INTRODUCTION
South America has witnessed a great expansion in rickettsiology and at least 6 Rickettsia species have been described in this continent for the first time since 2001 (Labruna, Reference Labruna2009; Pacheco et al. Reference Pacheco, Moraes-Filho, Marcili, Richtzenhain, Szabó, Catroxo, Bouyer and Labruna2011). Among these species Rickettsia parkeri has gained much attention. This bacterium, isolated initially by R. R. Parker from Amblyomma maculatum ticks in the USA (Parker, Reference Parker, Kohls, Cox and Davis1939), was found to be pathogenic to humans more than 60 years later (Paddock et al. Reference Paddock, Sumner, Comer, Zaki, Goldsmith, Goddard, McLellan, Tamminga and Ohl2004). Human disease caused by this Rickettsia was shown to be milder than the highly fatal Rickettsia rickettsii-caused Rocky Mountain spotted fever. It was also shown to be clinically distinct by the presence of a cutaneous eschar at the tick-bite site as well as regional lymph node enlargement (Paddock et al. Reference Paddock, Sumner, Comer, Zaki, Goldsmith, Goddard, McLellan, Tamminga and Ohl2004; Cragun et al. Reference Cragun, Bartlett, Ellis, Hoover, Tyring, Mendoza, Vento, Nicholson, Eremeeva, Olano, Rapini and Paddock2010). R. parkeri was recently found and/or isolated from Amblyomma triste ticks in Uruguay, Brazil and Argentina (Venzal et al. Reference Venzal, Portillo, Estrada Peña, Castro, Cabrera and Oteo2004; Silveira et al. Reference Silveira, Pacheco, Szabó, Ramos and Labruna2007; Nava et al. Reference Nava, Elshenawy, Eremeeva, Mastropaolo and Paddock2008). Actually, rickettsiosis due to R. parkeri has been recognized for almost 20 years in Uruguay but the infectious agent was misdiagnosed with Rickettsia conorii (Venzal et al. Reference Venzal, Portillo, Estrada Peña, Castro, Cabrera and Oteo2004; Conti-Diaz et al. Reference Conti-Díaz, Moraes-Filho, Pacheco and Labruna2009). Recently, human rickettsiosis caused by R. parkeri infection was also recognized in Argentina (Romer et al. Reference Romer, Seijo, Crudo, Nicholson, Varela-Stokes, Lash and Paddock2011).
Important aspects link human R. parkeri infections in the USA, Uruguay and Argentina; they are transmitted by either A. maculatum or A. triste, tick species that maybe confounded by their morphological similarities (Mertins et al. Reference Mertins, Moorhouse, Alfred and Hutcheson2010). Since A. triste is a species associated to flooded/marsh areas, human disease has distinct epidemiological features. Curiously in Brazil, in spite of the presence of R. parkeri-infected A. triste ticks (Silveira et al. Reference Silveira, Pacheco, Szabó, Ramos and Labruna2007), not a single human disease due to A. triste bite has been reported so far. Nonetheless, recent observations have shown that a Rickettsia genetically related to R. parkeri is circulating among other tick species, in different environmental context and is causing human disease clinically related to the disease caused by R. parkeri. Spolidorio et al. (Reference Spolidorio, Labruna, Mantovani, Brandão, Richtzenhain and Yoshinari2010) reported a novel spotted fever group (SFG) agent of human rickettsioses in São Paulo State, southeastern Brazil, in an Atlantic rainforest reserve. Hence the disease was named Atlantic rainforest spotted fever, caused by a Rickettsia genotype designated as Atlantic rainforest. Until this report, R. rickettsi was the only tick-borne Rickettsia species known to cause human disease in the country. DNA sequences of the Atlantic rainforest Rickettsia were amplified from the tick-bite lesion (eschar) and were shown to be a novel SFG strain closely related to R. parkeri, Rickettsia africae and Rickettsia sibirica. Even though the tick of this human case was not recovered, Sabatini et al. (Reference Sabatini, Pinter, Nieri-Bastos, Marcili and Labruna2010) detected high levels of Rickettsia genotype in Amblyomma ovale ticks of dogs and vegetation in another Atlantic rainforest in the state of São Paulo. Subsequently, the same Rickettsia genotype was identified in the state of Bahia (northeastern Brazil) from a patient with febrile illness which, among others, included eschar at a tick-bite location, a regionally enlarged lymph node, and a rash (Silva et al. Reference Silva, Eremeeva, Rozental, Ribeiro, Paddock, Ramos, Favacho, Reis, Dasch, de Lemos and Ko2011). Furthermore A. ovale, Amblyomma aureolatum and Rhipicephalus sanguineus ticks were shown to be infected with the Atlantic rainforest genotype on the southern coast of Brazil, where dozens of clinical cases of spotted fever rickettsiosis have been reported in humans, although without determination of the Rickettsia species (Medeiros et al. Reference Medeiros, de Souza, de Moura, Lavina, Belatto, Sartor, Nieri-Bastos, Richtzenhain and Labruna2011).
