Insecticide resistance and role in malaria transmission of Anopheles funestus populations from Zambia and Zimbabwe

In the recent past, malaria vector control has been primarily pyrethroid-based through indoor residual house spraying [IRS] or the use of long-lasting insecticide-treated bed nets (LLINs) or both. This intensive use of insecticides, together with pesticide usage in agriculture, has led to a dramatic increase in resistance in the mosquito vectors across the whole African continent [http://www.irmapper.com]. Anopheles funestus, one of the four major vectors of Plasmodium falciparum malaria in Africa [1],[2], is widely distributed throughout much of the African tropics and subtropics [3]. The earliest records of insecticide resistance in An. funestus are from the 1980s [4] and currently resistance in this species is known from Mali, Guinea, Ghana, Benin, Uganda, Kenya, Malawi, Zambia, Mozambique and South Africa [1]. A well-documented impact of insecticide resistance on vector control programme failure is the major malaria epidemic that occurred in South Africa in 1999/2000 when pyrethroid resistant An. funestus returned to South Africa after DDT was replaced with pyrethroids for indoor house spraying [5]. The situation returned to pre-failure levels after DDT was reintroduced and used in a mosaic system together with pyrethroids. A number of insecticide resistance studies on An. funestus from southern Africa have been documented, but studies from other regions are limited [1].

Anopheles funestus belongs to a group of at least 11 species, all of which are morphologically similar at the adult stage [6],[7]. Some members of the group can be distinguished on egg and larval characteristics (An. confusus, An. rivulorum and An. leesoni) but this is not always the case and those species belonging to the An. funestus sub-group (An. funestus, An. funestus- like, An. parensis, An. aruni and An. vaneedeni) are virtually identical [1],[6],[7]. The importance of species identification lies in the fact that of the 11 species, only An. funestus plays a major role in malaria transmission, this species being highly adapted to humans and human habitation with parasite infection rates sometimes as high as 22% [6]. The only other member of the group that has been implicated in low-level, localised transmission is An. rivulorum in Tanzania [8] and western Kenya [9] but in each case the parasite infection rates have been below 1%.

Malaria vector control programmes need to know that their interventions are targeting the major vectors so that scarce resources are not wasted on non-vector mosquitoes. As a result, various genetical techniques have been applied to members of the An. funestus group with the aim of providing accurate species identification. Initially, the banding patterns of the giant polytene chromosomes were used [10],[11] but these have limited practical value because An. funestus and An. vaneedeni share the same inversion arrangements resulting in the same banding patterns [10], as do An. funestus and An. funestus-like [12]. In addition, the technique has the disadvantage of being applicable only to half-gravid female mosquitoes. However, studies of the chromosomal inversion polymorphisms in Burkina Faso revealed the existence of two distinct forms that apparently do not mate in sympatry [13] and that are also significantly differentiated at the molecular level [14], supporting the hypothesis of two distinct species being present. Twenty years after the work of Green and Hunt [10],[11], a molecular assay was developed that allowed for the rapid and accurate identification of the five most common members of the An. funestus group [15] and this is the standard protocol used in many laboratories today.

While An. funestus has abundant levels of molecular and chromosomal polymorphism across its range, there are only a few population genetic studies of the species and these are limited in geographic scope compared with similar work on the Anopheles gambiae complex [1]. Phylogenetic studies using mitochondrial DNA (mtDNA) have revealed that there are two clades within An. funestus. Clade I is widespread throughout the continent and clade II is known only from southern Mozambique and Madagascar [16],[17]. It is estimated that these two clades evolved independently about 1 million years ago based on fixed differences and divergence in mtDNA [16]. However, the deep mtDNA divergence was not accompanied by corresponding nuclear divergence as measured by 10 microsatellite loci in the same study [16] so it is not clear what this genetic variation means in terms of species differentiation.

In the present study, we investigated the biological attributes of insecticide resistance and parasite infection rates that both impact on malaria vector control activities, in relation to the mtDNA clades in Zambia and Zimbabwe.

