Blood feeding patterns of Nyssomyia intermedia and Nyssomyia neivai ( Diptera , Psychodidae ) in a cutaneous leishmaniasis endemic area of the Ribeira Valley , State of São Paulo , Brazil

Introduction: The aim of this study was to identify the blood feeding sources of Nyssomyia intermedia (Ny. intermedia) and Nyssomyia neivai (Ny. neivai), which are Leishmania vectors and the predominant sandfl y species in the Ribeira Valley, State of São Paulo, Brazil, an endemic area for cutaneous leishmaniasis. Methods: Specimens were captured monthly between February 2001 and December 2003 on a smallholding and a small farm situated in the Serra district in the Iporanga municipality. The blood meals of 988 engorged females were tested using the avidin-biotin immunoenzymatic enzyme-linked immunosorbent assay (ELISA). Seven blood meal sources were investigated: human, dog, chicken, bovine, pig, horse and rat. Results: The results showed that among the females that fed on one or more blood sources, the respective percentages for Ny. intermedia and Ny. neivai, respectively, were as follows: human (23% and 36.8%), pig (47.4% and 26.4%), chicken (25.7% and 36.8%) and dog (3.9% and 0%), and the differences in the blood sources between the two species were statistically signifi cant (p = 0.043). Conclusions: Both species had predominant reactivity for one or two blood sources, and few showed reactivity indicating three or four sources. Many different combinations were observed among the females that showed reactivity for more than one source, which indicated their opportunistic habits and eclecticism regarding anthropic environmental conditions.

The subfamily Phlebotominae is an insect group that has worldwide distribution.It comprises approximately 500 species 1 whose females feed on a wide range of vertebrates, including mammals, birds, amphibians and reptiles in the Americas.
Many sandfl y species have the ability to transmit pathogens to animals, some of which are reservoirs for cutaneous and visceral leishmaniasis agents 2,3 .
The Ribeira Valley region has one of the highest incidences of cutaneous leishmaniasis cases in the State of São Paulo 4,5 .This region had a total of 1,174 cutaneous leishmaniasis cases registered between 1998 and 2006 and one of the highest average annual incidence coeffi cients in the State of São Paulo (47.71/100,000 inhabitants).The Serra district is situated in the Iporanga municipality, which has an incidence coeffi cient of 152.9/100,000 inhabitants, one of the highest in the Ribeira Valley region 4 .
Understanding sandfl y behavior and its implications for transmitting Leishmania in this area involves studying their blood feeding habits, as this information is relevant when estimating the degree of host-vector contact.
Highly specifi c and sensitive methods of detecting blood sources are necessary to identify the host-feeding patterns of blood-sucking insects.The use of the biotin-avidin enzymelinked immunosorbent assay (ELISA) method allows the detection of small volumes of blood from single or various sources 6,[7][8][9] .The aim of the present study was to report on the feeding patterns of Nyssomyia intermedia (Ny.intermedia) (Lutz  & Neiva, 1912) and Nyssomyia neivai (Ny.neivai) (Pinto, 1926), which are the predominant sandfl y species among the 19 species found in the Serra district 4,10 and the suspected transmission vectors of cutaneous leishmaniasis agents 11 .

Study area and sandfl y sampling
The Serra district (24º 33' 19.7'S, 48º 40' 42"W, 198m above sea level [ASL]) is situated on the banks of the Betari river, a

RESULTS
tributary of the Ribeira river in the municipality of Iporanga, an area containing hostels for tourists who visit the Parque Estadual do Alto Ribeira (PETAR), which is situated in the speleological province of the Ribeira Valley in the Serra de Paranapiacaba in the south of the State of São Paulo within the Atlantic forest domain.
The JRA smallholding contains a dwelling in which a family of four and a dog lived and a peridomicile with a pigsty, a chicken-house and a goose-shelter.The AG small farm contained two dwellings, one of which was temporarily inhabited by a family (two people).The other dwelling is permanently inhabited by one person, with a hen house, two pigsties, horses and cows in the proximity of the two houses.
The capture of phlebotomines on the JRA smallholding were performed monthly from February 2001 to December 2003 in the hen house, in the pigsty and on the porch of the house (with modifi ed automatic light traps) and from May 2001 to June 2002 in the peridomicile (with modifi ed black and white Shannon traps), as described by Galati 4,12 .
The captures on the AG small farm were performed monthly from March 2001 to December 2003 using modifi ed automatic light traps in a hen house, a pigsty and on the porch of one house 4 .
During a dispersion study, sandfl ies were also obtained from two other locations: a peridomicile situated 180m from the AG small farm and a cave located 2km from the farm.The captures were performed monthly on eight consecutive days from August to December 2003 using electric manual aspirators in the morning and at night in resting places (i.e., the nest of a dog's litter, on rocks and tree trunks, and in pigsties, hen houses, birds' nests, rat burrows, banana groves and cattle pens), automatic light traps in peridomiciles, domestic animal shelters, the forest, the forest's edge, pastures, banana groves, caves, or Shannon traps in a peridomiciliary area.A more detailed description is given in Galati 10 .

