Dr Michael Antoniou and Claire Robinson deconstruct the myths
The Genetic Technology (Precision Breeding) Bill, which removes regulatory safeguards from whole subclasses of genetically modified organisms (GMOs), is reaching its final stages in the UK Parliament’s House of Lords. In discussions on the Bill, we have noticed that UK Parliamentarians in both Houses are being misled with false, misleading, and biased information that ignores the concerns raised by the science underpinning gene editing technology. The false claims are not confined to the UK – variants of them are being used all over the world in debates about proposals to weaken the regulations around gene editing and other GM technologies.
Why is the UK government being misled on this crucial issue – and in turn, passing on misleading information to Parliamentarians? It’s largely because the government has chosen to rely for advice on people with conflicts of interest involving the agricultural GMO industry. The government has appointed GMO developers and people with links to GMO companies and pro-GMO lobby groups to sit on the key committees that advise it on how to regulate GMO products – the same products that that these people stand to profit from. They will also be in charge of deciding whether any given GMO is allowed into our food and fields.
Independent and critical scientists have been excluded from discussions in which key elements of regulation are planned, such as the definition of “precision bred” GM plants and animals and the new risk assessment for GM foods.
Conducting policy on this basis is equivalent to putting a pesticide salesman in charge of pesticide regulation or an oil magnate in charge of energy policy.
Here are some of the myths that are being promoted, along with our responses, which are given both in brief and in detail. (This information has been sent to all interested parties.)
1. Myth: Many mutations occur in natural reproduction. So there’s no need to worry about those that occur as a result of the gene editing process.
Facts in brief: In natural reproduction, certain regions of the organism’s DNA genetic material (genome) are protected from mutations (DNA damage). Also, the genetic variations that arise through rounds of natural reproduction are not random but are biased in the direction of adaptation and survival. This contrasts with the genome-wide mutations that are caused by gene editing processes, which are random and widespread. Some of these mutations could alter the composition of the gene-edited crop making it toxic or allergenic, resulting in a public health risk.
Facts in detail: In natural reproduction, certain regions of the genome are protected from mutations (DNA damage). A review by Dr Katharina Kawall describes how genic regions of the genome (i.e. areas that contain genes) of higher organisms are protected from mutation through different mechanisms, including heightened recruitment of DNA repair machinery.
Furthermore, the spectrum of genetic variations that arise through rounds of natural reproduction also does not appear to be random. A discovery that was originally made in bacteria but now has also been made in plants (Arabidopsis) is that in natural reproduction, genetic variations arise not randomly but in a directed manner that is “biased” towards survival. An important aspect of this directed non-random genetic variation is that certain genic regions that are important to the plant’s survival and adaptation are protected from mutation.
As the authors of this study state, “Mutations occur less often in functionally constrained regions of the genome”. Through mechanisms that are currently unknown, the organism (in this case, a plant) is sensing its environmental conditions and trying to adapt to them at a genetic and epigenetic (gene expression) level.
The authors of the paper conclude, “Epigenome-associated mutation bias reduces the occurrence of deleterious mutations in Arabidopsis, challenging the prevailing paradigm that mutation is a directionless force in evolution.”
These observations – of genic regions being protected from mutation and of the biased directed genetic variation inheritance occurring during natural reproduction – are strikingly at odds with the genome-wide large-scale random mutations, ranging from small DNA base unit changes to massive chromosomal rearrangements, that accumulate from the different phases of the gene editing process (tissue culture, cell transformation, and action of the gene-editing tool).
The widescale random mutations that arise from the gene editing process, taken as a whole, are far more likely to cause gene disruption and function and consequent altered biochemistry of the organism than the genetic variations that arise from natural reproduction. And as numerous scientists have warned, gene editing-induced altered biochemistry could include the possibility that a gene-edited crop will unexpectedly be toxic or allergenic.
Finally, the issue of unintended mutations from gene editing is not simply a numbers game. The quality, as well as the quantity, of mutations must be considered and analysed: not just how many there are, but what they do.
In conclusion, assertions that mutations arising from rounds of natural reproduction are random and widespread – and that as a result, the unintended mutations from gene editing processes are nothing to worry about – are incorrect.
