Interaction Between Insecticide Resistance-Associated Genes and Malaria Transmission in Anopheles Gambiae S. L. During a Cluster-Randomized Controlled Trial of A “lethal House Lure” in Central Côte D’ivoire

There is evidence that the Kdr L1014F and Ace-1 R G119S mutations involved in pyrethroid and carbamate resistance in Anopheles gambiae inuence malaria transmission in sub Saharan Africa. This is likely due to changes in behavior, life history, vectorial competence and capacity. In the present study, performed as part of a two-armed cluster randomized controlled trial (CRT) evaluating the impact of household screening plus a novel insecticide delivery system (In2Care EaveTubes), we investigated the distribution of insecticide target site mutations and their association with the infection status in wild An. gambiae s.l populations. selected with Plasmodium Kdr and using quantitative chain The alleles were compared between An. coluzzii and An. gambiae and between infected and uninfected groups for each species. specimens (p > 0.05).


Introduction
Anopheles gambiae complex mosquitoes are the main malaria vectors in sub-Saharan Africa [1]. Its remarkable vectorial capacity [2] is largely due to its propensity to blood feed on human and rest indoors [3]. This great ability to adapt to human behaviour led to the development of insecticide-based vector control measures targeting indoor biting. These measures are primarily long lasting insecticidal nets (LLINs) and indoor residual spraying (IRS) and are used to limit the human-vector contacts and reduce mosquito survival [4]. These insecticide-based vector control tools have been highly effective against malaria vectors by considerable reductions in disease burden [5] However, long-term effectiveness of both strategies is threaten by the emergence of insecticide resistance in malaria vector populations [6,7].
There are several mechanisms responsible for insecticide resistance of which metabolic and target site resistance are the most recurrent [8][9][10]. Metabolic resistance increases enzymes responsible for the insecticide degradation, while modi cation of the insecticide target site prevents the molecule from binding the site. The molecular basis of resistance mediated by target site mutations has been characterized in several mosquito populations [11][12][13]. For example, the G119S mutation in the Ace-1 R gene (a single amino acid substitution, from a glycine to a serine at the locus 119 in the acetylcholinesterase catalytic site) is responsible for organophosphate and carbamate resistance among malaria vectors in West Africa [14]. Likewise, the L1014F gene also called the Kdr-west mutation (an amino acid substitution, from leucine to phenylalanine in the voltage gated sodium channel gene, at the 1014 locus typically causing knock down resistance (kdr)) is responsible for pyrethroid and dichlorodiphenyltrichloroethane (DDT) resistance in mosquito populations [12].
Despite the rise of insecticide resistance, its operational signi cance has never been elucidated clearly. In many instances, insecticide-based tools seem to continue to protect against malaria [15][16][17][18] whereas a community trial of LLINs clearly demonstrated that resistance is having an impact [19]. Resistance is dynamic and therefore cannot be randomized to assess its epidemiological impact. Several studies have evaluated the association between single insecticide resistance genes mutation (Kdr or Ace-1 R ) and vectorial competence in An. gambiae [20][21][22]. Nerveless, this were laboratory assays utilizing mosquito colonies or wild strains infected with malaria parasites in laboratories. The coexistence of both Kdr and Ace-1 R in wild population of An gambiae s.l. is common in west Africa, including Côte d'Ivoire [23,24]. The impact of such association on vectorial competence has never been studied.
We took advantage of a two-armed cluster randomized controlled trial evaluating the impact of household screening plus a novel insecticide delivery system (In2Care EaveTubes) to capture mosquitoes in study villages around Bouaké by human landing catches, between May 2017 and April 2019. Mosquitoes were identi ed to species and then genotyped for Kdr L1014F and Ace-1 R G119S mutations using quantitative polymerase chain reaction assays and the frequencies of the two alleles were compared between An. coluzzii and An. gambiae and then between infected and uninfected groups for each species.

Study area
The trial was conducted from May 2017 to April 2019 in central Côte d'Ivoire. The methodology used in this study has been well described by Sternberg et al. [25]. Brie y, forty (40) villages were identi ed within a 60 km radius within the district of Bouaké. All households in the study villages received insecticide treated nets while half of the study villages (20) received screening (S) plus In2Care eave tubes (ET).

