a16z Podcast - Journal Club: Using CRISPR to Prevent Coronavirus and Influenza Infection

Episode Date: May 15, 2020

In this episode of a16z bio Journal Club, general partner Vijay Pande, bio deal team partner Andy Tran, and bio editor Lauren Richardson discuss a novel CRISPR-Cas-based anti-viral strategy.The discussion covers the differences between this newly developed prophylactic strategy, traditional vaccines, and anti-viral drugs; how this strategy can be engineered to target a huge range of coronavirus and influenza strains; and the next steps needed to go from paper to practice:“Development of CRISPR as an Antiviral Strategy to Combat SARS-CoV-2 and Influenza” in Cell (April 2020), by Timothy Abbott, Girija Dhamdhere, Yanxia Liu, Xueqiu Lin, Laine Goudy, Leiping Zeng, Augustine Chemparathy, Stephen Chmura, Nicholas Heaton, Robert Debs, Tara Pande, Drew Endy, Marie La Russa, David Lewis, and Lei Qia16z Journal Club (part of the a16z Podcast), curates and covers recent advances from the scientific literature -- what papers we’re reading, and why they matter from our perspective at the intersection of biology & technology (for bio journal club). You can find all these episodes at a16z.com/journalclub.

Hello, and welcome to the A16Z Journal Club. I'm Lauren Richardson, and today we're talking about a new approach to prevent viral infection. Specifically, we're discussing a recent article published in the journal Cell, titled Development of CRISPR as an antiviral strategy to prevent SARS-CoV-2 and influenza from the labs of Marie LaRusa, David Lewis, and Stanley Chi at Stanford. Our discussion covers the differences between this newly developed prophylactic antiviral strategy, traditional vaccines, and antiviral drugs, how it can target a huge range of viruses and the next steps needed to go from paper to practice. I'm joined today by A16Z general partner Vijay Ponday and bioteam

partner Andy Tran. But before we begin, here's my summary of the article. Prophylactic treatments are those that you take before exposure to prevent infection. In this article, the authors developed a strategy using CRISPR CAS as a prophylactic against coronavirus and influenza infection. Viruses inject their genetic material into host cells, which then hijack the cellular machinery to generate more viruses. The idea presented in this paper is to use the cleaving activity of CRISPR CAS to destroy viral genomes and thus prevent viral replication. As a reminder, CRISPR CAS is a two-part system. First, CASE enzymes are proteins that cut nucleic acids, either RNA or DNA. In this article, they were using a version of CAS

called CAS 13D, which recognizes and cuts RNA. This is appropriate since both influenza and coronaviruses encode their genome in RNA. The other part, the CRISPR element, are short, customizable RNAs that guide CAS to the correct sequences for cutting. The authors called their system Pac-Man, and it targets specific regions within either the SARS-CoV-2 genome, the virus that causes COVID-19 or the influenza A genome, earlier strains of which have caused past pandemics such as the Spanish flu and H1N1 and is responsible for many cases of seasonal influenza. The authors express CRISPR CAS and lung epithelial cells, which are the main targets of these viruses, and show that they were able to inhibit viral replication and

infection. So that's the high-level summary. Let's start with the big picture of what's happening with COVID-19 vaccine versus treatments. We are in this global arms race, if you will, to really develop therapeutics and vaccines as fast as humanly possible. There's been so much hype and anticipation to a lot of the drugs. One thing that is important is that these drugs are not miracle cures or silver bullets. What is really important to note is that to really get us back into having some peace of mind to return to normalcy, we really need ultimately to have a vaccine or prophylactic for the broader

a healthy population. The challenge for certain viruses, and we've seen this in other cases beyond COVID, is that some viruses are really hard to hit with vaccines. One that I've dealt with research wise is dengue. Dengue is a deadly example because it has antibody dependent enhancement, ADE, such that the second time you get it, it's that much worse, and it's really hard to come up with a vaccine for it. Yeah, the general idea is that your body creates antibodies against a pathogen, but instead of being effective antibodies, they actually help the pathogen. So the second time you see it, like in the case of dange, it actually helps enhance infection as opposed to preventing infection. So there is some whisperings in the coronavirus literature that this may be

happening, which would be a real blow to vaccine development. Yeah, definitely would make corona that much more difficult because many of us may have a primary infection already with minimal symptoms. And we may feel like, okay, now we're in great shape, right? For some viruses, it would mean we have antibodies, and therefore the next time we'll be easier. ADE turns this upside down. And actually, if it is true that coronavirus has this effect, then the worst is yet to come, unless we have some sort of way to handle it. Yeah, certainly makes the goals of serological testing the opposite of what you would expect. Instead of identifying people who have antibodies that could return to the workforce, you're identifying people who might be now extra susceptible to a secondary coronavirus

infection. So both vaccines and this CRISPR-based approach are prophylactics, but they have very different mechanisms of action. So let's start with just the key differences between what a vaccine does versus what this CRISPR prophylactic does. For a vaccine, typically what you do is you have some dead part of the virus that is injected into create some sort of antigen, something for your immune system to recognize, such that you develop antibodies and those antibodies suppress any future live infection. And this approach is tricky in that you have to think about what antigen you want to use, and that's sort of what the design is.

