Gene silencing through RNAi and antisense Vivo-Morpholino increases the efficacy of pyrethroids on larvae of Anopheles stephensi

Vector-borne diseases are among the main public health threats in the world. According to the WHO, 216 million cases of malaria and 445,000 deaths occurred in 2016 [1]. Although great results in malaria control have been achieved in the past decades, the occurrence of drug resistance in Plasmodium falciparum, in particular against artemisinin and other drugs [2‐7], and insecticide resistance in mosquito populations [1, 8‐11] are threatening the efforts for an effective control of the disease. Insecticides remain the core of all malaria control programmes, despite the diffusion of resistant vector mosquitoes, caused by their heavy use. Pyrethroids, and permethrin in particular, are widely used for indoor residual spraying (IRS) and for the treatment of bed nets. As such, it is of pivotal importance to understand the molecular mechanisms of detoxification in mosquitoes, both in sensitive and resistant populations/strains. This knowledge could lead to the development of strategies aimed to restore sensitivity in resistant populations and to avoid the evolution of resistances in sensitive ones [12‐17].

Several studies, carried out over the years, have identified a number of genes involved in the detoxification of xenobiotics in mosquitoes, such as Glutathione-S-transferase (GSTs) [18‐22], Epsilon glutathione transferase (GSTe) [23], Cytochromes P450 (CYPs) [22], Acetylcholinesterase (AChE1) [24] and ATP Binding Cassette (ABC) transporters [17, 25‐29]. Among these, in particular, the ABCG4 transporter, belonging to the G subfamily of ABCs, has consistently been shown to be up-regulated in response to permethrin treatment, suggesting an important role in detoxification against this insecticide in Anopheles stephensi larvae [22, 25, 26] and adults [28]. A first aim of this study was thus to verify whether the ABCG4 efflux pump plays a role in permethrin detoxification in An. stephensi. To this purpose, assays using siRNAs targeted on the ABCG4 mRNA, were performed to determine whether the inhibition of the expression of this gene increases susceptibility to permethrin.

It was emphasized that down-regulation through RNA interference (RNAi) has been achieved for various detoxification genes, inducing an increase of mosquito sensitivity to different classes of insecticides [14, 30, 31], but this technique has also been used for identifying new resistance candidate genes [32]. RNAi-based tools have also been tested for their biopesticide potential [33] and as sterility inducer [34]. These results highlight the potential of RNAi as a promising research tool towards the development of novel strategies in vector control. Thus, RNAi-based tools hold potential for possible field applications as larvicides [35].

As stated by the WHO, the larval source management (LSM) still represents the backbone of integrated mosquito control programmes, with a large-scale effectiveness, able to complement measures against adult mosquitoes and limit the residual transmission of malaria [36, 37]. In this backdrop, oligonucleotides for gene expression inhibition, like siRNA for RNAi, gained importance as a potential novel class of ecofriendly larvicides, which can target both insecticide-resistant and-sensible vectors [14, 30, 35]. However, the duration of activity of these molecules and their stability (persistence) in the field, such as the external water environment and the inner organism of the target larvae, has still to be checked [38, 39]. The “oral delivery” of dsRNA to Anopheles mosquito larvae, while demonstrating also a systemic knockdown effect of target genes [30], implies a partial degradation of the RNA oligonucleotides in the insect gut [40] and a decrease in their effect [41].

Another antisense gene knockdown technology is the antisense Morpholino (MO), which is based on the action of uncharged molecules able to induce a complementary-based block mRNA translation into protein without degradation of mRNA [39, 42]. The use of these oligonucleotides has achieved excellent results in applications requiring an extreme specificity in complex systems (e.g. embryo development) [39]. These highly stable synthetic oligonucleotides can be also conjugated with a delivery moiety, allowing cell-penetration and the in vivo-uptake. These conjugated molecules, Vivo-Morpholinos (Vivo-MOs), have already been used in cell culture treatment, or in studies in vivo through microinjection [43, 44], electroporation and also through oral administration [45], and bath-immersion [46, 47]. A recent study performed on adult of An. stephensi underlined the suitability of Vivo-MO oral delivery as an efficient method for gene knockdown in mosquitoes [41]. For this reason, in the present work the second aim was to confirm the potential of Vivo-MO through administration in water to larvae, verifying the biological effects (larval mortality) and the effects on gene expression.

