The Future of Cancer Medicine with Dr. Jay Bradner, President of Novartis Institutes for BioMedical Research

Anthony  00:05

Hello, you're listening to the third series of Theory & Practice all about science for the future. I'm Anthony Philippakis. 

Alex  00:12

And I'm Alex Wiltschko. 

Anthony  00:14

Today we are talking about the future of cancer medicine. Cancer is now the second most common cause of death worldwide, and more than one in three people in the US will be diagnosed with cancer at some point in their lives.

Alex  00:26

This podcast may not help many of those who are currently undergoing treatment, but it will provide hope and massive hope for the future. We expect most of the treatments that we discuss today to be commonplace in 10 to 20 years time. When we look back at the history of cancer treatment, the recent rapid progress is incredible. The first recorded treatment for cancer was surgery described 5000 years ago, in an ancient Egyptian manuscript. Surgery is still the mainstay of cancer treatment where the tumor is accessible, but it's a blunt instrument. You can debulk the tumor, but not necessarily get rid of the whole thing or prevent its spread or metastasis.

Anthony  01:06

The next major development in cancer treatment was a long time coming, just over 100 years ago, with the introduction of radiotherapy followed swiftly by the introduction of chemotherapy. These treatments allow greater targeting of cancer cells and access to cancers previously inaccessible by surgery. However, both radiotherapy and chemotherapy kill cancer cells. They are cytotoxic. And their effects include harming some normal cells. There's a delicate balance between using enough radiotherapy and chemotherapy to kill the cancer cells while still preserving our slowly healthy dividing cells. 

Alex  01:42

But in the last 20 years or so, many new treatments have come online, all of which aim to provide greater precision and targeting of cancer cells alone. We now know that cancer is a genetic disorder, an accumulation of molecular alterations in the genome, leading to the coding of proteins that allow, support or promote uncontrolled cell division. New treatments work at the genetic and molecular level but also are beginning to target the new micro-environment that cancers create. 

Anthony  02:11

Today we will focus on three broad categories of new developments in cancer treatments. First, there's targeted chemo therapies. These are drugs that specifically attack the altered driver genes that cause cells to divide rapidly. Second, we have cancer immunotherapy, which focuses on enabling the body's own immune system to target cancers. And then finally, there's the disabling the tumor microenvironment. The tumor stromal interactions that lead to altered supporting cells and new blood vessels that form to support cancer growth. 

Alex  02:44

To help us understand the true potential the future of cancer medicine, we're joined by Dr. Jay Bradner, who is the president of the Novartis Institute of Biomedical Research, known as NIBR. As with many of our guests in Theory & Practice, his expertise covers more than one discipline. He was a professor at Harvard Medical School, working with patients who had hematological cancers, and then undertook further training in organic chemistry and chemical biology, all foundations to become a drug hunter, to undertake the challenging quest to find new medicines.

Anthony  03:24

Hi, Jay, welcome to Theory & Practice. 

Jay  03:27

Thank you, Anthony. It's wonderful to be here. 

Anthony  03:29

Today, we want to talk about the future of cancer medicine. Can you start us off on this path by taking us through a rapid overview of each of the three broad areas of development in cancer treatments, and tell us why they open up this field? So maybe let's double click on the first of these targeted chemo therapies and oncogenetics. Walk us through what happened after the sequencing of the human genome, and how we started to discover the genetic drivers of cancer, and then what happened with actually making medications that were specific to those drivers. 

Jay  04:02

You know, here the story is inspiring, but it's also humbling. The genetic basis of cancer is now firmly established as a disease of acquired mutations to the genome. And in this post genomic era, where we're so easily able to access the guidance of genome sequencing, we have a clarified understanding of why patients have developed the cancer. The harvest of insights from this data set, ever-expanding, is strong, but there's much work to do. The cartography of the landscape of mutations is well underway. Genome sequencing is prevalent even at the bedside, but it's not exhaustive. This is a good foundation, in particular for early disease. We might now know what the major drivers are those that promote growth, but there's still so much to learn to translate this into therapeutics. How does recurrent disease work? How does drug resistant disease work? What are the mechanisms of these alterations and their relative contributions to the genesis of disease, as well as its progression. We need a creative and innovative, even artisanal mindset to reduce the practice first therapeutics for even the targets that we know. 

