Biobeef Blog

Thoughts of public sector animal geneticist - all views are my own

New York Times Reporting on Agriculture

Look I get it. The New York Times (NYT) does not like GMOs, industrial agriculture, factory farming or meat consumption. But I question the decision of such an influential media source to feature TWO front page articles detailing agriculture industry funding of agricultural scientists, Prof. Kevin Folta in 2015 and in 2022 Prof. Frank Mitloehner, who work doing public outreach in these fields. With the implication that they are “on the take” and promoting misinformation as a result of this funding. But what the NYT failed to show in these stories was that either of these public sector scientists, whose reputations the NYT has forever brought into question, ever made statements that were unsupported by peer-reviewed literature. There is a term for when someone attacks the character, motive, or some other attribute of the person making an argument, rather than the argument itself; it is called an “ad hominem” attack.

The most recent article featured one of my departmental colleagues at UC Davis, Professor Frank Mitloehner. The title of the article “He’s an Outspoken Defender of Meat. Industry Funds His Research, Files Show” was a little confusing, as it seemed to be suggesting the source of Professor Mitloehner’ s research funding was what was concerning, but in fact most of the article was about communication and outreach done by the Clarity and Leadership for Environmental Awareness and Research (CLEAR) Center at UC Davis, which receives almost all its funding from industry donations. Something that Professor Mitloehner has been very open about as detailed in his blog post, and on the CLEAR website. Prof. Mitloehner provides his response to the NYT article and accompanying Greenpeace hit piece here. The irony of Greenpeace, which itself has shamelessly ignored the scientific consensus  on the safety of GMOS and has engaged in ‘Tobacco-style’ PR on this topic for 30 years, accusing  Prof. Mitloehner of being  “a sock puppet” is projection at its finest.

The evidence that the NYT article provides to suggest that these communications are slanted is a statement by Matthew Hayek, an assistant professor in environmental studies at New York University. He states “Almost everything that I’ve seen from Prof. Mitloehner’s communications has downplayed every impact of livestock,” he said. “His communications are discordant from the scientific consensus, and the evidence that he has brought to bear against that consensus has not been, in my eyes, sufficient to challenge it.”

I am not sure exactly which specific scientific consensus Prof. Mitloehner’s communications are discordant from, but in presentations I have not heard Dr. Mitloehner make a statement that was not supported by peer-reviewed papers. I have heard him clearly state on multiple occasions that livestock are responsible for 14.5% of global emissions, and that cows and other ruminants account for 4% of US greenhouse gases (GHG) in concordance with the scientific consensus.

I am no GHG expert, but what really seems to be at the heart of people’s “beef” with Prof. Mitloehner’s communications is whether methane (CH4), as a short lived GHG, should be treated differently to CO2 which is a long-lived GHG in predictions of global warming impacts. This particularly impacts ruminants whose rumen-dwelling methanogenic bacteria produce (CH4) when breaking down otherwise indigestible cellulose. This is discussed on the CLEAR site.

This is actually NOT a settled science.  The commonly utilized ‘carbon footprint’ impact assessment makes use of the GWP100 metric, i.e., the global warming potential over a 100-year time horizon, while standardizing the atmospheric effects of all GHG to CO2-equivalents (CO 2-eq). It is typically claimed under GWP100 that methane (CH4) is, as a greenhouse gas, 28 times more potent than CO2. GWP100 became the standard metric more than 30 years ago when it was selected for in the Kyoto protocol. It has had a history of critiques in relation to characterizing the climate impacts of CH4 (Pierrehumbert, 2014). The authors of the Intergovernmental Panel on Climate Change Fifth Assessment Report themselves state that the GWP100 climate metric should not be considered to have any special significance.

A newer metric, GWP*, was developed by researchers at the Environmental Change Institute, School of Geography and the Environment, University of Oxford (Allen et al., 2018, Cain et al., 2019). These scientists argue that this metric more aptly represent how CH4 emissions translate into temperature outcomes at various points in time, by treating CH4 as a flow gas rather than a stock gas like CO2. This metric was not developed by the livestock industry. The New York Times article includes a quote “the use of that method [GWP*] by an industry “as a way of justifying high current emissions is very inappropriate.” However,  this metric is not necessarily “livestock friendly” under all conditions (Costa et al., 2021). In fact if livestock populations are increasing, the global warming impacts when using GWP* are higher than using the GWP100 metric. And Professor Mitloehner is not alone in suggesting that “expressing CH4 emissions as ‘CO2-equivalent’ emissions based on the GWP100 could misdirect attention from the need to reduce global net CO2 emissions to zero as quickly as possible” (Reisinger et al., 2021). Some have further argued that avoiding animal-sourced foods based on the GWP100 metric may result in trading a short-term climate benefit from reducing short-lived CH4 emissions, with a longer-term problem of increased CO2 and N2O emissions, making climate stabilization even more difficult.

Even the most recent IPCC technical report (2021; doi:10.1017/9781009157896.002) includes discussion of the GWP* metric

“New emissions metric approaches, such as GWP* and Combined GTP (CGTP), relate changes in the emissions rate of short-lived greenhouse gases to equivalent cumulative emissions of CO2 (CO2-e). Global surface temperature response from aggregated emissions of short-lived greenhouse gases over time is determined by multiplying these cumulative CO2-e by TCRE (see Section TS.3.2.1). When GHGs are aggregated using standard metrics such as GWP or GTP, cumulative CO2-e emissions are not necessarily proportional to future global surface temperature outcomes (high confidence).”

The best way to do environmental assessments of the livestock sector is not a settled science. All-too frequently assessments are stated in simplistic terms, making use of a myopic selection of metrics, and overlooking underlying heterogeneity and complexities. If the New York Times has problems with the way that Prof. Mitloehner and the CLEAR center communicate these topics then they should provide their science-based arguments as to why the approach he is using is incorrect. Not malign his reputation by implying he is putting out information that is “discordant from the scientific consensus” with no further elaboration as to what exactly he said that was incorrect, or which specific consensus his information is discordant from. Nor discredit his work solely because of the industry funding he clearly discloses. Because that is how science works, not through “ad hominem” attacks.

A Word About Funding Graduate Students

When I started as a professor at Davis 20 years ago, the cost of in-state graduate student fees was $5,037 per year (including health insurance of $966/year). Today they are $19,378 (including $5,472/year in health insurance). Add an additional $15,102 if the student is from out of state or international for a whopping $34.5K/year. To be clear the faculty advisor pays this cost, NOT the graduate student. And that does not cover a stipend for the student to actually live. A 50% teaching assistantship pays around $2,583 per month, or $30,995 per year. So faculty need to budget $19,378 + $30,995 = $50,373 per year for in-state graduate students and $65,475 per year for out of state and foreign students. Whereas a first year postdoc salary, where the scholar already has their PhD and who works 100% time and takes no classes, is $55,632 salary + $11,905 benefits = $67,537/year.

So there is now a perverse incentive to employ postdocs rather than train graduate students, especially foreign graduate students, which leaves the obvious question of who is going to train graduate students? When hiring a graduate student at 50% time is almost the same cost as hiring a 100% time postdoc something is wrong with the system. In times past, graduate student fees for students on research assistantships used to be paid by the State on so-called “19900” funds, a benefit that was quietly eliminated last year. And way back when I did my graduate training in the 1990s as a foreign student the University waived my out-of-state tuition, rather than charging my faculty advisor. I can foresee a time when it will be cost-prohibitive for faculty to train graduate students, especially foreign students.

This is especially true if faculty are using extramural grant funds that have been charged full overhead, which is currently set at 59.5%, and will be 60% in 2023. What this means is that for each $100,000 in grant funds, only $40,500 is available as direct costs to pay salaries and benefits, the remainder goes to indirect costs.  So faculty effectively need to bring in $95,810 of extramural research funds annually to cover the expenses of a single year of an in-state graduate student, of which a total of ~$65K does not go the student ($45,477 goes to indirect costs, and $19,378 goes to fees and health insurance). And that does not begin to pay research expenses associated with the student’s project.

One way out of this is to have the graduate student take a teaching assistantship which covers their fees and stipend for the quarters in which they are teaching. But by definition they are working 50% time teaching during that quarter which leaves little time for research.

When I started as a professor 20 years ago this month, graduate students were a very small part of the research budget, because for the most part their fees were covered by the State, and the faculty were typically only responsible for the stipend. Training graduate students is a core part of the University’s mission. But it really is becoming cost-prohibitive to take on new graduate students and given projected increases in overhead rates and fees, I do not see this situation improving.

DNA is NOT a drug. And regulating genome edited research animals as a drug is unworkable.

Investigational research animals that have been genome edited CANNOT enter the food supply in the United States, irrespective of the edits they carry, unless the researcher that has produced that animal has submitted an FDA Investigational New Animal Drug (INAD), and additionally has obtained a food use authorization which requires a TON of paperwork and also data collection. PERIOD. So when I read the headline, “Gene-edited beef cattle receive regulatory clearance in U.S.”, I felt the need to pen this BLOG. And it is a little wonky (in the political sense) because that is the nature of regulatory discussions.

While it is true that the FDA decided to exercise enforcement discretion for two slick founder animals produced by a company, Recombinetics, this “enforcement discretion” option is only available to developers. And this decision was not a “regulatory approval”, but rather it was a determination by the FDA that this product was low risk enough that it is not an FDA enforcement priority. In other words, a blind eye will be turned to the sale of this “unapproved animal drug” as it is deemed low risk. And that would be quite appropriate given the low risk of this product. But what is not obvious from the press coverage is 1) This decision is made on an individual animal basis (i.e. the data for EACH genome edited animal needs to be submitted to the FDA) and this particular decision was for two cattle and their future offspring, and 2) This path is only available to commercial developers who are bringing a product to market, not academic or university researchers who are researching genome editing.

If I, as an academic researcher, was to perform exactly the same edit and make a slick genome edited cow, the only way that I could have that animal enter commerce and the food supply would be to open an INAD, and then request a food use authorization. And I know exactly the amount of work involved in doing that as I have spent basically the entire pandemic trying to obtain food use authorization for some genome edited knock-out (gene deletion) research sheep and cattle we have produced at UC Davis. Spoiler alert – I did not obtain a food use authorization. However, I was eventually given permission to render them – meaning their carcasses could enter the animal feed chain (i.e. dog food).

