Current capabilities for human genome editing: report for the WHO expert advisory committee on developing global standards for governance and oversight of human genome editing

Current capabilities for human genome editing: report for the WHO Expert Advisory Committee on Developing Global Standards for Governance and Oversight of Human Genome Editing Güneş Taylor and Christophe Galichet April 2021 WHO/SCI/RFH/2021.06 © World Health Organization 2021 Some rights reserved. This work is available under the Creative Commons Attribution- NonCommercial-ShareAlike 3.0 IGO licence (CC BY-NC-SA 3.0 IGO; https://creativecommons.org/licenses/by-nc-sa/3.0/igo). Under the terms of this licence, you may copy, redistribute and adapt the work for non-commercial purposes, provided the work is appropriately cited, as indicated below. In any use of this work, there should be no suggestion that WHO endorses any specific organization, products or services. The use of the WHO logo is not permitted. If you adapt the work, then you must license your work under the same or equivalent Creative Commons licence. If you create a translation of this work, you should add the following disclaimer along with the suggested citation: “This translation was not created by the World Health Organization (WHO). WHO is not responsible for the content or accuracy of this translation. The original English edition shall be the binding and authentic edition”. Any mediation relating to disputes arising under the licence shall be conducted in accordance with the mediation rules of the World Intellectual Property Organization (http://www.wipo.int/amc/en/mediation/rules/). Suggested citation. Taylor G, Galichet C. Current capabilities for human genome editing: report for the WHO Expert Advisory Committee on Developing Global Standards for Governance and Oversight of Human Genome Editing, April 2021. Geneva: World Health Organization; 2021 (WHO/SCI/RFH/2021.06). Licence: CC BY-NC-SA 3.0 IGO. Cataloguing-in-Publication (CIP) data. CIP data are available at http://apps.who.int/iris. Sales, rights and licensing. To purchase WHO publications, see http://apps.who.int/bookorders. To submit requests for commercial use and queries on rights and licensing, see www.who.int/copyright. Third-party materials. If you wish to reuse material from this work that is attributed to a third party, such as tables, figures or images, it is your responsibility to determine whether permission is needed for that reuse and to obtain permission from the copyright holder. The risk of claims resulting from infringement of any third-party-owned component in the work rests solely with the user. General disclaimers. The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part of WHO concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted and dashed lines on maps represent approximate border lines for which there may not yet be full agreement. The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or recommended by WHO in preference to others of a similar nature that are not mentioned. Errors and omissions excepted, the names of proprietary products are distinguished by initial capital letters. All reasonable precautions have been taken by WHO to verify the information contained in this publication. However, the published material is being distributed without warranty of any kind, either expressed or implied. The responsibility for the interpretation and use of the material lies with the reader. In no event shall WHO be liable for damages arising from its use. The named authors alone are responsible for the views expressed in this publication. iii Contents Abbreviations ............................................................................................................................ iv Context ....................................................................................................................................... 1 1. Introduction to genome editing ........................................................................................... 1 2. Improving NHEJ and HDR outcomes ................................................................................ 3 3. Base and prime editing ....................................................................................................... 4 4. Precision editing of mitochondrial DNA ............................................................................ 6 5. Transcriptional modulators and epigenetic editing ............................................................ 6 6. Improving the sensitivity of ways to assess the outcome of genome editing ..................... 7 7. Alternative routes for heritable human genome editing using stem cells ........................... 8 8. Prenatal (in utero) somatic gene therapy or heritable genome editing ............................... 9 9. Conclusion and future scoping ......................................................................................... 10 References ................................................................................................................................ 11 iv Abbreviations ABE Adenine Base Editor CBE Cytidine Base Editor CRISPR Clustered Regularly Interspaced Short Palindromic Repeats dCas9 nuclease-dead Cas9 DSB Double-Strand Break GONAD Genome editing via Oviductal Nucleic Acid Delivery gRNA guide RNA HDR Homology-Directed Repair indel insertion or deletion MMEJ Microhomology-Mediated End Joining mtDNA mitochondrial DNA NHEJ Non-Homologous End Joining pegRNA prime editing guide RNA PGC Primordial Germ Cell REDIT RecT Editor via Designer-Cas9-Initiated Targeting RS-1 RAD51-stimulatory compound-1 SCGE Somatic Cell Genome Editing TALEN Transcription Activator-Like Effector Nuclease ZFN Zinc Finger Nuclease 1 Context To develop enduring global standards for governance and oversight of human genome editing, it is necessary both to define the genome editing technologies currently available and to appreciate the trajectory of the research underpinning them. The unprecedented rate of progress in this field necessitates regular revisions to benchmark reports such as that of the International Commission on the Clinical Use of Human Germline Genome Editing in 2020, and as such, this report focuses particularly on recent developments and highlights areas of current research efforts. At the outset, it must be stressed that despite all that has been learned in the last two decades about human genes, genomes and genetic variation, there are still many gaps in our knowledge. Multiple genes and “environmental” effects can influence the incidence and severity of many human disorders with an underlying genetic cause. Genetic background effects are well known in animal models, and also apply to humans. These can reflect one, several, or many gene variants that can map anywhere within the genome or be closely linked to the main causative “mutant” allele. Many genetic variants identified in humans have an unknown impact on phenotype - some will be in protein coding regions, others in RNA products of genes, and many will be in non-transcribed regions, where they could affect gene expression of single or even multiple genes. Even for diseases that are commonly referred to as monogenic, diseases attributable to the malfunction of a single gene, sound evidence for a causative role of a specific genetic variant would be needed prior to genome editing. This can come from family studies or population genetics, but ideally it requires knowledge of whole genomes, not just of the specific gene in question. Environmental effects can be due to the influence of nutrition in utero or postnatally, gut and other microbiota, exposure to pathogens and other harmful substances such as pharmaceuticals, and so on. It follows that, without further research designed to explore these variables, it might be hard to predict the exact outcome in attempts at human genome editing. This will be true for somatic as well as heritable applications. 1. Introduction to genome editing Genome editing encompasses a number of techniques to modify DNA within the genome of a targeted cell, thereby altering the information encoded at the target site. These techniques include the use of meganucleases, zinc finger nucleases (ZFNs), transcription activator- like effector nucleases (TALENs) and, most recently, CRISPR-Cas9.1 Each of these systems employs a different repertoire of molecules to alter DNA, and thus has different properties, which need to be carefully considered in any clinical research or applications. Genome editing technologies traditionally rely upon the generation, and subsequent repair, of double-strand breaks (DSBs) of DNA at user-defined target sites. It is this ability to 1 CRISPR: clustered regularly interspaced short palindromic repeats. 2 programme user-defined targets, in combination with improved DNA sequencing methods giving both reference and patient-specific genome sequences, that sets modern genome editing apart from previous efforts to alter genomes. DSBs are predominantly repaired via an error-prone mechanism known as non-homologous end joining (NHEJ), which regularly results in small insertions or deletions (indels). These indels can cause significant changes to the information encoded at the target site, thereby disruptively “editing” the underlying genome. However, under certain circumstances a higher-fidelity mechanism for DNA repair within cells called homology-directed repair (HDR) can be employed. HDR occurs with less frequency as it requires the presence of a repair template with high sequence similarity to the regions flanking the target site. Thus, if a donor DNA template carrying a specific edit is introduced together with the genome editing components, cells are able to repair DSBs at target sites via HDR, thus introducing the desired template sequence into the genome. Cells can also repair DNA DSBs by microhomology-mediated end joining (MMEJ), which makes use of homologies of just a few base pairs and leads to usually short (but sometimes quite extensive) deletions. Another cell-based repair mechanism that can be triggered during HDR uses the DNA sequence from the other chromosome (instead of the DNA template provided), leading to “gene conversion”. This can lead to repair of the mutant allele in a heterozygote, or if this happens after the allele has been repaired, in a homozygote. However, it can also spread over long distances and lead to loss of heterozygosity in adjacent regions of the chromosome, thus resulting in large stretches of identical DNA sequences in the chromosome pair. This could have significant consequences: for example, if the individual was heterozygous for a recessive mutation in a gene adjacent to the one being targeted, gene conversion could result in this now becoming homozygous, for either the normal or the disease-causing variant, the latter leading to a new, perhaps unexpected, phenotype. Of note, all genome editing events occur independently within targeted cells, whether by NHEJ or HDR. Consequently, screening for successful editing outcomes is vitally important, especially in any potential clinical context. While all genome editing platforms have been used extensively to model and understand human conditions in laboratory animals (1), in recent years, CRISPR-mediated genome editing has become the focus of most efforts to develop potential therapeutic tools for humans. Unlike ZFNs and TALENs, which are proteins engineered to have both a nuclease, usually a FokI domain, and a part made up of individual modules that each recognize specific DNA bases, CRISPR is a two-part molecular system, comprising a CRISPR guide RNA (gRNA) and a Cas9 DNA-cleaving enzyme that interact with each other. Partly because it is simple to generate specific gRNAs, the latter offers superior speed, efficiency and programmability compared to the other genome editing technologies. Since the publications first demonstrating the power of CRISPR genome editing in bacteria (2) and subsequently in mammals (3, 4), the scientific community has spent enormous efforts to further refine its effectiveness and develop a sizable repertoire of other CRISPR-based tools to address specific needs. For example, most pathogenic mutations found in the human genome must be corrected rather than simply disrupted in order to benefit patients, meaning that the rare HDR events are required rather than the disruptive NHEJ. Unfortunately, the 3 inefficiency of HDR in most cell types and the challenges associated with donor DNA template delivery have limited the therapeutic relevance of HDR to date. Consequently, technical improvements in HDR rate, component delivery and specific single nucleotide alterations have particular clinical significance, and most clinical trials using genome editing focus on cases where gene disruption by indel generation is therapeutic. Despite being the most recent innovation in the genome editing toolkit, CRISPR has already been used in the clinic. In a landmark case in February 2020, two patients with refractory myelomas and one with metastatic sarcoma had their own immune cells edited in vitro to allow these to better recognize and eliminate cancer cells upon reintroduction, with positive initial results reported (5). Below are summary overviews of current and contemplated uses of genome editing technologies in humans. However, these cannot be all inclusive, given the rapid rate of development in the field. Moreover, methods that would be very challenging to engineer and are likely to raise significant ethical and societal objections, such as gene drives in humans (which are being developed to manage important disease vectors such as mosquitoes and invasive species), are not included. 