Creating a Toolkit to Accelerate the Development of Gene Editing Therapies

Daniel Levine: Erik. Thanks for joining us.

Erik Sontheimer: Thank you for having me. It’s a pleasure to be here with you.

Daniel Levine: You recently co-authored a perspective piece in Nature, about the National Institutes of Health Somatic Cell Genome Editing program, its effort to accelerate the development of safe and more effective means of editing genomes, and the challenges it’s seeking to address. Perhaps we can begin with some orientation. When people talk about gene editing what’s meant by that term, and how would you contrast it to what people think of as gene therapy?

Erik Sontheimer: Those are excellent questions. What we mean by gene editing is making changes to the actual genome of a patient in terms of its therapeutic applications. Gene therapy is where you introduce a new copy of a gene that has therapeutic value. One analogy I’ve heard that I think can be useful is that if you have a car with four tires, and one of the tires goes flat, gene therapy is like sticking a fifth tire on top of the one that went flat. The flat tire is still there. You haven’t fixed it. You’ve just covered it up by putting another tire on that same wheel. Whereas with genome editing, you’re actually fixing the flat. You’re doing an edit to the native gene itself in the patient. I’m going to pick diseases at random. Let’s say you have a patient with cystic fibrosis in the lung and the CFTR gene is mutated in that patient. There are gene therapies that simply add a new copy of CFTR. That would be gene therapy, but not gene editing. Whereas, if you go in and repair the mutation in the defective copy in the patient that is genome editing.

Daniel Levine: The tools and technology to perform gene editing have been evolving rapidly. What can we do with the technology today and how big a gap is there between what exists today and the promise and potential to use this as an approach to address genetic diseases?

Erik Sontheimer: In terms of what is already advancing into the clinic, there are some very impressive things going on. It is somewhat limited in terms of the tissues that can be accessed and in terms of the nature of the edits that can be made. The tissues that can be accessed is an important issue because mutations that cause disease, in many instances, manifest themselves in a particular tissue. Cystic fibrosis is a good example: lung tissue is where the mutation matters the most. If you were to repair that mutation in some other tissue, like the spleen or something like that, you wouldn’t actually do anything useful. In many instances, the edit is focused on a particular tissue. There are some tissues that are easier to access, although still difficult, to do editing than others based on the delivery technologies that currently exist. With that in mind, there are clinical trials already happening where cells from the bone marrow, effectively stem cells that give rise to blood lineages, that can be edited outside the body and then reintroduced and re-engrafted. This would be for treating diseases such as sickle cell disease and beta thalassemia. Those are already in the works and working their way through clinical trials with some very impressive results. Another tissue that is already reachable, at least to the extent that the FDA is allowing clinical trials to go forward, is the retina. Editas Medicine is a company in Massachusetts that has a clinical trial ongoing in which editing is done in the retina for a form of childhood blindness. Finally, the liver—there are other clinical trials going on, including one from a company called Intellia Therapeutics, where the editing is being done in the liver. These are some examples of things that can be done in terms of tissues being accessed. The other thing that is more doable now, and that we hope to be able to get beyond, is the type of editing in which you take a defective gene and then simply break it. In other words, you disrupt it so that it can no longer be expressed and cause damage. That is a relatively straightforward edit to do. You’re simply introducing inactivating mutations. You’re not actually fixing a defective gene itself by doing gene inactivation in particular tissues, such as the blood lineages, the retina, and the liver. Those are the kinds of things where there’s already a lot of progress that has been made. With regards to the SCGE, the goal is to go beyond those things and go after some tougher targets because that’s where the NIH considered there to be the most unmet need. For instance, there are some groups that are targeting the musculature and there are other groups that are targeting the central nervous system. These are a few examples and there are many more within the consortium. It’s about trying to bring delivery tools to allow new types of tissues to be edited therapeutically, and also to enable the introduction of other flavors of edits that are not just about breaking a gene, but perhaps actually repairing a gene.

Daniel Levine: The NIH Common Fund, in 2018, launched the Somatic Cell Genome Editing Program, what was the need they saw and what was the mandate for the program?

