Matice Founder Jessica Whited on Harnessing Regenerative Species for Medical Breakthroughs - Ep. 198

0:10Hello, and welcome to the NVIDIA AI podcast. I'm your host, Noah Kravitz. Innovations are born from all sorts of origin stories, but how often do you hear one that involves a mother's love for her child, the axolotl salamander, and deep learning? That's a rare combo even for this show, but it's part of the lore behind Matisse Bioscience, a regenerative medicine company founded to harness the power of super regenerators, species in nature like the axolotl salamander that are amazing healers and possess some important similarities to humans. Matisse is using AI and super regenerators to help humans heal from injuries without scarring and to reimagine wound and skin care.

0:52With me today are Jessica Whited, a regenerative scientist at Harvard University and the co-founder of Matisse, and Matisse's director of bioinformatics and research, Mark Manessis. Jess and Mark are here to explain the link between axolotls and human skin care and talk about how AI is helping them advance the field. Jess, Mark, welcome and thanks so much for joining the NVIDIA AI podcast. Thank you, Noah, for having us. Yes, thanks, Noah. So can we start with the origin story I kind of alluded to? Jess, we were talking before we hit record that even though I read about it elsewhere in the media, as a parent, I kind of wanted to make sure it was okay to ask you about an injury your son suffered, but you gave me the okay, So maybe we can start there.

1:38Yeah, sure. So I am a regenerative biologist. And after I got my PhD at MIT, I decided that I was going to do postdoctoral studies. And I really wanted to go back to the drawing board and ask myself, what did I think were the biggest open questions in biology in that time period? And to me, the regeneration of large complex body parts was one such question. It's really fascinated scientists and non-scientists alike for forever. but it's been a hard problem to crack for logistical details that we don't have to get into here. And I decided that I would go work at Harvard Medical School and do a postdoc, which is a thing you do before you become a professor.

2:17And I established a colony of salamanders there. And then I moved on to start my own lab initially at Brigham and Women's Hospital. And we are all focused in the academic lab on understanding how salamanders regenerate legs at really a molecular and cellular and genetic level. And this gets me out of bed every single day. Like I'm super excited to understand it and to uncover these mysteries and to be a part of this community, et cetera. And in the longterm, I actually really do believe that if we don't know how this works as humans, then we will be in a much worse position for imagining future therapeutic approaches to stimulate limb regeneration in humans.

2:57So that is like, in my mind, the Holy Grail or one of them of regenerative biology and in regenerative medicine. And that's really what we're after. And it occurred to me that it is really also very cool to think that as part of this limb regeneration program, for example, these animals also don't scar and it's, it goes hand in hand, right? So like super regenerators have amazing regenerative prowess, but they also don't scar. And humans, well, like we're not so great at regenerating things like whole limbs, can't do it naturally. And we always scar when, when we get an injury, but I want to put an asterisk by that word always.

3:36And we'll revisit that later because when we're fetuses, we actually don't scar up to a certain point. So I was interested in this and I knew that this happened and we had bumped into the anti-scarring connection in our own work in the lab, but it wasn't what we were center of the place studying in my research lab. Okay. Everything changed one day when it was a day in July, several years ago, and I have twin sons, and they were eight years old at that time. And it was a Saturday. And actually, we just submitted a revision of a paper. And so I was super pumped, I'm going to have a great weekend.

4:12And my kid went out for bike ride. I wasn't with them. They were their dad and he was on a new bike and he actually just sort of lost control of the bike. And luckily there wasn't a car involved and he was on a bike path, but he just couldn't puzzle his way out of how to turn the bike or actually stop it. And he went head first into sort of like a wall. Yeah. Yeah. And so I got the text message and I was, I just got back from the gym, but I was at the grocery store picking up the groceries for the evening and it said, you know, this happened, we're in an ambulance, he's going to need a couple stitches, but it's, it'll be fine.

4:50And of course, my mom had exploded. And I was like, wait a minute, no, you know. So I rushed the hospital, the local hospital, and I saw him and I was just, you know, having a meltdown already about like, what happened? Is he going to be okay? Is he in that question? Is he going to be okay? Is you can really unpack that in many ways? Like, is he going to be okay? Is it life-threatening? And he definitely had like some severe injuries and a concussion and a broken nose, et cetera. And then it goes to the, is he going to be okay? Like what's the permanent damage to his face, you know? And once I established that, yes, it stinks that he has a concussion and he ended up going to Boston Children's Hospital and they confirmed that, et cetera.