But really, it's harnessing your own immune system to do the work and just kind of catalyzing it. And other innovative area is RNA vaccines. So vaccines where instead of giving an antigen, you give an RNA sequence that actually then your body then uses to produce the antigen, that is of value to your immune system to rev it up. On the other extreme, you have antivirals, like Tamiflu. Tamiflu is a small molecule. It's like a traditional drug that inhibits an enzyme that the virus needs. And by doing that, it kills the virus. What we're talking about today is interesting because it has some aspects of both.

It's prophylactic like a vaccine. But it has the complexity or at least the sophistication, something more like a drug, especially in the sense that it's not necessarily directly looking to sort of stimulate your own immune system, but to give you really new functionality to tackle viruses. And in particular, in some ways it makes sense because CRISPR in bacteria evolved to kill viruses. It's interesting to ask the question, could we modify this mechanism that nature has already come up with, not for bacterias to take advantage of it, but for us? It's almost the most natural application for CRISPR. Like CRISPR's been used for so many different things. But, I mean, it evolved to be antiviral. And so here we're using it an

antiviral application. Yeah, no, George Church has this great line that CRISPR isn't really gene editing. It's genetic vandalism. Yeah, that's right. Bacteria don't want to edit an infecting virus. They want to destroy it. There's a very natural sort of analogy between the CRISPR part being the hardware, the infrastructure, the platform, and the biinformatics part being the software. And that you can imagine as new viruses come up, once you sequence the viral genome, if they're already sort of caught in that conserve region, that's great. You have stuff already off the shelf. If they're not, then it's just a bioinformatic question to find a different conserved region and then just put that new payload in. But it's not starting from scratch. It's not discovery. It's literally just taking your chassis doing essentially almost like a software update for the new thing you have to deal with and rolling out that update. And this is a pretty dramatic paradigm shift from the way we traditionally think about either vaccines or therapeutics. One of the really exciting things about this work is that with this CRISPR-based approach, you can engineer a guide that specifically target certain regions of the SARS-CoV-2 genome. What they actually did is perform biofininformatic analysis that aligned a lot of the published sequences.

And then what they were able to find is that just by creating six CRISPR RNA guides, you can target about, 91% of the sequence coronaviruses. And that's really important because a lot of these regions that are conserved among broader coronaviruses, like the SARS, MERS, genus, for instance. The use of these multiple CRISPR RNAs is really a two-fer. You can increase the range, the number of coronaviruses that it can target, but you are also safeguarding against escape and mutation and resistance

because viruses are super fast to mutate. And so if you were just using one guide RNA, that the virus would only have to change one or two base pairs so that it was now resistant to that guide RNA. If you use six and they're targeted in these super highly conserved regions like they have selected, it's so much harder for the virus to evade all of those. So that's a real benefit to this approach is both the number of coronaviruses that they can target and the prevention against a rapidly evolving strain. Yeah. And it's beautiful the work that the paper showed. Six CRISPR RNAs cover 91% And literally, if you have 22 CRISPR RNAs, you can cover, you know, basically 100% of targeted coronaviruses.

And obviously, there's a lot of engineering issues to iron out. But once you have this, you could literally have, you know, one prophylactic that someone can take to cover, you know, all future coronavis. So this article is definitely a proof of concept paper. They were in cell culture. They were using lentivirus to deliver the CRISPR nucleases, the guide RNAs, and the SARS cov2 fragment construct. And lentivirus is a gene deliver. system that works great in the lab, but would not be feasible in a clinical setting. So I'm curious what you guys think the next steps are that are needed to validate the feasibility

of this approach. It's really important to test out the genome editing dynamics, kinetics, editing efficiency, all in a live context. They demonstrated the ability to cleave SARS-CoV-2 fragments and then also reduced replication in human lung epithelous cells, but they didn't have the authorization to handle the dangerous virus, so it's not a live virus. Yeah, one of the things I really enjoyed about this paper was that they were not only looking at SARS-CoV-2 and coronaviruses, but they were also looking at influenza. And they showed that this CRISPR prophylactic would work against a live influenza A virus. So that really helps kind of bolster the feasibility of this approach. Another key thing, of course, it's a CRISPR modality, which is

plagued with CRISPR-related issues, which is off-targets and the like. And of course, delivery is always a big challenge for a lot of CRISPR systems. How do you actually develop the appropriate delivery system or vector to really safely and efficiently get it into a good amount of human cells for this prophylactic to be actually feasible and usable would be key as well? Delivery, it's another opportunity for engineering and there's a lot of different avenues one can take. Since COVID has especially creates so many respiratory issues, there's a potentially natural form of delivery through the lung, and then in principle, maybe this could be breathed in. You can almost imagine the day when before you leave the house to go to the grocery store,

