, Volume 139, Issue 1, pp 41–52

Safe and fit genetically modified insects for pest control: from lab to field applications


  • F. Scolari
    • Department of Animal BiologyUniversity of Pavia
  • P. Siciliano
    • Department of Animal BiologyUniversity of Pavia
  • P. Gabrieli
    • Department of Animal BiologyUniversity of Pavia
  • L. M. Gomulski
    • Department of Animal BiologyUniversity of Pavia
  • A. Bonomi
    • Department of Animal BiologyUniversity of Pavia
  • G. Gasperi
    • Department of Animal BiologyUniversity of Pavia
    • Department of Animal BiologyUniversity of Pavia
SI-Molecular Technologies to Improve SIT

DOI: 10.1007/s10709-010-9483-7

Cite this article as:
Scolari, F., Siciliano, P., Gabrieli, P. et al. Genetica (2011) 139: 41. doi:10.1007/s10709-010-9483-7


Insect transgenesis is continuously being improved to increase the efficacy of population suppression and replacement strategies directed to the control of insect species of economic and sanitary interest. An essential prerequisite for the success of both pest control applications is that the fitness of the transformant individuals is not impaired, so that, once released in the field, they can efficiently compete with or even out-compete their wild-type counterparts for matings in order to reduce the population size, or to spread desirable genes into the target population. Recent research has shown that the production of fit and competitive transformants can now be achieved and that transgenes may not necessarily confer a fitness cost. In this article we review the most recent published results of the fitness assessment of different transgenic insect lines and underline the necessity to fulfill key requirements of ecological safety. Fitness evaluation studies performed in field cages and medium/large-scale rearing will validate the present encouraging laboratory results, giving an indication of the performance of the transgenic insect genotype after release in pest control programmes.


FitnessTransgenesisSITPopulation replacement


In the last decade, it has been widely demonstrated that efficient germ-line transformation for the generation of genetically modified (GM) insects is possible for several dipteran species (Handler and O’Brochta 2005). Such technical advances render feasible the application of transgenesis to implement pest management programmes against both agricultural pests and disease vectors (Benedict and Robinson 2008). Indeed transgenic approaches have been proposed aimed at improving the efficacy of (1) population reduction, mainly relying on the Sterile Insect Technique (SIT), and (2) population replacement strategies against agricultural pests and disease vectors (Scott et al. 2005; Handler 2002). The SIT is a species-specific and environment-friendly insect control approach that relies on the mass rearing, sterilization, and field release of large numbers of insects (Knipling 1955). The competition between released sterile and resident males for mating with wild females leads to the reduction of the reproductive potential and, ultimately, if continued releases of high-quality sterile males in overwhelming numbers over several consecutive generations are performed, a progressive reduction of the population size and eventually the total eradication of the pest population will occur. Currently, the SIT is the most widely applied control method against tephritid fruit flies (Enkerlin 2005; Klassen and Curtis 2005) and it is also being used for the control of the pink bollworm moth Pectinophora gossypiella (Saunders) and the codling moth Cydia pomonella (L.) (Bloem et al. 2005). Successful area-wide integrated pest management (AW-IPM) programmes including the use of SIT have led to the eradication of the New World screwworm Cochliomyia hominivorax (Coquerel) from several areas of Central and South America (Robinson et al. 2009). The eradication in 1997 of the tsetse fly Glossina austeni Newstead in Zanzibar, Tanzania, confirmed the feasibility of using an AW-IPM approach with a SIT component as a tool for creating tsetse-free areas, even if topographical and ecological conditions greatly influenced the success of the campaigns against tsetse (see Vreysen 2006 for a review). In spite of these successes, SIT is not appropriate for every disease vector, since key requirements are (1) the intensive rearing of large numbers of vector insects for mass release, (2) the availability of efficient sex-separation methods, (3) sterilization techniques able to produce large numbers of insects with minimal effects on fitness, (4) effective release methods and (5) efficient marking systems to identify released individuals. In the case of mosquitoes, several trials to control species as Aedes aegypti, Aedes albopictus, Culex pipiens, Culex quinquefasciatus, Anopheles albimanus and Anopheles gambiae were performed, but although many attempts resulted in a reduction in the mosquito population, very few achieved eradication in the release area or long term control (Benedict and Robinson 2003). Transgenic technology may enhance operational SIT programmes at three levels: genetic sexing, sterilization and monitoring (see Scolari et al. 2008a for a review). One example of the potential role of transgenesis in implementing pest control is given by RIDL (Release of insects carrying a dominant lethal; Alphey and Andreasen 2002; Alphey et al. 2002), a variant of the conventional SIT, in which both genetic sexing and ‘sterilization’ are achieved by the same construct. This population reduction method uses a strain of insects homozygous for a dominant lethal genetic system, so that the ‘sterilization’ of the released insects is induced not by irradiation but by homozygosity for a dominant lethal gene (Alphey 2007). Mating with wild individuals results in offspring that are heterozygous for the lethal gene leading to the death of all progeny and hence eventual suppression of the population due to a decrease in its reproductive capacity (Heinrich and Scott 2000; Thomas et al. 2000). Efficient repressible RIDL systems were first demonstrated in Drosophila models and in the Mediterranean fruitfly Ceratitis capitata using the tetracycline-repressible transactivator (tTA) to control expression of a toxic effector (Thomas et al. 2000; Gong et al. 2005). A RIDL system has been developed in Ae. aegypti which has the potential to overcome previous limitations of SIT in this species, providing an effective and safe method of control (Phuc et al. 2007).