As for our results about the RNAi, the expression analysis on the whole larval body are consistent with the data from previous studies, showing that the peak of ABCG4 over-expression occurs after 6 h of exposure to permethrin [26]. This transporter was chosen as a target for silencing because it demonstrated a strong up-regulation among the ABC transporters in An. stephensi in response to permethrin treatment, suggesting its involvement in the detoxification against this insecticide [22, 25, 26]. RNAi-based assays were thus performed to confirm this hypothesis, thus to assess whether the inhibition of ABCG4 has an effect on mosquito mortality, using two siRNA concentrations. At both concentrations, at 0.03 μg/μl and 0.06 μg/μl, siRNAs were able to induce an increased mortality at the two examined time-points (Fig. 1). This increase is correlated with the RNAi efficacy in down-regulating the expression of ABCG4 at both time points at both concentrations (Fig. 2). In this study, a detectable down-regulation using both the 0.03 μg/μl and the 0.06 μg/μl doses was achieved. In particular, after 24 h a down-regulation of around 5.74-fold, compared with insecticide-alone treatment, was achieved with the lower concentration; this downregulation in gene expression was associated with an almost 20% mortality increase. These obtained results are consistent with studies where RNAi targeted on efflux pumps and G-protein-coupled receptor of insecticide resistant larvae of Aedes aegypti and Culex quinquefasciatus induced an increased toxicity, respectively, of temephos and permethrin [14, 52, 53]. These results, and those of the current study, are also coherent with another work on the diamondback moth Plutella xylostella, where the silencing of another ABC transporter, ABCH1 (structurally similar to ABCG sub-family members), led to larval and pupal lethal phenotypes when exposed to the Cry toxin of Bacillus turingiensis [54]. By comparing the results obtained at the two times and at the two concentrations, the increase in larval mortality was not correlated with the dose of siRNA used or with the duration of insecticide exposure (Fig. 2). It is thus possible that at 6 h of exposure, with the 0.03 μg/μl siRNA dosage, a plateau effect is achieved.

With the prospect of applying gene silencing tools in the field, for the control of mosquito larvae, an important issue to be addressed is the method of delivery of interfering/downregulating molecules. The results using siRNA show that the administration in water, and thus the acquisition of the molecule via oral feeding, work successfully on An. stephensi larvae, as previously shown for other mosquito species [31] as well as for insect species belonging to different orders [55]. Despite this, the possible factors that could affect the downregulation and, therefore, the achievement of a high level of larval mortality, remain to be clarified. The efficiency of RNAi is highly variable in insect species and in different conditions: the critical factors that determine the ability of the target organism to uptake the oligonucleotides, its spread to tissues and cells, the activation of an autonomous RNAi machinery for mRNA degradation should specifically be investigated in the different species [56]. Moreover, the critical factors related to the molecule, e.g. its stability and integrity in the environment and in the larval gut lumen, needed to guarantee the cellular delivery of a sufficient amount of intact siRNA, have not yet been investigated in detail [40]. For example, about host-related factors, recent studies suggested that Diptera lack the transmembrane channel-mediated uptake mechanism, formed by the RNA channel transporters SID-1 and SID-2, that are involved in siRNA uptake in the gut lumen of other insects [55, 57‐60]. Despite this, the results with fluorescent siRNA demonstrate that this type of molecule actually diffuses into the larval mosquito body, possibly through an endocytosis-mediated mechanism, as suggested in other studies on larvae of Aedes aegypti and Anopheles gambiae, that demonstrated the spread of the RNAi effects, after oral administration, to tissues outside the gut [30, 31]. Anyway, the clear downregulation of ABCG4 expression in response to siRNA, observed at both six and 24 h of permethrin treatment (for both concentrations used), indicates that the molecule actually reaches mosquito cells, activating the degradation of target mRNA also in An. stephensi larvae.

