Written evidence submitted by Professor Michael Bonsall,
Dr Nina Alphey and Ms Charlotte Elves (GEN0032)

Mathematical Ecology Research Group, Department of Zoology, University of Oxford

Our cross-disciplinary research programme explores ecological and genetic aspects of novel methods to control insect populations.

Please find attached a brief policy note for consideration by the Committee. This document outlines the ecological risks of organisms containing gene drives and recommends an international approach to developing appropriate and proportionate guidance. Gene drives are ‘selfish’ genetic elements, which can spread or ‘drive’ themselves through a population of sexually reproducing organisms through a mechanism of biased inheritance. Synthetic gene drives can be created using genome editing techniques and have many potential applications in human and animal health, agriculture and biodiversity conservation. Examples include spreading through a wild mosquito population a trait that reduces the ability to transmit a disease of concern; the aim would be to reduce the human economic and social burden of disease, to reduce harm to livestock, or to protect threatened species. Our interest is specifically in ecological aspects, and we restrict our focus to non-human organisms. We recommend:

a co-ordinated international effort to produce over-arching principles for ecological risk assessment of gene drive technologies, supported by more detailed guidance for different applications;
a campaign of education to raise awareness among researchers and biosafety/bioethics committees;
a database of decisions on contained use of gene drives, to share good practice and promote consistency.

We provide the evidence presented here in a personal capacity and the content should not be regarded as official views of any of the organisations to which we are affiliated.

Ecological Risks of Gene Drive Technologies

Overview

Gene drive technologies offer enormous promise to improve human, animal and environmental health, yet also entail potential ecological risks.
There is a vast array of potential uses of gene drives, over a huge range of species, environments and goals (Figure 1).
Current regulation of genetically modified organisms provides useful precedents, but has some gaps in relation to gene drives.
We recommend a co-ordinated international effort to produce over-arching principles for environmental risk assessments, supported by more detailed guidance for different applications.
At national level, a campaign of education to raise awareness among researchers and biosafety/bioethics committees should be accompanied by a database of decisions on contained use, to share good practice and promote consistency.

Figure 1. Applications of gene drives

Background and Introduction

In sexually reproducing organisms, most genes have a 50:50 chance of being inherited by each of the organism’s progeny. This is known as Mendelian inheritance, after Gregor Mendel, whose 19th Century experiments demonstrated the laws of heredity. Gene drives are genetic mechanisms by which a gene (or genetic element, a portion of DNA) is inherited by more than half the progeny; this biased transmission, even if it is harmful to the organism, can allow the gene to spread or ‘drive’ itself through a population, and gives rise to the name ‘selfish genetic element’2. Gene drive mechanisms of diverse kinds exist in nature 2,3. Molecular and synthetic biology are now sufficiently advanced that engineered gene drives can be created, with increasing ease and design capability4.

Broadly, synthetic gene drives can be partitioned into two categories according to the nature of the trait that they drive into a target population: ‘suppression drives’ reduce the population of the target species (for example by damaging a gene with a function essential to survival or reproduction); ‘modification drives’ (or ‘replacement drives’) introduce a desirable trait (such as reduced capability to transmit a pathogen, or lower resistance to existing control measures) that is intended to persist in the population. Potential applications span human and animal health, agriculture and conservation (Figure 1), and include1,5:

eliminating populations of arthropods, such as mosquitoes, that transmit (or ‘vector’) human, animal or plant disease;
rendering vector populations unable, or less able, to transmit a pathogen (although this approach is harder to achieve than suppressing the population, the vector population would remain in its ecological community);
protecting threatened species by eliminating or modifying populations of disease vectors, or by modifying the threatened population to resist harm from natural enemies;
protecting human health by rendering animal reservoir populations immune to a pathogen;
controlling damaging invasive species;
enhancing sustainability of pesticides and herbicides by driving susceptible genes through populations that have become resistant;
engineering pests or weeds to become vulnerable to an otherwise harmless molecule that can then be used as a non-toxic pesticide or herbicide.

