Cancer gene therapy: innovations in therapeutic delivery of CRISPR-Cas9
Introduction
Cancer is among the foremost medical problems in the developed world today. In Canada alone, 2 out of 5 people will develop cancer in their lifetime, and 1 in 4 people will die of cancer [1]. Cancerous cells evolve rapidly and are heterogeneous, and are therefore extremely difficult to target specifically. The disease is traditionally treated using [a combination of] chemotherapy, radiation, and surgery. Although therapies involving cytotoxic chemical agents and radiation are sometimes effective, they can have mixed outcomes and cause severe health side effects [2]. Meanwhile, surgery is only effective in early cases where the cancer remains confined. As our understanding of cancer has progressed, research has shifted towards elucidating the molecular basis of cancer for the development of targeted cancer treatments. This is reflected in the prevalence of drugs which produce specific molecular alterations [2]. However, targeted therapies are also possible at the genetic level, forming the basis for the application of gene therapy to treat cancer.
Gene therapy research has only emerged in the past two decades. The aim of this relatively young field is to edit and deliver recombinant DNA for therapeutic purposes. Gene therapies are promising treatments for many diseases because they can be both quick to develop and specific at the molecular level. However, there have been setbacks at the clinical stage. During a high-profile 1999 gene therapy trial for ornithine transcarbamoylase (OTC) deficiency, participant Jesse Gelsinger died of an immune response to the adenoviral vector used for delivering the corrective gene [3]. Almost at the same time, a highly-publicized gene therapy trial to treat X-linked severe combined deficiency (X-SCID) with retrovirus was completed. The therapy seemed like a huge success at first, but some patients later developed cancer due to vector integration with nearby proto-oncogenes, among other genetic abnormalities [4]. Interest and confidence in gene therapy suffered as a result.
However, the field is making a comeback as recent clinical trials for various diseases are producing promising results. Currently cancer is the major target of gene therapy, making up about 65% of gene therapy clinical trials as of 2012 [5]. In 2016, out of 66 gene therapy clinical trials compiled by the Journal of Gene Medicine, 46 trials were targeted towards some form of cancer therapy [6]. Although our knowledge of genetic diseases such as OTC deficiency and X-SCID is extensive, efficient and safe transfer of recombinant DNA products has been the main setback in the past. Thus the key factor contributing to successes today are improvements in gene therapy delivery .
CRISPR-Cas9 technology
In the past five years genome editing technologies using clustered regularly interspersed palindromic repeats (CRISPR) in combination with CRISPR-associated systems (Cas) have revolutionized the field [7]. This technology has become so versatile and accessible that it has been adopted in academic labs worldwide and has been featured in more than 5000 publications on PubMedsince 2013. The deployment of CRISPR-Cas technologies has even given riseto human germline editing discussions [7]. As of 2016, two trials for cancertreatment have been announced in China and the United States, both of which will utilize CRISPR-Cas to engineer patient T cells in vitro to destroy cancer cells [8], [9]. Although gene therapy makes up less than 5% of interventional cancer studies worldwide at present, this development has considerable promise[10].
Implementation of this technology requires a Cas nuclease, such as the widely-used nuclease from the bacterium Streptococcus pyogenes (Cas9), to be expressed within the target cell. To function, Cas9 requires a guide RNA (gRNA) composed of a scaffold sequence for Cas9 binding and a spacer sequence that defines the target region [11]. Injecting naked plasmids encoding Cas9 and a gRNA into the bloodstream results in very low levels of gene editing in mice [12]. New delivery methods for CRISPR-Cas9 must be efficient, non-toxic, evade immune clearance, and in the case of cancer therapy, deliver specifically to tumor cells. Here we review the in vitro and pre-clinical advances in developing viral and non-viral vectors for CRISPR-Cas9 delivery made within the past few years, and discuss their implications for cancer gene therapy (Fig. 1).
Figure 1. Local and systemic delivery of CRISPR-Cas9 to target cancer cells. (a) Viral vectors and synthetic nanoparticles can be injected intravenously, for systemic administration; or intratumorally, for direct delivery. Vectors delivered systemically must travel through the endothelial wall and through tumor tissues before transducing or transfecting cancer cells. Alternatively, hydrogels can be used for sustained, local delivery of nanoparticles near the tumor site. (b) Adenoviral vectors deliver DNA coding for Cas9and gRNA into the cells by binding to specific molecules on the cell surface, triggering endocytosis. Synthetic nanoparticles can deliver Cas9 in protein form, complexed with gRNA. Nanoparticles are taken up by endocytosis, fuse with the endosomal membrane, and escape into the cytosol, releasing their payload into the cells.
