Safe and purposeful genome editing under harmonized regulation for responsible use: views of research experts

CRISPR-Cas9 revolutionized the precise editing of mammalian cells genome. The present study explores genome editing (GE) in the context of the Responsible Research and Innovation framework for emerging technologies, through semi-structured interviews with life sciences researchers worldwide. Our study demonstrates that for researchers in the field, GE technology is viewed as promising but also harboring unsolved challenges. These experts call for complementary research to improve the technology and increase knowledge of the genome function. They clearly do not support what they perceive as unsafe, unpredictable and irrelevant applications, and they view the lack of international harmonization of regulation in combination with cultural differences in public attitude as difficult challenges. Interviewees see public misconceptions as a problem while recognizing the need to foster a clear science-society dialogue with informed citizens. This study with scientists provides insight into the science-based priorities for GE to be a technology that can be responsibly applied.


Introduction
Clustered Regularly Interspaced Palindromic Repeats and its associated nuclease Cas9 (CRISPR-Cas9) was first presented as a revolutionary biotechnology tool for precise genome editing of mammalian cells in 2012 (Barrangou and Horvath 2017).Since then, numerous labs have shifted from traditional genetic modification systems to CRISPR-Cas9 genome editing (GE) technologies (Adli 2018), which are cheaper and more straightforward at equal or higher precision (Adli 2018;Camporesi and Cavaliere 2016).The rapid uptake of the technology is accompanied by ethical (Brokowski and Adli 2019), regulatory (Isasi, Kleiderman, and Knoppers 2016), policy (Nordberg et al. 2018) and societal (Howard et al. 2018) concerns including safety and ethical implications, making GE a highly relevant case within science and technology studies.
The research community's efforts to deal with the wider implications of GE has translated into events, guidelines and even requests for a moratorium for certain applications (de Lecuona et al. 2017).When participating in this discussion, scientists are effectively engaging in constructive technology assessment (CTA) and Responsible Research and Innovation (RRI) even though they may not be familiar with these concepts.Both CTA and RRI are based on the understanding that different actors can co-produce sociotechnical future scenarios (and imaginaries) to facilitate reflection over science and technology and how to use it responsibly for society (Borup et al. 2006;Jasanoff and Kim 2015).RRI builds on the ethical, legal and social implications (ELSI) framework developed by philosophers, social scientists and molecular biologists (Bucchi 2004;Felt 2018) and it is a public policy that envisages knowledge production between several actors for more responsible techno-scientific futures (Rip 2018).RRI enacts the four dimensions of inclusion, anticipation, responsiveness and reflexivity (Owen, Macnaghten, and Stilgoe 2012), which translate into a multi-stakeholder effort to anticipate scenarios, critically reflect and timely respond in an in-depth manner to positive and negative implications of emerging technologies (Felt 2018;von Schomberg 2013).
Discussing challenges, potential and applications is important for effective policy-making and responsive anticipatory governance with regards to GE as an emergent technology (Hurlbut 2018;Hurlbut, Saha, and Jasanoff 2015).Including multiple stakeholders promotes foresight and technology assessment in the name of "collective expectations" (Konrad et al. 2016), somehow addressing the premise of "technologies of humility" as a mean to deal with uncertainties as part of this debate (Jasanoff 2003).Within this theoretical framework for technology anticipation, exploring expert views in a particular topic with the aim of developing arguments and reasoning on technology assessment seems a significant approach.The discussion of the societal consequences of GE has a history of several decades, but the emergence of CRISPR-Cas9 is changing the technological potential in a way that makes the discussion much more urgent.This became evident in the course of the present study when the birth of the first genome-edited human babies was announced (Regalado 2018) and since then the significance and implications of the case (He Jiankui's case) have been explored (e.g.Meyer 2022;Wahlberg et al. 2021).
In the present study, we aim to understand how scientists using CRISPR-Cas technology reflect on the technology and its use.With this, we aim to gather and make sense of scientists current thoughts and concerns about GE.Using an in-depth qualitative approach, we explore what researchers see as realistic, desirable and concerning in developing and applying GE and the societal and governance consequences of its implementation for human life.The imaginaries constructed by researchers are structured into three overarching themes where they provide both problematization and ideas to deal with the development and the future of the technology from bench to bedside across species and across the world.

