OVERARCHING CONCERNS OF THE COMMITTEE
Five overarching concerns of animal biotechnology dominated discussions before the committee. The first was whether anything theoretically could go wrong with any of the technologies. For example, is it theoretically possible that a vector used for gene transfer could escape and become integrated into the DNA of another organism and thereby create a hazard? The second was whether the food and other products of animal biotechnology, whether genetically engineered or not, or from clones, are substantially different from those derived by more traditional, extant technologies. A third major concern was whether the technologies raise novel environmental issues, and a fourth was whether they raise animal welfare issues. Finally, there was concern as to whether the statutory tools of the various government departments and agencies involved are sufficiently well-defined and whether the technologic expertise and capacity within agencies are sufficient to cope with the new technologies, should they be deemed to pose a hazard. Before these issues are considered in the individual chapters that follow, the committee felt that it was important to articulate how it defines “concern”. The term “concern” is used throughout the report and is defined as “an uneasy state of blended interest, uncertainty, and apprehension.” The committee also attempts to put the new technologies—which form the focus of this report—into perspective and to discuss some of what it has learned from
past animal agricultural practices, and particularly from those technologies that have reached fruition in the past half century.
THE CURRENT STATE OF ANIMAL BIOTECHNOLOGY
Agricultural output of poultry and livestock in the United States exceeds $90 billion annually, of which around $11 billion consists of exports (USDA, 2001). There are currently about 9 million dairy cows, 5 million dairy heifers, and 85 million beef cattle and calves in the United States, and approximately 100 million hogs are slaughtered annually. However, trends in food consumption are changing. Even as the demand for red meat remains high, many consumers are changing their preferences from red meat to alternative protein sources. Americans consumed 82 pounds of chicken per capita in 2000 compared to 69.5 pounds of beef—a reversal of the situation a generation ago. Sales of farmed fish also have increased markedly as fish farming has become more productive and efficient. The main fish products traded domestically and internationally are shrimp (and prawns), Atlantic and coho salmon, and mollusks, but the market shares of tilapia, sea bass, and sea bream are increasing (Lem, 1999). Carp is, by far, the finfish type produced in largest quantity worldwide, with production about ten times that of salmon (FAO, 2000), but is primarily consumed domestically in Asian countries, rather than traded. Channel catfish constitutes the major species of finfish farmed in the United States (Lem, 1999). Per capita demand for high-quality meat and fish products is expected to increase both in response to rising world population and to improvements in the standard of living over the next 25 years (Pinstrup-Andersen and Pandya-Lorch, 1999). As a consequence of increased demand for meat and the deterioration and loss of agricultural land, there is pressure to utilize the potential for biotechnology to improve productivity in animal agriculture. As the techniques for producing transgenic animals become more efficient and as more is known about controlling how inserted genes are expressed, it is likely that the approaches soon can be integrated into agriculture. Indeed, the commercial production of transgenic fish, which is likely to occur worldwide, already is imminent.
Genetically engineered poultry, swine, goats, cattle, and other livestock also are beginning to be used as generators of pharmaceutical and other products, potential sources for replacement organs for humans, and models for human disease. The technology to produce foreign proteins in milk by expressing novel genes in the mammary glands of livestock already has advanced beyond the experimental stage, with some of the products currently in clinical trials (Colman, 1996; Murray and Maga, 1999). In theory, transgenic animals can provide milk that is more nutritious for the consumer, or that is enhanced for certain protein components that might be valuable for
manufacturing cheese or other dairy products. However, the largest investments in the technology to date have been made by pharmaceutical companies interested in producing enzymes, clotting factors, and other bioactive proteins in milk.
Companies also are interested in farm animals as possible sources of replacement organs for humans. Transplantation is an accepted and successful treatment for organ failure, but there is an enormous shortage of available human organs. As there are ethical and practical concerns related to the use of donor organs from primates, the pig, in particular, is being considered as an alternative. Unfortunately, humans express antibodies to a carbohydrate epitope (terminal 1,3-galactose residues) that is present on the surface of pig cells (Sandrin et al., 1993). As a result, the xenograft immediately becomes a target for acute rejection. To remedy this situation, pigs will be produced that lack the 1,3 galactosyl transferase enzyme (Tearle et al., 1996; Dai et al., 2002; Lai et al., 2002).
Although the mouse, because of its small size, short generation times, fecundity, and well-studied genetics has become the animal of choice for providing models for human disease, farm species might provide alternatives where the mouse is inappropriate. One possible future scenario is the creation of specific gene knockouts in farm animals in order to mimic human disease in a large animal model. For example, McCreath et al. (2000), have generated genetically-engineered sheep carrying a mutated collagen gene, and have suggested that such animals could serve as models for the human connective tissue disease osteogenesis imperfecta.
The development of such technologies and others yet to be conceived and their incorporation into agricultural and biomedical practice raises concerns about whether the end products can be consumed safely, whether there are likely to be unwanted effects on the environment, and whether animal welfare will be adversely affected. The goal of this report is to identify concerns that will aid the federal regulatory agencies in evaluating the possibility of such adverse outcomes. However, before proceeding further, it is perhaps helpful to understand what is meant by biotechnology and to appreciate how far such biotechnology already has been incorporated into current agricultural and biomedical practice. It also is clear that the concerns of the public are focused on some of the more recent technologic advances relating to gene transfer between organisms that would not normally interbreed and to assisted reproductive procedures, such as somatic nuclear cell transfer to create so-called clones (Eyestone and Campbell, 1999; Box 1.1). Many of these recent advances have not yet left the experimental stage, but it is clear that several, including transgenic finfish, which are soon likely to be commercialized, are likely to assume importance both in agriculture and medicine.
