Serologic evidence of the exposure of small mammals to spotted-fever Rickettsia and Rickettsia bellii in Minas Gerais , Brazil

Introduction: Sources of pathogenic Rickettsia in wildlife are largely unknown in Brazil. In this work, potential tick vectors and seroreactivity of small mammals against four spotted-fever group Rickettsia (R. rickettsii, R. parkeri, R. amblyommii and R. rhipicephali) and Rickettsia bellii from peri-urban areas of Uberlândia, a major town in Brazil, are described for the first time. Methodology: Small mammals were captured and blood samples collected. Ticks were collected from the surface of the host and the environment and posteriorly identified. Reactivity of small mammal sera to Rickettsia was tested by indirect immunofluorescence assay (IFA) using crude antigens from five Brazilian Rickettsia isolates. Results: Information was obtained from 416 small mammals (48 Marsupialia and 368 Rodentia). Forty-eight animals were parasitized and two tick species, Ixodes loricatus and Amblyomma dubitatum, were found on several host species, with a few tick-host relationships described for the first time. From the 416 tested sera, 70 reacted to at least one Rickettsia antigen (prevalence of 16.8%) and from these, 19 (27.1%) reacted to two or more antigens. Seroprevalence was higher for marsupials (39.6%) than for rodents (13.8%). Marsupial and Rhipidomys spp. sera reacted mainly (highest seroprevalence and titers) to R. bellii, and that of Necromys lasiurus mainly to R. rickettsii. Conclusions: Although the serologic assays poorly discriminate between closely related spotted-fever group Rickettsia, the observed small mammal seroreactivity suggests the circulation of Rickettsia in the peri-urban area of Uberlândia, albeit at low levels.


Introduction
Rickettsia spp.are intracellular Gram-negative bacteria.The genus comprises human and animal pathogens, with a few causing very severe, sometimes lethal disease.Spotted-fever Rickettsia is a group strongly associated with ticks as vectors and includes both pathogenic species such as Rickettsia rickettsii, R. conorii, R. parkeri, R. africae, as well as Rickettsia with undetermined pathogenicity.The world has witnessed the discovery of several Rickettsia species in the last two decades and description of new species can be expected [1].At the same time, the knowledge about the relationship between human disease and specific Rickettsia has also increased with the use of molecular diagnosis.Similarly, in Brazil, where only one spottedfever human rickettsial disease was known, a new disease caused by an R. parkeri-like agent was recently discovered [2,3].The range of this new agent is rather extensive along the Atlantic coast, and human infection may be hidden by other febrile diseases [4][5][6].These findings underscore the possible existence of other tickborne rickettsial agents in the country.
Brazil is a huge country with continental dimensions, and research of rickettsial disease is nonexistent in several locations, as is the case for Uberlândia, the second-most populous county of Minas Gerais state.Initial information of Rickettsia circulation on a given area may be provided by serology of some key hosts [7].Small mammals such as rodents are hosts for immature stages of several tick species and may amplify Rickettsia infection in tick populations; thus, they are an important study target for the epidemiology of rickettsiosis [8,9].We herein present tick species and a serosurvey of small mammals from peri-urban areas of Uberlândia to four spotted-fever Rickettsia found elsewhere in Brazil as well as Rickettsia bellii, a common Rickettsia in the country but with unknown pathogenicity [10].

Sampling location
Sampling was conducted in Uberlândia municipality (18°54′S; 48°15'W), Minas Gerais state, southeast Brazil.The city is the second-most populous of the state, with approximately 646,000 inhabitants.The region belongs to the Cerrado biome, a savannah, and has a tropical climate with two distinct seasons: a hot and rainy summer and a cool and dry winter.Small mammals were captured at nine locations surrounding the urban area.Locations were characterized by either native Cerrado, its fragments, or anthropogenically altered landscapes (Table 1).

Small mammal capture
Small mammals were captured from July 2011 to August 2012 along two to nine campaigns set by convenience at each of the nine sites distributed around the city.Overall sampling effort at each location is shown in Table 1.At locations 1 and 2, small Shermantype live traps (25 x 8 x 9 cm) one to two meters above ground on trees, and bigger ones (30 x 8 x 9 cm) on the ground, were distributed in grids for four consecutive nights at each campaign.In all other locations, Sherman (30 x 8 x 9 cm) and Tomahawk (45 x 16 x16 cm) live traps at a 3:1 proportion were placed in linear transects set at 10-to 20-meter intervals on the ground for five consecutive nights.Traps were baited daily with a mix of banana, peanut, and oat.Animals recaptured within the same campaign were immediately released and no sample was collected from these.Captured animals were anesthetized with ketamine cloridrate (Syntec, Santana de Parnaíba, Brazil) and acepromazine (Vetnil, Louveira, Brazil) and blood samples were collected.Animals were identified according to morphological characters, and in a sample, taxonomic identification was determined by karyotype analysis [11].Animals were marked with an ear code and released at the capture site after full recovery from anesthesia.Voucher rodent species were euthanized by deepening anesthesia, and carcasses were deposited at the National Museum of the Federal University of Rio de Janeiro (UFRJ).
All procedures were submitted to and approved by the ethics committee of the Federal University of Uberlândia (permit 071/11) and Brazilian Environment Institute (IBAMA) (permits 22629-1, 13373-1, and 10762-1).