Issues arising from new technology

In 2016, the National Academies’ report Gene Drives on the Horizon: Advancing Science, Navigating Uncertainty, and Aligning Research with Public Values3 remarked on the breath-taking pace of change in gene drive research and asserted a need for ecological risk assessment in the authorization of gene drives for environmental release, given their ability to persist in the receiving environment. Although there is a growing body of research on the molecular biology of gene drive systems, research addressing the ecology of gene drives lags behind and there are considerable knowledge gaps. Research on population dynamics, evolutionary processes and ecosystem dynamics will be essential. The report set out concerns regarding gaps in the current understanding of the relationship between a gene drive’s effectiveness and factors including:

fitness costs - reduced ability of the organism to reproduce or of its offspring to survive to reproductive maturity;
dispersal, and gene flow within and between populations – through interbreeding and mating with immigrants;
mating behaviours (e.g. variation in courtship activities);
population sub-structures (e.g. mix of ages, or sexes, or spatial structure); and
generation times.

Tying these concerns to the current regulatory framework in the USA, the report concluded that the environmental assessment and the environmental impact statements required by the National Environmental Protection Act, are incapable of properly characterizing, and capturing, the risks of gene-drive modified organisms. Report recommendations 6-1 to 6-3 centre on the development and application of appropriate ‘ecological risk assessment tools’.

In this policy note we focus on ecological considerations. An effective ecological risk assessment tool should identify ‘hazards’ and accurately predict the ‘risks’ of harmful effects arising from those, and avoid conflating the two (Box 1). Potential harms that might materialize following genetic modification of organisms include: creating new or more vigorous pests and pathogens; exacerbating the effects of existing pests through hybridization with related transgenic organisms; harm to non-target species, such as soil organisms, non-pest insects, birds, and other animals; disruption of biotic communities, including agro-ecosystems; irreparable loss or changes in species diversity or genetic diversity within species3,6.

Box 1. What is risk?

‘Hazards’ are substances or activities with potential to cause adverse effects to living organisms or environments7, regardless of how likely or unlikely those effects are. Environmental release of a gene drive is a hazard with a variety of potential effects (see list in main text). The risk, or probability of a harmful effect occurring, depends on the exposure of the population, species, habitat or ecosystem characteristic of interest to the hazard. ‘Risk’ characterisation calculates the likelihood that a particular substance (or activity) will cause harm in the light of the nature of the hazard and the extent to which people, animals, plants and/or the environment are exposed to it7. In one common formulation risk = hazard ⨯ exposure. Risk assessment entails hazard identification (the gene drive release), hazard characterisation (the harms that might occur), exposure assessment (dispersal, gene flow and ecological interactions) and risk characterisation (probability of harm)7.

Gene drives and gene editing

Among the drive mechanisms found in nature, attention has focussed on those based on endonucleases, i.e. proteins that cut a particular pattern of DNA, typically a sequence that is only expected to occur once in the genome. Homing endonuclease genes, or HEGs, are one such class of selfish genetic elements, found in microbes. Natural HEGs generally confer no advantage on the host organism, but spread themselves through populations, for example, by encoding a protein that cuts the DNA sequence and uses a cellular repair mechanism to copy themselves into the ‘empty’ chromosome in individuals that carry the HEG on only one chromosome of a pair (Box 1).

In Anopheles gambiae mosquitoes, the major vector of human malaria, researchers are using HEGs to reduce their capacity to transmit disease8,9, which occurs through females feeding on blood (male mosquitoes do not bite). One approach deploys a HEG that cuts, during the process of sperm formation, a short repeated sequence of DNA found on one of the sex chromosomes. Sperm that would generate a daughter (on fertilizing an egg) have their DNA ‘shredded’ by multiple cuts, while sperm that would produce sons are unaffected; in experiments this resulted in the male-determining chromosome being inherited by about 90% of progeny rather than the Mendelian 50%. Another approach places a HEG inside a gene that is essential for mosquito survival, or for female fertility, so that successful homing – copying the HEG into the normal version of the gene on the other chromosome of a pair – knocks out that essential gene. The resulting reduction in either total mosquitoes or female mosquitoes means there are fewer females to lay eggs that become the next generation and fewer females to bite people and transmit disease. These HEGs were adapted from nature and required years of effort to create a synthetic version with the desired target sequence and positioned appropriately in the genome9.

