Malaria is a blood-borne disease caused by various plasmodium species that live within erythrocytes. However, while malaria can, on rare occasions, be transmitted via blood transfusions or congenital transmission, it is predominantly transmitted through the bite of a female mosquito of the Anopheles genus. Consequently, vector control strategies have long been a major component of attempts to limit the spread of infectious diseases transmitted by mosquitoes, such as malaria. This proved to be somewhat successful between 2000 and 2015, as the number of people infected with malaria steadily dropped, due at least in part to the use of insecticide-treated bed nets. However, over the past few years, mosquito insecticide resistance has become increasingly evident. Here, we briefly explore the major insecticide resistance mechanisms found in mosquitoes and discuss how we think studies in fruit flies can improve our understanding of mosquito insecticide resistance and contribute to the identification of new targets for insecticides. Thus, the rationale for this mini-review is not to provide a thorough understanding of the subject of mosquito insecticide resistance, but rather to provide an overview of the research themes in the field to Drosophila melanogaster (Dm) biologists, and briefly draw attention to how Dm genetics and cell biology can be co-opted to address questions relevant to mosquito insecticide resistance. For more thorough discussions of the subject of mosquito insecticide resistance, we refer readers to a number of reviews that have been published in recent years on this subject (Dusfour et al., 2019; Susanna and Pratiwi, 2021; Suh et al., 2023).

Insecticide resistance usually develops gradually, as a result of several mechanisms that culminate in reducing the effectiveness of the insecticide. Nevertheless, there are four main insecticide resistance mechanisms referred to as target-site resistance, metabolic resistance, cuticular resistance and behavioral resistance; and many mosquito species can develop resistance through multiple mechanisms. We begin by briefly reviewing each type of insecticide resistance mechanism.

While the ideal model system for studying mosquito insecticide resistance and vector control are mosquitoes themselves, several reports lend credence to the proposition that Dm are also useful in studying the effects of mutations linked to insecticide resistance in mosquitoes. The well-developed genetic toolkit in Dm, make them an ideal insect model to first identify a broad class of genes that regulate a process with implications for developing insecticides to mosquitoes. Subsequently, more focused studies in mosquitoes examining multiple genes in the same class of genes, may identify the specific genes that regulate the process in mosquitoes. With regard to target-site resistance, both synonymous and nonsynonymous mutations in VGSCs have been associated with target-site resistance; but, in general, the significance of synonymous mutations are unclear. Previously assumed to be benign, growing evidence indicates that synonymous mutations can influence phenotypes by regulating gene expression and/or protein stability via an alteration in mRNA stability, a disruption of splicing activity, a change in the efficiency of miRNA binding, a perturbation of translational activity, and an interference with the function of long non-coding RNAs. Studies in Dm can test the ability of synonymous mutations in VGSCs to impact these cellular processes.

A gene that encodes for a VGSC in Dm is paralytic (para). This gene can encode for about 60 isoforms as a result of alternative splicing (Lin et al., 2009; Lin et al., 2012). As there are varied periods and patterns of expression of some of these isoforms, it would be interesting to determine how knock-in mutations in fruit flies that mimic the relevant synonymous mutations in mosquitoes regulate expression of these isoforms. Further, a Dm strain in which Para is tagged with GFP endogenously exists (Ravenscroft et al., 2020). The relevant knock-in mutations can be created in this strain to ascertain how Para localization and/or stability is affected. Lastly, the GAL4/UAS system can be exploited to overexpress constructs carrying the appropriate mutations (Fischer et al., 1988; Brand and Perrimon, 1993). Similar analyses can be performed for GABAR and AChE.

Fruit flies can also be used to gain a better understanding of how allelic drive can be used to reverse insecticide resistance. In a particularly noteworthy study, CRISPR/Cas9 gene editing of para was used to generate a series of common VGSC mutations found in mosquitoes and tested for susceptibility to various insecticides (Kaduskar et al., 2022). The study revealed that it was possible to replace the resistant allele with a native (susceptible) allele in population cages. Thus, this proof-of-principle study highlights how fruit flies can be employed to elucidate the potential of performing a targeted reversion of an insecticide-resistant strain to a wild-type susceptible state.

