Dynamics of multiple insecticide resistance in the malaria vector Anopheles gambiae in a rice growing area in South-Western Burkina Faso

Malaria transmission in tropical Africa is mainly dominated by Anopheles gambiae complex, including An. gambiae s.s., the most anthropophilic vector transmitting malaria in sub-Saharan Africa [1]. Formerly considered as a single species, it began early to accumulate genetic heterogeneity. In the 1980s, cytogenetic studies based on chromosomal inversion arrangements found five incipient chromosomal forms. The suspected role of these inversions was to restrict gene flow among the forms and to provide adaptation to different ecological settings [2, 3]. In Burkina Faso, Mopti and Savanna chromosomal forms dominated An. gambiae population structure [4]. These chromosomal forms appeared more or less genetically isolated in the field, presumably through prezygotic barriers since viable and fertile hybrids have been obtained in the laboratory [5‐7]. However, cytogenetic analysis is not a precise way to evaluate the degree of hybridization between forms because of the presence of cryptic 'heterokaryotypes'. Recent studies based on molecular markers such as X-linked ribosomal DNA suggested the existence of only two entities within An. gambiae, referred to as M and S molecular forms [8]. So far, in Burkina Faso and Mali, in savannah environments of West Africa, Mopti and Savanna chromosomal forms coincide respectively with M and S molecular forms [4]. As a main malaria major vector with high level of polymorphism, An. gambiae has been a subject of many investigations in West Africa, such as bio-ecology and insecticide resistance studies [9‐12].

In Burkina Faso, a study carried out in the mid-1980s in Vallée du Kou [13], showed that the Mopti chromosomal form was dominating. Molecular-based identification of An. gambiae s.s. populations conducted in this area in 1999 and 2000 confirmed that the M molecular form predominated throughout the year [11, 12, 14], although with some temporal variations.

The S molecular form occurred in low frequency until the end of the rainy season (October/November), when it peaked around 30% [12], as found previously by Robert et al [13]. A similar pattern was found in the same environment in Mali [7, 9]. More recently, in 2004 the S form was observed in Vallée du Kou towards the end of the rainy season at a frequency of 50% [15]. This study gave no results regarding the frequencies of the two forms throughout the malaria transmission season.

In West Africa, the main mechanism involved in pyrethroid-resistance in An. gambiae is caused by target site insensitivity through a knockdown resistance (kdr)-like mutation caused by a single point mutation (Leu-Phe) in the para-sodium channel gene [16]. Preliminary surveys done in Vallée du Kou in 1999 indicated that the Leu-Phe kdr mutation has been found almost only in the S form at high allelic frequency (0.95) compared to just 0.006 in the M form [12]. However, the spread of the mutation in the M population seemed an ongoing process in Vallée du Kou as it increased to a frequency of 0.02 in 2000. Nowadays, the population structure of An.gambiae and their pyrethroid resistance status are probably modified with the changing in agricultural practices needing intensive use of insecticides (cotton and vegetable cropping) and also the increasing of man made breeding sites as puddles throughout the village.

Furthermore, it has been recently noted in An. gambiae from the same area the occurrence of a single point mutation (glycine to serine at position 119) in the oxyanion hole of the acetylcholinesterase enzyme [17]. This mutation named ace-1 ^R mutation was associated with insensitivity of An. gambiae to organophosphates and carbamates [18].

In Burkina Faso, current chromosomal and molecular forms of An. gambiae s.s. are tightly correlated [4, 10]. Formerly Robert et al. [13] studying the distribution of An. gambiae s.s. cytotypes in the rice field area of Vallée du Kou in 1984 showed a predominance of the Mopti chromosomal form. With the progress in molecular genetic, this distribution has been updated in 1999 and 2000 by Diabate et al [11, 12], pointing out the predominance of the M molecular form corresponding to the Mopti chromosomal form. However the occurrence of the S molecular form (corresponding to the Savanna chromosomal form) has been observed toward the end of the rainy season. The results obtained in 2005 followed the same pattern of distribution, but the overall proportion of the S form has increased further reaching a maximum of 51% vs. 24% at the same period in 2000 [12]. Taking to account that both studies were performed in the same place (VK7) and during the same period (from July to November), it appears that the relative frequency of the S form has increased. Future studies to confirm this trend are encouraged.

The increase in the S form could probably be due to human activities creating new habitats for the S population in this area, such as house building using bricks of banco and muddles taken out from the soil. Such activities increase the number of temporary rain-filled breeding sites like pits, ponds and puddles that are favourable to S form development throughout the village.

In general, in West Africa, the S form is not well adapted to rice paddies, whilst the M form develops rather well [24, 25]. However, some of the rice paddies used to grow vegetables and some irrigation canals not well managed could constitute during the rainy season, suitable habitats for the S larvae.