To request an INAD, the sponsor (i.e. me in the case of a university professor), must provide the FDA information including an overview of the project objectives, a description of the constructs or genomic alterations in the animal, a description of the method(s) or technologies used to produce and deliver/introduce the genomic alteration, and a description any completed studies including any relevant information from other animals in which the same genes have been knocked out (e.g., references to published articles characterizing knockout mice). Additionally, a request for “a categorical exclusion from the requirements to prepare an environmental assessment” must be submitted to accompany the INAD file. When FDA reviews and approves new animal drugs, it has to comply with the requirements of the National Environmental Policy Act, which includes a review of environmental risks, if any, where required. Requesting this exclusion entailed certifying that no extraordinary circumstances existed whereby the gene-edited cattle could significantly affect the quality of the human environment. This required documenting the location of facilities where the animals were housed, containment measures, management practices and conditions for all facilities, animal waste and animal disposal, and a description of standard operating procedures, and submitting this information to allow for a determination to be made.

Upon obtaining an INAD, then an investigational food use authorization (FUA) must be requested to introduce the knock-out sheep and cattle into the food supply at the completion of the experiment, rather than incinerating them. For this, the information in Table 1 was requested.

The “Predicted Off-targets” in silico analysis included 10 sequences with three mismatches to the guide sequence in the reference cattle genome. There were no sequences with less than three mismatches other than our target, as per our guide design. Given the (graduate student) time and experimental expense involved with performing this off-target sequence verification for 10 loci for each of the animals, and the fact that these analyses were unrelated to the graduate student’s  scientific interest in these animals (meaning we would be doing these analyses solely for the FUA), the FUA was abandoned in favor of requesting a rendering use authorization.  This meant that the animals would be allowed to enter the animal feed chain, rather than the human food chain at the completion of the experiment.  To obtain rendering authorization required a separate application to the FDA Division of Animal Feeds. Similar information to that outlined in Table 1 for the FUA was requested, but this request did not require off-target sequence verification.

Such regulatory interactions are not a routine part of basic academic research in animal breeding and genetics. Neither is disposing of all animals by incineration, burial, or composting at the completion of a research project. This is particularly expensive when working with large livestock species, and is a unique expense associated with the use of gene-editing in food animals not incurred when researching conventional breeding methods. Further, these regulatory data are typically being requested prior to the conclusion of the experimental work, rather than at the completion of the research when manuscripts describing and detailing the methods and results are typically being compiled and written. As such, preparing these documents required that I dedicated a substantial amount of time solely for regulatory compliance, and pursuing a FUA would have required further additional experimental work. This experience has definitely dampened my enthusiasm about performing further research with genome edited livestock in the United States.

I should mention that there are also fees associated with all of this. As an academic institution, UC Davis is eligible for a fee waiver, however the Animal Drug User Product and Sponsor (ADUFA) Fees are substantial. And while these fees might be quite reasonable for an actual new animal chemical drug evaluation, it becomes hard to wrap your head around a half million dollar fee to get a single nucleotide polymorphism approved, when nature has made literally millions of unregulated SNPs in cattle genomes. UC Davis actually recently received an errant  ADUFA bill for over $140,000 as the fee waiver requests to the FDA were apparently misplaced, and that definitely got the attention of the University!At the end of the day, most academic researchers did not sign up to be regulatory scientists. And while I appreciate the need for such scientists to provide data to get approval for commercial products, it is not something that is of academic interest to me. I have little desire to undertake research to show that there are no substantial differences in nutritional or composition content of edible tissues between edited and non-edited animals unless that was the purpose of the edit (my upcoming paper on the meat and milk composition of the offspring of a genome edited polled bull not withstanding! Spoiler alert **** THERE WAS NO DIFFERENCE BECAUSE HORNS ARE UNRELATED TO MILK AND MEAT COMPOSITION OR SAFETY*****), nor in verifying the absence of hypothetical off-target sites in the genome so that our experimental knock-out animals can enter the food chain. I fear that this regulatory approach will effectively preclude public sector research into genome editing with food animals in the United States.

Public Acceptance of Animal Genomics and Biotechnology

Animal biotechnology is the application of modern molecular techniques to animals. Genetic engineering and cloning are two older forms of animal biotechnology , and genome editing is a more recent entrant. Animal genomics is the scientific study of structure, function and interrelationships of both individual genes and the genome in its entirety. Utilization of genomic information in breeding is often referred to as genomic selection (GS). In my view these two fields – biotechnology and genomics – face entirely different public acceptance issues. I wrote a conference proceedings paper for a forthcoming conference  in November 2021, and drew on some of my work in the past in this area. However the “My thoughts” section of the paper is new and published here for the first time. The entire paper  (which is not yet peer-reviewed) is available on my laboratory website . 

My thoughts

                There exists a considerable literature castigating “scientists” (typically meaning research professionals and bench practitioners) for poor communication with the public on the topic of genetic engineering and cloning, and more recently genome editing and GS. The contention seems to be that this failure to communicate uncertainty is what historically “provoked public alienation and fomented controversy” around these technologies, and that this will likely cause problems for genome editing and GS. I beg to differ. Unless these later two topics become politicized, or perhaps more importantly competing business interests develop an approach to monetize fear around these technologies by inflating public perceptions of risks and arousing opposition in an attempt to trigger a spiral of silence (Scheufele, 2014), they will be integrated into livestock breeding programs largely without public scrutiny in the same way as other breeding advancements have been. Artificial insemination has not been recently communicated to the public, and yet its use is routine. However, if they become targeted, both bench and social scientists will have a hard time being heard above the drone of misinformation on social media where science and politics are inextricably linked, similar to what we observed with communications around uncertainties and relative risks associated with COVID vaccines and treatments.

I use the following evidence and observations to support these assertions. There is no money to be made opposing GS. There is no “Non-GS Project” label. There are no large multinational companies controlling its use that can be used as a proxy for evil (e.g. Monsanto). I do not foresee a targeted campaign to preclude the use of GS in genetic improvement programs, in part because it is founded on naturally-occurring genetic variations, and in part because it is hard to problematize into a clean, dichotomous framing i.e. genomic bulls are “bad”, and conventionally-selected bulls are “good”. And while many of the same criticisms leveled against GE and cloning can be equally associated with GS (e.g. increasing the rates of inbreeding), these concerns are likewise associated with conventional selection programs.

Artificial insemination reduces genetic diversity, and conventional selection programs include traits like docility which could be considered a behavioral disenhancement. Layers are selected to not exhibit broody behavior. I am unaware of any campaigns to preclude the incorporation of temperament traits into breeding goals for ethical reasons, despite the fact this clearly alters the telos of the animal. Additionally, there are glaring disparities when it comes to the implementation of GS in the developing world, and even in small breeds; it is expensive to develop large populations of genotyped, phenotyped animals. It is not a scale-neutral technology, advantaging large breeds and genetic providers over small ones. Such inequality concerns would be problematic for a GE application, yet these concerns are rarely even discussed as it relates to GS, and they have not precluded the adoption of this technology. Genomic selection is not a perfect science, there are uncertainties and emerging issues (Misztal et al., 2021), but it is the most accurate tool we have to select the future performance of the offspring of an individual. The absence of an additional regulatory layer to the use of genomic testing has allowed the unfettered, uncontested and rapid adoption of GS in livestock breeding programs globally.   

Cloning is clearly unnatural, well at least somatic cell nuclear transfer (SCNT) cloning is unnatural in that it takes place in a laboratory. Cloning is actually rather common in nature, as evidenced by identical twins. Cloning elite animals has no obvious benefit to the consumer, and really is not that useful in breeding programs as it replicates the current generation rather than the next generation. It has had limited application in serving as a genetic insurance policy, and at times enabling the production of elite sires using less resources (Kasinathan et al., 2015).

By these metrics it would appear cloning is destined for market failure. And it has been effectively banned in the EU. In the Netherlands, the Dutch Animal Health and Welfare Act and Animal Biotechnology Decree prohibited the application of biotechnology to animals without a specific license. Criteria for being given a license included: the goal serves a public interest, has no unacceptable impacts on health and welfare of animals and does not raise any overriding ethical objections.  It is characterized as a ‘No Unless’ policy – no application of biotechnology to animals unless there is a very good reason for doing so. Since 2005, Denmark has required special licensing for animal biotechnology through the Act on Cloning and Genetic Modification of Animals. This legislation came about in large part due to ethical concerns surrounding the impact of biotechnological applications on animal integrity. This Act effectively limits the commercial use of animal cloning and genetic engineering to “creating and breeding animals producing substances essentially benefitting health and the environment”.

However, in countries where cloning is allowed, opposition to cloning has slowly faded, and it is being adopted where it is cost-effective – mostly in high-value recreational animals like bucking bulls and polo ponies. I would argue in countries where clones are not regulated differently to conventional breeding, and products from clones are not labeled as they are in fact impossible to differentiate from products from non-cloned animals –(despite the apparent green milk moustache in Figure 1!), there has been no way to effectively monetize fear around clones.

Figure 1. The Center for Food Safety depiction of cloned milk from their 2007 campaign against animal clones.

The Center for Food Safety, Consumers Union, Food and Water Watch, The Humane Society of the United States, the American Anti-Vivisection Society, the Consumer Federation of America and the Organic Consumers Association tried hard in the early days of cloning, but at the end of the day it is hard to create a convincing argument that a cloned product is somehow more dangerous than its identical progenitor. And in the absence of tracking or labeling requirements, it was just not possible to create a cost-effective “absence-labeling” campaign as was done with rBST and GMOs. 