CRISPR (Clustered regularly-interspaced short palindromic repeats) are segments of DNA from bacteria, involved in their immune systems. An accompanying protein, such as Cas9 (CRISPR-associated protein 9), acts as a pair of molecular scissors that will cut DNA at a specific target sequence determined by a small guide molecule (RNA). This system can be designed to target practically any specified sequence of DNA in a genome, to insert, edit or disrupt genes (Box 2), and researchers have done so in a range of organisms from yeast to monkeys, and in human cells4. The ease with which this can be done is fuelling calls to use such systems responsibly, especially as inadvertent creation of gene drives is now possible10,11.

Box 2. Harnessing DNA repair

All organisms have natural mechanisms to repair or destroy faulty molecules. When DNA is damaged by a break or cut, the cell’s machinery has broadly two methods to repair it.

The simplest, conceptually, is to stick the two broken ends back together, which is known as non-homologous end-joining (NHEJ). If this is done perfectly, it is impossible to tell afterwards whether a cut occurred. However, the process is often imperfect, for example with bases omitted (analogous to trimming frayed ends before re-joining), or even joining two ends from different breaks. NHEJ is error-prone and can result in an alteration to the DNA that is disruptive. One practical application is to use an endonuclease to cut a target gene causing a ‘knock out’, resulting in loss of the chosen gene’s function.

As most genes occur on paired chromosomes, the alternative method of repair is to copy and insert the sequence from the other, homologous, chromosome, known as homology-directed repair (HDR). When a gene that was present in only one copy is inserted onto the sister chromosome in this way, it is known as a ‘homing’ event (hence the ‘H’ in HEG). This allows the gene to spread through a population, even if it is harmful to individuals, as long as the drive advantage outweighs the fitness cost disadvantage12. With novel techniques using CRISPR systems, RNA-guided Cas9 can be used to cut a specific DNA sequence, and to trick the cell’s repair machinery into performing HDR using a co-delivered cassette of DNA instead of the homologous chromosome4. This enables precise genetic manipulation, and insertion of almost any chosen DNA sequence of suitable length, resulting in gene ‘knock-in’ (inserting a functional gene) or in gene editing (changing the sequence of an existing gene). This provides a mechanism to drive a desirable trait through a population.

Gene drive should not be confused with genetic modification (artificially altering genetic material); they overlap but are not the same. Genetically modified (GM) plants or insects currently in the field confer traits that are inherited in Mendelian fashion with no gene drive. Wolbachia are maternally-inherited parasitic bacteria that spread through an insect population by favouring the offspring of Wolbachia-infected females. Transfer of Wolbachia strains into mosquito species can reduce their ability to transmit some human pathogens13, and this is being tested as a non-GM gene drive to spread that transmission-reducing trait through wild populations14. Regulations for GM crops are well established. Population-suppressing GM insects are in trials for mosquitoes15-18 and for agricultural pests19-21, and regulations are evolving. In the USA, for example, requirements depend on the insect species and proposed trial site, including environmental assessments by the Department of Agriculture (USDA, for agricultural pests)22 or the Food and Drug Administration (FDA, for mosquitoes)23. The Wolbachia drive, lacking any genetic modification of the release organism, requires no compulsory risk assessment prior to release in many jurisdictions; voluntary assessments of risk have been published24.

