Triatomine Feeding Profiles and Trypanosoma cruzi Infection, Implications in Domestic and Sylvatic Transmission Cycles in Ecuador

Understanding the blood meal patterns of insects that are vectors of diseases is fundamental in unveiling transmission dynamics and developing strategies to impede or decrease human–vector contact. Chagas disease has a complex transmission cycle that implies interactions between vectors, parasites and vertebrate hosts. In Ecuador, limited data on human infection are available; however, the presence of active transmission in endemic areas has been demonstrated. The aim of this study was to determine the diversity of hosts that serve as sources of blood for triatomines in domestic, peridomestic and sylvatic transmission cycles, in two endemic areas of Ecuador (central coastal and southern highland regions). Using conserved primers and DNA extracted from 507 intestinal content samples from five species of triatomines (60 Panstrongylus chinai, 17 Panstrongylus howardi, 1 Panstrongylus rufotuberculatus, 427 Rhodnius ecuadoriensis and 2 Triatoma carrioni) collected from 2006 to 2013, we amplified fragments of the cytb mitochondrial gene. After sequencing, blood meal sources were identified in 416 individuals (146 from central coastal and 270 from southern highland regions), achieving ≥ 95% identity with GenBank sequences (NCBI-BLAST tool). The results showed that humans are the main source of food for triatomines, indicating that human–vector contact is more frequent than previously thought. Although other groups of mammals, such as rodents, are also an available source of blood, birds (particularly chickens) might have a predominant role in the maintenance of triatomines in these areas. However, the diversity of sources of blood found might indicate a preference driven by triatomine species. Moreover, the presence of more than one source of blood in triatomines collected in the same place indicated that dispersal of vectors occurs regardless the availability of food. Dispersal capacity of triatomines needs to be evaluated to propose an effective strategy that limits human–vector contact and, in consequence, to decrease the risk of T. cruzi transmission.


Introduction
The identification of species present in blood meals is an important strategy to understand the implication of hematophagous vector species in disease transmission and their ecological interactions. Constant environmental changes, disease outbreaks and increases in vectors' distribution pose challenges for understanding the transmission scenarios of Intestinal contents were extracted from triatomines (235 adults and 272 nymphs). Most of the samples (n = 320, 63.12%) were extracted within 20 days post collection, and 36.88% were extracted from 21 to 40 days post collection. Variation in time from collection to extraction depended on the availability of a field laboratory (rapid containment kit, Germfree, Ormond Beach, FL, USA) to process the insects or the transfer and delayed process in the laboratory at CISeAL. All intestinal content samples were kept at −20 • C until DNA extraction. Triatomines were collected from three transmission cycles: (a) In the intradomicile (indoors), triatomines were collected in any structure inside the house, such as rooms (e.g., wall, bed, between clothes), piles of material (accumulation of tiles or bricks), animal shelters (bird or rat/mice nests). (b) The peridomicile cycle was considered the space around the house (usually delimited by a fence). Triatomines in the peridomicile were collected in piles of material (e.g., bricks, wood) or animal shelters of domestic and nondomestic animals (e.g., chicken, pigeon and rat/mice nests, dog and cat beds). (c) The sylvatic cycle (outside household limits) was mainly associated with animal shelters of non-domestic animals (e.g., squirrel, rat and bird nests), although it included areas where human activities can take place (e.g., forest patches with trails, areas for agriculture and harvesting, etc.).

