Dispersion and oviposition of Aedes albopictus in a Brazilian slum: Initial evidence of Asian tiger mosquito domiciliation in urban environments

Aedes albopictus, originally considered as a secondary vector for arbovirus transmission, especially in areas where this species co-exist with Aedes aegypti, has been described in most regions of the world. Dispersion and domiciliation of Ae. albopictus in a complex of densely urbanized slums in Rio de Janeiro, Southeastern Brazil, was evidenced. In this study, we tested the hypotheses that 1) Ae. albopictus distribution in urban slums is negatively related to distance from vegetation, and 2) these vectors have taken on a domestic life style with a portion of the population feeding, ovipositing, and resting indoors. To do this, we developed an integrated surveillance proposal, aiming to detect the presence and abundance of Aedes mosquitoes. The study, based on a febrile syndrome surveillance system in a cohort of infants living in the slum complex, was performed on a weekly basis between February 2014 and April 2017. A total of 8,418 adult mosquitoes (3,052 Ae. aegypti, 44 Ae. albopictus, 16 Ae. scapularis, 4 Ae. fluviatilis and 5,302 Culex quinquefasciatus) were collected by direct aspiration and 46,047 Aedes spp. eggs were collected by oviposition traps. The Asian tiger mosquito, Ae. albopictus, was aspirated in its adult form (n = 44), and immature forms of this species (n = 12) were identified from the eggs collected by the ovitraps. In most collection sites, co-occurrence of Ae. aegypti and Ae. albopictus was observed. Key-sites, such as junkyards, thrift stores, factories, tire repair shops and garages, had the higher abundance of Ae. albopictus, followed by schools and households. We collected Ae. albopictus at up to 400 meters to the nearest vegetation cover. The log transformed (n+1) number of females Ae. albopictus captured at each collection point was inversely related to the distance to the nearest vegetation border. These results show that Ae. albopictus, a competent vector for important arboviruses and more commonly found in areas with higher vegetation coverage, is present and spread in neglected and densely urbanized areas, being collected at a long distance from the typical encounter areas for this species. Besides, as Ae. albopictus can easily move between sylvatic and urban environment, the entomological monitoring of Ae. albopictus should be an integral part of mosquito surveillance and control. Finally, key-sites, characterized by high human influx and presence of potential Aedes breeding sites, should be included in entomological monitoring.


Introduction
Different arboviruses, such as dengue, chikungunya, Zika and yellow fever are transmitted to humans by mosquitoes of the genus Aedes (Meigen 1818), particularly Ae. aegypti (Linnaeus, 1762) and Ae. albopictus (Skuse, 1894), two invasive and frequently sympatric species. Although Ae. albopictus is considered to have a low capacity to transmit pathogens (as arboviruses) to humans, it has been demonstrated the potential role of this species in dengue, chikungunya and Zika virus transmission and outbreaks [1,2]. The domestic form of Ae. aegypti is highly anthropophilic, predominating in urban and suburban areas, where households and humans are abundant. However, Ae. aegypti is often found in transition areas between highly urbanized and urban forest, which might serve somehow as a refuge [3,4,5,6,7]. Furthermore, Ae. albopictus is typically more common in areas with higher vegetation coverage and more scattered human populations, but it has also been described in transitional environments with relatively low vegetation cover and frequently coexisting with Ae. aegypti [3,5,7,8,9]. Knowledge of the species habitat and environmental determinants is essential for predicting Ae. aegypti and Ae. albopictus presence and abundance in an area, which might impact arboviruses transmission. In this study, we tested the hypotheses that 1) Ae. albopictus distribution in urban slums is negatively related to distance from vegetation, and 2) Ae. albopictus has taken on a domestic life style with a portion of the population feeding, ovipositing, and resting indoors.

