4

Likelihood of Heritable Genetic Modifications Presenting Harms to Food Animals or Humans

INTRODUCTION

Heritable genetic modification (HGM) of food animals involves use of molecular technologies such as gene transfer or genome editing to introduce desirable traits into organisms (see Chapter 1). Most of our foods, because of their long history of use, are generally recognized as safe. The absolute likelihood of harm being realized to an animal with a heritable genetic modification (an HGM animal) or to the consumer of a food derived from an HGM animal cannot be directly estimated. However, the likelihood of harm being realized relative to an appropriate comparator can be estimated to assess animal safety or the safety of foods derived from HGM animals. As discussed below, the estimation of likelihood of harm being realized from consumption of a food from an HGM animal as practiced to date has been based upon comparison of proximate analysis or more detailed analysis of amino acid composition, fatty acid composition, levels of key hormones, known allergens, bioactive compounds, and other markers between the HGM animal-derived food and an appropriate comparator. The comparison should not only include the mean value for a component, but also the observed range of that component for that food; consuming foods from HGM animals found outside the range may be more likely to result in a harm than consuming those that were conventionally produced. As noted below, other approaches might also be applied to assess the likelihood that biological mechanisms presenting harm to the HGM animal or to the consumer of an HGM animal product (Chapter 3) would become realized.

RISK ANALYSIS: KEY CONCEPTS AND APPLICATION TO HGM ANIMALS AND DERIVED FOOD PRODUCTS

Risk analysis was developed in the context of assessing the consequences of exposure to hazardous materials such as toxic chemicals, and the conceptual approach has been extended to other applications. In the context of HGM animals, risk analysis would include considering how the application of HGM technology might affect the health of an animal or the qualities of a food derived from an HGM animal such that a defined harm to the animal or the consumer becomes realized and then estimating the associated risk.

Key Concepts of Risk Analysis

Consideration of risk posed by HGM of food animals or by consumption of foods from HGM animals must be based on an understanding of key concepts of risk analysis. Risk analysis involves both risk assessment and risk management. Risk assessment is the use of a factual base to define the health effect of exposure of individuals or populations to hazardous materials (NRC, 1983). Historically, risk assessment was developed for estimating cancer risk posed by small, genotoxic molecules (NRC, 1983). Taking into account both scientific and other factors, risk management is the process of weighing policy alternatives and selecting the most appropriate regulatory action (NRC, 1996). The National Research Council assumed “organizational separation of risk assessment from regulatory decision-making” and noted that risk management “requires the use of value judgments on […] acceptability or risk […] and reasonableness of cost of control.” (NRC, 1983). Within this larger risk-analysis context, Understanding Risk: Informing Decisions in a Democratic Society (NRC, 1996) provided important definitions of hazard and risk, as presented below. The National Research Council (NRC, 2002a) pointed out that risk analysis for transgenic plants fulfills two distinct roles: (1) technical support for regulatory decision making, and (2) establishment and maintenance of regulatory legitimacy. These roles have parallels in the context of oversight of the development and use of HGM animals.

Risk Assessment

Risk assessment involves identifying harms and hazards, assessing the probability of exposure to the hazards, and assessing the probability of harm being realized given exposure to the hazards. In the context of the committee’s charge, this requires identifying potential harms to the HGM animal, harms to the consumer of foods derived from an HGM animal, or other defined endpoints that may emerge through other risk pathways. Conduct of the risk assessment process may result in either a qualitative or quantitative estimate of risk that reflects the likelihood of harm becoming realized as a result of taking the proposed action (in this case, regulatory approval of an HGM animal application). Consideration of risk must be based upon clear definitions of key underlying concepts of harm, hazard, exposure, and risk.

Within the context of risk assessment, a harm is a negative outcome that is realized by exposure to a hazard. In the context of food safety assessment, harm would be a negative outcome experienced by the HGM animal (e.g., reduced health) or the consumer of a product of the HGM animal (e.g., allergic response). Biological mechanisms that may present harms to the HGM animal or consumer of its derived food products are considered in Chapter 3.

A hazard is a substance or agent that, upon exposure, might result in a defined harm. In the context of HGM animals, a hazard is an HGM that might affect animal health or that changes the safety of a food product derived from the HGM animal to now pose harm to the consumer.

Exposure in this context would include the expression of the HGM and subsequent effects upon health and welfare. Exposure for a human follows from consumption of food derived from an animal with an HGM; the probability of exposure is a function of the degree to which the consumer incorporates that food into their diet.

Risk, R, is the risk of a defined harm to the HGM animal or to the consumer of an HGM animal product, which is the product of two probabilities: the probability of exposure, P(E), to the hazard, and the probability of the hazard being realized given that exposure has occurred, P(H/E).

Noting that R = P(E) × P(H/E), risk can be minimized by practicing risk management. Risk management should be part of how risk likelihood is estimated recursively, taking into account risk likelihoods given different suites of risk management measures.

Exact probabilities of likelihood and extent of harm given exposure, P(H/E), may prove difficult or impossible to determine for all types of possible harm; indeed, it may be impossible to know a priori all indirect causal pathways for all harms. Further, it may be necessary to classify levels of concern into qualitative categories (no, low, medium, high), based on integrating current and emerging scientific knowledge of gene function, biochemical pathways, allergenicity, immunology, disease ecology, and other fields.

Qualitative risk assessment can support a risk manager in priority setting and policy decision making. Qualitative assessments have contributed to the peer-reviewed literature on risk analysis and have served as tools to identify

and prioritize research needs (Coleman and Marks, 1999). For example, while encouraging the use of quantitative data to the extent possible, the Codex Committee on Food Hygiene (CAC, 1999) also values qualitative information for risk assessment. Qualitative risk assessment has proven useful in a wide range of applications, especially for systems that are difficult to model. Examples go beyond food safety assessment (Coleman and Marks, 1999) to include assessments of environmental mutagens (Lohman, 1999), zoonotic potential of emerging animal diseases (Palmer et al., 2005), and possible control measures (Wieland et al., 2011), prioritization of issues facing fishery management (Fletcher, 2005), and risk posed by consumption of wild game (Coburn et al., 2005).

In the context of HGM animals, risk assessment is effectively a hypothesis-generating and testing process, the null hypothesis being that a food produced from an HGM animal is not safe to eat relative to the range found within an appropriate comparator (Rudenko, 2024), in which case the analysis could lead to a conclusion that this HGM animal-derived food is not as safe to consume as the comparator food. To summarize and provide a supporting example, should an HGM affect critical biochemical pathways, affecting the homeostasis of the animal, critical harm endpoints might include decreased health status. The potential hazard, in this case, is the HGM that might adversely affect the animal’s health. Exposure of an animal to the hazard comes from expression of the HGM itself. To the HGM animal, the risk of harm being realized given exposure to the HGM will follow from the magnitude of the effects of the novel gene product or knockout of a gene upon phenotypes such as health, growth, and reproduction. In context, risk to the animal is the product of exposure to the HGM and the likelihood of harm being realized following exposure. Risk management during development of the HGM animal line and during production of HGM animals in agriculture will have a large bearing upon the assessment of risks to both the HGM animal and to the consumer of an HGM animal-derived food.

