An Assessment of Whole Blood and Fractions by Nested PCR as a DNA Source for Diagnosing Canine Ehrlichiosis and Anaplasmosis

Ehrlichiosis and anaplasmosis are tick-borne diseases. Ehrlichia canis and Anaplasma platys infect mainly white cells and platelets, respectively. The main DNA source for PCR is peripheral blood, but the potential of blood cell fractions has not been extensively investigated. This study aims at assessment of whole blood (WB) and blood fractions potential in nested PCR (nPCR) to diagnose canine ehrlichiosis and anaplasmosis. The 16S rRNA gene was amplified in 71.4, 17.8, 31.57, and 30% of the WB, granulocyte (G), mononuclear cells (M), and buffy coat (BC) samples. Compared to the WB, the sensitivity of the PCR was 42.86% for the M, and BC fractions, 21.43% for the G, and 33.33% for the blood clot (C). There was fair agreement between the WB and M, BC and C, and slight with the G. Fair agreement occurred between the nPCR and morulae in the blood smear. One animal was coinfected with A. platys and E. canis. This study provided the first evidence of A. platys infection in dogs in Paraíba, Brazil, and demonstrated that WB is a better DNA source than blood fractions to detect Ehrlichia and Anaplasma by nPCR, probably because of the plasma bacterial concentration following host cell lysis.


Introduction
Ehrlichiosis and anaplasmosis are important, emerging zoonotic tick-borne diseases caused by gram-negative, obligate intracellular bacteria from the Anaplasmataceae family. In the host cells, the bacteria reside in inclusion bodies (morulae), which provide a hospitable environment for survival [1,2].
helminthoeca. Ehrlichia and Anaplasma infections are transmitted through the salivary secretions of attached ticks. Ehrlichia canis is usually transmitted by brown dog tick (Rhipicephalus sanguineus) bites, which can also transmit E. ewingii and most likely Anaplasma platys [1]. The occurrence of the tick R. sanguineus parasitizing humans in Brazil [3] serves to warn the risk of transmission of such pathogens (A. platys and E. canis) to humans [4,5].
E. canis species mainly infect monocytes, which causes canine monocytic ehrlichiosis, and A. platys species infect platelets, which causes canine cyclic thrombocytopenia. 2 The Scientific World Journal The A. platys platelet tropism is unique among ehrlichialrelated organisms, even though all of these infections may result in thrombocytopenia [2]. E. canis is the main pathogen implicated in cases of canine ehrlichiosis in Brazil, but A. platys has recently been identified by PCR in samples from the South region with a prevalence ranging from 25.5% to 55% [1,5].
The diagnosis of canine ehrlichiosis and anaplasmosis relies on the cultivation, serology, PCR, and direct microscopic examination of stained blood smears to identify intracytoplasmic morulae. Smear diagnosis has low sensitivity, as there are few bacteria present in the samples, morulae can be visualized only during the acute phase, and the percentage of infected cells is usually less than 1% [6]. Additionally, the presence of A. platys is cyclical, and the bacteria are easily mistaken as nonspecific inclusion bodies and staining artifacts [1,7]. Serology is hampered by cross-reactions and cannot discriminate between a current infection and previous exposure to the pathogen. Moreover, antibody titers tend to persist for several months to years after treatment, making serology an unreliable tool for posttreatment diagnosis [8].
The first PCR-based diagnostic method for ehrlichiosis amplified the 16S rRNA gene and was reported by Iqbal et al. in 1994 [9]. Further improvements and the use of other target genes increased the sensitivity of the tests. The p30based nested PCR (nPCR) assay has been shown to be more sensitive than the 16S rRNA-based nPCR assay [10], possibly because E. canis contains multiple copies of the p30 gene but only one copy of the 16S rRNA gene [11]. As opposed to single-step PCR, nPCR amplification of the 16S rRNA gene has been used more often to detect E. canis and A. platys. In both single step PCR and nPCR, the peripheral blood is frequently used as a DNA source [1,5,12]. Only a single report has described the use of mononuclear cells as a DNA source [9].
There is a high prevalence of canine ehrlichiosis, but there are few reports on the identification of the infectious agents; therefore, a practical diagnostic technique that can be routinely used in veterinary medicine must be established. The nPCR assay may fulfill this requirement, but the blood fraction that serves as the best DNA source must be determined beforehand. The aim of the present study was to compare the effectiveness of whole blood (WB) and blood fractions-buffy coat (BC), granulocytes (G), mononuclear fraction (M) and blood clot (C)-by nPCR to diagnose canine ehrlichiosis and anaplasmosis.