2. Improving NHEJ and HDR outcomes As indicated in the Stadtmauer et al. report (5), when in vitro screening of the CRISPR- targeted cells is possible, the error-prone nature of NHEJ can be mitigated by growing colonies (clones) from single cells, screening these and selecting only those for further use that have their genome edited correctly. However, not all tissues or cell types are as amenable to screening as immune cells; thus efforts are being directed to improving NHEJ outcome predictability. Recent work has demonstrated that predictable and precise NHEJ-mediated deletions are possible using paired Cas9 gRNAs to generate adjacent DSBs (6). While off- target prediction precision has also improved, the generation of DSBs at non-target sites and consequently the possibility of other NHEJ events occurring still needs to be considered. Minimizing the time that the Cas9 is active within a cell, for example, by using short-lived versions of the protein, can help reduce off-target events, as can using specific variants of either the nuclease or the gRNAs that favour their activity at the cognate target DNA sequence rather than those that differ by one or two base pairs. Unlike HDR, even with improvements to outcome prediction, NHEJ-edited alterations are ultimately not user defined. However, HDR is usually favoured less frequently than NHEJ within cells and requires the presence of a donor sequence template carrying the intended replacement information. Advances have been made in the use of chemical inhibitors of key components of the NHEJ and microhomology-mediated end joining DNA damage repair pathways to shift the balance in favour of HDR. The use of SCR7, an inhibitor of DNA ligase IV involved in the NHEJ pathway, permits an increased incidence of HDR-mediated repair in human cell lines (7). Furthermore, introduction of RAD51, an essential protein for HDR, has also been shown to favour HDR occurrence. Provision of RAD51-coated oligonucleotides has also been used to favour integration of repair templates, as has introduction of RS-1 (RAD51-stimulatory compound-1), a compound known to stabilize the association of RAD51 4 to the DNA, into cells (8). In human fibroblast and induced pluripotent stem cells, direct fusions of Cas9 to a key factor for DSB resection and HDR initiation, CtIP, focuses its activity on the target site, thereby enhancing HDR (9). Competition between NHEJ and HDR is cell cycle dependent, with HDR being restricted to late S and G2 phases (10). Consequently, synchronizing cell cycle phases has also been shown to help increase the rate of HDR (11). It is worth noting that HDR repair only occurs in cells that are dividing (have an active cell cycle), thus this method cannot be used in differentiated cells that no longer divide (they are “post-mitotic”), such as neurons. Aside from shifting the balance from NHEJ to HDR, the availability of all necessary components at the target site is an important factor for determining high editing efficiency. Accordingly, multiple methods to maximize the presence of all components have been developed. One such method, REDIT,2 features a modified gRNA containing MS2 stem loop scaffolds, which, upon binding of the MS2 coat protein, bring the DNA donor template to the cleavage site (12). Another approach is the use of streptavidin-tethered Cas9 enzyme, which can be used in combination with repair templates containing biotinylated oligonucleotides, to favour incorporation of the template at the CRISPR on-target site (13). Furthermore, as reported in a 2019 study, two gRNAs in conjunction with a long single-stranded DNA template can dramatically improve large and complex gene edits in early mouse embryos (14). In light of these developments, we predict that combinations of multiple methods to further improve HDR rates will be increasingly employed. 3. Base and prime editing Many disease-associated human genetic variants are attributable to point mutations affecting single DNA base pairs, such as the transition mutations in the ageing disease progeria or transversion events leading to sickle cell anaemia, or small indels, such as the four base pair insertion underlying the majority of cases of Tay-Sachs disease. New tools allowing modification without DNA cleavage or DSB formation have also been gaining increased attention because of concerns regarding the possibility of large genomic rearrangements subsequent to DSB formation. Base editors were developed to circumvent the issues of random indels from NHEJ and ineffective HDR rates in CRISPR-mediated genomic editing, which hinder their effective use in correcting specific and small mutations. Base editors fuse enzymes with particular DNA-altering properties to Cas9, thereby focusing their action on CRISPR-targeted sites. The Cas9 used in base editing is usually a nickase (making a single strand cut in DNA) (15) or it is catalytically dead (nuclease-dead Cas9, or dCas9) (16), thereby avoiding the DSBs and DNA damage responses induced by classical Cas9. Furthermore, no donor templates are required, reducing the number of components that have to act together. Base editors have demonstrated efficiency in many contexts, including post- mitotic cells in vitro or in vivo, and in both early mouse and human embryos. 2 RecT Editor via Designer-Cas9-Initiated Targeting. 