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). The latter strain is dieldrin resistant due to the rdl A296S GABA-gated chloride receptor mutation. An. gambiae Kisumu (KCMUCo and NIMR) is a laboratory strain colonised in 1953 from Kenya. An. coluzzii N’gousso (LSHTM) is a laboratory-strain colonised in 2006 from field mosquitoes collected in Cameroon. PCR at LSHTM has confirmed, that this colony are An. coluzzii62 . An. stephensi SK (LSHTM) is a laboratory strain colonised from Pakistan in 198263. 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 Sudan64. The strain has been exposed to dieldrin and confirmed resistant due to GABA mutation (Ala296Ser).
In all three testing facilities, all life-cycle stages of colony mosquito populations were maintained under standard insectary conditions (25-27oC, 80% relative humidity, light:dark cycles of 12:00h each). In LSHTM mosquito larvae were reared in large white trays, with 12-hour 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-hour light-dark cycles, and fed with TetraMin (Tetra, U.S.).
Adult mosquitoes were kept in cages of ~30 x 30 x 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 (Figure 1)65. 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 and o 1ml 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 hour. Mosquito mortality was recorded every 15 minutes up to 1 hour. Surviving mosquitoes were supplied with 10% glucose and held for 72 hours, with mortality being recorded every 24 hours.
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 procedures66. 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 papers or control papers containing risella oil (Universiti Sains Malaysia, Malaysia). Knock-down was recorded every 15 minutes up to 1 hour. After the 60-minute exposure, mosquitoes were transferred back to the holding tubes and mortality was recorded after 24 hours.
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 minutes at 30 Hz and incubated overnight at 56oC. 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: AR-3T (5’-GTGTTAAGTGTCCTTCTCCGTC-3’; specific for An. arabiensis), GA-3T (5’-GCTTACTGGTTTGGTCGGCATGT-3; specific for An. gambiae s.s.) and IMP-UN (5’-GCTGCGAGTTGTAGAGATGCG-3’; common for all species)67. Each 20 μl reaction volume 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: 95oC for 5 minutes, followed by 30 amplification cycles (95oC for 30 seconds, 58oC for 30 seconds, 72oC for 30 seconds) and a final elongation step at 72oC for 5 minutes. 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.
The presence of the A296S Rdl mutation in An. arabiensis was determined using a TaqMan assay68. Each 20 μl reaction volume contained 20-40 ng of gDNA, 10 μl 2X PrimeTime® Gene Expression Master Mix (IDTTM, USA), 800 nM of primers SerRdlF (5’-TCATATCGTGGGTATCATTTTGGCTAAAT-3’) and SerRdlR (5’-TCGTTGACGACATCAGTGTTGT-3’) and 200 nM of probes WT2 (5’/HEX/TTACACCTA/ZEN/ATGCAACACG/3IABkFQ/3’) and Ser (5’/FAM/CACCTAATG/ZEN/AAACACG/3IABkFQ/3’). Prepared reactions were run on a Stratagene Mx3005P qPCR system with the following conditions: 95oC for 10 minutes, followed by 40 amplification cycles (95oC for 10 seconds, 60oC for 45 seconds), 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 determination was undertaken using BioRssay69 in RStudio v4.0.270. 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 discriminating concentration (DC) as per the Lees et al. method32. The DC was also calculated by multiplying the LC99 by two as per the WHO approach33. 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.