you take a hit off your CRISPR inhaler and go about your day instead of wearing a mask. And what do we think about safety for these things? And safety for delivery and safety for the CRISPR part. The CRISPR part seems relatively straightforward because it's going after RNA. It's not editing your genome. And it could be the delivery part. Now becomes the real question, can we come up with delivery where it could be done multiple times in a way that's safe? So A.V is typically the gold standard for gene therapy delivery. These adeno-associated viruses are approved gene therapies right now. These are what we call one-and-done treatments because your body will develop immunity against this

capsule. If this all works out well, this could be a universal coronavirus or flu vaccine. But if you think about readministration every single year, you probably want something that is less immunogenic. The other big problem is immunogenicity of CRISPR nucleases. A lot of the common CRISPR nucleases that we use do come from pathogenic sources. There are people that are developing new forms of nucleases, you know, CASX, for instance, that might have come from non-pathogenic sources that could be used in the less immunogenic fashion. CRISPR cas nucleus is definitely like a hardware platform, if you think of it in the molecular sense. All of the concepts that they showed here in this paper, you can apply it not only to CAS 13, but also different nucleases, or even lipid nanoparticles, non-viral approaches, different amphiphilic peptides

that are also shown to deliver CAS9 nucleases really well, could all be combined here. And so we can also think about how do we engineer better CRISPR system that utilize the same platform. But a lot of the stuff and the foundation that they develop can be plug-in-play. Do you think that the hurdles that we just discussed with safety, delivery, immunogenicity will be easier, faster, or more feasible to clear than the hurdles facing a vaccine for COVID-19? Or is it just anyone's guess right now? There's a couple different scenario. So one scenario is where antibody dependent enhancement, ADE, is actually a real serious, difficult problem that can't be cracked. if that's the case, then this is looking pretty good in comparison.

There's another scenario where a COVID vaccine becomes much more akin to like an influenza vaccine, and so a more traditional approach is work, and so then this might be harder. Unfortunately, with COVID, a lot is still unknown. What I'm thinking about is not just planning for what we can do to help COVID in 2020 and 21, but what would we do about the pandemic that could be in 2020? 2030 or 2035, if you look at the timing between these pandemics, you think about SARS and MERS and COVID, the years between them are becoming fewer and fewer. If that's the case, having a broad spectrum sort of programmable-ish approach that could be brought out very quickly.

That's particularly intriguing. Though validating this CRISPR approach might take longer and might span this current one, you know, if we get this moonshot right, we could dramatically save time for all future pandemics, and we could basically sidestep this really linear path of vaccine development. Okay, so we've discussed the scientific hurdles that a therapy like this would still have to overcome. But assuming they were overcome, are there strong business models and incentives for prophylactic treatments? Or does it suffer from some of the same headwinds as antibiotics and traditional vaccines? Yeah, I think if you think about the modern record of producing vaccines, it doesn't really inspire that much confidence. Because if you think about SARS-Cove-1,

MERS, Zika, Ebola, all really provoke these similar arms rays to make a vaccine, if you will. And to date, only the Ebola effort has been successful. And the vaccine was approved basically last year, five years after the epidemic really happened. Well, Lauren, you talked about comparing this to antibiotics. The thing about antibiotics is that we intentionally don't give them out because we want to avoid resistance. And that's what's made the economics of novel antibiotics so challenging. because if you have a great antibiotic, it goes in lockbox and doesn't get used.

This is actually upside down for a couple different reasons. One, it inherently is engineered to go after resistance and that a few mutations here and there are not going to make a big deal. And then secondly, I could imagine that it would be the type of thing. Instead of going in a lockbox, it would be heavily manufactured and distributed such that everyone would have it available so that I think there would be a real commercial advantage of doing something like this. So for those two reasons, I think this becomes almost the opposite.

of what we're seeing in antibiotics. I think, you know, thinking about this whole economic context, governments, an institution spends billions of dollars every year on nuclear weapons that they hope to never use. How about we spend a couple billion to build these plants and teams to equip ourselves to handle the next outbreak and pandemic? I think that's an excellent point. So it turned it into a national defense issue

and not just a health market health demands issue. It's interesting to think about what can we do either to have a response already, or to engineer something rapidly in response to something being a threat. I think the old-style sort of military top gun like war where it's going to be our fighter jets against other fighter jets becomes less and less of a reality, bioterrorism becomes probably a much more insidious threat.

I mean, these vaccines remain the best and virtually only weapon against these viruses and bioterrorism. So it's going to be a mission-critical defense mechanism going forward. So one last thing, and this maybe is more sci-fi, it's interesting to ask, could sort of a CRISPR approach create a true broad spectrum antiviral for like all RNA viruses? And especially given the nature of viruses and how they spread within a population, the ability to tackle these things early means that they don't spread, which means that we don't have these crazy pandemics anymore.

That would be the ultimate fantasy. Thank you both. And thank you for joining the A16Z Journal Club this week. To recap, this research shows that it is possible to program a CRISPR-based system to target both coronaviruses and influenza genomes to prevent infection. There are a number of challenges still to overcome, especially as these results are only in a cell culture model. But there is huge potential here to create a broad range prophylactic treatment for viral infection, and advances in engineering biology will help take us there. We'll continue to discuss related themes in other A16Z podcast episodes.