Founder Stories: Mark Fishman, Founding President of Novartis Institutes for BioMedical Research

Thanks, Jessica. Larry, thank you for inviting me here. I still have no idea what this conference is about, but I really appreciate participating in it. Your phenotype is the constellation of all the facets of your body and your mind, all the form and function. And it's assembled by your genome, your genetics, your environment, and a certain stochastic component of your embryonic development. So it's complex. What we know is that when you have extreme phenotypes that are very unusual, especially if the onset is at a young age, more often that has a tremendous genetic contribution. And what I'd like to do today is to illustrate how we might use those kinds of often genetically driven extreme phenotypes in both the discovery of new medicines and how to think of— use those to think about areas we've had very little success in, such as behavioral and psychiatric problems. I'm going to begin with the translation to new medicines. Now, it takes a very long time to go from a fundamental discovery to a new medicine. We actually analyzed this. I looked at all of the important new medicines from the discovery that people said was responsible, the basic science discovery, to the medicine. How long do you think that the median time was? 17. So here is 17. How far back did you go? We took all the modern medicines. Yeah, 25. 30 years is the median. Now, the one that I like the most was more than 100. This shows that in the picture on the— the side here, Hales, who's the first person to measure blood pressure. He did it by putting a pipe into the neck of a horse. All right, so one could say that our current antihypertensive regimen, which probably half of the audience is on, began in 1733. More people think that it began with the discovery of renin in 1898, which is a hormone that's secreted by the kidney that makes your blood pressure go up. And the first The first blockade of the renin system was synthesized in 1974 and was approved in 1981. So now you say, well, why does it take so long here? And it's an interesting question. I bet a lot of you think it's because pharmaceutical companies are dumb. That may be part of the story. But it turns out that most of the time is taken way before it even gets to a pharmaceutical company to begin to work on it, because we don't actually know that that biology is related to a disease. So our strategy in pharmaceutical discovery for many years was hampered by the fact that you knew what was only obvious to everyone at a certain time. And that's why there were only 5 really new drugs produced each year for many, many years because easiest was just to copy somebody else. So in principle, you need a strategy, a better strategy. Now I learned this one of the first days when I showed up at Novartis. Their base is in Basel, and I took the plane over. I was jet lagged. And they said to me, We were going through all the divisions at that time. So there was the pharmaceutical and there was me, the scientist. And they said, and everyone's going to give us their strategy. I had no clue what they were talking about. None. And so, but I had one advantage. I didn't even know what they meant by strategy. I'd been at Harvard for all those years. You don't have a strategy. You might have some tactics, but you don't have any strategy. In the case of what I assume was buyer's remorse, the board of directors, after they had hired me to run the new Novartis Institutes, asked me to go to Harvard Business School for 3 months. This was a near-death experience for both parties, I should point out. But I did learn one thing, which is that if you're in business, Or if you want to— I don't know if there's students here. If you want to start a consulting firm and someone asks you for a strategy, all you have to do is put up a 2x2 chart and then say, I'm going to be in the upper right-hand corner. So literally, as I'm walking across the room, I'm saying, OK, I got the chart. What are my axes for a strategy for drug discovery? I said, well, probably should meet an unmet medical need. There should be a patient who actually needs this. So it made sense. And we should probably have some understanding of the disease mechanism so it becomes tractable. And this turned out to actually be very effective. And it seems so obvious to me— so I actually published this, and several venture capitalists told me afterwards that they based their entire portfolio around this. I said, you gotta be kidding. This is pretty obvious. But the reason that it works is several. First of all, can you tell what's not on here? What's left off? OK, think pharmaceutical company. Thank you. Money. It turns out that we were— turned out to be the only research group in the industry— only research group, and I actually asked around— that did not evaluate the value of a project on its assumed monetary importance. Now, I didn't do that because I didn't know how to do it, but also those— what you call a net present value, what's the value over the lifetime of a drug— completely arbitrary. The assumption when they first made statins for cholesterol they were going to make $50 million a year because these drugs were hitting a very important part of the body's metabolism. When they hit $9 billion, people still were saying, well, God, I don't know where that came from. But the fact was, we cannot predict often how valuable they would be. So one thing I left out of here was that. The second was that it forced us to focus on rare diseases. Why? Because at that time especially, they were meeting unmet medical needs. People didn't make medicines for rare diseases. And yet, they were often the ones where we understood the disease the best, especially those where there was a genetic underpinning. I gave this to our team, and I asked them to evaluate all their projects. And sadly, in a— we had to kill 30% of the ongoing projects because they didn't meet this very simplistic algorithm. Why were they doing it? They were often copying other drugs, so no need. Or sometimes they'd be making drugs for something that wasn't even very