Life‐history attributes of juvenile Anopheles gambiae s.s. in central Uganda; implications for malaria control interventions

Abstract Malaria is among the leading causes of death in Uganda, and Anopheles gambiae sensu stricto (s.s.) is the predominant vector. Although current vector control interventions have greatly reduced the malaria burden, the disease persists. New interventions are needed in order to eradicate them. Evaluation of new tools will require the availability of well‐characterized test vector populations. Juvenile An. gambiae s.s. from Kibbuye and Kayonjo‐derived populations were characterized under semi‐field and laboratory conditions, given that various vector traits, including abundance and fitness are dependent on development profiles at this life stage. Ten replicates comprising 30 first instar larvae each were profiled for various life‐history attributes (egg hatching, larval development time, larval survivorship, pupal weight and pupation rate). All parameters were similar for the two sites under laboratory conditions. However, the similarities or differences between field and laboratory development were parameter‐specific. Whereas, larval survivorship and pupal weight were similar across seasons and laboratory in colonies from both sites, in the semi‐field settings, pupation rate and larval survivorship differed between seasons in both sites. In addition, the average larval development time during the wet season was longer than that of the laboratory for both sites. Availability of mirror field sites is important for future tool evaluations.


INTRODUCTION
Malaria, a mosquito vector-borne disease, is a leading cause of morbidity and mortality in sub-Saharan Africa. The African region contributes over 94% of global cases, in 2019, Uganda alone contributed about 5% of the global cases (WHO, 2020). More than 95% of the country lies in high malaria transmission areas. Malaria accounts for over 20% of hospital outpatient visits and up to 19% of inpatient admissions (President's Malaria Initiative, 2020), thereby imposing a huge burden on the country's healthcare system. Current malaria control interventions rely on indoor residual spraying of insecticides, usage of insecticide-treated bed nets and drug therapy. However, in spite of the control scale ups made over the past decade, malaria persists. Anopheles gambiae sensu stricto (s.s.) is the principal vector of malaria in Uganda, while Anopheles funestus and Anopheles arabiensis are considered secondary vectors (Presidential Malaria Initiative, 2016).
The progress in the fight against malaria has slowed significantly since 2015, with malaria vector and parasite counter adaptations to mainstream control measures and budget shortfalls being among the challenges contributing to malaria persistence (Guyant et al., 2015). There is, therefore, a need to develop additional interventions to supplement current malaria control efforts if we are to eliminate the disease.
Mosquito population size can vary greatly depending on several larval development and growth factors (Barreaux, 2018). Knowledge of life-history parameters of vector development attributes such as survivorship, size and growth rates among others provides important data for site characterization and modelling to predict the impact of control interventions. Although most malaria control efforts are targeting the adult stage, programmes that target aquatic immature stages are increasingly gaining interest to supplement the core indoor insecticide-based interventions (Derua et al., 2019). This is because the vector capacity, the intensity of transmission and fitness of adult mosquitoes are dependent on juvenile stages. Larviciding is one of the control efforts that will benefit from enhanced understanding of sitespecific juvenile life-history attributes of target vector populations.
For example, survivorship and development time of mosquito stages can help determine when to apply a control intervention in a given site. Malaria elimination has not been achieved despite decades of control efforts, and so innovative approaches that complement current methods are needed.
Anopheles gambiae s.s. juvenile stages development data exists but only from a handful of locations in Africa. In Burkina Faso, there were notable differences in terms of the phenotypic and physiological development of larvae reared in the insectary compared to semi-field conditions (Mouline et al., 2012). In Tanzania, Eliningaya et al. (2005) observed that the development and survival of mosquito larvae were higher in semi-field conditions than in the insectary. However, in the same study, Eliningaya et al. (2005), obtained similar pupation and adult emergence rates in the insectary and semi-field conditions. This was attributed to differences in light intensity and temperatures in the field compared to the insectary.
Mosquito life-history attributes in the East African region are expected to vary given the diversity in the climate, physical features and ecology. The region experiences the largest inter-annual rainfall variations in the world, although a drying trend in March-May rainy season has been observed since the 1980s. The strong, sometimes non-linear altitudinal gradients of temperature and moisture regimes, also contribute to the climate diversity of Eastern Africa (Camberlin, 2018). These variations in climatic and environmental conditions in the region may result in the adaptation of mosquito species, leading to changes in species composition and development traits and subsequent changes in the dynamics of mosquito-borne disease transmission (Afrane et al., 2012).
Life-history attributes of juvenile An. gambiae s.s. in Uganda are poorly described. Most studies have focused on the field ecology and behaviour of adult mosquitoes (Mutebi et al., 2014;White, 2008). To our knowledge, the only available data on life-history attributes of mosquitoes in immature stages comes from a few studies done on Aedes species in Uganda (Lutwama & Mukwaya, 1995;Sempala, 1981). The knowledge of life-history attributes of An. gambiae s.s. in Uganda is, therefore, warranted. In this study, we measured larval developmental time, larval mortality, pupae weight and rate of pupation under laboratory and semi-field conditions in two An. gambiae mosquito populations in Uganda. Well-characterized natural populations could provide ideal test sites for future evaluation of the effectiveness of various vectorbased control measures. Insights into the ecological, environmental and/or biological differences at the aquatic stage in nature could be additionally harnessed for malaria control.

