Transgenic Res (2022) 31:489–504 https://doi.org/10.1007/s11248-022-00315-9 ORIGINAL PAPER Larval mosquito management and risk to aquatic ecosystems: A comparative approach including current tactics and gene‑drive Anopheles techniques Robert K. D. Peterson  · Marni G. Rolston Received: 8 February 2022 / Accepted: 13 June 2022 / Published online: 7 July 2022 © The Author(s) 2022, corrected publication 2022 Abstract Genetic engineering of mosquitoes repre- aquatic environments that exceed current tactics for sents a promising tactic for reducing human suffering larval mosquitoes. As such, these new techniques from malaria. Gene-drive techniques being developed would likely comply with currently recommended that suppress or modify populations of Anopheles safety standards. gambiae have the potential to be used with, or even possibly obviate, microbial and synthetic insecti- Keywords Risk assessment · Larvicide · Culicidae · cides. However, these techniques are new and there- Anopheline · Malaria · Mosquito control fore there is attendant concern and uncertainty from regulators, policymakers, and the public about their environmental risks. Therefore, there is a need to Introduction assist decision-makers and public health stewards by assessing the risks associated with these newer mos- For the past 20 years, malaria in much of sub-Saha- quito management tactics so the risks can be com- ran Africa has primarily been managed by indoor pared as a basis for informed decision making. Pre- residual treatments of insecticides, long-lasting insec- viously, the effect of gene-drive mosquitoes on water ticidal bednets, and artemisinin-based combination quality in Africa was identified as a concern by stake- therapy (WHO 2020; Zhou et  al. 2020). Although holders. Here, we use a comparative risk assessment these tactics have been remarkably successful in low- approach for the effect of gene-drive mosquitoes on ering malaria deaths, these gains are threatened by water quality in Africa. We compare the use of exist- resistance, persistence, and resurgence. Consequently, ing larvicides and the proposed genetic techniques the World Health Organization (WHO) has called for in aquatic environments. Based on our analysis, we the research, development, and use of alternative tac- conclude that the tactic of gene-drive Anopheles for tics for malaria management to maintain and improve malaria management is unlikely to result in risks to on the successes in recent years (Derua et  al. 2019; WHO 2020; Zhou et  al. 2020; Antonio-Nkondjio et al. 2021). This article is part of the Topical Collection on “Risk assessment and regulation of gene drive mosquitoes”. Existing and new technologies for mosquito and malaria management pose benefits and risks R. K. D. Peterson (*) · M. G. Rolston  to human health and ecosystems. Genetically engi- Department of Land Resources & Environmental Sciences, neered mosquitoes represent a promising tactic for Montana State University, Bozeman, MT 59717-3120, USA reducing human suffering from malaria. This tech- e-mail: bpeterson@montana.edu nology includes gene-drive approaches that suppress Vol.: (0123456789) 1 4 90 Transgenic Res (2022) 31:489–504 populations of specific mosquito species (often the groups identified human and animal health, biodi- referred to as population suppression strategies), such versity, and water quality as major protection goals. as Anopheles gambiae (sensu lato), the vectors of Consequently––as one example––it is imperative to Plasmodium spp., the pathogen that causes malaria. understand and communicate the risks of mosquito Another approach known as population modification management tactics to aquatic environments and does not reduce mosquito populations, but, rather, it water quality, including risks to people and other non- limits the ability of mosquitoes to transmit Plasmo- target organisms. Therefore, our scope in this paper dium spp. but otherwise does not intentionally affect is to discuss these risks focusing on stressor identi- the mosquitoes (Bier 2021). fication and effects assessment of using gene-drive Currently, research and development of a gene- mosquitoes for malaria management compared to drive system for population suppression using the the existing non-gene-drive larviciding tactics (i.e., tac- doublesex locus (dsxFCRISPRh) has shown promise in tics directed at larval mosquitoes). We define “water experiments with caged, laboratory populations of quality” broadly as that which includes the abiotic An. gambiae (Kyrou et al. 2018; Connolly et al. 2021; and biotic characteristics that determine its suitabil- Hammond et al. 2021). Other forms of gene drive are ity for a particular purpose, including consumption by also being researched, including integral gene drives, people and other animals (USNOAA 2021). daisy-chain gene drives, and toxin-antidote recessive embryo (TARE) drives (Nash et al. 2018; Noble et al. 2019; Champer et al. 2020). Approach and risk characterization The techniques currently being researched that suppress or modify populations of An. gambiae have For the purposes of this paper, we define risk assess- the potential to be used with or even possibly obviate ment as a formalized basis for the objective evalua- microbial and synthetic organic insecticides. How- tion of risk in which assumptions and uncertainties ever, these technologies are new and therefore there are considered and presented (NRC 1983, 1996, is attendant concern from opinion leaders, regulators, 2009; National Academies of Sciences, Engineering, policymakers, and the general public about their envi- and Medicine 2016; WHO 2017). Both human-health ronmental risks (Scudellari 2019; Teem et  al. 2019; and ecological risk can be described in quantitative Connolly et al. 2021). Consequently, there is a press- terms as a function of effect (in many cases “toxic- ing need to assist decision-makers and public health ity”) and exposure (NRC 1983). Risk assessment, stewards by objectively assessing the risks associated therefore, is arguably the most established, robust, with relevant mosquito management tactics so that and science-based method available to estimate risk. the risks can be compared to each other as a basis for Consequently, it is a powerful tool for evidence-based informed decision making (United Nations 2020). societal decision-making. The optimal way to accomplish this is by using Risk assessment typically uses a tiered modeling the science-based framework of risk assessment approach extending from deterministic models (tier (NRC 1983, 1996, 2009), specifically comparative 1) based on conservative assumptions to probabilis- risk assessment. The purpose of comparative risk tic models (tier 4) using refined assumptions (SETAC assessment is to qualitatively and quantitatively com- 1994). Conservative assumptions in lower-tier assess- pare different environmental risks for the purpose ments represent overestimates of effect and exposure; of improved decision-making (e.g., Peterson and therefore, the resulting quantitative risk values typi- Arntzen 2004; Peterson and Shama 2005; Peterson cally are conservative and err on the side of safety. 2006; Peterson et  al. 2006; Davis et  al. 2007; Davis Although terminology may vary, risk assessments and Peterson 2008; Schleier et al. 2008; Schleier and typically follow these steps: (1) problem formula- Peterson 2013; Raybould and Macdonald 2018). tion, (2) analysis phase, and (3) risk characteriza- In workshop exercises associated with the use of tion (NRC 1983, 1996, 2009; SETAC 1994; EFSA gene-drive mosquitoes in Africa for malaria manage- 2010; National Academies of Sciences, Engineering, ment, participants identified general protection goals and Medicine 2016; EFSA et al. 2020). The problem and possible pathways of harm (Roberts et al. 2017; formulation establishes the goals, breadth, and focus Teem et al. 2019; Connolly et al. 2021). In particular, of the assessment, the analysis phase has an effects Vol:. (1234567890) 1 3 Transgenic Res (2022) 31:489–504 491 assessment and an exposure assessment, and the risk (Connolly et al. 2021) and engage in initial compari- characterization is a consideration of the joint prop- sons to currently used larvicidal tactics. erty of effect and exposure to determine risk or what By focusing on problem formulation and effects, additional data are needed to calculate risk or refine we can identify potential primary and secondary risk estimates (USEPA 1998a). The effect assessment effects, which are important concepts in ecological often includes an identification of the stressor and risk assessment. We define a primary effect as the dose–response or density-response relationships. A stressor acting directly on a receptor. The USEPA stressor (also referred to as a hazard) is the entity that (1998a) also terms this a “direct effect”. A second- has the inherent ability to cause harm, whether it be a ary effect is when the direct response on a recep- substance, organism, or activity. tor becomes a stressor to another receptor (usually On first glance, the risk assessment framework another life stage, species, or abiotic entity). The may not seem well aligned with this particular sys- USEPA (1998a) also terms this an “indirect effect.” tem and question because gene-drive mosquitoes for Previous scoping and problem formulation