Use of transgenic Aedes aegypti in Brazil: risk perception and assessment

Abstract The OX513A strain of Aedes aegypti, which was developed by the British company Oxitec, expresses a self-limiting transgene that prevents larvae from developing to adulthood. In April 2014, the Brazilian National Technical Commission on Biosafety completed a risk assessment of OX513A and concluded that the strain did not present new biological risks to humans or the environment and could be released in Brazil. At that point, Brazil became the first country to approve the unconstrained release of a genetically modified mosquito. During the assessment, the commission produced a comprehensive list of – and systematically analysed – the perceived hazards. Such hazards included the potential survival to adulthood of immature stages carrying the transgene – should the transgene fail to be expressed or be turned off by exposure to sufficient environmental tetracycline. Other perceived hazards included the potential allergenicity and/or toxicity of the proteins expressed by the gene, the potential for gene flow or increased transmission of human pathogens and the occupation of vacant breeding sites by other vector species. The Zika epidemic both elevated the perceived importance of Ae. aegypti as a vector – among policy-makers and regulators as well as the general public – and increased concerns over the release of males of the OX513A strain. We have therefore reassessed the potential hazards. We found that release of the transgenic mosquitoes would still be both safe and of great potential value in the control of diseases spread by Ae. aegypti, such as chikungunya, dengue and Zika.


Introduction
In April 2014, Brazil's National Technical Commission on Biosafety -the agency officially responsible for the assessment of the risks posed by genetically modified organisms in Brazil -assessed the potential risks of the release in Brazil of the transgenic OX513A strain of Aedes aegypti and concluded that such a release would be safe. 1 At that point, Brazil became the first country to approve the unconstrained release of a genetically modified mosquito. Two years later, however, the Zika epidemic had added to the general public's concerns over the release of mosquitoes and we therefore decided to re-investigate the perceived risks and update the commission's risk assessment.

Control of arbovirus vectors
Several arboviruses -for example chikungunya virus, dengue viruses and, more recently, Zika virus -cause much human suffering in Brazil. 2,3 In 2015, for example, there were more than 1.5 million suspected cases of dengue fever in the country. 2 While attempts to create effective or cost-effective vaccines continue, control of the mosquito vectors remains of the utmost importance. 4 Brazil has had some success in controlling mosquitoes, even early in the 20th century when no effective insecticides were available. 5 Mosquito control, by residents, health workers and other municipal workers supported by a heterogeneous and broad set of collaborators, remains the focus of the National Dengue Control Programme. 6 However, the failure of this programme to reduce vector populations to levels that could interrupt dengue transmission 7 has spurred Brazilian interest in dengue vaccines 8 and novel approaches to vector control such as the sterile insect technique. [9][10][11][12][13] The International Atomic Energy Agency describes the sterile insect technique as "a type of birth control in which wild female insects of the pest population do not reproduce when they are inseminated by released, radiation-sterilized males". 14 Although sequential releases of large numbers of the sterilized males should lead to a reduction in the size of the pest population, the irradiation used can reduce the released insects' competitiveness -and this appears to be a particular problem when the insects involved are mosquitoes. 15 Over the last decade, as an alternative to the sterile male technique, the genetic modification of mosquitoes, with the production of large numbers of males and females that carry a self-limiting transgene, has been investigated. 16 Once released in the environment, the male insects carrying the transgene -which have to be produced in the presence of a selective agent that blocks the transgene's expression -should, potentially, compete on equal terms with the wild males. [17][18][19][20] The OX513A strain of Ae. aegypti, which was developed by the British company Oxitec, expresses a self-limiting Abstract The OX513A strain of Aedes aegypti, which was developed by the British company Oxitec, expresses a self-limiting transgene that prevents larvae from developing to adulthood. In April 2014, the Brazilian National Technical Commission on Biosafety completed a risk assessment of OX513A and concluded that the strain did not present new biological risks to humans or the environment and could be released in Brazil. At that point, Brazil became the first country to approve the unconstrained release of a genetically modified mosquito. During the assessment, the commission produced a comprehensive list of -and systematically analysed -the perceived hazards. Such hazards included the potential survival to adulthood of immature stages carrying the transgene -should the transgene fail to be expressed or be turned off by exposure to sufficient environmental tetracycline. Other perceived hazards included the potential allergenicity and/or toxicity of the proteins expressed by the gene, the potential for gene flow or increased transmission of human pathogens and the occupation of vacant breeding sites by other vector species. The Zika epidemic both elevated the perceived importance of Ae. aegypti as a vector -among policy-makers and regulators as well as the general public -and increased concerns over the release of males of the OX513A strain. We have therefore reassessed the potential hazards. We found that release of the transgenic mosquitoes would still be both safe and of great potential value in the control of diseases spread by Ae. aegypti, such as chikungunya, dengue and Zika.
Policy & practice Safety of transgenic Ae. aegypti in Brazil Paulo Paes de Andrade et al.
dominant transgene that is able to kill, at larval stage, the mosquitoes in which it is expressed (Fig. 1). 21 In the presence of tetracycline, the transgene is either not expressed or is only expressed at a very low and non-lethal level. Once released in the environment, most males carrying the transgene die after about two days. 21