Mosquito collections and processing
Mosquito collection process was initially described by Sternberg et al. [25]. Each month, mosquitoes were sampled by human landing catches (HLC) both indoors and outdoors at four randomly selected houses in each of the 40 study villages. HLC were done per month during the trial, from 6 p.m to 8 a.m the following day. Mosquitoes collected were sorted and morphologically identi ed to species using key described by Gillies and Meillon [26] and counted. All malaria vectors stored for further analysis, but for the interaction study, only An. gambiae s.l., the main malaria vector in Côte d'Ivoire was considered.

DNA extraction
Polymerase chain reaction (PCR) assays were used to assess sporozoite prevalence in a monthly random sub-sample of up to 30 females per village. Mosquitoes were identi ed to sibling species and kdr L1014F and Ace-1 R G119S mutations detected. Genomic DNA was extracted from the head and thorax of individual females using cetyl trimethyl ammonium bromide (CTAB) 2 % as described by Yahouedo et al. [27].

Detection of Plasmodium infection
Plasmodium spp. (P. malariae, P. falciparum, P. ovale, and P. vivax) infection was detected by quantitative real-time PCR according to Mangold et al.[28]. The sequences of the primers were synthesized and supplied by Euro ns Genomics (Ebersberg, Germany) and were as follows: forward PL1473F18 (5′-TAA CGA AGA ACG TCT TAA-3′) and reverse PL1679R18 (5′-GTT CCT CTA AGA AGC TTT-3′). The reactions were prepared in a total reaction volume of 10 µl, which contained 2 µl of 5x HOT FIREPol® EvaGreen® qPCR Mix Plus (Solis Biodyne, Tartu, Estonia), 0.3 µl of each primer, 6.4 µl of sterile water, and 1 µl of DNA template. The real-time PCR mixture were preincubated at 95°C for 12 min followed by ampli cation for 50 cycles of 10 sec at 95°C, 5 sec at 50°C and 20 sec at 72°C with uorescence acquisition at the end of each cycle. Characterization of the PCR product was performed with melt curve analysis of the amplicons (95°C for 60 sec, 60°C for 60 sec, then 60°C to 90°C for 1 sec, with uorescence acquisition at each temperature transition. Plasmodium species were identi ed by melting curve generated at different temperatures (i.e., P. malariae: 73.5-75.5°C.; P. falciparum: 75.5-77.5°C; P. ovale: 77.5 to 79.5°C and P. vivax: 79.5 to 81.5°C).
Detection of Kdr L1014F mutation in An. gambiae s.l.
Detection of the Kdr L1014F mutation was performed using the TaqMan real time PCR assay as described by Bass et al. [30]. The reactions were carried out in a total reaction volume of 10 µl, which contained 2 µl of the 5x HOT FIREPol® Probe Universal qPCR Mix (Solis Biodyne, Tartu, Estonia), 0.125µl primer/probe mix, 6.875 µl of sterile water, and 1 µl of DNA template.
Primers Kdr-Forward (5'-CATTTTTCTTGGCCACTGTAGTGAT-3'), and Kdr-Reverse (5'-CGATCTTGGTCCATGTTAATTTGCA-3') were standard oligonucleotides with no modi cation. The probes were labelled with two distinct uorophores: VIC to detect the susceptible allele and FAM to detect the resistant allele. Ampli cations were performed on the LightCycler® 96 Systems real-time qPCR machine (Roche LifeScience, Meylan, France) with cycling conditions of 95°C for 10 min, followed by 45 cycles at 95°C for 10 sec, 60°C for 45 sec and 72°C for 1 sec. FAM and VIC uorescences were captured at the end of each cycle and genotypes were called from endpoint uorescence using the LightCycler® 96 software (Roche LifeScience, Meylan, France) for results analysis.
Allelic and genotypic frequencies for insensitive acetylcholinesterase phenotypes characterized by the G119S mutation were determined in An. gambiae s.l., using the TaqMan assay, according to Bass et al. [31]. The reactions were carried out in a total reaction volume of 10 µl, which contained 2 µl of the 5x HOT FIREPol® Probe Universal qPCR Mix (Solis Biodyne, Tartu, Estonia), 0.125µl primer/probe mix, 6.875 µl of sterile water, and 1 µl of DNA template. Primers Ace-1-Forward (5'-GGC CGT CAT GCT GTG GAT-3'), and Ace-1-Reverse (5'-GCG GTG CCG GAG TAG A-3') were standard oligonucleotides with no modi cation. The probes were labelled with two distinct uorophores: VIC to detect the susceptible allele and FAM to detect the resistant allele. Ampli cations were performed on the LightCycler® 96 Systems real-time qPCR machine (Roche LifeScience, Meylan, France) with cycling conditions of 95°C for 10 min, followed by 55 cycles at 92°C for 15 sec, 60°C for 60 sec and 72°C for 1sec. FAM and VIC uorescences were captured at the end of each cycle and genotypes were called from endpoint uorescence using the LightCycler® 96 software (Roche LifeScience, Meylan, France) for results analysis.

