Domestic, peridomestic and wild hosts in the transmission of Trypanosoma cruzi in the Caatinga area colonised by Triatoma brasiliensis.

The role played by different mammal species in the maintenance of Trypanosoma cruzi is not constant and varies in time and place. This study aimed to characterise the importance of domestic, wild and peridomestic hosts in the transmission of T. cruzi in Tauá, state of Ceará, Caatinga area, Brazil, with an emphasis on those environments colonised by Triatoma brasiliensis. Direct parasitological examinations were performed on insects and mammals, serologic tests were performed on household and outdoor mammals and multiplex polymerase chain reaction was used on wild mammals. Cytochrome b was used as a food source for wild insects. The serum prevalence in dogs was 38% (20/53), while in pigs it was 6% (2/34). The percentages of the most abundantly infected wild animals were as follows: Thrichomys laurentius 74% (83/112) and Kerodon rupestris 10% (11/112). Of the 749 triatomines collected in the household research, 49.3% (369/749) were positive for T. brasiliensis, while 6.8% were infected with T. cruzi (25/369). In captured animals, T. brasiliensis shares a natural environment with T. laurentius, K. rupestris, Didelphis albiventris, Monodelphis domestica, Galea spixii, Wiedomys pyrrhorhinos, Conepatus semistriatus and Mus musculus. In animals identified via their food source, T. brasiliensis shares a natural environment with G. spixii, K. rupestris, Capra hircus, Gallus gallus, Tropidurus oreadicus and Tupinambis merianae. The high prevalence of T. cruzi in household and peridomiciliar animals reinforces the narrow relationship between the enzootic cycle and humans in environments with T. brasiliensis and characterises it as ubiquitous.

Trypanosoma cruzi Chagas, 1909 (Trypanosomatida: Trypanosomatidae) (Moreira et al. 2002), the etiologic agent of Chagas disease, is a obligate protozoan parasite that is capable of infecting dozens of species of triatomine vectors (Hemiptera: Reduviidae: Triatominae) (Silveira & Rezende 1994, OPAS 2006) and hundreds of species of mammals belonging to over 70 types (Noireau et al. 2009, Zingales et al. 2012).

Infection is characterised by a series of very complex elements, which include bio-ecological determinants (relationship between vectors, parasites and reservoirs), cultural factors (types of homes and way of life) and socioeconomic factors (occupation of territories and labour relations). Human infection is the result of direct action of man over nature, occurring when people invade the enzootic cycle of T. cruzi (Dias 2000, Noireau et al. 2009).

The different strategies of parasite transmission between mammals and vectors with different biological and ecological characteristics have resulted in the maintenance of parasites in distinct cycles. This restriction, in turn, has resulted in the evolutionary transformation from foci of enzootic transmission in restricted habitats to complex networks interconnected by domestic and wild species with a wide geographical distribution (Araújo et al. 2009).

The roles of the different species of mammals in the maintenance of the parasite are not constant according to time or region (Ashford 1996, 1997). Despite the long list of mammalian species that are naturally infected by T. cruzi, the role that each reservoir species plays in the dispersal and/or maintenance of the parasite can be extremely variable in distinct transmission cycles of T. cruzi in nature. This variation is mainly due to the complexity of these processes, ecological inter-relationships and the high speed with which man modifies environments (Noireau et al. 2005, Ceballos et al. 2006, Jansen 2009).

The Brazilian Northeast Region plays an important role in the national context of the epidemiology of Chagas disease (Dias 2000). The Caatinga region is ranked third in triatomine diversity, containing 15 (24%) of the described species (Gurgel-Gonçalves et al. 2012). Triatoma brasiliensis Neiva, 1911, is the species of triatomine of greatest importance in the domestic transmission of Chagas disease in the region, thus its control is a priority (Silveira et al. 1984, Dias et al. 2000). Despite the fact that the household population can be controlled through the use of insecticides with residual action, homes are recolonised at high speeds, requiring permanent vigilance against new outbreaks (Diotaiuti et al. 2000, Borges et al. 2005).

It is known that the adaptation of triatomines to artificial ecotopes is what determines their condition as vectors of epidemiological importance of T. cruzi. This ability depends primarily on two features: the ability to feed on the sources of blood available in the house and the ability to survive the microclimatic factors (temperature and humidity) determined by household hideouts (Lent & Wigodzinsky 1979, Lorenzo et al. 2000). In addition, man-made environmental changes favour the dispersion of triatomines (Forattini 1980) and the attraction of insects to household by light sources (Carbajal-de-la-Fuente et al. 2007). For T. brasiliensis, it has been shown that the adaptation of these insects to the intradomicile lifestyle is favoured by microclimatic similarities between households and natural ecotypes of the species in the state of Ceará (CE) (Lorenzo et al. 2000).

This work aims to characterise the importance of domestic, wild and peridomestic hosts in the transmission of T. cruzi in a rural area of CE, with an emphasis on environments colonised by T. brasiliensis.

MATERIALS AND METHODS

Research area - The study was conducted in the municipality of Tauá (CE) (Fig. 1), a region that has historically presented a near totality of homes infested by triatomines, with T. brasiliensis being the main vector species. Eighteen sites distributed throughout the district of Carrapateiras and totalling 252 households (UDs) were selected. In this case, the term “UD” refers to the peridomestic and domestic environments, that is, human habitations and their surroundings, with all permanent and temporary buildings, accumulations of materials, fences, animal shelters etc.