5 Cytidine base editors (CBEs) are comprised of a catalytically impaired nickase Cas9 tethered to a cytidine deaminase, which act together with an uracil glycosylase inhibitor to make C-to- T or G-to-A conversions or transition mutations (17). Subsequently, adenine base editors (ABEs) were developed using laboratory evolution to permit A-to-G or T-to-C transversions by a novel enzyme that deaminates deoxyadenosine (18). As with all genome editing tools, variations of ABEs have already been generated, with one in vitro-evolved variant known as ABE8e demonstrating 1100-fold faster activity than earlier variants (19). RNA base editors have also been developed to alter RNA transcripts rather than DNA (20), and these have already been used to correct a specific cystic fibrosis point mutation in vivo (21). However, DNA base editors have been shown to have off-target activity on RNA (22, 23). Efforts to reduce this erroneous activity report delivering Cas9 ribonucleoprotein directly (24), delivering base editor messenger RNA instead of plasmid RNA (25), or using lentivirus-like particles (26) to reduce the incidence of off-target events. While there are many forms of potential off-target edits, researchers have both honed the physical base editor components and employed computational methods to select higher-fidelity routes to editing. For example, using machine learning models trained on data from 11 different CBEs and ABEs, a new computational resource called BE-Hive has been developed to accurately predict base editing efficiency and product distributions for the most commonly used base editors. As an extension of base editing, prime editing has recently been developed for mammalian cell genome editing. Like base editors, prime editors are fusions of Cas9 nickase with an active partner, but in this case a reverse transcriptase. They also utilize an engineered prime editing guide RNA (pegRNA), which both specifies the target site and encodes the intended edit. This combination of nickase, reverse transcriptase and pegRNA means virtually any small indel or substitution can be achieved using this strategy in mammalian cells (27). It is speculated that the rate of off-target events with this technique is limited due to the number of steps required to resolve an edit. Compared to Cas9-mediated HDR, prime editing appears to offer comparable or higher efficiencies with a reduced incidence of random indels without the need for additional donor templates due to the pegRNA. Although the alterations made are limited in size, the flexibility and apparent efficiency of this system makes it potentially very valuable. Base editing and prime editing offer complementary strengths, though base editors currently offer higher editing efficiencies and lower indel levels than prime editors at certain sites. However, when the target nucleotides are not ideally positioned for base editing, prime editors typically offer higher efficiency. At present, while base editors offer fewer indels and higher efficiency, prime editors excel in conditions requiring greater precision, targeting scope and versatility. Prime editing is being used to correct a variety of pathogenic mutations in cell lines, from correcting sickle cell mutations to corrective four base insertions for Tay-Sachs disease. Base and prime editing offer significant hope for future clinical advances and have exciting implications for biotechnological applications. 6 4. Precision editing of mitochondrial DNA Mutations affecting mitochondrial function have significant ramifications for the patients carrying them and are heritable. Approximately 1 in 5000 people are estimated to carry pathogenic mutations in either nuclear DNA or mitochondrial DNA (mtDNA) that result in disease or significantly increased disease risk in the future (28). Transition point mutations constitute the vast majority of pathogenic mtDNA events and therefore could in principle be corrected by CBEs or ABEs. However, while proteins can be tagged in a way that allows them to shuttle into mitochondria, this cannot be done for CRISPR gRNA components, necessitating the development of new methods to target mtDNA within the mitochondrion. A CRISPR-free method using DddA, an enzyme that converts C-to-T on double-stranded DNA without the need of DSBs, has recently been developed to circumvent this. The ability of DddA to deaminate double-stranded DNA is an important innovation, as previous CBEs and ABEs deaminate single-stranded DNA generated by repair mechanisms subsequent to Cas9- mediated DSBs. To reduce the toxicity of this DddA, split DddAs were linked to separate TALENs, meaning that DddA activity would only be present when brought together, avoiding a change of all Cs into Ts, thereby also utilizing another genome editing technology to provide sequence specificity (29). In mitochondria, this so-called DdCBE complex (two transcription activator-like effector repeats fused to half of DddA respectively and an uracil glycosylase inhibitor) edits mtDNA with high on-target editing to off-target editing and no detectable indels. 5. Transcriptional modulators and epigenetic editing Aside from direct modification of DNA, which is required for truly heritable changes, it is possible to modulate the other layers of information that connect DNA sequence to gene activity. Ultimately it is aberrant gene function that causes a pathology, and this can be at the level of an altered protein, DNA sequence or gene activity. Thus, tuning of gene outputs rather than direct editing of the genes themselves has become a research focus, since the advent of programmable genome editing technologies made site-specific targeting possible. dCas9 can be fused to gene activators (30), repressors (31), or modifiers of epigenetic marks – histone: acetylation (32), demethylation (33); cytosine: demethylation (34), methylation (35). These tools encompass direct transcriptional activators and repressors, but also complexes able to fine-tune chromatin landscapes, the contours of which ultimately regulate information flow through the genome. These chromatin landscapes contain so called epigenetic information vital for correct development and maintenance of cell fate. For example, dCas9-KRAB fusions are able to decommission regions that enhance gene expression by causing chromatin compaction, which then renders it inaccessible to transcriptional activators (36, 37). Our ever-increasing knowledge of the genome, and of such enhancing or repressing regions, advance the possibility of sophisticated solutions to diseases, especially where these can be attributed to subtle changes in activity of specific genes or even networks of genes. 