Annotated Bibliography Entry

Reference. Belgium VIB 2001
Title: Safety of Genetically Engineered Crops, March 2001
Authors: Custers, R. (ed) 
Publisher: VIB, Flanders Interuniversity Institute for Biotechnology, Rijvisschestraat 120, B-9052, Zwijnaarde, Belgium
Publication details: March 2001. 160p. 

Summary
          Issue 1: Food Safety and Toxicity
         Issue 2: Vertical Gene Flow
         Issue 3: Effects on Non-target Organisms 
         Issue 4: Allergenicity of Foods derived from Genetically Modified Organisms
         Issue 5: Horizontal Gene Flow
Conclusion
Table of Contents

 

Back to Belgium page

Summary

The VIB report discusses how the risk assessment of genetically modified (GM) crops works and on what data the conclusions of authorities are based.  The report also discusses the issues for the safety assessment of future GM crops.  The VIB report discusses five safety issues: (1) food safety (toxicity), (2) vertical gene flow (outcrossing, superweeds), (3) unwanted effects on non-target organisms, (4) allergenicity, and (5) horizontal gene flow. For each issue, an overview is presented of the data and information used in the risk assessment of GM crops.  

Back to top

Issue 1.  Food Safety and Toxicity

Food, toxicity and genetic engineering
Man has learned by trial and error to avoid poisonous plants. But our current food is still not without toxic substances. Some plants even have to be cooked before they can be eaten safely. Traditional breeding can lead and has led to changes in the levels of these substances leading to their present levels. There is no formal food safety assessment required for these traditional foods because we have a ‘history of safe use’ with these types of foods. This history of safe use is however, not backed up by empirical evidence. Like traditional breeding, genetic engineering has the potential to alter the toxicity of foods. A newly introduced protein can be toxic, but also random changes could occur resulting from insertion or pleiotropy, leading to changes that are relevant from a toxicological perspective. The phenomenon of insertion is not new, because it is also known from some traditional breeding techniques. In contrast to traditional crops, for genetically engineered crops we do require a formal food safety assessment, and for the reason that we do not have a history of safe use of these genetically engineered foods.

Genetically engineered crops and a new paradigm in food safety assessment
The food safety assessment of genetically engineered foods should determine whether the modified food is as safe as its traditional counterpart. As a starting point for the safety assessment the concept of ‘substantial equivalence’ was introduced as a means of establishing a benchmark of safe food.

The food safety assessment of genetically engineered foods is now considered to consist of four parts: (1) a molecular characterization of the insert, (2) determination of any unwanted direct toxicological effects as can be predicted from the nature of the inserted sequences, (3) determination of any unwanted indirect toxicological consequences resulting from the modification, and (4) a morphological and behavioral analysis of the plant under relevant field conditions. The introduction of modified foods has led to a shift in the food safety assessment towards a greater need for whole food safety assessment.

Direct toxicological effects as can be predicted from the transgene
Standardized in-vitro and in-vivo toxicity tests are used to test the direct toxicity of the products introduced by genetic engineering. In addition, homology searches in databases of known toxicants can be helpful in the safety assessment. The approaches used by companies to determine the toxicity of the introduced products have differed somewhat. Not exactly the same analyses have been performed in all cases.  But in all cases the data provided for the crops currently on the market, were accepted by the authorities and considered adequate. The approach of assessing the direct toxicity as can be predicted from the nature of the transgene can differ depending on the type of modification. The simplest is when only one gene product is added, and no interactions with other components in the plant are expected. The situation becomes more complicated when multiple genes are involved and when the modification results in changes in one or more pathways. Knowledge on the involved gene products and the eventual involved pathways are then necessary to design the proper analysis and tests.