Forced expression of the mosquito P450 enzymes, CYP6P9a and CYP6P9b, in Dm conferred tolerance to both permethrin and deltamethrin (Riveron et al., 2013). Nevertheless, some discrepancies have been observed between data obtained from overexpressing detoxification genes in Dm and Anopheles gambiae mosquitoes. In such instances, the natural tendency is to dismiss the observations in Dm. However, the interpretation of such results is complicated by the fact that the outcome of a GAL4/UAS overexpression experiment is largely dependent on the extent and domain of overexpression. Even in fruit flies, where the GAL4/UAS toolkit is well-established, contradictory results can be obtained when the same gene is overexpressed in a specific Dm tissue using different tissue-specific GAL4 lines. Therefore, it appears the yardstick for success when overexpressing detoxification genes should be whether a protective effect against insecticides can be observed at all, as a result of using many GAL4 lines – both tissue-specific and ubiquitous. In this regard, the Gene-Switch system in Dm may be particularly valuable as it allows a dose-dependent expression of a transgene in response to varying concentrations of the RU486 (mifepristone) drug (Roman et al., 2001; McGuire et al., 2004; Nicholson et al., 2008; Ke and Hsu, 2019). While some studies in which a mosquito cDNA was overexpressed in Dm to test the effect of an insecticide have yielded promising results (Pavlidi et al., 2012; Riveron et al., 2013), it is possible that mosquito cDNAs may code for proteins that are incapable of interacting optimally with other endogenous regulatory factors. Hence, when overexpression of a mosquito cDNA fails to rescue Dm from insecticide-induced toxicity, it may be necessary to also test the effect of overexpressing the closest fruit fly ortholog of the mosquito gene. Ultimately, however, any promising results in fruit flies will have to be confirmed in mosquitoes.

Several peptidases, regulators of lipid and carbohydrate metabolism, sodium-calcium exchangers, and signaling molecules are induced alongside GSTs, P450 genes, and esterases in insecticide-resistant strains (Vontas et al., 2005; Liu et al., 2007). A relatively less explored aspect of metabolic insecticide resistance are the signaling pathways that regulate expression of what we refer to as insecticide resistance effector molecules (i.e., GSTs, P450 genes, and esterases). Identifying upstream regulators of these effector molecules is crucial, as it will furnish knowledge about the possibility of dismantling metabolic resistance mechanisms. A breakdown of the metabolic resistance mechanism should, in principle, enable an erstwhile resistant insecticide to be effective again. In one informative study, it was shown that GPCR signaling upregulates the expression of some P450 genes, raising the possibility that inhibitors of this specific GPCR signaling cascade may suppress insecticide resistance (Li et al., 2014). The readily available transgenic RNAi strains and sophisticated genetic tools in Dm make it particularly suited for exploring the regulatory relationships between the multiple genes induced in insecticide-resistant strains.

When mitochondrial function is impaired, it results in the activation of a mitochondrial stress signaling cascade that culminates in the induction of genes required for conferring tolerance to mitochondrial distress. As an example, we recently showed that RNAi-mediated knockdown of isocitrate dehydrogenase 2 (IDH2) leads to an upregulation of genes involved in Fe-S cluster biogenesis, redox and protein homeostasis, GSTs, P450 enzymes, esterases, regulators of lipid and carbohydrate metabolism, transporters, and signaling molecules, among other genes (Murari et al., 2022). This highlights the remarkable similarity in gene expression profile resulting from activating mitochondrial stress signaling or insecticide resistance. Further, it is widely accepted that a combination of piperonyl butoxide (PBO) and pyrethroids is more effective at reducing mosquito populations than using pyrethroids alone (Gleave et al., 2021). As PBO inhibits P450 enzymes, this observation raises the possibility that knocking down genes induced as a result of insecticide resistance may enhance the insecticidal effect of an ineffective insecticide. This is a phenomenon referred to as synergism. Given the similarity between the gene expression profile associated with insecticide resistance and mitochondrial stress signaling, we hypothesize that disrupting some components of the mitochondrial stress signaling cascade identified as a result of IDH2 disruption could synergize with low doses of an insecticide to cause lethality.

Furthermore, we have found that disruption of IDH2 in Dm flight muscles culminates in the induction of a ferroptosis-like form of cell death that may provide opportunities for addressing the issue of insecticide resistance. Ferroptosis is a non-apoptotic form of cell death that is triggered by iron-dependent lipid peroxidation (Stockwell et al., 2017); however, it has not been explored extensively in Dm or mosquitoes. As emerging evidence indicates that there are likely to be Dm-specific aspects of ferroptosis, if similar observations are found in mosquitoes, it may provide an additional mechanism for combating insecticide-resistant mosquitoes.

For the most part, the molecular basis of cuticular resistance is still being unraveled; but some candidate genes have emerged that may regulate expression of cuticle proteins. As a case in point, the laccase gene, which encodes for a diphenol oxidase, is induced in fenvalerate-resistant strains of Culex pipiens pallens (Pan et al., 2009). This gene can be expressed in fruit flies exposed to fenvalerate to examine whether it regulates the expression of cuticle proteins or confers tolerance to the insecticide. Further, fruit flies can be exposed to various insecticides and a time series RNA-Seq experiment performed to assess the order of induction of genes involved in cuticle biogenesis. Subsequently, RNAi and overexpression analyses can be used to decipher the regulatory relationships between the genes and the degree to which they impact insecticide resistance.