Herein we report the isolation, molecular characterization and ecological features of the Atlantic rainforest genotype from the location of the human index case in southeastern Brazil (Spolidorio et al. Reference Spolidorio, Labruna, Mantovani, Brandão, Richtzenhain and Yoshinari2010).
MATERIALS AND METHODS
Study site
A survey was held in the area of the strain Atlantic rainforest index case, an Atlantic rainforest reserve (Estação Ecológica de Juréia-Itatins) in Peruibe municipality, São Paulo State, southeastern Brazil. This reserve is located by the seaside, approximately 200 km south of São Paulo city. Part of the reserve still has scattered houses embedded in the forest, and the index case occurred in one of them. The survey was led around this house (24°25′23·6″S; 47°03′06·6″W; elevation 23 metres above sea level). Samples were collected on 6 field campaigns (April, June, August, November, December 2010 and February 2011) over a 1-year period. The present study was previously approved by São Paulo State Forest Institute (letters 858/2010 and D86/2010 W), and by the Brazilian Environment Institute – IBAMA (Permit No. 22212-2).
Host-seeking ticks
Host-seeking ticks were collected from vegetation at a range of 10 km by either cloth dragging or visual inspection as described elsewhere (Terassini et al. Reference Terassini, Barbieri, Albuquerque, Szabó, Camargo and Labruna2010). For this purpose 2 trails and their ramifications in the forest and within the property of the index case house were sampled. Additionally, the vegetation on both sides of a small 4 km-long road (Estrada do Juquiazinho) and 2 trails in the forest by a beach (Praia do Juquiazinho) were also inspected at each campaign.
Small mammals
In each campaign small mammals were captured within a radius of 500 metres from the index case house with the aid of 75 live-traps (60 Sherman and 15 Tomahawk) during 4 consecutive nights. Trapped animals were then anaesthetized with ketamine and carefully examined for ticks. Tick-infested rodents were euthanized by deepening anaesthesia, and the spleen and lungs were collected and frozen in liquid nitrogen. Blood sample was collected from all animals by venipuncture. Carcasses of euthanized animals were sent to Zoology Museum of the University of São Paulo for identification and deposit of voucher specimens.
Dogs
All ticks from the dogs of 14 properties scattered within the reserve were collected on each campaign. It was observed whether animals were restricted to the properties by either fences or tied by ropes or whether animals had access to the forest at will or occasionally with their owners. Blood samples for serology were obtained from these dogs twice; once in the first campaign, and later in the last campaign.
Tick identification
Adult ticks were identified according to the protocol of Barros-Battesti et al. (Reference Barros-Battesti, Arzua and Bechara2006), whereas Amblyomma nymphs were identified following that of Martins et al. (Reference Martins, Onofrio, Barros-Battesti and Labruna2010). Due to the absence of a proper literature, Amblyomma larvae were identified by direct comparisons with laboratory-reared larvae from engorged females of A. ovale, A. fuscum, A. naponense and A. brasiliense (the origins of these engorged females have been described elsewhere) (Martins et al. Reference Martins, Onofrio, Barros-Battesti and Labruna2010). Some larvae were reared to the nymphal stage in the laboratory; then they were identified following the protocol of Martins et al. (Reference Martins, Onofrio, Barros-Battesti and Labruna2010), which confirmed previous larval identifications. Larvae and nymphs of the genus Ixodes were identified to species according to the protocol of Marques et al. (Reference Marques, Barros-Battesti, Onofrio, Famadas, Faccini and Keirans2004).