U.S. POLICIES FOR ASSESSMENT OF SAFETY OF PRODUCTS OF ANIMAL BIOTECHNOLOGY

As background for considering how the conceptual approaches of risk assessment could be applied to the assessment of food safety of future HGM animal-derived foods, it is useful to consider the overall context of U.S. policy for regulatory oversight of transgenic and genome-edited animals and to examine case studies in which the food safety of HGM animal products has been assessed leading to approval for human consumption in the United States for cloned animals, two products of classical gene transfer (the genetically modified AquAdvantage Atlantic salmon and the Revivicor GalSafe pig), and a genome-edited product (Acceligen SLICK cattle).

Historically, risk or safety assessments relevant to HGM animals conducted by U.S. regulatory agencies utilized conceptual approaches originally applied to evaluate the safety of HGM plants and microbes. In the context of the scientific uncertainties at issue in those times, the Food and Agriculture Organization and World Health Organization (FAO and WHO, 2008) developed the Codex Alimentarius guidelines for assessing the safety of biotechnology-derived plant, microbial, and animal products, focusing particularly on the possibility of unexpected, pleiotropic effects of transgene expression upon the composition of foods. Under these guidelines, if no biologically significant effects (see Chapter 5 for discussion of biological versus statistical significance) are identified, “substantial equivalence” of the biotechnology-derived and conventional products is demonstrated. A finding of substantial equivalence allows the biotechnology product to be regarded as being as safe as the conventional product. The first products considered under this framework were enzymes used for food processing (Pariza and Foster, 1983; Pariza and Johnson, 2001; Pariza and Cook, 2010).

Policies for the assessment of the food safety of products of animal biotechnology have been developed, both within the United States and internationally (Hallerman et al., 2024). Within the United States, under the terms of the Coordinated Framework for the Regulation of Biotechnology (OSTP, 1984, 1985, 1986), the scope of the Federal Food, Drug, and Cosmetic Act (FFDCA) was extended to cover the products of animal biotechnology. As the agency enforcing the FFDCA, the U.S. Food and Drug Administration (FDA) has regulatory authority over the products of animal biotechnology intended for use as food. In contrast to the Codex Alimentarius concept of “substantial equivalence,” FDA applies the concept of “no material difference” between an HGM product and its conventionally bred comparator. The term “material” is important legally, because it provides an opportunity for “biologically significant” changes, meaning that even though there is a difference between an HGM food product

and its comparator, there is no biologically significant difference between the safety of the respective products. Genetically engineered animals are regulated under the animal drug provisions of the FFDCA and must receive formal approval before they may be introduced into commerce.

This policy approach subsequently was applied to reach science-based decisions for several HGM animals, including ornamental Glofish (2003, not intended for use as food), cloning of livestock (2006), AquAdvantage Atlantic salmon (2015), three “biopharm” animals (ATryn goats in 2009, though their milk was not allowed in the human food supply; chickens expressing human lysosomal acid lipase in 2015, though the eggs were not intended to enter the human or animal food supply; and the GalSafe pig in 2020), and SLICK cattle (2022).

Regulatory review of an animal bearing an intentional genomic alteration (IGA, equivalent to the term HGM in the charge to this committee) by FDA is focused on safety of the animals, of anyone who consumes food products from the animals, and of the environment. While developers of HGM animals generally are required to have an approved application prior to marketing, for those HGMs that are found to pose low risk, FDA may not expect developers to seek approval (FDA-CVM, 2023). That is, FDA Center for Veterinary Medicine (CVM) may exercise enforcement discretion on a case-by-case basis if, after a risk-based review, it determines that the product or category of products poses low risk to humans, animals, and the environment.

APPLICATION OF RISK ASSESSMENT PRINCIPLES TO PRODUCTS OF ANIMALS DERIVED FROM SCNT

While not a method for generating HGM animals, somatic cell nuclear transfer (SCNT) is employed to generate whole animals from transformed cells (see Chapter 1). Public concerns were raised following the production of Dolly, the first cloned sheep, which led the FDA-CVM in 2000 to task the National Research Council (NRC) with performing an independent, scientific review of available data on the safety of cloning to identify science-based concerns and elicit information on clones and foods derived from those animals from the scientific community. In 2001, FDA-CVM requested that producers not introduce meat or milk from clones or their progeny into food or feed products until the NRC report was completed and the agency had completed its own review of the safety of those food products. Following its review, NRC (2002b) indicated that: (1) the most likely hazard, for the animal or products from that animal, came from the resetting (reprogramming) of the donor-nucleus epigenome; (2) any harms that might result from that hazard (reprogramming) would be observed early in the development of a cloned individual; and (3) there were no published data comparing the composition of meat or milk from clones with that from conventionally bred animals. Nonetheless, the report concluded that there was no evidence that foods derived from adult somatic cell clones, or their progeny, pose a hazard, that is, there was no evidence that they present a food safety concern.

FDA-CVM subsequently analyzed all available data relevant to assessing the health of clones and their progeny and food consumption risks resulting from food products from these animals. FDA-CVM developed a two-pronged approach to its risk assessment. The first component was a systematic review of the health of the animal clone and its progeny, premised on the hypothesis that a healthy animal is likely to produce safe food products provided that all other measures appear to be normal. This outcome would result in the finding that the clone is likely to produce edible products that pose no food consumption risks. The second component was analysis of food composition, that is, proximate analysis, plus fatty acids, essential amino acids, and essential nutrients (e.g., iron). This approach poses that foods from healthy animal clones and their progeny that are not materially different from corresponding products from conventional animals would present no additional risks. The analysis relied on comparison of individual components of edible foods following identification of appropriate comparators. FDA reviewed data on the health of juvenile and adult cattle, swine, sheep, and goat clones, including the composition of meat and milk from those animals and corresponding information on clone progeny. Assessment of animal health and food composition data was based upon the results reported in peer-reviewed literature reports and unpublished data from two companies engaged in cloning of food animals. On the basis of this assessment, FDA-CVM (2008) drew the conclusion that edible products from healthy cloned animals and their progeny posed no additional risks relative to corresponding products from contemporary conventional comparators.

The draft FDA-CVM risk assessment drew considerable public attention, generating over 30,500 comments, 100 of which were described as “substantive” (FDA, 2021), in large part because it had no precedent in regulatory determinations. Because there were no heritable genetic modifications in the cloned animals that were the focus of that assessment, any hazards would be due to incomplete or inappropriate DNA methylation (i.e., epigenetic processes). Given the existing practice of U.S. Department of Agriculture Food Safety Inspection Service (USDA-FSIS) inspection of meat animals before slaughter, the assessment noted that animals with observable abnormalities would not enter the food supply, so any potential hazards and associated risks would be limited. This aspect of the risk assessment for cloned animals set the precedent for subsequent ones for animals bearing HGMs, in that USDA-FSIS inspection of meat animals before slaughter, which is effectively a risk management technique, was considered a critical component.