Tick collection and identification
Hosts were thoroughly examined for ticks by gently rubbing fine forceps against the fur on the whole body surface.Located ticks were collected with the aid of forceps.Flat ticks were placed in alcohol, and engorged larvae and nymphs were kept alive for molting to the next stage at 27°C, 80% relative humidity, and 12:12 photoperiod.Samples of host-seeking ticks were collected from the vegetation by dragging at each location a white 1 x 2-meter cloth over 100 meters on animal trails as described before [12].Dragging was performed always in the morning (between 8:00 and 10:00 a.m.) and repeated in at least three different months in each location to sample during both dry (May, June, and August) and rainy (December, January, February, and March) seasons.The dragging time schedule was intended to avoid both morning dew that would soak the cloth and periods of the day with direct exposure of vegetation to sun (when ticks descend to lower levels, making them unavailable to drag sampling).Both rainy and dry seasons were sampled because Amblyomma ticks in Brazil have a one-year life cycle, with adults prevailing in the rainy season and immatures in the dry one [13].Collected ticks were stored in alcohol and identified based on dichotomous keys for adults and Amblyomma nymphs [14][15][16].Lack of keys for Neotropical Amblyomma larvae and immatures of Ixodes as well as damaged samples precluded identification of several ticks retained as either Amblyomma spp. or Ixodes spp.Several engorged Ixodes immatures were identified after molting to the next stage, and others by comparison to descriptions by Marques et al. and/or reference specimens kept in alcohol [17].

Serology for Rickettsia
Reactivity of small mammal sera to Rickettsia was tested by indirect immunofluorescence assay (IFA) using crude antigens from five Brazilian Rickettsia isolates (R. bellii strain Mogi, R. amblyommii strain Ac37, R. rhipicephali strain HJ5, R. rickettsii strain Taiaçu, and R. parkeri strain At24) available at the Faculty of Veterinary Medicine of the University of São Paulo, as previously described [7,18,19].Briefly, starting from the 1:64 dilutions, sera were diluted in twofold increments with phosphate-buffered saline (PBS), pH 7.4, until the last reactive dilution.Slides were incubated with fluorescein isothiocyanatelabelled goat anti-mouse IgG (Sigma, St. Louis, USA), goat anti-rat IgG (Sigma), or sheep anti-opossum IgG (CCZ, São Paulo, Brazil).For each sample, the endpoint IgG titer reacting with each of the five Rickettsia antigens was determined.Reaction was considered homologous when an endpoint titer to a Rickettsia species was at least fourfold higher than those observed for all other Rickettsia species.In this case, the Rickettsia species or a closely related species was considered the possible antigen involved in a homologous reaction.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 dilutions, as previously reported [6].

Ticks collected
Information on tick infestation of small mammals is presented in Tables 2 and 3. From the 416 captured small mammals, 48 were infested by 135 ticks (prevalence: 11.5%; mean intensity [mean number of ticks per infested host]: 2.8 ticks).Only two tick species (Amblyomma dubitatum and Ixodes loricatus) were found on animals, but other species among unidentified ticks (Amblyomma spp.or Ixodes spp.) cannot be ruled out.Ticks found were overwhelmingly in immature stages (larvae and nymphs).In fact, only two adults of I. loricatus were found and solely on D. albiventris.
Dragging on vegetation yielded ticks from five locations (locations 1, 2, 3, 8, and 9) and included 29 Amblyomma spp.larvae as well as 9 nymphs and 10 adults of Ambyomma cajennense sensu lato [20].All larvae and nymphs were collected in the dry season and most adults in the rainy season.