Serology
Canine, rodent and marsupial sera were tested by the indirect immunofluorescence assay (IFA) using crude antigens derived from 5 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. Reference Labruna, Horta, Aguiar, Cavalcante, Pinter, Gennari and Camargo2007). After isolation of Rickettsia at the study site (Atlantic rainforest genotype) sera were also tested against such antigens. Briefly, sera were diluted in 2-fold increments with phosphate-buffered saline (PBS), starting from the 1∶64 dilution. Slides were incubated with fluorescein isothiocyanate-labelled rabbit anti-dog IgG (Sigma, St Louis, MO, USA), goat anti-mouse IgG (Sigma, St Louis, MO, USA) and sheep anti-opossum IgG (CCZ, São Paulo, Brazil) for canine, rodent and marsupial sera, respectively. For each sample, the endpoint IgG titre reacting with each of the 5 Rickettsia antigens was determined. An endpoint titre at least 4-fold higher for a Rickettsia species than that observed for any other Rickettsia species was considered probably homologous to the first Rickettsia species or to a very closely related species (Labruna et al. Reference Labruna, Horta, Aguiar, Cavalcante, Pinter, Gennari and Camargo2007). In each slide, a serum previously shown to be non-reactive (negative control) and a known reactive serum (positive control) were tested at the 1∶64 dilution.
Haemolymph test and PCR
Live ticks were checked individually for the presence of Rickettsia-like organisms, using the haemolymph test with Giménez staining (Giménez, Reference Gimemez1964; Burgdorfer, Reference Burgdorfer1970), and then frozen at −80 °C. A sample of ticks (n = 8) that was found to be positive for Rickettsia-like organisms in the haemolymph test was selected for rickettsial isolation in cell culture (see below). From all other frozen ticks, DNA was extracted using the guanidine isothiocyanate phenol technique (Sangioni et al. Reference Sangioni, Horta, Vianna, Gennari, Soares, Galvão, Schumaker, Ferreira, Vidotto and Labruna2005) and tested for Rickettsia by 2 different PCR protocols. Firstly, all DNA samples were tested with primers CS-78 and CS-323 targeting a relatively conserved fragment of the citrate synthase gene (gltA) that occurs in all Rickettsia species (Labruna et al. Reference Labruna, Whitworth, Bouyer, McBride, Camargo, Camargo, Popov and Walker2004). Samples yielding a visible amplicon of the expected size (compatible with 401 bp) were then tested by a second PCR protocol with primers Rr190.70F and Rr190.701R, which amplify a 632 bp fragment of the 190 kDa outer membrane protein gene (ompA) (Roux et al. Reference Roux, Fournier and Raoult1996) from only some Rickettsia species belonging to the SFG. The PCR products were sequenced and submitted to BLAST analysis to determine their similarities to the relevant sequences from identified Rickettsia species (Altschul et al. Reference Altschul, Gish, Miller, Myers and Lipman1990). A similar procedure was used to test for Rickettsia in tissue samples of small mammals; however, for DNA extraction the DNeasy tissue kit (Qiagen, Chatsworth, CA, USA), was used according to the manufacturer's instruction.