Empirical data supported the view that products derived from cloned individuals or from their progeny are not distinguishable from those of conventionally bred animals. For example, Tian et al. (2005) reported data on more than 100 parameters comparing the composition of meat and milk from beef and dairy cattle derived from cloning to those of genetic- and breed-matched control animals produced through conventional reproduction. The cloned animals and comparators were managed under the same conditions, including receiving the same diet. The composition of the meat and milk from the clones was largely not statistically different from that of matched comparators, and all parameters examined were within the normal industry standards or previously reported values. Gu et al. (2019) compared carcass traits, meat quality, and chemical composition of tissues from progeny derived from cloned and conventionally bred pigs. Their data showed that the progeny of cloned pigs were not different in terms of carcass traits, meat composition, and biochemical composition from conventionally bred pigs. Hur (2017) reviewed the results of over 100 studies assessing the animal and food safety of food products derived from cloned animals and found no evidence that meat or milk derived from cloned animals or their progeny pose a risk to food safety in terms of genotoxicity, adverse reproductive effects, or allergic reactions.

APPLICATION OF RISK ASSESSMENT PRINCIPLES TO ANIMAL PRODUCTS OF CLASSICAL GENE TRANSFER

FDA-CVM’s Guidance for Industry 187 (FDA-CVM, 2009, 2015a, 2024), hereafter GFI 187, laid out a seven-step process for oversight of the products of classical transgenesis: (1) molecular characterization of the recombinant DNA construct; (2) molecular characterization of the genetically engineered animal lineage; (3) phenotypic characterization; (4) durability assessment; (5) food safety assessment; (6) environmental assessment; and (7) claims assessment. Only with successful passage through all seven steps would FDA license commercial production of a genetically engineered or HGM animal. Approval can be limited or revoked should adverse outcomes be observed.

Focusing on the food and feed safety step of the GFI 187 process, the risk-related questions involved can be divided into two overall categories (FDA-CVM-VMAC, 2010). The first category regards whether there is any direct toxicity, including allergenicity, via food or feed consumption associated with the expression product of the introduced transgene construct. The second category of questions addresses potential indirect toxicity associated with the transgene and its expressed product; that is, whether expression of the transgene (hazard) would affect physiological processes in the resulting animal (associated hazard) such that food consumption risks (probability of an unidentified harm due to consumption of animals with inherited transgenes) would be increased. Potential adverse outcomes that could result from the food exposure pathway can be identified by determining: (1) whether there are any biologically relevant changes to the physiology of the animal (assessed partly in GFI 187 Step 3: Phenotypic characterization of the genetically engineered [GE] animal), and (2) whether reasons for toxicological concern are suggested by any biologically relevant changes in the composition of edible products from the GE animal compared with those from the appropriate non-transgenic comparator.

Using the approach embodied in GFI 187, FDA has determined that food from two HGM animals—the AquAdvantage Atlantic salmon and the GalSafe pig—are as safe and nutritious to eat as food from non-HGM salmon and pigs. Consideration of these FDA determinations as case studies (Boxes 4-1 and 4-2) illustrate application of the risk assessment methodology and key findings for each case.

APPLICATION OF RISK ASSESSMENT PRINCIPLES TO GENOME-EDITED ANIMAL PRODUCTS

Given that the guidelines for the food safety assessment of recombinant DNA-bearing animals might not apply well to the products of genome editing, FDA-CVM (2017) issued draft revisions to GFI 187 to clarify its approach to the regulation of intentionally altered genomic DNA in animals. This guidance addressed animals whose genomes have been intentionally altered using molecular genetics technologies, which may include random or targeted DNA sequence changes, including nucleotide insertions, substitutions, or deletions, or other technologies such as gene editing that introduce specific changes to the genome of the animal. This revised guidance applies to the intentionally altered genomic DNA in both the founder animal in which the initial alteration event occurred and to the lineage of its descendants that contain the genomic alteration. FDA-CVM (2024) published a statement of its risk-based approach to oversight of HGM animals indicating that CVM will classify risk into three categories. Category 1 pertains to products for which FDA does not expect developers to consult with FDA prior to marketing an animal containing an IGA. Category 2 is for products for which FDA may not expect developers to submit an application for approval if, after reviewing data submitted about that product’s risk, FDA reviewers conclude that they understand the product’s risks for the specified intended use, any identified risks are appropriately mitigated, and there are no further questions for which FDA would need to see additional data. Category 3 products are those that FDA will review and may approve based on data requirements that are proportionate to the associated risk.

A case study on Acceligen SLICK cattle (Box 4-3) shows the application of risk analysis principles to assess the safety of food products from genome-edited animals.

OVERVIEW OF PAST SAFETY ASSESSMENTS FOR HGM FOOD PRODUCTS

The case studies of U.S. regulatory risk assessment for HGM food animals presented in the previous sections are generally aligned with food safety determinations that have been made by regulatory agencies internationally (Table 2-7). These determinations mostly have focused on changed food-product composition as a possible hazard to the consumer. Some regulatory determinations have embodied a formal risk assessment process, while others have not due to a finding of regulatory discretion when a product was determined to pose low risk (e.g., SLICK and POLLED genome-edited cattle).

In these cases, regulatory reviewers found no indications that expression of the HGM impacted the welfare of the animals or caused harm to consumers. Nonetheless, past regulatory practice has been criticized by some parties, as exemplified in public comments on past draft risk assessments expressing criticisms of the risk assessment methods used. For example, the risk assessment for the AquAdvantage salmon was criticized for limited sample size (VMAC, 2010). This critique raises the issues of what effects (i.e., what harm endpoints) are being screened for, what effect size is large enough to matter (which determines the appropriate sample size), and the distinction between statistical significance and biological significance (discussed further in Chapter 5).

Some regulatory determinations have not applied risk assessment because an HGM animal was found to not be a regulated article. Such was the case for three genome-edited fishes in Japan that have been produced by the Regional Fish Institute (Box 4-4). This lack of review precluded learning of any effects of knockout of the respective genes upon fish health and food product safety. Loss of myostatin function gives rise to the so-called double-muscling phenotype of several cattle and sheep breeds, as well as dogs (Aiello et al., 2018). Double-muscled cattle, however, are more susceptible to respiratory disease, lameness, nutritional stress, heat, and fatigue (Holmes et al., 1973), as well as reproductive difficulties (Wiener et al., 2002). Leptin functions as a signaling molecule in a negative feedback loop that regulates appetite, food intake, and metabolism to maintain homeostatic control of adipose tissue mass (Friedman, 2014). Multiple syndromes are caused by leptin deficiency, and leptin-deficient mice show abnormalities in multiple physiological systems (Bray, 1991), including defects in the neuroendocrine axis, infertility or subfertility, and markedly increased corticosterone levels. Lack of knowledge of the impacts of loss or function of the myostatin and leptin genes in the genome-edited fishes limits characterization of any harms that could be realized within the respective fish lines and estimation of any corresponding risks.

