Good to see all of you. I’m Ursheet Parikh. I’m a partner at Mayfield where I lead investments in human and planetary health. Mayfield is one of the founding venture capital firms in the Valley. It’s been around since 1969. We partner with entrepreneurs like Lucas very early, tend to be seed or Series A, sometimes Series B. But it’s a joy and pleasure for me to have Lucas in conversation today. Lucas is one of the founders of Mammoth Biosciences. It’s been a delight to see them just hit it out of the park and the velocity of how they have developed more and more things like unique CRISPR systems and how they’re enabling that ecosystem. With that, let me start out with something. Lucas has a nickname, and it’s the CRISPR Whisperer. And Lucas, how did that come to be?
Definitely it’s been a very rapid trajectory that we’ve taken since starting in Jennifer’s lab. But it really started maybe back with this concept actually that I think we encountered way back in middle school, which is something you might be familiar with, Pasteur’s quadrant, which is basically this idea of applied basic research. And once I started at Berkeley in Jennifer’s lab, that really was the perfect distillation of this concept, where if you’re doing very basic research trying to understand how CRISPR systems work – in this case protect bacteria – applications just flow naturally. And it’s a really unique section of biology where you can do that basic research and have a lead application. It snowballed as we started to understand this natural diversity. We’re also collaborating with Jill Banfield’s lab, which is metro focused on metagenomics and really just mining diversity naturally.
And then from there we were able to find some really cool new CRISPR systems that unlocked this whole new section of using CRISPR, specifically being able to use them beyond the liver and beyond XPR applications. And that was really one of the foundations of Mammoth, and from there it’s just been basically taking the things that we were doing in Jennifer’s lab, which were very hands on, and figuring out how do we actually build automation to do that on a scale that’s three or four orders of magnitude beyond what we could do as graduate students in the lab. And here we are today, now thinking about how we actually apply those for patients.
I think Lucas is being very modest because I think Lucas and his team at Mammoth have probably been one of the most prolific survey discoverers and developers of new CRISPR systems. I’d love, Lucas, for you to walk us through just that history – where were things in 2014 when the first set of CRISPR companies were formed? And how has the field developed, and what have been some of the cool new systems that you and your team have developed?
I’ll preface this by saying from my perspective it’s still very early days for gene editing and for CRISPR. We’re just scratching the surface of what the promise is. But if we rewind the clock back to 2012, or maybe even before, CRISPR specifically was this scientific backwater that no one in the mainstream media was that interested in. And Jennifer definitely was very focused on RNA biology and how this worked from a basic science perspective. And it really was just happenstance that these tools they were working with were very applicable towards genome editing and that they worked in mammalian cells.
As with a lot of times in science, the first thing that got discovered just got wind behind it and that started to move forward, and that’s really the premise of a lot of these first generation CRISPR companies that are based on Cas9. But that said, with the haste to get those things advancing into the clinic, there was definitely a missing piece of what else was out there for gene editing. And there’s actually hundreds of thousands of different CRISPR systems that just exist in nature. Probably just in the bacteria in this room, there’s probably thousands of CRISPR systems.
What we really decided to focus on was looking at that diversity and saying, all right, statistically the chance that we pick the first system and it’s the best system seems very improbable so let’s actually understand that diversity, see if we can overcome some of those limitations of the legacy system Cas9. We originally focused on ultra-compact CRISPR systems. We’ll probably talk about it today, but the biggest challenge with gene editing is how do you actually get it into the right tissues. And these ultra-compact systems are much easier to get into the target tissues using delivery like LMTs or AVs.
And yeah, from there we started to build on top of that, what we call CRISPR+, actually being able to not just use CRISPR as a pair of scissors, but to think about them as more precision tools that you can use to do epigenetic regulation, application or more precise gene correction. And now we’re really looking towards how we put all these pieces together in terms of delivery technology, in terms of the gene editing systems, to move beyond just editing liver, which is really where when you see all the progress in the news for gene editing. I would think of it more as a good demonstration of the utility of CRISPR, but falls short of curing most genetic diseases, which is ultimately the promise.
That’s really the timeline for the company and for the field overall. And going forward, yeah, it’s about progressively being able to treat more and more diseases that are moving away from rare genetic disease and actually thinking about more common and prevalent diseases that have huge unmet needs for patients.