Isolation of rickettsiae in cell culture
The shell-vial technique, as previously described by Marrero and Raoult (Reference Marrero and Raoult1989), and modified by Labruna et al. (Reference Labruna, Whitworth, Bouyer, McBride, Camargo, Camargo, Popov and Walker2004), was used to isolate rickettsiae from adult ticks previously found to be positive for Rickettsia-like organisms in the haemolymph tests. Briefly, cultures of Vero cells were inoculated with tick-body homogenates and each tick, split into 2 samples, was incubated at both 28 °C and 32 °C. The percentage of Vero cells infected with rickettsiae was monitored by the use of Giménez staining of cells scraped from each inoculated monolayer. After the establishment of each isolate in the laboratory (i.e. at least 3 cell passages, with the prevalence of infected cells exceeding 90%), rickettsial DNA was extracted from the infected cells (Labruna et al. Reference Labruna, Whitworth, Bouyer, McBride, Camargo, Camargo, Popov and Walker2004). The extracted DNA was tested in a battery of different PCR protocols, using primer pairs targeting fragments of the rickettsial genes gltA and ompA as stated above, and for a 856-bp fragment of the rickettsial 135 kDa outer membrane protein (ompB) (Roux and Raoult, Reference Roux and Raoult2000), and a 834 bp fragment of the gltA gene (Labruna et al. Reference Labruna, Whitworth, Bouyer, McBride, Camargo, Camargo, Popov and Walker2004). All PCR products were purified and sequenced in an automatic sequencer (ABI Prism 310 Genetic; Applied Biosystems/Perkin Elmer, Foster City, CA, USA) according to the manufacturer's protocol. Partial sequences obtained were subjected to BLAST analysis, to determine their similarities to the relevant sequences from identified Rickettsia species.
Data analysis
Prevalence and mean intensity of tick infestation on main host species were determined according to the method of Bush et al. (Reference Bush, Lafferty, Lotz and Shostaka1997). The association between canine seropositivity to Rickettsia and access to forest or with infestation with at least 1 A. ovale tick was evaluated by two-sided Chi-square test. The association between dog parasitism with at least 1 A. ovale or R. sanguineus tick, and access to forest or restriction to a household area was evaluated in the same way. Values were considered significant at P < 0·05.
RESULTS
Ticks
Eight tick species were collected from the vegetation (Table 1). Regarding the animals, 38 dogs were examined repeatedly according to animal availability at each field campaign, totalling 119 canine examinations for the whole study. Amblyomma ovale was the main tick species of this host (743 adults, 32 nymphs and 9 larvae were collected), followed by R. sanguineus (178 adults and 19 nymphs). Generally, dogs more restricted to household surroundings were the ones that harboured R. sanguineus ticks. Mixed infestation prevalence of dogs with the aforementioned tick species encompassed 11·8% of the canine examinations. Other tick species collected from dogs were A. naponense (18 nymphs), A. brasiliense (27 nymphs and 2 larvae (reared to the nymphal stage)), A. fuscum (5 larvae) and A. cajennense (2 nymphs). Sixty-one larvae from dogs were retained as Amblyomma spp., but they were likely to represent A. brasiliense and/or A. naponense.
Host-questing ticks collected on the vegetation of the Atlantic rainforest Reserve ‘Estação Ecológica de Juréia-Itatins’ in Peruibe Municipality, state of São Paulo, Brazil, from April 2010 to February 2011
* Refers to a larval cluster that was reared to the nymphal stage in the laboratory.
† Through morphological examinations, these larvae were likely to represent A. brasiliense and/or A. naponense.
A total of 65 small mammals were captured by trapping during the study. Ticks collected from these small mammals are presented in Table 2. The prevalence and mean intensity of tick infestation on main host species are presented in Table 3. Dogs displayed the highest prevalence and highest mean intensity of infestation with adults of A. ovale. On the other hand, the rodent E. russatus was the main host for A. ovale subadults, particularly nymphs.
Tick species collected on small mammals from the Atlantic rainforest Reserve ‘Estação Ecológica de Juréia-Itatins’ in Peruibe Municipality, state of São Paulo, Brazil, from April 2010 to February 2011
N, number of captured individuals.
* Could not be identified to species level through morphological examination.