2. Myth: Compared with gene editing, far more mutations occur in chemical- and radiation-induced mutagenesis breeding, techniques that have been safely used for decades in crop breeding. Most of our crop plants have a mutagenesis-bred ancestor. We don’t regulate these techniques, so why regulate gene editing?
Facts in brief: Mutagenesis breeding has been practiced for decades, but was introduced before scientists understood the risks of creating widespread genetic damage in plant genomes. It has an unknown safety profile as no controlled studies have been done. It produces large numbers of deformed and non-viable plants – part of the reason why it has been little used and has produced relatively few crop varieties when compared with traditional plant breeding. Just because one risky technology (mutagenesis breeding) in the past escaped regulation doesn’t mean another risky technology (gene editing) should do so. Two wrongs don’t make a right.
Facts in detail: Chemical- and radiation-induced mutagenesis breeding techniques involve exposing plant material (e.g. seeds, plant cells in tissue culture, growing plants) to chemicals or radiation to cause mutations. These procedures increase rates of mutation by a thousand to a million-fold in the hope that one or more mutations will produce a desirable trait and this plant can then be bred on to produce a line of plants with that trait.
While these techniques have been used for decades, they were introduced at a time when scientists and regulators did not understand the potential consequences of inducing widescale mutations in plants. If they were introduced today, they would certainly attract the attention of regulators. Just because one risky technology (mutagenesis breeding) in the past escaped regulation doesn’t mean another risky technology (gene editing) should do so. Two wrongs don’t make a right.
The safety profile of mutagenesis breeding remains unknown because no one has carried out controlled feeding studies with mutagenesis-bred plants. It’s clear that they are not acutely toxic, but it is not known if there are more subtle effects of consuming them over the long term.
What is known, however, is that the mutations produced by mutagenesis breeding are different in quality from those produced by gene editing, in that gene editing can cause mutations in parts of the genome that are protected from mutations in mutagenesis breeding. But any implications that this difference may have for the food safety of mutagenesis-bred and gene-edited plants are not understood.
Current EU law (EU Directive 2001/18/EC) classes mutagenesis breeding techniques as GM methods that produce GMOs, but it exempts them from the requirements of the GMO regulations (safety assessment, traceability and labelling) because of an assumed history of safe use. Gene editing has no such history of safe use, as the technology is new in agriculture.
It is clear that chemical- and radiation-induced mutagenesis breeding techniques are highly risky for the plants themselves. They produce large numbers of deformed, infertile and non-viable plants, so they are wasteful and time-consuming to use.
No doubt this is an important reason why these techniques have not been widely used in plant breeding. The UN Food and Agriculture Organisation and the International Atomic Energy Agency keep a database of plant varieties that have been generated using chemical- and radiation-induced mutagenesis breeding and by cross-breeding with a mutant plant. The database contains only around 3400 varieties, including ornamental plants. In contrast, Norway’s Svalbard Global Seed Vault contains over 1 million crop varieties. In 2009 it was estimated to contain one-third of the world’s most important crop varieties. Such figures dwarf the number produced by mutagenesis breeding.
Another important factor that has restricted the use of these techniques is that they require specialized facilities and materials. Chemical-induced mutagenesis uses highly toxic chemicals, which are governed by strict regulations and need careful handling. For example, the most common chemical mutagen used is ethyl methanesulfonate, a probable carcinogen. Radiation-induced mutagenesis requires a specialised facility with a gamma-radiation source that must be carefully shielded. It has a greater tendency than chemical-induced mutagenesis to destroy plants’ viability, so it is less commonly used.
In conclusion, mutagenesis breeding produces genetic alterations that are different in not just quantity but also quality compared to unintended DNA damage from gene editing, and has been only of marginal importance in crop development.
3. Myth: Mutations are not bad or harmful, but drive evolution and are therefore to be welcomed.
Facts in brief: Most mutations are harmful and living organisms have evolved mechanisms to minimise them. Legislators worldwide have established laws to minimise the exposure of living organisms to mutagenic (causing mutations) agents.
Facts in detail: Most mutations are harmful. Living organisms have evolved mechanisms to minimise mutation frequency in genic regions during natural reproduction. Those genetic variations that do arise are not random, and result in biased, directed inheritance towards environmental adaptation.