Anthony  05:18

So say a little bit more about that artisanal mindset. 

Jay  05:20

Drug hunting is a funny thing. It's the ultimate interdisciplinary science. It's one part chemistry, one part biology…equal parts, genetics, and computation. And as all of these fields are so deep and mature, familiarity with each of these technology-driven, laboratory-based sciences is essential to practice drug discovery. And one might think, in this moment of artificial intelligence and machine learning where a car can almost drive itself down the street, that there would be technology solutions to all the challenges of drug discovery. We have 5600 drug hunters, about 270 drug discovery projects, and it's my dream that we could input the protein sequence of a mutated cancer gene, and that a computer would output the chemical identity, the structure of a targeted therapy, but we're not yet there. And so the experience of working at the interfaces of these high technology fields of study is really more artisanal. It's the search for a lead. It's the optimization of the molecule by, you know, careful sculpting and structuring of, of the molecules so that like a molecular locksmith, we might complete the process of drug discovery with a highly potent, highly selective drug molecule that just fits perfectly inside the active site of a cancer protein. This work is high technology for sure. And moving very, very quickly, faster than ever before, we can sometimes create a drug in just 18 months from declaring war on a target. But for some targets, like KRAS, that killed my dad, the G12D allele in pancreatic cancer, we're going on 35 years without effective medicines. So this artisanal nature of the work cannot be overstated.

Anthony  07:16

You know, let me talk through a story that you were intimately involved in. In 2010, your lab discovered a protein inhibitor called JQ1. So it's something called a bromodomain inhibitor. And you describe it as playing the role of tricking the cancer cell into thinking it was no longer cancer. So maybe you can tell us a little bit about that story.

Jay  07:37

You know, I trained first as a medical oncologist, a blood cancers doctor, and then retrained in organic chemistry and chemical biology, really interested to be a part of a field of study that could try to make more creative medicines for advanced cancers, than what I was trained to prescribe: conventional chemo therapies and the like. So, I cut my teeth as a young chemist trying to make molecules that would turn genes on and off: gene control molecules. Where we have such molecules like modulators of the estrogen receptor… these are very important medicines, say for breast cancer, but it's a really short list of molecules that target genes switching. And the reason for that is most of the proteins that mediate the turning on of cancer genes and the turning off of healthy cellular genes are considered undruggable. They are not enzymes; they lack pockets; it's like calling a locksmith to get through a wall, they show up and they'd say, Where's the hole? Yeah, I have these pins these tools. And so when I was a professor, I started a laboratory to undertake this challenge to try to make molecules that could target undruggable proteins, try to make methodologies for even studying undruggable proteins in a drug discovery context. And out of this work came the first inhibitor JQ1 as you said, name for Jun Qi, our head of chemistry, at that time, that bound to an otherwise unapproachable protein called BRD4. And when it was administered to cancer cells, it would switch off the growth genes in certain cancers and actually turn on some of the genes of being a normal cell. This molecule had the characteristic of tricking cancer, to forget that it was cancer.

Anthony  09:31

Walk us through how it ended up actually being a really transformative thing.

Jay  09:35

Now this JQ1 molecule has taught us a lot about the function of BRD4 in so many different cancers by making the molecule available. Probably more than 500 labs have worked with this molecule. And these insights gave us some sense of where this drug might work, what its limitations might be, what combinations might be considered. We brought a drug like version of JQ1 into clinical investigation ultimately through Roche pharmaceuticals, and this class of BRD4 inhibitors that emerged because there's probably, you know, probably a dozen of them being studied today is still a work in progress. But I'm hopeful for two things: that BRD4 inhibition will prove relevant to the benefit of cancer patients. And that secondly, it has challenged two of the paradigms of drug discovery, first, the perception of drug-ability, the ability to drug Protein-Protein interactions, and then secondly, to challenge the highly secretive and private mindset we sometimes take towards science because sharing the molecule at a stage, which is normally a deep dark trade secret, really opened up a field of study for expedited consideration. 

Anthony  10:48

Yeah, sure, I'd love to dive into a little bit, you described it that you open sourced JQ1, and more specifically, you decided not to patent it, and made it available to the whole scientific community. Walk me through why you did that. And then what the impact of it was?

