Founder Stories: Janice Chen, Mammoth Biosciences and Next Generation CRISPR Platforms

Good morning, and thank you to the organizers of Hello Tomorrow for giving me the opportunity to share about our work at Mammoth Biosciences, building the next generation CRISPR platform. I'm so happy to be presenting after Roger, who's given you a wonderful overview of not just the field of biotechnology, but how CRISPR technology has played a critical role in enabling new opportunities within the field. So Mammoth was founded in 2018 with the vision of advancing CRISPR discoveries into revolutionary products that will advance disease detection and treatment. This work really builds off of the foundational discoveries of these technologies originating from the lab of Nobel laureate and my co-founder, Professor Jennifer Doudna at UC Berkeley. And we're building on these foundational tools to really develop the next generation of CRISPR. And you've heard about all of the things that CRISPR has already done in the previous talk, and the field is yet still evolving in a number of different ways. But today I'll take you on a journey of how we use this CRISPR technology to enable life-saving therapeutics and diagnostics. So as you heard from the previous talk, you know, these first-generation CRISPR enzymes such as Cas9 has really transformed and paved the way for genome editing. But there's still fundamental limitations that hold the field back today. And some examples of that are there's still barriers to be able to deliver these molecules directly to the patient for in vivo editing. We've seen examples of where we can take cells out of the body, edit them, bring them back into the patient, and yet it's still a very costly, expensive, and really sometimes dangerous procedure for the patient. And so being able to deliver these CRISPR enzymes directly to organs will be a transformative area for this technology. Furthermore, there's also barriers to being able to use these first-generation enzymes for diagnostics applications. And I'll talk a little bit about that in the upcoming slides. And finally, as some of you know, there is a very complex IP landscape with the Cas9 protein in particular, which can make commercialization in some cases quite challenging. And so all of these fundamental limitations are things that having a next-generation CRISPR system can help overcome. So at Mammoth, we think about this in a wholly new approach, which is really how can we leverage the diversity of nature and uncover what nature has already evolved to develop and optimize novel CRISPR systems. So we think about CRISPR like a search engine for biology. And in this way, we can program CRISPR to perform a number of different functions at a particular site, and we do that by simply modifying the letters of this guide RNA molecule, which is analogous to the search bar. So by simply typing in the letters of the particular site in a genome that we want to target, we can program the CRISPR protein to do a number of different things. For example, we've been able to program CRISPR to read or detect DNA or RNA molecules. We've also been able to program CRISPR to write or edit particular genomic sequences. We've also been able to use CRISPR to knock out genes. And we've even been able to modulate gene expression using the CRISPR system. So in order to expand this toolbox of CRISPR, as I mentioned, we have to go and look at nature and what has already been evolved. So what's really remarkable is that we can uncover these novel Cas proteins and look at ways that have— CRISPR comes from this native bacterial immune system that exists to fight against foreign viruses. And so in this way, we can look at a number of diverse environments and actually collect samples from places like forests, lakes, thermal vents, or even places like a volcano to identify new bacteria that contain these CRISPR systems. And so we use a process called metagenomics, which essentially allows us to get the genomic sequences of these diverse bacteria. And once we get those sequences, we can take them through our bioinformatics and machine learning algorithms to then identify the particular locus that contains a novel Cas protein. So starting from the sequence that exists in a computer, we can then reconstitute those enzymes in a laboratory take them through a high-throughput experimental screening process, and then identify the conditions by which are necessary to make these enzymes function and learn about all the unique properties that these enzymes have. And through this approach, we've been able to uncover tens of thousands of new CRISPR enzymes, such as the ones represented by the families listed here, like Cas14, CasPhi, Cas12, Cas13, all of these completely new properties that have been discovered just because we've been taking this discovery-first approach. And what we've learned through this process is that when we can understand how these enzymes work, it unlocks a number of new applications that have great opportunities within things like biomanufacturing, diagnostics, therapeutics, research tools, agriculture, and so much more. And in a lot of ways, our— the ability to use CRISPR really is limited by our imagination. So I'm now going to focus a little bit on how our CRISPR proteins are ideal for genome engineering. And again, a lot of this work is thinking about how can we use this next-generation CRISPR enzyme to overcome some of the limitations that I discussed before around being able to multiplex and target multiple sequences within a genome. Furthermore, being able to develop these enzymes that are clear of the complicated IP landscape, being able to leverage their small size that will open up new ways of delivery, and ultimately being able to provide a higher fidelity system that can avoid some of the off-target