP3R [21, 29, 39] and msp2 full-length F and R [21], with inner amplification using a newly designed primer pair (MSP2 In-F 5′-TATGGACTATCCGTGAGCAG-3′ and MSP2 In-R 5′-CCAAGTTTGAGCTTGTATGAAAG-3′). A negative control containing distilled water instead of a DNA template in the PCR mixture and positive controls containing previously sequenced and verified A. platys, A. phagocytophilum, E. canis, and E. chaffeensis DNA were included in each PCR run. The PCR products were electrophoresed on 2% agarose gels to confirm the expected size of the amplified nucleotide fragments, i.e., 385 bp for A. platys, 365 bp for E. canis, 389 bp for E. chaffeensis, 407 bp for E. ewingii, and 334 bp and 1,100 bp (745 bp for nested PCR) for A. phagocytophilum. To corroborate the PCR results, randomly selected PCR products showing a strong positive reaction were purified with the QIA quick PCR purification kit (QIAGEN GmbH, Hilden, Germany), and DNA sequencing was performed directly on the purified PCR products using the fluorescence-labeled dideoxynucleotide technology of the ABI PRISM 3730 automated DNA sequencer.

PCR targeting the A. platys groEL gene

PCR targeting the A. platys groEL gene was conducted for further evaluation of the A. platys strain using an A. platys-specific primer pair for the groEL gene (GroAplatys-35s and GroAplatys-550as) [3]. Hemi-nested PCR for the A. platys groEL gene was also conducted using a newly designed primer (PlatGro-kc-as 5′-CCATCTGTGCTTTGATTTGG-3′). This PCR primer was derived from conserved regions on the basis of a multiple alignment of A. platys groEL sequences obtained from GenBank. Additional primers were designed to amplify both A. platys and A. phagocytophilum strains (ApGro-1s 5′-TAGTGATGAAGGAGAGTGAC-3′ and ApGro-284as 5′-TTCATTACCTTGTAGCCATCC-3′; 285 bp). These primer sequences were derived from conserved regions based on a multiple alignment of A. platys and A. phagocyophilum groEL genes. A diagram showing the primer locations and orientations was presented in Fig. 2. For each reaction, 2 µl of a DNA extract was used for PCR amplification in a total volume of 25 µl, containing 10 pmol of each primer. The reaction was visualized in a 2% agarose gel, and the positive PCR products (516 bp for inner amplicons and 1121 bp for outer amplicons) extracted from the gel were sequenced as described above.

Fig. 2.

Locations and orientations of the primers with respect to A. platys groEL gene. The numbers in each parenthesis represent the primer position relative to the nucleotide sequence of the A. platys groEL gene (GenBank accession number AF478129). *Reverse orientation. **The primer sequences of ApGro-1s and ApGro-284as were derived from the sites homology to multiple nucleotide sequences of A. phagocyophilum groEL genes (AF172158, EE473209, AF482760, KU519287).

Data analysis

A statistical analysis was performed with Systat 13® software (Systat Software, Inc., Chicago, IL, U.S.A.). Cohen’s Kappa coefficient was used to evaluate the agreement between the serological test (SNAP 4Dx) and PCR results and is reported with the 95% confidence interval (CI). Pearson’s χ2 test was employed for univariate analysis of the categorical data. Hematological parameters were evaluated with t-tests for comparison of two independent groups to detect significant differences. Statistical significance was considered to be indicated by a P-value less than 0.05.

RESULTS

16S rDNA PCR results for DNA extracted from dog blood and ticks

Among the 315 ticks examined in this study, Rhipicephalus sanguineus was the dominant tick species, represented by 36 nymphs and 270 adults (122 females, and 148 males). Haemaphysalis hystricis was also found (a total of 9 ticks, including 2 engorged females, 6 males, and 1 nymph), but only in Communities 12 and 13 (Fig. 1). Most of the R. sanguineus nymphs (30/36 nymphs; 83.3%) were collected from mountainous communities visited in the fall (Communities 9 and 14). The PCR results for the dogs and ticks are summarized according to their respective communities (Table 1). In Table 1, only the PCR results from the 306 R. sanguineus specimens are shown because all nine H. hystricis were PCR negative for the various pathogens tested in this study. None of the samples examined in this study tested positive for A. phagocytophilum, E. ewingii, or E. chaffeensis.

