A serological and molecular survey of Babesia vogeli , Ehrlichia canis and Rickettsia spp. among dogs in the state of Maranhão, northeastern Brazil

This study evaluated exposure and infection by tick-borne agents ( Babesia vogeli, Ehrlichia canis and Rickettsia spp.) in 172 dogs in rural areas and 150 dogs in urban areas of the municipality of Chapadinha, state of Maranhão, northeastern Brazil, using molecular and serological methods. Overall, 16.1% of the sampled dogs (52/322) were seroreactive to B. vogeli, with endpoint titers ranging from 40 to 640. For E. canis, 14.6% of the dogs (47/322) were seroreactive, with endpoint titers from 80 to 163,840. Antibodies reactive to at least one of the five species of Rickettsia were detected in 18.9% of the dogs (61/322), with endpoint titers ranging from 64 to 4,096. High endpoint titers were observed for Rickettsia amblyommii . Three (0.9%) and nine (2.8%) canine blood samples were PCR-positive for Babesia spp. and E. canis. The ticks collected from urban dogs were all Rhipicephalus sanguineus sensu lato, whereas the rural dogs were infested by R. sanguineus s.l, Amblyomma cajennense sensu lato and Amblyomma ovale. One A. ovale tick was found to be infected by Rickettsia bellii. This study provides an epidemiological background for controlling and preventing canine tick-borne diseases in a neglected region of Brazil.


Introduction
The emergence and reemergence of arthropod-borne diseases have been a challenge for veterinary and human medicine.Arthropods and the infections transmitted by them are expanding their zoogeographic boundaries due to climate changes and increased accessibility to certain environmental niches (SHAW et al., 2001).Vector-borne pathogens such as Babesia vogeli, Ehrlichia canis and spotted fever group Rickettsia cause diseases in dogs in many parts of the world.Dogs infected with these organisms may develop either subclinical infection or clinical signs of pyrexia, lethargy, pallor, hemorrhagic diatheses, icterus and vasculitis.Infection may be fatal in some dogs (HII et al., 2012).
The genus Rickettsia comprises a number of distinct serotypes that have been classified into separate species.Some of them are pathogenic and others are apparently nonpathogenic for humans (YU & WALKER, 2003).They are typically transmitted by arthropod vectors, which include ticks, mites, fleas or lice (BURGDORFER, 1988).The most pathogenic species, Rickettsia rickettsii, is the agent of Brazilian spotted fever in southeastern Brazil, where it is transmitted by Amblyomma ticks (LABRUNA, 2009).The other tick-associated Rickettsia species reported in Brazil are Rickettsia parkeri, Rickettsia amblyommii, Rickettsia rhipicephali, Rickettsia monteiroi and Rickettsia bellii.Among these, only R. parkeri has unquestionably been associated with human illness (LABRUNA et al. 2011).While several recent studies have revealed circulation of different Rickettsia species in some areas of Brazil, no study has been conducted in the state of Maranhão.
Taking into account that dogs are frequently infested by ticks, are companion animals and, depending on their lifestyle, may have access to different environments, they are able to exchange ectoparasites with other animals species and introduce pathogens into certain places, many of which causes zoonoses (SZABÓ et al., 2010).In this context, the present study (cross-sectional) evaluated exposure and infection by B. vogeli, E. canis and Rickettsia spp. in dogs in rural and urban areas in the state of Maranhão, northeastern Brazil, where information on tick-borne diseases is very scarce.

Materials and Methods
Ethics committee.The present study was approved by the Bioethics Committee for Animal Experimentation of the State University of Maranhão, Brazil.
