Genome editing on finfish: Current status and implications for sustainability

Novel genome editing techniques allow for efficient and targeted improvement of aquaculture stock and might be a solution to solve challenges related to disease and environmental impacts. This review has retrieved the latest research on genome editing on aquacultured finfish species, exploring the technological progress and the scope. Genome editing has most often been used on Nile tilapia ( Oreochromis niloticus Linnaeus), followed by Atlantic salmon (Salmo salar Linnaeus). More than half of the

Norway accounts for over 50% of the world's total production of Atlantic salmon (Salmo salar Linnaeus). 4 Despite being highly economically viable and providing working opportunities and export revenues, salmon production is subject to controversies rooted in the challenges the industry faces related to environmental impacts and animal welfare, thus hindering sustainable development. 5,6 The development of a more efficient aquaculture requires increased utilization of available genetic resources. 7 This includes use of valuable genetic material within selective breeding as for example marker-assisted breeding. 8 Genetic resources are also very useful for introduction, removal or single base exchange using genome editing (GE). 5,9,10 The use of GE demonstrates some promising possibilities for improvement of the aquaculture stocks, 11 with impacts for sustainable and efficient aquaculture. 5 The first approaches using genome editing included techniques as zinc finger nucleases (ZFN) and transcription activator-like endonucleases (TALEN). At present, the most novel method, the clustered regularly interspaced palindromic repeats (CRISPR) system, dominates. This system offers the possibilities of making small changes by fixing alleles and changing trait loci. 9 The CRISPR system is at present considered to be the most efficient, targeted and affordable genome editing technique. [12][13][14] Further expansion of the aquaculture production, with the aim to meet future need for food and economic growth, requires contribution to sustainable development. Sustainable development was originally defined by the Brundtland Commission as the '[…] development that meets the needs of the present without compromising the ability of future generations to meet their own needs'. 15 In 2015, the UN set out the 17 common sustainable development goals (SDGs).
These were based on the thoughts from the Brundtland Commission -and are common guidelines on how to achieve a sustainable world.
The goals are integrated in each other, emphasizing that everything depends on everything, and provide a balance where the three dimensions of sustainable development, environmental, economic and social, co-exist. 16 According to Stockholm Resilience Center, food connects all the SDGs. 17 Aquaculture and fisheries are both crucial for future food security, and '[…] offer development pathways to contribute to a more prosperous, peaceful and equitable world'. 1 It is therefore also of crucial importance that new solutions like genome editing can be used in sustainable manners.
Here, we present findings from a systematic review on the current status of genome editing in aquacultured finfish species, hence extending previous reviews. 5,9,18,19,20,21,22 As published in the previous reviews, 5,9,22 there is still a high focus on reproductive traits, but this has recently been expanded to include genes related to other production traits such as disease resistance.
The geographical origin of the research and innovation activities using GE on aquaculture finfish has also been reviewed. In addition, we have compared the number of reports wherein genome editing is used on a specific fish species with the commercial relevance of the species in aquaculture. In the systematic review, several of the identified studies have included some discussion of technical barriers by genome editing including off-target effects, which is highlighted here with the potential solutions. These challenges are also of regulatory relevance and need to be addressed by concrete regulatory approaches. 23,24 Regulatory approaches and concerns have just been briefly discussed in previous studies. 7,21,25 Here, we describe the regulatory approaches in the main countries researching genome editing on aquacultured finfish, and whether the countries have included non-safety factors, as contribution to sustainability, socio-economic and ethical aspects, in assessment of genetically modified organisms (GMOs). Norway is one of the countries which have included non-safety criteria in the regulation of GMOs. Here, we briefly elaborate on how the Norwegian impact assessment regulation can be used for a sustainability assessment of genome edited aquacultured finfish species.

| Genome editing technologies
Since the discoveries of the DNA structure and function, further research has focused on the ability to modify gene sequences.