Let’s take the ultra-compact systems. What is Cas9 size and what would be maybe top two or three examples from Mammoth of the ultra-compact systems? And what would become the applications that unlock these new systems?
The first Cas9 systems, to put some numbers on it, was about 1,300 to 1,400 amino acids, in terms of bacterial proteins, a very large bacterial protein, not compatible with most of the viral delivery methods that we use. The newer systems that we use, a few of them would be CasPhi, which was for us, it’s one of our larger systems, but even that, it’s about half the size of those normal Cas9 systems, so that’s about 700 amino acids. And then our most compact systems go down to about 400 amino acids. You’re talking about something that’s about a third or less of the size of the original CRISPR systems.
What that enables, again, going back to the key challenge of how do we actually deliver on this promise for gene editing, is it allows you to use delivery vectors that we have, namely things like AAV, which have a very, very constrained packaging limit. And it allows you to not just fit the CRISPR system, but fit all the other things that are needed for creating the bespoke therapy for a particular target.
One thing to think about and to keep in mind with CRISPR is that CRISPR is the heart of the therapy, but there’s really a whole ecosystem of other things around that that’s required to get this to patients, from the genetic understanding to the regulatory elements, and then of course the delivery vector. When we think about enabling CRISPR, for us it’s thinking about how do we actually innovate on that core component, the CRISPR system, so that we can better use all these other components that are required to push therapies forward.
Pre Mammoth, most of the CRISPR companies went with Cas9, and they decided to just build the application out of that. But there was also the tendency to want to keep it just to themselves and do it all their own. And one of the core founding premises of Mammoth was that this is going to be a platform that enables an ecosystem like Microsoft for PCs, or Apple with mobile phones. There’s a core set of things you do well, there’s a subset of applications you develop, but you really focus on the few things you do well and then partner with everybody else along the way out on that.
And you developed a wide range of applications. As you were doing the ultra-compact system on edit, what else did you end up finding, like diagnostics, as potentially interesting applications of CRISPR and how they go beyond healthcare? And I’d love to come back to the epigenetic and the CRISPR+ stuff, but while ultra-compact systems are happening, you were essentially developing CRISPR for beyond therapeutics.
At its core, CRISPR is a way to program a protein to a sequence of interest. And that can be in the context of a genome, it can be in the context of in vitro diagnostics. This also goes back to our work at Berkeley and Jennifer’s lab – one of our co-founders, Janice, was very interested, again, in the basic science of how a CRISPR system cuts its DNA target. And in doing that, and running radio-labeled gels and very old school biochemistry techniques, we were able to start to see that some CRISPR systems have this feature where they don’t just cut the target that you program them to, but they turn on this collateral activity that allows you to detect a signal. That led immediately to development of CRISPR for diagnostics, which was a field that really hadn’t been contemplated before.
But beyond that, when you think about it in the genomic context, we’re largely focused on human health internally. To Ursheet’s point, the applications of CRISPR go well beyond that. Of course many of the companies that you might hear about are using CRISPR for one application or another. Strain engineering is definitely a key area, especially when you’re talking about more exotic organisms that don’t have the robust molecular biology tools that we have for things like OI and as well as agriculture. I think agriculture is actually where we’re going to see probably some of the biggest and most immediate impacts for gene editing.
And yeah, it is important for us as a company to enable other groups to do this and really build a system of applications beyond what we can do internally. Because even with our rapid growth, there’s only so much we can do, even within genetic disease there’s only so much we can do, and so making sure that the tools that we develop get disseminated more broadly is definitely a key thesis for us as a company.
One of the things that in the early days of Mammoth was to align on the values and the mission, and the idea being that if the world had X number of products without Mammoth, with Mammoth it should be 10x or more. And that could only happen if we have an ecosystem and platform kind of mindset, which was actually pretty different from what most people were thinking about at that time. Going back to building and developing new CRISPR systems, you spoke about size. What are some of the other attributes that have made these CRISPR systems better?
As we think about CRISPR, the 1.0 technologies were just making a cut in the genome. In order to be able to do other types of edits, you really need a repertoire of different systems. Size is one of the foundational things that a lot of things are built upon, but specificity, especially as you move to some of these other applications, is also a very, very important one. When we talk about treating diseases that have some kind of standard of care already, those safety burdens become higher and higher, where there’s really no tolerance for any off-target cutting that might happen. Mining through the natural diversity, we’re able to find systems that are much more accurate.