On the other hand, population replacement aims at introducing a resistance mechanism to prevent disease transmission (James 2005; Rasgon and Gould 2005). This approach requires both a mechanism for resistance and a method to spread the gene into a population. Mechanisms of resistance have been developed in several mosquito species, such as RNAi to reduce transmission of dengue in Ae. aegypti (Franz et al. 2006), artificial peptides (SM1) to inhibit malaria development in A. stephensi (Ito et al. 2002) and expression of cecropin to impair malaria development in A. gambiae (Kim et al. 2004). Different methods for spreading a gene into a population are currently under investigation, for example Wolbachia symbionts (Rasgon and Scott 2003; Sinkins and Godfray 2004), engineered underdominance (Davis et al. 2001), fitness manipulation (Hahn and Nuzhdin 2004), multiple independently assorting loci (Schliekelman and Gould 2000), meiotic drive systems (Mori et al. 2004; Huang et al. 2007) and transposable elements (TEs; Boete and Koella 2002; O’Brochta et al. 2003; James 2005). Population replacement is currently under development and several aspects still need to be optimized (Benedict and Robinson 2003).

The substantial technological progress achieved since the early hypothesis that pest insects could be controlled through the application of modern genetic engineering technology, has led to an evaluation of the feasibility of proceeding with the initial testing of many transgenic insect lines in the near future. Laboratory-based discovery research is now underway on a variety of genetic engineering strategies in which the basic molecular, genotypic, physiological and behavioral features of transgenic insects are being evaluated and compared with those of their wild-type (wt) counterparts. In this laboratory phase, the biological assessment of transgenic lines includes (1) the determination of the genotype and phenotype of the offspring and the evaluation of their stability in subsequent generations and (2) the estimation of performance characteristics such as viability, life-span, fecundity, fertility, and mating competitiveness. The modification of one or more of these biological features may produce a negative load on the transgenic insect, preventing, in the long term, its exploitation for the improvement of control strategies. Moreover, since the success of a genotype in transmitting its genes to the next generation is influenced by both environmental conditions and its genetic background, laboratory tests need to be supplemented by field cage and medium/large-scale rearing, which represent the second and third phases of transgenic line testing, respectively. Only an integrated use of the data acquired from these three steps will give an indication of the transgenic insect genotype performance after release, permitting the rational design of competitive transgenic insects for use in genetic control programmes (Scott et al. 2005).

Fitness costs of insect transgenesis

Transgenic technologies, typically mediated by transposable elements, may impact fitness at several levels and with a wide range of consequences, as illustrated in Fig. 1. First, since germ-line transformation through microinjection implies the random integration of the transgene into the host genome, it is possible that the transgene may insert in transcriptionally active regions (Thibault et al. 2004). The new insertion may interfere with, or even disrupt, the normal function of one or more genes, compromising the fitness of the target individual due to the negative effects of insertional mutagenesis (Mackay 1989; Rasgon and Gould 2005). Second, the strength of transgene expression can be modified by the chromosomal environment surrounding the insertion site (position effect), possibly due to local chromatin arrangement or close cis-acting regulatory elements, sometimes resulting in transgene silencing or aberrant expression (Venken and Bellen 2005; Williams et al. 2008). Third, fitness reduction can be also attributed to the ‘hitchhiking effect’, that is the fixation of recessive deleterious genes in close proximity to the transgene insertion site (Maynard Smith and Haigh 1974; Marrelli et al. 2006). Finally, foreign protein products of transgenic systems, for example fluorescent transformation markers, may accumulate in large quantities causing toxicity in the compartments where they are expressed and therefore resulting in a fitness load for the GM insect (Liu et al. 1999; Marrelli et al. 2006). However, even if many causes of fitness reduction can be identified, fitness costs are not always clearly understood in terms of their causes and their actual costs. Since their effects may reduce the usefulness of transgenic strains in implementing control strategies, a precise and effective assessment of these issues is mandatory for strains developed to be used in field programmes.
Fig. 1

Potential fitness costs of transposon-mediated transgenesis in dipteran species. Four possible effects are depicted

How can fitness costs affect control strategies?