Seroreactivity to Rickettsia
From the 416 tested sera, 70 reacted to at least one Rickettsia antigen (16.8% overall prevalence) (Table 4).Seroprevalence was higher for marsupials (39.6%, 19 reactive animals from 48) than for rodents (13.8%, 51 reactive animals from 368).With the exception of location 8, all other locations had animals with seroreactivity to Rickettsia spp.However, only three animals were captured in this residential park (location 8).
Forty-six (65.7%), eight (11.4%),five (7.1%), and six (8.6%) serum samples were reactive to, respectively, one, two, three or five, and four species of Rickettsia.Homologous reaction (endpoint titer to a Rickettsia species at least fourfold higher than those observed for the other Rickettsia species) to R. rickettsii, R. parkeri, R. rhipicephali, and R. bellii was detected in, respectively, nine, four, four, and twenty animals (Table 4).Marsupials and Rhipidomys spp.reacted overwhelmingly to R. bellii, whereas N. lasiurus reacted to R. rickettsii.Endpoint titers of small mammals ranged from 64 to 1,024 against R. amblyommii and R. bellii, and 64 to 2,048 against R. rickettsii, R. parkeri, and R. rhipicephali.Two Rhipidomys spp.were the hosts that exhibited the highest titers to all five antigens; all animals from these genera were captured at location 1 (Glória farm).

Discussion
With the exception of two adult ticks on D. albiventris, ticks of solely immature stages were found on small mammals.Previous reports on ticks of small mammals from various regions of Brazil showed a similar tendency [6,19,21,22].Two tick species, A. dubitatum and I. loricatus, were recovered from small mammals, whereas only A. cajennense sensu lato, both nymphs and adults, were found in the environment.This apparent disconnection of environmental and host infestation may be explained by limitations of the sampling method of the environment, differing microenvironmental preferences of the host and parasite, and/or the lack of attraction of these small mammal species to specific ticks.In fact, A. cajennense sensu lato in Brazil are known to feed mainly on large mammals such as horses and capybaras in the adult stage [23].Further, samplings by cloth dragging, which were restricted to mornings, as well as temperature variations may have caused sampling bias with reflections on tick species and numbers collected.
Irrespective of this disconnection between ticks from the environment and on small mammals, rodents are a common source of microorganisms for arthropods worldwide and thus may contribute to the transmission of several infectious agents to other hosts during the tick's feeding and in the subsequent developmental stages of the tick [24].For this reason, rodent-tick associations herein described are evaluated with care.
Finding A. dubitatum nymphs on all tick-infested mammals with the exception of G. agilis was a noteworthy observation.In fact, A. dubitatum on C. marinhus, H. megacephalus, N. lasiurus, and Rhipidomys spp.is, to our knowledge, the first report of these host-tick relationships.This tick species (A.dubitatum = A. cooperi) [25] has been shown to harbor R. bellii, a species with unknown pathogenicity, and at least two other undefined Rickettsia species [26][27][28].At the same time, it is a tick species with strong association in all parasitic stages with capybaras (Hydrochoerus hydrochaeris) [29].Capybara, the biggest rodent in the world, also withstands high A. cajennense infestations, and this host and this tick together provide the major epidemiological background for Brazilian spotted fever (BSF) in southeast Brazil [30].Nonetheless, the initial Rickettsia source for ticks and capybaras is still undetermined [23].In fact, even though BSF has not been diagnosed in Uberlândia region, capybaras are common and were seen in several of the studied locations.Thus, infestations of A. dubitatum on small mammals and later on capybaras together with A. cajennense should be investigated further in endemic areas as a possible bridge for R. rickettsii.
Ixodes loricatus is a tick species common on New World Marsupials but found on Sigmodontinae rodents as well [25,31].This tick species frequently harbors R. bellii, and both ticks and opossums (Didelphis spp.) are common in BSF-endemic areas [6,32].There are two reports of R. rickettsii isolation from opossums from the states of São Paulo and Minas Gerais, Brazil [33,34].Thus, I. loricatus is another tick species that deserves attention as a potential pathogenic Rickettsia vector among small mammals.A homologous reaction was determined when an endpoint titer to a Rickettsia species was at least fourfold higher than to those observed for the other Rickettsia species.In this case, the Rickettsia species involved in the highest endpoint titer was considered the possible antigen involved in a homologous reaction (PAIHR); b NHR: non-homologous reaction Finding I. loricatus and A. dubitatum on small rodents and marsupials, as seen in our work, indicates the possibility of ticks feeding on these different hosts throughout their life cycles, with disease agent transmission from one to the other.In this regard, it is worthwhile to mention that it was previously observed that opossums from both endemic and non-endemic areas displayed titers to R. rickettsii at least fourfold higher than to any of the other rickettsial antigens [6,32].Moreover, it was shown that, under laboratory conditions, R. rickettsii was capable of infecting opossums without causing illness [34].In these experiments, animals developed a rickettsemia capable of causing infection in guinea pigs and ticks, although the infection rate in ticks was low.
Overall, serological prevalence of small mammals to the five Rickettsia antigens from our work (16.8%) was low, but within the wide range described in the country.However, due to the lack of antigens until a few years ago, there are only few reports on small mammal sera reactivity to several Rickettsia species in Brazil [6,19,21,32,36].Unfortunately, these reports are not enough to establish a standard serological profile for circulation of Rickettsia species in small mammals.At the same time, low prevalence of seroreactivity of small rodents, as seen in our work, must be evaluated with caution.Szabó et al., for example, observed an almost absolute seroconversion of dogs to Rickettsia spp.Atlantic rainforest strain, agent of spotted-fever human rickettsioses in Brazil, at forest sites that harbor infected Amblyomma ovale ticks, whereas seroprevalence was half in the case of Euryorysomys russatus, the small rodent that feeds A. ovale juveniles [6].Although other factors may be involved, it is possible that dogs, by crossing longer distances, enhance the chance of encounter with infected ticks.Thus, at endemic sites, seroprevalence of small mammal population with restricted home range might be decreased in comparison to other animals even if such hosts have an important role in disease epidemiology.
In our case, a probable homologous reaction of small mammals was found against four of the five Rickettsia antigens, overwhelmingly against R. bellii but also against R. rickettsii, R. parkeri, and R. rhipicephali.Rickettsia bellii has been found in several tick species in the country, attaining high prevalence in both I. loricatus and A. dubitatum ticks, which may explain our serological results [6,10,27,32,37].Reasons for the other probable homologous reactions, however, are less straightforward and deserve further investigation, particularly in the case of R. rickettsii.The aforementioned Rickettsia species may be circulating but in lower levels, or serologic results may reflect cross-reactivity with closely related Rickettsia from the spotted-fever group and to which small mammals were exposed to.
Non-homologous reactions to antigens from several Rickettsia species from our work may be a result of infection with unknown Rickettsia species and/or mixed infections.In fact, antigenic cross-reactivity among Rickettsia of different species occurs [38].In this regard, it is important to note that several Rickettsia species may be present at one location and even in the same tick, and thus animals may be exposed to more than one species simultaneously or over time with an unknown effect on seroreactivity [6].
In the few existing reports from Brazil, small mammal seroreactivity to Rickettsia and BSF relationship is not very clear.Reactivity was observed in BSF-endemic areas, in areas of silent focus, and also in apparently non-endemic areas [19,21,36].In this regard, it was observed in the state of São Paulo, Brazil, in areas with or without a history of recent confirmed cases of human BSF, that rather than prevalence of opossum serum reactivity, the height of titers was related to endemicity [32].In our case, overall titers of small mammals were low, with only two Rhipidomys displaying higher titers, but none of them exhibiting a homologous reaction, suggesting a lack of BSF endemicity or its restriction to specific, non-sampled spots.