Samples and Cell
Fractionation. Blood was collected from 21 dogs bearing suggestive clinical signs of either ehrlichiosis or anaplasmosis (petechia, ecchymosis, fever, and anorexia) and harboring ticks. Some animals also had intracytoplasmic morulae, as indicated by direct examination of blood smears and/or hematological parameters suggestive of ehrlichiosis and anaplasmosis. The dogs were selected from the veterinary hospital Universidade Federal

Hematology, Direct Examination of Blood Smears and Cell
Fractionation. Routine platelet counts, packed cell volume, and other hematology parameters were performed at the hospitals referred to above. The reference values were those described in Jain (1993) [15]. WB smears were stained with a hematoxylin-eosin-based rapid stain (Panótico rápido, Laborclin, Brazil) and observed by microscopy (100X objective, under immersion oil). The M-and G-enriched samples were obtained from 4 mL of WB with the SepCell kit (LGC Biotecnologia, Brazil), according to the manufacturer's instructions. The BC fraction was collected from 1 mL blood that was centrifuged at 12,000 g for 10 min.

Nested PCR (nPCR).
The first round of PCR used 0.5 to 1.0 µg of the genomic DNA, and the primers ECC and ECB were designed to amplify a 478 base-pair (bp) fragment of the Ehrlichia 16S rRNA [13]. The second round of PCR used a 1.0 µL aliquot of the first reaction as a template and the EHCA sense/EHCA antisense [14] and EHPL sense/EHPL antisense (João Pessoa Araújo Jr.: pers. comm., 2010) primers, which were designed to amplify a 389 bp fragment for E. canis and 384 bp fragment for A. platys, respectively. Separate reactions were used to detect each species individually. The primers are described in Table 1. The primer design was confirmed with the software Primer 3 (http://fokker.wi.mit.edu/primer3/input.htm). The reaction mix contained 1X reaction buffer (50 mM KCl, 20 mM Tris-HCl (pH 8.4), and 0.1% Triton X-100), 1.75 mM MgCl 2 , 0.2 mM dNTP mix, 1 µM PCR primers, 0.625 U Taq DNA polymerase, and autoclaved ultrapure water to a final volume of 25 µL. The thermocycle was as follows: 94 • C for 10 minutes followed by 40 cycles at 94 • C for 60 seconds, 60 • C for 60 seconds, 72 • C for 60 seconds, and a final step of 72 • C for 4 minutes before holding at 4 • C. Ultra-pure autoclaved water was used as negative control in each PCR batch. The genomic DNA from confirmed E. canis and A. platys cases was used as positive controls for the E. canis 16S rRNA and A. platys 16S rRNA genes, respectively. Ten microliters of the final products were electrophoresed at 90 volts for approximately 1 hour in 1.5% agarose gels containing ethidium bromide in Tris-Borate EDTA (TBE).  Table 2 shows the results of hematological, blood smear (direct examination), and nPCR on the WB, G, M, BC, and C samples from 21 dogs exhibiting clinical signs of ehrlichiosis. From each group, negative samples were detected. In seven animals (46.6%), identification at species level failed, as there was no amplification in the second PCR. Among them, the blood smears of five dogs were positive by direct examination and two displayed cytoplasmic inclusions. Seven dogs (33.3%) were positive by nPCR and direct examination of blood smears (presence of morulae); inclusions within platelets were found in two blood smears. Out of the 14 blood smear-negative animals, eight (63.6%) had at least one blood fraction positive for Ehrlichia or Anaplasma by nPCR, corresponding to 57.1% false negatives by direct examination. The WB DNA samples from 66.6% (6/9) thrombocytopenic and 42.85% (3/7) anemic animals were positive by nPCR.

