Content Title

(page/chapter no/table/figure)

Comments:

1. Background

The tone and content of this section would benefit greatly from a higher degree of neutrality and factual delivery.

The general statement that all new technologies have dual-use potential is not correct, nor appropriate in its generality. However, it is correct that GEd technologies have dual-use potential and therefore involve both safety & security issues.

4. Application of Genome Editing Technologies in the Indian Context – page 10

The statement “biotechnology offers safe and sustainable solutions to many environmental challenges” may be the hope of those engaged in regulating, researching or funding such biotechnology approaches – it is though merely an opinion statement. It either needs to be backed up by explicit research data of how particular environmental challenges and problems are being safely and sustainably solved (on a sufficiently long time scale), or simply be removed.

It is our opinion that this and similar statements have no place in a regulation text and should be removed, as they are giving a wrong sense of certainty and are completely irrelevant for the context of risk assessment.

Paragraph 3 page 10

The statement that Genome Editing Technology is offering to increase crop yield and productivity (optimally) by protecting them from various biotic and abiotic stresses and various other traits is again a statement of unsubstantiated and unsupportable opinion. Resistance traits as well as traits related to stress tolerance are complex multi-gene traits embedded in a gene regulation mechanism overlapping with other stress-related pathways and morphological characteristics. This statement is giving the wrong picture of what the biological and practical realities are and is unscientific in itself – it should thus be removed.

Furthermore, such statements fail to recognise that in order to assure sustainable and safe food production, a systems approach is required. Without resilient food production systems that minimise stress and maximise biological support networks, any food production system will remain highly vulnerable, in particular in times of climate change.

Page 11

This page unfortunately has the same listings as were produced for genetic engineering technologies some 20 years ago. The offering of “new promises” was made then with very little results.

Genome Editing should neither be regarded nor be treated as a wonder-technology. Like previous genetic engineering S&T, this technology is limited by the biological realities of living organisms. Altering one gene may – and usually does – impact various traits and regulatory pathways within an organism.

Scientifically, it is unsupportable to portray wishful thinking as “promises”, and in particular to do so without acknowledging the risks and uncertainties going hand in hand with the technology.

Whilst research for human gene therapy is a crucial area of investigation, genome editing also holds high risks, as research has shown in animal and cell studies.

5. General Considerations for Risk Analysis of GEd Organisms and Products Derived Thereof

– page 12

Whilst GEd organisms may differ from GE organisms in some respects, the statement that GEd organims “differ” from GE organisms in “many” respects is misleading – in particular in the context of risks.

First of all, Genome Editing is a new technology with limited experience and no established record of safe use in any species or for any purpose. The same is true for GE. On the contrary, there has been very little safety research into GEd organisms and the impacts of Genome Editing on the genome and the phenotypical changes of GEd organisms. Hence there is little and insufficient data for building any claims of safety.

Accidental process-induced introduction of superfluous and/or foreign DNA is found both in GE organisms as well as in GEd organisms.

For example: Genome edited cattle – modified to have no horns – were found to have about 4000 kbp of superfluous DNA present originating from the bacterial plasmid vector, including two antibiotic resistance genes. This was not established by the producer of the cattle, but by researchers in the regulatory agency FDA. Norris et al. 2020, Nature Biotechnology 38, pages 163–164(2020), doi: 10.1038/s41587-019-0394-6

Other cases of unintended integration of non-host DNA include: mice (e.g. Ono et al., 2015 and Ono et al. 2019); Plants (e.g. Jacobs et al., 2015, Li et al., 2015); fish (Gutierrez-Triana et al., 2018); fruitflies  (Drosophila melanogaster), nematodes (C. elegans), yeast and other fungi (eg Aspergillus); and planktonic crustaceans (Daphnia magna).

Whilst the insertion of vector DNA into DNA double-strand breaks is in itself a matter of concern, it is of perhaps even greater concern, that any other trace DNA present in the culture medium may be inserted into the host DNA. Ono et al., 2019, for example identified the presence of goat DNA and bovine DNA in the genome of the genome edited mice. This depended on whether goat or foetal calf serum had been used as a culture medium in the experiments. In fact, even retrotransposons had been transferred. In this context it becomes obvious that genome editing may unintentionally become a mechanism for horizontal gene transfer of not only foreign DNA but pathogens alike.

