2.9 GM Animals
2.9.1 Transgenic animals and targeted mutants
Transgenic animals (TA) express, or may express if properly induced, proteins encoded by cDNA or genes usually appended to heterologous transcription regulatory elements (TRE). These fusion genes are commonly referred to as transgenes. It is important to note that the expression of proteins, which are primary gene products, may or may not be the only acquired features of a given TA. If the transgene-encoded protein is an enzyme and if its substrates are present within the cell, then secondary gene products also will be synthesized. TA may accumulate these products in tissues or biological fluids in which they are not normally present. Expression of peptide hormones such as growth hormone (GH) in
TA radically changes carcass composition and induces systemic physiologic alterations. For example, mice that express GH are larger than their nontransgenic littermates, have elevated levels of insulin, and die prematurely due to liver and kidney damage. In addition, transgenic expression of hydrolytic enzymes may result in the disappearance or modification of molecules normally present in animal tissues or biological fluids. Closely related to TA are animals into which genetic alterations have been introduced, which in turn result in the suppression of endogenous gene expression. These are termed "targeted mutants" (TM) and include the so-called "knockout" or gene-disrupted animals. Both TA and TM are genetically modified animals and are produced and propagated for the purpose of gaining or losing functions with respect to the wild type or standard animal.
2.9.1.1 Examples
The relevance of TA to the field of nutrition may be better illustrated by a few examples, including pigs and mice that express human proteins in their milk or in their urine; pigs with unique fat/muscle ratios due to the expression of peptide hormones; sheep with altered carcass composition due to transgenic expression of hormones; mice that have altered milk oligosaccharides and glycoproteins; the production of lactose-free milk or milk with reduced lactose content due to elimination of α-lactalbumin; and tilapia overexpressing GH.
The meat or milk from the animals listed in these examples contains measurable molecular differences when compared with meat or milk derived from nontransgenic animals. As these products become accessible to consumers, both nutritionists and dietitians will need to assess

the overall impact of their inclusion into the food chain or their utilization to achieve desired nutritional results.
A transgenic animal is one that carries a foreign gene that has been deliberately inserted into its genome. The foreign gene is constructed using recombinant DNA methodology . In addition to a structural gene, the DNA usually includes other sequences to enable it

 to be incorporated into the DNA of the host and to be expressed correctly by the cells of the host.


Transgenic sheep and goats have been produced that express foreign proteins in their milk.
Transgenic chickens are now able to synthesize human proteins in the "white" of the eggs.
These animals should eventually prove to be valuable sources of proteins for human therapy.
2.9.1.2 Consequences of the use of transgenic technology
Products obtained from transgenic plants are already in the marketplace, and many vegetables and fruits are being targeted for modifications and improvements through the use of transgenic technologies. These are not necessarily focused on improving the nutritional quality of food products, but often on increasing efficiencies in the production of cultivars, improving resistance to insects and pathogens, and changing or enhancing organoleptic characteristics. Because of the fast pace at which TA are evolving, it is conceivable that a significant portion of food products on supermarket shelves will become targets for functional, compositional, or nutritional improvements. Products derived from TA and TM may be modified in predictable or unpredictable ways. For this reason, they represent an interesting subject of study for researchers in the field of nutrition. At present there are no available TA-derived products in the marketplace; therefore, a review on the subject must deal with the current status of relevant technology as well as with its potential, which will necessarily be framed by regulations, issues pertaining to the ownership of animals and technologies, and consumer perceptions.
2.9.2 Transgenic animals, an evolving concept
The production of TA is based on scientific principles established primarily through research in the areas of molecular biology and embryology. The fast pace of discoveries and technical improvements in these areas is reflected by the constant evolution of prevalent definitions. Page et al.define TA as "a result of the incorporation of a foreign gene such that it becomes an integral part of the natural chromosomal makeup of the animal." It is proposed

that definitions in which the key elements are the incorporation of exogenous DNA into the germ line of animals and the preservation of such genetic material in subsequent generations.
At present, exogenous DNA can also be incorporated transiently into specific tissues, thus producing TA that express heterolo-gous proteins without incorporating foreign DNA into their germ lines. In addition, homologous proteins that are expressed in certain tissues of a given species can be transgenically expressed in tissues in which they are not normally found. An example of this case is a murine enzyme (α-galactosyltransferase), which is normally expressed in liver and transgenically expressed in lactating mammary glands of mice. For purposes of this review, TA are defined as the nonhuman animals that have incorporated a fusion gene and permanently or transiently express at least one primary gene product encoded by the introduced DNA. This necessarily implies that at least one tissue, cell, or biological fluid has been modified either qualitatively or quantitatively as a result of the expression of the genetic constructs or transgenes.
2.9.3 Generation of transgenic animals
2.9.3.1 Genetic constructs
Several reviews describe the state of the art in TA production through the years.
These reviews address relevant technologies from different perspectives, including molecular biology aspects; the production of Pharmaceuticals in TA; the general use of TA as research tools; and the fields of biotechnology, agriculture, and nutrition. All of the mentioned reviews include sections on two fundamental aspects of transgenic technology: the generation of recombinant DNA suitable for transgenic expression and the techniques employed to introduce it into animals. Methods for the identification of TA and propagation of transgenic embryos also are discussed in several of these reviews.
2.9.3.2 Regulatory elements

General descriptions of the structural features of a transgene can be found in Rosen and
Btirki and Ledermann. Figure 23 is a schematic representation of transgene architecture.
Transgenes contain TRE, which control the expression of the protein-encoding DNA sequences. These elements regulate the initiation of transcription of adjacent DNA into mRNA, which is subsequently translated into protein. TA have been excellent tools to identify TRE such as promoters, enhancers, and silencers and to determine if their action is equivalent in different species.
It should be noted that the term "promoter" has been used as a synonym of TRE. There are TRE that are tissue specific; therefore, the expression of the adjacent DNA is targeted to certain tissues or organs. Examples of these are the lactogenic
TRE, which enhance DNA transcription during late pregnancy and lactation and target gene expression to the epithelial cells of the lactating mammary gland. Other TRE are inducible; that is, they can be "activated" by applying external stimuli. One example is the metallothionein
TRE, which is induced to promote transcription by the addition of heavy metals to the animal's diet. Another is the phosphoenol-pyruvate carboxykinase TRE, which is not active during fetal development, but is turned on after birth and can be regulated by the amount of dietary carbohydrate. Other TRE direct expression to many tissues simultaneously.
Most transgenic experiments result in insertion of the transgene into chromosomal DNA. For this reason, transgene expression may also be affected by the transcriptional state of surrounding genomic DNA. Dominant control sequences known as locus control regions (LCR) are used to "shield" the transgene from such effects (Figure 23C).
Transgenes can also be protected from position effects by other type of boundary elements known as matrix attachment regions
(MAR) or scaffold attachment regions. In addition to the mentioned elements, it is possible to engineer sequences that have differential susceptibilities to methyl-ation in different species.
Hypermethylated regions have been shown to inhibit gene transcription whereas hypomethylated DNA does not.

Figure 23 Schematic representations of different fusion genes for transgenic expression. (A) Simple construct containing a short transcription regulatory element, a cDNA, and polyadenylation encoding sequence. (B) More complex construct showing a larger regulatory element including an enhancer and strategically placed introns. (C) A construct with "shielding" elements, in this case, locus control regions
(LCR), which contains a full genomic sequence. On occasions the full genomic sequence including transcriptional regulatory elements have been successfully used.
2.9.3.3 Protein-encoding DNA
The core fragment of the transgene is the DNA sequences that encode protein.
Sometimes the only available DNA is a cDNA, which is generated from mRNA libraries by reverse transcription (Figure 23A). In contrast, most genomic DNAs contain intervening sequences called introns, which are not translated. Introns tend to stabilize and promote protein expression and can be engineered into a fusion gene (Figure 23B). Alternatively, genomic DNA with all its introns is preferred (Figure 23C). On the other hand, there are practical limitations to the size of the transgene. In general as the transgene becomes larger
(i.e., >20 kb) the rate of production of TA is decreased. This is thought to be due to fragmentation of DNA during the microinjection procedure (Yun, personal communication).
In some instances, this has precluded the routine use of genomic DNA. In some cases,