Discussion
The direct examination of stained blood smears to detect Ehrlichia in dogs has a low sensitivity rate (3 to 9%). In fact, E. canis morulae are difficult to detect in blood smears because this organism is usually present in very low concentrations [6]. In contrast, PCR has proven to be more sensitive for detecting Ehrlichia; for a 16S rRNA-based PCR assay is able to detect E. canis DNA from a rickettsemia, which is equivalent to one infected monocyte in 10 36 cells [1,5,12]. In addition to the large sensitivity differences inherent to the techniques, genotypic variants have been reported for E. ruminantium, and A. platys infects a wide range of host cells [1,2,17].  As expected, our study demonstrates that nPCR is more sensitive for detecting Ehrlichia than the direct examination of stained blood smears of dogs with suggestive clinical signs. Our results show that a 50% false negative rate may occur when only direct examination is used for diagnosis. In contrast, all animals with morulae in the blood smears were positive by nPCR for at least one of the WB or fraction samples.
The nPCR was able to detect Ehrlichia or Anaplasma DNA in 71% of the samples from dogs with suggestive clinical signs. This rate is slightly higher than that registered elsewhere in Brazil [1,5,12]. As previously reported [1,5], E. canis (46.6%) positivity in WB was higher than for A. platys (6.6%).
In seven (46.6%) of the samples, there was no amplification in the second PCR, and the positives were recorded as Ehrlichia spp. As the primers used were specific for E. canis and A. platys, the presence of other Rickettsiales, such as A. phagocytophilum, E. chaffeensis, and E. ewingii, should not be disregarded because they can also form cytoplasmic inclusions [18,19]. Furthermore E. ewingii was already reported in dogs in Brazil [20].
Coinfection with E. canis and A. platys was observed in an animal with a positive blood smear and that was positive for E. canis in the WB sample by nPCR. Cytoplasmic inclusions in the platelets were not observed, possibly due to low A. platys load [7]. It is worth mentioning that this is the first evidence for the involvement of A. platys in canine anaplasmosis in the State of Paraiba, Brazil.
The blood fraction samples that were positive for A. platys by nPCR were WB and C (dog no. 14) and B and C (dog no. 12). Despite the small sample size, the results suggest an increased likelihood of finding A. platys DNA in the BC fraction, which is more enriched with platelets than the other samples.
Contrary to previous reports [21,22], we found that there was no statistical association between thrombocytopenia (P = 0.596), anemia (P = 0.299), and the WB nPCR results. Similar to a previous report [1], anemia occurred in only 26.6% cases. These results demonstrate that thrombocytopenia is not sufficient to diagnose either canine ehrlichiosis or anaplasmosis. Santos et al. [22] also observed a high incidence of E. canis infection among nonthrombocytopenic dogs. In contrast, other diseases including immune-mediated thrombocytopenia, neoplasia, inflammatory diseases, or other infectious agents can provoke thrombocytopenia [23]. The differences in prevalence may reflect the diversity in strain pathogenicity or a selection bias because thrombocytopenic dogs are more likely to be tested for ehrlichiosis.
Peripheral blood has been the main source of Ehrlichia DNA for PCR assays because collection of this sample is less invasive than spleen and bone marrow collection. The use of serum samples for nPCR to detect E. canis has been suggested previously [24]. Our results support that whole blood is the best source for Ehrlichia DNA in PCR assays. Indeed, the Kappa value indicates a weak correlation between nPCR results from the WB samples and those obtained with the G, M, BC, or C samples; the PCR sensitivity from the M and B samples was only 42.9%. Therefore, our data and the literature support the use of WB as the best choice for DNA source for PCR Ehrlichia spp. detection. This is the first assessment of the use of different blood cell fractions as DNA sources to diagnose canine ehrlichiosis and anaplasmosis by PCR. Although the pathogens only infect leukocytes and platelets, WB is a better DNA source than any of the cellular Ehrlichia-enriched host cell fractions. A possible explanation may be based on the assumption that WB samples contain not only intracellular Ehrlichia but also organisms released by host cell lysis that are not found in the fractions. In support of this hypothesis, the 16S rRNA gene was successfully amplified by Mylonakis et al. [25] by nPCR in sera samples from naturally infected dogs. Hence, these authors recommend serum-based PCR analysis for the early diagnosis of CME when WB samples are not available. Furthermore, it was demonstrated that E. chaffeensis reached concentrations of ∼10 8 bacteria/mL in the plasma of SCID mice two weeks after infection [26]. There are no similar studies for E. canis or A. platys, but it is reasonable to assume that a similar scenario occurs in dogs infected with these pathogens, especially in the acute phase of the disease, when symptoms are severe, and platelet counts are usually reduced.
In conclusion, the present study demonstrates that canine WB is better than other cellular blood fractions as a DNA source to detect Ehrlichia and Anaplasma by PCR, most likely because of the bacterial concentration in the plasma following host cell lysis.