 |
| About Bioline | All Journals | Testimonials | Membership | News |
|
||||||
|
||||||
Online Journal - URL: http://www.bdt.org.br/bioline/by Hierarchical risk assessment of transgenic plants: proposal for an integrated system. Strandberg, B., Kjellsson, G. and Lokke, H.
Department of Terrestrial Ecology
Date received: 18 March 1998
Code Number:BY98002 SUMMARY The increasingly high number of applications to release or market transgenic plants calls for an effective and uniform treatment. Regulatory authorities in most countries are evaluating the safety of transgenic plants; however, internationally standardised and consistent schemes, test procedures and concepts have not been fully developed. This paper discusses the procedure of risk assessment of transgenic plants and proposes a three- step hierarchical framework for the assessment required before marketing is permitted. This approach is parallel to the hierarchal test systems which have been established for handling risks from toxic compounds. The relevance and complexity of information obtained can be increased but often at high cost and time consumption. The approach requires increasing information from the individual level through population and ecosystem levels to regional scale. It is proposed that experimental tests of the transgenic plants should be conducted at each tier, and the evaluation following the analysis should be based on acceptance criteria. The use of the principle of substantial equivalence of the transgenic plant and the receiver plant is suggested, but the concept should be further developed. Follow-up mechanisms such as monitoring programmes may be logical consequences of the principles of risk assessment, and may contibute to the learning process of handling risks of transgenic plants. Keywords: acceptance criteria, ecological risk assessment, genetically modified plants (GMP), hierarchical approach, monitoring, stress tolerance, toxic compounds, transgenic plants INTRODUCTION
Regulatory authorities in most countries are evaluating the safety of transgenic plants or other organisms. In 1987, the Society of Environmental Toxicology and Chemistry suggested research priorities in environmental risk assessment including risks arising from the use of biotechnology in plant breeding (Fava et al., 1987). In most countries risk assessment of every new transgenic plant is required as is the case in the European Union (Council Directive 90/220/EEC) and in the United States (7CFR part 340, Final Rule, 31. March 1993). Furthermore, parties ratifying the Convention on Biological Diversity are required to take steps to regulate and manage the risks associated with the use and release of living modified organisms resulting from biotechnology. The international Biosafety Protocol to the Convention on Biological Diversity is in the final stage of negotiations (Burgiel, 1998) but the data requirements need further elaboration. Concerns about genetically modified plants, or transgenic plants, centre on six issues of ecological, agricultural and health relevance:
2: will transgenic crop plants become weeds of agriculture? 3: will transgenic plants become invasive of natural habitats? 4: will engineered genes be transferred to related weeds or wild relatives (the hybrid offspring may then become invasive)? 5: will transgenes affect non-target organisms? 6: will transgenes affect ecosystems, biodiversity or regionally adapted genetic variation?
Risk assessment of genetically modified plants is commonly managed by separate administrative bodies, of which some address the assessment of the risks to human and domestic animals (issue 1, above) and others address the ecological risk assessment including agricultural aspects (issue 2-6). The present paper mainly focuses on ecological risk assessment of higher plants belonging to the taxonomic groups Gymnospermae and Angiospermae. The ecological risk assessment includes agricultural aspects and health aspects will also be mentioned. Those aspects relevant to human food safety issues have been addressed by the Organisation for Economic Cooperation and Development (OECD) and the World Health Organisation (WHO). Parker and Kareiva (1996) gave a thorough historic background for ecological risk assessment based on petitions for regulation in US to that date, and also described the situation in some other countries. Initially, Parker and Kareiva found a lack of agreement between ecologists and members of industry regarding terms such as 'weediness' and 'invasiveness', and a tendency to focus exclusively on weediness in an agricultural context and neglect potential ecological effects on natural ecosystems. The US regulation petitions, especially the earlier ones, gave a verbal reasoning about weed traits and the most consistently cited source of evidence was Baker's list of weed traits common to many weed species (Baker, 1965). While Baker's characters have value in examining existing weeds, they are unlikely to be useful when predicting whether an introduced species will become a weed (Williamson, 1993; 1994; Perrins et al. 1992a, b). However, the situation may be changing towards petitions presenting more quantitative data and including more extensive and scientifically based discussions. Some of the early weaknesses could be attributed to a simple learning process, as argued by Parker and Kareiva (1996), not only for the members of industry but also for the ecologists as well, who were slow to provide more than vague guidance on risk assessment (Regal, 1993). Towards tiered risk assessment In the stepwise procedure addressed in EU Directive 90/220/EEC the possible adverse effects of the transgenic plants are tested in stages of increased scale and decreased safety measures from contained experiments in the laboratory through small- and large-scale trials to commercial release. Each step needs separate regulation. This stepwise procedure was implemented to bridge the gap between simple, reproducible tests, which yield scientific easily understandable results but reveals relatively little about complex ecological effects. In the following, a hierarchical system for risk assessment of transgenic plants before marketing is proposed, being consistent with the idea behind the step-by-step procedure and applicable to this procedure. Thus, the proposed system represents the final body of information needed in order to make decisions about marketing and it includes data obtained from contained experiments and the preceding field trials. Hierarchical test systems are well-known from risk assessment of toxic compounds (e.g. van Leeuwen, 1995; Suter, 1993). Rissler and Mellon (1993) were first to argue for a tiered protocol for ecological risk assessment of transgenic plants, consisting of 3 tiers. The first Tier involved a basic analysis of existing information on the parental plant, and the second and third Tiers required experimental data generated on the transgenic plant itself. This procedure is similar to the one recently proposed by Kjellsson (1997) in which the complexity and relevance are suggested to increase from individual, through population to ecosystem level. Also, Illucea (1996) emphasised that risks should be addressed through a hierarchy of measures, regulation, management or other means of control. In parallel, the Nordic Working Group on Food Toxicology and Risk Assessment (NNT) has suggested a stepwise evaluation strategy for novel foods (NNT, 1991; Knudsen and Ovesen, 1995). The procedure suggested by NNT consists of four steps. The first step requires data, corresponding to the basic information for ecological risk assessment, on origin and composition of the parental plant, exposure to consumers, chemical data and existing toxicological data. The second step involves data on new traits or effects, and secondary/compositional changes of the plant. For plant based foods which are not cleared during step one and the step two procedures, a testing programme is recommended which includes nutritional, metabolic, and toxicological studies in vivo and in vitro. This third step includes nutritional and metabolic studies in rats, testing extracts for genotoxicity, combined 90 days toxicity/reproduction study in rats, and combined chronic toxicity/carcinogenicity study in rats. These studies are to be performed in a logical sequence by taking advance of the results from the preceding studies in the selection and the design of the following studies. Finally as the fourth step, data from human studies with volunteers having meals prepared from the engineered plants should be evaluated (Knudsen and Ovesen, 1995). One of the intentions behind the hierarchical approach proposed by Rissler and Melon (1993) was to stimulate the discussion on assessment and to make a basis for future modifications reflecting gained experiences. Since 1991 the number of releases has increased exponentially, and the need for regulations and guidelines in biotechnology have been felt world-wide at country, regional and international levels (Gaugitsch & Torgersen, 1995; Illueca, 1996). We will formulate an approach for risk assessment of transgenic plants which fulfil the need for:
ii: obligatory basic information involving experimental data, iii: standardised test protocols specifying the experimental conditions relative to the inserted trait, iv: clarifying the acceptance criteria, and v: experimental tests on the transgenic plant at each tier.