Interview guide design, and recruitment of participants
The interview guide included open-ended questions combined with a ranking exercise representing hypothetical applications of GE (Supplementary material).The interview guide was triangulated within the research team and tested in two pilot interviews with researchers.
Participants were recruited using convenience and snow-balling non-random sampling as well as purposeful sampling.They were recruited from different parts of the world to ensure geographic diversity of perspectives and experience.All participants are researchers from biomedical (BSR) and animal sciences (ASR) holding at least a PhD and with previous and/or current experimental practice and published records on genetic engineering, preferably on GE.Recruited researchers could be working with GE in human/animal cell models or in vertebrate organisms.From a total of 32 invitations, 22 researchers accepted to be interviewed, including researchers from Europe, Brazil, North America and China.The interviews were done online or by telephone and lasted 60-120 min, taking place between July 2018 and March 2019.

Interviewing and thematic analysis
Audio records of interviews were collected.Transcription was achieved resorting to an automated method -AUDITwhich simultaneously allowed preservation of participant's anonymity (Ramos, PADFR 2021).Transcripts were imported to NVivo 11 software (NVivo, QRS International (Americas), Burlington, MA) for thematic analysis.
Our study followed a hybrid inductive and deductive coding process.Initially, a subset of transcripts were coded manually and separately by the first and last authors.Through a first triangulation step, a codebook was sketched (Supplementary material).A second coding step was performed by the first author covering all transcripts.A sub-coding process retrieved additional refined codes stemming from some of the first ones.Following this coding process, the three authors worked together identifying patterns in the data leading to the generation of themes.

Ethics compliance
This study received ethics approval from the Ethics Committee of our institution, in May 2018 (process number 265/2018/CETI).All participants signed an informed consent form at the time of their participation in the study (Supplementary material).No personal data other than name was retrieved and the results are presented using pseudonyms.

Findings
The main story emerging from the qualitative analysis of the interviews with scientists is one about how to develop gene editing as a technology so that it can safely be used outside the laboratory, about its acceptance and about the challenges of governance internationally in a diverse context in terms of societal acceptance and access.In the following, the results of the analysis are organized according to the overarching themes technology, governance and society.
Technology: technical challenges of GE need to be solved to make it safe enough to proceed to in vivo applications Of the many different specific challenges of GE identified by the interviewees, the vast majority have implications for safety, with consequences for the different applications of the technology.The main challenges identified are the efficacy of introduction of genetic traits in species, the delivery of the CRISPR-Cas components, the efficiency of the homology-directed repair (HDR) pathway to introduce gene segments as well as the fidelity of its insertion (adverse on-target effects) and finally the undesired effects of off-targets of the CRISPR-Cas components and the genome complexity of organisms.Figure 1 shows an overview of the challenges and how they relate to each other, and to the potential implementation of GE technology in different contexts.Whereas the interviewees repeatedly highlight the need for researchers to improve technology in order to overcome these challenges, they also express that if the technology is considered as "not safe enough" it is preferably to explore GE in vitro and avoid using it for in vivo applications.
Technology development and application in multiple organisms GE technology has been advancing at a fast pace.As noted by one of the interviewees: "[…] the beauty of Cas9 is that there are no technical hurdles, there are virtually no technical hurdles to getting the genome to edit." (Megan-ASR).This opens the door to a wide range of applications: "[…] with the technical level that we have now it's clear that [it] can be applied to all kind of species."(Jay-BSR).With the possibility to correct multiple genes simultaneously, CRISPR-Cas9 reduces the time for gene correction: And that means not work only at one at a time.But 10 genes or just any number we want, let's say something between one and 10 […] to finally become faster.Faster in terms of introducing more genetic correction […].(Jay-BSR) A prevailing view among the interviewees is that CRISPR-Cas9 has facilitated the implementation of GE which they point out might broaden possibilities of the application of the technology.
Efficiency, specificity and fidelity crucial for safety Efficiency of GE is fundamental: "[…] to increase a lot the ability of the system to get into cells and not only to get into cells but to correct the sequence as we wanted" (Ruben-BSR).Likewise, improving specificity of Cas9 and fidelity of DNA template integration are seen as important steps for future CRISPR technology development: So, there's the good where the foreign DNA is inserted at the target site, then there's the bad where the foreign DNA inserts itself somewhere else in the genome and then there's the ugly, the ones where the DNA is inserted on the target site, however insertion is not precise.(Irvin-BSR)