Materials and methods
The study was conducted in Manguinhos (22˚52' 44,2 S 43˚14' 42,0 W), a low income urban slum complex comprised by 16 different densely urbanized communities. Within an area of 261 square kilometers, a population of 36,160 inhabitants (138 inhabitants per km 2 ) live in 10,816 households [10], characterized by a crowded housing, narrow alleys, inadequate sanitation, irregular domestic water supplies and haphazard waste management. Violence and constant police incursions make Manguinhos a difficult neighborhood for research activities and entomological monitoring. Low vegetation is common in Manguinhos, although there are some green delimited areas in the community, such as the Fiocruz campus (Fig 1), a river, and other waterways.
A mosquito surveillance integrating a large-scale dengue infant cohort study [9] was conducted from February 2014 to April 2017 in the study area. Ethical clearance was obtained from the Ethical Committee in Research (CAAE: 13202113.1.0000.5240) from the National School of Public Health, Oswaldo Cruz Foundation, Ministry of Health, Brazil. Each participant signed informed consent. Adult mosquito collections were performed weekly integrated to the cohort study. We used portable backpack aspirators in several collection sites: households, schools and key-sites (such as junkyards, thrift stores, factories, tire repair shops and garages), the last two defined as non-residential properties suitable for the maintenance of vector infestation. Households were visited after the report of fever in any children followed-up in the cohort study. Schools and key-sites, selected in strategic areas in the vicinity of the fever cases, were characterized by high human influx and presence of potential Aedes breeding sites. Adult mosquito sampling was performed for 15-20 minutes in each collection site. All sampling was performed by the same technicians throughout the whole study period. Aspirations were positive when at least one mosquito was collected. Adult mosquitoes were counted, sexed and identified to species level using the taxonomic key of Consoli and Lourenço-de-Oliveira [11] and stored in freezer (-80˚C). Oviposition was monitored placing 45 ovitraps weekly in the schools from October 2015 to May 2016, totaling 806 observations. Wooden paddles were collected weekly and inspected for the presence of eggs, which were counted and hatched to identify larvae species. Collection points were geo-referenced, and distance to the nearest vegetation cover was measured using QGIS 2.18. Shape files are publicly available and free to use at Rio de Janeiro's Municipal Data Repository (http://www.armazemdedados.rio.rj.gov.br/). We modeled the log transformed (n+1) abundance of collected Aedes female mosquitoes and the distance from each collection point to the nearest vegetation patch in the study area using a simple linear model. Analyses were carried out in R and RStudio [12,13].
Both species were collected during all seasons. Aedes aegypti was more abundant during the wet season, peaking in December with abundance declining abruptly in May. Aedes albopictus peaked in June declining afterwards. For both species, more specimens were collected in the key-sites and schools. Throughout the study period Ae. aegypti was present in all three different collection sites, while Ae. albopictus maintained a positive mean only in key-sites. In schools and households, Ae. albopictus was less often found (Fig 2). Both Ae. aegypti and Ae. albopictus were strongly correlated (R = 0.87, p < 0.05). The number of Ae. albopictus adult females retrieved in the urbanized area fitted an exponential regression curve with the distance to the nearest vegetation border (R = 0.65; p < 0.0001). The log transformed (n+1) number of females Ae. albopictus captured at each collection point was inversely related to the distance to the nearest vegetation border (Fig 3). The regression curve parameters estimated were: y = 260.5629 Ã exp (-0.2545 Ã x), where y represents the distance to the nearest vegetation border and x the number of female Ae. albopictus retrieved. Regression curve parameters estimated for the linear regression curve: y = a+bx, were a = 253.3969 and b = -43.4897. The number of Ae. aegypti adult females retrieved near the vegetated areas also fitted an exponential regression curve, with the number of females decreasing exponentially with the distance to the nearest vegetation border (p < 0.0001). The number of females of Ae. aegypti captured at each collection point weighted according to the number of collections performed at each site, was inversely related to the distance to the nearest vegetation border (Fig 3). The regression curve parameters estimated were: y = 307.6928 Ã exp (-0.2368 Ã x), where y represents the distance to the nearest vegetation border and x the number of female Ae. aegypti retrieved (Fig 3).