effects that we see with the earlier proteins. So one remarkable discovery over the last couple of years is that, you know, there are these ultra small CRISPR systems that exist in nature. And I think this really changed the way people thought about how CRISPR can work. And so just an example of some of these enzymes that were discovered, we have Cas14. As you can see, it's about a third of the size of the original Cas9 enzyme. And also CasPhi, which actually comes from huge bacteriophages. And so it really makes us think about how these systems have evolved over the last billions of years. But again, what's remarkable is that even though they're small, that, that, that doesn't mean they don't work. And so one of the key things is understanding how do these ultra small enzymes perform in a gene editing scenario. So what we're showing here is simply a head-to-head comparison of the CasPhi enzyme, one of these smaller systems compared to the old Cas9 protein. And we've been really, really excited to see that this enzyme has equivalent or better editing efficiency in T cells. And as you heard from the previous talk, this is a really important application for things like immunotherapy in terms of being able to modify T cells in patients to target cancer cells. Furthermore, we can also use this CasV enzyme for multiplexing applications like the one shown on the right, and that means we're able to target multiple genes within that genome in one shot. And this opens up new opportunities for creating a universal CAR-T therapy, again, really critical for immunotherapy applications. So again, we have a small enzyme. It works very well. So why is that important? Well, again, a major limitation in the field of gene editing today is being able to deliver these molecules to specific organs of interest without— directly to the patient. And one of the key technologies used in this approach are viral vectors. Also known as AAVs, which are extraordinarily powerful. They're also programmable in terms of being able to target specific organs, but they have limitations in the payload that they can actually put into that particular vector. And so the larger enzymes simply cannot fit into this vector, but when we have a smaller system like Cas14 or CasPhi that are able to not only fit within that specific cargo, you can also include other modules to that protein, like additional guide RNAs for multiplexing, different fusion proteins that allow us to do certain functions at particular sites that don't even involve cutting. So again, this opens up a new opportunity for what we can actually bring in terms of therapeutic molecules to the patient. So now I'm going to switch gears a little bit and talk about how this platform has also really revolutionized the field fields of molecular diagnostics. And I think the diagnostics example is just a really beautiful summary of what happens when the discovery of this detection capability in CRISPR has been able to open up a new method for molecular detection. And so this is fundamentally a programmable platform. It has rapid enzymatic recognition by the CRISPR protein itself, very high molecular accuracy, and this translates into the ability to create simple and specific assays. So our detector system is what we call our CRISPR diagnostics platform. And essentially this uses the same key principle as I already discussed, where we have a guide RNA molecule that is programmed to target any sequence we're interested in detecting. And so once that guide RNA and specialized Cas proteins form a complex, they can then find that matching DNA or RNA sequence And once they find that target, they actually become this nonspecific cutter, which is able to cleave these reporter molecules, providing a real-time signal that you found your target sequence. So I'm just going to show a brief video to further describe how this process works. So imagine you're inside a test tube, and you're watching these molecules work. And so here's the Cas protein that binds to that guide RNA molecule, and this complex is now searching through you know, billions of letters of DNA to find that target sequence. So now that it's found that sequence, it makes a cut. But these proteins don't stop there. Now that it's active, it will now continue to cut these reporter molecules that actually produce, you know, a color change indicating that the molecule has been found. So this really is the core technology for how we can use CRISPR not just for gene editing, but also for molecular detection. So today in diagnostic testing, we have a choice. We can have a high cost but also very accurate test like a PCR test. And, you know, now that we're so familiar with testing with respect to COVID, but we also have the option of a, you know, low-cost disposable, but a test that is not very accurate like a rapid antigen test. Fundamentally, what CRISPR has the ability to do is to bridge these two worlds and offer the high molecular accuracy into the hands of patients and consumers. And so at Mammoth, we really think about this entire continuum of testing. All of the testing from laboratories through the home are really critical in terms of supporting diagnostic testing. And CRISPR has the ability to be integrated in all of these different formats. Formats. So within a laboratory, you can use CRISPR to simplify the workflow and increase throughput of testing. And at home, at the home, you can also implement and integrate these CRISPR chemistries into low-cost devices that can be disposable and provide a sample-to-answer result like a rapid antigen test, but with much, much higher accuracy. So Since, you know, the company started, we've been focused on building our CRISPR platform. And really, when the pandemic hit in early 2020, we very quickly reconfigured the detector platform to target the