Prevalence and mean intensity of tick infestation on animals from the Atlantic rainforest Reserve ‘Estação Ecológica de Juréia-Itatins’ in Peruibe Municipality, state of São Paulo, Brazil, from April 2010 to February 2011
* Prevalence: number of tick-infested animals / number of examined animals × 100; Mean intensity: mean number of ticks per infested host.
# N, number of canine examinations, since most dogs were re-sampled during the study period.
+ larvae morphologically compatible with A. brasiliense and A. naponense.
Unplanned tick collections during the study yielded A. ovale (1 adult) on a domestic cat, Dermacentor nitens (26 adults, 1 nymph) and A. cajennense (4 adults) on 6 horses, A. ovale (1 adult) on a coati (Nasua nasua) and A. fuscum (3 males) on a lizard (Tupinambis merianae). All of these animals had direct contact with the forest, except for the 6 horses, which were restricted to a mixed overgrowth pasture adjacent to the forest.
Serology
Overall 88·6% of the dogs reacted to at least 1 rickettsial antigen. Generally, canine endpoint titres were highest against R. parkeri, with 1 animal displaying an endpoint titre as high as 32 768. It was found that 7 dogs and 1 dog respectively had endpoint titres to R. parkeri and R. rickettsii, at least 4-fold higher than those observed for the other Rickettsia species, indicating a possible homologous reaction to these Rickettsia species or closely related species. Three of these dogs, when sampled again at the end of the study, retained 4-fold higher titres to R. parkeri (Table 4). Using the same criterion, 3 E. russatus were exposed to R. parkeri or a closely related species, 2 opossums and 1 Akodon sp. to R. rickettsii or a closely related species, and 1 Monodelphis sp. was exposed to R. bellii or a closely related species. Comparative serology conducted simultaneously with Rickettsia sp. Atlantic rainforest genotype (isolate P-51; see below) and R. parkeri (strain At24) antigens with sera from animals from this study yielded identical endpoint titres.
Seroreactivity to five Rickettsia species of animals from the Atlantic rainforest Reserve ‘Estação Ecológica de Juréia-Itatins’ in Peruibe Municipality, state of São Paulo, Brazil, from April 2010 to February 2011
† An homologous reaction was determined when an endpoint titre to a Rickettsia species was at least 4-fold higher than those observed for the other Rickettsia species. In this case, the Rickettsia species (or a very closely related species) involved in the highest endpoint titre was considered the possible antigen involved in a homologous reaction (PAIHR).
Data analysis
Canine seropositivity for Rickettsia spp. was significantly associated with access to forest (P < 0·001) and with infestation with at least 1 A. ovale tick (P < 0·010). Infestation of dogs with A. ovale ticks was significantly associated with access to forest (P < 0·001), whereas no significant (P = 0·777) association was found between dog infestation with R. sanguineus ticks and either access to forest or restriction to household area.
Haemolymph test and isolation of rickettsiae
Rickettsia isolation was attempted in 8 ticks (6 A. ovale and 2 R. sanguineus) displaying Rickettsia-like structures in the haemolymph. Rickettsia were successfully isolated and established in Vero cell culture from 4 A. ovale ticks only at 28 °C. Rickettsial DNA samples amplified by PCR from these isolates were sequenced and submitted to Blast analyses. Two of the isolates (designated as P-51 and P-240), each from 1 of 2 A. ovale ticks were identified as Rickettsia sp. strain Atlantic rainforest, since their gltA (1078 bp), ompB (761 bp) and ompA (567 bp) gene fragments were 100% equal to corresponding sequences (GenBank Accession numbers GQ855235, GQ855236, and GQ855237, respectively) of the original strain Atlantic rainforest recently reported to be infecting a human in the same study site as the present study (Spolidorio et al. Reference Spolidorio, Labruna, Mantovani, Brandão, Richtzenhain and Yoshinari2010). One isolate (designated as P-10) from a third A. ovale tick was identified as R. bellii since its gltA fragment (1070 bp), was 100% identical to the corresponding sequence of R. bellii in GenBank (CP000087). As expected, no ompA or ompB PCR products were generated from this isolate, since the primer pairs used for these two genes are known not to work for R. bellii. Surprisingly, the