Sample collection and processing
Engorged female specimens were captured, conditioned in refrigerated Petri dishes and transported to the Laboratory of Public Health Entomology/Phlebotominae of the Epidemiology Department, Public Health School of the São Paulo University.
Parts of the abdomen and the thorax of each engorged female were placed in a microtube and stored at -20•C until the ELISA test could be performed.The species were identifi ed by the head and genitalia as described by Andrade Filho 13 , who distinguished them based on the number of rings, the width of both: the base of the terminal knob and the apex of the individual sperm ducts situated at the junction of the spermathecae, and the length of the individual and common ducts.The generic name abbreviations follow Marcondes 14 .

Blood meal analysis
The hosts of the blood-fed phlebotomines were identifi ed using a biotin-avidin ELISA in accordance with Marassá 7,8 .The feeding behavior was investigated using host-specifi c anti-human, antichicken, anti-dog, anti-rat, anti-horse, anti-pig and anti-bovine sera.

Statistical analysis
For the statistical analysis, Fisher's exact test was used to compare the differences between the blood sources of the two sandfl y species.
The species that were identified in the blood ingested phlebotomines captured on the JRA smallholding and the AG small farm and their respective percentages, as determined by the biotin-avidin ELISA, are listed in Tables 1, 2 and 3. On the JRA smallholding, 141 engorged females were captured, 83.7% of which were reactive (Table 1).Among those insects that fed on one or more blood sources, the respective percentages for Ny.intermedia and Ny.neivai, respectively, were human (23% and 36.8%),pig (47.4% and 26.4%), chicken (25.7% and 36.8%) and dog (3.9% and 0%), and the difference in the blood source frequency between the two species was statistically signifi cant (p = 0.043).
For the specimens captured on the porch, no statistically signifi cant difference with regard to the blood source (p = 0.345) was found between the two species, and the proportions of those insects that fed on the respective sources were as follows: human (41.7%), pig (25%), chicken (26.2%) and dog (7.1%) for both sandfl y species combined.
No statistically signifi cant difference between the blood sources found in the Ny.intermedia and Ny.neivai that were captured on the porch was observed (p = 0.400); the frequencies of the blood sources that were found in these two species were as follows: pig (46%), human (20.6%), chicken (16.9%), horse (11.7%), bovine (3.6%) and dog (1.2%).