Fig. 1

: geographical location of the study sites (Mutuca and Cachoeira do Júlio, highlighted in red), municipality of Tauá, state of Ceará. Source: alunosonline.com.br/geografia/ceara.html (accessed in 4 June 2009) and archives of the Chagas Disease Control Program of Health Department of the State of Ceará.

Tauá is located in the hinterland of Inhamuns (6º00’11”N 40º17’34”S) at an altitude of 402.7 m, 320 km from the capital, Fortaleza. The average temperature varies between 26-28ºC, with an average rainfall of 597.2 mm3, with a rainy season from February-April (IPECE 2012). In this region, the vegetation of the savannah is highly degraded, with physiognomic patterns marked by the secondary succession of predominant shrub and tree savannah. The predominant vegetation is deciduous and garranchenta, growing in shallow and stony soils, with an extreme water deficit for much of the year (Oliveira 2006). The economic activities are restricted to agricultural and extensive cattle activities, with very low incomes and negligible productivity (Oliveira 2006). In this context, crystalline basement rocks, mainly granitic rocks with significant fractures, provide habitats for small mammals, reptiles and insects. Amongst them is T. brasiliensis, for which this environment is its natural habitat (Forattini 1980).

Mammal capture - Wild - Rock formations that could be natural shelters for T. brasiliensis are identified with the presence of small wild animals such as rodents and marsupials, a fact confirmed by an active search for traces of these insects and mammals. Three areas were selected: two in Mutuca, that were 2.6 km apart each other and identified as: Pedra da Cruz (MPC) (adjacent to residences - 94 m and more susceptible to human intervention) and Seu Evangelista (ME) (with intermediate distance to the residences - 211 m and less susceptible to human intervention) and Cachoeira do Júlio (CJ) (distant residences - 370 m and virtually no human intervention at 5.4 km away from MPC). In these places, the traps were placed in transects. The traps were latticework traps with trigger hooks in small (33 x 14 x 10 cm - 250 units), medium (33 x 12 x 14 cm - 46 units), large (46 x 16 x 19 cm - 11 units) and large tomahawk (50 x 21, 5 x 20 cm - 9 units) sizes, suitable for catching animals of different sizes and weights. The traps remained in the field for 16 days and were deployed in four distinct campaigns (February/2009, August/2009, February/2010 and August/2010) of four days and nights each. The periods chosen reflect the maximum drought period (February) and the immediate following period (August), in which we would expect to observe the behaviour and population structure of different animals.

The bait used was a mixture of bacon, oatmeal, bananas, peanut butter and corn on the cob (D’Andrea et al. 2006). The traps were baited in the evening and were revisited at dawn. Lactating or cub-bearing marsupials were not sampled. These and other non-rodents were returned to the wild at the capture site after the experimental procedure.

For convenience, sampling was defined by the total trap catch per day of effort during the period of exposure (sum of traps by type multiplied by the exposure period) (Torres et al. 2008). The ultimate goal was to search for a general representation of fauna infected with T. cruzi, whether the animals were recaptured or not.

The captured animals were anesthetised with a combination of ketamine and xylazine intramuscularly with 1 mL syringes. The following data were collected: species type, location of capture, sex and age (adult/juvenile). In addition, blood samples were collected by puncturing the tail vein and collecting a maximum of 1 mL of blood from rodents and 3 mL from marsupials, considering the weight of each animal.

Domestic and peridomestic - The domestic and peridomestic sample was defined as including all dogs, cats and pigs encountered during the visits to residential localities of Mutuca and CJ. For ovine and caprine specimens, the herd records available in the Agency of Agricultural Defense of Ceará State were used to define sampling based on finite and known prevalence populations, separated by animal species, sex and age (individuals less than 6 months of age were classified as young and those older than 6 months as adult). For this study, only animals under six months of age were used in attempts to identify recent caprine contact with T. cruzi.

In the domestic and peridomestic environments, the collection of blood from dogs and cats was performed by femoral vein puncture, while for pigs, sheep and goats it was performed by the jugular vein puncture. The same data were collected as described above for wild animals, including the environment frequented by the animals (wild, peridomestic and indoor) and how long they had lived in that house or its surroundings. All procedures were performed under the guidance of responsible professionals and followed the general recommendations as to the integrity and well-being of the animals and signed consent forms were obtained from the owners or guardians.

Diagnosis of infection with T. cruzi - Parasitological - Two sheets of thick films were prepared by using mammalian specimens and stained with Giemsa stains. One hundred fields per drop were observed by optical microscopy with a magnification of 1,000X (FIOCRUZ 2008).

Serological - Three methods of serological diagnosis were used: immunofluorescence assay (IFA), according to Camargo (1966), immunoenzymatic assay (ELISA indirect) and immunochromatographic rapid test and dual path platform test (RT-DPP).

The reactions were developed with IgG conjugated to fluorescein isothiocyanate (Sigma). For differential diagnosis, all sera were tested for infection with Leishmania spp by the IFA method and for Leishmania infantum using the quick test for diagnosis of canine visceral leishmaniasis (CVL) (DPP) specific Bio-Manguinhos for dogs (MS/SVS 2006, 2011).

