Introduction

The scale-up of key diagnostic, treatment and vector control interventions, particularly long-lasting insecticidal nets (LLINs) and indoor residual spraying (IRS)1, has led to a global decline in malaria incidence. Approximately, 1.7 billion malaria cases and 10.6 million malaria deaths were averted between 2000 and 20201,2, with LLINs and IRS accounting for 68% and 10% of these achievements, respectively2. Insecticidal products currently used in malaria vector control predominantly target indoor host-seeking or resting Anopheles vectors, inducing lethality following exposure. However, widespread deployment of these interventions has placed high levels of selection pressure on mosquito populations, leading to the evolution and spread of insecticide resistance and coincident declines in malaria control1,3. Insecticide resistance has become established in malaria vector populations to four classes of insecticides that have been historically used for public health use: pyrethroids, carbamates, organophosphates, and organochlorines4,5. Pyrethroid resistance is of particular concern, because until recently, it was the only class of insecticide utilised in LLINs6.

In response to the threat of insecticide resistance, substantial investments have been made in the development and repurposing of new insecticides and chemical classes, with novel modes of action, to improve malaria vector control and potentially mitigate further resistance selection. IRS operational strategies in sub-Saharan Africa currently use organophosphate (Actellic 300CS; containing pirimiphos-methyl, an acetylcholinesterase inhibitor) or neonicotinoid products (SumiShield 50WG or Fludora Fusion; containing clothianidin, a nicotinic acetylcholine receptor agonist)7,8. Two generations of LLINs containing a pyrethroid insecticide and a second partner chemical, piperonyl butoxide (PBO; an insecticide synergist)9, a pyrrole (chlorfenapyr; an oxidative phosphorylation uncoupler)10 or an insect growth regulator (pyriproxyfen; a juvenile hormone analogue)11, have been evaluated, with strong epidemiological evidence to support the use of PBO-LLINs (Olyset Plus)12 and chlorfenapyr-LLINs (Interceptor G2)13 to control malaria transmitted by pyrethroid-resistant vector populations. However, reports of incipient resistance to these new insecticides are beginning to emerge14, soon after they have been deployed, highlighting the urgent need for additional chemicals, with distinct target sites, to incorporate into effective resistance management strategies15.

Broflanilide (tradename TENEBENAL) is a novel insecticide discovered by Mitsui Chemicals Agro, Inc16, which has been formulated as a wettable powder for IRS (VECTRON T500). It has a unique chemical structure characterized as a meta-diamide, that acts as a non-competitive antagonist (NCA) of the gamma-aminobutyric acid (GABA) receptor of chloride channels in the insect inhibitory nervous system, causing mosquito mortality by hyperexcitation and convulsion17. Broflanilide has been classified by the Insecticide Resistance Action Committee (IRAC)17 as a GABA-gated chloride channel allosteric modulator (IRAC Group 30). Broflanilide has low mammalian toxicity and a good safety profile; antagonist activities of meta-diamides are considerably lower in human GABAARα1β2γ2S, mammalian GABAARα1β3γ2S and human GlyR α1β receptors than in insect RDL GABA receptors, due to the presence of A288G in human GlyR α1β18. Furthermore, physiochemical data indicate that broflanilide is stable to hydrolysis and soil photolysis, giving it the potential for long-lasting application in IRS but low environmental persistence; it has low solubility in water (0.71 mg/L at 20 °C), its vapor pressure of 6.6 × 10–11 torr and Henry’s law Constant of 3.0 × 10–14 atm-m3/mol suggest that volatilization is not a major dissipation pathway and finally soil adsorption coefficient (KF) values of 113 to 248 mL/g indicate low mobility in soil. Evaluation of the residual efficacy of VECTRON T500 has been coordinated between the London School of Hygiene and Tropical Medicine (LSHTM), the Kilimanjaro Christian Medical University (KCMUCo), the National Institute for Medical Research (NIMR) in Tanzania, Mitsui Chemical Agro Inc (MCAG) and the Innovative Vector Control Consortium (IVCC). To date, broflanilide has demonstrated effectiveness against pyrethroid-susceptible and -resistant vector populations19,20. In the IRS product, VECTRON T500, broflanilide has shown prolonged effectiveness in experimental hut trials in Tanzania and Benin in comparison with World Health Organization (WHO) prequalified products19,20. VECTRON T500 is now under evaluation in non-inferiority community trials in both countries to provide data in support of its evaluation by the WHO Prequalification Unit, Vector Control Product Assessment Team (PQT/VCP) as a new IRS product for insecticide resistance management21.

