Background
Bubaque is the most populated island and the main commercial centre of the Bijagos archipelago, a group of 88 remote islands and islets situated off the coast of Guinea-Bissau, West Africa. A majority of the approximately 30,000 Bijagos islanders maintain a lifestyle supported by subsistence agriculture, with the port on Bubaque acting as the major hub for the import and export of goods, and the movement of people between the islands and the mainland. Bubaque consists primarily of dense forest and areas of cultivated agricultural land surrounded by stretches of beach and mangrove swamp, although the majority of the island’s approximately 11,000 inhabitants are concentrated in a semi-urbanized area built around the port. The climate is tropical, with a heavy wet season between June and November, and a dry season from December to May.
Malaria is a leading public health concern in Guinea-Bissau, with an estimated parasitological prevalence ranging from 3 to 30% [
1,
2], and a peak in transmission towards the end of the wet season (October–November) [
3]. Although several entomological surveys have been completed in and around the capital city of Bissau, there is limited knowledge of vector species composition, sporozoite rate, and insecticide resistance on the Bijagos islands. To our knowledge there has been only one previous mosquito survey at a single site on Bubaque that reported a mixture of primarily
Anopheles gambiae sensu stricto (s.s.) (hereafter
Anopheles gambiae)
, Anopheles coluzzii,
An. gambiae/
An. coluzzii hybrids, and a low number of the salt-tolerant species
Anopheles melas [
4]. In mainland Guinea-Bissau,
An. gambiae,
An. coluzzii, and their hybrids are the dominant malaria vectors [
5], with
An. gambiae having the highest sporozoite rate [
5].
Anopheles melas often breeds at high densities in coastal regions of West Africa [
6,
7], and sporozoite-positive females have been reported in coastal areas of Guinea-Bissau [
5].
Malaria vector control on Bubaque currently relies on the use of long-lasting insecticidal nets (LLINs). Government-led LLIN distributions have been implemented on a 3-year cycle since 2011, and household surveys suggest high uptake (> 90%, Anna Last, unpublished data). Although LLIN-based vector control has contributed to impressive reductions in malaria prevalence across Africa, selection for pyrethroid resistance has the potential to slow and perhaps even reverse this progress [
8]. As a result, the WHO recommends regular insecticide resistance monitoring in areas with widespread LLIN use.
The island ecology of the Bijagos archipelago creates a potential for differences in malaria vector composition and seasonality compared to mainland Guinea-Bissau. To design/maintain an appropriate malaria control programme for the islands, a comprehensive understanding of the malaria vectors and their dynamics is needed. An entomological investigation on the Bijagos archipelago is presented here, describing the major malaria vectors on the island of Bubaque, their seasonal variation and sporozoite rates, and a characterisation of resistance to pyrethroid insecticides.
Discussion
The island of Bubaque displays strong seasonal variation in malaria vector species composition and abundance. During June/July, the onset of the rainy season,
An. gambiae was the dominant endophilic species, although significant proportions of
An. coluzzii and hybrid forms were also found.
Anopheles melas adults constituted the smallest proportion of overall catches. The relative dominance of
An. gambiae over
An. coluzzii in this period indicates a reliance of the major portion of the wet season vector population on the availability of temporary, rain-dependent larval breeding sites (
An. coluzzii tends to be associated with more permanent water bodies [
18,
19]). In contrast, during November/December,
An. melas was the dominant species (comprising > 85% of catches). This is likely a result of its capacity to maintain year-round populations by breeding in the island’s abundant littoral habitats and mangrove swamps (Bubaque is reasonably small, comprising a high ratio of coastal biome to inner-island area). Breeding of
An. gambiae and
An. coluzzii is limited to sites of low salinity, which are generally rare in the dry season. Catch rates per household suggest a greater absolute abundance of endophagic
An. gambiae s.l. mosquitoes during the wet season (~ 8-fold higher per household catch rates compared with November/December), consistent with the increasing availability of temporary rain-dependent larval habitats supporting expanding mosquito populations.
A previous small-scale wet-season sample collected on Bubaque, reported a high proportion of
An. gambiae mosquitoes (89.5%) (molecular forms not specified), with a lower proportion of
An. melas (8.9%) [
4,
20]. This is consistent with wet season surveys on mainland Guinea-Bissau, where
An. gambiae and
An. coluzzii were found to dominate [
5,
21,
22]. Interestingly, Marsden and colleagues also reported catching a single
Anopheles arabiensis adult on Bubaque (1 out of 67
Anopheles adults caught) [
4,
20].
Anopheles arabiensis has also been found in the city of Bissau [
23], and in the northern inland regions of Guinea-Bissau, where there is an abundance of its preferred dry shrubland and savannah habitats [
24]. The lack of
An. arabiensis in the present survey in either adult or larval collections likely reflects very low densities of this species on Bubaque, but may also result from a bias introduced by the indoor trapping method used in the large majority of mosquito collections;
An. arabiensis tends to be less endophagic and endophilic than
An. gambiae and
An. coluzzii [
25]. Larval collections closely reflected
Anopheles species composition inferred from adult trapping, with a majority of
An. gambiae and a smaller proportion of
An. gambiae/
An. coluzzii hybrids and
An. melas. Several culicine species were also identified including the arbovirus vector
Aedes aegypti.
