Unraveled

Short placeholder heading

Subscribe to Unraveled

Unraveled is a podcast produced by Dana-Farber that explores the mysteries of the science of cancer. Veteran broadcaster Ken Shulman digs into the cutting-edge science that is transforming cancer research and providing hope for scientists and cancer patients.

Season 2

In Season 2, we explore the theme of momentum. The momentum that takes therapies from test bench to bedside, and then back to test bench for fine-tuning. The momentum provided by researchers, who defy headwinds and even gravity to keep science moving forward. And the momentum that doctors harness, with medicines and care that can transform once lethal cancers into treatable conditions. Listen to the trailer below.

Season 2 Trailer

Season 2, Episode 1: BCL-2 Inhibitors: Driving Cells to Destruction

The story begins decades ago with a man named Stanley Korsmeyer, who led the molecular oncology program at Dana-Farber from 1998 until his death in 2005. He discovered that B cell cancers like CLL over-produced a protein called Bcl-2, and interfered with apoptosis, or programmed cell death. But how that went from an interesting discovery to a game-changing cancer drug is a story of persistence and momentum, and it's the first episode of season two of Unraveled.

Season 2, Episode 1 Transcript

Ken Shulman 00:06
If you remember your high school physics, you'll remember Isaac Newton and his laws of motion. Several of those laws talk about something called Momentum. For Newton, momentum was a product of mass and velocity. Essentially, you multiply how much a thing weighs by how fast it's moving. In ideal circumstances with no headwinds or gravity to fight against, momentum is a constant, it doesn't change. But we almost never live in ideal circumstances. In most cases, if you want to preserve momentum, you need to apply force and to maintain that force. I'm Ken Shulman, welcome to the second season of Unraveled a Dana-Farber Cancer Institute podcast.

Last season, we visited some of the hottest topics in cancer research. We featured some of Dana-Farber, his most compelling scientists, physicians and patients. This season our theme is momentum, the momentum that takes therapies from test bench to bedside and then back to test bench for fine-tuning; the momentum provided by researchers who defy headwinds and even gravity to keep science moving forward; the momentum that doctors harness with medicines and care that can transform once lethal cancers into treatable conditions.

In the spirit of momentum, to keep the ball rolling, we'll pick up this first story in the present. Meet Brian Kimball. He's a retired physician assistant living in New Hampshire. In 2014, following a routine checkup and blood test, Kimble was diagnosed with CLL, chronic lymphocytic leukemia, it's a blood cancer, a mutation in the B cells, the cells that produce antibodies. CLL is a slow-growing cancer. Most people including Kimball experience few symptoms at the outset. But over time, these mutant B cells can muscle out their healthy counterparts in the bone marrow. With sufficient numbers, they can stem the production of red blood cells; they can damage organs. More than 4000 people die each year from CLL. And Brian Kimball, he had a pretty good run, but in 2020, six years after his initial diagnosis,

Brian Kimball 02:29
I began to feel fatigued. When I was just doing routine things, walking my dog on trails, I just didn't have the endurance that I would typically have. And so I had some lab work done. And lo and behold, my numbers had really changed dramatically.

Ken Shulman 02:48
The numbers Brian Kimball mentions of the numbers of lymphocytes, platelets and red blood cells in his bloodstream. They showed the cancer had progressed a lot, but I'll spare you the suspense. Brian's doing fine. He had a year of treatment at Dana-Farber, and he's back hiking in the woods with Barley, his golden retriever. He gardens in the spring bikes and kayaks in the summer, and cross-country skis in winter.

Brian Kimball 03:13
So all these things don't feel difficult. I feel as good as I've ever felt.

Ken Shulman 03:19
So how did Brian Kimball get his groove back? That's the real story. A story that starts decades ago at Dana-Farber with a man named Stanley Korsmeyer. Korsmeyer led the molecular oncology program at Dana-Farber from 1998 until his death in 2005. He studied something called apoptosis. That's a term coined from the ancient Greek. It literally means falling off or falling away. But in biology, it's trade talk for programmed cell death. In the years before Korsmeyer came to Dana-Farber, he discovered that B cell cancers like CLL, overproduced a protein called BCL-2. BCL is short for B cell lymphoma. Korsmeyer saw that this BCL-2 protein, the one that cancer shifted into overdrive, interfered with apoptosis, with programmed cell death. Essentially, the BCL-2 protein kept the cancer cells from dying when they were supposed to. It's one of the many things cancer does to survive and prosper. In hindsight, the discovery that BCL-2 saved cells from self-destruction proved to be a game changer. But at the time was just an interesting discovery. One of many, at least that's what it sounded like to Tony Letai.

Anthony Letai, MD, PhD 04:39
I remember back when I was a graduate student getting my PhD at University Chicago a long time ago, I actually had the opportunity to see Stan speak, and being sort of callow and ignorant, I listened to him speak. It's like programmed cell death. I mean, who cares, cells live, the cells die, it doesn't, you know, it's just like, it's not that interesting. into me. And I just totally forgot about it.

Ken Shulman 05:04
Tony Letai came to Dana-Farber as a fellow in 1998. The same year Stan Korsmeyer arrived there. Three years later, Letai signed on to work in Korsmeyer's lab.

Anthony Letai, MD, PhD 05:14
Stan was also a wonderful human being. And honestly, I chose his lab originally more, because I thought he was a great guy, even before I really was mature enough to understand the quality of what was going on in his lab.

Ken Shulman 05:28
Korsmeyer's character might have drawn the tie to the lab. But it was the work that kept him there... the work and some very good timing.

Anthony Letai, MD, PhD 05:35
It was just at the point when I entered his lab, that was about the year 2000. People start to think about, well, if BCL-2 is keeping cells alive, and maybe preferentially keeping cancer cells alive, maybe if we get rid of it, or antagonize it, or inhibit it in some way, maybe we can kill cancer cells selectively. So that's when it really, the mood change to one of actually targeting it. Rather than just learning about it.

Ken Shulman 06:01
The mood may have changed from inquiry to intervention. But it wasn't just on a whim, there was momentum. And Letai working in Korsmeyer's lab helped get it rolling.

Anthony Letai, MD, PhD 06:12
That was like making this tumor in this mouse, I knew I could switch it on and off, I inserted a switch in it. And I made the point of showing that if I switched off BCL-2, the leukemia would go away, just to show the field that like, this is fine. It's okay. There was this like bump in the road. But it's a totally legitimate effort. If you shut off BCL-2, you can get cancers to go away. So we should keep doing this.

Ken Shulman 06:40
Our cells have myriad mechanisms for growth, repair. And when the time comes for self-destruction. Cells turn off or fall off, when they've completed their mission when they were down, or when they suffered genetic mutations that can't be fixed, mutations that then make them a threat.

Anthony Letai, MD, PhD 06:57
The way that works is there's things inside, every cell in our body also has within it something called mitochondria. You may remember hearing about these in biology class a long time ago. And we often think of them as like the power plant of the cell. And that's true, they're really necessary for making energy and other things. But for the purposes of apoptosis, why they're so significant is they're sort of a bag of poison.

Ken Shulman 07:21
When a cell reaches the end of its useful life, a group of proteins gang up and poke a hole in the mitochondria, releasing a toxic booby trap that destroys the cell. Now, I realize that might sound a little dicey. Fortunately, we've got a countdown, a multi-step process for how these proteins are diploid. A process that makes sure our cells really mean it when they say, "Never more."

Anthony Letai, MD, PhD 07:48
One of the important things that comes out is a protein called cytochrome C. It has a day job of helping make energy for the cell. But the job in this case is it interacts with a bunch of proteins that have just been sitting around doing nothing. One called APAF-1, another one called Caspase-9, and it forms something that we call the apoptosome.

Ken Shulman 08:08
The apoptosome is a protein complex that activates another set of proteins called proteases. The handoffs continue, as these proteases activate yet another set of proteins, enzymes, this time, that hack the cell's DNA to pieces, so it can never replicate. And just in case, this method fails, these same enzymes activate something called Eat Me signals.

Anthony Letai, MD, PhD 08:32
And then there's cells in our body that are what we call professional phagocytes. These phagocytes come along, they recognize these Eat Me signals, and what do they do? They literally eat up the cell. They surround it, engulf it, and then digest it. And that, is the end.

Ken Shulman 08:50
It's an exquisitely designed system and exquisitely effective, except when it isn't. Sometimes with CLL and other cancers, the final countdown gets postponed. How, as Korsmeyer and other researchers dug deeper into the process, they saw that BCL-2, the protein that seemed to block cell death, doesn't work alone. There's a whole scrum of proteins that determine a cell's fate.

Anthony Letai, MD, PhD 09:17
And he was the first person to figure out that there's not just BCL-2, but a whole family, some of which are pro-apoptotic, some of which are anti-apoptotic, keeping the cell alive like BCL-2, and exactly how they interact and what controls them was really the subject of a lot of Stan's research. And I think most of you acknowledge him, you know, a real worldwide leader in that.

Ken Shulman 09:39
So it's not a lone-wolf operation. It's a strategic standoff, a chess match between two teams, the team trying to end the cell and the team trying to keep it alive. Any significant shift in numbers on either side tips the balance, and CLL has figured out a way to game the system. How does it do that? It drives tumor cells to produce high quantities of BCL-2. And these proteins, these excess BCL-2 proteins, snag the pro-death proteins that would otherwise have entered the cell. It's like a game of capture the flag. The pro-death proteins are neutralized, imprisoned, unable to complete their mission. And the cancer cells live on, the CLL thrives. How can you defeat an enemy that just gobbles up your best troops? Well,

Anthony Letai, MD, PhD 10:29
The stakes are really high now in the cell, because it has a lot of pro-apoptotic proteins that are all bound up by these anti-apoptotic protein. So it's a state that we call primed for death. These are cells where you just got to tickle the anti-apoptotic proteins a little bit and get them to release the pro-apoptotic proteins. Then they go puncture the mitochondria and let the poison out.

Ken Shulman 10:51
Well, then, all that was needed was a feather, something to tickle the BCL-2 proteins to get them to let the dogs out.

Matthew Davids, MD, MMSc 10:59
I don't know that I knew enough to be skeptical at the time. It was also new and exciting to me. And I saw potential for clinical targeting. So maybe I was just naive. But I thought it seemed very feasible. If you have the right drug and the right target, you can hit it hard and that you can have an effective therapy.

Ken Shulman 11:17
That's Matthew Davids. He directs clinical research at Dana-Farber's lymphoma division. He's also Brian Kimball's doctor, you remember the CLL patient with a golden retriever in New Hampshire. Davids came to Dana-Farber as a first year fellow in 2008. He fell in love with the study of apoptosis. And soon he signed on with Tony Letai's lab. Letai had picked up the momentum that Korsmeyer started. He was moving the research forward. Still, at the time, few in the field believed it was possible to hit the BCL-2 target, to find a drug that could inhibit the pathway CLL used to evade cell death.

Matthew Davids, MD, MMSc 11:57
It was a time where actually there was very little fanfare in the field, because there had been a lot of failures over the years. You know, there was incredible science, but there was a skepticism that we could ever target that pathway therapeutically because of some of the prior drugs that had tried and failed.

Ken Shulman 12:11
There was skepticism. But there had also been a lot of progress. And there's been progress since then. Scientists know a lot more about how CLL keeps tumor cells alive. Stan Korsmeyer discovered that BCL-2 captured pro death proteins. And he also discovered that there was a whole family of BCL-2 proteins. Some of these BCL-2 proteins were pro-survival. Some were pro-death. The pro-death proteins have names like BIM and BAK and BAD. And they share one feature that proved to be important, the BH3 domain. The domain, it's a structural unit on proteins, it's made of amino acids. In this case, the BH3 domain works like a sort of docking station. Now, lots of proteins have domains and many have several, but BIM and BAK and BAD and their pro-death posse have only one, the BH3 domain. And that BH3 domain is what gets them in trouble gets them caught. Because the pro-survival BCL-2 proteins also have a BH3 domain, and they use that domain to snag the pro-death proteins. Like attracts like. CLL the cancer takes advantage of that shared attraction. It drives tumor cells to produce excess BCL-2 proteins. And these BCL-2 proteins capture the pro-death posse at the BH3 domain. So you have the shock troops, the pro-death proteins ready to puncture the mitochondria and release the poison. But they're in the clutches of the BCL-2 proteins. What you need now in Letai's words, is the feather to tickle those BCL-2 proteins and free the troops. So how would you do that? Imagine for a moment that you want to play pickleball but you look across the room and see your dog has your racket in its mouth. You offer him a ball, a chew toy, a treat. Nothing works. He wants the racket. But it turns out it's not the whole racket he wants. It's the handle with the leather grip. So you go downstairs and find the racket you broke last week. It's just a handle now, but it does have the same leather grip. You offer that racket fragment, the handle to the dog. He wags his tail, drops the racket and sinks his teeth into the surrogate. That's sort of what Letai and Davids and others did. They used another racket handle, a synthetic handle, fragments of the proteins that would get the BCL-2 protein to open its jaws, clamp down on the counterfeit, and free up BIM and BAD and company to do their pro-death thing, to kill the cancer cells. These small fragments are called BH3 mimetics, because the fragments look and feel like the BH3 domain, not the racket, just the handle, the part that got caught in the BCL-2 trap. It was literally a bait and switch. According to Matthew Davids, it wasn't just a way to defeat CLL. It was a new way to fight cancer.

Matthew Davids, MD, MMSc 15:33
It's actually a very different mechanism of action compared to most drugs in oncology.

Ken Shulman 15:38
Most cancer drugs are small molecules that bind with the receptor on the surface of a tumor cell, or they bind to a pocket within the protein. The BH3 mimetics, the synthetic racket handle work differently.

Matthew Davids, MD, MMSc 15:52
That's more of a mechanical phenomenon that targets the protein, that displaces that pro-apoptotic molecule. So it's a whole it's really actually a whole new paradigm of cancer treatment.

Ken Shulman 16:01
In 2011. The year Matthew Davids joined the Dana-Farber faculty, he led a clinical trial for a promising BH3 mimetic. The drug, known at the time as ABT 199, had performed remarkably well in a CLL clinical trial in Australia. Davids treated some of the first patients in the US with a drug that would later be named Venetoclax. The results were well, staggering.

Matthew Davids, MD, MMSc 16:27
It was amazing. I mean, these were patients who had run out of all treatment options, you know, they were probably going to go to hospice, and sort of as a last resort, they came on this clinical trial and upwards of 80% responded, in some cases, very durably. I have just a handful of patients, but I have some patients still from that era who are still on Venetoclax. now, eight, nine years later, in some cases. It's just incredible to see how durable the responses can be.

Ken Shulman 16:49
Davids presented his results in 2012. Four years later, in 2016, the FDA approved Venetoclax for patients with CLL. It was the first BH3 mimetic to be approved for treatment in cancer. And much of the science that led to the drug was done at Dana-Farber. That was one reason Brian Kimball decided to make the trip from New Hampshire down to Boston.