The changing in the vector population structure may presumably increase the malaria transmission level. Even though the malaria vector population was dominated by the M form throughout the year, this form had a low sporozoite rate because of the low parity rate observed within its population [14]. During the sympatry period (October), the sporozoite index in the S form was not statistically different to that of the M form, but these are preliminary results that would need to be confirmed on a bigger sample and during different periods of the year.

The main mechanism conferring resistance of An. gambiae to pyrethroids in West Africa, the Leu-Phe kdr mutation, did not vary in the S form (93%) compared to its frequency in 1999 and 2000 [12], as expected.

The spread of the kdr mutation is an ongoing process in the M form as its allelic frequency in 2005 was five fold higher than in 2000. Indeed six years ago, the kdr mutation was found occurring only in the S form and some investigations conducted in VK7 during the rainy season 1999 failed to identify this mutation in the M one [11].

This difference persisted because of the strong reproductive barriers between the two forms relevantly pointed out by some authors [4, 26, 27]. In November 1999, the kdr mutation was identified for the first time in the M form at a very low frequency (0.006) with only one heterozygous RS among 161 specimens tested [12]. The following year, the frequency in the M form increased to 0.02 and all individuals kdr-positive were only heterozygous (RS). Now, the proportion of homozygous resistant in the M form is increasing which enhances the overall frequency of this mutation.

The high frequency of the kdr mutation in the S form is presumably due to the long-term and extensive use of insecticide for cotton crop protection, DDT in the 1960–1970s replaced by pyrethroids in the 1980s [11]. Then, its spreading from the S to the M form through introgression [28] is a recent and ongoing process, limited in savannah environments of West Africa to the place where An. gambiae M/Mopti and S/Savanna forms are found in sympatry at high densities [29].

Cotton crops are located just on the outside of rice fields. Insecticides applied on cotton during the rainy season may drift or be washed to the rice field during heavy rains conferring a selection pressure even for the M form. Moreover, during the last decade some rice producers also started to grow vegetables in the paddies using the same cotton's insecticides.

These agricultural uses of insecticides, mainly pyrethroids, exert the main selective pressure on the mosquito populations because of the quantity applied (e.g. six rounds of treatment each two weeks on cotton crop during the rainy season) and its action on the larval stage on large population of both sex. The domestic use of pyrethroids, through coils and ITNs, is more selective (it acts only on the anthrophilic fraction of biting females) and probably plays a secondary role on resistance selection in this rural area.

The present results showed that the ace-1 ^R mutation is mostly present in the S form and less frequently in the M one. This finding suggests that the ace-1 ^R resistance allele is evolving along the same pathway like the kdr mutation in this area. As for kdr, it occurred probably prior in the S form and may acquire by the M form through introgressive hybridization [30]. The selection of the ace-1 ^R mutation in the S form could be related to the increasing use of organophosphates in cotton treatment in mixture with pyrethroids since the end of the 1990s. This insecticide resistance management (IRM) strategy was implemented at a large scale in West Africa to manage the pyrethroid resistance of the main cotton pest, Helicoverpa armigera [31].

With the exposure of the M populations to insecticide pressure, the ace-1 ^R mutation began to spread within this form. Similarly to the kdr mutation, it will probably increases in frequency within the M populations in the coming years. Additional studies are crucial to determine precisely the origin of this gene among the M form and gene flow pattern between the two forms in natural populations. Alternatively the reproductive fitness associated with this mutation in An. gambiae both S and M population remains to be evaluated [32].

The reported change in malaria vectors population structure is mainly driven by human activities and will call for modified malaria control strategies. The increasing of the S form proportion and the emergence of the ace-1 ^R mutation concomitantly with the Leu-Phe kdr mutation among the same populations of An. gambiae s.s. is an atypical ecological pattern in an irrigated rice growing area. With the expansion of agricultural practices such as vegetable growing, the application of pesticides with different mechanisms of action is rising. This may favour the development of multiple resistance mechanisms in An. gambiae [33, 34], which is a dynamic process that needs to be carefully monitored at the molecular form level and through designed spatial and seasonal surveys.

Further studies are needed to determine: (i) the phenotypic effects, particularly when the two mutations occur concomitantly and taking into account if metabolic-based resistance is present and (ii) the operational impact of both mutations on the efficacy of pyrethroid or organophosphate/carbamate based vector control.

Until recently, several studies conducted in savannah environment of Ivory Coast (West Africa) showed that pyrethroids treated nets still achieved a good control of kdr resistant An. gambiae either in experimental huts [35] or in field trials [36]. Indeed recent paper from N'Guessan et al. in southern Benin [37] indicated that the kdr target insensitivity present at high frequency in M/Forest population of An. gambiae is associated with the decreased efficacy of ITNs and pyrethroid based IRS. Some alternatives to pyrethroids on ITNs are therefore necessary. Preliminary studies using organophosphate and carbamate treated nets in experimental huts have already shown good results in areas of kdr resistance [38, 39].