well understood. OK, so we decided to go ahead and try this out in several regards on the rare disease. Now again, let me emphasize, why would we look at a rare disease? Well, think about this. If you have a population, all of which have one mechanism for a disease, you can very effectively test your drug if that hits that mechanism. And then later, it may be that you will find subsets of a population with a more complex disease that will have that particular mechanism activated. Does it make sense? So let me give you an example. We had an antibody we made to an inflammatory molecule. Inflammation is whenever you hurt yourself or there's injury and it gets red and swollen and painful. And that is often by an interleukin-1. That's a molecule that's released all the time. So we had an antibody to that. Now traditionally, this molecule would have been tested, this new drug would have been tested in patients with a big disease, big market, rheumatoid arthritis, for example. That's a very complex disease. Some patients are burned out. Some patients have different mechanisms. So we didn't want to start in that kind of complicated disease. Instead, we started in a very rare disease called CAPS syndrome, cryopyrin-associated periodic syndrome. A disease often onset in childhood, sometimes later in life, which has too much interleukin-1. That's their problem. It has a mutation in the system that makes interleukin-1 that causes the cells to pour out interleukin-1 beta. The patients suffer from having generalized inflammation. This shows the back of a woman that had this syndrome. The inflammatory byproducts go to the kidney, can cause amyloid in the kidney, and are often lethal. Often, patients can't sleep. They are uncomfortable— arthritis, eye problems. It's a very profound problem, all caused by too much interleukin-1. We had an antibody to interleukin-1. If it didn't work in these patients, it wasn't going to work. And we gave it to— we didn't know exactly the dose. We gave our first patient. One dose, and she was better in 6 hours, completely, completely cured of her symptoms, and stayed better for 6 months. One shot. Now, this is not, by the way, a sales and marketing dream, as, as was pointed out to me. You've got a, you know, a few thousand patients in the world, one shot, they're better, like And this was before they had the ridiculous days of millions of dollars for— they weren't enthusiastic. But there were other diseases that had similar basis. For example, systemic juvenile idiopathic arthritis. This is a little more common. It basically is also a genetic disease, hits the kids like malaria. And they, again, got better. Again, fairly rare. While we were doing the work, a paper came out that pointed out the same pathway, however, is activated in gout. That's common. I mean, everyone from Ben Franklin on down, who was carried to the— as you probably know— the Constitutional Convention on a litter because his gout was so bad. Uric acid crystals, especially the big toe, awful inflammation, activates the same pathway, works really well there. But none of this was actually what I had in mind as the long-term plan. As a cardiologist, I knew there was one remaining risk factor that we had not hit in cardiovascular disease, and that's inflammation. If you have high levels of inflammation, as measured by CRP, for example, in your blood, you're at the same risk as if you smoke or have diabetes. Even if you have no other risk factors. And so I wanted to give it to patients who had atherosclerosis. So now, um, the problem with this kind of a trial is that now you're getting into that broad population, and even though you may select patients who have evidence of inflammation, they're still pretty heterogeneous. Plus, the rate of heart attacks is fortunately going down. So you're talking about a 10,000-person, $800 million to $1 billion trial. Because it had worked all the way through then, this drug had worked all the way through, this is the kind of thing you take to the board and to the CEO. And they said, OK. But I should tell you, as we were doing this trial, about 4 years into it, I got a call from the CEO who said to me just a few words. He said, this better work. And it did. So this is actually a remarkable drop, very significant drop, in the rate of heart attacks and strokes and in angina in patients. And amazing, we had to do another trial simultaneously, which we really hadn't, had sort of had to guess at ahead of time. They wanted us to do, the FDA wanted us to do a trial to make sure, because it's so many patients, if they're going to get this prophylactically, it's important that they don't get cancers. It turned out that amazingly, it dramatically reduces the incidence of lung cancer. Now lung cancer is an inflammatory disease in many patients, smoking, et cetera. So now they're happy. And this actually is a strategy that works. What we're looking for now as we seek new drugs are drugs that hit key nodes in important pathways, like here. And sometimes that can actually come from very basic observations in model organisms. One of the favorite model organisms is the fruit fly. Other are yeast, and I'll come back to zebrafish in a minute. But these are very powerful in they offer you genetic models. And if you can work out something in a genetic model system, maybe you can transfer it to a human genetic model. System. So let me show you an interesting fly eye. This is the normal fly eye, composed of these regular ommatidia. This is one in the mutation of the tuberous sclerosis gene. Now this is— I'll come back to in a second. Tuberous sclerosis is a disease, an uncommon disease of children. And they knew at the time They found the gene. And in flies, when that gene was knocked out, you got this overabundance of cells, like a tumor. The pathway was worked out that causes this overgrowth. And it turned out that the tubular sclerosis protein is an inhibitor of another protein called mTOR. And this was worked out first in yeast and then in flies. Now this is very interesting because we know kids who have this tuberous sclerosis protein loss manifest overgrowth of cells