Triatomine Selection
Analyses were carried out in 507 selected samples of DNA purified from the intestinal content of five species of triatomines (60 P. chinai, 17 P. howardi, 1 P. rufotuberculatus, 427 R. ecuadoriensis and 2 T. carrioni) previously collected from 2006 to 2013 in the central coastal (CC) (n = 171) and southern highland (SH) (n = 336) regions of Ecuador ( Figure 1). Intestinal contents were extracted from triatomines (235 adults and 272 nymphs). Most of the samples (n = 320, 63.12%) were extracted within 20 days post collection, and 36.88% were extracted from 21 to 40 days post collection. Variation in time from collection to extraction depended on the availability of a field laboratory (rapid containment kit, Germfree, Ormond Beach, FL, USA) to process the insects or the transfer and delayed process in the laboratory at CISeAL. All intestinal content samples were kept at −20 °C until DNA extraction. Triatomines were collected from three transmission cycles: (a) In the intradomicile (indoors), triatomines were collected in any structure inside the house, such as rooms (e.g., wall, bed, between clothes), piles of material (accumulation of tiles or bricks), animal shelters (bird or rat/mice nests). (b) The peridomicile cycle was considered the space around the house (usually delimited by a fence). Triatomines in the peridomicile were collected in piles of material (e.g., bricks, wood) or animal shelters of domestic and non-domestic animals (e.g., chicken, pigeon and rat/mice nests, dog and cat beds). (c) The sylvatic cycle (outside household limits) was mainly associated with animal shelters of non-domestic animals (e.g., squirrel, rat and bird nests), although it included areas where human activities can take place (e.g., forest patches with trails, areas for agriculture and harvesting, etc.). Geographic origin of 507 triatomines included in the analysis of blood meal sources. Each point indicates a locality and the different shapes and colors the triatomine species that were analyzed in each locality, as described in the legend. Central coastal corresponds to Manabí province and southern highlands to Loja province, from which 171 and 336 samples, respectively, were analyzed. Black lines represent the province limits. Each point indicates a locality and the different shapes and colors the triatomine species that were analyzed in each locality, as described in the legend. Central coastal corresponds to Manabí province and southern highlands to Loja province, from which 171 and 336 samples, respectively, were analyzed. Black lines represent the province limits.

DNA Extraction and Natural Infection with Trypanosomes
DNA was extracted from the homogenized intestinal content with the DNAzol reagent for the isolation of genomic DNA (Invitrogen TM , Waltham, MA, USA) in a DNA extraction hood that was sterilized with UV light before and after each use. All DNA samples were stored at −20 • C until use. Infection with T. cruzi was assessed by amplification of the variable domain of the minicircle of kDNA by the presence of a 330-bp product [23,24]. A negative sample (without DNA) and a positive control with DNA of T. cruzi were used in each PCR assay. Amplification products were detected in a 2% agarose gel stained with 1X SYBR Safe (Thermo Fisher Scientific, Waltham, MA, USA).

Sensitivity of the Different Primer Sets
The best performance of primer sensitivity using human DNA serial dilutions was achieved for the primers FACytb, where the amplification of a 355-pb fragment of cytb was obtained using up to 1 pg of DNA. For the two other primer sets (L14841-H15149 and L1091-H1478), the expected bands were obtained using up to 10 pg of DNA.

Amplification and Identification of the Sources of Blood Meals
Intestinal content DNA samples were amplified with the primer set FACytb, following the amplification conditions described in Table S1. Preparation of the mix was carried out in a pre-PCR hood sterilized with UV light before and after each use. A negative sample (without DNA) and a positive control with human DNA were used in each PCR assay. The expected 355-bp fragment was visualized in 2% agarose gels after SYBR Safe staining. Amplification products were sent for sequencing to Macrogen, Inc. (Seoul, Korea). Forward and reversed sequences were aligned with Mega V6.0 (https://www.megasoftware.net/ manual.pdf). Consensus sequences were compared with DNA nucleotide sequences in NCBI BLAST using the megablast query. Matches were based on results achieving ≥95% maximum identity and cover. Statistical analysis was carried out with Chi 2 and Fisher test to compare blood sources, transmission cycles and geographical regions.

Analyses of Trophic Networks of Triatomines and Their Blood Meals
To visualize the associations between the triatomines and the vertebrates which they feed on, we conducted trophic network analyses. We scored a matrix of interactions by counting the number of individuals of each species of triatomine (n = 5) that fed on a respective vertebrate species (n = 16). This matrix was analyzed with the functions plotweb and visweb using the library bipartite [27] in R version 3.6.2 [28].