Proposed new system We propose a tiered hierarchical approach for risk assessment before marketing (Fig. 1) including all above mentioned issues (1-6), in which experimental tests and model simulations of varying complexity are found at each tier. Generally, the hierarchy consists of three levels of increasing complexity. Such an approach is relevant to some risks e.g. the invasiveness to natural habitats (issue 3), which might be tested at increasing level of complexity, while others should be assessed at one specific level e.g. gene transfer (issue 4). If genes through introgression are transferred to weeds or wild relatives, then the weediness in agroecosystems (issue 2) and the invasiveness to natural habitats (issue 3) of the hybrid should be tested following the three step tiered procedure.
At each tier in Figure 1, an evaluation of the available information and test data is made and decision on acceptance and marketing, further assessment, or rejection is taken. The evaluation at each stage is carried out by the competent international and national authorities under guidance of scientific expertise. Decisions should be based on specific criteria for acceptance. Such criteria should be based on science, but they may also include political considerations, e.g. on ethics and socio-economy. Scientific criteria of acceptance can be expressed as defined quantitative differences between the transgenic plant and the receiver plant on e.g. growth rate, survival in competition experiments, content of toxic substances, etc. International organisations (OECD, UN, EU, etc.) are currently developing principles for acceptance. The criteria for acceptance or rejection of release should be decided by these international fora with the assistance of scientific expert panels. The specific standardised tests and delimitations of criteria for risk assessment of transgenic plants is an area for further research and development. Some types of data may result in rejection based on one single criteria while other data may result in rejection based on an overall evaluation. Tier 1 comprises a range of background information on the transgene, the parental plant, and on effects on human health and the environment. Some of the required basic information is already available in scientific research literature, monographs from OECD, national environmental agencies, etc. This information on the receiver or parental plant should be supplemented by simple standardised and replicable tests on the transgenic plant and its hybrids with wild relatives compared with the receiver plant addressing both general and trait-specific risks in short-term experimental designs. Tier 1 tests should include changed competitiveness, recruitment (i.e. survival and reproduction) and toxicity. Such tests could easily be run in parallel to the efficacy tests routinely conducted by the companies. Transgenic plants offer a range of possibilities for insect and pathogen control, and in this regard have the safety advantages of reducing the possibility of spray drift and ground water contamination. As a prominent example, a series of plants have been transformed to express toxins from Bacillus thuringiensis (B.t. protoxins). Cannon (1996) has reviewed the recent developments in the biology and molecular genetics of B. thuringiensis into the context of biological control and applied entomology. Plants engineered to express toxic substances like insecticides may present risks to non-target organisms, and pharma- ceuticals, diagnostics, or other industrial chemical-producing crops may threaten wildlife feeding on the plants. At Tier 1 plants which express toxins or pharmaceuticals should be tested for effects on a number of non- target organisms, by use of laboratory procedures. The Tier 1 analysis and evaluation separate the plants and food products into three categories: 1) those that do not show a higher risk than their non-transgenic counterparts are considered to pose low risk and are accepted according to the principle of substantial equivalence, 2) those that present a risk which exceeds the acceptance criteria are rejected and the analysis ends, and 3) those that are insufficiently described in relation to potential environmental risks. The latter category moves to Tier 2. The Tier 2 comprises advanced testing of the transgenic plant for which environmental risks are anticipated. The level of complexity, and the length of the test period are increased at Tier 2 at which changes in relative ecological performance in different competitive environments and effects on ecosystem (issue 6), including advanced testing of non-target organisms (issue 5), are addressed. The Tier 2 tests are followed by a risk analysis similar to that performed at Tier 1. At the third Tier, we propose that risks at regional scale should be analysed i.e. risk aring from use of the plant on large and adjoining areas. Data from monitoring programs and literature will be useful. The environment will experience the total area grown with different transgenic plants together in various combinations over time, while the regulators are dealing with them on a product-by-product basis. The assessment of transgenic plants case by case could introduce properties which pose low or no risk at the local, small scale, but entail a considerable risk to the environment when used at a large, regional scale; especially the risk of (local) extinction of non-target organisms (issue 5) should be addressed. Furthermore, climatic differences between regions may raise different questions in different countries. This aspect has pronounced relevance for transgenic plants with altered tolerance to environmental stress. The level and type of hazard that exist may change greatly with different climatic conditions e.g. from mountainous Middle Europe to a Mediterranean lowland situation. We suggest to include this view on the risk assessment as the final step of the hierarchical risk assessment scheme, as depicted in Figure 1. After step 2 or 3 in risk assessment, when marketing and commercialization is accepted and extensive use can be foreseen, adequate monitoring programmes need to be set up for new, unfamiliar combinations of species and traits. These should allow detection of transgene escape from cultivated fields (both directly and through hybridisation), and early detection of any major changes in habitats prone to invasion. Although major hazards to the environment are unlikely at this stage of the assessment, risk assessment cannot possibly foresee all hazards that may occur in a complex natural environment. Consequently, monitoring should be done to minimize the risks, and help to indicate if management measures are needed to remedy any unforeseen and undesirable situations in the environment e.g., loss of species, unexpected changes in plant cover. It