The food safety assessment of GM animals and derived products can largely be performed along the lines that have already been established for the evaluation of GM plants and derived products for the consumer (Codex Alimentarius Commission, 2003). This means that the initial step of the food safety assessment will be a comparative safety assessment of the GM animal with its appropriate comparator, including a food intake assessment, followed by a full risk characterization.
5.2.1 Comparative safety assessment
Because of their history of safe use, it is generally acknowledged that traditional food products may serve as a baseline for comparison for the assessment of the safety of GM foods. This is generally referred to as the concept of substantial equivalence. The basic idea is that novel genetically modified organism (GMO)-derived food products should be at least as safe as the traditional products that they may replace in the diet.
Substantial equivalence is not a safety assessment in itself; rather it represents the starting-point which is used to structure the safety assessment of a GM food relative to its conventional counterpart. Since its inception, the concept has evolved (FAO/WHO, 2000). More recently, it has been suggested to replace it with the broader Comparative Safety Assessment (CSA) concept (Kok and Kuiper, 2003). The Consultation agreed that the CSA approach is equally applicable to GM animals as it is to GM plants.
The CSA is basically a two-tiered approach. The initial step comprises a thorough comparison with the closely related conventional counterpart to identify any differences that may have safety implications for the consumer. This comparison includes both phenotypic characteristics as well as a compositional analysis. The phenotypic analysis should also include comparative health parameters. The compositional analysis will focus on key substances in the animal products under scrutiny and will be subject to changes according to the latest scientific state of the art.
The second step of the CSA comprises the toxicological and nutritional evaluation of the identified differences between the GM animal and its conventional counterpart. As a result of this second step additional testing may be required and can result in an iterative process in order to obtain all relevant information for the final risk characterization.
In the initial step the information that should be considered includes:
the transformation process of the genetic modification, including the sequence of the inserted material before and after the transformation event;
the copy number and site(s) of insertion;
sequence analysis of the site(s) of insertion, i.e. flanking regions;
stability of the integration (multiple generations);
the safety of any newly expressed proteins, including assessment of allergenicity;
occurrence and implications of unintended effects;
the role of the new GM animal food in the diet; and
the potential influence of processing or spoilage on the new GM food product.
More precise criteria for the molecular characterization are dealt with in the Codex guidelines (Codex Alimentarius Commission, 2003) and are currently being discussed for further elaboration within the OECD.
In terms of classical risk assessment, the CSA approach means that for complex food products, including GM animal-derived products, it will in general be necessary to do both hazard identification and characterization as well as the food intake assessment in order to be able fully to characterize the risk.
FIGURE. Schematic overview of the risk assessment process
5.2.2 Hazard identification and characterization
Hazard identification and characterization are typically the first steps in any risk assessment. Any differences found as a result of the CSA serve as comparable to the hazard identification and hazard characterization steps in a traditional risk assessment paradigm. However, for complex GMO-derived foods, the hazard identification and characterization steps will not be as readily completed as in the case of well-characterized single chemical compounds because of the variety and magnitude of unintended effects that may occur when testing complex food products.
5.2.2.1 Molecular characterization
An extensive molecular characterization of the inserted genetic material construct will generally be required, both before and after the insertional event. The molecular characterization should furthermore comprise an analysis of the copy number and a sequence analysis of the flanking regions of the place of insertion in order to identify any unintended effects.
5.2.2.2 Safety of the gene product
The safety of the gene product must be assessed on a case-by-case basis. Depending on the knowledge of the expressed product, the assessment may range from a limited evaluation process of the available data on the protein, such as amino acid sequence and expression rates in different tissues, to, in the case of less well-documented proteins, extensive toxicity testing including animal studies. In theory, production of GM animals may lead to the introduction of many new proteins without a history of safe use into the human diet. The assessment of the novel proteins should be based on current knowledge of toxic substances, including a search for sequence homology with known toxins, and the function of the novel protein. In the case of unknown proteins, a full classic toxicological safety assessment procedure will form part of the evaluation.