The assessment of the presence in genetically-engineered strains of any possible transgene-related fitness cost is essential, as negative side-effects of transgenesis may reduce the utility of such strains both at mass rearing and field release levels. Any successful release programme depends upon the ability to mass rear healthy transgenic laboratory colonies, which require standardized maintenance procedures and specialized assays to be validated. Insertional mutagenesis, transgene silencing or aberrant expression and toxicity of the transgene products may impact the viability of a strain. In addition to reduced viability, impaired stability of the strain represents a notable problem at the mass rearing level, especially in the event of breakdown or wt invasion, but also at field release stage. The transposase produced by endogenous transposable elements (TEs) may indeed destabilize integrated heterologous elements and transgenes inserted by TE-mediated germ-line transformation. The potential risk of cross-mobilization or non-homologous recombination resulting in the instability of transgenes integrated into the genome has to be carefully evaluated, since the interaction of TEs can result in vector destabilization (Sundararajan et al. 1999). TE cross-mobilization may greatly impact the stability of such transformation systems and, together with eventual movements of endogenous TEs, induced by the interaction between several factors within the host genome and the vector, might result in mutations or reduced viability of the individuals. Horizontal transfer of TEs has been well documented (Kidwell 1992a, b; Simmons 1992; Robertson and Lampe 1995; Jordan et al. 1999). Because genetic barriers between species are not completely impenetrable to gene flow, the ability of a gene vector system to function also in non-target organisms is of great ecological concern (Atkinson et al. 2001; Handler 2002). Therefore, the potential of transgene remobilization has to be assessed and represents a priority for the risk assessment of using transgenic insects in the field. The environmental concerns evoked by the use of transgenic insects in the field mainly refer to their high mobility, dispersive capacity and ability to produce large numbers of progeny. For example, if a genetically modified insect carries characteristics beyond the intended modification, ecological problems may arise from its colonization potential and its capacity to displace not only conspecifics but also other species, and from the possibility of horizontal gene transfer. Given this scenario, the SIT is a good candidate for minimizing the potential ecological risks related to the release of transgenic organisms. The sterility of the released insects will act as a biological safety mechanism that prevents vertical transmission of the transgenes, which will be eliminated from the ecosystem once the SIT programme is terminated. However, even in such applications, great care should be taken to use as many safety mechanisms as possible to prevent the undesired spread of transgenes. Different methods, especially aimed at generating non-autonomous transposon insertions, have been recently developed and they will be described below. Besides stability and safety, fertility, fecundity and resistance to infections are also fundamental requirements of an efficient and economically-sustainable rearing facility. These attributes are also fundamental in the field, where fitness loads may compromise the performance of released insects reducing their ability to survive, disperse and compete for mates with their wt rivals, thus impairing the efficiency of pest control strategies.