DISCUSSION
Knowledge of the sandfly feeding pattern is relevant to understanding the host animals that are involved in the leishmaniasis transmission cycle and can also help to identify the conditions that favor increases in the population density of vectors, particularly in the anthropic environment (e.g., the increased risk of infection among humans and domestic and synanthropic animals).
Nyssomyia intermedia s. lat.has been implicated as a vector of the cutaneous leishmaniasis agent in the southeastern and southern regions of Brazil 15,16 .Now, Ny. neivai is again considered to be a distinct species from Ny. intermedia 17 , and Ny.intermedia has been found to be more closely associated  13,18,19 where the present study was performed 4 .
Although studies that focus on the feeding habits of Ny. intermedia 20 and Ny.neivai 21,22 have been performed, the present research analyzed their feeding pattern in a sympatric area.
In the peridomicile of the JRA smallholding, Ny. neivai was more attracted to humans and chickens, corroborating the observations of Casanova 21 .The Ny. intermedia and Ny.neivai found on the porch of the smallholding presented a distinct pattern of blood source preference compared with those found on the AG small farm.On the JRA smallholding, both species took more human blood, which could be attributed to the greater number of people who permanently reside in the dwelling and to the greater diversity in blood meal combinations, which may be the result of the smallholding's proximity to the residual forest, hen house and pigsty, thereby providing the opportunity for and ease of contact with humans.
On the AG small farm, the majority of the blood meals identifi ed in the peridomicile and on the porch comprised pig blood.This observation might be attributed to the large number of pigs that were present on this site during the captures between March 2002 and December 2003.The confi nement of these animals and their large areas of exposed skin favor the concentration of male and female sand fl ies for courtship and provide blood sources for the females, as observed with regard to henhouses for Lutzomyia longipalpis 23 .
The aspirations that were performed on the AG small farm in resting places in the peridomicile and on the rock walls at the forest edge did not reveal any exclusive blood meal preference.The variety of food sources suggests the same eclectic feeding behavior that was observed in the peridomiciliary environment and its association with domestic animal hosts 24 .
On the AG small farm, the blood meal analysis also showed a mixture of ingested blood, as demonstrated by the proportion of females that obtained blood from more than one source.
The blood meal analyses demonstrated that Ny. intermedia and Ny.neivai tended to feed on multiple hosts.The variety of food sources that were identifi ed corroborated their eclectic feeding habits and their opportunistic behavior.However, this eclectic feeding pattern was restricted to domestic animals for the specimens collected in the peridomiciles, on porches, and on the slopes at the forest's edge.This behavior may be associated with the fact that these two species breed in peridomiciliary sites, as observed for Ny.neivai 25 (at the time identifi ed as Lu.intermedia s. lat.) and Ny.intermedia 26 .
Some large mammals, such as swine, do not appear to be Leishmania reservoirs because there has been only one published report of Leishmania amastigotes in a lesion on the ear of a pig living in the peridomicile of a dwelling of a human case of cutaneous leishmaniasis 27 and one Brazilian report of infection caused by Leishmania infantum.Although these pigs produced Leishmania antibodies, neither the presence of the parasite nor its DNA was detected 28 .Similarly, no such role can be attributed to cattle with regard to Leishmania donovani, the agent of visceral leishmaniasis in Bangladesh 29 .
However, domestic dogs are frequently reported to be naturally infected with Leishmania braziliensis, the agent of cutaneous leishmaniasis with the widest distribution in South America [30][31][32][33][34][35]36 . Oftn this scenario occurs in old colonization areas that also have a high prevalence of human infection by this parasite, sometimes with signifi cant association between the occurrence of human cases and the presence of infected dogs in the domicile.Under these circumstances, the dog has been incriminated as the infection source or as the reservoir of the parasite [30][31][32][33][34][35] .In the same manner, equines, which may have high natural infection rates in some areas, may be an infection source for humans and other animals 30,33 .However, the role of these animals as reservoirs of this parasite has not been well established 37,38 .
In the present study, the low rate of dog blood reactivity may not refl ect a lack of attraction between the two sandfl y species investigated and dogs but instead may refl ect the scarcity of this blood source; only one dog was present in one of the households (JRA smallholding), and evidently, a dog entered areas of the AG small farm.
Despite the greater number of females examined, no rat blood was detected; therefore, it seems that the populations of Ny. intermedia and Ny.neivai in this study area are rarely attracted to rodents, which contrasts with the findings of Afonso 39 for Ny.intermedia in the State of Rio de Janeiro.
A portion of the 19.1% of specimens that were nonreactive can be attributed to other animal sources that were not investigated.However, non-reactivity may also have been observed because many females with dark abdomens and increased volumes were actually full of eggs rather than blood.It is also possible that the amount of blood or its state of degradation prevented the identifi cation of its source.
The great availability of avian hosts, which are refractory to the development of this parasite, and some larger mammalian hosts, which are less effective at transmitting the agent of cutaneous leishmaniasis than rodents 38 , may dilute the transmission of this parasite in these areas.This scenario may explain the absence of human cases of cutaneous leishmaniasis at these two sites, despite the high density of these vectors observed by Galati 4,10,12 .
In conclusion, although the risk of humans contracting cutaneous leishmaniasis in the described scenario is small, the opportunistic feeding behavior of Ny. intermedia and Ny.neivai suggests that actions should be taken at the local level to mitigate the vulnerability of human populations exposed to these vectors of cutaneous leishmaniasis agents, particularly in the presence of synanthropic mammals that may be reservoirs of this parasite.In any case, the presence of these sandfl ies in high densities is a source of annoyance to both residents and domestic animals.

TABLE 1 -
The blood meal distribution identifi ed between January 2001 and December 2003 (using automatic CDC light traps) by source and sandfl y species in the JRA smallholding, Serra district in Iporanga municipality according to the collection site.

TABLE 2 -
Blood meal distribution identifi ed by source and sandfl y species according to the collection site at the AG small farm, Serra district, Iporanga municipality, between March 2002 and December 2003.

TABLE 3 -
Blood meal distribution identifi ed by source and sandfl y species according to the collection site at the AG small farm, Serra district in Iporanga municipality (captured with aspirators in resting places) between August and December 2003.
with the areas of Atlantic forest located in the coastal States of Espírito Santo, Rio de Janeiro and São Paulo, whereas Ny. neivai is more common in the whole Southern region and westward from the Atlantic plateau in the State of São Paulo.However, these species are sympatric in the Ribeira Valley region of São Paulo