Insecticides targeting the GABA-gated chloride receptors have been highly effective, being used extensively across Africa in the 1960s–1970s4, for agriculture and public health, including cyclodienes (dieldrin), phenyl pyrazoles (fipronil) and isoxazolines (fluralaner)22. However, the 2001 Stockholm Convention on Persistent Organic Pollutants prohibited the use of cyclodienes due to their slow degradation and environmental persistence23. Despite this ban, there is evidence indicating that the mutations conferring resistance to dieldrin (rdl) have persisted decades later in malaria vector populations24,25. Whilst some in silico studies have demonstrated that the mode of action of broflanilide is distinct from other NCAs targeting the GABA-gated chlorine channel, including dieldrin, fipronil, lindane and α-endosulfan26,27, little biological evidence has been generated to date with mosquito strains, to establish whether there is any cross-resistance to broflanilide via the mutation in the GABA-gated chloride receptor leading to dieldrin resistance (rdl)26,27,28.

The aim of this multi-centre study was to expand upon initial discriminating concentration (DC) testing of broflanilide20,28, defined as the concentration of insecticide that in a standard period of exposure, is used to discriminate the proportions of susceptible and resistant phenotypes in a sample of a mosquito population29, using additional Anopheles vector species, to facilitate prospective susceptibility monitoring of this new insecticide; and to investigate the potential of the A296S rdl resistance mutation in the GABA receptor gene to offer cross-resistance between broflanilide and dieldrin.

Results

Broflanilide discriminating concentration determination

This multi-centre study tested a total of 7370, 2–5 day old, unfed female Anopheles mosquitoes from six colony strains (An. gambiae Kisumu KCMUCo, n = 1812; An. gambiae Kisumu NIMR, n = 860; An. coluzzii N’Gousso, n = 1640; An. arabiensis KGB, n = 873; An. arabiensis SENN, n = 476; and An. stephensi SK, n = 1709), across different concentrations of broflanilide (Fig. 1).

Figure 1
figure 1

Multi-centre broflanilide discriminating concentration testing study: experimental design. Figure created using BioRender.com.

Full size image

A clear mortality-dose response following broflanilide exposure was evident with all insectary strains tested (Fig. 2). Table 1 details lethal doses (%) of broflanilide required for mortality of all six Anopheles insectary strains, with corresponding DCs presented in Table 2, calculated according to two methodologies30,31. Of the five insecticide-susceptible mosquito strains tested (An. gambiae Kisumu KCMUCo, An. gambiae Kisumu NIMR, An. coluzzii N’Gousso, An. arabiensis KGB and An. stephensi SK), the lowest DC was observed for An. gambiae Kisumu NIMR (LC99 × 2 = 1.126 μg/ml [95% CI 0.197–2.78 μg/ml]; LC95 × 3 = 0.7437 μg/ml [95% CI 0.0882–2.2338 μg/ml]), followed by An. stephensi SK (LC99 × 2 = 4.72 μg/ml [95% CI 2.08–8.04 μg/ml]; LC95 × 3 = 1.95 μg/ml [95% CI 0.5652–4.17 μg/ml]) (Table 2). By comparison, the highest DC was recorded for An. gambiae Kisumu KCMUCo (LC99 × 2 = 54.00 μg/ml [95% CI 28.00–134.00 μg/ml]; LC95 × 3 = 17.82 μg/ml [95% CI 10.41–33.00 μg/ml]) (Table 2), indicating substantial variation in mortality-dose response between the two insectary colonies derived from the same original stock (An. gambiae Kisumu). The insecticide-resistant mosquito strain (dieldrin-resistant: An. arabiensis SENN) presented an intermediate DC as determined by the method of Lees et al.31 and as determined following the WHO approach32 (LC99 × 2 = 3.76 μg/ml [95% CI 0.92–7.96 μg/ml]; LC95 × 3 = 1.33 μg/ml [95% CI 0.1506–4.11 μg/ml]), when compared to the other insectary colonies (Table 2).