Brian Kimball 17:13
I mean, when I was first diagnosed, I started following all the big names in CLL research, and Dr. Davids was among them. And I was elated to discover that he was right in Boston and that I was able to get an appointment with him. I remember going in with a long list of questions that he very patiently went through and answered every single one of them and I promised myself I wouldn't do that to the next time. So,

Ken Shulman 17:44
Kimble kept that promise. But in 2020, when it was clear, Kimble needed treatment, David had a question for him. What treatment did he prefer? There were two options. The first was something called the BTK inhibitor. It binds to the BTK protein on the malignant cell shuts the cell down and eventually kills it. The second option was a two drug combination. The first drug was Venetoclax, the BH3 mimetic that frees up the pro-death proteins. The second drug was Obinutuzumab, that's a monoclonal antibody that binds with the protein on the surface of the tumor cell. Both options were viable. But the first treatment with BTK inhibitors could go on indefinitely. The second treatment, that two-drug combo, would only last 12 months. Kimball also thought the list of potential side effects for option two looked a little less daunting. Now the two-drug combo was relatively new. But Davids and his team had more than a decade of experience with Venetoclax.

Brian Kimball 18:49
What I knew at the time was that that combination of therapy had only been approved about 15 months prior to my needing it. And so I felt fortunate to be in a setting where they had a lot of experience already at that point and using that combination, and that they were getting the outcomes that the trials had shown that they would.

Ken Shulman 19:11
Kimball started the two drug-combo treatment in the fall of 2020. At first, he came to Boston every week as Davids gradually upped the dosage of Venetoclax. Then he came down once a month, and then once every three months. At six months, just halfway through the treatment, he had already reached the MRD threshold, the minimal residual disease level where cancer cells can't be detected even by highly sensitive molecular tests. By November 2021, the treatment was done.

Brian Kimball 19:42
I have to say it was not a difficult year. The hardest part was just getting from New Hampshire to Boston and being where I needed to be at the appointed time. Truly that that was all. It was not, not a hard regimen to follow.

Ken Shulman 20:01
In a perfect system once again, according to Isaac Newton, momentum is preserved. There's no resistance, no forces to bring a moving object to a halt. In that perfect world, Brian Kimball's story has the happiest of endings. Now, his story is pretty darn happy in this world too. But the story is not over. Almost every therapy can be improved, made more precise, less toxic; the dosage can be altered; it can be administered as a frontline therapy, or in cases of relapse, or following a stem cell transplant; it can be combined with other drugs and apply to other cancers. And then there's that other kind of resistance, the things that tumor cells teach themselves in order to hack through our best laid treatment plans. In short, cancer doesn't rest. So neither can science or doctors.

Jacqueline Garcia, MD 20:55
So my job is to understand the science, translate it into a question that really could be answered in the context of a clinical trial.

Ken Shulman 21:04
That's Jacqueline Garcia. She's a physician at Dana-Farber, and she designs and directs early translational clinical trials. Part of that job is asking patients if they want to participate, to try out a therapy that, however promising, is still in development.

Jacqueline Garcia, MD 21:20
I take them through a diagnosis, I take some through a relapse, and I take some through end of life. And it really helped me to appreciate one, was this worth patients time? Did I take something away? Did I add? Did I ask a good question? And so I think that I find that I often put a lot of responsibility on myself in a good way to make sure that whenever I'm designing a clinical trial, which is truly a human experiment, that I'm making sure I move the field forward, and I move the experience forward for the patient to give them a chance that they might not have had before.

Ken Shulman 21:52
Garcia specializes in acute myeloid leukemia. AML is a common form of adult leukemia. It causes bone marrow to produce immature cells called blast cells. These blast cells crowd out red blood cells and can lead to bone marrow failure. As the word acute implies, things can go south very quickly with AML if it's not treated. In her clinical trials, Garcia was able to confirm that AML, just like CLL, was highly dependent on BCL-2. It was the same mechanism. The BCL-2 proteins captured all the pro death proteins and kept the cancer cells alive. Garcia's work with AML patients led the FDA to approve a combination Venetoclax therapy for the disease in 2020. Momentum. Research and researchers keep pushing a therapy towards new frontiers. But there's also pushback AML does depend on BCL-2 to keep its tumor cells alive, just like CLL the disease Brian Kimball was treated for. But unlike CLL, which is more homogenous, AML presents differently in different patients.

Jacqueline Garcia, MD 23:10
One of the things that remains difficult in acute myeloid leukemia is that it's a heterogeneous disease. So you can't just hit it with one therapy, you really have to be selective and thoughtful about the mutations that are present, the functional characteristics, the age of the patient. So there might be ways for us to personalize therapy and this will be an initial look or pilot test to see if this particular assay — it's called dynamic BH3 profiling — can really help us to select a therapy for a patient.

Ken Shulman 23:40
So how can physicians come up with bespoke therapies? How can they know precisely what drug or what combo of drugs will work in a disease that comes in so many shapes and sizes? One potential solution is called dynamic BH3 profiling. It's something Tony Letai's been working on for years. Dynamic BH3 profiling is a snapshot taken in real time, that shows how close a cancer cell is to self-destruction. The technique was developed in Tony Letai's lab. Working in vitro loci applies protein fragments called peptides to the mitochondria in cancer cells. The peptides are close cousins to BH3 mimetics. That's the active ingredient in Venetoclax. If these peptides cause the poison sack to spill its guts, that means the cell was already close to the cell death cliff. Letai does the same experiments combining the peptides with other drugs.

Anthony Letai, MD, PhD 24:39
When the drug's on board, does it make the mitochondria more sensitive to our peptides that we're putting on there? And if the answer is yes, that means that drug moved the cell closer to the cliff's edge. And the reason why this matters — and this has been a matter of 10 years of work here in the lab — the reason why that matters is it turns out that is an excellent predictor of that drug being an effective drug in vivo, in real life.

Ken Shulman 25:07
And that's the idea, a targeted treatment for cancer, individual drugs or cocktails mixed for individual patients. Meanwhile, the research trials and treatment continue. Jacqueline Garcia is busy with a new trial with a combination therapy for myelodysplastic syndrome, a blood cancer common in older adults. She says the preliminary results are promising. Matthew Davids is taking a close look at the way some CLL cells develop resistance to Venetoclax and to other BCL-2 inhibitors. How they switch their addiction from BCL-2 proteins and ramp up production of other anti-death proteins that also keep the tumor cell alive.

Matthew Davids, MD, MMSc 25:50
You can find when you take samples from patients later on when they've been on Venetoclax for a while, they don't depend so much on BCL-2 anymore, they started to rewire their cells to depend on other proteins like MCL-1 and BCL-XL. And so we're very interested in my lab now and trying to find ways in CLL to actually rewire the cells back to being BCL-2 dependent to re-sensitize them to Venetoclax treatment.

Ken Shulman 26:17
The story of Venetoclax is a story of momentum. The motion begins almost 40 years ago with Stanley Korsmeyer, and his work on apoptosis. And the science continues to surge forward today, with new chapters being written by physicians like Matthew Davids and Jacqueline Garcia. And as different as the plot and characters might be, every new chapter still echoes the first chapter and says Tony Letai, each one echoes the person who penned that first chapter, Stan Korsmeyer.

Anthony Letai, MD, PhD 26:49
Just by being a decent guy who was really good at science, he basically created an influence an entire field and still influenced people. I mean, I still to this day, and I think you could talk to other trainees of his who would say the same thing. When I'm confronted with a question with how to run my lab or what to do in a particular circumstance, I often just first ask myself, okay, what would Stan do?

Ken Shulman 27:14
Stan Korsmeyer died in 2005 of lung cancer at the age of 54. That's three years younger than Tony Letai is today.

Anthony Letai, MD, PhD 27:23
Which is a very sobering thought. Because, you know, I always saw him as like such a role model and such a senior figure, and to be older than him right now, it's very ironic. And I have, I was thinking about this earlier today, which is I think, like a lot of his trainees, I often wish that I wish I could just have five minutes to tell Stan what I'm doing and what's going on in this field that he founded because I'm sure he would love it.

Ken Shulman 28:03
Next time,

28:04
We tested and found no activity whatsoever, no toxicity, no activity, and we thought we had everything wrong.

Ken Shulman 28:11
We learn how a tragedy in medicine turned into triumph. I'm Ken Shulman, and this is Unraveled, a Dana-Farber Cancer Institute podcast.

Season 2, Episode 2: Thalidomide and its Second Act in Multiple Myeloma

Momentum isn't always one way. And it's not always constant. Sometimes it shoves you sideways, sometimes it stops you in your tracks. And sometimes only sometimes, it drives you to write one of the most astonishing second acts in all of medicine. That's the story of thalidomide, its second act in multiple myeloma, and the promise of protein degradation.

Season 2, Episode 2 Transcript

Ken Shulman 00:13
In most cases, when we think about momentum, we tend to think about forward momentum, a locomotive roaring down the track, always heading north. In physics, momentum is agnostic — forward, backward, up, down. It doesn't matter. But it matters for us. We like to believe in progress, in a universe where things get better. It's one reason why so many people find comfort in science, because it feels like science is always moving forward, and that we're the ones pushing it forward. We take a theory; we test it. If all goes well, we use it to wipe out smallpox or power a submarine. It's like that in cancer research to. Or it's usually like that,

Ben Ebert, MD, PhD 00:54
in most cases, we think about cancer research, understanding the biology of a cancer identifying a target, that we want to inhibit, developing a drug against that target and testing it clinically. This went completely backwards.

Ken Shulman 01:10
So, momentum isn't always one way. And it's not always constant. Sometimes, it shoves you sideways. Sometimes, it stops you in your tracks. And sometimes, only sometimes, it drives you to write one of the most astonishing second acts in all of medicine.

Ken Shulman 01:38
I'm Ken Shulman, and this is Unraveled, a Dana-Farber Cancer Institute podcast.

Ken Shulman 01:51
If there's a template for cancer research, it goes more or less like this. Someone in the lab locates a target in a cancer cell, a protein, a gene that helps the cell proliferate. Then someone else builds a molecule, a drug that can hit and neutralize that target. And finally, someone else tests the drug to see if it's effective and if it's safe. There are variations in the script — speed bumps, and hairpin turns, and more than a fair share of crashes. But for the most part, research is a neat play in three acts, discovery, development, and deployment.

Ben Ebert, MD, PhD 02:27
One interesting aspect of this story is that nobody set out to do this. We had a drug that worked.

Ken Shulman 02:34
That's Ben Ebert. He's chair of medical oncology at Dana-Farber. He, and his lab study hematological malignancies. Those are cancers that form in the bone marrow or in the immune system. And that drug he's talking about, the drug that worked on cancer, is thalidomide Yeah, that thalidomide, the anti-nausea drug that cause thousands of horrible birth defects in the 50s and 60s. You see, scientists didn't stop studying thalidomide after the tragedy it caused. If anything, they stepped up their efforts and dug even deeper in the 70s and 80s, to see if they could figure out what the heck went wrong. And while they didn't find that fatal flaw, at least at first, they did unearth a few surprises.

Ben Ebert, MD, PhD 03:19
It was found then to have some potential anti-cancer properties called anti-angiogenic properties, blocking blood vessel formation, and that prompted a whole series of clinical trials in cancer with thalidomide, which were almost entirely negative, except it was also used in multiple myeloma. And in multiple myeloma, it had striking activity and ultimately was FDA approved.

Ken Shulman 03:45
Let's break that one down. In the lab, in vitro, thalidomide showed potential to block the formation of blood vessels. And because many tumors are angiogenic, because they promote the formation of new blood vessels to fuel their own growth, doctors in several countries did clinical trials with thalidomide for cancer patients. And those clinical trials were almost entirely negative, except in multiple myeloma, where the results were dramatic. By the time Ben Ebert first joined Dana-Farber in 2001, doctors across the country were treating their multiple myeloma patients with thalidomide. But it was a sort of gift horse drug. No one knew exactly how or why it worked. They only knew it did work. There were a few theories, the main one being that it was anti-angiogenic, that it blocked the formation of new blood vessels, that it starved tumors. But that didn't explain why it only worked in this one form of cancer.

Ben Ebert, MD, PhD 04:44
And the ability of a drug to work in a very specific genetic subtype means there has to be a very specific mechanism of drug activity that is making that work. So, I was very interested in understanding why these drugs work.

Ken Shulman 04:57
Many researchers were intrigued by thalidomide and its success in the lab with multiple myeloma. But in order to understand how this one's infamous drug was being proposed as a cancer fighter, we need to know a little bit about the disease.

Ken Shulman 05:15
Multiple myeloma is a B-cell cancer. It affects a mature form of B cells, a form known as the plasma cell. Plasma cells are where we get our antibodies. They're a key element in our immune system. Like all blood cells, B cells start out in our bone marrow. It's where they develop and mature, and it's where, occasionally, trouble starts.

Ken Anderson, MD 05:38
In myeloma, there is one clone of cells that produces one abnormal clone of protein.

Ken Shulman 05:46
That's Ken Anderson. He directs the Jerome Lipper Multiple Myeloma Center and Lebow Institute for Myeloma Therapeutics at Dana-Farber.

Ken Anderson, MD 05:55
Therefore, the disease is too many plasma cells in the bone marrow producing monoclonal protein which we can measure in the blood or urine,

Ken Shulman 06:05
A runaway clone, making copies of itself, crowding out healthy cells in the bone marrow. It's how a lot of blood cancers work, how they erode our health.

Ken Anderson, MD 06:15
The complications of our disease include low red blood cell counter anemia, bone disease, lytic disease or thinning of the bones, osteoporosis, high blood calcium or kidney disease.

Ken Shulman 06:30
Ken Anderson has studied multiple myeloma since the early 1980s. His research and clinical trial work have helped develop more than a dozen drugs and close to 30 treatments for a disease that, when he started his career, was considered a death sentence.

Ken Anderson, MD 06:45
In my lifetime, we've seen myeloma go from a disease where patients only lived a few months to now many times patients having normal lifespan. So, it's been quite remarkable.

Ken Shulman 07:01
Early on, Anderson was fascinated by the data surrounding thalidomide and myeloma. He wanted to know how the drug worked. He also wanted to know if he could help make it better. He led preclinical studies on two variants of thalidomide. At the time these variants were called CC 5013 and CC 4047. The variants promised to be both more effective and safer than thalidomide. And they kept that promise. Today they're called lenalidomide and pomalidomide. Hundreds of thousands of patients all over the world have benefited from these two compounds. They're used all across the clinical spectrum as frontline therapy in combination with other drugs, and after relapses and bone marrow transplants. Now, back in the 80s and 90s and the early aughts, Anderson was one of many researchers who still hadn't unraveled the mechanism that made these drugs work in myeloma. But he got a lot closer. In the lab, he saw that lenalidomide and pomalidomide kept cancerous cells from binding to the bone marrow. He also saw that the drugs stem the growth of new blood vessels; that they were in fact, anti-angiogenic just as advertised. But that, according to Anderson at the time, wasn't what made them effective. He'd seen something else.

08:23
Most importantly, we showed that these agents, which are now called immunomodulatory, agents, up regulated the host or patient's own T cell and natural killer cell immune responses against their own myeloma. And they downregulated these T regulatory cells, which cause immunosuppression.

Ken Shulman 08:52
Our immune system is an intricate choreography of cells and signals. It's designed to destroy invasive threats like viruses, bacteria, and cancer. It's also designed to leave healthy cells alone. So, our immune system responds to signals to attack and to signals to stand down. Anderson saw that these thalidomide variants upregulated the attack signals and downregulated the stand down signals. The drugs told our immune system to open fire, while at the same time squelching any voices calling for surrender. To the best of anyone's knowledge, lenalidomide and pomalidomide were immunomodulatory drugs. They were drugs that modulated the immune system. Anderson says that system is still our first and best defense against disease.