PROSPECTIVE APPLICATION OF RISK ANALYSIS

Past assessments illustrate application of risk assessment principles to HGM animals and provide the context for a prospective consideration of risks posed by HGM animals that may be presented for consideration by regulatory authorities. The sections below examine various risk pathways that might be at issue, along with a qualitative evaluation of the likelihood that harm might be realized. Critical unknowns that might be addressed by research conducted over the medium to long term were also identified.

Potential Harms of HGM to Animal Welfare

As noted in Chapter 3, harms to a food animal may be driven by hazards induced at the molecular level by either classical gene transfer or genome-editing techniques. Possible harms to the animal include alterations that reduce fitness, increase disease susceptibility, or reduce animal welfare. Harms to the animal could stem from: (1) alterations at the intended or unintended genomic site(s) that could change phenotype in an undesirable manner; (2) integration of exogenous DNA such as plasmids, virus, or other genes that result in an undesirable gain of function; (3) conformational changes in the three-dimensional chromatin structure of the DNA that could alter expression of genes and thereby change phenotype in an undesirable manner; (4) epistatic effects of the intended or unintended HGMs on the expression of other genes, resulting in changes in traits of the phenotype; and (5) pleiotropic effects stemming from alteration of a gene or insertion of a transgene with effects on multiple traits, resulting in a new or undesirable relationship between phenotypic traits. The presence of unanticipated alterations or their consequences does not inherently present a safety concern to the HGM animal. Indeed, potential harm related to gene transfer or genome editing should be considered in the broader context of potential harm associated with all animal breeding approaches, including selective breeding. The likelihood of unintended genetic alterations to occur in the founder generation of an HGM line is considered to be moderate. Should negative phenotypic impacts upon animal welfare be observed in prospective founders or in their descendants, those animals would not be used for development of a commercial line.

Harms to the animal may also result from the perturbation of metabolic processes underlying physiological balance, or homeostasis (Chapter 3). The choice of promoters used in gene constructs has a large bearing on the effects of the expression of a transgene, often by design outside of the normal homeostatic feedback of the host, possibly leading to unwanted phenotypes, for example, in growth-hormone transgenic pigs (Pursel et al., 1989) and several species of salmon (Hallerman et al., 2007). Knockout of a gene using genome editing also could have follow-on effects, depending upon the function of that gene. For example, effects would vary based upon which molecule in a pathway was targeted. Knockout of a receptor molecule can block a signaling pathway and would impact the expression of multiple genes, which in turn would disrupt multiple other signaling pathways. Moreover, while pathways are typically represented as linear processes, there are well-established examples of genes that are involved in crosstalk among biochemical pathways to coordinate pathway processes. For example, the AMPK pathway is activated by glucose deficiency and inhibits the mTOR pathway, which is responsible for cell growth. In such cases, harms could manifest as negative effects upon the phenotype of the HGM, for example, cataracts, lameness, stillbirth, or spinal deformity (Matika et al., 2019; Epstein et al., 2020; Guo et al., 2023). In contrast, HGMs could also impact animal welfare in positive ways. Improvement of disease resistance, for example, would reduce animal suffering and mortality. As noted in Chapter 3, the impact of HGMs on animal welfare is particularly difficult to assess.

Management of risk to HGM animals would follow from detailed molecular and phenotypic evaluation though the process of development of the HGM line (see Risk Management section below; see also Chapter 5). Affected individuals would be culled before commercialization of an HGM line and any harms would be limited to the small numbers of animals within the experimental line. Understanding of mechanisms underlying any issues of homeostasis may be possible by analysis of the whole-genome DNA sequence, and perhaps also by application of transcriptomic, proteomic, and metabolomic workflows during the process of development of an HGM line (as discussed in Chapter 5).

Potential Harms to Consumers of Food Products from HGM animals

Direct harm endpoints to the consumer of products of HGM animal products (discussed in Chapter 3) might include toxicity, changed composition sufficient to affect consumer nutrition, or allergenicity and food intolerance. Indirect harms due to HGM could result from harm to the health of the animal, for example, in greater bacterial load in milk due to infections, or poorer health of an animal that still passes USDA-FSIS inspection and enters the human food supply.

The consumer’s exposure would stem from eating a food product from an HGM animal after the expected or foreseeable consumption of that food (FAO and WHO, 2008). Exposure should be evaluated in the context of the

total diet and the assessment based on the customary dietary composition for a population of consumers, which will vary with the importance of the food in the diet(s) of a given population(s). The diets of human consumers being highly varied, the exposure to any hazard posed by any particular HGM animal product would be expected to be small.

To the consumer, the risk of harm being realized given exposure will follow from an estimation of the concentration(s) of biologically active substance(s) or nutrients in a food derived from an HGM animal and any known factors that influence their bioavailability (FAO and WHO, 2008). The risk of harm being realized given exposure is minimized by the cooking of animal products prior to consumption. Because humans eat varied diets, any change in composition of a particular HGM animal-derived food would be likely to prove inconsequential to the well-being of the consumer.

Risk to the consumer, then, is the compound product of dietary exposure to a food product from an HGM animal and the likelihood of harm being realized following exposure.

Toxicity

While there is potential risk for toxicity caused by classical gene transfer from a non-food species, this is not considered likely in the genome-edited animal. The question would be whether edible products of the HGM animal have been screened adequately to detect the presence of unanticipated compositional changes that might pose toxicity (NRC, 2002a).

As described in Chapter 3, while some plants protect themselves from herbivory by synthesizing anti-nutritional or even poisonous compounds to ward off predation by herbivores, insects, and microorganisms (e.g., alkaloids, terpenoids, phenolics, protease inhibitors, glucosinolates, and volatile organic compounds), traditional food-animal species do not express toxins. The few food animals that pose toxicity do so because they sequester toxins taken up from plants that they consume. These include puffers and other fishes of Order Tetraodoniformes; octopus and shellfishes expressing tetrodotoxin; and some birds, including ruffed grouse (Shumway, 1995; Ligabue-Braun and Carlini, 2015; Kotipoyina et al., 2020; NIOSH, 2024). Toxicity of animal products also might stem from mycotoxins or bacterial contamination or from toxic materials in feeds, but do not arise from the animals themselves, and are not any more at issue for HGM animals than other animals. On this basis, the potential risk of an unintended introduction of a toxin into an HGM animal is expected to be negligible. The application of classical toxicological and new spectroscopic assays for assessing toxicological endpoints and ensuring that toxicity poses a low level of food safety concern is discussed in Chapter 5.