"minigenes" that contain a subset of the genomic introns are used in the transgene construct.
Although naturally occurring primary gene products are the most frequent targets for transgenic expression, it is also possible to produce chimeras or fusion proteins. Signal sequences that target the nascent polypeptide to intracellular compartments, secretory signals, or entire reporter genes can be engineered into the protein-encoding DNA fragment.
2.9.3.4.1.1.1.
Incorporation of the transgene
The most commonly employed technique to introduce transgenes into animals is microinjection of the fusion gene into the male pronucleus of embryos as reviewed. These embryos are then implanted in pseudopregnant females and the resulting offspring are assessed after birth for the presence of the transgene. It is not possible to regulate the number of copies or the sites in which the transgene will be inserted into the chromosomal DNA of the recipient. Likewise, the influence of the surrounding chromatin cannot be predicted. That is the reason why shielding elements such as the LCR and MAR described above are used in transgene construction. The number of copies of the transgene may or may not affect the final level of transgene-encoded protein produced.
Because of their short generation times, mice have been frequently produced using embryo microinjection. The use of this technique for the preparation of larger transgenic farm animals with significantly longer gestation periods requires long waiting periods before the results of an experiment can be evaluated. For this reason, techniques have been developed to determine if microinjected embryos at advanced stages of development have incorporated the transgene prior to implantation into surrogate mothers.
A second technique currently used in freshwater fish and marine organisms is electroporation. Eggs are incubated in the presence of the transgene and electrical pulses are applied. Alternatively, sperm can be electroporated, thus acquiring the transgene. This sperm is then used to fertilize eggs, which results in transgenic organisms. Electroporation is also used to introduce transgenes into somatic mammalian cells and pluripotent embryonic stem cells.
These cells can be cultured and tested for transgene incorporation. Cells that carry the transgene can then be introduced into host morulae, which are then implanted into recipient mothers. The resulting TA are chimeric; that is, the transgene is present in some, but not all, of the cells of the progeny including gametes. Homozygous animals for the transgene can be produced by crossbreeding selected animals. Embryonic stem cells are frequently used to generate targeted gene mutations.

TA can also be generated by infection with retroviruses and retroviral vectors.
Embryos or pluripotent cells can be transfected using this technique. An interesting application of retroviral transfection described and involved the direct transfection of animals through the teat canal using such vectors. In this case, the exogenous DNA was incorporated only into the genome of mammary gland cells and was not transmitted through the germ line. Transfection takes place during hormone-induced mammary gland differentiation and transgene-encoded proteins can be found in the milk. The transgene is transcribed at least for the duration of lactation.
Finally, once cells containing transgenes or targeted mutations are available, it is possible to obtain derived TA or TM by nuclear transfection. In this technique nuclei are obtained from cell cultures of transfected cell lines such as fibroblasts and are transferred into enucleated oocytes. These oocytes are then implanted into pseudopregnant animals.
This demonstrates that somatic cells can be used to generate TA (in the sense that they still contain a human-made fusion gene). This technology can expedite the development of a productive herd that would be comprised of clones from an original TA. The recent generation of transgenic rats and mice by testis-mediated gene transfer may not only provide a more efficient means of TA production but also allow TA production in currently "resistant" species. This method generates transfected sperm by direct injection of
DNA-liposome complexes into the male testis. These animals are then mated to normal females to generate the TA.
2.9.4 Transgenic animals and the field of nutrition
TA can be studied from different perspectives depending on their actual and potential contributions to the field of nutrition. Most of the available reports in the scientific literature describe the use of TA as experimental systems or models for scientific studies or as production prototypes. In only a few instances are protein Pharmaceuticals being produced by TA in commercial or near-commercial scale. From the perspective of nutrition and its allied fields, TA can be classified as:
(1) models for the study of nutritional stages and metabolic diseases,
(2) sources of modified food products,
(3) bioreactors that produce ingredients for nutritional products.

Occasionally, the difference between animals used as bioreactors and animals as sources of modified food products is subtle and may be illustrated by the following example.
Milk from a TA that produces high levels of K-casein could be directly consumed or used to prepare cheese. In this case the product is derived from the TA with no significant handling other than traditional food processing. In contrast, if K-casein is purified from milk and is used as an ingredient to improve the function or nutritional characteristics of food or nutritional products, then the TA is being used as a bioreactor. The latter is the approach most favored for the production of Pharmaceuticals in milk.
2.9.5 Transgenic animals and transgenic mutants as models for the study of nutrition
Several reviews describing TA as disease and biochemical models have been published. Some of these are described. It is discussed models for cardiovascular disease based on alterations to renin-angio-tensin and other systems. Some of these models could be useful to study diet effects on hypertension, cardiac hypertrophy, and thrombosis. Several models for the study of endocrine disorders, including diabetes, were reviewed that TA and
TM models of lipoprotein metabolism and atherosclerosis. This review catalogues mice with altered lipoprotein transport proteins, lipases, and receptors and is perhaps one of the most illustrative documents on the potential of TA as nutritional models. It is reviewed transgenic models for endocrine disorders, and also described that the history and rationale behind animal models focused on the study of obesity, allowing the reader to compare models obtained through traditional genetic selection techniques with those generated through the use of transgenic technology. In addition, animals that express GH transgenes may be models for the progression of type II diabetes. These animals have normal glucose levels but elevated insulin levels. Although they are not diabetic, they are hyperinsu-linemic. This situation parallels that seen in humans who are destined to become diabetic. In a similar manner, animals that express GH antagonists have normal levels of glucose but reduced insulin levels. The effect of nutritional status and either elevated or depressed levels of GH may help determine the glucose/insulin levels in patients with type II diabetes. TA that were originally developed to study muscle mass or growth rates have also turned out to be interesting nutritional models. Table 17 lists examples of TA and TM models.
Models of tissue differentiation and function that are relevant to the field of nutrition also have been generated. A glycosyltransferase, a 1-3/4 FT, was expressed in small intestinal epithelia cells and the presence of glycoconjugates synthesized by the enzyme was confined to crypt cells.
Double TA simultaneously expressing the simian virus 40 tumor

antigen and the transferase exhibited increased synthesis of the secondary gene product, thus providing a marker that responds to proliferative status of the cells. It is believed that glycoconjugates in mucosal cell surfaces function as receptors for pathogenic microorganisms. In this experiment the cell surface glycoconjugates of epithelial cells of the small intestine were remodeled. By altering the putative receptors of intestinal pathogens, researchers can verify if susceptibility to intestinal infections decreases or increases concomitantly. This suggests that models can be exquisitely fine-tuned to target effects to specific tissues or cells. Perhaps one of the most exciting features of some models of nutrition status related diseases is that many of them are diet sensitive. This allows the researcher to quickly evaluate the effects of diet changes and their impact on concomitant variables.
The applications described in this section are just a few of the areas related to nutrition in which TA and TM models are being generated. It is precisely in their roles as models that TA and TM have fulfilled the expectations of scientists. In addition, the ectopic expression of a transgene in the developing embryo may prove to be lethal or the TA may fail to thrive postpartum. Even in those cases, TA generate knowledge about the function of gene products and elicit research in new lines of inquiry.
Table 17 Examples of mouse models produced through transgenic or targeted mutagenesis technologies
Primary gene product
Brown adipocyte uncoupling protein
Expression of human triglyceride lipase
Mutant thyroid hormone receptor
Apoliprotein E
Bovine growth hormone
A2 Adenosine receptor
Alpha 1B adrenergic receptor
Cholesteryl ester transfer protein and apolipoprotein B
Breast cancer oncogenes
Human renin and angiotensinogen a 1-3/4 Fucosyltransferase a-Galactosidase
Source of transfected brown fat tumors
Lowering HDL cholesterol levels
Resistance to thyroid hormone
Diet sensitive atherosclerosis
Diet (carbohydrate) sensitive growth hormone in plasma
Thyroid hyperplasia, hyperthyroidism
Growth stimulation, induction, other
Applica tion malignancy
Cholesterol feedback regulation, LDL induction
Effect of diet in cancer onset
Atherosclerosis
Proliferative state of epithelial small intestine cells
Fabry's disease
Reference
(year)
Rossetal.(1992)
Buschetal.(1994)
Wongetal.(1997)
Plumpetal.(1992)
McGrane(1988)
Ledentetal(1992)
Ledentetal(1997)
Liu etal. (1997)
Raoetal.(1997)
Sugiyama.(1997)
Bryetal. (1996)
Ohshimaetal(199
7)
HDL-high density lipoprotein. LDL-low density lipoprotein.