5:34And that there was a surgery that he was going to need to repair the inside of his space and they hooked us up with a plastic surgeon and all of this was great, but I just couldn't shake that. Oh, he's got all these cuts like right in the center. Right. And then he was going to require surgery right in the center. And that was going to leave another mark. Right. And so instantly my, my mom instinct was to say, you know, I've been working on this problem for a long time in academia as a part of a, of a research team. And all of this work that we're doing is really exciting. But there's some more obvious, I think, shorter term benefits to humans that could be reaped from this kind of research that we're doing.

6:13And that's when it just hit me that now is the time to do this, like sooner rather than later. So when I was much older than your son, and I'll keep this very brief, but it was moving day. My now wife and I were moving in together for the first time in Brooklyn, in New York City. And long story short, I snapped a piece of Ikea furniture trying to pull it out of the moving truck. It hit me in the face, blood everywhere, nowhere near, no internal damage or anything like that. But went to the ER and needed stitches. And my mom, I called my mom because that's what you do. And my mom is a retired nurse practitioner.

6:54And she suggested to me that I ask if there was a plastic surgeon who could do the stitches because that mom's instinct, like I was going to be okay, but okay having, you know, a scar on my upper lip. I was so grateful for that and felt so lucky my mom could do that. But now I'm sort of imagining you and we can laugh because it's much later and everything's okay. Imagining you saying, hold on, I can regenerate his face. I was kind of like, oh, it was more like, oh, crap, why haven't we already invented it? Right. Right, right. You know, so that's how Matisse was born. And, you know, that feeling that I had really does fuel my everyday work with respect to the company.

7:36Yeah. I am really driven to make products that can help people and that, you know, people like me could have went to the store and instead of just buying, you know, the stuff that's currently out there and a bandaid, I could actually buy something that might have biological activity and actually influence the outcome. Right. So, yes. I hope somewhere there are Mother's Day cards with, you know, hand-drawn you as a superhero. Not to pump you up here, but, you know, this is where these stories come from. So, super regenerators, are the axolotl, is that the sort of whatever the name is, the apex predator equivalent of the super generators, are, you know, regenerating an entire limb?

8:18I understand that we don't know how that happens, but talk a little bit, if you will, about this class of species. When we get into vertebrates, which means like they're a lot closer to human beings from an evolution standpoint, you'd be hard pressed to find something better than a salamander, I would say. Now axolotls are a kind of salamander and they're very popular right now for many other reasons, you know, like Minecraft and axolotls are cute. And unfortunately they're critically endangered in this and that. But the reason why we prefer to use axolotls as opposed to other salamanders is because they actually have a shorter generation time than many other salamanders, but it's still actually quite long compared to other research organisms in modern biology.

8:58So it's still about a year. And just for reference, like a fruit fly, which is what I worked on in grad school, that's two weeks, a mouse is about six weeks. So while you can do things in a salamander, an axolotl, like knockout genes, for example, in order to actually get a whole lot of animals that have the same mutation, that's still like a multi-year endeavor. But that is the reason why we choose axolotls over many other salamanders, which have basically all been shown of the ones that it's been studied in to regenerate limbs. And salamanders can also regenerate tails, including the spinal cord, which we know humans cannot, parts of their heart, parts of their brain, you know, their gut, a friend who studies ovaries, lungs, like a lot of different tissues they can regenerate.

9:42So I think that's one of the reasons why I call them a super regenerator. Yeah. When they regenerate, are they not carbon copies, but are they functionally 100 percent, you know, to the capacity of the of what they're replacing these different body parts? And yeah. Yeah, they are. Now, I would say that, you know, from a scientific standpoint, we're also still just sort of peeking behind the curtain about what might be molecularly different from the original one. And how many times can they do this? And like, are they using the same cells every time, et cetera? Like those are more nuanced questions that I think a lot of that the jury's still out on.