Changed food composition

Harm to a consumer might conceivably follow from changed composition of a food product, that is, as the consequence of introduction of a new nutrient(s) or related substance(s), alteration of either the quantity or bioavailability of a nutrient(s) or related substance(s), addition or heightened concentration of undesirable substance(s) (e.g., allergens), or alteration of the interaction(s) of nutritional or health relevance of these substances (FAO and WHO, 2008). As noted in Chapters 3 and 5, the risk associated with changed food composition may prove difficult to quantify, as the composition of any animal-derived food is also a function of how the animal was produced, its genetic background, stage of development, reproductive status, and other factors. Trott et al. (2022) asserted that the requirement to perform a compositional analysis of edible products from the recombinant-DNA animal outlined in Codex Alimentarius (FAO and WHO, 2008) for traits that are not designed or expected to impact food product composition provides little data useful for food safety assessment. Lacking criteria to determine which key components in an animal-derived food might pose a human hazard if consumed, and what levels of effect size would matter to food safety, the authors noted that it is not possible to calculate the sample sizes needed to demonstrate significance. In addition, it is expensive to generate and maintain conventional comparators and doing so may prove impractical in animals for which conventional comparators (non-HGM animals of the same genotype) are not routinely produced. Hence, the comparative food safety assessment approach adds significant expense to HGM animal experiments, especially for large livestock species. Against the background of these considerations,

therefore, to assess whether a food is safe for human consumption and whether nutrient composition is meaningfully altered, it is important to compare the levels of key nutritional components of HGM animal-derived food products to the range or variation of those levels in those same products that are sold in commerce, for example, in terms of proximate analysis, key nutrients, and other substances that may be present intentionally for which there may be a priority concern for food safety.

Because plant-derived foods can contain endogenous components and antinutrients that can be toxic to humans and animals when present in high concentrations (Herman et al., 2009), the International Life Sciences Institute Crop Composition Database was established as a resource for data on these compounds (Ridley et al., 2004; Alba et al., 2010; Sult et al., 2016). However, animal products do not typically contain such compounds and hence are not routinely subjected to extensive compositional analyses (Lema et al., 2007). Data on the composition of common meat and milk products are limited (Roseland et al., 2015, 2017, 2024; Williams et al., 2017). In a unique dataset relevant to the context of this report, Trott et al. (2022) assembled data for protein, crude fat, ash, amino acid, mineral, and vitamin composition of milk and selected raw beef products from genome-edited and conventionally bred animals. Nutrient values in the meat from control and genome-edited offspring fell within the range found in beef composition databases; milk composition varied within controls, between control and genome-edited offspring, and over time, and all values except sulfur percentage were within ranges reported in peer-reviewed literature.

Health status—manifested as the phenotype of the animal—has proven to be a reliable indicator of animal product safety and is used in combination with other antemortem and postmortem inspections to determine that food from animals produced through conventional breeding is safe to eat. The health status and phenotype of the HGM animal provide the most reliable indicator of the safety of foods derived from that HGM animal.

Due to risk mitigation during the process of HGM line development and testing, the likelihood of unintended, biologically significant changes in the composition of food products from fully developed HGM animal lines is low. Additional compositional testing of food products from GM animals in the absence of a relevant hypothesis for assessing risk is not justified.

Allergenicity

As noted in Chapter 3, food allergies arise when a protein or glycoprotein in food elicits a heightened immunoglobulin E response of the immune system in a consumer. Four of the nine foods or food groups that account for over 90 percent of food allergies in the United States include animal products—namely, cow milk, eggs, fish, and shellfish (Taylor et al., 1999)—although 160 other foods have been identified as causing food allergies (Hefle et al., 1996). Hence, assessment of the potential for novel expression or increase or decrease of allergenic compounds in an HGM product is relevant to risk assessment. No single piece of information or assay result provides sufficient evidence to predict allergenicity, and the European Food Safety Agency Panel on Genetically Modified Organisms (EFSA GMO Panel, 2022) recommended adopting a weight-of-evidence approach. The panel also called for modernization of key elements of allergenicity risk assessment, including consideration of clinical relevance, route of exposure, threshold value for food allergens, update of in silico tools, and better standardization and integration of test materials and in vitro and in vivo protocols (see Chapters 5 and 6).

Given that human allergens are well described and documented, it is unlikely that HGM animals would be unintentionally produced with increased allergenic potential. Genomic screening of N1 (i.e., first-generation HGM) animals for genetic alterations and phenotypic assessment of allergenic response using a range of assays in a weight-of-evidence approach would be expected to be sufficient for ensuring that genes affecting known allergenic potential remain unchanged (unless intentionally modified). In fact, gene editing has been posed as a means for reducing allergenic potential (Brackett et al., 2021; Wang et al., 2024). There may be instances where an animal product may contain more than one allergen, and although the HGM might be designed to remove one of them, physiological compensation may result in elevated levels of other allergens, which then should also be measured in the resulting HGM animal product. The risk likelihood posed by allergenicity of HGM animal-derived foods is low to moderate and would be small if effective risk management is practiced (see Risk Management section). In the context of allergenicity oversight, FDA provides detailed requirements for manufacturing and labeling of foods bearing defined allergens.

Potential Harms to Humans Through Other Pathways

Disease-resistant animals, viral mutation, and pathogen spillover

A major goal for the application of animal biotechnology is the development of disease-resistant HGM animal lines (see Chapter 2). Development of disease-resistant or disease-resilient animals can reduce livestock losses; however, it is also important to consider potential risks. These animals may provide selective pressure on pathogens, causing the emergence of new pathogen variants with altered host range or virulence (see Chapter 3). The likelihood that a particular pathogen could evolve increased transmissibility or pathogenicity is low to medium, but the impact upon human populations could be large. This harm endpoint is not novel to HGM animals, but the intermediate steps along the risk pathway to that endpoint could be complex, ambiguous, and not well characterized. Several viruses that cause human infection come from poultry and livestock, among which coronaviruses, rabies virus, and influenza viruses are the most frequently reported (Chen et al., 2021). The importance of this possible impact is evidenced by the fact that several major human disease outbreaks have followed spillover from livestock, including poultry (e.g., avian influenza virus; see Chapter 3), pigs (e.g., influenza, Japanese encephalitis, Nipah, and coronaviruses; McLean and Graham, 2022), cattle (e.g., Rift Valley fever, tick-borne encephalitis, and cowpox viruses; McDaniel et al., 2014; and human coronavirus OC43; Forni et al., 2017), camels (e.g., Middle East respiratory syndrome virus; WHO, 2022), and alpacas (e.g., human coronavirus 229E; Forni et al., 2017).