Study area.During March 2010, domestic dogs were sampled in rural and urban areas of three municipalities, Anapurus (03°40'S and 43°06'W), Chapadinha (03°44' S and 43°21'W) and Mata-Roma (03°37'S and 43°06'W), within the Chapadinha region, state of Maranhão, northeastern Brazil.The Chapadinha region is characterized by a humid tropical climate, in which the rainy season predominates from December to May, while the dry season extends from July to November.The annual average temperature is higher than 27°C, with a maximum of 37°C and minimum of 21°C; the mean annual rainfall ranges from 1,200 to 1,400 mm.The area is within the Cerrado biome, with a flora composed of two diverse types of vegetation: deciduous and seasonal forests (SELBACH & LEITE, 2008).
Sampling procedure.Blood samples were collected from 150 dogs living in urban areas, and from 172 dogs living in rural areas of the Chapadinha region.Of these, 123 (50 urban and 73 rural) were in the municipality of Chapadinha, 100 (50 urban and 50 rural) in Anapurus, and 99 (50 urban and 49 rural) in Mata Roma.Owned and apparently healthy dogs were sampled by convenience directly from their residences, according to accessibility of the place.All the blood samples were collected from the cephalic or jugular vein.The whole blood and blood serum from each dog were stored separately in microtubes at -20°C until the laboratory analyses were performed.At the time of the blood collection, all the dogs had a healthy appearance.A questionnaire focusing on potential risk factors for tick-borne diseases [age, gender, living place (urban or rural) and hunting practices] was given to each dog owner.At the time of blood collection, the dogs were examined for presence of ticks.Any ticks present were collected and taken to the laboratory for taxonomic identification.The results regarding the ticks that were infesting the dogs of the present study have been reported elsewhere (COSTA et al., 2013).
Serological analyses.Canine serum samples were tested individually by means of the indirect immunofluorescence assay (IFA), using B. vogeli-infected canine erythrocytes as the antigen.The Jaboticabal strain of B. vogeli was used for this in accordance with the method described by IICA (1987).Serum samples were also tested by means of IFA using E. canis-infected DH82 cells as the antigen.The São Paulo strain of E. canis was used for this, as previously described (AGUIAR et al., 2007b(AGUIAR et al., , 2008)).Reactions were performed using fluorescein-conjugated anti-dog IgG (Sigma-Aldrich, St. Louis, MO, USA).Serum was considered to contain antibodies reactive to each agent if it displayed a reaction at the dilution 1:40.Samples that reacted at the screening dilution (1:40) were then titrated using serial two-fold dilutions to determine endpoint titers.
Antibodies reactive to Rickettsia spp.were assayed by simultaneously using five Rickettsia isolates from Brazil: R. bellii strain Mogi, R. amblyommii strain Ac37, R. rhipicephali strain HJ5, R. rickettsii strain Taiaçu and R. parkeri strain At24, as previously described (LABRUNA et al., 2007).Samples that reacted at the screening dilution (1:64) were then titrated using serial two-fold dilutions to determine endpoint titers.Serum showing a Rickettsia species titer at least four times higher than those observed for the other Rickettsia species was considered to be homologous to the first Rickettsia species or to a very closely related genotype (LABRUNA et al., 2007;PIRANDA et al., 2008).
For all reactions, a nonreactive canine serum specimen (negative control) and a known reactive canine serum specimen (positive control) were included on each slide.These sera were obtained from serum bank of the Laboratory of Parasitic Diseases in the Department of Preventive Veterinary Medicine and Animal Health (VPS) at the Faculty of Veterinary Medicine and Zootechny (FMVZ), University of São Paulo (USP).