Evidence for a systemic RNAi in dipterans has already been obtained in other studies which indicated an efficient internalization and biodistribution of dsRNA in An. gambiae cultured cells [61], and a strong RNAi effect throughout the development of mosquitoes in Aedes aegypti larvae, after feeding [34]. Aspects that are still to be clarified regarding the actual amount of siRNA that reaches larval tissues, after the administration through the breeding water, with the various possible “challenges”, such as pH of the water and gut lumen, and the potentially variable conditions of the gut lumen of larvae in field conditions [40]. The exposure of oligonucleotides to these challenges, as noted in different reports [31, 62], implies that oral feeding-based delivery of interfering RNA molecules determines a less effective knockdown compared to microinjection-based administration. On the other hand, oral administration of siRNA is less time-consuming than microinjection, and more suitable for high-throughput screenings. In addition, it is likely the sole possible way of delivery for field applications for mosquito control.

Recently, to achieve the oral administration of RNAi-inducing molecules to be tested as biopesticides, different approaches have been developed with the aim of delivering intact dsRNAs/siRNAs into cells [63]. The study of abiotic and biotic transport methods was designed to prevent the degradation of RNA molecules, more frequently in the field settings than in lab conditions [64, 65]. Liposomes were tested as abiotic carriers [66, 67], as well as hydrogel- [61], carbon quantum dot- [68] and chitosan-nanoparticles [35, 69]; the last ones, for example, are currently considered as the most economical and environmentally safe system for the delivery of dsRNAs to insect larvae [64, 70]. On the other side, the alternative biotic delivery was achieved through the Escherichia coli and the Pichia pastoris expression systems, for a cheaper large-scale administration of RNAi inducing molecules to third instar larvae of mosquitoes [71, 72].

Among the antisense gene knockdown technologies, antisense Morpholinos (MOs) present characteristics that should guarantee key advantages for future field applications. They can ensure highly specific antisense activity, and are able to bind the target RNA sequence having less interaction with unintended mRNAs, compared with knocking-down RNAi methods based on protein/catalytic activity (e.g. siRNA). Their stability is due to their molecular structure, and to the covalently linked delivery moiety, that make them more durable in water, and more easily internalized by intestinal cells [42]. Moreover, their lack of electrostatic charge (neutral charge) minimizes the interaction with proteins, hence implying a reduced toxicity and immunogenicity; at the same time, being chemically different from a “normal” DNA or RNA molecules, they are resistant to nucleases, which implies a higher stability [39, 42].

The results here obtained showed an increased susceptibility to permethrin after Vivo-MO treatment, congruent with the results obtained using siRNA; the role of the ABCG4 transporter in permethrin detoxification in An. stephensi is thus further supported. The addition of Vivo-MO to the LD₂₀ dose of permethrin led to a significant increase of larval susceptibility, until reaching a 50% mortality at the second time point. In particular, after 24 h, the 0.406 μg/μl Vivo-MO dose determined an increase in larval mortality of 30% compared to permethrin alone, while the 0.203 μg/μl dose, at the same time, determined an increase in mortality of 13% (Fig. 4b). The reached effect appears dose- and time-dependent, with mortality of the two increasing doses shifting from 10 to 23% at 6 h and from 35 to 52% at 24 h. As shown by our results, administration of Vivo-MO in absence of permethrin did not cause any detectable/significant effect in term of mortality and ABG4 gene expression. It could be concluded that the molecule is non toxic if administered alone: its effects were detrimental to mosquito larvae when administered with permethrin. This feature has primary importance to avoid negative effects on non-target organisms: specifically-designed Vivo-MO should have no effects on non-target species.