Vertebrate Diversity as Source of Blood for Triatomines
Of the 507 samples analyzed (Table S2), 82.05% of the samples (n = 416) were identified to blood meal species due to the sequences that matched with the ones in Genbank with an identity ≥ 95%. For 10.4% (n = 53) of the samples, PCR products were not obtained, and for 1.4% (n = 7), non-legible chromatograms or the presence of multiple peaks were present. For 6.1% (n = 31) of the samples, the obtained sequences matched with a reported species in GenBank, but the identity was <95%; among these last samples, we found birds (Gallus gallus, Gallus lafayetii), primates (Homo sapiens), rodents (Rhipidomys leucodactylus, Simosciurus stramineus, Proechimys semispinosus, Rattus satarae and Rattus rattus) and three species of insects that do not occur in the study area (Rhodnius pictipes, Kleidocerys resedae and Tiniocellus spinipes). None of the sequences with identity < 95% were included in the analysis.
For the 416 triatomines where blood meals were identified, sixteen vertebrate species were observed (Table 1). Trophic patterns of the five studied vector species are summarized in Figure 2. The class Mammalia represented 67.8% (n = 282) of the analyzed samples, with ten species, including five species of rodents (Aegialomys xanthaeolus, Mus musculus, Rattus norvegicus, R. rattus, S. nebouxii (previously known as Sciurus stramineus)), one carnivore (Canis lupus familiaris), two artiodactyla (Capra hircus, Sus scrofa), one didelphimorphia (Didelphis marsupialis) and one primate (H. sapiens). Four species of the class Aves were identified (G. gallus, Campylorhynchus fasciatus, Columba livia and Leptotila verreauxi decolor) and represented 31.2% (n = 130) of the analyzed samples. One species of the class Reptilia (Boa constrictor imperator) and class Amphibia (Rhinella marina) (previously known as Bufo marinus) was identified in <1% (n = 4) of the samples. Overall, the most abundant vertebrate hosts identified as sources of blood for triatomines were humans (H. sapiens) (n = 170, 40.9%), chickens (G. gallus) (n = 115, 27.6%) and rodents (S. nebouxii and R. rattus) (n = 89, 21.4%).  numbers in red indicate the number of triatomine specimens per species with a given blood meal species, and the intensity of the background color is proportional to these numbers (higher the frequency, darker the background); (b) bipartite graph of the trophic network.
All five triatomine species analyzed were found to have ingested human blood. Only two triatomine species-R. ecuadoriensis and P. chinai-ingested chicken blood and R. ecuadoriensis was the only species of triatomine that had ingested blood of the four recorded bird species (G. gallus, C. fasciatus, C. livia and L. verreauxi). In P. howardi, blood meal from the amphibian (R. marina) and the reptile (B. c. imperator) was found. Only human blood meal was identified in the few specimens of T. carrioni and P. rufotuberculatus ( Figure 2).