Statistical analysis
To analyse the distribution of Kdr L1014F and Ace-1 R G119S genotypic and allelic frequencies, data collected in the same study arm between May 2017 and April 2019 were compared between species. The association between genotypic and allelic frequencies for these mutations and infection status were determined using Pearson Chi-square test in R software (version 4.0.3). Both Kdr L1014F and Ace-1 R G119S combined genotypic frequencies distribution within infection status in each species were also included. The Fisher test was used when individual number available for test was less than 30. The signi cance threshold was set at 5%. Odds ratios were computed to assess the strength of difference or association between resistance alleles and infection status. The allelic frequencies were tested to Hardy-Weinberg equilibrium (HWE) conformity using the Exact HW test and calculated as follows: NB: Kdr L1014F and Ace-1 R G119S mutations comprise three genotypes expressing different allelic variants on the targeted loci. RR indicates the resistant homozygous genotype; RS, the heterozygous genotype and SS, the susceptible homozygous. The resistant (R) and susceptible (S) alleles are possible versions of these genes.  . 11223). Verbal and written informed consents, using local language, was obtained from all participants (mosquito collectors and household heads) prior to their enrolment in the study. Mosquito collectors were vaccinated against yellow fever and the project provided treatment of con rmed malaria cases free of charge for any study participant according to national policies.

Results
Genotypic and allelic frequency distribution in Anopheles gambiae s.l. species Out of 1,392 mosquitoes analysed in PCR, 1255 were successfully identi ed to species (< 10% failure rate). Both An. gambiae (n = 624; 49.7%) and An. coluzzii (n = 631; 50.3%) were found. For each species, the proportions of infected vs uninfected were similar ( Fig. 1). There were no signi cant differences in the allelic frequency of Kdr or Ace-1 R between the control and Eave tube areas for each species (p 0.05) ( Table 1).

Insecticides resistance genes and infection status
Genotypic and allelic frequencies of Kdr L1014F and Ace-1 R G119S genes among infected and uninfected individuals are shown in Table 3. Regardless of the species and study arms, there were no signi cant differences in genotypic or allelic frequencies between infected and uninfected individuals (p 0.05) ( Table 3).