all over their body. It's not a malignancy per se. It's really just an overgrowth. You can see them in the skin. You can see it growing in the kidneys. And this is what they often die from, because they're very vascular and they bleed. You can see it in the lung. And you see it in the brain, these subependymal giant cell astrocytomas. These are benign in that they don't metastasize, but obviously a big problem in that they invade. They swell. And the kids often have autism and other cognitive problems. So we know that the— that what the kids are missing is this protein. And we know that it activates a pathway downstream because you've lost this inhibitor. And what we had in hand at the time was an inhibitor of the downstream protein. It's called a finitor, hitting tor. You may have heard of it, other name, rapamycin. Rapamycin is named because it comes from the island of Rapa Nui, which is Easter Island. It's from a yeast cell on Rapa Nui. So rapamycin inhibits mTOR. And so I said, well, we should try— Afinitor in the kids who have tuberous sclerosis. Now, this had been made inside Novartis before I got there, this Afinitor, for a completely different reason. It was actually being used— they wanted to use it as an immunosuppressant and thought maybe there might be some uses in other systems. It wasn't working on anything very well at the time. So I said, well, at least let's try it on a disease where it's completely logical it should work. And they said, no, we have bigger markets. And I said to them, well, I tell you what, how about if you get on a plane with me and we go to Cincinnati where there's a tuberous sclerosis clinic and you make rounds with me for a morning. And if at the end of that time you still say to me you're not willing to try it, OK. I didn't mean I would give up, but they didn't know that. So we did. We went to Cincinnati. And we saw not only the kids who were having seizures and whose parents had to sleep in the bed next to them because they would have a seizure during the night or couldn't really use language very well or had to get regular MRIs. And so when we left there, they said they would give it a try. And it really worked beautifully. It was— and this shows you one of the patients, for example, after 3 months of treatment with— for their tumor in the brain, which is extremely refractory, you can see it shrunk to almost nothing. And Another thing that came out of this is also very interesting, is that once you have a cure for a disease, you find there are many more patients with the disease than before. So they thought there were 10,000 with this. It turns out there's at least 30,000, maybe 50,000. We've seen this time and time again, that we treat a rare disease, and physicians who had been reluctant to make a diagnosis now are eager to do so. So this strategy, as I mentioned, what I thought was a relatively simplistic strategy of focusing on what were mechanistically understood rare diseases and bringing to bear very good basic science, worked very well. We were able to make— at that point, most companies would make 1 or 2 new medicines a year. By the time I left, we'd been in In about a decade, we made 90 for 130 indications. So it worked pretty well. It was very, very gratifying. I'd like to move on now to, though, a more— the next step that I'm concerned about, which is how do we apply this logic to behavior and psychiatric illness, autism, schizophrenia, depression? We had no luck in those fields. We don't really understand the basic biology of it at all. And I think part of the problem is we can't even describe the phenotype well. We use terminology that's now left over in these medical big thick tomes, if you've ever seen them, that are just made up by committees. We don't really know how to define the problem. So I'd like to focus for a few minutes on a fundamental science question that leads into a medical question, which is that of what drives social behavior? Are there genes that drive it? Are there diseases that are problems of those genes? I'm not the first person to think of this question. This is Charles Darwin. His second book, which no one ever reads, focused on a couple of problems that he didn't think fit into his understanding of natural selection very well. And this, Larry, is for your social biology question. One was he didn't understand the role of beauty. But more importantly, he didn't understand why animals should gather together socially. Because he didn't, of course, know about genetics then, although he does talk about genetic selection. It's a— he's using the term slightly differently because we didn't have a notion of a gene at that time. But he didn't understand why natural selection, which should work on the individual, would drive animals to collective behavior. How could you modify an individual in a way that would be beneficial to the whole group without harming that individual. It's a very— you can see why it doesn't really work in a standard theory of natural selection. And there are clearly some animals that like to be alone, more or less. These are pictures I took in Zambia, by the way. That's a different story. And there are other animals that just like to be together. They just literally seem— if you've ever been with hippos, they just seem to like to be together. They often form, as animals form groups. For example, birds, when during periods of migration, will form very ornate groupings. What's the basis of that? How do they avoid hitting each other? How do they know where to go? Who's the leader? Who's the follower? We don't have any real sense of that. In the 1930s, there was a scientist in England who went through the bogs of England following— he was a birder. And he recorded his entire life experience in a book on this problem. And he came up by page 30 with what he thought was the answer. He said, it has to be telepathy. Humans avoid hitting each other often in different situations. They form complex groups doing amazing things. But, you know, it looks pretty similar to what the animals are doing. But I'm going to show you a movie now of when there's a problem. With social behavior.