Erik Sontheimer: It saw multiple needs and there are multiple sub-programs within the program. The biggest need that it saw was to address the challenge of delivery. Because again, delivery is one of the main barriers to implementation beyond a few tissues that are already somewhat accessible. In terms of the numbers of grants that were awarded and the dollar amounts of the grants, the biggest chunk, by a significant margin, is to address the challenges of delivery. Those are not the only ones though. In addition, they have separate initiatives for other aspects of genome editing. Even though they’re smaller, they’re still very important core components of the Consortium. One of them is to expand the repertoire of genome editing platforms. Examples of this would be to explore genomic diversity in additional categories of microbes to find, not just new Cas9 or Cas12, but other things that we can’t even anticipate ahead of time, that could be very useful as a genome editing platform. In addition, things like base editing and prime editing, both of which were developed in the lab with David Liu at the Broad Institute, are examples of next generation editing platforms. We at the Consortium want to continue to develop these and develop new capabilities from them. In addition to delivery, we have expanding the repertoire of editing platforms. We have another program about tissue systems for detecting and understanding potential adverse events. I know that’s a bit of a mouthful, but, in a nutshell, what it means is that we want to be able to explore, investigate, and understand things that can go wrong but not have to wait until clinical trials happen. There are platforms in development that are described, for instance, as lab on a chip, organ on a chip, or related things referred to as organoids that can be used as surrogates for human tissue. It’s not human tissue in the sense that it came out of a person’s body, but it has many of the characteristics of human tissue. You can do analyses to understand whether there are problems associated with off-target editing or unintended genomic damage at the target site. So, to recap, we are working on delivery, expanding the repertoire of editing systems, and biological systems for understanding the adverse events. Then last but not least, an important feature of the consortium gets back to concerns that have been raised over the last 10 or 15 years about preclinical biomedical research in general. That is rigor and reproducibility. There have been exercises where biotech companies and pharma companies have tried to reproduce many of the preclinical studies that have been reported in the peer reviewed literature. The frequency with which those reproductions were done, and this was back in 2012, was shockingly low. So, over the years Francis Collins and other leadership at the NIH have been doing many things across different funding initiatives to try to improve rigor and reproducibility. The SCGE is a very unique program in the sense that we are forcing ourselves to do third-party testing of the technologies that are being developed for the sake of reproducibility. The groups that are funded for the delivery program are developing their own technologies in their own laboratories. In addition to doing that, a mandated milestone near the end of year three is to send their materials to a testing center and the testing center has to show that they can also make the technology work and get a comparable level of editing as well. Only if you pass that reproducibility milestone, do you advance to the fourth and fifth years of funding where things move into large animal testing.

Daniel Levine: This is a 6-year, $190 million program. Can you give some sense of the scope in terms of investigators, institutions, projects involved?

Erik Sontheimer: Yes. I don’t have the exact numbers right in front of me. Roughly speaking, I believe there are 45 different grants that are being issued. Some of these grants have multiple investigators and can be across multiple institutions. I believe there are in the neighborhood of 70 or 80 different institutions, nearly all of them are in the United States. There are some participants in the Consortium in Canada and Europe, but it’s mostly based in the United States.

Daniel Levine: The program, as its name suggests, is limited to somatic editing as opposed to germline editing. The effects of somatic editing is limited to the individual whose genome is edited and those changes can’t be passed onto future generations. What’s the significance of that limitation? Is it a matter of ethics, of safety, or is it intended to address some potential public concerns about the technology?

Erik Sontheimer: It’s really all of the above. I believe this is based on decisions that have been made at the highest levels of the NIH with Francis Collins and other NIH leadership. Effectively they have said germline genome editing is something that we do not currently support. The NIH and Francis Collins have joined calls for a temporary moratorium on it. This is partly to understand the safety aspects of it, and partly in recognition that it needs to be a conversation not just about what scientists want to do, but a broader societal dialogue that determines to what extent, if at all, such research will go forward. Yes, you’re right. Somatic is actually in the name, and the program is limited to somatic tissues. What we mean by that is, not only do we not have explicit goals to accomplish germline editing, we are required to analyze whether any unintended germline editing has happened. If that happens, it is deemed unacceptable. So, we are not trying to do it, and we are making sure that we don’t do it without meaning to.

Daniel Levine: CRISPR Cas9 has become synonymous with gene editing, but there were technologies that existed prior to perform gene editing and new ones have emerged. How big of a toolkit exists, and what do we know about the versatility of these tools? How big of a tool set will we need?