BOX 1.1 A Definition of Cloning
The verb “to clone” and the noun “clone” have a range of meanings and interpretations. The noun is derived from the Greek word klōn, meaning a twig. Its original use in English was to describe asexually produced progeny, and it has been in familiar use in horticulture for centuries. “To clone” in this context, therefore, means to make a copy of an individual. “Clone” was later adopted into the parlance of modern cellular and molecular biology to describe groups of identical cells, and replicas of DNA and other molecules. Monozygotic twins are clones, but the term has recently become popularized in the media to mean an individual, usually a fictitious human, grown from a single somatic cell of its parent. The first reports of animal cloning were in the late 1980s and were the result of the transfer to anucleated oocytes of nuclei from blastomeres (cells from early, and presumably undifferentiated, cleavage-stage embryos), a technique that is referred to as blastomere nuclear transfer or BNT, in this report. Cloning of sheep, cattle, goats, pigs, mice, and, more recently, rabbits and cats, by transplanting a nucleus from a somatic, and presumably differentiated, cell into an oocyte—from which its own genetic material has first been removed—was achieved about a decade later (Wilmut et al., 1997; reviewed by Westhusin et al., 2001), leading to the speculation that humans also could be cloned. It is important to note that somatic cell nuclear transfer (SNT) also can be used to produce embryonic stem cells, giving researchers the opportunity to obtain undifferentiated stem cells that are genetically matched to the recipient for research and therapy, which is independent of the discussion here regarding the use of SNT for reproductive cloning of animals. Neither BNT nor SNT result in an exact replica of an individual animal, although the progeny are very similar to each other and to their donor cell parent. Any genetic dissimilarity is likely due to the cytoplasmic inheritance of mitochondria from the donor egg, which possesses its own DNA, and to other cytoplasmic factors, which seem to have the potential to influence the subsequent “reprogramming” of the transferred somatic cell genome in such a way that spatial and temporal patterns of gene expression in the embryo are affected as it develops (Cummins, 2001; Jaenisch and Wilmut, 2001). For these reasons, many scientists have objected to the use of the term clone in the context of somatic cell nuclear transfer. The committee acknowledges this shade of meaning and has attempted to make the appropriate distinction when the term clone is used. Nevertheless, clone is now so widely accepted as a synonym for somatic cell nuclear transfer—not just by the public at large—but also by embryologists and other biologists, that the committee has retained it rather than attempt to replace it with a more precise, but cumbersome, phrase.
THE ORIGINS OF BIOTECHNOLOGY IN ANIMAL AGRICULTURE
Biotechnology literally is technology based on biology; it is the application of scientific and engineering principles to the processing or production of materials by biologic agents to provide goods and services. The application of biotechnology to animals has a long history, beginning in Southwest Asia after the last ice age, when humans first began to trap wild animal species and to breed them in captivity, initially for meat and fiber and later for transport and milk. Of the approximately 48,000 mammalian species, fewer than 20 have been successfully domesticated (Diamond, 1999). Other than cats and dogs, only five of these species (cattle of the Bos genus, whose ancient
ancestor is the now extinct auroch; sheep derived from the Asiatic mouflon species; goats, which are descended from the benzoar goat of West Asia; pigs derived from captured wild boars; and horses, which originated from now extinct wild horses that roamed the steppes of Southern Russia) are found worldwide (Diamond, 1999; Box 1.2). As pointed out by Hale (1969) and Diamond (1999), the animals that have been successfully domesticated and farmed share and exhibit a unique combination of characteristics. They are relatively docile, are flexible in their dietary habits, and can grow and reach maturity quickly on a herbivorous diet, and breed readily in captivity. They also have hierarchical social structures that permit humans to establish dominance over them, and are adapted to living in large groups. They do not include species that generally have a tendency to be fearful of humans or disturbed by sudden changes in the environment. Our ancestors no doubt based their selection methods for improving their herds and flocks on how easy the animals were to farm, as well as on potential agricultural value. In turn, the animals are adapted to thrive in a domesticated environment.
BOX 1.2 Progression of Technologies Incorporated into Modern Animal Agriculture1
Vaccinations and other health technologies2
Freezing of semen4
Sire testing and selection5
Use of antibiotics in feed to increase gain6
Embryo splitting and cloning from blastomeres8
In vitro maturation/in vitro fertilization of oocytes and in vitro culture of resulting embryos
Use of hormones to control ovulation in farm animals and to induce spawning in fish zygotes9
Hormonal sex reversal and production of monosex fish stocks
Chromosome set manipulation10
Steroid administration to improve weight gain
Bovine somatotropin (BST) to increase milk production in dairy cows
The fact that the modern breeds of these species differ so markedly from their progenitor species is a reflection of how quickly directed breeding can act. The modern Holstein, which dominates the contemporary United States dairy industry, little resembles its ancestors of only a half-century ago. Milk production per cow increased almost threefold between 1945 and 1995 (Majeskie, 1996), largely as a result of breeding from select bulls. There has been an accompanying drop in the number of cows, land devoted to dairy production and in manure produced. On the downside, the cows have a tendency towards lameness, are considerably less fertile than in the 1940s, and are frequently maintained in a herd for no more than 2–3 years or 2–3 lactations (Pryce et al., 2000; Royal et al., 2000), and represent a very narrow genetic lineage (Weigel, 2001). The export of these animals and their lineages to Europe and elsewhere is assuring the globalization of both the benefits and drawbacks of the American Holstein. Analogous changes are ongoing in the swine industry, where the pressure to produce lean, fast-growing animals of uniform size is leading to the abandonment of old breeds (Notter, 1999). Paradoxically, unless the old livestock breeds are eaten, sheared or milked, they will not survive.
