Before CRISPR genome editing materials can execute a therapeutic change, they must reach their destination, which can seem as forbidding as a well-fortified castle, and remote, too, particularly if the therapeutic change is to be executed in vivo. For example, if CRISPR genome editing materials are to be administered via systemic injection, they must overcome a succession of obstacles on their journey to a specific cell type or cellular compartment.
The first obstacle is the circulation, where multiple hazards await, compelling developers of CRISPR therapeutics to stuff their genome editing cargo into a protective shell. The cargo, which may include Cas9 protein, mRNA, or plasmid DNA, needs to be electrostatically compatible with its packaging—the inner surface of a nonviral nanoparticle, for example.
The nanoparticle shell can protect cargo from hazards such as aggregation, immune clearance, removal by the kidneys, or premature release. And when decorated with targeting ligands, the nanoparticle shell can guide its cargo to the right cell types. Cargo release can become even more precise if targeted tissues are subjected to volleys of focused thermal energy, electromagnetic energy, or ultrasound.
A well-aimed bombardment or a coating that includes cell-penetrating peptides could help the encapsulated genome editing materials overcome the next obstacle, the plasma membrane of the target cell—the outer wall of the castle. Here, the genome editing materials may burrow through directly, or they may effectively pass through a gate, protective shell and all, via endocytosis. But then the endosome itself will pose another obstacle. If endosomal escape should fail, the particle containing the genome editing materials will fail as a trojan horse. It will ignominiously roll along the lysosomal degradation pathway.
Finally, after they brave the circulatory frontier, breach the plasma membrane, and (if necessary) escape the endosome and the encapsulating nanoparticle, genome editing proteins and plasmids must somehow traverse the cytosol and enter the nucleus, the cell’s castle keep. For protein, the final leg of the trip is relatively straightforward and may be facilitated by add-ons such as nuclear localization sequences. For plasmid DNA, the final leg is more of a loop, since the DNA must be transcribed into mRNA, which must leave the nucleus to be translated into protein, which must, like the plasmid DNA, make its way to the nucleus.
As daunting as all these obstacles may seem, many of them are efficiently negotiated by viral vectors, specifically, vectors based on lentiviruses and adeno-associated viruses. Viral vectors, however, have several drawbacks. Often, they have limited capacity, which means that multiple viral vectors may need to be deployed in parallel or in staggered formation as a means of distributing the genome editing load. Viral vectors have also been associated with instances of insertional mutagenesis and carcinogenesis, as well as with heightened immune responses. Finally, viral vectors do not always lend themselves to large-scale production.
In hopes of avoiding these drawbacks, scientists are developing nonviral vectors, lipid- or polymer-based nanocarriers. Nonviral vectors have already shown promise as platforms that can deliver CRISPR-Cas systems relatively safely. Nonviral vectors have also been extensively engineered to accommodate varied cargos, including large cargos, while lowering barriers to systemic delivery. For example, the incorporation of hydrophilic molecules such as poly(ethylene glycol) (PEG) can prevent fouling and decrease immunogenicity, and the incorporation of polycations such as polyethylenimine (PEI) can cause endosomal deformation, facilitating endosomal escape. But even for nonviral vectors, barriers remain, such as the need to improve targeting. Finally, nonviral vectors still need to adequately address their biggest challenge: efficacy.
Lipid- and polymer-based nanoparticles
Two common nonviral delivery vehicles, lipid- and polymer-based nanoparticles, are being refined by researchers. For example, a research collaboration between Tufts University and the Chinese Academy of Sciences reported “effective and very fast” CRISPR-Cas9 genome editing in vitro and in vivo was achieved with bioreducible lipid-Cas9 messenger RNA (mRNA) nanoparticles.
The lipid-based nanoparticles, recently described in Advanced Materials, are made of synthetic lipids that incorporate disulfide bonds. When the nanoparticles enter the cell, the disulfide bonds break upon exposure to the cell’s reductive environment. The nanoparticles fall apart, and the contents are released into the cell.
The researchers applied the new method in a mouse model, where they hoped to reduce the presence of a gene coding for PCSK9, the loss of which is associated with lower LDL cholesterol. “Intravenous injection of a Cas9 mRNA/sgRNA nanoparticle effectively accumulates in hepatocytes,” the researchers wrote, “and knocks down proprotein convertase subtilisin/kexin type 9 level in mouse serum down to 20% of nontreatment.”