Erik Sontheimer: It would generally be the case for scientists in the field that the bigger, the better in terms of the breadth of tools that are available. Everybody acknowledges that there is no one particular platform and there is no one particular tool that is going to serve every need for every type of genome edit in every potential tissue. The more options we have the better off we will be. There is a very substantial part of the Consortium that is focused on CRISPR Cas9. There are others that are focused on CRISPR Cas12. One of the funded groups in expanding the repertoire of genome editing, is Jillian Banfield and Jennifer Doudna at UC Berkeley, where they are identifying CRISPR Cas-Phi, CRISPR Cas14, and various other new flavors with probably more that will continue to come out over the years. The more microbial genomes we sequence, the more we will likely find additional platforms. Then finally, David Liu’s lab at the Broad Institute that developed base editing and prime editing, and there is a group at Yale that is not doing CRISPR at all, but rather a type of synthetic chemistry that makes a nucleic acid analog that can induce editing. CRISPR is the biggest piece of it and CRISPR Cas9 is the biggest piece of that, but we all expect that there will be additional platforms coming forward that we can’t even imagine right now. We want to make sure that we are leaving ourselves the space to identify those and put them to good use.

Daniel Levine: You mentioned some of the clinical trials that are going on right now, as with gene therapies, it appears the eye and liver are the low hanging fruit. Why are these easier to target than cells and other parts of the body?

Erik Sontheimer: The eye and liver. Those are the ones that are truly in vivo editing. In addition, it would be the blood system where the editing happens outside the body, ex vivo, and is then reintroduced. Getting back to the in vivo with the eye and with the liver, let’s start with the eye. There is already a substantial body of research in terms of successfully doing gene therapy in the eye. There is an FDA approved gene therapy, called Luxturna, that is having great success for other forms of inherited blindness. It effectively capitalizes on technology that was already developed for gene therapy, but now uses it for genome editing where the viral vector that you’re introducing directly into the patient by local injection into the vitreous humor on top of the retina, instead of delivering a copy of the gene, you’re delivering a copy of the editing machinery itself. So, it is gene editing rather than gene therapy. There are platforms already in existence for that purpose and they can be repurposed for genome editing. From the standpoint of the liver, it is much the same. There are technologies that are already developed for RNAi therapies as well as lipid nanoparticles. The liver is a place where a huge fraction of the body’s blood flow naturally goes through. If you introduce something by subcutaneous or intravenous injection, a lot of it is going to go straight to the liver. We’ve learned a lot about how to get uptake in the liver. These are places that are best positioned to capitalize on earlier generations of therapeutics simply applied to genome editing. I think that’s the reason why we’re seeing these come along first.

Daniel Levine: As with gene therapies, gene editors need a vector to carry them where they need to go to do their job. You mentioned the challenge of delivery being among the largest. As you think about how big a challenge that is, where does developing the right vectors come in and does the ability to target desired cells and tissues ultimately depend on the selection of the right vector?

Erik Sontheimer: It certainly can. It’s worth noting that vectors, for instance, in the sense of something called adeno associated virus, which is an FDA approved vector for certain forms of gene therapy, those are definitely important modalities. That’s how the delivery is done for the studies in the eye that I just described. There are major components of the SCGE that are trying to improve on these current vectors either to make them more specific for a particular tissue or to expand the range of tissue specificity. For instance, some of these vectors are particularly good at going to the liver and others might be better at going to the musculature. Trying to identify forms of these that can target one tissue or another is an important effort. Another important effort is to overcome some of the current limitations of how much room you have in that vector to code for the editing machinery. Developing new vector capabilities is definitely an important part of it, but many other efforts don’t use viral vectors at all. There are industry initiatives and there are funded groups within the SCGE that are doing the direct injection or introduction of Cas9 protein loaded with a guide. So, there’s no vector at all. You’re just putting the machinery in, in a form that can actually reach a target cell and get into the nucleus and execute the edit. There are others that are using messenger RNA technology. Of course, we’re hearing a lot of that in the form of the COVID vaccines that many of us are receiving. The trial that is ongoing in the liver, courtesy of Intellia Therapeutics, uses messenger RNA delivery. Protein delivery and messenger RNA delivery have a couple of very important advantages, which is that they are much more transient. The editor is only in there for a very short period of time. It’s expected that will greatly reduce the possibility of adverse events arising from immune system activation because these are bacterial proteins that you’re introducing. So yes, vectors are important, but they are not the only game in town for delivery, LNP delivery and messenger RNA delivery are also very much in play.