The dog (Canis familiaris), on the other hand, provides an interesting example of the range of phenotypes that can be derived by selection within a single species. Dogs are believed to have originated in several separate domestications from wolves (Canis lupus and Canis rufus) and coyotes (Canis latrans) before the domestication of livestock. They have undergone remarkable modifications in size and behavior over short periods of intense selection and to provide the diversity observed in modern breeds. This reflects the enormous pool of genetic variation within the species (Wayne and Ostrander, 1999), but (possibly) also the fixation of new mutations into different genetic lineages. Inbreeding of dog breeds, as of domestic livestock, has led to a major narrowing of intrabreed variability (Zajc et al., 1997).
The same kinds of selective pressures that molded the large farm animal species has led to the creation of the modern breeds of farmed fowl, which include chickens, ducks, geese, and turkeys domesticated for their meat, eggs, and feathers. As in the dairy industry, there has been a remarkable improvement in the productivity of the poultry industry over the last 60 years. Between 1940 and 1994, yearly egg production per laying hen increased from 134 to 254, mainly as a result genetic selection. The broiler industry has shown similar
gains (Pisenti et al., 1999). In 1950, a commercial bird took 84 days to reach a market weight of 1.8 kilogram. By 1988, this market weight was reached by only 43 days (Pisenti et al., 1999) on about half the amount of feed (Lacy, 2000).
Scientific breeding, combined with better nutrition and veterinary care, clearly has produced breeds of animals that are remarkably productive, although sometimes strikingly different in habits and appearance from those farmed early in the twentieth century. The practice has also led to a loss of many breeds of livestock and fowl, and a decline in genetic diversity within the breeds that survive. For example, it has been estimated that there were several hundred specialty lines of chicken in North America at the beginning of the last century, whereas the number of commercial hybrid strains now available through suppliers is fewer than 10 (North and Bell, 1990).
Aquatic animals, including finfish and shellfish, now are farmed, and specific breeds that have been selected for growth and other traits are established now in the largest industrial sectors of aquaculture, such as channel catfish, rainbow trout, and Atlantic salmon. The growth and quality of such animals are also amenable to genetic engineering through modern biotechnology. Genetically engineered or highly selected aquatic species present special problems in terms of confinement, as the features that might make them attractive commercially might pose risks to the genetic base of their wild relatives with which they can interbreed (Hallerman and Kapuscinski, 1992b).
Insects also have been domesticated for farming. The two best-known examples are the honeybee and silkworm; considerable genetic gains in productivity have provided strains of these insects far removed from the ancestral species from which they were derived. Attempts to develop strains of honeybee with improved resistance to pathogens and silkworms that produce proteins other than silk are on the horizon. Insects, like fish, are especially difficult to confine so that “escapes” are almost inevitable. In addition, insects, including ones that can be engineered transgenically, are likely to continue to be used as part of biocontrol programs for pest insects and invasive plant species and, as such, might be intentionally released into the environment. There will almost certainly be attempts to replace or to infiltrate native populations with insects that have been engineered in such a manner that they are less of a pest or unable to transmit pathogens (Hoy, 2000). Private-sector companies already have begun to farm recombinant proteins (antibodies, cytokines, enzymes, and bioactive peptides) from insect larvae. Whereas the U.S. Department of Agriculture’s Animal and Plant Health Inspection Service (APHIS) regulates the release of insects for pest management, it is unclear which agency is responsible for protecting against accidental release of insects from mass rearing factories. Horizontal gene transfer, disruption of ecosystems, and native species extinctions are among the potential hazards that arise from permanent releases of transgenic arthropods into the environment (Hoy, 2000).
BOX 1.3 Examples of Technologies that are Experimentally Established but not Yet in Widespread Use in Animal Agriculture
Production of sexed semen
Production of transgenic animals by direct gene transfer
Production of transgenic animals through genetic engineering of sperm
Cloning of adult animals by somatic cell nuclear transfer to produce “copies”
Cloning of animals by somatic cell nuclear transfer to achieve genetic engineering
The traditional kind of biotechnology emphasized at the beginning of this section relies upon natural breeding procedures to select valuable phenotypes from the variation in the existing gene pool of a species and is beyond the purview of this report, even though it has contributed so successfully to modern-day production agriculture. It is firmly entrenched in our agricultural communities, and many are generally conversant with its benefits and risks. Importantly, other forms of research-driven biotechnologies, based on improved insight into reproductive physiology and endocrinology, embryology, genetics, and animal health also have made their way into standard farming practices over the last 75 years (Box 1.2). A few of the procedures listed extend the boundaries of biotechnology to the development of organisms that have a combination of traits generally not attainable in nature through conventional breeding and are not themselves without controversy. Some of those listed are perceived by both scientists and lay people as endangering human health or as adversely affecting animal welfare or the environment. Certain of the technologies even can have unintended, long-term consequences on the economics of agriculture itself. Finally, some of the concerns raised about the technologies in Box 1.2 are quite relevant to those listed in Box 1-3. Although several of these technologies remain experimental and have not yet become a part of standard agricultural practice, others (e.g., commercialization of transgenic fish) are undergoing government review for commercial approval. It is these newer technologies on which this report is focused. For these reasons, it is worthwhile discussing Box 1.2 and some of the issues that these technologies have raised before moving on to the ones associated with Box 1.3.