Fitness studies on transgenic insects: the current state of the art

Mediterranean fruit fly

The SIT has proven successful in reducing, controlling and eradicating populations of the medfly, C. capitata, worldwide. The development and implementation of the SIT against this species has been so rapid and effective that it has been validated and practiced on industrial and area-wide scale (Hendrichs 2000; Klassen and Curtis 2005). In addition, the medfly was the first non-drosophilid insect to be transformed (Loukeris et al. 1995) and this achievement opened the way to the genetic transformation of many other pest insects that are targets of SIT programmes (Handler and McCombs 2000; Handler and Harrell 2001; Handler and O’Brochta 2005; Koukidou et al. 2006; Condon et al. 2007a). Different medfly strains are currently available that should increase the efficacy and cost effectiveness of the SIT both at the mass-rearing, releasing and monitoring stages, since they provide (1) genetic marking for the identification of transformed insects (Zwiebel et al. 1995; Handler et al. 1998; Michel et al. 2001); (2) male-specific fluorescent sorting (Scolari et al. 2008b); (3) sexing for male-only strains (Fu et al. 2007; Condon et al. 2007b); (4) reproductive sterility through embryonic lethality (Gong et al. 2005; Schetelig et al. 2009a). A potential candidate for practical improvement of medfly SIT monitoring procedures, after further contained field trial tests, is represented by a recently developed transgenic sperm-marked strain (Scolari et al. 2008b). Different lines carrying two testes-specific markers were generated by fusing the promoter of the spermatogenesis-specific medfly β2-tubulin gene with the reporter genes encoding a red or green fluorescent protein, respectively. In a preliminary set of laboratory experiments, the mating ability of the transgenic homozygous males in comparison to wt was estimated. Twenty-five transgenic males were tested at a 1:1:1 ratio competing with wt males for copulation with wt females. These assays, run in ten replicates for each line, showed that several transgenic lines presented reduced fitness, probably due either to a genetic bottleneck that occurred during their establishment, to the expression of the transgene, or to its particular insertion into the genome. However, the males of one of these lines (#1260_F-3_m-1) showed no negative load in obtaining copulations. This line is particularly promising because its males did not show a reduction also for other fitness components. Indeed the transgenic adults from this line were as longevous as wt and their copulation latency was not significantly different from that of their wild counterparts. They were also able to successfully transfer their fluorescent sperm to females and had great success in siring progeny when competing with wt males. Hence this transgenic marker does not generally cause a fitness reduction, since the genetically modified males were able to transmit their genes to the progeny in a competitive manner. Moreover, line #1260_F-3_m-1 shows both red and green fluorescence in the sperm, possibly due to a position effect on the red fluorescent body marker also present in the construct, indicating that the strong expression of fluorescent proteins is not detrimental to the sperm’s functionality. These results are particularly important for the use of such transgenic flies in SIT monitoring procedures, because, after their sterilization and field release, it will be possible to detect fluorescence in flies captured within the target area. In addition, sperm marking will facilitate the assessment of the mating status of trapped females, which in turn, will allow the mating efficiency of the released sterilized males to be monitored. Another proof that the generation of highly competitive medfly transgenic strains is achievable is represented by the fitness tests performed on males expressing the reproductive sterility system based on transgenic embryonic lethality developed by Schetelig et al. (2009a). Based on previous studies on D. melanogaster (Horn and Wimmer 2003), this approach aims at achieving medfly reproductive sterility by transgene-based embryonic lethality without radiation. It relies on the transmission of a transgene combination that causes embryo-specific lethality in the progeny. To specifically direct the effect of the transgenes to the embryonic stage, newly isolated medfly promoter/enhancer elements of cellularization-specifically-expressed genes drive the expression of the tetracycline-controlled transactivator (tTA). The expressed heterologous transactivator then activates the expression of the lethal effector gene hidAla5 (Bergmann et al. 1998) and leads to embryonic lethality. When transgenic males carrying this system mate to wt females, all progeny die during embryogenesis. Complete embryonic lethality implies no fruit damage from developing larvae and avoids transgene spread into the wild population. This system allowed the development of several lines, one of which (LL #67) showed complete embryonic lethality and high competitiveness with wt in both laboratory and field cage tests. In laboratory assays, wt females were crossed with wt males and transgenic males from this newly generated strain in different ratios (1:1:1, 1:1:3, 1:1:5, and 1:1:9; wt females: wt males: transgenic males). The reduction of progeny compared to wt-only controls showed that the transgenic line is highly competitive and the fertilization success of its males was higher than that of the wt males starting from ratio 1:1:5. For the ratio 1:1:9 a total progeny rate of only 0.4% was calculated, while a wt control at ratio 1:10:0 resulted in little reduction of overall progeny, showing that a transgenic male from this line, in competitive conditions, can perform comparably or even better than a wt male. Moreover, field cage data showed that both non-irradiated and irradiated males from this line were at least as, if not more, competitive as wt non-irradiated males. The high competitiveness of the transgenic lines and the complete embryonic lethality, causing reproductive sterility without the necessity to use irradiation, has a notable potential for improving the efficiency of operational medfly SIT programmes, opening the way to future studies that will compare the fitness costs of transgenic lethality and radiation-based sterility.