Viral vectors
By far the most popular tools for delivery of CRISPR-Cas9-mediated gene therapy today are viral vectors, which made up 66% of gene therapy trials as of 2012 [13]. The biggest challenge of employing viral vectors is to ensure that they are specific in their target and in their tropism (i.e., affinity for a select cell type) [14]. To compound the specificity problem, CRISPR-Cas9 gene editing is also subject to off-target effects which have been extensively analyzed [15], [16].
Adeno-associated viruses (AAVs), adenoviral vectors (AdVs), and retroviruseshave comprised the bulk of gene therapy trials to date. AdVs hold the advantage of carrying larger constructs than AAVs, with a higher level of associated proteinexpression, and have been employed to carry Cas9 with resulting targeting efficiency comparable to rates achieved with transcription activator-like effector nucleases (TALENs) [17]. Retroviruses convert their viral RNA genome to DNAvia reverse transcriptase and integrate into the genome. However, as the X-SCID trial and other, more recent gene therapy trials have demonstrated, integrating viruses can be dangerous due to the risk of insertional oncogenesis[18]. In the OTC trial, by contrast, adenoviruses were fatal due to their high immunogenicity. Various safety mechanisms have been proposed for both of these issues and are reviewed elsewhere [19], [20].
Another way to reduce the risks of insertional oncogenesis and immunogenicity is to use AAVs in place of AdVs. AAVs have a comparatively low incidence of integration in the genome and persist as episomes in primates [21]. In addition, they cannot replicate without the aid of a helper virus. Although most of the population carries antibodies for some strains of AAV (AAV-1 and AAV-2), prevalence of antibodies for other strains such as AAV-5, AAV-6, AAV-8 and AAV-9 is much lower and therefore these alternatives may be suitable for gene therapy [22]. Combining AAVs with CRISPR-Cas9 promises a solution which allows diverse and lasting gene editing with minimal immunogenic effects.
AAVs are the smallest of viral vectors, with a genome size of about 4.7 kB.Therefore it can be difficult to package Cas9, with a length of about 4.2 kB, let alone its gRNAs. However, Cas9 can be split and then functionally reconstituted by, for example, adding auto-processing intein domains [23]. In fact an in vitro approach has been developed to modulate the expression level of full and truncated Cas9 protein by combining different domains in this way [24]. Furthermore, a recent study using AAV-9 to deliver split Cas9 in mouse models did not incite any immune response to the viral vector itself, although antibodies were made against the Cas9 protein [25].
Lastly, alternatives to AAVs, AdVs and retroviruses are emerging. For instance, Sendai virus is an attractive option as, unlike the main three viral vectors, it forms no DNA intermediate during its viral cycle. This means it has the least risk of integration into the genome. It has also been successfully used to deliver CRISPR in vitro[26]. Although viral vectors have been shown to deliver CRISPR both in vitro and in vivo, whether this strategy will specifically destroy tumors in an animal model remains to be seen.
Non-viral vectors
The issues of safety, immunogenicity, and payload size in viral vectors have prompted research into alternative methods of DNA delivery in vivo. Non-viral vectors made using lipid bilayers or cationic polymers such as polyethyleneimine(PEI) make up a class of synthetic vectors which have long been used to deliver DNA to cells in vitro, and generally have larger genetic payloads and fewer toxic effects compared to viral vectors [27]. Unlike with viral vectors, synthetic vectors do not contain immunogenic pathogen-associated molecules, and patients are unlikely to have pre-existing immunity [28]. In spite of these advantages, synthetic vectors were used in less than 6% of gene therapy trials as of 2012 [13]. The primary disadvantage of synthetic vectors is their low gene transferefficiency compared to viral vectors, which limits their utility for both systemic and intratumoral delivery [29]. Advances in the design of lipid and polymericvectors have increased the feasibility of delivering DNA and RNA for therapy in vivo (reviewed in Ref. [27]); however, delivery of CRISPR-Cas9 presents its own unique difficulties.
Although synthetic lipid or polymer nanoparticles can easily be made to carry a payload much larger than 4.2 kb, which is the size of Cas9, nanoparticles must have a small volume in order to be delivered systemically through the endothelial gaps in blood vessels [29]. This means that the concentration of DNA inside of the nanoparticles must be very high in order to deliver CRISPR-Cas9 intravenously [30]. In addition, as with viral vectors, delivery of DNA encoding CRISPR-Cas9 runs the risk of random integration into the genome [31]. Plasmid delivery can increase the risk of non-specific as the plasmids persist in the cell long enough for these off-target effects to take place [32]. To address these problems, several recent studies have utilized non-viral vectors to deliver Cas9 protein rather than DNA in vivo which leads to more transient expression and does not cause insertional oncogenesis via genome integration. Delivery of Cas9 protein in mouse models using lipid and polymeric nanoparticles has proven especially promising.