CONCERNS REGARDING EXTANT TECHNOLOGIES
There are well-established guidelines for the application of technologies that maintain animal health, such as standard vaccination against viral and bacterial diseases. Indeed considerable efforts are being made to expand the
range of such technologies in order to prevent epidemic spread of disease in flocks and herds, which are particularly at risk when farmed under intense conditions (BBC, 2001). Even the therapeutic use of antibiotics to treat animals that have bacterial infections or are in danger of becoming infected seems not in itself to be controversial, except when antibiotics of medical importance to humans are employed.
Subtherapeutic Use of Antibiotics
The U.S. Food and Drug Administration (FDA) approved antibiotics as feed additives for farm animals in 1951. Their use since has been extended to fish farming, particularly with the global spread and dramatic increase of aquaculture in tanks and pond-like structures where antibiotics are used for prevention and control of disease rather than to enhance growth (NRC, 1999). The treated animals are found to grow more quickly and utilize feed more efficiently than animals on regular feed. At least 19 million pounds of antibiotics are used annually for subtherapeutic purposes in animal agriculture, and generally are added to feed and water (NRC, 1999). Some of these compounds, used on livestock, including penicillin, tetracycline, and fluoroquinolone used on livestock, also are prescribed to treat human illnesses, and the practice has been shown in a few instances to contribute to antibiotic resistance of human pathogens (Chiu et al., 2002; Molbak et al., 1999). It now is generally accepted in the scientific and medical communities that antibiotic resistance can be exacerbated by the widespread improper use of antibiotics. What remains controversial is whether agriculture contributes sufficiently to the problems associated with resistant pathogens to justify a complete curtailment of their use as growth promoters (DANMAP, 2000; Stephenson, 2002). A recent report from the National Research Council (NRC, 1999) failed to find a definitive link between the agricultural use of antibiotics in animal feed drinking water and antibiotic resistance of human pathogens. The report states, “The use of drugs in the food production industry is not without some problems and concerns, but does not appear to constitute an immediate public health concern.” Since that report was released, additional information, raising further concerns, has been released (Fey, 2000; Gorbach, 2001). Consequently, the practice remains under intense scrutiny and is opposed by some scientific and medical organizations.
Assisted Reproductive Procedures
Artificial insemination (AI), and the later, associated use of frozen semen, sire testing and sire selection are all part of a combinatorial approach to improve the genetic quality of farmed species. AI, when first introduced into agriculture,
elicited an enormous outcry from farmers, the press, and religious groups. It was claimed to be against the laws of God, a repugnant practice that would lead to abnormal outcomes, and economically unsound (Herman, 1981; Foote, 1996). It gradually has become an accepted practice in agriculture, as well as in human and veterinary medicine. The ability to freeze semen and maintain a high degree of fertilizing ability after thawing extended the power of AI, since a few select bulls could be utilized to inseminate many females in different geographic areas. Such bulls could be tested, not only for fertility, but also for their ability to sire progeny that produced copious amounts of milk. By maintaining accurate records, breeding value estimations of particular bulls could be calculated. The result was the remarkable increase in milk production, noted earlier. On the other hand, the process is leading to potentially destructive inbreeding since many of the select bulls are related. Inbreeding coefficients among modern Holsteins and Jersey breeds are now about 5 percent and rising (Weigel, 2001). The outcome might be inbreeding depression and broad susceptibility to the epidemic spread of disease. There also has been a remarkable recent loss of fertility, with successful pregnancies resulting from first insemination dropping from more than 40 percent to as low as 20 percent or less in some herds as milk yields have risen (Pryce et al., 2000; Royal et al., 2000).