work malaria management are still in research and devel- on gene-drive mosquitoes has identified potential opment stages. Therefore, there is little to no experi- primary and secondary effects (Roberts et  al. 2017; ential information on potential stressors, effects, and Teem et  al. 2019; Connolly et  al. 2021) (Fig.  1). exposure. However, the stepwise nature of risk assess- Obviously, there will always be limited knowledge of ment allows for a logical process whereby risk issues secondary effects posed by a stressor because the pos- can be presented, compared, and considered (Peter- sibilities could represent a virtually uncountable num- son and Arntzen 2004; Wolt et  al. 2010; Raybould ber. However, scientifically reasonable and probable and Macdonald 2018; Raybould et al. 2019; Romeis secondary effects are a much lower and practically et al. 2020). In addition, genetically engineered Aedes manageable number. Regardless, the concept of pri- aegypti mosquitoes have been produced, assessed for mary and secondary effects is important for our pur- risks, approved by regulatory agencies, and commer- poses because we are dealing with stressors that can cially used (Harris et al. 2011; MCTI-CTNBio 2014; be shown to have no or very low inherent toxicity to Carvalho et  al. 2015; USFDA 2016; USEPA 2020a, non-target organisms, including humans. This is espe- b), although the techniques and modes of action are cially germane to gene-drive mosquitoes because not different than what is being developed for gene-drive only will the engineered proteins most likely be inher- An. gambiae. This demonstrates, however, that risk ently non-toxic to non-target organisms, but they will assessment and regulatory approaches are amenable also most likely be produced by the mosquitoes and to genetically engineered mosquitoes. This paper will will be very low concentrations in the environment. explore via a comparative, qualitative risk assess- ment framework the risks of using existing larvicides Comparative risk assessment versus those of gene-drive mosquitoes to aquatic environments. An obvious advantage of comparative risk assess- ment is that we can evaluate if the new tactic (in this case, gene-drive mosquitoes) has the potential to Conduct of the assessment pose increased risk compared to current tactics (in this case, larvicides). Although obvious, this ability The fact that there currently are no gene-drive sys- is underused, but is particularly powerful because it tems for malaria management that are sufficiently allows risk to be evaluated within the context of exist- advanced to be presented to regulatory authorities ing management systems for pests. Comparative risk presents fundamental constraints on the thoroughness assessment is also fundamental as a starting point of risk assessments that can be done. For example, in the safety assessment of genetically engineered the inherent ability of a genetically engineered protein organisms, termed “substantial equivalence” (Codex to cause harm is not yet known for a gene-drive An. Alimentarius Commission 2003). Furthermore, this gambiae. However, the framework is still valuable concept is embedded in the safety standard suggested because we can focus on the problem formulation and by James et al. (2020), which recommends that gene- effect assessment (especially stressor identification) drive mosquitoes should be released in the field only Vol.: (0123456789) 1 3 492 Transgenic Res (2022) 31:489–504 Fig. 1 A conceptual map of stressors, primary effects, and engineered protein is toxic to both the target larvae and non- secondary effects associated with larvicides and genetically target organisms even though all current projects suggest that engineered mosquitoes for malaria management in sub-Saha- the protein will not be toxic ran Africa. *denotes the hypothetical case that the genetically if they “…will do no more harm to human health than This is because of the regulatory distinction between wild-type mosquitoes of the same genetic background effects on individuals and populations. In most cases, and no more harm to the ecosystem than other con- there will be no effects on populations even though ventional vector control interventions.” there might be effects on individuals, but there is some evidence of secondary effects on non-target Larvicides as the comparator populations with repeated use (Hershey et  al. 1998; Lawler 2017; Brühl et al. 2020). Because larvicides are the comparator in this assess- As mentioned above, the effects assessment in ment, some background on this mosquito manage- the analysis phase of a risk assessment identifies if ment tool is warranted. When used according to prod- a stressor has the inherent ability to cause harm. For uct labels, current larvicides will deleteriously affect conventional larvicides, this is a relatively straight- some aquatic non-target organisms (discussed in forward process because the stressor is a known detail below). However, these effects most likely will toxin and the toxic mode of action is well under- not produce unacceptable risks according to current stood and studied as well as the doses necessary to regulatory thresholds (USEPA 1991, 1998b, 2006). causes morbidity and mortality (Fig.  2). However, Vol:. (1234567890) 1 3 Transgenic Res (2022) 31:489–504 493 Fig. 2 Potential primary and secondary effects of larvicides that the active protein is not toxic to non-target organisms. (left) and gene-drive mosquitoes (right) associated with trophic The effects would apply mostly to Anopheles coluzzii and An. levels for aquatic ecosystems. Secondary and tertiary consum- funestus because they are the only species that occupy semi- ers are grouped together because the effects would apply to permanent and permanent water bodies both levels. The “Gene-Drive Mosquitoes” graphic assumes for gene-drive mosquitoes, the transgene may encode at a characterization of risk––as has been the case for proteins that cannot be identified as causing “harm” larvicides. For these new tactics such as gene-drive to any other organism except for the intended effect mosquitoes, the problem formulation, stressor identi- on the target organism. This fact challenges the fication, and effect assessment arguably will be more notion that complete risk assessments are needed, or important (Fig. 2) to the final estimate of risk. can even be done, for some of these products. This The purpose of our paper is to comparatively exam- is because if there is no inherent ability of the pro- ine issues associated with the risks to water quality tein to cause harm (i.e., stressor) to any other organ- from current vector management tactics and from gene- ism, there is conceptually no need (country-specific drive mosquito tactics. Although gene-drive mosquito regulatory requirements notwithstanding) to engage systems for malaria management are still in research in the stepwise risk assessment process in which esti- and development stages with several engineered genes mates of exposure to the stressor are compared to being investigated, it is highly likely that the result- dose–response relationships (Peterson and Arntzen ing proteins will not have conventional insecticidal 2004). Risk assessment traditionally relies on esti- properties. As such, they should not impose risks on mating or using actual environmental exposures to water quality and non-target aquatic organisms that are the stressor and comparing those to effects to arrive greater than current larvicides. Indeed, the risks might Vol.: (0123456789) 1 3 4 94 Transgenic Res (2022) 31:489–504 be appreciably lower (Fig. 2). However, it is important Africa is essential because the ecology of these spe- to stress that in most cases gene-drive mosquitoes will cies is critical to understand when assessing primary be used within an existing Integrated Pest Manage- and especially potential secondary effects. Although ment (IPM) system for malarial mosquito management we focus on the An. gambiae complex for most of (WHO 2017, 2020). Therefore, multiple tactics such as this paper, because of the current status of gene-drive larvicides and gene-drive mosquitoes will be used con- research and development, we also are including the currently and assume that implementation within an An. funestus complex because of its importance in IPM system has as a goal ensuring that risks from all malaria transmission and possible future targeting tactics are acceptable. efforts. A notable difference between current larvicides Four primary malaria vectors belong to two main and gene-drive mosquitoes is that the mosquitoes mosquito complexes in Africa. The An. gambiae (stressor) can multiply in the environment (up to the complex is comprised of nine species (Sinka et  al. point at which total population numbers decline over 2012; Barrón et al. 2019) and the An. funestus com- time, which is the purpose of the population suppres- plex has 13 species (Ogola et al. 2018). Three of the sion gene-drive tactic). This attribute should be part most important vectors occur within the An. gambiae of the risk assessment, but its uniqueness should not complex (or An. gambiae sensu lato (s.l.)): An. gam- be construed as necessitating a separate risk assess- biae sensu stricto (s.s. or S-form), An. arabiensis, and ment. Indeed, and arguably, risk assessments primar- An. coluzzii (M-form). ily should be based on the effects of and exposure to Larvae of An. gambiae and An. arabiensis exploit the product, not the process by which the product was similar habitats. Both species prefer small, sunlit, produced. temporary, vegetation-free habitats, which are com- mon during the rainy season (Githeko et  al. 1996; Choice of larvicides Gimnig et al. 2001; Koenraadt et al. 2004). Although both anopheline species develop quickly in warm Unlike gene-drive mosquitoes, there is a relatively water, a strategy which prevents desiccation in their large amount of data on toxicity and exposure for ephemeral habitats, An. arabiensis is better adapted conventional larvicides. This is because of global to hot, dry conditions (Githeko et al. 1996), develop- regulatory requirements for chemical and biological ing approximately one week faster than An. gambiae pesticides as well as years of commercial use after the (Schneider et  al. 2000). However, the eggs and first pesticides have been registered (WHO 2013, 2017). instars of both species are relatively resistant to desic- In this paper, we discuss the larvicides methoprene, cation (Beier et al. 1990; Koenraadt et al. 2003). Bacillus thuringiensis israelensis, and Lysinibacillus Larvae of these two species adapt quickly to tem- sphaericus (= Bacillus sphaericus) to provide exam- porary, anthropic habitats. During the rainy season, ples of risk issues associated with current products. human-made breeding sites include temporary pools Although there are other larvicides, such as mono- created during construction (Khaemba et  al. 1994), molecular films, pyriproxyfen, spinosad, diflubenzu- borrow pits, drinking water vessels, and tire ruts ron, temephos, and novaluron, we will not evaluate (Gimonneau et  al. 2012; Etang et  al. 2016). During these out of concerns for brevity, because they are the dry season, preferred anthropic habitats include either not currently used for mosquito management in brick-making pits (Carlson et  al. 2004) and perma- Africa, or because they are unlikely to be used in the nent dams (Khaemba et  al. 1994). Other production near future (Choi et al. 2019; Derua et al. 2019). Sim- sites consist of early-season rice fields without well- ilarly, we will not evaluate biological controls, such developed vegetation and wells. as larvivorous fish. Anopheles coluzzii and An. funestus are also pri- mary malaria vectors in Africa, and they exploit very Target species: the Anopheles gambiae and different habitat types than An. gambiae and An. ara- Anopheles funestus species complexes biensis. Larvae of these species are associated with large, permanent, complex, and stable habitats (Etang Knowledge of habitat and food preferences of the two et  al. 2016). They are commonly found in water main species complexes of malaria mosquitoes in bodies dominated by floating plants, overhanging Vol:. (1234567890) 1 3 Transgenic Res (2022) 31:489–504 495 vegetation, and algae and are tolerant of shade (Gim- 1991; Lawler 2017). However, fish are susceptible nig et  al. 2001; Gimonneau et  al. 2012). Preferred to methoprene exposure at relatively high concentra- habitat includes slow-moving water along rivers and tions that exceed application rates for mosquito man- natural ponds (Gimnig et al. 2001), as well as water agement (Brown et al. 1998, 2002; Smith et al. 2003; bodies related to anthropogenic activities such as Hurst et  al. 2007); it is moderately toxic to rainbow mature rice fields. The rate of development for An. trout, Oncorhynchus mykiss, and bluegill sunfish, coluzzii is slower, but this species exhibits strong Lepomis macrochirus. predator-avoidance behavior, an important strategy Methoprene is classified as highly toxic to the because predators are more common in the perma- planktonic crustacean Daphnia magna. It has adverse nent, complex habitats where they occur (Gimonneau effects on freshwater amphipods, Gammarus sp. et al. 2010). (Breaud et al. 1977), lobster (Walker et al. 2005), blue Abundance of An. coluzzii and An. funestus crab, Callinectes sapidus (Horst and Walker 1999), peaks during and immediately after the rainy season fiddler crab (Stueckle et  al. 2008), shrimp (Brown (Gimonneau et  al. 2012), and Kudom (2015) docu- et al. 1998; Wirth et al. 2001; Ghekiere et al. 2007), mented that An. coluzzii larvae can coexist with An. a mayfly species, Callibaetis pacificus, non-biting gambiae in temporary habitats such as footprints midges (Chironomidae), and a dytiscid beetle, Lacco- and tire tracks during this period. However, they philus sp. (Norland and Mulla 1975). are sustained throughout the dry season by breed- In a long-term study on experimental ponds where ing in permanent water bodies with high levels of each site was treated at three-week intervals six times organic material (Kudom 2015). In fact, populations over a season, Hershey et  al. (1998) concluded that of many anophelines increase early in the dry season, methoprene had a negative effect on aquatic insect when larval habitats are more stable and less prone predators at treated sites. These impacts were consid- to flooding (Kweka et al. 2012, 2015). Warm, ephem- ered to be both direct and indirect through food and eral pools tend to have greater exposure to sunlight, interaction webs, as the chemical acted to cause mor- which supports the growth of microorganisms and tality to the predator populations, but also decreased provides an important food resource for foraging lar- the availability of prey. Pinkney et al. (2000) observed vae (Minakawa et al. 1999; WHO 2013; Kweka et al. that methoprene applied to experimental ponds had 2015). no significant impact on non-target arthropods com- pared to control treatments. Larvicides: methoprene In a reasonable worst-case (i.e., tier-1) risk assess- ment, Davis (2007) found that acute and chronic Methoprene is a chemical that mimics the juvenile exposures to methoprene did not exceed USEPA reg- hormone of certain insects. It hinders normal matura- ulatory levels of concern for Daphnia magna, blue- tion of early mosquito instars, and, therefore, larvae gill sunfish, or rainbow trout. In a review focused on that consume methoprene are unable to reach adult- environmental safety, Lawler (2017) concluded that hood (USEPA 1991, 2006). Application timing of the rates of methoprene used for mosquito manage- methoprene is critical; it works best when the insects ment have no detectable effects on the majority of are at earlier developmental stages (Gordon and Bur- freshwater and marine invertebrates evaluated. Fur- ford 1984) because late instars, pupae, and adults are ther, Lawler (2017) stressed the important distinc- not affected. tion between outcomes from laboratory toxicological Methoprene degrades quickly in soil, groundwa- studies (i.e., effects) and field studies and actual envi- ter, exposed water, and vegetation. Half-lives in water ronmental exposures (i.e., risk). range from 30 h in clean water to 60 to 70 h in sew- age. As much as 80% will degrade within 13  days Larvicides: Bacillus thuringiensis israelensis after application (USEPA 1991). The ecotoxicology of methoprene is reviewed thor- Bacillus thuringiensis (Bt) is a soil bacterium. Its oughly by Lawler (2017), and therefore we will only insecticidal property is the result of a crystalline summarize here. Methoprene is practically non-toxic by-product (endotoxin) of sporulation that affects an to terrestrial vertebrates and amphibians (USEPA insect’s microvillar lining when consumed (Mittal Vol.: (0123456789) 1 3 4 96 Transgenic Res (2022) 31:489–504 2003). The insecticide most likely creates an infec- Two formulations of Bti had no effect on non- tion court for secondary infection by other bacteria target invertebrates, including the amphipod Hya- that are common in the insect’s midgut (Broderick lella azteca, in test ponds that had a Bti concentra- et al. 2006) as well as other toxic mechanisms (Cac- tion of 100  mg/L (Gharib and Hilsenhoff 1988). cia et al. 2016). Bt is a highly regarded insecticide Milam et  al. (2000) found that treatments of Bti because its many strains target specific insect spe- were much more damaging to An. quadrimaculatus cies or narrow groups of insects. Consequently, than sentinel species, including Ceriodaphnia dubia, it is well known that Bt endotoxins are practically Daphnia magna, Daphnia pulex, and Pimephales non-toxic to mammals, fish, and birds (Mittal 2003) promelas. In a laboratory assay, Olmo et  al. (2016) and they break down quickly in the environment observed dose–response toxicity in two copepod and (USEPA 1998b). three cladoceran species. Hershey et  al. (1998) con- Bacillus thuringiensis israelensis (Bti) is the strain ducted a large-scale study using 27 experimental of Bt that is used for mosquito management. Bti is ponds in Minnesota, USA. The focus of their study practically non-toxic to mammals, birds, and fish was to determine the impact of multiple aerially (Mittal 2003) and is not persistent (Hajaij et al. 2005), applied direct applications of granular methoprene although it is toxic to some aquatic receptors, includ- and Bti on non-target invertebrates. Bti and metho- ing non-biting midges (Chironomidae). Ali (1981) prene significantly lowered numbers of chironomids, found that applications of Bti to experimental ponds