Available online at www.sciencedirect.com

www.elsevier.com/locate/tvjl

The Veterinary Journal 179 (2009) 145–148

TheVeterinary Journal

Short Communication

Molecular evaluation of the incidence of Ehrlichia canis,Anaplasma platys and Babesia spp. in dogs from Ribeirao Preto, Brazil

Flavia Santos a, Juliana S. Coppede a, Andre L.A. Pereira a, Letıcia P. Oliveira a,Patrıcia G. Roberto a, Roberta B.R. Benedetti c, Lenise B. Zucoloto c, Flavia Lucas d,

Lucia Sobreira d, Mozart Marins a,b,*

a Unidade de Biotecnologia, Universidade de Ribeirao Preto (UNAERP), Ribeirao Preto, SP, Brazilb Heranza – Biotecnologia, Ribeirao Preto, SP, Brazil

c Curso de Medicina Veterinaria, Centro Universitario Barao de Maua, Ribeirao Preto, SP, Brazild Curso de Medicina Veterinaria, Centro Universitario Moura Lacerda, Ribeirao Preto, SP, Brazil

Accepted 17 August 2007

Abstract

Canine monocytic ehrlichiosis caused by Ehrlichia canis is endemic in many regions of Brazil. Since thrombocytopenia is a commonfinding in infected dogs, many clinicians tend to use it as an indication for antibiotic treatment. Polymerase chain reaction (PCR) andnested PCR were used to study the presence of E. canis, Anaplasma platys and Babesia spp. in thrombocytopenic and non-thrombocy-topenic dogs from Ribeirao Preto, Brazil. Despite the high prevalence of E. canis infection among thrombocytopenic dogs, 46.7% of thethrombocytopenic dogs studied were either infected with Babesia spp. or A. platys or not infected with any of the three pathogens. Therewas a high incidence (25.4%) of E. canis infection in non-thrombocytopenic dogs. Although infection with E. canis should be consideredin thrombocytopenic dogs, the final diagnosis needs to be confirmed by complementary tests such as blood smears and PCR to avoid theunnecessary use of antibiotics.� 2007 Elsevier Ltd. All rights reserved.

Keywords: Ehrlichia canis; Anaplasma platys; Babesia spp.; Nested PCR; Canine monocytic ehrlichiosis

In Brazil, the first case of canine monocytic ehrlichiosis(CME) was reported in 1973 in a dog from Belo Horizon-te, Minas Gerais (Costa, 1973). Today, the disease isfound throughout the country and is considered to beendemic in many regions. After routine clinical examina-tion of suspected animals, and in the absence of any char-acteristic clinical sign of other diseases, most veterinarianstend to rely on thrombocytopenia to diagnose CME, fol-lowed by antibiotic treatment with no identification ofother possible causative agents. Although a common find-ing in CME-infected dogs, thrombocytopenia can also be

1090-0233/$ - see front matter � 2007 Elsevier Ltd. All rights reserved.

doi:10.1016/j.tvjl.2007.08.017

* Corresponding author. Address: Unidade de Biotecnologia, Uni-versidade de Ribeirao Preto (UNAERP), Ribeirao Preto, SP, Brazil.Tel.: +55 16 3603 6892; fax: +55 16 3603 7030.

E-mail address: marins@heranza.com.br (M. Marins).

a manifestation of other diseases, as well as in infectionswith other Ehrlichia spp. and parasites (Harrus et al.,1999).

The area around Sao Paulo State is considered to beendemic for Ehrlichia canis. In a study conducted in thetown of Botucatu, located 240 km from Ribeirao Preto,nested polymerase chain reaction (PCR) was used to detectE. canis infection in thrombocytopenic and non-thrombo-cytopenic dogs (Bulla et al., 2004). Among dogs with aplatelet count <100,000, 63% were found to be infectedand the authors suggested that animals with severe throm-bocytopenia should be screened for ehrlichiosis. However,most veterinarians apply the reference parameter of<200,000 as an indication of thrombocytopenia. Using thisvalue as a reference, the number of infected animals falls to45.2% in that study.