SARS-CoV-2 virus. So in the early days, we collaborated with one of our scientific advisory board members at UCSF who had some of the first COVID samples from the cruise ships that had docked in Oakland, as well as some local community transmission that was happening early in the Bay Area. And so we were very quickly able to demonstrate by simply, you know, reconfiguring this detection platform, we were able to use CRISPR to detect COVID in patient samples. And in fact, this was one of the first studies in the world to actually demonstrate CRISPR for the detection of COVID-19. And that initial proof of concept has since kicked off a number of collaborations and partnerships with academic groups, industry, and other government agencies. To really accelerate the development of CRISPR for COVID-19 diagnostics. So one of the key products that we've been pushing forward is the ability to leverage the simplicity of CRISPR into a high-throughput format that can increase the throughput of testing in laboratories to support, you know, the testing that's required throughout the pandemic. And when we've been able to look at how this high-throughput test called Detector Boost compares to your traditional RT-PCR workflow, we found that this system was able to run a full 96-well plate of samples at half the amount of time that's required for your traditional PCR workflow. And this is really important because it's much faster, it's simpler, you can run more samples in a given time, but also in a clinical evaluation comparing these two systems, we're able to show that you actually don't lose performance. And so this is really an exciting next step for where the technology could be in terms of being able to provide an alternative solution for your traditional PCR testing. In the same way, we can integrate these chemistries onto a decentralized testing platform as I had described earlier. And so these are some examples of early prototypes that we've developed at Mammoth, being able to start from a you know, nasal swab and provide a sample-to-answer result in less than 40 minutes. And while this is still in the early stages of development, we're really excited to continue to further develop decentralized testing, which is so critically important for patients and consumers to be in control of making decisions about their health, not just for COVID, but infectious diseases and other indications beyond that. So I just wanted to spend the last few minutes talking about, you know, really kind of new and evolving situation with the SARS-CoV-2 variants in particular. You know, this is an article from the New York Times from just a few days ago highlighting kind of the unexpected and concerning number of mutations in the new Omicron variant. And what I really like about this article is that it's actually focused on the three-dimensional structure of the spike protein within the SARS-CoV-2 variant and showing kind of the number of mutations in the context of its structure. And so, you know, this has certainly alarmed many, many people and scientists around the world, and we're still learning a lot more about the impact of this variant. But this really underscores the importance of genomic surveillance when it comes to being able to control an infectious virus. And so when we look at a genomic level of where these mutations, you know, this is comparing 3 different variants of concern. First, Alpha, Delta, and you can see there's some key mutations in the spike region of that protein. And you can see strikingly that this Omicron variant contains significantly more mutations within that spike region, in particular that receptor binding domain that interacts with potentially immune response. And so, Understanding how these variants evolve is really, really critical in helping making policy decisions and informing public health guidance to help stay in control of this pandemic. So when we looked at this problem, we wanted to understand whether CRISPR could play a role in helping us more quickly identify the emergence of these two, of these new variants. And in fact, currently the gold standard for genome surveillance involves whole genome sequencing. And while it's incredibly powerful technology that tells you every single base within that genome, it is limited by this long turnaround time, which can typically take 1 to 2 days, as well as the requirement of expensive laboratory instrumentation. And so when we think about CRISPR as a potential alternative, this offers a much faster time to result, just under 2 hours, and detecting key mutations that are present in virtually all circulating lineages of SARS-CoV-2. So in again evaluating how the detector system compares to whole genome sequencing, what you can see here is that the performance of detector in detecting these key mutations actually matched completely with the whole genome sequencing results. Furthermore, we can sort of assign a classification of these lineages based on just 3 mutations, and we're able to show complete agreement with what we had expected with that particular lineage. And so this really offers a new alternative for identifying key mutations and helping inform clinical and public health procedures. So I just want to, in closing, bring it back to our vision of being able to deliver on the promise of CRISPR technology. We are entering a world of precision medicine, and CRISPR is going to play a critical role in that new evolution. And I think You know, at Mammoth, our goals really are to use this CRISPR technology to develop permanent cures for diseases that have no treatment, as well as being able to detect anything, anywhere in the context of diagnostics. And with that, I would like to thank the entire Mammoth Biosciences team and our partners and sponsors for making this work possible. Thank you for your attention.