Substantial equivalence
To determine whether there are unwanted indirect effects resulting from the modification, the concept of substantial equivalence is used as a principle to guide the assessment. To test for such effects it has been determined that a comparison should be made between the modified crop and its non-modified counterpart. This is a rather indirect method looking only for symptoms. The comparison looks at substances that are relevant from a toxicological, nutritional, or wholesomeness point of view. There are no standardized lists yet of what substances to compare. For the crops currently on the market, such analyses have been made and showed no significant changes. But, if significant changes would be found, this would indicate that an unwanted effect has taken place resulting either from (1) pleiotropy, (2) insertion, or (3) somaclonal variation. The finding of such a significant change triggers further analysis to find the actual cause, but this may prove difficult. Probably, in not all cases an explanation will be found. In future perhaps new methods (proteomics, metabolomics, micro-array) will be available that are more able to make the comparison between the modified and the non-modified foods.

Improving the food safety assessment
To further improve the food safety assessment a proposal is made for an overall safety testing procedure.  This procedure starting from the four points mentioned above should give a more explicit overall approach to the analysis. It is a two-step procedure leading to a judgment of the wholesomeness of the modified food. It should also make more explicit in what cases a history of safe use is enough to establish safety.

Back to top

Issue 2:  Vertical Gene Flow

Safety of Genetically Engineered Plants: an Ecological Risk Assessment of Vertical Gene Flow

Gene flow and hybridisation
Sexual reproduction among crops, weeds and wild plants is possible on the conditions that (1) the crop and the weed or the wild relative are within a distance that pollination can occur, and (2) the plants are sexually compatible. Sexual compatibility is hampered by the existence of external and internal reproductive isolation barriers that determine whether or not viable hybrids are formed. In practice this means that for a number of crops like maize, potato and tomato there are no wild relatives in Europe with which successful hybridisation could occur. But it should be realized that genetic engineering, like classical breeding, has the potential to change reproductive isolation barriers. The experience with transgenic plants like oilseed rape shows that gene flow from these plants does take place as predicted from the knowledge on the sexual reproduction of these plants. Even though isolation barriers can limit the chances of hybridisation, in many cases chances of hybridisation are not zero and for risk assessment purposes the chances then have to be considered to be one.

Codes to help risk assessment
Gene flow indices (Dpdf) consisting of the three factors: (1) dispersal of pollen (Dp), (2) dispersal of diasporas (Dd), and (3) frequency of distribution (Df), can help in the risk assessment of genetically engineered crops, especially in the case of field trials. Through the combination of the three codes a risk classification of crops in a certain region can be made. The higher the risk class, the more stringent safety measures will have to be taken to prevent outcrossing. What currently is missing is a sort of risk factor for genes: a measure for the hazard posed by a particular gene. If such a factor would exist it would truly help risk assessment. It is however difficult to foresee such a measure because a gene may be hazardous in one plant but harmless in another.

Weediness
Genetic engineering, like conventional breeding, is able to alter characteristics of crops resulting in a crop or its wild relative becoming more weedy. In many crops, especially the ones that have been domesticated to such an extent that they are no longer able to compete with wild species in natural habitats, the addition of a few genes is very unlikely to turn the crop into a weed. In crops that are still very close to their wild and sometimes weedy variants, the addition of one gene might be enough to trigger weediness. Herbicide tolerance is not a major concern from the viewpoint of weediness. Only during application of the herbicide is there a selective advantage for the plants possessing this characteristic. However, to prevent acceleration of the selection of tolerant weeds and possible future problems with (herbicide tolerant) volunteers it will be important to apply the necessary crop- and herbicide rotation.