Recent research focused on polymer-based nanoparticles includes work that is being advanced at the University of Wisconsin-Madison, where laboratories headed by Shaoqin “Sarah” Gong and Krishanu Saha are collaborating on customizable synthetic nanoparticles for the delivery of Cas9 nuclease and a single-guide RNA (sgRNA). “We describe the synthesis of a thin glutathione (GSH)-cleavable covalently crosslinked polymer coating, called a nanocapsule (NC), around a preassembled ribonucleoprotein complex,” the scientists and their colleagues reported in a recent Nature Nanotechnology article.
The NC stays intact in the bloodstream, but falls apart inside the target cell, where disassembly is triggered by cytosolic glutathione. According to the scientists, the NC is synthesized by in situ polymerization, has a hydrodynamic diameter of 25 nm, and can be customized via facile surface modification.
“NCs efficiently generate targeted gene edits in vitro without any apparent cytotoxicity,” the article’s authors declared. “Furthermore, NCs produce robust gene editing in vivo in murine retinal pigment epithelium tissue and skeletal muscle following local administration.”
A comprehensive view
At the Ohio State University, researchers led by Yizhou Dong, PhD, are developing cell-specific and multifunctional drug delivery systems and evaluating these systems in animal models for treating genetic disorders, infectious diseases, as well as cancers. Earlier this year, Dong and colleagues published a review article, in Trends in Biopharmaceutical Sciences, that encompassed ex vivo and in vivo techniques. While discussing the latter, the article also highlighted several viral and nonviral systems.
“AAV serotypes with different tropisms have the ability to target different organs,” the authors noted. “[And] a number of biomaterials, such as lipid, polymeric, and inorganic nanoparticles (NPs), have been developed for transient expression of the CRISPR systems in vivo.”
When asked by GEN to highlight potential advantages of nonviral vectors, Dong noted that nonviral delivery systems possess unique properties such as repeat administration and minimal genotoxicity. “Varieties of nonviral delivery systems are needed,” he continued. “Tissue and cell specificity, biodegradability and biocompatibility, [and] safety are among the important factors to be considered for selection of delivery materials.”
Recent work in the Dong laboratory has emphasized the development of mRNA carriers. “We developed lipid-like nanomaterials that efficiently encapsulate mRNA-encoding gene editing components and guide RNA,” he detailed. “The formulated nanomaterials can protect the payloads from degradation in serum. Meanwhile, these nanomaterials are able to deliver these payloads mainly into the mouse liver and induce genome editing of specific genes.”
In their review, Dong and colleagues emphasized that delivery systems need to be on target in two senses. First, the genome editing materials carried by the delivery systems need to reach “many types of cells, such as neurons, cardiomyocytes, and immune cells.” Second, once the materials are delivered, they should affect only the intended gene loci “so as to prevent large deletions and complex rearrangements.”
“Many factors such as therapeutic targets, delivery systems, indications are important to decide payload type,” Dong remarked to GEN. “Nonviral delivery systems need to match payload type. Different payload types have distinct physicochemical and biological properties, which would provide advantages for specific applications.”
Traditionally, the bugle call “Charge” told the cavalry when to spring into action. Soon, focused ultrasound may do the same for vehicles that transport and deliver CRISPR-based therapeutics. That is, ultrasound could be used to disrupt nanomaterials and induce them to release their cargo when and where desired.
At Penn State, researchers led by Scott Medina, PhD, developed ultrasound-sensitive fluoroprotein nanoemulsions that can be acoustically tracked, guided, and activated for on-demand cytosolic delivery of proteins. To date, the researchers’ work has emphasized the delivery of antibodies. The researchers also have indicated that they intend to work with CRISPR constructs and achieve ultrasound-controlled genome engineering of cells in complex three-dimensional tissue microenvironments.
The new delivery system depends on a family of fluorous tags, or FTags, discovered by Medina and colleagues. “FTags transiently mask proteins to mediate their efficient dispersion into ultrasound-sensitive liquid perfluorocarbons, a phenomenon akin to dissolving an egg in liquid Teflon,” the scientists explained in ACS Nano. “We identify the biochemical basis for protein fluorous masking and confirm FTag coatings are shed during delivery, without disrupting the protein structure or function.”