It is worth noting that a lucrative pet cloning industry has emerged in the absence of regulatory oversight of non-food applications of cloning. In fact, Barbara Streisand recently took on two puppies cloned from her dead dog for the fee of $50,000.  If there is a direct benefit, at least in the mind of the person cloning their pet dog or bucking bull, then people are willing to overcome their hesitations regarding cloning. And as to the entry of these clones into the food supply, it is mostly a moot point. Undoubtedly products from cloned livestock – elite breeding stock at the end of their productive life, and even bucking bulls at the end of their bucking career have entered the food supply on a limited scale.  And considering that the US exported 190 million dollars’ worth of bovine semen in 2018, it is more than likely that there are offspring of clones running around globally.

And so we come to genome editing, the new kid on the block. And its fate is currently uncertain. Public perception is still forming around this technology, but I have a sinking feeling that genome editing will suffer the same fate as GE animals for the following reasons. Firstly, competing market forces have already started to conflate the two technologies. The Non-GMO project has come out with the following announcement “GMOs are now being created with newer genetic engineering techniques, some of which do not involve transgenic technologies. The Non-GMO Project is committed to preventing these new GMOs from entering the non-GMO supply chain.” The National Organic Standards Board voted to exclude all genetic modification and manipulation from organic production in 2016, including genome editing. And Greenpeace in their 2021 position paper entitled “Danger Ahead. Why genome editing is not the answer to the EU’s environmental challenges”, warns that the use of so-called gene (or genome) editing techniques like CRISPR-Cas could not only exacerbate the negative effects of industrial farming on nature, animals and people, but it could effectively turn both nature and ourselves (through the food we eat) into a gigantic genetic engineering experiment with unknown, potentially irrevocable outcomes.“ And so we again have a situation where activist groups and the natural and organic food industry will monetize fear and run a campaign of misinformation to suggest that genome edited animals are “unsafe”, whilst animals with naturally occurring genetic variants are “pure” (and also more expensive!).

Secondly, irrespective of the nature of the genome edit, the proposed regulatory approach to genome edited animals is the same as for GE animals, in both the EU and the United States. Even SNPs and deletions are being treated as drugs in the US. The absence of one intentionally altered base pair among 3 billion in the bovine genome thus results in an unsaleable new animal drug. By capitulating to this regulatory logic and tacitly agreeing that the emperor is wearing clothes, we replicate the situation where only large companies will be able to afford the regulatory and IP costs of bringing a genome edited animal product to market. Hitherto, the IP in livestock breeding has been primarily protected by secrecy and use of cross-breeding (Bruce, 2017). Small companies and academic laboratories will be unable to make use of a technology that originally resulted from public research funds. They will again be relegated to the sidelines, unable to afford even experimental work in large animals as all milk, meat and eggs from all genome edited “investigational animals” are unsaleable, and the animals themselves have to be composted, buried, or incinerated. There is then little incentive for public sector scientists to stick their neck out doing public communication around a technology they cannot use. Especially when doing so will likely result in hostile freedom-of-information act requests, and reputational defamation by front groups financed by the natural and organic food industry such as U.S. Right To Know (Kloor, 2015).

Figure 2. There are four species of transgenic fluorescent GloFish® available in six colors. Photo courtesy GloFish.

At the end of the day, I am not convinced widespread public opposition is what is preventing the adoption of new animal biotechnologies. The prevailing narrative repeated verbatim is that the public outright rejects GMOs. But that is not observed in actual purchasing behavior when GMO products are available. For example, GloFish® (Figure 2) are marketed to aquarists in the US, where they are now sold in every state in the nation, as well as throughout Canada. Sales represent approximately 15% of US aquarium fish sales. Although some authors raised early environmental and ethical concerns about GloFish (Rao, 2005), these concerns have waned over time.

GloFish is subject to enforcement discretion in the US. This is not a determination of “safety” under the Federal Food, Drug, and Cosmetic Act but is instead a determination that, based on risk, FDA does not believe it would be a good use of its limited resources to act against sponsors for the marketing and distribution of these unapproved products. Its sale is prohibited in other jurisdictions, including Europe, Australia, and Singapore. The success of this product suggests that consumers are willing to purchase GE animals, at least as aquarium pets. Alan Blake, CEO of the company marketing GloFish, wrote regarding public acceptance that consumers will purchase a product that they desire, irrespective of the breeding method that was used to produce it. In his words, “It is not about the process [of genetic engineering], it is about the product” (Blake, 2016).

Similarly, the Impossible Burger, a soy-based food product is proudly GMO with it recombinantly produced, bleeding leghemoglobin, has been a market success. Ironically the same anti-GMO groups that targeted GE in agriculture; GMO Watch, Consumer Reports, and the Center for Food Safety, went after Impossible Burger for using GMO heme and soy. They perpetuated the same fearmongering around GMO in Impossible Burgers as they had used around GMO in corn – claiming it hurt rats in a feeding study. And Impossible Food fought back, Rachel Conrad, Chief Communications officer wrote, “Finally, we’d like to request that Consumer Reports disclose its anti-GMO agenda in full transparency, and the biases of its activist employees. For years Consumers Reports, and fellow anti-GMO ideologues have been waging a PR war against GMOs — whether in vaccines, insulin, cheese or more recently the Impossible Burger.” And likewise, the PinkGlow GE pineapple that contains lycopene, a pigment that gives some produce its red color has been success, fetching a premium of as high as $50 per pineapple.

These GE applications might be considered frivolous, after all we can live without fluorescent aquarium fish and pink pineapples. But they are market successes because 1) they were allowed to come to market, and 2) they are products that the customer wanted with at least a perceived benefit. One thing is for sure – if products are not commercially available because it is cost-prohibitive, or even impossible to get regulatory approval, then the public will not be able to indicate their acceptance by purchasing them. That has essentially been the situation for GE food animals for the past 35 years (Van Eenennaam et al., 2021). And for GE food in Europe more generally, although there is of course a glaring incongruity there. In 2018 alone, the EU imported more than 30 million metric tons (MT) of soybean products, 10 to 15 million MT of corn products, and 2.5 to 4.5 million MT of rapeseed products, mainly for livestock feed. The EU’s main suppliers are Argentina, Brazil and the United States. The share of GE products of total imports is estimated at 90-95 percent for soybean products, 20-25 percent for corn, and less than 20 percent for rapeseed (USDA Foreign Agricultural Service, 2018), suggesting GMO feeds are a resounding market success! Again the lack of a requirement to label milk, meat and eggs from animals fed GE feed avoided demonization of these products and has facilitated the widespread use of GE feed in animal production systems globally.  

I’ll leave you with a 60 second video of a delicious meal we prepared recently featuring the first approved GE food animal – the AquAdvantage salmon. Delicious.

Contemplated Regulatory Framework (Part #4)

This is part 4 of a 1, 2, 3, 4 part series on Regulation of Genetically Modified Animals

I am in perhaps a somewhat rare position regarding the contemplated USDA regulatory framework, as I actually have several animals from amenable species (sheep, cattle) on the ground that were developed using “techniques that use recombinant, synthesized, or amplified nucleic acids to modify or create a genome” for “agricultural purposes”.  I have not published these data and they are the dissertation research projects of graduate students, and so I will only say that these animals were produced using CRISPR/Cas9 genome editing and a sgRNA to produce a gene knockout. This was done by introducing a Cas9-sgRNA ribonucleoprotein (RNP) into single cell zygotes following in vitro fertilization of in-vitro matured oocytes.  The resulting double strand break (DSB) was repaired by the non-homologous end joining (NHEJ) pathway, and subsequent embryo transfer to surrogate dams. In other words, each animal is a unique individual, and there was no foreign template or intergeneric combination of genetic material introduced into the editing process at any time. None of the genome editing target genes were intended to impact animal health, nor was the application intended to have an animal health claim. So as I read through the contemplated regulatory framework, I cant help but personalize how it would affect my research, and how it compares to the existing regulations.

By way of background, it is perhaps important to emphasize that the animals that make up the research herds and flocks at land-grant universities are typically entering the food supply at the end of experiments.  Our dairy cattle getting fed different experimental rations keep getting milked, the calves from a beef cattle crossbreeding study are finished in our feedlot and enter commerce at the end of the study, and our poultry produce eggs that are sold to the public, along with meat products from all our animal facilities (except the horses!).  There we sell foals at the UC Davis foal auction. The food animals are actually processed at our USDA-inspected slaughter facility, and butchered by meat science students. The income from the “meat lab” sales goes back into the operational account for the animal facilities.

So back to the genome-edited knockout animals I have on campus. Under the existing FDA regulations, they are considered unapproved animal drugs and cannot enter the food supply without a new animal drug approval or prior authorization. According to the FDA, I would need to apply for an investigational new animal drug (INAD) for these animals, and then request a Food Use Authorization (FUA) for these experimental animals to enter the food supply at the end of the experiment. I have actually already started to apply for that and the FDA have requested an answer to the following questions before they can even open an INAD, let alone tell me what information I need to provide to them to potentially obtain a FUA for my experimental gene knockout sheep and cattle. I have started on this paperwork, but it will take some time to prepare it all and submit it through their electronic portal.

  • Description of the project, including the goal of the intentional genomic alterations (IGAs), any relevant information from other animals in which the same genes have been knocked out (e.g., references to published articles characterizing knockout mice), and a description of the planned studies
  • Description of how the IGAs were produced, including detailed experimental design
    • Include information such as which Cas9 nuclease variant was used (e.g., SpCas9, SpCas9-HF1) and any associated information regarding the variant’s fidelity, criteria used to design and select guide RNAs, and any controls utilized to mitigate the potential for unintended alterations
  • Available molecular characterization data, including a description of the methods used (e.g., PCR, Sanger sequencing, long-range PCR, etc.)

So under the USDA contemplated regulatory framework, published in the Federal Register 12/28/2020, these knockout animals would clearly fit under the “amenable species modified or developed using genetic engineering and intended for agricultural purposes and human food”. If they were knockout plants they would have an up-front exemption from regulation under the SECURE revision as they “result from natural cellular repair of a targeted DNA break without any introduced DNA to direct the repair”.