The specific pathogen and the type of HGM (e.g., whether the animals are disease-resistant or disease-resilient) will influence the level of risk associated with possible viral evolution in livestock. However, managing the risk of pathogen spillover from disease-resistant HGM animals requires two complementary approaches, discussed in more detail in Chapter 5. Briefly, the first approach focuses on HGM experimental design to minimize the likelihood of pathogen escape (e.g., by “stacking” multiple different edits and targeting multiple host genes; see Chapters 5 and 6). The second approach requires monitoring of domesticated animals, wild animals, and human populations for signs of disease spillover and outbreak. This approach should also encompass effective biosecurity at HGM animal production facilities and monitoring of high-risk human populations (i.e., farm workers). Research is needed to investigate how disease-resistant or disease-resilient HGM animals might influence pathogen evolution, and how this influences viral fitness, including the ability of viruses to replicate and transmit in both edited and unedited hosts. This is considered necessary to know.

Transfer of antibiotic resistance genes and subsequent effects

As noted in Chapter 3, unintended promotion of antibiotic resistance is a public health concern (Manyi-Loh et al., 2018). Antibiotic-resistance genes often have been included in expression vectors utilized in classical transgenesis because antibiotic resistance is useful for selecting plasmid-bearing E. coli cells when a large quantity of expression vector is prepared for a gene transfer experiment and for selecting transformed cultured cells in anticipation of an SCNT-based production of an HGM animal. Researchers are cognizant of the issue of antibiotic resistance in the design and execution of current experiments producing new lines of HGM animals. Selectable markers other than antibiotic resistance are now preferred, and a directed search can be used to detect antibiotic-resistance genes in whole-genome sequences of early-generation HGM animals. Such individuals would be removed from further development of HGM lines, and any hazard to animal or human health would thereby be addressed. Thus, the risk posed by integration of antibiotic genes into the genomes of transgenic animals is indirect and low, and the risk is dwarfed by the hazards posed by widespread antibiotic use in animal production. Efforts to reduce the occurrence of animal disease (e.g., porcine respiratory and reproductive syndrome virus-resistant pigs) have the potential to reduce antimicrobial use in animal production and thereby to reduce the risk of resistance emerging.

Mobilization of animal viruses via xenotransplantation

As noted in Chapter 2, transgenic and genome-edited animals, some lines with multiple transgenes and edits, have been produced for purposes of xenotransplantation of organs, tissues, or cells to human patients who are

unable to secure transplants from human sources. The intimate contact between porcine cells, tissues, or organs with the body of the recipient poses the hazard that xenotransplantation may be associated with transmission of potentially zoonotic porcine viruses to patients and perhaps also to others, especially family and caregivers (Chapter 3). Cross-species viral infection would require a series of events, that is, expression of infectious virus, localized infection of host cells, spreading infection in the host, persistent viremia, and expression of disease (NRC, 2002b; Chapter 3). To pose a risk to society, such an infection would require transmission to close contacts and epidemic transmission. Each step is associated with low likelihood. However, the impact of spillover of pathogens among species can range from a few, local cases to a pandemic scale. Hence, multifaceted risk management will be needed. Development of some pig donor lines has involved knockout of endogenous retroviruses. Control measures for porcine viruses include high biosecurity at farms producing xenotransplantation lines, along with the use of vaccines and antiviral drugs, and perhaps also the use of RNA interference in the development of xenotransplantation lines (Mao et al., 2023). Groenendahal et al. (2023) identified all microorganisms and parasites posing harm to human xenotransplant recipients and developed a risk-based framework to manage risks, noting that most risks can be eliminated with practice of high biosecurity programs. After a xenotransplantation event, the patient, family, and caregivers should be monitored for relevant porcine-derived viruses using polymerase chain reaction-based, immunological, and other methods (Denner, 2022). Research needs pertaining to assessing risks posed by xenotransplantation are discussed in Chapter 6.

Horizontal gene transfer

Another concern regarding the consumption of GM food is that horizontal gene transfer (HGT) could compromise consumer health (Pontiroli et al., 2007; Nawaz et al., 2019; Chapter 3). However, HGT is not a harm per se, but rather an exposure pathway through which potential hazards may move from one organism to another. Hence, risk assessment should consider both the severity and likelihood of negative consequences (Keese, 2008). A chain of sequential events affects the probability of HGT posing a consumer health risk, that is, an encounter between the genetic material from the donor organism and the consumer, donor genetic material entering the cell and the nucleus, integration of donor DNA or RNA into the recipient organism’s genome, expression of a novel trait in the recipient organism, the persistence of this trait, and, if integrated into the germline with a suitable promoter, transmission to offspring (Van Elsas and Bailey, 2002). For all these reasons, even with occurrence of HGT, the probability of the other processes also occurring is near zero, reducing the overall possibility of harm being realized to near zero (Vega Rodríguez et al., 2022). If the probability of HGT from consumed food to the consumer was significant, then all the genes in that food would have a probability of moving into the consumer’s genome and causing harm, not just the modified genes. No evidence of HGT through the food consumption pathway has been observed.

Risk Management Practices for Animal-Derived Foods

Context

Within the risk analysis process, risk is a function of the probability of exposure to a hazard. Therefore, risk can be minimized by practicing risk management. Risk management might be applied through the development of the HGM animal line, the agricultural production process, and at slaughter to limit the hazards that may be present, and if present, to limit exposures.

Risk management during the process of development of HGM animals

Applications of gene editing for enhanced trait development in food animals will follow sequential layers of risk mitigation before a product of an HGM animal line is made available for human consumption. Risk management for the use of clustered regularly interspersed short palindromic repeats (CRISPR) technology to introduce genes from other species (effectively, transgenesis) will utilize the same sequential layers of risk management,

but may also require environmental risk assessment as detailed in previous National Research Council (NRC, 2002a, b) recommendations and GFI 187 (FDA-CVM 2009, 2015b, 2017, 2024) regarding risks associated with transgenesis, particularly as related to aquaculture species.

Several methods have been developed to identify intended and unintended edits caused by DNA editing methodology (Lee et al., 2024; Lopes and Prasad, 2024), although the detection of unintended edits is problematic (see Chapter 5). Although use of these methods for generating food animals can bring about changes to DNA that are unexpected, the developmental pipeline for integrating genome-edited animals into a production system has several layers of assessment, described below, that ensure the ultimate safety of both the animal and the consumer (Figure 4-1). In most cases, these additional steps make it unnecessary to apply additional extensive analyses for deep characterization of intended and unintended editing at the DNA level for animals ultimately entering the food production system.

Level 1: Using advanced CRISPR methodology to develop genome-edited animals with fewer unintended alterations

The first level of risk mitigation is at the design stage. The methods for developing genome-edited animals are evolving rapidly (Chapter 1) to increase precision and reduce the incidence of unintended edits. After these

methods are demonstrated in model systems and humans, they are likely to be applied to food animals. Likewise, multiple algorithms have been designed to identify unintended edits by guide RNAs, including CRISPOR and Cas-OFFinder (Bae et al., 2014). These algorithms can be used during a CRISPR-Cas9 editing design phase to choose guide RNAs that will have little to no potential for unintended edits.