Molecular analyses.DNA was extracted from 300 µL of whole blood from each dog using the Wizard ® genomic DNA purification kit (Promega corporation, Madison / USA), in accordance with the manufacturer's instructions.DNA samples were tested by means of the polymerase chain reaction (PCR) using the sets of primers described in Table 1.For Babesia spp., a single PCR was used (DUARTE et al., 2008); for Ehrlichia canis, a nested PCR was performed, with the second reaction specific for E. canis (DAWSON et al., 1996;MURPHY et al., 1998).Reactions were performed in a final volume of 25 µl, containing 10 mM of Tris-HCl (pH 8.3), 50 µM of KCl, 1.5 mM of MgCl 2 , 0.2 mM of each deoxynucleoside triphosphate, 1.5 U of Taq DNA polymerase (Invitrogen), 11 pmol of each primer and approximately 100 ηg of canine genomic DNA.The thermocycling conditions for the reactions were 94°C for 3 min, followed by 30 cycles at 94°C for 1 min, with annealing at 56°C (Babesia spp.) or 60ºC (first nested reaction) or 55°C (second nested reaction) for 30 s, and extension at 72°C for 1 min; a final extension step at 72°C for 3 min was used.The amplified products were viewed under ultraviolet light after electrophoresis on agarose gel stained with SyBr gold (Invitrogen).
The ticks collected from the dogs were individually subjected to DNA extraction by means of the guanidine isothiocyanatephenol technique (SANGIONI et al., 2005).They were then tested by means of PCR using the primers CS-78 and CS-323 (Table 1), targeting the citrate synthase gene (gltA) of Rickettsia spp.(LABRUNA et al., 2004a).The PCR products from the ticks were purified using ExoSap (USB) and were sequenced in an automatic sequencer (Applied Biosystems/Perkin Elmer, model ABI Prism 310 Genetic, California, USA), in accordance with the manufacturer's protocol.The partial sequences obtained were subjected to BLAST analyses (ALTSCHUL et al., 1990) to infer the closest similarities to samples in GenBank.
Statistical analysis.The minimum number of sampled dogs in each urban or rural area was statistically calculated considering each canine population, representing one fifth of the human population.Statistical associations of seropositivity to tick-borne pathogens with potential risk factors (independent variables) were tested by means of the chi-square or Fischer`s exact test, when it necessary.The odds ratio (OR) was calculated with 95% confidence limits.All analyses were performed using the Epi Info software, version 6.04d (CDC, Atlanta, GA, USA).

Results
A total of 322 dogs were sampled, comprising 148 males and 122 females; 86 dogs were < 1 year old, 187 dogs were 1-3 years old, and 49 dogs were > 3 years old.In order to increase the robustness of the statistical analyses, canine seroreactivity to B. vogeli, E. canis or Rickettsia spp.among the three municipalities (Anapurus, Chapadinha, and Mata Roma) were grouped, as presented below.
The serological results from the dogs according to whether they lived in urban and rural areas are shown in Table 2. Significantly, more dogs were seroreactive to B. vogeli in the urban areas than in the rural areas (P = 0.0119).The chances of dogs becoming seroreactive to B. vogeli in urban areas were about twice (OR: 2.27) their chances in rural areas.Regarding to E. canis, there was no significant association between urban and rural dogs (P = 0.4486).On the other hand, significantly more dogs were seroreactive to Rickettsia spp in rural areas, where the chances of dogs becoming seroreactive to Rickettsia spp were 18.09 times the chances shown by dogs living in urban areas.
Overall, 16.1% of the sampled dogs (52/322) were seroreactive to B. vogeli, with endpoint titers ranging from 40 to 640.For E. canis, 14.6% of the dogs (47/322) were seroreactive, with endpoint titers from 80 to 163,840.Antibodies reactive to at least All 61 dogs that were seroreactive to Rickettsia spp.presented reactions with R. amblyommii, for which at least 12 dogs presented endpoint titers that were four times higher than the endpoint titers showed for the other four Rickettsia species.These 12 dogs were considered to have been infected by R. amblyommii or a very closely related genotype (Table 3).Among the 61 dogs that were reactive to R. amblyommii, 54 were reactive to R. rhipicephali The proportions of dogs that were seroreactive to B. vogeli and E. canis were statistically similar (P > 0.05) for dogs of both genders and for different age groups.For Rickettsia spp, dogs of different age groups had similar seroreactivity (P > 0.05); on the other hand, significantly more female dogs were seroreactive to Rickettsia spp.than were male dogs (P < 0.05).Hunting practice, reported for 26.4% of the dogs (85/322), was not statistically associated (P > 0.05) with seroreactivity to B. vogeli or E. canis; however, it was significantly associated (P = 0.00674) with seroreactivity to Rickettsia spp.The chances of hunting dogs becoming seroreactive to Rickettsia spp was 2.33 times greater than the chances of dogs with no hunting practice.