Embryo recovery and transfer provides the opportunity for a particularly valuable animal to parent many more offspring in her lifetime than would be otherwise possible (Seidel, 1984). The embryos also can be frozen and then either stored or transported before they are used to initiate a pregnancy. It is a relatively common technology and has been used to produce an estimated 40,000 to 50,000 thousand beef calves every year (NAAB, 1996). The approach is to induce, by using hormones, the maturation and release of more than a single egg from the ovaries (superovulation; Driancourt, 2001). Then, the animal usually is inseminated with semen from an equally select bull, and the embryos are collected and transferred individually, or in pairs, to the reproductive tract of less valuable cows, which carry the calf to term. Modern technologies also provide the possibility of freezing the embryos and determining their gender prior to transfer. The main concern with this technique, as with the AI-associated technologies discussed above, is that it can lead to narrowing of the genetic base of the breed, in this case involving both parents. A related technique is to use a needle to aspirate immature oocytes from the ovaries (in the case of livestock the oocytes often are taken from slaughtered animals at an abattoir) and to mature the oocytes for about one day in a culture containing hormones. At the stage when the oocytes reach a point midway through the second division of meiosis, they are fertilized with live sperm. In rare instances, fertilization is achieved by a single sperm or sperm head, which is injected through the tough outer zona pellucida of the oocyte, either beneath the zona or directly into the cytoplasm (intracytoplasmic injection, or ICSI). Whatever method is used for fertilization, the resulting zygotes usually are then cultured
until the embryo reaches a more advanced stage of development. In humans, of course, these combined techniques form the basis of in vitro fertilization procedures and have resulted in hundreds of thousands of normal infants, but the techniques also have become an important means of producing embryos for experimental purposes in agricultural research (First, 1991). Importantly, in vitro maturation of oocytes underpins cloning and transgenic technologies (see Chapters 2 and 6), where large numbers of competent, matured oocytes are needed to provide the many eggs necessary for nuclear transfer and pronuclear injection, respectively (see Chapter 2). In vitro fertilization also is used commercially to preserve the genome of particularly valuable animals that have infertility problems such as blocked oviducts or that respond poorly to superovulation (Boland and Roche, 1993), a technique described below. This commercial application of IVF is a relatively uncommon, with about 4,000 calves born from its use annually (NAAB, 1996). Few concerns have been raised about this technique, which essentially is identical to that employed for in vitro fertilization in humans, although some animal welfare issues have been raised (Chapter 6).
In order to manage breeding programs more intensively, control over the reproductive cycles of livestock by hormonal intervention has increased. In general the technologies are relatively benign and involve injecting the animal with hormones, usually to stop progression through the existing estrous cycle and sometimes to mimic the events that lead to selection of one or more mature follicle(s) that will ovulate. Superovulation is a technique designed to mature a cohort of follicles simultaneously, with result that several eggs are ovulated simultaneously (Nebel and Jobst, 1998; Britt, 1985). A hormone treatment analogous to that used to produce a timed ovulation in the large farm animals is used to induce gonadal maturation in fish (Mittelmark and Kapuscinski, 2001). None of these techniques have raised public health concerns, since the hormones are similar or identical to those in normal reproduction and the amounts used within the physiologic range.
Splitting or bisecting embryos became an esoteric but well-established practice in the 1980s in order to provide zygotic twins (or, in modern parlance, clones; Boland and Roche, 1993; Heyman et al., 1998). The pieces of the embryo—usually “halves,” which are genetically identical in terms of both their nuclear and mitochondrial genes (see Box 1.1)—are placed in an empty zona (the protective coat around early embryos) before being transferred to different recipient mothers to carry them to term. It is estimated that only a very small number of the calves (1 to 2 percent of those resulting from embryo transfer in the United States and Canada) are produced in this manner (NAAB, 1996). Nevertheless, these animals have been introduced into commercial herds, and have produced progeny; their milk and meat are consumed by the public.
Cloning by nuclear transplantation from embryonic blastomeres (blastomere nuclear transfer, or BNT; see Box 1.1) is an expensive procedure
that also has its origins in the 1970s (Willadsen and Polge, 1981; Willadsen, 1989). What distinguishes it from somatic cell nuclear transfer, the technology that led to the creation of Dolly and much of the controversy over human cloning, is the stage of development at which the nuclei are transferred (Wilmut et al., 1998). In the older procedure, the cells or blastomeres used were from the so-called morula stage of cell development (although some were from the cleavage stage and others from the blastocyst stage) when the embryo still is an undifferentiated mass and its cells presumed still capable of forming all tissues of a fetus.
The cloning technologies of embryos splitting (EMS) and embryonic nuclear transfer (NT) were introduced into dairy cattle breeding in the 1980s. The Animal Improvement Programs Laboratory of the USDA’s Agricultural Research Service (ARS) is responsible for tracking the performance of dairy cattle throughout the U.S. Recently, working with the Holstein Association, they evaluated the performance of cloned Holsteins produced by EMS and NT (H.D. Norman, USDA–ARS, personal communication). The numbers of EMS and NT clones were documented by gender and birth year. All NTs were from embryos rather than adult cells. Through 2001, there were a total of 2,226 EMS (754 males and 1,472 females) and 187 NT (61 males and 126 females) Holstein clones registered. Of female EMS clones, 921 had yield records, and 551 had noncloned full siblings with yield records. Of the 126 female NT clones, 74 had yield records, but only 11 had noncloned full siblings. These familial relationships were used to compare the performance of cloned and noncloned full siblings for standardized traits and genetic evaluations as part of the national evaluation program. These standardized traits included total milk yield, fat content (by weight and pecent), protein content (by weight and percent), somatic cell score, and productive life (in months). Also calculated were yield from contemporaries and predicted transmitting ability. Norman and his colleagues concluded that the numbers of clones have decreased for EMS males and for all NT clones over the past decade. Animals that were selected for cloning were slightly superior genetically to the contemporary population mean for yield traits; the yields of NT clones were similar to, and those of EMS clones were slightly less than, those of their noncloned full siblings.