09:40
It's potent. It's selective. It's adaptive. We could never design it ourselves. But we can exploit it and take advantage of it and get the immune system to do what it should have done in the first place, which is recognize the tumor cell for the foreign invader it is and reject it,

Ken Shulman 09:59
And that was that or should have been that. In the early 2000s, the drugs worked, they seem to be safe. And a drug thalidomide that had brought sorrow to thousands of families had been repurposed decades later into a therapy that was saving thousands of lives. Talk about a second act. It's just that, well, thalidomide, and lenalidomide and pomalidomide were still a mystery to Anderson and to everyone else. They had theories on how the drugs worked. They had some evidence, both in the lab and in the clinic. There was anti-angiogenesis; there was immunomodulation. Most importantly, patients got better. But none of those theories and none of that evidence came close to explaining one key point. How was it that these agents were so effective against multiple myeloma, and so completely ineffective against so many other blood and solid cancers?

Ken Shulman 10:59
Let's assume you're a researcher, in this case, a cancer researcher. You already know the drill the template where you look at a disease, find its Achilles heel, and see if you can invent a bow and arrow that will strike that heel and stick to it. For Ben Ebert, that template made no sense in 2009, when he started looking more closely at thalidomide and its chemical cousins. He still wanted to know why these drugs only worked on multiple myeloma and a few related cancers. But the standard template wouldn't work here because they had the bow and arrow, and they have the disease. What they didn't have was a target, an Achilles heel. So, what does the researcher do when they have to go off script, when they have to start at the end of the line and work back to the beginning? You shift the training to reverse; you start from your last piece of information.

Ben Ebert, MD, PhD 11:50
So, the first step is to understand what protein the drug binds to. And we did that with a drug called lenalidomide.

Ken Shulman 11:59
Ebert wasn't alone on this backwards trip. Several researchers including Dana-Farber's Bill Kaelin, the 2019 Nobel Laureate in medicine, were grappling with the same problem. They wanted to know how the drugs actually worked. In 2010, one year in, Ebert and Kaelin, and everyone else got an assist from a group at the Tokyo Institute of Technology. This group discovered that thalidomide binds to a protein called cereblon. Cereblon plays a role in myriad cell processes. And we're still not clear on everything it does. But along with discovering that this protein, cereblon, was the one that binds with thalidomide, the Tokyo group also discovered that cereblon played a role in fetal limb development. That bond, the thalidomide latching onto cereblon in utero, explained the birth defects back in the 50s and 60s, but it didn't explain how thalidomide shut down myeloma or myelodysplastic syndrome, another cancer Ebert was working on and that responded well to the drug. Still, the cereblon discovery was a great leap forward... or backwards. Either way, after the Tokyo discovery, things stalled. It turns out that reverse engineering a drug is complicated and exasperating.

Ben Ebert, MD, PhD 13:20
I would say that the postdoc who worked on this project came into my office at least once per week and asked to work on a different project because it was too hard and he wasn't getting anywhere, and really worked on it for a couple of years without progress before everything opened up.

Ken Shulman 13:34
And everything did open up eventually, but not before their assumptions were turned on their heads. The first assumption to be upended was that these drugs killed cancer cells by simply goosing the immune system, by upregulating those attack signals and downregulating the surrender of signals. It wasn't that Ken Anderson was wrong. Thalidomide and lenalidomide and pomalidomide did boost T cells and natural killer cells, but something else was going on with myeloma. And it had to do with cereblon. Cereblon, the protein that bonds with thalidomide and its derivatives, is a receptor on an enzyme complex called ubiquitin ligases. These ligases are part of a system that breaks down or degrades proteins we no longer need. Let's take a closer look at that. Our bodies have a system for degrading proteins. The degrading is done by a protein complex called a proteasome. It's like a big molecular recycling machine. Now how does this proteasome, this recycling machine, know which proteins to degrade? That's where the ubiquitin ligase comes in. The ubiquitin ligase bonds with the protein that needs to be degraded. It then recruits a chain of proteins, a chain called ubiquitin. Together, ligase and ubiquitin tag the selected protein for pickup. The proteasome, the recycled machine, sees the tag and does the rest. It's like sticking on a molecular post-it note, with ubiquitin as the paper and the ubiquitin ligase as the glue. So, in this case, thalidomide was binding with cereblon, with a receptor on a ubiquitin ligase. Remember? The glue on the post-it note. Something was going on, something out of the ordinary. Was it possible that thalidomide wasn't modulating the immune system? Was it possible that this drug was somehow pushing the ubiquitin ligase to tag proteins for degradation? Now, cancer researchers had fantasized for a long time about breaking down runaway proteins that drive tumors, about a drug that would induce these ligases and ubiquitin to tag a tumor cell and toss that cell out on the curb for the proteasome to take it apart. But they couldn't figure out a way to realize that dream. And anyway, that wasn't how cancer drugs worked.

Ben Ebert, MD, PhD 15:56
Almost all of our drugs inhibit an enzyme. They block the activity of an enzyme. So, we assumed that thalidomide and lenalidomide block the activity of this ubiquitin ligase.

Ken Shulman 16:08
Ebert and his colleagues knew that almost all cancer drugs work by blocking the action of an enzyme. It made sense to assume that thalidomide and its variants did the same thing — that they kept the ubiquitin ligase and ubiquitin from sticking to a target. And Ebert was relatively sure he could chart the downstream effects to connect the dots between blocking the post-it notes and stopping the spread of cancer. But he didn't have to because the assumption was wrong. The drugs didn't inhibit the ubiquitin ligases. They enhanced them, expanded their range, and drove them to stick to new targets.

Ben Ebert, MD, PhD 16:46
But what turned out to be the case is that thalidomide and lenalidomide bind CRBN, this ubiquitin ligase, and induce the enzyme to ubiquitinate and target for degradation new proteins that are not normally degraded by that enzyme. So, it refocuses this ubiquitin ligase to target new proteins for degradation.

Ken Shulman 17:11
The process Ebert is describing — ubiquitination — was well known and extremely common. As the name suggests, ubiquitin for ubiquitous, these regulatory proteins can be found in almost every living tissue. What was new here was that with the help of these drugs, the ubiquitin ligases, the glue, was targeting and sticking to proteins that Ebert and his colleagues and almost everyone in the field thought were unreachable. More importantly, the proteins the ligase were sticking to were proteins myeloma needed to survive. They were the Achilles heel. In the lab, in the presence of these drugs, the ligases were binding to two proteins on myeloma cells, two transcription factors, proteins that helped immature B cells differentiate into plasma cells.

Ben Ebert, MD, PhD 18:01
And those transcription factors are not druggable targets. They're not proteins that we thought we could drug. They are essential for the survival of multiple myeloma cells. So, when they are degraded, the multiple myeloma cells die. And they're not normally degraded at all by CRBN. But when the drug is present, these proteins are very rapidly and very efficiently degraded.

Ken Shulman 18:22
Not only did these drugs enable the ubiquitin ligase to bind with new targets, they enabled the ligase to bind with the two proteins that were driving this particular form of cancer, multiple myeloma, and a few other cancers that were also dependent on that protein. This explains why early trials with thalidomide were so disappointing. It only worked in cancers that displayed the specific proteins. The research project that once felt like a dead end now look like a four-lane highway. Several postdocs in Ebert's lab signed on to help him and his stalwart grad student. There was tangible momentum, but there were still some bumps in the road. Inspired by their success in vitro, the team moved on to replicate those results in a mouse model. And they hit a wall.

Ben Ebert, MD, PhD 19:10
We tested that lenalidomide and found no activity whatsoever. No toxicity. No activity. And we thought we had everything wrong.

Ken Shulman 19:18
The team scratched its collective head. How could this one be explained? Mice are usually a reliable model and cereblon is cereblon, a receptor on a ubiquitin ligase. But it turns out that cereblon isn't always cereblon. Ebert and his team took a closer look at the protein — at the receptor that bothered with the drug. They examined its DNA. And they discovered that mice and humans have a different sequence in the gene that codes for cereblon. It's only a slight difference. But it's big enough that mice respond to thalidomide in a very different way than humans do. Then Ebert and his team rewrote that mouse DNA sequence. The drug worked just as it did with humans. cells. This misstep had produced something important. Now, Ebert and his team had a mouse model they could use to test other drugs that bonded with cereblon. Historically, finding out about the two DNA sequences was enormously significant, because it helped explain how regulators in Europe had failed to see the dangers in thalidomide. They did test the drug on mice and rats. But they didn't know that mice have a slightly different genetic sequence for cereblon. They didn't know that mouse cereblon is totally unresponsive to thalidomide. They could have administered any dose and still not seen a reaction.

Ben Ebert, MD, PhD 20:37
And actually, the amazing thing about its first use was, one of the striking things was, that when it was first tested, and preclinical models and mice and in rats, it had no toxicity at all, and was thought to be a wonder drug. It was totally safe, and had these really useful properties like inhibiting nausea, and therefore was developed rapidly for use in pregnant women, because it was thought to be so safe.

Ken Shulman 21:06
Ben Ebert and his Dana-Farber colleague, Bill Kaelin published their findings separately in 2014. Lots of people noticed, including a graduate student in Basel, Switzerland named Eric Fischer. He'd been working on thalidomide since 2010. And he was familiar with the tragedy having grown up in Germany, where many of the drug-induced birth defects occurred.

Eric Fischer, PhD 21:29
So, I was studying ubiquitin ligases, as it relates to their ability to control DNA repair processes. And it turns out, it's the same family of ubiquitin ligases, a member of which is the one that thalidomide binds, and so I became immediately intrigued.

Ken Shulman 21:48
Eric Fischer studies the ubiquitin proteasome system. He also directs Dana-Farber's Center for Protein Degradation. But in grad school in 2014, Fischer wasn't thinking about oncology. He was studying how ubiquitin ligases help repair DNA after it's been damaged by excess sunlight. But the more he read about thalidomide, the more he grew intrigued. He saw how the ubiquitin ligase, in the presence of lenalidomide, bonded with two proteins that were vital to myeloma, and how that same ligase through ubiquitination, led to the destruction of these proteins and shut down the cancer. It was a whole new approach to fighting the disease, an approach researchers had dreamt about but didn't think was possible or practical.

Eric Fischer, PhD 22:35
The concept was around. But many people questioned the clinical utility because people had these thoughts that, if you start degrading things, you would probably really mess things up. And there would be unexpected toxicities and things would go awfully wrong.

Ken Shulman 22:56
But lenalidomide, as Ebert demonstrated, didn't slash and burn through random proteins. It was highly specific. It modulated the ubiquitin ligase. So, it would target just two proteins and spare the rest. That was proof of concept, at least that protein degradation therapy was practical. But possible? The two targets, the transcription factors that lenalidomide hit, were notoriously smooth, as smooth as cue balls. There didn't seem to be any place for a drug to stick a landing. The two target proteins in question the Achilles heels, were nicknamed Icarus, and Ilos. Their formal names were IKZF-1 and IKZF-3. IKZF is short for Icarus family zinc finger. The proteins are named after a domain, the zinc finger domain. The zinc finger is a common protein domain, lots of proteins have them including Icarus, and Ilos.

Eric Fischer, PhD 23:53
It's almost like Lego, you put a few of those together, and you get a functional protein. And a C2H2 zinc finger, is one of these folds, is a fold that's most commonly really known for binding to DNA.

Ken Shulman 24:08
The C2H2 zinc finger that's its full name is involved in many cellular processes. Sure, in theory would have made a great target for drugs, but in practice.

Eric Fischer, PhD 24:19
People would have considered it entirely impossible to bind to a C2H2 zinc finger with a small molecule. And that's where, for this sometimes-used term, undruggable, comes from but the nature how thalidomide acts, and how it utilizes the fold of the E3 ligase, to recognize this specific loop in the C2H2 zinc finger makes it unique and really provided this proof of principle that you could target those proteins, and that is likely something we can repeat.

Ken Shulman 24:54
And that was the last stop on this upstream trip, when thalidomide or one of its cousins was on board, the ubiquitin ligase was able to latch on to a target, no one thought it could grab — a C2H2 zinc finger, to boldly go where no drug had ever gone before. It wasn't the final frontier, says Fischer. But it was an important one.

Eric Fischer, PhD 25:15
It removed these barriers, and it removed this classification as something is undruggable. Because we've seen so many examples now of things we considered undruggable that are actually quite druggable. And I think that thalidomide example is a great example where we hadn't really considered an entire family of proteins, these transcription factors, as druggable.

Ken Shulman 25:40
And this is where the momentum shifts, where the story stops analyzing what has happened and starts imagining what could happen. Thanks to research at places like Dana-Farber, we can now target ligase receptor called cereblon. And when we do that, we can stop multiple myeloma cells from proliferating. We can stop a few other cancers, as well. But there are hundreds of these ligases, and there are other types of cancer driven by proteins that might be candidates for targeted degradation. Looking forward, after his 40 years of research, Ken Anderson thinks we're only at the beginning of a long, promising journey.

Ken Anderson, MD 26:22
These degraders I believe strongly, once they have shown their efficacy and safety, primarily in cancer, will likely be very useful quite beyond our field. You could degrade a protein that's essential for bacteria or viral replication. You could degrade proteins that are aberrant in neurologic diseases, storage diseases, rheumatologic diseases. So, it's really opened up a whole new class of therapeutics.

Ken Shulman 27:05
Next time, we'll shine a spotlight on bench-to-bedside research at Dana-Farber and on a discovery that has helped millions of breast cancer patients around the world.

27:14
You know, the laboratory work drives us to clinic, the clinic drives us back to the laboratory for more discovery or understanding resistance or how to get things to work better. And that synergism is really part of the blood and the mission of the Dana-Farber to have that research and clinical proximity.

Ken Shulman 27:33
I'm Ken Shulman, and this is Unraveled, a Dana-Farber Cancer Institute podcast.

Season 2, Episode 3: Stop the Division: CDK-4/6 Inhibitors and the Cell Cycle

Cancer is often a problem of cell division; cancer cells keep doubling and doubling, faster and faster. Eventually, they crowd out the healthy cells we need to survive. So researchers proposed a question: Why not stop that land grab? Why not find a way to jam the gears of the cell cycle to stop cancer cells from dividing? Today, we have drugs that do just that: They're called CDK-4/6 inhibitors. The story of those drugs, the momentum that brought them from bench to bedside, was written largely at Dana-Farber Cancer Institute — and keeps being written today. It's the story we're telling in episode three of Unraveled.

Season 2, Episode 3 Transcript

Ken Shulman 00:06
That thing we call the cell is the smallest self-sustaining biological unit. A unit that can feed itself, generate energy, and most importantly from a survival standpoint, reproduce. And that reproduction, it's done by cell division, by cells dividing into two identical copies of themselves. With cell division, an amoeba becomes two amoebae. In more complex organisms, like say, a palm tree, a bark cell becomes two bark cells. Cell division is how life sustains itself. It's how we humans grow, how we replenish our liver and spleen. How we replace the cells we lose when we skin our knees. The cell division is also, unfortunately, how cancer sustains itself. With cell division, the cancer cells keep doubling and doubling, faster and faster. Eventually, they crowd out the healthy cells we need to survive. So, to ask the obvious question, why not stop that land grab? Why not find a way to jam the gears of the cell cycle, to stop cancer cells from dividing?