And yeah, another key thing that we think about is this portfolio approach. When you think about Cas9, there’s a lot of applications that have been built on that, but there’s a lot of applications that you can’t do with that system. Generally we don’t want every company to have one little area that they can work on, but for us, we actually have the breadth of tools to be able to widely use these systems, whether it is for epigenetic editing or for editing plants or for diagnostics. But you have all of those systems under one roof and you can actually choose the best system for the job as opposed to just being stuck with whatever you started with.
I think that is a very interesting and important attribute. Because I think that fundamentally works with the platform model. If you are developing an application, in this case an application to get you a diagnostic or therapeutic, you want to be able to use the best tools available to you to solve the problem, rather than only the tool that you had licensed out. And structurally that is why partners find Mammoth to be a really good partner, because they know that the innovation engine at Mammoth will keep on finding more tools. And if they’re trying to solve the problem for a specific disease, they will be getting the best tools.
In contrast what has happened is that there have been companies that have taken each version of the incremental development in CRISPR and have become dedicated to it. And from an application developer or a therapeutic company perspective, that becomes a big challenge, in that they may start working with a partner only to find that the version of the tool they need is actually sitting with somebody else.
And then I think from an investor perspective, either private investors or public investors, you end up having to worry about whether you are picking the right tool, or the values of the people who created it. Are they in on this company, or are they onto that next version or the next shiny object? And this is fundamentally this thing about building things to last for the long haul. So now, Lucas, introduce CRISPR+. Where do you draw the line from all of that and how that is working to where the future is headed?
CRISPR+ is a term we came up with a few years ago to really capture this evolution of CRISPR beyond making double stranded cuts as a genome editing tool, recognizing that most diseases cannot be cured by just making a break in a genome. A lot of this is thinking about CRISPR as this method to search out a sequence, really this honing system, and to fuse on different modules that are going to change the effect that you’re getting. This can be things like epigenetic editing, where you’re actually permanently changing the methylation of the genome, which can silence genes, or turning them on. It’s also more precise methods to not just correct things that you can knock out, but actually write in sequences that you want. This is really one of the key pieces of how we develop this to more broadly address genetic disease.
One of the dirty secrets of the CRISPR+ field though is that for a lot of these companies, you’re taking a large system, a Cas9 system, and you’re now fusing on even more machinery, even more baggage onto those systems. You end up with these systems that are so enormous that they really have pretty limited applicability except in the ex vivo applications.
It’s really been a natural transition for us as a company where we’ve got these very, very small CRISPR systems which leave open all this payload in traditional delivery vehicles where we can now fit that machinery in and efficiently deliver it to tissues in vivo. And it wasn’t the original or only reason that we developed those small systems, but being on the forefront of that, it’s really positioned us to be able to push those forward into the clinic.
As you look at the crystal ball going forward, what are some of the diseases that you think you will have therapies in the clinic in the next three to five years?
Of course in the field overall, most are closely watching and anticipating that we’ll have the first commercial sickle cell programs, again, a very exciting milestone for the field, but still in an ex vivo application. We’re seeing a lot of activity around rare genetic diseases in the liver. But really where we as a field want to be moving is towards these much more prevalent, serious unaddressed diseases, especially as we think about degenerative diseases in the brain, from the CNS or skeletal muscle or the heart.
Eventually as we get these proof points, especially with rare disease, it’s about moving more towards preventative medicine, things that maybe aren’t as acute but that we know 10 years down the line or 20 years down the line, you’re going to have some kind of adverse health consequence. Necessarily we need to build up the safety profiles of the systems before moving into those, but that’s really what gets me excited. And for me a lot of the initial programs for both Mammoth and otherwise are really to try and develop the systems so you can go and go after things like curing Alzheimer’s or Parkinson’s, things that many people have had personal experiences with.
It is definitely very, very fabulous. Lucas, what’s the secret? How is it that Mammoth’s able to do these things so much better? And clearly the conversation that everyone’s talking about is generative AI, a lot of large language models and how AI is going to try to change everything. And I’d love to hear a little bit about what enables your success in this research and research product area. And how much of it is technology, how much of it is culture? And what kind of technology?