Fragile X syndrome affects an individual like my son Harrison, where he has globally impaired across all areas of brain function in terms of his ability to move physically, think cognitively, interact socially, communicate. In all those areas, he's impaired.

One, two, one. Good job. Which number is 2?

He's got some ADHD where he gets impatient and he doesn't stay on task very long. He's got a lot of movement. He has the flapping and the rocking. Part of his body generally has to be in motion. There's a kid the same age across the street, and he's running, throwing baseballs, hitting baseballs in you know, tennis balls in the street with baseball bats, talking and socializing with other kids.

Those types of activities for Harrison, being with other peers, should be a fun part of his life, are the most painful for him because they are so hard and out of his routine that it becomes overwhelming.

Okay, ready to go home? Yeah!

I worry a lot that We're going to get old at some point, and how is he going to be taken care of?

If there was a medication that could be discovered that would help slow down or stop the disease process in Harrison, that would take me to a place in my thoughts that I haven't been able to go to. My greatest hope for Harrison and Fragile X syndrome, for all people affected, would be for them to develop as much of a normal life as you and I have. To live on your own, to make decisions, to have relationships. That would be a full, rich quality of life.

So he has one form of autism. Autism has— is defined phenotypically. Again, it's an extreme. You have decreased communication, decreased social interaction, and repetitive behavior. [INAUDIBLE] Now, the standard approach, the standard thought now about autism is there are cortical problems, problems in the higher brain centers, and those are revealed by certain genetic changes that we have seen. In some patients. The question I had was, is the essential genetics of social behavior the problem? Not really necessarily higher function. So yes, there are some patients who have very severe changes in the brain, but there are also many who really don't see very much else going on in the brain. Is this a fundamental problem, perhaps in a more primitive way. And maybe we could understand this from more primitive organisms. And we turned to the fish for this. Now, I have a fondness for fish, not just because of my name, but because about 20 years ago, we had started— we'd done a first genetics screen in an animal called the zebrafish, which is a— you probably raise some of them in an aquarium. But their embryos are transparent. —and we'd been interested in the genetics of organ formation and found many genes and many pathways that were involved in setting up organs. And when I left Novartis, I wanted to work with the zebrafish again. And this field I thought was one where maybe not all my students were, so maybe I could actually get a foothold. Now fish actually have, again, beautiful collective behaviors. It's believed that these— are to help avoid predators, to find food. They have certain energetic biases that help as they're going through the water. And it's been described, the kind of formations have been described but not understood very well, as a balance of attractive forces and repulsive forces and forces of alignment. So how could we actually begin to get a feeling of this, so we set up systems using zebrafish where we would put 6 cells, 6 fish in a tank and just watch them. Now I'm gonna show you what wild type, the wild type is normal, look like in a second. But you'll see they just continuously move. And describing it is almost impossible. I can tell you if they're sort of schooling together or they're shoaling, but that's not really what they're necessarily doing. So we set it up with computer vision and with machine learning, and we'll let the computer tell us, are there patterns that we're not seeing? And then that becomes a certain— you can do a certain type of statistics with that and get an answer. And then they say, well, OK, what genes do you want to knock out or affect? And I said, we'll knock out about 100 genes individually that are related to autism and see if this very human disorder The genes for human disorder have an effect when they're lost in the fish. So first let me show you what the normal look like. You just watch them. And you see they're moving. And they're kind of hard to describe what they're doing, right? I mean, they're— now there's one guy going over here. And there's a bunch over here. And this is— they're just hanging out, right? So we do this for half an hour. These are hard experiments. Now let me— so we went through about 100, and we found several that were quite interesting. For example, here's one mutation. You can see the fish just don't get together. You can watch them as long as you want to. They basically ignore each other. So they're dispersed. It's as though they have no attractive force at all. On the other end, we had fish mutations where they just wouldn't leave each other alone. Right? So— so this means we do have genes that drive genetics behavior, and now we're trying to understand where in the brain these are acting