New Genetics and Society 5
Interviewees stress that the type of application defines the required level of fidelity of CRISPR-Cas9: "As we are talking about the somatic cells, we need to control a lot of insertions; we cannot do this in germ cell lines for now" (Lisa-BSR).In the same way, participants mention that it is easier to deliver CRISPR-Cas9 to certain type of cells in vitro than in vivo: For example, defective genes in muscles and nervous system tissue can be corrected with amazing creativity and efficiency in cell culture using the new abilities of CRISPR.However, those corrections often cannot be delivered to tissues in an effective way that is meaningful for patient health.(Charles-BSR) Broadly, the interviewed CRISPR experts highlight diverse technical challengeswith regards to efficiency of repairing genes, specificity of the CRISPR components used and fidelity of the integration of the new DNA materialthat must be solved in view of safety of the process before its application in patients.In this way, technical aspects relate to risk acceptance, further developed in another sub-section.

Delivery and off-target effects influences safety
Several interviewees refer to anatomical characteristics making specific tissues particularly apt targets -"a possible entry point for in vivo GE […] will be in organ systems that are easily reachable and well and anatomically well-defined like the eye" (Irvin-BSR) -, suggesting that these organs will be the first on which somatic GE will be applied.For in vivo delivery of GE technology, viral vectors rather than nanoparticles are preferred, and the need to avoid a potential immune response from patients is emphasized.So, for participants that see delivery as a technical challenge, they consider that before moving to clinical in vivo applications, essential considerations on "how to deliver CRISPR?" should be included: […] basically new methods should be found and presently [we] have an efficient viral approach where you have to inject enormous titers of viral particles, that is still not very efficient […] it inflates the tissue with these second viral vectors copies.That's totally unexplored whether this will be for used for human therapy […] So, delivery for gene editing [in] somatic cells; that is one big thing.(Jay-BSR) Participants also identify off-target effects (OTEs) as fundamental technical challenges that need to be solved before moving to clinical therapy.OTEs happen when a gene correction meant to happen in one gene ends up taking place in another gene affecting critical functions.For example, if the affected gene has impacts in cell death and cell proliferation, it may lead to complicated problems related with tumor formation and progression.Therefore, specificity of Cas9 nuclease cutting function and fidelity of the insertion of the DNA replacing material are identified as the main causes for OTEs.Irvin-BSR highlights the need to tackle: "Not only to the off-target activity at the nucleases themselves but to something that we want to act through, which is the importance of the specificity […] of which your foreign DNA is inserted into the genome and the fidelity".In general, participants note that a system to detect OTEs is essential: "so we're going to have to figure out a way to somehow assess off-targets or develop reagents that we're confident do not have off-target for the use in the clinic" (Megan-ASR).Not only GE for clinical applications but also germline GE needs OTEs assessment as pointed out by Benjamin (BSR): "[…] birth of an in vitro genome-edited embryo.I mean, it can be done but until the off-target and also many other scientific problems associated with this [are solved] it [is] still really [a] far away approach to do this.".Simon (ASR) alludes to He Jiankui's case of the genome-edited babies raising the question if the researchers had even been successful in knocking-out the intended gene (CCR5) to emphasize OTEs relevance for safety: "[…] we don't know the effects of injecting CRISPR in human people or sperm or oocyte.And I think we even don't know for example, that the Chinese [researcher] performed the knockout of the CCR5 […].And … could it have [had] another unexpected effect?".Another important observation is the need to assess effects in a wider context.This includes over time and across organs: Delivery to the right type of cells and the right gene, the latter hampered by the possibility of OTEs, are thus additional technical challenges to be solved before the application of the technology in view of safety.