146 F. Santos et al. / The Veterinary Journal 179 (2009) 145–148

In a similar investigation in Rio de Janeiro, located712 km from Ribeirao Preto, Macieira et al. (2005) con-firmed the higher prevalence of E. canis in thrombocytope-nic dogs when compared to non-thrombocytopenicanimals. However, among thrombocytopenic dogs, lessthan one-third of the animals (26.8%) were infected withE. canis. Analysing blood smears, the authors suggestedthe presence of other parasites such as Babesia spp. andAnaplasma spp. Therefore, thrombocytopenia and otherclinical signs used as an indication for E. canis infectioncan no longer be considered sufficient to establish the diag-nosis of CME. Moreover, in view of the high incidence ofother parasites in Brazil, the correct identification of theparasite (or of other diseases as the cause of thrombocyto-penia, especially when E. canis has been ruled out) contin-ues to be a problem for the clinician. In the present studywe used PCR and nested PCR to determine the incidenceof E. canis, Babesia spp. and Anaplasma platys in dogsfrom Ribeirao Preto.

The animals were seen at the clinics of two private Veter-inary University Hospitals between October and December2005. EDTA-anticoagulated whole blood was collectedfrom 221 dogs and an aliquot of each blood sample wasused for the extraction of genomic DNA using the GFXGenomic Blood DNA Purification Kit (Amersham Biosci-ences). The remaining sample was used for manual plateletcount (Jannini and Jannini-Filho, 1984). The dogs weredivided into two groups: thrombocytopenic (<200,000platelets/lL) and non-thrombocytopenic (P200,000 plate-lets/lL).

Nested PCR was used for the detection of E. canis andA. platys (Bulla et al., 2004; Dawson et al., 1994). All prim-ers were based on those previously described in the litera-ture and sequences deposited in GenBank. Briefly, in thefirst reaction the newly designed forward primer Apla-sense 5 0-CTCAGAACGAACGCTGGCGGCAAGC-3 0

and the previously described reverse primer ECB 5 0-CGTATTACCGCGGCTGCTGGC-3 0 (Dawson et al.,1994) were used to amplify the 16S rRNA gene from bothspecies. In the case of E. canis, these primers generated afragment of 481 bp, encompassing position 1–481 of thesequence deposited in GenBank (AF162860), while for A.

platys a fragment of 476 bp was generated, encompassingposition 1–476 of the sequence deposited in GenBank(AF286699).

In the second reaction, 1 lL of the first reaction wasused with the species-specific primers designed to amplifyspecific targets of the 16S rRNA gene: ECA 5 0-CAAT-TATTTATAGCCTCTGGCTATAGGAA-3 0 (Wen et al.,1997) and HE-3 5 0-TATAGGTACCGTCATTATCTTC-CCTAT-3 0 (Dawson et al., 1994) for E. canis, which gener-ated a 389-bp fragment, encompassing position 49–437 ofthe GenBank sequence, and the previously described senseprimer E. platys primer 5 0-GATTTTTGTCGTAGCTT-GCTA-3 0 (Kordick et al., 1999) and the newly designedreverse primer Aplastintrev3 5 0-GGTACCGTCATTATC-TTCCC-3 0 for A. platys, which generated a fragment of

382 bp, encompassing position 47–428 of the sequencedeposited in GenBank.

A single PCR was performed for the detection of Babesia

spp. using primers that amplify 18S rRNA gene targetsfrom a wide range of Babesia species: the previouslydescribed sense primer BabgenF 5 0-GAAACTGCGA-ATGGCTCATTA-3 0 (Baneth et al., 2004) and the newlydesigned reverse primer Babesiarev1 5 0-CCATGCTGAA-GTATTCAAGAC-3 0, which generated a 642-bp fragment,encompassing position 81–722 of the GenBank sequence(AY072926). As a positive control, cytdogFWD 5 0-CAT-CAGTCACCCACATCTGC-3 0 and cytdogREW 5 0-CCATGAATGCTGTGGCTATG-3 0 primers were usedfor amplification of the mitochondrial cytochrome b gene,which produced a 198-bp fragment. PCR and nested PCRwere performed in a PT-100 thermocycler (MJ Research)consisting of an initial step at 95 �C for 10 min, followedby 40 cycles at 95 �C for 1 min, 60 �C for 1 min and 72 �Cfor 1 min. After a final step at 72 �C for 10 min, the reactionwas stored at 10 �C. A 10-lL aliquot of each reaction wasanalyzed by agarose gel electrophoresis. The amplifiedproducts were purified and sequenced in an automatedDNA sequencer (model 377, Applied Biosystems). Fig. 1illustrates an example of the results obtained.