Selective advantage and improved fitness
If a trait provides a selective advantage this means that after the outcrossing of the trait to a wild relative it has a good chance of being accumulated in the wild population. Selection results in the trait being preferentially attained. The fitness of a plant is commonly measured by the number of successful offspring compared to the offspring of other plants and it is always defined in relation to environmental variables. The concepts of selective advantage and fitness are used in risk assessment to help to determine the risks associated with a transgene. Traits related to the success of gene flow, resistance to biotic or abiotic stress might result in selective advantages or serious fitness improvements, if the absence of the trait in nature is an important determing factor in the existing ecological balances. Whether there is a real risk in specific cases can only be determined through thoughtful analysis and experimentation. It should be kept in mind that many of the above mentioned types of traits can also be obtained through classical breeding methods and that their possible ecological effects should be assessed in the same way. Experience from conventional breeding and the introduction of exotic genes and genomes forms the proof of the fact that the introduction of new varieties and new crops has the potential to influence the natural flora. However, reports of fitness advancement for hybrids in natural ecosystems are rare. Conversely, gene flow from domesticated crops has mostly made the wild plants less competitive. The experience from nontransgenic practice directs us to take great care.

Ecological view and consequences for risk assessment
For the future risk assessment a rational stepwise approach is necessary, taking into account the knowledge of the crop and its wild relatives, knowledge on the biogeographical situation, and knowledge on the transgene. The testing of the transgenic crop should follow a step-by-step procedure evaluating data of a first step before stepping into a next phase. The more risky the crop and/or the transgene, the more stringent the testing scheme will have to be before the transgenic crop can be allowed to be grown commercially on a large scale. But in the end one dilemma will remain: even after the most careful risk assessment process, only a mass release will bring all effects to the surface. The small-scale field trials do not allow to investigate the ecological risks of widespread commercialization. Therefore in order to achieve sustainability in cultivating

Back to top

 

Issue 3.  Effects on Non-target Organisms

Effects on Non-target Organisms of the Release of Genetically Modified Crops into the Environment

Non-target effects
Non-target effects are defined as the unwanted (negative) effects of crops or their accompanying farming practice on organisms living in or around the agricultural field that are not intended to be hurt. Although the topic here is the impact of transgenic organisms on non-target organisms, the impact of new modified varieties produced using traditional breeding may be as significant, if not more so. Non-target effects could arise from a number of factors: (1) the gene product affects organisms for which the plant was not designed, (2) there may be effects as a result of changed agricultural practice. Farming, especially in Europe, is known to have profound effects on the environment since large portions of the countryside are farmed. The intensification of agriculture has led to changes in the farmed environment. The use of chemicals also has had a clear effect on the organisms living in the farmed environment. There are for instance examples known of the use of certain chemicals leading to a decline in bird life.

Genetically modified crops
Only a limited amount of genetically modified traits have reached the market and are now being grown on a large scale. Non-target effects are not limited to these transgenic plants, but they may exacerbate the problem as they are specifically designed to modify farming practice. Changes could be both positive or negative. Changes can also be transferred to the natural environment if the transgenic trait is transported to wild relatives through cross-pollination. An impact analysis for examining the likely effects on non-target organisms should consider: (1) those species reliant on the crop itself, whether through using it for food or shelter, (2) those plants and animals that live within the field and might be damaged if changes are made to the crop that modifies their habitat or their ability to survive, (3) plants and animals living in the field margin or hedges and walls, if the management of the crop modifies the size, extent or susceptibility to herbicides and pesticides of this field area, (4)  soil and soil organisms may be affected by the changes in plant variety or management. Here, we do not give specific recommendations of what species to consider, but in determining them, the following factors are relevant: (1) how far should one go in testing the effects on non-target organisms? Should we test all organisms, or only a limited number that is considered useful or relevant from a biodiversity point of view? (2) Specific species and their number may differ from region to region, and  (3) Tests performed should be relevant and validated.