7 6. Improving the sensitivity of ways to assess the outcome of genome editing As previously mentioned, a major challenge to genome editing, specifically in the human context, is that of delivery and the subsequent monitoring of editing outcomes. This applies to CRISPR-NHEJ/HDR and also to the use of base and prime editors. While in vitro somatic editing involving many cells can provide plenty of material for assessment of editing outcomes, in early stage embryos this is significantly hampered. The success of a single editing event is best achieved in early stage embryos, as any edits propagate through the entire organism. Therefore, methods allowing detection of off-target events in early embryos, with a limited number of cells available, must be refined and made more sensitive. Recent advances permit genome-wide analysis to be performed on single cells, thereby enabling the detection of off-target events (38). However, the depth of sequencing (number of times each, usually amplified, segment of DNA is read) needed for obtaining sufficiently accurate sequence data may still be an issue for certain sequencing platforms. Recent evidence demonstrating that Cas9-induced DSBs can lead to large chromosomal abnormalities in 16% of cells in genome-edited human embryos is particularly sobering (39, 40). This Cas9-induced large chromosomal abnormality rate is reproducible and comparable to that in other cellular contexts (41-46). Improving methods to screen out, reduce, or ideally eliminate these aberrant events may be feasible; alternatively, methods that avoid DSB generation entirely ought to be employed. From gRNA generation to evaluating the action and outputs of CRISPR-mediated genome editing, in silico analysis of genome editing outcomes is a rapidly expanding and necessary field for development. Websites and software have been created to allow the generation of specific gRNA that avoids or minimizes off- target events, often utilizing a scoring system to rank potential off-target sites according to number of mismatches. The potential for mosaicism (where not all cells have the same genetic alteration), off-target effects and multiple possible genetic outcomes at the on-target site in any given genome editing experiment are difficult to identify, distinguish and ultimately quality control for. This challenge must be addressed before widespread and versatile therapeutic advances can be made. When using optimized methods, the frequency of off-target events can be so low that they are difficult to find. Indeed, they will be absent in the large majority of cells, which can be apparent when single somatic cells are expanded clonally in vitro after genome editing, or when analysing an animal that has developed from a single-celled zygote that had been subject to genome editing. However, off-target events can become a problem with in vivo somatic genome editing where the editing has to take place in many millions of cells, and it will be a challenge for any assessment method to find rare cells with off-target edits. It then becomes essential to have sufficient in silico and preclinical data to make it very unlikely that they will not have a deleterious effect. Such an event would happen, for example, if a tumour suppressor gene were inactivated. It would be critical to avoid any inappropriate on-target or off-target events in any attempt at making a heritable change. If the genome editing is carried out in zygotes or very early 8 embryos, then the chances of these may be low, and refined assessment methods could in theory be used to screen for embryos that only have the desired edits. However, mosaicism, which can result from delayed activity of the genome editing components even if these are introduced at the one cell stage, becomes a significant problem, because any biopsy of a few cells taken from the preimplantation embryo may not be representative of the remainder. Methods such as base editing, where a high degree of fidelity and efficiency might be achievable, could provide a solution to this problem. Ethical, societal and regulatory issues would need to be addressed before such research is planned. Of note, research to improve accuracy and efficiency of genome editing techniques in early embryos may require a large number of human embryos. It may also involve the creation of research embryos, which is currently only legally permitted in a small number of countries. 7. Alternative routes for heritable human genome editing using stem cells An alternative route to heritable human genome editing would involve editing stem cell precursors of gametes in such a way that these can be expanded clonally and carefully screened prior to the generation of gametes and embryos. Experiments involving the use of CRISPR in primordial germ cells (PGCs), including in vitro genetic modification prior to implantation into germ cell-free gonads, have been successful in some animal models, such as birds (47). In addition, recent advances in deriving PGC-like cells from human induced pluripotent stem cells (48) and differentiating these towards oocytes or spermatogonia in culture could provide a future route to heritable genome editing. Increasing work, including that of Kyle Orwig and a number of others, focuses on genome editing of spermatogonial stem cells, which can then be used to derive functional sperm after being reintroduced into testes (preferably treated to deplete the endogenous spermatogonial stem cells). This in vitro editing and subsequent grafting approach has already been used in pigs (49), mice (50-53), rats (54) and macaques (55-57). Genetic depletion of endogenous spermatogonial stem cells has also been employed (58). The feasibility of these culture systems for humans remains to be proven, but with advanced studies into the evolutionary conservation of mammalian spermatogenesis (59), the relevant research appears to be progressing rapidly (60). Genome editing methods involving the use of gamete precursor cells and primordial germ cells introduce novel safety and ethical issues, which will need to be examined separately from those of genome editing in somatic and germ cells. These methods will involve new and, in a sense, artificial methods of reproduction using gametes from cultured cells. Research on ways to obtain in vitro-derived gametes is progressing at pace in order to improve basic understanding of how these interesting and important cells develop. If successful, such research could provide essentially unlimited numbers of oocytes and early embryos for other types of research, including on heritable genome editing. 