Jay  11:04

At the time that we did this research, and still to this day, I'm just constantly frustrated by the pregnant pause between knowing or believing that a medicine might help a cancer patient and the delay to bring a proof of concept molecule into clinical investigation, armed with the best guidance a field of study can deliver. And at the time, I was coming up in science, a field that I'm much less familiar with that of computer science and information technology, and still is just exploding and exploding, and in particular, I believe, because of a unique characteristic of computer scientists: they crave connectivity. They are different than academic biologists and chemists who crave scientific priority to all scientists wanting to connect to a good idea, but the art of academic science is to have the idea connected uniquely and only to them. We, not caring as much about that, decided to do a social experiment. In the context of this research, we might download the practices of open-source coding to the field of drug discovery, which is, you know, is historically very secretive. And we were able to do this in part because we had a willing institution at Dana Farber, a cancer charity, that was open to the idea that this could be best for patients and best for science. I think it's too early to say whether this model can be scaled across the whole field of drug discovery, cancer or otherwise. But what I can share with you is that the learnings were fast and furious, that the productivity of our lab in its research was exceptional, that the conductivity that we enjoyed to this day, opens new threads of investigation adjacencies, to scientists, that I or my lab might never have come into contact with. And for sure, by the time bromodomain inhibitors enter the clinic, they had the power of a thousand postdocs behind them, different than the way many drug discovery campaigns are performed.

Anthony  13:13

You know, at the time you were in academia. Now you lead Novartis Institutes for BioMedical Research, how you think about it now?

Jay  13:20

Well, about six years ago, I was approached by the Swiss pharmaceutical company, Novartis, to lead Research and Early drug development to be the president of NIBR, which is our internal innovation engine. I will admit, never having led anything before, I was curious as to why they would approach me in my early 40s, really having just reduced to practice the idea of an independent academic lab. And it’s clear at that time, and today, that Novartis believes that this open posture to drug discovery can be a powerful vehicle of expedited value creation in every dimension, value to the scientific community value, most importantly, to the patient. And ultimately, by being first and best armed with actionable information, valuable even to the business of Novartis. 

Anthony  14:12

Amazing.

Alex  14:13

That's really excellent. I mean, one thing that that comes to mind, I thought I'd just share, maybe as a side note is computer science and machine learning is kind of looking a little bit more like drug development these days. As the models that we're training are getting bigger and bigger, they take more and more resources to train. And it strikes me as similar to the amount of resourcing that's required to do the kinds of investigations that you're doing. And so I think, just as you might have learned from computer science and the open source ethos, there's some lessons coming to computer science on the horizon in terms of the resources required and the kind of research that's happening. So we might look to your model soon in computer science and the learnings from your social experiment when you feel like they've concluded. So when you know the answer, when you judge how it's turned out, please let us know. Because I think we have something to learn.

Jay  15:06

Alex, you know, thanks for that. It's been so long since I've had an in-depth discussion, especially with experts in computer science around the open science. It's interesting what you just say, that computer science today - now fast forward 10 years from our little experiment - is in one respect, radically open, that I can access AlphaFold almost the minute that it appears, disclose it to the world for all of its impact on protein folding.

Alex  15:36

You can even read the code, not just run it, you can read its innards, you know, it's incredible!

Jay  15:40

Read its innards and change its innards. But in another respect, as you so well articulate…Compute is actually rather closed, that as so many of these computational approaches require compute at unprecedented scale, only those that could afford or could access, the power of compute, can really harvest the fruits of these labors. Drug Discovery isn't quite at that point. It's still just artisanal enough, that it's accessible to so many, in fact, I think that the pervasive accessibility of molecular biology, the creation of an antibody, the use of CRISPR, genome editing, genome engineering, the many modalities that enable this, I actually think that drug discovery has never been more accessible, even to artisanal academic laboratories.

Alex  16:30

Wow, that's a really interesting connection. Let’s switch gears and move onto oncoproteins. There's several ways of blocking the actions of these proteins produced by oncogenes. And one of them is targeted protein degradation and it approaches to use molecular glues, and you're an expert in these. And I'm curious if you could tell us what are molecular glues, and how might they work to stop cancer progression?