Enzymes like polymerases, ligases and restriction endonucleases
provide the ability to make changes through cutting and ligating, and the polymerase chain reaction (PCR) offers isolation of fragments.
Repairing lethal DNA breaks is inherent in cells endogenous machinery. Thus, combining the possibility to both introduce breaks at the desired sequence and cellular self-repair is the foundation for GE. 26 During the last 20 years, several new techniques for modifying DNA have emerged, both oligonucleotide-directed mutagenesisbased techniques (ODM) and nuclease-mediated site-specific mutagenesis techniques. In this review, we focus on targeted alterations of the fish genome and the site-specific nucleases (SSN), while also recognizing ODM-related activities such as RNA interference (RNAi). There are four categories of site-directed nucleases: meganucleases, ZFN, TALEN and CRISPR. 27 ZFN is composed of modular DNA recognition proteins. 22 When associated with restriction enzyme FokI, the complex can be designed to recognize specific chromosomal sequences of 9-18 nucleotides, and at dimerization, the FokI enzyme can induce double-strand breaks (DSB). 26 Use of ZFN was established in 1996 and its use within research increased from 2003. The method was hampered by difficulties of design and validation of proteins for specificity in the complex. In addition, ZFN had low efficiency with very few mutations in F0 generation (parent generation), leading to low transmission to F1 generation (first filial generation). These challenges lead to a newer tool emerging in 2010/11, TALEN. As with ZNF, TALEN is using the restriction enzyme FokI and the cleavage requires dimerization. TALEN is, however, easier to design and validate than ZFN and recognizes fewer nucleotides, thus being more efficient than ZFN. The protein design, synthesis and validation are, however, still not efficient enough which hampers widespread use of this tool. All the site-directed nucleases use the organisms repair system to induce either site-specific mutations (insertion or deletion, indels) or insertions of new sequences. 27 The most recent technology, CRISPR/Cas nucleases emerged as late as 2012/13 26 and are molecular features of bacteria and archaea for recognition, thus | 3 protection against virus infection. 28 This system is RNA-mediated and performs sequence-specific detection and silencing of foreign nucleic acids. The CRISPR system is organized with the Cas proteins (CRISPR-associated proteins) encoded in operons and 'CRISPR arrays consisting of genome-targeting sequences (called spacers) interspaced with identical repeats'. 29 The repeats are short fragments from foreign nucleic acid that has entered the cell (e.g. by infection of viruses). 26 In the genome editing system, guide RNAs (gRNA) lead the CRISPR system to the target DNA sequence and cleave the target site by the nuclease. The first studies of the CRISPR/Cas system were performed in 1987, while the first publication on CRISPR system for GE was published in 2012. 29 The nuclease-mediated site-directed techniques ZFN, TALEN and CRISPR induce a DSB at a specific site in DNA. This stimulates natural repair mechanisms. One repair mechanism is non-homologous end-joining (NHEJ), which induces random point mutations, inserting or deleting material (indels). Alternatively, if a donor DNA strand homologous to the sequences bordering the DBS is provided, a homologous directed repair (HDR) will happen. The type of donor determines the type of repair, insertion or replacement of a sequence within the DBS, correction of a base or deletion of a sequence. 9,27,30 The mutations lead to either knockout (KO) or knock-in (KI) of a gene or DNA sequence.

