Founder Stories: Mathai Mammen, CEO of FogPharma

This is the BioCentury Show. Brought to you by BioEquity Europe. BioEquity is Europe's premier international showcase for financial dealmakers and biopharma executives to meet rising biotechs.

Hello and welcome to the BioCentury Show. I'm Simone Fishburne, editor-in-chief of BioCentury. I am joined by Matthijs Mammen, incoming chairman, president, and CEO of FOG Pharma. Dr. Mammen was previously EVP of Pharmaceuticals R&D at Johnson Johnson, where he headed Janssen's global R&D activities. Prior to that, he was at Merck, And before that was co-founder and SVP of R&D at Theravance. So Matthijs, we've known each other for a number of years. I think it's fair to say you are entering industry veteran territory in the youngest possible sense of that. And I know that you've got insightful, original thoughts about topics in R&D and industry, and we will get to those. But I want to start with your upcoming role at FOG Pharma, which officially kicks off on June 12th. That's right. So tell us, Matthijs, why FOG Pharma, what you were looking for coming out of J&J, and why that company is in the sweet spot for you.

Yeah, first, thank you. Hello, Simone. It's wonderful to be here. I really appreciate it. So why Fog Pharma? So just to step back and put some context around the, the response, you know, in my life there are various things that I'd really like to see achieved in our ecosystem. You know, at the center of that are great medicines, but underneath that there are certain aspirations for certain science, certain medicines, certain impact, and I had an absolutely brilliant time at both Merck and J&J, but I believe that to create some of these more foundational pieces that ultimately would lead to patient impact and to serve a career aspiration of running a company as its CEO, I wanted to go do that. I was willing to start from scratch or take something very substantial, but in all cases, what I was seeking is one of a short list of of achievements that if you could do X, it might change the world. So with fog, there are several aspects to that. One is there's a growing frustration in our field right now with intracellular protein-protein interactions. There are a lot of these that are on the table as high-confidence, well-validated targets, but they immediately go off the list to most companies because they're not tractable, they're not druggable. And if anything, I've seen over the last 10 years that list grow. And it's unfortunate because there are patients that would benefit if you could actually make drugs out of some of these. And I saw with fog the best possible approach that I know of right now to these intracellular protein-protein interactions. And what it is, is it's a helix-based technology that's these small proteins that are designed to cleverly enter the intracellular compartment and deal adequately with proteins interacting with other proteins. Also, FUG has an absolutely wonderful board and set of investors that have the sort of long-term potential that I'm seeking in general. Like, I would like to build a company that's sustainable, goes through all elements of research and development and ultimately commercial, and do it in ways that I believe are forward-facing, very modern, that take advantage of technology and new practices, and ultimately have patients at the very center, have impact at the very center.

So just digging a little bit into the angle for now about running the company and so on, you know, it's considered quite a coup often for a relatively small company to snag someone who's been in a big pharma, especially at the executive team level. But I often wonder, and one day I'm actually gonna do the numbers on this, but I would do often wonder to what degree this really, really creates more successful outcomes. And so, you know, how much of the pharma industry experience translates to smaller companies? And, you know, you've been prior to J&J and Merck in a startup that you created. So you've actually gone in both directions. So how do you look at this? And are there aspects where the large company experience actually doesn't bring a benefit or can be adverse? And how do you think the, you know, your board, and when you're in that situation on future boards, should be thinking about these hires?

It's an absolutely brilliant question, right? Because it's not obvious. There are aspects of being in a small company environment, an entrepreneurial environment that play well in a large pharmaceutical company. Company, aspects that don't, and vice versa. When you grow up and only have very large company experiences, it brings a great deal of perspective on what future issues and problems can be on ways that might be best in class of doing various aspects of R&D and commercialization of products. But at the same time, you're optimizing for different things. Cash per se, for instance, cash flow per se isn't usually as rate limiting or activity limiting an activity in a large company. So to become disciplined on use of resources very efficiently, to sequence activities potentially that you might run in parallel because that is more cash efficient, these aspects are sometimes hard to get used to. But on the flip side, there's, you know, like if you step back, I would say, you know, what prevents like an idea from becoming really high impact in the world? And I think from a large company perspective, a J&J perspective, I see 3 things. One is you might not know what biology to tackle in the first place. Like that's a big target selection is an issue. The second would be that you just can't drug it. So, you know, you know what to do, but you can't get it done and you can't get it done with a drug-like medicine. It's the third aspect that's often in the way that I think a lot of companies that are young companies don't adequately take into consideration, and that is product adoption and attraction that a very useful approved product might make in the market. And that can be because the groundwork hasn't been laid. Maybe the mechanism or the modality are unfamiliar. Maybe the disease is unfamiliar and the sort of finding patients element of that is a major issue. And a lot of that can be accounted for and taken into consideration very early in the R&D process. And I know how important that is. And maybe that's a perspective I wouldn't have if I hadn't been at Merck and J&J for the period I have. And so that's— maybe there are aspects to that that are perspective, knowing who to turn to for various, various aspects that are quite downstream from the discovery of a compound that I hope to bring to the table when I'm at FUG Pharma.