is important that cheap and effective methods are developed. A simple indicator of invasion is increased frequency of the plant species outside the cultivated area. If unintended release is indicated by the monitoring procedures, identification of the suspected transgene should be performed by DNA analyses (PCR and RFLP) or other adequate techniques (e.g. Kjellsson et al., 1997). Experimental data at all tiers should be provided by tests described in detailed test guidelines and protocols and all testing and studies should be conducted using the receiver species as controls. According to the principle of substantial equivalence used for food safety evaluations, a similar principle of ecological and/or agricultural substantial equivalence should be the basis for comparisons of the GMP and the receiver plant in question. Compilations of current test methods and suggestions for their use in risk assessment of transgenic plants are available for main processes such as: Competition, establishment, pollination and gene- transfer (Kjellsson & Simonsen, 1994, Kjellsson, Simonsen & Ammann, 1997). The study by Linder and associates (Linder & Schmitt 1995) on oil-modified Brassica napus focusing on effects of oil modification on seed dormancy, germination and early establishment is an example of a trait specific test suitable at Tier 1. The test design and model by Parker, Kareiva and associates (Parker & Kareiva, 1996; Kareiva, Parker & Pascual, 1996), the PROSAMO (Planned Release of Selected and Manipulated Organisms) test procedure developed at Silwood Park (Crawley, 1991; 1992; Crawley et al., 1993, Rees et al., 1991) and the mesocosmos test (Stockey & Hunt 1994) are examples of Tier 2 test designs. These test designs are not fully replicable and should be further improved and standardised. The competitive environments, which the plants are planted into should, however, represent generally applicable conditions such as particular vegetation densities, habitat type and specific climatic conditions. When tests are performed, extrapolation and predictions of results beyond the data available become an important issue. Especially at Tier 2, mathematical modelling can be used to integrate data, and by sensitivity analysis make it possible to identify key-parameters for the processes analysed. Some areas of modelling which are especially suitable in risk assessment of transgenic plants include: two-species competition models, models of gene migration between populations and inbreeding, and spatial models of vegetation dynamics in relation to invasion. It is essential that model data based on greenhouse or laboratory tests are validated by comparable field trials. However, the reliance on results from modelling in actual risk assessment should be used with caution. Even the most complex model can only to a certain degree simulate the complexity of ecosystem processes. Tier 1: Background information and basic testing A thorough knowledge on the characteristics of the receiver plant is central to risk assessment of transgenic plants. The main reasons for this are: 1) It gives the required background for evaluation of the effects of inserted traits; 2) It gives valuable information on possible gene-transfer and hybridisation; 3) It indicates areas where adverse effects exist or may develop; and 4) It points to processes and life-stages that could be important to include in test procedures and protocols. In risk assessment procedures by government or interstate authorities such as EU, US, OECD and UN (EU Directive 90/220/EEC; NRC, 1989; OECD, 1992; UNEP, 1996) basic information (Table 1) is currently required although the requirements have generally been stipulated in a very broad sense. The suggested requirements, presented in Table 1, include additional and more precise types of basic information, which are commented on below. We find that by keeping this information separate and by making requirements for each species/cultivar only once during the procedure, i.e. at the Tier 1, this will speed up the procedure and make it more manageable. Table 1. Background information on the receiver plant. Main review data required in pre-release risk assessment of genetically modified plants based on the EU Council Directive 90/220/EEC. Suggestions for additional information requirements or more precise information are indicated by an asterisk.
--------------------------------------------------------------------------
Main category Subjects
-------------------------------------------------------------------------
Taxonomy/phylogeny, Family, genus, species, subspecies,cultivar
life-history traits and (incl.
physiology synonyms) Related genera and species with which gene
exchange may occur
Life-form (herb, shrub, tree) *
Persistence (annual, perennial) *
Reproductive strategy (monocarpic, polycarpic)
Carbon fixation cycle (C3 and C4 plants),
nitrogen fixation *
Biogeography Centre of origin
Natural habitats
Known weediness
Genetic composition Karyotype information: Chromosome number, Ploidy
level, etc. *
Genetic diversity *
Biochemistry Toxic compounds
Pollination and Breeding system (sex distribution,
gene-transfer incompatibility, selfing)
Pollen production, quantitatively *
Pollen dispersal vector (pollination syndrome)
Pollen dispersal distances
Introgression and Hybrid detection
hybridization Hybrid vigour *
Known environmental requirements of hybrids *
Recruitment Seed production *
Seed dispersal vector
Seed dispersal distances
Seed bank *
Vegetative reproduction Clonal growth and fragmentation
---------------------------------------------------------------------------
The background information for the receiver plant on taxonomy, biogeography, phylogeny and genetic composition represent baseline knowledge which is essential for evaluation of evolutionary and genetic aspects of transgenic plants. These types of data are included in environmental biosafety reviews in most countries (OECD, 1995). We also find that genetic diversity data and data on variation of ploidy levels are important to include for assessment of risks of gene transfer and hybridisation. Information on physiology of the receiver plant, which generally has not been considered, should be provided in order to indicate plant growth potential in different stress situations (e.g., heat, drought, low access to nutrient). The life-history traits indicate some basic life-cycle determinants such as survival rate, generation time and persistence in the vegetation and also affect the structure of the vegetation. We also find it important that data on the competitive ability of the receiver plant are included as these relate to the risk of direct invasion in ecosystems. The information on pollination processes including gene-transfer and introgression gives valuable indications on the probabilities for genome dispersal, either