Among the 322 blood samples subjected to molecular analyses, three (0.9%) were positive in Babesia PCR, comprising two urban dogs in the municipality of Chapadinha and one rural dog in Anapurus.E. canis-specific nested PCR was positive in nine dogs (2.8%): four in urban areas and five in rural areas.Because of logistic problems during the study, DNA sequences of these PCR products neither were nor determined.
Parasitism by ticks was found in 59.9% of the dogs (193/322), among which 40.9% (79/193) were urban and 59.1% (114/193) were rural.A total of 929 specimens were collected and identified  -: not reactive at a serum dilution of 1:64; *It was determined that the reaction was homologous when the endpoint titer for one Rickettsia species was at least four times higher than the titers observed for the other Rickettsia species.The Rickettsia species involved in the highest endpoint titer was considered to be the possible antigen involved in a homologous reaction (PAIHR).

Discussion
This study showed that 16.1%, 14.6% and 18.9% of the dogs in the Chapadinha region were exposed to infection by B. vogeli, E. canis and Rickettsia spp., respectively, when analyzed by serologic tests.While these agents had been repeatedly reported infecting dogs in other parts of Brazil at a wide range of prevalence values (TRAPP et al., 2006;LABRUNA et al., 2007;MAIA et al., 2007;PINTER et al., 2008;SAITO et al., 2008;FURUTA et al., 2009;GUIMARÃES et al., 2009;SILVA et al., 2010a, b;VIEIRA et al., 2011), data for the state of Maranhão were very scarce or absent.Recently, B. vogeli was reported infecting dogs and R. sanguineus ticks in another region of Maranhão (SILVA et al., 2012).Conversely, but there had not been any previous reports of Ehrlichia spp or Rickettsia spp on dogs in this state.On the other hand, infection by Ehrlichia sp. on domestic cats was recently reported in in São Luís Island, Maranhão (Braga et al., 2012).
R. sanguineus ticks are the only known vector of B. vogeli and E. canis (Dantas-Torres, 2008).Costa et al. (2013) reported that infestations by R. sanguineus occurred at similar prevalence rates among urban and rural dogs in the present study.This explains the statistically similar prevalence rates for seroreactivity to E. canis among urban and rural dogs.On the other hand, the presence of anti-Babesia canis antibodies was likely to be more frequent in urban areas than in rural areas in the present study, even though up to 11% of the rural dogs were seroreactive.The relatively higher B. vogeli seroprevalence in urban dogs than in rural dogs could be correlated with possibly higher density of R. sanguineus ticks in urban areas, where these ticks would find more suitable areas for development of their nidicolous off-host stages (LABRUNA & PEREIRA, 2001;SZABÓ et al., 2001;PASSOS et al., 2005).This density might have less effect on E. canis infection because this agent tends to induce lifelong infection (MCCLURE et al., 2010).
Serological results for Rickettsia spp.were much higher among rural dogs than among urban dogs, with higher endpoint titers for R. amblyommii, which was considered to be the agent possibly infecting these dogs in rural areas.R. amblyommii has been reported infecting several Amblyomma species in Brazil, including A. cajennense s.l.(LABRUNA et al., 2004b, c;OGRZEWALSKA et al., 2008OGRZEWALSKA et al., , 2011)).The dogs in the present study were found to be infested by Amblyomma ticks only in rural areas (COSTA et al., 2013).Therefore, their serological status was possibly due to their infestation by Amblyomma ticks (mainly A. cajennense s.l.) in rural areas.Similarly, the statistical association of hunting practices with seroreactivity to Rickettsia spp. was possibly associated with higher exposure to Amblyomma ticks while hunting, since these ticks are primarily associated with wildlife in Brazil (LABRUNA & PEREIRA, 2001).