In the IFA, the cut-off values for serum titres in domestic mammals were ≥ 1:40 for dogs and cats and > 1:40 for the other species. The results were expressed as the highest serum dilution at which specific fluorescence is still observed. The reaction reading was taken using a fluorescence compound microscope with a high intensity light source (ultraviolet light). The cut-off for the ELISA was established by taking the average of the optical density of the negative control and adding 20% to that value. The quantification of the reaction was determined by spectrophotometer readings at a wavelength of 450 nm. The data were only considered suspicious or ambiguous in one method and in reactions with dilution cut-offs. With the Rapid Test on DPP® LVC, we followed the manufacturers’ instruction accompanying the CVL-Bio-Manguinhos RT KIT DPP®. All reactions included two positive controls and two negative controls. The prevalence of infection was calculated by obtaining the percentage of individuals who exhibited positive results in both serological tests.

Molecular - Polymerase chain reaction (PCR) multiplex - Whole blood samples were placed on filter paper (Whatman Nº 1), dried, wrapped in aluminium foil and isofilm and stored individually at -20ºC. Samples were subjected to multiplex PCR for the diagnosis and characterisation of T. cruzi in TcI or TCII.

DNA was extracted from the samples on filter paper was extracted from 1 mm punched samples in individual microtubes. In each tube, we added 100 µL of deionised water and the sample was heated at 70ºC for 10 min and centrifuged for 3 min at 17,900 g (Machado et al. 2000). The supernatant containing the DNA was stored at 4ºC and used for PCR.

As a positive control, we used the default Colombian and Y strains belonging to groups TcI and TCII, respectively. After PCR, 3 µL of the product was visualised on 6% polyacrylamide gel and stained with 0.2% silver nitrate to reveal the presence of specific bands.

Characterisation of the infestation and natural infection with T. cruzi - All households in the selected locations were manually searched by agents for endemic infestation in a thorough manner using the Manual of Technical Standards Campaign for the Control of Chagas Disease of the Ministry of Health (MS/SUCAM 1980). All surfaces, internal and external walls, furniture, utensils and other miscellaneous objects (MS/SUCAM 1980) were observed. In addition to this routine information, we noted the specific capture location of any triatomines. The intradomicile location was considered to be a single location, while the peridomicile location was divided into multiple types of attachments: chicken coops, sties, barns, firewood, stones, tiles and bricks and other.

Indices of infestation (% of positive UD’s for triatomines in relation to surveyed) of natural infection (% of triatomines infected with T. cruzi in relation to those examined) and of colonisation (% of positive houses with nymphs inside the home in relation to defined number of positive houses in households) were calculated.

The field research for the wild environments of triatomines was performed manually at dusk, with the aid of lanterns, in the same places where mammals were captured in the February/2009 and August/2009 campaigns for four consecutive nights. The insects were separated by location for later identification of species, developmental stage, T. cruzi infection status and identification of dietary sources.

Identification of dietary source - Cytochrome b (Cytb) - The identification of the food source was performed specifically to complement the list of wild animals that share the environment with T. brasiliensis.

Total DNA was extracted from the abdomen of insects (previously preserved in 70% alcohol) using the HotShot protocol (Truett et al. 2000). We used the primers L14841 5’-AAAAAGCTTCCATCCAACATCTCAGCATGATGAAA-3’ and H151495’-AAACTGCAGCCCCTCAGAATGATATTTGTCCTCA 3’; these universal primers are designed for vertebrates (Kocher et al. 1989). They amplify a 305 bp fragment and do not amplify any triatomine DNA present in the animal tissue.

The PCR was performed in a final volume of 25 μL, containing 40-50 ng of genomic DNA, 2.5 μL of 10X buffer, 2.0 μL of 2.5 mM dNTP, 0.75 μL of 50 mM MgCl2, 2.5 mL of each primer at a final concentration of 10 pmol and 0.2 μL of Taq polymerase 0.5 U/μL (Invitrogen). For each PCR reaction, a negative control (no DNA) was run in parallel.

The amplifications were performed using a thermal cycler (Eppendorf Mastercycler) with the following conditions for the reaction sequences: initial denaturation 95ºC for 5 min, 95ºC for 30 s, primer annealing at 58ºC for 30 s, extension at 72ºC for 1 min, return to step 2 and repeat 35 times and a final extension at 72ºC for 6 min. The amplified products were observed in 8% polyacrylamide gel stained with 0.2% silver nitrate.

For samples found to be positive by PCR, the products were purified using the QIAquick PCR Purification Kit (Qiagen) according to the manufacturer’s protocol. After purification, the quality and concentration of DNA were measured by spectrophotometry using the Nanodrop ND 1000 and the sample purity was determined by their 260/280 nm absorbance ratio, followed by sequencing (ABI 3730xl DNA; Applied Biosystems).

To identify the host species associated with triatomine, the sequences obtained were compared with sequences deposited in GenBank using the BLASTN search.

E thics - The project was submitted to the Animal Ethical Committee of the Federal University of Ceará (protocol 103, October 2011) and approved by the Chico Mendes Institute for Biodiversity Conservation of the Ministry of Environment through the Biodiversity Authorization and Information System (SISBIO) (case 31693-1, authentication code 46619742).