Figure 2
figure 2

Linear relationships between probit-transformed mortality rates and log-dose of broflanilide for different Anopheles insectary strains (left), with 95% confidence intervals (right).

Full size image
Table 1 Lethal concentrations (%) of broflanilide (μg/ml) required for mortality of six Anopheles insectary strains.
Full size table
Table 2 Estimated broflanilide discriminating concentrations (μg/ml).
Full size table

Pairwise Bonferroni comparisons indicated significant differences in dose-mortality responses between An. arabiensis KGB and An. gambiae Kisumu KCMUCo (p = 0.00076), An. arabiensis SENN and An. gambiae Kisumu KCMUCo (p = 0.00121), An. arabiensis SENN and An. gambiae Kisumu NIMR (p = 0.00875), An. stephensi SK and An. gambiae Kisumu KCMUCo (p < 0.0001) and An. gambiae Kisumu KCMUCo and An. gambiae Kisumu NIMR (p < 0.0001).

Dieldrin cross-resistance testing

To confirm the resistance profiles of both An. arabiensis colonies, initial WHO susceptibility tests were performed on An. arabiensis SENN (dieldrin-resistant) and An. arabiensis KGB (dieldrin-susceptible). A total of 115 KGB individuals were exposed to the discriminating concentration of dieldrin (0.4%) with 100% mortality observed after 60 min (Fig. 3A); demonstrating that this strain was susceptible to dieldrin.

Figure 3
figure 3

An. arabiensis KGB (A) and An. arabiensis SENN (B) mortality after exposure to dieldrin in WHO susceptibility tests. Error bars represent 95% confidence intervals (CIs).

Full size image

Average 24-h mortality of 112 An. arabiensis SENN tested using 0.4% dieldrin impregnated filter papers was 12.1% (95% CI 1.17%-22.9%) (Fig. 3B). For the 107 An. arabiensis SENN mosquitoes tested with 4% dieldrin impregnated filter papers (10X discriminating concentration), average 24-h mortality was 16.5% (95% CI 1.33–31.73) (Fig. 3B); confirming that this strain was highly resistant to dieldrin. All An. arabiensis SENN which survived 4% dieldrin exposure possessed the A296S mutation. All An. arabiensis SENN tested in dieldrin bioassays were confirmed as An. arabiensis by species-specific PCR.

Resistance ratios between An. arabiensis SENN (dieldrin-resistant) and An. arabiensis KGB (dieldrin-susceptible) were 1.13 [95% CI 0.7879–1.63] at the LC50, indicating an absence of cross-resistance between broflanilide and dieldrin. A subset of An. arabiensis SENN (n = 290) tested against different broflanilide concentrations were screened for the presence of rdl A296S to investigate the potential for this resistance mechanism to mediate cross-resistance against broflanilide. There was no association between rdl A296S genotype and survival or death following exposure to any broflanilide concentration (Fisher’s exact test = 0.6019). All An. arabiensis SENN screened for rdl A296S were confirmed as being An. arabiensis by species-specific PCR.