fourth tick specimen from which we isolated rickettsiae was shown to be infected by both strain Atlantic rainforest and R. bellii. In this case, the infected cells were proved to be infected by strain Atlantic rainforest through the ompA PCR, which generated a product 100% identical to this strain (GQ855237). On the other hand, analysis of the nucleotide sequence from the gltA PCR amplification of this tick revealed dual electropherogram peaks at numerous base positions throughout the nucleotide sequence. Further analysis revealed that the superimposed sequences were consistent with a combination of the known R. bellii sequence and the known Atlantic rainforest sequence. Then, this isolate DNA was tested by a new PCR protocol using primers (forward: 5′-ATCCTGATTTGCTGAATTTTTT-3′; reverse: 5′-TGCAATACCAGTACTGACG-3′) that were designed to be specific for a 338-bp fragment of the R. bellii gltA gene. In this case, the sequence of the PCR product was 100% identical to R. bellii (CP000087), confirming that R. bellii was co-infecting the Vero cells together with strain Atlantic rainforest. In summary, the rickettsial stocks generated in the present study derived from A. ovale ticks, comprising 2 pure strain Atlantic rainforest isolates, 1 pure R. bellii isolate, and 1 mixed culture. These isolates and the mixed culture were deposited at the rickettsial collection of the Faculty of Veterinary Medicine of the University of São Paulo, Brazil. Partial sequences (gltA, ompB, ompA) from strain Atlantic rainforest, and partial gltA sequence of R. bellii (designated as strain Peruibe) generated in this study were deposited into GenBank and assigned nucleotide Accession nos. JQ906783-JQ906786, respectively.
Rickettsial DNA in ticks and hosts
The overall prevalence of rickettsial infection in A. ovale ticks from dogs was high (14·2%), with 30 ticks (12·9%) being infected by strain Atlantic rainforest, and 4 ticks (1·7%) infected by R. bellii (Table 5). One single tick was infected by both strain Atlantic rainforest and R. bellii, as stated above. The prevalence of the infection by strain Atlantic rainforest in A. ovale ticks from vegetation was much lower (6·3%). Strain Atlantic rainforest was also found in 2% of the R. sanguineus ticks collected from 2 dogs. One of these dogs also had a strain Atlantic rainforest-infected A. ovale tick, whereas the other one did not, although it was found to be infested by A. ovale ticks. The highest rickettsial infection rate (87·5%) was found for R. bellii among I. loricatus ticks parasitizing D. aurita.
Rickettsial infection in adult ticks collected from hosts and vegetation in the Atlantic rainforest Reserve ‘Estação Ecológica de Juréia-Itatins’ in Peruibe Municipality, state of São Paulo, Brazil, from April 2010 to February 2011
+ PCR of adult ticks that moulted from engorged nymphs collected on dogs.
* Strain Atlantic rainforest.
** One A. ovale tick was infected with both Rickettsia sp. strain Atlantic rainforest and R. bellii.
Rickettsial DNA was searched for in spleen and lungs from 15 tick-infested rodents (9 E. russatus, 4 Akodon spp., and 2 Oxymycterus spp.), and 3 marsupials (1 Micoureus demererounae and 2 Monodelphis sp.). No rickettsial DNA was found in the internal organs of these hosts.
DISCUSSION
In the present study the Atlantic rainforest spotted fever rickettsia, a human pathogen in Brazil, was isolated in cell culture for the first time. Previous phylogenetic analysis inferred from the genes gltA, ompA, ompB, sca4 and htrA (17 kDa protein) revealed that the Atlantic rainforest genotype (also referred to as strain Bahia) is closely related to the human pathogens R. parkeri, R. sibirica and R. africae (Spolidorio et al. Reference Spolidorio, Labruna, Mantovani, Brandão, Richtzenhain and Yoshinari2010; Silva et al. Reference Silva, Eremeeva, Rozental, Ribeiro, Paddock, Ramos, Favacho, Reis, Dasch, de Lemos and Ko2011). More recent works have suggested that these 3 Rickettsia species could just represent geographical variants of a single species (Walker and Ismail, Reference Walker and Ismail2008; Goddard, Reference Goddard2009; Pacheco et al. Reference Pacheco, Arzua, Nieri-Bastos, Moraes-Filho, Marcili, Richtzenhain, Barros-Battesti and Labruna2012). Until taxonomy of such group is clarified, Atlantic rainforest genotype is considered to be a distinct rickettsial genotype, pathogenic for humans.
