Diversity, natural infection and blood meal sources of phlebotomine sandflies (Diptera, Psychodidae) in the western Brazilian Amazon

BACKGROUND The state of Rondônia (RO) is a hot spot for human cases of cutaneous leishmaniasis. Many sandfly species in RO are putative vectors of leishmaniasis. OBJECTIVES This study examines the diversity patterns and the presence of Leishmania DNA and blood meal sources of sandflies in RO. METHODS A sandfly survey was performed between 2016 and 2018 in 10 municipalities categorised into three different environment types: (i) Conservation Unit (CUN) - comprised of preserved ombrophilous forests; (ii) Forest Edge (FE) - small forest fragments; and (iii) Peridomicile (PE) - areas around dwellings. FINDINGS A total of 73 species were identified from 9,535 sandflies. The most abundant species were Psychodopygus davisi (1,741 individuals), Nyssomyia antunesi (1,397), Trichophoromyia auraensis (1,295) and Trichophoromyia ubiquitalis (1,043). Diversity was the highest in CUN, followed by the FE and PE environments. One pool of Ps. davisi tested positive for Leishmania braziliensis, reinforcing the possibility that Ps. davisi acts as a vector. The cytochrome b (cytb) sequences were used to identify three blood meal sources: Bos taurus, Homo sapiens and Tamandua tetradactyla. MAIN CONCLUSIONS Our results demonstrated that sandflies can switch between blood meal sources in differing environments. This study enhances the knowledge of the vector life cycle in RO and provides information relevant to leishmaniasis surveillance.

Brazil is also home to a range of putative vectors such as Nyssomyia intermedia, Nyssomyia neivai, (3) Nyssomyia shawi, (4) Psychodopygus carrerai, Psychodopygus hirsutus hirsutus (5)(6)(7) and Trichophoromyia auraensis (Mangabeira). Leishmania (Leishmania) infantum Nicolle is the causative agent of VL, and its primary vector is Lutzomyia longipalpis (Lutz & Neiva), which is distributed throughout Brazil. (2) Other species associated with the transmission of Le. (Le.) infantum include Lutzomyia cruzi (Mangabeira), (8) Mg. migonei (9) and Pintomyia fischeri. (10) CL occurs in all regions of the Brazilian Amazon. Its incidence rates are the highest in the states of Pará (PA) and Amazonas (AM). (11) In contrast, VL is endemic to the states of Mato Grosso (MT), Mato Grosso do Sul, Roraima, Tocantins and PA. (11) The zoonotic transmission risk and incidence of this disease are increased by certain factors, including (i) the presence of an ombrophilous forest, (ii) a variety of blood meal sources (e.g., tapirs, opossums, sloths and armadillos), (iii) a variety of Leishmania species and (iv) the presence of sandfly vector species. (5,12,13) In major transmission foci, high incidences of CL might also be the result of human exposure to the environment through activities, such as hunting and fishing, or the result of recent deforestation caused by the construction of hydroelectric power plants and roads. (14) RO is located in the western Amazon basin; it borders the states of Acre (AC), AM and MT and it shares an in-ternational frontier with Bolivia. The state is comprised of open ombrophilous forests that have been decreasing in area since the 1970s as a result of agriculture activities. RO has the third highest incidence of CL in north Brazil, having registered approximately 12,000 cases between 2007 and 2018; according to the Brazilian Health Ministry, this incidence rate is the result of intense zoonotic transmission. (11) No human cases of VL have been reported in RO; however, this state is thought to be at risk for VL because canine cases have recently occurred in the municipality of Cacoal. (15) Previous studies have reported that human cases of CL in RO (16) Despite the presence and high diversity of known CL vectors, such as Bi. flaviscutellata, Ny. antunesi, Ny. umbratilis and Th. ubiquitalis, (4,13) few studies conducted in RO have demonstrated the natural infection of sandflies. (4,7,13,17) The presence and high abundance of Leishmania in RO suggest that sandflies maintain their transmission cycles in the this region. (4,7,13,17) The epidemiological knowledge of sandflies can be improved by studying their blood meal sources (18,19) and identifying putative reservoirs. The mitochondrial genes, such as cytochrome b (cytb), have been used as molecular markers for the detection of their blood meal sources. (18)(19)(20) Thus, the present study aimed to characterise the sandfly fauna and identify their blood meal sources, as well as to assess the natural infection caused by Leishmania in RO.