Vertebrate Blood Meal Source per Geographic Area, Triatomine Species, Developmental Stage and Transmission Cycle
The blood meal distribution per species in the two regions is presented in Figure 3a.
Overall, blood meal vertebrate diversity appeared higher in CC (14 blood meal sources) than in SH (eight blood meal sources), but this difference was not significant (Fisher exact test, p > 0.05). A similar observation was made for R. ecuadoriensis, the only studied species present in both regions; of a total of 12 blood meal hosts, 11 were identified in CC and 6 in SH (Fisher exact test, p > 0.05). For P. howardi, a species only found in CC, five sources of blood were identified, including three mammals (Human, R. rattus and S. scrofa), one species of amphibian (R. marina) and one species of reptile (B. constrictor imperator); these last two are non-typical blood sources for triatomines.  To refine our analysis, we examined the sources of blood meals of the triatomines according to the microenvironment in which they were captured, as presented in Figure 4. In the intradomicile, triatomines were mostly found in rooms (n = 56, 67.4%), particularly in bedrooms (in the bed, between clothes and on the wall). As expected, in this transmission cycle, the main source of blood was human (54.2%), although species of other orders were found, such as Galliformes (G. gallus), Rodentia (R. rattus), Artiodactyla (C. hircus) and Carnivora (C. lupus familiaris). Triatomines were also found in piles of material (bricks or wood) stored inside the house. Most of the triatomines found in this microenvironment had Rodentia blood (R. In the SH region, human was the most abundant source of blood for triatomines, particularly for the specimens collected in the intradomicile (P. chinai, P. rufotuberculatus, R. ecuadoriensis and T. carrioni), but also it was frequently found in the species that infested the peridomestic and sylvatic cycle (P. chinai, and R. ecuadoriensis). In the CC region, human blood meal was found in all three transmission cycles but was particularly abundant in R. ecuadoriensis from the peridomicile and sylvatic cycles (Figure 3a). For R. ecuadoriensis, chicken (G. gallus) and rodents (S. nebouxii, R. rattus) are other important sources of blood. For P. chinai, rodents (R. rattus) and, to a lesser extent, ungulates (Artiodacyla-C. hircus), Pathogens 2021, 10, 42 8 of 17 carnivores (C. lupus familiaris) and birds (G. gallus) were also detected as important sources of blood (Figure 3a).
In both regions, sources of blood meal were identified in adults and nymphs. Overall, no significant distribution of the main blood meals (human, chicken (G. gallus) and rodent (R. rattus and S. nebouxii)) was observed between adults and nymphs (Fisher exact test, p = 0.11; Figure 3b), even if a higher diversity of hosts was observed in nymphs (15 blood meal sources) than in adults (10 blood meal sources), but this was not significant (Fisher exact test, p > 0.05) (Figure 3b).
Blood meal sources were identified in triatomines collected in the intradomicile, peridomicile and sylvatic cycles. In total, considering three categories of blood meals, the two most frequent ones (human and G. gallus) and a third category grouping all the others, the observed distribution was significantly different between the transmission cycles (Chi 2 = 96.14, df = 4, p < 0.0001). Humans tend to be overrepresented in blood meals of triatomines captured in the intradomicile cycle (the observed frequency was 32.7% greater than the expected) and, to a lesser extent, in those captured in the sylvatic cycle, where the observed frequency was 17.8% greater than expected. In the peridomicile, G. gallus were overrepresented in the blood meals of triatomines and the observed frequency was 90.4% greater than expected.
To refine our analysis, we examined the sources of blood meals of the triatomines according to the microenvironment in which they were captured, as presented in Figure 4. In the intradomicile, triatomines were mostly found in rooms (n = 56, 67.4%), particularly in bedrooms (in the bed, between clothes and on the wall). As expected, in this transmission cycle, the main source of blood was human (54.2%), although species of other orders were found, such as Galliformes (G. gallus), Rodentia (R. rattus), Artiodactyla (C. hircus) and Carnivora (C. lupus familiaris). Triatomines were also found in piles of material (bricks or wood) stored inside the house. Most of the triatomines found in this microenvironment had Rodentia blood (R. rattus); however, blood from other sources, such as human, goat (Artiodactyla) and dog (Carnivora), was also found. Another microenvironment present in the intradomicile (although less frequent) was animal shelters (chicken and rodent nests). Triatomines collected in this microenvironment presented blood from chicken and human. In the few triatomines collected indoors by the inhabitants (triatomines attracted by light), only human blood was detected (Figure 4).
In the peridomicile, most of the studied triatomines were collected in animal shelters (mostly chicken nests and some others in rodent nests). In these microenvironments, chicken or rodent blood meals were the most frequent, but human blood meals were also detected. Triatomines collected in piles of material (mainly bricks and woods) presented different blood meal sources such as R. rattus, humans, S. scrofa, B. c. constrictor and R. marina (Figure 4).
In the sylvatic cycle, all studied triatomines were collected in animal shelters, mainly squirrel nests (87.0%). A total of 11 different blood meal sources were identified, with human and the squirrel, S. nebouxii, being the most abundant. A small proportion of triatomines had blood meals of domestic and non-domestic species of the order Rodentia (R. rattus, A. xanthaeolus and M. musculus), Carnivora (C. lupus familiaris), Didelphimorphia (D. marsupialis), Galliformes (G. gallus), Columbiformes (L. v. decolor, C. livia) and Passeriformes (C. fasciatus) (Figure 4).