9 8. Prenatal (in utero) somatic gene therapy or heritable genome editing As researchers refine their control over genome editing technologies for uniform results in complex postnatal and adult tissues, parallel efforts are being made to advance the application of genome editing to fetuses or even embryos in utero. Fetal surgery was first developed during the 1980s to deal with fatal or severe anatomic diseases such as spina bifida. Techniques leveraging structures such as the umbilical vein to deliver systemic therapies are in place. Advances in early detection of conditions in utero allow the possibility to treat fetuses prior to the onset of advanced clinical signs and further tissue growth. Treatments could be surgical, genetic, or via stem cell transplantation, or could adopt a mixture of approaches. Genome editing signals the advent of in utero gene therapy and can make use of the same broad list of methods as for postnatal somatic genome editing. However, the delivery of the genome editing components or of edited genetic material (for example, when combining genome editing with stem cell treatments) into the developing embryo or fetus remains a challenge. Moreover, while the risk of unintended edits to the germ cells in fetal testes or ovaries is probably low, this possibility will need to be explored if heritable changes are to be avoided. Alternatively, if this is the desired outcome, then specific methods (such as use of viral vectors “pseudotyped” to target germ cells) will need to be developed. Developments in this area hold significant hope for disorders, especially severe single gene disorders, from muscular dystrophy to thalassaemia and haemophilia. Phase 1 clinical trials for stem cell transplantation, haematopoietic stem cells for thalassaemia, and enzyme replacement therapy for lysosomal storage disorders using genome editing are under way and should have improved outcomes compared to postnatal interventions. A number of consensus statements on this field have been published by the International Fetal Transplantation and Immunology Society. In mice and rats, methods allowing editing of the genome without the need for ex vivo embryo manipulation have been developed, the first being termed GONAD (genome editing via oviductal nucleic acid delivery) (61, 62). The GONAD method delivers genome editing components directly to the oviduct of early pregnant dams using a glass micropipette. Subsequent developments improving this method, resulting in higher efficiency of editing, are termed i-GONAD (63-66). Model-specific variants of i-GONAD have recently been developed to utilize this approach in other organisms (67-69). While GONAD and i-GONAD utilize injection or electroporation as the method of delivering genome editing components into the embryo, other methods using viruses, such as adeno-associated viruses, have also been developed and used efficiently in pre-implantation mouse embryos (70). Notably, each of these techniques currently carries either a high risk of embryo or fetal loss or low specificity of outcomes (71). Integration of multiple techniques, such as tethering of gRNAs to Cas9 in an i-GONAD approach, may be an avenue for improved control over outcomes in this or other in utero contexts for somatic or heritable genome editing. 10 9. Conclusion and future scoping Genome editing technologies represent a rapidly expanding toolkit of molecular systems to alter DNA, RNA and chromatin landscapes. They can induce random mutations to disrupt genes, be used to introduce deliberate changes with increasingly fine control, and potentially even reshape transcriptional landscapes to modulate clinically implicated gene networks. As our understanding of the genome deepens, these tools are being modified for uses in more and more specialized contexts to gain the intended, potentially therapeutic, outcomes. A recent report by Saha et al. (72) on the United States National Institutes of Health Somatic Cell Genome Editing (SCGE) programme, which also gives a current overview of methods, describes the establishment of a toolkit to “compile new genome editors, delivery technologies and methods for tracking edited cells in vivo, as well as newly developed animal models and human biological systems, along with validated datasets”. This SCGE toolkit3 has been developed specifically with somatic genome editing in mind, and while there may be some overlap, it is unlikely to be generally applicable to heritable genome editing. One area that needs immediate attention is the method of delivery of genome editing components to the target cells and tissues. The components can be delivered directly by injection or electroporation (or similar methods) into target cells; indirectly, by introduction of in vitro “repaired” cells to the organism; or intrauterine, by transplacental means. Alternatively, viruses or nanoparticles can be used to introduce genome editing components into cells, both in vitro and in vivo. Other methods also exist, each with advantages and disadvantages in comparison with the other options (73). However, a reoccurring stumbling block for all applications of genome editing technologies, particularly those to be applied in vivo, is the inefficiency of the delivery methods currently available. While applications in early embryos bypass this issue to some extent, as they have a limited number of cells to begin with, the problem of being able to target only a small proportion of a specific cell type in vivo restricts the potential contributions of these technologies in adult clinical conditions. Investigations into smaller, hypercompact phage-encoded CRISPR alternatives are currently under way, but further efforts are required before this issue is resolved. Delivery of genome editing components should be regularly highlighted as requiring urgent innovation to help make clinical aspirations a more tenable reality – a particularly significant consideration for public dialogues surrounding the use of genome editing technologies in humans. Finally, the methods detailed here have not been judged for their suitability for use in clinical research or applications. All such decisions must involve review and oversight processes that can make risk–benefit judgements based not only on scientific and clinical evidence, but also on local ethical and societal views. 3 https://scge.mcw.edu/toolkit. 11 References 1. Galichet C, Lovell-Badge R. Applications of genome editing on laboratory animals. Laboratory Animals. 2021;23677221993141. doi:10.1177/0023677221993141. 2. Jinek M, Chylinski K, Fonfara I, Hauer M, Doudna JA, Charpentier E. A programmable dual-RNA-guided DNA endonuclease in adaptive bacterial immunity. Science. 2012;337(6096):816-21. 3. Wang H, Yang H, Shivalila CS, Dawlaty MM, Cheng AW, Zhang F, et al. One-step generation of mice carrying mutations in multiple genes by CRISPR/Cas-mediated genome engineering. Cell. 2013;153(4):910-8. 