The present survey detected rather high infection rates of canine A. ovale ticks with Atlantic rainforest genotype. A very high seroprevalence rate of dogs to SFG Rickettsia, with titres being highest against R. parkeri and Rickettsia sp. Atlantic rainforest genotype antigens, was also observed. Moreover, infected A. ovale ticks were found on most of the dogs that had access to the rainforest. In fact a highly significant correlation was observed between seropositivity of dogs, access to the rainforest and A. ovale infestation. It is thus certain that canine seropositivity is a result of infestations by A. ovale ticks acquired during incursions into the forest. Such observation is further strengthened by the detection of infected A. ovale ticks questing on vegetation of trails inside the forest.
In a previous study conducted in another Atlantic rainforest reserve of the state of São Paulo, Sabatini et al. (Reference Sabatini, Pinter, Nieri-Bastos, Marcili and Labruna2010) reported Atlantic rainforest genotype infecting 13·6% of the A. ovale ticks collected from dogs, and 8·8% of A. ovale ticks collected from vegetation. In the present study, a higher infection rate by the Atlantic rainforest genotype was also found amongst A. ovale ticks taken from dogs (12·9%) in comparison to those picked from vegetation (6·3%). These differences might be attributed to Rickettsia activation (Hayes and Burgdorfer, Reference Hayes and Burgdorfer1982), which results in higher ricketsial load in fed ticks and, consequently, in higher chances of detection by PCR. Although Rickettsia activation for Atlantic rainforest genotype is for now speculative, the mean annual temperature of the studied area is at least 17 °C below a dog's temperature. Thus it is reasonable to assume that Rickettsia replication is enhanced in ticks feeding on dogs. Whatever the case, determining whether the disease agent can be transmitted to humans by ticks questing on the vegetation but after several hours of parasitism or immediately from dog ticks that accidentally detached and may transmit the already re-activated Rickettsia is an important feature to be addressed by forthcoming research.
Alternatively, higher infection rates on feeding ticks could be related to horizontal transmission via dogs. Such horizontal transmission could be between co-feeding ticks on a rickettsemic dog, which is the classic route of horizontal transmission of several other Rickettsia species (Philip, Reference Philip1959; Piranda et al. Reference Piranda, Faccini, Pinter, Pacheco, Cançado and Labruna2011). However, horizontal transmission could also occur via copula in ticks, as reported for other Rickettsia and tick species (Hayes et al. Reference Hayes, Burgdorfer and Aeschlimann1980), or between co-feeding ticks on non-rickettsemic hosts (Zemtsova et al. Reference Zemtsova, Killmaster, Mumcuoglu and Levin2010); however, the significance of the role of this latter route appears to be much less efficient than it is for some tick-borne viruses (Labuda and Nuttall, Reference Labuda and Nuttall2004).
Horizontal transmission could also be the source of infection of the R. sanguineus ticks (2%) that were found to be infected by the Atlantic rainforest genotype in the present study, since these ticks co-infested dogs with A. ovale ticks. Interestingly, Sabatini et al. (Reference Sabatini, Pinter, Nieri-Bastos, Marcili and Labruna2010) also found a 1·9% infection rate among R. sanguineus taken from A. ovale-infested dogs. Anyhow, the R. sanguineus–strain Atlantic rainforest relationship is a matter of concern since dog parasitism by both tick species is a common feature and R. sanguineus is a nidicolous tick that completes its whole life cycle in human dwellings (Guglielmone et al. Reference Guglielmone, Szabó, Martins, Estrada-Peña, Barros-Battesti, Arzua and Bechara2006).