Among the 175 dogs and 306 R. sanguineus, positive results of nested PCR for the 16S rRNA gene were obtained for A. platys in 47 dogs (26.9%) and 71 ticks (23.2%) and for E. canis in 15 dogs (8.6%) and only 3 ticks (1%). The origin of the sequenced positive amplicons is given in Table 1. The sequencing analysis (Table 2) indicated that the E. canis-positive nucleotide fragments were 100% identical to the E. canis strains from Taiwan (GenBank accession number: DQ258496), Lima (DQ915970), and Portugal (EF051166), while the A. platys-positive fragments were 100% identical to the A. platys strains from Okinawa (GenBank accession number: AY077619), Guangzhou (AF156784), the Philippines (JQ894779), and Sommieres, France (AF303467). No nucleotide variation was observed among the sequenced PCR amplicons.

Table 2.

Comparison of A. platys and E. canis nucleotide fragment sequences from this study with other A. platys and E. canis strains

Comparisons of positive test results between SNAP 4Dx and 16S rDNA PCR, and results of A. platys groEL gene PCR for dogs and ticks positive to A. platys 16S rDNA PCR.

Difference in seroprevalence among communities were depicted in Fig. 1. Among the 16 communities, 44% (7 communities) were classified as communities with low seroprevalence (range: 0–20% for Anaplasma sp. and 0–7.1% for E. canis), and 56% (9 communities) were classified as showing high seroprevalence (range: 28.5–64.3% for Anaplasma sp. and 15.8–50% for E. canis). Seropositivity for both E. canis and Anaplasma sp. was observed in most of the communities, except for Communities 8, 11 and 16, where the local dogs were only seropositive for Anaplasma sp., and for Communities 2, 4 and 9, where no dogs showed seropositivity to SNAP 4Dx testing (Table 1). However, contrary to our expectation, a number of dogs were positive to nested PCR for A. platys 16S rRNA gene regardless of the local seroprevalence. While there was moderate agreement between the test results from the 16S rRNA gene PCR and SNAP 4Dx analyses for E. canis (Kappa value: 0.525; 95% CI 0.315 to 0.735), the agreement between the A. platys 16S rRNA gene PCR assay and the SNAP 4Dx Anaplasma sp. results was very poor, at −0.060 (95% CI −0.199 A. platys groEL gene [3] and hemi-nested amplification of this gene using a newly designed primer (PlatGro-kc-as) were conducted for all of the DNA samples that were positive for A. platys 16S rDNA. Among the 47 dogs and 71 ticks that were positive for the A. platys 16S rRNA gene, only 20 dogs and 11 ticks were positive for the groEL gene. The sequencing analysis of the positive amplicons showed them to be closest to the A. platys strain identified in R. sanguineus from the Philippines (JN121382), with 100% DNA identity (Table 2). Among the DNA samples negative for the A. platys 16S rRNA gene, none were positive for the A. platys groEL gene.

All of the DNA samples that were positive for the A. platys 16S rRNA gene but negative for the A. platys groEL gene (27 dogs and 60 ticks) were negative for the Anaplasma spp. groEL gene according to PCR using the newly designed primers ApGro-1s and ApGro-284as. A partial sequence of each gene (E. canis 16S rRNA and A. platys 16S rRNA and groEL genes) was submitted to GenBank under the accession numbers KY565476, KY565477, KY565478, KY565479, KY581623 and KY581624.