10:17But when you see the animal who has regenerated a limb, you cannot tell which limb was ever amputated and it is functioning and using it. It's articulated. It's got innervation, et cetera. So it's basically perfect. And we want to know how that happens, like exactly how that happens at the molecular level. I feel selfish. I want to ask you about how this translated into forming the company, but I have a million questions about regeneration now. How common or is there sort of a percentage ratio of species in nature that do regenerate versus those that don't? And really what that's getting to is I'm trying to sound a little bit smarter than just asking, why don't humans regenerate limbs or parts of our hearts or brains?

11:02Yeah, I would say that actually regeneration is much more common across the animal species if we take a holistic view of it in the species that have been studied. And then there's a whole lot of species for which we actually don't know the full extent of their regenerative abilities. But certainly humans and other, let's say mammals, not all, but most other mammals are known to have very restricted regenerative abilities when we're talking about a large body part. But we also have to remember humans have a lot of regenerative capabilities when we're thinking about like individual tissues, like our skin, for example, or the lining of our gut.

11:37And these kinds of tissues have cells being born all the time and they're getting replaced, right? And then what I alluded to earlier, which is that fetal humans, as well as fetal other kinds of mammals, for about half of fetal life, we don't scar. And then we transition into the scarring mode and the jury is still out. We don't really know for sure why. And there's a lot of speculation about the reasons why, and some of it is supported by some kinds of experiments. But I would say that we still don't understand all the reasons why this happens. We can talk about the most common explanations, and they relate to, oh, maybe humans have like an enhanced immune response that makes it impossible for this to happen naturally, and maybe that can be overcome.

12:22Another possibility that people like to talk about is humans might have more breaks on the system because we have evolved to possibly have more what are called tumor suppressors, which are kinds of genes that basically keep cells in check from multiplying too much. And there are some similarities between the limb regeneration program and tumors, but you know, the truth is we don't really know the reasons why humans cannot naturally regenerate limbs. But the premise is that since we can regenerate skin and as fetuses perfectly, that we can sort of consider this a latent capability. And that if we could just stimulate the human cells and give them the instructions in some way to go back and be able to do the thing they used to be able to do, then we could potentially affect the outcome in a way we care about when we're talking about scars.

13:14And so is that kind of Matisse's operating principle, for lack of a better way to say it? Yeah. So Matisse's operating principle is that we can use science and computational biology to go in and take these huge data sets from super regenerators, including axolotl, but not limited to axolotl. We can sift through all of them and make some predictions about what the molecules are that are most likely to impart this non-scarring outcome or this outcome of enhanced healing, less scarring, let's say. And then what we do is we synthesize them chemically. So basically we're making them in a lab and then we are bringing them into the lab space in Cambridge and there they get tested on human cells to see if these molecules are giving the instructions to the human cells to tell them to do things that would be pro-regenerative anti-scarring.

14:15And from that, we can tweak it further and we can pump all this information back into the algorithm. Right. So I felt like this is where I'm supposed to make sure the audience knows that you're using, is it synthesized salamander molecules? Yeah. Yeah. You could say that. So it's like we're actually using a peptide-based approach. And so a protein is an example of a peptide, but usually the way we use it in, in biology, the term peptide often can mean a shorter version of a longer protein. And so what we're doing is we're taking and we're predicting what proteins are likely to be having these effects that we desire.

14:53And then we are imagining the parts of the proteins that will impart that effect. And then those protein parts are these peptides and they're, you know, there are certain amino acids are in length. So amino acids are the building blocks of peptides, like bases are the building blocks of DNA. We basically have them synthesized chemically, and then we import them into the lab. So you mentioned having a massive data set, which is where the algorithm comes in, the AI. You're gathering the data from studying the salamanders in the lab? Right. So we're gathering the data from data sets that have already been generated for other purposes.

15:34Yes. So like the data sets that are in the world about how salamanders are regenerating limbs, for example. And then let's talk about what you do with the data then. Right. So I'm happy to talk a little bit about that. And then I would love it if Mark could offer his perspectives. But basically the kind of data that we're looking at is the data about which genes are expressed. So meaning they're being transcribed into mRNA, which is basically the instructions for making proteins, right? So we're asking which genes are turning on and turning off and in certain cells that we care about during this process.