For the USDA animal health safety review I would need to show that “the animal under review is found to pose no greater risk to animal health than the animal from which it was derived”. This is where is gets a little unclear.  The term “the animal from which it was derived” suggests that the animals are being derived from a known unmodified animal, presumably through genetic engineering of a cell line derived from that animal, followed by somatic cell nuclear transfer cloning of the edited cells. But my animals are derived from editing multiple zygotes, so each edited animal has its own unique genomic sequence i.e. there is no animal from which these animals were derived, and conceptually editing outcomes differ between all of these different individuals, especially because the DSB was repaired using the error-prone NHEJ pathway.  So does each individual animal have to have a different evaluation?

The safety review requires “a molecular characterization of the modification and an understanding of the process by which it was introduced, that the intended change was made and that there were no unintended disruptions of endogenous genes, unintended DNA insertions, or off-target changes if the genome was modified without inserting DNA.” It is not hard to prove that the intended change was made, we typically do that with a PCR amplification of the target locus, and sequencing of the amplified PCR product. But to answer the rest of those questions will require, at a minimum, whole genome sequencing of every animal, and proving a negative i.e. proving no unintended disruptions of endogenous genes, unintended DNA insertions, or off-target changes if the genome was modified without inserting DNA.  Both the sequencing and the bioinformatics to try to answer these questions would be very expensive, well beyond the cost of a sheep, or even a cow. This is especially true when every animal has a unique genotype, and there is no way to differentiate off-target changes from spontaneous de novo mutations.

Further, at a minimum, the animal health risk assessment would include an evaluation of the following issues:

  • Molecular Characterization: What is the genetic modification(s) in the animal, how was the genetic modification(s) introduced, and how does the genetic modification(s) alter protein or ribonucleic acid (RNA expression)?
  • Animal Health: Is there scientific evidence that the modified animal could plausibly, either directly or indirectly, increase susceptibility of livestock, including of the animal itself, to pests, non-infectious diseases, or infectious diseases of livestock, including zoonotic diseases? Is there scientific evidence that the modified animal could plausibly increase the spread of pests or infectious diseases of livestock, including zoonotic diseases?  When a plausible pathway to such an increased risk is identified, further analysis would be conducted to evaluate the pathway.  When an animal health claim is made or a modification is known to adversely affect animal health, the review would assess the animal health claim.
  • Environmental Factors: Is there scientific evidence that introduction of the modified animal into the environment may result in environmental impacts that would warrant review pursuant to the National Environmental Policy Act (NEPA) or other statutes?

And in addition, the FSIS review would include an evaluation of the following issues (theoretically pre-slaughter, although obtaining meat for compositional analysis from animals pre-slaughter is obviously problematic):

  • Evaluation of expressed substances: Is there scientific evidence that the genetic modification could result, directly or indirectly, in toxins, chemical residues, or other potentially deleterious substances in meat or poultry products?
  • Allergenicity: Is there scientific evidence that the genetic modification would directly or indirectly alter the allergenic potential of meat or poultry products derived from the animal?
  • Food storage and processing: Is there scientific evidence that meat or poultry products derived from the modified animal could mislead consumers regarding wholesomeness or the need for appropriate storage (e.g., meat that maintains a red appearance even when spoiled)?
  • Compositional analyses of key components: Is there scientific evidence that meat or poultry products from the modified animal are compositionally (e.g., nutritionally or functionally) no different than meat from conventional animals, such that it meets any regulatory definition, standard of identity or other labeling requirement, and consumer expectations for the applicable product?

Needless to say, providing answers to all of these questions for each individual knockout animal would be an unsurmountable hurdle for an academic laboratory. Throughout the USDA contemplated regulatory framework the term “developer” is used. For example, “Under the contemplated regulatory framework, developers could request that USDA conduct a risk-based and science-based safety review focused on animal health”. Historically, developers of genetically engineered plants have been large multinational companies who have the resources to provide the regulatory data to support commercialization and ultimately deregulation of their products. Research universities are not developers in this sense, but we still produce amenable species modified using genetic engineering and intended for agricultural purposes and human food.

One of the questions the USDA poses in the request for comments is,

“How often does a start-up company or not-for-profit university or research organization modify or develop an animal using genetic engineering?”

My answer would be never.  Researchers at universities often modify animals using genetic engineering but never has one developed and commercialized an “amenable species modified or developed using genetic engineering and intended for agricultural purposes and human food.” The two food approvals for genetically engineered animals were both obtained by companies, albeit small companies. AquaBounty for the “AquAdvantage” salmon, and Revivicor for the “GalSafe” pig.  To date neither of these products has been sold for food in the United States, although AquaBounty is getting mighty close.

In closing, as long as animals produced using genetic engineering, even those that could have been produced using conventional breeding,  are subjected to unique regulatory scrutiny not required of identical products produced using conventional breeding, research in food animals using genetic engineering for agricultural applications will be cost-prohibitive in the United States.

Another question the USDA poses is,

“Should USDA exempt certain types of genetic modifications of amenable species intended for agricultural use from regulation?  If so, what types of modifications and why?”

I would suggest that the SECURE up-front exemptions for certain types of modifications in plants, specifically a modification (1) resulting from natural cellular repair of a targeted DNA break without any introduced DNA to direct the repair;  (2) that is a targeted single base pair substitution; or (3) that introduces a gene known to occur in the plant’s gene pool, or causes a change in a gene that corresponds to a version of that gene present in the organism’s gene pool should be extended to genetically engineered animals.  These exemptions were primarily designed to exclude from regulatory oversight organisms that could have been achieved using conventional breeding. Just like the Coordinated Framework calls for.

If you have comments on the contemplated regulatory framework, there is a 60 day public comment period that closes 2/26/2021. I would encourage all people working in this field to post comments. Comments can be posted Federal eRulemaking Portal: Go to  Supporting documents and any comments that are received on this Advance Notice of Proposed Rulemaking may be viewed at

In the words of one commentator

“It, of course, remains to be seen what will be the Biden Administration’s approach to ag biotech; however, given the consistency of approach to the evolving ag biotech regulatory structure by both the Obama and Trump Administrations, it would not be surprising if the Biden Administration also continued the evolution of a more science-based, evidence-based, and risk-based approach to regulation of ag biotech – including genetically engineered animals. It would behoove stakeholders that would support such consistency of regulatory approach to submit detailed comments that would assist USDA in developing such a notice of proposed rulemaking (NPRM).

As with all opportunities to provide comments on Federal agency proposals, it is important to submit concise, substantive, and well-supported and well-documented comments to the administrative docket. In this case, given the trans-Administrations nature of the action, it is important that interested stakeholders use this opportunity to best advantage to urge the Biden Administration to continue a 21st Century approach to regulating ag biotech.”         Keith Matthews, Wiley Rein LLP

Regulation of Genetically Modified Animals Part #3

This is part 3 of a 1, 2, 3, 4 part series on Regulation of Genetically Modified Animals

So what is the USDA’s proposal for the regulation of genetically engineered animals? In a nutshell, it proposes moving regulation of food animals that are genetically engineered for agricultural purposes such as human or animal food, fiber, and labor from FDA to USDA. This means that intentional genomic alterations in food animals would no longer be automatically and mandatorily regulated by the FDA as “drugs”. That is good news, because genetic variation between individuals cannot reasonably be considered a drug. All life on Earth is made up of genetic alterations. It is the very foundation of all selection programs, and indeed evolution itself. In some ways it is going back to the future, as APHIS oversight of  agriculture and forestry products developed by modern biotechnology was envisioned in early discussions of the regulation of biotechnology.

APHIS would conduct a safety assessment of animals that have been modified or developed using genetic engineering subject to the Federal Meat Inspection Act (FMIA), and the Poultry Products Inspection Act (PPIA) with a specific eye to alterations that may increase the animal’s susceptibility to pests or diseases of livestock, including zoonotic diseases, or ability to transmit the same. The Food Safety and Inspection Service (FSIS) would also conduct a pre-slaughter food safety assessment to ensure that the slaughter and processing of certain animals modified or developed using genetic engineering would not result in a product that is adulterated or misbranded, as they do with animals produced using conventional breeding.

This is a definite improvement over the FDA approach. Period. Clearly the approach being proposed has been used in plants, and has allowed at least some genetically engineered crops to come to market, albeit mostly those developed by large biotechnology companies. However, there are still some logical inconsistencies in the proposal as it stands. And I realize that perfect is the enemy of good enough, but I can’t help but look at this from my perspective as an academic working in livestock improvement, who has seen the promise of genetic engineering wither on the vine. Regulatory evaluations have included “Alice-in-Wonderland” evaluations that include questions that have no right or wrong answer. If there is no hypothesis to test it is not possible to do a power analysis or design a sensible experiment. Studies that product developers have conducted to try to address these questions have been used by groups opposed to the technology to suggest the whatever data is provided indicates unacceptable risks as discussed here. At least the USDA proposal is seeking to identifying “plausible” risks, suggesting the need to test at least some hypothesis, rather than a fishing expedition.

First, lets look at what genomic alterations are covered. Specifically, those that are introduced into animals of the “amenable species” (cattle, sheep, goats, swine, horses, mules, other equines, fish of the order Siluriformes (catfish), chickens, turkeys, ducks, geese, guineas, ratites, and squabs) modified or developed using genetic engineering that are “intended for agricultural purposes” such as human or animal food, fiber, and labor. FDA would continue its review of amenable species modified or developed using genetic engineering intended for non-agricultural purposes, including medical and pharmaceutical purposes (other than veterinary biologics), and gene therapies; and in non-amenable species. So genomic alterations in non-food animals remain with the FDA, and so does non-agricultural genetic engineering. Sorry dog and cat people – you remain with the FDA, irrespective of the nature of your edit.

And what is “genetic engineering” in this case? It is defined to mean “techniques that use recombinant, synthesized, or amplified nucleic acids to modify or create a genome”.  Note how this is a technique-based definition. Already, we have run afoul of the 1986 Coordinated Framework for Regulation of Biotechnology which states, exercise of regulatory oversight should be product-risk based…”and should not turn on the fact that an organism has been modified by a particular process or technique.”

