September 1, 2014 (Vol. 34, No. 15)
The programmable, site-specific nuclease systems at the heart of genome editing keep evolving—even the newest system, named CRISPR (for clustered regularly interspaced short palindromic repeats) or CRISPR/Cas (for CRISPR-associated system).
Although it was introduced just two years ago in a landmark paper by Emmanuelle Charpentier, Jennifer A. Doudna, and colleagues, CRISPR has already inspired researchers around the world to refine and elaborate the technology. The “CRISPR craze,” as the journal Science calls it, shows no sign of abating.
At the same time, more established nuclease systems—ZFNs (for zinc finger nucleases) and TALENs (for transcription activator-like effector nucleases)—refuse to be sidelined. These technologies, along with CRISPR, occupied center stage at a recent FASEB conference. This event, entitled “Genome Engineering—Cutting-Edge Research and Applications,” not only covered the engineering of custom nucleases, it also explored emerging principles of synthetic biology, clinical translation of genome engineering, and other topics.
ZFNs, artificial restriction enzymes generated by fusing a zinc finger DNA-binding domain to a DNA-cleavage domain, remain central to the first gene-editing technology engineered to recognize specified targets in mammalian cells. However, these complex proteins are not trivial to engineer; screening and optimization is an iterative process.
An expert in ZFN engineering, Sangamo Biosciences designs these enzymes to use in the development of novel therapeutics. One particular advance was cited at the FASEB event by the company’s senior director of therapeutic gene modification, Michael Holmes, Ph.D.
“Recently, we demonstrated in mice that we could introduce ZFNs that targeted the first intron of the endogenous albumin gene,” said Dr. Holmes. “Albumin is a very highly expressed protein in the liver, so we consider the gene to be a “safe harbor” locus. Modifying a small number of these loci allows very high levels of transgene expression.”
Using the albumin locus to convert the liver into a protein secretion factory provides flexibility for different payloads; modifying less than 1% of liver cells can produce therapeutic levels of relevant proteins.
The human factor IX gene, a gene that occurs on the X chromosome and is mutated in hemophilia B, may be inserted into mouse albumin loci. Doing so results in secretion of 40–60% of completely functional factor IX, essentially correcting the clotting disorder. The same approach was used in a mouse model of hemophilia A, which is caused by mutations in factor VIII, resulting in secretion of up to 40% of normal factor VIII levels.
Physiological levels of proteins from genes involved in lysosomal storage diseases, such as Hurler, Hunter, Gaucher, and Fabry, could also be expressed in mice, further validating the approach. The albumin locus may have advantages over adeno-associated virus (AAV) delivery. The AAV vector is not integrated into the liver genome; as the liver turns over, the AAV episome is likely to be lost.
According to Greg Davis, Ph.D., principal R&D scientist, Sigma-Aldrich Biotechnology, key ZFN enhancements in recent years are cleavage domains, termed enhanced high-fidelity FokI domains. Because these domains work only as heterodimers, they enhance specificity by blocking an individual ZFN from homodimerizing with itself on DNA and cleaving potential off-target sites.
Whereas ZFNs require two binding events to create the double-strand DNA break, the CRISPR technology initially required just one binding event. To improve the specificity of the CRISPR/Cas9 technology, paired nickases seemed like a tractable approach.
If two CRISPRs bind close together and both nick an opposite strand of DNA, a double-strand break will occur. At Sigma, Cas9 mutations were undertaken to create nickases to cleave only one strand of DNA. Close nicks result in a double-strand break. Nickases allow more permissive spacing than the use of FokI, providing increased flexibility for practical research applications, such as disease SNP modeling.
Sigma has a large lentivirus vector library, used currently for shRNA libraries, and is applying it to CRISPR. shRNA screening is limited to the exome of the genome, the protein-coding genes or roughly 1% of the genome, whereas CRISPR targets the entire chromosome. A good delivery tool, lentivirus is applicable to both arrayed and pooled high-throughput screening applications.
RNA interference (RNAi) represses activity, but for activation and epigenetic studies, modified ZFNs, TALENs, and CRISPRs may be applicable. A natural extension to the CRISPR platform is to inactivate the nuclease activity on the Cas9 protein, turning the CRISPR into a DNA-binding protein. Then activators, such as VP16 and VP64. and repressors, such as the KRAB domain, can be fused to dead Cas9. This allows enzymatic activity to be localized to a specific part of the chromosome, permitting the study of genetic regulation at specific loci.