T. cruzi Infection and Blood Meal Diversity in the Different Transmission Cycles
Of the 416 triatomines with identified blood meals, 65.9% were infected with T. cruzi. Table 1 and Figure 5 summarize the infection rates observed according to each blood host species and transmission cycle. The infection rate was similar in triatomines fed with mammal and bird blood (64.2% and 69.2%, respectively, Chi 2 = 0.04, p > 0.05). Similarly, adults presented a T. cruzi infection rate (67.6%) not different from that of nymphs (64.5%) (Chi 2 = 0.31, p > 0.05). rattus); however, blood from other sources, such as human, goat (Artiodactyla) and dog (Carnivora), was also found. Another microenvironment present in the intradomicile (although less frequent) was animal shelters (chicken and rodent nests). Triatomines collected in this microenvironment presented blood from chicken and human. In the few triatomines collected indoors by the inhabitants (triatomines attracted by light), only human blood was detected ( Figure 4). In the peridomicile, most of the studied triatomines were collected in animal shelters (mostly chicken nests and some others in rodent nests). In these microenvironments, chicken or rodent blood meals were the most frequent, but human blood meals were also detected. Triatomines collected in piles of material (mainly bricks and woods) presented different blood meal sources such as R. rattus, humans, S. scrofa, B. c. constrictor and R. marina (Figure 4).
In the sylvatic cycle, all studied triatomines were collected in animal shelters, mainly squirrel nests (87.0%). A total of 11 different blood meal sources were identified, with human and the squirrel, S. nebouxii, being the most abundant. A small proportion of triatomines had blood meals of domestic and non-domestic species of the order Rodentia (R. rattus, A. xanthaeolus and M. musculus), Carnivora (C. lupus familiaris), Didelphimorphia (D. marsupialis), Galliformes (G. gallus), Columbiformes (L. v. decolor, C. livia) and Passeriformes (C. fasciatus) (Figure 4).

T. cruzi Infection and Blood Meal Diversity in the Different Transmission Cycles
Of the 416 triatomines with identified blood meals, 65.9% were infected with T. cruzi. Table 1 and Figure 5 summarize the infection rates observed according to each blood host species and transmission cycle. The infection rate was similar in triatomines fed with mammal and bird blood (64.2% and 69.2%, respectively, Chi 2 = 0.04, p > 0.05). Similarly, adults presented a T. cruzi infection rate (67.6%) not different from that of nymphs (64.5%) (Chi 2 = 0.31, p > 0.05). Triatomine infection rates with T. cruzi collected in the domicile and sylvatic cycle reach around 60% (Chi 2 = 0.02, p > 0.05); these values were significantly lower than those of peridomestic cycle (Chi 2 values of 4.31 and 5.43, respectively, p < 0.05). In the domestic cycle, the infection rate of triatomines with mammal blood meals (61.4%) or bird blood meals (53.8%) were not significantly different (Fisher exact test, p > 0.05). Additionally, the infection rate in triatomines with rat blood (80.0%) tends to be higher than those with human blood (53.3%), but this difference is not statistically significant (Fisher exact test, p > 0.05).
In the peridomestic cycle, eleven species of vertebrates were identified as sources of Triatomine infection rates with T. cruzi collected in the domicile and sylvatic cycle reach around 60% (Chi 2 = 0.02, p > 0.05); these values were significantly lower than those of peridomestic cycle (Chi 2 values of 4.31 and 5.43, respectively, p < 0.05). In the domestic cycle, the infection rate of triatomines with mammal blood meals (61.4%) or bird blood meals (53.8%) were not significantly different (Fisher exact test, p > 0.05). Additionally, the infection rate in triatomines with rat blood (80.0%) tends to be higher than those with human blood (53.3%), but this difference is not statistically significant (Fisher exact test, p > 0.05).
In the peridomestic cycle, eleven species of vertebrates were identified as sources of blood meal (seven mammal, two bird, one reptile and one amphibian species); no significant differences in infection rates were observed between the principal sources of blood meals, chickens (71.1%), rats (68.4%) and humans (83%) (Fisher exact test, p > 0.05). In this cycle, T. cruzi was detected in one of the two amphibian blood meals and in both samples with reptilian blood meals (Table 1, Figure 5).
In the sylvatic cycle, species of vertebrate hosts (7 species of mammals, 4 species of birds) were found as sources of blood for triatomines, with mammals being the most frequent source, with a 58.8% T. cruzi infection rate. The mammal group included domestic and synanthropic animals, such as dogs, opossums, mice and rats, and sylvatic ones, such as yellowish rice rats and squirrels. Surprisingly, human blood was found in 48.4% (n = 78) of the sylvatic triatomines, with a T. cruzi infection rate as high as 62.8%. Overall, the T. cruzi infection rate in birds (domestic and wild) reached 73.1% (Table 1, Figure 5). A non-statistical difference was found when we compared human, squirrel and bird blood meals (Fisher exact test, p > 0.05).