Discussion

The development of novel insecticide formulations for IRS whose efficacies are not compromised by pre-existing cross-resistance in vector populations, is crucial to sustain current gains in malaria vector control33. This multi-centre study builds upon initial broflanilide DC testing performed with single insecticide-susceptible insectary colonies, using six mosquito strains, representing major Anopheles species; An. gambiae and An. arabiensis are sympatric malaria vectors across sub-Saharan Africa34, An. coluzzii is a pervasive malaria vector species in West Africa35 and An. stephensi is the primary urban vector species in the Indian subcontinent36, which has become an invasive rural species in the Horn of Africa37 and has recently been detected in Nigeria38. Study results demonstrated significant heterogeneity in mortality-dose responses following broflanilide exposure, between Anopheles species (e.g. An. stephensi SK vs. An. gambiae Kisumu KCMUCo), within Anopheles species (e.g. An. arabiensis KGB vs. An. arabiensis SENN) and even between the same insectary strain maintained at different testing facilities (An. gambiae Kisumu KCMUCo vs. An. gambiae Kisumu NIMR). Across all vector species tested, the ranges of DC generated by this study were 1.126 μg/ml to 54.00 μg/ml (LC99 × 2) or 0.7437 μg/ml to 17.82 μg/ml (LC95 × 3). These estimates provide an initial benchmark for broflanilide susceptibility monitoring, as part of ongoing VECTRON T500 community trials in Tanzania and Benin21. Further studies will be required on different Anopheles species and populations in order to identify what will be the definitive DC for broflanilide susceptibility monitoring in conjunction with the use of VECTRON T500 in malaria vector control programmes.

Our findings align with previous studies which estimated the LC95 × 3 = 11.91 μg/ml [95% CI 8.253–21.618]28 with An. gambiae Kisumu LITE and LC95 × 3 = 210 μg/ml [95% CI 115.5–423.3]20 with An. gambiae Kisumu CREC. The difference between the DCs in these studies may be explained by a difference in the method for coating bottles used in bioassay testing. In the current study, technical grade broflanilide was dissolved in acetone with 800 ppm Mero® (81% rapeseed oil methyl ester), as recommended by the commercial manufacturer. Although the role of Mero® in the pickup and uptake of some insecticides is not yet fully understood, it is known that it prevents insecticide crystallization, which can inhibit absorption across the insect cuticle, allowing broflanilide to remain in an amorphous state throughout bioassay testing. The addition of Mero®, therefore, increases the efficacy of broflanilide in bottle bioassays with mosquitoes, i.e., it decreases the concentration of broflanilide needed for lethality. Similarly, the efficacy of clothianidin in bottle bioassays has also been shown to be enhanced by the inclusion of Mero® when coating bottles39.

The differences between bioassay results using the same mosquito strain (An. gambiae Kisumu) maintained at two separate testing facilities (KCMUCo and NIMR) raises some interesting questions regarding direct comparability of insectary colony data. Differences in mosquito rearing conditions, including larval rearing conditions (e.g. crowding, access to nutrition)40, time of testing (e.g. night or day)41,42, temperature and humidity43, mosquito age44 and physiological status45 can have a significant effect on observed bioassay mortality. Whilst every effort is made to maintain standardized test conditions, according to WHO protocols30, even differences of 4 °C during holding periods can have a significant effect on mosquito mortality43, with lower temperatures associated with reduced mortality. Finally, an unascertainable amount of variation between the An. gambiae Kisumu strains maintained at different testing facilities may be attributable to long-term genetic divergence, and in turn, differences in relative colony fitness since these mosquito populations have been maintained in separate facilities for more than a decade. These observations support periodic in-depth strain characterization at both phenotypic and genotypic levels, as has been reported for recently colonized insecticide-resistant colonies46,47, to strengthen future laboratory screening of new insecticides.

A secondary objective of this study was to investigate whether there was any biological basis for cross-resistance to broflanilide via the A296S mutation in the GABA-gated chloride receptor leading to dieldrin resistance (rdl). Despite the ban of dieldrin decades ago, the A296G and A296S rdl mutations have persisted in some contemporary vector populations at high frequencies48,49,50. We observed no evidence for cross-resistance to broflanilide in this study supporting its deployment in areas of pre-existing rdl; indeed, lower concentrations of broflanilide were required to induce complete mortality of the dieldrin-resistant An. arabiensis SENN strain, compared to its susceptible counterpart, An. arabiensis KGB, which may in part be explained by the fitness costs associated with highly insecticide-resistant populations, as shown in previous field studies51,52. This was further reinforced by a lack of association between the A296S rdl mutation and the outcomes of broflanilide bottle bioassays. Three additional amino acids, which surround the broflanilide binding pocket in the GABA receptor, have been identified that can disrupt insecticide binding: G331, I272 and L27628. Screening of the Ag1000 genome data has failed to identify any naturally-occurring mutations in these amino acids in Anopheles field populations28. These genetic regions warrant inclusion in newly developed amplicon-sequencing panels53, which are being rolled out to monitor insecticide resistance across Anopheles vector populations, in conjunction with standard insecticide susceptibility monitoring.