MATERIALS AND METHODS
Study areas -RO is located in the northern region of Brazil; it borders AM to the north, MT to the east and AC to the west and shares an international border with Bolivia to the southwest (Fig. 1). It has an area of approximately 238,000 km² and contains 52 municipalities.
The sandfly fauna was collected in three environments: Conservation Unit (CUN) -characterised by large areas of ombrophilous rain forest; Forest Edge (FE) -characterised by small forest fragments near urban areas; and Peridomicile (PE) -areas around dwellings that are situated near small forest fragments and contain enclosures where domestic animals are raised. Collections in the CUN environment were conducted between 2016 and 2017 at three places ( Fig. 1): the Jaru Biological Reserve (REBIO Jaru), which has a territory that covers six municipalities ( Sandfly collection and identification -In the CUN environments, collections were made along two different trails and sampling was performed twice in 2016 and twice in 2017. Six Hoover Pugedo (HP) light traps were set along each trail (12 HPs/reserve) and collections were made between 06:00 p.m. and 07:00 a.m. for five consecutive days.
Collections in the FE and PE environments were made from 2016 to 2018 at five locations within each municipality. At each location, one trap was set in the FE environment and two traps were set in the PE environment, using a total of 15 traps per municipality.
Male sandflies were clarified in 10% potassium hydroxide (KOH), washed in 10% acetic acid and slidemounted in Berlese fluid. Females were divided into engorged and non-engorged specimens, and their heads and genitalia were clarified and slide-mounted as above. The thorax and abdomen of each female were stored in a microtube with 96% ethylic alcohol for further molecular analysis. Species identification was carried out using the morphological characters described by Galati. (2) Molecular detection of Leishmania -Females were sorted according to species abundance, collection location and environment type and were separated into pools of 2-20 specimens. DNA extraction and polymerase chain reaction (PCR) assays were performed by targeting kDNA and hsp70, as described elsewhere. (4,21) The Th. ubiquitalis males and the Le. amazonensis reference strain IOC/ L0575 (IFLA/BR/1967/PH8) were used as positive controls and ultrapure water was used as the negative control.
Sandfly blood meal sources -Engorged females were separated according to species, municipality, and environment type. During DNA extraction, three samples were used as negative controls: one sample containing DNA-free water and two samples containing a female sandfly with no blood present in the gut. DNA extraction was carried out using the phenol/chloroform protocol described by Sambrook and Russell. (22) A PCR was carried out using the primers cytb 1 and cytb 2, which are complementary to the conserved region of the cytb gene in vertebrate mitochondrial DNA. (23) The PCR amplification was carried out in a 50 µL reaction volume containing 25 µL (1X) Go Taq Colorless (Promega ® , Madison, WI, USA), 1.5 µL of each primer (cytb 1 and cytb 2, 10 µM each) and 5 µL of DNA (< 250 ng). The amplifications were performed in a thermocycler (Veriti ® -Applied Biosystems, Foster City, CA, USA) with an initial denaturation of 95ºC for 5 min, followed by 35 cycles of denaturation at 95ºC for 30 s, annealing at 53ºC for 30 s and extension at 72ºC for 1 min, with a final extension at 72ºC for 6 min. Amplified products were purified using the QIAquick Purification Kit (Qiagen, Hilden, Germany) and submitted to the Fiocruz Sequencing Facility (Rio de Janeiro, RJ, Brazil).