and how they're acting. We don't know the answer to that yet. The next question would be, can we use this to discover new medicines? And here we were actually fortunate. Let me show you, it was a little bit of guesswork. This shows fish now in a tank. You're just looking from the side, normal fish tank. Again, wild type means they're normal. See, they swim, they go up to the top and go around. Okay? And another mutation, again one that was related to autism that had a very curious behavior. —which is they stayed on the bottom of the tank. It's as though they're digging. They're trying to escape. Right? And it's what we call in genetics 100% penetrant. I mean, every fish did this. For a variety of reasons, we tested— we screened different drugs. And we found one that completely restored normal behavior. And what's interesting is this drug, which is the only anti-schizophrenia drug that's used in refractory patients, is quite toxic. It has other problems. Takes 3 weeks to act in people, and it took 3 weeks to act in the fish. So I think we at least have a systematic way we can approach this and begin to see, can this very primitive organism mirror enough of the issues we have with social behavior in a species-specific way that we can actually analyze what's going on and begin to discover new medicines. Okay. I'm going to finish with something else I've been doing recently, just take a couple of minutes on this, which is to think about how we are going to train the next generation of physician scientists in this kind of, uh, way of thinking. And here, when I say this kind of thinking, what I mean is today we go from the basic observation to a new drug. But what if we actually tried to go back and use the patient to think about what the basic science of the problem would be? Now, this is not a new idea at all, nor is it mine. Mike Brown, Joe Goldstein, looked at patients with hypercholesterolemia. That's where we have all the statin inhibitors from. The basic science that grew out of this on receptor-mediated endocytosis, sterol-regulated transcription, opened whole fields. You just heard about Coley, who saw a patient whose sarcoma got better when it got infected, saying to himself, huh, maybe there's something about the immune system that can kill cancers. Kind of a little wacko injecting pus into people's tumors, and it wasn't too well accepted. But it was the launching of a new field. Or Peter Nowell down at Penn, who made a mistake when he was checking blood samples from a patient with CML, chronic myeloid leukemia. And he ran it under tap water. And the chromosomes expanded. And he was able to see in CML that there was a funny loss of chromosome 22. Actually, it was translocated to chromosome 9. Now, at this time in the '60s, we didn't know that cancer was genetic. And he said, look, in this patient population, there's a chromosomal change. Maybe it's genetic. And then others found that this so-called Philadelphia chromosome was at the junction of BCR and ABL2 genes. David Baltimore, Owen Witte. And from that, we were able to find out that ABL was overactivated and make Gleevec, which was actually a Novartis drug that targeted that, which was the first targeted therapy. All began, again, with a patient, in this case, a pathological idea. So my hope was that we could begin to take residents and train them in this way at the Mass General where I was doing some work. So why aren't they doing this? I mean, when I was trained, we actually did spend some time thinking about the underlying biomedical problem. It's because the patients are very, very sick now because insurance companies won't let them stay in. The medical record, like Epic, takes 2 hours to fill out for every hour you spend as a physician taking care of a patient. There's more concern with billing than there is with the patient care. When patients are very complicated or they can't figure out what's going on, you don't have time to think about it, you call in consultants. So many patients are seen by 10, 12 consultants, no one ever synthesizes the picture. So an uncommon extreme phenotype or something that's unusual, forget it. No one's going to spend time doing it. And as a consequence, The number of young physician scientists is going down, and the old guys is hanging in. And that's not good. So it's hard, and there's no time for it. So I was asked whether I would think of a way of trying to rectify this situation by Katrina Armstrong, who's the chief of medicine. Now, we came up with a method that really wasn't quite my idea. It was the idea of two residents who came to me, Victor Fedorov and Lauren Seitel. And they said, we'd like to take a look at these very complicated patients. And what they said to me, being a little bit naive, is we'd like you to fund us to do any work we want to do with them. My first response was like, who are these guys? But we talked it through and said, look, this isn't about money. It's about trying to actually have the time. And we worked together to set this up. And then residents would come and spend a week. And I'd pick a patient from a list that was given to me, someone that had no diagnosis had been able to be made. And I would call my friends