RESULTS

Parasitological diagnosis - The parasitological diagnosis by thick smear was negative for all 317 domestic and peridomestic animals tested (106 ovine, 83 caprine, 53 dogs, 41 cats and 34 porcine) as well as for all 112 wild animals tested (83 Thrichomys laurentius, 11 Kerodon rupestris, 5 Rattus rattus, 4 Didelphis albiventris, 3 Monodelphis domestica, 2 Galea spixii, 2 Wiedomys pyrrhorhinos, 1 Conepatus semistriatus and 1 Mus musculus).

Serological and molecular diagnosis of infection with T. cruzi - Serology of domestic mammals - In the 53 dog samples investigated for infection with T. cruzi, 74% (39/53) and 85% (45/53) were seropositive in ELISA and IFA techniques, respectively (Tables I, II). For the 41 samples from cats, tests for anti-IgG T. cruzi showed that 51% (21/41) had an IFA result of ≥ 1:40. Table I shows the diagnoses of T. cruzi, Leishmania spp and L. infantum determined by different techniques: IFA, ELISA, RT-DPP LVC and PCR multiplex in domestic mammals, according to age group. Among the dogs positive for T. cruzi, ages ranged from eight months to 15 years, with an average of three years.

TABLE I

Abundance of capture and seroprevalence of natural infection with Trypanosoma cruzi, Leishmania spp and Leishmania infantum in peridomestic and domestic mammals in two locations in the city of Tauá, state of Ceará, 2012

Species	Abundance n (%)	Mutuca n/n (%)					Cachoeira do Júlio n/n (%)					Total n/n (%)
		ELISA T. cruzi	IFA T. cruzi	IFA Leishmania spp	ELISA LVC	RT-DPP	ELISA T. cruzi	IFA T. cruzi	IFA Leishmania spp	ELISA LVC	RT-DPP	ELISA T. cruzi	IFA T. cruzi	IFA Leishmania spp	ELISA LVC	RT-DPP
Ovis aries	106 (33.4)	-	0/50	-	-	NA	-	1/52	-	-	NA	-	1/102 (1)	-	-	NA
Capra aegagrus hircus	83 (26.2)	-	0/34	-	-	NA	-	0/49	-	-	NA	-	0/83 (0)	-	-	NA
Canis familiaris	53 (16.8)	31/41	35/41	25/41	17/41	7/41	8/12	10/12	6/12	4/12	5/12	39/53 (74)	45/53 (85)	31/53 (58)	21/53 (40)	11/53 (21)
Felis catus	41 (13)	-	18/34	16/34	-	NA	-	3/7	6/7	-	NA	-	21/41 (51)	22/41 (54)	-	NA
Sus domesticus	34 (10.6)	-	1/22	-	-	NA	-	1/12	-	-	NA	-	2/34 (6)	-	-	NA
Total	317 (100)	31/41 (76)	54/181 (30)	41/75 (55)	17/41 (41)	7/41 (17)	8/12 (67)	15/132 (11)	12/19 (63)	4/12 (33)	5/12 (42)	39/53 (74)	69/313 (22)	53/94 (56)	21/53 (40)	11/53 (21)

IFA: indirect immunofluorescence assay; Leishmania spp: Leishmania braziliensis and L. infantum; LVC: canine visceral leishmaniasis; NA: does not apply; RT-DPP: immunochromatographic rapid test/dual path platform test.

TABLE II

Diagnosis of Trypanosoma cruzi, Leishmania spp and Leishmania infantum determined by techniques of indirect immunofluorescence assay, ELISA, immunochromatographic rapid test/dual path platform test-canine visceral leishmaniasis and polymerase chain reaction PCR multiplex in domestic mammals, according to age group in the city of Tauá, state of Ceará, 2012

Species	Age group (years)	T. cruzi n/n (%)	Leishmania spp n/n (%)	L. infantum n/n (%)	T. cruzi and Leishmania spp n/n (%)	T. cruzi and L. infantum n/n (%)	T. cruzi (title = 1:40) n/n (%)	PCR multiplex (TcI) n/n (%)
Dog (Canis f amiliaris)	0-1	2/16 (12.5)	0/16 (0)	1/16 (6.2)	2/16 (12.5)	1/16 (6.2)	1/16 (6.2)	4/9 (44.4)
	1-3	9/20 (45)	1/20 (5)	0/20 (0)	4/20 (20)	4/20 (20)	0/20 (0)	3/18 (16.6)
	3-15	9/17 (53)	1/17 (6)	0/17 (0)	4/17 (23.5)	2/17 (12)	0/17 (0)	4/16 (25)
	Total	20/53 (38)	2/53 (4)	1/53 (2)	10/53 (19)	7/53 (13)	1/53 (2)	11/43 (25.6)
Cat (Felis catus)	0-1	0/14 (0)	0/14 (0)	NA	5/14 (36)	NA	2/14 (14)	0/9 (0)
	1-3	1/16 (6)	2/16 (12.5)	NA	3/16 (19)	NA	1/16 (6.2)	0/8 (0)
	3-15	0/11 (0)	5/11 (45)	NA	1/11 (9)	NA	2/11 (18)	0/6 (0)
	Total	1/41 (2.4)	7/41 (17)	NA	9/41 (22)	NA	5/41 (12)	0/23 (0)

Leishmania spp: Leishmania braziliensis and L. infantum; NA: does not apply.