tipulids, ceratopogonids, and brachycerans in treat- significantly lowered numbers of non-target chirono- ment ponds. Disruption of food webs and interaction mids. At the highest treatment rate of 4,000  g/ha, webs was hypothesized to have occurred in many of there was a 54 to 92% reduction in chironomid abun- these reductions because predators seemed to decline dance. In golf-course ponds at a treatment of 3,000 g/ with prey. However, populations rebounded in the ha, there was a 30 to 67% chironomid reduction, but years after the treatments. Niemi et al. (1999) found numbers returned to pre-treatment levels 14  days changes in insect diversity in Bti-treated ponds, and after treatment (Ali 1981). Charbonneau et al. (1994) reduced total insect numbers in ponds treated with found that although Bti caused high mortality of chi- both methoprene and Bti. Lawler et al. (1999) found ronomids in a laboratory, a much lower and statisti- that Bti and methoprene had no measurable impact on cally non-significant mortality was observed in the sentinel amphipods in ephemeral mangrove swamps field. Similarly, Duchet et al. (2015) did not observe on Sanibel Island, Florida, USA when treated with any effects on two chironomid species and Lagadic Bti granules at 5.6 kg/ha and a methoprene liquid for- et  al. (2016) observed no immediate or long-term mulation applied at 10.65 ml AI/ha for the control of effects on chironomid community structure after Aedes taeniorhynchus. Davis and Peterson (2008) did application of Bti. not observe any overall deleterious effects on non- However, a series of recent studies in Europe sug- target arthropods in a field experiment with a single gest repeated use of Bti has secondary deleterious application of Bti. effects on predators (Jakob and Poulin 2016; Poulin Ecological effects have been noted for Bti used and Lefebvre 2018), primarily through reducing chi- for black fly and mosquito management. Merritt ronomid populations. Allgeier et al. (2019) and Brühl et  al. (1989) observed few changes in indices used et al. (2020) observed significant reductions in adult to measure treatment effects of Bti used for black fly chironomid emergence rates after Bti applications in management in a Michigan river. Drift samples taken mesocosm and field studies. In a microcosm experi- at a control and treatment site did not differ for chi- ment, Bordalo et al. (2021) also observed deleterious ronomids, baetids, gammarids, or hydropsychids, but effects on stream benthic invertebrates, including chi- there were some treatment effects on perlid stoneflies ronomids. It is important to note that in many of these and elmid beetles. Similar results were observed in studies, the location evaluated received 30 to 50 aerial 10 stream trials measuring stream insect density of Bti applications per year, an exceptionally high fre- selected taxa (Lawler 2017). Molloy (1992) observed quency of application for Bti. However, WHO (2013) that Bti applied for black fly control within a New has recommendations that include a maximum of 24 York stream affected filter-feeding chironomids, but applications per year. not surface-dwelling or tube-dwelling members of Vol:. (1234567890) 1 3 Transgenic Res (2022) 31:489–504 497 the same family. Caddisflies and mayflies showed no modification. In the following paragraphs, we discuss positive or negative response to Bti treatments. secondary effects that apply to both current and gene- drive approaches. Larvicides: Lysinibacillus sphaericus Immature mosquitoes as food for predators Lysinibacillus sphaericus (= Bacillus sphaericus) is a soil bacterium that has a similar insecticidal action as One secondary effect of population suppression is Bti (Mittal 2003). For L. sphaericus, the insecticidal the potential reduction of beneficial species that feed agent is in the spore cell wall and is a by-product of on the larvae and pupae of An. gambiae (sensu lato) spore production (Mittal 2003). When the agent is (Fig.  1). Many invertebrate species and larvivorous consumed by the mosquito larva, it degrades the lin- fish feed on the aquatic larval and pupal life stages ing of the midgut. The insecticide is more effective of mosquitoes (Service 1977; Ohba et al. 2010; Dida against Anopheles and Culex species than Aedes et  al. 2015). Predatory invertebrates may be respon- species (Mittal 2003), and it remains more active in sible for as much as 90% of the mortality of imma- eutrophic waters than Bti (Lawler 2017). ture mosquitoes in certain aquatic habitats (Service Brown et  al. (2004) found no toxicity to non-tar- 1971, 1973, 1977). In the wetlands of western Kenya, get Australian fauna including the fish Pseudomugil Ohba et al. (2010) found that 54.2% of 330 potential signifier and the shrimp Leander tenuicornis. Merritt predators had ingested immature stages of An. gam- et  al. (2005) observed similar results in a three-year biae, including Odonata larvae (70.2%), Hemiptera study in two habitats in which 138 invertebrate taxa (62.8%), Amphibia (41.7%), and Coleoptera (18%). were exposed to L. sphaericus. Results indicated few However, there is little evidence that aquatic pred- impacts on taxa categorized into functional groups. ators rely solely on immature mosquitoes for survival. Rather, larval and pupal stages of mosquitoes serve Secondary effects: larvicides and gene-drive as one of many food sources for predators. After an mosquitoes extensive literature review of An. gambiae preda- tion in Africa, Collins et al. (2019) suggested that no Although all substances are toxic depending on the predators have been found to be closely associated or dose, it is clear that proteins expressed in a gene-drive dependent on An. gambiae larvae, and that this mos- system to suppress or modify mosquito populations quito complex is probably not an essential part of any for malaria management would not be similar to lar- ecosystem food web. Roberts et al. (2017) concurred, vicidal active ingredients. They would most likely suggesting the loss of An. gambiae from a particular be practically non-toxic to non-target organisms and aquatic habitat is unlikely to cause ecological harm, would challenge the current situation with pesticides even though many invertebrates and fish prey on this that there are deleterious effects other than those species. Likewise, Derua et  al. (2018) found that caused by a reduction in the population of the target long-lasting microbial larvicides (Bti and L. spha- population. Further, as proteins expressed in mosquito ericus), which reduce immature populations of An. larvae, they would almost certainly be expressed at gambiae and An. funestus, have no ecologically sig- environmental concentrations that are orders of mag- nificant impact on the abundance or diversity of non- nitude lower than conventional larvicides (Connolly target invertebrates or vertebrates in the western high- et al. 2021). lands of Kenya. Consequently, the focus in most cases would be Another important consideration for ecological on the secondary effects associated with population risk is that in sub-Saharan Africa two of the three pri- suppression of the target organism (in this case, spe- mary malaria vectors prefer small, ephemeral, sunlit cies in the An. gambiae or An. funestus complex). water bodies that do not support predator populations It is important to note that the goal of both conven- (Carlson et al. 2004; Diabate et al. 2005; Gimonneau tional larvicides and the gene-drive systems discussed et al. 2010, 2012). Aquatic predators typically require here is to lower the population of the pest mosquito more time to develop than mosquito larvae, and there- to reduce malaria (Fig. 1). Indeed, that is the point of fore occur in more permanent habitats (Kindlmann the management tactic unless the focus is population and Dixon 1999; Terhorst et  al. 2010). Therefore, Vol.: (0123456789) 1 3 498 Transgenic Res (2022) 31:489–504 mosquito larvae in ephemeral habitats such as hoof would not be reduced through feeding (Marten 2007; prints or road ruts exhibit higher survival because Connolly et al. 2021). there are fewer predators (Munga et  al. 2006). The Bacteria, protozoa, and other primary producers seasonality of An. gambiae combined with the may serve as secondary food sources for mosquito ephemeral nature of its larval habitats likely results larvae and therefore may be affected by reduced num- in predation that is limited to opportunistic general- bers of larvae. Gimnig et al. (2002) suggested that if ist predators (Collins et  al. 2019), and does not dis- algal resources are depleted, An. gambiae larvae will proportionately and adversely affect any specific non- feed on available bacteria, but bacterial abundance target species. Overall, the current weight of evidence was not significantly affected. Östman et  al. (2008) suggests that a reduction in An. gambiae and closely found that protozoan densities and diversity increased related mosquito larvae most likely would have a neg- dramatically after floodwater mosquito populations ligible impact on predator abundance. Moreover, the were significantly reduced by Bti treatments. species complex does not seem to play a key role in Somewhat related to the food and detritus issue is ecosystems (Collins et al. 2019; Connolly et al. 2021). the secondary effect of numerous dead An. gambiae larvae having