And, you know, let's stay there and talk a little bit about data science, which is an area that you've emphasized as central to drug development. As well as to how a company operates, both at J&J and in various conversations between us and writings of yours. So, you know, you talked a little bit about capital efficiency. Um, obviously you have to invest a certain amount in data science to really yield the best from it. So how do you plan to implement data science in the running of FOG Pharma? And in times like this where there is a sort of capital crunch. Is that a luxury or a necessity?

I don't think it need be expensive, first, can I say. It's more a frame of mind and an openness to do an activity in a new and modern way. And the beauty of a pharma is that there isn't a decades-long established way of doing everything in the company. So it feels like a wonderful, amazing blank sheet with an amazing forward-looking group of people. And data science, as we've talked about before, is not a discovery thing or a development thing. It's a whole company thing. And so, and it's about approaching it as such. Like, are there, for certain problems that companies need to solve, are there appropriate data sets? Are there clean enough datasets? Are there algorithms and data science methodologies that might help do whatever we're doing more effectively? And as I've often said, like, you know, the best thing you can possibly do with data science is to make something more likely to happen that you want to happen. It's that probability of getting to the end that's by far the most important piece. And next down would be doing it faster, and next down would be doing it less expensively. So all three of that same exact framing, um, I'll plan on using at Fogg Pharma going forward. So data science will be very much part of the company, of course.

And, and are there ways that we've sort of covered in the past a little bit in our conversations that this extends even to functions like HR or hiring, or, you know, maybe you can spell that out?

Yeah, it can. Like, you know, at, um, I know at J&J, for instance, the, the company explored a relationship with, uh, a company in California that's effectively a data science company called Findem. And it's just an example, but it uses a new methodology on a couple billion, or actually several billion people in the world, of data that helps infer function apart from what a person might report as their skills, what can be inferred from the totality of data available in the world. So there are all sorts of things one can do with data science, even to upend recruitment. And so those kinds of things I think companies need to be open to in the future. But it applies everywhere, Simone, everywhere in the company that anyone does anything where people lean on others with experience, they use heuristics, they're using some sort of pattern recognition in the world to decide what to do. And I just want to be open to data science methodologies, combined with the right data for being able to do that to assist companies being successful.

So I want to segue from here into something that is, I think, adjacent to data science, but you can talk about what— how you see the differences, which is machine learning, AI. And, you know, I don't think any conversation is complete today without talking about ChatGPT, the democratization of generative AI. On the one hand, it's on everyone's lips, everyone's lips. And at the same time, among some people, there's a lot of eye rolling going on. So I want to ask you, to what degree do you think this technology is going to have an impact on drug development? How much attention should we be paying to it there? What areas can it be leveraged in? And in particular, because I said you've got these two, you've got the eye rollers, you've got the people excited. What's the line between being a Luddite and an overenthusiast in your mind?

Yeah, so, you know, I think the, the, if you look at the J&J experience, it's been very much practical. You know, the team there under Najat Khan and the whole of the team under R&D, and I would say the whole of the team under J&J has been very practical in that there are a selection of problems that make sense to try to solve using data science methodologies. Because again, A, the data exists and B, the methodologies exist. And it's not—

So where are they? Are they early in drug development?