Definition of A. platys infection

The DNA samples that were positive for A. platys 16S rDNA according to PCR were divided into two groups based on the results of PCR for the A. platys groEL gene: one group positive for both the 16S rRNA and the groEL genes, while the other was positive for 16S rDNA but negative for the groEL gene. Because the nucleotide fragments amplified via nested PCR targeting the A. platys 16S rRNA gene were also 100% identical to unclassified Anaplasma spp., such as an A. platys-like Anaplasma sp. detected in goats in southern China (accession number: JN558821) and the Anaplasmataceae agent detected in brown brocket deer in Brazil (accession number: KF020572), we considered it possible that the nucleotide fragments may not conclusively verify the identity of the template DNA as A. platys. Accordingly, we defined A. platys as being confirmed by PCR for both the A. platys 16S rRNA and groEL genes, while positivity only for A. platys 16S rDNA indicated a probable unknown Anaplasma sp.

Distribution of dogs and ticks carrying E. canis, A. platys, and a probable unknown Anaplasma sp.

Among the 175 dogs examined, 8.6% (15/175) were PCR positive for E. canis, 11.4% (20/175) for A. platys and 15.4% (27/175) for an unknown Anaplasma sp., including 3 dogs co-infected with E. canis and A. platys. The E. canis-positive dogs came from Communities 1, 3, 6, 7, 11, 12, 13, 14 and 15 (Table 1). Although 67% (10/15 dogs) of the dogs carrying E. canis were detected via SNAP 4Dx, 13% (5 dogs) of the dogs carrying E. canis dogs in Communities 3, 7, 11 and 13 were negative according to SNAP 4Dx. Three ticks (1%) positive for E. canis as well as the 3 co-infected dog samples were collected from Communities 13 and 14. The three E. canis-positive ticks, including 2 fully engorged females and 1 engorged nymph, were removed from two dogs, both of which were infected with E. canis.

There was no significant association between the SNAP 4Dx results for Anaplasma sp. and A. platys subclinical infections (i.e., dogs positive for both the A. platys 16S rRNA and groEL genes) (P=0.076; χ2 test); 60% (12/20 dogs) of A. platys-infected dogs were not detected by SNAP 4Dx. Notably, in Community 9, although 54% (6 of 11 dogs) of the local dogs were infected by A. platys, none of the dogs tested positive for Anaplasma sp. according to SNAP 4Dx. Interestingly, although 18 of the 26 ticks collected from Community 9 were removed from these 6 A. platys-positive dogs, almost all (17) of these 18 ticks were positive only for an unknown Anaplasma sp. Overall, there was a significant association between the SNAP 4Dx Anaplasma sp. results and the presence of infesting ticks carrying A. platys (P=0.008; χ2). The geographical origins of the 11 A. platys PCR-positive ticks and the PCR results for their host dogs are summarized in Table 3. These A. platys positive ticks comprised 6 females, 2 males and 3 nymphs and were collected from 8 dogs from 6 communities. A. platys was identified in 2 females and 1 male tick removed from dogs that were negative for A. platys infection. Co-infection of A. platys and E. canis was not observed in the 11 ticks, despite the fact that 2 of the 8 host dogs were infected with both A. platys and E. canis (Table 3).

Table 3.

Ticks carrying A. platys and their host dog infection status

Tick ID (Host dog ID- tick number)	Community ID	Stage	Host dog status
			A. platys 16S rDNA	A. platys groEL	E. canis 16S rDNA
V437	9	female	+	+	–
V230-17	11	female	–	–	–
V230-29	11	female	–	–	–
V521-18	12	female	+	+	–
V481	13	female	+	+	+
V493-2	13	female	+	+	–
V496-1	14	male	+	+	–
V498-3	14	nymph	+	+	+
V498-5	14	nymph	+	+	+
V498-21	14	nymph	+	+	+
V261-2	16	male	–	–	–