Sigma has formed a partnership with Cleveland Clinic’s Molecular Screening Core to develop a CRISPR core. This new core is part of the Case Comprehensive Cancer Center at Case Western Reserve University and is open for researchers at Cleveland Clinic and other biomedical research centers in the Cleveland area.
One of core’s first users, Paul Tesar, Ph.D., associate professor, department of genetics and genome sciences, Case Western Reserve University School of Medicine, uses gene-editing technology to study neurogenic and neurodevelopment disorders. Dr. Tesar’s group focuses on areas of the brain that impact oligodendrocytes, which make myelin.
Specific mutations cause oligodendrocytes in the brain to produce myelin less effectively. The result: leukodystrophies, a class of pediatric congenital disorders. To better understand these diseases, the laboratory uses genome editing and aspires to create corrected oligodendrocytes from pluripotent stem cells. Once modified, the oligodendrocytes would help correct the disease in these patients.
“Gene-editing technology continually changes and can be overwhelming to a new user,” commented Dr. Tesar. “The new CRISPR core provides a rapid and effective way to access the technology regardless of level of expertise, and builds a community of researchers. Protocols, vectors, support, and technical expertise are in place to help you get exactly what you need for your particular experiment.”
An ongoing effort is improving the specificity and reducing the off-target effects of the first-generation CRISPR platform. This platform consists of a guide RNA (gRNA) and an endonuclease, Cas9. The gRNA, which consists of approximately 100 nucleotides (the first 20 of which are complementary to the target cleavage sequence), complexes with Cas9 and directs it to a specific target site where Cas9 cleaves the sequence. Mismatches can occur at the 5′ end, the gRNA region responsible for targeting. Efficient cleavage occurs with up to three mismatches; the activity is lost if there is a fourth mismatch.
Hypothesizing that there was more binding energy than was necessary to recognize and cleave the on-target site, the Joung laboratory truncated the gRNA by three nucleotides to make the gRNA more vulnerable to mismatches, and, therefore, more specific.
“This is a very simple strategy for increasing the specificity of the first-generation system,” stated J. Keith Joung, M.D., Ph.D., associate chief of pathology for research, associate professor of pathology, Massachusetts General Hospital.
A second approach created a system that essentially doubled the length of the recognition sequence, borrowing a trick from the ZFN and TALEN platforms. A mutant of Cas9, dead or dCas9, with inactivated nuclease activity, was used to create a dimer system. Then FokI and dCas9 domains were fused to create a FokI-dCas9 fusion protein. Cleavage only occurs when there are two recognition sequences with appropriate spacing and orientation between them.
This approach doubled the length of the binding site relative to the original CRISPR system (44 base pairs versus 22). Since the chance of finding related sites that long is very low in the human genome, specificity is increased.
“The current gene-editing techniques fundamentally change how one can do biological research. Incorporating the epigenome further expands the power of these technologies; to be able to rationally up- or downregulate the expression of specific genes in the cell would have broad implications,” concluded Dr. Joung.
The CRISPR platform is positioned to be used to develop treatments for disease that other modalities cannot. Dr. Joung, along with Feng Zhang, Ph.D., Jennifer A. Doudna, Ph.D., George Church, Ph.D., and David R. Liu, Ph.D. are co-founders of Editas Medicine, which was launched in late 2013 to utilize TALEN and CRISPR genome-editing technologies to develop novel human therapeutics.
“Years and years of developing ideas and aspirations about genome repair are now intersecting with a tool that works robustly. We have a super-saturation of ideas. And now, given our fundamental knowledge of the genome, complementary technologies, and CRISPR, these ideas have the potential to crystallize into therapies,” added Katrine Bosley, CEO, Editas Medicine. “The robustness of CRISPR—in many cell types, against many targets, and in many, many scientists’ hands—is a key reason for the rapid expansion of its use.”