Risk perceptions and assessments
In almost all countries, any living modified organism derived from modern biotechnology is very strictly regulated and subjected to a long and detailed risk assessment. 22 In Brazil, which is a Party to the Cartagena Protocol on Biosafety to the Convention on Biological Diversity, such risk assessment is the sole responsibility of National Technical Commission on Biosafety. 23 The commission consists of 27 members and an equal number of surrogates -chosen by various federal government agencies and civil organizationsincluding highly qualified scientists from across the country. As stipulated in the Cartagena Protocol, any risk assessment of a genetically modified product has to be supported with so-called hard data and preferably with hard data collected in the country that intends to use the product. In 2011, the commission was asked to conduct a risk assessment of the biosafety of the OX513A strain of Ae. aegypti as a precursor to the experimental, small-scale release of adult males of the strain in Brazil's Bahia state -first in Juazeiro and then in Jacobina-that would take place from 2012 to 2014. The commission used data collected in these releases and in field experiments conducted in the Cayman Islands and Malaysia, along with some relevant laboratory results, to produce a comprehensive risk assessment. 1 In April 2014, on the basis of this assessment, the commission approved the OX513A strain for unconstrained release throughout Brazil, primarily as a method of dengue control.
Risk assessments by the commission, like those by most agencies tasked with assessing the risks posed by the release of a genetically modified organism, consist of four main steps, which comply with the Cartagena Protocol: (i) problem formulation; (ii) risk characterization; (iii) risk classification; and (iv) an overall safety evaluation. 24,25 In problem formulation, specialists outline the release scenario and list all the associated hazards -as perceived by themselves, by other specialists and by the general public. Each perceived hazard is then assessed in the context of the planned release, using any relevant data that are available and considering each plausible route by which the hazard may cause harm. In this risk characterization, any perceived hazards that cannot plausibly lead to harm are excluded. In the subsequent risk classification, the risk of each of the remaining perceived hazards occurring is estimated. Finally, the assessors estimate the overall risk posed by the planned release and consider whether the organism should be considered safe for the environment.
Although international guidelines on the assessment of food safety 26 do not apply directly to mosquitoes, they can still be useful in assessing the allergenicity and toxicity of novel proteins created as the result of the genetic modification of such insects.