The proposed rule goes on to clarify it would “not include conventional breeding methods such as directed breeding, artificial insemination, embryo transfer, selective breeding, cross breeding, genetic backgrounding for purposes of studding, or other practices commonly available to and employed by producers.” I have worked in animal genetics my entire career, and have not the slightest idea what “genetic backgrounding for purposes of studding” means. Seriously. NO IDEA. I am also not sure the difference between directed breeding and selective breeding, but maybe I do not get out enough.

Onwards to some other concerns, there does not seem to be any clear distinction between genome edited animals without intergeneric DNA combinations that could have been achieved using conventional breeding, albeit likely less efficiently that can be achieved using traditional approaches, and genetically engineered animals harboring an rDNA transgene. This distinction was clearly made in the SECURE revision of APHIS’ biotechnology regulations for plants; but not in this contemplated regulatory framework for animals. This seems strange.

The proposal reads “The regulatory framework that USDA is considering would be conceptually similar to the recently updated USDA regulations for the movement of organisms, notably plants, modified or developed using genetic engineering, (i.e. SECURE).  However, due to the differences in experience, biology, and breeding practices of animals as compared to plants, there would be some differences between these regulatory frameworks.  For example, although SECURE includes up-front exemptions from the regulations for certain types of modifications [i.e. plants without intergeneric DNA combinations that could have been achieved using conventional breeding], we envision that all amenable species modified or developed using genetic engineering and intended for agricultural purposes would be subject to permitting requirements for their import, interstate movement, or environmental release until they have undergone an expedited safety review or an animal health risk assessment and been determined not to pose an increased risk to animal health.  We do seek comment on this issue.“

Well I certainly have some comment on this issue!  What are the “differences in experience, biology, and breeding practices of animals as compared to plants” that make a SNP in a plant eligible for an up-front exemption from the regulations, but a SNP in an animal not? Crops naturally produce allergens, toxins, or other anti-nutritional substances, and some rare safety issues that have been associated with conventional plant breeding, such as allergens in Kiwi fruit, or high levels of solanine in potatoes. I have a hard time coming up with analogous examples from animal breeding despite intensive selection for traits of interest. So what is uniquely hazardous, or even risky, about genomic alterations that could have been achieved using conventional breeding in animals, but not plants, that makes cisgenics, SNPs and deletions ineligible for an up-front exemption from the regulations? At the end of the day – both kingdoms provide food, and different regulations for the different kingdoms makes little scientific sense.

I have many more specific comments on this proposal, but they all will come back to this basic point. Regulation should be proportionate to risk, and agnostic to method. If regulations are being proposed that mandate a more onerous pathway for identical products produced one way as compared to another, or differ between kingdoms, or require additional testing only of products produced using one method, then they will tilt the scale to the less onerous pathway. This may not always be in the best interests of society. For over 30 years now animal geneticists have had little ability to employ genetic engineering in animal breeding programs. This comes with an opportunity cost as detailed here.

In contemplating  an improved regulatory approach for genetically modified animals, perhaps it is time to ditch the process-based trigger which requires additional regulatory scrutiny of plants and animals that could have been achieved using conventional breeding, and rather take the advice of the 1996 Coordinated Framework, and that is that regulatory review should be confined to organisms deliberately formed to contain an intergeneric combination of genetic material from sources in different genera (aka foreign or transgenic DNA that could plausibly produce a toxin or an allergen), and that oversight should be exercised only where the risk posed by the introduction is unreasonable, that is, when the value of the reduction in risk obtained by additional regulatory oversight is greater than the cost thereby imposed.

Regulation of Genetically Modified Animals Part #2

This is part 2 of a 1, 2, 3, 4 part series on Regulation of Genetically Modified Animals

The “Guidance for Industry #187” entitled, “Regulation of Intentionally Altered Genomic DNA in Animals” was published in the Federal Register in 2017, and a public comment period followed. I wrote about my concerns regarding this approach in a previous BLOG. The FDA took comments on this draft revised guidance during the 90-day public comment period which closed June 19, 2017. As I wrote in a BLOG in January 2019, and what remains true today, is that there has been ZERO formal response to the many comments by public sector scientists working in this field who submitted detailed comments, and who see that this regulatory approach will make food animal research using genome editing cost prohibitive, and effectively preclude the use of gene editing in food animal breeding programs.

To date, there have been two food use approvals for genetically engineered animals in the US, ever. One, the AquAdvantage salmon, the founder event of which occurred in 1989, and which has still yet to be sold in the US despite approval in 2015 as discussed here. AquaBounty estimated it has spent $8.8 million on regulatory activities to date including $6.0 million in regulatory approval costs through approval in 2015, $1.6 million (and continuing) in legal fees in defense of the regulatory approval, $0.5 million in legal fees in defense of congressional actions, $0.7 million in regulatory compliance costs (~$200,000/year for on-going monitoring and reporting including the testing of every batch of eggs), not to mention the $20 million spent for maintaining the fish while the regulatory process was on-going from 1995 through 2015.

And on December 14, 2020 the FDA announced it second food animal approval for an Alpha-gal (galactose-α-1,3-galactose) knockout “GalSafe” pig. The press release announced the “First-of-its-Kind Intentional Genomic Alteration in Line of Domestic Pigs for Both Human Food, Potential Therapeutic Uses.” This wording “intentional genomic alteration”, at first made me think this was approval for a genome edited knockout. But then I read an article quoting a spokesperson for the developer saying they have been working on the GalSafe pigs since 2007. That is pre-genome editing in food animals. I searched PubMed and realized that this approval was for a traditional gene knockout approach to achieve a homozygous founder event that was followed by cloning, aka old school genetic engineering, and that was first published in the peer-reviewed literature in 2003.

A petition calling for a harmonization of the U.S. regulatory approach to gene editing in food species so that both plant and animal breeders have access to gene editing innovations to introduce useful sustainability traits like disease resistance, climate adaptability, and food quality attributes into U.S. agricultural breeding programs was launched at the Plant and Animal Genome Conference in 2018, and signed by over 300 scientists, and was shared with the federal regulatory agencies in June 2018.

Comments from scientists included:

“It is long-standing U.S. policy to regulate an item derived from biotechnology as a product, not by the process through which it was produced. We must base oversight based on assessed risk, not on the basis of political considerations.”

“Regulatory uncertainty and inconsistency are barriers to innovation. Importantly, it prevents small to medium businesses from having a pathway to market for products that could be of immense benefit, including the welfare of animals. Draft Guidance for Industry #187 is out of step with crop plants and is not consistent with a science-based risk based regulatory framework. It is disappointing that this Draft undermines a regulatory system that has, until now, been regarded as robust and well respected. The Draft is not fit for purpose.”

In the meantime, USDA announced the SECURE (Sustainable, Ecological, Consistent, Uniform, Responsible, Efficient) revision of APHIS’ biotechnology regulations that was published in the Federal Register on May 18, 2020 following a detailed response to the public comments. It includes up-front exemptions from the regulations for certain types of modifications, specifically a modification (1) resulting from natural cellular repair of a targeted DNA break without any introduced DNA to direct the repair;  (2) that is a targeted single base pair substitution; or (3) that introduces a gene known to occur in the plant’s gene pool, or causes a change in a gene that corresponds to a version of that gene present in the organism’s gene pool.  These exemptions were primarily designed to exclude from oversight plants without intergeneric DNA combinations that could have been achieved using conventional breeding, albeit likely less efficiently. Food (e.g high oleic oil from company Calyx) from genome edited plants has already entered the US market.

The FDA doubled down on their new animal drug regulatory approach in a 2020 Nature Biotechnology correspondence entitled “Genome editing in animals: why FDA regulations matters” There, FDA Center for Veterinary Medicine (CVM) Director Steven M. Solomon made the case for

“why it is necessary for there to be regulatory oversight of intentional genomic alterations in animals, even when the intended modification seeks to replicate a naturally occurring mutation.” He then specifically distances his argument from intentional genomic alterations performed in organisms from other kingdoms that we eat, i.e. plants and microbes, with the statement “Readers should note that our statement here relates to intentional genomic alterations in animals; we are not commenting on alterations in plants or other organisms.”                      FDA CVM

As I wrote in my BLOG earlier this year, Nature Biotechnology published its own editorial response the FDA’s correspondence countering that, “the origin of a DNA arrangement (conventional breeding, recombinant DNA or gene editing) makes little difference to an animal. The genomes of domestic cattle contain millions of natural variants: the 1000 Bull Genomes Project found >86.5 million differences (insertions, deletions and single nucleotide variants) among cattle breeds. According to prominent researchers in the field, none of these variants has been shown to produce ill effects on consumers of milk or meat. Amidst this background of innocuous variation, how can the presence of one identifiable variant justify the costs and delays of mandatory FDA oversight?

Nature Biotechnology finishes with the argument that,

“A cautious, process-based regulatory route keeps the FDA out of trouble and lowers litigation risks for CVM’s lawyers. But the agency could still alter course without reversing direction completely.

Mandatory oversight could be phased out to a system whereby the agency exercises discretion over which gene-edited animals are regulated according to the hazard represented by the introduced trait. This would be consistent with USDA policy and longstanding US regulatory policy. It would give the animal biotech sector a chance to bloom. And it would counter the narrative of fearmongers who would taint all gene-edited animals as hazardous to public health and injurious to animal welfare.”                                          Nature Biotechnology

And so we come to December 2020, and the UDSA advance notice of proposed rulemaking “Regulation of the Movement of Animals Modified or Developed by Genetic Engineering “ which was published in the Federal Register on 12/28/2020. What is the USDA proposing? And how does it differ from the FDA’s approach? Does the contemplated regulatory framework improve things? More on that in the next BLOG.


Regulation of Genetically Modified Animals Part #1

This is part 1 of a 1, 2, 3, 4 part series on Regulation of Genetically Modified Animals

Ever since the US Food and Drug Administration (FDA) announced its plans to regulate genomic alterations in genome edited animals as veterinary drugs in January 2017, I have been a vocal critic of this regulatory approach. I am a livestock geneticist, and have seen the negative impact that this expensive and unpredictable regulatory approach has had on the development of genetically engineered animals. Even today, there has not yet been the sale of a single food product from a genetically engineered animal in the United States, although AquaBounty is getting close with their approved, fast-growing AquAdvantage Atlantic salmon. And just this month a second product, a gene knock-out pig was approved by the FDA, although the company has also not yet sold any pork product to consumers.