| Genome editing in aquacultured finfish
As well as being an important research tool, CRISPR could provide an efficient way to expedite genetic improvement of farmed animals.
Aquatic animals are easy to work with compared to many terrestrial species due to high fertility rates, short generation time and external fertilization. 9 In 2015, Ye et al. 21 reviewed different fish breeding methods and pinpointed CRISPR system as promising for '[…] efficiency, precision and predictability […]' in fish aquaculture. This was later followed up by Zhu and Ge 22 which published a study on recent advancements in genome editing on finfish, focusing on reproductive traits. 22 Other possibilities were later presented by Gotesman et al. 19 where genome editing and RNAi were pointed out as useful therapy tools for combating pathogens in aquaculture. A concomitant review by Elaswad and Dunham 18 described how different genetic and genomic tools for disease reduction in aquaculture could be achieved by the CRISPR/Cas system. They also highlight the possibility for knock-in (KI) procedures and to the benefits by the combination of genome editing and selective breeding. 18 The increased speed of technology development within gen(ome) sequencing has aided the rapid development of genome editing technologies. Houston and Macqueen 20 reviewed the exploitation possibilities from sequencing and annotation of the Atlantic salmon genome. They build from Lien et al. 31 which was part of the Salmon Genome Project and had a special focus on the ecology, physiology and evolution of the salmon genome as well as highlighting further possibilities by genome editing. Wargelius 5 focused on sustainability issues related to Atlantic salmon production and other relevant solutions that genome editing may offer. Subsequently, Gratacap et al. 9 published a review on current technical possibilities that genome editing offers for aquaculture species globally. The latter publication listed 21 studies where genome editing was used (successfully) on different aquaculture species (including one oyster species) and categorized the solutions according to traits. To present the current and future status of use of genome editing on aquaculture finfish, we have performed a systematic literature review.

| ME THOD
The methodological approach used for the systematic literature search is based on relevant items from the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). 32 Using GS, the retrieved articles were often duplicated since they were from different websites and often composed of newspaper/magazine articles or master theses, while WoS allowed for more precise search (e.g. no newspaper/magazine articles or master theses). Different search strings were also used (see Appendix 1). One search string contained a list of the major aquaculture finfish species given by FAO. 33 These 20 fish species made up 84,2% of total aquaculture production worldwide. 33 An updated list was published in May 2020 1 after the first searches were performed, but it did not contain any significant changes compared to the list of 2018.

| Search strategy
The initial identification of articles was mostly based on titles.
After identification, each of the abstracts was screened for exclusion records (see Appendix 1). Different exclusion criteria were made because the two databases yielded different types of output lists. This was followed by merging all retrieved scholarly articles having an experimental approach into one list, and any duplicates were removed.

| Grouping of data
The strategy for grouping the data was done inspired by Catacora- Vargas et al. 34 in order to identify the direction and location of the genome editing field associated with aquaculture finfish species. The review articles were used as supplements in the current work and were not analysed to the same detail as the experimental articles.
The data in the experimental articles were grouped after: species, objective of the study, trait, gene(s), type of genome editing results (NHEJ/HDR) and institutional affiliation of 1st author. The search for technical barriers in the articles was done through searching for relevant terms in all articles and then coding relevant paragraphs in NVivo 12.6.0 software, followed by analysis of the coding book.

| RE SULTS
The  Table 1 presents the resulting papers included in this review. The CRISPR/Cas system dominated the field of genome editing on aquacultured finfish ( Figure 1).
We found two scholarly publications using ZFN, 35,36 one study using TALEN 37 and two studies using both TALEN and CRISPR. 38

| Species and traits
The search included the 20 most exploited aquaculture finfish species globally. 33 Table 1 lists the results according to species and area of interest, while Figure 2 shows the distribution of species. The two most studied species are Nile tilapia (Oreochromis niloticus Linnaeus) and Atlantic salmon. Today, the main traits that are selected for in aquaculture in the United States, Europe and China through breeding are growth, disease resistance, processing yields and product quality, reproductive traits, feed conversion efficiency, morphology and tolerance to environmental stressors. 7,87,88 It could therefore be expected that these traits would appear in the studies retrieved in this review. Reproduction (maturity/fecundity) and development were the most studied traits, found in this systematic review, see Figure 3. This also included sex determination and sterility. Then came growth, pigmentation, disease resistance, use of trans-GFP and omega-3 metabolism. The traits studied mirrors the most important traits in modern breeding, where, for example, omega-3 content in fish can be considered important for product quality for human consumption.
In Table 1, we have included categorization of what areas of interest the different papers indicate to have. Considering the CRISPR field of research to be quite young, we acknowledge that areas of interest in each study is/are focused on key issues such as maturity/ fecundity -thus being overlapping. However, we have attempted to assign each study the field of interest we consider most prominent -for example being technology development or final productoriented such as production of sterile fish for aquaculture.