Study area
Mosquitoes were collected from Kibbuye Village in Mukono District (0.2835 N, 32.7633 E) and Kayonjo Village in Kayunga District (0.9860 N, 32.8536 E) in Central Uganda (Figure 1). Both districts experience two rainy seasons and two dry seasons per year. The first rainy season is generally from March to June, followed by a dry season from July to September. The second shorter rainy season runs from October to November and is followed by a dry period from December to February. The districts experience an average annual rainfall of about 1435 mm (Data-Africa, Uganda Mukono/ Kayunga, 2015). Temperatures typically vary from 16.7 C to 27.8 C, although the water temperatures of mosquito breeding habitats could be a degree higher or lower than the environmental temperature. The study sites have rich flora that include forest and swamp vegetation, savannah short grasses and thorny bushes. One village was selected from each district based on several factors, including human settlements, vegetation type, the prevalence of An. gambiae s.s. mosquitoes and malaria endemicity. The two districts experience high malaria incidences (up to 150 confirmed malaria cases per 1000 population/year) and are located in areas of high mosquito densities of An gambiae s.s. (Ministry of Health, 2014). The selected districts have reportedly high levels of insecticide resistance associated with knockdown resistance (kdr) mutation Vgsc-L1014S in An. gambiae s.s. (Lynd et al., 2019). Ethical approval was obtained from Uganda Virus Research Institute Ethics Committee (GC/127/16/11/348) and the Uganda Council for Science and Technology (HS 1328). Informed consent was obtained from the sub-county leaders, village leaders and household owners before gravid females were collected.

Field collections and processing
Laboratory and semi-field-reared samples were used to generate various population-level mosquito development life attributes ( Figure 2).

Indoor collection of adult mosquitoes
Ten houses from each study village were randomly selected for indoor collections of adult mosquitoes in each district. Collections were made at the end of the rainy season in January 2017 ( Figure 1). The collections were used to establish mosquito colonies as a first step for the study (Figure 2). Informed consent to sample from the village was obtained from sub-county leaders, village leaders and household owners, respectively, before embarking on the collection of gravid females. Collections were made in the morning hours from 06:00 to 09:00 AM. Coordinates of the selected houses were recorded using a handheld global positioning device (Garmin GPSMAP ® 64s, Garmin. Olathe, Kansas, USA). Ten gravid females identified using morphological keys (Gillies & Coetzee, 1987) as An. gambiae were collected from each house using a mouth aspirator. Mosquito samples collected from each house were individually gently placed into separate 250 ml paper cups fitted with a net at the top as previously described (Coluzzi & World Health Organization, 1973). The cups were placed in a cage and immediately transported to the insectary at Uganda Virus Research Institute, Entebbe, Uganda for further rearing. The gravid females in the insectary were fed on 10% glucose for 3-4 days to attain full egg development. A forced-egg laying method was used to induce the females to lay eggs (Morgan et al., 2010). Each female that oviposited was killed by freezing at À80 C for 2 min, transferred into a labelled tube containing 80% alcohol and stored at À80 C.

Molecular species identification of mosquitoes
Each field-caught female (F o ) that oviposited first generation (F 1 ) egg batches were subjected to species diagnostic Polymerase Chain Reaction (PCR)-analysis for molecular species identity confirmation as described by Wilkins et al. (2006).