Greater understanding of the genome
According to participants, next-generation sequencing can be used to detect OTEs, but they also stress the limitation of current methods focusing only on the most likely predicted off-target sites: Currently, what is done is actually [that] people are [using] the predictive path.It's made by sequencing, however, nobody is […] interested about the possible sites where it can be interacting or about additional controls […] you cannot predict all the changes.(Liam-BSR) The need for greater understanding of the genome, and the function and interconnectedness of genes, is also highlighted: Do we have enough knowledge about the genome to let's say: 'OK, for complex diseases we need to edit these in these genes' […] We need a whole research infrastructure […] on finding causal variants but people have been looking for [that] almost 30 years already.[And] still, we only have a couple handful of genes that we know.(Owen-ASR) In general, participants also mention that greater understanding of genome complexity and genetic diversity evolves due to research with non-human species: "You can address questions in a mouse, but I have to say the cliché, a mouse is New Genetics and Society 7 not a man.Therefore, if you can do in complementary species experiments, complement those results all together or if you can focus on other species, I think the insights will be better."(Irvin-BSR).

Risk acceptance
Again, the relevance of technical challenges in terms of safety and the level of risk that may be accepted varies with applications.About this, Vince-BSR, says: "[…] if you define in vitro as something that cannot survive outside the lab, well perhaps is relatively safe and not so much of an ethical concern.But if it's real organisms that can survive in the environment … then it might be different".Nevertheless, controlling for every factor when using GE technologies is considered utopic despite the absolute need of minimizing side effects and Steve (BSR) completes: "[…] of course you want the translation of the technology to really [be in] clinical use for human traits […] but these have to be really controlled and be in certain place like only in research use, in the laboratory use."Reflections over interventions in germline cells also express worries about genetic enhancement of future humans as noted by another interviewee: We cannot foresee what happens in the future, the future generations.We should never do stuff where we don't know what will happen.Once we delve [into] all the technology for germline modification that can be used to make artificial enhancement elements, that could create the nightmare scenario of designer babies.(Phillip-BSR) The importance of considering the type of application in dealing with technical challenges, seems a common concern for these experts, independently of their field of research.This concern around technical aspects and the emphasis on their impacts on the safety of applications provides robust support to the concerns already raised by STS scholars and which motivated the call for a global observatory to follow the main outcomes of research into GE of organisms (Jasanoff and Hurlbut 2018).Questions about how to take technological challenges into account at the regulatory level and how this may limit the acceptable applications of GE were also addressed by the interviewees, as explored in the following section.
Governance: purpose of GE interventions is important for regulation of the technology and criteria must be defined in order to implement it When discussing different applications, it is clear that for the interviewees, acceptability is directly related to how purposeful an application is perceived to be.They also stress the importance that criteria are established for when GE can be applied.Figure 2 offers an overview of the different topics explored by researchers at this stage.Participants suggested different questions that should be answered in the first place.
I think that the big question would be like: When is it OK to modify a genome and when is [it] not?You have a child that has a particular mutation that you know that increases the chances of having cancer.But that doesn't mean that will have cancer.Because we have other environment factors.So … when is it OK to say 'OK, so this is a situation where I can change the genome of the embryo and this is not the case'.So to draw that line -I think it's gonna be particularly hard to draw that line.[…] not just disease susceptibility, it is also diseases, different types of diseases.So for example is it OK to change everyone's CCR5 to avoid HIV infection?Is it OK to change other types of receptors?(Jay-BSR) The excerpt illustrates the criteria that the life sciences experts envision to delimit acceptable purposestherapy, prevention of diseases and reduction of disease susceptibilityas a first step towards governance of GE.

Access to genome editing
In addition to the question of eligibility, some interviewees also raise the question of accessibility to GE with regards to social justice.An interviewee illustrates that GGE "[…] might be creating inadvertently a new class of human beings.The ones New Genetics and Society 9 that are genetically modified and the ones that are not.And one cannot at this stage foresee if this will introduce a societal aspect" (Irvin-BSR).He continues by saying that editing of humans may result in prejudice and asks: "Are there people that will be genetically modified?Is there a status that people will acknowledge?Is something that is supposed to be private?[…] Will there be any discrimination in favor of people that underwent germline genetic modification?"The interviewee finishes by saying that access might be dictated by economic constraints, which may result in medical tourism: And related to that, will some wealthy people have access to these technologies almost by default?Just because they have money and they can travel from the countries where these technologies are not available to a country where these technologies are available.(Irvin-BSR) The general feeling expressed here is that accessibility to the technology is key and that avoiding/preventing different kinds of discrimination of citizens or group of citizens will be crucial.