Serology of peridomestic mammals - The diagnosis of natural infection by T. cruzi in sheep, goats and pigs, as determined by multiplex PCR and IFA. The positive sheep come from CJ and most suspect animals [59.4% (19/32)] are from Mutuca. Among the caprine suspects, 64% (7/11) were from CJ. Pigs positive for T. cruzi were a female of 10 months from CJ and a year-old male from Mutuca.

Wild mammals - The success of the total catch was 9.6%, with an overall effort of 1,165 trap-nights in four stages and three locations. The numbers of species collected in MPC, ME and CJ were five, five and four, respectively. Among rodents and marsupials, a total of nine species were captured, corresponding to 112 samples (Table III).

TABLE III

Abundance of wild mammals and synanthropic captured by place and time in the city of Tauá, state of Ceará, 2009 and 2010

Species	Beginning of rain n/n (%)				Beginning of drought n/n (%)				Abundance
	MPC	ME	CJ	Total	MPC	ME	CJ	Total
Thrichomys laurentius	11 (61)	18 (75)	12 (100)	41/83 (49.4)	17 (81)	8 (53)	17 (77.3)	42/83 (50.6)	83 (74)
Kerodon rupestris	3 (17)	2 (8.3)	-	5/11 (45.4)	4 (19)	-	2 (9.1)	6/11 (54.5)	11 (10)
Rattus rattus a	-	2 (8.3)	-	2/5 (40)	-	3 (20)	-	3/5 (60)	5 (4.5)
Didelphis albiventris	-	1 (4.2)	-	1/4 (25)	-	3 (20)	-	3/4 (75)	4 (3.5)
Monodelphis domestica	1 (5.5)	-	-	1/3 (33.3)	-	1 (7)	1 (4.5)	2/3 (66.7)	3 (2.6)
Galea spixii	2 (11)	-	-	2/2 (100)	-	-	-	0/2 (0)	2 (1.8)
Wiedomys pyrrhorinos	-	-	-	0/2 (0)	-	-	2 (9.1)	2/2 (100)	2 (1.8)
Conepatus semistriatus	1 (5.5)	-	-	1/1 (100)	-	-	-	0/1 (0)	1 (0.9)
Mus musculus b	-	1 (4.2)	-	1/1 (100)	-	-	-	0/1 (0)	1 (0.9)
Total	18/54 (33.3)	24/54 (44.4)	12/54 (22.3)	54 /112 (48.2)	21/58 (36.2)	15/58 (25.8)	22/58 (38)	58/112 (51.8)	112 (100)

a: intradomiciliary capture at the request of a resident; b: captured in nest of casaca-de-couro (Pseudoseisura cristata); CJ: Cachoeira do Júlio; ME: Seu Evangelista; MPC: Pedra da Cruz.

In 2009, the predominant capture of adult males [44% (32/73)] and young males [30% (22/73)] was observed, but the presence of young and adult females was also observed, with a success catch by site and study period of 23.6% (73/309), representing 65% of all animals sampled. In 2010, 79.5% (31/39) of animals captured were young males, with successful capture rate of 4.6% (39/853), even with an effort that was 2.7 times greater than in 2009.

By correlating rainfall and capture effort within our wild animal capture data, we found that at the end of the drought period (February 2009/March 2010), the number of animals taken was 10% less than the number of animals captured in the early dry season (August 2009/2010), with a predominance of young male animals (45% - 23/51), followed by adult males (33% - 16/51). At the beginning of the dry season (August 2009/2010), the capture of young male animals was higher, corresponding to 49% (30/61).

Table III shows that there was no difference between the number of species caught in the wild, the period of rainfall incidence and the abundance of the two main species found.

PCR multiplex - The 124 samples (43 dogs, 11 goats, 23 cats, 36 ovine and 11 porcine) considered to be positive or doubtful for T. cruzi by serological diagnosis were subjected to molecular characterisation using multiplex PCR, including 11/115 (8.8%) amplified fragments to characterise the presence of DNA from TCI (all dogs) (Table II). For the 24 samples (22 T. laurentius and 2 R. rattus) captured in February/2009, no amplified fragments were characterised by the presence of DNA from T. cruzi.

Domiciliary triatomine infestation and infection by T. cruzi - Two hundred fifty-one UDs were surveyed in February and March of 2009 and in the 18 sites of the study, 39% (97/251) were positive, 12.3% (31/97) of which were in households, 29% (73/97) in peridomicile areas and 7.2% (7/97) in both environments. All locations investigated exhibited the presence of insects, with an average infestation rate of 40%, ranging from 16-100%.

The house construction was in good condition: 64.5% (162/251) were constructed from masonry with plaster, 27% (68/251) masonry without plaster, 4.4% (11/251) without clay plaster and 4% (10/251) with clay plaster. In peridomestic environments, 437 attachments were studied and 18% (79/437) tested positive. Among those, there was a predominance of infection in chickens at 47% (37/79); tiles, bricks and stones accounted for 24% (19/79) and other ecotopes accounted for 16% (13/79).

In the household survey, 749 insects were captured. Of those, 369 (49.3%) were T. brasiliensis, 377 (50.3%) were Triatoma pseudomaculata and three (0.04%) belonged to other species. The number of T. pseudomaculata captured in the peridomicile environments was 25 times higher than in households. Contrarily, the presence of nymphs in homes [13/15 (86.7%)] demonstrates their ability to colonise that environment.