Data analysis -Interpolation and extrapolation curves (iNEXT) were used to evaluate sample coverage and compare diversity indexes (Shannon and Simpson) between environments. Comparisons were made using Hill numbers expressed as order q values, and the data were analysed in the R program. (7) The sequences (hsp 70 and cytb) were analysed using the Phred, Phrap and Consed software programs, (24) with the minimum value defined as Q = 30. The consensus sequences were submitted to the Basic Local Alignment Search Tool (BLAST) (http://blast.ncbi.nlm.nih.gov/ Blast.cgi) and compared with the sequences obtained from the National Center for Biotechnology Information (NCBI) GenBank database (http://www.ncbi.nlm.nih. gov/genbank/).

RESULTS
A total of 73 species and 14 genera were identified from 9,535 individuals (♀4,089/♂5,446) ( A sample coverage analysis indicated that sandfly populations were sufficiently represented in all environments. The CUN, FE and PE environments yielded 5,847, 2,111 and 1,457 individuals and 68, 58 and 47 species, respectively, at 99% sample coverage (Table I).
Species diversity was the highest in the CUN environments, followed by the FE and PE environments. The CUN environments yielded the highest Shannon index (H') = 19.5 common species and Simpson index (1/D) = 6.9 dominant species, followed by FE with H' = 14.5 common species and 1/D = 6.9 dominant species and PE with H' = 10.3 common species and 1/D = 5.2 dominant species (Table I, Fig. 2).
A total of 1,755 females were divided into 274 pools representing 35 species. The PCR targeting of kDNA and hsp70 identified one pool of Ps. davisi infected with Le. (Vi.) braziliensis (query cover = 100%, identity = 100%, GenBank accession KX573933.1) (Fig. 3). The infected pool was collected from an FE environment in the municipality of Monte Negro.

DISCUSSION
The epidemiological pattern of leishmaniasis in RO is characterised by a zoonotic or sylvatic transmission cycle in which humans might acquire infection via exposure to sandflies in the Amazon rainforest. (5,13,16) A total of 73 species were registered in this study, which demonstrated a higher level of species richness than the previous surveys conducted in RO. (7,13,25) Sand-  fly diversity was the highest in protected environments. The CUN environment exhibited the highest levels of species richness was found in the CUN environment, followed by the FE and PE environments; these findings corroborate those of the previous studies conducted in the Amazon region. (4,21,26,27) Although sampling methods differed between environments, it was still possible to perform reliable diversity comparisons because the sample coverage was 99% in each environment. Our data demonstrated that sandflies might serve as biodiversity indicators. The species richness was reduced by 10 species in the FE environments and by 21 species in the PE environments relative to the CUN environments. This indicates that the reduction of forests to small fragments affects sandfly composition primarily by eliminating the species that occur in minor abundance (rare species). (26,27) The sandfly species Ny. antunesi, Ps. davisi and Th. ubiquitalis were abundant in all environments. Humans generally come in close proximity to forest fragments while engaged in agriculture or activities like hunting and fishing. These activities increase the risk of exposure to Leishmania vectors in possible transmission foci, (25,28) and that risk is exacerbated by the abundance of Ny. antunesi, Ps. davisi and Th. ubiquitalis in the areas we studied. One pool of Ps. davisi tested positive for Le. (Vi.) braziliensis DNA in the municipality of Monte Negro. This finding is significant because Le. brazilien-sis is responsible for 50% of human CL cases in the rural population of Monte Negro. (25,28) Ps. davisi is an abundant species in this region, (25) as well as in other parts of RO, (4,7,13) and Ps. davisi individuals have been previously found to be infected with Le. (Vi.) braziliensis (5) and Le. (Vi.) naiffi. (13) The discovery of this infection in the current study further supports the evidence that Ps. davisi might act as a vector in RO.