While GMO regulations provide some precedents, new or distinct features of gene drives pose additional challenges. The CRISPR-based molecular technology is very much easier and quicker to use than older methods, e.g. HEGs, and have a vast range of potential targets (at all scales - DNA sequences, species, and objectives). GM insects contain self-limiting constructs, which because of their disadvantages to the insects will quickly disappear from the population unless large-scale releases are sustained. In contrast, a self-sustaining gene drive might spread through an entire population, and possibly beyond, potentially from a single small release or escape. Current regulations were designed for managed populations (agricultural settings for engineered plants or modified insect pests), but gene drive applications are typically designed for wild populations, which may change matters to be considered and their priority or weighting.

Contained use – escape risk

The primary ecological concern is spread of a gene drive through one or more populations, possibly an entire species, following inadvertent release. Modified individuals could escape from a laboratory, or from a ‘contained’ trial. This includes invasion into other lab stocks, which might not be managed so securely as the gene drive strains, as well as into wild populations.

The consequences of this could include:

individual-level phenotypic (novel trait) effects;
population/species-level effects;
wider ecological consequences;
effects on public confidence in science, genetic technology and its regulation.

Biosafety / bioethics committees should be aware of the potential risks relating to gene drives. The CRISPR molecular technology makes it too easy to create a gene drive without intending to, and gene drives constructed inadvertently pose the highest risk of escape into the environment because of unawareness of the need for containment. Steps should be taken to encourage researchers to submit all proposed CRISPR-based genetic experiments to the committee, and further detailed consideration given where the molecular design is such that a gene drive mechanism might be created.

With such measures in place, current regulations and guidance for contained laboratory use of GM organisms are reasonably appropriate for gene drive systems. The main focus for improvement should be on:

Education and information, to raise awareness;
Guidance or best practice on risk assessments;
Making information on existing decisions openly available, to promote consistency and facilitate some calibration against other studies and other committees.

Open (intentional) release

A key consideration when assessing the ecological risks is the invasiveness of the gene drive. Depending on the details of the drive mechanism, a drive may be ‘global’, in the sense that it could spread throughout a population from very low initial numbers, and hence potentially to all populations however tenuously connected, and even to closely related species. Other kinds of drive mechanism have frequency-dependent genetics, and can only spread if they are present above some threshold frequency (in practice, this is complicated by natural variation in demography, environmental conditions, spatial distribution, etc). Threshold-based designs offer the prospect of confined trials, where release of sufficient numbers into a carefully selected, isolated target population can allow effects to be studied there, but any dispersal into non-target populations should be rare and far below the threshold and so unable to spread there. With ‘global’ gene drives there may be no such thing as a confined field trial. In terms of practicality and cost-effectiveness, ‘local’ drives will need larger numbers of insects released than ‘global’ drives to achieve the intended benefits.

Factors affecting the spread of gene drive systems include not just the genetics (and the genetic invasiveness of the drive system), but also the population dynamics, including the effect of spatial distribution of the organisms and how interconnected populations are. To assess the potential benefits and risks of gene drive systems, it is necessary to understand both population genetics and dynamics and how they interact. There might be an establishment phase, where the gene drive increases to high frequency in the initial (sub-)population, then a spreading phase, where the trait diffuses through the target population and perhaps beyond. The time for the driven trait to reach fixation (100% gene frequency) depends on the population dynamics. We are increasingly aware that populations are genetically interconnected more than previously appreciated. There are very few genuine islands that are genetically isolated (desert oases, and physical islands are potential examples). Because field trials are not containable, regulators have a unique challenge as the step-wise process recommended for trials for GM insects, for example (see below), cannot be followed for global gene drive systems. Mathematical modelling is particularly valuable for informing safe regulation, as it can be used to help regulators understand how a potential gene drive trait could spread, how the population genetics are impacted and, where the aim is disease control, how it affects the disease dynamics.