The variability in mortality-dose response to broflanilide, evidenced in this study, strongly advocates for further broflanilide DC testing, using additional insecticide-susceptible Anopheles colonies, particularly an An. funestus strain; this vector species predominates across southern sub-Saharan Africa and plays an increasing role in malaria transmission in areas where other vector species have been controlled by insecticidal interventions54. Unfortunately, this was not feasible for inclusion in this multi-centre study, due to notorious difficulties rearing this particular species under controlled insectary conditions55. Broflanilide testing using wild Anopheles populations is also needed to demonstrate the efficacy of this novel insecticide to control pyrethroid-resistant vectors and to assess variability in the tolerance of these populations to broflanilide. Previous laboratory studies have demonstrated a lack of cross-resistance between mechanisms of resistance possessed by pyrethroid-resistant insectary strains and broflanilide20,28. However, a plethora of complex coinciding, insecticide resistance mechanisms can be found in natural Anopheles populations56,57,58,59,60,61, which are not adequately reflected in genetically homogenous insectary colonies.

Conclusions

This multi-centre study, using six Anopheles insectary colonies, representing major malaria vector species, determined the putative discriminating concentration for broflanilide to range between LC99 × 2 = 1.126 to 54.00 μg/ml or LC95 × 3 = 0.7437 to 17.82 μg/ml. Comparison of the susceptibility of dieldrin-resistant and -susceptible An. arabiensis colonies provided no phenotypic or genotypic evidence for cross-resistance to broflanilide via the A296S rdl mutation in the GABA-gated chloride receptor leading to dieldrin resistance. Use of the adjuvant Mero® increased broflanilide efficacy, highlighting the need to standardize bottle bioassay testing for this new insecticide. Differences in bioassay results using the same mosquito strain (An. gambiae Kisumu) maintained at two separate facilities raised issues regarding direct comparability of insectary colony data and emphasizes the need for periodic in-depth strain characterization to strengthen future laboratory screening of new insecticides. Our study findings provide a benchmark for broflanilide susceptibility monitoring as part of ongoing VECTRON T500 community trials in Tanzania and Benin.

Methods

Mosquito strains

Six Anopheles insectary colonies were used for this multi-centre evaluation: susceptible An. gambiae Kisumu (KCMUCo and NIMR), susceptible An. coluzzii N’Gousso (LSHTM), susceptible An. stephensi SK (LSHTM), susceptible An. arabiensis KGB (LSHTM) and dieldrin-resistant An. arabiensis SENN (LSHTM). An. gambiae Kisumu (KCMUCo and NIMR) is a laboratory strain colonised in 1953 from Kenya. Susceptible An. gambiae Kisumu at KCMUCo is routinely characterized three to four times in a year, with respect to body weight, wing length and both phenotypic and genotypic resistance profiles. Prior to this study, An. gambiae Kisumu KCMUCo was confirmed as susceptible to alpha-cypermethrin (pyrethroid) and bendiocarb (carbamate) in WHO tube tests62; species identification was confirmed by PCR63 and screening for L1014S, L1014F and G119S-Ace-1 did not detect the presence of any insecticide resistance associated mutations in this colony. Susceptible An. gambiae Kisumu at NIMR is routinely characterised every three months, with respect to phenotypic and genotypic resistance profiles. Prior to this study, An. gambiae Kisumu NIMR was confirmed as susceptible to deltamethrin and permethrin (pyrethroids) in WHO tube tests62; species identification was confirmed by PCR63 and screening for L1014S and L1014F did not detect the presence of any insecticide resistance associated mutations in this colony. An. coluzzii N’gousso (LSHTM) is a laboratory-strain colonised in 2006 from field mosquitoes collected in Cameroon. CDC bottle bioassays using deltamethrin and permethrin have established this colony as pyrethroid-susceptible64; PCR has confirmed species identification and that this colony lacks L1014S and L1014F mutations65 . An. stephensi SK (LSHTM) is a laboratory strain colonised from Pakistan in 198266. CDC bottle bioassays using deltamethrin and permethrin have established that this colony is pyrethroid-susceptible64 and species identification is regularly confirmed on the basis of morphological features67. An. arabiensis KGB (LSHTM) is a laboratory strain colonised in 1975 from Zimbabwe. An. arabiensis SENN (LSHTM) is a laboratory strain colonised in 1969 from Sudan68. This strain has been exposed to dieldrin and confirmed resistant due to the GABA-gated chloride receptor mutation (Ala296Ser). Further characterization of the latter two strains is described below.