Geoffrey Shapiro, MD, PhD 01:07
There was an area that was fraught with quite a bit of frustration. And one of the problems with those initial drugs is that there was very little therapeutic window between what was happening in cancer cells and normal cells, and they would stop the proliferation of normal cells as well.

Ken Shulman 01:25
And there's the rub. Because it isn't enough to design a drug that keeps cancer cells from dividing. The drug also needs to pass over healthy cells, like an Old Testament angel who smites the tumor but spares the healthy cells, so they can reproduce. Today we have those drugs, they're called CDK inhibitors. And the story of those drugs, from bench to bedside, was written largely at Dana-Farber.

Ken Shulman 01:55
I'm Ken Shulman, and this is Unraveled, a Dana-Farber Cancer Institute podcast.

Ken Shulman 02:07
In 1665, an English polymath named Robert Hooke put a cross section of cork under a microscope. The pattern that he saw — repetitive contiguous shapes, laid out row after row — reminded him of the cells he'd seen in monasteries, the cells were monks lived in prayed. So that was the name he gave to those shapes, he called them cells. Over the next few centuries, scientists found out what the cell was made of and what it was made for. The cell nucleus was discovered in 1831. Cell Theory, the idea that all living organisms are made up of cells, was proposed in 1839. Still, the biology of the cell cycle, the nuts-and-bolts mechanics of how cells divide, remained Uncharted well into the 20th century.

Geoffrey Shapiro, MD, PhD 02:55
But it was clear in the early 1990s, that the cell cycle was driven by a set of enzymes called cyclin dependent kinases, which govern the transition of cell cycle phases, one to the next.

Ken Shulman 03:09
That's Geoffrey Shapiro, Senior Vice President for Developmental Therapeutics at Dana-Farber. We'll get to the cyclin dependent kinases in a minute. Shapiro came to Dana-Farber in 1991, is a postdoc, he wanted to study cell cycle biology. It was a good place for that. Dana-Farber's David Livingston, whom we profiled last season, and who has since passed away, had done groundbreaking work in the field. In particular, Livingston studied how missteps in the cell cycle could lead to cancer. As Jeffrey Shapiro explains, the cell cycle that process by which one cell becomes two is marked by four phases.

Geoffrey Shapiro, MD, PhD 03:50
There's S phase for DNA synthesis, M phase for mitosis when cells split in two, and then the gap phases G1 and G2, which fall in between, and people joke around that the gap phases represent gaps in time between the S and M phases, but also represent gaps in our knowledge.

Ken Shulman 04:08
It takes about a day for a cell to divide, to cycle from G1 to S to G2 to M. In between each of these phases, there's a checkpoint, a stop sign that halts the cell cycle until the cell shows it's ready to move on. David Livingston did a lot of work on one of those stop signs, the RB protein. RB is short for retinal blastoma. The RB protein is the stop sign where the cell pauses between phase G1 and S phase, from mature cell to copying its own DNA.

Geoffrey Shapiro, MD, PhD 04:42
So, when cells decide that they are going to enter S phase, the RB protein has to be inactivated — that stop sign has to be inactivated. And that's done by phosphorylating the RB protein, adding phosphates on, and the enzymes that phosphorylate RB are the cyclin dependent kinases.

Ken Shulman 05:03
And there they are, again, the cyclin dependent kinases. So, the key moments of the cell cycle, the moments when the cell moves from one phase into the next, are regulated by a series of enzymes known as cyclin dependent kinases, CDK for short. Cyclin dependent kinases work in concert with another set of proteins called, you guessed it cyclins. Together, the cyclins and the CDK enzymes pin a phosphate tail on the RB gene, on the stop sign between cell phases. And that phosphate tail takes the sign down, turns the traffic light from red to green, the cell cycle moves forward.

Ken Shulman 05:47
Now, surprise, surprise, it turns out that there's a whole family of cyclins and an equally large number of CDK enzymes, each with its unique biochemistry and structure. The ones that Livingston and Shapiro and others were interested in were the D family cyclins, and their corresponding enzymes CDK four and six. These were the ones that pin the tail on the retinoblastoma protein, that took down the stop sign between phases G2 and S. So why were doctors Livingston and Shapiro interested in those specific proteins. It wasn't just that they were traffic cops on the road to cell division. It was because those proteins cyclin D1, and CDK 4 and 6, tend to show up in large numbers in several cancers. They team up to help strip down stop signs and cancer cells aiding and abetting cancers, runaway growth. So, researchers knew that certain cancers needed these particular proteins cyclin, D1 and CDK 4 and 6, to divide and grow. Just like they knew that healthy cells also needed the same proteins. At least they thought they knew that. Dana-Farber scientist Peter Sicinski, wanted to test that dogma. In his lab, together with his students and postdocs. Sicinski ran a series of experiments. In the first one, he created a group of mice born without the gene for cyclin D1. That meant the mice couldn't produce the cyclin D1 protein. He repeated the experiment cutting out cyclins D2 and D3. What he wanted to see was whether the mice could develop and grow without the cyclins.

Peter Sicinski, MD, PhD 07:25
And what we have found is that mice lacking individual cycling proteins, cyclin D1, cyclin D2, cyclin D3, develop relatively normally. They were viable. If Cyclin D1 was required for proliferation of every cell type that these mice would never be born. These mice would be, as we say, embryonic lethal. And yet mice lacking Cyclin D1 were born.

Ken Shulman 07:48
And there was more. The bioengineered mice, the ones without the gene for cyclin D1, were also resistant to certain breast cancers, tumors just didn't form. Sicinski and his students did the same experiment, this time with mice that lacked the CDK 4 gene, and they got the same result. These mice also resisted tumor genesis. Breast cancer wouldn't take root. It was all very interesting, a great day at the lab. But there were miles to go before any of this could translate into therapy. It was one thing to prevent the formation of tumors. But to be an effective therapy, a drug needs to block and shrink tumors that have already formed. So, one of Sicinski's postdoctoral fellows engineered a group of mice whose gene for cyclin D1 could be turned off at any stage in life. These mice were born with a working Cyclin D1 gene. They were allowed to mature. The researcher then induced breast cancer in the mice.

Peter Sicinski, MD, PhD 08:47
And she waited until these mice developed breast cancer. And once these mice developed breast cancer, she turned off Cyclin D1. And she found that there was no obvious effect on animal physiology, however, the proliferation of breast cancer cells was completely blocked.

Ken Shulman 09:04
At the same time, another lab shut down CDK 4 in mice, and that seemed to block growth in lung cancer.

Peter Sicinski, MD, PhD 09:11
So collectively, these results indicated that Cyclin D. CDK 4/6 activity is not required for proliferation of the vast majority of normal cells. However, it is essential not only for initiation, but also for maintenance of tumors. And that was the first time that I became, maybe not convinced, but hopeful that chemical inhibition of CDK 4/CDK 6 might have a therapeutic effect.

Ken Shulman 09:41
Like the cell cycle, the drug development cycle has several phases. The initial phase usually takes place in the lab. If that work goes well there may be animal trials. If those go well there are trials with patients. At one point, industry gets involved. In order to attract industry, researchers usually need to identify a target, to single out a protein or receptor that if shut down, will stop, or at least slow the progression of disease. And that target has to be something they can hit. Researchers now knew that blocking either cyclin D1 or CDK 4 and 6 would shut down the cell cycle, keep the cell at the stop sign so it couldn't divide. Now it was a question of which protein to target. Cyclins because of their molecular structure are very hard to hit. So, they decided to take aim at the CDK enzymes. The first generation of small molecules that could inhibit CDK enzymes showed up in the early 2000s. They were great at stopping cancer cells from replicating in the lab. The trouble was, Geoffrey Shapiro tells us, they also stopped a lot of other cells from replicating in patients.

Geoffrey Shapiro, MD, PhD 10:51
And one of the problems with those initial drugs is that there was very little therapeutic window between what was happening in cancer cells and normal cells, and they would stop the proliferation of normal cells as well. And there were toxicities and side effects associated with those initial drugs that were developed.

Ken Shulman 11:10
Still, there was enough momentum in the project to attract the drug companies. Soon there were several second- and third-generation compounds in development, precise and targeted drugs that could single out specific enzymes and spare most others.

Geoffrey Shapiro, MD, PhD 11:25
So palbociclib, ribociclib, and abemaciclib were the first three compounds, palbociclib being the first of them, that were highly selective for CDKs 4 and 6, so they only inhibit CDKs 4 and 6 and because of their selectivity, they are less toxic than the more promiscuous drugs.

Ken Shulman 11:46
The drugs went into Phase One trials with patients who had solid tumors, they seemed safe, even tolerable. But overall, the results were lackluster. They didn't seem to improve patient outcomes. By 2007. Pfizer, which at the time made palbociclib was going to shelve the project. Shapiro and his Dana-Farber colleagues had a different notion.

Geoffrey Shapiro, MD, PhD 12:08
We actually went to Pfizer and asked them if we could test the drug in lymphoma. So, we were interested in the subset of lymphoma called mantle cell lymphoma. And there was interest in mantle cell lymphoma, because that's a cyclin D1, CDK 4 driven disease.

Ken Shulman 12:26
Many hematological malignancies, many cancers of the blood are dependent on cyclin D2 or D3. But mantle cell lymphoma needs cyclin D1 and CDK 4r to proliferate. And CDK 4 was the enzyme the Pfizer drug shut down.

Geoffrey Shapiro, MD, PhD 12:42
So, we said to Pfizer, let's test the drug there and see what happens. And by the way along the way, we will get tumor biopsies for you. And we'll have pre- and on-treatment tumor biopsies, and we will design assays that really demonstrate that this drug inhibited CDKs 4 and 6, will really prove to you that drug was getting into tumor cells and hitting its target.

Ken Shulman 13:06
Shapiro study involved 17 mantle cell patients at Dana-Farber, all were at an advanced stage and had received chemotherapy treatments, a few have undergone bone marrow transplants. As a proof-of-concept experiment, the mantle cell trial was wildly successful, the palbociclib hit its CDK 4 enzyme target in every patient. But even though the drug hit the bullseye in all 17 patients, only five patients experienced significant long-term health benefits. Still 5 out of 17 was a pretty good outcome, especially in a preliminary trial with patients at such a critical stage. But the real positive result was how accurate the drug was. It hit the bull's eye every time — 100%. So, when another company came up with a different molecule, a molecule that targeted a different mantle cell enzyme and produced far better outcomes in that disease than palbociclib, Pfizer didn't blink. Neither did Shapiro. Their eyes opened even wider when they read a 2012 paper out of UCLA, a paper that showed that palbociclib, Pfizer's CDK 4 and 6 inhibitor seemed to arrest growth in the most common form of breast cancer.

Geoffrey Shapiro, MD, PhD 14:20
And so, with that preclinical data coupled with the target engagement data that we had shown in patients, Pfizer became interested in saying okay, let's look at this now in ER positive breast cancer.

Erica Mayer, MD, MPH 14:38
The way we take care of breast cancer patients today is so dramatically different than when I was a fellow in training or earlier on in my career.

Ken Shulman 14:47
That's Erica Mayer. She directs clinical trials in breast cancer at Dana-Farber, where she also sees patients. Clinicians divide breast cancer into several subgroups. The largest subgroup, by far, is hormone receptor positive breast cancer. HR positive breast cancers — they're sometimes called ER positive breast cancers — are driven by hormones, by estrogen and progesterone. These HR positive cancers respond well to hormone therapy, to drugs that cut off the supply of hormones the tumors need to grow. When Erica Mayer joined the Dana-Farber faculty in 2006, hormone therapy had long been the gold standard of treatment for HR positive breast cancer, both in newly diagnosed cases and in cases where the cancer had metastasized, had spread to other parts of the body.

Erica Mayer, MD, MPH 15:37
The hormone, or what we sometimes call endocrine therapy, is the bedrock, the backbone of therapy for hormone receptor positive disease, and it is also one of our most effective therapies for patients with advanced or metastatic breast cancer, if the cancer has come back in the body as stage four or metastatic disease. The backbone medicines include drugs like tamoxifen, or a category of drugs called aromatase inhibitors, we call them AI. There’re three drugs in that group, letrozole, anastrozole, and exemestane. These are all oral pills. They're taken daily, very well tolerated. And these pills have been around for decades. They're taken by millions of women around the world.

Ken Shulman 16:19
Mayer had read about the roles that cyclin D1 and CDK 4 and 6 played in the cell cycle. She knew about the stop sign, the RB gene, and how these enzymes took that sign down. But none of that was front and center on her radar as she traveled to Texas in 2012 for the annual San Antonio Breast Cancer Symposium. There, she watched Richard Finn of UCLA present results from a study called Paloma.

Erica Mayer, MD, MPH 16:48
And Dr. Finn presented these results in December 2012 and showed the patients who got the palbociclib had a quite dramatic improvement in how long the drugs were able to control the cancer before the cancer eventually became resistant and started growing again. It was basically a doubling in the time that the drugs could control the cancer. This was very remarkable.

Ken Shulman 17:11
The Paloma study was a phase two clinical trial of palbociclib, Pfizer's CDK 4 and 6 inhibitor. In the UCLA lab, Finn had already shown that palbociclib could shut down growth in several breast cancer cell lines. Now in the phase two Paloma study, Finn recruited 165 patients with metastatic HR positive breast cancer. He treated one patient group with hormone therapy, with an aromatase inhibitor called letrozole. He treated the second group with a combination of letrozole and palbociclib, the CDK 4 and 6 inhibitor. The second group did markedly better.

Erica Mayer, MD, MPH 17:50
And it really kind of cracked open this whole field of using this category of drugs, CDK 4/6 inhibitor. It really felt like an electric atmosphere in the room that day to see what felt like the dawn of an era.

Ken Shulman 18:06
Mayer was certainly ready to seize that new day. Right there in San Antonio, she and a Dana-Farber colleague sat down with their partners at Pfizer. The UCLA team had shown how palbociclib was effective in patients with metastatic breast cancer in cancers that had spread. But what about patients whose breast cancers hadn't yet metastasized?

Erica Mayer, MD, MPH 18:27
If it works for metastatic disease, maybe it will work for patients who don't have metastatic disease — early-stage patients where using a drug like palbociclib perhaps might prevent patients from ever getting metastatic disease in the first place.

Ken Shulman 18:42
Mayer and her colleagues recruited 160 patients with early-stage hormone receptor positive breast cancer at Dana- Farber and at other centers around the country, patients whose cancers had not yet spread. These patients were already receiving adjuvant endocrine therapy, hormone treatments that follow surgery or chemo or both. In this Dana- Farber led trial, the patients also took palbociclib for two years.

Erica Mayer, MD, MPH 19:09
The trial was not designed to see how well it was working, because you can't do that with a small trial. But it was designed to just see, can we do this? Is this feasible? Can patients tolerate this drug, patients who have gone through their breast cancer treatment? They've had surgery, radiation, chemo, and they are trying to get back to normal life? Can they do that and take this new drug?