Recombinant adeno-associated virus (rAAV) vectors enable insertion, deletion, or substitution of DNA sequences into the genomes of mammalian cells. The technique builds on Capecchi and Smithies’ Nobel Prize discovery that homologous recombination (HR), a natural high-fidelity DNA repair mechanism, could be harnessed to perform precise genome alterations in mice. rAAV-mediated genome editing improves the efficiency of this technique to permit genome engineering in any pre-established and differentiated human cell line. Such lines, unlike those of mice, have low rates of HR.
rAAV is effective at delivering donor templates and stimulating HR at the site of nuclease-induced cleavage sites. rAAV alone has long been known to increase the levels of HR and affect specific gene edits, but by combining with nucleases, the levels of HR are even further stimulated. Something unique about the single-stranded nature of the rAAV payload combined with its effective delivery to the nucleus makes it a particularly efficient donor.
rAAV is an optimal tool for the generation of highly precise knock-ins, and has applicability with all of the nuclease-based gene-editing platforms. Recently, rAAV was shown to be over 10-fold more effective as a donor than either an oligo or a plasmid. The nuclease-targeted cut occurred at some distance, over 20 bp, from the desired genetic change, a more common situation than having the nuclease cut targeted right on top of the mutation. In these situations, the composition of the donor can play a very important role.
According to Eric Rhodes, chief technology officer, Horizon Discovery Group, as the platforms for undertaking gene editing become faster, cheaper, and more efficient, the use of gene editing will increase enormously. New technologies will likely be discovered, and hybrid systems that incorporate the best aspects of more than one platform will emerge.
Tools to Understand Disease
Considered a realistic goal, the current focus in HIV therapy is cure. Lens epithelium-derived growth factor (LEDGF) is now considered a possible second target along with CCR5.
Many aspects of the HIV life cycle are unknown at the cell biology level. When the HIV lentivirus enters and infects the cell, its genome is integrated into the host genome. LEDGF/p75 is a cofactor of HIV-integrase (IN). It serves as the chromosome “docking factor” for the incoming virus and facilitates integration after docking has occurred. This modular protein has a domain ensemble at its N-terminus that attaches to chromatin and a C-terminally located integrase-binding domain (IBD) that binds the integrase dimer present in the virus.
Knockdown of LEDGF using RNAi inhibits HIV replication, but it is difficult to achieve complete knockdown of this very abundant protein, and very small residua of LEDGF that survive RNAi can fulfill the cofactor function. Gene editing, by definition, removes all possible protein from the cell.
In 293T and Jurkat human cell lines, TALENs were used to delete the entire LEDGF/p75 gene (PSIP1), and in another instance to delete just the exons that code for the IBD. At a MOI (multiplicity of infection) of 1.0, a marked delay was seen in HIV replication. HIV-1 was fully rescued by the expression of a chromatin-tethered IBD protein but not by an IN binding mutant.
However, assembly of infectious particles was normal. Potency of allosteric integrase inhibitors (ALLINIs) for rendering produced virions noninfectious was also unaffected by total eradication of LEDGF/p75 from cells. This fundamental information provides more understanding of how the virus replicates the cellular level.
“These experiments helped us learn about the role and life cycle of LEDGF. It is now straightforward to knock out genes in human cells using the TALEN and CRISPR technologies,” explained Eric Poeschla, M.D., professor, Mayo Clinic. “You just design and assemble them in the lab in a week or two, and—this is the key—they work predictably. These technologies are going to be continually optimized. It is exciting to think what will be available two years from now.
Next-Generation Gene Editing
Historically, targeted genetic modification in whole animals was limited to just a few strains of mice. This constraint arose from the fact that conventional gene targeting relies on antibiotic selection of embryonic stem cells (ESCs), which were only established and validated in a handful of mouse lines.
Nuclease technologies, such as zinc finger nucleases (ZFNs), transcription activator-like effector nucleases (TALENs), and clustered regularly interspaced short palindromic repeats (CRISPRs), on the other hand, have been rapidly developed into a universal tool for genome engineering.
“These nucleases are designed to introduce sequence-specific double-strand breaks within the genome,” says Xiaoxia Cui, Ph.D., vp of R&D at Sage Labs. “The presence of double-strand breaks at the exact target site activates cellular mechanisms to repair the breaks, leading to precise gene modification. Nucleases mediate gene editing directly in embryos in such an efficient rate that neither ES cell manipulation nor antibiotic selection is necessary.
Virtually any strain of any species is now potentially receptive for nuclease-mediated gene targeting and, in the meantime, overall timelines to create an animal model are drastically reduced, according to Dr. Cui.
Last July, Sage Labs obtained a license on CRISPR/Cas9 technology from the Broad Institute with the rights to use the system to engineer cell lines and animal models for their clients as well as distribute validated CRISPR reagents. The SAGEspeed CRISPR/Cas9 reagents were launched in the same month.