“Modern” cloning involves taking an unfertilized egg, removing its chromosomes, and introducing the nucleus from a differentiated cell of the animal to be cloned, which is frequently an adult (Box 1.1; Wilmut et al., 1997; Polejaeva et al., 2000; Kuhholzer and Prather, 2000). The introduced nucleus is reprogrammed by the cytoplasm of the egg and directs the development of a new embryo, which is then transferred to a recipient mother to allow it to develop to term. The offspring formed will be identical to their siblings and to the original donor animal in terms of their nuclear DNA, but will differ in their mitochondrial genes and possibly also in the manner their nuclear genes are expressed or biochemically engineered (see Box 1.1 and Chapter 2). Cloning
from blastomeres, the older of the two procedures, has been reported to result occasionally in large calves (and lambs), the so-called large offspring syndrome (LOS; Young et al., 1998; Sinclair et al., 2000). Analogous, though possibly more serious, abnormalities might be associated with cloning from somatic cells and are discussed further in Chapters 2 and 6 of this report.
Among the most contentious technologies used in animal agriculture is the use of steroid hormones to increase the rate of weight gain and to reduce accumulation of fat deposits of young heifers and steers as part of the “finishing” process prior to slaughter (Heitzman, 1976; Lammers et al., 1999). The steroids are administered by slow release from a plastic implant embedded beneath the skin of the ear, which provides “physiologic” circulating levels of the hormone in the bloodstream. The hormones used are mainly Zeranol (in Ralgro™), a naturally occurring fungal metabolite (zearalenone) with estrogenic action; estradiol, progesterone, and testosterone, or mixtures of these steroids (in various Synovex™ formulations); and trenbolone (Doyle, 2000). Concern about these hormones is probably, in part, a legacy of diethylstilbestrol, which was eventually banned from use in the poultry and beef industry because of its adverse effects on humans. However, the amounts of present-use compounds consumed from meat derived from treated cattle are small, and numerous scientific studies generally have indicated that these residues exist at such low concentrations that they pose little risk to consumers (Doyle, 2000; Lange et al., 2001; United States Mission to the European Union, 1999; Henricks et al., 2001), provided good veterinary practices are employed (e.g., using the correct number of implants and placing implants correctly in the ear cartilage), although the U.S. Geological Survey has recently documented the presence of hormones in a number of streams and rivers (some of these hormones likely come from implants; Kolpin et al., 2002). Despite the scientific evidence for safety, the European Union implemented a ban on U.S. beef imports, valued at over $100 million per year in 1989 (Andrews, 1997). A concern that has not been extensively examined so far is whether these hormones pose any sort of environmental threat through their leaching into soil and water. For example, two recent studies have shown that a commonly used androgenic growth promotor—trenbolone—has been found in groundwater near cattle feedlots, and that this growth promotor has androgenic effects (Gray, et al., 2001; Schiffer, et al., 2001).
The use of bovine somatotropin (BST) to increase milk yield from dairy cows has had a similar checkered history and is the subject of trade disputes. Currently banned in Europe even for experimental studies, BST was approved by the FDA for use in U.S. dairy cattle in 1993 because testing had revealed no concerns regarding consumer safety (Juskevich and Guyer, 1990; Bauman, 1999). The Monsanto product, Posilac™, now is widely used throughout the U.S. dairy industry, where milk production can be increased as much as 30 percent in well managed, appropriately fed herds, without adversely affecting the quality or composition of the milk. The BST, which is almost indistinguishable in sequence from the natural hormone, is present in low concentrations in milk, but has no biologic activity in humans. The level of IGF-1, the hormone induced by BST, is somewhat elevated but within the “physiologic range” for cows and is probably digested along with other milk proteins in the adult stomach, although it might have biologic activity in the intestine of neonates (Burrin, 1997). In its assessment, the FDA did not report that BST or IGF-1 pose any risk either in humans or animals that consume cows’ milk. As with other technologies that increase productivity, a concern frequently raised is why more milk is needed when the developed world appears to have more than enough of the product. One answer is that increased productivity translates into fewer animals, producing less waste and utilizing less land—an extremely important consideration for future land management use. The greatest concerns about BST are probably in the area of animal welfare. High-yield milking cows show a greater incidence of mastitis than lower-producing cows, but studies have shown that mastitis is not exacerbated by BST administration (Judge et al., 1997). Another concern—a practical one for the dairy industry—is a recent trend to breed heifers only once and then to sustain milk production for as long as 600 days by using BST. Lengthening lactation via BST in second calf and older cows is a larger contributor to having fewer calves per lifetime in the herd than first-calf heifers. The result has been a shortage of replacement heifers for producers, since only one calf is born during the milking life of the animal (Harlow, 2002).
Marker-assisted selection involves establishing the linkage between the inheritance of a particular trait—which might be desirable, as in the case of milk yield—or undesirable, as in susceptibility to a disease, with the segregation of particular genetic markers. Thus, even if the gene that controls the trait is unknown, its presence can be inferred from the presence of the marker that segregates with it. This technology, which is particularly important for studying
complex traits governed by many genes, has only recently become a factor in animal breeding and selection strategies (Georges, 2001). Its use likely will increase exponentially as the industry incorporates the data from the various genome sequencing projects and as the density of useful, segregating markers increases on the chromosomes of the species. Initially, animals will be screened for genes that control simple traits, such as horns, which are undesirable in cattle, and halothane sensitivity, which segregates with metabolic stress syndrome in pigs. With time, easily identifiable markers will be chosen that accompany the many genes controlling more complex traits such as meat tenderness and taste, growth, calf size, and disease resistance. The approach has enormous potential for improving the quality of agricultural products, disease resistance, and other traits but could be misused (Dekkers and Hospital, 2002). For example, stringent selection of prime animals could potentially narrow genetic diversity even more than is evident at present. Use of the technique also could maximize short-term gain in productivity but at the expense of longer-term improvement due to what has been termed polygenic drag (Dekkers and Van Arendonk, 1998; Dekkers and Hospital, 2002). In essence, the cumulative effect of genes with effects too small to be exploited in a marker-assisted selection program could contribute more to increasing desired traits than genes with major effects. However, marker-assisted selection might be a powerful measure to counter inbreeding by providing genetic measures of heterozygosity, encouraging breeding strategies that maintain diversity at the majority of sites in the genome, and allowing the genetic potential of rare breeds and wild ancestors to be utilized and incorporated into mainstream agriculture.