Discussion
In our study, the entomological survey followed reports of febrile children, in a routine entomological surveillance. Both Ae. aegypti and Ae. albopictus were collected indoors in an urban endemic area for dengue, Zika and chikungunya, with the number of females of similar magnitude (Table 2). Although Ae. albopictus is typically not commonly found in densely urbanized slums [5], we identified 44 adults and 12 immature forms of this species during the three-year survey in Manguinhos. Aedes albopictus has adapted well to suburban and urban environments, and has been described as the sole vector in urban areas in China and Italy [16,17]. Moreover, there is evidence of a geographical variation in the behavior of this species, with gravid females captured indoors in Italy [18]. The collection of Ae. albopictus adults in densely urbanized slums as Manguinhos complex seems to emphasize the dispersion (a statistical term that describes the distribution of organisms over a landscape) [19] of this species, and could be indicative of an increased establishment of this species in anthropogenic-influenced environments, as occurred for Ae. aegypti [1,20]. In addition, our results support the evidence of an initial domiciliation by Ae. albopictus, defined as the process by which a species occupy niches in the anthropic environment (feeding, resting, and perhaps mating indoors) [21]. Most of the Ae. albopictus adults were collected from key-sites, typically described as highly favorable to Ae. aegypti infestation as shown by the positive and strong correlation between both species [22]. The surveillance and monitoring of such areas are essential to inform vector control strategies [22,23,24]. The fragile infrastructure of key-sites favors Ae. aegypti proliferation [8], but has not yet been linked to Ae. albopictus production. In fact, Manguinhos complex promotes high vector infestation levels through poor sanitation, interrupted water supply and high human population density. In locations where Ae. albopictus was collected (5.1% of sites surveyed), Ae. aegypti was also present in 84.6% (11 of 13) of the locations, showing a clear pattern of co-occurrence, i.e. Ae. aegypti was present in most of the sites where Ae. albopictus was collected. In a previous entomological survey in the same area, both species co-occurred at the transition zone between the forest and the densely populated region [5].
In another entomological survey, low numbers of immature Ae. albopictus were found in Favela do Amorim, one of the 16 slums that composes Manguinhos complex and which also surrounds the forested area in the Fiocruz campus [25]. In our study, the presence of Ae. albopictus adult females and males, together with eggs and larvae, led us to conclude that this species may be establishing itself in the slums of Manguinhos.
The finding of Ae. albopictus inside the households, where febrile cases were reported, clearly indicates that this species has a tendency toward domesticity that may not be as strong as that of Ae. aegypti, but that nevertheless could be of epidemiological importance. Indoor residence of this species highlights the need of maintaining entomological and epidemiological surveillance in vulnerable areas. This is of utmost importance, since we demonstrated the presence of Zika virus in engorged Ae. aegypti mosquitoes in a key-site where Ae. albopictus was found, in the same densely urbanized slum, before the first case of autochthonous Zika virus disease was diagnosed in Rio de Janeiro city [24].
Since both Ae. aegypti and Ae albopictus share the same larval habitats, it has been suggested that their coexistence may be a transient phenomenon, that should be followed by the reduction or displacement of one of the two species through interspecific competition during larval stages [1,26,27] or through asymmetric reproductive interference via interspecific mating. This last circumstance is also known as satyrization, that is a form of reproductive interference where males of Ae. albopictus mate with females of Ae. aegypti resulting in no offspring and permanent sterilization of the cross-mated females [28,29]. However, our monitoring of the current study areas during the last 15 years suggests that the two species may have reached a relative steady state of coexistence in urban areas of Manguinhos, Rio de Janeiro. In addition, we recently showed the lack of major competitive displacement of Brazilian Ae. albopictus males (including Manguinhos strain) to satirize Ae. aegypti females, suggesting that the low satyrization potential of Brazilian Ae. albopictus males may account for the lack of displacements of Ae. aegypti [29]. A previous study showed that this coexistence shows large seasonal fluctuations in both pupal productivity and interspecific competition in the study area, favoring Ae. albopictus over Ae. aegypti. Even though the study shows a clear advantage for Ae. albopictus, seasonal fluctuation of the interspecific competition effects over Ae. aegypti are not sufficient to displace this species in the study area [7]. It has been shown that A. albopictus is superior to Ae. aegypti in resource competition, maintaining greater population growth at higher combined densities [7,30,31], as well as producing greater survivorship during periods of low food availability [32].
In the present paper, we have found a similar pattern for spatial distribution of Ae. albopictus females within the urban area, with mosquitoes collected at almost 400 meters to the nearest vegetation area. Previous studies showed that gravid Ae. albopictus are capable of dispersing at least 800 m in urban areas [33], and that their larvae showed competitive advantages over Ae. aegypti [34,7]. Our results suggest that Ae. albopictus, a competent vector for important arboviruses, including dengue (DENV), chikungunya (CHIKV), Zika (ZIKV) and yellow fever (YFV) [35,36,37], may spread into neglected and densely urbanized areas, if close to vegetated areas. In addition, as this species tends to shelter outside houses [4], are capable of dispersing great distances inside forests near human dwellings and can easily move between sylvatic and urban environments [38], there is an urgent need to establish entomological surveillance protocols targeting this species.
The results obtained in this study show the global importance of maintaining entomological monitoring of Ae. albopictus as a part of surveillance and control programs. This is especially true in Brazil and elsewhere in the Americas where Ae. albopictus might participate in the spillback of arboviruses to enzootic cycles much in the same way as happened to YFV in the last few centuries [1,20,39]. In fact, preliminary evidence shows that ZIKV might be already circulating among neotropical nonhuman primates in Brazil [40]. Besides, this arbovirus has already been detected in wild-caught Ae. albopictus from Bahia, Brazil [41]. Thus, entomological surveillance studies integrate with host-seeking behavior of Ae. albopictus should be investigated inside densely urbanized slums in order to determine whether the presence of Ae. albopictus in slums near vegetated border has an epidemiological importance in the transmission dynamics of these arboviruses.

Conclusions
Densely urbanized slums favor the permanent circulation of mosquitoes, humans and viruses. Continuous longitudinal monitoring is essential in these vulnerable areas in spite of all challenges such as limited access, violence and floodings. Moreover, key-sites, with high human concentration, mobility, and presence of potential Aedes breeding sites, should be included in entomological monitoring. Concomitantly, due to the increasing evidence confirming Ae. albopictus as an efficient viral vector, it would be necessary to extend the entomological monitoring for Ae. albopictus mosquito species. This species has been shown to be a primary vector for arboviruses in different countries and has progressively established in urban areas. However, it has not been a target of surveillance programs in Brazil yet. Finally, this study points out the great importance of integrated studies, since they reinforce the virological, entomological and epidemiological approaches.
Supporting information S1