| Geographical origin of genome editing research compared to major finfish producing countries
In our analysis of the literature, we investigated the institutional affiliation of the 1st author for each study to determine the geographical location of the research, see Figure 4. China is still on the top. 9,22 Others are the United States, Norway, UK, Japan, Egypt, Czech Republic, Republic of Korea, India, France and the Philippines. Some of the papers have been credited two countries because the 1st author had two institutional affiliations at the time of publication.
China has produced most publications (29), followed by the United States (9) and Norway (7).
For countries with aquaculture production, the choice to consider genome editing as an approach may depend on the type of challenges the country/region faces, regulative conditions, knowledge about the species and wild relatives and consumers acceptance of GM/GE foods. Moreover, Wargelius 5 argued that a prerequisite for genome editing is that the species genome is fully sequenced and annotated. Considering these proposed criteria, we expect there to be some correlation between the species importance in present aquaculture production, for how long they have been produced, first selective breeding study (history of aquaculture), and to whether genome editing has been approached for this species.
According to FAO, 1 Asia is the major aquaculture producing region according to volume (88.69% of global production), and China is the largest country with a total of almost 58%. America produces 4.63%, Europe 3.75% and Africa 2.67%. It is evident that China as the most producing aquaculture country is also the one doing most research on the use of genome editing on aquacultured finfish.

TA B L E 1
Overview of genome editing in aquaculture finfish with respect to fish species, field of interest, specific trait and gene(s), additional remarks, genome editing system and institutional affiliation of 1st author. Abbreviations can be found listed alphabetically below the Growth Cleveland et al. 66 Cleveland et al. 67

TA B L E 1 (Continued)
Norway, the third most important country identified in our study, produces 1,65% of the total volume. Norway does however account for over 50% of the world's total production of Atlantic salmon. 4 This history of aquaculture could also be compared to the species used in studies of genome editing to see whether there is a correlation between history of farming and the interest in novel tools like genome editing, see Table 2. Nile tilapia is the species which according to the review of Houston et al. 8  The second most studied species with regard to genome editing was through our retrieval, the Atlantic salmon. All these articles had their first author affiliated to a Norwegian institution, except one using Norway, yet it is already the species which globally has the most exploited traits for breeding programmes. 87 The genome of the Atlantic salmon was published as a bacterial artificial chromosome-based map first, 101 and later a high-quality whole genome of the Atlantic salmon was published by31 as part of the Salmon Genome Project.

| Technical challenges and off-target mutations by using CRISPR technology in finfish
The use of genome editing on finfish, either for commercial use or in research, brings technical challenges that should be considered.
Some of these are off-target mutations and mosaicism in the F0 generation. Considering the discussion from the papers identified in this review, there is a further need to identify the presence of off-target and other unintended effects. This may imply to use recent developments as next-generation sequencing and multi-omics approaches, as seen approached in Jin et al. 53 These methods need to be sensitive enough to distinguish between natural variation and mutations introduced by genome editing.

| Effect of ancestral whole-genome duplication
Another challenge relevant when discussing teleosts, and especially salmon, is ancestral whole-genome duplication (WGD) events and particularly the salmonid-specific 4th round (Ss4R).