a deleterious effect on water quality. To Effects on the food of larval mosquitoes our knowledge, there have been no studies of this for current larvicides. Gene-drive population suppression Another secondary effect of population suppression would reduce the population, resulting in increasingly could be an increase in algal blooms (including toxic fewer larvae and therefore negate specific concerns algal blooms), which might adversely affect wildlife. about water quality due to extensive larval mortality. Algae and other primary producers are important Conversely, with a larvicide, there would be dead lar- larval food sources for anopheline mosquitoes (Con- vae in the water and concentrations of the larvicide nolly et  al. 2021). Kaufman et  al. (2006) suggested each time it is used. that algal biomass on water surfaces is important for larval development of An. gambiae, and Gimnig et al. Effects of engineered proteins and nucleic acids (2002) found that An. gambiae larval grazing reduced algal biomass and abundance in an experiment using Another potential secondary effect is that dead gene- an artificial habitat with rainwater seeded with cow drive mosquito larvae will differentially contaminate dung. The presence of algal mats also serves as an the water compared to non-gene-drive larvae. Based attractant for ovipositing Anopheles females (Bond on the techniques currently being investigated, it is et al. 2005). Both An. gambiae and An. funestus have unlikely that the DNA, RNA, or proteins responsible been positively associated with algae (Minakawa for population suppression in gene-drive mosquitoes et al. 1999; Gimnig et al. 2001), despite their different would negatively affect water quality any more than habitat preferences. However, this association may non-gene-drive mosquitoes. Of course, the engi- also reflect the growth of inedible algal forms, such neered proteins responsible for the desired effect as filamentous green algae, which is indigestible for in the gene-drive mosquitoes would be assessed for most invertebrates (Martin and Kukor 1984). Studies fundamental toxicity and allergenicity as is currently linking reductions in An. gambiae larvae to increases done with transgenic products, with positive toxicity in algal blooms might be irrelevant because habitat or allergenicity likely leading to a regulatory rejection used by this species is temporary and may not support (EFSA 2010; EFSA et al. 2020; Connolly et al. 2021). healthy communities of primary producers (Teem Given the likely impact of the population suppression et al. 2019). However, larvae of An. coluzzii and An. strategies, which would be to reduce the production funestus occur in more complex, permanent habitats of offspring (i.e., larvae), the “contamination” due (Gimnig et  al. 2001; Gimonneau et  al. 2012) and to gene-drive larvae would be less than that of non- might play a greater role in reducing algal blooms. gene drive larvae, or gene-drive larvae from popula- Regardless, a decline in mosquito larvae would not tion modification strategies. However, in none of the affect toxic algal blooms because the cyanobacteria larval types would the effect of the “contamination” that comprise these blooms are toxic to many ani- be any greater than that of non-genetically engineered mals, including mosquito larvae, so cyanobacteria mosquitoes in the environment. Vol:. (1234567890) 1 3 Transgenic Res (2022) 31:489–504 499 Niche replacement (Roberts et al. 2017; Teem et al. 2019; Connolly et al. 2021). We conclude that the tactic of gene-drive An. A substantive reduction of larval An. gambiae popu- gambiae for malaria management is unlikely to result lations could also result in an ecological niche open- in risks to aquatic environments that exceed current ing up for other vector species that transmit malaria larviciding tactics. Although these systems currently or other diseases. Studies have documented mosquito are in research and development stages, it is likely management which reduced populations of anophe- that the resulting proteins will not have insecticidal line mosquitoes in East Africa and resulted in higher properties that are mechanistically similar to current densities of other species, likely because of preferen- larvicides. As such, they should not impose risks on tial elimination of adults and consequently popula- water quality and non-target aquatic organisms that tion reduction (Gillies and Smith 1960; Gillies and are greater than current larvicides. In fact, the risks Furlong 1964; Bayoh et  al. 2010). Anopheles gam- might be lower (Fig.  2). Our conclusions directly biae is the most efficient vector of malaria (Lindsay relate to the important regulatory concept of “sub- et al. 1998), in part because it has a very effective bio- stantial equivalence” (Codex Alimentarius Commis- logical response to competition. It reduces its larval sion 2003). Furthermore, they are consistent with developmental time in the presence of competitors the recommended safety standard of James et  al. without an increase in larval mortality or a reduction (2020), who recommend that gene-drive mosquitoes in body size, but the effect depends on water volume should be released only if they “…will do no more (Paaijmans et al. 2009). This strategy results in higher harm to human health than wild-type mosquitoes of competitive success compared to An. arabiensis or the same genetic background and no more harm to An. coluzzii, which share aquatic habitats with An. the ecosystem than other conventional vector control gambiae but have lower rates of malaria transmission. interventions.” Therefore, any reduction in An. gambiae abundance It is important to reiterate, however, that in most should translate to reduced risk of malaria, since the cases gene-drive mosquitoes will be used within an competitors most likely to replace it are not as effi- existing IPM system. Consequently, IPM tactics such cient vectors. as larvicides and gene-drive mosquitoes will be used The An. gambiae complex is comprised of many concurrently and regulators will need to ensure that morphologically indistinguishable species, which risks from all tactics are acceptable. means hybridization potentially occurs. If gene flow Traditionally, risk assessment relies on estimat- between species includes the gene construct of gene- ing or using actual environmental exposures to the drive mosquitoes, malaria transmission may be fur- stressor and comparing those to effects to arrive at ther reduced, as naïve species in the complex are a quantitative characterization of risk. However, for exposed and eventually genetically modified (Roberts gene-drive mosquitoes, the problem formulation, et al. 2017). Under such conditions, this management stressor identification, and effect assessment may tactic should result in fewer inputs over time, includ- be more important to the final risk estimate (Fig. 2), ing potentially requiring fewer larvicide applications. especially in these early days when there is no body In addition, McArthur et  al. (2014) determined that of experiential use data. gene-drive An. gambiae larvae have the same mortal- Mosquito and malaria management should always ity rate as wild-type larvae, suggesting there should use IPM. This approach is also referred to as Inte- not be an increase in the accumulation of phenotypes grated Mosquito Management (IMM) and Integrated in the environment. Vector Management (IVM) when concerned with mosquito vector management. IPM is a comprehen- sive approach to managing pests that is economically Conclusion and ecologically sustainable (Peterson et  al. 2018). Although using multiple tactics and integrating those Because of workshops with stakeholders that identi- tactics are not an absolute requirement for a success- fied concerns about aquatic environments and water ful, sustainable IPM program, they are commonly a quality, we have used a comparative qualitative feature of IPM. The concept of ecological sustainabil- risk assessment approach for aquatic environments ity includes resistance by the pest to the management Vol.: (0123456789) 1 3 500 Transgenic Res (2022) 31:489–504 tactic, and, therefore, an overall goal of IPM is to following a cluster randomized trial. Sci Rep 11:17101. manage resistance. This is especially salient because https:// doi. org/ 10. 1038/s 41598- 021- 96362-z management tactics such as contemporary synthetic Barrón MG, Paupy C, Rahola N et al (2019) A new species in the major malaria vector complex sheds light on reticu- insecticides, biological insecticides, and gene-drive lated species evolution. Sci Rep 9:14753. https://d oi. org/ approaches obviate long-term issues of broad-spec- 10.1 038/ s41598- 019-4 9065-5 trum toxicity and environmental residuality of pesti- Bayoh MN, Mathias DK, Odiere MR et  al (2010) Anopheles cides. Consequently, because resistance development gambiae: historical population decline associated with regional distribution of insecticide-treated bed nets in by pests is arguably the most significant contempo- western Nyanza Province, Kenya. Malar J 9:62. http:// rary risk with management tactics (Peterson et  al. www. malar iajou rnal.c om/ conte nt/9/ 1/ 62 2018), the discovery and development of new tactics Beier JC, Copeland R, Oyaro C et  al (1990) Anopheles gam- is critical to long-term management success (WHO biae complex egg-stage survival in dry soil from larval development sites in western Kenya. J Am Mosq Control 2020). Provided initial regulatory safety assessments Assoc 6:105–109 and field applications and monitoring are success- Bier E (2021) Gene drives gaining speed. Nat Rev Genet. ful, gene-drive mosquitoes will undoubtedly be an https://d oi. org/ 10. 