Hi, Dennis.

Hello.

I'm here with some questions, of course. I don't let you go before you answer all of them, of course. Now, Rutger was already telling us a little bit about the future, but Can you also look into a crystal ball? Where do you see CRISPR in 5 to 10 years?

Absolutely. I think, you know, CRISPR has been a rocket ship, right? This technology has really since 2012 been demonstrated to be useful in gene editing. And not even 10 years later, here we are with some of these, you know, Cas9 molecules actually in clinical trials. So that's sort of the first 10 years is really showing, okay, we have this enzyme that can cut. A particular gene and provide a therapeutic output. And I think the next 5 to 10 years really will be actually bringing these first-generation products to market, but also thinking about how can we use CRISPR in more creative ways, such as, you know, not actually cutting the genome, but in fact just doing precision surgery on a particular sequence. So, you know, there are ways to actually change the nucleotides without actually cutting the genome, which can be potentially toxic, but also just thinking about ways we can more creatively turn on and off genes, again, without the potential concerns around off-targets. So I think all of these different ways, especially with these new enzymes, are going to open up new opportunities, again, not just in the therapeutic space, but also new ways for detection within agriculture. There's just so many opportunities.

And we have a lot of questions also coming from the floor, so let's try to get through most of them in our time. And the time is actually not running, just for For the crew to understand, now the time is running. Can CRISPR-based therapeutics possible in curing diseases which are not monogenetic? Your thoughts about the complexity involved taking on account the off-targets.

It's a fantastic question. So certainly the, you know, the current trials really are targeting monogenetic diseases because there's, you know, one mutation or one sequence. Sequence that is being targeted. And that's sort of the initial proof of concept to show that this technology can be used in a therapeutic application. As we look at polygenic diseases, of course, this increases the complexity for creating, you know, really robust therapeutics, but it can be done. I think it's still early in terms of the safety and efficacy of this technology for other more complicated diseases. It also really depends on understanding the particular biomarker that's important to change in the context of the disease itself.

I was going to say, we have to go through the haze to really see where we are going. Can CRISPR technology be applied to explore mRNA and plasma exosomes?

Yeah, so there's been a lot of studies using CRISPR for RNA editing, for instance. So we didn't talk about that too much, but there's enzymes that specifically can edit and modify RNA sequences. Really exciting space because you're not changing the genome itself, but you're changing this mRNA molecule that can be transient. So there's different advantages of being able to have a transient edit versus something that could potentially be permanent.

And we have, I think, a lot of people from the scientific field. So how well do we know the occurrence rates of false positive target sites among Cas enzymes, including CasPhi?

Yeah, so the off-target is, of course, a really important issue to understand. I think one of the challenges in the field is that there isn't sort of a unified standard for how we measure off-targets. There are a number of tools that have been developed and have been really useful for us to get an understanding of off-target effects, but it's not sort of a mandate for, you know, this is the gold standard for how we measure off-targets. We know that Cas9 has different off-target effects, not just from, you know, creating single, you know, base mutations in other areas of the genome, but there have been studies that have shown there's entire chunks of the DNA that can be removed, which is incredibly concerning from a therapeutic standpoint. With CaspV, those studies of understanding the off-target effects are underway. But certainly in an in vitro context, in a test tube, we know that the off-target effects are quite minimal because of the level of precision that we select for in our discovery process.

And there's always the question about money. How is the cost comparison between classical methods and detector? And have you thought to the application outside of the medical area?

Yeah, so in terms of developing a product that will be adopted by the end user, the economics are absolutely critical part of that equation. So when we have been putting together the concept of using CRISPR for diagnostics, making sure that we're competitive with the best performing assays on the market is a key part of that. So in the second part of your question, I'm sorry.

Oh, okay. Let me just go back. Oh, the application outside the medical area.

Application outside of medical areas, yeah. So agriculture, many of the people here I think really understand the importance of being able to, for instance, you know, create sustainable crops. And so the ability to have a tool that can easily edit, you know, genomes of certain plants to make them more resilient is an area of great interest, not just for Mammoth, but I think the field at large in terms of the applications of CRISPR.

And one last question before we are out of time. Could CRISPR-based diagnostic tests compete with antigen lateral flow tests by its production price?

Yeah, so certainly it's really early for CRISPR diagnostics, and so there's still a lot of proof to be shown in the scale-up, whereas antigen tests have been around for decades. It's been used in many, many different contexts and, you know, has been quite successful in this pandemic. So it's still to be shown, the scalability of where the CRISPR diagnostics platform will be. But in these kind of initial proof of concept studies, we're very excited about the potential to disrupt what antigen tests have been able to do.

Excellent. Thank you so much, Janice, for taking the time and also answering questions. Again, Janice, thank you so much.

Thank you very much.