Confirms process-based regulation is necessary
A new study by Chinese researchers has been hailed in a lobbying post as showing that plant gene editing via CRISPR is precise, predictable, and controllable.
In fact, however, it shows the opposite – that the CRISPR process, taken as a whole, causes large numbers of off-target mutations.
The clear conclusion for those of us without an interest in promoting GMOs is that process-based regulation of CRISPR’d plants, which takes account of the inherent risks and shortcomings of the GM procedures used to make them, is needed.
However, the lobbyist uses the findings of the new study to argue that GMO regulations should be changed to product-based: That is, they should only consider the intended product and ignore the process by which the plant was generated.
Such lobbyists aim to dismantle the EU's process-based GMO regulation, which acknowledges the inherent uncertainties of the GM process, and replace it with product-based regulation such as that of North America, where only the intended resulting product is assessed for risk.
The lobbyist's reasoning
The lobbyist writes that the study shows that tissue culture and "spontaneous" mutations are the source of most nucleotide (base units of DNA) variations – and not the genome editing process itself.
The lobbying post is written by the CEO and co-founder of a plant biotechnology company. He points out that the authors of the new study found that the safest breeding approach, harvesting seeds from non-GM parent lines, introduces around 30 to 50 spontaneous mutations into the next generation in rice. The authors also observed around 200 tissue culture-introduced somaclonal variations per rice plant. Somaclonal variation is the genetic variation among the progeny of a plant regenerated from that plant's cells grown in tissue culture.
The lobbying post adds that the authors found, "in stunning contrast to many conventional breeding technologies", that editing of the target genes was "exceptionally precise" because they could not find any off-target mutations in 47 out of 49 rice plants edited by certain CRISPR tools.
Ramming the lobbying point home with capitalised shouty words, the lobbyist writes, "This study shows why it is WRONG to call older breeding technologies safer than genome editing”. He states that it should remind regulators to “regulate genome-edited products, not genome editing itself”. He concludes, “This work would be a VALUABLE REFERENCE for regulatory agencies and scientific community."
Is the lobbyist right? As is always advisable, we looked at the study itself to see what the results tell us. We were gobsmacked that this study is being used – including by the authors themselves – as showing the safety, precision, and predictability of CRISPR.
The study authors analysed gene-edited plants engineered with two different CRISPR nucleases for off-target and on-target effects.
In a commendable "first" for these types of studies, the authors separated out the different elements of the gene editing process and used whole genome sequencing (WGS) to count the number of off-target and on-target effects that each of these elements induced in the plants.
The different elements of the gene editing process that were examined were:
* Tissue culture only
* Agrobacterium infection (used to carry the CRISPR gene-editing tool into the plant cells) only, and
* Introducing the different CRISPR gene editing tools (Cas9 backbone and Cpf1 backbone) via Agrobacterium infection, but without including the guide RNA molecules that direct the editing tools to the desired site to induce a double-strand break in the DNA.
The effects looked for were single nucleotide variations (SNVs), small and large insertions/deletions (indels), and genome rearrangements.
In addition, as a baseline, the authors analysed non-GM wild type plants from three consecutive generations for "spontaneous" mutations.
The researchers found that seed saved from wild type non-GM rice plants had 30 to 50 spontaneous mutations per plant.
They found few off-target effects from the CRISPR editing tools – only 2 out of 49 plants had them.
However – and this is the crucial aspect – they found a vast number of mutations from the tissue culture (200 per rice plant). In the authors' words, "Our results clearly show that most mutations in edited plants are created by the tissue culture process, which causes approximately 102 to 148 single nucleotide variations (SNVs) and approximately 32 to 83 insertions/deletions (indels) per plant."
Tissue culture is an obligatory part of making gene-edited (or older-style transgenic) plants.
The results also showed that "Agrobacterium infection is mutagenic with a preference for introducing indels". The Agrobacterium infection increased the number of mutations over and above the number caused by tissue culture.