Compare that to the progression of genetically engineered crops which have been commercialized for more than 22 years, and in 2019 alone they were grown on 190.4 million hectares by 17 million farmers in 29 countries. Clearly there is something different about the trajectory of genetically engineered food products developed in the animal kingdom as compared to the plant kingdom. Yet at the end of the day, products from both kingdoms end up on the plate of omnivorous consumers. Is there something inherently risky about genetically engineered food animals that explains this discrepancy?

Recently, the United States Department of Agriculture (USDA) came out with an advance notice of proposed rulemaking which was published in the Federal Register on 12/28/2020. This proposal contemplates a different regulatory approach for genetically engineered animals of certain “amenable” food animal species that are “intended for agricultural purposes”. There is a lot to unpack in this document, and so first some background information might be helpful. This is going get a little wonky.

Thirty five years ago, the first paper documenting the production of genetically engineered food animals was published in 1985. A year earlier, a December 1984 Notice “Biotechnology regulation; coordinated framework” came out in the Federal Register, with the introduction reading, “Only forty years ago, DNA was discovered to be the repository of genetic information…”.

In that document, USDA outlines it plan to use its existing  regulatory framework to regulate agriculture and forestry products developed by modern biotechnology. In 1986, the White House, Office of Science and Technology Policy (OSTP) published the Coordinated Framework for Regulation of Biotechnology. This document states, “This framework has sought to distinguish between those organisms that require a certain level of federal review and those that do not. This follows a traditional approach to regulation. Within agriculture, for example, introductions of new plants, animals and microorganisms have long occurred routinely with only some of those that are not native or are pathogenic requiring regulatory approval.” The document goes on to clarify that for genetically engineered plants, the “regulated article” would be defined as any organism or product altered or produced through genetic engineering, if the donor organism, recipient organism, or vector or vector agent belongs to a group of organisms designated by the proposed regulations as having plant pests or any organism or product which USDA determines is a plant pest. A similar approach was envisioned for the regulation of genetically engineered animals.

The Coordinated Framework was subsequently updated in 1992 to include “Exercise of oversight in the scope of discretion afforded by statute should be based on the risk posed by the introduction and should not turn on the fact that an organism has been modified by a particular process or technique”. Additionally, it was clarified that “(O)versight will be exercised only where the risk posed by the introduction is unreasonable, that is, when the value of the reduction in risk obtained by additional oversight is greater than the cost thereby imposed.”

Back in 1992, genetic engineering typically involved random genomic insertion of intergeneric DNA combinations (i.e. transgenic construct) that expressed a protein which resulted in a desired outcome in the target species. For example, fast growth in the case of the AquAdvantage salmon, disease-resistance in the case of a mastitis-resistant cow, or decreased phosphorus in the manure of the “EnviroPig”.

For a number of years, regulatory oversight of the movement of both genetically engineered plants and animals rested with the USDA. Early work with transgenic goats undertaken by my colleagues at UC Davis in the early late 1990s and early 2000s was done in consultation with the USDA’s Animal and Plant Health Inspection Service (APHIS). Under the Animal Health Protection Act (AHPA), APHIS is authorized, among other things, to prohibit or restrict the importation and interstate movement of live animals to prevent the introduction and dissemination of diseases and pests of livestock within the United States.  The AHPA broadly defines the terms “livestock” as “all farm-raised animals”, and “animal” as “any member of the animal kingdom (except a human)”. APHIS currently uses its plant pest authorities under the Plant Protection Act to assess and regulate the movement of genetically engineered plants into the environment.

A chapter in a 2004 PEW report which came out almost 20 years after the first genetically engineered livestock were reported in 1985 entitled “Regulating Genetically Engineered Animals”, read

“because GE animals are so new and are still largely being used only in research, the agencies likely to oversee them have not yet established clear overall or product-specific policies for regulating them under existing laws. Regulators, researchers, developers, and potential consumers are thus currently navigating in uncertain waters, and the discussion of regulatory policies in this chapter is necessarily somewhat speculative.”

In that PEW report the terms “genetically engineered” and “transgenic” were used synonymously, as genetic engineering at the time involved the random insertion of recombinant DNA (rDNA) constructs.

Then in 2009, the FDA Guidance for Industry #187 entitled, “Regulation of Genetically Engineered Animals Containing Heritable rDNA Constructs”, announced FDA’s intent to regulate all genetically engineered animals modified by rDNA techniques, including the entire lineage of animals that contain the modification, under the new animal drug provisions of the Federal Food, Drug, and Cosmetic Act (FD&C Act). In that act, a new animal drug is defined as “an article (other than food) intended to affect the structure or any function of the body of … animals.”

The FDA clarified that they considered the rDNA construct and its expression product in a genetically engineered animal to be the drug, not the genetically engineered animal itself. Although the FDA’s regulatory evaluation was based on attributes of the product, the method used to produce the genetic change, that is rDNA techniques  versus other breeding methods, was the trigger for regulatory oversight. In other words, the trigger for regulatory oversight was based on the process designed to produce the genetically engineered animal, not on the risks associated with the specific characteristics of the animal or its food products (milk, meat or eggs).

And so it continued until January 2017, when the FDA came out with its updated draft “Guidance for Industry #187” entitled, “Regulation of Intentionally Altered Genomic DNA in Animals”. This guidance proposed to regulate all food animals whose genomes have been intentionally altered using modern molecular technologies including gene editing technologies which may include random or targeted DNA sequence changes including nucleotide insertions, substitutions, or deletions, or other technologies that introduce specific changes to the genome of the animal as veterinary drugs.

So there has been a gradual metamorphosis over the past 30 or so years from a proposal for USDA to oversee the regulation of  genetically engineered food animals that pose “unreasonable” risks regarding introduction and dissemination of diseases and pests of livestock under APHIS, to a 2009 draft guidance requiring mandatory FDA premarket new animal drug approval for all animals modified by recombinant DNA techniques; to an updated 2017 draft guidance requiring mandatory FDA premarket new animal drug approval of any “intentionally-introduced genomic DNA alteration in animals produced using modern molecular technologies”, irrespective of risk or even the presence of a heritable rDNA construct. I think that is called regulatory creep. A set of administrative rules giving rise to unintended consequences, because of their broad industry reach.

The Cost of Precautionary, Process-based Regulation (3 of 3)

This is part 3 of a 123 part series on Genetic Engineering in Livestock

In 1996, the year of Dolly  the infamous first somatic cell nuclear transfer (SCNT) cloned mammal and the promulgation of the United States Coordinated Framework for Regulation of Biotechnology , a review paper found that of the 289 published papers in PubMed when searching for the term “transgenic livestock”, the vast majority of original reports focused on growth enhancement, and that 24% of them were review articles. This proportion has not changed much, with that same search today returning approximately 30% review articles. What happened to the promise of transgenic livestock that almost one quarter of the publications in the field are reviews rather than original research articles, and only a single transgenic food animal has been commercialized in 35 years (as discussed in Part 2 of this series)?

The livestock breeding sector is distinct from the plant breeding sector where transgenic plants are grown by over 17 million farmers globally. Some of this is due to biological differences between the two kingdoms, including the mode of reproduction (e.g. plants can self-pollinate), and the number of progeny per reproduction cycle. A major difference between animal and plant breeding is that the former places greater emphasis on population improvement, with product development consisting primarily of multiplication of these improved genetics through outcrossing, whereas in the latter greater emphasis is placed on selection of an improved product in the form of a recognizable plant variety, which is often also the source of parents for the next breeding cycle .

One transgenic plant transformation event (often protected by IP) can be amplified in seed multiplication, but in animals the founder event will have to occur in elite stock in the breeding nucleus and then be transmitted down to the commercial sector through multipliers, a process that can take decades depending upon the generation interval of the species.

Additionally, desired traits will need to be introduced into several breeds or lines in animal breeding programs because they often make extensive use of heterosis (or hybrid vigor) through crossbreeding. For example, a typical modern broiler chicken breeding program is shown in Figure 1.

Figure 1. A typical modern broiler chicken breeding program, represented as a pyramid where each level represents a generation, and the approximate timeline to move genetics from top of the broiler breeding pyramid to the consumer. (Modified)

Genetic improvement and pedigree selection for desired traits occurs at the top of the pyramid across four lines. Within the pedigree segment there are specific male and female lines, with the males typically selected for heritable growth and production traits and the female lines selected for early growth and conformation. The commercial broiler (5th generation) at the bottom of the pyramid is a four-way cross derived from the cross of a male and female parent line.

Genome editing offers an efficient approach to introduce useful genetic variation into livestock breeding programs through targeted inactivation of gene function, and/or through allele introgression in the absence of undesired linkage drag. Genome editing involves the use site-directed nucleases (e.g. TALENS (Transcription Activator-Like Effector Nucleases) or the CRISPR (Clustered Regularly Interspaced Short Palindromic Repeats)/CRISPR-Associated Systems (CAS)) to efficiently introduce a double stranded break (DSB) at a predetermined location in the genome. The DSB can be repaired by the cell’s natural DNA repair mechanism (non-homologous end joining; (NHEJ)), often resulting in single nucleotide changes, deletions or small (1-2 nucleotide) insertions at the DNA cut site. In this case, although the location of the cut site is very precise, the exact change that occurs when the DNA is repaired is random and so a number of different outcomes representing minor sequence changes are possible.

Alternatively, repairs can be directed to introduce, delete, or replace a series of letters in the genetic code using a nucleic acid template. This essentially enables the introduction of known, desired alleles or haplotypes via homology directed repair (HDR) based on what is understood about naturally-occurring genetic variation in the target species. It is now possible to alter animal genomes without necessarily incorporating transgenic genetic material. To date SCNT cloning of edited somatic cells, especially HDR donor-template defined alterations, has been the primary method to produce livestock carrying nuclease-mediated genetic changes, especially HDR donor-template defined alterations, in their genomes.