Ken Shulman 19:32
Apparently, they could. The trial known as DFC 13559, showed that patients could fold palbociclib therapy into their daily lives, that they could return to work and take care of their families while taking the drug.

Erica Mayer, MD, MPH 19:46
And while this trial was going on, palbociclib continued its clinical development in subsequent Paloma studies for metastatic breast cancer and became approved for metastatic breast cancer in 2015.

Ken Shulman 20:01
In that same year 2015, Mayer began to work on the PALLAS trial, a much larger phase three trial involving 6,000 patients and over 400 cancer centers around the world. This was the trial designed to see whether palbociclib could help prevent cancer from spreading in the first place. One patient group received a combination of letrozole, the hormone therapy, and palbociclib. The second group received letrozole alone. The answer to the question seemed to be no. After five years of follow up, patient outcomes were not improved by the addition of palbociclib. But another study, conducted almost in parallel, with another CDK 4/6 inhibitor called abemaciclib was positive. That drug seemed to prevent metastases. And in late 2021, the FDA approved abemaciclib for adjuvant use, for cases where the breast cancer has not yet spread.

Erica Mayer, MD, MPH 20:59
And, you know, we're really delighted that now we have one of these drugs available to us in the adjuvant setting, as well, to hopefully improve our ability to cure this disease. The approval of the CDK 4/6 inhibitors for metastatic breast cancer, initially in 2015 with palbociclib, a little later in 2017 for abemaciclib and ribociclib, has dramatically changed our treatment paradigms for metastatic breast cancer, and we know are helping patients have longer periods of disease control and helping patients live longer with this disease.

Ken Shulman 21:37
Doctors and researchers have good reason to be delighted with CDK 4/6 inhibitors, with these drugs that stop cancer cells from dividing and block tumor growth. The drugs help patients live longer, longer is better, but it's not forever. Eventually, even with these therapies, metastatic breast cancer cells find a way to start dividing again.

Geoffrey Shapiro, MD, PhD 21:59
These treatments, unfortunately, like most kinase inhibitors, these are not curative treatments, and the cancer cells will always evolve ways around them.

Ken Shulman 22:08
Geoffrey Shapiro says a lot of his lab work today looks at drug resistance. At the way cancer cells learn to evade capture or control. He wants to understand the mechanisms of persistence so he and others can overcome them. But it's tricky work. Cancer cells can be relentless, as pesky as a nighttime mosquito, as clever as Houdini and handcuffs. There's a whole repertoire of escape moves that Shapiro and his colleagues need to identify, analyze, and neutralize.

Geoffrey Shapiro, MD, PhD 22:37
But the one that my laboratory is focused on is an overexpression of CDK 6. So, it's very interesting in breast cancer cells, in ER positive breast cancer, there's a lot more CDK 4 than CDK 6. But we're finding as cells become resistant to these drugs, they begin to express very, very high levels of CDK 6. And the levels are high enough to overcome what the inhibitor can handle. So, there's more CDK 6 than the drug can bind to. And so that extra CDK 6 in the cell, actually helps the cell overcome the cell cycle arrest, and drive the cell into S phase, and the cells become resistant to the drug.

Ken Shulman 23:18
drug resistance is the next chapter in the CDK 4/6 story. And Shapiro is not alone in writing it. He's helping Peter Sicinski and his postdocs and students design laboratory trials, this time to figure out why some forms of breast cancer don't respond to CDK, 4/6 inhibitors from the get-go. And he's supporting Erika Mayer, as she takes a closer look at the PALLAS study to better understand why palbociclib fell short there. And why abemaciclib, a similar drug hit the mark. It's a team effort, an assault launched on a number of fronts.

Geoffrey Shapiro, MD, PhD 23:52
I think the environment here has been very good for conducting that work and has really facilitated the success in the field. You know, the laboratory work drives us to clinic, the clinic drives us back to the laboratory for more discovery or understanding resistance or how to get things to work better. And that synergism is really part of the blood and the mission of the Dana-Farber.

Ken Shulman 24:14
We've spoken this season and frequently about momentum, how research generates new research forming a constant loop of revision and review. We've traced the drug development cycle from laboratory to clinical trials, from bench to bedside, and almost just as often from bedside back to bench. But Erica Mayer reminds us of another element in the cycle, a key element that helps sustain this momentum. The patients who participate in trials.

Erica Mayer, MD, MPH 24:42
And so that's given me the opportunity to include my own patients from clinic in this really important work. and being able to on a very granular level, sit with my patient and talk about how they're doing see their progress, talk about their experience on a trial and at the same time, see the data presented in front of 10,000 people, read the article in the New England Journal, and then bring that back to my patient and say, hey, look, this is the trial you are involved in and it's on the front page of our top medical journal. You know, that's an incredible feeling. And you know, I think my patients feel that too. They're really happy and excited to be part of this progress.

Ken Shulman 25:30
Next time, a team of Dana-Farber, doctors and researchers locate a genetic mutation that causes lung cancer and find a way to counter it.

25:39
There were a subset of patients who had dramatic responses. People called it a Lazarus-like effect. They would be on their deathbed, they would take Iressa, and they would get up and they would walk away.

Ken Shulman 25:52
I'm Ken Shulman, and this is Unraveled, a Dana-Farber Cancer Institute podcast.

Season 2, Episode 4: Making it Personal: Targeted Cancer Therapies

By the early 2000s, researchers at Dana-Farber and elsewhere knew that a certain protein appeared in many tumors taken from lung cancer patients. Based on that data, doctors started treating those patients with a drug that inhibited that protein. Unfortunately, for reasons that weren't clear at the time, most of the patients didn't respond — but there was a small percentage of patients, about one out of ten, who did. People called it a "Lazarus-like effect."

So how can doctors know which of their patients is Lazarus? How can they know which ones will respond to a drug and which ones won't? In other words, how can they interpret data so it makes sense? That's the subject of this episode of Unraveled about precision medicine and the EGFR discovery. It's an episode that begins with a riddle and ends with a roadmap.

Season 2, Episode 4 Transcript

Ken Shulman 00:00
This twenty-first century, this age we live in, sinks and swims in a sea of data. Sixty percent of all voters prefer this. Twenty percent of all consumers buy that. Data and statistics shape political campaigns and marketing strategies. They inform many of our choices, from the most trivial to the most important. They even give meaning to our lives.

But statistics alone can't tell you which path to choose. It's not just that they can be false or misleading. It's that they're incomplete. It's not enough to know that, say, four out of five dentists surveyed recommend a certain brand of sugarless gum for their patients who chew gum.

If we're planning on using that data to drive strategy, or to make sense of our world, we need to know a lot more. Where do those four dentists live and practice? Who are their patients? And what's up with the holdout? Why won't that stubborn fifth dentist sign on? I bet you think you know what's coming next. An episode on big data and artificial intelligence. Well, that episode is in the hopper. We'll close this second season of Unraveled with it.

Today's episode also has to do with data. But it's about the difference between data and knowledge. The way we can move from what we see to what we know.

Here's an example: By the early 2000s, researchers at Dana-Farber and elsewhere knew that a certain protein appeared in many tumors taken from lung cancer patients. Four out of five patients, actually. So, based on that data, doctors started treating those patients with a drug that inhibited that protein, a drug called Iressa. Unfortunately, for reasons that weren't clear at the time, most of the patients didn't respond.

But there was a small percentage of patients, about one out of ten, who did.

Matthew Meyerson, MD, PhD 01:55
People called it a Lazarus-like effect. They would be on their deathbed, they would take Iressa, and they would get up and they would walk away.

Ken Shulman 02:04
So how can doctors know which of their patients is Lazarus? How can they know which ones will respond to a drug and which ones won't? In other words, how can they interpret data so it makes sense? That, of course, is the question. And it's the subject of this episode about precision medicine, an episode which begins with a riddle and ends with a roadmap.

I'm Ken Shulman and this is Unraveled, a Dana-Farber Cancer Institute Podcast.

Lung cancer is, unfortunately, the leading cause of cancer death. In the U.S., more than one hundred thousand people succumb to the disease every year. Two million die of lung cancer each year worldwide.

There are two major categories of lung cancer. The largest of these, by far, is non-small cell lung cancer. And the largest subgroup in non-small cell lung cancer is adenocarcinoma. Adenocarcinoma is an aggressive form of cancer; it's almost always incurable. Five-year survival rates are low.

How low? It depends. It depends on whether the cancer has spread when first diagnosed. And on a lot of other factors. Besides, survival rates are statistics, not certainties. They're data gleaned in the rear-view mirror. Not read in the crystal ball. Now these survival rates aren't meaningless. They can suggest, with reasonable accuracy, that four out of ten patients with lung adenocarcinoma won't live past the five-year mark. But they can't tell an individual patient if they're one of those four. Like most statistics, this one applies to groups. Not to single patients.

Adenocarcinoma, like all cancers, shows up as a glitch in the growth program. The signals that regulate cell replication get crossed or squelched or distorted. In the resulting chaos, cells start dividing on their own. And the proteins that normally slow down or stop the frenzy are silenced.

Bruce Johnson, MD 04:17
So what happens is that it's a little bit like putting your foot on the accelerator.

Ken Shulman 04:22
That's Bruce Johnson. He ran the lung cancer program at Dana-Farber from 1999 to 2013. He's still at Dana-Farber now as senior advisor to the President.

Cell division, he tells us, is triggered by a switch. By a receptor — by many receptors, actually — on the cell surface. In normal circumstances, in healthy growth, an agonist — another protein — binds with that surface receptor. And that surface receptor, now activated, emits a flurry of signals that in turn trigger even more signals downstream. The entire cascade of signals orchestrates the process by which one cell becomes two. Now in this type of cancer, Johnson explains, a switch called EGFR — the epidermal growth factor receptor — just goes rogue. It turns on all by itself.

Bruce Johnson, MD 05:13
As I said before, it's like putting your foot on the accelerator and leaving it on there. And the term we use for it is constitutive activation, meaning that the receptor is turned on. It doesn't need an agonist, it doesn't need a signal to turn on, number one. Number two is that there's not a signal that tells it to turn off. So it sends a signal to the cell that it should have unmitigated growth.

Ken Shulman 05:36
35 years ago, when Johnson started his career, there were no approved treatments for non-small cell lung cancers, including adenocarcinoma. The only treatable lung cancer was small cell lung cancer.

Bruce Johnson, MD 05:48
And there we use conventional chemotherapy. And we were doing that going back in the 1960s and seventies. We didn't have any therapy for non-small cell lung cancer. And in the 1990s we got chemotherapy. And it was sort of stagnant for about 15 years.

Ken Shulman 06:03
With chemotherapy, lung cancer patients got a few extra months before the cancer returned. But as time passed, and as the field evolved, researchers gradually traced the network of pathways that generated and sustained lung cancer. Much of their work was piecemeal — single tiles or discoveries that only made real sense after they were set into the mosaic with all the other tiles.

One of these tiles was the EGFR protein, the epidermal growth factor receptor. Researchers noticed that close to 90% of lung cancer samples they examined in the lab contained unusually large quantities of EGFR. The same protein shows up big in a lot of other cancers too.

EGFR is one of those switches — one of those surface receptors — that start the ball rolling in cell division. So it made sense, at the time, to conclude that a mutant EGFR protein could be driving lung cancer. Therefore, it also made sense to treat lung cancer patients with drugs that would inhibit the EGFR protein. By inhibiting the EGFR protein, doctors hoped to stop or even reverse tumor growth.

So doctors, including Bruce Johnson, began treating lung cancer patients with an EGFR inhibitor called gefitinib. The responses were dramatic, like Lazarus from the grave — but those responses weren't widespread. Only one in 10 patients showed improvement. The remaining nine showed no response at all.

Researchers and doctors were justifiably baffled, because they couldn't see any connections or shared characteristics linking that lucky 10 percent.

Bruce Johnson, MD 07:43
There were certain clinical subgroups that were more likely to respond than others. So the things that we had noticed were that women were more likely to respond than men. People who had never smoked were more likely to respond than those who had a smoking history. It was also observed that folks from Japan were more likely to respond to these molecules.

Ken Shulman 08:07
So that's a good one. A woman, a non-smoker, and a person from Japan walk into a bar. Where do you go with that? And what on earth, if anything, did these three groups have in common?

2001 is a watershed year for cancer science and treatment. It's the year that a drug called Gleevec was approved for a relatively rare blood cancer called chronic myelogenous leukemia. Now lots of drugs are approved every year. And all of them, in some way, are novel. But Gleevec was quantum leap.

Bill Sellers, MD 08:42
Chemotherapy mostly is directed at the idea that cells are either proliferating too fast or they're just making too much DNA. And we just need to disrupt that.

Ken Shulman 08:55
That's Dana-Farber doctor Bill Sellers.

Bill Sellers, MD 08:58
But that doesn't really say anything about the cause of the cancer. What's the root cause?

Ken Shulman 09:05
The root cause. Cancer, as we've said many times, is the result of a misprint in the cellular growth program. A genetic typo that sends the cell making machinery into a dangerous overdrive, where it churns out hordes of maverick cells.

Chemotherapy attacks these maverick cancer cells and, hopefully, kills them.

Gleevec, on the other hand, keeps these cells from being born. If chemo is a machine gun, cutting down everything in its path, Gleevec is a sniper's rifle. A targeted therapy that shuts down the machinery that makes the rogue cells in CML.

How does it work? It blocks the fusion of two genes located in the so-called Philadelphia chromosome. This is the fusion that drives CML. Blocking the fusion stops the disease.

Bill Sellers, MD 09:54
The world was on fire about this clinical result because essentially every patient was responding and there was minimal toxicity. And, you know, this first idea of a smart molecule rather than just sort of bluntly trying to kill cancer with chemotherapy.

Ken Shulman 10:10
In late spring of 2001, Sellers and his Dana-Farber colleague Matthew Meyerson were on a plane together from Geneva. They were returning from a conference they'd attended in Switzerland. It was an exciting time for cancer science. The field was energized by Gleevec — a discovery that opened a brave new world for research and therapies. And thanks to the human genome project, which was nearing completion, we had a pretty good idea about the structure of our DNA.

But the cancer genome — the galaxy of genetic mutations that drove cancer — was almost completely uncharted. And that was the galaxy that Gleevec had opened for exploration. It was a rare moment, when the universe seemed to be made not of atoms, but of possibility.

Still, there was reason for caution. Skeptics pointed out that Gleevec had worked in a hematological malignancy — in a blood cancer, and a rare one at that. Sellers and Meyerson wanted to find out whether something similar could be done with solid tumors.

The pair talked almost nonstop for the entire eight-hour flight to Boston.

Matthew Meyerson, MD, PhD 11:19
And Bill said, you know, we can't sequence the whole genome. It's way too expensive. At that time, he said, we should just sequence something that's really important.

Ken Shulman 11:29
That's Matthew Meyerson. Today, he holds the Charles A. Dana Chair in Human Cancer Genetics at Dana-Farber. In 2001, he and Sellers were relatively low-ranking faculty members.

Matthew Meyerson, MD, PhD 11:41
And he said, you know, Gleevec tells us the receptor tyrosine kinases are really important. Let's sequence the receptor tyrosine kinase genes. And I said, oh, Bill, that's a great idea, so we decided to do it.

