Abstract
The introduction of the CRISPR-Cas9 system in 2013 has revolutionized experimental genetics in mice and rats. This commentary gives an overview on the use of CRISPR either for gene editing in the germline or for editing and beyond editing in somatic cells. Future perspectives are opened by emerging CRISPR technologies that could enable genome engineering at larger scale.
Since the first knockout mouse was described more than three decades ago, engineering of the mouse genome has proven extremely useful for understanding the function of genes in vivo.1–3 On the time scale of engineering technologies I would like to highlight 1987, 2002 and 2013 as years with key discoveries that opened up new routes for research. The year 1987 witnessed the first gene targeting by homologous recombination in mouse embryonic stem cells that established a paradigm for thousands of knockout lines that followed. 4 The technology was later refined further into conditional gene targeting using Cre/loxP recombination. 5 The year 2002 saw the sequencing of the mouse genome, giving us a complete view and access to all genes, greatly enabling their engineering at individual or large scale. 6 The third key development was the application of the CRISPR-Cas9 nuclease system for gene editing in mammalian genomes in 2013, including its use in rodent zygotes. 7 , 8 Looking back from here, we can now reflect on eight years of CRISPR-Cas9 applications in mouse and rat zygotes and consider – where do we stand and where do we go?
To answer these questions it is helpful to subdivide the applications of CRISPR-Cas9 into three groups. The first group includes the complete or conditional knockout of genes in the germline as previously achieved through gene targeting in embryonic stem cells. However, quantitatively this approach is boosted by CRISPR-Cas9 due to the significantly reduced time and effort required for gene editing directly in one-cell embryos. Furthermore, the protocols for use of CRISPR-Cas9 in zygotes have evolved in several steps since 2013 and facilitate the production of germline mutants. First, the classical approach of using homologous recombination targeting vectors with large homology regions was complemented by the use of synthetic oligonucleotides for introduction of small precise sequence modifications. 8 Second, knockout alleles can now be easily produced without targeting vectors using guide-RNAs (gRNAs) that cause frameshift mutations by small deletions or insertions (indels) or by the deletion of gene segments using pairs of gRNAs. 9 Therefore, more edited lines can be generated in a shorter time with fewer resources, greatly benefiting individual researchers as well as enabling large scale mutagenesis as pursued by the International Mouse Phenotyping Consortium or commercial providers. 10 , 11 In addition, the delivery of CRISPR reagents became more convenient and reliable by replacement of in vitro transcribed RNAs with recombinant Cas9 protein and synthetic gRNAs from commercial vendors. 12 Higher homologous recombination frequencies can be obtained by using linear double-stranded DNA donor molecules (HMEJ, TILD) or by using long single-stranded DNA donors (Easi-CRISPR).13–15 Another milestone was the establishment of protocols for batch electroporation of zygotes as an easy alternative to serial embryo microinjections, initially restricted for use of Cas9/gRNA complexes and oligonucleotides passing the zona pellucida.16–18 The latest advance also enables the delivery of targeting vectors with a size of several kilobases (kb) in combination with zygote electroporation. For this purpose, the donor is packaged into AAV viral particles that can be added to the culture medium in high numbers, pass the zona pellucida, transduce into the zygotes and are beneficial for homologous recombination by their single-stranded and end protected genome. Although donor delivery via AAV requires extra efforts for virus preparation it offers the perspective of raising the knockin frequencies to levels of ≥50% as reported in several pioneer studies.19–21 Under these conditions half or more of the pups born from manipulated zygotes represent the desired founder mutants, requiring only a limited number of embryos to be handled within one or two working days. These high frequencies were previously reported only for the generation of knockout alleles via indel formation. If the AAV technology can be validated by a wider community across a wider range of target genes, it will make also the production of knockin mouse models more reliable. Another versatile application of CRISPR technology in zygotes is the introduction of mutant alleles into genetic backgrounds other than the commonly used C57BL/6 strain, such as NSG or NOD mice.22,23 Chromosome engineering for the deletion, inversion or duplication of megabase segments, initially performed using CRISPR-Cas9 in embryonic stem cells, can also be achieved directly in zygotes.24,25 Beyond its application in mouse zygotes, CRISPR-Cas9 had an even greater impact for rat genetics since, in the absence of reliable embryonic stem cells in the pre-nuclease era, nuclease technologies, first shown with zinc-finder nucleases and TALEN, enabled targeted manipulation of the rat genome,26,27 enabling the same types of genetic models to be created in rats as have been built in mice. 28 Seen from a perspective before 2013, these developments alone represent a revolution, but the utility of CRISPR is even broader because of its additional application in somatic cells and for genome-wide engineering, as discussed in the following sections.