between populations or by crossings to closely related species. This information will also be valuable when monitoring programmes are set up after acceptance at Tier 2 or 3. The main type of information for assessment of the risk of establishment and reproduction in new habitats will come from data on life-history and recruitment (e.g. reproductive strategy, seed production, seed dispersal and seed bank) and the competitive interactions in the environment. Some of these data will be available at the Tier 1 level, but additional data will be required for assessment at higher levels. When existing information is used, it is essential that exact reference to the source is provided. Cases where plant species occur as weeds in cultivated soil or as escapes in natural or seminatural habitats will indicate potential problems which may be enhanced by the transgene. It also may indicate possible spread of the gene by introgression. Determination of hybrid vigour and the associated changes in growth and fertility must also be included in the required data. The character of the genetic insert (position in genome, stability, redundancy, etc.) and the techniques used are types of information which are generally supplied by the industry and which may be of relevance to the reproductive genetics and production performance of the transgenic plant. However, for risk assessment purposes this information is less important. The use of genetic markers also pose special problems (e.g. spread of antibiotic resistance, herbicide tolerance, etc.), which will have to be addressed. Two issues in transgenic plants should be specially addressed at Tier 1 in the hierarchical approach: 1) Plants expressing toxic substances, and 2) Plants having increased tolerance to environmental stress. For these issues, reference to response mechanisms, environmental effects and test procedures are given in Tables 2 and 3, respectively. Plants expressing toxic substances Toxic compounds produced by plants may pose a risk for human health and the environment. As a basis for the assessment at Tier 1, extensive residue chemistry data of the toxic compounds must be available (see Table 2). The potential exposure of particularly vulnerable population groups, e.g. children, old people or indivivuals with allergy, and the likely effects of processing should be addressed. In particular, the potential occurrence of allergic reactions to novel proteins or other constituents of novel food should be explored. The degradation of the toxins by gastric fluid in vitro should be documented. The health hazards to humans includes both acute and chronic effects, e.g. acute poisoning, neurological changes, carcinogenic effects, reproduction failures, effects measured over full lifetime or for generations, including inherital malformations, as well as more specific effects, such as hormonal disturbances. Information and data valid for these assessments normally arise from standardised toxicological laboratory tests in vivo with mammalians, in particular rats or mice, or by use of in vitro assays. Information are also included from accidents, epidemiological observations or from further sources, such as chemical or biological monitoring in the environment or in workplaces. Similar hazards to domestic animals may be included in the evaluation. For transgenic plants producing toxins acting as pesticides (fungicides, insecticides, rodenticides or repellents) the uniform principles for evaluation and authorizations of plant protection products in the EU may be used as a guidance for settring up test programmes with criteria of acceptance of this group of plants (EC, 1997). For the assessement of the hazard potential of other toxic compounds produced by plants, than those acting as pesticides, a testing scheme is suggested which is similar to the requirements for new chemical substances. For chemical compounds, the types and amounts of the required data vary according to the amount of chemical that is expected to be marketed. For plant produced compounds which are identical to existing chemicals, the necessary data for evaluation may be available without further testing. Thus, in the EU and in the U.S.A., information about any new chemical must be submitted to by producers/importers as part of their notification of the chemical before bringing it to the market. These issues are as well addressed by the Novel Food Directive of the European Union (EU, 1997). Before assessing the risk to wildlife and aquatic organisms, it should be considered if exposure is likely. However, avian and chronic tests with relevant test species should be conducted at Tier 1 in cases of risk of exposure. Similarly, the impact on selected aquatic organisms should be conducted in cases where exposure to the aquatic environment is likely. When exposure is likely, risks to wildlife should be assessed even in cases when human or livestock exposures are not conceivable. For plants with toxic compounds which may appear in pollen, testing of the risks to pollinators should be mandatory at Tier 1. The studies should include larvae which may be more sensitive than adults. In case of genetically engineered plants producing insecticides, non-target and beneficial insects should be studied using relevant species such as parasitic wasps, ladybird beetles, or green lacewing. Such studies may also be relevant for plants producing pharmaceuticals. The toxic compounds produced by the genetically engineered plants may enter the soil environment through dead plant material and cause significant mortality to soil dwelling organisms such as earthworms and springtails. Reduced reproduction or other sublethal effects may change the structure and functioning of the soil ecosystem. Therefore, data should be provided at Tier 1 on the effects of such toxic compounds produced by genetically transformed plants. For the Tier 1 assessment of risks to the environment, information on the fate of toxins is required. The fate of plant produced toxins should be elucidated with emphasis on degradation in soil and water, interactions with soil and sediment particles, and mobility in the environment. The dissipation and behaviour of the toxic compounds should be investigated by use of guidelines modified from those developed for the ecotoxicological assessment of pesticides and industrial chemicals. Further, tests should provide sufficient data to evaluate the possibility of accumulation of residues of the active substances and of relevant metabolites, degradation and reaction products. Data should also be provided to evaluate the mobility and leaching potential of the toxic compounds and other relevant substances. Table 2. Transgenic plants expressing toxic substances(pesticides, pharmaceuticals or toxic bulk chemicals). Suggested data requirements and testing for marketing at Tier 1.