Ken Shulman 11:54
We've talked about kinases in previous episodes. About those enzymes that help drive cell division. Gleevec, the game changing targeted therapy, stops cancer by shutting down a receptor tyrosine kinase. EGFR — the growth factor present in large numbers in lung cancer and many other cancers — is also a receptor tyrosine kinase.

Now, there are about 500 kinases in the human genome. A small minority of these belong to the tyrosine group. That minority includes the enzyme that Gleevec shuts down. And, as mentioned, it includes EGFR.

But even that small minority — the tyrosine kinase group — was still too big a field for Meyerson and Sellers to scan. They needed to fine tune their search even more, to maximize their resources and to leverage their chances of hitting paydirt — of producing actionable data, knowledge that could lead to effective targeted therapies for cancer.

Matthew Meyerson 12:52
We realized, well, actually, it's too expensive to sequence all the receptor tyrosine kinase genes. We need to sequence small pieces of them. And we found the pieces that have the most mutations. And we started our sequencing in those pieces that have the most mutations. So that's how we started the project.

Ken Shulman 13:11
Gene sequencing was a lot slower and a lot more expensive 20 years ago than it is today. Sellers and Meyerson narrowed their search further, focusing on just three cancers: prostate, lung, and glioblastoma. They sifted through close to 2000 exons — those pieces of DNA that code for proteins. Of the many kinases they sequenced, four emerged that were of interest. Four genes that coded for some of the most common aberrant proteins that flourished in those three cancers.

One of those four genes was the EGFR gene.

The two junior faculty members weren't working in a vacuum. There was a lot of information available about EGFR. But none of it was conclusive. None of it showed a clear path towards treatment, or hinted at why only one in ten patients responded to EGFR inhibitors. The information came in scattered, like single tiles in a mosaic that couldn't yet be read.

Here's what Meyerson and Sellers did know. They knew from clinical reports that EGFR inhibitors like gefitinib were far more effective in patients from Japan than they were in patients from the United States.

And like Bruce Johnson and everyone else, they didn't know why that was.

Now the next part of this story will appear quite differently to different people. As in an optical illusion where you either see a human face or you see a camel. If you believe in fate, this part will read as destiny. If you believe in coincidence, it will read as dumb luck.

So here it is. In order to sequence DNA from lung cancer patients — in order to analyze the genes that coded for the receptor tyrosine kinase in EGFR — Meyerson and Sellers needed tumor samples. Lung cancer tumor samples. The lung cancer samples that the pair found to work with at Dana-Farber, the ones they would sequence, just happened to come from Japan. From the country where patients, for reasons still unknown, seemed to respond to gefitinib in greater numbers. The very country where the EGFR inhibitors worked best.

The thing is, Meyerson and Sellers didn't choose to work on Japanese samples. The samples just happened to be there. They'd been brought to Dana-Farber by a visiting surgeon from Tokyo. He left them behind when he returned home.

Matthew Meyerson, MD, PhD 15:31
And we had a bunch of lung cancer samples from him, and we were supposed to use them for another project. And that project never quite got off the ground. So when we started this project, we said, 'You know what, this guy has been waiting patiently. We should use his samples.' And that was actually the rationale.

Ken Shulman 15:48
So Meyerson and Sellers started sequencing the DNA from the Japanese samples. They found several mutations in the gene for EGFR. The human epidermal growth factor receptor.

The discovery looked promising. These mutations were located in a region of the gene that Meyerson had already linked to other mutations linked to lung cancer. So the site felt right. There was also an amino acid residue on the site, similar to another residue linked to yet another mutation tied to cancer.

And there was more. It wasn't just that this first sample, this first patient with the EGFR mutation, was from Japan.

Matthew Meyerson, MD, PhD 16:27
It was a woman patient. Women had responded better to gefitinib than men. It was a patient with lung adenocarcinoma, the subtype that had responded best to gefitinib. And finally, this patient was a nonsmoker, and nonsmokers had responded better. So we immediately thought, okay, this is probably the explanation for why gefitinib is working in these patients because they have EGFR mutations.

Ken Shulman 16:54
So we're back to the bar joke. Except this time it has a punch line. Japan. Female. Non-smoker, with adenocarcinoma. This one patient, the one with the EGFR mutations in her DNA, just happened to straddle all the disparate groups that had responded to gefitinib. To the EGFR inhibitors. Even if you don't believe in fate, it's hard not to see an invisible hand behind this discovery.

And the discovery? It wasn't fully confirmed, but from their work, Meyerson and Sellers were pretty sure that the patients who responded to gefitinib — to the EGFR inhibitor — would all display the same genetic mutation.

Whether or not the drug worked didn't depend on whether the tumors showed large amounts of the EGFR protein. Because many tumors showed large amounts of the EGFR protein. And most of those patients didn't respond to the drug. Successful treatment, it seemed, depended on whether or not the patient showed a mutation in the EGFR gene, even though the drug targeted the EGFR protein.

Let me repeat that. It didn't matter whether patients had the EGFR protein in their tumors, because patients weren't responding to the therapy based on the makeup of their tumors. They were responding to it based on the makeup of their DNA.

Ken Shulman 18:20
We spoke earlier about the difference between data and knowledge. Between information that's interesting and information that's actionable. It's interesting, for example, to learn that the EGFR genetic mutation that Meyerson and Sellers observed in the lab shows up in approximately 40% of patients from Japan and East Asia. And that the same genetic mutation only shows up in about 10% of U.S. patients.

But what do you do with that? Give the drug to all Asian patients and hope the coin comes up heads instead of tails?

In order for that information to become actionable, we need a way to get the drug to the patients we know will respond to it, and not waste the time of patients — who may not have a lot of time to waste — with therapies we're pretty sure won't work.

In short, we need a screen.

Pasi Jänne, MD, PhD 19:08
Until that time, lung cancers were treated as one size fits all. And this helped identify a subset of lung cancers that were defined by this genetic alteration and that was predictive of efficacy of treatment.

Ken Shulman 19:23
That's Pasi Jänne. He directs the Lowe Center for Thoracic Oncology at Dana-Farber. In 2003, Meyerson and Sellers sat down with Jänne and with Bruce Johnson to share their findings. At the time, Jänne was a fellow in Johnson's lab.

Johnson was enthusiastic and wanted to collaborate. Under his direction, Jänne started collecting lung cancer biopsy samples from patients treated at Dana-Farber. He and Johnson also studied lung cancer cell lines to see which of them responded to gefitinib, the EGFR inhibitor.

There was one cell line that Johnson had brought to Dana-Farber from the National Cancer Institute, where he'd previously worked. Cell line NCI H3255.

Pasi Jänne, MD, PhD 20:08
And in the in the course of the studies, we helped identify that there was one outlier cell line that was particularly sensitive to an EGFR inhibitor, and we ultimately decided to sequence those cell lines, the sensitive and resistant ones, and noted that the outlier cell line actually had an EGFR mutation.

Ken Shulman 20:28
It was the same observation that Meyerson and Sellers had made. Jänne and Johnson treated several hundred lung cancer patients with gefitinib. As clinicians in multiple locations had already observed, some of these patients responded dramatically to the inhibitors, while others did not respond at all.

Pasi Jänne, MD, PhD 20:48
And so I helped identify those individuals, along with Dr. Johnson. And then, as we were putting this story together, we ultimately studied those tumors from the patients that were both sensitive and resistant to the medicine clinically, and noted that individuals whose tumors responded to the medicine all harbored EGFR mutations where that was not the case in the patients whose tumors did not respond to the medicine.

Ken Shulman 21:14
This was confirmation. And this was actionable. Working in parallel, both research pairs had found a predictive biomarker. They'd located a piece of information that, if properly harvested, could predict, with reasonable certainty, whether a specific drug would work in a specific patient.

Meyerson and Sellers and Johnson and Jänne published their findings in Science Magazine. And it didn't take Dana-Farber very long to act.

Pasi Jänne, MD, PhD 21:42
The original publication happened in April of 2004, and by August of 2004, we had set we had set up a system where we would be able to sequence patients' tumors before they started on any therapy. Our colleagues in pathology set up a clinical test, and then we were able to order it as a test and test individuals' tumors, whether or not they had the mutation, and allowed us to then choose the EGFR inhibitor therapy. And that's still done today. And it's one of the most common molecular tests that is done in lung cancers around the world today.

Ken Shulman 22:20
Today, when a patient with lung adenocarcinoma comes to Dana-Farber for treatment, tissue from their tumors is scanned for nearly a dozen mutations. There are targeted treatments available for almost 40% of all lung cancers. Some of these can even substitute for chemotherapy. Gefitinib, the original EGFR inhibitor, has been replaced by second generation inhibitors, drugs that are far more aggressive with the mutant protein and, at the same time, far less toxic for the wild or healthy form. End result: Tumors shrink faster. Remission lasts longer. And side effects are milder.

It's a thrill for doctors like Pasi Jänne to see those results. In patients. And even more in their scans.

Pasi Jänne, MD, PhD 23:05
A scan is something you can see with your own two eyes. You can see a before and after … person was feeling better. It's a remarkable feeling to see something that you have been involved in impact another human being. And, you know, being in a cancer center, you know, we have a very focused mission and that is to help our patients. But to be able to see that with your own two eyes, it's great. It's fantastic.

Ken Shulman 23:32
The article in Science, the article on EGFR mutations, is a sort of Magna Carta of medicine. Since 2004, the year it was published, it's been cited more than 10,000 times. If you're wondering what that means, consider that for most papers, anything over 150 citations is like batting 300. 10,000 citations is like winning the Triple Crown ten years in a row.

And the work behind the article, the work by Johnson and Jänne and Meyerson and Sellers and others, would signal the beginning of a promising era in cancer therapy: the era of precision medicine, in lung cancer, and in many other cancers with solid tumors. An era where doctors test for predictive biomarkers. Where they screen individual tumors for specific pieces of genetic information, information that can guide doctors towards tailor-made, timely treatments.

Matthew Meyerson, MD, PhD 24:27
We knew, in the case of Gleevec, a really clear connection between a molecular alteration — in this case, the Philadelphia chromosome and the response to therapy. I think that the EGFR discovery really brought that same concept to what we call the solid tumors, you know, the era of, you know, lung cancer, breast cancer, colon cancer, prostate cancer. And it said, okay, we can build molecularly targeted therapies, in particular in lung cancer. And it led to a flood of new targeted therapies.

Ken Shulman 25:02
For lung adenocarcinoma patients, precision medicine has changed the landscape — and stretched the horizon. With the latest generation EGFR inhibitors, remission can last as long as two years. And it's not unusual for some patients to survive for three years, five years, and even more.

But. And there's always a but. While this era of precision medicine is well underway, it is nowhere near completed. There are other cancers and other drug sensitivities to be linked to other genes.

And, even in precision medicine, there's drug resistance. EGFR inhibitors are good. But they are not forever. Eventually, no matter how robust the response, cancer finds a way to outfox or outflank or simply outlast most small molecules and antibodies. Much of Pasi Jänne's work today is on drug resistance.

Pasi Jänne, MD, PhD 25:54
And so we're trying to understand is, how do you best combat resistance? Do you wait for it to happen and then figure out why resistance happens and then develop a new strategy to overcome resistance? Or do you develop a… sort of a cocktail of medicines from the very beginning to delay or prevent resistance from happening in the first place? And so we're trying to model many of those scenarios in the laboratory and then ultimately take those findings and test them through clinical trials.

Ken Shulman 26:25
And there it is again. The bench to bedside loop at Dana-Farber that supports and encourages this brand of breakthrough. The commitment to harvest information, to gather isolated shards and tiles, and over time, working as a team, to transform those fragments into a mosaic. Into a big picture that shows us not only how cancer works, but also how we can stop it.

In 2005, one year after the landmark publication in Science, Bill Sellers left Dana-Farber for an 11-year stint at Novartis. There, as head of cancer drug discovery, he brought almost 40 drugs into phase one trials. Ten of these drugs are approved for treatment now — an astonishing result for oncology, where only one out of every 20 phase one therapies makes it to market. Sellers returned to Dana-Farber in 2016 as a senior faculty member. He recalls, with fondness, the work he did with Meyerson.

Bill Sellers, MD 27:21
Good thing we were young. And I think also it was good that Matthew was bold, too. So to take a step like this for our careers at sort of a junior level required some boldness. And I think this is why the partnership was exciting to me, because I know for myself, I don't think I would have had the guts to do it by myself.

Ken Shulman 27:41
Meyerson, too, is grateful for the teamwork.

Matthew Meyerson, MD, PhD 27:44
You know, one of the really special pieces of this story, I would say, is partnership. We established early on a joint group that ended up being a wonderfully fruitful joint group.

Ken Shulman 27:57
Meyerson and Sellers and Jänne and Johnson press on at Dana-Farber. Looking for cures, not the cure. They know, from experience, that there's no magic bullet.

They know that discoveries that lead to new knowledge can also lead to new challenges. And that those challenges grow more complex at each new level. But they also know, almost twenty years after they published their EGFR findings, that there is real progress. That there's momentum. And that the best way to maintain that momentum is by working together.

Matthew Meyerson, MD, PhD 28:31
And all of us, you know, this, you know, EGFR discovery. We made our first discoveries in 2003. We published the paper in 2004. But all four of us are continuing, you know, Bruce, Pasi, Bill Sellers and I, we're continuing to work together closely today. And so I think that that partnership piece is a really special thing that has enabled us, and it's also a really special feature of Dana-Farber science in particular. That's been a great part of our scientific culture.

Season 2, Episode 5: With a Little Help From My Friends: Combination Immunotherapy

If there is such a thing as a holy grail in cancer research, a secret spell or golden ring that can ward off any and all forms of the disease, it probably lives somewhere in the realm of immunotherapy. A new category of immunotherapy drugs, called checkpoint inhibitors, take the brakes off the immune system and let the T cells do their job: attack cancer. They've been approved for more than 25 different types of cancer.

However, the average response rate for checkpoint inhibitors is 20 to 30%, which means that it's not working for 70% of the people who take it. So that's the project now: to extend the benefits of immunotherapy to all cancer patients, instead of to just a few. And the best way to do that, it seems, is by doubling down, by pairing immunotherapy with other therapies so patients can benefit from both. It's called combination immunotherapy, and it's what this episode of Unraveled is about.

Season 2, Episode 5 Transcript

Ken Shulman 00:03
If there is such a thing as a holy grail in cancer research, a secret spell or golden ring that can ward off any and all forms of the disease, it probably lives somewhere in the realm of immunotherapy.

Gordon Freeman, PhD 00:16
The thinking for over 100 years it'd be if you could just get the immune system to attack the cancer. That would be wonderful. And so lots of ways were tried, but none of those would work.

Ken Shulman 00:33
That's Dana-Farber's Gordon Freeman. He may not have found that golden ring, but he got a very good look at it. Gordon Freeman discovered the PD1/PDL1 pathway, it's a pathway involving two proteins, one on the surface of a T cell, our first line of defense, the other on the surface of a cancer cell. Freeman saw that when these proteins bind, they trigger a checkpoint in the T cell. As a result, instead of attacking the cancer cell, the foreign invader, the T cell shuts down. And instead of being killed, the tumor cell slips past unharmed, to grow and multiply.