2.9.6 Transgenic animals as sources of functionally modified food products and as bioreactors
The development of transgenic tomatoes constitutes perhaps one of the most exciting recent events in the area of food science. A business intelligence report suggested that products from TA would be on the market at the same time transgenic tomatoes were introduced. As of
June 1998 products from TA have not yet reached the marketplace. Several reasons may be responsible for the apparent lag of commercial introduction of products from TA, but a few seem to be particularly important:
(1) As opposed to plants, TA were first targeted as bioreactors for pharmaceuticals and these products are subjected to lengthy regulatory processes;
(2) expression of certain transgenes in farm animals has resulted in deleterious effects;
(3) the regulatory hurdles for a transgenically-produced food or nutritional product are unknown, whereas pharmaceuticals have been produced from a variety of sources including genetically-modified organisms;
(4) it is possible that the current state of the technology and peculiarities of animal systems require longer periods of time for their development.
2.9.6.1 Carcass and meat improvement
The prospective contributions and problems of TA as improved livestock are summarized in a review and compares the potential of TA as an option to obtain lean beef with other approaches. List of traits that could be improved through the use of transgenic technology. Those relevant to nutrition are
(1) efficiency of meat production,
(2) improved quality of meat,
(3) changes in the quantity and quality of lipids,
(4) efficiency of milk production,
(5) improved quality of milk,
(6) resistance to parasites and pathogens,
(7) improved quality of poultry and their eggs.

In addition, significant efforts are ongoing to improve fish and aquaculture in general. For many years farmers have selected animals that grow faster and convert feed into carcass weight more efficiently. Administration of anabolic hormones has also been used as a strategy to enhance growth and carcass quality; however, animals may develop health problems when subjected to chronic hormonal treatments.
The advent of TA allowed for the determination of effects resulting from the expression or suppression of single genes and for the methodical analysis of the effects of peptide hormones from different species. For these reasons, the most prevalent theme in the area of improved livestock through transgenesis is the expression of GH and related proteins.
In addition to GH, somatostatin, which inhibits the release of GH, insulin-like growth factor
(IGF-I), which acts on peripheral tissues, and GH release factor, which stimulates GH synthesis and release, have been expressed in experimental and farm animals. Several fusion genes containing different promoters have been expressed in different species. Table 18 summarizes some of these experiments and their results.
Some interesting secondary effects, which frequently include infirmities of different types, have been reported in TA expressing the mentioned proteins. These and other unexpected deleterious effects in TA will be discussed below. In addition, some of these animals can be studied as models of certain endocrine stages, as is the case of GH transgenic lambs, which were hyperglycemic. Mice are generally used as models to predict transgene viability in other species. However, mice do not always accurately emulate transgene expression effects that occur in larger animals. This is exemplified by mice that express bovine and human GH, which show accelerated growth rates and are larger than control mice. That has not been the case when GH is expressed in pigs, which may gain weight faster and be more efficient in converting feed into meat, but do not grow larger than control littermates. However, one report does describe GH transgenic pigs with enhanced growth rates. Perhaps the most remarkable feature in both transgenic pigs and lambs expressing GH is that the carcasses are significantly leaner than those of control animals. Recently, a mouse lacking the GH receptor gene has been generated and exhibits a dwarf phenotype. Nutritional studies on this animal may help in evaluating the role of GH as it relates to weight gain, body composition, and so forth. These examples suggest the potential impact of TA in the field of nutrition. Such an impact is further illustrated by the outstanding review of Pursel and Solomon on carcass composition in transgenic swine. In their review, transgenic and control animals are compared with regard to several biochemical composition parameters. For example, transgenic pigs expressing bovine GH not only contained less fat, they consistently contained

lower percentages of saturated fatty acids than control animals. In addition, a larger percentage (36%) of the total fatty acids of transgenic pigs were polyunsaturated fatty acids; in control animals the average was 19%.
There are many conceivable ways to alter meat quality through the expression of transgenes. For example, lipid composition could be altered by the transgenic expression of enzymes that act on fatty acids; the inducible expression of proteases could be used to tenderize meat prior to slaughter; proteins such as gelatin could be overexpressed; and muscle proteins from some species could be expressed in others to yield firmer, softer, or more digestible meat. As long as normal function is not significantly altered, the amino acid profile of certain tissues may be altered by overexpressing proteins or by exchanging endogenous genes for selectively mutated ones. However, as in the case of GH, technical subtleties and concomitant changes cannot be evaluated until TA have been generated, and in some cases, until the transgenes have been passed to subsequent generations. Although a significant amount of experience and data are being accumulated in this area, advances in the basic understanding regarding control of transgene expression are still required.
Table 17 Transgenic expression of growth hormone (GH) and related proteins
Primary gene product
Bovine GH
Human GH
Porcine GH
Human IGF-I
Ovine GH
Human GH
Bovine GH
Human IGF-I
Bovine GH
Tilapia GH
Rainbow trout GH
Transgenic animal
Mice
Mice
Pig
Mice
Lamb
Pig
Pig
Pig
Pig
Tilapia
Carp
Promoter Highlight Reference
(year) hMTA-IIA hMTA-IIA hMT-IIA
MT-I oMT-IA
MT-I
MT-I
MT-I
PEPCK hCMV
RSV
Increase in growth rate
Increase in growth rate
Increased weight gain
Somatic growth gain
Body fat as low as 1/5 of controls
Decrease in carcass fat
Palmiter et al.
(1982) 75
Palmiter et al.
(1983) 41
Vizeetal.
(1988)
76
Mathewsetal.
(1988)
77
Wardetal.
(1989)
78
Purseletal.
(1989) 67
Decrease in carcass fat Purseletal.
(1989) 67
Elevated IGF-I
41% reduction in backfat
Purseletal.
(1989)
67
Wieghart et al.
(1990)
8 depth
F1 82% larger than control Martinez et al.
20-40% faster growth
(1986)
13
Chenetal.(1993) 7
9

hMT-IIA-human metallothionein IIA. oMT-IA-ovine metallothionein IA. MT-I-mouse metallothionein IA.
PEPCK-rat phosphoenolpyruvate carboxyki-nase. hCMV-human cytomegalovirus enhancer-promoter.
RSV-rous sarcoma virus.
2.9.6.2 Modified milk and milk ingredients
Milk is a food product that has been extensively modified via expression of specific transgenes and, in many cases, without deleterious consequences for the TA. Several pharmaceutical proteins are already being produced in TA milk and are at different stages of development, from concept testing to preclinical and clinical evaluation. Examples of these are human a-antitrypsin in sheep, human plasmino-gen activator in goats, human protein C in pigs, and human IGF-I in rabbits. Milk can be modified or improved from different perspectives:
(1) as an important dietary component for mass consumption,
(2) as a raw material for dairy products, or
(3) as a source of ingredients for nutritional products including infant formula.
The fundamentals of transgene expression in lactating mammary glands and the fine points of the control of gene expression are discussed in a recent account. The architecture and function of milk protein genes was reviewed and published a comprehensive review on milk improvement through the use of transgenic technology. This review includes a historical background of the field and contains tables summarizing the primary transgene products and regulatory elements used to express proteins in the milk of TA. Human milk is considered by many to be the gold standard of infant nutrition. For this reason, it is not surprising that several groups are attempting the transgenic expression of human milk proteins and other human milk constituents. Human milk proteins, which are candidates for transgenic expression in milk, were catalogued and published a seminal comparative review of human and bovine milk. Nonprotein constituents of human milk, such as oligosaccharides, also have been targeted for expression in milk of TA. One of the most comprehensive reviews on these carbohydrate structures and their putative functions was published.
Table 18 Transgenic expression experiments of p and K caseins (CN) in milk of transgenic mice
Caseinorigin
Promoter Reference
(year)