In a recent paper on CRISPR-cas9-mediated HDR  (Skryabin et al. 2020; Science Advances doi: 10.1126/sciadv.aax2941) authors reported not just about the unintended insertion phenomena (both for NHEJ & HDR) but also the difficulty of detecting these. The authors state in their abstract: “Nevertheless, the rapidly evolving technique still contains pitfalls. During the generation of six different conditional knockout mouse models, we discovered that frequently (sometimes solely) homology-directed repair and/or non-homologous end joining mechanisms caused multiple unwanted head-to-tail insertions of donor DNA templates. Disturbingly, conventionally applied PCR analysis, in most cases, failed to identify these multiple integration events, which led to a high rate of falsely claimed precisely edited alleles. We caution that comprehensive analysis of modified alleles is essential and offer practical solutions to correctly identify precisely edited chromosomes.”

5. General Considerations for Risk Analysis of GEd Organisms and Products Derived Thereof

– page 12

The Draft Regulation states further: “Genome editing is a precise molecular method of mutation leading to deletion or addition or substitution of target base pair(s) in the native genes/ nucleic acid sequences.”

Though genome editing is often referred to as being “precise”, this is a very limited view and is not correct, in particular when attempting to directly or indirectly equate precision with predictability and then safety.

What is it that is regarded as precise, and precise at which level? “Precise” was also the term that was chosen to describe the genetic engineering technologies used in the 1990s. The term was not correct then, nor is it now.

Firstly, GEd organisms as well as GE organisms both hold the risk of unintended process-induced mutations and off-target effects. Here it has to be borne in mind, that the so-called precision of genome editing is merely meant to express a measure of efficiency at the very basic level of nucleotides. Most DNA breaks will occur at the target site whilst a small proportion may occur elsewhere in the genome. These off-target “cuts” or breakages are thought to occur most likely in places with clear homologies to the target site. 

Experiments have shown CRISPR/Cas9 may cut DNA even with 2-3 nucleotide mismatches be­tween the DNA sequence and the guide-RNA, albeit -as mentioned above- with lowered efficiency. There does not seem to be a hard rule as to how many nucleotide mismatches are tolerated by the mechanism, as this also depends on the species, cell type, the actual nuclease variant and the experimental conditions.

Whilst there is an increasing reliance on the use of algorithms to calculate and predict the potential off-target sites, according to the degree of homolo­gy (based on the number and position of mismatch­es), there is also increasing concern about this.

In fact, the sole reliance on algorithms to accurately predict the potential off-target sites or regions for off- or on-target effects has come into question re­peatedly, as only whole genome sequencing, an in­creasingly affordable technology, would be able to pick up some of the mutational effects observed. This does not only refer to extensive mutations de­linked from the actual cutting site (Kosicki et al. 2018, Nature Biotechnology 36 (8):765-+. doi: 10.1038/nbt.4192.), but also to the integration of vector backbone DNA derived from the plasmid used in the original transgene construct, and for ex­ample observed in genome editing experiments with oilseed rape (Braatz et al. 2017, Plant Physiology 174 (2):935-942. doi: 10.1104/pp.17.00426).

Akcakaya et al. (Akcakaya et al. 2018, Nature 561 (7723):416-+. doi: 10.1038/s41586-018-0500-9.) find that many studies re­porting no or few off-target effects (mutations) will have failed to identify actual off-target effects due to the limitations of the “in silico” (i.e. computer modelling) predictions of potential off-target sites.

Secondly, GEd organisms (as GE organisms before) hold the risk of unintended process-induced on-target or near-target effects. Researchers from the Sanger Institute, UK, for example, reported ev­idence of significant on-target mutations, such as large deletions — of up to 9.5 kb — and complex rearrangements around the DNA breakage site (Ko­sicki et Al. 2018 – see above). Additionally, they found mutations (deletions, rearrangements and even insertions) away from the target site, i.e. not physically linked to or running on from it.

Whilst the implications of these specific and complex re­arrangements have not been investigated, such re­arrangements constitute a clear risk, as they can alter gene expression, give rise to further muta­tions during reproduction, as well as disable or alter the sequence of genes at the site of rearrangement.

