Is Outdoor-Resting Behaviour in Malaria Vectors Consistent? Short Report From Northern Ghana.

Background Recent reports of a change in the resting behaviour of malaria vectors, from predominantly indoor resting to outdoor resting following blood feeding, have been attributed to selection pressure from use of vector control tools such as indoor residual spraying and long-lasting insecticide-treated nets. Recent studies have observed vectors resting predominantly outdoors in settings where anti-vector tools are extensively deployed. This present study examined if the outdoor resting behaviour in the vector population, is random or indicative of a consistent preference of one resting site over the other. Methods Mark-release-recapture (MRR) experiments were conducted with outdoor-resting Anopheles gambiae and An. funestus mosquitoes collected from pit shelters, animal houses and granaries in two villages in Northern Ghana. Mosquitoes were marked with uorescent dyes and released indoors. The experiments were controlled with indoor-resting mosquitoes, which were marked and released outdoors. Species of all recaptured mosquitoes were identied and assessed for consistency in their resting behaviour.


Background
Vector control with indoor residual spraying (IRS) and long-lasting insecticidal nets (LLINs) has contributed largely to reduction in malaria incidence and mortality [1]. These tools are both insecticide and indoor-based and target vectors that feed indoors (endophagic) and rest indoors (endophilic) [2][3][4].
However, recent reports are indicating that vector control tools are inducing behavioural changes in mosquitoes such that some vector species have shifted from indoor-resting behaviour to resting outdoors (exophilic). This was as documented from studies in Ethiopia [5], Ghana [6,7], Kenya [8,9)], Libreville [10] and Tanzania [11,12]. Such behavioural diversi cation in vector population can be detrimental to vector control [13,14], which mainly target vectors resting indoors. Therefore, efforts to understand emergence of such behaviour are highly essential for elimination agenda.
Mark-release-recapture (MRR) studies are a standard experimental method extensively employed to investigate mosquito biology, ecology, life history and behaviour, measuring parameters including dispersal and ight distance; survival, population size and density, as well as blood feeding, host seeking and reproductive behaviour [15][16][17][18]. MRR experiments have been used particularly for eld and laboratory studies involving mosquito species such as Aedes [19,20], Culex [21,22] and Anopheles [15,16,18]; to collect data highly relevant in understanding pathogen transmission, gene ow and development and optimization of vector intervention tools [16,18,23].
Outdoor resting behaviour in vector populations may be triggered by heavy use of insecticide-based indoor intervention in settings where control tools are widely implemented [11,24,25]. Vectors in these settings avoid contact with insecticides and rest outdoors where no insecticide is used [11,[24][25][26]. Such behaviour promotes outdoor transmission as vectors can bite unprotected humans outdoors [11,27].
Northern Ghana has documented outdoor resting in An. gambiae sensu lato (s.l.) population [6, 7] but it is not clear whether this behavior only occurs randomly in the vector population. This present study examined if the preference for outdoor resting in the vector population was consistent or not. MRR experiments were conducted using eld-collected mosquitoes of unknown age in two villages in northern Ghana, where coverage of IRS and LLINs are high [7,28,29].

Study sites
The MRR experiments were conducted in two villages in northern Ghana, Kpalsogu (9.33 0 N, 1.02 0 W) and Libga (9.35 0 N, 0.51 0 W) during the rainy season between July-November 2017 and dry season in December 2017-January 2018. The region has a unimodal rainfall pattern with monthly density between 150-250 mm and a forest vegetation zone with mean daily temperature ranging between 25-30°C; and an average relative humidity between 65-75% [30]. . This region has previously documented malaria incidence rate of about 40% in under-ve children (29,31) and LLINs coverage is > 70% in both villages [7,29]. Kpasolgu has been under IRS implementation since 2008 [7]. Extensive farming is practiced in both villages, occurring throughout the year as the villages are close to irrigation dams.

Mosquito collection
Live outdoor-resting adult female Anopheles mosquitoes were collected from pit shelters, granaries and animal resting shelters. The pit shelters were positioned about 5-10m outside six randomly selected compounds in each village. The mosquitoes were carefully collected using mouth and prokopack electrical aspirators so that they remain alive. Collections were carried out during the early mornings (05.00-07.00).
To compare the trend in behaviour of mosquitoes when resting indoors and also as a control for the experiments, indoor-resting mosquitoes were also collected from sleeping rooms in the compounds where the pit shelters were situated, using mouth and prokopack aspirators at the same time the outdoor mosquitoes were collected.
Mosquito processing, marking and subsequent release.
Captured mosquitoes were immediately transported to the insectary for morphological identi cation as female An. gambiae s.l. and An. funestus complex [32]. Following identi cation, indoor and outdoor mosquitoes were carefully kept in separate cages in the insectary, irrespective of their abdominal status. The mosquitoes were fed with 10% glucose until experiment time at dusk of the same day between 18:00-19:00.
The marking of mosquitoes was carried out 3 hours to release time. Batches of 15-20 mosquitoes from the same compounds were kept in the same paper cup. The mosquitoes were lightly dusted with red uorescent dye and left in the cup to rest and feed on 10% glucose until when transported to the study sites for release. The indoor (control) specimens were similarly marked with green dye.
At least 200 mosquitoes were released in ve sleeping rooms (indoors) of the same compounds where the pit traps were situated. This was done to assess if they would remain indoors to rest or return outdoors where they were originally collected from. The control indoor-collected mosquitoes were released outdoors to ascertain if an alternate trend in the resting behaviour would be demonstrated.