K
-CN bovine
K -CN bovine
K
-CN
Own*
Goat-p-CN
Own rabbit p-CN goat Own p-CN bovine
Own
Rijnkles et al.
(1995)
25
Gutierrez et al.
(1996) 26,94
Baranyietal.
(1996)
95
Persuyetal.
(1992)
96
Rijnkles et al.
(1995) 25 p-CN bovine p-CN bovine
Sheeplactoglobulin
Bovine αlactalbumin
Hitchinetal.
(1996)
97
Jengetal. (1997)
98
*Was not expressed in mammary gland. "Own" denotes that regulatory elements were casein promoters of the same species.
The mammary gland is a privileged tissue in terms of its ability to synthesize milk proteins and lactose during late pregnancy and lactation. This unique character reflects a degree of biochemical insulation that is tightly controlled by hormones. Caseins, lactoglobulins, and whey proteins are among the primary gene products that are exclusively synthesized during lactation. Their expression is regulated by lactogenesis-responsive TRE.
The most abundant protein in the milk of experimental animals, such as mice, rats, and rabbits, is the whey acidic protein (WAP). WAP is a soluble protein whereas other proteins such as K-casein occur both in solution or as components of micelles. TRE of WAP genes from several species have been identified and used to generate TA that contain transgeneencoded products in their milk. The WAP TRE, pioneered by researchers at the National
Institutes of Health, has been used to express proteins in lactating mammary glands of animals. Other TRE that are frequently used to generate TA are from the α-lactalbumin, (3lactoglobulin, and different casein genes). Similar to the case of GH transgenic expression as a central theme of studies on carcass improvement, caseins and WAP are perhaps the dominant topics of transgenic expression targeted to the mammary gland. Because of the abundance of
WAP and caseins in milk, significant efforts have been undertaken to determine the nature and location of the regulatory elements that direct their synthesis. Discovery and characterization of powerful regulatory elements have the potential to impact the efficiency of expression of other proteins in addition to caseins or WAP. In addiiton, caseins are targets as functional ingredients that may facilitate cheese manufacturing and possess important biological activities that are relevant to infant nutrition.
Table 15 summarizes experiments in which caseins have been expressed in the milk of TA. Unfortunately, most of these experiments have been carried out in mice and no transgenic farm animal producing CN has

yet been reported. For this reason, it is difficult to assess the impact of TA in the field of nutrition.
As mentioned above, a rather ambitious and significant target for the transgenic modification of animal milk is the emulation of human milk. The production of proteins and secondary gene products characteristic of human milk has been pursued by several groups, including our own. The current state of the technology is illustrated by Table 19.
From this table it is apparent that only a few human milk proteins have been expressed in TA.
Furthermore, only one, lactoferrin, has been expressed in a large animal. The expression of lysozyme could be viewed from at least two different perspectives:
(1) the decreased microbial load, reduction of handling costs and processing intensity, or the increase of shelf life
(2) the synthesis of a human milk component that is desirable to have as an ingredient of dairy products including infant formulas.
This further exemplifies the difference between a TA producing an improved milk versus one producing an ingredient. The expression of fucosyltransferases is aimed at demonstrating the feasibility of synthesizing secondary gene products and remodeling of milk glycoconjugates and not at producing large quantities of the enzyme. Only catalytic amounts are needed to alter the oligosaccharide and glycoconjugate profile of milk. In each of these cases, available reports describe the technical foundations of a specific aspect of milk modification. Even using modern techniques for embryo propagation, commercial introduction of TA-derived milk products for mass consumption may not be available for more than 3 years. While pursuing emulation of human milk, two main types of tasks must be accomplished:
(1) synthesis of characteristic human milk components
(2) elimination of undesirable animal milk components.
TM, therefore, may play as important a role as transgenic expression of milk proteins.
Even though the present review is not focused on TM, it is important to note that technology is evolving at a fast pace in this area. For example, advances in ribozyme technologies have provided another method of reducing or eliminating specific proteins. Ribozymes are RNA molecules capable of hydrolyzing RNA, thus interfering with the translation of mRNA transcripts. It is reported the reduction of bovine α-lactalbumin content in the milk of transgenic mice through the expression of a second transgene encoding an α-lactalbumin specific ribozyme. With the use of these technologies, nonhuman animal milk proteins that are deemed undesirable may be reduced or eliminated. Other components are deemed un-

desirable for certain nutritional applications. Lactose (Gal(31-4Glc), which is the most abundant milk saccharide, is one of such components.
Table 19 Examples of human milk proteins expressed in transgenic animals
Animal Human milk protein
Lactoferrin
Lactoferrin
Bile salt stimulated lipase
Lysozyme
Fucosyltransferase a1
-3/4 FT
Fucosyltransferase a1
-2FT I*
Cow
Mouse
Mouse
Mouse
Mouse
Mouse, rabbit
Reference
(year)
Krimpenfort(1993)
99
Krimpenfort(1993)
99
Stro¨mqvist et al.
(1996) 100
Magaetal. (1994)
101
Prieto et al. unpublished data
Prietoetal. (1995)
1

Figure 24 Fluorophore assisted carbohydrate electrophoresis (FACE) image of labeled free oligosaccharides from different milk samples. Oligosaccharide extracts were prepared as previously described.
1
The equivalent to 5 |j,L of milk were labeled, electrophoresed, and imaged utilizing a Glyko (Novato, CA USA) system developed for the analysis of Olinked oligosaccharides and according to manufacturer's directions. Lactose, the most abundant saccharide in the mixtures, was allowed to run off the gel. Lane numbers indicate the source of the milk samples with the exception of lane 3, which is a molecular weight standard consisting of linear glucose polymers. Lane 7 is a standard of the human milk saccharide 2'fucosyllactose. Lane 1, sow; 2, sheep; 4, human; 5, rabbit; 6, cow; and 8, human. Positions of glucose polymers are indicated at the right side of the image: G3 = three glucose units; G4 = four glucose units; etc.
17

2.9.6.3 Remodeling of glycoconjugates in transgenic animals
Lactose is a secondary gene product; its presence in milk reflects the activity of the lactose synthetase complex, which is comprised of a-lactalbumin and Gal(31-4 galacto-syltransferase. In a similar fashion, the synthesis of free saccharides in solution and carbohydrate moieties of glyco-lipids and glycoproteins requires the expression of active glycosyltransferases in the appropriate cell compartments. One of the criticisms that has been raised to the production of glycoproteins in the lactating mammary glands of TA is that co- and post-translational protein modifications vary from animal to animal. A good example was provided. This group transgenically expressed human bile salt stimulated lipase (BSSL) in mouse milk and compared its glycosylation with those of authentic human BSSL and BSSL expressed in C127 mouse cells, CHO cells, and
Escherichia coli. Although transgenically-expressed BSSL was active, it was demonstrated that O-glycosylation was either reduced or absent. Because the proteins expressed in C127 cells were significantly glycosylated, it was concluded that the differences were not due to the species but to the characteristic glycosylation machinery in the lactating mouse mammary gland. Differences in the glycoconjugates present in milk of different species are dramatically illustrated in Figure 25 . Free oligosaccharides in solution are synthesized by the sequential elongation of lactose. It is apparent from this figure that human milk contains a uniquely rich and complex oligosaccharide profile when compared with those from other species. The structures of these molecules and their potential functions have been reviewed. Glycosyltransferases are the primary gene products responsible for the synthesis of oligosaccharides and other glycoconjugates. It is hypothesized that by expressing human glycosyltransferases in lactating mammary glands of mice, oligosaccharides that are not present in mouse milk would be synthesized and would accumulate in the milk of TA. This hypothesis was proven when transgenic mice expressing the human fucosyltransferase a 1-2 FT-I, under the control of WAP regulatory elements, synthesized 2'fucosyllactose (Fucal-
2Gal(31-4Glc), which subsequently accumulated in their milk. In addition, endogenous glycoproteins were remodeled as they acquired fucose residues. Figure 25 summarizes the results of these experiments. Figures 25A and 25B are chromatographic profiles of oligosaccharide extracts from milk of a control and a TA, respectively. The presence of a major additional component (2'fucosyllactose) in the milk of TA is clearly
18

demonstrated. Figure 23C is a Western blot analysis using a lectin specific for a 1-2 fucosylated residues and demonstrates the presence of a fucosylated glycoprotein only in TA milk. Since then, we have expressed other glycosyltransferases in animal mammary glands. In one of these experiments a murine a 1-3 galactosyltransferase, which is normally expressed in liver, was transgenically expressed in the lactating mammary glands of mice. The primary product of the trans-gene was a homologous protein and it synthesized a previously undescribed free oligosaccharide with the structure Galal-
3Gal (31-4Glc. These TA pass the transgene through the germline and it has been possible to obtain double TA through crossbreeding.
The removal of lactose from milk also has been pursued, and different strategies have been proposed. By decreasing or deleting the levels of a-lactalbumin, which cause decrease the lactose content of milk in TM. However, the total milk output substantially decreased. This may be caused by the absence of lactose in the "milk compartment" of the mammary gland. It is believed that lactose is the major osmolite in milk that is responsible for the movement of water from the plasma into the milk.
It may be possible to produce glycoproteins with "humanized" or modified glycosylation patterns in the milk or other tissues of TA and TM.
19