Epidemiological Importance of the Identification of the Vertebrate Sources of Blood for Triatomines
Triatomines feed on a wide diversity of vertebrate species, and the detection of these hosts in a given area is essential to understand the interactions between triatomines and vertebrate hosts and to identify important reservoir hosts of CD. In Ecuador, a great deal of information has been obtained about the diversity and ecological conditions of triatomines and T. cruzi [16,17,29,30], which is crucial to define target species of triatomines for their control. Although insecticide spaying was successful to control T. dimidiata populations, limited effectiveness was observed when applied to the control of other domiciliated species [31]. Likewise, the continuous reinfestation of houses constitutes a challenge for effective control measures to avoid human-vector contact.
To find alternative and comprehensive measures to fight against CD transmission, information about the contact between the triatomine vectors and the vertebrate hosts is essential. The identification of the sources of blood meal, and, at the same time, the detection of T. cruzi infections, contributes to: (1) defining the rate of human-vector contact, which informs us about the risk of CD transmission, (2) determining the most frequent sources of blood, other than humans, that allow vector populations to thrive, (3) the infection rate of the vectors that is considered a second risk index of CD transmission, (4) the potential mammal reservoirs that could be targeted for CD control, (5) the knowledge of the dynamics of transmission through the cross-analysis of the place of capture of the vectors, the source of the blood meal of each vector and the proximity of the potential hosts, and (6) the vector species composition and their ability to transmit the parasite.
The present study answered some of the questions listed above. Among others, this study described the great diversity of vertebrate hosts that are sources of blood for the main vector species in Ecuador, the high T. cruzi infection rate of triatomines, even when they are captured in the nests of fowls (laying hens or other birds), which are refractory to T. cruzi infection, and the high frequency of human blood meals in vectors captured indoors but even in those captured outside the home.

Diversity of Mammal Hosts and Its Implication in T. cruzi Transmission
Mammal species are recognized as frequent blood sources and the most important because these species can be infected and transmit T. cruzi. In this study, mammals contributed 67.8% of the identified feeding sources; of these, human, black rat (R. rattus) and squirrel (S. nebouxii) represented together 91.8%. The two first species were found in triatomines collected in the domestic, peridomestic and sylvatic cycles, while squirrels contributed above all to blood sources in the sylvatic cycle.
Human blood has been previously found in the three principal triatomine genera implicated in T. cruzi transmission (Panstrongylus, Rhodnius and Triatoma) [32][33][34][35]; however, the high proportion of human blood meals in sylvatic triatomines in both regions, while they have different ecological conditions, was unexpected. Contact of humans with sylvatic triatomines has been proposed to be related to outdoor activities (e.g., camping, harvesting) or to the movement of triatomines from sylvatic to domestic settings and vice versa [36][37][38]. In the present context, such human activities and possible human-vector contact outside the home as well as triatomine dispersal need to be further studied.
Endemic and wild species of rodents are important blood sources for triatomines, particularly in the sylvatic cycle [36,[39][40][41]. In this study, apart from humans, triatomines mostly fed from squirrels (S. nebouxii) and a few other rodent species. In previous studies, sylvatic triatomines have been searched in burrows, tree holes, under trunks and rocks, but they were mainly collected in nests built in trees, particularly in squirrel nests [18,19,42]. Although in CC, the blood meals of sylvatic triatomines were more diverse and squirrels did not seem to be an important source of blood, in SH, these results support the importance of S. nebouxii in the maintenance of triatomine populations in forest patches around rural communities [43].
Synanthropic rodents such as rats (R. rattus) have been found to be key sources of blood for triatomines in some areas such as Bolivia, Mexico (where rats play a major role as feeding source and should be the main reservoir of peridomicile colonies of T. longipennis) or Guatemala (where rodent control was applied as part of a strategy for an integrated vector management) [36,[44][45][46]. However, in Panama, Argentina, Venezuela and Colombia, rats are not an important source of blood for triatomines [32][33][34][35]. In our studied areas, the results showed a low frequency of feeding from rat blood (15.6% of mammal blood meals), although rats have the capacity to circulate between sylvatic and domestic cycles. The contact of rats with triatomines might be less frequent than expected, even if they are widely distributed; thus, in Ecuador, their role in T. cruzi transmission would be more dependent on local ecological characteristics.
The opossum is considered a natural host species for the parasite and it is an important reservoir in some areas in Colombia and Brazil [35,[46][47][48]. However, based on this and a previous study [21], opossums seem to have a limited role in triatomine maintenance and T. cruzi transmission in Ecuador, although this is an abundant species.
The role of dogs also differs by region. In general, in Chagas-disease-endemic areas, there is a high frequency of dogs infected with T. cruzi and their presence in dwellings has been associated with higher risk of human infection [47,49]. In Colombia, dogs have been found as an important source of blood for T. maculata [35]. In other areas, such as the USA (Texas and Louisiana) and Brazil, the T. cruzi seroprevalence of dogs ranged from 7 to 76% [50][51][52][53]. In Northern Argentina, infected dogs constitute a risk factor and have been used as sentinels for T. cruzi domestic transmission [54,55]. However, in other areas, dogs might not have such an important role, although they have been found to be a less frequent source of blood [56,57]. In areas of Ecuador in which Chagas disease is endemic, dogs are present in the majority of domiciles [16,17]; nevertheless, the results of this study indicate that dogs are not a major source of food for triatomines, despite their availability.
Other mammals, such as the mouse (M. musculus), the brown rat (R. norvegicus), goats (C. hircus) and pigs (S. scrofa), might be occasional sources of food and could play a limited role in T. cruzi transmission in Ecuador.