Various measures have been proposed to manage or mitigate ecological risks 1,25:

Physical / barrier containment – laboratory containment and biosecurity for early-stage investigations, to prevent escape into the environment;
Ecological containment – geographical location of experimental work and testing in an inhospitable environment where escaped organisms would be unable to survive or find mates;
Molecular containment – arrange that the drive strain is unable to have effect in susceptible populations, for example by targeting a gene that is only present in the target population (extensive monitoring would be needed to confirm this) or testing gene drive effectiveness in a laboratory population by targeting an inserted transgenic sequence so that any escapes are harmless;
No drive – for gene editing applications separate the Cas9 and guide so that drive does not occur, and conduct field trials to investigate ecological effects using strains that are engineered to have the desired trait but without the gene drive mechanism to spread it;
Build local rather than global drives - for example, elements which cannot spread unless they exceed a threshold frequency in the population, or multi-part genetic designs that drive initially then decay away and so have limited time to spread26;
Reversibility – design, build and evaluate a ‘reversal drive’ before a modification gene drive system is released into the field, designed to overwrite changes spread by the initial drive if required (although this would not necessarily reverse any ecological consequences suffered in the interim period).

It might be useful to draw comparisons between gene drives and bio-containment guidelines applying to human or animal pathogens, where there are commonly agreed criteria that help decide what level of containment measures are needed to use in a lab environment. A similar framework for gene drive containment could help biosafety and ethics boards understand what measures would be appropriate for particular gene drive studies. With gene drives, there is a blurring of traditional lines between lab and field studies and full release, so the environmental risks should be considered at an early stage.

Resistance

As with any new technology, there is the possibility of evolution of resistance. Resistance does not pose a direct ecological risk in itself. The effects of resistance to vector or pest population control are mostly economic or social, in that resistant populations continue to cause harm to agriculture or forestry or transmit disease. There can be indirect ecological effects, for example, when reduced efficacy of chemical insecticides or herbicides tends to lead to more frequent use and greater quantities of active ingredient. Also, when there is strong selection for resistance, other unrelated genes that are in close proximity on the chromosomes can be favoured by ‘hitch-hiking’ effects, which might in turn reduce genetic diversity. However, this would be true of any random mutation that provided a strong advantage in a particular environment and so became widespread.

In the case of gene drives, some of these same factors will be relevant. The primary impact of resistance will be to prevent or slow the spread of the desirable trait or population-reducing change and hence to delay or inhibit (biologically or economically) achievement of the ultimate aim. An endangered species may become extinct before a disease tolerance trait manages to spread, for example, even if resistance only delays it.

To enhance the chances of success of a proposed gene drive programme, thought should be given at an early stage to the mitigation or management of resistance. The most obvious source is a change in the target DNA sequence such that the target site is not recognised and the endonuclease will not cut it, thus disabling the drive mechanism. Molecular aspects to reduce such possibilities should be considered when selecting genes to target, and the guide RNA that will recognise them. Designing a construct that cuts at multiple sites within one gene would reduce the chance of a mutation being sufficient to negate all the cuts and thereby escape the homology-directed repair that causes the gene drive. However, repair by non-homologous end-joining, is imperfect and can itself generate new mutations in DNA sequence; poor design could potentially result in creation and drive of resistant mutations through a population. If the desired trait is not an inherent part of the gene drive mechanism itself, but is included separately within the genetic construct as a distinct cargo gene, that cargo could become unattached from the driver, so the gene drive spreads through the population but does not achieve the desired effect. These kinds of resistance are a concern for the effectiveness, and cost-effectiveness, of gene drive technology, but are not relevant when assessing risks of detrimental ecological effects.

In the case of releasing vectors where the driven trait is inability to transmit disease, attention should be given to the possibility of resistance evolving in the pathogen. Malaria parasites are sophisticated organisms with complex multi-stage life cycles, and are likely to represent a greater risk of evolving resistance to evade pathogen-blocking or transmission-reducing traits than viruses are. There is no reason to suppose that adaptation conferring resistance to the gene drive mechanism would be any more likely to result in increased virulence or pathogenicity than any other randomly occurring mutation. Mathematical modelling could be very useful to explore the extent to which changes in pathogen traits in response to a gene drive system, and the fitness advantages (or disadvantages) of those changes, could impact on the population genetic composition and the consequences for disease dynamics.