Jay  16:58

Well, what a molecular glue is, is a molecule that serves to bring to big protein or RNA biomolecules together to glue proteins together. You could think of these molecules in the disease and drug discovery context as having a capacity to short circuit pathways causing disease. Now, it's a little bit what's old is new again, because of course, nature invented this first. A medicine actually from Novartis called rapamycin, and other medicines like Cyclosporine and FK-506: they actually are natural products from soil organisms, and they can suppress the immune system by short circuiting pathways of signal transduction the communication of a stimulus to the nucleus of a cell. Rapamycin glues two large proteins together that normally would never come together. One example that you referred to as targeted protein degradation, my lab created the first chemistry to bring proteins involved in protein destruction, the ubiquitin proteasome pathway into proximity of proteins that we might want to eliminate for the treatment of cancer, often called pro-tax, these targeted protein degraders, are proving to be a new modality of therapy and cancer and beyond. But some of these molecules are really big, and they're hard to make them orally bioavailable, they lack good drug like properties, they have off target effects. And so the group that I'm leading presently at NIBR is trying to make small format, I think 400 500 daltons (for the scientists listening today) cell-permeable, orally bioavailable, well-behaved drug molecules that can fuse cancer target proteins to the garbage disposal system, that can fuse RNA, to the splicing systems so as to eliminate in one instance, the Huntington disease protein in Huntington's disease. It's a new area, but it's exciting. You know, Alex, because our field has been stuck in a rut, believing that a small molecule is a key and a keyhole that binds and activates or binds and inhibits, but bringing the modularity that proteins in the cell possess to a small format drug like small molecule, it really opens avenues to unprecedented therapeutics. 

Alex  19:31

So instead of hiring a locksmith to pick the keyhole, you just tagged the door for demolition with a small molecule or something like that. I don't know if I've got some metaphor, right.

Jay  19:41

I think that's not bad, you know, putting a trash sticker on a protein. But I will say this: something I've noted in science and in my early career to date is that nothing moves faster than the lagging strand of innovation. As these molecular glues were first described by ourselves and Phil Chamberlain and others in the context of protein degradation, they're now almost systematically considered in protein degradation. But I can tell you that our consideration at NIBR of molecular glues is much broader than protein degradation, as I shared, to influence splicing and spinal muscular atrophy and Huntington's Disease. Intramolecular glues that put just a very specific little bit of superglue into the middle of a historically undruggable protein called SHP2, a protein phosphatase, a protein that escaped drug discovery for 30 years. But our drug hunters were able to do it, by reconsidering the property of what such a molecule might do - not binding to its active site, like a key and a keyhole, but rather gluing it in an unproductive inhibited conformation. I'm crazy about molecular glues; it's not all we do at Novartis. This isn't Elmers. But I do think there's a chance to make really important mechanistically distinct molecules with this concept in mind.

Alex  21:04

That's excellent. One thing that strikes me about the molecular glue concept is you're using the body, you know, grabbing the trash disposal mechanisms and saying please dispose of this. There's another approach in cancer treatment, which is to recruit the body's own immune system to attack cancer cells. And as in oncogenetics, there's many developments that we could talk about, but maybe you could talk about CAR T-cell therapy. So how does this work?

Jay  21:31

This is a powerful approach that grew up in government labs and in academia, the idea that you could take an immune cell out of a cancer patient and make a living and bespoke therapy to infect that cell in a petri dish with a virus that delivers a gene that doesn't exist in nature that turns the T-cell into a cancer hunting and cancer killing medicine. This is the CAR chimeric antigen receptor. It's a gene that directs the T-cell to bind cancer, and it grows as it kills cancer. It's a serial killer. Now, it sounds like science fiction, but only through iterative innovation principally in government labs, in academia, over about 15 years, this idea really grew up. And it seemed to us at Novartis, this is now going back about eight years, that really gravity had changed at the University of Pennsylvania, where Carl June and Stephan Grupp, had tweaked the formula just enough to push patients with refractory widespread B cell cancers, blood cancers into remissions. And so we entered into a paradigmatic collaboration that continues to this day to try to take these artisanal learnings and turn them into a real world medicine. The first CAR T-cell therapy would emerge from this out of our joint laboratories called chimeric, which targets the CD19 protein on B cells. And all these years later, there are patients like Emily Whitehead, you might have heard of, who remained cancer free, despite having been failed by so many lines of therapies that have been effective for other children. 