1038/ s41576- 021-0 0386-0 important tactic within IPM programs for malaria Bond JG, Arredondo-Jimenez JI, Rodriguez MH et al (2005) Oviposition habitat selection for a predator refuge and management. food source in a mosquito. Ecol Entomol 30:255–263 Bordalo MD, Machado AL, Campos D et  al (2021) Acknowledgements We thank B. Dass, A. Faraji, H. Que- Responses of benthic macroinvertebrate communi- mada, and J. Wolt for initial reviews of the manuscript. ties to a Bti-based insecticide in artificial microcosm streams. Environ Pollut 282:117030. https:// doi.o rg/ 10. Funding The Foundation for the National Institutes of 1016/j.e nvpol. 2021. 117030 Health, the Montana Agricultural Experiment Station, and Breaud T, Farlow J, Steelman C, Schilling P (1977) Effects Montana State University supported this work. of insect growth-regulator methoprene on natural popu- lations of aquatic organisms in Louisiana intermediate Open Access This article is licensed under a Creative Com- marsh habitats. Mosq News 37:704–712 mons Attribution 4.0 International License, which permits Broderick NA, Raffa KF, Handelsman J (2006) Midgut bac- use, sharing, adaptation, distribution and reproduction in any teria required for Bacillus thuringiensis insecticidal medium or format, as long as you give appropriate credit to the activity. Proc Natl Acad Sci 103:15196–15199 original author(s) and the source, provide a link to the Crea- Brown MD, Thomas D, Kay B (1998) Acute toxicity of tive Commons licence, and indicate if changes were made. The selected pesticides to the Pacific blue-eye, Pseu- images or other third party material in this article are included domugil signifer (Pisces). J Am Mosq Control Assoc in the article’s Creative Commons licence, unless indicated 14:463–466 otherwise in a credit line to the material. If material is not Brown MD, Carter J, Thomas D, Purdie DM, Kay BH (2002) included in the article’s Creative Commons licence and your Pulse-exposure effects of selected insecticides to juvenile intended use is not permitted by statutory regulation or exceeds Australian crimson-spotted rainbowfish (Melanotaenia the permitted use, you will need to obtain permission directly duboulayi). J Econ Entomol 95:294–298 from the copyright holder. To view a copy of this licence, visit Brown MD, Watson T, Carter J, Purdie DM, Kay BH (2004) http:// creati vecom mons. org/ licen ses/ by/4. 0/. Toxicity of VectoLex (Bacillus sphaericus) products to selected Australian mosquito and nontarget species. J Econ Entomol 97:51–58 Brühl CA, Després L, Frör O, Patil CD, Poulin B, Tetreau G, Allgeier S (2020) Environmental and socioeconomic References effects of mosquito control in Europe using the biocide Bacillus thuringiensis subsp. israelensis (Bti). Sci Total Ali A (1981) Bacillus thuringiensis serovar. israelensis (ABG- Environ. 724:137800 6108) against chironomids and some nontarget aquatic Caccia S, Di Lelio I, La Storia A et al (2016) Midgut microbi- invertebrates. J Invertebr Pathol 38:264–272 ota and host immunocompetence underlie Bacillus thur- Allgeier S, Kästel A, Brühl CA (2019) Adverse effects of mos- ingiensis killing mechanism. Proc Natl Acad Sci USA quito control using Bacillus thuringiensis var. israelen- 113:9486–9491. https://d oi. org/1 0.1 073/p nas. 15217 sis: Reduced chironomid abundances in mesocosm, 41113 semi-field and field studies. Ecotoxicol Environ Saf Carlson JC, Byrd BD, Omlin FX (2004) Field assessments in 169:786–796 western Kenya link malaria vectors to environmentally Antonio-Nkondjio C, Doumbe-Belisse P, Djamouko-Djonkam disturbed habitats during the dry season. BMC Public L et  al (2021) High efficacy of microbial larvicides for Health 4:33. https:// doi.o rg/ 10.1 186/ 1471- 2458-4-3 3 malaria vectors control in the city of Yaounde Cameroon Carvalho DO, McKemey AR, Garziera L et al (2015) Suppres- sion of a field population of Aedes aegypti in Brazil by sustained release of transgenic male mosquitoes. PLoS Vol:. (1234567890) 1 3 Transgenic Res (2022) 31:489–504 501 Negl Trop Dis 9:e0003864. https:// doi.o rg/1 0. 1371/ journ (Skuse) and Tanytarsus curticornis Kieffer (Diptera: al.p ntd.0 0038 64 Chironomidae) in coastal wetlands. Ecotoxicol Environ Champer J, Lee E, Yang E et  al (2020) A toxin-antidote Saf 115:272–278 CRISPR gene drive system for regional population modi- EFSA (2010) Guidance on the environmental risk assessment fication. Nat Commun 11:1082. https:// doi. org/ 10. 1038/ of genetically modified plants. EFSA J 8:1879. https:// s41467- 020-1 4960-3 doi.o rg/ 10. 2903/j. efsa. 2010. 1879 Charbonneau CS, Drobney RD, Rabeni CF (1994) Effects EFSA, Naegeli H, Bresson J et  al (2020) Adequacy and suf- of Bacillus thuringiensis var. israelensis on nontarget ficiency evaluation of existing EFSA guidelines for the benthic organisms in a lentic habitat and factors affect- molecular characterisation, environmental risk assess- ing the efficacy of the larvicide. Environ Toxicol Chem ment and post‐market environmental monitoring of 13:267–279 genetically modified insects containing engineered gene Choi L, Majambere S, Wilson AL (2019) Larviciding to pre- drives. EFS2 18. https:// doi. org/1 0. 2903/j.e fsa. 2020. 6297 vent malaria transmission. Cochrane Database Syst Rev. Etang J, Mbida Mbida A, Ntonga Akono P et al (2016) Anoph- https:// doi.o rg/ 10. 1002/1 46518 58.C D012 736. pub2 eles coluzzii larval habitat and insecticide resistance in Codex Alimentarius Commission (2003) Codex Alimentarius - the island area of Manoka, Cameroon. BMC Infect Dis Guideline for the Conduct of Food Safety Assessment of 16:217. https:// doi. org/ 10. 1186/s 12879-0 16- 1542-y Foods Derived from Recombinant-DNA Plants Gharib A, Hilsenhoff W (1988) Efficacy of 2 formulations of Collins CM, Bonds JAS, Quinlan MM, Mumford JD (2019) Bacillus thuringiensis var israelensis (H-14) against Effects of the removal or reduction in density of the Aedes vexans and safety to non-target macroinverte- malaria mosquito, Anopheles gambiae on interacting brates. J Am Mosq Control Assoc 4:252–255 predators and competitors in local ecosystems. Med Vet Ghekiere A, Fockedey N, Verslycke T, Vincx M, Janssen CR Entomol 33:1–15. https:// doi. org/1 0. 1111/ mve. 12327 (2007) Marsupial development in the mysid Neomysis Connolly JB, Mumford JD, Fuchs S, Turner G, Beech C, North integer (Crustacea: Mysidacea) to evaluate the effects of AR, Burt A (2021) Systematic identification of plausible endocrine-disrupting chemicals. Ecotoxicol Environ Saf pathways to potential harm via problem formulation for 66:9–15 investigational releases of a population suppression gene Gillies M, Furlong M (1964) An investigation into the behav- drive to control the human malaria vector Anopheles iour of Anopheles parensis Gillies at Malindi on the gambiae in West Africa. Malar J 20:170 Kenya coast. Bull Entomol Res 55:1–16 Davis RS, Peterson RKD (2008) Effects of single and multi- Gillies M, Smith A (1960) The effect of a residual house- ple applications of mosquito insecticides on nontarget spraying campaign in East Africa on species balance in arthropods. J Am Mosq Control Assoc 24:270–280 the Anopheles funestus group. The replacement of A. Davis RS, Peterson RK, Macedo PA (2007) An ecological funestus Giles by A. rivulorum Leeson. Bull Entomol risk assessment for insecticides used in adult mosquito Res 51:243–252 management. Integr Environ Assess Manag 3:373–382. Gimnig JE, Ombok M, Kamau L, Hawley WA (2001) Charac- https:// doi. org/1 0. 1002/ ieam. 563003 0308 teristics of larval anopheline (Diptera: Culicidae) bab- Davis RS (2007) An ecological risk assessment for mosquito itats in western Kenya. J Med Entomol 38:282–288 insecticides. Thesis. Montana State University. Bozeman, Gimnig JE, Ombok M, Otieno S et al (2002) Density-depend- Montana ent development of Anopheles gambiae (Diptera: Culici- Derua YA, Kahindi SC, Mosha FW et al (2018) Microbial lar- dae) larvae in artificial habitats. J Med Entomol 39:162– vicides for mosquito control: Impact of long lasting for- 172. https://d oi.o rg/ 10. 1603/ 0022-2 585-3 9.1.1 62 mulations of Bacillus thuringiensis var. israelensis and Gimonneau G, Bouyer J, Morand S et  al (2010) A behav- Bacillus sphaericus on non-target organisms in western ioral mechanism underlying ecological divergence in Kenya highlands. Ecol Evol 8:7563–7573. https:// doi. the malaria mosquito Anopheles gambiae. Behav Ecol org/1 0. 1002/ ece3.4 250 21:1087–1092. https:// doi. org/1 0. 1093/b eheco/ arq114 Derua YA, Kweka EJ, Kisinza WN et al (2019) Bacterial lar- Gimonneau G, Pombi M, Choisy M et al (2012) Larval habitat vicides used for malaria vector control in sub-Saharan segregation between the molecular forms of the mosquito Africa: review of their effectiveness and operational Anopheles gambiae in a rice field area of Burkina Faso, feasibility. Parasites Vectors 12:426. https:// doi.o rg/1 0. West Africa. Med Vet Entomol 26:9–17. https:// doi.o rg/ 1186/ s13071-0 19-3 683-5 10.1 111/j.1 365- 2915. 