Relative risk
The answer to the question, "Will this product have an impact on the environment?" is almost always "Yes". But it is not the most important question, says the Ecological Risk Assessment Group from Oxford. Instead we should be asking, "Will it result in more or less harm than the current conventional alternatives?"
There is much debate about the potential costs and benefits of genetically modified organisms (GMOs). Promised benefits include reduced use of artificial inputs (for example, fertilizers, herbicides and pesticides) and all the environmental advantages associated with this, as well as the quick delivery of designer products (such as oilseed rape with altered oil composition). It is now widely accepted that moving towards more sustainable agriculture is a desirable goal, part of this process being the reduction of inputs. However, possible costs of the use of GMOs include their potential to disrupt ecosystems. There are currently gaps in our knowledge which prevent us from weighing up these costs and benefits. The agricultural and wider ecosystems are so important to our wellbeing that it is imperative that scientists address these issues.
When answering the question 'Will this product have an impact on the environment?' we would almost always answer 'Yes', but to us this is not the most important question. What we should be asking is, 'Will the GMO result in more or less harm than the current conventional alternatives?' It is therefore important to establish whether the commercialization of GMOs will reduce chemical inputs and their unintended effects and so benefit the environment. Against such potential benefits we must weigh the severity of undesirable ecological consequences attributable to the GMOs themselves.
Some crops, such as oilseed rape, have a number of wild relatives with whom they could cross-hybridize. There is no doubt that such hybrids will form (see, for example, ref. 1), but how will they behave in natural habitats? One undesirable consequence could be the ecological release of the crop or hybrid from the regulating influence of its natural enemies, resulting in increases in their abundance or distribution. Could this occur, resulting in ecosystem disturbance? There are a variety of genetically manipulated food crops at various stages of development. Of these we have selected two groups to illustrate our views: herbicide-resistant and virus-resistant plants. The former group contains crops already grown in North America so representing a group from which data can now be collected on a large scale. The latter represents a group widely planted in China and close to market elsewhere, but for which some ecological questions remain unanswered.
About a decade ago, it was thought that most GMOs would suffer reduced fitness because most genetic change lessens fitness and an additional gene would incur a metabolic cost in the absence of selection pressure for that gene. In reality, fitness costs of herbicide resistance have been notoriously difficult to detect (but see ref. 2). Nevertheless, at best such genes will only be neutral in the absence of the selection pressure (the herbicide). More important, it has always been recognized that stress-tolerance traits may confer advantages to the plant, resistance to viral pathogens being one of these. So one crucial issue is whether transgenic hybrids (with neutral or possibly advantageous genes) are likely to become more abundant in natural habitats. The ecology of a GMO is not fundamentally different to that of any other organism. Hence, ecological theory provides an appropriate framework with which we can assess the potential impact of a GMO. The fundamental criterion for assessing invasiveness is that populations must have an finite rate of increase greater than zero in order to persist (the finite rate of increase is equivalent to the average number of seeds produced per seed sown). This parameter is likely to be very context specific. Field trials in which the ecological context can be manipulated and replicated provide the means by which invasiveness can be estimated (see, for example, ref. 3). No evidence so far suggests increased invasiveness of GMOs containing herbicide-tolerant genes, but further tests are still ongoing.
Virus-resistant plants present a different set of concerns, in that a hybrid could be released from the regulating influence of its pathogens. The study of plant–pathogen interactions has produced two further considerations. First, from the few data that are available, virus-resistance or -tolerance genes appear to be very rare or absent in wild relatives of economic crops4. Second, there are a variety of constructs that confer resistance to viruses (involving genes derived from mammals, yeast, plants and the viruses themselves). These diverse approaches to conferring resistance undoubtedly differ in the genetic risks they present. However, the 'pathogen-derived' transgenes are likely to be substantially different to resistance genes derived from plants and the pathogens have no co-evolutionary history with these resistance mechanisms5. These considerations highlight the need for a greater understanding of the mechanisms and ubiquity of virus resistance, and for the relevant ecological experiments to estimate potential invasiveness. Large-scale field experiments in which a variety of genotypes are placed in a wide range of ecological contexts are required to build up our basic knowledge. Such case-by-case assessment is required before the development of general statements concerning this particular class of GMOs. It is not so much that dangers have been highlighted here, but that there is an absence of ecological knowledge to make an adequate assessment of the risks involved.
Ecology also has a role to play in considering the benefits as well as the risks. It is only with commercial scale releases of genetically modified crops that the hypothesis that GMOs reduce pesticide input can be truly examined. There is a clear need for dispassionate data on this issue. Population genetics is also making a key contribution, providing a link between the short-term ecological issues and the long-term evolutionary questions.
In our view, the scientific way forward is to highlight the knowledge gaps that have recently become apparent, including relevant data on all agronomic alternatives. This may illustrate that GMOs fare better than conventional methods in long-term sustainable land usage. We should be optimistic, as the development of ecological theory over the past twenty years puts us in a strong position to conduct the relevant experiments to fill identified knowledge gaps and to construct the appropriate management practices to minimize any environmental impact. Ecologists should provide better qualified statements concerning the relative risks, both between classes of GMOs, and in comparison to the conventional alternatives.
Rosie Hails
2. Bergelson, J., Purrington, C. B., Palm, C. J. & Lopez Gutierrez, J. C. Costs of resistance: a test using transgenic Arabidopsis thaliana. Proc. R. Soc. Lond. B 263, 1659-1663 (1996).
3. Crawley, M. J., Hails, R. S., Rees, M., Kohn, D. & Buxton, J. Ecology of transgenic oilseed rape in natural habitats. Nature 363, 620-623 (1993).
Ian Cooper, Andy Lilley, Mark Bailey
Ecological Risk Assessment Group, NERC Institute of Virology and Environmental Microbiology, Mansfield Rd, Oxford OX1 3SR, UK
REFERENCES
1. Scheffler, J.A. & Dale, P.J. Opportunities for gene transfer from transgenic oilseed rape (Brassicae napus) to related species. Transgen. Res. 3, 263-278 (1994).