The results of these two research projects (Scolari et al. 2008b; Schetelig et al. 2009a) enable us to make some comments and draw important lessons for future studies of fitness of genetically modified insects. First, they illustrate the advantage of creating many independent transgenic lines in order to select only those that are competitive. Second, as recently described, it is possible to use the site-specific integrase system from phage phiC31 to add additional functional transgenes into these fit transgenic medfly lines at the same pre-evaluated genomic position (Schetelig et al. 2009b). In both the sperm-marked and the reproductively sterile medfly described above, donor plasmids containing an attB site, with additional markers and transposon ends were integrated into the attP sites already present in the genome of the transgenic flies by phiC31 integrase-mediated recombination. Then, newly developed medfly transposase-encoding ‘jumpstarter’ strains were crossed to both transgenic strains resulting in the post-integrational excision of transposon ends, leaving stably integrated transgene insertions that cannot be further remobilized. This system will allow the combination of several transgene-encoded advantageous traits at evaluated innocuous genomic sites to generate optimized strains for pest control that minimize ecological risks due to potential transgene remobilization and fulfill the supplementary qualifications regarding transgene stability in mass rearing and release. In conclusion, different transgenic systems developed in the medfly model species are now available and allow the planning of a possible transfer of transgenic technology from the laboratory to the field. They will indeed improve SIT applications by simplifying monitoring and sterilization procedures, thus providing powerful tools for the first ecological insights into the fitness of transgenic insects and their dispersion in the field.

New world screwworm

Cochliomyia hominivorax (Coquerel; Diptera: Calliphoridae) is one of the most important pests of livestock in neotropical regions and the first insect to be controlled using the SIT (Wyss 2000). Successful AW-IPM programmes including the use of the SIT have actually led to the eradication of this species from the southern USA, Mexico, Central America, Panama and some islands in the Caribbean, representing undoubtedly the most successful example of application of the SIT (Robinson et al. 2009). This species was transformed in 2004 (Allen et al. 2004a) and several strains carrying enhanced green fluorescent protein (EGFP) are maintained as stable colonies. They were used to evaluate post-larval fitness costs associated with transformation, including average pupal weight, adult emergence, male ratio, and mating competitiveness (Allen et al. 2004b). Fecundity, fertility, larval productivity, and longevity (Allen and Scholl 2005) indicated an almost general absence of consistent fitness loss in comparison to the wt control. One of these strains, CLAY, with the brightest fluorescent phenotype, which in mating competition and performance tests matched the non-transgenic parental strain, was also successfully cryopreserved, providing the possibility for its use in mass production (Handler et al. 2009). This appears to be particularly promising for the incorporation of this strain into future SIT programmes, but sterilization, through an early expressed conditional lethal system, and consequent fitness tests on the final strain, still need to be performed before its consideration for release.


In the last decades several release trials were conducted with sterile male mosquitoes of different species (reviewed in Benedict and Robinson 2003), but operational difficulties in rearing, imperfect sex-separation, compromised male competitiveness, inefficient marking systems and limited sterility in eggs laid by wild females negatively impacted the effectiveness of the SIT campaigns (Klassen and Curtis 2005). There are currently no large scale SIT programmes in operation against any mosquito species (Benedict and Robinson 2003), although a recent attempt to develop SIT against Ae. albopictus in Italy showed a decrease of about 36% in the number of viable eggs in the release area (Bellini et al. 2007). On this background, the global health and economical problems due to mosquito borne-diseases has stimulated an exceptional effort aimed at generating new molecular tools and a better understanding of the biology and genetics of mosquitoes that culminated in notable technological advances in the genetic manipulation of several species (Catteruccia et al. 2000; Grossman et al. 2001; Perera et al. 2002). These encouraging results have multiplied the possible approaches to accomplish eradication of local vector populations. Particularly, the development of sexing systems for sex separation at early developmental stages to obtain male-only mosquito strains (Catteruccia et al. 2005) may greatly help SIT or its derivative, the RIDL (Alphey et al. 2002; Scott et al. 2002). RIDL technology is progressing rapidly and a strain of Ae. aegypti is now available for field testing. This transgenic strain (LA513A) carries a late-acting tetracycline repressible RIDL system, whose delayed lethal expression allows larval competition to occur (Phuc et al. 2007). The LA513 RIDL system resulted in death at the larval/pupal boundary with penetrance of lethality up to 97%, encouraging the testing of this line for mosquito control (Yakob et al. 2008). Moreover, the RIDL technique has the potential to be applicable to a wide range of mosquito vectors, and it has also been developed in Ae. albopictus, an important vector of dengue and chikungunya (Wilke et al. 2009). A recent improvement of the RIDL system showed the possibility to engineer late-acting, repressible, tissue-specific, and female-specific transgene expression to produce a flightless phenotype in female Ae. aegypti (Fu et al. 2010). In these newly developed strains, the promoter derived from the Ae. aegyptiActin-4 gene leads to the expression of tTA in a stage-, tissue-, and sex-specific manner, combining late-acting lethality with effective female sterility. This effective sterility is due to the inability of the flightless females to produce their distinctive wing-beat frequencies that act as sexual cues to males. The female incapacitation due to the flightless phenotype may be considered as an equivalent of lethality for a RIDL strain. The development of this technology has all the advantages of the classical RIDL, including late-acting lethality, but permits any life stage to be released and has therefore the potential to facilitate the control of Ae. aegypti and other mosquito species.