PCR amplification
Two legs, as DNA template from each mosquito, were directly dropped in an aliquot of PCR reaction mix consisting of; 1 U of Taq DNA polymerase (Invitrogen), 0.3 mM MgCl 2 , 0.08 mM dNTPs, 1 μM of each primer (Wilkins et al., 2006) and buffer (Invitrogen, Life Technologies corp. Carlsbad, CA, USA) at 1Â concentration in a distilled water (dH 2 0) topped-up 25 μl reaction volume. Primers (Eurofins genomics) consisted of Universal IMP-UN as forward primer and respective reverse primers; ME-3T for An. merus, QD-3T for An. quadriannulatus, GA-3T for An. gambiae and AR-3T for An. arabiensis for potential An. gambiae candidate sibling species prevalent in the region. The PTC-100 MT thermocycler (MJ Research Inc, Watertown, MA, USA) was used for amplification following the Wilkins et al. (2006) protocol. PCR products were separated by electrophoresis through 1% agarose gels and visualized by ultraviolet illumination after gel staining with ethidium bromide.

Colony establishment
Following molecular species identification of F o adult females, oviposited eggs from confirmed An. gambiae s.s. were used to establish the colonies (Figure 2). The colonies were reared following protocols described by Benedict (2007). Two colony lines (one for each site) were established for Kayonjo and Kibbuye sites. The F 1 colonies were used to rear mosquitoes for life-history investigations. Emerging containing 500 ml of dH 2 o. Larvae were reared at temperature ranges of 24 C-28 C. The larvae were given a daily portion of 10 mg/day of fish food (Tetramin ® Germany, Teyra GmbH Company) as previously described by Bayoh and Lindsay (2004), Kirby andLindsay (2009), Bock et al. (2015). Emerging adult mosquitoes were maintained in 30 cm Â 30 cm Â 30 cm holding cages at 60%-70% relative humidity at 24 C-28 C. The adults were fed on 10% glucose solution from soaked cotton pads. The colonies were maintained up to the 6th generation (F 6 ) before the commencement of the life-history attribute studies. At the F 6 generation, egg production and mosquito survivorship had become similar between subsequent generations by that stage indicating colony stability. The purity of the colony as An. gambiae s.s. was ascertained by molecular identification of a sample of 10 dead males and females from each generation.

Experimental set up in the insectary
A total of 30 first instar larvae of An. gambiae s.s., obtained within 3 h of hatching from the colonies, were used in replicates of 10 for each study site. The number of larvae used was similar to that used by Bayoh and Lindsay (2003). The rearing procedure for colony establishment was as described above. The rearing water was changed every 2 days by transferring the larvae from one dish into another with fresh distilled water. The temperature of water in the rearing dishes was maintained at 26 AE 1 C in a thermostat-controlled room. The relative humidity varied between 64% and 70% as measured by a hygrometer.
The larval dishes were inspected daily every 6 h and dead larvae were removed using a pipette, counted and recorded. The number of individuals transforming into the next larval stage was recorded daily until pupation. Emerging mosquitoes that died were recorded as part of pupal mortality.

Determination of life-history attributes of juveniles
Percentage of eggs hatching: This was determined as the number of eggs that hatched out of the total number of eggs that were placed in a dish for hatching expressed as a percentage. Larval survivorship: This was determined as the number of fourth instar larva that developed into pupae stage out of initial first instar cohort in all replicates (300). Larval development time: This was determined as the average time spent for larvae to develop from the first instar stage to pupae.

Molecular Identification
Laboratory studies The stage duration was determined when 50% of individuals in one stage had transformed into the next immature stage (Bayoh & Lindsay, 2003). Pupal weight was recorded as the average weight of the pupae in each of the 10 replicates. The weight was measured using an electronic balance (model Mettler PE 200) to the nearest 0.01 mg. Pupation rates were recorded as the total number of pupae collected per time (measured in days) taken to develop from the fourth instar larva to the pupation.