Communities with a high frequency of dogs positive for the unknown Anaplasma sp. were observed: Community 2 (60%: 6/10 dogs), Community 11 (26.9%: 7/26 dogs), and Community 15 (57.1%: 8/14 dogs) (Table 1). Overall, 60 ticks (19.6%) from 9 communities were positive only for the A. platys 16S rRNA gene, including 15 nymphs, 22 females and 23 males, with a significantly higher prevalence being observed in nymphs (41.7% in nymphs vs. 18% in females and 15.5% in males, P<0.05). None of the dogs positive for the unknown Anaplasma sp. were positive for Anaplasma spp. according to SNAP 4Dx. The hematological profiles (such as platelet numbers and hematocrit values) of the dogs positive for the unknown Anaplasma sp. were within the normal range. The mean platelet number and mean hematocrit for this group were 256.4 × 103/µl (SD 105.5) (P=0.603, mean for dogs negative for A. platys 16S rDNA and dogs positive for A. platys groEL gene: 244.1 × 103/µl, SD 133.9) and 42.96% (SD 7.2) (P<0.05, mean for dogs negative for A. platys 16S rDNA and dogs positive for A. platys groEL gene: 39.29%, SD 8.6), respectively, while the corresponding values for A. platys-infected dogs were 200.6 × 103/µl (SD87.7) (P<0.05, mean for non-infected dogs: 251.6 × 103/µl, SD 133.271) and 34.64% (SD 6.6) (P<0.05, mean for non-infected dogs: 40.5%, SD 8.5).

DISCUSSION

This study was conducted based on the findings of our previous community-based study [38] to clarify the factors contributing to the marked differences in the seroprevalence of canine anaplasmosis among communities with equally high tick infestation levels. The results demonstrated that 8.6% (15/175 dogs) and 11.4% (20/175 dogs) of the dogs were carrying E. canis and A. platys, respectively, while R. sanguineus, the single dominant species in the study area, harbored E. canis at a frequency of only 1% (3/306 ticks) but A. platys at a frequency of 3.6% (11/306 ticks). The A. platys strain prevalent in this area appeared to belong to the same clone as an A. platys strain identified in the Philippines, with 100% identity. The results suggest that R. sanguineus plays a greater potential role in the transmission and maintenance of A. platys than as a reservoir of E. canis. A number of DNA samples positive for A. platys 16S rDNA unexpectedly tested negative for the A. platys groEL gene according to a hemi-nested PCR analysis, which resulted in an incomparable prevalence when calculated based on the 16S rRNA and groEL PCR results independently. One possibility for the occurred discrepancy might be because Anaplasma sp. that is not known to infect dogs or a potential novel Anaplasma sp. is prevalent in the study area. Dogs infected with this probable unknown Anaplasma sp. were not seropositive for Anaplasma sp. according to the SNAP 4Dx test and lacked the apparent hematological abnormalities accompanying A. platys infections; the prevalence was 15.4% (27/175 dogs) in dogs and 19.6% (60/306 ticks) in R. sanguineus, with some enzootic communities being observed.

The prevalence of A. platys in dogs according to PCR varies across geographical regions worldwide, e.g., 33% in Italy [30], 24.7% in Grenada [37], and 8.3% in Arizona [10]. A previous study reported a prevalence of 8.9% (4/45 pet dogs) in northern Taiwan [5], which was lower than the prevalence (11.4%) observed in our study. This may be due to differences in sampling, because most of the dogs (133/175) examined in the present work came from an environment with a high level of tick infestation, which has been linked to a high prevalence of A. platys in dogs [5, 18]. Our results confirmed this previous finding by revealing that subclinical A. platys infections were more frequently observed in communities with high tick infestation.

However, our study also revealed some controversial aspects regarding the detection and determination of A. platys infection. First, the target sequence for the PCR identification of A. platys may affect the A. platys prevalence observed via PCR considerably, as an 11.4% prevalence was determined based on PCR positivity for both the A. platys 16S rRNA and groEL genes, while a 26.9% prevalence was detected according to PCR positivity for only the 16S rRNA gene. The prevalences calculated based on the 16S rRNA and groEL PCR results were not comparable. As the reason for this, it is feasible to assume the presence of an unknown Anaplasma sp. that appeared to be carried by dogs and R. sanguineus was enzootic in the study area. The unknown Anaplasma sp. could be a novel species closely related to A. platys. This issue will be further discussed later in this section. Second, our findings also indicated the limited ability of SNAP 4Dx to detect A. platys infections. In contrast to a previous study [10], there was no significant correlation of the test results from the SNAP 4Dx test for Anaplasma sp. and the PCR test for A. platys. By using the new product, SNAP 4Dx Plus test, we have been experiencing similar discordant test results between PCR and serology (data not shown). Compared with the prevalence in the dog population examined in the previous study, in which E. canis infection was highly enzootic (greater than 36%), the PCR prevalence of E. canis observed in dogs in our study area was only 8.6%. It has been implied that persistent infection with A. platys may be more efficiently detected via SNAP 4Dx testing for Anaplasma sp. in the case of sequential or concurrent infections with A. platys and E. canis than in cases of infection by A. platys alone [12]. In addition, our results showed that infestation by A. platys-carrying ticks was positively associated with seropositivity for Anaplasma sp. according to SNAP 4Dx. Further work is required to address the mechanisms of seropositivity for Anaplasma sp. according to SNAP 4Dx in A. platys infections.

