The CRISPR toolbox gets bigger and better

Researchers are honing genome-editing systems, discovering new ones, and envisioning versatile CRISPR tools yet to come

Lauren Solomon, Broad Communications
Credit: Lauren Solomon, Broad Communications

Genome editing tools based on CRISPR systems are revolutionizing the world of life-sciences research and could one day spawn novel medicines — but there is still a long way to go. While experiments using CRISPR-based tools are revealing how genes operate in health and disease, researchers are hunting for new tools to add to this biological “swiss army knife” while making existing ones more precise and versatile.

“We are probably still in the bronze age or, perhaps optimistically, the iron age of these technologies,” said David Liu, director of the Merkin Institute of Transformative Technologies in Healthcare and core institute member at the Broad Institute of MIT and Harvard, professor of chemistry and chemical biology at Harvard University, and HHMI investigator. “Despite substantial recent advances, there are many important ways these agents still need to be improved to maximize their effectiveness and safety as research tools and as human therapeutics.”

Over the past decade, scientists have built various CRISPR-based genome editing systems that can be programmed to target and edit specific stretches of genetic code, with potential applications in biological research, therapeutic development, clinical diagnosis, and agriculture. The current systems, however, are far from perfect: they don’t always hit their intended targets precisely, they often make undesired edits, they are difficult to deliver into living cells, and they can trigger an unwanted immune response.

To help overcome CRISPR’s limitations, Liu and other researchers are thinking creatively to improve CRISPR-based tools and imagine new capabilities and uses of this innovative technology.

Cas9 gets a makeover

The CRISPR system was discovered in bacteria, where it functions as an adaptive immune system that recognizes and rapidly destroys foreign DNA from invading viruses. DNA-cutting enzymes disable these invaders by finding and cleaving specific sites in the viral genome.

As part of the CRISPR-based editing system adapted from bacteria, Cas9 is the most widely used CRISPR enzyme. However, natural Cas9 lacks the precision needed for many experimental or therapeutic applications, in part because the double-stranded cuts it makes can lead to undesired and uncontrolled genetic changes, and because the mechanism used by the Cas9 enzyme to find targets that match the RNA guide is imprecise. The Cas9 molecule is also large, unable to cross the cell membrane on its own, and capable of triggering an undesired immune response if delivered inappropriately — all of which hinders its therapeutic potential.

In labs across the Broad and beyond, scientists are working to overcome these obstacles.

For most applications, it is crucial for Cas9 to locate and selectively edit the correct targets within the cell. In the Broad’s Genetic Perturbation Platform, institute scientist John Doench and colleagues use machine learning in collaboration with scientists at Microsoft Research to predict which guide RNAs are the best at homing in on specific sites. Careful engineering based on these data can minimize undesired “off-target” genetic changes.

“We’ve already been able to sort out some of the rules underlying which guide RNA sequences help Cas9 function best in the cell,” said Doench. “Predicting where a given Cas9 might also cut — its off-target effects — has been much harder, but we’re starting to see some early fruits from this effort.” Because billions of sites in the genome are vulnerable to off-target cuts, solving this problem will require a combination of experimental and computational approaches.

Another limitation of Cas9, along with some other CRISPR enzymes, is that it requires a specific stretch of base letters near a target site in order to do its DNA-cutting work. This requirement limits its range to places in the genome harboring this so-called “PAM” (for protospacer-adjacent motif) sequence of DNA. To help address this challenge, Liu’s team used a guided evolution system to create new proteins in the lab, producing a new “xCas9” enzyme with loosened PAM requirements. The xCas9 enzyme can reach many more sites in the genome with greater specificity than the original Cas9. Like Cas9, xCas9 can be harnessed to turn genes on, cut DNA, or edit single letters of DNA when converted into base editors.

Others are working to better control Cas9 activity and limit its off-target effects through chemistry. Broad associate member Amit Choudhary, an assistant professor of medicine at Harvard Medical School who holds an appointment at Brigham and Women’s Hospital, leads efforts to develop tools that can switch genome editing on or off, by developing synthetic molecules that inhibit CRISPR enzymes or turn them on. Choudhary and his team plan to then apply these Cas9 control systems to disease-relevant models, for example, making edits only in certain cells, such as beta cells involved in diabetes, or developing containment strategies for anti-malarial “gene drives” in the mosquito.

Adding some new genome editors to the toolbox

To create a system for editing DNA that doesn’t rely on actually cutting the DNA double helix or on unpredictable DNA repair processes, Liu’s team recently developed a new kind of genome editing agent known as a DNA “base editor.” Base editors use a disabled Cas9 enzyme and base-modify enzymes directly convert a single target letter of DNA into a different one.

“It’s been wonderful to see what people in other labs have done with base editing technology,” said Holly Rees, a graduate student in Liu’s lab who, along with Nicole Gaudelli and Alexis Komor, was key in developing the new base editor. “Capabilities like these open up a whole new world of possibilities for life science research.”

Using a suite of base editors created in Liu’s lab, researchers can now correct all four of the so-called “transition” mutations — C to T, T to C, A to G, or G to A — that together account for almost two-thirds of all single-letter mutations known to cause disease, including many that cause serious illnesses for which there are no current treatments.