Alex  23:17

And how important do you think it'll be for cancer treatments going on into the future? 

Jay  23:21

Well, fast forward to 2022, we just recently announced a new and homegrown next generation of CAR T-cell therapy because there's so many challenges with this medicine. Each medicine is invented bespoke patient by patient - that makes it very hard to manufacture. It's a three week process; it's very difficult to get those cells to the patients at the time that they need them. It's also very expensive. And so we worry that this effective therapy might not reach all the patients that need it. And so we did what we do best: we stripped it to the nuts and to the bolts and gave it a full reconsideration. It's early days. But our current innovation in CAR T-cell therapy is something we call T charge. This is a one day manufacturing process, which would make it more accessible to many more patients, also less expensive to prepare. Secondly, and most importantly, by having the T-cells outside of the body for a short period of time, they retain some of their potency, their stem cell-like characteristics. And we hope that this will translate into much more durable responses for patients. There's a lot of work to do in this area. It doesn't yet work for solid tumors. We don't as yet have an off the shelf cell that we can give to any patient. They're still created in a bespoke fashion patient by patient. But it is an important and exciting, but I will tell you plainly, very challenging, new paradigm of cancer therapy.

Anthony  24:45

All right, so at the start of the show, we outlined three revolutions in cancer, the first being Imatinib, Gleevec, and the rise of targeted chemo therapies. And then you walked us through CAR T-cell and the dawn of immunology oncology. Now we'd like to go into the third of the revolutions, which is medications that target the tumor microenvironment. And here going back to 2001, we saw the launch of Bevacizumab or Avastin. What's angiogenesis all about? And what is the new generation of drugs targeting the tumor microenvironment doing and why are we excited about it? 

Jay  25:24

I can remember very vividly hearing Judah Folkman present his stunning research into the growth of blood vessels essential to the development of tumors, especially metastatic tumors, and the discovery of a critical mediator VEGF and the discovery of an antibody that might target this Bevacizumab that you just invoke. It was so exciting. Bevacizumab and VEGF inhibitor therapy remains an important component to some advanced cancers, but it's fair to say that the experience of developing VEGF antagonists for our field has been very humbling. It has shown that the idea of blocking the tumor microenvironment can matter, and that angiogenesis can be important. But it's also provided some tough learnings about the lack of translatability of some preclinical models of cancer, where those mechanisms were so dramatic and the only relevant model system of cancer in the patient. We don't actively work on angiogenesis at Novartis at this time, but we do believe in the opportunity of targeting the microenvironment, in particular, the immune cells that percolate out of these vessels, to hopefully eradicate cancer cells, recognizing them with all of their dozens of mutations as different than the host body. Could we provoke them more effectively, to eliminate cancer? Now, I will say that Novartis is a relative newcomer to this field, but our program is led by an extraordinary scientist Glenn Dranoff. And now about eight years into the reconsideration of immuno-oncology cell therapy. We have a pipeline of medicines, some antibody therapies to targets that have been well considered going back to the 90s, but what I'm most excited about are some of the new targets that we've discovered in more predictive models of cancer that have led to new types of therapies. And listen, it's very early days. But I'm hopeful that building on the remarkable contributions of PD-1 therapy and CTLA-4 therapy, that there will be other nodes beyond the so called checkpoints that might amplify immune impact on eradicating cancer in the tumor microenvironment. Now, Anthony, I will tell you, this is tough sledding, that the denominator of immuno-oncology medicines is massive, and the numerator of dramatically impactful drugs remains rather small. It is my bias, admittedly, that we must consider cell autonomous drivers of cancer with asymmetric weight, but I'm proud that we are in the pursuit in the hunt for new microenvironmental targets and therapies.

Anthony  28:11

I mean, stepping back all of this progress over the last 20 years, how has our understanding of cancer change through these experiences?