Cas9 has recently been successfully delivered to mammalian cells using cationic lipid nanoparticles (liposomes) [32]. Although Cas9 itself is not sufficiently anionic for delivery with cationic lipids, complexing Cas9 with a gRNA molecule increased the negative charge on the protein [32]. Transfection of human cells in vitro showed that protein delivery was slightly more efficient than plasmid delivery, and resulted in a tenfold decrease in nonspecific genome editing. Injection of the Cas9:gRNA:lipid complex into the mouse inner ear resulted in 20% transfection of mouse inner ear cells, with no detectable toxicity [32]. A follow-up study was able to improve the transfection efficiency by utilizing bioreducible lipids, which degrade in the reductive environment of the cell, allowing the cargo to be released after endosomal escape [33]. Using this method, the transfection efficiency of human cells with CRISPR-Cas9 in vitrowas increased from approximately 40% to over 70% [32], [33]. Although in vivoediting with CRISPR-Cas9 was not attempted, the lipid nanoparticles were used to deliver a different genome editing protein, Cre recombinase to mouse brain cells near the site of injection [33].
Nanoparticles made of palindromic DNA, known as ‘nanoclews’, have also been used to deliver CRISPR-Cas9 to tumor cells in vivo. The nanoclews were complexed with the cationic polymer PEI to offset the negative charge of the DNA backbone [34]. In this study, intratumoral injection of nanoclews carrying Cas9 protein and gRNA in mice resulted in 25% transfection of tumor cells 10 days after injection. This is comparable to the transfection efficiency observed by Zuris et al. in the mouse inner ear; however, it is unclear whether 20–25% transfection efficiency will be sufficient to cause tumor regression.
For systemic delivery of CRISPR-Cas9, zwitterionic amino lipids (ZALs) were optimized to carry very long RNA molecules [30]. Separate ZAL nanoparticles carrying Cas9 mRNA and gRNA, respectively, were administered to mice intravenously, resulting in gene editing in liver, lung, and kidney tissues. Quantification of gene expression in the liver revealed that up to 3.5% of hepatocytes had been successfully edited with CRISPR-Cas9 [30]. These results indicate that lipid nanoparticles can be used to deliver CRISPR-Cas9 systemically; however, these nanoparticles were not targeted to any specific tissue, and therefore transfected several tissues with relatively low efficiency. To target cancer systemically, synthetic nanoparticles need to be designed to specifically transfect cancer cells. Nanoparticles made from a variety of different materials are available in order to achieve this goal (reviewed in Ref. [35]).
For the treatment of solid tumors, an alterative to systemic injection is local delivery using hydrogels: scaffolds formed out of naturally occurring or synthetic biocompatible polymers, which can be complexed with DNA, RNA or protein [36]. Hydrogels can be inserted during surgery or injected with a needle near the site of injection, providing local and sustained release of therapeutic genes[37].The hydrogels themselves induce very little toxicity in healthy tissues, and are poorly immunogenic [36], [37], [38]. Although this strategy has not yet been applied to CRISPR, several studies have utilized hydrogels to deliver RNAs for cancer treatment. Gold or synthetic nanoparticles coated with siRNA or miRNA and loaded onto an injectable hydrogel scaffold were shown to cause tumor reduction and prevent macro-metastases [37], [39]. Finally, gold nanoparticleshave been replaced with with synthetic particles made from poly(beta-amino ester), which dissolves in the acidic tumor microenvironment, facilitating release of the payload [38]. By coating these nanoparticles with siRNA and loading them into an injectable hydrogel, the authors were able to achieve 70% gene silencingin murine tumors compared with 20% using intravenously injected nanoparticles [38]. These results suggest that hydrogels could be used to aid the delivery of CRISPR-Cas9 by synthetic nanoparticles.
We have seen that a variety of synthetic vectors including liposomes, bioreducible lipids, nanoclews, ZALs and hydrogels show promise for delivery of CRISPR-Cas in vitro. Overall, studies utilizing synthetic vectors to treat cancer using CRISPR-Cas9 in vivo must be carried out before the potential of non-viral vectors can be realized.
Conclusion
Genome editing with CRISPR-Cas has revolutionized the field of gene therapy in just a few short years. However, the specific limitations of CRISPR-Cas9, including the large size of Cas9 and its potential for off-target effects, pose challenges to vector design. As with other forms of gene therapy the primary concerns include efficiency and specificity of delivery. Here we have described very recent in vitro and pre-clinical work done to deliver CRISPR-Cas9 using both viral and non-viral methods, as well as the benefits and downsides associated with each delivery method. Viral vectors are very efficient, but safety issues have limited their usage and their payload is small. Non-viral vectors are being rapidly developed at present, especially nanoparticles. These nanoparticles can deliver either Cas9 protein or mRNA, however, their transfection efficiency is still much lower compared to viral vectors. We have highlighted potential methods to surmount the weaknesses of each method. When brought into the clinic these developments will allow CRISPR-Cas9 to target cancer efficiently throughout the body without harmful side effects, allowing the enormous potential of cancer gene therapy to be realized.