4. Cong L, Ran FA, Cox D, Lin S, Barretto R, Habib N, et al. Multiplex genome engineering using CRISPR/Cas systems. Science. 2013;339(6121):819-23. 5. Stadtmauer EA, Fraietta JA, Davis MM, Cohen AD, Weber KL, Lancaster E, et al. CRISPR-engineered T cells in patients with refractory cancer. Science. 2020;367(6481):eaba7365. 6. Guo T, Feng Y-L, Xiao J-J, Liu Q, Sun X-N, Xiang J-F, et al. Harnessing accurate non-homologous end joining for efficient precise deletion in CRISPR/Cas9-mediated genome editing. Genome Biology. 2018;19(1):170. 7. Maruyama T, Dougan SK, Truttmann MC, Bilate AM, Ingram JR, Ploegh HL. Increasing the efficiency of precise genome editing with CRISPR-Cas9 by inhibition of nonhomologous end joining. Nat Biotechnol. 2015;33(5):538-42. 8. Pinder J, Salsman J, Dellaire G. Nuclear domain 'knock-in' screen for the evaluation and identification of small molecule enhancers of CRISPR-based genome editing. Nucleic Acids Res. 2015;43(19):9379-92. 9. Charpentier M, Khedher AHY, Menoret S, Brion A, Lamribet K, Dardillac E, et al. CtIP fusion to Cas9 enhances transgene integration by homology-dependent repair. Nature Communications. 2018;9(1):1133. 10. Heyer WD, Ehmsen KT, Liu J. Regulation of homologous recombination in eukaryotes. Annu Rev Genet. 2010;44:113-39. 11. Lin S, Staahl BT, Alla RK, Doudna JA. Enhanced homology-directed human genome engineering by controlled timing of CRISPR/Cas9 delivery. Elife. 2014;3:e04766-e. 12. Wang C, Cheng JKW, Zhang Q, Hughes NW, Xia Q, Winslow MM, et al. Microbial single-strand annealing proteins enable CRISPR gene-editing tools with improved knock-in efficiencies and reduced off-target effects. Nucleic Acids Research. 2021. 13. Gu B, Posfai E, Rossant J. Efficient generation of targeted large insertions by microinjection into two-cell-stage mouse embryos. Nat Biotechnol. 2018;36(7):632-7. 12 14. Pineault KM, Novoa A, Lozovska A, Wellik DM, Mallo M. Two CRISPR/Cas9- mediated methods for targeting complex insertions, deletions, or replacements in mouse. MethodsX. 2019;6:2088-100. 15. Mali P, Yang L, Esvelt KM, Aach J, Guell M, DiCarlo JE, et al. RNA-Guided Human Genome Engineering via Cas9. Science. 2013;339(6121):823-6. 16. Qi LS, Larson MH, Gilbert LA, Doudna JA, Weissman JS, Arkin AP, et al. Repurposing CRISPR as an RNA-guided platform for sequence-specific control of gene expression. Cell. 2013;152(5):1173-83. 17. Komor AC, Kim YB, Packer MS, Zuris JA, Liu DR. Programmable editing of a target base in genomic DNA without double-stranded DNA cleavage. Nature. 2016;533(7603):420-4. 18. Gaudelli NM, Komor AC, Rees HA, Packer MS, Badran AH, Bryson DI, et al. Programmable base editing of A•T to G•C in genomic DNA without DNA cleavage. Nature. 2017;551(7681):464-71. 19. Lapinaite A, Knott GJ, Palumbo CM, Lin-Shiao E, Richter MF, Zhao KT, et al. DNA capture by a CRISPR-Cas9-guided adenine base editor. Science. 2020;369(6503):566-71. 20. Cox DBT, Gootenberg JS, Abudayyeh OO, Franklin B, Kellner MJ, Joung J, et al. RNA editing with CRISPR-Cas13. Science. 2017;358(6366):1019-27. 21. Montiel-Gonzalez MF, Vallecillo-Viejo I, Yudowski GA, Rosenthal JJ. Correction of mutations within the cystic fibrosis transmembrane conductance regulator by site- directed RNA editing. Proc Natl Acad Sci U S A. 2013;110(45):18285-90. 22. Grünewald J, Zhou R, Garcia SP, Iyer S, Lareau CA, Aryee MJ, et al. Transcriptome- wide off-target RNA editing induced by CRISPR-guided DNA base editors. Nature. 2019;569(7756):433-7. 23. Zhou C, Sun Y, Yan R, Liu Y, Zuo E, Gu C, et al. Off-target RNA mutation induced by DNA base editing and its elimination by mutagenesis. Nature. 2019;571(7764):275-8. 24. Richter MF, Zhao KT, Eton E, Lapinaite A, Newby GA, Thuronyi BW, et al. Phage- assisted evolution of an adenine base editor with improved Cas domain compatibility and activity. Nature Biotechnology. 2020;38(7):883-91. 25. Gaudelli NM, Lam DK, Rees HA, Solá-Esteves NM, Barrera LA, Born DA, et al. Directed evolution of adenine base editors with increased activity and therapeutic application. Nature Biotechnology. 2020;38(7):892-900. 26. Lyu P, Lu Z, Cho SI, Yadav M, Yoo KW, Atala A, et al. Adenine Base Editor Ribonucleoproteins Delivered by Lentivirus-Like Particles Show High On-Target Base Editing and Undetectable RNA Off-Target Activities. Crispr j. 2021;4(1):69-81. 13 27. Anzalone AV, Randolph PB, Davis JR, Sousa AA, Koblan LW, Levy JM, et al. Search-and-replace genome editing without double-strand breaks or donor DNA. Nature. 2019;576(7785):149-57. 28. Schaefer A, Lim A, Gorman G. Epidemiology of Mitochondrial Disease. 2019. p. 63- 79. 29. Mok BY, de Moraes MH, Zeng J, Bosch DE, Kotrys AV, Raguram A, et al. A bacterial cytidine deaminase toxin enables CRISPR-free mitochondrial base editing. Nature. 2020;583(7817):631-7. 30. Perez-Pinera P, Kocak DD, Vockley CM, Adler AF, Kabadi AM, Polstein LR, et al. RNA-guided gene activation by CRISPR-Cas9-based transcription factors. Nat Methods. 2013;10(10):973-6. 31. Gilbert LA, Larson MH, Morsut L, Liu Z, Brar GA, Torres SE, et al. CRISPR- mediated modular RNA-guided regulation of transcription in eukaryotes. Cell. 2013;154(2):442-51. 32. Hilton IB, D'Ippolito AM, Vockley CM, Thakore PI, Crawford GE, Reddy TE, et al. Epigenome editing by a CRISPR-Cas9-based acetyltransferase activates genes from promoters and enhancers. Nat Biotechnol. 2015;33(5):510-7. 33. Kearns NA, Pham H, Tabak B, Genga RM, Silverstein NJ, Garber M, et al. Functional annotation of native enhancers with a Cas9-histone demethylase fusion. Nat Methods. 2015;12(5):401-3. 34. Xu X, Tao Y, Gao X, Zhang L, Li X, Zou W, et al. A CRISPR-based approach for targeted DNA demethylation. Cell Discov. 2016;2:16009. 35. Vojta A, Dobrinić P, Tadić V, Bočkor L, Korać P, Julg B, et al. Repurposing the CRISPR-Cas9 system for targeted DNA methylation. Nucleic Acids Res. 2016;44(12):5615-28. 36. Sripathy SP, Stevens J, Schultz DC. The KAP1 corepressor functions to coordinate the assembly of de novo HP1-demarcated microenvironments of heterochromatin required for KRAB zinc finger protein-mediated transcriptional repression. Mol Cell Biol. 2006;26(22):8623-38. 37. Thakore PI, D'Ippolito AM, Song L, Safi A, Shivakumar NK, Kabadi AM, et al. Highly specific epigenome editing by CRISPR-Cas9 repressors for silencing of distal regulatory elements. Nat Methods. 2015;12(12):1143-9. 38. Smith RH, Chen YC, Seifuddin F, Hupalo D, Alba C, Reger R, et al. Genome-Wide Analysis of Off-Target CRISPR/Cas9 Activity in Single-Cell-Derived Human Hematopoietic Stem and Progenitor Cell Clones. Genes (Basel). 