Chromosome Set Manipulation in Mollusks and Finfish
Altering the chromosome complement of an animal can be a useful way of rendering that animal infertile, and is exploited widely in the production of fish and mollusks. Well-timed application of high or low temperatures, certain chemicals, or high hydrostatic pressure to newly-fertilized groups of eggs can interfere with extrusion of the second polar body (the last step in meiosis), resulting in “triploid” individuals with three, instead of the usual two, chromosome sets (e.g., for oysters; Allen et al., 1989). A later treatment can suppress the first cell division of the zygote, resulting in “tetraploid” individuals with four sets of chromosomes. Crossing tetraploids, which are fertile in some species, with normal diploids can then produce large numbers of triploids (Scarpa et al., 1994). Such chromosome set manipulations have been applied to cultured marine mollusks to produce confined stocks of triploids that are unable to reproduce. This application is of particular importance, as some of the shellfishes most suited to aquaculture are not indigenous to a given area and can pose ecologic risks to native species should they or their larvae escape
confinement and enter natural ecosystems (USDA, 1995). Induction of triploidy reduces the likelihood that an introduced species would establish self-sustaining populations, because such animals are theoretically sterile. For example, the triploid Suminoe oyster (Crassostrea ariakensis) is being assessed for oyster production in the Chesapeake Bay, where diseases complicate restoration of the native Eastern oyster (C. virginica). Should triploidy prove an effective means for reproductive confinement, culture of sterile Suminoe oysters could support the recovery of the declining Chesapeake oyster production industry.
Another benefit of producing sterile mollusks is in maintaining product quality throughout the year. The meat quality of oysters is high just before they spawn, but low after spawning. The product quality of reproductively sterile, triploid oysters remains high year-round. Hence, triploid stocks of Pacific oyster (Crassostrea gigas) provide a tangible benefit to aquaculturists, and now make up almost half of commercial production in the Pacific Northwest.
Unfortunately, repeatable induction of 100 percent triploidy on a commercial scale poses a considerable technical challenge. Non-triploid larvae within batches of larvae easily can go undetected if their frequency is low (USDA, 1995). Should triploidy be desired for purposes of maintaining product quality and the species is indigenous to an area, no harm is posed. If, on the other hand, triploidy is to be utilized for reproductive confinement purposes, the presence of reproductively fertile individuals—even in low numbers—might establish progeny and a self-sustaining population. There also are indications that a small percentage of triploid oysters can progress to a “mosaic” state, with diploid cells arising within the background of triploid cells, leading to the possibility that they could produce viable gametes (Calvo et al., 2001; Zhou, 2002).
Triploidy often has been used to reduce the likelihood that introduced finfish species would establish self-sustaining populations. Use of all-female triploid stocks has been suggested as a means of achieving reproductive confinement of transgenic fishes, including Atlantic salmon (the leading candidate for commercialization). As with mollusks, however, repeatable induction of 100 percent triploidy poses a considerable technical challenge, and commercial net pen operations produce hundreds of thousands of salmon, with many escaping (Hallerman and Kapuscinski, 1992b; Carr et al., 1997; Fiske and Lund, 1999; Volpe et al., 2000).
Another technology used on finfish is to farm monosex fish stocks (Beardmore et al., 2001), which are preferred by producers either because one gender grows faster or larger than the other (e.g., males in catfish and tilapia, females in rainbow trout), or because certain species (e.g., tilapia) attain sexual maturity before reaching harvest size. Monosex populations have been established in several ways, but most reliably through hormone-induced gender reversal. All-male fry can be produced by direct administration of testosterone in feed, or all-females by administration of estrogens. Monosex stocks also can be
produced indirectly by gender reversal and progeny testing to identify XX males for producing all-female stocks, as in trout (Bye and Lincoln, 1986) and salmon (Johnstone and Youngson, 1984), or YY males for producing all-male stocks, as in tilapia (Beardmore et al., 2001).
LIMITS OF THE REPORT
The above examples illustrate that a spectrum of earlier biotechnologies already has become integrated into agricultural practice. The introduction of new technologies does not mean that there are no concerns or even dangers posed by their use, or that there is universal acceptance among the public. The experience of the last 50 years, if nothing else, illustrates that there must be continued vigilance even after technologies have been approved. Conversely, it should be recognized plainly that increases in agricultural efficiency brought about by new technologies, such as those discussed above, undoubtedly have contributed to a more abundant, cheaper, more varied and lower cost food supply, and to enormous savings in agricultural land use.