Birds and Their Role in Triatomine Population Maintenance
Birds do not have a direct role in T. cruzi transmission because they are refractory to the parasites; they are able to destroy the trypomastigote form through the action of their immune system. However, they are a source of blood that plays an important role in the survival and maintenance of triatomine populations [12], particularly in rural areas where raising chickens is frequent. Experimental studies in T. infestans have demonstrated that this vector species prefers feeding from dogs than from chickens [58]. In Ecuador, having chickens was found to be a major determinant of triatomine infestation and R. ecuadoriensis is the vector species that showed a preference for this microenvironment [16,17,59]. Most chickens are kept free-ranging, except for laying hens, whose nests are located outside against the house walls so to be easily monitored by the inhabitants [60]. Large colonies of triatomines have been found in these nests. Although chickens are the expected source of blood for triatomines collected in chicken nests (63.6%), we also found human (31.8%) and rodent blood meals (4.6%). Moreover, triatomines that contained chicken blood were found infected by T. cruzi. These last findings show that triatomines, adults and nymphs, move from one microenvironment to another; thus, the dispersal of triatomines between the intradomicile, peridomicile and sylvatic cycles might play an important role in transmission. According to the present results, other birds, such as doves, pigeons and others, might contribute to the maintenance of triatomine populations in the sylvatic cycle.

Triatomine Species and Blood Meal Preference
Rhodnius ecuadoriensis is the most abundant and widely distributed species found in the two studied regions; moreover, it colonizes all three transmission cycles [16,17]. In both regions, given that blood meals taken from humans are frequent, regardless of the transmission cycle where the bugs were collected (27.6% in peridomicile to 58.1% in intradomicile), it can be argued that this species has remarkable anthropophilic behavior. The most conclusive observations are those made in specimens collected in the sylvatic cycle: in CC, although a variety of hosts is observed (10 species of vertebrates counted), the human food represents 57.3%; in SH, they are slightly less represented (41.5%) since this species is often found in the nests of squirrels and meals from them reached 46.8%. Among these insects, around half are adults who have thus moved to rely on humans for food, while the nymphs feed from their congeners engorged with human blood, as discussed in the next section.
In P. chinai, whose presence is well established in domestic and peridomestic cycles in SH only [17], of the 51 specimens studied, around half have also fed on humans. R. rattus seems to be the second most preferred species by P. chinai, when collected in intradomicile, while this is not the case for R. ecuadoriensis, which shares the same region and same villages. Detection of other sources of blood, such as the dog, goat, rat and chicken, however, shows flexibility in food choices, a characteristic conducive to the adaptation of species to new environments.
What is striking for P. howardi compared to other species is the diversity of diet observed in specimens collected in the peridomicile, which included humans, rats, reptiles and amphibians; no chicken blood meals were detected, even though chickens are abundant in the peridomicile, but the sample size was small.
All these results might indicate that although triatomines have a diverse diet, some species showed blood source preferences that can have, as in the case of R. ecuadoriensis, important implications in T. cruzi transmission to humans.