They throw out— They can be all over the place. Like for example, for rare diseases, one very important problem is being able to locate a patient. And so there, there was a very good example. There's an ongoing example finding patients with pulmonary arterial hypertension using ECGs, which are collected in the course of diagnosing these patients. But there, there's a recognition that only a very small fraction of the information in an ECG is actually gleaned by cardiologists and other physicians. And these data science methodologies, these algorithms were able to glean more and point to a patient has potentially having pulmonary arterial hypertension. Like that, there are all sorts of patient finding problems. And this applies both in the drug development phase and ultimately in the commercial phase. You have to know you have an illness. The doctor has to know you have an illness in order to get a medicine. You know, the, the use of data to further define a patient as being responsive to a particular mechanism. This involves among all patients that have a named illness, which ones are responsive to a certain mechanism. To find mechanistically homogeneous populations is a massive job of our entire industry. That is a data science problem that will make drug development obviously more likely to be effective because signal is not lost in noise, A, but B, the effect sizes, if you concentrate all your responders, are going to be higher. You know, the chances that you can prove superiority over some other medication are higher. The overall effect for the company is better as both the business. For the patients, of course, it's better because if you take the medicine, you're apt to respond. And if you're not apt to respond, you're not a candidate for the medicine. So you don't have particular side effects. So that's in the precision medicine element. There are others that have to do with external control arms that lessen the cost potentially and allow more exploration of more standards of care. In a drug development program that can also be really, really effective. These are all drug— I used examples exclusively from drug development and further downstream. I haven't used examples yet from upstream because I think those are still coming. You alluded to machine learning, AI, and large language models. I think they're going to be implemented everywhere. But one of the ways that we're seeing right now it's an AI methodology that created AlphaFold. And that is a phenomenal achievement given, early in my life, I remember following this space of protein folding extremely closely. I was very close to it and I was a big fan of the various people like Harold Ceraga and others that were doing work and trying to understand this problem. And sure enough, This methodology that started with a blank sheet and used transformer architecture in conjunction with a lot of data has created a methodology for predicting structure of protein awfully well. Like that, we'll see achievements wherever there's adequate data and with these advanced architectures. Large language models, as you said, and maybe I'll stop with that, is they're the latest in that this remarkable notion that the entire purpose was to try to predict next words after sequences of words, has somehow led in conjunction with reinforcement learning from human feedback and various methods of doing that to this incredible thing that we're seeing with ChatGPT-4. I think what we're learning there is we think of language as very complex and difficult. There are different languages in the world, but perhaps the structure of language was a bit more accessible than we imagined. And these large language models have effectively shown us that. Other problems of science and medicine, I don't know if they'll fall as easily as language has. But let's see. Let's see what people are able to come up with as the innovation is fast and furious now in that field as well.

Well, thank you for that very, very thorough answer. We're going to take a short break now. And when we come back, we're going to talk a little bit about R&D.

The 23rd BioEquity Europe Conference heads to Dublin, Ireland in May 2023 with a program focused on how to create a new playbook for biotech success in Europe. In 2022, over 800 industry leaders from 30 countries joined BioEquity in Milan, Italy for 2+ days of strategic panels, one-to-one partnering meetings, and C-level networking. More than 150 biotechs selected by BioCentury presented their story to investors and potential partners. Don't wait. Two of the last three in-person BioEquity conferences sold out. Register today and join BioCentury EBD and our Dublin regional host committee at the industry's premier C-suite and investor conference. Visit bioequityeurope.com for more information.

Welcome back. I'm at the BioCentury Show with Matthijs Mammen. Let's talk a little bit about some R&D specifics, in particular where we are with immuno-oncology. So, you know, there seems to be a little bit of dampening enthusiasm in immuno-oncology, and that may be only on the heels of the massive the previous years. So we've seen very few new targets in IO coming out of AACR, and that is a metric we cover each year. So this year, sort of the very few, I think it was only one real IO target, was a surprise to us. I can tell you that anecdotally, people tell us that investors and VCs are less bullish on investing in IO. There is one specific issue we could talk about, which is checkpoint inhibitors. This is something I spoke about previously on the BioCentury Show with Gilead CEO, Madad Parsi. I know you know him. And the idea is, you know, Gilead has a program in TIGIT, but aside from that, despite tons of money and activity in next-generation checkpoints after PD-1, only LAG-3 has had a product come to market. So, you know, what is the difficulty with this next generation of checkpoints? Are we not doing the experiments right? Or is there less there than we anticipated?

I would like to broaden immunooncology as being well beyond checkpoints, though. Of course, fair. Checkpoint is an angle. But stepping back for a moment, we've learned something totally remarkable that has been hypothesized and explored for decades. So there's a long history in immunooncology. But we've learned that inhibiting a particular checkpoint in PD-1, PD-L1 and another in CTLA-4 does in fact augment the immune system enough to recognize tumors as foreign and clear them.

Okay, but we learned that years ago.