Level 2: Applying advanced sequencing technology for characterization of changes inherited by commercial progenitors

The second level of risk mitigation is at the molecular level, with characterization of changes in the DNA of the founder animal and, more importantly, the modified DNA sequence that is intended to be transmitted across generations. DNA sequencing technologies can be used to identify target-specific alterations and to perform whole-genome sequencing (Chapter 5). However, any sequence analysis requires careful consideration to distinguish between natural variation, intended alterations, and unintended alterations brought about by the editing machinery. Hence, any molecular analysis should be viewed as a first-pass analysis to identify regions of the genome for further characterization, rather than as a complete safety-assessment tool for animal or human health concerns.

Level 3: Genome annotation methodology identifying functional consequences of genomic changes

The third level of risk mitigation is based on genome annotation and variant effect prediction (Hunt et al., 2018) to determine the impact of genome changes on the expected phenotype. Variant effect prediction is a computational method that uses evolutionary conservation, functional annotations, and physicochemical differences to infer the likely effects of genetic variants upon an organism. The method requires that functional annotation and the functions of variants in humans and model organisms have already been determined (Chapter 5). Considering that the purpose of genome engineering in food animals is to enhance production efficiency and resiliency, strategies that result in lethality, gross anatomical or physiological effects, including diminished reproductive effects, will not advance beyond the initial proof-of-concept research phase, because no obvious market or utility for such animals exists. Human clinical trial experience shows, however, that some very subtle or infrequent events may not be detected in phase 3 trials (which typically have thousands of participants for small-molecule drugs and hundreds for biologics), but are detected only when the drug or biologic has been commercialized, underscoring the importance of risk evaluation and mitigation strategies or phase 4 trials.

Level 4: Performance testing by breeding companies

The fourth level of risk mitigation in food animals focuses on the phenotype of the breeding stock that founds the commercial line. The founders are unlikely to be sources of food (because of their defined role) and as such can be subject to detailed phenotypic characterization to uncover unanticipated abnormalities. Commercialization of an HGM food animal requires modification of elite high-performance breeding stock already subject to selective breeding. At this production stage, the modified gene is just one of many that contribute to expression of polygenic traits (traits that are influenced by many, perhaps hundreds of genes) that are selected for in commercial breeding programs, such as feed efficiency, fertility, health, and growth rate, depending upon the breed and producer-level objectives. If expression of the HGM as phenotype in the descendants of HGM founders conflicts with commercial objectives, the gene will be eliminated from the population and the risk thereby eliminated.

Level 5: Federal inspection

The fifth level of risk mitigation occurs through federal regulation of food products sold commercially for human consumption. Because of the history of concern over the safety of meat products (NRC, 2002b), the United States has in place a meat and poultry inspection infrastructure (Table 4-1) that is effectively a risk management system. Risk mitigation related to food occurs through federal regulation of food products sold commercially for human consumption. The USDA-FSIS enforces the Federal Meat Inspection Act, the Poultry Products Inspection Act, and the Egg Products Inspection Act. As a result, meat and poultry are inspected at harvest. To enter the commercial food chain, all amenable food animals must be healthy and ambulatory during mandatory antemortem

TABLE 4-1 Federal Inspection of Animal-Derived Food Products

NOTES: USDA-FSIS = U.S. Department of Agriculture Food Safety Inspection Service, FDA = U.S. Food and Drug Administration.

^a USDA-FSIS applies for ictalurid catfishes only, i.e., channel catfish and its hybrids; NOAA Fisheries provides voluntary, fee-for-service inspection of marine fish, shellfish, and fishery products.

inspection (although “ambulatory” does not apply to fishes). Carcasses and meat products (cattle, sheep, swine, goats, and siluriform catfishes) are evaluated by on-site, online USDA-FSIS inspectors to ensure that these products are safe, wholesome, and properly labeled. Likewise, the USDA-FSIS is responsible for inspecting domestic poultry (chickens, turkeys, ducks, geese, guinea fowl), ratites, squab, and egg products. FDA, authorized by the Federal Food, Drug, and Cosmetic Act, the Public Health Service Act, and the Code of Federal Regulations Title 21, oversees the safety of milk, milk products, shell eggs, other meats (e.g., game and other livestock species), fish, shellfish, and other aquatic animals, as well as other foods. Both agencies also ensure that good food safety practices are employed during harvest and production of food products (e.g., mandatory Hazard Analysis and Critical Control Point [HACCP] plans). These plans include mitigation strategies with validation testing to reduce or eliminate biological, chemical, or physical hazards. Approaches implemented to ensure the safety of food products derived from conventionally bred animals would also be applied to those derived from biotechnology.

Healthy animals, whether HGM or conventionally bred, would be expected to yield food products that are safe for consumption. Therefore, monitoring for welfare would provide a useful approach to risk management. In response to consumer concerns about the treatment of farm animals, most livestock production industries in the United States have developed and implemented certification programs with science-based animal care guidelines (discussed in more detail in Chapters 3 and 5).

Focusing on the example of mammalian meats, the USDA-FSIS is responsible for ensuring that meat and meat products intended for human consumption are safe and properly labeled and packaged. USDA-FSIS ensures food safety at slaughter through mandatory robust antemortem and postmortem inspection procedures, offline verifications to determine whether establishments implement effective sanitation and food safety systems, and verification sampling programs (Stumps, 2022). Facilities identify and present animals for on-site antemortem inspection on the day of slaughter. In antemortem inspection, USDA-FSIS personnel examine livestock and poultry prior to harvest to determine whether they show any signs of disease or abnormalities and are fit for human consumption. Only animals that pass antemortem inspection are allowed to enter the food chain. The purpose of postmortem inspection of livestock and poultry is to protect public health by ensuring that carcasses, including meat and internal organs, were produced from healthy animals and do not have defects, lesions, or disease, or any external contaminants (e.g., hair follicles, soil, fecal matter, ingesta, or milk). Lymph nodes, small intestines and organs (heart, lungs, and liver) are inspected for any abnormalities associated with diseases. USDA-FSIS veterinarians (i.e., professional inspectors) are available to assist in serious diagnoses that are outside the scope of an online inspector. In addition, proper disposition of carcasses is required in any case in which inspectors detect abnormal conditions (e.g., presence of visible tumors) that require additional analysis. USDA-FSIS veterinarians can submit tissues for laboratory analysis to obtain additional diagnostic information as necessary. In these thorough antemortem and postmortem inspection procedures, USDA-FSIS verifies regulatory compliance through verification of HACCP and sanitation regulatory requirements. Validation of the HACCP system includes review of the scientific or technical information (e.g., scientific journal articles) to show that the HACCP plan, including any

programs to support the hazard analysis decisions, effectively control the hazards. Hence, any emerging scientific evidence that the products of animal biotechnology could pose risk would be taken up by the inspection system implemented at points of slaughter.

In implementation of the Food Allergen Labeling and Consumer Protection Act of 2004, commercially available food products are labeled to alert consumers of potentially allergenic content. Regulatory review of products of HGM animals expressing gene products with allergenic potential might lead to the requirement that the products be labeled when marketed so that vulnerable consumers would thereby be warned of any allergenicity hazard.