Figure 25 Analysis of control and transgenic mouse milk samples. Mouse milk samples from control (A) and transgenic (B) mice were prepared as previously described
1
and were assayed for free oligosaccharide content using high performance anion exchange chromatography. Protein pellets from the same samples were subjected to SDS-PAGE using a 10% polyacrylamide gel. After transferring to a polyvinylidene difluoride membrane, the glycoproteins were visualized using Ulex europaeus agglutinin I, a lectin specific for fucose a1 -2 linkages (C). The first lane corresponds to the control milk sample analyzed in (A), and the second column is the transgenic sample from (B).
2.9.7 Unexpected effects in transgenic animals
Table 20 summarizes some deleterious effects observed in TA after birth. Some of these unexpected findings have led to further research in the area of gene control and physiologic changes due to chronic expression of certain proteins. It is also not uncommon to find that embryos fail to develop when certain genes are expressed using systemic regulatory elements. A case in point is the expression of glycosyltrans-ferases and enzymes that affect glycosylation in general. The expression of Nacetylglucosaminyltransferase I was lethal to embryos whereas the expression of a sialic acid specific esterase under the metallothionein TRE also resulted in early arrested
20

development. The overexpression of (31-4-galactosyltransferase, also under the metallothionein TRE, resulted in impaired mammary gland development. In addition, experiments in which the human al-3/4-FT glycosyltransferase was expressed in different mouse tissues yielded different results with no apparent deleterious effects for the animals. Expression of the enzyme was directed by either the rat intestinal fatty acid binding protein TRE or the WAP promoter. The secondary gene products and remodeled glycoconjugates synthesized by the enzyme remained localized to their targeted tissues. However, the affects of the transgene expression may be speciesspecific. As previously described, we were able to express the a 1-2 FT-I in mice under the control of the WAP TRE. When the same fusion gene was used to generate transgenic rabbits, the transgene affected milk production and resulted in lactose-free milk. Several generations of these animals are now under study to understand the effect of the enzyme on rabbit lactation. These examples strongly indicate that many factors influence the final results and the viability of different TA species, even when the same construct or primary gene product is expressed. Other effects may simply be due to the limited capacity of tissues to express proteins, specifically, the lactating mammary gland. Ec-topic expression, such as in the case of WAP controlled expression of erythropoietin or extemporaneous expression of WAP in pigs, are interesting from the point of view of development, morphogenesis, and protein function. As indicated above, some of the effects are species-specific. This is dramatically illustrated by the fact that transgenic sheep express mouse WAP in several tissues other than the lactating mammary gland. The examples in Table 17 and those discussed in this section suggest that in-depth research on gene control in different targeted species is still required.
21

Table 20
2.9.8 Alternatives to transgenic technology
Transgenic organisms have captured the imagination of scientists and industrialists alike. However, biotechnology has evolved in other areas. Taking into account the relatively long development times for transgenic production of food products and ingredients, other options may be considered to solve specific problems.
An example is the production of human (3-casein. This human milk protein is present in several forms or variants containing from zero to five phosphate groups attached to serine and threonine residues. Bovine phosphorylated (3-casein that contains all the variants has been produced in the milk of TA.
Preparations of human (3-casein containing all the variants have also been obtained from E. coli by coexpressing the casein and human casein-kinase genes. In addition, (3-casein has been produced in transgenic potatoes) but the protein was not phosphorylated. These examples indicate that TA may not be the obvious choice to produce large quantities of human milk proteins. From the standpoint of nutrition, milk composition can also be altered by traditional methods such as changes in animal diet. When the goal is to produce specific modifications in animal food products for commercial purposes, all options should be considered and analyzed from at least four perspectives:
(1) timeline for commercial scale development,
(2) regulatory issues and concerns,
(3) economic feasibility,
22

(4) adaptability of current production methods.
In some instances TA and TM may not represent the best available options.
2.9.9.
Considerations for the assessment of the safety of genetically modified animals used for human food or animal feed
2.9.9.1 General
2.9.9.1.1 GM crops
Most of the GM crops that are on the market have been modified to enhance agronomic performance or to facilitate plant breeding. These crops contain one or two foreign genes, coding for proteins that are expressed in rather low quantities in plant tissues. Examples of introduced traits are
• Herbicide resistance, enabling GM crops to sustain application of herbicide on their leaves contrary to weeds, which perish under the action of the herbicide.
• Insect resistance, by the introduction of insecticidal proteins from bacteria.
• Male sterility, which facilitates hybrid breeding.
It is expected, however, that the range of traits will become diversified in future. In addition, some future GM crops will have undergone more profound genetic changes.
Whole new metabolic pathways, for example, may be introduced by the introduction of multiple genes. One example is the ‘golden rice’, which has been modified to produce provitamin A in its kernels, which naturally do not contain this provitamin. These new developments in crop plant biotechnology are reviewed in more detail elsewhere.
The foreign DNA in these crops is commonly introduced by two methods. First, plasmids of the bacterium Agrobacterium tumefaciens are employed, which are integrated into the plant genome. Second, DNA is introduced by ‘bombardment’ with accelerated DNA-coated particles. For both methods, cells or tissues from plants are cultured in vitro and regenerated into whole plants after genetic transformation with the foreign DNA.
23

2.9.9.1.2 GM animals
Many of the GM animals for food purposes that have been modified to enhance performance, have been transformed with growth-related genes, such as those for growth hormone (somatotropin), growth hormone releasing factor, and insulin-like growth factor. In addition, GM animals have been created with novel enzymes in their intestinal epithelium to increase the efficiency of feed utilisation as an alternative to the use of feed additives. Examples include animals expressing phytase to increase the uptake of organically bound phosphorus and animals expressing bacterial enzymes catalysing the synthesis of the essential amino acid cysteine.
A common method of genetic modification of animals is the microinjection of foreign DNA into egg pronucleus (mammals) or egg cytoplasm (fish). Another technique involves the genetic modification of mature cells followed by nucleus transfer into eggs.
Recently, the use of autonomous artificial chromosomes complemented these techniques for inheritable genetic modifications. These chromosomes may contain several features common to natural chromosomes, including centromeres, telomeres, and satellite DNA. The target foreign DNA has been introduced into these chromosomes. The advantage of this system is that the randomness and uncertainties inherent to foreign gene insertion are avoided. These artificial chromosomes are replicated as separate entities and passed through the germline to following generations.
In addition to the methods that target germline modification, gene therapy can be employed for the genetic modification of somatic cells within an animal. Different types of vector can be used, including retroviruses, which are integrated into the animal’s
DNA, and adenoviruses and plasmids, which do not integrate but remain autonomous.
Gene therapy was recently employed to target the transient local expression of plasmids carrying the gene for growth hormone releasing factor in pigs. The plasmid DNA containing the gene of interest had been applied by the technique of electroporation to pig muscles. Serum levels of growth hormone and insulin-like growth hormone were increased in the weeks following the modification.
24

2.9.9.2 Safety assessment of GM foods and feed
2.9.9.2.1 Substantial equivalence
2.9.9.2.1.1 GM crops
The principle of substantial equivalence encompasses the comparison of the
GMO to its conventional counterpart. Substantial equivalence is often mistaken for the outcome of a safety assessment. It is rather the starting point of the assessment, though, and based upon the degree of equivalence, additional safety experiments are undertaken.
The comparison can involve the phenotypic and compositional characteristics of the novel crop. The compositional analysis usually involves the analysis of macro- and micronutrients, antinutrients, and toxins. In addition, the comparison should involve plants harvested at a sufficient number of locations representative for the commercial cultivation during at least two growing seasons. There is, however, no standard recipe for this comparison, because of the divergent nature of food crops as well as the various genetic modifications. The GM crops are therefore evaluated on a case-by-case basis.
For the GM crops that are on the market now, with one or two foreign genes introduced, the principle of substantial equivalence has worked well. It is envisaged, however, that future GM crops will be the result of more intricate modifications, for example by the introduction of multiple genes and/or metabolic routes that are new to the recipient crop. Unintended effects of genetic modification appear more likely for such profoundly altered crops. Further refinement of the comparative approach of substantial equivalence is needed to keep abreast of these developments.
Currently research is carried out to develop analytical methods to analyse the changes brought about in profoundly altered GM crops. These methods do not focus on singular compounds, but record a whole spectrum of compounds, proteins or RNA present in a crop to screen for possible differences. Examples are liquid chromatography coupled to nuclear magnetic resonance (LC–NMR) to detect metabolites, twodimensional gel electrophoresis to detect proteins, and cDNA microarray hybridisation to detect RNA. The identity of these differences should then be further explored, and the implications for product safety assessed.
25