around. The world who were scientists or physicians to be advisors and just give advice. And the team would then synthesize what they thought was the pathway. And they would give a lecture at the end to the entire medical service with me and then consultants. And it got off to a good start. And then Victor and Lauren went skiing in Canada and were killed in an avalanche. And I thought it was going to fall apart. I thought that, you know, because this was really their brainchild. But it was really amazing to me, because of that, the other residents stepped up. And we've been doing it for about 3 years now. And it's really taken off. It's a very popular rotation. And it's called Pathways. And I'll just tell you about 2 patients, the kind of thing that they see. So sometimes we see patients.— who you might see in other situations in something called, for example, the NIH Undiagnosed Disease Network. You know, just somebody where you actually can get an idea through, for example, through genetics. So this was a young man, 18-year-old proband here, as we call them, came in because he had very painful mouth ulcers and aplastic anemia, no red cells. And actually, his bone marrow was in general failing. It turned out that he had his father, uncle, and a grandmother, and a sibling who had the problem with the mouth ulcers. And these were really painful. I mean, he couldn't eat or drink or anything for this. And he had this curious aplastic anemia. In this case, the residents came up with a pathway they thought might be involved. And sent off to do exome sequencing, get the sequence. It turned out that they— we had guessed wrong on the pathway. But we found another pathway that was activated, STAT5 pathway. It turned out to be a dominant mutation. Now that was very important. They were about to give this patient a bone marrow transplant. It turns out if you have a mutation in this pathway, you don't do well with the transplant. So we stopped that. And also, a new drug had just came out that affects that pathway that we wouldn't have guessed at. Hitting the JAK-STAT pathway. And you can see when we give the drug here, red cells go right up. The young red cells called reticulocytes go up very quickly. White blood cells go up. And he's fine. The patient has been able to avoid a transplant. Some of them we don't have an answer for, but the— resident has picked it up as a future topic for their own research. So this one, I'm going to show you a temperature chart. And this temperature chart is from a woman who came in with— and unbeknownst to her, she came in for a different reason— her temperature dropped from a relatively normal temperature down 10 degrees every night. She didn't know what was going on. She was shivering, was completely unaware of it. But her organs were beginning to fail, probably many organs, partly as a consequence of this. Now, your temperature in general goes down by about a degree at night. So the circadian rhythmicity is right, it's just that it's really exaggerated. So you can see very regular. And the residents looked at this. One of them actually was interested in neurology as well. And came up with a hypothesis that this was a problem in the preoptic area of the brain where it controls the rhythmicity of temperature. But the question is then, why would this have happened? What I haven't told you is that this woman had chronic HIV/AIDS. And she had not— she was a prostitute and never kept to her treatment over her life until she came into the hospital right around the time— so basically her fluctuations, her temperature hypothermia started at the time we gave her drugs for her AIDS. And the reason turns out to be that her immune system, which had been suppressed by the AIDS, now came back. And as you've heard today, your immune system can make antibodies to lots of different things. And what her immune system is unfortunately making an antibody to are cells in the brain that control, uh, temperature. This is actually an experiment that was done in rodents, not by us, by other people, that show exactly that. If there's certain cells, if you you inactivate or activate them, you can change the rodent's temperature tremendously. So it's basically now up to the resident who's interested in going to neurology to prove— which is, I think most people now agree this is probably the case— that what's going on here is she has new antibodies that have come back that are hitting part of the brain which control the temperature. So as I've explained to the residents, the goal of these sessions is not what you see— read about in a clinical pathological conference where people are showing off what they know. It's to find out what we don't know and to say, can we use newer insights, newer technologies— exome sequencing, aptamers, single-cell sequencing— to make breakthroughs in questions that have been sitting dormant for many, many years by focusing again on these phenotypic extremes can we then help the individual patient that we're taking care of, and can we open new frontiers? Thank you. So the answer to your question about what the Gold Lab Symposium is about is that you hear talks you would never hear anywhere else. Oh, there you go.