In another major CRISPR advance, altering RNA rather than DNA is now possible thanks to a class of RNA-targeting CRISPR enzymes first characterized by scientists in the lab of Broad core institute member Feng Zhang, who is also an investigator at the McGovern Institute, an associate professor of brain and cognitive sciences and biological engineering, and the James and Patricia Poitras Professor of Neuroscience at MIT. These CRISPR-associated enzymes, which are in the Cas13 enzyme family, precisely target and cut RNA molecules and could one day be used to treat diseases without permanently altering the genome. A temporary fix aimed at RNA offers a safer way to make corrections in the cell, which could be used to recover lost protein function to treat a disease without tampering with the genome itself.

Zhang and colleagues have designed variants of Cas13 that bind specific sites on RNA without cutting the molecule, serving as precise, programmable platforms for targeting RNA transcripts. The enzymes can be used in a variety of applications, such as fluorescently imaging RNA molecules in living cells or inducing single-letter RNA changes.

Using Cas13 enzymes, Zhang’s team has also developed a rapid, inexpensive diagnostic system that can detect single molecules of RNA or DNA, enabling tests for pathogens, tumor DNA, or any genetic signature of interest. Known as SHERLOCK (shorthand for Specific High-sensitivity Enzymatic Reporter unLOCKing), the system takes advantage of CRISPR’s off-target tendencies by adding extra synthetic RNA that releases a signaling molecule after it is cut by the CRISPR enzyme. The team developed a miniature paper version of SHERLOCK that displays the signaling molecules on tiny strips of paper to indicate presence of a target molecule, which could be developed into a field-ready diagnostic for use during an infectious disease outbreak.

Getting CRISPR where it needs to go

Actually delivering CRISPR tools into the right cells in patients or in animal models is challenging. The protein complex is too large to cross cell membranes. Typically, to get around this obstacle, the DNA encoding Cas9 and its guide RNA is delivered into a cell — rather than the entire enzyme — and the cell uses its own native protein-making machinery to produce the enzyme and guide RNA.

But Liu’s team recently developed a method to package the entire Cas9 gene-editing complex within an envelope of lipids and deliver it directly into a cell. The approach results in much more precise genome editing, and the team, which included researchers from Massachusetts Eye and Ear, demonstrated its potential by using it to prevent hearing loss in a mouse model of human genetic progressive deafness.

Additionally on that front, Choudhary is working to develop miniature, synthetic genome editors that are 100-fold smaller than the commonly used CRISPR-associated enzymes and can therefore enter the cell more easily.

The path ahead for CRISPR

The scientists agree that for genome editors to reach their full potential, they need to be more selective, specific, and efficient. In addition, they each have some ambitious aims for CRISPR in the near future.

“In the past five years, a number of CRISPR-based tools have emerged, from base editors to epigenetic modifiers. Yet, we still lack efficient approaches for precise gene insertion,” said Zhang. “To realize the full potential of CRISPR-based therapeutics, we will need to develop approaches to efficiently excise and replace sections of the genome.”

Zhang added that tools to insert a single base at a targeted site would also be useful, specifically for the thousands of genetic diseases where loss of a single letter of DNA leads to a dysfunctional protein that can cause disease.

Liu hopes that within ten years, in addition to switching out single letters of DNA, scientists will have the ability to change virtually any DNA sequence of any length into any other DNA sequence in a living cell or animal, with high efficiency, high DNA specificity, and minimal byproducts. “I believe that this science fiction-sounding goal is within our grasp, given the current rate of progress and intense activity in this field,” he said.

Doench would like to see techniques to efficiently deliver proteins or nucleic acids to specific cell types in living animals and people, a capability that would be transformative in the CRISPR arena and beyond. “For CRISPR to be a widely used therapeutic approach,” he said, “that's one of the nuts that needs cracking.”

Research associate Ruth Hanna, who works alongside Doench, stressed the importance of developing better cellular and animal models of complex disease, so that researchers can properly assess the effects of CRISPR edits on the function of proteins and cells.

The scientists have been developing new ways to use CRISPR-based tools to “knock out,” or prevent the activity of, multiple genes in a cell at once, which enables large-scale studies of gene function. These combinatorial screens can be applied to a variety of diseases and cell types, such as hunting for “synthetic lethal” pairs of genes that, when targeted with drugs, can deliver a fatal two-pronged hit to cancer cells. But first, scientists need better cellular models of diseases like cancer.

“Just having CRISPR tools isn’t enough,” said Hanna. “Building better preclinical models is critical if we want to use genome editing to understand what goes awry in disease and how we might be able to intervene in the clinic.”

In some cases, genome editing itself can help create the improved cell model. Broad scientists Zuzana Tothova and Ben Ebert recently developed a new cancer model that reflects human genetics. Here they used CRISPR-Cas9 to insert cancer-driving mutations into human blood stem cells. By transplanting those edited cells into mice, the scientists built customized mouse models that demonstrated the progression of leukemia. This approach could one day be applied to other cancer types.

Choudhary would like to see more applications of chemistry-based approaches to give current CRISPR enzymes new abilities. “Nature has already provided us with parts necessary for precise genome editing,” he said. “I would love to see chemists putting these parts together in new ways to develop synthetic genome editors.” In addition, he hopes to see tools that use CRISPR to make precise and site-specific chemical modifications of DNA, giving the opportunity to expand the genetic alphabet beyond the natural nucleotides, A, T, C, and G, and generate unnatural proteins or semi-synthetic cells with novel functions.

Zhang is also interested in further exploring the natural diversity of CRISPR systems to discover more enzymes with adaptable features. “It’s critical for researchers to keep exploring CRISPR biology, and natural systems in general,” he said, “because we really don't know what else is out there. There is still so much more to discover and learn.”