However, the same situation may also arise for unin­tended off-target site mutations, as the action of CRISPR/Cas9 would work under the same rules for both. Off-target sites though have not yet been investigated for com­plex rearrangements, a fact that needs urgent atten­tion, given that the risks are likely to be the same.

Thirdly, even if the genome editing technology has produced the DNA changes as intended, this does not necessarily mean that the outcome in predictable or safe. Here we are addressing the unintended effects of intended on-target mutations.

For example, in a recent publication in Nature Communications is indicat­ing that intended on-target indel mutations – set by the error-prone NHEJ repair mechanism – may have very unexpected and indeed problematic consequences (Tuladhar et al. 2019, Nature Communications. 10, Article number: 4056. https://doi.org/10.1038/s41467-019-12028-5). Tuladhar investigated the consequences of intend­ed knock-out mutations, in particular of (intended) frameshift mutations induced by indel mutations. The re­searchers looked at the processing of the resulting RNAs, their translation into proteins as well as the impact on gene regulation. They reported: “Here, by examining the mRNA and protein products of CRISPR targeted genes in a cell line panel with presumed gene knockouts, we detect the production of foreign mRNAs or proteins in ~50% of the cell lines. We demonstrate that these aberrant protein products stem from the introduction of INDELs that promote internal ribosomal entry, convert pseudo-mRNAs (alternatively spliced mRNAs with a PTC) into protein encoding molecules, or induce exon skipping by disruption of exon splicing enhancers (ESEs). Our results reveal challenges to manipulating gene expression outcomes using INDEL-based mutagenesis and strategies useful in mitigating their impact on intended genome-editing outcomes.”

The important news here is the gen­eration of new Internal Ribosomal Entry Sites (IRES) leading to the production of truncated proteins and the alteration of pseudo-mRNAs resulting in protein coding RNAs. Whether this may be a common or rare phenomenon, the occurrence of such CRISPR/Cas induced indel mutations, has serious implications for safety as well as predictability, clearly indicating the need for rigorous risk assessment.

These findings are also a re­minder that CRISPR/Cas9 is a new technology that due to its ease of use, has found wide-spread appli­cation without the necessary time to establish all the consequences and risks of that use.

Finally, with regards to safety and predicability it is important to recognise and establish, that “precision” on one level does not equate to precision or predictability on another level.

The impact of the changes at the nucleotide level on gene regulation, metabolism, cell communication, interaction with other organisms in different environmental settings cannot be deduced in a linear way. Only investigation can give reliable answers.

In conclusion, it is a fallacy to equate ‘precision’ or predetermination of location with ‘safety’ and ‘predictability’. Furthermore it is incorrect to misinterpret “high efficiency” as ‘precision’, as shown by all the examples above.

Furthermore, the size of a mutation cannot be correlated with the impact or safety outcome of such a mutation. It is well known that point mutations on their own can have detrimental effects. The application of genome editing affecting a multitude of genes simultaneously either due to multiplexing or the presence of multiple copies of such genes adds to the various risk scenarios already identified above, as does the possibility of sequential use of genome editing. Combined, – even if “only” SDN1 is being applied – significant changes can be made to any organism of choice.

Page 13

The fact that a mutation may also occur under natural conditions does not make it a safe mutation or a mutation that would last through evolutionary pressures and selection. The issue of ‘detectability’ or lack of detectability is equally not a measure of safety. Indeed, genetically engineered organisms are also only detectable when it is known what is being looked for. Equally, SDN1, SDN2 and SDN3 derived GEd organisms will be detectable if the relevant sequence information is deposited and known.

It should also be borne in mind, that genome editing makes the whole genome accessible to change, unlike to conventional mutational breeding, as some areas of the genome are under normal circumstances more protected than others. This has been detailed in Kawall 2019 (Fronties in Plant Science, 24 April 2019, https://doi.org/10.3389/fpls.2019.00525). Furthermore, genome editing will affect all copies of a gene, whilst conventional mutational breeding will only affect one individual copy.

