Research and analysis

ACRE advice concerning Defra’s consultation on the regulation of genetic technologies

Published 29 September 2021

ACRE has provided below detailed advice concerning key aspects of Defra’s consultation on the regulation of genetic technologies, and was assisted in this by two co-opted members who have provided invaluable scientific and technical expertise in the areas of plant genetics and food safety (Prof Huw Jones, IBERS, Aberystwyth and member of ACNFP) and in animal genetics (Dr Huw Jones, member of FAnGR).

The purpose of this advice was to assist Defra in its consideration of the issues raised by the consultation, including the scientific aspects of certain safety considerations. Defra have published the full Government response to the consultation.

Scope

ACRE notes that the scope of the Defra consultation includes organisms produced by gene-editing (GE) and other genetic technologies. ACRE understands this to encompass a wide range of methods, including but not limited to, genome editing, epigenome editing and transcript editing. Furthermore, our increasing knowledge of the proportion of genetic material which does not contain genes (i.e., non-coding DNA) will in the future extend the range of targets for editing.

Use of the term GE in the discussion below should be considered to include all these alternative approaches. These technologies will continue to evolve and increase in efficiency and accuracy, and it will remain necessary for regulators to assess whether or not genetic changes introduced by these technologies could have arisen naturally and/or through traditional breeding.

Traditional breeding and GE: scientific context

ACRE has previously considered many of the issues which are raised by this consultation (mainly in the context of plant breeding but also considering applications in animals and micro-organisms). ^{[footnote 1][footnote 2]} Most relevant are ACRE’s conclusions on a range of site-directed (or targeted) mutagenesis technologies such as those now commonly referred to as gene-editing.

ACRE’s view is that an organism produced by gene-editing or another genetic technology would not pose a greater safety risk than a traditionally bred or naturally occurring version of that organism, as a result of how it was produced.

Underpinning this view is the key principle that it is the final characteristics of an organism which determine whether it presents any safety risks, regardless of the method used to produce that organism. Furthermore, ACRE’s view is that where GE introduces genetic alterations and combinations that are of the type that are selected for in traditional breeding, the environmental release of these organisms should not be regulated in the same way as the environmental release of genetically modified organisms (GMOs).

Other expert groups, governments and organisations have commented on the importance of developing scientifically credible regulatory options. They and ACRE have cited an evidence base which demonstrates that the genetic material of organisms is not fixed and immutable, but is frequently subject to high levels of natural variation and selection, occurring in real time within species and individuals, and sometimes in response to environmental stimuli.

Traditional breeding relies on the dynamic nature of genetic material and makes use of a range of genetic changes, including major structural variations which occur naturally within species and their close relatives ^{[footnote 3][footnote 4]} that are used in breeding programmes. Examples of these are exchanges, rearrangements and duplications of extensive stretches of DNA containing potentially hundreds of genes that have taken place during breeding for disease resistance in tomato^{[footnote 5]} and wheat^{[footnote 6]} and during the domestication of peach varieties^{[footnote 7]}.

Genetic change that is exploited in breeding programmes can also result from the activity of transposable elements (TE) found naturally in plant genetic material. For example, in maize, the selection of varieties which allowed flowering under long-day conditions was responsible for helping maize to be grown over a broad range of latitudes globally. This trait was found to be the result of TE insertion^{[footnote 8]}. In rice, TE activity was shown to be responsible for altering the expression of a gene which gave plants tolerance to aluminium^{[footnote 9]}. Two TEs independently inserted into the promoter region of the orange Ruby gene, resulting in convergent evolution of the blood orange trait^{[footnote 10]}. And TE insertion was responsible for the oval shape typical of the Roma tomato variety^{[footnote 11]}.

Safety aspects

ACRE’s discussions have included the issue of unintended or ‘off-target’ effects which are known to sometimes occur during gene-editing, and which have been raised as a potential safety concern by some consultation respondents. Generally, two different types of unintended/off-target effects sometimes occur during gene-editing. The first type is the editing of nucleotide sequences other than those at the intended target site. The second is the unintentional introduction of DNA from a different species to the one undergoing gene editing. Whilst raised as potential safety issues, these occurrences are not necessarily harmful and can usually be removed by segregation in subsequent breeding steps. A useful review of the regulatory approaches to these two issues has been published recently^{[footnote 12]}. It is helpful to consider them separately as there are different issues associated with them:

The phenomenon of unintended introduction of DNA from other species has been observed most notably during the production of gene-edited animals. After DNA strands are cut by the gene editing ‘machinery’ at a predetermined site, their repair is driven by the cell’s own mechanisms. If left to chance, the repair process is random, with nucleotides from within the cell or the surrounding media occasionally being incorporated in the repair. Since bovine or caprine serum is frequently used in culture media for mammalian embryos, this has led to reports of bovine or caprine sequences being incorporated into target sites^{[footnote 13]}.

This effect can be mitigated to some extent by using a repair template, but some unintended insertions although uncommon have been reported in the past (below), which may suggest that some level of genome screening at target sites prior to establishing pregnancies may be useful if a high level of insertion accuracy is required. Avoiding the need to use bovine and caprine serum in culture media would also be expected to reduce the level of risk for target site ‘contamination’, and research is on-going in that area.