Wild T. brasiliensis and natural infection by T. cruzi - Among the 166 specimens of wild T. brasiliensis captured, 131 (79%) were nymphs and 35 (21%) were adults. Of these, a fifth stage nymph and one adult male were infected with T. cruzi.

The presence of T. brasiliensis in the wild was confirmed in the same places described for the capture of small mammals, focusing on the huge conglomerate rocks that present numerous opportunities for these and other animals to take shelter. After darkness, triatomines of all developmental stages left their hiding places and stayed on the surfaces of the stones. There was little observable movement of these insects, leading to easy capture. Interestingly, triatomine bugs present an aggressive stance, attempt to attack the captors and may even pursue them for several minutes with a distended proboscis, ready to sting (Fig. 2). The surface temperature of the rocks falls from 65ºC in the early afternoon to approximately 30ºC at night. Triatomines return to their hideouts approximately 21-22 h later. On some occasions, we observed adults flying, but in proportion to the number of insects visible, flying was not a routine behaviour on these occasions.

Fig. 2

: attack of adult (A) and nymph (B) of Triatoma brasi-liensis (with extended proboscis) in the wild environment during the day. Source: L Diotaiuti (October/2010).

Identification of the dietary source - Cytb - The food sources of T. brasiliensis captured in their wild environments were identified from 35 DNA samples (n = 35). Of these, 16 (9 nymphs and 7 adults) confirmed the presence of the studied fragment (bandwidth ~305 bp) corresponding to the DNA of Cytb. These samples belonged to the most anthropic environment, MPC, and to the least, CJ. The species described in Table IV as food source of insects were identified.

TABLE IV

Dietary sources of wild Triatoma brasiliensis, identified by cytochrome b in the city of Tauá, state of Ceará, 2009

Dietary source	n (%)
Galea spixii	5 (31.3)
Kerodon rupestris	4 (25)
Capra hircus	3 (18.8)
Tropidurus oreadicus	2 (12.5)
Tupinambis merianae	1 (6.2)
Gallus gallus	1 (6.2)
Total	16 (100)

DISCUSSION

This study demonstrated a high serum prevalence of infection by T. cruzi in dogs living in the Caatinga region and this result is in agreement with other authors (Alencar 1987, Herrera et al. 2005, Xavier et al. 2007, Lima et al. 2012). Age appears to be important in determining disease prevalence because it was observed to increase with increasing age of the animals (Table II).

Dogs are considered by Gürtler et al. (1998) and Cohen and Gürtler (2001) to be an important risk factor for the transmission of T. cruzi to humans. This finding was based on the characteristics of permanent and prolonged infection and high infectivity (Gürtler et al. 1996) in triatomines compared to children or adults. The prevalence rate of 38% (20/53) in dogs in the study area is important in the context of T. cruzi circulation, even after progress has been made in reducing household populations by triatomine control actions. This fact was emphasised by the characterisation of TcI in dogs (Table II), showing a lineage related to the wild enzootic cycle, strengthening the close relationship between the zoonotic cycle of T. cruzi and the human populations in question (Xavier et al. 2007, Lima et al. 2012, Zingales et al. 2012).

The occurrence of mixed infection with L. infantum [13% (7/53)] and Leishmania spp [19% (10/53)] in dogs was expected because visceral leishmaniasis is endemic in the municipality of Tauá.

In cats, we obtained results that support the occurrence of mixed infections, where 22% (9/41) of samples reacted to T. cruzi and Leishmania spp, peaking at 36% (5/14) in smaller animals more than a year old. Among those animals surveyed, four were six-nine months old and two were found living in the same house with three dogs over five years old that had positive serology for T. cruzi and/or L. infantum.

Comparing our results for dogs and cats, we find that 38% (20/53) of dogs had positive serology for T. cruzi, while only 2.4% (1/41) of cats did, with significant differences in their age distributions (Table II). Dogs have been described as frequent sources of blood for triatomines, sometimes even more so than humans (Gürtler et al. 1996), chickens or cats (Gürtler et al. 2009).

The management of domestic animals should be taken into account when defining the risk area for the transmission of T. cruzi. According to their displacement or confinement, it is possible to determine these animals’ degree of exposure to the transmission cycle and their proximity to homes (Roque & Jansen 2008). The use of pets as sentinels in the identification of risk areas for the transmission of T. cruzi has been proposed in countries such as the United States of America, Venezuela, Mexico and Argentina (Shadomy et al. 2004, Crisante et al. 2006, Estrada-Franco et al. 2006, Gürtler et al. 2007). As such, depicting the profile of infection in these animals allows the real and imminent risk of transmission to humans to be determined.

Goats and sheep accounted for 58.3% (185/317) of our samples and we obtained an infection prevalence of only 1% (1/102) among sheep. This prevalence proved to be extremely low compared to other studies (Marzochi et al. 1987, Rozas et al. 2005), confirming only the results obtained by Herrera et al. (2005), who identified a 2% (1/56) prevalence rate in samples from caprines in the state of Piauí. Perhaps the choice of animals up to six months of life explains the low prevalence rate found in our study.

These animals remain loose to eat and drink during the day and are collected in corrals in the evening. These corrals surround the homes, are constructed of native wood and generally contain stony places where the presence of T. brasiliensis is easily verified.