In all three testing facilities, all life-cycle stages of colony mosquito populations were maintained under standard insectary conditions (25–27 °C, 80% relative humidity, light:dark cycles of 12-h each). In LSHTM mosquito larvae were reared in large white trays, with 12-h light–dark cycles, and fed NISHIKOI staple fish food pellets (Nishikoi, UK). In KCMUCo and NIMR, mosquito larvae were reared in large white round bowls, with 12-h light–dark cycles, and fed with TetraMin (Tetra, U.S.).

Adult mosquitoes were kept in cages of ~ 30 × 30 × 30 cm at varying densities, with 10% glucose provided ad libitum. In LSHTM, colony cages were maintained by regular blood feeding using a Hemotek feeder. In KCMUCo and NIMR, colony cages were maintained by regular blood feeding on Guinea Pigs.

Broflanilide discriminating concentration testing

A discriminating concentration is defined as the concentration of insecticide that in a standard period of exposure, is used to discriminate the proportions of susceptible and resistant phenotypes in a sample of a mosquito population29. Discriminating concentration testing of broflanilide was undertaken using the CDC bottle bioassay method, but with minor modifications to the published guidelines (Fig. 1)64. Probit analysis was used to determine thirteen concentrations of broflanilide for testing (100, 46.4, 21.5, 10, 4.6, 2.2, 1, 0.46, 0.22, 0.1, 0.046, 0.022 and 0.01 μg/ml). Technical grade broflanilide (Mitsui Agro, Inc., Japan) was dissolved in acetone with 800 ppm Mero®; the adjuvant Mero® was used to ensure the insecticide was distributed evenly throughout each bottle and to prevent crystallisation of broflanilide during the conduct of bioassays. Control bottles consisting of acetone alone and acetone + 800 ppm Mero® were run in parallel during each bioassay.

Each Wheaton 250 ml bottle and cap was coated using 1 ml of insecticide solution by rolling it and inverting the bottle. In parallel, control bottles were coated with either 1 ml acetone or 1 ml acetone + 800 ppm Mero® per bottles. Once coated, all bottles were covered with a cotton sheet and left to dry in the dark overnight; and were washed thoroughly and re-coated before every test. During bioassays, replicates of 20–25, two-to-five-day old, unfed female mosquitoes were exposed for 1-h. Mosquito mortality was recorded every 15 min up to 1 h. Surviving mosquitoes were supplied with 10% glucose and held for 72-h, with mortality recorded every 24-h.

WHO insecticide susceptibility tests

To confirm the resistance profiles of both of the An. arabiensis colonies used in this study, WHO susceptibility tests were performed using An. arabiensis SENN (dieldrin-resistant) and An. arabiensis KGB (dieldrin-susceptible) to measure dieldrin susceptibility, following standard procedures69. Replicates of 20–25, two-to-five-day old, unfed female mosquitoes were released into WHO holding tubes. After acclimatization of the mosquitoes for one hour in the vertical position, mosquitoes were blown into exposure tubes containing WHO dieldrin (0.4% and 4%) impregnated filter papers or control papers containing risella oil (Universiti Sains Malaysia, Malaysia). Knock-down was recorded every 15 min up to 1-h. After the 60-min exposure, mosquitoes were transferred back to the holding tubes and mortality was recorded after 24-h.