Insect resistant crops versus insecticides: non-target effects
Insecticides can have a significant effect on non-target organisms depending on the insecticide, the time and way of application. Not all insecticides are as selective as one would hope for. Bt crops are the most well known genetically modified insecticidal crops. Bt is highly selective against certain classes of insects. The action of one particular Bt protein can be against one or a few insects and the possibility exists that the Bt crops hits sensitive insects that live in or around the crop. An intensive discussion has emerged after it was shown that larvae of the Monarch butterfly were sensitive to the Bt protein present in Bt maize. A number of laboratory and field studies have been performed to test whether the Bt protein would actually pose a serious risk to the Monarch. The results of these studies point towards a limited risk.  At this moment there is no comparison made between the non-target effects of the modified versus the nonmodified practice. It will be difficult to do so, but it would probably be a fair way of weighing the risks versus the benefits. When comparing the two practices the following may be relevant to consider for both practices: (1) the insecticides used, (2) its type of use, (3) direct toxicity to relevant organisms, (4) total amount of active ingredient, (5) type of exposure to non-target organisms, (6) biodegradability, and (7) effectiveness.

Non-target effects of herbicide tolerant plants
Herbicide tolerant plants may have an effect on the type and number of weeds present in a field. This shift could have an effect on non-target organisms. On the other hand the genetically modified practice replaces a practice in which often more and different herbicides are used. In that case the number of weeds may be greater, but on the other hand the herbicides themselves could have an unwanted effect. Like for insecttolerant plants the comparison between the two practices is not explicitly made.

Other modifications and their effects on non-target organisms 
In the future many other types of modifications will be introduced ranging from plants with changed nutritional characters to plants that resist abiotic stress like drought. All of these crops can have an impact on non-target organisms and they should be considered in the manner proposed above. Especially the introduction of a genetically engineered crop into a region where this crop was never able to grow before, like the introduction of a conventionally bred crop into a new habitat, will have an impact on the organisms livingthere. The experience with the current genetically engineered crops, engineered to replace conventional agricultural practices, has not revealed indications of dramatic effects on non-target organisms of the growing of these new crops.

Back to top

Issue 4. Allergenicity of Foods derived from Genetically Modified Organisms

Food allergy and food allergens  
Food allergy can be defined as a food hypersensitivity – an adverse reaction to food – in which the reactions are primarily immunologically mediated. A sensitizing food or food constituent, when eaten, is partly degraded, absorbed and triggers a reaction in the immune system. IgE antibodies play an important role. Food allergy associated clinical symptoms involve the oral cavity, the gastrointestinal tract, the skin, the respiratory tract and the circulatory system. The symptoms may range from mild to very severe anaphylactic shock. Food allergy is present in about 1.5-5% of the general population, corresponding to 8-10% in the paediatric population, and around 1% in adults. Many children ‘outgrow’ their allergies. Most (major) food allergens have a number of features in common: they are glycoproteins with a molecular weight between 10-70 kD, are often relatively stable to acid- and heat-treatment, and relatively resistant to digestive breakdown. Of the worldwide documented food allergies over 90% are caused by 8 food or food groups: peanuts, milk, eggs, soybean, tree nuts, fish, crustacea and wheat.

Allergenic potential of foods derived from genetically modified organisms 
Since only a limited number of traditional foods or food groups are known to cause allergies there is, from a scientific viewpoint, no serious indication to expect that new transgenic foods will more frequently result in allergies, but there still may be changes with allergenic consequences that have to be considered in a safety assessment. In case new genes are added to a food crop, the source of the gene is of ultimate importance, for it is possible that genes coding for allergens could be transferred. If a source has an unknown history of safe use in food production one has to be cautious as well. Also random changes could occur that may have consequences that are relevant from the viewpoint of allergenicity. On the other hand, genetic engineering can also be used to eliminate allergens from foods. 