The development of gene transfer technology for a series of vector species also made it possible to design gene drive systems to spread genes that can block the transmission of pathogens (Abraham et al. 2005; Ito et al. 2002; Moreira et al. 2002). To this end, the use of transposable elements, Wolbachia, meiotic drive genes, and homing endonuclease genes have been proposed (Sinkins and Gould 2006; Windbichler et al. 2007). The next step will consist in the large scale field release of GM mosquitoes carrying a desired trait such as malaria refractoriness and the ability to spread it to a large fraction of the wild-type vector population (James 2005). However, before each of these control measures can be transferred into the field for practical applications, it is essential to assess the fitness and population ecology of genetically modified mosquitoes (Marrelli et al. 2006).

The first fitness study on genetically modified mosquitoes evaluated the persistence of transgenic Anopheles stephensi (Liston; Diptera: Culicidae) over several generations (Catteruccia et al. 2003). Homozygous mosquitoes from these lines, expressing fluorescent protein markers from a ubiquitous actin promoter, were established in a 1:1 ratio in the same cage with wt mosquitoes to compare their mating competitiveness. The frequency of the transgene was followed over subsequent generations. While longevity, feeding rate, egg-laying rate and larval developmental time compared favorably with wt, there was a sharp decline in the frequency and eventual loss of the transgene allele over eight or fewer generations for all the tested lines. Apart from the effect of gene disruption in one of these lines due to transgene insertion into a coding region, this low fitness could be attributed to the fixation of deleterious alleles during inbreeding to establish the homozygous lines (hitchhiking effect).

In another research project, the reproductive and developmental fitness of transgenic Ae. aegypti lines containing either the Hermes or MOS1 transposable elements were examined (Irvin et al. 2004). Life-table characteristics of transgenic Ae. aegypti were compared to non-transgenic mosquitoes and the results showed that fitness costs in all three tested transgenic lines were severe for every considered demographic parameter. Across all gonotrophic cycles, each transgenic strain exhibited significantly reduced survivorship for all life stages, and mortality was greatest for the transition from egg to larvae. Adult longevity and fecundity were also lower. Even if there was not an equally negative load on all fitness components across the three transgenic strains, the final outcome of the impact of transgenesis on each fitness measurement was detected when demographic parameters of each strain were determined. As in the previously described research project (Catteruccia et al. 2003), fitness reduction may be due to the maintenance of the transgenic mosquitoes as homozygotes, a condition that could be associated not only with inbreeding depression but also with founder effects, indicating that fitness loads may derive from causes other than the expression of the transgene itself (Li et al. 2008). Additional negative effects could be generated by the activity of a Hermes transposase which could have induced insertion of this element by cut-and-paste transposition into multiple somatic nuclei during development.