Conclusions
The relatively low prevalence of small mammal seroreactivity and the observed homologous reactions suggest that in the peri-urban area of Uberlândia, there is a circulation of Rickettsia, albeit at low levels.Among these, R. bellii and probably one or more spotted-fever group Rickettsia species may be present.In this context, apart from opossums and I. loricatus, both Rhipidomys spp.and A. dubitatum deserve further investigation as well as molecular identification of Rickettsia in ticks.Last but not least, a follow-up is mandatory as endemicity may vary over time for reasons yet unknown.
(scholarship to M.G.Coelho).The authors acknowledge Zoonosis Control Center of Uberlândia and Laboratório de Ecologia de Mamíferos (LEMA), Universidade Federal de Uberlândia for logistic collaboration.The authors are indebted to Laboratório de Biologia e Parasitologia de Mamíferos Reservatórios Silvestres do Instituto Oswaldo Cruz (FIOCRUZ) for the identification of several small mammals.

Table 1 . Location, vegetation type, and trapping effort of small mammal capture from July of 2011 to August 2012 in the peri- urban areas of Uberlândia, State of Minas Gerais, southeastern Brazil Location Coordinates Vegetation Trapping effort (trap/night)
bValley-side marshes where the water table reaches or almost reaches the surface during the rainy season

Table 2 .
Prevalence and mean intensity of tick infestation of small mammals captured from July 2011 to August 2012 in the peri-urban areas of Uberlândia, State of Minas Gerais, southeastern Brazil Rhipidomys spp.).The number of individuals captured from each species of Rodentia is shown in Table2.The capture of 416 small mammals included one D. albiventris captured three times and a G. agilis and two Hylaeamys megacephalus captured twice, all in different campaigns.
a Mean number of ticks per infested animal

Table 3 .
Tick species found on small mammals captured from July 2011 to August 2012 in the peri-urban areas of Uberlândia, State of Minas Gerais, southeastern Brazil

Table 4 .
Homologous and non-homologous seroreactivity to Rickettsia species of small mammals captured from July 2011 to August 2012 in the peri-urban areas of Uberlândia, State of Minas Gerais, southeastern Brazil