---------------------------------------------------------------------------
Subject Effects or processes Examples of tests or data
---------------------------------------------------------------------------
Human health: *Acute toxicity Acute rat test
*Allergy Animal skin test
*Mutagenicity Chronical test with rats
*Carcinogenicity Chronical test with rats
*Reproduction Chronical test with rats
*Metabolism in mammals Feeding studies with radio-
labelled toxins
Degradation in Uptake from gastric fluid In vitro test for degradation
gastric fluid
Metabolism in Uptake, translocation and Studies with radio-labelled
plants metabolism toxins
Risks to wildlife Acute and chronic effects Feeding studies with birds and
at population level mammals
Pollinators Ecotoxicity Honey bee test
Beneficial Ecotoxicity Tests with selected species
invertebrates
Soil organisms Ecotoxicity Tests with earthworms and
springtails
Bioaccumulation Octanol/water partioning
coefficient
Bioaccumulation in fish
Tests for bioaccumulation
Environmental Degradation Tests in different soils
fate Test for aerobic and anaerobic
degradation in soil
Mobility in soil Leaching in soil Laboratory soil column studies
Effects on aquatic *Toxicity to algae Growth test
species: *Toxicity to fish 14-days toxicity test
*Toxicity to Daphnia Reproduction test
---------------------------------------------------------------------------
Plants having increased tolerance to environmental stress. An increased tolerance to harsh conditions in the environment may allow plants to grow in areas where this has not been possible before. This has advantages in agricultural practice, reducing the need for e.g. artificial watering and other measures that have to be taken for sustaining agriculture in climatically extreme areas. Hence, some of the biotechnological development is currently towards plants with increased resistance to e.g. salt, drought, cold and low nutrient conditions. A general review on stress tolerance in relation to breeding is available (Blum 1988). Resistance to environmental stresses raises some new questions and environmental concerns that demands for specific procedures in risk assessment of stress tolerant transgenic plants. The case of a transgenic plant with increased tolerance to drought will be discussed and, suggestions for data and type of tests are presented. The particular hazards, which can be foreseen for a drought-tolerant plant, relate to the potential invasion and possible disruption of environmental stability in natural ecosystems, inflicting upon normal development such as successional progress. The spread and establishment of drought tolerant plants into dryland habitats with sparse but highly adapted vegetation could lead to scarcity or even extinction of the local floras. The invasibility is habitat dependent and mesic communities are often considered to be most susceptible to invasion (Rejmanek, 1989). However, great differences between regions exist. Vulnerable areas to drought- tolerant plants in particular are the Mediterranean region (S. Europe, N. Africa), California and steppe and semidessert areas in N. America, Eastern Europe, Australia and elsewhere. Of course, some of these areas have already suffered from a massive invasion of foreign species during past centuries - so essentially these problems are not new. However, the invasion of a highly productive crop species presents special problems even in the cultured area itself as the inserted trait may disperse from crop to weedy relatives by introgression. Consequently, identification of habitats prone to invasion of specific species will be needed. Futhermore, characterisation and analysis of the vulnerability of different types of habitats, following the set-up of monitoring programs will be needed. When we turn to the elements that we need to know more about, clearly the actual mechanisms controlling the drought resistance become essential to analyse before release. Much of this information will be available for most applications of cases from the industry, but additional test data may be required for a proper risk assessment based on the functional responses to drought. The basic set of data that should be required at Tier 1 in addition to the general review data is given in Table 3. These data include: plant growth measures such as growth rate and biomass allocation, morphometric measures, factors affecting the reproductive rate and plant competitive interactions. When tests are performed, conditions relevant to the stress situation should be defined by the environmental authorities. Plants modified to increased tolerance to drought will show differential allocation to organs as a way to enhance water uptake, decrease evaporation and preserve water resources. Consequently, measures which relate to plant growth and biomass allocation (e.g. root:shoot) become important. The stomatal resistance of leaves determines the water evaporation from the plant. Stomatal resistance differs according to water stress (i.e. water potential) during the day, and maximum resistance values are usually specific but differ between species and ecotypes (e.g. mesophytes and xerophytes). Some chemical compounds (e.g. proline) are known to be produced and accumulated in plant tissue after exposure to environmental stress (e.g. drought, low temperatures and salt) and thus may be used as biomarkers. Furthermore, some of these compounds may have direct toxic effects on wildlife (and humans) which have to be considered. Environmental changes selectively affect important stages in a plant's life cycle and will ultimately determine the probabilities for increase or extinction of the invasive population. The density and survival rate of seeds in the soil (seed bank) together with the seed germination rate determines the input of new plants to the ecosystem each year, which is partly influenced by environmental stress (e.g. drought and vegetation density). The specific type of seed bank survival determines the probability for long-term survival in a site. The reproductive output of the individual plant (i.e. flowering and seed production) and the consequent changes in plant population size are essential measures of the success of the genotype in the specific environment. Competitive interaction, between an invasive transgenic plant and the indigenous vegetation at the habitat, is a crucial factor determining survival and reproductive success of the transgenic in the new environment. The competitive performance of the transgenic plant in relation to reference species may be analysed in two-species tests or in multi-species environments and be compared to the performance of the unmodified plant. Changes in reproductive success of the transgenic plant will indicate possible changes in competitive ability in natural habitats as well. Table 3. Transgenic plants with altered (increased) drought resistance. Suggested data requirements and type of tests at Tier 1. Tests are performed on both the transgenic and the unmodified plant under the same conditions of environmental stress.(Reference of tests methods to Kjellsson & Simonsen, 1994).
---------------------------------------------------------------------------
Subject Functional response to stress Type of tests - examples
---------------------------------------------------------------------------
Plant growth Conservation of resources Growth-rate test
during periods of stress Total plant biomass
Morphometric Differential allocation to Mean leaf area (MLA),
measures plant organs Root:shoot ratio
Sexual versus vegetative
reproduction
Transpiration Closing of stomata, etc. Leaf water potential
Stomatal resistance
Stress-induced Production, accumulation or Proline, etc.
chemical compounds activation leading to Heat-shock proteins
protection of plant
Reproductive rate Reduced in stressful Seed germination
environments Plant survival
Flowering
Seed production
Seed survival (seed bank)
Plant-environment Increased in stressful Transplant and greenhouse
competitive environments tests: Dry, low
interactions competitive and dry, high
competitive
---------------------------------------------------------------------------
Tier 2: Advanced testing The different types of hazards which can be foreseen will to a large extent depend on the inserted trait. In many cases sufficient data will be available at the Tier 1 evaluation. If the evaluation of the Tier 1 data results in unacceptable risks, the environmental authorities may decide that the transgene plant will not be marketed. However, in cases where the information is insufficient to make a satisfactory risk evaluation leading to acceptance and possible marketing, more testing will be needed, and the evaluation proceeds to Tier 2. At Tier 2 further documentation may be needed for plants producing toxins to assess the risk to indigenous flora and fauna, or the potential for ground water pollution as shown in Table 4. To elucidate the impact on wildlife, field studies may be set up to test conceptual models about how the feeding rate on crop plants and patterns of alternative food items will influence the target and non-target population dynamics. For beneficial invertebrates, and for soil organisms field experiments should be carried out to study potential risks to arthropod communities such experiments should include the use of sweep sampling, pitfall traps, or sampling of soil cores followed by extraction of soil dwelling animals. Such sampling protocols may generate 100,000 or more individual specimens to be processed per year. Special care must be taken to quantify migration of species in and out from open test areas. If possible, the studies should include caged experimental units, or barriers for organisms living in the soil or on the soil surface. If the genetically modified plant produces toxins which are likely to leach from the soil, and thereby pose a risk to ground water contamination as revealed at the Tier 1 evaluation, lysimeter or field studies are needed to further study the mobility and the degradation kinetics under natural environmental and agricultural conditions. If the toxins show potential leaching, the plants should not be released for placing on the market. Table 4. Transgenic plants expressing toxic substances (pesticides, pharmaceuticals or toxic bulk chemicals). Suggested testing at Tier 2.