Jay  28:20

In some ways, our understanding of cancer biology has changed quite radically. We have a full map of early stage disease. We're still learning what changes exist beyond the coding or the gene regions of the genome. Trailing this guidance, we have a better understanding of the function and mechanism of circuits in the cell, but there's still so much more to learn: recurrent disease, drug resistance, non obvious changes in parts of the genome, changes to the epigenome. As a drug hunter, I'm most interested in and impatient for three harder to access insights. The first: we've talked about on drugability, you know, having drug most of the low hanging fruit, many targets are still beyond the reach of conventional approaches, and how might we imagine new types of drugs for these undruggable targets? Second, Achilles heels: beyond what's mutated, what non obvious, unaltered proteins promote the progression of cancer? We think about undermining cellular identity - what's the estrogen receptor for pancreatic cancer, synthetic lethality, what genes become essential only to the cancer when specific genes are mutated? And we think about the immune system as what genes exist to protect us safe from viruses, but actually hurt us when we develop cancer. And third is prevention. You know, optimism is a real watch out for me, I suppose practicing stem cell transplantation the creative work of drug hunting our real setup for a mindset that imagines serially what might be possible. But there is a doomsday scenario that we haven't discussed that treating established widespread, impossibly, genetically diverse solid tumors is just not approachable with targeted and immune therapy. It's also our responsibility as scientists to imagine that doomsday scenario, that after years of neglect, we might learn what we know already that the cure for cancer is never better contributed than before the cancer has become widespread. But perhaps even before clinically evident disease exists. The targets our field selects are drivers of cancer. What targets might we select to prevent it, to nip it in the bud, to intervene after one mutation not to wait for the dozens that most advanced cancers feature? What might these targets be? And how would we approach developing medicines in this space? Where studies will be longer than patent lives and of unprecedented expense. This is where our field needs to go to solve and un-drugability to establish functionally the Achilles heels and to move to earlier lines of therapy, not just in drug development, but in target nomination.

Anthony  31:15

You know, I'm glad you brought up prevention. And I can't resist just asking at least one question about it. Right now, there's tremendous excitement about liquid biopsy as a technology. How excited are you about it? And where do you think the most important approaches to cancer prevention will come from? 

Jay  31:34

I should disclose that we're not a diagnostics company. Although we are a high volume user of these measurements in our clinical trials. This makes me either the wrong person to ask as not a dyed-in-the-wool expert or the perfect person to ask as I have no skin in the game. So with that disclosure aside, I believe that the capacity to survey for cancer associated genetic changes before there's radiographically evident disease, clinically symptomatic disease, could be a game changer for the more effective prevention of cancer, such as blocking clonal hematopoiesis of indeterminate potential from progressing to acute myeloid leukemia, or the preemptive treatment of cancer to consider a stage zero of cancer, you know, usually speak in stages 1, 2, 3 and 4 of clinical presentation. Is there a stage zero, where there's a KRAS, G12C allele circulating in the body and you don't know where and you tried with a PET scan or a CAT scan? But you might ask what a KRAS, G12C inhibitor, eliminate that circulating gene, and prevent a patient from developing a cancer that a year or two from now would be much more challenging to treat. This research is obvious but not easy. And that creates challenges for our field. But I'm very hopeful that with improvements in the robustness, the use cases, and the introduction of liquid biopsy into investigational studies, that this field of science can have a huge impact.

Alex  33:13

I want to ask a kind of a personal and motivation question. Developing cancer treatments isn't for the faint of heart. I mean, we know that you have to look at almost a thousand compounds just to find one that you can go forward with for clinical trials. You've described your own feelings towards the science using words like opportunity and optimism and hope. But it also seems like you need endurance and resilience. I'm interested in where your endurance and resilience come from?

Jay  33:43

Well, Alex, I haven't reflected on that in some time. I think many are drawn to cancer research because it's such an impossible challenge to society. To individuals that affects almost all of our families at some form, it's a puzzle of science. But that's not what drives me in this research. You know, you want your time in science, you want your time on Earth to matter for something. And at least in a career in science, for your time in science to really matter. You have to take choices about where to invest your energy to download the right tools to ultimately be a contributor, and I chose cancer because it was so upsetting to me. As a younger man in medical school and training as a young oncologist to face the impossible challenge that cancer patients face. It's not unique to cancer patients, but the tragedy that befalls patients who face a cancer diagnosis, the heartbreak even of those who might in the end be cured was, for lack of a better word, so upsetting. I suppose you could have one or two reactions to it to not want to practice that type of science do orthopedics and fix knees and get everybody back on the slopes. But my reaction was, I suppose more, you know, visceral and leaning into that discomfort and using it as a source of energy, and it's not hard to remain motivated. As we get older, we surely are, it hits closer and closer to home. Now in my current role at Novartis, I have an opportunity to think more broadly in cancer, and honestly, sometimes the best insights for cancer therapy arise from solutions innovated in neuroscience or immunology. I don't want to lead you astray. I think that I'm equally desperate to find first effective therapies for Alzheimer's disease and to bring vision back to those who lacked sight with optogenetic gene therapies. But at my core, I'm hopeful that this time in science will contribute to new inroads to the to the treatment of cancer.