Focused ultrasound also resounds in the work of Columbia University’s Kam Leong, PhD, and Elisa Konofagou, PhD. These scientists recently received a $3.2 million grant from the NIH to help them achieve focused ultrasound-mediated delivery of gene-editing elements to the brain for neurodegenerative disorders.
The scientists’ overall objective is to develop a noninvasive, focused ultrasound–mediated technology for delivering AAV vectors and nonviral polyplexes carrying CRISPR elements to the brain and to evaluate the efficacy on two major neurodegenerative disorders, Alzheimer’s and Parkinson’s diseases. The new technology will build on the researchers’ previous work, which demonstrated that focused ultrasound, in conjunction with monodispersed, gas-filled microbubbles, could deliver therapeutic payloads to a specific region of the brain through intravenous injection in both rodent and nonhuman primate models.
Unique identifiers for individual nanoparticles
Variations on a nanoparticle design meant to improve tissue targeting are synthesized by the thousands without too much difficulty. But determining which variations will perform better in vivo is a laborious one-at-time or few-at-a-time undertaking—unless the designs are evaluated en masse.
An efficient means of evaluating nanoparticle designs has been developed by scientists at Georgia Tech led by James Dahlman, PhD. “My lab has developed high-throughput DNA barcoding approaches to test how many distinct nanoparticles deliver nucleic acid drugs in a single animal,” he tells GEN. “In this method, nanoparticle 1 is formulated so it carries DNA barcode 1, and nanoparticle 100 is separately formulated so it carries DNA barcode 100.
“After a quality control check is performed, nanoparticles are pooled together and co-administered into the animal. After waiting for drug delivery to occur, we isolate the cell types of interest (both on- and off-target cells), and we use next-generation sequencing to quantify how much of each barcode ended up in the tissue. In a typical experiment, we quantify how ~100 nanoparticles deliver drugs to ~20 cell types, which is equivalent to performing 2,000 in vivo experiments in a single mouse.”
Dahlman asserts that predictions for the in vivo delivery of individual nanoparticles are far better if nanoparticles are screened directly in vivo, instead of in vitro, which is the “traditional” approach. “By testing several thousands of nanoparticles in vivo,” he continues, “we have identified nanoparticles that deliver nucleic acid drugs to many new cell types, including several we were surprised to observe.”
Dahlman and colleagues have launched Guide Therapeutics, a company that uses barcoding to deliver gene therapies to new tissues. “The results have been very exciting,” Dahlman states. “In both the lab and the company, we are cautiously optimistic that barcoding can be used to identify nanoparticles with more specific delivery profiles, thereby reducing side effects by reducing delivery to off-target cells.”
A breath of fresh CRISPR
Genome editing materials can’t just breeze into cells. Or can they? Even cells so well defended as lung and airway cells may admit wisps of genome editing proteins such as Cas nucleases. All that’s needed is an inspired delivery method. One possibility is the aerosolization of amphiphilic peptides.
Amphiphilic peptides, which combine hydrophilic and lipophilic properties and facilitate the translocation of proteins across membranes, are being evaluated for various applications, including genome editing. In fact, scientists from the University of Iowa, in collaboration with scientists from Feldan Therapeutics, recently used engineered amphiphilic peptides to deliver genome editing nucleases and ribonucleoproteins to cultured human airway epithelial cells and mouse lungs.
The University of Iowa’s Paul B. McCray, Jr, MD, asserts that the overall advantage of the approach is its simplicity. “The shuttle peptide is mixed with the Cas-RNP in a saline solution, and then the Shuttle/Cas-RNP solution is then applied directly to cells or administered locally to the airways,” he explains. “The shuttle peptides facilitate rapid delivery (seconds) of the protein cargo, thereby allowing a rapid onset of effect.”
Using reporter mice, McCray and colleagues evaluated the editing efficiency of a shuttle peptide that had been noncovalently combined with Cas9 RNP. The scientists found that nasal instillation generated an editing efficiency up to 13%. “We are currently working to improve this editing efficiency,” McCray reports. He adds, however, that a 10–15% genome editing efficiency should be sufficient to restore therapeutically relevant levels of CFTR protein activity in airway epithelia of cystic fibrosis patients.