Purpose affects acceptability
Editing of germline cells is generally not seen as acceptable by researchers unless as Vince (BSR) points out "[…] if one can prevent a life-threatening disease, perhaps I could give it a thought.But if it's just to increase the person's skills or something like that, then I'm against it by all means".When doing the ranking exercise (figures S1 and S2), researchers expressed preferences for applications that would help many people: So I'm gonna again put the food things and food security and getting rid of human pathogens that affect the most people up at the top.I really think that's moral imperative is to do [as] most good [as] possible.I guess cancer and genetic diseases would be next because lots of people get cancer and have genetic diseases.(Bruce-ASR) In general, interviewees also consider some applications like GE horses irrelevant (figures S1 and S2).Moreover, an application that initially is reasonable may also represent a slippery slope: Once you start it never ends so even if you're trying to correct the genetic disease that's fine, but the next year, someone would say 'but my child is not intelligent enough.This is a disease'.Just a normal intelligence.'We know how to improve it, please do this'.So then there's the enhancement.(Jay-BSR) The existence of alternatives is another criterion that researchers consider: "In the midst of existing highly effective prenatal genetic diagnosis methods, there is very little (near zero) legitimate need for CRISPR germline correction techniques to be used [for gene editing human embryos]" (Charles-BSR).The interviewees see a gap between technology potential and regulation with implications for misuse of the technology as suggested by Michael (BSR): "I think if there would be a way under to control that it's not misused you know, abused, um … for eugenic reasons.Then I think it could be OK.But who is controlling that and who gets who gets control and who draws the line so … I think that's an issue".
The purpose of the application of GE is viewed as major aspect for the acceptance of the technology in a way that also provide the foundation for the regulation of the technology.

Criteria to harmonize regulation
The diverse regulatory landscape of GE may open the way for the genetic tourism issue: […] let's say that that editing is allowed in North America or in China or South America or whatsoever.And that we would be able to buy genetic material from those parts whereas gene editing in Europe is still not allowed.I mean, [this] could happen, that is mismatch […] in legislations related to editing.(Owen-ASR) To tackle this, researchers evoke the need to apply regulatory criteria by defining: "[…] for which application would it be allowed and for which applications not?
[…] like one of the conditions could be: Is there an alternative or not?" (Owen-ASR).
Researchers acknowledge that stakeholder interaction helps in establishing such criteria when drawing guidelines for the use of the technology as noted by another interviewee: […] I don't say that we should use or take advantage of the extreme situation of a given technology, but we should know the limitations and then the society such [as] politicians and ethical people should draw the limits of where we should use the technology.(Frank-BSR) The importance of a regulatory harmonization by involving different stakeholders in the process is thus a common view highlighted by interviewees.RRI means collaborative work between the different societal actors in science and innovation.For the participants in this study, a desirable harmonization of regulation will be constructed together with stakeholders.The call for a more inclusive public engagement and involvement of stakeholders in science policy regarding GE (lltis, Hoover, and Matthews 2021) is backed here by researchers in the life sciences.The perceptions and attitudes of the general public will influence regulatory actions, which has consequences for international harmonization, as further discussed in the next section.
Society: cultural differences influence public perception, and scientists have a responsibility to engage in dialogue Society as an overarching theme is presented in Figure 2. When the interviewees reflect on GE technology in society, two main topics emerge: cultural differences and public attitude.Cultural differences between countries will affect public attitude to gene editing and indirectly also regulation worldwide.Some New Genetics and Society 11 researchers point out that different ethics pattern worldwide influence public receptivity: […] we have still this very western centric view of the world.[…] there was a very nice interview on TV […] it was about germline editing in humans.And someone asked the Chinese scientist, or talked about the European conservative view there and the guy was just laughing.(Michael-BSR) Likewise, religion defines some of researchers' perspectives about their ethics towards implementation of GE technologies: […] one thing that is for me personally is a bit of an issue … I mean, I'm a Christian and so, how should I compare with this gene editing?Is this a bit like playing God?But then to this extent are we doing it anyway already with our less classical ways of breeding?[…] I guess religious aspect is also important to somehow cover as well in the ethical debate […] even in cultures where there are a lot of Christian principles in it that has formed over all the centuries […] People think about it, even if they are not religious.(Owen-ASR) Ethics and public perception play into in the distinction of concepts and applications: There's a huge difference between germline and somatic because there's ethical concerns with germline editing in humans.Who has the right to make their child […] what traits are they allowed to change, right?Because you're getting into eugenics.And as a society in general, I think we shivered to think about that because it's associated with nationalism and racism and bigotry.It's been represented in a lot of science fiction movies as what the future will be.So when that's ethically allowed and what traits are being changed, … maybe genetic diseases will be allowed at first, but … starting to alter humans for other reasons.We don't practice genetic improvement on our own species, typically.So that's why I see it as very distinct from germline editing in animals.(Bruce-ASR) There may be a sequence to the acceptability of gene editing: "[…] it will be the first use in vegetables people will start to think that is not that dangerous and actually [it will be] helping them.And then [they] will probably accept better the use of CRISPR-Cas9 to therapy, drug and probably this will be done in the 5, 10 years' time" (Paul-BSR).
The influence of culture shapes the attitude of citizens as perceived by the CRISPR experts.Moreover, when ethical questions are at stake, interviewees highlight how the public discussion of GE is affected by associations with more general problems.