We have learned how to implement that right now with the checkpoints. And now I think what that's motivated is a different approach to using the immune system to kill tumors, and that is use of CD3 bispecifics, CAR-Ts, and And I think we're going to see directed therapies from other components of the immune system, whether one puts chimeric antigen receptors in them or learns how to pull them into a tumor. And this, and then the TCR technologies are on top of that. So the whole industry, I feel, has been tremendously motivated to use the immune system right now for cancer. Checkpoints themselves, it's complex. There are various both accelerators and brakes to the immune system that's very condition or environment specific. You can be on the one hand thinking you should be able to find a whole series of these, but on the other hand, you shouldn't be surprised that any of them work that well in a universal sense. I continue to be amazed by the PD-1 data and how broad it is right now that you can employ it broadly irrespective of tumor. I think there, the next great advance is going to be to identify why certain tumors don't respond to like a PD-1 or CTLA-4. And what we've seen and what the companies across the ecosystem are seeing is that there's different reasons for that in different contexts. Why are tumors cold? And so The warming up or heating up of otherwise cold tumor continues to be of interest. In situ immunization, so to speak, like having more antigens to respond to or greater numbers of antigens to respond to does seem effective. We saw some data recently from an mRNA from Moderna with Merck on an mRNA vaccine that elevates the concentration of particular antigen in the tumor microenvironment, and that's been enough to turn certain colder tumors hot and bring in an immune response where there wasn't one before.

So when we look back at this era, is— are we sort of entering a phase now? Because we really definitely did go into a phase of everybody going after the next checkpoints. And yes, they were building CAR-Ts alongside, and we now have CAR-Ts and now we're improving on those. So are we sort of entering a phase now where we're sort of still working out with the tools that we've got and some that we're optimizing now how to really harness this? Is it a phase of understanding the biology or understanding how to construct the clinical trials? Should we not really be expecting big breakthroughs for a while as we flesh out these various details of how to properly harness it?

I think we're in the hard work phase of things now. PD-1 and CTLA-4 are clearly big breakthroughs. Breadth of response, et cetera. But I think to what I was referring to as warming up or heating up a cold tumor, I feel like there's not going to be one magic solution there. And we're going to see different experiments that try different things there across the board. And that's going to be a hard work phase of getting immune checkpoint inhibitors to be useful. The other is in the clinical trials, we still don't have, I mean, we have microsatellite stability instability. We have the beginnings of a tumor mutation burden. But in fact, like, the immune system does need to see several or more antigens to respond to in a tumor. How do you know which antigens, which antigens to promote, might depend on your HLA context as a person with the tumor? That's another hard work phase for the industry to try to have precision biomarkers for immuno-oncology. That doesn't really exist right now in a, in a high stringency way. We have these looser pointers that a patient may or may not be responsive. And then the use of individual components of the immune system to enact function, I think, is continuing to be a very important area. But even there, as you know, it's early, early days. Like when we transfect T cells with CARs, it's not one T cell type, it's Many dozens, maybe hundreds. And what the role is of each of these in both efficacy and unwanted side effects is still gonna— it's gonna take years to play out. So the fine-tuning hard work phase of CAR T-cells that are autologous right now still will continue over the next decade plus. And then off-the-shelf treatments will come with time, but those are a lot harder to do than I think was originally hoped.

Well, talking of hard work, I'm gonna jump now to neuro. So cancer is obviously the biggest, well, maybe not obviously, but it is by some way the biggest draw of investment dollars and of, you know, pharma programs and so on. Second after that, by most of our analyses, is neurology. But the bulk of that and the bulk of the enthusiasm there is in neurodegeneration. And no question that we've got a big social burden coming our way, in particular with Alzheimer's disease there. But neuropsychology has not had— neuropsych has not had the same kind of investment. We all know that there's a massive mental health burden across this nation, across the globe. And I don't see the same amount of fervor there. And I want to— I know J&J did do some work in that area. Some people tell me reimbursement, reimbursement is the problem. And so companies aren't motivated to go down that because that's the problem. Others tell me trial sizes, lack of— we just don't have the biology yet. There seem to be as many reasons as not to do it as people can find. So help me think through when this paradigm gets changed.

Let's talk about this because there's not, again, like one thing I'll say that if you could do X, it will revolutionize.

Somebody would have done it, I guess, if that were the case, right?