Page 13

Whilst it is being acknowledged in the present DRAFT Regulations that current genome editing tools (here SDNs) are not “completely” error-free, it is failing to portray the full extent of the unintended effects and impacts both in quality and quantity. Indeed, it is brushing aside, and is under-stating the unintended effects by the way they are being portrayed, in particular to only mention them in a minor way right at the end.

Additionally, the text fails to address most of the points we have raised above, and does not even mention unintended on-target effects and impacts.

Given that this present draft text is meant to address the issue of risk assessment and safety, it unfortunately falls seriously short of identifying the risks and giving them the space and depth required and placing them at the centre stage, where indeed they should be – given the obligations and duties at hand.

6. Tiered approach for the risk assessment – page 14

This view also acknowledges that GEd is a new set of technologies and that there is no history of safe use. Therefore, comprehensive and rigorous risk assessment is required in line with the precautionary principle.

Figure 2, page 14.

The depiction of the risk scenarios in genome editing is based on assumptions and selective view and not on evidence. There is no reason to assume or to assert that the unintended effects and the off-target effects are less for SDN1 than SDN2 or SDN3. Nor is there reason to assume or assert that the consequences and/or risks of such effects are less for SDN1 than SDN2 or SDN3.

This kind of approach is unacceptable.

Table 1, page 15

“Grouping of the GEd organisms”

Given the scientific evidence presented above and given the fact that GEd is not only a new technology but a continuously changing and growing technology, there is no justification for such grouping, where Group 1 intended to be excluded from proper risk assessment as indicated in Table 2 page 20.

There is no scientific justification to treat edits/deletions/insertions of one or a few basepairs any different than those with 20 or 42 or 177 (arbitrary numbers). A single base pair change can result in as severe consequences as those related to the other figures. Complexity or lack of complexity again is no guarantee for safety, unless established through rigorous scientific investigation and evidence.

There is also a difference between assumed lack of complexity and actual complexity, which is not being addressed here. Knock-outs hold their own risks, which again require testing rather than assumption of safety, or assumption of knowledge. Given that process-induced mutations and on- and off-target effects can happen at any level of SDNs, all SDN categories should be investigated and risk assessed to the same standard.

Furthermore, the use of transformation technologies to deliver the SDN to the cell as well as tissue cultures required (including use of protoplast for plants) has its own set of risks for process-induced genome-wide mutations, which has not been covered at all in the risk assessment (e.g. Wilson et al. 2006. Biotechnology and Genetic Engineering Reviews, 23:1, 209-238, https://doi.org/10.1080/02648725.2006.10648085).

Page 16

Paragraph 1:

The fact that a GEd organism may be “similar” to a naturally occurring variant, cultivated or wild species or mutants does not establish safety. The statement that products from “such organisms” have a “well established history of safe use” is somewhat worrying, as it is not only a question of “products’ of such organsims, but the behaviour and interaction of that organism itself; and furthermore to assume wild species (and all their products and compounds) have a history of safe use.

The rationale that follows in the following sentences of that paragraph is in our assessment unsupportable. Group I GEd do not “mimic” such naturally occurring mutants – (though here the chemical/radiation induced mutants are no longer mentioned, why?). Firstly, that something is similar does not make it safe, that a mutation may occur in nature does not make it common or probable, or safe, nor does it take into account that GEd is a process with process induced mutations and on- and off-target effects.  Furthermore, to assume that history of safe use for cultivated germplasm or conventional breeding or conventional mutational breeding means “history of safe use” for GEd Group I organisms is outright wrong and perverts the meaning of “history of safe use”.

We already covered the point of indistinguishable and/or undetectable earlier, again showing that this supposed indistinguishability is not a proof for safety, and it does not take into account process-induced mutations, nor is it correct to assume there is not a way for detection, traceability, transparency or labelling.

In conclusion, Group I GEd organisms have no history of safe use, and there is no scientific nor other justification to declare them as safe and treat them any differently than SDN2 or SDN3 or Group II & III.

7. Regulatory approval – Page 20

Based on all the above points, we strongly object on scientific grounds to the regulatory matrix developed in Table 2, in particular to the exclusion of Group I plants from RCGM, GEAC and Statutory Market regulation. Without a robust risk assessment and approval there will be undue risks to farmers, environment as well as human health.