Assessment of the potential allergenicity of (new) food proteins
Current strategies to assess the potential allergenicity of genetically engineered foods are based on a case-by-case screening using a decision tree as suggested by IFBC/ILSI. The strategy starts from looking at the source of a particular gene. In the approach different immunochemical, in vivo, and physico-chemical analyses are used to assess the potential allergenicity. What tests are actually performed depends on the source of the genes involved and the results of particular tests. The approach has a rather good positive predictive value in case of (known) allergens for which sera of allergic patients exist. Its positive and negative predictive value in case of potential allergens for which no sera of allergic patients exist is less good, but it is generally considered to result in a reasonable certainty of no evidence of allergenicity. Questions with regard to its positive and negative predictive value are based upon concerns regarding some statistical aspects in cases where sera of allergic patients are very difficult to obtain, and on the fact that the physico-chemical analysis in the ‘right hand arm’ of the decision tree is mainly based upon some general characteristics that most major allergens have in common. The fact that there are exemptions in both ways (there are proteins that possess these characteristics that are not allergens, and there are proteins that do not possess these characteristics that are allergens) undermines its predictive value. This is why recently an FAO/WHO expert consultation has proposed alterations to the original decision tree to take into account the latest knowledge and techniques. Important alterations are that the endpoints of the assessment now is an estimate of the likelihood of allergenicity and that additional targeted serum screening and animal models have been taken up in the assessment. Whether the alterations indeed lead to a greater positive and negative predictive value will depend on the actual performance of the animal models.

From assessment of potential allergenicity towards allergy risk assessment 
Another concern with the current approach is that it is focused mainly on the allergenic potency. It does only in a minimal sense take into account that there are differences between minor and major allergens, in some cases there even do not have to be relevant risks on health effects. Also, even though a certain protein is found to have allergic potency, it does not have to result in the sensitization of consumers as a result of too low concentrations in the food or factors in the food processing. From the current approach mainly based on hazard identification it is therefore proposed to go towards a real risk assessment and risk management approach. In such an approach data on the allergenic potency relative to known allergenic potencies of existing allergens and on the expected exposure would be included. In such an approach perhaps a distinction could be made between minor and manageable risks and major risks that could better be avoided.

Back to top

Issue 5.  Horizontal Gene Flow

The Biosafety of Antibiotic Resistance Markers in Plant Transformation and the Dissemination of Genes through Horizontal Gene Flow

Microbes and genes 
Microbes and especially bacteria are an integral part of human life. For instance one gram of human faeces contains approximately 100 billion bacteria of more than 50 genera. Bacteria, but also bacterial genes are present everywhere in our environment. Also antibiotic resistance genes are present in large amounts. These genes are required for the self-protection of microbes that produce antibiotic substances themselves.  Antibiotic resistance genes have become key tools in genetic engineering to enable researchers to select transformed clones that have taken up the fragment of DNA under study. This fragment and the antibiotic resistance marker gene are physically linked for this purpose.

The antibiotic resistance marker genes used in plant genetic engineering
Also in plant genetic engineering antibiotic resistance marker genes are widely used to be able to select the plants that have taken up the desired DNA fragment. The most important ones that are used are bla TEM1 (ampicillin and amoxicillin resistance), aad (streptomycin/spectinomycin resistance), npt-II (kanamycin/neomycin resistance), npt-III (kanamycin/neomycin/amikacin resistance), hpt (hygromycin resistance), and cat (chloramphenicol resistance). The npt-II marker gene is the one that is mostly used in the plants on the market and in field trials, but also bla TEM1 appears in a marketed crop.

The use of antibiotics versus natural antibiotic resistance 
Kanamycin and neomycin are rather toxic antibiotics that are of no clinical importance. Ampicillin and amoxycillin are widely used in human and animal chemotherapy but their use is declining. Resistance to all these four antibiotics is widespread in nature. The spread of the bla TEM1 resistance gene through horizontal gene transfer to different species of bacteria is well documented. The antibiotic resistance genes are present in large resistance reservoirs from which the gene can be easily picked up by bacteria when there is a need to do so.