In addition, in recognition of the harmful effects of most mutations, regulators worldwide have established strict laws to protect humans and other living organisms from mutagenic effects of (for example) chemicals and radiation. Thus the notion that mutations are somehow good and to be welcomed is not supported by the scientific understanding of how genetic variation arises, including within an evolutionary context.
In reality, the science supports those who express reservations about the safety of unintended mutations resulting from the gene editing process in plants.
Another vital consideration in this discussion is that risk increases with scale. Deliberately inducing random mutations throughout the genome of an organism through genetic engineering techniques (including gene editing) in a plant to be released at large scale, potentially in many global locations, is very different in scale (and therefore carries heightened risk) from a mutation arising in a single organism in nature and then being selected for fitness over evolutionary time.
For a scientific discussion of this issue, see this article and the scientific paper it references.
In the case of gene variations or mutations arising in nature or through natural crop breeding, the demands of adaptation to environmental conditions, as well as the experience and knowledge of the characteristics of a crop plant on the part of human plant breeders, will select for only those variations or mutations that enhance crop performance and are evolutionary for the plant – and that do not lead to the production of toxins or allergens that could harm human or animal consumers.
Even a single mutation in a single gene could make a crop plant toxic or allergenic. This possibility is amplified enormously given that multiple gene functions will be altered by the gene editing process and that plants are expected be modified by multiple gene editing events, with each bringing its own unintended spectrum of mutations and alterations in gene function. Hence caution, extensive testing, and robust regulation of gene-edited crops and animals are necessary.
4. Myth: Research in animals shows that many new variants (that were not in parental animals) occur in natural reproduction and many of these are in genes. Therefore off-target mutations that occur in gene editing are not a concern.
Facts in brief: Genetic variation (including in genes) through natural reproduction has evolved to bring about adaptation to environmental conditions. But this is different from the random genome-wide mutations that occur during laboratory genetic modification procedures (including gene editing), which are not governed by evolutionary forces or driven by the need to adapt to environmental conditions. These mutations, as well as the epigenetic changes that can arise from genetic modification procedures such as gene editing, can predispose animals to disease. Long-term and transgenerational research is needed to check that gene-edited animals do not possess them.
Facts in detail: Genetic variation (including in genes) does indeed arise through natural reproduction. This process has evolved to bring about adaptation to environmental conditions. However, such natural genetic variation is different from the random genome-wide mutations that occur during laboratory genetic modification procedures (including gene editing), which are not governed by evolutionary forces and are therefore not driven by the need to adapt to environmental conditions. Genetic modification procedures are governed by the necessarily narrow and restricted understanding and commercial imperatives of the genetic engineer.
We should also bear in mind that genetic variation can result in predisposition to disease, as revealed by thousands of genome-wide association studies. Genetic variation resulting from natural reproduction, which is governed by evolutionary forces, will select against disease predisposition variants over long timescales. The genome-wide mutations arising from gene editing procedures are not subject to such evolutionary pressures and therefore could give rise to a spectrum of gene variants that would predispose gene-edited animals to unanticipated diseases.
In addition, epigenetic changes inevitably take place during human or animal embryo manipulation. This includes during the procedures used to generate transgenic or gene-edited animals. These epigenetic changes can result in alterations in gene expression patterns, which can predispose the animal to disease. Furthermore, such epigenetic changes can be inherited by future generations, meaning that they can also suffer.
5. Myth: Gene-edited pigs have been engineered to contain a single gene knockout (gene disruption) aimed at preventing infection by a virus causing Porcine Reproductive and Respiratory Syndrome (PRRS). This is a boon to animal welfare that should be welcomed and facilitated by removing onerous regulations.
Facts in brief: Given that new strains of any given virus are generated at a rapid rate, it is likely that variants of the PRRS virus (PRRSV) will quickly be selected for that bypass the single-gene block that has been engineered into the pigs. Also, regulations must require research to be carried out pre-commercialisation to ensure that the genetic engineering changes made to the genome won’t affect the animals’ health and wellbeing. Overall, the solution to the problem of livestock diseases is not to genetically engineer animals so that we can continue to raise them in inhumane conditions, but to improve the conditions in which we rear them.