In this regard, a distinction should be made between GM animals developed for food purposes and GM animals developed for pharmaceutical, xenotransplantation or industrial purposes. While all GM animals and their derived products in the latter category will not be intended for food purposes, such animals may enter the food supply either directly or indirectly and therefore their hazards should be assessed.
So far, the number of different genes that is used for the production of GM food animals is still rather limited when compared with plants, but this situation may change with the progress of genome sequencing programmes that are likely to provide a wealth of data on important animal physiological pathways.
5.2.2.3 Allergenicity
In the case of newly expressed proteins in the GM animal, the allergenic potential of the protein will need to be assessed. For the production of specific well-characterized proteins by the GM animal, it needs to be established whether post-translational modifications are comparable to the same substances being produced by more traditional sources in order to assess potential altered toxicological or allergenic properties of the newly synthesized proteins.
It has been recognized that there is no single parameter that can predict the allergenic potential of a substance. Recently, a strategy to assess allergenicity of biotechnology products has been formulated (FAO/WHO, 2001; Codex Alimentarius Commission, 2003), which relies on the following parameters: source of the gene, sequence homology, serum testing of patients known to be allergenic to the source organism or to sources distantly related, pepsin resistance, the prevalence of the trait and assessment using animal models.
The Consultation agreed that the strategies and methodologies for the allergenicity testing in GM animals will not differ fundamentally from those currently in use for the assessment of GM plants. It was recognized that animal models for allergenicity testing, even those that are not yet validated, may be of value to identify potential allergens. It is recommended that additional efforts should be directed to the further development and validation of these models.
5.2.2.4 Gene transfer
The DNA construct used to change the genetic make-up of the animal should be considered within an assessment, especially if the gene or its promoter is derived from a viral source, since horizontal transfer or recombination may occur. Additionally, bacterial host-derived materials may include additional sequence fragments unrelated to the target gene (National Research Council, 2002). Inadvertent introduction of such sequences into the germline of a GM animal not only has the potential for creating unintended genetic damage but can also contribute by recombination to the generation of novel infectious viruses. A well-known example is the generation of a replication-competent murine leukaemia virus (MLV) during the growth of a vector containing a globin gene (Purcell et al., 1996).
There is also potential for horizontal transfer of the gene construct: food-ingested foreign DNA may not be completely degraded in the gastrointestinal tract of mice and pigs (Chowdhury et al., 2003; Schubbert et al., 1997; Schubbert et al., 1998). For the food safety assessment, it is prudent to assume that DNA fragments may survive the human gastrointestinal tract and be absorbed by either the gut microflora or somatic cells lining the intestinal tract.
Assessment of the safety of the genetic construct should include marker genes. Commonly used marker genes are genes that code for antibiotic resistance. Risk assessment of these selectable genes should focus on gene transfer to micro-organisms residing in the gastro-intestinal tract of humans or animals. However, as the potential of this gene transfer cannot be completely ruled out, the safety assessment should also consider information on the role of the antibiotic in human and veterinary medical uses.
In general, the Consultation advocated avoiding the use of any unnecessary DNA sequences including marker genes in the genetic construct.
5.2.2.5 Unintended effects
Potential unintended effects represent a significant concern with GM animals and these effects highlight the difficulty of establishing generic considerations instead of case-by-case considerations. Unintended effects can be divided into insertional effects, related to the place of insertion of the transgenic fragment, and secondary effects, related to the nature of the expression products of the introduced genes. The major approach to detect any unintended side effects in the GM animal is phenotypical, including a compositional analysis to compare the new food organism with the conventional counterpart.
In general, the compositional analysis should be performed on the basis of validated scientific methods. Strategies for the compositional analysis of food products derived from GM animals will not differ fundamentally from those of plants, where key substances are identified and analysed per species. Furthermore, in order to be able to interpret the data from the compositional analysis of individual animal products adequately, insight into the natural variation in the relevant macro- and micronutrients and antinutrients, if present, will be required.