Transgenic plants as a potential contributor to the spread of (antibiotic resistance) genes
There is no dedicated mechanism of horizontal gene transfer from plants to bacteria. For bacteria to be able to pick up a plant gene a number of requirements have to be met: (1) plant cells (i.e. of decaying plants) should release DNA fragment in the environment of at least the size of an average gene, (2) the DNA should persist in the environment for a longer period of time, (3) bacteria should be able to take up the released DNA, (4) the DNA should be stably established in the recipient cells, and (5) the establishment should at least be neutral so that the transformed cells are not counter-selected. The uptake and expression of plant DNA by bacteria is therefore a multiple step process in which each step can be a limiting factor, making it a very improbable event. It has never been found under natural conditions. Calculations show that the number of bla TEM1 antibiotic resistance genes present in soil is larger than the number of genes present in for instance transgenic maize that contain these genes. The genes in the soil are also present on highly transmissible plasmids, where in the plants they are very much less transmissible. It can be concluded that it is many factors more likely that a bacterium acquires an npt-II, bla TEM1, or hpt gene from a plasmid present in the different resistance reservoirs than from a transgenic plant.

Horizontal transfer of DNA ingested by mammals
Mammals daily eat DNA in large amounts and there is large history of safe intake of DNA. Nonetheless one has tried to determine the fate of DNA in the intestinal tract. M13 DNA was fed to mice to test this. No M13 phages could be found in the faeces of the treated animals, but surprisingly M13 DNA fragments could be found in the faeces, in the blood stream and even integrated in some mice cells including cells of the foetuses borne by the pregnant female mice. There still are many questions surrounding this finding and no further results have been reported. It opens up the question of the potential mutagenic role of ingested DNA.

Back to top

Conclusion

What is the debate on genetically engineered crops all about? The issues that are subject of the debate on genetically engineered crops can be divided into the following categories:

1. Biological questions

Are genetically engineered crops dangerous for human health? What about food safety? Are there risks for our environment and the organisms living in our countryside? Do they pose a threat to biological diversity? How can they help attain sustainability? All these questions are repeatedly asked and perhaps they are not adequately answered or perhaps also new questions keep arising as the technology develops further and is applied to an ever increasing scale.

2. Societal questions

What do genetically engineered crops mean for the power- and dependence relationships in the food chain? What about patents and ownership? How important is freedom of choice and how can it be guaranteed? What do genetically engineered crops mean for (our relationship with) the developing countries? What about the economics of genetically engineered crops?

Underlying the two debates described above there are questions relating to the direction in which our society wants to develop. What are the goals our society strives for? What are the problems that we face today in our agriculture and environment and how can these problems be solved or circumvented? What developments or technologies can seriously contribute to the solving of these problems? And what criteria do we use to evaluate and weigh these different developments against each other? Often these goals and criteria are not made explicit, but implicitly they play an important role in the debate, because they determine the attitudes of different stakeholders in the debate.

Back to top

 

Table of Contents   

A. Introduction 4
1. Setting the scene 4
2. What is the debate on genetically engineered crops all about? 4
3. What does this report do and what does it not do? 5  

B. Some information on breeding, genetic engineering and agricultural practices 6
1. Conventional breeding and selection 6
2. Genetic engineering of plants 7
3. Changes in plants obtained by genetic engineering or by conventional 8
4. Differences between conventional breeding and genetic 11
5. Environmental effects of current agricultural practices 12

C. Summary 13

D. Discussion 20

E. Safety aspects of genetically engineered crops 26
1.Toxicity and Food Safety of Genetically Engineered Crops 27
Jan Pedersen, Folmer D. Eriksen, Ib Knudsen

2.Safety of Genetically Engineered Plants: an Ecological Risk Assessment of Vertical Gene Flow 60
Klaus Ammann, Yolande Jacot, Pia Rufener Al Mazyad  

3. Effects on Non-target Organisms of the Release of Genetically Modified Crops into the Environment 88
Julian Kinderlerer

4. Allergenicity of Foods derived from Genetically Modified Organisms 108
André Penninks, Leon Knippels, Geert Houben  

5. The Biosafety of Antibiotic Resistance Markers in Plant Transformation and the Dissemination of Genes through Horizontal Gene Flow 135
Philippe Gay

Annex 1: Overview of referees 160

 

Back to top

Web site: http://www.vib.be/biotechnology/more_biosafety.htm