There are several problems, and I would put the same is true with the area of infectious diseases like antibacterials. So there, in both those cases, let's handle them together if you don't mind. There's the problem of being imprecise in how one runs a clinical trial. In depression, we take all comers effectively that have various forms of depression and put them in one trial. We know that they're not mechanistically homogeneous. We know that for sure. Just the same, by analogy, if you have a lung infection, even if you know it's bacterial, you should not— you cannot lump all those as one infection, but yet we run at best case gram-positive versus gram-negative lung infections. These are old-school ways of doing things. They're the opposite of precision medicine. In both those cases, you should be able to bring precision medicine to the table. For neuropsychiatry, as you know, Bill Martin and the Janssen team have been trying to do exactly that. He was doing that actually in his prior company as well. But the basic notion there is try to identify with a particular mechanism, a mechanistically singular form of a condition. So it need not be just drawn from depressed patients, but it can be from any number of patients with a neuropsychiatric condition, so long as they respond to the same mechanism and try to find a response signature that identifies them as such. And within the J&J context, they have something like that is my understanding now, and they're continuing it for kappa opioid receptor agonists and other mechanisms as well. So the world right now for psychiatry, the solution is going to be identifying a precision signature so you can practice precision psychiatry, and then you will commercially be able to run a superiority study versus today's agents. That's the only path I think is that precision approach. And once you're superior, just like happened in oncology, you'll remember back in the day, oncology was never going to be an interesting field because there was too many generics that worked reasonably but not great. But how to ever show that your agent was better took the revolution of precision medicine. Same exact things are going to happen in psychiatry and same exact things can happen with infectious disease. If you knew for sure that you were going to be unresponsive to today's cephalosporins and so forth, and you knew that your patient that you were defining was a potential responder to your new antibacterial, you could again run superiority studies. Non-inferiority studies in psychiatry and infectious disease, not inferior to something that costs like a quarter a day. I mean, there's not much of a commercial, you know, attractiveness to that, and the patients aren't better off either. So in both those cases, precision medicine is the answer. Easier said than done to find those mechanistically homogeneous populations, but that has to be the goal.

We're almost out of time, but I want to end on this question because you've talked very persuasively about precision medicine. You've talked about some advanced technologies in particular in cancer, although we know that they're going to extend beyond that. But I want to talk about— end on this question of how these fare against the mission to increase access, get more medicines and more trials into community practices and community hospitals. In particular, I think a really worrying example perhaps came up in an analysis that we published last week about the ODAC on AstraZeneca's Lynparza for prostate cancer. And it came up that only 50% of prostate cancer patients in academic medical centers are routinely tested for the mutation in question. In community centers, that number's 30%. So it really seems that we're creating these medicines and these opportunities, but we are not comprehensively and certainly not equitably enabling their uptake. So what needs to happen to change that?

It's such a massive passion area of mine.

That's why I thought I'd end on the line.

So thank you for that. We have to be able to, you know, we have to identify patients for medicines. But being practical again for a moment, I think the first wave has to recognize that patient treatment, physician practice is not going to change overnight. So you have to look at the tests and work as they're done today and do your very best to use the data that's being collected to point those physicians to the right medications. That's, that's job one. And so to use the example that you just gave, you know, it's one of the reasons that motivated the J&J team and other teams in the, in the industry right now to look at histopathology slides to infer mutation status. Because as you said, it's only a small fraction of patients that are fully sequenced. Unfortunately, it's a low number in the United States. It's a low number in Europe. And in community centers, it's even lower. And in the world, it's even lower. But 100% of people, if they're biopsied in their cancer, do get pathology slides made. So to the extent, just like the ECG example and PAH before, to the extent that these are data-rich offerings, but we're only gleaning a very small fraction of the information, to that extent, let's do our very best to make sure that we're inferring as much as we can from those big information-rich tests that are being run, like a pathology slide, and try to point people at particular therapies. I actually hold out a lot of hope for that with the advent of machine learning methodologies. If there's enough training data, it's absolutely stunning how much data is in these ordinary tests that are being run that we're not actually using. And that will hasten the delivery of products into the patients that actually need them around the world. And let me end on this last piece. And it's a little bit of a vent, but it's the Inflation Reduction Act piece. I do think that the volume of medicines in the world need to continue to have maximum impact, need to still be oral products. You need to be able to get to oral products to impact billions of people easily. I'm sure that some will argue against me there, but I believe that. And I believe that all the biologics and all those are massively useful and will continue to make a big impact in the world. But the biggest of all will be pills and oral products. So it bothers me a lot that that aspect of it right now, given your question around global access, is going to be more disincentivized given the IRA. And we have to try to do something about that.

Well, thank you very much. I know we could talk a lot more and we probably will. The IRA is definitely something that we are covering at BioCentury in, in great depth. And yes, there are a lot of concerns about that particular aspect. And it's going to be interesting really to see whether and how companies adapt, whether there is— there's obviously going to be some pushback on the law, but that doesn't mean it will actually be effective. So certainly we feel like we're at an inflection point at the same time as the biology and the technologies are unfolding. This, as ever, has been a fascinating conversation. Matthijs, thank you very much. We look forward to getting you back another time.

Thanks very much for the opportunity. Thanks, Simone.

The BioCentury Show is brought to you by BioEquity Europe. BioEquity is Europe's premier international showcase for financial dealmakers and biopharma executives to meet rising biotechs.