In our study, the frequency of R. sanguineus showing PCR positivity for E. canis was only 1% (3 of 306 ticks) (2 fully engorged females and 1 nymph); however, 8.6% (15/306) of the dogs were positive according to PCR. The prevalence of E. canis among ticks has been reported to be 2.0% in Malaysia [26]. In contrast to the results for E. canis, the frequency of R. sanguineus carrying A. platys was 3.6% (11/306 ticks), including 1 male and 2 female ticks that were removed from dogs not showing PCR positivity for A. platys. Notably, although 4 of the 11 A. platys-positive ticks were removed from 2 dogs infected with A. platys and E. canis concurrently, E. canis was not detected in these 4 ticks, despite the fact that one of them was an engorged female (Table 2). This may suggest that the results regarding E. canis-carrying ticks may not be due to the contamination of host dog blood. Because the tick population in this study did not include ticks in the questing stage, it is still unclear whether R. sanguineus could act as a natural reservoir or vector for A. platys transmission. It has been suggested that R. sanguineus tropical sp. is a competent vector for E. canis transmission, while that R. sanguineus template sp. is not [25]. Further studies on the vector competency of the R. sanguineus species complex are required to understanding the maintenance and transmission of E. canis and A. platys in Taiwan.

The prevalence of a novel Anaplasma sp. that is genetically close to A. platys has been reported in ruminants and deer from other regions of the world [22, 27, 36, 40]. However, to the best of our knowledge, this is the first report that a potential novel Anaplasma strain closely related to A. platys may be prevalent in dogs. The primer pair ApGro-1s and ApGro-284 was used in the Anaplasma groEL PCR with the aim of amplifying a broader range of Anaplasma groEL sequences as well as ruling out the possibility of A. phagocytophilum. GroEL sequences similar to other Anaplasma sp., but not to A. platys or A. phagocytophilum, have been detected in animals from which the 16S rDNA sequence close to A. platys was amplified simultaneously [40]. This study showed that the infected dogs exhibited no antibody reaction to Anaplasma sp. in SNAP 4Dx test, and the clinical significance of this organism appeared to not be associated with the hematological abnormalities commonly observed in A. platys infections. Further work is required to pursue the true identity of this unknown Anaplasma sp. through genetic and pathological analyses.

This study showed that the prevalence of A. platys, most closely related to an A. platys strain (JN121382) from the Philippines, predominated over that of E. canis in dogs and R. sanguineus in Nantou County of central Taiwan. This study also suggested that R. sanguineus plays a role as a vector and reservoir of A. platys, while its role as a reservoir of E. canis is suspect. PCR targeting the A. platys 16S rRNA gene should be avoided as a routine diagnostic method for A. platys infection because our results indicated that it is likely that an unknown Anaplasma sp. closely related to A. platy is prevalent among asymptomatic dogs and may affect the accurate detection of A. platys in this area. The unknown Anaplasma strain could be important for further identification. Further work is required to address the genetic analysis of A. platys and Anaplasma sp. prevalent in the study area as well as the acquisition process of Anaplasma sp. and E. canis by ticks in the natural environment.

CONFLICT OF INTEREST

The authors declare no conflict of interest.