Life-table experiments and cage competition were employed to assess the fitness of transgenic A. stephensi carrying two transgenes that prevent the transmission of the rodent malaria parasite Plasmodium berghei (Moreira et al. 2004). All the developed lines carry the same marker and the same promoter driving the expression of two different anti-parasitic effector proteins (SM1 and PLA2, respectively). The SM1 lines showed no evident fitness load in comparison to the non-transgenic mosquito controls, while the PLA2 lines presented a remarkable fitness reduction. Unlike the previous two studies (Catteruccia et al. 2003; Irvin et al. 2004), in this case the lines were maintained as heterozygotes by repeatedly backcrossing transgenic mosquitoes with wt mosquitoes from laboratory population cages. In this way, they maintained the genetic background of the transgenic lines similar to that of the controls, allowing a direct correlation between fitness load and transgene presence, also avoiding hitchhiking of any deleterious gene present near the transgene insertion site. Moreover, marker and promoter expression was limited to specific cell types, minimizing negative effects due to foreign protein accumulation in many tissues. The authors could therefore conclude that transgenesis itself does not necessarily confer a fitness cost. Subsequently, the hypothesis that, when fed on Plasmodium-infected blood, transgenic mosquitoes expressing SM1 would be more fit than wt because of their reduced parasite prevalence was tested in cage-invasion experiments using a starting population of equal numbers of wt and SM1-transgenic A. stephensi mosquitoes, maintained on P. berghei-infected blood (Marrelli et al. 2007). It was shown that, when fed on Plasmodium infected blood, the transgenic malaria-resistant mosquitoes had a significant fitness advantage over wt in terms of higher fecundity and lower mortality. This suggests that the fitness advantage of SM1-transgenic mosquitoes fed on infectious P. berghei may be due to the blockage of ookinete invasion of the mosquito midgut, an hypothesis that is in contrast to other studies (Hurd et al. 2005; Ahmed and Hurd 2006). The difference is probably due to different approaches to render the mosquitoes Plasmodium-refractory. In the study by Hurd et al. (2005), Plasmodium-refractiveness was due to the activation of the endogenous innate immune cascade which renders refractory mosquitoes no more fit than their non-refractory counterparts; this hyperimmune response may imply fitness costs that countered the advantage of not being infected. Conversely, the expression of the SM1 gene is apparently harmless and, moreover, SM1 inhibits development at a very precocious stage of Plasmodium development in the mosquito, preventing midgut invasion and activation of the mosquito immune system, indicating that the mode and timing of transgene expression have a crucial importance (Marrelli et al. 2007). However, since the conclusions drawn from Marrelli et al. (2007) come from a population genetics modeling approach that investigated the dynamics of transgene spread in mosquito cage populations, these results need to be supported by tests on natural mosquito-Plasmodium combinations that better resemble field conditions (Boëte 2005). Additionally, heterozygous SM1-transgenic mosquitoes have the highest fitness but the related fitness reduction in homozygotes prevents transgene fixation, which would be optimal for more practical applications for disease control in the field (Lambrechts et al. 2007). A subsequent research project identified a fitness load in three independent homozygous SM1-transgenic mosquitoes where the effector protein expression was under the control of the blood-inducible vitellogenin promoter (Li et al. 2008). The authors attributed such disadvantages to a lower mating success of homozygous transgenic males which competed less effectively for females than wt males, therefore generating fewer progeny; transgenic larval development was slower and in some lines fecundity was also lower. This phenomenon may be the result of a hitchhiking effect or insertional mutagenesis and not due to a negative effect of the transgenic protein, as although homozygous transgenic mosquitoes were replaced by wt, transgenic heterozygous mosquitoes persisted with wt.

Rodrigues et al. (2008) recently developed four transgenic lines of Ae. fluviatilis expressing the mutated phospholipase A2 enzyme (PLA2m) which was shown to significantly inhibit P. gallinaceum oocyst development. Their preliminary experiments to assess the stability of the transgene revealed the absence of fitness costs, especially in relation to fertility and survivorship. Further studies aimed at evaluating other fitness-related traits such as the body size and the expression of major enzyme classes involved in the metabolism of xenobiotics showed that (1) both male and female transgenic individuals were larger than wt controls, suggesting that this trait may have an impact on the overall fitness, and (2) no significant difference in the activity of enzymes related to metabolic insecticide resistance was detected in transgenic mosquitoes (Santos et al. 2010). Studies like this are of particular importance to avoid unsuccessful implementation of control strategies in the field and contribute to emphasize the necessity to develop many transgenic lines in order to choose the best homozygous line for: (1) stronger effector gene expression; (2) easier mass rearing; (3) more efficient introgression of transgenes in the field; (4) direct assessment of eventual hitchhiking effect (Li et al. 2008). In addition, to distinguish between fitness costs associated with the effector molecule or with insertional mutagenesis, site-specific integration may be very useful also for mosquito species. In a recent paper (Amenya et al. 2010), hemizygous mosquitoes from different transgenic lines of An. stephensi generated by injection of piggyBac-based constructs carrying docking sites for integration by the phiC31 system were analyzed for several fitness parameters, such as fecundity, fertility, larval-to-pupal developmental time, larval viability, pupal sex-ratio, adult wing-span and adult longevity. The results indicated that transgenesis generally implies negative effects on adult survivorship, but each transgenic line showed peculiar variations in life-table parameters when compared to the wt control. Further experiments will allow the stable and efficient integration of additional transgenes into attP docking sites whose location and impact on fitness are already known, mitigating the variation in expression induced by random integration.