Fostering dialogue between scientists and society
The need for science-society dialogue is acknowledged: "[…] what is really needed is a societal debate about gene editing involving society and what you can do with the technology and then this is so I think that we can make an informed decision whether they would like or not" (Owen-ASR).Because implementation is at different stages, this debate will be different for human versus animal gene editing: I think for humans; the debate should be "if".[…] I think [it] is really complicated, especially with humans.In animals I mean, I think that it's too late for that debate on if it should be used because it is being used (Jane-ASR) Such a discussion in the public sphere is acknowledged by some researchers as being dependent "on who's starting this discussion, is it the large industry or is it a this is the scientist?Or is it a govern?[…] the public have different perceptions or trust in different types of organizations" (Vince-BSR).It also seems to be focused on specific cases of GE applications and leaving others out.Quoting another interviewee: "[…] I don't see so many people discussing gene drive and gene drive is much closer to application than […] the birth of an in vitro human embryo."(Selma-BSR).Some interviewees call for a cautious and honest approach when discussing this with non-expert citizens: […] some recognition that we may not have all the answers and, an honest evaluation of the risks […] being open about those risks.[…] I think there are going to be lots of shades of grey and so that's why I think we need to have some humility in terms of our own abilities and our own understanding of biology in general.[…] Share both the facts and the uncertainties with the broader public.(Megan-ASR) Educating the public to tackle misconceptions, hype and fear of GE is essential in researchers' general opinion: The more important thing is that you educate people, because the [official issue] must not only come from experts because obviously experts tend to have "bigger limits", so [probably] experts take the technology to an extreme [somehow].We should educate people to know about it and to understand [and to] also [have] judgment about it.Judgment based on knowledge, not judgment based on belief.So I think it's important that you can educate people because at the end not all the laws are made by the [scientists].(Paul-BSR) Overall, scientists highlight the need to equip people with the necessary scientific understanding through dialogue in order for them to be able to participate in decision-making processes upfront.Language shapes this dialogue, as has been shown recently (Hartley et al. 2023).How technologies, governance and society are intertwined within responsible science and innovation and how anticipation is a key aspect of technology assessment are the main lines of how the interviewed experts discuss GE as an emergent technology.The safety of applications is transversal to all the research questions and the aggregating theme identified in this study.OTEs are the most worrying technical challenge New Genetics and Society 13 for safety and interviewees consider crucial to develop methods to detect OTEs before clinical application.Literature highlights the need to avoid OTEs due to the critical consequences it may have for genome integrity and regulation (Cathomen et al. 2019).The difficulty of setting a "safe enough" standard has been a challenge since CRISPR discovery (EASAC 2017;NASEM 2017).Interviewees also consider that lack of genome knowledge increases the risk for undesired effects in human cells and that inefficiency in delivering CRISPR-Cas components may prevent a desired systemic effect of gene correction.These researchers believe that technology development will solve the issues of efficiency of CRISPR-Cas correction and fidelity of target DNA.Their belief is likely shaped by recent developments in optimization of genome editing repair pathways and the appearance of revolutionary editing strategies (Anzalone et al. 2019).Our study is concordant with a recent survey to biomedical researchers, which identified off-target mutations, efficiency levels for therapy and difficulties in targeting specific tissues in vivo as likely obstacles for the use of GE technologies in humans (Rocha, Braga, and Mota 2020).