2011.0 0957.x Diabate A, Dabiré RK, Kim EH et  al (2005) Larval develop- Githeko AK, Service MW, Mbogo CM, Atieli FK (1996) Rest- ment of the molecular forms of Anopheles gambiae (Dip- ing behaviour, ecology and genetics of malaria vectors in tera: Culicidae) in different habitats: a transplantation large scale agricultural areas of Western Kenya. Parassi- experiment. J Med Entomol 42:548–553 tologia 38:481–489 Dida GO, Gelder FB, Anyona DN et  al (2015) Presence and Gordon R, Burford IR (1984) Effects of methoprene, a juvenile distribution of mosquito larvae predators and factors hormone analogue, on the larval and pupal stages of the influencing their abundance along the Mara River, Kenya yellow fever mosquito, Aedes Aegypti. J Insect Physiol and Tanzania. Springerplus 4:136. https://d oi. org/1 0. 30:279–286 1186/s 40064-0 15-0 905-y Hajaij M, Carron A, Deleuze J, Gaven B, Setier-Rio M-L, Vigo Duchet C, Franquet E, Lagadic L, Lagneau C (2015) Effects of G, Thiéry I, Nielsen-LeRoux C, Lagneau C (2005) Low Bacillus thuringiensis israelensis and spinosad on adult persistence of Bacillus thuringiensis serovar israelensis emergence of the non-biting midges Polypedilum nubifer Vol.: (0123456789) 1 3 5 02 Transgenic Res (2022) 31:489–504 spores in four mosquito biotopes of a salt marsh in south- Kenya highlands. Parasites Vectors 8:356. https:// doi. org/ ern France. Microb Ecol 50:475–487 10. 1186/s 13071-0 15-0 968-1 Hammond A, Pollegioni P, Persampieri T et  al (2021) Gene- Kyrou K, Hammond AM, Galizi R et  al (2018) A CRISPR– drive suppression of mosquito populations in large cages Cas9 gene drive targeting doublesex causes complete as a bridge between lab and field. Nat Commun 12:4589. population suppression in caged Anopheles gambiae https:// doi. org/1 0.1 038/ s41467- 021-2 4790-6 mosquitoes. Nat Biotechnol 36:1062–1066. https://d oi. Harris AF, Nimmo D, McKemey AR et  al (2011) Field per- org/ 10. 1038/n bt.4 245 formance of engineered male mosquitoes. Nat Biotechnol Lagadic L, Schäfer RB, Roucaute M, Szöcs E, Chouin S, de 29:1034–1037. https:// doi.o rg/1 0. 1038/n bt.2 019 Maupeou J, Duchet C, Franquet E, Le Hunsec B, Ber- Hershey AE, Lima AR, Niemi GJ, Regal RR (1998) Effects of trand C, Fayolle S, Francés B, Rozier Y, Foussadier R, Bacillus thuringiensis israelensis (Bti) and methoprene Santoni J-B, Lagneau C (2016) No association between on nontarget macroinvertebrates in Minnesota wetlands. the use of Bti for mosquito control and the dynamics of Ecol Appl 8:41–60 non-target aquatic invertebrates in French coastal and Horst MN, Walker AN (1999) Effects of the pesticide metho- continental wetlands. Sci Total Environ 553:486–494 prene on morphogenesis and shell formation in the blue Lawler SP (2017) Environmental safety review of methoprene crab Callinectes sapidus. J Crustac Biol 19:699 and bacterially-derived pesticides commonly used for Hurst TP, Kay BH, Ryan PA, Brown MD (2007) Sublethal sustained mosquito control. Ecotoxicol Environ Saf effects of mosquito larvicides on swimming performance 139:335–343. https://d oi.o rg/1 0.1 016/j.e coenv. 2016.1 2. of larvivorous fish Melanotaenia duboulayi (Atherini- 038 formes: Melanotaeniidae). J Econ Entomol 100:61–65 Lawler SP, Jensen T, Dritz D, Wichterman G (1999) Field effi- Jakob C, Poulin B (2016) Indirect effects of mosquito control cacy and nontarget effects of the mosquito Larvicides using Bti on dragonflies and damselflies (Odonata) in the temephos, methoprene, and Bacillus thuringiensis var. Camargue. Insect Conserv Divers 9:161–169. https:// doi. israelensis in Florida mangrove swamps. J Am Mosq org/1 0.1 111/i cad. 12155 Control Assoc 15:446–452 James SL, Marshall JM, Christophides GK et al (2020) Toward Lindsay SW, Parson L, Thomas CJ (1998) Mapping the ranges the definition of efficacy and safety criteria for advancing and relative abundance of the two principal African gene drive-modified mosquitoes to field testing. Vector- malaria vectors, Anopheles gambiae sensu stricto and Borne Zoonotic Dis 20:237–251. https:// doi.o rg/ 10. 1089/ An. arabiensis, using climate data. Proceedings: Biol Sci vbz. 2019.2 606 265:847–854 Kaufman MG, Wanja E, Maknojia S et al (2006) Importance of Marten GG (2007) Larvicidal algae. AMCA Bulletin No 7 algal biomass to growth and development of Anopheles 23:177–183 gambiae larvae. J Med Entomol 43:669–676 Martin MM, Kukor JJ (1984) Role of mycophagy and bacte- Khaemba BM, Mutani A, Bett MK (1994) Studies of anophe- riophagy in invertebrate nutrition. Current Perspectives line mosquitoes transmitting malaria in a newly devel- in Microbial Ecology: Proceedings of the Third Interna- oped highland urban area: a case study of Moi University tional Symposium on Microbial Ecology, MIchigan State and its environs. East Afr Med J 71:159–164 University, 7-12 August 1983 257–263 Kindlmann P, Dixon AAF (1999) Generation time ratios— McArthur CC, Meredith JM, Eggleston P (2014) Transgenic determinants of prey abundance in insect predator–prey Anopheles gambiae expressing an antimalarial peptide interactions. Biol Control 16:133–138 suffer no significant fitness cost. PLoS ONE 9:e88625. Koenraadt CJM, Paaijmans KP, Githeko AK et al (2003) Egg https://d oi.o rg/ 10. 1371/ journ al. pone. 008862 5 hatching, larval movement and larval survival of the MCTI-CTNBio (2014) Technical Opinion no. 3964/2014: malaria vector Anopheles gambiae in desiccating habi- Request for Opinion on Commercial Release of Geneti- tats. Malar J 2:20. http://w ww.m alari ajour nal. com/ conte cally Modified Microorganism. National Biosafety Tech- nt/2/ 1/ 20 nical Commission, Brasilia, Brasil Koenraadt CJM, Githeko AK, Takken W (2004) The effects of Merritt R, Walker E, Wilzbach M et al (1989) A broad evalu- rainfall and evapotranspiration on the temporal dynamics ation of BTI for black fly (Diptera: Simuliidae) control of Anopheles gambiae s.s. and Anopheles arabiensis in in a Michigan river: efficacy, carry and nontarget effects a Kenyan village. Acta Trop 90:141–153. https:// doi. org/ on invertebrates and fish. J Am Mosq Control Assoc 10. 1016/j. actat ropica.2 003.1 1.0 07 5:397–415 Kudom AA (2015) Larval ecology of Anopheles coluzzii in Merritt RW, Lessard JL, Wessell KJ et al (2005) Lack of effects Cape Coast, Ghana: water quality, nature of habitat and of Bacillus sphaericus (VECTOLEX®) on nontarget implication for larval control. Malar J 14:447. https:// doi. organisms in a mosquito-control program in southeast- org/1 0. 1186/ s12936- 015- 0989-4 ern Wisconsin: a 3-year study. J Am Mosq Control Assoc Kweka EJ, Zhou G, Munga S et  al (2012) Anopheline larval 21:201–212. https:// doi. org/ 10. 2987/8 756-9 71X(2005) habitats seasonality and species distribution: a prereq- 21[201:L OEOBS] 2.0. CO;2 uisite for effective targeted larval habitats control pro- Milam C, Farris J, Wilhide J (2000) Evaluating mosquito con- grammes. PLoS ONE 7:e52084. https:// doi.o rg/ 10. 1371/ trol pesticides for effect on target and nontarget organ- journ al. pone. 005208 4 isms. Arch Environ Contam Toxicol 39:324–328 Kweka EJ, Munga S, Himeidan Y et al (2015) Assessment of Minakawa N, Mutero CM, Githure JI et al (1999) Spatial dis- mosquito larval productivity among different land use tribution and habitat characterization of anopheline types for targeted malaria vector control in the western Vol:. (1234567890) 1 3 Transgenic Res (2022) 31:489–504 503 mosquito larvae in western Kenya. Am J Trop Med Hyg gambiae under semi-field conditions in western Kenya. 61:1010–1016 Acta Trop 109:124–130. https:// doi. org/ 10.1 016/j.a ctat Mittal P (2003) Biolarvicides in vector control: challenges and ropica. 2008.0 7. 010 prospects. J Vector Borne Dis 40:20–32 Peterson RK (2006) Comparing ecological risks of pesticides: Molloy D (1992) Impact of the black fly (Diptera: Simuliidae) the utility of a Risk Quotient ranking approach across contol agent Bacillus thuringiensis va israelensis on chi- refinements of exposure. Pest Manag Sci 62:46–56. ronomids (Diptera: Chironomidae) and other nontarget https:// doi.o rg/ 10. 1002/p s.1 126 insects - results of 10 field trials. J Am Mosq Control Peterson RKD, Arntzen CJ (2004) On risk and plant-based Assoc 8:24–31 biopharmaceuticals. Trends Biotechnol 22:64–66. https:// Munga S, Minakawa N, Zhou G et  al (2006) Association doi. org/ 10. 1016/j. tibte ch. 2003.1 1.0 07 between land cover and habitat productivity of malaria Peterson RKD, Shama LM (2005) A comparative risk assess- vectors in western Kenyan highlands. Am J Trop Med ment of genetically engineered, mutagenic, and con- Hyg 74:69–75 ventional wheat production systems. Transgenic Res Nash A, Urdaneta GM, Beaghton AK et  al (2018) Integral 14:859–875. https://d oi. org/ 10. 1007/ s11248-0 05- 1411-8 gene drives for population replacement. Biol Open Peterson RKD, Macedo PA, Davis RS (2006) A human-health bio.037762. https:// doi.o rg/ 10.1 242/ bio. 037762 risk assessment for West Nile virus and insecticides National Academies of Sciences, Engineering, and Medicine used in mosquito management. Environ Health Perspect (2016) Gene Drives on the Horizon: Advancing Science, 114:366–372. https://d oi.o rg/1 0. 