assessment
As part of its risk assessments of the OX513A strain in 2014, the Technical Commission listed the potential hazards. Although some of the listed hazards were perceived by specialists, most had been voiced by the general public. Among the main perceived hazards were the unexpected survival of at least some of the mosquitoes that carried the transgene and/or their progeny as the result of partial penetrance and/or the presence of tetracycline in the environment, allergenicity and/or toxicity of two new proteins expressed, possibly in the mosquitoes' saliva, as a result of the genetic modification, the vertical and horizontal flow of the transgene and its consequences, the participation of transgenic mosquitoes in the transmission of dengue viruses and the occupation, by other vector species, of breeding sites made vacant by the intervention. 1 Only one of the listed perceived hazards -that is, survival in the presence of tetracycline -was considered plausible, and even that hazard was deemed unlikely since concentrations of tetracycline found in freshwater are usually more than two orders of magnitude lower than that needed to block the transgene's expression in laboratory-bred insects. 1 In addition, only 50% of the progeny of any female mosquitoes carrying the transgene into adulthood would carry the lethal gene -and they all die at the larval stage if not also protected by exceptionally high concentrations of tetracycline.
Although the occupation of breeding sites cleared of Ae. aegypti by other vector species -for example Ae. albopictus -was deemed a negligible risk by the entomologists involved in the risk assessment, it remained an area of concern among a few of the commission's non-specialists.
The commission approved the nationwide release of adult males of the OX513A strain on the condition that Oxitec investigated tetracycline concentrations in potential breeding sites by means of periodic literature reviews and also investigated the occupation of cleared breeding sites by Ae. albopictus. The plans for the commercial releases of males of the OX513A strain throughout Brazil have not been changed since the 2014 risk assessment. It is expected that the releases will take place mainly in cities or other densely populated areas, in association with other control measures and under continuous surveillance by the relevant municipal health authorities and Oxitec. At the time of writing, a large-scale but still pre-commercial release is in progress in Piracicaba county in São Paulo state. The data already collected during this release (Fig. 2) confirm the effectiveness of such releases, support the encouraging results previously obtained in the Cayman Islands, Malaysia and the Brazilian town of Juazeiro, 19,27 and the commission's earlier conclusion that such releases are safe.
As male mosquitoes do not feed on blood, the intention is to release only male Ae. aegypti. However, there is a small margin of error during the separation of males and females before field release and therefore up to 0.2% of the released insects may be females. 20 Although the bites of the released females are clearly perceived as a risk by the general public, they are relatively rare and most of the released females will be dead within two days, too soon for them to become infectious even if they have taken a blood meal that contains chikungunya virus, dengue virus or Zika virus. 29 Even if, as may occasionally happen, a female carrying the transgene survives long enough to become infectious, there is no evidence that it will be any worse as a vector than a wild Ae. aegypti. 30 As genomic analyses have revealed many examples of gene transfer between distinct taxa, 31 horizontal gene transfer is always a controversial issue in discussions about the unanticipated risks of releasing a genetically modified organism. However, such gene transfers appear to be very infrequent. There is evidence of horizontal gene transfers from invertebrates to mammals but all such transfers that have been detected appear to be transposon-dependent and very infrequent. 32-34 It does not seem possible that the piggyBac transposon used in the construction of the OX513A strain of Ae. aegypti could mediate the transfer of genetic information from the mosquito to other genomes. This transposon does not codify transposase and, since transposase is not codified in the mosquito's genome, the insert cannot be remobilized.
Even if a transposase were present, there seems to be no risk of genetic transfer between the OX513A strain's transgene and the Zika virus because the viral genome is made of single-stranded ribonucleic acid, which is not a substrate for transposases. When, in an effort to understand the Zika outbreak in Brazil, various isolates of the Zika virus were sequenced, no evidence was found for transposon insertion or any other form of horizontal gene transfer. 35

Conclusion
After circulating around the world, the Zika virus is rapidly spreading through the Americas. 36 The detection of an association between microcephaly in a baby and maternal infection with Zika virus [37][38][39][40][41] has greatly increased the perceived importance of the primary vector of Zika virus -that is, Ae. aegypti -among both the general public and policy-makers. 42 The results of our reassessment of the safety of releasing males of the genetically modified OX513A strain of Ae. aegypti, as well as the results of a related assessment by the United States Food and Drug Administration, 43 indicate that releases of such mosquitoes still offer a safe and potentially effective way of reducing wild populations of the vector. The risk-assessment components adopted by both the commission and the United States Food and Drug Administration 43 are those recommended by the World Health Organization. 44 In the ongoing battle against the Zika outbreak, the Brazilian health managers' choice of control methods has to be based on careful risk assessments and not on risk perceptions. Public perceptions -especially with the advent of fast and global communications via the Internet -are often inaccurate and the general public is often misled by conspiracy theories and