Semi-field experiment setup
Semi-field life-history experiments were set up at Kibbuye and Kayonjo study sites during both the dry and wet seasons. Dry season studies ran from July to September 2017. Ten semi-field habitats were made using plastic washbowls (35 cm diameter Â 20 cm deep). In the bowls, 3 kg of soil from the study site was mixed with 3 L of screen filtered pond water to form mud. Pond water used as semi-field habitats was obtained from one large pond in each study site that contained Anopheles mosquitoes. The mud was left to dry to mimic the natural soil lining of habitats. The method used for setting up semi-field habitats is a modification of Gimnig et al. (2002) in that there was no variation in the larval density in the replicates and the larvae were not given food supplements. The larvae fed on algal biomass found in the artificial habitats. The pond water was filtered using screens and topped-up daily to replenish the amount lost due to evaporation. Ten semi-field habitats were set up for each site and left to stand for 6 weeks before the introduction of larvae. Each washbowl was covered with insect netting to prevent other mosquitoes from ovipositing in the artificial habitats. An. gambiae s.s. larva hatched from the established colony in the insectary were transferred to the field site. A total of 30 first instar larvae (within 6 h after hatching) were introduced into the artificial habitats in the field. The field life-history attributes were recorded and captured as earlier described in the laboratory. In addition, the mean physio-chemical conditions of water in semi-field habitats were measured daily for pH and conductivity using a standard por- Comparison of life-history attributes between laboratory and semi-field experiments for Kayonjo dry season

Comparison of life-history attributes between laboratory and semi-field conditions for Kayonjo site wet season
There was a statistically significant difference in some life-history attributes between laboratory and field-reared mosquitoes from 1.5 days. The overall mean survival rate from egg hatching to adult emergence was 68.3% (Table 2).

Field studies during the dry season
The percentage of eggs hatched was 61 AE 2.3% out of the 720 eggs from the insectary parent stock. The mean duration of the specific larval stages varied in the dry season; the least was 1.1 AE 0.1 days in the pupal stage and the highest was 2.3 AE 0.1 days in the fourth instar. The pupal stage lasted for a mean period of 1.2 AE 0.14 days. On average, it took 8.9 AE 0.5 days for An. gambiae s.s. to develop from egg to adult.
Pupation rate and pupae weight was 9.7 AE 1.2 and 3.6 AE 0.8 mg, respectively. The mean survival rate was 84% for the pupae stage and the overall mean survival rate from egg to adult emergence was 61% (Table 2). Semi-field habitat water temperature averaged at about 20.6 C, with the temperatures of habitat at night ranging from 12.5 C to 24.5 C and 15 C to 30.5 C during the day (Table 3).

Comparison of attributes between wet and dry season in Kibbuye site
There was a statistically significant difference in some life-history attributes between field experiments conducted during the wet and dry season at the Kibbuye study site (F (4,11) = 8.509, p = 0.02; Wilk's Λ = 0.244 at 95% confidence interval). The pupation rate was signifi-

Comparison of attributes between laboratory and dry season in Kibbuye site
There was a statistically significant difference in some life-history attributes between laboratory and field-reared mosquitoes from

Comparison of life-history attributes between laboratory and wet season in Kibbuye site
There was a statistically significant difference in some life-history attributes between laboratory and field-reared mosquitoes from Kibbuye in the wet season (F (4,15) = 3.056, p = 0.0001; Wilk's Λ = 0.025 at 95% confidence interval). Larval development time was significantly longer

Water parameters between sites and across seasons
Generally, the water parameters had limited or no effect at all on the life-history attributes between sites and across seasons. Conductivity  (Table 3).

Comparisons between sites
Comparison of growth parameters between sites under laboratory conditions  (Kaufman et al., 2006;Rejmankova et al., 2000). It is possible that some of these factors were sub-optimal during the wet season. For both seasons, more mortality was recorded in second instar larvae hence this is the weakest link in the cycle, and therefore, control interventions need to take this into perspective during design and evaluation.
The pupation rates mirrored trends seen in larval development above in that they were higher during the dry season than in the wet season for the Kayonjo site. This could be partly because, during the dry season, water temperatures tend to rise during certain periods of the day, contributing to faster growth (Kirby & Lindsay, 2009 Water physico-chemical parameters are weakly correlated with the attributes at the time of measurement. However, water chemical parameters change over time and may affect the number of larvae in the habitat. Our study period may not have been long enough to conclusively detect any effects. There is still a need to understand the tolerance and influence of physico-chemical parameters on life-history attributes to come up with more conclusive baseline information for the geographical area. In

CONFLICT OF INTEREST
The author declares that there are no competing interests.

AUTHOR CONTRIBUTIONS
Charles Batume drafted the first draft of the manuscript and con-

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.