Anthony  35:59

With that, you know, I'd like to thank you for coming and talking to us, Jay. It's been such an incredible conversation. Really inspiring.

Alex  36:06

Thanks so much.

Jay  36:07

Thank you both. Enjoy this series. And I'm honored to be a part of it.

Alex  36:15

Huge thanks to Dr. Jay Bradner. We usually take time at the end of each episode in the spirit of our old regular in person meetups in Boston many years ago before Omicron, to discuss a big problem, the nail and possible solutions, and the hammers, inspired by what we just heard. So Anthony, do you have a hammer or a nail this week?

Anthony  36:37

I have a nail. And it's a topic that Jay in this interview just touched upon at the end, and I thought it'd be worth going into in greater detail. And it's the topic of using so called liquid biopsy for cancer prevention. Let me explain a little bit what this is about. You know, as Jay said, the real Holy Grail of cancer is not to cure it once it's happened, but to actually prevent it from becoming metastatic in the first place. And as just a reminder, you know, cancers go through what they call stages of growth. And a stage four cancer is one that's already metastasized. And at that point, it's very, very difficult to treat. But if you can catch a cancer before it's metastasized, then you have the opportunity of curing it through surgery by just cutting it out. And so there's this really important question of how do we do screening approaches to find cancer in its earliest stages. And, you know, we have colonoscopies, which have been very impressive for colon cancer, and really saved a lot of lives. We also do screening for breast and prostate cancer. And there, we think they help. But again, the data is a little bit more equivocal. But I want to talk today about a new approach that's called liquid biopsy. And it relates to kind of recent developments in genomics and sequencing. So to kind of start this off, there was an incredible discovery a few decades ago by a researcher in Hong Kong, named Dennis Lo, that if you look at women who are pregnant, and you draw blood, and you look at the pieces of DNA that are just floating around the bloodstream, that are not inside of a cell, they're just free DNA, that is from cells that died, you see that a small fraction of those are actually from the fetus, not from the mom. Not so surprising. But it was kind of amazing, because if you take just a big hunk of blood, and start to sequence all of the free DNA that's in it, you get rid of all the cells first, you can actually start to be able to get a readout of what the DNA of the fetus looks like. And this turned out to be a very valuable technology called non-invasive prenatal testing, where, in first trimester of pregnancy, you can start looking for aneuploidies, like Down syndrome, and then that is something that people can act upon. And it was a tremendous breakthrough. Because before that, you really had to rely on these indirect methods like imaging, where you kind of make a score of the probability of having Down syndrome. And then if it's high, then you have these invasive procedures, where you have to go in and draw amniotic fluid and that can lead to miscarriage. So this was really kind of a breakthrough in prenatal testing, and led to several different companies and products that are very widely used on the market today. Alright, so you might say, this is about cancer. Why we talk about pregnant women? Well, there was a very interesting finding by Illumina, which is the company that makes the sequencers and they also do a lot of the diagnostic tests. And they started to find a lot of examples where it wasn't so much that the DNA from the fetus was showing lots of aneuploidies and messed up chromosomes, but rather, it was actually the DNA from the mother showed some fraction of the chromosomes were greatly deranged. And in a lot of those cases, they saw that they were actually catching an early cancer, that when you then imaged, you could find it. 

Alex  40:03

Oh, okay, so let me just see if I can understand that. So, inside of our bodies, we've got an immense number of cells each has a copy of our genome. And cancer happens when one of those copies of the genome gets screwed up in such a way where the cell is allowed to divide uncontrollably. And the insight here is that the genome are some fragments of DNA from those cancerous cells, end up outside of cancer cells circulating in the bloodstream. But we can still tell that it's DNA from the original person, it's just mutated in such a way that could be dangerous to them.