2.9.9.2.1.2 GM animals
The strategy for the safety assessment of foods derived from GM animals has been discussed within the OECD and the FAO/WHO. In these discussions, it was concluded that the food safety assessment of GM animals should comprise:
• Molecular characterisation of the inserted foreign DNA.
• Safety of foreign gene products.
• Unintended effects of the insertion of foreign DNA, e.g. effects on animal health and carcass composition.
• The effect of disease resistance brought about in transgenic food animals on consumer’s exposure to disease-causing agents.
The American authorities have formulated additional points that need to be considered in evaluating the food safety of GM animals, particularly the potential risks emanating from the use of retroviral sequences, including the risk of recombination with wild-type viruses.
These points for consideration were extended recently in an interim report of the
Canadian government in preparation of guidelines for the food safety assessment of transgenic livestock and fish. This report also considers that the principles for the safety assessment of GM plants can be applied to that of GM animals, although more elaborate guidance may be required.
The Canadian report suggested the inclusion of the following additional items in the safety assessment:
• The well-being of the transgenic animal.
• Compositional (‘nutritional’) analysis
• analysis of nutrients and bioactive compounds
• sampling of raw and cooked material, tissues and the whole-ground carcass
26

• meat quality impact of ‘large animal syndrome’
• Compositional analysis: novel methods needed in addition to targeted analysis.
• Toxicology
• screening for unintended effects by molecular profiling techniques and study of the GM animal’s behaviour and development genes.
• attention paid to the possible consequences of under- and over-expression of
The issue of the transgene copy number in homozygous animals versus heterozygous animals called for further consideration. With regard to the well-being of transgenic animals, a recent paper discusses the evaluation of animal welfare in the context of GM technology.
One of the safety considerations in the assessment of GM crops is the possible effect that the genetic modification may have had on the levels of intrinsic antinutrients and toxins. Although food animals rarely produce antinutrients and toxins, safety evaluators should be aware of exceptions to this general rule. Tetrodotoxin, for example, occurs in puffer fish, which is consumed in Japan. Another, more commonly encountered example is that of the enzyme thiaminase, an important antinutrient that degrades thiamine (vitamin B1). This enzyme is present in viscera of some commercial fish species. Thiaminase has been implicated in massive extinctions of larvae from prey fish both in Northern America and the Bothnic area, known as Cayuga syndrome or
Early Life Stage Mortality Syndrome. The larvae would have previously suffered thiamine deficiency during their embryo stage due to maternal consumption of forage fish containing thiaminase. Feeding livestock with thiaminase-rich fish has been associated with adverse effects, such as paralysis and mortality in dogs and loss of equilibrium and mortality in salmon. In addition, thiamine deficiency in a Thai human population has been associated with the consumption of raw fermented fish containing thiaminase. The effect of the genetic modification event on the expression levels of thiaminase in GM fish should therefore be taken into account.
27

Studies have been reported in literature on the comparative carcass analysis of growth-enhanced GM animals. GM pigs genetically modified with various types of growth hormone, for example, yielded leaner meat (i.e. lower fat content) than control pigs, while meat lipids contained less saturated fatty acids and more unsaturated fatty acids.
It was tested the safety of growth-enhanced tilapia, which had been genetically modified with tilapia growth hormone. Macaques were administered tilapia growth hormone by intravenous injection of 1 g/kg daily for 30 days. Blood samples were tested for parameters that are likely influenced by growth hormone activity. No effects of tilapia growth hormone were observed, corroborating the previously observed lack of activity on in vitro cultured rabbit cartilage. In addition, no adverse effects were observed on the blood parameters of human volunteers, who consumed GM tilapia twice daily during 5 days. It was fed carp transgenic for growth hormone to mice, which were subsequently analysed for reproductive toxicology, blood chemistry, and histopathology. No differences were found between the feeding of transgenic and non transgenic carp to the mice.
2.9.9.2.2 Allergenicity
2.9.9.2.2.1 GM crops
As all food allergens are proteins, the potential allergenicity of newly introduced proteins has been a consideration in the safety evaluation of GM crops. A decision tree has been drawn up to aid the evaluation for potential allergenicity of a transgenic protein. This decision tree was recently refined (Fig. 26). The first step in this refined tree comprises the comparison of the novel protein’s structure with the structures of known allergens by computer-assisted alignment of their amino acid sequences. The direction following the first step through this decision tree depends on, among others, the source of the foreign gene. If the gene source is a known allergen, than its reaction with sera from patients that are allergic to the specific source should be tested. In case of a negative result, or if the gene source has no history of allergenicity, (further) testing involves sera from patients that are allergic to organisms broadly related to the gene source. Depending on the outcome, further testing may be required, involving in vitro digestion with the stomach protease pepsin (as most food allergens are stable to
28

digestion) and in vivo animal testing. If any of these steps yield a positive outcome, the
GMO should be considered likely to be allergenic.
Fig. 26. Decision tree for the assessment of potential allergenicity of proteins introduced in GMOs.
Another consideration that may also arise from the modification, is that of the changes brought about in the properties and occurrence of allergens that are intrinsic to recipient organism. Experimental crops have been created, for example, with decreased levels of intrinsic allergens.
2.9.9.2.2.2 GM animals
Allergenicity is also an issue of concern in GM animals. Animal products, such as shrimp and cow’s milk, contain a number of well known and lesser known intrinsic allergens, e.g. tropomyosin (shrimp). The effect of the genetic modification on their
(allergenic) properties and tissue levels should therefore be determined.
29

Another point of consideration is the apparently incomplete digestion—and intestinal uptake—of orally administered proteins and peptides with hormonal activity in fish. This accounts for the fact that fish growth is enhanced by experimental feeding of, for example, animal pituitaries containing growth hormone. As fish and animals can be turned into feed for cultured fish, the possible effects of ectopic expression of protein hormones in GM animals warrants further consideration.
2.9.9.2.3 Gene transfer
2.9.9.2.3.1 GM crops
Antibiotic resistance genes are employed as markers used in the development and selection of genetically modified plants. Their purpose may be twofold. First, the
DNA vector carrying the DNA of interest needs to be produced in sufficient quantity prior to the genetic modification. To this end, plasmids containing the gene of interest and an additional antibiotic resistance gene are introduced into bacteria. The antibiotic resistance allows the bacteria harbouring the plasmid to grow on antibiotic-containing media, which ensures production of the plasmid. The plasmid can subsequently be purified from the bacterial culture for use in the transformation of plants. Secondly, plant cells that have successfully been transformed with DNA containing both the gene of interest and an antibiotic resistance gene are able to survive on antibiotic-containing media, whereas non-modified cells perish. This facilitates the selection of successfully transformed plant cells, which subsequently can be regenerated into plants.
Antibiotic resistance genes have no purpose in the GM crops, but are used only in the initial selection process. Safety concerns have been expressed, however, over the risk of transfer of these genes to micro-organisms residing in the gastrointestinal tract of humans and animals consuming GM crops. Although the probability for such a transfer appears low, no data to date have been published on transfection rates in vivo.
Consequently, a precautionary approach has been taken by various governments with regard to the use of antibiotic resistance genes. The target antibiotic, for example, should not be clinically important. In addition, the naturally occurring resistance to the antibiotic should be taken into account. Based on these considerations, the kanamycin resistance gene npt II was approved and has since then become the most common antibiotic resistance gene in commercial GM crops. Recently adopted EU legislation, however, prohibits the use of antibiotic resistance genes.
30

2.9.9.2.3.2 GM animals
The documented use of antibiotic resistance genes in GM food animals appears rare and may therefore be less of a concern. It can be envisioned, though, that in the production of DNA vectors for the transgene, ‘contaminating’ bacterial antibiotic resistance genes may originate. The question then arises how the accidentally introduced antibiotic resistance genes may come into contact with pathogenic microorganisms that might subsequently be transformed with these genes. Intestinal epithelial cells, for instance, may shed into the lumen, degrade and thereby expose their DNA to the intestinal microflora.
Another concern that has been raised specifically over GM animals is the use of retroviral sequences. This may especially pertain to transgenic poultry, because
(disabled) retroviruses are employed to successfully transform eggs, whose blastoderms contain numerous cells, which precludes the use of nuclear micro-injection or nuclear transfer. It has been reported that inserted retroviral sequences in animals recombine with wild type viruses giving rise to new retroviruses. Recombination of wild type retroviruses with endogenous retroviral sequences containing the transgenes may pose an increased risk over recombination with naturally incorporated latent retroviruses.
Novel methods can help circumvent this risk of recombination by the use of artificial retroviruses, in which the retroviral genome has been replaced by DNA vectors lacking sequences that are prone to recombination. In addition, the insertion of DNA (retroviral and non-retroviral) may hypothetically lead to the activation of adjacent latent viruses, although it has been argued that animal cells are capable of suppressing the activity of endogenous viral sequences. The targeting of foreign DNA to specific integration sites in the animal genome, as recently demonstrated in sheep, would circumvent this hazard.
2.9.9.2.4 Postmarket surveillance
2.9.9.2.4.1 GM crops
The EU recently adopted legislation on the cultivation of GM crops that requires the post market surveillance of the GM crops for any unanticipated adverse effects appearing in the long term (EU Directive 2001/18). Applicants seeking to introduce GM crops to the EU market are required to submit a draft monitoring program for approval.
In addition, new legislation on novel food and animal feed (including GM food and
31