In 2020, a paper reported that the genomes of two polled (hornless) Holstein bulls produced in 2015 were found to include unintended DNA^{[footnote 14]}. While one allele in the genome of the bulls included the target insertion as expected (leading to the hornless characteristic being present), the second allele was found to also include a fragment of DNA from the carrier plasmid^{[footnote 15]}. As part of the original genome editing project the bulls’ genome had been screened only for off-target insertions of the introduced ‘polled’ allele, and none had been found. However, the authors had not screened for DNA from the carrier plasmid. If it had been included in the initial screening it would likely have allowed its detection and appropriate action to be taken.

Although retained inadvertently, it is important to note that the presence of the plasmid DNA had no observable effects on the bulls’ natural development (or expression of the polled characteristic) or that of the progeny of one of the bulls which were produced and screened in a later study^{[footnote 16]}. Further, it is important to note that the use of a carrier plasmid was a requirement for the specific editing technology used in the study. A carrier plasmid would not be required when using CRISPR-Cas9, which is the technology now most commonly used.

We now consider the issue of off-target effects: nucleotide changes that occur at places other than those intended. Of the three editing methods developed to date, CRISPR-Cas9 offers the highest level of target specificity by typically using a 20-nucleotide guide RNA that is complementary to desired target sites. Although facilitating a very high level of specificity, the DNA sequences being targeted may not be unique in the genome of interest and the potential to introduce off-target effects has been highlighted as a potential concern in some studies on plants^{[footnote 17]} and animals, even when using CRISPR-Cas9.

Some sequences will be highly repeated within genomes, and as a result selecting appropriate targets as well as an appropriate editing technology will both play important roles in ensuring high specificity^{[footnote 18]}. Many recent genome editing studies on animals have reported no incidences of off-target insertions when using CRISPR-Cas9^{[footnote 19][footnote 20]}. Where off-target insertions have been identified the number of incidences has tended to be low and they have occurred at sites that could be predicted using available software^{[footnote 21]} which could be monitored using low level screening if required^{[footnote 22]}. This may be particularly relevant during the deployment of novel technologies such those utilising relaxed PAM sites^{[footnote 23]}.

It is also important to note that research is still on-going into new methods and approaches that could further reduce the risk of off target insertions and unintended insertions and some, such as modifying the delivery system^{[footnote 17]}, use of single stranded DNA repair template^{[footnote 23]}, and prime editing approach have already shown promise^{[footnote 24]}.

In conclusion ACRE’s view is that there is no scientific reason to suppose that off-target sequence alterations in GE organisms will result in greater safety concerns than those which result from other forms of mutagenesis. Indeed off-target effects introduced by GE methods are very significantly rarer than those produced during the course of conventional methods of plant breeding^{[footnote 17]}. Similarly, there is good evidence to suggest that GE-induced mutagenesis in animals is no greater than the background rate of de novo mutation^{[footnote 25]}. However, from a regulatory perspective, it is clear that most countries who have reviewed their assessment procedures for GE organisms, have also acknowledged the importance of providing a degree of assurance either that no exogenous DNA has been inserted, and/or that no unacceptable risks are presented by the new trait.

A further issue that has been raised by consultation respondents is the possibility that GE organisms, and/or the genetic material of GE organisms, might become preferentially established in the environment and cause harm. ACRE notes that in plant breeding over millenia, humans have tended to select against the ability to persist in the environment so that farmed crops are now generally unable to do so. Whereas in animal breeding a combination of contained management systems and high levels of domestication mean that very few ‘wild’ crossable populations remain especially in the UK. In our view this makes the transfer of genetic material to wild animal populations very unlikely. Furthermore, GE relies on the creation of the same types of genetic variation that are selected for in traditional breeding.

ACRE’s view therefore is that there is no scientific reason that the application of GE technology for crop and livestock improvement would lead to a greater likelihood of persistence of these organisms in the wider environment, which is rarely observed for traditionally bred organisms.

For some aquaculture species ACRE note that the difference between wild and farmed species is less, as their domestication is a more recent event compared to other livestock species. However, the development of genetic sterility through genome editing is showing promise and could provide an added safeguard against accidental crossing^{[footnote 26]}.

ACRE has also considered how this issue might apply to the potential environmental release of GE micro-organisms and notes that different parameters are likely to apply under this scenario; for example, given the high levels of horizontal gene transfer that are known to occur between micro-organisms in natural environments, gene flow is more likely to be a realistic possibility. This should be an important consideration in the context of developing any new regulatory procedures for micro-organisms produced using gene editing and/or other genetic technologies.

ACRE believes that the basic principle of whether an organism could be produced by traditional breeding can provide an initial assessment of whether that organism should be classed as a GMO. This is because of the long history of safe use of traditional breeding methods in plants and animals and a modern understanding of genome dynamics (see above). However, a scientifically credible regulatory approach for GE organisms, which relies on a comparison with what is possible using traditional breeding methods, must account for the similarities and differences between plants, animals and micro-organisms, and the methods that have been developed to exploit genetic variation within them.