Discussion
This study evaluated the effects of the Kdr L1014F and Ace-1 R G119S genes on Plasmodium sp. infection status in natural An. gambiae s.l. populations. The presence of both An. coluzzii. and An. gambiae in similar proportions in this longitudinal study was consistent with previous studies in the area of Bouaké [24,32] but it contrasts with another study conducted in adjacent areas within Bouaké which found An. coluzzii to be predominant [33]. The difference observed is likely due to the study sampling period covering both rainy and drying seasons in our study compared to rainy season only [33]. We observed no difference in infection rate between An. gambiae and An. coluzzii. This aligns with previous studies conducted in Burkina Faso and Senegal [21,34], which reported equivalent Plasmodium susceptibility to these species. Our current results demonstrate that both sibling species are equally dangerous vectors of human malaria in the central region of Côte d'Ivoire.
With regard to resistance genes, there were no signi cant differences in the allelic frequency of Kdr or Ace-1 R between the control and Eave tube areas regardless of the species. This is because Kdr was already close to xation in An. gambiae s.l. species prior to the eave tubes intervention (> 80%) [24] leaving tiny window for further selection. Also, the insecticide deployed in the eave tube trial was a pyrethroid (beta-cy uthrin) [35] which could not induce a selection pressure on the Ace-1 R since this gene is associated with organophosphate and carbamate resistance [14,24].
We found signi cantly higher Kdr L1014F and Ace1 R G119S genotypic and allelic frequencies in An. gambiae than in An. coluzzii, which was in agreement with observations of Koukpo et al.[36] in Benin and by Zogo et al. [37] in Côte d'Ivoire. There were 59 times greater probability of encountering Kdr L1014F resistance allele of An. gambiae relating to An. coluzzii, whereas the frequency of Kdr L1014F heterozygous individuals was reversely higher in An. coluzzii (42.95%) than in An. gambiae (1.12 %). This clearly highlighted a deviation from Hardy-Weinberg expectations within both malaria vector species for the Kdr L1014F mutations. It is possible that evolutionary factors affect mosquito population structure through the excess use of insecticides. These factors induce the selection of rare and existing mutations in natural population of both species which become later variably widespread [38].
Furthermore, Ace-1 R G119S allelic frequency in An. gambiae was signi cantly higher than in An. coluzzii, although the amplitude was moderate. The low proportion (< 10%) of homozygous resistant (RR) genotypes observed in An. gambiae and An. coluzzii population could indicate the high tness cost associated with Ace-1 R G119S gene [39,40]. Conversely, this tness cost associated with Ace-1 R seems to be resorbed by the duplication of this gene which induced various heterozygous genotypes by increasing their proportions [41]. Further studies focusing on Ace-1 R genotype distribution, including the duplication in An. gambiae s.l. is needed. Our study showed that in areas where Kdr L1014F and Ace-1 R G119S coexist in An. gambiae s.l., the frequency of individuals bearing the Kdr L1014F RR genotype appeared signi cantly higher in An. gambiae than in An. coluzzii. By contrast, the frequencies of those bearing the Kdr L1014F heterozygous genotype were reversely signi cantly higher in An. coluzzii than An. gambiae, con rming the trend when this genotype is in isolation. This is the rst study evaluating the distribution of individual An. gambiae s.l. bearing both mutations inside them. It calls for further studies to better understand the genotypic structure of their combinations.
The vectorial competence in association with resistance genes was investigated. We found no evidence of association between Plasmodium infection status and Kdr L1014F or Ace-1 R G119S genes. These results were similar to those found in Guinea where these target site mutations (Kdr L1014F or Ace-1 R G119S ) were not associated with Plasmodium infection in wild An. gambiae [42], but that the phenotypic resistance was rather associated with infection. By contrast, a study in Tanzania found a link between Kdreast and Plasmodium infection in wild An. gambiae [43].
However, the non-association between Plasmodium infection status and resistance genes under natural condition contrasts with several other studies reporting that resistance associated genes affect vector competence to transmit Plasmodium parasites [20,21,44]. Reasons for the difference could be three-fold: (i) These contrasting results could derived from studies that used colonies maintained in laboratory over years, which can decrease resistance, including loss of genetic diversity [45,46]. (ii) Some genetic susceptibility studies do not take account of additional factors in uencing competence in natural vector population; e.g. mosquito blood feeding rate, age at infection, longevity, exposure to insecticide and other pathogens that could in uence mosquito immune status [47][48][49][50][51]. Natural infection study also implies the effects of ecology and behavior on vectorial competence [52,53]. (iii) Resistance is a package encompassing mutations plus metabolic components with different functions; therefore isolating one from the other, may not be representative of the phenotypic resistance. The absence of association between genotypes in combination (Kdr L1014F-Ace-1 R G119S) with infection status in An. coluzzii or An. gambiae requires further attention by control programmes, given that this is now common observation in many parts of west Africa [13,24].

Conclusion
We saw no signi cant association of the Kdr L1014F and Ace-1 R G119S mutations alone or in association with infection status in wild An. gambiae and An. coluzzii demonstrating similar competence for Plasmodium transmission within Bouaké areas.
Nevertheless, the frequencies for the Kdr and Ace-1 R genotypes and alleles were signi cantly higher in An. gambiae than in An. coluzzii. Additional factors in uencing competence in natural vector population and those outside alleles or genotypes measurements contributing to resistance should be consider when establishing link between insecticide resistance and vector competence. Combination of Kdr L1014F and Ace-1R G119S genotypic frequencies between infected and uninfected groups, in each study arm, Error bars represent 95% con dence intervals. SET: Screening plus In2Care Eave Tubes. For all combined genotypes, the two rst alleles refer to Kdr genotypes and the two last refer to Ace-1R genotypes.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.