Some technologies in Box 1.2 bridge the gap between what is an already established commercial practice and what is new (Box 1.3). For example, cloning from blastomeres (Box 1.1) in reality is little different from nuclear transfer from somatic cells, listed in Box 1.3, except that the transferred nuclei might not have to be so extensively reprogrammed in the cytoplasm of the recipient oocyte. Similarly, chromosomal set manipulation remains partly experimental and partly an active commercial technology.
Box 1.3 is a partial list of technologies that either are very close to being commercially available (pending approval from regulatory agencies) or are predicted to emerge from experimental to commercial use quite soon. The first one listed, the production of single sex sperm, is achieved through a cell sorting procedure that depends upon the higher DNA content of female sperm (Johnson, 2000; Lu et al., 1999). The technology is not expected to raise any new concerns and, provided the procedure can be scaled up, is likely to be highly beneficial in the dairy industry, where there is a surfeit of low-value bull calves, and to the beef industry, where males have a higher production value than females. The remaining technologies, however, might be more worrisome to the public and to the regulatory agencies, and it is these that are addressed in this report.
In terms of the types of technologies discussed, the scope of the report had, of necessity, to be limited. Three criteria are emphasized in this report:
The first criterion is immediacy of technologic commercialization, particularly if the products already are impinging on the regulatory system. It is clear that some of these technologies (e.g., commercial production of transgenic finfish) already are beyond the experimental stages of development. In addition, some biopharmed
drugs are in Stage 3 clinical trials and decisions must be made soon about the disposition of the livestock involved.
A second criterion is the potential impact of the technology. Some new procedures seem unlikely to raise concern (e.g., the sperm sexing discussed in the previous section) or represent relatively minor changes in practice. Other technologies might be broadly adopted, yet the possible harm they could cause and the overall benefits to society are difficult to evaluate.
A third criterion is whether there is sufficient information available about the technology to evaluate concerns properly. Indeed, the committee explicitly acknowledges that there are uncertainties associated with the application of each of the technologies discussed in this report. Unresolved scientific uncertainty interferes, not only with attempts to determine how best to apply emerging technologies to animals, but also how to predict the impacts of their application. Some hazards (see Box 1.4) remain theoretical, uninvestigated, poorly characterized, or even unknown. Such uncertainties present significant challenges to scientists and policy makers who wish to estimate the likelihood and distribution of harms and benefits resulting from application of those technologies. For example, some outcomes of applications of the technologies listed in Box 1.3, such as production of transgenic animals by gene transfer, are very difficult to predict. Uncertainties range from mere inexactness and unreliability to those that are fundamentally unknowable a priori (Funtowicz and Ravetz, 1992). Clearly, technologies that pose high stakes and high uncertainties pose fundamentally different challenges than those posing low stakes and little uncertainty. For this reason, for each concern discussed in this report, the committee has attempted, where possible, to specify (1) what is known, (2) the certainty with which it is known, (3) what is not known, (4) what is suspected, and (5) the limits of the science.
The committee also recognizes that there likely are either species or categories of species of animals not discussed specifically regarding concerns associated with biotechnology. Two examples of categories include companion animals and wildlife. While there are likely to be unique concerns that emerge with both categories, the concerns identified in the report regarding applications of the technologies (Chapter 2), environmental issues (Chapter 5), and animal welfare issues (Chapter 6) are all relevant and should be included in any considerations of wildlife and companion animal species.
BOX 1.4 Harms, Hazards, and Risks
The charge of the committee was to identify, but not to quantify, risk issues concerning products of animal biotechnology, and to provide criteria for selection of those risk issues considered most important that need to be addressed or managed for the various product categories. In order to provide criteria for selection of risk issues, it is important to understand how risk is determined. As outlined in Chapter 5 and as set forth by NRC (1983; 1996), a hazard: is an act or phenomenon that has the potential to produce harm, and risk is the likelihood of harm resulting from exposure to the hazard. This committee used the NRC (1996) definition of risk to develop a set of working steps to prioritize concerns. Because risk is the product of two probabilities: the probability of exposure, and the conditional probability of harm given exposure has occurred, the steps in risk analysis are to: (1) identify the potential harms, (2) identify the potential hazards that might produce those harms, (3) define what exposure means and the likelihood of exposure and 4) quantify the likelihood of harm given that exposure has occurred. (The committee notes that risk analysis in other fields can and does include additional steps in risk assessment; see Kapuscinski, 2002). Multiplying the resulting probabilities then was used to prioritize risk. While absolute probabilities are difficult to determine at this time, relative rankings from high to low are possible based on available evidence for each category. The risks, harms, and hazards are different for each chapter because the issues are different (i.e., a hazard resulting in an animal wellbeing concern might not be an environmental or human health concern).
Discussion of concerns regarding impacts of GE mice on the environment and human health also are limited in this report for several reasons. GE mice are not part of the animal production system for human food, and laboratory mice are highly unlikely to escape the confines of animal facilities because of their economic value and the generally high-quality care given to laboratory rodents. While mice might be a high risk for escape, might feralize easily, and might carry many different transgenes, the functionality of the transgenes used in mice rarely has been for a construct that will increase fitness in natural environments. Thus, the overall risk for most constructs is expected to be low. If mice were developed to be resistant to pest control measures (pesticides) or to be more disease resistant, then risks would be much higher. However, the use of mice in this way seems quite unlikely.