feed) has been drafted by the European Commission, which envisions the same postmarket surveillance for GM food and feed. Traceability of the GM food and feed should be warranted by the labelling of the specific GM components throughout the entire production chain.
2.9.9.2.4.2 GM animals
The same would apply to GM animals as to GM crops. The tracing and postmarket surveillance of any future GM animal products in the EU will be facilitated by two factors: the EU.
1. Traceability-systems for animal products are currently being implemented in
2. Ingredients derived from animals are less generally present in food- and feedproducts than some vegetable ingredients (e.g. maize starch and soybean oil). This can be in part accounted for by anti-BSE measures for animal feed.
A difference between GM crops and GM animals is that some GM food animals may be regarded veterinary medicines by some regulatory authorities. An introduced gene that codes for a growth hormone, for example, is comparable to the administration of exogenous growth hormone, which is a veterinary medicine. These GM animals would therefore be subject to ‘pharmacovigilance’, which is the mandatory post market surveillance for medicines. Frameworks for pharmacovigilance are in place in Western nations and abroad, e.g. adverse reaction reporting systems. Technical requirements for pharmacovigilance of veterinary medicines in Japan, Northern America, and EU will be harmonised in the near future. These requirements include the reporting of adverse effects of veterinary medicines to the authorities, which is mandatory for the manufacturer and voluntary for others (e.g. animal owners, veterinarians).
32

2.9.9.3 Regulation of GM foods and feed
2.9.9.3.1 GM crops
The cultivation, trade, and food and feed uses of GM crops are subject to national regulations in most nations and the safety assessment of the GM crop is part of the admission procedure. For cultivation, for example, the environmental safety will be assessed, but this is beyond the scope of this article. Depending on the envisioned purposes for the novel crop, its food and feed safety will also need to be assessed.
Although the national regulations are somewhat different, the underlying principles are the same. The internationally acknowledged comparative approach of
‘substantial equivalence’ is an important guiding principle in the safety assessments under these regulations. Initiatives are underway to harmonise the regulations and the inherent safety assessments of GMOs internationally.
The Task Force on the Safety of Novel Foods and Feed of the Organisation for
Economic Co-operation and Development (OECD) is developing guidance documents, for example, on which compositional parameters should be compared between the GM crops and appropriate comparators. So far, guidance documents for soybean and canola have been completed, while those for potato, sugar beet, wheat, maize, rice, sunflower, and cotton are underway. These guidance documents serve as minimum recommendations, but are not legally binding.
The FAO/WHO Codex Alimentarius has developed international standards for food safety that its member states should implement into their national legislation.
Codex standards, for example, have been developed on residues of pesticides, contaminants, and veterinary medicines in food. These standards are legally binding and can be referred to in the case of international disputes on the food safety of traded foods and commodities. Efforts towards harmonisation of the national food safety assessments of GMOs are made by the Codex Ad Hoc Intergovernmental Task Force on Foods
Derived from Biotechnology. Expert consultations on specific items (substantial equivalence, allergenicity) have been organised by a special working group for this task force. Guidance has been drafted for the risk assessment and risk analysis, and special attention has been devoted to detection and traceability of GMOs. It is expected that within a few years these documents will be completed and adopted.
33

2.9.9.3.2 GM animals
The regulatory experience with market applications for GM food animals is very limited compared to that with GM crops. To our knowledge there are actually two cases where regulatory approval has been sought for the market admission of GM food animals. These are the Bresatec pig in Australia and the AquAdvantage salmon in the
USA.
The GM pig developed by Bresatec (currently Bresagen) is transgenic for a growth hormone gene that can be ‘switched on’ by adding zinc to the diet, which allows for increased production of meat (muscle tissue) at an increased feed efficiency. The
Bresatec company provided data to the Australian authorities that its GM pork meat were substantially equivalent to ordinary pork. At that point in time, however, the
Australia New Zealand Food Authority and the Australian Genetic Manipulation
Advisory Committee were unable to process Bresatec’s application because food safety of GM animals fell outside their regulatory oversight. Subsequently, because Australian supermarkets expressed a reluctance to market GM pork, Bresatec reportedly abandoned this project.
The AquAdvantage salmon developed by Aqua Bounty is transgenic for a salmon growth hormone gene under control of a promoter from the ocean pout’s antifreeze protein gene. The transgene is expressed in the salmon liver and provides for year-round secretion of growth hormone. The growth of the young AquAdvantage salmon is 4–6 times as high as for ordinary salmon, yet mature AquAdvantage salmon at harvest are the same size as ordinary salmon, weighing approximately 3 kg (Fig. 27).
Furthermore, the size of the GM salmon at sexual maturity did not increase above that of typical salmon after four generations of breeding. Growth hormone levels in edible tissues are reportedly within the range found within non-GM counterparts.
34

Fig. 27. Second generation transgenic Atlantic salmon containing the antifreeze protein promoter linked to a salmon growth hormone gene (AquAdvantage salmon).
The large (transgenic) and two control siblings are shown for comparison. Despite the increased growth rate of the GM sibling, both GM-and non GM-salmons will reach the same size at sexual maturity.
Aqua Bounty has submitted an application to the American authorities for the market approval of the AquAdvantage salmon. Contrary to agronomically enhanced
GM crops, GM animals with improved agronomic traits are regarded ‘new animal drugs’ by the Americans, comparable to exogenous growth hormone administered to animals. The application was therefore filed to the Center for Veterinary Medicine of the Food and Drug Administration as a ‘new animal drug application (NADA)’. The safety assessment will include many aspects, including the environmental risks associated with the different options for fish cultivation (open pens, closed tanks), risks of waste discharge from the fish farms, and food safety. The environmental risks are discussed elsewhere.
Food safety studies that have been carried out in preparation of the NADAapplication for the AquAdvantage salmon include:
• Level of expression of the introduced gene.
• Serum levels of peptide- and steroid-hormones.
• Bioavailability of the salmon growth hormone to humans.
• Comparison of the composition [macronutrients, micronutrients (vitamins, minerals), amino acids, fatty acids] between GM and non-GM salmons.
• Potential allergenicity of the GM salmon compared to ordinary salmon.
35

Several papers on the AquAdvantage salmon in their presmolt phase have been published, including data on metabolic activity, body composition and body composition as influenced by food deprivation. It has been noted, for example, that the presmolt transgenic fish bodies contained more moisture and less protein, lipid, and ash than controls of the same weight.
As noted above, the FAO/WHO Codex Alimentarius Task Force is preparing guidance documents for the safety evaluation of GMOs. In anticipation of future developments in food biotechnology, an Expert Consultation will be convened on the topic of genetically modified fish.
2.9.10 Regulatory and legal issues
The utilization of TA to produce food products and ingredients on a commercial scale does not seem to be achievable in the near term. It is difficult to predict how many years will be required for this to occur. In the meantime, the technology has been divulged and discussed beyond the specialized press. It is plausible that drugs produced from TA bioreactors will pioneer their way to market before food products. Regulatory agencies and consumer groups are already taking steps to prepare themselves for such an event. In the United States, the Food and Drug Administration (FDA) has issued a notice entitled "Points to Consider in the Manufacture and Testing of Therapeutic
Products for Human Use Derived From Transgenic Animals." This notice is an important first step to initiate discussions on regulatory issues pertaining to TA and illustrates the approach of the FDA to this particular tool of biotechnology. However, regulatory actions are not given in vacuum. Consumer attitudes toward biotechnology and political pressures result in governmental actions that may influence the development of applications of TA. The reader is referred to a study conducted on the attitudes of consumers toward biotechnology in The Netherlands. Furthermore, it is reported that the Dutch government limited cloning experiments designed to propagate
TA. It is clear that a current limitation of TA is the time required to generate a productive herd. This limitation can be circumvented by cloning already available TA.
The impact of the mentioned governmental decision may be significant for companies based in The Netherlands. Even definitions in regulations may create controversy. A way to circumvent the long waiting period for natural or induced lactation of a transgenic cow is by expressing transgenes in the lactating gland using viral vectors.
36