As for the relative expression of ABCG4 gene (Fig. 5a, b), Vivo-MOs act differently from siRNAs, making mRNA unavailable for translation, but without degrading it: mRNA molecules, bounded and inhibited by the oligonucleotides, likely accumulate inside the cell without being used [38]. It was actually observed that ABCG4 mRNA amount increased after combined permethrin-Vivo-MO treatment (Fig. 4a, b), while siRNA-permethrin treatment lead to a reduction in ABG4 mRNA content in the larvae (Fig. 2a, b). It is possible to suggest that, after the permethrin + Vivo-MO treatment, the larval organism is increasingly stimulated to produce ABCG4 transporter to expel the insecticide, but the mRNA overproduced has no effect, as it has been seen from the mortality results (Fig. 3a, b): the more Vivo-MO molecules bind the target mRNA, making it inactive, the greater the production of mRNA. The expression of ABCG4 gene, as can be seen from the comparison between the two time points, with the two concentrations of Vivo-MO, is stable (expression at six and 24 h is not different), but is dose-dependent, i.e. with the higher dose of Vivo-MO the amount of mRNA detected is higher. It is thus possible that the administration of a higher dose of Vivo-MO in the presence of the same dose of permethrin would lead to an increased larval susceptibility to permethrin. However, the possibility that inhibition of ABCG4 could lead to the activation of other defensome genes responsible for detoxification processes and cellular defence (an issue worth of further investigations [17, 22]), should be considered. Due to the lack of ad-hoc antibody against the ABCG4 of mosquitoes (or other insects), quantification of its expression at the protein level is challenging. Therefore, or goal was to determine the effect of this nucleotide on larval mortality (a sort of phenotypic effect). In future studies, it would also be interesting to design a scramble Vivo-MO control, not complementary to the target gene, in order to assess whether the phenotypic effect is really due to the lack of the transporter or to any other phenomena. For example, in previous studies, it has been shown that permethrin can strongly interact with nucleic acids, intercalating DNA bases (it is prone to bind to G-C base pairs) [73]. One might thus suggest that the effects that were recorded could have partially been determined by this, or similar, effects of Vivo-MOs.

In the wild, mosquito larvae are exposed to the residues of insecticides used in agriculture or in the control of adults, where these residues flow to the breeding sites [7]. These low amounts of insecticides, creating a toxic stress, could induce tolerance by the upregulation of genes involved in the xenobiotic metabolism/detoxification of chemicals [74, 75]. Notably, this low dose insecticide exposition is regarded as one of the major causes for the onset of resistant forms in mosquitoes [76, 77]. Considering the above phenomenon, it was decided to perform Vivo-MO assays using a low permethrin dosage (LD₂₀), which might represent a condition present in breeding sites in treated areas.

Since prior to this study no bioassays had been carried out with Vivo-MO on mosquito larvae, the doses and the timing of administration applied in this study represent a first methodological reference for antisense MO application through direct feeding in mosquito larval stages. It will certainly be interesting to verify in the future if a higher dose of the molecule could lead to a further increase in the susceptibility to permethrin. The possibility to prolong the knocking-down with sequential Vivo-MO treatments against the same ABC transporter should be tested, as well as a multiple knockdown targeting different detoxification genes, in order to avoid that the MO effect declines over time, or to block compensatory effects [41].

In this study, the effects of two different types of mRNA-targeting oligonucleotides, designed to inhibit the expression of the defensome gene ABCG4, and administered through oral delivery in water, were determined on larvae of An. stephensi. The obtained results shown that, targeting ABCG4 gene for silencing through both techniques resulted in an increased pyrethroid efficacy. In conclusion, these results open the way toward the possibility to exploit the inhibition of this gene in the context of integrated programmes for the control of An. stephensi mosquitoes and thus malaria transmission. Of course, prior to field application, several issues should be addressed. First, the potential effects on non-target organisms, starting with in silico studies on the specificity of siRNA or Vivo-MO oligonucleotides, but also addressing their potential “side effects” not related with mRNA targeting. Secondly, the stability and methods of delivery of these molecules in field conditions. Finally, field application would obviously require a cost-effective method for production of the oligonucleotides.