Researchers have already produced a number of gene-edited farm animals for biomedical research applications, and also for agricultural purposes (see Table 2 in my recent review paper). The latter group includes animals carrying no novel DNA sequences (e.g. Porcine Reproductive and Respiratory Syndrome (PRRS) virus-resistant CD 163 knockout pigs ; myostatin knockout sheep and cattle with increased lean muscle yield ; and knockout sheep with increased wool length and yield ; intraspecies allele substitutions (e.g. hornless dairy cows due to an allele substitution at the POLLED locus ); intraspecies gene insertions also known as cisgenics (e.g. cows with increased resistance to tuberculosis due to knock-in of bovine NRAMP1 gene); and animals with interspecies allele substitutions, analagous to traditional transgenic applications (e.g. domestic pigs (Sus scrofa) carrying an allele from the African Swine Fever-resistant African warthog (Phacochoerus africanus).

Not surprisingly, the traits that have been targeted by researchers with gene editing (animal health and well-being, product quality and yield) are common to the breeding objectives of traditional selection programs. Health and welfare traits are also of particular interest to the general public, especially projects like the female-only layer chickens. As with earlier genetic engineering approaches, whether breeders will be able to employ gene-editing in commercial farm animal genetic improvement programs will very much depend upon global decisions around regulatory frameworks and governance.

A chance to rethink

In early 2017, the United States Food and Drug Administration (FDA) released its updated draft “Guidance for Industry #187” and changed the title to “Regulation of Intentionally Altered Genomic DNA in Animals”. This guidance proposes to regulate all food animals whose genomes have been intentionally altered using modern molecular technologies, including gene editing technologies, as new animal drugs. This includes many of the same nucleotide insertions, substitutions, or deletions that could be obtained using conventional breeding.  No longer is it the presence of a transgenic rDNA construct that triggers mandatory premarket FDA regulatory oversight prior to commercial release, but rather it is the presence of any “intentionally altered genomic DNA” in an animal’s genome that initiates oversight.

This runs against the wording of the 1986 Coordinated Framework which indicated regulatory review was only required for organisms deliberately formed to contain an intergeneric (i.e. transgenic) combination of genetic material. It also runs counter to the OSTP policy,  the 2016 U.S. National Academies of Sciences, Engineering, Medicine report, which recommended a “product not process” regulatory trigger. In addition it is contrary to the USDA approach to the regulation of gene edited food plants, and is also out of step with decisions being made by other regulatory agencies in a number of countries around the world (e.g. Argentina), with implications on global trade. Mandating premarket regulatory approval for deletions, mutations, and the conversion of one wild-type allele to another wild-type allele in the same species (cisgenic) that could have been obtained using conventional breeding is not scientifically defensible, given the known genetic variation that exists naturally in livestock genomes.

Figure 2 shows the global regulatory landscape for gene-edited animals. If those animals do not carry novel DNA (transgenic DNA), meaning that the DNA alteration could have been achieved using conventional breeding, then a number of countries are not requiring additional regulatory oversight. To date only a single entity, a large global animal genetics company, has announced plans to commercialize a gene edited food animal. The University of Missouri signed an exclusive global licensing deal for potential future commercialization of PRRS virus-resistant pigs with UK-based Genus plc. This company has also entered into a strategic collaboration with Beijing Capital Agribusiness Co. Ltd, a leading Chinese animal protein genetics business, to research, develop, register and market elite pigs that are resistant to PRRS virus in China.

Figure 2. Global picture of whether additional regulations will be triggered for gene-edited animals that do not carry transgenic DNA

In a February 2020 correspondence entitled Genome editing in animals: why FDA regulations matters published in Nature Biotechnology, FDA Center for Veterinary Medicine (CVM) Director Steven M. Solomon  made the case for “why it is necessary for there to be regulatory oversight of intentional genomic alterations in animals, even when the intended modification seeks to replicate a naturally occurring mutation.” He then specifically distances his argument from intentional genomic alterations performed in organisms from other kingdoms that we eat, i.e. plants and microbes, with the statement  “Readers should note that our statement here relates to intentional genomic alterations in animals; we are not commenting on alterations in plants or other organisms.”

This suggests there is something uniquely risky about intentional genomic alterations in animals. However, the history of plant breeding is rife with examples where antinutrients in plants have inadvertently been increased through conventional selective breeding efforts (e.g. think solanine in potatoes, or psoralens in celery). I have a hard time coming up with an analogous food safety concern from conventional animal breeding. Dr. Solomon then goes on to support his argument with the a case of a naturally-occurring genomic variant that was deleterious to animal health and that increased in frequency due to conventional selection. “There is a particularly compelling example of the risks of occult genomic alterations in cattle produced by traditional breeding techniques: a high incidence of bovine leukocyte adhesion deficiency (BLAD) syndrome, a lethal autosomal recessive disease, in Holstein calves.”

In an unusual move, Nature Biotechnology published its own editorial response the FDA’s correspondence countering that, “If the BLAD case history tells us anything, it is that the origin of a DNA arrangement (conventional breeding, recombinant DNA or gene editing) makes little difference to an animal. The genomes of domestic cattle contain millions of natural variants: the 1000 Bull Genomes Project found >86.5 million differences (insertions, deletions and single nucleotide variants) among cattle breeds. According to prominent researchers in the field, none of these variants has been shown to produce ill effects on consumers of milk or meat. Amidst this background of innocuous variation, how can the presence of one identifiable variant justify the costs and delays of mandatory FDA oversight?

The US Food and Drug Administration is sticking to its plan to carry out mandatory premarket review of all gene-edited livestock, irrespective of trait risk. It should rethink.

The US Food and Drug Administration (FDA) Center for Veterinary Medicine (CVM) continues to argue that every animal created by gene editing should be subject to mandatory premarket review and substantial safety testing. This position, first outlined in a 2017 draft guidance, presupposes that all gene-edited animal are potentially hazardous — despite the evidence and the views of many researchers. It canonizes a precautionary stance to gene-edited animals, irrespective of the trait engineered, even if the animal is indistinguishable from one created by conventional breeding. It is out of step with the practice of the US Department of Agriculture (USDA), which recently deferred from regulating gene-edited products. And it goes against decades of US policy focusing on appropriate product risks and benefits through the Coordinated Framework.”

Course Correction editorial, Nature Biotechnology, February 2020

I have spent a career in animal genetics and genomics. For a good chunk of that time, it has not been possible to use the tools of molecular biology to introduce useful genetic variation like disease resistance  into livestock breeding programs through transgenesis. Genome editing offered a new opportunity to use molecular tools in a way that would not necessarily introduce transgenic DNA. It was hoped that this would help avoid the regulatory costs that effectively strangled transgenic applications like the USDA ARS’ mastitis-resistant cow,  and University of Guelph’s “Enviropig“. Precluding these transgenic animals from coming to market is associated with a forgone opportunity cost as we have not cured mastitis, and pigs still can’t digest phytate. Genetic solutions that could have helped to address these animal disease and phosphorus pollution problems have just been sidelined.

There was some hope that gene-edited animal alterations that could have been achieved using conventional breeding, would not be treated differently to conventional breeding from a regulatory standpoint. This was the decision of the USDA regarding genome edited plants, and that of several countries (Figure 2). However, the decision in some countries including the US, is to regulate gene-edited animals based on the fact they are gene-edited, rather than product novelty or risk. I fear this will lead to another long list of forgone genetic innovations in US animal agriculture, and advantage countries with a product-risk focused regulatory approach.

US Regulation of Transgenic Food Animals (2 of 3)

This is part 2 of a 1, 2, 3 part series on Genetic Engineering in Livestock

When AquaBounty sought to commercialize the first transgenic food animal in the mid 1990s (first produced in 1989), there was no official regulatory approach in place. Former CEO of AquaBounty technologies Dr. Ron Stotish in a 2012 abstract entitled “AquAdvantage salmon: pioneer or pyrrhic victory” (Transgenic Research 21: 913-914) wrote :

AquaBounty consulted FDA and other government agencies in hopes of identifying a regulatory process that could be employed to review and approve the AquAdvantage salmon for food use in the United States. AquaBounty established an Investigational New Animal Drug [INAD] file with the Center for Veterinary Medicine in 1995, well in advance of any clear regulatory paradigm. Between 1995 and 2009, the sponsor [AquaBounty] conducted a variety of GLP studies aimed at meeting what was hoped to be the eventual regulatory requirement for an application of this nature. Although there was informal consultation and communication between the sponsor and CVM staff during this time, it was not until 2009 that CVM [FDA Center for Veterinary Medicine] released Guidance Document 187, codifying requirements for consideration of an application for a transgenic animal.

This 2009 Guidance Document 187 was entitled “Regulation of Genetically Engineered Animals Containing Heritable rDNA Constructs”. The Federal Food Drug and Cosmetic Act (FFDCA), defines a drug as an “article intended for use in the diagnosis, cure, mitigation, treatment, or prevention of disease in man or other animals;” and “articles (other than food) intended to affect the structure or any function of the body of man or other animals.” A “New Animal Drug includes a drug intended for use in animals that is not generally recognized as safe and effective for use under the conditions prescribed, recommended, or suggested in the drug’s labeling, and that has not been used to a material extent or for a material time.” Using this definition the FDA considered the “regulated” article to be “the rDNA construct in a GE animal that is intended to affect the structure or function of the body of the GE animal, regardless of the intended use of products that may be produced by the GE animal.”

In that guidance the FDA defined GE animals as animals with both heritable and non-heritable rDNA constructs. The classification of transgenes in animal genomes as drugs meant that each animal lineage derived from a separate transformation event (or series of transformation events) was considered to be a separate new animal drug, subject to a separate new animal drug approval. This determination meant that all unapproved GE animals, their offspring, and their food products were “deemed unsafe”. The FDA exercised enforcement discretion for GE animals of non-food-species that are raised and used in contained and controlled conditions such as GE laboratory animals used in research institutions. In other words, researchers working with the literally millions of GE laboratory animals did not have to go through the INAD and NADA  (New Animal Drug Application) requirements.