A big part of it is the team that we’ve been able to assemble, as well as the commitment to innovation. Going back to the early commentary, there is a tendency for some companies to take something that’s working and just harvest it, just think about how to monetize it. We definitely take a longer term perspective, and this isn’t just the internal management team. But it does require a cohort of investors that are supportive of that and that are thinking on long time scales of how we make an impact on the species. That’s the foundation.
And then to think about it more technically, a lot of it’s just the science that goes on internally. Going back to how we were doing this even in an academic setting, a lot of it is very, very creative approaches to finding the systems, the models that you’re using. There’s of course a lot of excitement around AI recently, but most of us know in biology that these genomic data sets are really built for machine learning and it’s just something that just happens naturally, that we use these tools and develop these tools. That’s been a key part of it.
As with any AI or ML approach, it’s about having very clean, good data sets to start with. And that really was the foundation, is this metagenomic engine that started this, and now it’s much more around protein engineering and continuing to evolve and adapt systems. But no, technical things aside, a lot of it is the culture of the company of really what we set out to do and keeping that vision in mind as we continue to grow and not letting our success distract from the long term vision.
I think you’ve brought up the culture and the long term mindset we’ve made. Changing gears, what has been your growth journey as from scientist to entrepreneur, manager and leader? I’d love to get some look back on your journey and both what are your learnings and what is your advice for anyone that’s really take the benchtop science and translate that into real products?
It’s definitely been an interesting and awesome opportunity to be able to do this. Most of the time that people spend in graduate school doesn’t prepare them to actually manage and interact with people. You learn the science, you’re a technical expert, but a lot of it is progressively, and by no means are me or any the founders perfect at this, but progressively getting better and better at orchestrating the team and taking that scientific knowledge, using that to actually motivate and build credibility with the group.
It is a challenge, and not all PhDs make it through that hurdle. And one of the big areas that probably we should adapt is the education system. The graduate education system is a much bigger focus on just communication generally. It’s important for entrepreneurship and of course leading teams and building teams, but it’s really important for of course anything we do pretty much, and just also educating the broader public about what the opportunity is of the tools that we’re building or the broader synbio community. For us it’s just been also of course working with mentors and having coaching, being open to feedback and really taking that feedback to heart.
What is your process when you are trying to recruit or bring in someone? And how do you start with a benchmark of excellence for the role? Because there’s so much that you don’t know. In our opinion, a company’s fundamentally limited by the learning ability of the founders. Independent of the role they take in the company, their influence is really broad, and it does require that they really develop a broad aperture and effectively bring their own infrastructure for learning very quickly. I’d love to hear what your personal process is, and then what you look for.
Early on in the company when I didn’t have much experience in hiring, my initial instinct was to just hire the people that seem the smartest, just the technical experts. But if you just push a bunch of technical experts together without actually curating that group, it very quickly crumbles and falls apart. A lot of it is hiring, maybe indexing less now on that and indexing more on is the person actually passionate about the company? Are they going to be committed? Are they agile and able to learn quickly? Because most of the stuff we’re doing, you’re not going to find someone off the shelf who’s done it before. And then of course are they going to work well within a group and actually be able to build something bigger than themself as an individual.
Often in interviews, I’ll ask someone to teach me about something that I have a good sense of. And how they respond to that and how eager they are to really dig into what they know is a good signal for how well they’re going to be able to adapt to a startup like Mammoth.
We have about 10 minutes left. I can continue asking lots of questions, but if any of you guys have questions, raise your hands.
Thank you very much, Lucas. Now I’m quoting Jennifer. She said that a cure is not a cure if people that need it cannot afford it. Which are the bottlenecks next to get the point that everyone can afford CRISPR therapies?
Definitely it’s not an easily solvable challenge for sure. It is important to have that perspective over the lifetime of the technology as well. Build the roads to your first model and then try and build more afterwards is definitely something that I think resonates here. But it’s not just a question about the intervention. A lot of this is that the healthcare system is built to have this chronic care model where you’re giving someone a pill every morning or giving them an injection every month. And CRISPR does have the potential in terms of the curative things to actually condense that treatment cycle into one treatment. The biggest cost factor is going to be also about the regulatory strategy of how do you get multiple successive things into the clinic without having to reinvent the wheel.