In contrast to previous investigations (Bulla et al., 2004;Macieira et al., 2005), a marked increase in the incidence ofE. canis infection was observed among non-thrombocyto-penic dogs. We believe that the reason for this conflictingresult is the nested-PCR procedure. Most authors use asa substrate in the second reaction, which is species specific,only the positive reactions of the first PCR whose primersdetect a wide range of species. In contrast, we screenedpositive and negative samples of the first reaction in thesecond amplification. To ensure the absence of contamina-tion, we also used the negative controls of the first reactionin the second reaction and they were always negative.Eighty-six (38.9%) of the 221 animals studied were positivefor E. canis, with 57 (66.3%) being classified as thrombocy-topenic and 29 (33.7%) as non-thrombocytopenic. More-over, we observed thrombocytopenic dogs that wereinfected with A. platys or Babesia spp. either as a singleinfection or as a co-infection with E. canis. A. platys andBabesia spp. were identified in 33 (14.9%) and 18 (8.1%)of the animals studied, respectively. Twenty-seven(81.8%) A. platys- and 17 (94.4%) Babesia spp.-positive ani-mals were classified as thrombocytopenic. Among the 221animals studied, 57/107 (53.3%) of thrombocytopenic ani-mals and 29/114 (25.4%) of non-thrombocytopenic animalswere positive for E. canis infection. Considering co-infec-tion with A. platys or Babesia spp. together with E. canis,the number of infected animals increased from 57 (53.3%)to 85 (79.4%) among thrombocytopenic animals, corre-sponding to an increase of 26.1%. However, the increasewas only 4.4%, from 29 to 34 (29.8%), in non-thrombocy-topenic animals.

Although the incidence of E. canis infection was higheramong thrombocytopenic dogs than among non-thrombo-

Fig. 1. Representative electrophoretic agarose gel showing the results of the nested PCR assay for detection of E. canis and A. platys 16S rRNA gene andof the single PCR assay for detection of Babesia spp. 18S rRNA gene. (a) Lane (1) 50-bp DNA ladder (Invitrogen), lanes (2,5) negative water control,lanes (4,7) positive control showing 198-bp products of mitochondrial cytochrome b gene, lane (3) a 389-bp product of a positive E. canis sample, and lane(6) a 382-bp product of a positive A. platys sample. (b) Lane (1) negative water control, lanes (4,5) positive control showing 198-bp products ofmitochondrial cytochrome b gene, lane (2) a 642-bp product of a positive Babesia spp. sample, lane (3) a negative Babesia spp. sample, and lane (6) 50-bpDNA ladder (Invitrogen).

F. Santos et al. / The Veterinary Journal 179 (2009) 145–148 147

cytopenic animals, 50 (46.7%) thrombocytopenic animalswere not infected with E. canis. Infection with A. platys orBabesia spp. was observed in 19 (38%) of these animals.The tick Rhipicephalus sanguineus is the common vectorfor all three pathogens studied here, a fact probably facili-tating co-infection. Positive PCR results for two or threeof the pathogens in the same animal were observed and weremore common in thrombocytopenic dogs. The chi-squaretest used to determine the statistical significance of the asso-ciation between thrombocytopenia and infection with thesepathogens clearly indicated a greater tendency toward E.

canis infection among thrombocytopenic animals comparedto non-thrombocytopenic ones (P < 0.001). This is evenmore significant when co-infection is considered.

Despite differences in the nested-PCR protocol betweenthis study and those conducted by Bulla et al. (2004) inBotucatu and by Macieira et al. (2005) in Rio de Janeiro,the findings of the three studies regarding thrombocytope-nic animals confirm that this cannot be used alone as anindicator of E. canis infection. Our results clearly demon-strated that, even in an endemic area such as RibeiraoPreto, the use of thrombocytopenia alone is not sufficientfor the diagnosis of E. canis infection and screening forother pathogens such as A. platys and Babesia spp. is nec-essary to reach a definite diagnosis.