Strong seasonal variation in abundance of
An. gambiae s.l. species has been reported elsewhere in coastal regions of West Africa. A survey in The Gambia found reductions in overall
An. gambiae s.l. numbers during the dry season, with catches dominated by
An. melas [
7]. Similarly, in a coastal region of southwest Nigeria the population of
An. gambiae was highly seasonal and dependent on rainfall, while
An. melas was able to maintain a relatively constant dry season population [
6].
Anopheles gambiae and
An. coluzzii exhibit strong reproductive isolation throughout the majority of their sympatric range, with rates of hybridization typically < 1% [
26,
27]. However, higher rates of hybridization have been found in West Africa, particularly along the coastline of The Gambia [
28] and Guinea-Bissau (where hybridization rates of up to 25% have been reported) [
28‐
30]. A large proportion of hybrids were found on Bubaque, (hybrids accounting for > 27% of total catches, although 30.6% of
An. gambiae, An. coluzzii and hybrid catches during June-July), with rates exceeding those commonly found on mainland Guinea-Bissau [
5] and in The Gambia [
28].
Anopheles gambiae had the highest
P. falciparum sporozoite rate among all
An. gambiae s.l. mosquitoes sampled, with 13.9% and 20% positivity in the dry and wet season, respectively, although the dry season sporozoite rates should be viewed with caution due to the very small sample size. The greater population density and sporozoite positivity of
An. gambiae in the wet season suggests it is the primary wet season vector on Bubaque. Although
An. gambiae also displayed the highest sporozoite rate of the
An. gambiae s.l. species in the dry season, a far greater density of
An. melas (with a 20-fold higher median per trap catch rate in this period) combined with a positive sporozoite rate of 6.1%, suggests that this species may be responsible for the majority of dry season transmission. The sporozoite rates found in the present survey are similar to those observed in a previous wet season survey on mainland Guinea-Bissau that included sampling on several of the islands of the Bijagos, where average sporozoite rates of 12.6% in
An. gambiae and 11.1% in
An. melas were recorded [
5]. The higher sporozoite rates in
An. gambiae,
An. gambiae/
An. coluzzii hybrids, and
An. melas, at the onset of the dry season correlates with months of peak malaria prevalence. However, there was a lower average dry season sporozoite rate across the wider
An. gambiae s.l., complex (9.7% in the wet season compared to 6.8% in the dry season) combined with ~ 8-fold lower indoor dry season catch rates. A tentative calculation of sporozoite-positive
An. gambiae s.l. during both seasons, gives a median of 0.78 positive females per trap per night in the wet season, compared with 0.068 in the dry season.
Analysis of the wet-season samples from Bubaque revealed
kdr-w allele frequencies of 36% in
An. gambiae, 35% in
An. coluzzii and 42% in
An. gambiae/
An. coluzzii hybrids. This contrasts with a survey in Bissau performed over a decade ago, where
kdr-w allele frequencies of 7% and 0.8% were recorded in
An. gambiae and
An. coluzzii, respectively [
31]. The sharp increase in
kdr-w frequencies is consistent with a response to strong selection associated with the wide-spread distribution of LLINs in Guinea-Bissau since 2011.
Parallel comparisons between distributions of the
kdr allele and pyrethroid resistance detected by in vivo assays are important for providing functional data on resistance, and may uncover additional mechanisms. Resistance can arise through a variety of processes, including metabolic resistance resulting from the amplification or upregulation of genes encoding detoxification enzymes. CDC-bottle bioassays were used in the present study to measure the resistance of field-caught
An. gambiae s.l. females to discriminating doses of permethrin and α-cypermethrin. Although this represents a limited subset of insecticides used in LLINs, it represents two of the most commonly used pyrethroids in LLINs on Bubaque (Anna Last, unpublished data). Full susceptibility to permethrin and partial resistance to α-cypermethrin was found. This permethrin susceptibility contrasts with studies in the neighbouring Gambia, where marked resistance has previously been described, particularly in coastal areas [
32]. FSusceptibility to α-cypermethrin was restored upon pre-exposure of mosquitoes to the insecticide synergist piperonyl butoxide (PBO), implicating mixed-function oxidases as an additional resistance mechanism. Mixed-function oxidase-based insecticide resistance is not uncommon among
An. gambiae populations in West Africa, with biochemical assays and microarray studies implicating metabolic resistance in Benin [
33], Nigeria [
33], and Ghana [
34]. Next generation LLINs (incorporating PBO) are currently being trialled in Africa in areas of high metabolic resistance, and have shown field efficacy [
35], although, beneficial effects of PBO-nets may be restricted to areas with high levels of pyrethroid resistance [
36].
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.