PCR screening for rdl

Genomic DNA was extracted from 290 An. arabiensis SENN which underwent broflanilide bioassay testing and 219 An. arabiensis SENN which were exposed to dieldrin. Individual mosquitoes were homogenized in a Qiagen TissueLyser II (Qiagen, UK) with sterilized 5 mm stainless steel beads for 5 min at 30 Hz and incubated overnight at 56 °C. DNA was extracted using DNeasy® 96 Blood and Tissue Kits (Qiagen, UK), according to the manufacturer’s protocol.

Individual mosquitoes were identified to species-level using species-specific PCR primers for An. gambiae s.s. and An. arabiensis (Table 3)63. Each 20 μl reaction contained 20–40 ng of gDNA, 10 μl HotStart Taq 2X Master Mix (New England Biolabs, UK) and 25 pmol/ml of primers AR-3T, GA-3T and IMP-UN. Prepared reactions were run on a BioRad T100™ thermal cycler with the following conditions: 95 °C for 5 min, followed by 30 amplification cycles (95 °C for 30 s, 58 °C for 30 s, 72 °C for 30 s) and a final elongation step at 72 °C for 5 min. PCR products were visualised on 2% E-gel agarose gels in an Invitrogen E-gel iBase Real-Time Transilluminator. A Quick-Load® 100 bp DNA ladder (New England Biolabs, UK) was used to determine band size. PCR products of 387 bp or 463 bp were indicative of An. arabiensis or An. gambiae s.s., respectively, relative to positive controls; no-template negative controls were included with all reaction runs.

Table 3 PCR primer and probe sequences.
Full size table

The presence of the A296S rdl mutation in An. arabiensis was determined using a TaqMan assay70. Each 20 μl reaction contained 20–40 ng of gDNA, 10 μl 2X PrimeTime® Gene Expression Master Mix (Integrated DNA Technologies, USA), 800 nM of primers SerRdlF and SerRdlR and 200 nM of probes WT2 and Ser (Table 3). Prepared reactions were run on a Stratagene Mx3005P qPCR system with the following conditions: 95 °C for 10 min, followed by 40 amplification cycles (95 °C for 10 s, 60 °C for 45 s), and lastly a dissociation curve. No-template negative controls were included with all reaction runs. The presence of a wild-type individual was indicated by a substantial increase in HEX signal, the presence of the A296S rdl mutation was indicated by a substantial increase in FAM signal; increase in both signals indicated a heterozygote.

Data analysis

Discriminating concentration (DC) determination was undertaken using BioRssay71 in RStudio v4.0.272. Mortality-dose regression analysis using a generalized linear model was performed per mosquito strain. Lethal doses for 50%, 95% and 99% (LC50, LC95 and LC99) with 95% confidence intervals were calculated. The LC95 value was multiplied by three to determine the DC as per the Lees et al. method31. The DC was also calculated by multiplying the LC99 by two as per the WHO approach32. Differences in dose-mortality responses between strains were evaluated using pair-wise comparisons with Bonferroni correction. All other statistical analyses were conducted in GraphPad Prism 9.4.0.

Ethics approval

Ethical approval for the study was obtained from the London School of Hygiene and Tropical Medicine (LSHTM; ref#26035) and the National Institute for Medical Research (NIMR) in Tanzania (NIMR/HQ/R.8a/VOL.IX/3520). KCMUCo and NIMR obtained approval from the Animal Welfare and Ethical Review Board of LSHTM (ref#2019-14) for use of animals for mosquito maintenance. Study procedures and reporting are in accordance with the Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines. All study procedures were performed in accordance with relevant guidelines and regulations.