4. State of the art in GM animal production

4.1 Introduction

A safety assessment of foods derived from GM animals must be grounded in an understanding of transgenesis methods, anticipated applications and possible outcomes from transgene integration and expression. Against this background, the Consultation considered the potential hazards of transgene expression on the animal and on human health-related environmental issues. The Consultation also discussed future perspectives on development, use and oversight of transgenesis for the purposes of food animal production.

The Consultation also noted that genetically modified insects are being produced, although not currently for food production. Although issues posed by genetically modified insects need discussion, they were beyond the scope of this Expert Consultation.

4.2 Techniques and applications

4.2.1 Techniques

A number of techniques can be utilized for transferring genes into animals (Houdebine, 2003). They differ in their suitability for different classes of animals, in their efficiency of transformation, and in their implications for risk assessment.

Utilization of the gene transfer approach depends upon knowledge of a gene encoding a product conferring a trait of interest. The gene to be transferred is incorporated into an expression vector that also contains genetic elements to control its expression. The use of different types of expression vectors poses methodological advantages for different classes of animals, and also affects the likelihood of subsequent genetic or immunological hazards being realized. Biotechnologists may purposely transfer into a host a:

• Fusion gene - a gene encoding a product of interest with an element that will regulate its expression in the host.

• Transposon - a DNA element capable of excising itself from one location in the genome and inserting itself into another location, which has been modified to contain the fusion gene.

• Retrovirus - a virus that can integrate itself into the genome and become expressed through the host cell’s replicative processes, and that has been modified to contain the fusion gene.

Many expression vectors contain marker genes. Some marker genes are simply reporters for successful gene transfer, while others encode gene products so that transgenic individuals can be selected for, e.g. by application of antibiotics.

Common methods for introducing an expression vector into the host include:

• Microinjection - direct injection of the expression vector into fertilized eggs or host cells using a fine glass needle.

• Electroporation - introduction of the expression vector into fertilized eggs or host cells by application of pulses of electricity to induce transient pores in the membrane of host cells.

• Particle bombardment - coating the expression vector on to gold particles, and introduction in host cells by bombardment with the particles.

• Cell transformation, followed by cloning - since it is more straightforward to add or knock out genes for cultured cells than for fertilized eggs, nuclei from successfully transformed cells can be transferred into enucleated eggs and implanted into surrogate dams to generate somatic cell cloned animals, which are also transgenic.

• Transformation of gametes - genes may be introduced into oocytes or spermatocytes, and the transformed gametes used for fertilization, generating a whole animal.

Application of any one of these methods will result in the successful transformation of a small percentage of the animals so produced. Transgenic individuals can then be identified and bred to develop a transgenic line.

4.2.2 Applications and their potential benefits

Transgenic animals expressing one introduced gene have been or might be developed for a variety of applications, posing a range of possible benefits to food production or human health (Table 1). Such animals are at various stages of development. Early applications for approval of transgenic animals for food production involve several species of fishes expressing introduced growth hormone genes.

Production of transgenic agricultural mammals is challenging and expensive, especially because of their low reproductive rate and internal fertilization and development. Many transgenic founder individuals are mosaic for the transgene, i.e. they have it in some, but not all, of their cells. For these reasons, development of transgenic agricultural mammals has lagged. However, nuclei from transformed cultured cells or transformed cells from a mosaic animal can be used as donor material for somatic cell nuclear transfer-mediated cloning, producing individuals that are transgenic in all cells. This approach might eventually be used to produce transgenic lines for food production, as is already being applied for development of pigs intended for xenotransplantation, where several transgenes will have to be expressed and several host genes knocked out.

4.3 Food-related hazards associated with producing GM animals

The discussion below addresses hazards associated with transgenesis methods and the environmental release of GM animals that have a bearing on food safety. The possible hazards of transgenesis discussed below must be placed into perspective by considering their relative likelihood and the degree of harm they pose. We note that these hazards are not unique to transgenesis.

4.3.1 Transgenics

Introduction of a transgene into an animal is not a precisely controlled process, and can result in a variety of outcomes regarding integration, expression and stability of the transgene in the host.

The desired outcome generally is stable integration of a single copy of the transgene into a single location in the genome, and not in a functional gene or a regulatory element. However, other outcomes are frequently observed, including integration of multiple copies of the transgene at one locus or insertion of the transgene at multiple locations in the genome. Insertion of the transgene into a host gene may turn the host gene off, sometimes affecting the viability or health of the host. Insertion of a transgene sometimes can affect expression of another gene(s). A transgene may become rearranged before integration, thereby becoming non-functional. During the process of transgenesis, undesired DNA sequences may become inserted into the genome, such as marker genes or selectable markers from the expression vector or contaminating bacterial DNA left over from vector production. Hazards stemming from insertional events or genetic instability can be identified by screening and managed by culling individuals that have undesired events during the course of development of the transgenic line.