That left the small group of mostly public sector agricultural researchers working with food animals saddled with the requirements needed to get a new veterinary drug approved if any of their work was to ever reach the market. These requirements include a seven-step regulatory process in which the agency examines the safety of the recombinant DNA (rDNA) construct to the animal, the safety of food from the animal, and any environmental impacts posed (collectively the “safety” issues), as well as the extent to which the performance claims made for the animal are met (“efficacy”).

Molecular characterization of the rDNA construct determines whether it contains DNA sequences from viruses or other organisms that could pose health risks to the GE animal or to those eating the animal. Molecular characterization of the GE animal lineage determines whether the rDNA construct is stably inherited over multiple generations. Phenotypic characterization assesses whether the GE animals are healthy, whether they reach developmental milestones as non-GE animals do, and whether they exhibit abnormalities. A durability assessment reviews the sponsor’s plan to ensure that future GE animals of this line will be equivalent to those examined in the pre-approval review. If the GE animal is intended as a source of food, FDA assesses whether the composition of edible tissues differs and whether its products pose more allergenicity risk than non-GE counterparts.

In the meantime, all investigational GE animals, their littermates, and surrogate dams and their products were deemed “unsafe” and had to be disposed of by “incineration, burial, or composting.” My colleague, Dr. Matt Wheeler, at the University of Illinois has been working on transgenic pigs for over twenty years. His work has included expressing the bovine lactalbumin gene in the milk of transgenic pigs (1998), which enhances lactation performance and preweaning piglet growth rates (2002). This work has been carried out under an INAD, which initially authorized the rendering of a certain subset of experimental animals. A misunderstanding over the regulatory status of cogestating one INAD line of transgenic pigs with littermates of another INAD line of transgenic pigs resulted in the FDA determining that Dr. Wheeler had “failed to follow protocol”. One Friday, Dr. Wheeler relayed to me, that the FDA swept down on his laboratory and confiscated his computer and laboratory note books. They also locked him out of his own laboratory for several days,

The experience he says, “made him do some soul searching as to whether he wanted to continue doing work with transgenic livestock.”

Such actions have done little to build trust with the regulated academic  community. Dr. Wheeler estimates he has incinerated thousands of transgenic pigs and their littermates and surrogate dams as unapproved new animal drugs during the course of his research, and guesses the added costs of being under INAD requirements to be in the vicinity of USD$2 million. Fortunately academic research institutions and small companies are exempt from additionally paying the recurring annual INAD review fee, which can be thousands of dollars.

The genetically modified AquAdvantage salmon (top) gets to market weight in half the time required for its traditional counterpart. AquaBounty.

To continue on with the AquAdvantage saga, Dr. Stotish wrote in 2012 “by mid 2010 [AquaBounty] had concluded all the necessary research, submitted all required regulatory studies, and received the results of the CVM reviews indicating satisfaction with the submitted data in addressing all requirements for approval. The Center convened a Veterinary Medicine Advisory Committee [VMAC] to review the results of the CVM assessment and conclusions on September 20, 2010. ” Little did the company know that there would be a 5 year delay and to complete “radio silence” from the FDA following the 2010 VMAC meeting until approval in November 2015. The salmon has had a long road to market.

The VMAC meeting for the AquAdvantage salmon was held in September 2010. During the course of the meeting the VMAC members discussed the strengths and weaknesses of the data (4). Ultimately the consensus document of the VMAC charged with providing advice and recommendations to the FDA found (1) “no evidence in the data to conclude that the introduction of the construct was unsafe to the animal,” (2) that the studies selected to evaluate whether or not there was a reasonable certainty of no harm from consumption of foods derived from AquAdvantage salmon were  “overall appropriate and a large number of test results established similarities and equivalence between AquAdvantage Salmon and Atlantic salmon,” and (3) that the AA salmon did grow faster than their conventional counterparts.

The potential environmental impacts from AA salmon production were mitigated by the proposed conditions of use which limits production to FDA-approved, physically-contained fresh water culture facilities. The eyed-egg production site is located on Prince Edward Island (Canada), and the grow-out of market fish is proposed to occur in Panama with multiple biological (all female, triploid), physical (land-based tanks with fencing/screening), and geographical (hydroelectric dams with no fish passage, thermal-lethal downstream temperatures) redundant containment measures. The VMAC concluded that although they “recognized that the risk of escape from either facility could never be zero, the multiple barriers to escape at both the PEI and Panama facilities were extensive”.

Activist disinformation campaign

Disinformation that was created to misinform about the nutritional data of the AquAdvantage salmon

The data package that AquaBounty voluntarily made publicly available in advance of the 2010 VMAC in a good-faith attempt to increase transparency in the regulatory process, was used by special interest groups like the Center for Food Safety who misrepresented the data by suggesting that “FDA’s 2010 data release showed that GE salmon have 40% higher levels of the growth hormone IGF-1, which increases the risk of cancer. And finally, wild salmon has 189% higher levels of omega-3 fatty acids than GE salmon can produce.” This false interpretation stands in stark contrast to the FDA finding which was “AquaBounty (ABT) salmon meets the standard of identity for Atlantic salmon as established by FDA’s Reference Fish Encyclopedia. All other assessments of composition have determined that there are no material differences in food from ABT salmon and other Atlantic salmon.” This type of vilification of technology and fearmongering by special interest groups with no interest in truthfulness does nothing to benefit society.

The U.S. Food & Drug Administration ultimately approved AquAdvantage salmon for sale in November 2015. But an obscure rider was attached to a budget bill by Alaska senator Lisa Murkowski in December of that same year, effectively blocking the FDA from allowing salmon into the U.S. In the meantime Canadian regulatory authorities approved the fish in 2016, and sold 5 tons of fillets (4.5 metric tons) there in 2017, and the same in 2018. “The people who bought our fish were very happy with it,” Stotish is quoted in a press article. “They put it in their high-end sashimi lines, not their frozen prepared foods.

It is worth noting that the US imported 339,000 metric tons of salmon in 2016, worth more than $3 billion. The vast majority of that came from farmed Atlantic salmon raised in floating cages off the coasts of Canada, Chile, Norway and Scotland, and flown to the United States. Atlantic salmon raised in land-based tanks like the AquAdvantage salmon is rated “best choice” by Seafood Watch, and  the carbon footprint of salmon produced in land-based closed systems is <50% lower than that of salmon produced in conventional marine net pen fish farms in Norway and delivered to the U.S. by air.

In 2018, FDA approved an application by AquaBounty for a growout facility in Indiana, offering an opportunity to produce US-raised Atlantic salmon. And finally on March 8, 2019, 30 years after the founder fish was produced in Canada in 1989, the FDA  deactivated this import alert pertaining to the GE salmon.  It still has not reached US consumers, despite recent court victories.

AquaBounty estimated it has spent $8.8 million on regulatory activities to date including $6.0 million in regulatory approval costs through approval in 2015, $1.6 million (and continuing) in legal fees in defense of the regulatory approval, $0.5 million in legal fees in defense of congressional actions, $0.7 million in regulatory compliance costs (~$200,000/year for on-going monitoring and reporting including the testing of every batch of eggs), not to mention the $20 million spent for maintaining the fish while the regulatory process was on-going from 1995 through 2015 (David Frank, AquaBounty; pers. comm; January 2020).


While the AquAdvantage fish has been awaiting regulatory approval, salmon breeders in Norway have been busy selecting for fast-growing salmon. The genetic gain for growth-rate in salmon has been estimated at 10–15% per generation. Farmed salmon have been exposed to ≥12 generations of domestication and was the first fish species to be subject to a systematic family‐based selective breeding program which began in 1975. Studies on farmed salmon in the 9-10th generation of selection showed their size relative to wild fish was 2.9:1 under standard hatchery conditions, and 3.5:1 under hatchery conditions where growth was restricted through chronic stress. In other words, selective breeding programs have produced genetically-distinct lines of fast-growing salmon, and since the 1970s, tens of millions of these fertile farmed salmon have escaped into the wild. Glover estimated that over three to four decades, introgression of farmed salmon into Norwegian wild salmon populations ranged from 0% to 47% per population, with a median of 9.1% .

Presumably these fast-growing salmon pose the same risks as the AquAdvantage salmon, although conventional breeding is not regulated and the latter was associated with decades of delay, millions of dollars in new animal drug regulatory approval costs, and all fish must be raised as sterile triploid females in land-based containment tanks and is subject to ongoing regulatory compliance costs.

This might sound like an argument to just let conventional breeding do its job. As a geneticist I have to agree,  and this is absolutely an option, albeit a slower one, for traits that vary in the target population. But for some traits like disease resistance, having the ability to bring in useful genetic variation enables breeders to introduce novel traits that cannot be achieved by conventional breeding. The problem is that it has not be possible to bring transgenic food animals to market, even those that address traits of interest to the consumer like disease-resistance and animal welfare traits. All of it is blocked.

Having a different regulatory standard triggered by the process (i.e. rDNA versus other breeding methods) used to make a product , rather than on the risk associated with the product itself  runs counter to the The United States “Coordinated Framework for the Regulation of Biotechnology,” which is technically agnostic towards the technology or process under review.  According to the Office of Science and Technology Policy (OSTP) in the Federal Register in 1992:

“Exercise of oversight in the scope of discretion afforded by statute should be based on the risk posed by the introduction and should not turn on the fact that an organism has been modified by a particular process or technique … (O)versight will be exercised only where the risk posed by the introduction is unreasonable, that is, when the value of the reduction in risk obtained by additional oversight is greater than the cost thereby imposed ”

This suggests that the U.S. should only exercise regulatory authority over organisms — plant or animal — based on the risks they pose.  This is irrespective of the breeding technique used to produce them, and used only when the risk posed is unreasonable, which is clarified to mean the cost of regulatory oversight is not greater than the reduction in risk obtained by that oversight. it is difficult to see how this applies to transgenic animals given the history of the fast-growing salmon.  It was hoped that new breeding methods like gene-editing would be treated more like conventional breeding. Unfortunately, that is not how it happened, as discussed in Part 3 of this series.

« Older posts

© 2023 Biobeef Blog

Theme by Anders NorenUp ↑