How to assess fitness parameters

To limit the negative side-effects of transgenic manipulation on target insects and on the environment to an acceptable level, laboratory and field experiments aimed at determining fitness parameters, mating competitiveness, and transgene stability need to be performed. Laboratory experiments may be divided in two broad categories: (1) life-table and (2) competition experiments. Life- and fertility-table experiments are powerful tools for analyzing and understanding the impact that an external factor (in our case represented by transgenesis) has upon the growth, survival, reproduction, and rate of increase of an insect population (Landahl and Root 1969; Bellows et al. 1992). They include the estimation of parameters such as reproductive rate, survival and development time of immature stages, mating success, blood-feeding success, adult survival. Competition experiments need to be articulated in three phases. In the first phase, cage competition in laboratory has to be performed, in which the strains carrying the transgene/s are placed in equal frequencies with individuals from the target population; in this way, mating behaviour studies may be conducted aimed at identifying the age when a male or female becomes sexually active or receptive, the capacity to locate mates, competition for mates, mate choice and sperm use mechanisms. In the second phase, the cage competition experiments have to be performed at the release site, using wild insects collected directly from the field. This represents the first step towards the local environment where the transformed insect will be able to compete with wt individuals. The third phase comprises the release of the genetically modified insects into large replicate outdoor enclosures, where competition experiments between them and individuals collected directly from the field will occur. These intermediary testing grounds between the laboratory and the field within the target area may be defined as semi-field systems, which are enclosed environments located within the natural ecosystem of the target pest and exposed to ambient conditions, where all features necessary for the pest lifecycle completion are present (Knols et al. 2002). For example, in the case of transgenic mosquitoes, semi-field systems involve large outdoor cages delimited by nets, and with aquatic larval habitats, blood hosts for adult females, sugar sources (plants) for adult individuals, resting sites, and environmental features (e.g. swarm markers to stimulate mating). The results obtained from these different assays have to be monitored over multiple generations, compared to the control cages and interpreted in the light of the genotype of the transgenic insects and the ecological characteristics of the environment (Scott et al. 2005).

Conclusions and perspectives

In this review we describe the critical steps of the process through which transgenesis may help in implementing pest management programmes, stressing the importance of producing fit and competitive transformed insects, as summarized in Fig. 2. Several observations can be derived from the previous studies on fitness assessment of transgenic insects. First, the generation of several transgenic lines is to be preferred, since insertional mutagenesis caused by transposon-mediated germline transformation may randomly impact the overall fitness of the individual. It is therefore essential to generate multiple transgenic lines, compare them and select the fittest. The availability of numerous independently obtained lines of the same construct is also fundamental because the strength of transgene expression can be modified by chromatin surrounding the insertion site. To avoid this side-effect of germ-line transformation, the use of insulator elements for modulating position effects could be considered (Markstein et al. 2008). As it has been repeatedly observed that inbreeding can have a strong impact on fitness, it is particularly promising that homozygous transgenic strains can have a performance similar to that of wt, especially as the mass rearing of homozygous lines is often required. The development of transgenic homozygous lines will therefore (1) simplify the rearing procedures, reducing the costs of strain maintenance, (2) enable the direct assessment of eventual hitchhiking effects, and (3) permit a strong expression. The availability of phiC31-mediated site-specific recombination systems for medfly (Schetelig et al. 2009b) and Ae. aegypti (Nimmo et al. 2006) represents a notable step forward in transgene manipulation technology. The possibility to generate stabilized lines thanks to these post-integration transgene stabilization approaches, combined with the possibility of adding further functional transgenes at innocuous genomic positions, will permit the development of many different strains without the need for further germ-line transformation with its associated potential fitness costs, multiplying the experimental opportunities. Fitness tests aimed at determining parameters such as survival rate, dispersive ability, mating competitiveness, sperm transfer and sperm functionality, motility and mobility of promising stabilized transgenic lines will highlight the fittest, which will become ideal targets for further applications. A few examples may include (1) the set up of novel studies on functional genetics and genomics, (2) the achievement of easier mass-rearing and improved line stability, (3) the optimization of transgenic insect performance in field for the improvement of pest control strategies, both in terms of population suppression and replacement. Finally, future evaluation of GM insect fitness will require routine semi-field assays, followed by release of GM insects in isolated areas that have been carefully characterized with respect to the genetic and ecological structure of local insect populations, combined with disease transmission data. Only if GM insects fulfill key requirements of fitness and ecological safety, assessed by actual field use, will they be useful for improving the existing control strategies.
Fig. 2

Schematic representation of the process for generating fit and safe homozygous transgenic lines for practical use in pest control strategies


The authors thank Gerald Franz and Jorge Hendrichs for support and advice and Ernst A. Wimmer for input and suggestions.

Copyright information

© Springer Science+Business Media B.V. 2010