Expression of the transgene ideally should have no undesired effects on the expression of other host genes or health of the host. Other outcomes, however, have been observed. The transgene can be silenced by methylation or through other mechanisms. Because expression of the transgene often is controlled by novel regulatory elements outside the host’s normal homeostatic feedback mechanisms, expression of the transgene can have pleiotropic effects, that is, effects upon multiple traits of the host. Notable pleiotropies have been observed among animals expressing introduced growth hormone genes, and have included pigs, sheep and fish exhibiting a range of morphological or metabolic abnormalities. Other pleiotropies, such as increased carcass yield, may be positive. Ectopic expression of the transgene may occur in tissues, sexes or life stages where it is not expected, and may affect the health of the host and the safety of its food products. Hazards stemming from transgene expression can be identified by screening and managed by culling individuals with undesired expression phenotypes during the course of development of the transgenic line.

The use of viral and transposon vectors poses the hazard that the transgene might subsequently move within the genome. Work with Drosophila suggests that transposons may have a greater probability of movement following crossing into a new background strain. Even though the vectors were engineered to lack all the DNA sequences needed to be packaged into virions or to transpose, there is a theoretical possibility that they could recombine with other DNA sequences in the genome, such as endogenous transposable elements, or with exogenous viruses or transposons, thereby gaining infectivity or mobility. Development and use of well-designed vectors will reduce the likelihood of these hazards.

The development of pigs for xenotransplantation involves knocking out expression of molecules that elicit immune response in humans and adding molecules that make the surface of pig cells more like that of human cells. This raises the possibility that pigs might become more susceptible to human viruses. This could provide an alternative host for spread of human disease, and could also give rise to a new evolutionary pathway for adaptation of pig viruses to humans. This hazard could theoretically be minimized by using pig breeds lacking known endoviruses for development of xenotransplantation lines and by maintaining such lines in strict quarantine.

4.3.2 Cloning

Cloning may be used to propagate GM animals and raises its own issues. The Consultation, however, did not address the risks associated with cloning per se (especially somatic cell nuclear transfer).

Somatic cell nuclear transfer-mediated cloning requires reprogramming of the genome from a differentiated cell to allow it to drive embryogenesis. This results in some degree of altered gene expression, especially early in the life of the cloned individual. The effects of altered gene expression and of reproductive manipulations needed for the cloning process may result in high rates of prenatal and postnatal mortality and of morphological or physiological abnormalities in cloned individuals, which may in turn affect animal health and food safety (National Research Council, 2002). Observation of the limited numbers of offspring of cloned animals produced to date suggests that they may be phenotypically normal.

4.3.3 Environmental considerations that can affect food safety

Different GM animals pose different potential environmental benefits and risks (National Research Council, 2002; Pew Initiative on Food and Biotechnology, 2003; Scientists’ Working Group on Biosafety, 1998). This discussion does not address all environmental issues but, rather, focuses on the connection between environmental entry of GM animals and food safety. The potential spread of GM animals or their transgenes in the environment is an environmental hazard that provides a route for entering into the human food supply.

The potential entry of GM animals into the food supply via the environment will vary owing to different predispositions of the animals to enter the environment and spread, differences in the farming system’s ability to reduce animal escape, and differences in whether humans hunt or fish for the same species. Some farmed animals are often transported and sold alive, posing additional routes for accidental entry into the environment. Escaped GM fish and shellfish, or their descendants, could be harvested without being detected and subsequently eaten by people. The current status of development of GM animals suggests that food safety managers might be faced with this issue first for GM fish and shellfish, and somewhat later for some kinds of GM poultry such as ducks and quail.

Key species or taxa of GM animals can be ranked in terms of their ability to become feral, likelihood of escape from captivity, mobility and historical reports of ecological community disruption Such a ranking from high to low for North America (National Research Council, 2002). would consist of insects, shellfish, fish, mice-rats, cat, pig, goat, horse, rabbit, dog, chicken, sheep and cattle. On a regional scale, the relevance of this ranking to food safety will change depending upon whether these animals are widely eaten by humans. Furthermore, the rankings will vary among regions owing to different environmental conditions, but the same risk factors apply.

Assessment of the potential environmental spread of GM animals or their transgenes should be done case-by-case for each combination of integration event (i.e. transgenic line) and local environmental conditions. The assessment should compare the GM animal with its conventional counterpart, i.e. unmodified animals derived from the same genetic source. The assessment should estimate the probability of movement of the GM animal or its transgene(s) into the environment, given the estimated rate of escape. This involves assessing whether:

• the GM animal, compared with the conventional counterpart, has a lower, equal or higher potential for gene flow to any wild or feral relatives found in the receiving ecosystem. Recent research suggests that the transgene could be purged within a few generations or could spread through the natural population and possibly affect its abundance (Muir and Howard, 2001, 2002). For potential entry into the food supply, purging would be a safer outcome. However, these animals may still enter the food supply because the purging process is likely to take one or more generations;

• the GM animal, compared with the conventional counterpart, has a lower, equal or higher potential to invade and establish itself as an alien species, particularly when the receiving ecosystem lacks wild or feral relatives.