2020;11(12). 39. Zuccaro MV, Xu J, Mitchell C, Marin D, Zimmerman R, Rana B, et al. Allele- Specific Chromosome Removal after Cas9 Cleavage in Human Embryos. Cell. 2020;183(6):1650-64.e15. 14 40. Alanis-Lobato G, Zohren J, McCarthy A, Fogarty NME, Kubikova N, Hardman E, et al. Frequent loss-of-heterozygosity in CRISPR-Cas9-edited early human embryos. bioRxiv. 2020:2020.06.05.135913. 41. Adikusuma F, Piltz S, Corbett MA, Turvey M, McColl SR, Helbig KJ, et al. Large deletions induced by Cas9 cleavage. Nature. 2018;560(7717):E8-e9. 42. Kosicki M, Tomberg K, Bradley A. Repair of double-strand breaks induced by CRISPR-Cas9 leads to large deletions and complex rearrangements. Nat Biotechnol. 2018;36(8):765-71. 43. Cullot G, Boutin J, Toutain J, Prat F, Pennamen P, Rooryck C, et al. CRISPR-Cas9 genome editing induces megabase-scale chromosomal truncations. Nat Commun. 2019;10(1):1136. 44. Owens DDG, Caulder A, Frontera V, Harman JR, Allan AJ, Bucakci A, et al. Microhomologies are prevalent at Cas9-induced larger deletions. Nucleic Acids Res. 2019;47(14):7402-17. 45. Lee H, Kim JS. Unexpected CRISPR on-target effects. Nat Biotechnol. 2018;36(8):703-4. 46. Przewrocka J, Rowan A, Rosenthal R, Kanu N, Swanton C. Unintended on-target chromosomal instability following CRISPR/Cas9 single gene targeting. Ann Oncol. 2020;31(9):1270-3. 47. Ioannidis J, Taylor G, Zhao D, Liu L, Idoko-Akoh A, Gong D, et al. Primary sex determination in chickens depends on DMRT1 dosage, but gonadal sex does not determine secondary sexual characteristics in adult birds. bioRxiv. 2020:2020.09.18.303040. 48. Mitsunaga S, Shioda K, Hanna JH, Isselbacher KJ, Shioda T. Production and Analysis of Human Primordial Germ Cell-Like Cells. Methods Mol Biol. 2021;2195:125-45. 49. Webster D, Bondareva A, Solin S, Goldsmith T, Su L, Lara N, et al. Targeted Gene Editing in Porcine Spermatogonia. Front Genet. 2020;11:627673. 50. Sato T, Sakuma T, Yokonishi T, Katagiri K, Kamimura S, Ogonuki N, et al. Genome Editing in Mouse Spermatogonial Stem Cell Lines Using TALEN and Double- Nicking CRISPR/Cas9. Stem Cell Reports. 2015;5(1):75-82. 51. Wu Y, Zhou H, Fan X, Zhang Y, Zhang M, Wang Y, et al. Correction of a genetic disease by CRISPR-Cas9-mediated gene editing in mouse spermatogonial stem cells. Cell Res. 2015;25(1):67-79. 52. Shinohara T, Kanatsu-Shinohara M. Transgenesis and Genome Editing of Mouse Spermatogonial Stem Cells by Lentivirus Pseudotyped with Sendai Virus F Protein. Stem Cell Reports. 2020;14(3):447-61. 53. Sato T, Ogawa T. Generating Genetically Engineered Mice Using a Spermatogonial Stem Cell-Mediated Method. Methods Mol Biol. 2019;1874:87-98. 15 54. Chapman KM, Medrano GA, Jaichander P, Chaudhary J, Waits AE, Nobrega MA, et al. Targeted Germline Modifications in Rats Using CRISPR/Cas9 and Spermatogonial Stem Cells. Cell Rep. 2015;10(11):1828-35. 55. Fayomi AP, Peters K, Sukhwani M, Valli-Pulaski H, Shetty G, Meistrich ML, et al. Autologous grafting of cryopreserved prepubertal rhesus testis produces sperm and offspring. Science. 2019;363(6433):1314-9. 56. Shetty G, Mitchell JM, Meyer JM, Wu Z, Lam TNA, Phan TT, et al. Restoration of functional sperm production in irradiated pubertal rhesus monkeys by spermatogonial stem cell transplantation. Andrology. 2020;8(5):1428-41. 57. Shetty G, Mitchell JM, Lam TNA, Phan TT, Zhang J, Tailor RC, et al. Postpubertal spermatogonial stem cell transplantation restores functional sperm production in rhesus monkeys irradiated before and after puberty. Andrology. 2021. 58. Ciccarelli M, Giassetti MI, Miao D, Oatley MJ, Robbins C, Lopez-Biladeau B, et al. Donor-derived spermatogenesis following stem cell transplantation in sterile NANOS2 knockout males. Proc Natl Acad Sci U S A. 2020;117(39):24195-204. 59. Shami AN, Zheng X, Munyoki SK, Ma Q, Manske GL, Green CD, et al. Single-Cell RNA Sequencing of Human, Macaque, and Mouse Testes Uncovers Conserved and Divergent Features of Mammalian Spermatogenesis. Dev Cell. 2020;54(4):529-47 e12. 60. Tan K, Song HW, Thompson M, Munyoki S, Sukhwani M, Hsieh TC, et al. Transcriptome profiling reveals signaling conditions dictating human spermatogonia fate in vitro. Proc Natl Acad Sci U S A. 2020;117(30):17832-41. 61. Gurumurthy CB, Takahashi G, Wada K, Miura H, Sato M, Ohtsuka M. GONAD: A Novel CRISPR/Cas9 Genome Editing Method that Does Not Require Ex Vivo Handling of Embryos. Curr Protoc Hum Genet. 2016;88:15.8.1-.8.2. 62. Takahashi G, Gurumurthy CB, Wada K, Miura H, Sato M, Ohtsuka M. GONAD: Genome-editing via Oviductal Nucleic Acids Delivery system: a novel microinjection independent genome engineering method in mice. Sci Rep. 2015;5:11406. 63. Ohtsuka M, Sato M, Miura H, Takabayashi S, Matsuyama M, Koyano T, et al. i- GONAD: a robust method for in situ germline genome engineering using CRISPR nucleases. Genome Biol. 2018;19(1):25. 64. Kobayashi T, Namba M, Koyano T, Fukushima M, Sato M, Ohtsuka M, et al. Successful production of genome-edited rats by the rGONAD method. BMC Biotechnol. 2018;18(1):19. 65. Takabayashi S, Aoshima T, Kabashima K, Aoto K, Ohtsuka M, Sato M. i-GONAD (improved genome-editing via oviductal nucleic acids delivery), a convenient in vivo tool to produce genome-edited rats. Sci Rep. 2018;8(1):12059. 16 66. Gurumurthy CB, Sato M, Nakamura A, Inui M, Kawano N, Islam MA, et al. Creation of CRISPR-based germline-genome-engineered mice without ex vivo handling of zygotes by i-GONAD. Nat Protoc. 2019;14(8):2452-82. 67. Kobayashi Y, Aoshima T, Ito R, Shinmura R, Ohtsuka M, Akasaka E, et al. Modification of i-GONAD Suitable for Production of Genome-Edited C57BL/6 Inbred Mouse Strain. Cells. 2020;9(4). 68. Kaneko T, Tanaka S. Improvement of genome editing by electroporation using embryos artificially removed cumulus cells in the oviducts. Biochem Biophys Res Commun. 2020;527(4):1039-42. 69. Shang R, Zhang H, Bi P. Generation of mouse conditional knockout alleles in one step using the i-GONAD method. Genome Res. 2021;31(1):121-30. 70. Yoon Y, Wang D, Tai PWL, Riley J, Gao G, Rivera-Pérez JA. Streamlined ex vivo and in vivo genome editing in mouse embryos using recombinant adeno-associated viruses. Nat Commun. 2018;9(1):412. 71. Palanki R, Peranteau WH, Mitchell MJ. Delivery technologies for in utero gene therapy. Advanced Drug Delivery Reviews. 2021;169:51-62. 72. Saha K, Sontheimer EJ, Brooks PJ, Dwinell MR, Gersbach CA, Liu DR, et al. The NIH Somatic Cell Genome Editing program. Nature. 2021;592(7853):195-204. 73. Yip BH. Recent Advances in CRISPR/Cas9 Delivery Strategies. Biomolecules. 2020;10(6):839.