Policy considerations

Regulations will need to be flexible so that they can keep up with the rapid pace of technological developments. As gene drives can be so diverse – in their construction, application, objectives and environmental context – gene drives designed for release into the environment will need to be assessed on a case-by-case basis, by species, by construct and focused on the particular receiving environment. It would be preferable to regulate the product or phenotype (the manifestation of the novel trait) not the process of modification6.

When identifying potential harms and assessing their risks, regulators must judge the proposed gene drive strategy against suitable alternatives. The appropriate comparison might be with some combination of the current control method, an idealised version of current technologies (doing what is currently done but better), or no action. However, it would not be appropriate to compare a proposed gene drive system to an idealised risk-free alternative that does not exist in practice. For example, EU regulation of GM plants has been criticised for overzealous interpretation of the Precautionary Principle (an approach to decision-making under conditions of scientific uncertainty that is sometimes characterized as ‘better safe than sorry’)27. The whole of a proposed gene drive release programme should be considered for the comparison, not just the nature of the modified organism. This should include production of modified organisms, delivery into the environment, monitoring, concurrent use of other methods such as chemical controls, and other changes in practice such as no-till crop management. Ideally, there should be consistency across regulatory regimes, including matters such as what timeframe to consider in identifying potential harms and benefits, and how to apply discounting to convert monetary amounts at different time points into comparable present values.

Potential geographic spread and relative lack of reversibility are concerns for gene drive. These are also relevant to biological control, which involves the release of natural enemies to counter the impact of invasive pest animals or plants. Biological control therefore provides some precedents on which to draw, in formulating questions for regulators to consider and for suggesting experimental evidence that should be provided as part of a permit application.

Risks should not be considered in isolation, but evaluated against some level of acceptable risk and, if not exceeding that, weighed up against the prospective benefits to decide whether it is appropriate to take those risks in order to achieve the likely benefits. Regulation of GM plants varies between regimes as to whether regulators can, or do informally, take account of benefits when performing environmental risk assessments. Ideally risks should be viewed in the context of benefits, but no requirement for a formal cost-benefit analysis should be imposed. This would likely be too costly, for a small organisation applicant and for a regulatory body utilising limited public funding, and would be more stringent than for other methods (chemical producers are not required to demonstrate epidemiological outcomes for vector control products). Economic and social benefits need to be more clearly articulated when thinking about risk. Public attitudes typically take account of perceived ‘value’ of technology, for example, opposition to GM crops is greater than that to GM vaccines.

Gene drives could potentially spread across international borders, which may happen if there is gene flow between interconnected populations. There is a need to consider implications for sovereignty issues and international regulatory approval mechanisms. There is currently no real recourse in international law for such matters, except possibly by invoking trade rules if a released gene drive caused economic damage.

In developing countries regulatory capacity may be lacking, understanding of the relevant science may be inadequate, and ability to make impartial assessments may be impaired. There is therefore a need to have international guidance or standards.

There are existing precedents concerning GM insects - a field in which gene drive research is relatively advanced - for which several guidance frameworks have been produced. In 2014 the World Health Organization (WHO) issued guidance for the testing and regulation of GM mosquitoes28, and in 2012 the European Food Safety Authority published a regulatory framework for GM animals29. In support of the development of GM insect technologies, both of these documents recommended a tiered approach to underpin environmental risk assessment, progressing step-wise from laboratory studies (focussed on the molecular biology and simple ecological processes), through contained or confined trials, to pilot implementation. At each step, scientific evaluation would be accompanied by risk assessment, risk management and risk communication.

The challenge is that gene drives can and will be developed in huge range of species and applications across diverse sectors – including conservation, agriculture, and human health. It is questionable whether, and if so how, a single common statement could be developed in practice that would have widespread recognition. Ideally a core set of principles could be agreed, with expansion of guidance and practical examples relating to particular fields. International co-ordination will be essential – with such a broad range of applications, no single body could have credible authority, but there are some possibilities: for example, the World Health Organization WHO for human health, the United Nations Food and Agriculture Organization FAO for agriculture, the International Plant Protection Convention IPPC for agricultural plant pests, the International Union for Conservation of Nature IUCN. Such bodies should be encouraged to co-ordinate, and encourage consistency.