The incidence of E. canis infection among thrombocyto-penic animals (<200,000 platelets/lL) was 45.2% and53.3% in Botucatu and Ribeirao Preto, respectively,whereas in Rio de Janeiro the incidence was only 26.8%.Moreover, in Botucatu and Rio de Janeiro the incidenceof E. canis infection among non-thrombocytopenic dogswas reported to be 1.4% and 3.5%, respectively. However,

using our protocol we observed a high incidence of 25.4%in the region of Ribeirao Preto. According to Macieiraet al. (2005), these contrasting results might be associatedwith differences in the prevalence of the disease, occurrenceof the tick vector, dog populations studied and peculiaritiesin the host-parasite relations in different regions of Brazil.

These findings indicate the need for further studiesregarding CME in Brazil, and the need for surveillanceand further investigation of dogs as a potential reservoirof E. canis and other blood parasites. Although E. canis

does not show a high zoonotic potential, dogs may serveas a reservoir for other life-threatening species of the Ana-plasmataceae family such as Ehrlichia chaffeensis (Walker,2005), the causative agent of human monocytic ehrlichio-sis. In conclusion, our results demonstrated the positiveimpact of PCR-based assays on the diagnosis of CMEand other related zoonotic diseases in the veterinary field.

Acknowledgments

This study was supported by grants from Fundacao deAmparo a Pesquisa do Estado de Sao Paulo (Grants 04/14196-5 and 05/03953-2). We thank the personnel of thetwo University Hospitals for providing the isolates in-cluded in this study, Dr. S.C. Franca and the staff of theBiotechnology Unit-UNAERP for general support, andK. Markendorf for revision of the English text.

References

Baneth, G., Kenny, M.J., Tasker, S., Anug, Y.S., Varda, L., Amos, S.,Susan, E., 2004. Infection with a proposed new subspecies of Babesia

148 F. Santos et al. / The Veterinary Journal 179 (2009) 145–148

canis, Babesia canis subspp. presentii, in domestic cats. Journal ofClinical Microbiology 42, 99–105.

Bulla, C., Kiomi Takahira, R., Pessoa Araujo, J., AparecidaTrinca, L.,Souza Lopes, R., Wiedmeyer, C.E., 2004. The relationship between thedegree of thrombocytopenia and infection with Ehrlichia canis in anendemic area. Veterinary Research 35, 141–146.

Costa, J.O., 1973. Ehrlichia canis infection in dog in Belo Horizonte-Brazil. Arquivos Escola Veterinaria UFMG 25, 199–200.

Dawson, J.E., Stallknecht, D.E., Howerth, E.W., Warner, C., Biggie,K., Davidson, W.R., Lockhart, J.M., Nettles, V.F., Olson, J.G.,Childs, J.E., 1994. Susceptibility of white-tailed deer (Odocoileus

virginianus) to infection with Ehrlichia chaffeensis, the etiologicagent of human ehrlichiosis. Journal of Clinical Microbiology 32,2725–2728.

Harrus, S., Waner, T., Bark, H., Jongejan, F., Cornelissen, A.W.C.A.,1999. Recent advances in determining the pathogenesis of caninemonocytic ehrlichiosis. Journal of Clinical Microbiology 37, 2745–2749.

Jannini, P., Jannini-Filho, P., 1984. Interpretacao Clınica do Hemograma.Sao Paulo, Sarvier.

Kordick, S.K., Breitschwerdt, E.B., Hegarty, B.C., Southwick, K.L.,Colitz, C.M., Hancock, S.I., Bradley, J.M., Rumbough, R., McPher-son, J.T., MacCormack, J.N., 1999. Coinfection with multiple tick-borne pathogens in a walker hound kennel in North Carolina. Journalof Clinical Microbiology 37, 2631–2638.

Macieira, D.B., Messick, J., Cerqueira, A.D.M., Freire, I., Linhares, G.,Almeida, N., Almosny, N., 2005. Prevalence of Ehrlichia canis

infection in thrombocytopenic dogs from Rio de Janeiro, Brazil.Veterinary Clinical Pathology 34, 44–48.

Walker, D., 2005. Ehrlichia under our noses and no one notices. Archivesof Virology 19 (Suppl.), 147–156.

Wen, B., Rikihisa, Y., Mott, J.M., Greene, R., Kim, H.Y., Zhi, N., Couto,G.C., Unver, A., Bartsch, R., 1997. Comparison of nested PCR withimmunofluorescent-antibody assay for detection of Ehrlichia canis

infection in dogs treated with doxycycline. Journal of ClinicalMicrobiology 35, 1852–1855.