---------------------------------------------------------------------------
Subject Effects Type of tests or data
--------------------------------------------------------------------------
Wildlife Impact on populations Field studies and/or
mathematical modelling
Beneficial invertebrates Impact on populations Field studies and/or
mathematical modelling
Soil organisms Impact on populations Field studies
Mobility in soil Leaching Lysimeter studies or
field studies
---------------------------------------------------------------------------
The studies should be designed to detect any significant impact caused by the transgenic plants on wildlife, invertebrates, or soil organisms, respectively. Control experiments should be conducted using the parental plant, and criteria for acceptance of the transgenic plant should be developed based on the Tier 2 test protocols. If adverse effects are not present or the risks to single populations or communities are negligible, the engineered plant should be released for placing on the market, unless use on large, adjoining areas is anticipated. In such case, the evaluation of the plant should proceed to Tier 3. The data that should be required for risk assessment of drought resistance at Tier 2 include more advanced and time consuming tests concerning short- term invasion and effects of transgenic plants in habitats outside the field area (Table 5). Table 5. Transgenic plants with altered (increased) drought resistance. Suggested data requirements and type of tests at Tier 2. Tests are performed on both the transgenic and the unmodified plant under the same conditions of environmental stress. (Reference of tests methods to Kjellsson & Simonsen, 1994).
---------------------------------------------------------------------------
Subject Functional response Type of data and tests
- examples
---------------------------------------------------------------------------
Identification of Potentially vulnerable Biodiversity measures
invasible habitats and environments to invasion (e.g. species content
environmental conditions and, diversity index)
Vegetation cover
Soil type (texture)
Microclimate and year-
to-year variation
---------------------------------------------------------------------------
Habitat invasibility Fate of drought tolerant Mesocosm tests
plant and hybrids Transplant experiments
in the field
Modelling
---------------------------------------------------------------------------
Impact of plant Species composition, Biodiversity measures
invasion^1 population changes and (e.g. species content
vegetation structure and, diversity index)
Population census and
growth rate estimates
Vegetation cover (e.g.
LAI)
Modelling
---------------------------------------------------------------------------
^1 : Long-term monitoring is covered at Tier 3.Habitat types which are prone to invasion by the transgenes should be identified (e.g. type of vegetation and management), including the environmental conditions (e.g. summer drought, open vegetation and soil type) which make them vulnerable specifically to drought resistant plants. Extensive fate studies should be done where the performance (growth rate, yield, reproductive output) of the plants with transgenes are compared to conventional varieties. These studies should initially be made in enclosed greenhouse conditions (e.g. mesocosms with vegetation profiles) and secondly, if required, be extended to field conditions. Modelling may be employed for extrapolation purposes to decrease the need of data from slightly different environmental conditions. Initial impacts of plant invasion on vegetation structure and cover should be monitored in short- term trials (i.e. < 5 years) but needs to be specifically addressed in long-term monitoring programmes. Tier 3: In case of use on large and adjoining areas (regional scale) Whenever widely use of the transgenic plants are foreseen, climatic differences between regions need to be taken into consideration in the risk assessment procedure. We propose that these aspects are focused at the final step of the test procedure (Tier 3). The regional aspect has, however, pronounced relevance for transgenic plants producing toxins, e.g. with insecticidal properties, and for plants resistant to herbicides. Use of plants producing the same toxic compound or mixture of toxic compounds on large, adjoining areas may pose a risk for extinction of herbivore species feeding on crops, or their predators. A similar risk might arise from widespread use of herbicide resistant plant. In this case the animals feeding on the weeds and their predators may be subject to local extinction. Growing herbicide resistant plants over large areas may also result in regional elimination of a number of annual weeds, which are dependent on the agricultural treatment. The background information at Tier 1, which includes taxonomy, phylogeny, biogeography and genetics, is mandatory for evaluation of gene stability and potential hybridisation in a regional context. Special measures should be taken in centres of origin, or in regions where many closely related species exist, to reduce genetic contamination by new (inserted) traits. This should be done in order to conserve the original gene pools which are also agronomically important. Well known examples of problematic cases include: potatoes in South America, maize and squash in Middle America and rice in South East Asia (Colwell, 1994). In cases of identified risks, the area grown and the frequencies at which the trangenic plant species are grown should be restricted according to specific management plans. This implies among others that areas grown with normal and transgenic plants should be balanced, and that the extent of refuge areas for the herbivores and other interacting organisms should be taken into consideration. DISCUSSION The principle of precaution which is fundamental to risk assessment of transgenic plants is intended to assure safety. The increasingly high number of petitions for regulation of transgenic plants calls for an effective and uniform treatment. The proposed hierarchical risk assessment scheme is an attempt to meet these requirements giving a flow of information and indications on where specific test data are required. The level of sufficient information will depend on the plant, the inserted trait and the use. Generally, experimental tests of the transgenic plant should be conducted at each tier. The test conditions should be defined relative to the inserted traits. In order to obtain uniform evidence and base decisions on comparable data it is highly important that a range