For example, micro-organisms (non-GMO) used in industrial processes and in agriculture have traditionally been first identified in strain collections, and often then further developed by a process of selection and screening for desired traits. A modern example of this process is the work of AgBiome plc whose significant microbial collection was the basis for the sequencing of 80,000 microbial genomes, leading to the identification of numerous strains with useful properties such as the production of crop protection substances.

The genetic basis for microbial traits is often relatively easy to identify, compared to changes produced in plants and animals because of the small sizes of microbial genomes and the relative ease with which many micro-organisms can be genetically engineered. A range of techniques, that are broadly equivalent to the new GE technologies in terms of what they can be used for, have been available for many years for bacteria, although their utility varies considerably between species and even between different strains of the same species.

The various CRISPR-Cas systems are all microbiological in origin and have been shown to be effective in many cases for GE in heterologous species. Thus, it is increasingly likely micro-organisms with desirable traits will be developed for industrial, agricultural, or human health uses. Using targeted mutagenesis approaches such as gene-editing is likely to prove more effective than the earlier screening/selection methods referred to above.

Many of these cases will inevitably go on to be released into the environment either as a deliberate consequence of their application, or an inevitable side-effect of it (for example, excretion of bacteria in a probiotic drink, or of bacteria which have been engineered to work as a vaccine). There are still gaps in our knowledge of microbial ecology, in particular in the details of the mechanisms and extent of horizontal gene transfer in the natural environment, and what determines the likelihood of a particular strain evolving into a novel pathogen. It may therefore be important to consider micro-organisms separately to plants and animals when developing new regulatory procedures for GE organisms.

1. ACRE advice: New techniques used in plant breeding ↩

2. Why a modern understanding of genomes demonstrates the need for a new regulatory system for GMOs ↩

3. Coletta, R., et al. How the pan-genome is changing crop genomics and improvement. Genome Biol 22, 3 (2021). ↩

4. Weber, N., et al. Crop genome plasticity and its relevance to food and feed safety of genetically engineered breeding stacks Plant Physiol 160:1842-1853 (2012) ↩

5. Young ND & Tanksley SD. RFLP analysis of the size of chromosomal segments retained around the Tm-2 locus of tomato during backcross breeding. Theor Appl Genet 77: 353–359 (1989) ↩

6. Jones, SS et al. The development of disease resistance in wheat. Annu Rev Phytopathol 33: 429–443 (1995) ↩

7. Guan, G et al. Genome structure variation analyses of peach reveal population dynamics and a 1.67 Mb causal inversion for fruit shape. Genome Biology 22:13 (2021) ↩

8. Huang C, et al. ZmCCT9 enhances maize adaptation to higher latitudes. Proc Natl Acad Sci U S A. 2018; 115: E334–41 ↩

9. Yokosho K, et al. Retrotransposon-mediated aluminum tolerance through enhanced expression of the citrate transporter OsFRDL4. Plant Physiol. 2016; 172:2327–36 ↩

10. Butelli E, et al. Retrotransposons control fruit-specific, cold-dependent accumulation of anthocyanins in blood oranges. Plant Cell. 2012; 24:1242–55 ↩

11. Xiao H, et al. A retrotransposon-mediated gene duplication underlies morphological variation of tomato fruit. Science. 2008; 319:1527–30 ↩

12. Lema et al. Regulatory Assessment of Off-Target Changes and Spurious DNA Insertions in Gene-Edited Organisms for Agri-Food Use. Journal of Regulatory Science 9 (2021) 1–15 ↩

13. Ono et al., (2019) Communications Biology 2, article 57 ↩

14. Carlson et al, (2016) Nature Biotechnology, 34, p 479–481 ↩

15. Norris et al, (2020) Nature Biotechnology, 38, p 163–164 ↩

16. Young et al., (2019) Nature Biotechnology, 38, pages 225–232 ↩

17. Graham et al., (2020) Plant Genome Editing and the Relevance of Off-Target Changes. Plant Physiol. 183:1453-1471 ↩ ↩² ↩³

18. Burkard et al, (2019) Plos Pathogens. 23 (DOI:10.1371/journal.ppat.1006206) ↩

19. Henning et al., (2020) Scientific Reports, 10:22309 ↩

20. Zhou et al, (2019) Frontiers in Genetics. vol. 10, article 215 ↩

21. Dong et al., (2019) Genes, Genomes, Genetics. vol. 9, p 3645 - 3651 ↩

22. McFarlane et al., (2019) Frontiers in sustainable food systems. vol. 3, article 106 ↩

23. Ren, Q. et al., (2021). PAM-less plant genome editing using a CRISPR–SpRY toolbox. Nature Plants 7, 25–33 ↩ ↩²

24. Anzalone et al., (2019) Nature, vol. 576, p 149–157 ↩

25. Ayabe et al (2019) J Reprod Dev. 65 (1): 1–5 ↩

26. Güralp et al 2020; Scientific Reports 10:18042 ↩