Discussion
Food derived from GE crops, in vitro GE research and GE for human therapy emerge in the interviews as desirable applications with no cause for concern.As regards human therapy, our interviewees' view that this will be hindered by public acceptance is in contrast with the evidence of high acceptance of GE for therapy purposes in some public surveys (75-90%) (Delhove et al. 2020) and citizens' vision of human GE therapy as highly likely in the next 20 years (So, Sladek, and Joly 2021).Such divergences between experts and non-experts might relate to the culture of risk perception, where scientists base their views on the scientific method whereas the general public has a more intuitive understanding that transcends science and includes societal, political, regulatory or ethical factors (Gaskell et al. 2007).
Some interviewees are worried over unequal access to GE therapies, including concern that this may lead to edited and non-edited classes of individuals.This social justice argument has been further developed by scholars who see that differentiated access to GE therapies may lead to aggravation of discrepancies and marginalization of vulnerable citizens (Halpern et al. 2019;Hildebrandt and Marron 2018) as well as to irresponsible medical tourism only affordable to highincome citizens traveling to places where GE regulation is more relaxed (Meagher et al. 2020;Rosemann et al. 2019).Our interviewees generally consider GGE undesirable and worrying, but they also think it is unlikely and even less realistic for enhancement purposes.The interviewees emphasize that although technically easier to perform than somatic, germline editing is unlikely to satisfy safety considerations regarding physical, psychological and societal consequences for future generations.Again this is concordant with the literature where scholars explore reasons against GGE based on safety, responsibility, accountability and potential threats posed to future generations (de Wert et al. 2018;Howard et al. 2018).Nevertheless, interviewees generally acknowledge that a certain level of risk must be accepted for GE either in somatic and germline, and that it can be estimated based on preclinical animal testing which also has been suggested in the literature (Matthews and Iltis 2019;Polcz and Lewis 2016).Interviewees also stress that GGE should be avoided where alternatives such as pre-implantation diagnosis exist, considering it only acceptable for cases of life-threatening diseases and where these alternatives are not enough.Autosomal dominant and polygenic diseases have been identified as cases where there is no alternative to GGE (Gyngell, Douglas, and Savulescu 2017).Once again, scientists and general public perception seems to be discrepant since GGE for medical reasons earns considerably high public support for prevention of diseases of multiple levels of severity (Jedwab et al. 2020).
Overall, the interviewees welcome GE regulation and highlight the need to establish criteria and define crucial concepts like disease, susceptibility, enhancement and prevention.International health agencies defend that defining disease is important when regulating therapy and disease prevention (WHO 2021).A recurrent issue in the interviews is the lack of harmonization in how GE is regulated.It is known that GE regulation and legislation vary worldwide (Baylis et al. 2020) with a risk of safety and responsibility being dismissed in countries with less oversight (Boggio et al. 2019).Public opinion will influence regulation, and participants of the study emphasize the need to inform and educate the public to avoid misinformation and allow a rational discussion of GE applications.
Creative and effective forms of engagement between scientific experts, decision-makers and the public are necessary and may create an opportunity for newer participatory mechanisms of citizens with real repercussions for their own lives under democratic governance of science (Jasanoff 2003).Interviewees acknowledge that ethics, religion and regulation differ between countries and these cultural and normative differences affect attitude towards gene editing but their view is still distant from a culture of "civic science" (Wirz, Scheufele, and Brossard 2020) based on an early engagement of the public.However, when interviewees speak about the need for education, this is closer to a dialogue model that also accounts for cultural and experimental knowledge, because focus is on making citizens knowledgeable to participate in decision-making, contrasting with a deficit model where scientific expertise is the dominant knowledge (Reincke, Bredenoord, and van Mil 2020).Whereas the view that further development of the technology will overcome many of the present problems predominates, some interviewees also highlight the importance of recognizing our limited knowledge, in a way that comes closer to the "technologies of humility" concept (Jasanoff 2003).Contrary to deference to scientific authority, trusting science and its actors is a precursor of a democratic view (Howell et al. 2020) and this is a crucial factor influencing society questions about safety (Braun and Meacham 2019).The view that a broad stakeholder engagement with an informed public will be more useful reflects a culture of global public engagement for decisionmaking about GE (Jasanoff and Hurlbut 2018).The anticipation of scenarios in a collective manner builds on inclusive deliberation embedded in democratic principles and supposes a process of discussion and negotiation between actors (Owen, New Genetics and Society 15 Macnaghten, and Stilgoe 2012) and participants in this study recognize that this is crucial for the emergent character of GE technology.
Limitations and strengths of the study Our group of participants displays a mostly European, biomedical and/or animal science profile.This is a consequence of the non-random sampling method used and we recognize that other profiles of researchers could be sought for these interviews (e.g.medical or clinical background).Nonetheless, we achieved enough sample breadth and depth.
Finally, it is worth stressing that half of the interviews took place right after the He Jiankui's case.This might have influenced researchers' view and therefore we note a possible pre-and post-interpretation of the "birth of an in vitro genomeedited human embryo" ranking case (Supplementary material).