A special case is hemizygous GM animals that are intended for marketing. If the trait is inherited in a recessive way, so that it is expressed only in a recessive homozygote, it may be prudent to assess both the hemizygote as well as the homozygous animal for potential deleterious effects of the specific trait.
In future, compositional analysis may also be based on unbiased profiling of the GM food product and the conventional counterpart. Techniques for the profiling approach are now under development and can be divided into three subsections: genomics, proteomics and metabolomics to screen for differences in the GM animals in relation to the gene transcription products, proteins and metabolites, respectively. At the moment, however, none of these techniques is yet validated and ready for routine use in risk assessment.
5.2.3 Food intake assessment
Whereas traditional risk assessment uses “exposure assessment” to indicate exposure to hazards, the Consultation agreed that “food intake assessment” is a more appropriate term for the case of food products. Food intake assessment addresses complex foods and not individual chemical compounds, and their entry into the food supply can affect diets and also overall consumption patterns.
The goal of a food intake assessment is to assess the amount of food or food ingredient an individual or population group may consume. No exact criteria have been formulated so far for the factors that need to be considered in a pre-market intake assessment of a complex novel food product. Some food intake paradigms make assumptions based on per capita production while others use per capita distribution. An intake assessment may also consider the cooking and food preparation process used. Some governments have instituted tracking of animal-derived food and from this data set post-market consumption data may be determined. Food intake assessments will also include an estimate of the extent to which current food products will be replaced by the GM animal-derived novel food product. Thus, the accuracy of the intake assessment for GM animal-derived foods is dependent upon the available data on consumption patterns of consumer groups of interest and the validity of the underlying parameters. Specific consumer groups may refer to different age groups, but also to more vulnerable groups such as pregnant or lactating women or specific patient groups.
Food intake assessment will be based not only on available consumption data, but also on our knowledge of the bioavailability in the gastro-intestinal tract of specific food components under investigation. Probabilistic mathematical models for integrating food consumption and distribution may in specific cases be used in a comparative approach to estimate future intakes more precisely.
5.2.4 Integrated toxicological evaluation
Following the phase of hazard identification, hazard characterization and food intake assessment, an integrated toxicological evaluation will combine all the information in relation to the food safety of the complex GM animal-derived food product. This integrated toxicological evaluation needs to identify food safety issues that may require additional investigation, including traditional toxicity testing.
In general, it will not be possible to test complex animal products by classical toxicological animal studies in the way they are routinely used to test single compounds. Classical studies measuring physiological responses relative to dose are complicated if the laboratory animal is receiving doses of the GM animal’s edible tissue. If the genetic modification would result in the expression of novel proteins or if the compositional analysis revealed an alteration in an endogenous protein product or metabolite, the traditional toxicological approach would require the concentration of the product to be elevated in the laboratory animal’s diet to the extent that the diet will often become unbalanced. This might result in toxicological observations that are unrelated to the product under investigation. The limitations of standard toxicity testing applied to whole foods have also been discussed (Codex Alimentarius Commission, 2003).
On occasions where the genetic modification results in an increase in a specific (exogenous) protein, for instance directly derived from the gene construct, traditional testing would still be valid to assess that protein. Alternatively, there may be instances wherein endogenous protein levels in the GM food are increased well above the physiological level in the given animal species and it might be prudent in specific cases to (also) test this elevated protein in animal studies.
5.2.5 Integrated nutritional evaluation
In order to identify nutritional issues that need further investigation, an integrated nutritional evaluation will need to combine all the information related to the nutritional aspects of the complex GM animal-derived food product. This has to be done in addition to the integrated toxicological evaluation.