1289/e hp. 8667 Navigating Uncertainty, and Aligning Research with Peterson RKD, Higley LG, Pedigo LP (2018) Whatever hap- Public Values. National Academies Press, Washington, pened to IPM? Am Entomol 64:146–150. https:// doi.o rg/ D.C. 10.1 093/a e/ tmy049 Niemi GJ, Hershey AE, Shannon L et  al (1999) Ecological Pinkney AE, McGowan PC, Murphy DR et  al (2000) Effects effects of mosquito control on zooplankton, insects, and of the mosquito larvicides temephos and methoprene on birds. Environ Toxicol Chem 18:549–559. https:// doi. insect populations in experimental ponds. Environ Toxi- org/1 0. 1002/ etc. 56201 80325 col Chem 19:678–684. https://d oi.o rg/1 0. 1002/e tc.5 6201 Noble C, Min J, Olejarz J et al (2019) Daisy-chain gene drives 90320 for the alteration of local populations. Proc Natl Acad Poulin B, Lefebvre G (2018) Perturbation and delayed recovery Sci USA 116:8275–8282. https:// doi. org/ 10.1 073/ pnas. of the reed invertebrate assemblage in Camargue marshes 171635 8116 sprayed with Bacillus thuringiensis israelensis: Bti Norland RL, Mulla MS (1975) Impact of Altosid on selected impact on reed invertebrates. Insect Science 25:542–548. members of an aquatic ecosystem. Environ Entomol https:// doi. org/ 10.1 111/1 744-7 917. 12416 4:145–152. https://d oi.o rg/ 10.1 093/ ee/4. 1. 145 Raybould A, Macdonald P (2018) Policy-led comparative envi- NRC (1983) Risk assessment in the federal government: man- ronmental risk assessment of genetically modified crops: aging the process. National Academy Press, Washington, testing for increased risk rather than profiling phenotypes D.C. leads to predictable and transparent decision-making. NRC (1996) Understanding risk: information decisions in a Front Bioeng Biotechnol 6:43. https:// doi. org/ 10. 3389/ democractic society. National Academy Press, Washing- fbioe. 2018. 00043 ton, D.C. Raybould A, Holt K, Kimber I (2019) Using problem formu- NRC (2009) Science and decisions: advancing risk assessment. lation to clarify the meaning of weight of evidence and National Academies Press, Washington, D.C. biological relevance in environmental risk assessments Ogola EO, Fillinger U, Ondiba IM et  al (2018) Insights into for genetically modified crops. GM Crops Food 10:63– malaria transmission among Anopheles funestus mosqui- 76. https:// doi.o rg/ 10.1 080/ 21645 698. 2019. 16216 15 toes, Kenya. Parasites Vectors 11:577. https:// doi.o rg/ 10. Roberts A, de Andrade PP, Okumu F et al (2017) Results from 1186/ s13071- 018- 3171-3 the workshop “Problem Formulation for the Use of Gene Ohba S-Y, Kawada H, Dida GO et  al (2010) Predators of Drive in Mosquitoes”. Am J Trop Med Hyg 96:530–533. Anopheles gambiae sensu lato (Diptera: Culicidae) larvae https:// doi.o rg/1 0. 4269/ ajtmh.1 6- 0726 in wetlands, western Kenya: confirmation by polymerase Romeis J, Collatz J, Glandorf DCM, Bonsall MB (2020) The chain reaction method. J Med Entomol 47:783–787 value of existing regulatory frameworks for the environ- Olmo C, Marco A, Armengol X, Ortells R (2016) Effects of mental risk assessment of agricultural pest control using Bacillus thuringiensis var. israelensis on nonstandard gene drives. Environ Sci Policy 108:19–36. https://d oi. microcrustacean species isolated from field zooplankton org/ 10. 1016/j.e nvsci.2 020. 02. 016 communities. Ecotoxicology 25:1730–1738. https:// doi. Schleier JJ, Peterson RKD (2013) A refined aquatic ecological org/ 10.1 007/ s10646- 016- 1708-9 risk assessment for a pyrethroid insecticide used for adult Östman Ö, Lundström JO, Persson Vinnersten TZ (2008) mosquito management. Environ Toxicol Chem 32:948– Effects of mosquito larvae removal with Bacillus thur- 953. https:// doi. org/ 10.1 002/ etc. 2126 ingiensis israelensis (Bti) on natural protozoan commu- Schleier JJ, Davis RS, Shama LM et  al (2008) Equine risk nities. Hydrobiologia 607:231–235. https://d oi. org/ 10. assessment for insecticides used in adult mosquito 1007/ s10750- 008- 9387-z management. Hum Ecol Risk Assess Int J 14:392–407. Paaijmans KP, Huijben S, Githeko AK, Takken W (2009) https://d oi.o rg/ 10.1 080/ 108070 3080 193481 2 Competitive interactions between larvae of the malaria Schneider P, Takken W, McCall PJ (2000) Interspe- mosquitoes Anopheles arabiensis and Anopheles cific competition between sibling species larvae of Vol.: (0123456789) 1 3 5 04 Transgenic Res (2022) 31:489–504 Anopheles arabiensis and An. gambiae. Med Vet Ento- USEPA (2020b) Human Health and Environmental Risk mol 14:165–170 Assessment for the New Product OX5034 Containing Scudellari M (2019) Self-destructing mosquitoes and sterilized the Tetracycline-Repressible Transactivator Protein Vari- rodents: the promise of gene drives. Nature 571:160–162. ant (tTAV-OX5034; New Active Ingredient) Protein, a https:// doi.o rg/ 10. 1038/ d41586- 019- 02087-5 DsRed2 Protein Variant (DsRed2-OX5034; New Inert Service MW (1971) Studies on sampling larval populations Ingredient), and Genetic Material (Vector pOX5034) of the Anopheles gambiae complex. Bull World Health Necessary for Their Production in OX5034 Aedes Organ 45:169–180 aegypti. U.S. Environmental Protection Agency, Wash- Service MW (1973) Mortalities of the larvae of the Anopheles ington, D.C gambiae Giles complex and detection of predators by the USFDA (2016) Finding of No Significant Impact (FONSI) precipitin test. Bull Entomol Res 62:359–369 In Support of a Proposed Field Trial of GE male Aedes Service MW (1977) Ecological and biological studies on Aedes aegypti mosquitoes of the line OX513A In Key Haven, cantans (Meig.) (Diptera: Culicidae) in southern Eng- Monroe County, Florida Under an investigational new land. J Appl Ecol 14:159–196 animal drug exemption. U.S. Food and Drug Administra- SETAC (1994) Final report: Aquatic Risk Assessment and Mit- tion, Washington, D.C igation Dialogue Group. SETAC Press, Pensacola, Fla USNOAA (2021) Water quality describes the condition of the Sinka ME, Bangs MJ, Manguin S et al (2012) A global map of water, including chemical, physical, and biological char- dominant malaria vectors. Parasites Vectors 5:69. https:// acteristics, usually with respect to its suitability for a par- doi.o rg/ 10. 1186/ 1756-3 305-5- 69 ticular purpose such as drinking or swimming. In: U.S. Smith DG, Wilburn C, McCarthy RA (2003) Methoprene pho- National Oceanic & Atmospheric Administration. https:// tolytic compounds disrupt zebrafish development, pro- flori dakeys.n oaa. gov/o cean/ waterq uali ty.h tml. Accessed ducing phenocopies of mutants in the sonic hedgehog 9 Sep 2021 signaling pathway. Mar Biotechnol 5:201–212. https:// Walker A, Bush P, Wilson T et al (2005) Metabolic effects of doi.o rg/ 10. 1007/s 10126-0 02- 0062-5 acute exposure to methoprene in the American lobster, Stueckle TA, Griffin K, Foran CM (2008) No acute toxicity Homarus americanus. J Shellfish Res 24:787–794 to Uca pugnax, the mud fiddler crab, following a 96-h WHO (2017) The evaluation process for vector control prod- exposure to sediment-bound permethrin. Environ Toxicol ucts. World Health Organization, Geneva 23:530–538. https://d oi. org/1 0. 1002/ tox. 20363 WHO (2020) World Malaria Report 2020: 20 Years of Global Teem JL, Ambali A, Glover B et  al (2019) Problem formu- Progress and Challenges. World Health Organization, lation for gene drive mosquitoes designed to reduce Geneva malaria transmission in Africa: results from four regional WHO (2013) Larval source management: a supplementary consultations 2016–2018. Malar J 18:347. https:// doi. org/ measure for malaria vector control: an operational man- 10. 1186/ s12936-0 19- 2978-5 ual. World Health Organization, Geneva Terhorst CP, Miller TE, Levitan DR (2010) Evolution of prey Wirth EF, Lund SA, Fulton MH, Scott GI (2001) Determina- in ecological time reduces the effect size of predators in tion of acute mortality in adults and sublethal embryo experimental microcosms. Ecology 91:629–636 responses of Palaemonetes pugio to endosulfan and United Nations (2020) Report of the ad hoc technical expert methoprene exposure. Aquat Toxicol 53:9–18. https:// group on risk assessment. United Nations Environment doi.o rg/1 0.1 016/S 0166- 445X(00) 00157-0 Programme, Montreal, Canada Wolt JD, Keese P, Raybould A et al (2010) Problem formula- USEPA (1991) Reregistration eligibility document isopropyl tion in the environmental risk assessment for genetically (2E,4E)-11-methoxy-3,7,11-trimethyl-2,4-dodecadien- modified plants. Transgenic Res 19:425–436. https:// doi. oate (referred to as methoprene). U.S. Environmental org/ 10. 1007/s 11248- 009-9 321-9 Protection Agency, Washington, D.C Zhou G, Lo E, Githeko AK et al (2020) Long-lasting microbial USEPA (1998a) Guidelines for ecological risk assessment. larvicides for controlling insecticide resistant and out- U.S. Environmental Protection Agency, Washington, D.C door transmitting vectors: a cost-effective supplement for USEPA (1998b) Bacillus thuringiensis. U.S. Environmental malaria interventions. Infect Dis Poverty 9:162. https:// Protection Agency, Washington, D.C doi.o rg/ 10. 1186/s 40249-0 20-0 0767-3 USEPA (2006) Larvicides for mosquito control. U.S. Environ- mental Protection Agency, Washington, D.C Publisher’s note Springer Nature remains neutral with regard USEPA (2020a) Summary of the Data and Information Related to jurisdictional claims in published maps and institutional to Vectorial Capacity Presented for the New Product affiliations. OX5034 (EPA File Symbol: 93167-EUP-E). Memo from Amanda A. Pierce to Eric W. Bohnenblust. U.S. Environ- mental Protection Agency, Washington, D.C Vol:. (1234567890) 1 3