Facts in detail: Given that new strains of any given virus are generated at a very rapid rate, it is likely that variants of the PRRS virus (PRRSV) will relatively quickly be selected for that bypass the single-gene block that has been engineered into the pigs. The speed of emergence of new strains of SARS-CoV-2 during the COVID pandemic gives a clue as to just how soon this will happen.
To date, four independent but functionally related PRRSV receptors have been reported, through which the PRRSV infects cells. The principal receptor is CD163, the protein that has been knocked out in the gene-edited pigs. There are already many different variants of PRRSV in existence and new variants are appearing all the time through cycles of infection. Therefore in the absence of CD163, it is highly likely that a new PRRSV strain will rapidly be selected for that infects through a CD163-independent pathway, possibly through the other three known receptors.
So there is good reason for concern that the gene-edited pigs will only prove resistant to infection for a relatively short time span.
Furthermore, the long-term and intergenerational health consequences of animals lacking CD163 function needs to be investigated. CD163 is known to scavenge waste products (haptoglobin-haemoglobin complex) resulting from red blood cell breakup, which takes place at relatively low levels under normal physiological conditions but is markedly enhanced in various disease conditions, which the animals may encounter.
Robust regulation should require that such research is carried out before gene-edited animals are commercialised.
Overall, however, the gene editing/genetic modification approach to animal diseases is misguided, since it is recognised that the overcrowded and stressful conditions in industrial pig and poultry farming are breeding grounds for pathogens for these animals.
In the case of the avian flu virus that affects poultry, the relationship of intensive farming to this disease is well documented. For instance, a Compassion in World Farming report (references to the scientific literature are included) states, “Wild birds have taken much of the blame for the spread of H5N1 across the world but this report… unveils evidence showing that the development of highly pathogenic strains of bird flu lies at the door of factory farming. The past two decades have seen a complete transformation in the poultry industry with a 300% increase in production across the world. This increase is, by and large, thanks to reliance on intensive farms and a transnational production system. Poultry production is now a global affair. An intensive poultry farm provides the optimum conditions for viral mutation and transmission – thousands of birds crowded together in a closed, warm and dusty environment is highly conducive to the transmission of a contagious disease.”
The report adds, “Whilst not denying that wild birds, backyard and free range farms naturally play a role in the spread of the current epidemic, continuing to focus containment measures on poultry is misguided and ignores the overwhelming evidence pointing towards other often more important routes of spread – namely the global trade in live poultry and poultry products. The spread of H5N1 from China to Europe, Africa and the Middle East correlates with major road and rail routes rather than bird migratory routes or seasons.”
So the solution to the problem of livestock diseases is not to genetically engineer the animals so that we can continue to raise them under inhumane conditions, but to change the system of husbandry in which we rear them.
6. Myth: Whole genome sequencing, which can accurately identify the full spectrum of unintended mutations at both off-target and on-target edit sites, including inadvertent insertion of foreign DNA, is too impractical and onerous to undertake for gene-edited plants.
Facts in brief: Multiple reference genomes derived from whole genome sequencing of major crop plants are now in the public domain. Whole genome sequencing has yielded important information about the unintended effects of gene editing on the genome. In a study on gene-edited rice using the CRISPR/Cas gene editing tool, whole genome sequencing was used to investigate unintended mutations arising from different aspects of the gene editing procedure. The researchers found that the gene editing procedure (taken as a whole, namely, tissue culture and Agrobacterium-mediated cell transformation) resulted in several times more unintended mutations than were found in rice propagated through natural pollination (see Fig 2b in this study).
7. Myth: Plant genomes contain extensive regions of “repeat sequences”, so it is difficult to obtain an accurate picture of the whole genome. This is because standard genome sequencing technology produces short lengths of DNA sequence reads that then need to be linked up accurately, using computational tools. This can be difficult, given that stretches of repeat sequences can be present at different chromosomal locations.