Thus, although we have not identified an important prevalence in sheep in accordance with other studies, the link between sheep, goats and triatomines is an intimate and persistent one, not only in the wild environment, but also around homes, where these animals represent a significant food source for these insects and are thus vulnerable to the transmission cycle of T. cruzi.

The sample of pigs represented 10.6% (34/317) of our study, with a prevalence rate of 6% (2/34) and suspected infection rate of 26.5% (9/34). These animals are common in the study area, but the collection of blood samples was limited due to the difficulty of capturing the animals. The vast majority of pigs live in small herds and while they frequent peridomicile areas for feeding and sleeping, they do not allow humans to approach with ease. Confounding other limitations, such as the difficulty of access to the animals that exist in peridomicile environments, goats and sheep are released during the day, which makes it difficult to collect blood samples.

Pigs were recently included on the list of peridomestic animals that may be important in the transmission cycle of T. cruzi, especially in prevalence areas of acute Chagas disease (Roque & Jansen 2008, Roque et al. 2008), where high prevalence are described, but with a low incidence of parasitosis.

T. laurentius, caviomorph rodents that are mainly associated with rocky sites in semiarid regions, were the most abundant species at all sites studied by us, which is in agreement with Herrera et al. (2005), Xavier et al. (2007) and Roque et al. (2008). The greatest species diversity was in ME, considered by us to be susceptible to human interference at an intermediate level. We note that the capture of R. rattus occurred in households and that of M. musculus in a nest of casaca-de-couro (Pseudoseisura cristata). The capture of D. albiventris occurred only in places where we also expected greater human modifications to the environment. The ability of these animals to adapt to this environment is recognised and cited as one of the factors that make them an important reservoir of T. cruzi.

Contrary to what is found in the literature, we found that the areas in the locality of CJ that had the lowest richness of mammals were the most isolated areas and distant from human action. Most studies of small mammals have shown that greater human action corresponds to the greater simplification of fauna, with a predominance of synanthropic animals such as marsupials and rodents that can roam between the human and wild environments (Mills & Childs 1998, Roque et al. 2008). There was a huge difference in the capture success between the four campaigns, even with an increase in the capture effort. It is known that the density of small mammals is mainly related to the availability of water. According to data from the National Institute of Meteorology, the amount of precipitation for Tauá was regularly above average between 2007-2009, indicating that water was available in the environments inhabited by these animals, even during droughts. This fact is reflected in the ease of capture of these populations in 2009, when males, females, adults and juveniles were sampled. In 2010, there was a significant decrease in precipitation and, consequently, an abrupt decrease in the animal populations sampled, with a predominance of young males. Although T. laurentius has been described as important in the transmission cycle and/or maintenance of T. cruzi in the semiarid region (Herrera et al. 2005, Xavier et al. 2007, Roque et al. 2008), none of our samples amplified fragments that characterised the presence of T. cruzi DNA, confirming the negative cytological examination. These data show that these animals, although the most abundant, do not participate in the amplification cycle of T. cruzi in the sampled area.

Only 6.2% (7/112) of wild animals captured were marsupials (4 D. albiventris and 3 M. domestica) and none were infected by T. cruzi. The fact that we captured 71% (5/7) of these animals, mostly adults, at the beginning of the dry season suggests dispersion or further exploration of the environment because the traps were placed at ground level.

Locations with greater diversity of environments provide greater opportunities to be exploited by hosts of T. cruzi. These habitats are occupied by different species of mammals with different degrees of susceptibility that exert different pressures on the parasite and, consequently, shape an epidemiological scenario of greater diversity for T. cruzi (Rozas et al. 2007, Xavier et al. 2007, Jansen & Roque 2010). Thus, even without the detection of parasitosis by direct examination and multiplex PCR, the capture of rodents and marsupials that are recognisably important in maintaining and/or amplifying the T. cruzi cycle in environments that have been known to be infested by T. brasiliensis leads us to reflect on the animals’ exposure to the parasite. Regarding the degree of human interference in these locations, it will be possible to define areas of interface between domestic and wild environments that pose a risk in the transmission of T. cruzi.

T. pseudomaculata, predominantly a peridomestic species, was found to colonise the interior of homes successfully through the predominance of nymphs in that environment (87%). This ability rests primarily on T. brasiliensis, capable of colonising 12.3% of the intradomiciliary environments investigated, even with over than 60% of homes investigated having good construction conditions.

In the peridomicile environment, some ecotopes have been more attractive for the development of the insects (Cecere et al. 2004). In our study, chickens were the most infested ecotopes, followed by tiles, bricks and stones, agreeing in part with Sarquis et al. (2006), who found greater infestations in goat sties and chicken perches. Diotaiuti et al. (2000) demonstrated that perches (33%), chickens (24%) and goat and sheep pens (14%) are the most important ecotopes. This result shows the similarities and particularities of the epidemiological profile of each region and the influences of environmental, socioeconomic and cultural factors. Interventions to prevent human disease will be more effective if these differences are known and taken into consideration.

The prevalence of nymphs of T. brasiliensis in all environments shows the importance of this species as a vector of T. cruzi. No other species captured were parasitised and T. brasiliensis was positive for T. cruzi in the three investigated environments, including indoors, posing a real risk of transmission of the parasite. In a study performed in CE, Lorenzo et al. (2000) showed that microclimatic similarities between wild habitats (rocks) and hiding places in indoor walls favour the adaptation of T. brasiliensis to the artificial environment. Thus, indoor locales are an attractive environment for T. brasi-liensis because they provide appropriate environmental conditions, such as availability of stable food sources throughout the year (Diotaiuti 2007).