The challenge of using fruit flies to study behavioral insecticide resistance lies in the fact that a paradigm for studying it has to be established first, as there is some debate about whether behavioral insecticide resistance mechanisms should be classified as such, or simply regarded as avoidance mechanisms (Zalucki and Furlong, 2017). Nevertheless, one area in which research in fruit flies could prove informative for dissecting the mechanism of behavioral resistance to insecticides, is by examining whether the insecticide in question can trigger an aversive response in the olfactory system. It has been shown that pyrethrum extracts from flower heads of Tanacetum cinerariifolium, which have long been used as insect repellants, evoke olfactory responses to cause aversion in both Dm and Drosophila suzukii (Liu et al., 2021). Further studies revealed that pyrethrin, the major component of pyrethrum, activates at least three distinct odorant receptors – Or7a, Or42b and Or59b – to elicit the insecticide repellent response. A similar set of experiments in Aedes aegypti mosquitoes revealed the importance of AaOr31 in conferring pyrethrum repellence in mosquitoes (Liu et al., 2021). While the result of this olfactory system-induced aversive response is beneficial, conceivably, a comparable phenomenon could occur as a result of insecticide spraying to trigger an aversive response which culminates in behavioral resistance.

The extensive arsenal of genetic tools for probing the Dm olfactory system make it an ideal system to uncover a possible role of the olfactory system in behavioral resistance. The Dm olfactory system senses airborne molecules via the activation of receptors located on olfactory receptor neurons (ORNs). Each ORN expresses G-protein-coupled receptors which elicit a unique odorant response profile for that neuron. Each olfactory receptor functions together with an obligate co-receptor dubbed Orco (olfactory receptor co-receptor). As a start, fruit flies can be exposed to an insecticide and the phenomenon of behavioral resistance examined in orco mutants. If the behavioral insecticide resistance response is abrogated in orco mutants, it will mean further studies dissecting the effect of eliminating individual olfactory receptors may uncover which ones are required for behavioral insecticide resistance.

Established in 2005 through a $50 million grant from the Bill and Melinda Gates Foundation to the Liverpool School of Tropical Medicine, the Innovative Vector Control Consortium (IVCC) has been spearheading efforts to identify new and improved insecticides. In this respect, more than 4.5 million compounds have been evaluated for potential use as insecticides, which resulted in the identification of six classes of new active ingredients that have implications for malaria vector control (see https://www.ivcc.com/research-development/insecticide-discovery-and-development/). Other screening efforts have involved the use of malaria vector species (Lees et al., 2019; Lees et al., 2020). Nevertheless, fruit flies (Dm) can be a complementary model system for identifying new molecular targets for insecticides. As previously alluded to with respect to metabolic resistance, studies in fruit flies can identify how the multiple genes induced in response to insecticides are regulated; and how this can be exploited to suppress insecticide resistance. Moreover, there are many other pharmaceutical products that may be useful insecticides but have not yet been studied extensively in fruit flies. These include pymetrozine and flonicamid, which are modulators of chordotonal organs; semicarbazones, some classes of diamides and several other modulators of nerve and muscle function; and growth and development targets such as ecdysone receptor agonists and inhibitors of respiration. Studies in fruit flies are likely to provide useful information about their roles, but with the ultimate aim of confirming any promising results in mosquitoes, and working with the IVCC to further exploit any promising findings to prevent disease transmission.

Many respiratory inhibitors proposed as insecticides exert their effect largely by altering reactive oxygen species (ROS) production or other mitochondrial processes that do not involve a disintegration of the oxidative phosphorylation (OXPHOS) complexes. Examples include several inhibitors of complex I (CI), such as fenazaquin, pyridaben, fenpyroximate, pyrimidifen, tolfenpyrad and tebufenpyrad. Nonetheless, the broad specificity of these CI inhibitors has limited their use. As a case in point, tebufenpyrad and pyridaben alter mitochondrial dynamics in rat dopaminergic neuronal cultures (Charli et al., 2016). Their broad specificity has also meant that concentrations used in the field, during clinical trials, had to be chosen judiciously. As sub-lethal levels of mitochondrial ROS can activate compensatory stress responses (Owusu-Ansah et al., 2008; Owusu-Ansah and Banerjee, 2009; Owusu-Ansah et al., 2013; Murari et al., 2021; Murari et al., 2022), this may explain, at least in part, why a clinical trial in Tanzania exploring the effectiveness of fenpyroximate and abamectin-treated durable wall liners as a control mechanism for malaria was not successful (Mpangala et al., 2021). We postulate that a more favorable outcome could be achieved with respiratory chain inhibitors that block aspects of the assembly of the OXPHOS system. We have been studying the mechanism of mitochondrial CI assembly in Dm and found that severe RNAi-mediated disruption of multiple CI subunits is lethal (Garcia et al., 2017). We hypothesize that a similar phenomenon may occur in other dipterans such as mosquitoes that could be exploited to develop new insecticides.