16:12Adding on to that, we do take a multi-omics approach to this. So we not only look at the level of expression, the mRNA expression and a fine green resolution such as a single cell, as well as more larger resolutions that's just like a bulk expression. We also look at different combinations of information, such as epigenetic control, as well as translatomic information. So, respectively, the epigenetic control, right, that shows us which different types of genes are actually accessible during the time of wound healing, as well as, you know, for the translatomics, we see which mRNAs are actually being transcribed to proteins during these wound healing stages.

17:01Because they could be giving out these mRNA signals, but they might not necessarily be being translated into proteins. Right. And I'm probably going to get this wildly wrong, but is the goal to figure out which proteins, I'm going to get the language wrong, which proteins are being transcribed to and then to figure out how to synthetically replicate that in human molecules? Right, right, right. So the idea, yeah, so the idea is to take this multi-omics approach, right, to basically identify first the protein, the proteins that are most important during the wound healing stages, right, of different regenerative species, right?

17:41What are the crosses between these regenerative species as well as differences between non-regenerative and regenerative species to identify the proteins that are really regulating this kind of wound healing program and then taking those proteins and turning them into peptides to be tested within the lab. So we've had a couple of episodes on the show about drug discovery, AI and drug discovery, and I was actually reading about it recently. It makes me think of that there are similarities in the process to some extent anyway. Are there? Yeah, yeah, yeah. I think it's a very, very similar process.

18:22I think you've also, you know, with the advent of a lot of next generation sequencing technology, I think there's very similar approaches to kind of taking this multi-omics approach to sequence different genetic information to elucidate these like molecular mechanisms that lead to wound healing. But I think what, you know, one of the few things that really kind of set us apart is first, we combine this kind of multi-omics approach with, you know, this really deep understanding of the biology, which Jessica has first kind of started scratching the level. She's an expert in axolotl regeneration, and so is her lab, right?

19:04So this is one of the big things that they really fundamentally know about. And second, we combine also trying to identify the different chemical makeups that determine what a wound healing peptide would look like. So we don't only kind of gather data from the axolotl, but rather we also try and figure out the structural and chemical information necessary to define what a subtype of a wound healing peptide would look like versus a non-wound healing peptide. And then are you able to use an algorithm, use computer modeling to sort of virtually try out different peptides to predict what effects they would have on a human?

19:56And again, my reference point is, well, besides just talking about algorithms, probably my reference point, like I said, is drug discovery. But it almost seems like there's kind of two, there's more than two, but sort of two parts of the process we're talking about now. Now there's understanding, you know, processing the data and looking for ways to understand what's happening in the oxalotl and these super regenerators on the protein level. And then trying to figure out how do you sort of translate that is not quite the right word, I know. But translate that into an approach for, you know, positively impacting human skin.

20:35Right. Absolutely. Right. So, yeah. So those are pretty much, yeah, the kind of two approaches we have. which is first finding that protein and then kind of defining this subset of proteins, right? And so, you know, to kind of test this out in Silico, we have kind of, you know, a couple of different algorithms that we've developed to try and do that, which is first we've used some supervised learning methods to take all of the data sets we've generated from testing all of the different peptides we've had. So, you know, we look at all of our successes and our failures in our lab. And we basically try to correlate all of those successes and failures to basically the structural information of the protein, as well as the genetic information of that protein.

21:29So, you know, what we would have that is this kind of statistical model that says, given this set of structural and genetic information, right, we train a statistical model to do this, what type of output would it have in the test that we run in the lab, right? So, yeah, so that's one of the supervised methods we do, but we also have different kind of unbiased methods where we try to see basically from all of the successes we have, what type of peptides do they actually look like, right? What types of kind of features do they have? And can we basically try and put these features and bias these features in the creation of our peptides?

22:11How would you describe your success so far? What stage of the journey are you at? And is the end product, a sort of like a skin cream, a topical of some sort? Sure. I can take that one. So, and then Mark can add what he would like. We have already discovered several peptides that actually quite a few that act to do the things that we really care about in human skin. And three of the things that we care about are the speed at which cells move to cover a wound surface. So after a wound, there's going to be some removal of cells in the sort of most superficial layers of the skin called the epidermis.