Facts in brief: It is feasible to compile an accurate whole genome sequence, despite long stretches of repeat sequences, as evidenced by the huge body of human and animal, as well as plant, whole genome sequences available in the public domain. Any remaining difficulties of accurately aligning whole genomes through areas of long repeat sequences are solved by the advent of long-read DNA sequencing, which provides unequivocal placement of any long stretches of repeat sequences.
Facts in detail: Long stretches of repeat sequences are not unique to plants but are present in human and animal genomes. It is quite feasible to compile an accurate whole genome sequence, despite long stretches of repeat sequences, as evidenced by the huge body of human and animal, as well as plant, whole genome sequences available in the public domain.
Any remaining difficulties of accurately aligning whole genomes through areas of long repeat sequences are solved by the advent of long-read DNA sequencing. As the name implies, long-read DNA sequencing provides continuous sequence reads of up to over 1.5 million DNA base units and would provide unequivocal placement of any long stretches of repeat sequences. Several companies offer a long-read genome sequencing service, making this technology readily available. Therefore the claim that whole genome sequencing of plants is impractical as an analytical tool to ascertain a more comprehensive molecular genetic characterisation of a gene-edited plant is not supported by the state of the science.
8. Myth: Molecular compositional profiling methods (“omics” methods: gene expression-profiling transcriptomics, protein-profiling proteomics, and small biochemical molecule-profiling metabolomics), which some are asking to be carried out to ascertain if a gene-edited plant is equivalent to its non-gene-edited parent except for the intended gene editing outcome, would not be useful. They would just generate a large body of data that would be impossible to make sense of.
Facts in brief: Molecular profiling multi-omics methods are now used by thousands of research groups around the world to gain a more comprehensive and deeper insight into the function of biological organisms and systems. The computational tools (bioinformatics and statistical methods) used to analyse the data from these molecular profiling methods are well established and allow scientists to make good sense of the data generated and its health and disease implications.
Facts in detail: Molecular profiling multi-omics methods are now used by thousands of research groups around the world to gain a more comprehensive and deeper insight into the function of biological organisms and systems in both healthy and disease states. The computational tools (bioinformatics and statistical methods) used to analyse the data from these molecular profiling methods are well established and allow scientists to make good sense of the data generated and its health and disease implications. These computational tools are constantly evolving, as is the ability of scientists to interpret the results in a meaningful way.
If the implication is that the field of plant molecular biology has not advanced sufficiently to allow meaningful interpretation of molecular profiling datasets, this suggests that until the plant molecular biology field is in a position to make sense of molecular profiling datasets, it should not be venturing down a route in which it manipulates the genome of plants with unintended genetic and biochemical outcomes.
But of course, this is not the case, as the scientific literature has many examples of researchers doing exactly that.
Indeed, Dr Antoniou’s own research group published a paper in 2016 where they used a multi-omics approach (proteomics and metabolomics) to demonstrate that a glyphosate-tolerant GM maize was not substantially equivalent to its non-GM relative. The large-scale protein and metabolite alterations that were detected were unintended outcomes of the GM transformation process.
Although the health implications of these changes are unknown, the findings graphically illustrate how a GM transformation process (of which gene editing is an example) can dramatically alter the composition of a plant, with potential downstream health consequences for the consumer.
In light of the above, we challenge UK-based researchers to conduct multi-omics studies of established lines of gene-edited plants compared to their non-gene-edited parents, where the plants have been grown at the same time and in the same location, and, based on the data obtained, make the claim that the gene-edited plant is equivalent to the non-gene-edited parent, apart from the intended change.
Gene-edited products should remain regulated
The government has always said it would be led by the science in its policymaking. In the context of the new bill, we have to ask ourselves what science the government has been led by.
It is evident from the science presented above that the government is not paying attention to the objective science that graphically illustrates that gene editing is neither precise nor predictable. The technology consists of treating the organism with the gene editing tool and hoping that something useful emerges from it. Without detailed characterisation of the outcomes at both a gross and deeper molecular level (which should be made publicly available), no one can claim that a gene-edited crop or animal is as safe as one that is naturally bred.
For the reasons listed above and many others, we advocate that gene-edited products remain regulated under current GMO legislation and are evaluated under a process- and a product-based regulatory system. Only such a regulatory approach is true to the science that underpins gene editing technology and can foster the public trust sought by GMO developers.