Anthony  40:39

That's exactly right. You know, and it's not so surprising that DNA from cancer cells would be floating around the bloodstream. Because cancer cells, as they rapidly divide, often grow too fast for the nutrients around them and actually end up dying. So these DNA from the cancer cells is often present at higher quantities than you might expect. So this led researchers at Illumina and elsewhere to kind of ask the question, well, could we actually try and make a new screening technology for cancer? And this actually gets into some of the business side of the show - kind of took an interesting path, they actually created a new company, that's called Grail. And the name comes from the fact that cancer screening really is kind of a holy grail of medicine, and Illumina, seeded it, incubated it, and then spun it out. In full disclosure, GV was an investor, and many other groups were as well. And still early days, but at least, this is a very widely applicable field, where a lot of groups are now trying to develop liquid biopsies, not just Grail. But there are several companies trying to catch cancer in its early stages. And some of the trials are starting to read out. And the evidence looks pretty good, still early days, but there is now actually a test on the market that's called Gallery by Grail and other ones will be following suit rather soon. And again, we don't know. But this has the potential to be a whole new era in cancer screening.

Alex  42:04

That's fascinating. So it seems like this is something that's actually getting out there to patients, either now or soon. What's the kind of timeline - if you can just like sketch out – that this can impact people's cancer care today?

Anthony  42:16

Yeah, well, so like I said, there's a test on the market called gallery that you can order, it's generally not covered by insurance. And I think it's fair to say, the evidence base for it is still building up. And one of the challenges, and this goes back to our subjects on clinical trials, is these are really hard trials to run. Because you have to take a large cohort of people, only a small number will actually develop cancer over a period of like five years. And then you have to do a randomized trial, where you show that you can detect and actually this is quite interesting is the bar for proof is quite high, you have to not just show that you can detect cancer early, but that you can actually save lives. And actually, this goes back to an earlier generation of screening. And this is actually a good example of why screening for cancer is so hard. I think almost every physician at some point in their training or practice has the experience of getting a chest X ray on a patient who has a cough. And, lo and behold, you see a lung nodule, and you take them to surgery and cut out the lung nodule and see that it was cancer, and you feel like you saved a life because you got to the cancer before it metastasized, and were able to remove it. So then, of course, that begs the question, well, why don't we start doing routine chest X rays on everybody? And so they ran the trials of this. And sure enough, you do find more cancers when you do this. But you don't actually save more lives. And it's a little complicated to think about why that happens. But I think in general, the belief is that by the time that the cancer is big enough to be visible, probably too late - in the sense that the cat’s already out of the bag. And there's the other side of this too, which is that although removing a lung nodule is a relatively benign procedure, it's not the biggest surgery in the world, there still is risk of bleeding, infection complications. And so for every life that you save, because you caught a cancer early, there are other people that you will do harm to from the surgery itself. So the benefit has to be bigger than the harm. And so, the first trials were all negative - it doesn't work. And then they said, Alright, well, what if we do it in a higher prior population with people who are really heavy smokers? That didn't work. And then they did, Well, what if we do CT scans instead of chest X rays? Well, that didn't work. What about high resolution CT scans in special selected populations - it took decades. And finally, we now have a lung cancer screening protocol. But it took a very, very long time to actually figure out the right population that actually benefited from this. And so, how long will it take before the dust really settles on liquid biopsy? I'm not sure. My guess is that over the next five years, we'll start to have a pretty good evidence base for whether or not this approach is actually saving lives or just finding cancers early.

Alex  44:58

That's fascinating stuff. Thanks for sharing that story there. It really seems like we're kind of bringing the future into the present a little bit here.

Anthony  45:06

Excellent. Thanks so much, Alex. It's great to talk to you today. 

Alex  45:09

Always awesome, man. 

In the next episode, we'll be discussing protein folding. Later in the series, we'll be discussing diverse topics such as the future of psychiatry. If you've got any questions for us, or for our guests email theoryandpractice@gv.com or tweet at @GVTeam. We'd love to hear from you.

Anthony  45:28

This is a GV podcast and a Blanchard house production. Our science producer was Hilary Guite. Executive producer Rosie Pye with music by DALO. I'm Anthony Philippakis.

Alex  45:46

and I'm Alex Wiltschko. 

Anthony  45:47

And this is Theory & Practice.