of test protocols should be available for obligatory testing. When data and test results are evaluated in a hierarchical test system, it is mandatory that acceptance and rejection criteria are defined for further decisions. The principle of substantial equivalence of the transgenic plant and the receiver plant should be applied but needs further development to be internationally accepted. The type of criteria will depend on type of test data (quantitative or qualitative) and level of decision and may include: statistical significant differences, decision levels decided by expert panels, etc. All information on health effects to human should be available at Tier 1. Additionally, Tier 1 comprises basic information on the receiver plant, on the transgene, and on basic ecological tests with the transgenic plant. In the former hierarchical approaches (e.g. Rissler & Melon, 1993) the Tier 1 analyses are based only on information on the parental plant and the transgene. This procedure might not result in any problems as long as the traits of the transgenic plant do not clearly separate from the parental traits. However, there may be considerable differences in traits between a transgenic plant, e.g. having obtained resistance against environmental stress, and the parental plant. As a consequence, we find it necessary to include simple tests on the ecological fitness of the transgenic plant at the basic level (Tier 1). The seed companies could run such tests in parallel to the efficacy test routinely conducted. The probability of genetic escape by pollen should be treated as a separate issue in ecological risk assessment addressed at the basic level (Tier 1). If genes through introgression are transferred to weeds or wild relatives, then the hybrids and backcross generations should be tested together with the transgenic plants following the tiered procedure. The Tier 2 evaluation is necessary to further explore results from Tier 1 which indicate effects. Further, Tier 2 testing is a prerequisite to perform a proper Tier 3 evaluation. If use of the transgenic plants on large, adjoing areas is foreseen, Tier 2 and 3 analyses of effects on non-target organisms, ecosystems, biodiversity and regionally adapted genetic variation are needed. Widespread use of transgenic plants, primarily insect and herbicide resistant plants may pose risks to non-target organisms. These risks, which will arise at the regional scale from growth of different transgenic crops on several adjacent fields, are not included in the case-by-case risk assessment of today. At present, risk assessment deals primarily with short-term effects of the inserted traits, and therefore, the tests and analyses will not always detect effects which may only occur after considerable time. Consequently, we propose that long-term effects should be looked for in monitoring programs. Transgenic plants are not yet on the market in many countries, but will be it in a few years. Therefore, today is the last chance to get base-line data of the pre-release conditions. When monitoring is done at regular intervals, information on genome dispersal, occurrence in different habitats and environmental effects caused by the transgene invaders, should be possible to obtain. As shown in Figure 1, set-up of monitoring programs and restrictions are included in the risk assessment scheme as feedback/follow-up mechanisms. Such feedback mechanisms are always needed when considering hierarchical risk assessment. In hierarchical test systems relevance and complexity i.e. "naturalness" of information can be increased but often at high cost and time consumption. Tests and analysis, therefore, will not always detect possible long-term effects. Long term effects should be addressed with the caveat that monitoring itself is costly and effects to the environment may be irreversible. Changes in the environment that are caused directly by transgenic plants or by trangenes introgressed into hybrids may, however, easily be revealed by presence of the inserted gene. However, the dilemma remains: When long-terms effects are detected, it will perhaps be too late to fully remedy the damages. A management plan can only expect to reduce some of the hazards, not fully remove the organisms which caused them. ACKNOWLEDGEMENTS
The authors want to thank Juliane Albjerg, Holger Pedersen, Jan Grundtvig Hojland and Hans Erik Svart for valuable comments on previous versions of the manuscript. An anonymous reviewer provided pertinent commments and suggestions. REFERENCES Anonymous 1994 Assessment Criteria for Determining Environmental Safety of Plants With Novel Traits. Regulatory Directive Dir94-08, Government of Canada, December 16, 1994. Baker, H. 1965 Characteristics and Modes of Origin of Weeds. Baker, H.G & Stebbins, G.L. (eds.) The Genetics of Colonizing Species. Proceedings of the First International Union of Biological Sciences Symposia on General Biology. Pp 147-171. Academic Press, N.Y. Bartsch, D., Sukopp, H., & Sukopp, U. 1993 Introduction of plants with special regard to cultigens running wild. Wohrmann, K., & Tomiuk, J. (eds.) Transgenic organisms: Risk assessment of deliberate release. Pp. 135-151 Birkhauser, Basel. Blum, A. 1988. Plant breeding for stress environments. CRC Press, Boca Raton, Florida, 223 pp. Burgiel, S. 1998 The "Biosafety Protocol" - Fourth session of the ad hoc working group. BINAS News 4(1), 5-9. http://binas.unido.org/binas/news/98issue1.html Cannon, R.J.C. 1996 Bacillus thuringiensis use in agriculture: A molecular perspective. Biol. Rev. 71, 561-636. Colwell, R.K. 1994. Potential ecological and evolutionary problems of introducing transgenic crops into the environment. Krattiger, A.F., & Rosemarin, A. (eds.) Biosafety for sustainable agriculture: Sharing biotechnology regulatory experiences of the western hemisphere. Pp 33-46. ISAAA & SEI, Ithaca. Crawley, M. 1991 Predicting the Ecological Implications of Biotechnology. Roy, D.J., Wynne, B.E. & Old, R.W. (eds.) Bioscience Society. Pp 133-150. John Wiley & Sons. Crawley, M. 1992 The Comparative Ecology of Transgenic and Conventional Crops. Casper, R. & Landsmann, J. (eds.) Proc. 2nd. Int. Symp. Biosafety Results of Field Tetsts of Genetically Modified Plants and Microorganisms. Pp. 43-52. Biol Bundesanstalt Fur Land- und Forstwirtschaft, Braunsweig. Crawley, M.J. 1995 Long term ecological impacts of the release of genetically modified organisms. CEP Pan-European conference on the potential long-term ecological impact of genetically modified organisms. Proceedings Strasbourg, 24-26 November 1993. Pp. 43-66. Council of Europe