Recapture of released mosquitoes
The mosquitoes were left for 48 hours before recapture was initiated. This offered su cient time for the mosquitoes to redistribute in their new environment and enabled us to assess if the mosquitoes opted to return to their initial collection location.
To recapture, mosquitoes were sought from sleeping rooms, animal houses, clay pots, granaries and pit traps deploying light traps, exit traps, pyrethrum spray catches as well as mouth and prokopak aspirations. All rooms and compounds in each of the study sites were sampled to recapture the released mosquitoes. Recapturing was done for 7 consecutive days before the next release event. In all, twelve (12) rounds of MRR experiment were conducted within the study period.
Processing and identi cation of recaptured mosquitoes All recaptured mosquitoes were morphologically identi ed as marked and unmarked. Only the marked Anopheles mosquitoes were selected for data collection and assessed for change in resting behaviour. All unmarked mosquitoes were counted and recorded.
The marked recaptured mosquitoes were processed to discriminate the sibling members of An. gambiae s.l. and An. funestus complex. DNA was extracted from these mosquitoes and molecular species were genotyped using the protocols as previously described [33,34].

Ethical considerations
Ethical approval was obtained from the Ghana Institutional Review Boards (IRB) of Noguchi Memorial Institute for Medical Research (NMIMR). Village heads were also engaged in discussions where concerns on release of mosquitoes into sleeping rooms were clari ed. MRR experiment presents no potential risks as mosquitoes were released in the compounds where they were initially captured and a proportion was to be recaptured.
Following all the 12 MRR experiments conducted, a total of 3, 950 Anopheles mosquitoes were recaptured, of which 3,919 (99.2%) were unmarked and 31 were marked (0.8%). These unmarked mosquitoes were notably not among the mosquitoes released for the MRR experiments. Subsequent analyses of the outcome of the MRR experiments were focused only on the marked mosquitoes that were recaptured.
Recapture of the marked mosquitoes was achieved from nine (9) out of twelve release experiments. Thirty (31) marked test mosquitoes were recaptured from 4,460 released. This indicates a recapture rate of 0.7%. The breakdown of the recaptured test mosquitoes comprised 25 An. gambiae s.l. and 6 An. funestus complex (Figure 1), from which twenty eight (28) were resting outdoors, and three (3) indoors. The majority of the mosquitoes recaptured outdoors were An. arabiensis (11, 39%), followed by An. coluzzii (7, 25%) and An. funestus s.s. (6, 21%). An. gambiae s.s. constituted the least outdoor-recaptured species (4, 15%) ( Table 1). The three mosquitoes recaptured indoors were mainly An. coluzzii. None of the control samples that were released outdoors was recaptured from any location, indoor or outdoor. The aim of this study was to examine if outdoor resting behaviour of the malaria vector populations described in the study settings [6, 7] was consistent or it was just a random occurrence. The study observed that the majority of mosquitoes returned outdoors to rest after being released indoors. This pattern of preference for outdoor resting in mosquitoes was consistently demonstrated from nine out of the twelve experiments, despite a low recapture rate. As outdoor resting behaviour in vector population could counteract control efforts [27,37,38], research assessing the extent of this behaviour is well-timed and necessary to guide decision making in malaria control programmes.
Switch to outdoor-resting behaviour in the vectors may be as a result of avoidance behaviour from indoor interventions [38,39], which were heavily used in the study sites. Most sleeping rooms had LLINs and they were also sprayed with pirimiphos methyl, the insecticide that was being used for IRS at the time of the experiments [7]. These mosquitoes might be avoiding contact with the insecticide indoors and chose to rest outdoors, where they may be safe, as there was no outdoor intervention in the villages. This could have also accounted for the predominant outdoor resting preference documented in the vectors in a previous study from this setting [6] and others all over Africa [8,12,36,37], which prompted this investigation.
The highly endophilic vectors: An. gambiae s.s., An. coluzzii and An. funestus s.s. were recaptured outdoors after being released indoors. This demonstrates a likely switch to outdoor resting behavior because these vector species may be avoiding contact with insecticide indoors and exiting to rest outdoors [40][41][42]. Change from endophily to exophily in response to control intervention is increasingly being documented in these highly endophilic vectors [8,12,25,35,44,45]. This could have negative impact on malaria control as it can promote outdoor and residual transmission in these settings [11,38,45].
Anopheles arabiensis was found to predominate the mosquito species found resting outdoors from the recaptured test mosquitoes. This may be because An. arabiensis is known to be highly exophilic and zoophilic (preference for animal blood meal) [46][47][48], where it tends to stays outdoors to rest and feed on animals when there is intervention indoors [4,49]. This vector species has also been observed to display insecticide avoidance and early-exiting behaviour [12,27,42], which make them di cult to control [11].
In this study, the recapture rate was much lower than most MRR studies reported. Several reasons could have accounted for this. One of such reasons is age of mosquitoes [16,18,50], which was unknown in this study. Plausibly, if the mosquitoes were old, they may have died within the period of recapture.
Indeed, senescence is a factor associated with reduction of mosquito daily survival in the wild [51,52]. Age was also previously suggested to be responsible for the low recapture rate in a similar experiment [16]. Moreover, predators such as spiders were particularly common in the study areas, and were found inside animal houses that provide favourable environment for the mosquitoes to rest. Other factors including emigration from the study area, climate condition, stress and negative effect of the experimental procedures could have also contributed to this low recapture success as previously documented in other MRR experiments [16,18,23].

Conclusions
This study showed that preference for outdoor resting in malaria mosquitoes may be an emerging consistent behaviour. There is a need for further studies to establish this observation in settings where interventions are extensively deployed. Furthermore, a probe into the genetic basis underlying this behavioural change will also be highly essential. This is important as malaria control moves to the elimination phase in sub-Saharan African countries. Composition of released and recaptured marked test and control mosquitoes

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.