Because we’re building this as a platform, the roads to the first program, again, will be expensive, but the successive ones should be able to stand on that. The FDA, they’re very smart people and they’re trying to figure this out and grapple with it. But there needs to be some kind of platform approval process that also allows you to move into rare and rare genetic diseases. No silver bullet, but of course, as everyone knows in healthcare, the cost is usually not the cost of the therapy, but it’s actually all the investment that goes in the front. And a lot of that is driven by what’s required by the regulatory process approved that it’s safe. I’d say thinking more creatively about that, especially after we have these initial successes, is going to be important to drive the cost down of these therapies.
Can you expand on that at all on the regulatory side? Is the FDA engaging?
They’re willing to talk about it now. It’s still too early in the technology life cycle for them to be thinking about that because it’s still this one-off thing. But even if you look at one genetic disease, you often have mutations scattered across the disease target. It’s really rare that you have just one perfect mutation. And that’s of course what most of the programs are focused on. What you want is a way to get a broader approval for the platform, and then with CRISPR being a platform, you can just change out the guide. And having some kind of minimal safety requirement of actually understanding that that guide is actually the same as in terms of safety from the perspective about targets and things as the other guide, and then being able to get a leg up on the program and actually more broadly address the patient population.
It’s still too early, and whether it’s the FDA or MedSafe or EMA that drives this, it’s going to play out of course as most regulatory policies do over the course of years and years. But yeah, all of that said, the access issue is definitely important, but there’s also treatments that need to be delivered immediately, and that’s the first step that needs to be taken even to broaden that access subsequently.
When I’m looking at regulators, this is what I advise everyone to think about. Recognize that they have a hard job. They don’t get to market fast. They do end up having to take a lot of the pain for anything that’s done prematurely. Having said that, they went through this not because they want to stop you. Essentially most of regulators, they’re scientists, they’re engineers. They care. They want things to get to people. That’s why they entered that field.
We do make a common mistake that a lot of people do, is not talk to regulators early enough. And because if you really think about it, one of the best things you can do is develop relationships. If there is negative news, you might as well get it. You want to go ahead and engage in that kind of conversation. And over time you do see a good change in mindset emerging. And often what happens is I think emerging companies actually make that as a bigger mistake than some of the larger companies. Because the larger companies will often have a lot of dedicated people for talking to regulators so they get to input things.
The advantage you have as an emerging company is that your senior leadership and your best scientists can actually engage in the conversation. The regulators actually like talking to you because they are not talking to layers and layers of corporate BS. They are actually talking to people who will acknowledge, help them understand by breaking the core issues down around that.
Yeah. Lucas talked about how AAV imposes delivery constraints on how big the Cas9 can deliver. Could you alleviate that by using mRNA? mRNA and LNP, now they’re talking about delivering self replicating DNA.
It’s a great technology. The challenge there, especially when we’re talking about programs in the clinic, is it’s really constrained to the liver. And it’s a pretty small subset of diseases. The liver is a great place to target. It’s this production place for a lot of things. But that said, it’s still pretty constrained, and viral delivery, specifically AAVs, are still really the gold standard for everything ex peripatetic.
If I’m completely honest, delivery from the technical hurdles is more challenging than gene editing. We build something that goes and cuts a sequence in the genome, and that’s great for us that we’ve got the easy task. But delivery is really hard to get systems that work effectively and safely across different models as well. Our perspective has been, we’ll be agnostic in terms of the delivery technology. And as, LNPs mature and hopefully you can move ex peripatetically, we’re ready to use those. But what we have today is that our delivery vectors, in terms of the treatments that we want to push forward, let’s take what we’ve got, versus compounding risk with a bunch of new elements that are going to be more risky altogether.
Well we have time for one last question.
If you could spawn one CRISPR system that trumps the current limitations of the other ones, which would it be and why?
Well I’m biased, but I would say it’s our nanocas system, and largely just because pretty much everything we test about it addresses a lot of concern. Off-targets, the safety of the system, we’re talking about regulatory, is definitely one of the things that gets brushed under the rug, and especially in ex vivo applications. And Cas9 is definitely pretty sloppy in terms of where it cuts in the genome. That’s what scares us about CRISPR, is that you’re making any unintended consequences. The nanocas system, not only is it a third of the size, but also has the specificity advantage, it has the ability to work in all these CRISPR+ applications. But that said, I still think that repertoire approach is pretty critical towards addressing the diversity of genetic diseases out there.
Well, thank you all, guys. This was great. Thank you, Lucas.