Ken Shulman 01:22
That's what Freeman saw and mapped in the year 2000. And his map led to the development of a new category of drugs called checkpoint inhibitors. These are drugs that block that PD1/PDL1 pathway, and other similar pathways, drugs that take the brakes off the immune system and let the T cells do their job. Today, checkpoint inhibitors have been approved for more than 25 different types of cancer; countless patients have benefited from them. However, an average response rate is 20 to 30%, which means that it's not working for 70% of the people who take it. And so the real goal now is to bring benefit to those 70% where it's not working yet.

Ken Shulman 02:11
So that's the project now: To extend the benefits of immunotherapy to all cancer patients, instead of to just a few. And the best way to do that, it seems, is by doubling down, by pairing immunotherapy with other therapies so patients can benefit from both. Sound logical? Sure. Sound easy? Not on your life. It's called combination immunotherapy. And it's what this episode is all about.

Ken Shulman 02:46
I'm Ken Schulman and this is Unraveled, a Dana-Farber Cancer Institute podcast.

Ken Shulman 03:03
The human immune system is almost indescribably complex. There are T lymphocytes and B lymphocytes, neutrophils and basophils, chemokines and cytokines and macrophages and many other elements, each with their own roles and responsibilities. The system's great variety is one of its greatest assets. The deep arsenal enables it to fend off any number of threats, viruses, bacteria, and quite often, cancer.

Ken Shulman 03:32
But that same variety, that teaming network of cells and factors and signaling that makes up our immune system, also makes it very hard to read. And to blur the picture even more, the immune system doesn't stand still, when it fights cancer. Its response changes as the tumor develops and evolves the tissue and proteins and blood vessels surrounding the tumor, what we call the tumor microenvironment. That changes as well often in response to treatment.

Ken Shulman 04:02
The long and the short of it: The cancer that doctors treat today may be very different from the cancer they treated a year ago in the same patient. All this is to say is that sometimes trying to recruit the immune system in the war against cancer is almost like playing four-dimensional chess.

Patrick Ott, MD, PhD 04:25
I think that PD1 inhibition in some sense was a godsend, right?

Ken Shulman 04:30
That's Patrick Ott. He directs clinical sciences at the Center for Immuno-Oncology at Dana-Farber. He also sees patients in the clinic. Ott says that checkpoint inhibitors, the ones that take the brakes off the immune system, are more or less a miracle.

Patrick Ott, MD, PhD 04:46
Knowing about the complexity — if you're a person who thinks like glass half empty, the more you understand and comprehend the tools that in some way that the tumor has to defend an immune attack — you're gonna think like, how can immunotherapy ever be effective but for some reason, I would argue that we don't fully understand why. If you block that one inhibitory pathway, which is PD1/PDL1, one really converts an ineffective, anti-tumor, immune response to an effective one. /p>

Ken Shulman 05:21
Like the immune system that it constantly tries to outwit, cancer has an equally eclectic bag of tricks to draw from. It uses those tricks to evade detection and destruction. The PD1/PDL1 pathway, the pathway cancer uses to deactivate our T cells, is a mainstay in that repertoire, a top 10 maneuver. And as Ott points out, targeting that single maneuver is sometimes enough. But sometimes it isn't. Because there's so much more going on in the ground war between cancer and the immune system.

Ken Shulman 05:57
Cancer uses other pathways to shut down our defenses. Sometimes instead of, and sometimes in addition, to PD1. Chemotherapy can reset the entire immune table with dying cells releasing all sorts of signals into the bloodstream. Those signals can set off a new cascade of immune responses. And then there's drug resistance, the way tumors learn to blunt the effective therapies we hit them with. In this broader context, a 20 or 25% response rate with immunotherapy does look like a miracle. But if it's going to fulfill its promise, immunotherapy needs backup and follow up.

Patrick Ott, MD, PhD 06:37
So it's great to see a 20% response rate. The next step is well, why did the other 80% not respond? Right? And then the beginning of sort-of the notion in the field was, well, now that we have this proof of principle, we basically use PD1 inhibition in some way that's called as a backbone of therapies, where you basically treat a patient with PD1 inhibition and then add on other things.

Ken Shulman 07:08
So how do we know which other things to add on to PD1 therapy? Let's take melanoma, a disease that Ott specializes in. Melanoma also shuts down T cells. It targets a different receptor called CTLA4. Like PD1 , CTLA4 is a checkpoint on the T cell. When triggered CTLA4 deactivates the T cell. Cancer cells waltz by.

Ken Shulman 07:37
Freeman's work sparked a surge in research in PD1 and on pathways that regulate our cancer immune response. And that surge led to a wave of clinical trials of checkpoint inhibitors, drugs that could take the brakes off the immune system and help the T cell get its mojo back.

Ken Shulman 07:57
In 2011, the FDA approved CTLA4 inhibitors for treatment in metastatic melanoma. It was the first checkpoint inhibitor to gain approval. And those CTLA4 inhibitors work pretty well alone. More than 25% of melanoma patients experience long-lasting remissions or cures.

Ken Shulman 08:25
In 2014, three years after it approved the CTLA4 inhibitors, the FDA approved the first PD1 inhibitors. It was an even bigger deal. Both pathways shut down T cell activity, but CTLA4 tends to shut them down at the early stage of disease. PD1 works its black magic later on, when T cells start to give up after a long battle with cancer, and PD1 controls a lot more cell signals than CTLA4 does. So in 2014, researchers started trying the two drugs out together.

Patrick Ott, MD, PhD 09:05
And that improved outcomes really dramatically. So PD1 inhibition alone, CTLA4 inhibition alone, independently, were effective, but the combination was even more effective. And that became essentially a standard of care for melanoma in the first line, but also was approved in other cancers including lung cancer, bladder cancer and kidney cancer. So as an example, when you combine two immune checkpoint inhibitors, you can improve the efficacy.

Ken Shulman 09:35
That's one combination immunotherapy success story, where two drugs are better than one. In this case, you combine two checkpoint inhibitors, one that works on PD1, the second that blocks CTLA4. This two-drug combo improved outcomes in a broad range of patients across a broad range of cancers. Success stories like this one spawned thousands of clinical trials with teams of scientists and clinicians and drug companies all searching for that elusive one-two punch. But not all of these trials lead to a knockout victory.

Patrick Ott, MD, PhD 10:12
There have been molecules that were targeted by or that were investigated by many different companies. And they were, you know, successes and early trials where there was like, you know, encouraging response rates. But there has certainly been a good number of stories where initially there was quite a bit of enthusiasm in these initial trials that didn't pan out later on when there was more rigorous, randomized, later-stage trials.

Ken Shulman 10:45
Gordon Freeman's discovery of the PD1/PDL1 pathway was a milestone in cancer research and therapy, a quantum leap forward, put in any cliche you want — it was that important. Important because this discovery demonstrated, among other things, that immunotherapy can work, that it can help our immune system, as exquisitely intricate and interconnected as it is, stare down many forms of cancer. Immunotherapy is already a big part of that help. But almost everyone in oncology believes combination immunotherapy can be an even bigger help.

Ken Shulman 11:28
To that point, there were over 5000 ongoing clinical trials for combination immunotherapy in the United States alone. There are trials that pair immunotherapy with chemotherapy, with another immunotherapy, or with drugs that block the formation of blood vessels. There are trials that combine immunotherapy with precision therapy, with drugs that target specific genetic mutations. And there are trials testing on immunotherapy prior to and following surgery. That's a lot of trials, a lot of possible combinations, outcomes and outliers. In physics, we know that every action generates an equal and opposite reaction. In medicine, it's not quite that linear. There are just too many unknowns, too many variables, especially when testing two drugs instead of one. And with all these variables, all these possible combinations of therapies, how can clinicians know which ones are worth pursuing?

Ken Shulman 12:31
Sometimes you just have to go with what you see. And that's what F. Stephen Hodi did. He's a doctor at Dana-Farber, where he directs the Melanoma Center. He also directs the Center for Immuno-Oncology. In the early 2010s, Hodi observed how some of his metastatic melanoma patients were doing remarkably well after they'd received checkpoint inhibitors. A prospering patient is always a welcome sight. But Hodi observed something else.

F. Stephen Hodi, Jr., MD 13:04
When you looked at their biopsies, we actually saw that the immune system was attacking the blood vessels feeding the tumor deposits and seemed to spare normal blood vessels. And so it was kind of an interesting thought that the immune system may be selective against the blood vessels feeding tumors very specifically.

Ken Shulman 13:22
That bite requires a bit of commentary. Cancer has a whole bag of dirty tricks it uses to survive. One of these is something called angiogenesis. Angiogenesis is the creation of new blood vessels. In most circumstances in a healthy organism, angiogenesis is a helpful process. It's how we build capillaries to bring oxygen to organs and tissues. It's how we channel blood to cuts and scrapes as they heal. With cancer, though, that helpful process becomes harmful. Because many solid tumors also promote angiogenesis. They secrete growth factors, factors that help form dedicated blood vessels to redirect blood flow to tumors. This redirected blood helps the tumors thrive and grow. Again, all this happens at the expense of healthy tissue.

Ken Shulman 14:14
What Hodi saw in these patient biopsies was that after treatment with checkpoint inhibitors, the patient's immune systems seem to be attacking the newly formed blood vessels, the ones that tumors had created to feed themselves. Once again: After these metastatic melanoma patients were treated with a checkpoint inhibitor, their jazzed up immune systems didn't just go after the tumor cells. They went after the blood vessels the tumor is built to hijack the body's blood flow. And there was more: As the immune system took aim at the blood vessels that fed the tumor, it was sparing the blood vessels that fed healthy tissue. To Hodi, it looked like the immune system could tell the difference.

F. Stephen Hodi, Jr., MD 14:58
It was a little bit of a surprise. It's also just intriguing that the immune system was able to distinguish the blood vessels specific to the tumor. And that maybe those blood vessels were biologically different. They express different proteins, they function different than normal blood vessels.

Ken Shulman 15:15
Hodi wasn't sure how the immune system could make that distinction. But he believed it had something to do with timing and signaling.

Ken Shulman 15:26
When you cut yourself, a number of things need to happen so the skin can repair itself. First, you get hemostasis. The body stops the flow of blood to the wound so the healing can begin. As part of that healing, the immune system swoops in to sweep away dead skin cells, and to get rid of bacteria and other infectious agents. The next step in healing is angiogenesis, laying down the blood vessels that will supply the new skin cells. But before our bodies can take that step, the inflammation — the immune response — needs to subside.

F. Stephen Hodi, Jr., MD 16:03
So tumors take advantage of that, because they need new blood vessels to grow. And they want to avoid immune recognition to grow. And so the interplay between angiogenesis factors,blood vessel-forming factors in the immune system, seems to kind-of be a two-way street, and their kind of crosstalk with one another.

Ken Shulman 16:25
So the tumors seem to wait for the right moment to lay down their new blood vessels right after the immune system finishes its clean-up. Any earlier and the blood vessels will be attacked. Any later and new skin cells and other blood vessels have already begun to take shape. It's all about timing. When the tumor says the time is right, they secrete something called VEGF — that's vascular endothelial growth factor. VEGF is a signal protein that kicks off angiogenesis. And part of that kickoff, part of ensuring that the tumor has an adequate supply of blood, involves protecting these new blood vessels from immune attack.

Ken Shulman 17:11
From previous research, and from his own observations, Hodi suspected that VEGF also did that. That the growth factor was also an immunosuppressant, that it deflects or diffuses the immune response, so the blood vessels can take root.

F. Stephen Hodi, Jr., MD 17:25
We found that patients who had higher levels of VEGF in their blood before receiving CTLA4 blockade didn't do as well as patients who had lower levels, and so there was a lot of initial pieces being put together — guilt by association — but as you put the pieces of the puzzle together, kind of indicate that maybe this would benefit from further investigation.

Ken Shulman 17:46
VEGF, the blood vessel growth factor, seem to be muting the immune response, even after patients received the checkpoint inhibitors, in this case CTLA4 therapy. Hodi wondered whether that immune response might be more robust if VEGF was also targeted. So he led a clinical trial with combination therapy, a one-two punch with a checkpoint inhibitor that would energize a patient's T cells, and with an anti-angiogenic drug that would keep the tumor from laying down new blood vessels. That anti-angiogenic drug would also keep VEGF from dampening the immune response.

Ken Shulman 18:26
It was a very compelling hypothesis, and the trial results showed it to be spot on. Hodi saw that after some patients received the checkpoint inhibitor — this was the CTLA4 checkpoint — their immune systems didn't just rev up their T cells. They also produced antibodies, antibodies that shut down VEGF and other angiogenic growth factors. Again, as a secondary effect of the checkpoint inhibitor, the drug that takes the brakes off the T cells, other immune cells were creating antibodies to block the formation of new blood vessels. In essence, the checkpoint inhibitor was also anti-angiogenic.

Ken Shulman 19:07
There were interesting crossover results from the VEGF inhibitor as well, from the drug that shut down the blood vessel growth factor. Hodi saw that when VEGF was targeted, the blood vessels around patient tumors seemed to change a bit.

F. Stephen Hodi, Jr., MD 19:22
They're more permissive of immune cells to cross from the blood into the tumor microenvironment. Some of the biochemistry changes their morphology, their shape changes and allows the immune cells to cross into the tumor microenvironment better, which might make immune therapy better because immune cells are gonna be right in the tumor microenvironment.

Ken Shulman 19:38
Let's repeat that one. Targeting VEGF didn't just help cut off the tumor blood supply. It helped T cells and their immune partners do a better job. It opened the door for tumor-fighting cells to infiltrate the tissue surrounding the tumor. Now that's a winning combination: two drugs with two different targets in two different mechanisms, teaming up to fight cancer on two different fronts, and giving each other a helping hand.

F. Stephen Hodi, Jr., MD 20:07
These combinations of checkpoint blockade and the angiogenesis inhibitors have been tested in a variety of cancers even beyond melanoma, with some really interesting and important change of practice for kidney cancer, endometrial cancer, liver cancer; I've been trying lung cancer. So these kind of concepts of have broadened generally to the possibility of synergistic effects between blocking blood vessel chemicals and taking the brakes off the immune system.

Ken Shulman 20:37
Hodi's trial and its encouraging results emerge from extensive tumor analysis and bloodwork. He started with what he saw, with physical evidence that then led to a hypothesis he could test in patients. It wasn't just a case of taking two drugs that work on their own and seeing if they might work better in tandem.

F. Stephen Hodi, Jr., MD 20:57
We kind-of very early on had a suggestion of what the possible mechanisms are in that blood vessel formation and all the biochemistry involved with that actually were important for the immune system and that the two cross talk with one another, and that you could find ways to maybe put them together to help patients.

Ken Shulman 21:20
And Hodi did find a way to help patients: boosting the immune response with one drug and blocking angiogenesis with a second. And it wasn't just one in one making two; the two drugs together also changed the tumor microenvironment. Together, they made it more conducive to an immune response. It was a very impressive result. But Hodi knows that the better part of the combination immunotherapy story has yet to be written.

F. Stephen Hodi, Jr., MD 21:51
We're learning different aspects of the immune system, the role of other immune cells besides T cells in terms of the fight against cancer; we're learning how all those immune cells interact with one another, how antibodies are involved in fighting cancer. And I think [we've] still made great strides — are still, I think, fairly early in our understanding of how to apply these principles to improve therapies.