International standards are not regulatory instruments, they only become effective when the principles they prescribe are adopted by countries within their national legislation or requirements. However, international bodies can provide guidance to national governments, regulators or agencies, and set out standards that countries can adopt if they choose. A notable example is the widely adopted standard for wooden packaging material used to ship goods between countries (Box 3). We recommend that international standards on gene drive regulation should be developed and designed to be relevant, easy to implement, and attractive to many countries.

Box 3. Internationally adopted standards for wooden packaging

Plant pests have increasing opportunities to take advantage of global trade to invade new areas, potentially causing economic and environmental harm. Phytosanitary (plant health) standards have been developed to prevent the international transfer of insects and diseases that could affect plants or ecosystems. The International Plant Protection Convention (IPPC)30 operates under the auspices of the Food and Agriculture Organization (FAO) of the United Nations (UN). Its International Standards for Phytosanitary Measures (ISPMs) are science-based and are recognised by the World Trade Organization (WTO), aiming to minimise pest risk without creating unjustified barriers to trade. ISPMs are adopted by many countries, typically through national plant protection organisations. ISPM 15 Regulation of wood packaging material in international trade requires wood packaging used in international trade, such as pallets, to be made from debarked wood, be treated (heat treated or fumigated), and stamped or branded to indicate compliance. These measures are required to be applied within an official certification system and import controls are recommended to monitor compliance. ISPM 15 has been adopted by many countries, giving this international standard broad global reach and in effect harmonising regulatory requirements across much of the world.

Conclusions

The suitability of current environmental and ecological risk assessments of gene drives is being considered, in terms of their ability to meet legitimate regulatory aims and whether current requirements are achievable and desirable. Assessments must be non-discriminatory and proportionate, to protect biodiversity and minimise harm to human or animal health and the environment, whilst not being so onerous or restrictive as to quash the achievement of the potential benefits or discourage research and innovation into biotechnology solutions to significant societal or economic problems. Questions must be resolved over what information can reasonably be expected by an effective regulatory system, and how responsibility for the collection and analysis of this data should be distributed across governmental agencies, research institutions, applicants for release permits and other parties. Such issues are under scrutiny, by policy-makers in the UK31,32 and elsewhere. Research scientists are actively involved in the debate about suitable regulation of gene drive technologies and related societal issues1,25.

The applied use of gene drives beyond the laboratory poses new issues that are not addressed adequately, or with sufficient clarity or scope, by existing guidance on genetically modified organisms. As gene drives blur the distinction between a field trial and uncontained implementation, and modified organisms do not respect geopolitical boundaries, an international approach is preferable for developing and maintaining guidance for regulation, and best practice for ecological and environmental risk assessments.

International bodies such as the WHO, UN and others, already have mechanisms for interacting, setting up joint working groups, for example, and recognising each other’s standards in international agreements. International guidance and standards for responsible development, testing and implementation of gene drive technologies should be achievable in principle. The relevant international organisations will need both the funding and an appropriate mandate to achieve them in practice.

Recommendations:

A campaign of education and information to raise awareness among researchers and their institutions.
An open database to provide transparent precedents for decisions by biosecurity and bioethics committees on contained use of gene drives (where the ecological risks arise from unintentional escape), and to promote consistency across organisations and countries. Include a range of good practice examples, with an indication of the information considered and reasoning behind the resulting decision.
A collaboration between international bodies (such as WHO – Tropical Disease Research, UN – FAO, World Bank - Hazard Management Unit) to develop high-level ecological risk assessment principles for developing, testing and implementing gene drive technologies, and create a framework to enable development of supporting guidance relevant to particular fields in human health, animal health, agriculture and conservation.

January 2017