Initial gene transfer and genome-editing research provides proof of concept for the efficacy of an HGM. In these early stages of HGM animal line development, costly and exhaustive methods of risk assessment are not warranted, particularly if the modified animal line is contained and not intended for consumption. In this case, the probability of exposure to the hazard is near zero, thereby mitigating the risk of harm resulting from exposure. Practical risk management measures for the HGM animal for the subsequent development of an HGM animal line would include propagation only of animals that exhibit healthy, productive phenotypes, as well as the use of best animal husbandry practices and veterinary care.

TOWARD REFINEMENT OF RISK ASSESSMENT APPROACHES

Several issues require further consideration in order to refine application of the risk assessment process to HGM animals.

Comparator groups

As noted above in the context of the food safety assessment for the AquAdvantage salmon, animals of the same age, breed, sex, and husbandry conditions might be utilized for experimentally rigorous, scientific comparisons to quantify the effect of an HGM. However, in the regulatory context of assessing food safety, comparison of the qualities of an HGM animal product with those of the full range of similar food products already in commerce for that species is recommended. In Chapter 5, however, it is noted that there are no Organisation for Economic Co-operation and Development compendia on the composition of animal products as there are for certain crops. For HGM animals, developers and regulators might obtain useful data for comparators from the peer-reviewed published literature and the unreviewed “gray” literature. Animal breeders, as well as food companies and commodities groups, may have additional data of value. Hence, research is recommended to develop compendia of: (1) data currently existing on important parameters in key species; (2) analysis of data gaps; and (3) generation of critical “missing” data (discussed further in Chapter 6).

For comparisons among products in the context of food safety assessment, focusing upon biological as opposed to statistical significance is recommended. That is, if a suitably large number of observations does not demonstrate statistical significance, then a difference between HGM animals and comparators is not biologically significant. The issue of determining the appropriate numbers of HGM animals and comparators is especially important for species with low fecundity and long reproductive cycles.

Identification of the most appropriate comparator and numbers should reflect the a priori level of concern based on the nature of the intended HGM, for example, gene knockout, altered compositional profile, or disease resistance.

Linking phenotype to underlying physiology

Regulatory assessment of the safety of HGM animal-derived foods is based in part on changes to the physiology of the animal and the composition of its edible products (FDA-CVM, 2009, 2015b, 2017, 2024). In cases where changes are made to a gene with a known relation to biochemical pathways, changes to fluxes through those pathways can be measured using targeted biochemical assays. However, there will be cases where changes to a gene will have unanticipated biological consequences. In such cases, the use of transcriptomic, proteomic, and/or metabolomic methods (as discussed in Chapter 5) may yield useful insights. Such applications have not yet been elaborated for assessment of the consequences of heritable genetic modifications of animals (Chapter 6).

Most genetic modification experiments target well-studied genes (Chapter 2), but too often, there is a view that an alteration to an intergenic region would have no consequence because there is no functional gene at that chromosomal location. This view ignores consideration of any regulatory sequence that may lie there, which may cause a conformational change in the tertiary structure of the DNA molecule and thereby affect gene expression. The National Academies of Sciences, Engineering, and Medicine (NASEM, 2016, p. 494) concluded that the size and extent of the HGM itself has relatively little relevance to its biological effect and consequently its food safety risk. Yet, some discussions and regulatory approaches have assumed that a change of one base pair or a deletion of any size could occur naturally and is therefore safe and should be exempt from regulation (Gould et al., 2022). However, a change could be made in a single base pair that gives rise to a stop codon, perhaps preventing the production of an enzyme that catalyzes a key step in a metabolic pathway. Such a change could result in a desired new trait and also in an unintended metabolic shunt toward greater production of an undesirable metabolite with potentially negative consequences (Enfissi et al., 2021) for the animal or the consumer. Hence, it is important to link an HGM to the expressed animal phenotype.

Benefits

Risk assessment has classically been applied to estimate the likelihood of a defined harm becoming realized; in the context of this report, the harms in question would be realized to the HGM animal or to the consumer of an HGM animal product. However, animals with HGMs might be produced to present health benefits to production animals or nutritional benefits to the consumer (see Chapters 2 and 3). Considering exposure and likelihood of benefit being realized given exposure, the conceptual structure of risk analysis can be applied to estimate the benefits of HGM animal products to a population of consumers. As this committee was charged with examining harms, rather than benefits, the committee noted that under the aegis of GFI 187 (FDA-CVM, 2015b), issues of consumer benefit would be considered under the claim assessment aspect of the regulatory protocol. If food from an HGM animal differs from that of its non-engineered counterpart, for example, having a different nutritional profile, that difference would be material information that would have to be revealed in labeling (FDA-CVM, 2015b). An HGM animal-derived food whose composition is within the range of variation as the same food item obtained from conventionally bred animals does not present harm to consumers.

KEY FINDINGS

Risk assessment is the process of identifying potential harms to the HGM animal or to the consumer from consumption of a food product derived from an HGM animal. Risk assessment involves identifying hazards, assessing the probability of exposure to the hazards, and assessing the probability of harm being realized given exposure to the hazards.

1. The likelihood of an HGM causing unintended impacts upon the founder generation of an HGM line is low to moderate, and such risk would be near zero in the advanced generation to be commercialized. With the exception of allergens, the risk of harm being realized given exposure to a bioactive compound is minimized when animal products are cooked prior to consumption. Because humans eat varied diets, any change in composition of a particular HGM animal-derived product would be likely to prove inconsequential to the well-being of the consumer. The potential risk of an unintended introduction of a toxin into an HGM animal is negligible.
2. To assess whether an HGM-derived food is safe for consumption and provides expected nutrients, it is important to compare the levels of key nutritional components of HGM animal-derived foods to the range or variation of those levels in those same conventional foods that are sold in commerce. The health status and phenotype of the HGM animal provide the most reliable indicator of the safety of foods derived from that HGM animal. Due to risk mitigation during the process of HGM line development and testing, the likelihood of unintended, biologically significant changes in the composition of food products from fully

2. developed HGM animal lines is low. Additional compositional testing of food products from GM animals in the absence of a relevant hypothesis for assessing risk is not justified.
3. Noting that the phenotype of an animal has been and remains a reliable indicator of the safety of foods derived from it, and that phenotype is under selection in development of the HGM line, the likelihood of harm being realized from an HGM animal-derived food is expected to be quite low, especially against the background of the larger effects of genetic background (breed) and rearing conditions upon the composition of an animal-derived food product. The likelihood that a particular pathogen could evolve increased transmissibility or pathogenicity is low to moderate, but the impact upon human populations could be large. Addressing this gap in the knowledge is considered necessary. The risk posed by integration of antibiotic genes into the genomes of transgenic animals is indirect and low.

REFERENCES