| REGULATIVE FRAMEWORKS IN COUNTRIES DOING GENOME EDITING ON FINFISH -CRUCIAL FOR USE?
How to regulate genome edited organisms as plants and animals has during the recent years been discussed. Regulative issues concern both whether genome edited organisms should be regulated under present regulative frameworks for GMOs or if they should be exempted, and whether the regulation is according to product or process. 105,106 Compared to older GMOs, the newer genome edited organisms can be generated without use of transgene sequences. 105 This is a common topic of discussion, even though insertion of desired sequences is possible using HDR, as shown in four of the retrieved papers of this study. 55,58,72,78 Regulative concerns could affect the use of genome editing in applied research with the goal for commercial use. 9 It has been argued that GMO regulation may hamper research and innovation of genome edited organisms due to the excessive regulatory requirements placed on GMOs. 107 Ishii and Araki 105

| CONTRIBUTION TO SUSTAINABLE DEVELOPMENT
Our review outline that some of the challenge's aquaculture is experiencing, like disease and genetic contamination in wild stocks, has possible solutions through genome editing. In Norwegian aquaculture, an expansion of the salmon farming industry requires transition to a more sustainable production. This final section will therefore discuss how the different solutions retrieved in this review can contribute to a more sustainable salmon production, based on how the contribution of genome edited organisms to sustainable development is evaluated under the Norwegian Gene Technology Act (GTA).
The GTA is a unique regulation that requires, besides assessment of risk to the environment and health, consideration of the ethics, social utility and contribution to sustainable development of GMOs.
This, in addition to the urgent need for innovation and new solutions in aquaculture, is reflected in the focus of the Norwegian studies retrieved in this review, where the aim is to generate a fish more appropriate for a sustainable aquaculture. 25,57,58,59,60,61,62 In support of the research on the germ cell free Atlantic salmon, other studies have looked at growth and maturation, 115  The control questions regarding ecological boundaries and global effects on biodiversity should therefore be taken into wide consideration when evaluating genome edited organisms. All the control questions should also, according to the Norwegian Act, consider both the product and process, to ensure that sustainability is regarded throughout the whole production line/supply chain. The impact of aquaculture on nature environment is also to a large extent the driving force for proposing use of genome editing. However, solving ecological issues cannot have a negative impact on society and/or economy; therefore, all aspects must be evaluated.
The first control questions relevant for aquaculture finfish regarding global impacts and ecological boundaries ask whether the biological diversity is affected globally, whether the ecosystem way of function is affected and whether it will affect energy utilization, climate gases and pollution. Here, the research on reproduction and development is important. Sterile fish will not be able to reproduce with wild stocks after escape, and hence, the impact on environment will be reduced. In Norway, the issue with escaped fish is highly urgent. Güralp et al. 61 have recently published a method using a combination of genetic sterility and rescue, which may allow large scale production of sterile salmon. 61 A sterile fish will not only aid this issue, but it would also be considered a prerequisite for using genome edited fish in ocean pen production. Here, we do, however, want to emphasize the need for more research on how such a sterile salmon would impact wild relatives and surrounding biodiversity when it escapes. 119 Disease resistance could aid any aquaculture sector globally, and it would aid both the economic efficiency of the production, but also animal welfare and the impact on wild stocks, thus both biodiversity and responsible productions aspects of sustainability. Increased welfare is, alongside with sustainability, assumed an important argument for application of genome editing in aquaculture, especially in a country like Norway where ethical responsibility is implemented in the Act. 120 In addition, the Norwegian Animal Welfare Act states that all animals, including fish, have intrinsic value independent of their utility for humans, and shall be treated well and protected from unnecessary pain and strain. 121 Any implementation of genome editing in aquaculture has to consider this and elaborate how animal welfare should be considered for the species in question.
Secondly, the control questions include questions on the distribution of benefits and risks between generations and rich and poor.
Anticipation of both the potential beneficial and adverse consequences of using genome editing in aquaculture is difficult because there is no former use to refer and learn from. Regarding GMOs, the standard implication is often that even though we remove an issue, for example disease resistance, some other issue will follow, as for example a new pathogen implying that one need to consider a longer timeframe when assessing potential impacts.