Triatomine Dispersal and Access to Vertebrate Hosts
To understand the dynamics of transmission in an endemic area, knowing the movements of triatomines is very informative. In this context, the current cross-study that considered the places where each triatomine was captured with its blood meal offers valuable information, even if the study was limited to a single meal per triatomine. Indeed, several triatomines had multiple meals (1.4%) that were reflected by multiple peaks in the chromatograms, showing an overlap of cytb sequences. Although the resolution of the multiple meals by molecular cloning was not applied, detecting multiple sources of blood could facilitate the understanding of the ecological interactions of triatomines [6,61].
In this study, triatomines collected in chicken nests in the peridomicile and squirrel nests in the sylvatic cycle presented unexpectedly a high diversity of blood sources. Furthermore, triatomines collected in all transmission cycles, including sylvatic, showed the presence of human and chicken blood, showing that these triatomines had fed elsewhere before arriving in these places. The dispersal capability of triatomines has been previously reported. In Argentina, the dispersion distance range of adult T. infestans has been recorded from 0.1 up to 2 km [62,63]. In Mexico, an elevated dispersal between domestic and sylvatic areas was suggested for T. dimidiata [64]. In the Andean valley in Bolivia, the evaluation of dispersal distances of T. infestans by the capture-mark-recapture method in the field demonstrated significant movement in adults of up to 168 m and in nymphs of 34 m [65].
Unexpected blood meals of nymphs that differ from the available host in the place of capture are not easy to explain. Nymphs have a limited capacity for dispersion due to the lack of wings, and in some cases, it is not possible to assume that the nymph located its meal by walking, so other strategies such as cleptohematophagy or cannibalism, which have been previously reported under experimental conditions for some triatomine species, can be proposed [66][67][68]. Alternatively, nymphs could be passively dispersed by transportation by animals or humans, as has been reported for Triatoma brasiliensis, Rhodnius prolixus and Triatoma pseudomaculata [69][70][71].
Host accessibility is a major factor that shapes the patterns of the sources of blood for triatomine, and it varies according to the studied area [72]. In Panama, anthropogenic landscape disturbance changed the host community structure and increased R. prolixus infection with T. cruzi [15]. Conversely, in Argentina, domestic animals and their availability throughout the year affect triatomine´s host-feeding choices and even human-vector contact rates [73]. In Ecuador, triatomines have accessibility to domestic animals such as dogs and chickens [16]. Although chickens are an important blood source for R. ecuadoriensis, humans constitute the most important blood source for all five vector species analyzed. In Ecuador, the dispersal and dynamics of triatomines, particularly R. ecuadoriensis, need to be further evaluated to determine the factors that influence the human-vector contact in the domestic cycle and the possible scenario that can explain human blood meals in specimens collected in the sylvatic cycle.

Conclusions
Humans are the main source of blood for the current sample of triatomines collected in Ecuador, indicating that human-vector contact is more frequent than previously thought. This result, together with the high infection rates with T. cruzi, indicates a high risk of transmission of the parasite in the central coastal and southern highland regions of Ecuador. Moreover, the high frequency of chicken blood in peridomestic triatomines indicates that this species is fundamental in the maintenance and thriving of triatomine populations in household environments. The triatomine R. ecuadoriensis, which is the most abundant species in both studied regions and widely distributed in the domestic and sylvatic cycles, appeared to prefer human and chicken blood, having a principal epidemiological role. Moreover, other species that colonized peridomestic environments, such as P. chinai and P. howardi, might prefer human and rodent sources of blood but not chicken, taking second place in the risk of transmission. Both adults and nymphs, who are similarly infected with T. cruzi, come into contact with humans; even nymphs, which have limited dispersal because they cannot fly, can acquire human blood by other mechanisms such as cleptohaematophagy on their congeners. The dispersion profile of triatomines is obviously very complex. They are influenced by the availability of vertebrates, their own trophic preferences, but also the environmental disturbances that lead them to move and adapt to new sources of blood. The current results show that adults move from the outside to the inside of the house and vice versa, and the same applies for nymphs in some situations, such as that of the nests of laying hens located against the outside walls of houses. Both adults and nymphs therefore represent a risk of transmission. In this context, additional studies will provide a better understanding of the dispersal patterns of vector species, but it