Furthermore, Ny. antunesi and Th. ubiquitalis might act as vectors in RO. Both species were found in high abundance in this as well as other studies conducted throughout the state, (4,13,17,25) and the susceptibility of these species to natural infection by Leishmania has been demonstrated in the two studies performed in Porto Velho. (4,17) Furthermore, both species are suspected vectors in AM and MT which border RO. (21,29) Other abundant species found in RO included Ny. whitmani, Ps. carrerai carrerai, Ps. complexus, Ps.  hirsutus hirsutus and Th. auraensis. Ny. whitmani has already been recorded in abundance in RO. (13,25) Ny. whitmani and Ps. hirsutus hirsutus were found in high abundance in the FE environments, suggesting that these species are confined to degraded environments; both species have been associated with dense forest environments (13,25) and with environments impacted by anthropic activities. (25,30) Neither species has been found infected with Leishmania in RO; (4,5,7,17) however, both species are putative vectors in the Amazon region because they have been found infected with Leishmania in PA. (6) In RO, Ps. carrerai carrerai occurs mainly in dense forest environments. (6,7,13) Only three studies conducted in central RO have demonstrated the predominance of Ps. carrerai carrerai and Ps. complexus in this region. (13,25) Ps. carrerai carrerai has been found to carry promastigote flagellates identified as Le. (Vi.) braziliensis (5) and carrying Leishmania DNA. (7)  Trichophoromyia auraensis is abundant primarily in the municipalities of Guajará-Mirim and Porto Velho and in central RO (13,17) where Leishmania DNA has been detected in females. (7,17) In our study, Th. auraensis was found abundant only in the CUN environments; however, the natural infection of Th. auraensis by Leishmania spp has been reported (30) and Th. auraensis has been found in abundance in both FE and PE environments in AC of western Brazil. (19,30) In our analysis of blood meal sources, blood was taken from the stomachs of 15 engorged females, and the PCR amplification of the cytb gene led to the identification of the DNA belonging to humans (H. sapiens), domestic animals (Bos taurus) and sylvatic animals such as anteaters (T. tetradactyla). The DNA belonging to Bos taurus and H. sapiens was present in samples from every collection made in the PE and FE environments, and this DNA was found in the blood taken from 13 female specimens belonging to the species: Bi. flaviscutellata, Ny. antunesi, Pa. dendrophyla, Ps. carrerai carrerai, Ps. davisi and Ps. hirsutus hirsutus. The T. tetradactyla DNA was present in the blood sample taken from two Ny. antunesi females captured in the FE environments. Although 15 engorged females represents a small sample size relative to other studies, our findings are significant because previous studies targeting the cytb gene have detected the DNA of only domestic animals, such as cats, dogs, chickens, bovines, equines and pigs. (18)(19)(20) Preserved environments showed the highest variety of blood meal sources for sandflies, and a large variety of blood meal sources guarantees the maintenance of the gonotrophic cycle. (12) Anteaters (Tamandua spp) act as possible reservoirs for some Leishmania species, such as Le. (Vi.) guyanensis, (12) and in degraded natural habitats, the scarcity of sylvatic reservoirs, such as ant-eaters, might cause sandflies to migrate to the PE environments to find new blood meal sources. (4,18,19,27) In our study, the PE collection points were close to small forest fragments, and the availability of blood meal sources in the form of domestic animals, such as Bos taurus, might have attracted sandflies from the forest fragments.
Few studies have examined the importance of blood meal sources in the Brazilian Amazon. Recently, in the municipality of Rio Branco (AC), the blood collected from the intestinal contents of the two specimens of Ps. davisi was subjected to PCR targeting the cytb gene; this led to the identification of Gallus gallus as a blood meal source. (19) The current study improves our knowledge of blood meal sources by demonstrating that vectors, such as Ny. antunesi and Ps. davisi, feed on humans and bovines in the PE environments and feed on sylvatic animals, such as anteaters in the FE environments.
Our study largely corroborates the findings of previous studies concerned with the transmission cycle of leishmaniasis in RO. However, the fact that sandflies are using humans and domestic animals as blood meal sources indicates that the transmission profile might be changing in the PE environments. These findings can be used to enhance the epidemiological surveillance of leishmaniasis in RO.