Mitochondrial CI is the most elaborate component of the OXPHOS system (Fiedorczuk et al., 2016; Agip et al., 2018; Rhooms et al., 2019). Mammalian CI has 45 subunits organized into two domains of the complex, oriented almost perpendicularly to each other, and referred to as the matrix and membrane domains ( Figure 1 ) (Agip et al., 2019). There are three distinct functional modules of CI dubbed the N, Q, and P modules. The N module contains the flavin mononucleotide (FMN) cofactor and accepts electrons from NADH in the mitochondrial matrix. The Q module is situated between the N module and the membrane domain and transfers electrons to ubiquinone. The proton-pumping P module is essentially the membrane domain; and can be further sub-divided into a proximal P_P and distal P_D module [reviewed in (Vartak et al., 2014; Formosa et al., 2018; Rhooms et al., 2019)]. A total of 14 core subunits contain all the catalytic centers of CI and are conserved from the ancestral enzyme in bacteria to the eukaryotic enzyme. Seven core subunits (NDUFS1, NDUFS2, NDUFS3, NDUFS7, NDUFS8, NDUFV1 and NDUFV2) are encoded by the nucleus, translated in the cytoplasm, and imported into the mitochondrion; while the other core subunits are encoded and translated within the mitochondrion (ND1, ND2, ND3, ND4, ND4L, ND5 and ND6). The 31 remaining subunits are referred to as accessory or supernumerary subunits, as they are not directly involved in performing the bioenergetics functions of CI. During CI assembly, specific subcomplexes consisting of a few CI subunits form largely independently of each other and merge in a stereotypic fashion en route to forming the mature complex (Stroud et al., 2016; Garcia et al., 2017; Formosa et al., 2018). CI assembly factors (CIAFs) are proteins that are usually found in association with specific subcomplexes and assist with the assembly process; but they are subsequently released when assembly is complete. The N, Q and P modules are synthesized from specific subcomplexes or assembly intermediates that can be tracked by immunoblotting or complexome profiling techniques (Guerrero-Castillo et al., 2017; Formosa et al., 2020; Murari et al., 2020; Szczepanowska et al., 2020).

Most of the mammalian accessory subunits are conserved in Dm (Garcia et al., 2017), but the extent to which each accessory subunit contributes to CI assembly differs between organisms. For instance, while RNAi-mediated disruption of NDUFA12 potently impairs CI assembly in Dm flight muscles and can cause lethality, a CRISPR-mediated knockout of NDUFA12 in human cells has minimal effects on CI assembly (Stroud et al., 2016; Garcia et al., 2017). We anticipate that such differences in roles of CI subunits in CI assembly can be exploited to make new arthropod- or diptera-specific CI assembly inhibitors, which can ultimately result in the development of novel insecticides.

There is an ever-present need to identify novel mosquito insecticides that are more effective and selective, and entail new mechanisms of action. This may require homology modeling for identifying novel chemistries and insect-specific modes of regulation of already established targets, as was performed for the housefly VGSC (O’Reilly et al., 2006). Alternatively, it may be necessary to screen for novel sites in genes that are well established insecticide targets, as was performed for the isooxazoline insecticide, A1443, a ligand-gated chloride channel antagonist (García-Reynaga et al., 2013). It is also crucial to uncover entirely novel biochemical targets, as was the case with specific classes of diamides that were found to modulate insect ryanodine receptor function (Sattelle et al., 2008). Molecular targets that regulate developmental pathways in insects, such as cell death, autophagy, molting, and basic organelle function could also be exploited. In fact, the importance of autophagy in the fat body, and ecdysone signaling in regulating molting and other aspects of mosquito physiology, has been described extensively, highlighting their potential utilization as targets of insecticides (Bryant and Raikhel, 2011; Childs et al., 2016; Shaw and Catteruccia, 2019; Werling et al., 2019; Brown et al., 2020; Ekoka et al., 2021; Maharaj et al., 2022). While, undoubtedly, studies in fruit flies can contribute to our understanding of all these putative aspects of mosquito insecticide development, we propose that developing insecticides that disrupt organelle function to cause cell death, especially those that inhibit mitochondrial CI assembly, via a mechanism distinct from how CI assembly is regulated in humans, holds great promise for identifying new and insect-specific targets of insecticides. However, we note that at best, Dm should only be regarded as a complementary model system for studying mosquito insecticide resistance as any promising leads would have to be confirmed in mosquitoes.

EO-A wrote the review and got feedback from KFBH and DV. All authors contributed to the generation of the figure. All authors approve the submission.