No Rickettsia-infected ticks were found among the 217 R. sanguineus s.l. and 16 A. cajennense s.l.specimens collected from the dogs.The absence of rickettsial infection among the R. sanguineus ticks was corroborated through serological analysis, since this tick was the only species on the urban dogs, which were almost all seronegative for Rickettsia spp.Because canine seroreactivity to R. amblyommii was strongly associated with rural areas or hunting practices, it is reasonable to speculate that our limited sample of only 16 A. cajennense s.l.ticks was too small, thus precluding findings of R. amblyommii-infected ticks, given that R. amblyommii had already been reported infecting A. cajennense s.l.ticks in Brazil, Costa Rica and Panama (LABRUNA et al., 2004b;BERMÚDEZ et al., 2011;HUN et al., 2011).Conversely, one out of the 26 A. ovale ticks was shown to be infected by R. bellii.Indeed, R. bellii is the most common rickettsial agent infecting Amblyomma ticks in Brazil, including A. ovale (LABRUNA et al., 2011).None of the dogs of our study presented any serological evidence of R. bellii infection.It is possible that R. bellii is not infective to dogs, since a previous study also found no serological evidence of rickettsial infection among dogs that had been infested specifically by R. bellii-infected Amblyomma ticks (PINTER et al., 2008).
Regarding the two distinct habitat types (urban and rural areas) sampled in the state of Maranhão, it was shown that canine seroreactivity to E. canis was similar in urban and rural areas, whereas seroreactivity to B. vogeli was higher among urban dogs, and seroreactivity to R. amblyommii was higher among rural dogs.Interestingly, a very similar scenario was recently reported in the state of Pará (a state neighboring Maranhão), where Spolidorio et al. (2013) found anti-B.vogeli antibodies in 59.6% of the urban dogs, and in 29.1% of the rural dogs (P < 0.05).For E. canis, the seroprevalence was similar among urban dogs (15.7%) and rural dogs (16.6%); for Rickettsia spp, rural dogs presented significantly higher (P < 0.05) prevalence (40.3%) than urban dogs (21.1%), with highest endpoint titers for R. amblyommii.Similarly, in the state of Mato Grosso, in the central-western region of Brazil, Melo et al. (2011) reported similar seroreactivity to E. canis among urban dogs (74.3%) and rural dogs (67.5%) (P > 0.05), whereas seroreactivity to Rickettsia spp. was significantly higher among rural dogs (75.6%) than urban dogs (19.3%) (P < 0.05), also with highest endpoint titers for R. amblyommii.Indeed, this common scenario reported for dogs in three different Brazilian states is primarily related to the tick species found on the dogs, with predominance of R. sanguineus in urban areas and presence of Amblyomma species in rural areas.This scenario should be considered to be the epidemiological background for controlling and preventing canine tick-borne diseases in these neglected regions of Brazil.FAPEMA) for financial support.The cooperation of the dog owners in offering their valuable time and their animals for sampling is highly appreciated.

Table 1 .
Primer pairs used in the present study for detecting tick-borne agents.

Table 2 .
Results from indirect immunofluorescence assays (IFA) for Babesia vogeli, Ehrlichia canis and Rickettsia spp.among dogs in urban and rural areas of the Chapadinha region, state of Maranhão, northeastern Brazil.
*The proportions of seroreactive dogs for each antigen were compared between dogs in urban and rural areas by means of the chi-square test; CI: confidence interval; OR: odds ratio; N: number.

Table 3 .
Results from indirect immunofluorescence assays (IFA) for five Rickettsia species among dogs in rural areas of the Chapadinha region, state of Maranhão, northeastern Brazil.