Now it works. Hi, my name is David. I'm from In Recovery. I wish my medical director was here, who was actually Larry's Yale roommate. I have a question regarding addiction and extreme cases of addiction, and if this kind of study can be applied to that as well, because from the your talk, I'm looking at autism and it seems like chronic drug use seems to cause almost a mild version of that. I'm just wondering how research can be applied to that.

The question is about addiction. As you know, about 70,000 people dying from that. This is, I would say, probably the single biggest medical crisis we have in the country at the moment. We know something about addiction in the sense of having some sense of the central pathways that do it. We've actually set up and weren't the first to do it. A former postdoc of mine, Randy Peterson, set up a very nice zebrafish addiction model. Because addiction is basically activation of the same pathways that are very useful for other types of behavior. It's the addiction part that's the problem, not the actual underlying behavior of going for food. And I think that there is enough known now about the addiction pathways that we can begin to interrupt them. There are several ways of doing that, and they're not exactly the kind of work I talked about here. But basically, when you take, let's say, morphine and it hits the mu receptor, there is an activation of the pain killing part, which is important. But there's the active— it goes through a separate pathway, which is what I was talking about, pathways. Separate pathway for the, the, the tolerance, the addictive properties, the stopping. So actually those can be separated at the level of the cell. And the trick people are trying to do now is to separate out the use of the morphine or morphine-like compounds for pain and activate only that pathway without activating the other pathways. And at the level of just cell biology and early animal experiments, they're getting there, but it's still going to take more work.

Yeah, it's a very enjoyable talk. Um, I'm thinking about the great, uh, insights that your trainees have had in treating these individual patients who were undiagnosed previously, and then you find an approach that actually leads to a therapy for them, and that's really incredibly exciting. But there's so many other thousands, tens of thousands, hundreds of thousands of people out there in— with other diseases. How do you industrialize this? What are the next steps to take this sort of demonstration project that you've started and spread it more broadly? And I guess a follow-on question could be, how do we pay for that as a society?

The last one, I have no idea. It's not more expensive, though. Let me start with that. Because if you keep a patient in a hospital for months having 10 consultants and they're not getting better, that's a very expensive program. So we published some of this in the New England Journal of Medicine, which is read by like 10 people. So whatever. And this is a habit of mind. So my goal first is a very modest one, is to get the residents to think like this when they leave this service, when they go out to the rest of it. So then that's one. Two is a lot of hospitals and institutions have told us they're interested in this and actually have come to watch us. And three, actually, I shouldn't make fun of it. There is one venture capital person who's extremely interested in education that actually is interested potentially in underwriting some of this. So I think it is a way— the other thing that's important about it, which you probably know, we're losing physicians by age 50 to burnout.

It's very—

my wife's a pediatrician. And I mean, she watches her colleagues just leaving because they're just tired. They're filling out forms all day long, insurance forms, calling up— she takes care of patients with cystic fibrosis— calling up the insurance company to plead with them to let them pay for the new drug that she knows will work. What kind of a system is that? Right? So my hope is that some of these folks go off and some of them actually help— that the time they spend looking at the science makes them happier physicians. I've also started, which I didn't have time to talk about, a new combined master's MBA program in biotechnology at Harvard to try to train folks who are going to be the political leaders of the future in thinking about these things scientifically too. So it is a lot of effort from a lot of people, but I do think it will percolate through society. Can I ask one last question? Yeah. What did they want to teach you in 3 months at Harvard Business School? All right, well, there's 2 different kinds of things you learn at Harvard Business School. There's the finance, economics, that's fantastic. I mean, things like, that's a language. And then there's management and leadership. And I still don't know what that is. I mean, everybody who teaches it says management is strategy, management is tactics, management is getting along with people. So everyone has their own book. If that's true, it can't be right. So that's the other reason to work on social behavior. Thank you so much.