Another important aspect regarding future generations is the preservation of the wild salmon stocks in Norway. Norway holds approximately 25% of the total world population of Atlantic salmon, which has encouraged the preservation of this species. 122,123 In this context, a sterile genome edited fish is not only a solution, but should be a prerequisite for use. Other considerations to be made are whether genome editing allows for intensification or maintenance of the aquaculture production volume. If the former, is that representing a threat or benefit for the opportunities of future generations?
The knowledge earned from studies of genome editing in one species can be used, albeit to a certain degree, in another. The research performed can therefore be useful for other countries with other aquaculture related challenges, including poorer countries with less resources to conduct this kind of expensive research on their own.
This transfer of knowledge depends on transparency of the process and the product.
Finally, the control questions are summed up in questions on how the ecological impacts and distribution between generations and rich/poor affect the economic growth. These questions are not directly related to the solutions proposed, but an economic analysis that is outside the scope of this review. We will, however, go briefly through how economic traits could contribute to sustainability.
Pigmentation can be an economic trait, as seen for common carp in various colours, but also a tool in development and use of genome editing like CRISPR/Cas9. Regarded to be a commercial and ornamental trait, this modification will affect goals related to economy through social interest as for example aesthetic value.
Both pigmentation and the use of trans-GFP have been applicated in studies aiming at developing CRISPR or TALEN as tools for aquaculture. The sustainability contribution of this use of genome editing will therefore depend on the knowledge generated from the activities. It could, however, also have importance for biosafety as the lacking pigmentation can be used to identify escaped genome edited fish.
In studies looking into growth, eight out of ten studies had aquaculture as main focus (Table 1). Increasing growth for increased production efficiency is valuable for reducing feed costs, but could have implications for welfare, as seen with bone defects after sp7 and mstn KO in common carp. 13,14 In Norwegian salmon production, growth has for long been an important trait in breeding efforts, and here the process is regarded a success. Increased growth can therefore not be regarded as priority in the development of a sustainable production in Norway.
Omega-3 is especially relevant in Norwegian aquaculture, as sufficient amounts of omega-3 fatty acids sustain health benefits for both fish and humans. 59 As described by Datsomor et al. 59,60 LC-PUFAs in the feed is an important contribution to omega-3 synthesis in the salmon. This could lead to less need for live feed and/or fish oil in the feed, which would be of economic and ecological benefit. 124 Efforts within the genome editing field have also been aimed to generate omega-3 producing plants for use in fish feed. 125 This could be an alternative for approaching the issue more directly, alternatively in combination.
Lastly, we want to express the necessity for modifications, additions and changes to be made for the sustainability guidelines to be adapted for evaluation of GE and GM animals, and aquaculture finfish species more specifically, as seen for herbicide tolerant crops in Catacora-Vargas 126 and by the Norwegian Biotechnology Advisory Board. 127 We find it necessary not only to adapt the questions to evaluation of living GM/GE animals, but also to specify the core ideas and evaluation questions. It does, however, give a brief idea of the complexity of addressing genome editing solutions as sustainable because they might (contribute to) solve environmental issues. More study is needed on how to evaluate sustainability in relation to genome editingo fish, in addition to (experimental) study of the effect of genome edited finfish on environment, economy and society.

| CON CLUS ION
We have found that the main traits researched are reproduction and development, growth, pigmentation, disease resistance, use of trans-GFP and study of the omega-3 metabolism. Compared with previous reviews, we find that there are other genes targeted in more recent studies. Reproduction is still the most targeted trait, but there is also an increase in other traits such as disease re- All the solutions found in this review can contribute to sustainability in each their own way. We emphasize the importance of prioritizing environmental sustainability in this regard. Biodiversity is of crucial importance to any food production system, also aquaculture. Its preservation should therefore be of main interest to both breeders, policy-makers and consumers. Evaluating the effect of a GMO on sustainability is required by law in Norway, and description for assessment has been developed for this specific term. These are, however, not fit for a thorough evaluation of live animals and should be revisited.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.