Daniel Levine: Testing is another important component of the SCGE: the development of animal models and means of testing human biological systems to detect unintended consequences of gene editing. Where are we in terms of having what we need to do this?

Erik Sontheimer: That’s a good question. I think in terms of human biological systems, like liver-on-a-chip and organoids, there’s been tremendous progress. But we’re nowhere near where we need to be in order to understand how tissues and cells will respond to the editing events that we are trying to execute. From the standpoint of animal systems, one important aspect of the Consortium is that it’s a Common Fund program. Basically, what that means is that it comes out of the office of the director of the NIH and it’s not driven by a particular institute that has a focus on a particular set of diseases or set of target tissues. It’s intended to be a program that will have benefit across all the institutes of the NIH. For that reason, they do not specify a particular disease gene target, they do not specify a particular preclinical model, or anything like that. A lot of the animal studies that need to be done, especially in small animals, are using reporter systems that are advantageous for detecting editing events and understanding them. What that means is that if you have a successful editing event, you might activate the expression of a fluorescent protein that you can detect easily by microscopy. Just to cite one example: for the animal systems in question, the initial phase with small animals is exclusively being done with reporters that will help us understand the efficiency and the efficacy of the edit that we’re doing. For preclinical development, the FDA requires a phase that uses large animal systems for both safety and efficacy studies as well. So, the later years involve testing that will be done either in pigs or in non-human primates. There it’s just technologically much more difficult to create and establish reporter systems within a reasonable timeframe. So that will likely not be initially a component of it. It might during later years, but initially it will probably not go through reporters, but through editing of native loci and the genomes of those test subjects.

Daniel Levine: One other important thing the SCGE is doing is seeking to create common metrics and standards. Why does this matter and how will it get others to adopt them?

Erik Sontheimer: A couple things. First of all, this is set up as a consortium. The goal of doing it as a consortium is to try to maximize the degree to which the impact and the advances of the consortium can be greater than the sum of its parts. We have people who are expanding the repertoire of editing platforms and we also have people who are figuring out new delivery routes. That means that it would be a pretty natural thing for somebody who identifies a new editing platform to say: “This group over here at this university is developing a new vector for targeting the central nervous system that would be a perfect route for delivering my new editing platform. Let’s collaborate and see whether we can get the synergies from these different initiatives.” Then, in addition, somebody’s delivery module or editing platform might be really good for one tissue and it might raise the question, could it be equally good for a different tissue? This type of interoperability, if it can be baked into the system to the maximum extent possible, can really accelerate development. Having common metrics and common readouts will allow these things to be made more interoperable and will also allow the comparisons between different systems to be directly comparable. That’s what is meant by standards and interoperability. Then putting all of these together into a single database, referred to as the SCGE Toolkit, which will eventually be public facing, will allow others to go and access that information and that data and enable them to explore these comparisons directly themselves.

Daniel Levine: Ultimately the idea here is to accelerate the development of gene editing therapies. What do you think of the potential for SCGE to do that?

Erik Sontheimer: I think that it’s high. I hope other people agree that it’s high. That’s the whole goal. If it turns out not to be high, then it would be difficult to describe the program as a success. This is publicly funded. It’s supported by taxpayers. It could well be that it’s subject to patents in the usual way, like academic research often is. Public research is intended to be to the benefit of the taxpaying public. That is the goal of the SCGE as well.

Daniel Levine: Erik Sontheimer, co-chair of the Somatic Cell Genome Editing Consortium steering committee, co-lead author of a recent Nature article, and a professor in the RNA Therapeutics Institute and the Program for Molecular Medicine at the University of Massachusetts Medical School. Erik, thanks so much for your time today.

Erik Sontheimer: Thank you.

This interview has been edited for clarity and readability.

Thanks to Pfizer, Inc., Bluebird, and Novartis Gene Therapies for their support of this podcast, part of our Platforms of Hope: Advances in Gene Therapy and Gene Editing series.