The second category includes CRISPR applications that go beyond mouse germline mutagenesis to enable gene manipulation directly in somatic cells in vivo or ex vivo, something intractable before 2013. Gene editing for the therapy of monogenetic diseases in somatic cells is an exciting new opportunity. Either Cas9 or base editors have been successfully used for therapeutic gene editing in the liver, muscle, or the whole body of newborn or adult mice.29–31 Currently, CRISPR reagents are mostly delivered to somatic cells by AAV vectors, but many in the field are seeking to replace AAV with non-viral delivery of components through various sorts of nanoparticles.32–34 In the US such developments are systematically supported by the Somatic Cell Genome Editing funding program of the NIH, including rodent testing centres for development of reporter systems and somatic editing tools. 35 The other application area of CRISPR in somatic cells aims not for gene repair but for knockout or precise mutagenesis of target genes for cancer modelling or functional CRISPR screens in primary cells either in vivo or ex vivo. 36 The establishment of Cas9 transgenic mice was instrumental for this research because it is difficult to deliver Cas9 using a single AAV vector that offers, besides the large Cas9 coding region, only very limited space for promoter and polyA regions. In contrast, the use of Cas9 transgenic mice or cells requires only the delivery of single AAV vectors with a small gRNA expression unit. Furthermore, mice with doxycycline inducible Cas9 transgenes can be employed to generate indel based sequence barcodes in somatic cells for the tracing and reconstruction of cell lineages. 37 Beyond gene editing, Cas9 variants without nuclease activity, that should not create double-strand breaks, can be used to bind the nuclease-dead dCas9 protein with specific gRNAs to selected genomic target regions. Fusion proteins of dCas9 with regulatory protein domains then allow programmable activation, silencing or the epigenetic repression of a target gene.38–43
The third class of CRISPR applications is just emerging or not yet realized but theoretically feasible with future potential, in particular if editing technologies will be further streamlined. A potential future application could be efficient multiplex gene editing for, for example, studying polygenic traits. However, CRISPR-Cas9 can be efficient for targeting single loci but targeting two or more loci in a single step in a reasonable fraction of zygotes is currently not routine. Nevertheless, in an ideal world of CRISPR technology Cas9 could process multiple target loci at high efficiency. Further research will be required to clarify whether nuclease efficacy can be further boosted or the enzyme may reach its limit on mammalian genomes that exceed the size of its natural host genome by ∼1000-fold. Other elusive CRISPR applications are methods for replacement of large (10–1000 kb) exogenous DNA segments that are presently lacking protocols for assembly and site-specific transfer into the genome. Such methods, once available, would allow the easy humanization of mouse genes by replacement of, for example, a 50 kb segment by its human counterpart in a single step in zygotes. Finally, for the potential global engineering of an entire mouse genome, as exemplified by the Yeast 2.0 project, 44 it would be necessary to efficiently modify an inbred strain at many locations simultaneously. One could, for example, imagine redesigning a strain with new traits such as resistance to viral pathogens, or repairing genetic defects of lab strains that were fixed homozygously by inbreeding. In the C57BL/6 mouse genome hundreds of genes are compromised by missense mutations in protein coding regions or the intronic integration of endogenous retroviral elements.45,46 In a first endeavour in this direction we are running a pilot study for the repair of 46 loci in the C57BL/6N genome. To this end CRISPR technology could liberate mouse geneticists from the historic origin of inbred strains, an endeavour that was intractable before the CRISPR era.
Taken together, the initial phase of CRISPR mediated gene editing in mice and rats has already brought us unprecedented freedom and new ways of modifying the genome in somatic and germ cells. In the coming years the universe of CRISPR applications will likely being further expanded and make genetic modelling an even more fascinating scientific endeavour.