Our data show that An. funestus clade II is not confined to southern Mozambique and Madagascar [16],[17] and can be found as far north as Nchelenge District in Zambia, being 2,000 km further north than the current distribution of Chibuto in Mozambique [17]. We do not have data on seasonal abundance of the two clades, so it is possible that clade II also occurs in the Honde Valley, Zimbabwe, but was not detected due to its very low frequency at both times of collection. The frequency of clades I and II in Nchelenge District remained stable over the two years of sampling. Neither the insecticide resistance data nor the sporozoite infection rates provided clear evidence that the clades are anything other than a polymorphism in a single panmictic population.

The Plasmodium falciparum infection rates for the 2013 Zimbabwe (2.3%) and Zambia (2.1%) populations were similar overall, with no significant difference between clade I (8/346 - 2.3%) and clade II (1/76 - 1.3%) (Pearson’s chi-square = 0.34, P = 0.56). However, there was a 4-fold increase in sporozoite rates in Nchelenge District in 2014, corresponding with an increase in reported malaria cases (ICEMR data, personal communication). A slight increase was also seen in the 2014 Zimbabwe sample. Of the eight households sampled almost every single inhabitant had had at least one bout of malaria during the 2013–2014 summer. This recent increasing trend of infective (and resistant) An. funestus being associated with increased number of malaria cases has been reported in other countries, such as Tanzania [24], Kenya [25] and Senegal [26]. These studies underline the importance of An. funestus as a major malaria vector that is sometimes underestimated by researchers and control programmes in Africa.

At present, malaria vector control programmes in Africa rely heavily on chemical methods, with long-lasting insecticide treated bed nets (LLINs) and indoor house spraying with residual insecticides (IRS) being the most widely implemented. However, the alarming increase in insecticide resistance in the main malaria vectors to many of the insecticides used in control, such as pyrethroids, carbamates and DDT, poses a very serious concern [27],[28]. The susceptibility tests clearly showed that both these An. funestus populations are highly resistant to pyrethroids and carbamates while remaining susceptible to DDT and the organophosphates. The malaria vector control interventions at both localities include distribution of LLINs and in the Honde Valley, the houses had been sprayed with lambda-cyhalothrin in December 2013. Neither intervention had any apparent impact on malaria transmission.

The so-called resistance intensity test carried out here was not a standard method of measuring resistance. While the CDC bottle bioassay method [29] recommends that exposures be continued beyond the diagnostic time in order to assess resistance intensity, they recommend that experiments be terminated at 2 hours and the manual [29] does not give criteria for evaluating resistance intensity. At the 2-hr interval, both An. funestus populations showed mortality of less than 20% on the WHO diagnostic dose of 0.05% deltamethrin indicating an extremely high level of resistance. What this means in the field, however, is difficult to judge since these tests do not take into account repellent effects or whether an individual mosquito will indeed rest continuously for 8 hours, or even 1 hour on a treated surface. What is clear however, is that pyrethroid-resistant An. funestus is quite capable of withstanding prolonged exposure to the pyrethroids used for treating bed nets and for house spraying.

The problem of insecticide resistance in An. funestus is no longer a localized one. The resistant population first detected in South Africa [30] and southern Mozambique [20] has now spread to northern Malawi [31] and is affecting the region as a whole, including the localities sampled in this study. The evidence that we have indicates that the resistance mechanisms are mediated by metabolic monooxygenase enzymes, well documented in other An. funestus populations [1]. This metabolic system does not incur a fitness cost [32] and even in the absence of insecticide pressure, the resistance is not lost, facilitating the geographical spread of resistance alleles. However, the resistance in An. funestus populations is by no means uniform across southern Africa with recent studies in south-eastern Zambia showing resistance to DDT as well as pyrethroids and carbamates [33].

This places yet more constraints on the vector control programmes operating in the region with only the organophosphates having a generalised killing effect on the An. funestus populations. While both Zambia and Zimbabwe are changing their IRS policies and will be using an organophosphate for IRS in the coming transmission season, the need for additional, affordable, new tools for vector control [28] has acquired an urgency as never before.