The authors concluded that the off-target mutations caused by two of the three CRISPR editing tools were "largely negligible when compared to spontaneous mutations or mutations caused by tissue culture and Agrobacterium infection in edited plants. The resulting knowledge is likely to serve as an important reference for plant researchers and regulatory agencies."
What this means
Based on these findings, it doesn't matter if the CRISPR tools do not introduce many mutations. The end result will still be lots of mutations from other aspects of the CRISPR genetic manipulation process. When making CRISPR'd plants, tissue culture is always used and Agrobacterium infection is commonly used to deliver the editing tool – the latter element adding to the spectrum of mutations introduced by the tissue culture.
Inevitably therefore, CRISPR'd plants are going to end up with numerous mutations.
This is not to suggest that the number of mutations is all-important – it's what the mutations do that matters. That can only be established by further experiments – "omics" analyses to establish compositional changes and then long-term animal feeding studies to establish biological effects. These could include unexpected toxicity or allergenicity.
Dr Janet Cotter of the consultancy Logos Environmental commented on the new study findings, "The researchers found no off-target effects in the majority, but not all, cases. That is why regulation is needed – to ensure there are no potentially harmful effects from the genome-editing process."
Dr Cotter explained:
1) The paper finds that good design of the CRISPR tool can minimise, but not eliminate, off-target effects. The authors state, "We only found true off-target mutations in two T0 lines expressing Cas9 protein with Cas9-J-sgRNA01."
2) Off target effects are not the only genetic errors that CRISPR can cause. There may also be unexpected "on-target" effects – for example, deletions and re-arrangements of the host DNA, or interference with gene regulation.
3) Genome-edited animals may be more susceptible to genetic errors created by genome-editing, but genetic errors have also been documented in many plants (e.g. Zhu et al., 2017; Wolt et al. 2016). Wolt et al. has a table of off-target effects in plants.
Thus far, this is a well carried out study that breaks new ground in terms of separating the gene editing process into its different elements and working out which element caused how many of certain types of mutations.
Sadly, however, the authors fall into the same trap that bedevils so many researchers in this highly politicised field. Seemingly aware that their results are truly alarming and back the case of those who advise caution over gene-editing techniques, the authors must have asked themselves how they were going to spin the results so that they don't sound so bad.
And this is where they step over the boundary that should separate science from hype and objective investigation from advocacy. They conclude, in line with the GMO industry and the biotech industry CEO lobbyist, "Our data support a recent call to 'Regulate genome-edited products, not genome editing itself'."
This bizarre statement is entirely at odds with the authors' own findings. Their results show that the CRISPR process, taken as a whole, induces hundreds of procedurally induced mutations. They show that the process is inherently problematic. Thus regulation based only on what the end product is intended to be is obviously diametrically opposed to what the science – including the authors' own results – is telling us.
So while the biotech industry lobbyist and the study authors are right in saying that the new paper will be a valuable reference for regulatory agencies, they are wrong in claiming that it backs the case for product-only-based regulation. On the contrary, the paper provides the clearest evidence for process-based regulation, such as the EU currently has in place.
1. Tang X et al (2018). A large-scale whole-genome sequencing analysis reveals highly specific genome editing by both Cas9 and Cpf1 (Cas12a) nucleases in rice. Genome Biology 19:84. https://genomebiology.biomedcentral.com/articles/10.1186/s13059-018-1458-5
2. Zhu C, Bortesi L, Baysal C, Twyman RM, Fischer R, Capell T, Schillberg S and Christou P (2017). Characteristics of genome editing mutations in cereal crops. Trends in Plant Science 22:38–52.
3. Wolt JD, Wang K, Sashital D and Lawrence-Dill CJ (2016). Achieving plant CRISPR targeting that limits off-target effects. The Plant Genome 9: doi: 10.3835/plantgenome2016.05.0047
Cotter J and Perls D (2018). Gene-edited organisms in agriculture: risks and unexpected consequences. Friends of the Earth USA. http://foe.org/wp-content/uploads/2018/09/FOE_GenomeEditingAgReport_final.pdf
Report: Claire Robinson