The complexity of the peridomicile environment is such that the impact of the renewal of attachments between one spraying and another, as well as the preference of insects per type of renovated attachment, directly affects the rate of colonisation (Oliveira-Lima et al. 2000). Moreover, the action of pyrethroid insecticides is limited because they do not reach all hiding locations and the pesticide’s residual action is reduced by bright light, high temperatures, wind and rain, further facilitating the process of colonisation of the insects in the Northeast Region (Diotaiuti et al. 2000, Oliveira Filho et al. 2000). This condition is further amplified as a result of contiguity between natural and artificial ecotopes, overlapping habitats and facilitating interaction between triatomines and humans, increasing the risk of transmission of T. cruzi (Borges et al. 2005, Moncayo & Silveira 2009, Sarquis et al. 2012). According to Gürtler et al. (2005), the failure to conduct entomological surveillance in areas infested by T. infestans in Argentina, including the lack of comprehensive and regular insecticide coverage and a lack of supervision, have led to a rapid recovery of the insect population after two-three years and to a restoration of household transmission.

The ease with which T. brasiliensis invades and colonises a wide variety of peridomiciliary structures allows the animals and, therefore, the infection to transit between these environments, connecting the cycle of T. cruzi with households. In our study, the peridomicile environment was the locale in which insects had the highest natural infection at 14% (22/157), even though chicken areas were the most infested ecotopes. This fact emphasises the need to identify the ecological and biological aspects of the relationship of triatomines with the different sources of peridomestic infection to implement entomological surveillance due to the importance of houses in the re-colonisation of habitats (Abad-Franch et al. 2005, Noireau et al. 2005, Guhl et al. 2009).

Complementing the list of animals associated with T. brasiliensis in wild environments, the blood ingested by triatomine was identified using molecular techniques. Among the food sources identified, 56% (9/16) belonged to rodents that live in the same environment, while 18.7% (3/16) were goats associated with rock formations and raised in a semi-extensive way. The association with reptiles and birds (25%, 4/16) may have limited natural infection of triatomines in sylvatic cases, as we recorded (Table IV).

Another behaviour that demonstrates the ecological valence and opportunistic character of T. brasiliensis is their aggressiveness in seeking a blood meal after leaving their shelters. Fig. 2 illustrates one of these attacks that we witnessed, where the insect was able to pursue its prey for food, sometimes even during the day.

In the Caatinga region, life and agricultural production are highly dependent on plant resources. The natives from the region draw numerous products and services from the land to make life possible in the semiarid area. The use of soil, terrain, flora and fauna by humans means that there is a close relationship between human life and diverse organisms. The secular process of occupation of the area has contributed to the general deterioration of its vegetation, resulting in a profound modification of the primary vegetation cover (Meunier & Ferraz 2005). Only with the development of strategies for regional conservation through the sustainable use of natural resources in the region and by seeking to preserve the diversity and stop the increase in desertification will it be possible to maintain ecological services required for rural populations and their livelihoods (Leal et al. 2005).

Preserving ecosystems and maintaining their diversity should generally reduce the prevalence of infectious diseases (Keesing et al. 2006, 2010). The increase in species diversity may reduce the risk of infection because it reduces the chances of the animal population becoming simplified through the selection of one or more species that have the ability to maintain or amplify a parasite and favour transmission (Ostfeld & Keessing 2000, Pinto et al. 2006).

A reflection on human behaviour as the primary responsible factor affecting environmental change on our planet and all of the consequences that this act represents fit in this context. Governments and societies have key roles in building solutions so that sustainable development is strengthened as a paradigm for all relevant actors in the economic, social and environmental areas. Only by reaching a consensus on the complementarity of these three pillars of development can we accomplish the sustainability needed and ensure the supply of current human needs without compromising the future of the next generation.

As in all biomes, survival in semiarid regions is closely related to the use of natural resources and to the development of regional conservation strategies. Only then is it possible to maintain ecological services as required by rural populations and their livelihoods. Preserving ecosystems, maintaining their diversity, increasing the stagnation of desertification and appreciating different ways of life will contribute to increasing the possibilities for answers that allow for a balance between health, environment and sustainable development.

The high prevalence of infection by T. cruzi observed in domestic and outdoor animals strengthens the close relationship between the enzootic cycle and human populations. The creation of peridomestic animal environments close to human habitations is important to ensure the safety of these “goods” and is a characteristic of life in semi-arid northeast regions. Chagas disease in the Northeast Region has peculiar characteristics when compared to the vast endemic areas that have T. infestans as their main species. The status represented by this species as the most important vector in the transmission of Chagas disease has failed to take into account native species, diminishing their importance. This perspective is indeed erroneous because, in the locale where indigenous species are prevalent and where T. infestans is present, these species are responsible for the domestic transmission of the disease. Adding to this consideration is the difficulty of controlling these species because they are in their natural environment and exert relentless pressure to colonise free environments. In this context, the presence of the parasite around homes is a valid risk factor for the transmission of Chagas disease in this region and this transmission may be aggravated if control measures are flawed, interrupted or discontinued.