22:52And so you want to encourage cells to sort of migrate in there and reform that barrier, which is really important because it's one of the main functions of your skin as a barrier, but also it's important for maintaining like, you know, water content underneath the wound, for example. And then the other thing that we've been solving for is the amount of inflammation and the kind of inflammation. So humans tend to have this like profound inflammatory response that lasts for quite a while after a skin injury. And you will notice this because there's reddening, for example, and sometimes wounds can also be infected.

23:30And salamanders tend to shift into this anti-inflammatory response a bit sooner. So they have an inflammatory response that's actually also critical for regeneration. but they shift into an anti-inflammatory response a bit sooner. And then the third thing is that in the slightly deeper layer of the skin called the dermis, which is underneath the epidermis, there are these cells that live there called fibroblasts. And fibroblasts, one of their main functions is to make the collagen and other extracellular matrix molecules that sort of like hold your skin together. And after an injury like a cut, for example, a lot of these fibroblasts will be cued, molecularly cued to turn into a different kind of cell called a myofibroblast.

24:11And myofibroblasts have some kinds of like tensile properties. So they're kind of contractile, but the important thing for the scar is that they spit out a different kind of collagen that has a disordered sort of appearance. And that is one of the reasons why as the scar matures that you can notice that his texture is different, for example. And that's one of the key things that the human eye can detect about a scar, right? The redness and the texture, especially. And so what you'd really like to do is to prevent the formation of myofibroblasts. And so the salamander is really great at doing that.

24:46And so if you are giving them instructions that say, Hey, fibroblasts, stop what you're doing. Do not turn into myofibroblasts. That to me is like an ideal kind of molecule. Right. And we can talk a little bit about, about that process. Cause this is called fibrosis when especially when you're talking about uh fibrosis that happens inside the body and that's a longer term target for the company because a lot of that is also driven by the formation of these myofibroblasts so any peptides that we find for topical application on the skin which is a shorter term path to actually human use could in the future possibly be repurposed and a lot more tests would have to be done and that's much more highly regulated to determine if they could be used inside the body to combat organ fibrosis.

25:32Yeah. Well, so I'm glad you brought that up because I was wondering kind of starting with, is, so you're, are you targeting as your first sort of product, a topical treatment of some sort? Yes, absolutely. And yes, that's, yes, that's because of where the company started from, but also it's strategic and so far it's like, we can get that to market a lot faster. Right. And, and is there a timeframe you're looking at? Um, so this sort of depends on several different, um, more business aspects, variables of the company. Um, suffice it to say, like we've, we've already done some tests with humans and we also have been working through formulations for a while now.

26:10So like I have some on my dresser at home, for example, that checking out, um, because there's a lot more than just the peptide that goes into this. And you definitely want to, you know, of course we're testing it on humans and looking to see like, does it actually establish that barrier faster after something like a dermabrasion, for example, which is, you know, takes off a layer of the skin and et cetera. So we, we are doing those things in humans, you know, with the ones that we have the most evidence for, and this sort of like later in the pipeline, but in the formulation, there's other things that matter also to people who are going to use it.

26:43Like, how does it feel on the skin and, you know, et cetera, all of it. Right. So we're working through those things right now. And I can't give you an exact timeline about when you'll be able to go to a drug store and pick it up, but in my brain, I'm hoping that it's probably somewhere in 2024, I would say. I'm hoping. Right, right. Great. And then, so then that was kind of a step towards the next thing I was thinking about, which you brought up, which is kind of longer term plans that there are in the science you're doing in your findings in the proteins and peptides and in the techniques you're using, it sounds like there's potential for regenerating more than just skin or the top layer of skin.

27:26Yeah, that is the case. But Mark, you want to elaborate? Yeah, well, I just wanted to add to something previously too, just to kind of be clear, because I think there's a part of this that's a little bit not as intuitive until we discuss about it, which is that why we necessarily turn the proteins into peptides. And it has to do exactly with what, you know, that first, you know, our first market strategy, which is targeting kind of skin concerns, right? Because if you think about it, why not, you know, why do we go through all of this trouble of once we've found the proteins that are really important to these wound healing programs why do we go through the trouble of turning them into peptides right and one of the reasons that we do that is because you know by turning them into peptides allows us for better kind of skin penetration since they are smaller okay and shorter and so that allows us to actually get the peptides you know towards to this to the cell of interest rather than just like the suit, you know, the more superficial layers of the skin.