Even though rickettsia was not found in the internal organs of the rodent E. russatus, this animal had the highest prevalence and mean intensity of A. ovale nymphs infestation among the small vertebrates examined so far. Moreover this rodent also displayed a high seroprevalence rate with titres highest against R. parkeri. These findings suggest a role of this animal as an amplifier host for the Atlantic rainforest genotype to A. ovale ticks, a condition yet to be evaluated experimentally. On the other hand, even though dogs (the main host for the adult stage of A. ovale in the study region) are susceptible to rickettsial infection (as demonstrated by a high infection prevalence and very high serological endpoint titres to R. parkeri), only adult-stage A. ovale ticks feed on dogs. Because transovarial transmission rates are likely to be low or absent when the primary rickettsial infection of ticks occurs during adult feeding (Burgdorfer and Brinton, Reference Burgdorfer and Brinton1975; Piranda et al. Reference Piranda, Faccini, Pinter, Pacheco, Cançado and Labruna2011), the epidemiological influence of dogs as amplifier hosts of the Atlantic rainforest genotype for A. ovale ticks is questionable.
Our mixed cell culture of 2 Rickettsia species (Atlantic rainforest genotype and R. bellii) from a single A. ovale specimen is noteworthy. While multiple infection by 2 or more Rickettsia species have been only rarely reported in individual ticks, all previous reports were confirmed solely by PCR performed directly on tick DNA extracts. These previous reports include R. parkeri and ‘Candidatus R. andeanae’ infecting an A. maculatum specimen (Varela-Stokes et al. Reference Varela-Stokes, Paddock and Engber2011), R. rickettsii and R. amblyommii infecting an Amblyomma americanum specimen (Berrada et al. Reference Berrada, Goethert, Cunningham and Telford2011), R. bellii and R. rhipicephali infecting a Dermacentor variabilis specimen (Wikswo et al. Reference Wikswo, Hu, Dasch, Krueger, Arugay, Jones, Hess, Bennett, Kramer and Eremeeva2008), and a triple infection by R. bellii, Rickettsia montanensis and R. rickettsii in a D. variabilis specimen (Carmichael and Fuerst, Reference Carmichael and Fuerst2010). Therefore, our finding represents the first mixed culture of 2 Rickettsia species from a single naturally infected tick specimen.
In conclusion, the Atlantic rainforest genotype circulates at relatively high levels among A. ovale ticks in the focus of human rickettsiosis in Peruibe. This tick species attaches to dogs that enter the forest and are taken to households by the host. An accidental bite by these ticks falling from dogs or eventually from the vegetation may transmit the Atlantic rainforest genotype to human beings and cause a clinical infection. Several aspects of the maintenance and circulation of the pathogen inside the forest shall be highlighted by further field data coupled with experimental data involving host and parasite target species. Further, even though the relationship of the Atlantic rainforest genotype, A. ovale, and the Atlantic rainforest biome is clear, it would be interesting to explore a possible relation of A. ovale and Atlantic rainforest genotype in other biomes as well. This tick species has a wide range from Mexico to Argentina (Guglielmone et al. Reference Guglielmone, Estrada-Peña, Mangold, Barros-Battesti, Labruna, Martins, Venzal, Arzua and Keirans2003) including other biomes in Brazil such as the Cerrado (Szabó et al. Reference Szabó, Olegário and Santos2007) and Amazon (Labruna et al. Reference Labruna, Camargo, Terrassini, Ferreira, Schumaker and Camargo2005).
ACKNOWLEDGEMENTS
The authors are indebted to Instituto Florestal from Secretaria do Meio Ambiente of São Paulo State for permits to conduct this study.
FINANCIAL SUPPORT
This research was supported by Fundação de Amparo à Pesquisa do Estado de São Paulo – FAPESP (grant no. 2010/50113-8) and Conselho Nacional de Desenvolvimento Científico e Tecnológico – CNPq (fellowship to M.P.J.S. and Academic Career Research Fellowship to M.B.L.).
