The dynamics of pyrethroid resistance in Anopheles arabiensis from Zanzibar and an assessment of the underlying genetic basis

Abstract

Background: The emergence of pyrethroid resistance in the malaria vector, Anopheles arabiensis, threatens to undermine the considerable gains made towards eliminating malaria on Zanzibar. Previously, resistance was restricted to the island of Pemba while mosquitoes from Unguja, the larger of the two islands of Zanzibar, were susceptible. Here, we characterised the mechanism(s) responsible for resistance on Zanzibar using a combination of gene expression and target-site mutation assays.

Methods: WHO resistance bioassays were conducted using 1-5d old adult Anopheles gambiae s.l. collected between 2011 and 2013 across the archipelago. Synergist assays with the P450 inhibitor piperonyl-butoxide were performed in 2013. Members of the An. gambiae complex were PCR-identified and screened for target-site mutations (kdr and Ace-1). Gene expression in pyrethroid resistant An. arabiensis from Pemba was analysed using whole-genome microarrays.

Results: Pyrethroid resistance is now present across the entire Zanzibar archipelago. Survival to the pyrethroid lambda-cyhalothrin in bioassays conducted in 2013 was 23.5-54.3% on Unguja and 32.9-81.7% on Pemba. We present evidence that resistance is mediated, in part at least, by elevated P450 monoxygenases. Whole-genome microarray scans showed that the most enriched gene terms in resistant An. arabiensis from Pemba were associated with P450 activity and synergist assays with PBO completely restored susceptibility to pyrethroids in both islands. CYP4G16 was the most consistently over-expressed gene in resistant mosquitoes compared with two susceptible strains from Unguja and Dar es Salaam. Expression of this P450 is enriched in the abdomen and it is thought to play a role in hydrocarbon synthesis. Microarray and qPCR detected several additional genes putatively involved in this pathway enriched in the Pemba pyrethroid resistant population and we hypothesise that resistance may be, in part, related to alterations in the structure of the mosquito cuticle. None of the kdr target-site mutations, associated with pyrethroid/DDT resistance in An. gambiae elsewhere in Africa, were found on the islands.

Conclusion: The consequences of this resistance phenotype are discussed in relation to future vector control strategies on Zanzibar to support the ongoing malaria elimination efforts on the islands.

Figure 1

Anopheles larval collection sites from Zanzibar.

Figure 2

Dose–response curves for An. arabiensis exposed to lambda-cyhalothrin on Zanzibar. An. gambiae were exposed to a range of concentrations of lambda-cyhalothrin in WHO susceptibility tests. The lethal concentration for 50% mortality (LC₅₀) was calculated by fitting a generalised linear model (GLM) using a binomial logit-link function. The LC₅₀ is shown above each plot. For Tumbe, two additional curves are shown specific for An. arabiensis (red dashed) and An. merus (orange dashed).

Figure 3

Susceptibility of An. gambiae to lambda-cyhalothrin in Zanzibar between 2011 and 2013. Mosquitoes were exposed to lambda-cyhalothrin (0.05%) in WHO susceptibility tests between 2011 and 2013. Blue (Unguja) and red (Pemba) percentage mortalities with 95% binomial confidence intervals are given one-hour of exposure to insecticide. Darker bars represent synergist assays using one-hour pre-exposure with piperonyl butoxide (PBO) performed in 2013.

Figure 4

Volcano plots for expression between An. arabiensis from Pemba against samples from Dar es Salaam and Unguja. Expression plots for the 268 candidate probes significantly up-regulated in (i) An. arabiensis from Pemba collected in 2012 against UNG and (ii) DAR. Probes for the three candidate P450s are highlighted on each plot.

Figure 5

Quantitative PCR validation of candidate P450s in An. arabiensis from Zanzibar. A) Quantitative PCR (qPCR) analysis of CYP4G16, CYP6Z2 and CYP6Z3 was performed on An. arabiensis collected from Pemba (Mangwena) and Unguja (Chuini & Mwera) in 2012. The mean ± SEM for three ddCq values relative to Unguja are presented. (B)CYP4G16 expression in An. arabiensis collected in 2013. Four groups of mosquitoes were included in the analysis: (i) Pujini exposed to lambda-cyhalothrin (0.05%) (Pujiniλ) (ii) Pujini unexposed to insecticide (Pujini_untreated) (iii) Mwera (iv) Chuini. The mean ddCq values ± SEM of five biological replicates are presented. NS = non-significant. *p <0.05 one-sided Student’s t-test.

Figure 6

Co-expression of CYP4G16 with genes significantly up-regulated in Pemba associated with fatty-acid metabolism. The expression levels between Pemba and Unguja are compared from those in the microarrays performed on 2012 collected An. arabiensis samples against those from qPCR performed on 2013 samples. For the 2013 qPCR data, the fold-changes are calculated from ddCq(Pemba)/ddCq(Unguja).