The nutritional analysis should focus on the potential replacement of nutritionally important food products by the novel GM animal-derived food products with possibly altered characteristics. The information for the nutritional analysis will largely be derived from the initial CSA, including the compositional analysis (especially macro-, micro- and antinutrients) and the estimated consumption rates. Detected alterations in the GM animal-derived food products compared with the traditional counterpart will be assessed by evaluating the significance of the compositional differences for the consumer in general and also, in specific cases, for specific consumer groups. Nutritional aspects of GMO-derived foods may become of increasing significance when the number of compositionally altered food products on the market increases. Therefore, the nutritional assessment of GM animal-derived food products is dependent on current consumption data of animal-derived food products in distinctive consumer groups and with respect to geographical and demographical differences. Special consumer groups perhaps worthy of consideration include children, pregnant or lactating women, elderly persons and the immuno-compromised.
Micronutrients are vitamins and minerals that are essential for normal physiology and biochemical functioning. Both deficiency and excess of a micronutrient can cause health problems which emphasizes the importance of this class of compounds. Macronutrients include dietary lipids, proteins and carbohydrates and these classes of compounds are present in the food and diet in substantial quantities. Assessment of the replacement factor of important animal-derived sources of micro- and macronutrients by GM animal products in the event of altered composition with relation to these nutrients is therefore of major importance. Bioavailability of the important micro- and macronutrients from GM animal-derived tissues is also of significant importance in this respect.
5.2.6 Risk characterization
Risk characterization is the final step of the risk assessment process and involves integrating the outcomes from the full toxicological and nutritional evaluations in order to reach an overall conclusion about the safety of the food.
The baseline for the safety of novel food products derived from GMOs, including GM animals, in all cases will have to be the assessment that the novel GM animal-derived food products is at least as safe as its conventional counterpart. If any questions remain after the initial CSA with respect to the safety of the GM animal-derived food products, additional tests may be required, including animal studies with the whole product or selected tissues/extracts. If, after a full safety assessment, the safety standard, i.e. as safe as its conventional counterpart, cannot be satisfied, the GM animal-derived product should not be approved for marketing. This risk characterization should be established on a case-by-case basis for food products derived from GM food animals.
In general, potential safety issues should be addressed adequately through a rigorous pre-market assessment, as the feasibility of post-market studies is very limited at present.
However, post-market surveillance may be an appropriate risk management measure in specific circumstances. Its need and utility should be considered, on a case-by-case basis, during risk assessment and its practicability should be considered during risk management.
The use of post-market surveillance as an instrument to gain information on the potential long-term or unexpected adverse and beneficial effects of food, either GM animal-derived or traditional, should be further explored. Post-market surveillance could be useful in certain instances where clear-cut questions require, for instance, a better estimate of food intake and/or nutritional consequence of a GM animal-derived food.
For GM animal-derived medicinal substances, existing pharmacovigilance schemes will apply to monitor any unforeseen unintended side-effects of the isolated medicinal substances. The same would apply in a veterinary sense with respect to the GM animal itself when modified with respect to the production of hormonal or disease-prevention substances: pharmacovigilance schemes could help to detect unintended side-effects of the introduced expression product to the GM food animal that were not detected in the pre-market phase. To this end, the GM animals should then be included in such pharmacovigilance schemes on the basis of “internal” administration of the specific veterinary substance.
It is important to note that product tracing and related control systems may be less straightforward in the case of chimeric organisms as different parts of the food animal will have different genetic constitutions and this may severely complicate analytical control of product tracing systems.
Depending on the questions and risk management needs underlying the establishment of post-market surveillance systems the product information conveyed to the consumer may, however, require adjustment. In order to enable consumers to relate potential adverse, e.g. allergenic, effects to a GM animal-derived food product, it may be necessary not only to label the product as GM animal-derived, but also to provide information on the specific GM animal source, for instance by including on the label the unique identifier code specific for a single integration event.
For the establishment of adequate product tracing systems and in order to distinguish safety evaluated products from unevaluated ones, it will be necessary to have the information on the sequence of the insert and on the flanking regions as well as reference materials of the transgenic animal and its conventional counterpart. It is recommended to use a relatively small size sequence for detection which is expected to be detectable in all kinds of products, including processed products.