This technique can, at a minimum, be employed to determine if it is worthwhile to initiate germline incorporation experiments for a given transgene.
TA have elicited controversy and discussion in areas such as politics, ethics, and law. The reader is referred to two articles that discuss issues regarding intellectual property and patent law as it applies to TA.
2.9.11 Particular opportunities in the field of nutrition
The advent of TA allows experts in the field of nutrition to think about ways to improve milk, meat, and eggs for nutritional purposes. Present generations are witnessing an unprecedented environment in which foods can be substantially modified without the need of lengthy genetic selection and crossbreeding cycles guided by
Mendelian genetics. It is now possible to conceive or conceptualize the ideal milks to be used in different applications, such as production of cheese, infant formula manufacture, or products for undernourished children, and the ideal meat for large scale production of ground beef. These functional characteristics may be some of the first modifications of animal tissues and fluids that reach the market because they have undisputed commercial value. These possibilities force researchers in academia and industry to reexamine old concepts and standards based on the inherent limitations of currently available production methods.
2.9.12 Limitations and conclusions of the present review
One of the main objectives of the present review was to provide a panoramic view of the technology involved in the generation of TA. It was also an objective to illustrate the technology's current situation and potential by invoking examples and referring the reader to more specialized accounts. Scientific literature was used as the primary source of information. Three reports regarding regulations and business issues are cited because it was deemed that such issues will inescapably impact the development and expansion of transgenic technology. Several interesting experiments and models can be found in patent applications and they may represent some of the most advanced aspects of the technology. It was decided not to include them in this review because in most cases reproducibility and scope of the examples are not addressed. It is the bias of the reviewers that TA are already valued models for the study of
37

metabolic pathways, diseases, and infirmities and that traditional approaches such as diet management can now benefit from the existence of these models. In their review on transgenic dairy cattle, it is pointed out a paradox by indicating "Unfortunately, this area of investigation has received less attention in laboratories than it has in review articles." Although this may be true, it is important to consider that the fine line between science and technology is crossed back and forth in areas such as the subject of the present review. Companies and institutes do not disclose all the advances in the area because scientific divulgence becomes secondary to the submission of patent applications. In addition, scientific publications become a marketing tool for nascent technologies and efforts have to be made to separate facts from proposals or from overoptimistic portrays of the future. The following conclusions may also permeate the review as preexisting biases:
(1) some aspects of transgenic technology, such as gene expression control, are still under basic scientific investigation;
(2) most TA described up to this point are prototypes;
(3) each case requires its own evaluation, because results from laboratory animals do not necessarily predict those that would be obtained from farm animals; disease.
(4) TA and TM are powerful tools to study the effect of nutrition in health and
2.9.13 The fate of tDNA and expressed products fed to animals
2.9.13.1 Survival in the digestive tract
The rapid disappearance of DNA fragments in the rumen of steers was demonstrated many years ago using the methodology available at the time. There is no reason to suppose that the DNA inserted into modified plants is any less susceptible to the action of the rumen flora and its nucleases. Moreover, the daily intake of total DNA by an adult ruminant is small and foreign DNA even smaller when compared to the total dry matter intake. In the case of Bt maize fed to dairy cows, the amount of Bt DNA ingested has been estimated to be 54 g/d compared to a total of 54 g/d total DNA, equating to a GM DNA of 1 to one million of total dietary DNA.
38

The consequences and potential risks associated with the presence of expressed proteins are of particular concern in any safety assessment. The protein products of genes coding for insect resistance and herbicide tolerance have been fed without consequences to rodents in acute tests of toxicity at a very high doses, typically in excess of 100,000 times the level that humans or animals would consume. In the case of expressed proteins, in vitro digestion of the CP4 EPSPS protein expressed in soybean resistant to glyphosate and measured by Western blot showed a rapid breakdown by proteolytic enzymes. Similarly, the PAT protein (phosphinothricin acetyl transferase) conferring glufosinate tolerance to maize proved to be heat labile and highly and quickly degraded in simulated human and animal gastric fluid. The Bt protein almost disappeared during the ensiling process or could not be detected in the silage. The foliar application of Bacillus thuringiensis (Bt) has been used for over three decades as a commercial product and has a well established safety record. Rapid in vitro degradation of Cry 1 a ( b ) protein in simulated gastric juice has been observed. Similarly, neomycin phosphotransferase (NPT II) protein is highly sensitive to in vitro enzymatic hydrolysis.
Further, in chronic in vivo experiments it has been shown that transformed tomatoes bearing Cry 1 a ( b ) gene can be considered as safe for rats fed the product for 91 days, on the basis of their performance and the absence of an effect on the fresh weight of selected organs.
All these data have shown that expressed proteins from tDNA do not produce harmful effects in the digestive tract or elsewhere after ingestion by the animals.
However, recent studies demonstrated the presence of fragments (<199 base pairs, bp) of plant DNA, but not recombinant DNA, in animal and in particular in poultry tissues.
These data indicated that plant DNA is not completely destroyed in the digestive tract and residues can be absorbed through the intestinal wall as demonstrated. This is contrary to the commonly-held assumption that only simple, low molecular weight nutrients can cross the intestinal wall of mammals.
2.9.13.2 Plant DNA in animal products
At present, there is no specific Community regulation on GM derived feeds and their subsequent effects on the quality of animal products. Consequently, studies on the effects of feeding GM plants to farm animals on the quality of their products would not be required by the Company seeking to market a GM variety unless there was a
39

recognised cause for concern. However, studies have been made in a variety of commercial and independent laboratories to evaluate any potential effects on the quality of milk, meat and eggs produced by farm animals. Feeding dairy cows with silage made either with Bt maize or with glyphosate tolerant maize did not lead to modifications in the protein content or to the different protein fractions in milk (Table 21). These data agreed with results observed by several other authors comparing the consequences of the use of glyphosate tolerant soybean, or Bt maize compared with their isogenic parents in the diet of dairy cows. In particular, results demonstrated that the properties of milk measured in terms of clotting time and fresh cheese yield were not affected by feeding transformed maize.
Table 21. Effect of feeding dairy cows with silage prepared with genetically modified maize on milk quality.
It has been recently estimated that farm animals and especially ruminants eating forages will ingest only a small amount of nucleic acid from their diet. It has also been shown that foreign DNA fragments introduced experimentally in the stomach could cross the intestinal wall and be recovered in the liver, the spleen and even in the lymphocytes of the mice several hours after gavage. The possible transfer of plant DNA to cattle and poultry tissue and products has been intensively studied in the case of animals fed control and Bt maize (Table 22). Fragments (<200 bp) derived from plants chloroplasts could be detected in duodenal juice, then in blood lymphocytes of both groups of control and experimental dairy cows eating maize silage and a faint signal has been observed in milk. Opposed to that, no plant DNA has been found in tissues of beef cattle. But more importantly Bt DNA which was detected in duodenal juice of cows fed
Bt maize, was never found in blood, milk and any of the tissues investigated in the
40

experiment. Similar fragments from maize chloroplasts were found in all chicken and layers tissues (muscle, liver, spleen, kidneys) examined, but no fragments were detected in either excreta or eggs. Experiments conducted in the United States also confirmed the presence of plant chloroplast DNA in animal tissues of a variety of species, but no polymerase chain reaction (PCR) products indicating the presence of the gene or fragments of the gene encoding Bt were found in milk, tissues and eggs.
Table 22. Detection of plant chloroplast DNA (199 bp) by PCR in tissues of control animals or animals fed Bt maize or EPSPS soybean
Not detectable; (+): faint signal; +: positive signal; ++: strong positive signal.
It is now recognised that in vivo uptake of DNA fragments by mammalian cells after oral administration is a common occurrence. However, there is no reason to suppose that tDNA is any more likely to be absorbed than any other part of the total plant DNA. The reason that chloroplast DNA is more readily detected simply reflects the copy number of the genes involved and the sensitivity of the PCR method employed. Each vegetative cell in a transformed maize plant may contain several hundred copies of a plastid gene, but only a single copy of a Bt or Pat gene. It is inevitable, as the sensitivity of detection methods is further improved, that PCR
41

products from tDNA will be detected in milk, meat and possibly eggs. While there is also no evidence that DNA from a plant source has ever been incorporated into the mammalian genome, the presence of at least 200 genes of evident bacterial origin leaves this a possibility. However, such incorporations occur over an evolutionary time scale and are of no significance to a safety assessment.
42