Conclusions and future perspectives
In summary, our study shows that for researchers in the field, GE technology is viewed as promising but with many unsolved challenges, in need of further research and of internationally harmonized regulation.They call for technology improvement and increased knowledge about genome function.They clearly do not support what they perceive as unsafe, unpredictable and irrelevant applications, but there are some discrepancies with the view of non-scientist citizens.Whereas public misconceptions are seen as a problem by the interviewees, there is also recognition that fostering a clear science-society dialogue is needed, and that science does not have all the answers.
This study with scientists that work daily with the technology provides insight into the science-based priorities for GE to be a technology that can be responsibly applied in humans.Contextualizing this study within the RRI framework, we note that a slice of the research community is represented and additional stakeholders are needed to confer inclusivity in order to better anticipate unintended consequences of resulting innovations.We are complementing the present study by interviewing ethicists and policy-makers, with focus on governance aspects of this technology.

Figure 1 .
Figure 1.Technology as an overarching theme comprising multiple challenges that may influence safety.The different issues around technology brought up by the interviewees, and how these issues relate to each other are illustrated.Delivery of genome editing components such as CRISPR-Cas9 depends on different factors and different type of cells may be influenced by the efficacy of gene correction.Together, these influence efficiency of technology which needs to be improved by different basic research methods.Fidelity and control of DNA target integration and specificity of Cas9 cutting influence the appearance of off-target effects which need to be detected and monitored.Genome complexity in combination with these technical issues determine if the technology is to be considered safe enough for application in vivo.The arrows indicate relationships between technical aspects.
[…]if you[do]  knock-in or knock-out [in] cell lines maybe the only phenotype you see is only a small part you can get from cells but you do not know what happens for long times growing for this kind of cells and organs.(Dean-BSR)

Figure 2 .
Figure 2 .Governance and society are overarching themes comprising six subthemes with ethical and societal dimensions.Helping many people is the most important purpose identified over a set of applications that range from purposeful to irrelevant.Cultural differences involving ethics and religion may shape public attitudes towards the technology and together these may influence regulation worldwide which matters for governance.Public attitude concerns citizens' perception and acceptance of the technology which involve mainly public education, misconceptions and science-society dialogue features.This brings issues of social justice to the table particularly related with discrimination of individuals based on socio-economic status.The arrows indicate how issues influence each other and the colors of arrows and boxes indicate how subthemes are interrelated.
Our analysis was structured into research questions related to the technology itself (challenges and expected realistic developments of genome editing technology) and its application (what applications are likely, what are desirable and what are concerning).