“The delivery of Cas-RNP protein instead of DNA or RNA limits the duration of cell exposure to the editing machinery and allows immediate onset of genome editing,” McCray maintains. “The shuttle peptides, which are composed of natural amino acids, and the Cas-RNP are rapidly and naturally eliminated by cells. The production of shuttle peptides by chemical synthesis is feasible and scalable as well as the production of Cas nucleases and synthetic gRNAs.”
He adds that the method is not tissue or cell specific, and that it allows delivery of Cas-RNP to lungs and other accessible tissues, as well as to cells in culture such as hard-to-modify human natural killer cells.
Presence of an absence
To emphasize “nonviral” is to invoke, however faintly, its opposite: “viral.” So, before this article closes, what else can be said about nonviral particles as alternatives to viral delivery systems? Earlier, it was noted that nonviral particles often lack the efficiency of viral vectors, which have had the advantage of having evolved over hundreds of millions of years. Yet nonviral vectors, which have the benefit of rational design, are catching up in some respects, and are even demonstrating advantages in others.
In any event, viral vectors aren’t resting on their evolutionary laurels. For example, they’re acquiring improved tissue targeting capabilities. So, will both nonviral and viral vectors find their niches? When GEN put this question to Dahlman, he responded, “I believe that some payloads (for example, Cas9, which ideally is expressed for short periods of time) will likely thrive using nonviral delivery, whereas other payloads (for example, DNA encoding Factor, which requires very long-term expression) will benefit from viral vectors.
“As I have studied more and more genetic diseases, it is almost always the case one type of delivery system is very likely to ‘win’ when compared to another delivery system. As a result, I don’t see a lot of competition between the viral and nonviral delivery systems.”
Boosting Delivery Rates, Scaling Up Production
By John D. Lewis, PhD
In the 17 years since the Human Genome Project was completed, we have made great strides in our understanding of the genetic causes of disease. For a great number of them, simply increasing or decreasing the production of a single gene product has the potential to be transformational if not curative for patients. Although we have developed the tools to manipulate gene expression and edit genes in the research setting, the real challenge has been to develop drugs based on these technologies that can have the same impact in human patients.
The key roadblock is intracellular delivery. Viruses do a great job of delivering genetic material inside cells, but there are significant limitations in their application to human disease. They are extremely challenging to manufacture and purify at scale. They have limitations in their cargo-carrying capacity. They have preferences for certain cell types and not others. Perhaps most importantly, since they are viruses, they stimulate the immune system when they are administered the first time—meaning multiple dosing is typically not feasible.
By using nanoparticles, it is possible to overcome the limitations of viral formulations. Nanoparticles can be formulated with lipids or polymers that are naturally occurring and can be dosed repeatedly due to their lack of immunogenicity. Nanoparticles have few limitations when it comes to cargo capacity and target cell preference.
Most importantly, nanoparticles can be manufactured using highly scalable techniques such as microfluidics. Microfluidic manufacturing relies on precise mixing of the nanoparticle constituents in a tightly controlled but continuous process that occurs in a microfluidic chip.
A process developed at the research scale can easily be scaled up to clinical or commercial scale using similar hardware and identical microfluidic architecture. This ability to rapidly design and manufacture genetic medicines that may be specific to an emerging disease (such as COVID-19) at a massive scale or to the genetics of an individual patient’s cancer at a personalized scale opens up a whole new world of possibilities for curing previously untreatable diseases.
Since the first approval for a gene therapy in the United States in 2017, there has been an explosion in research and development activity in the gene therapy and gene editing space. Recognizing the limitations of viral delivery systems, many biotechnology companies are adopting nonviral formulations to bring their promising medicines to market. I expect that demand for nonviral nanoparticle technologies to rise exponentially over the next few years.
Despite the obvious advantages of nanoparticles, poor intracellular delivery compared to viral systems is their main limitation. Entos Pharmaceuticals has developed a nanoparticle platform (Fusogenix) that can solve this limitation by incorporating a viral fusion protein into a lipid nanoparticle that allows the nanoparticle to deposit its cargo directly into the cytoplasm, bypassing the endocytic pathway. We manufacture our nanoparticles using the Precision Nanosystems microfluidic platform.
John D. Lewis, PhD, is the CEO of Entos Pharmaceuticals.