Press, Strasbourg,. Crawley, M.J., Hails, R.S., Rees, M., Kohn, D. & Buxton, J. 1993 Ecology of transgenic oilseed rape in natural habitats. Nature 363, 620-623. Dale, P.J., & Kinderlerer, J. 1995 Safety in the contained use and the environmental release of transgenic crop plants. Tzotzos, G.T. (ed.) Genetically modified organisms - a guide to biosafety. Pp. 36-63. CAB International, Wallingford. EC, 1997 Council Directive 97/57/EC of 22 September 1997 establishing Annex VI to Directive 91/414/EEC concerning the placing of plant protection products on the market. Official Journal of the European Communities, L265, Pp. 85-109 EU, 1997 Regulation on Novel Food and Novel Ingredients (EC No 258/97) on 15 May 1997 Fava, J.A., Adams, W.J., Larson, R.J., Dickson, G.W., Dickson, K.L. & Bishop, W.E. (eds.) 1987 Research Priorities in Environmental Risk Assessment. Workshop Report, August 16-21, 1987, Breckenridge, Colorado, Society of Environmental Toxicology and Chemistry (SETAC), pp. 103. Gaugitsch, H., & Torgersen, H. 1995 Streamlining regulations, keeping high safety standards: Revised criteria for the assessment of releases of genetically modified organisms (GMOs) into the environment. Ambio 24, 47-50. Gliddon, C. 1995 The potential role of monitoring in risk assessment and in fundamental ecological research. CEP Pan-European conference on the potential long-term ecological impact of genetically modified organisms. Proceeings Strasbourg, 24-26 November 1993. Pp. 253-263 Council of Europe Press, Strasbourg. Illueca, J. 1996 Speech for Trondheim meeting on invasive species. Sandlund, O.T., Schei, P.J. & Viken, A. (eds.) Proceedings Norway/UN Conference on Alien Species. The Trondheim Conferences on Biodiversity, 1-5 July 1996. Pp 13-17. Kareiva, P. 1993 Transgenic plants on trial. Nature 363, 580-581. Kareiva, P., Parker. I.M. & Pascual, M. 1996 How useful are experiments and models in predicting the invasiveness of genetically engineered organisms? Ecology 77: 1670-1675. Kjellsson, G. 1997 Principles and procedures for ecological risk assessment of transgenic plants. Kjellsson, G., Simonsen, V., & Ammann, K. (eds.) Methods for risk assessment of transgenic plants. II. Pollination, gene- transfer and population impacts. Pp. 221-236. Birkhauser Verlag, Basel. Kjellsson, G. & Simonsen, V. 1994 Methods for risk assessment of transgenic plants. I. Competition, establishment and ecosystem effects. Birkhauser Verlag, Basel, 214 pp. Kjellsson, G., Simonsen, V. & Ammann, K. (eds.) 1997 Methods for risk assessment of transgenic plants. II. Pollination, gene-transfer and population impacts. Birkhauser Verlag, Basel, 308 pp. Knudsen, I. and Ovesen, L. 1995 Assessment of novel foods: A call for a new and broader GRAS concept. Regul. Toxicol. Pharmacol. 21, 365-369. Linder, C.R. & Schmitt, J. 1995 Potential persistence of escaped transgenes: performance in transgenic, oil-modified Brassica seeds and seedlings. Ecological Applications 5, 1056-1068. Lipton, J., Galbraith, H., Burger, J. & Wartenberg, D. 1993 A Paradigm for Ecological Risk Assessment. Environ. Manage. 17, 1-5. Madsen, K.H. & Poulsen, G.S. 1997 Inserted traits for transgenic plants. Kjellsson, G., Simonsen, V., & Ammann, K. (eds.) Methods for risk assessment of transgenic plants. II. Pollination, gene-transfer and population impacts. Pp. 203-219. Birkhauser Verlag, Basel. NNT (Nordic Working Group on Food Toxicology and Risk Assessment), 1991 Food and New Biotechnology - Novelty, Safety and Control Aspects of Foods Made by New Biotechnology. Nordic Council of Ministers. Nord 1991:18. NRC, 1989 Field testing genetically modified organisms: Framework for decisions. Committee on scientific evaluation of the genetically modified microorganisms and plants into the environment. Board on biology, Commission on life sciences, National Research Council. National Academy Press, Washington, 170 pp. OECD, 1992 Safety considerations for biotechnology. OECD, Paris, 50 pp. OECD (Organization for Economic Co-operation and Development), 1993 Safety Evaluation of Foods Produced by Modern Biotechnology: Concepts and Principles. OECD, Paris. OECD, 1995. Commercialisation of agricultural products derived through modern biotechnology: survey results. OECD Environment Monograph 99. Organization for Economic Co-operation and Development, Paris, 178 pp. Parker, I.M. & Kareiva, P. 1996 Assessing the risks of invasion for genetically engineered plants: acceptable evidence and reasonable doubt. Biological Conservation 78, 193-203. Perrins, J., Williamson, M. & Fitter, A. 1992a Do annual weeds have predictable characters? Acta Oecologia 13, 517-533. Perrins, J., Williamson, M. & Fitter, A. 1992b A survey of differering views of weed classification: implications for regulation of introductions. Biological Conservation 60, 47-56. Rejmanek, M. 1989 Invasibility of Plant Communities. Drake, J.A., Mooney, H.A., diCastri, F., Groves, R.H., Kruger, F.J., Rejmanek, M. & Williamson, M. Biological Invasions: a Global Perspective. SCOPE. Pp 369-388. John Wiley & Sons Ltd. Regal, P.J. 1993 The true meaning of 'exotic species' as a model of genetically engineered organisms. Experientia 49, 225-234. Rees, M., Kohn, D., Hails, R., Crawley, M. & Malcolm, S. 1991 An Ecological Perspective to Risk Assessment. MacKenzie, D.R. & Henry, S.C. (eds.) Biological Monitoring of Genetically Engineered Plants and Microbes. Pp 9- 24. USDA and Clemson University, SC. Rissler, J. & Melon, M. 1993 Perils Amidst the Promise. Ecological Risks of Transgenic Crops in a Global Market. Union of Concerned Scientists, Cambridge, M.A. 92 pp. Stockey, A. & Hunt, R. 1994. Predicting secondary succession in wetland mesocosms on the basis of autecological information on seeds and seedlings. J. Appl. Ecol. 31, 543-559. Suter, G.W. 1993 Ecological risk assessment. Lewis Publishers, Boca Raton, 538 pp. UNEP, 1996 International technical guidelines for safety in biotechnology. United Nations Environmental Programme, Nairobi, 31 pp. van Leeuwen, C.J. 1995 General Introduction. van Leeuwen, C.J & Hermens, J.L.M. (eds.) Risk Assessment of Chemicals: An Introduction. Pp 1-17. Kluwer Academic Publishers, Dordrecht. Williamson, M. 1993 Invaders, weeds and the risk from genetically manipulated organisms. Experientia 49, 219-224. Williamson, M. 1994 Community response to transgenic plant release: predictions from British experience of invasive plants and feral crop plants. Molecular Ecology 3, 75-79. Copyright remains with the author.
Published by Bioline Publications. The following images related to this document are available:Line drawing images[by98002a.gif] |
| |||||||||