4.3.3.1 Status of methods for estimating potential environmental entry

The best methods for reliably characterizing potential environmental entry have not yet been standardized. The net-fitness methodology (Muir and Howard, 2001, 2002) provides a systematic and comprehensive approach based on contemporary evolutionary and population biology (National Research Council, 2002; Pew Initiative on Food and Biotechnology, 2003). It involves a two-step process: (1) measuring fitness-component traits covering the entire life cycle for GM animals, their conventional counterparts or wild relatives, and crosses between the two; and (2) entering the fitness data from step 1 into a simulation model that predicts the fate of the transgene across multiple generations. There is a need to validate this method’s predictions; initial experiments are under way towards this end. For example, a transgenic fish project is testing the validity of the net-fitness model’s predictions (summary at: www.reeusda.gov/crgam/biotechrisk/biot00ls.htm). There is also a need to add stochasticity, elaborate additional features and improve the user-friendliness of the simulation model. The data needed to apply this methodology have yet to be obtained for most GM animals; efforts to gather such data have recently begun for a few cases of GM fishes.

4.3.3.2 Confinement of GM animals

If the combined conclusions from the environmental hazard characterization (discussed above) and the food safety assessment (discussed in section 5) are that GM animals or their transgenes will spread in the environment to a degree that poses a risk to the human food supply, then risk managers should consider the need to apply confinement measures to prevent or reduce escape of GM animals or their viable gametes into the environment. The primary focus of these measures should be to ensure that release does not occur. If this cannot be assured, then it can be complemented by the use of methods to ensure that any escaped individuals cannot reproduce.

Biological, mechanical and physical/chemical confinement measures are available for fish and shellfish produced in different aquaculture systems (Agricultural Biotechnology Research Advisory Committee, 1995; Scientists’ Working Group on Biosafety, 1998; Kapuscinski, 2003). Biological confinement measures typically involve disrupting the animal’s ability to reproduce, such as sterilizing fish and shellfish by induction of triploidy, i.e. resulting in individuals with three sets of chromosomes rather than the normal two sets. Mechanical confinement involves application of some kind of implement to prevent or reduce escape of animals from the aquaculture system (e.g. screens in effluent pipes of land-based fish tanks or ponds) and physical confinement involves making an aqueous escape pathway lethal by changing a physical attribute of the water (e.g. heat effluent water to lethal temperature, and then cool down before discharge). Other confinement systems could be developed for GM terrestrial animals.

In many cases, there is a need for multiple confinement measures because no single measure is fully effective. For instance, all-female triploid populations of salmon can be used to ensure that any individuals that might escape physical confinement are unable to reproduce in the wild. There is also a need for robust verification of confinement measures.

There is a need to develop and validate better methods for reliably inducing reproductive sterility of GM animals, in particular GM fish and shellfish. Improved methods could include repeatable protocols for induction of triploidy (applicable to animals other than birds and mammals) and new methods for inducing sterility via chemical treatments or gene transfer, for instance, by inserting anti-sense genes to disrupt key steps in the endocrine pathway controlling reproductive development.

4.3.3.3 Monitoring for environmental entry and spread

In future, specific GM animals may gain approval for widespread production, either with or without approval to enter them in the human food supply. In such situations, it will be important to consider whether or not to apply post-market monitoring for unexpected environmental spread of the GM animals and their transgenes that pose food safety hazards. Methods for detection of such GM animals and their transgenes in the environment are likely to involve the application of two well-established bodies of scientific methodologies: (1) diagnostic, DNA-based markers and (2) sampling protocols that are adequate (in terms of statistical power) and cost-effective. However, there is a need to develop fully appropriate protocols for the application of these methods to post-market detection of environmental spread of GM animals and their transgenes. Monitoring can also be helpful to assure confinement of GM animals during research and development.

4.4 Future perspectives

Realizing the full range of potential benefits from use of GM animals will depend on advances in technical aspects of their production. For instance, innovative molecular methods will be needed in order to address current limitations with respect to low frequency and randomness of integration, gene silencing and epistatic and pleiotropic effects of transgenes. These innovations may include using improved expression vectors to target transgenes to specific places in the host genome or incorporating transgenes on to bacterial or yeast artificial chromosomes and their introduction into the host. Greater experience with anti-sense gene expression and homologous recombination-based gene knockout techniques will allow the turning off of targeted genes. Other advances, especially development of embryonic stem cells or primordial germ cells from additional species, will facilitate a greater range of genetic modifications of those animal species. Identification and the transfer of genes will promote beneficial food uses, and also necessitate their assessment with respect to food safety and environmental impacts (e.g. insects with food safety implications, such as honey bees). Transfer of new genes and advances in gene transfer and cloning techniques will facilitate developments contributing to human health by means of new animal models of human disease, expression of pharmaceutical proteins and development of genetic lines for xenotransplantation. Although many potential benefits from GM animals can be anticipated, these will present more challenging conditions for risk assessment, risk management and risk communication.

TABLE 1. Examples of application of gene transfer to animals

Sources: (NRC, 2002; Kapuscinski, 2003).