CRISPR-Cas is a relatively simple, yet effective technique for genome editing in mammalian cells. The RNA-guided Cas endonucleases of bacterial origin have been optimized to create efficient tools for the manipulation of DNA genomes of mammalian cells. This has accelerated fundamental research and enabled the clinical translation of novel therapeutic strategies to successfully treat genetic and other diseases [1, 2]. In fact, the considerable diversity of the prokaryotic CRISPR-Cas systems is currently driving a true biotechnological revolution. Most of these systems require two small RNAs, CRISPR RNA (crRNA) and trans-acting CRISPR RNA (tracrRNA), which were combined into a single guide RNA (gRNA) to simplify the use . Other systems do not require tracrRNA, e.g. the Cas12a system is guided by only a single crRNA [4, 5]. The prototype Cas9 enzyme cleaves double-stranded DNA (dsDNA), but Cas12 is additionally able to edit single-stranded DNA (ssDNA) and Cas13 cleaves exclusively RNA [4, 6, 7]. The distinct cleavage specificities of different Cas homologs trigger different DNA repair pathways of the host cell, for instance non-homologous end joining (NHEJ), that result in distinct editing lesions . Thus, selection of the appropriate CRISPR-Cas system is key and depends on the specific application that one has in mind. Selection of the appropriate gene delivery system is another critical variable that should be addressed early.
Vector systems based on retroviruses have been a popular delivery tool because they exhibit the unique property of integration into the genome of the host cell, which ensures the prolonged presence of the transgene. A replication-defective vector based on the human immunodeficiency virus (HIV) is called a lentiviral vector (LV). LVs are an efficient tool for transgene delivery to achieve stable and long-term expression in eukaryotic cells and have been successfully used for clinical gene therapy applications [9, 10]. LV-based vectors have a particular advantage over other retroviral systems in that they can infect both dividing and non-dividing cells. For some difficult-to-transduce cell types such as stem cells, LVs can greatly improve the transduction efficiency and thereby increase the probability of transgene integration into the genome of the target cells. However, safety concerns were raised after some initial oncogenesis events were described in trials with retroviral vectors [11, 12]. Following the development of leukemia in a few patients, the gene therapy field shifted more towards the use of LV. Initially, there was a similar concern that LV could also be oncogenic. LV-induced insertional mutagenesis was described in proliferative hematopoietic stem cell (HSCs) and tumors were observed in LV-treated mice [13, 14]. Researchers pursued different strategies to optimize these vectors and a safe third-generation LV was developed for clinical use [15–17]. For instance, whereas HIV encodes 5 essential proteins, only 3 genes (gag, pol, rev) were retained in this LV, which are actually expressed from separate plasmids to avoid the chance of generation of replication-competent virus by recombination [18–20]. The requirement for the HIV transcriptional regulator tat was circumvented by the use of a novel, constitutively active promoter and the HIV env (envelope) function was replaced by the G protein of vesicular stomatitis virus (VSV-G) to allow transduction of other cell types. In addition, all genes were codon-optimized to boost their expression, which also reduced the sequence similarity to the replication-competent virus. A complete 4-plasmid system was created by inclusion of the transfer plasmid that produces a transcript that is selectively packaged in the LV particles, followed by conversion into dsDNA that eventually is integrated in the genome of the transduced cells (Fig. 1A). Note that the LV RNA genome is packaged as a dimeric RNA in assembling virion particles.
A) We used a third-generation lentiviral vector that consists of 4 components: the transfer plasmid encodes the transgene cassette, the envelope plasmid encodes the Vesicular Stomatitis Virus (VSV) glycoprotein (G), and two packaging plasmids that encode Rev and Gag-Pol. Ψ is the packaging signal on the vector RNA. RRE is the Rev-responsive element and cPPT is the central polypurine tract, signals that play an important role in enhancing transgene expression and vector transduction, respectively. B) Lentiviral vectors that encode the CRISPR-Cas components: the CRISPR endonucleases RfxCas13d, CjCas9, SaCa9, AsCas12b and SpCas9 and their respective gRNA scaffolds were cloned in LV to yield constructs 1+, 2+, 3+, 4 + and 5+, respectively. The U6 promoter and gRNA scaffolds were deleted to make constructs 1-, 2-, 3-, 4- and 5-.
The LV can in principle transfer any chosen CRISPR-Cas system into any target cell type. However, LVs have a limited packaging capacity for the transgene RNA transcript and previous studies revealed that the LV titer decreases with increasing transgene size [21–23]. Kumar et al. demonstrated that there is a reduction in transduction titer of approximately 3-fold per kilobases (kb) increase in RNA cargo size for inserts over 5 kb, whereas the titer dropped approximately 100-fold for inserts between 6 and 12.5 kb . Likewise, Canté-Barret et al. demonstrated that the length of the viral RNA is a critical factor that affects the efficiency of LV-mediated gene transfer . Whereas the LV transduction efficiency on human and mouse stem cells dropped dramatically for viral RNAs that approached 6 kb, a modest reduction of their size of 0.6 kb rescued the transduction efficiency by more than threefold. Consistently, Sweeney et al. showed that the LV titer drops with increasing genome size . In general, genome sizes close to or exceeding that of the natural HIV template (9.8 kb) result in decreased packaging capacity and consequently a low transduction efficiency. The relatively large size of Cas systems thus presents a serious challenge for packaging in LV systems. For instance, a simple combination of the Streptococcus pyogenes endonuclease (SpCas9) with a gRNA cassette already makes a transgene of at least 4.2 kb. Additional regulatory elements like promoters, enhancers, regulatory RNA motifs and GFP or another reporter gene do easily make a vector that is significantly larger than 9 kb. The precise size limitations for packaging of CRISPR/Cas systems in LV systems have not been determined yet. Therefore, the aim of this work was to assess the impact of Cas genome size on LV production, the packaging of the transgene RNA and the transduction efficiency. We therefore packaged five Cas genes encoding endonucleases of difference size, with and without the matching g/crRNA cassette. We included in this study the prototype Staphylococcus aureus Cas9 system (SaCas9, 3.2 kb RNA) that produces indels at a similar efficiency as SpCas9 , but also the Cas9 system from Campylobacter jejuni (CjCas9, 3.0 kb RNA) that efficiently edits genes in mouse muscle and eye tissue , and the Cas12a and Cas13d systems that represent smaller Cas enzymes with just 3.9 and 2.9 kb RNA, respectively [27, 28].