Ken Shulman 22:25
So, combination immunotherapy: it's complicated. And here's a further complication. In addition to finding out which therapies might complement each other, which drugs or treatments can work in tandem, doctors also need to figure out the optimal order of those treatments. Because these treatments change the landscape. Some chemotherapy suppress T cell activity; others give T cells an added boost. Either way, the reaction to chemo will change the way a tumor reacts to immunotherapy and vice versa. It's not just a question of doubling down.

Leena Gandhi, MD, PhD 23:03
There's often a very standard kind of way to approach clinical combinations where we just put everything together and see what happens.

Ken Shulman 23:10
That's Lena Gandhi. She directs the Center for Therapeutic Innovation at Dana-Farber.

Leena Gandhi, MD, PhD 23:12
Which I think is really hard and not the right way to be approaching combinations of immunotherapies. Not only do we need to understand what is the effect of one drug on the tumor immune environment versus both, but we need to understand whether combining them together can be inferior if we're doing it in the wrong sequence.

Ken Shulman 23:34
Gandhi has led pivotal studies in combination immunotherapy, both in industry and more recently at Dana-Farber, where she's worked since 2020. She's interested, among other things, in therapy sequence, in discovering which treatment should lead and which should follow and after how long, and which treatments should be administered together.

Ken Shulman 23:56
She cites one study in lung cancer. Patients were treated with radiation and after that, with PD1 checkpoint inhibitors. The variable was the interval between the two therapies. Some patients got the checkpoint inhibitor just two weeks after finishing their radiation treatments. Some waited longer between the two. The longest interval was six weeks. The hypothesis was that the longer wait would produce better results. Physicians had long observed that radiation seems to temporarily alter our body's immune response, that it seems to reduce inflammation in the treated area. But the hypothesis was wrong.

Leena Gandhi, MD, PhD 24:37
In fact, initially, we thought that bringing in PD1 inhibitors immediately after radiation may not be so optimal because we often see the upregulation of T regulatory cells or other kinds of T cells that suppress T effector cells or killer T cells from actually killing the tumor. In point of fact, though, when we — at least in lung cancer — when we did those studies combining radiation with PD1 inhibitors, we saw that actually bringing the PD1 inhibitor even earlier after radiation actually had a better long-term effect.

Ken Shulman 25:11
And why was the hypothesis wrong? Why did the patients who waited only two weeks after radiation do better than the patients who waited six weeks? Well, the answer is, we're not sure.

Leena Gandhi, MD, PhD 25:23
It just highlights again how little we know actually about how all these different inputs are working together. And that radiation may not be the same as chemotherapy plus radiation, and all those kinds of things, and that we really need, honestly, to understand what's happening at the tumor microenvironment level to better design studies and combinations and sequences that might really get at what that environment is telling us it needs.

Ken Shulman 25:48
What would it take them to understand what the tumor and everything around it are saying, to decipher the language spoken in the region that Gandhi likes to call the tumor immune microenvironment?

Leena Gandhi, MD, PhD 26:01
Really what we need to know are some three key things, which is, what are the tumor cells doing in the immune microenvironment? What are the immune cells doing, and which immune cells? and that means understanding the proteins that are being expressed by tumor cells versus immune cells, and by understanding where and which kind of immune cells are acting next to the tumor cells, because all of that we think actually matters in generating a tumor immune response.

Ken Shulman 26:27
And then there's the matter of toxicity. Doctors have been treating cancer patients with multiple chemotherapies for decades, with excellent results and mostly manageable side effects. It's different with combination immunotherapy.

Leena Gandhi, MD, PhD 26:42
The toxicities are actually not as predictable when you put those two different kinds of drugs together as when you put two chemotherapy drugs together. I think it's really important that we, as doctors and researchers, think very carefully about what kinds of drugs we're putting together.

Ken Shulman 27:04
Just as two therapies can complement each other, can act both as single agent and as a productive partner. Those same two therapies can generate adverse effects, side effects that are stronger than the ones they might produce on their own. In lung cancer, for example, certain targeted therapies can cause an inflammation called pneumonitis. Immunotherapies can cause that same inflammation; doctors need to be careful combining the two. But Gandhi points to other combinations where outcomes are improved, and side effects are manageable.

Leena Gandhi, MD, PhD 27:40
One of the reasons that chemotherapies and combinations with immunotherapy has really taken off in many cancers is because we might have expected to see a lot worse side effects, like much more inflammation and therefore worse effects from the immunotherapy, and maybe worse effects from the chemotherapy. In fact, we didn't see that. What we saw is mostly you see the effects of the chemotherapy and the predictable side effects. And you can see predictable side effects of immunotherapy. But they weren't unexpected and they weren't necessarily more.

Ken Shulman 28:08
And even in cases where doubling down on therapies has doubled down on side effects, doctors are finding ways to minimize those side effects. They do it with better timing.

Leena Gandhi, MD, PhD 28:19
And there's good examples actually in melanoma, where combining BRAF inhibitors, which can have certain skin and liver toxicities, with PD1 inhibitors and CTLA4 inhibitors, were very toxic when given together, but you could sequence them in a way that was less toxic. And it just again emphasizes we need to know understand more about what each drug is doing to figure out the best way of combining them.

Ken Shulman 28:43
It's hard to predict when combination immunotherapy will bat 500, when it will work in 50, 60, or even 90% of patients. The list of possible combinations, sequences, and outcomes is almost infinite. We'll need 1000s of additional clinical trials. We'll need the foresight to know which trials to run and the hindsight to know what those trials told us. We'll need new therapies in addition to new combinations of existing therapies. And we'll need better information intelligence, an immune atlas that pulls data about specific cancers, specific therapies, and perhaps most important of all, how those cancers change in response to those therapies.

Ken Shulman 29:39
All these efforts are underway at Dana-Farber and institutions around the world. It's a lot of work, a lot of information to gather and process, a lot more than in many other areas of research. In comparison, targeted therapy, finding a genetic mutation that causes cancer, then shutting that mutation down, is relatively straightforward.

Leena Gandhi, MD, PhD 30:03
What we did with targeted therapy was a very simple process. One mutation that's called an oncogenic driver can be the one thing you have to target for a good tumor cell killing effect.

Ken Shulman 30:16
Still, Gandhi believes combination immunotherapy will eventually work in a majority of patients. We have the capacity, we have the technologies. And we have a detailed view of cancer and its many incarnations that was unavailable even five years ago.

Leena Gandhi, MD, PhD 30:34
But we can apply that same data of really understanding all of the mutations that are involved, all of the cell types that are involved, and all of the proteins that are involved and where they are, and actually use that information to guide how we develop therapies. But it's eminently doable. I think we have all the tools, we just have to apply them now in a higher order way to a tumor immune response.

Season 2, Episode 6: Big Data, AI, and Cancer Research

As our universe of knowledge expands, sometimes incrementally and more often, exponentially, its new dimensions create new challenges. The more we know, the better our tools are, the more choices we have. And that's great, but it's hard to distill mountains of patient data, clinical trials, therapies, discoveries, and a boatload of externalities. Hard to distill all of these into the best possible research and care. Oh, and don't forget the human genome in its almost infinite complexity.

It can be overwhelming when you try to think about it absent, sort of having some help. But what kind of human mind could provide that help? Could it be that the help we need, the intelligence we need isn't human at all.

In this episode, we look at how artificial intelligence (AI) experts, physicians, and ethicists are working together to apply AI to cancer research.

Season 2, Episode 6 Transcript

Ken Shulman 00:06
It's been an interesting season here at Unraveled. We looked at CDK inhibitors, the drugs that gum up the gears in a cell's reproduction machine. And we learned how those drugs, which were originally developed to treat leukemia, now help millions of breast cancer patients around the world. We heard about targeted therapies with the story of the EGFR gene. In that one, doctors took aim at a genetic mutation found in some lung cancer patients. And when they did, they helped usher in the age of precision medicine. We studied a family of drugs — thalidomide and its derivatives — that doctors had used for decades to treat myeloma, even though they didn't understand how that drug worked. And then we listened as Dana-Farber, doctors and researchers retrace the steps they took to solve that mystery. Above all, Season Two has been the story of momentum, the story of an expanding universe of knowledge, strategies and care.

Ken Shulman 01:15
But as that universe expands, sometimes incrementally and more often, exponentially. Its new dimensions also create new challenges. The more we know, the better our tools are, the more choices we have. And that's great, but it's also hard to manage. It's hard to distill mountains of patient data, clinical trials, therapies, discoveries, and a boatload of externalities. Hard to distill all of these into the best possible research and care. Oh, and don't forget the human genome in its almost infinite complexity.

Eli Van Allen, MD 01:53
We've gone from looking at genes one gene at a time to now we're looking at, you know, cellular transcriptional readouts of single cells in tens, if not hundreds of thousands of them across tens to hundreds of patients tumors. I mean, the dimensionality is... it's complicated. And it just gets overwhelming when you try to think about it absent, sort of having some help.

Ken Shulman 02:24
Overwhelming, unless we get some help. But what kind of human mind could provide that help? Could it be that the help we need, the intelligence we need, isn't human at all.

Ken Shulman 02:42
I'm Ken Shulman, and this is Unraveled, a Dana-Farber Cancer Institute podcast.

Ken Shulman 02:56
Unless you've been living in a cave for the past 10 years, or in a sensory deprivation tank, you know something about artificial intelligence. Terms like neural networks, or machine learning may not live in your everyday vocabulary. But you've heard them, perhaps even used them, maybe more than once. And more than using those words, you've used those things. You've chatted with bots to help find your waylaid luggage; you've interacted and maybe even yelled at digital hosts about your cable bill. There are programs that keep your company or that suggest the TV series we think you'd like. There's even a bot that composes passable essays on command, a bot that recently typed its way into medical school. Sound scary? It does to me. But for many doctors and researchers, AI, the software and algorithms and bots that make up the thing we call artificial intelligence, sounds like the help they need.

Eli Van Allen, MD 03:55
There's no reason why we couldn't take those same kinds of advances and the same kinds of energy and steer it into, you know, the most pressing matters of cancer medicine.

Ken Shulman 04:06
That's Eli Van Allen. He's a Dana-Farber, doctor who specializes in computational cancer genomics.

Eli Van Allen, MD 04:12
It's a field that's growing whereby we're trying to take all of the data from genomics and oncology, basically, every patient's cancer genome, that we're sequencing at Dana-Farber and elsewhere and learn as much as we can from these data, and increasingly using sophisticated algorithms to find patterns, find these exquisitely complicated needles in exquisitely complicated haystacks and really try to push the field forward.

Ken Shulman 04:40
Okay, time for a spot of commentary. The Cancer Genome, that's the exquisitely complicated haystack, the teeming web of mutations, proteins, enzymes, chromatin remodelers, immune cells, and environmental factors that govern tumor growth. And exquisitely complicated needles? Those are the patterns, the irregularities, repetitions, and abnormalities, the things that machines are so much better at spotting than we are. And the things that let doctors ask the next questions.

Eli Van Allen, MD 05:17
Can we perturb these things? Can we drug these things? Can we somehow find the right patients to give these drugs to and everything in between?

Ken Shulman 05:26
Van Allen grew up as a self-described computer geek in Southern California. He went to college at Stanford planning to study computer science.

Eli Van Allen, MD 05:34
Sort of a funny thing happened, though, along the way. So, while I was studying this, and really sort of preparing myself for a career in this domain, taking computer science classes and learning about the field, some of my friends approached me to say, well, you know, we're interested in starting a nonprofit. It's a camp for kids whose parents had cancer. And I thought it sounded interesting. I thought I wanted to help. But I didn't really know what I was getting myself into.

Ken Shulman 05:59
That summer camp experience changed Van Allen's life. He decided to study medicine, cancer medicine, in particular. But he took his interest in computers and artificial intelligence with him to medical school, and to his residency. A lot of people scratch their heads.

Eli Van Allen, MD 06:17
There were people who would hear my interest and say, Okay, well, you know, we can send you to talk to the folks building electronic health records and sort of that field of informatics, which is very important, but not really something that I was particularly excited about on a personal level. And then there were the folks who said, well, certainly there's a lot of opportunity for translational research and sort of like understanding foundational biology using computers, but that's not done by people like you, Eli.

Ken Shulman 06:45
Van Allen came to Dana-Farber as a first-year fellow in June 2010. It was one week after the LA Lakers beat the Celtics in the NBA Finals. He remembers the faculty booing him in good fun when he introduced himself as a Laker fan. He also remembers running into a few headwinds. When he said he wanted to work in oncology and in AI, he wasn't quite sure how we might merge these twin passions. Then one of his mentors, Bob Mayer, introduced him to Levi Garraway. Garraway was the director of the Joint Center for Cancer Precision Medicine. It's a cutting-edge collaboration between the Broad Institute, Brigham and Women's Hospital, and Dana-Farber.

Eli Van Allen, MD 07:26
And he was really living and breathing what I was aiming for. He was at the forefront of precision cancer medicine, and using genomics to sort of make treatment decisions for cancer patients across the board. But when I met with him within about two minutes, I still remember thinking, man, I don't understand 80% of the words coming out of this guy's mouth, but this is cool. This is for me; I know this is going to be the thing.

Ken Shulman 07:53
And just like that, Van Allen joined the Garraway lab at Dana Farber. His new boss gave him a daunting assignment.

Eli Van Allen, MD 07:59
When I started in Levi's lab, he sat me down and he said, look, we're about to start generating massive amounts of genomic data for each of our cancer patients. And I need you to come up with a way to find the needles in the haystack. Which of these mutations in each patient's genome is something we can act upon. Clinically, we can give a drug for that's actionable.

Ken Shulman 08:21
This was new territory, not just for Van Allen, but for the field. The information was there, but it was a single kernel, a single needle, buried in bail upon bale of data. The Garraway lab rose to the challenge, they built a tool the first of its kind, a tool, Van Allen calls an interpretation algorithm for cancer genomics to use at the point of care,

Eli Van Allen, MD 08:47
Which is a long-winded way of saying, here's a tool we built to help doctors use this complicated genomic information to find the right drugs for the right patients.

Ken Shulman 09:00
That algorithm was able to spot patterns camouflaged among ungodly gobs of patient data. And in those patterns was critical information about genetic mutations that were driving cancer.

Eli Van Allen, MD 09:12
One of those instances was actually in a patient with something called anaplastic thyroid cancer. This is a devastating disease for which the response rates to any therapeutic are basically nil. There really hasn't historically been a therapeutic strategy. And we had this one case of this one patient who had this profound response, and a very unexpected one, to a targeted therapy.

Ken Shulman 09:42
The patient he's speaking about had responded to a drug called everolimus. It's an mTOR inhibitor, a small molecule that binds to the mTOR receptor on the surface of a tumor cell. That bond when it works, blocks the cell cycle it keeps tumor cells from replicating. Everolimus is really effective in anaplastic thyroid cancer. Van Allen and his colleagues were stumped as to why it worked in this patient, but the algorithm singled in on a genetic mutation that made this patient's cancer vulnerable to the mTOR inhibitor. And it didn't take weeks or days or even hours to find.

Eli Van Allen, MD 10:19
The old-fashioned way to do this would have been to actually have to spend just extraordinary amounts of time trying to fish through this massive haystack. But with the algorithm we made, it took it about 10 seconds. It basically within 10 seconds, like, that's it, that explains it. And in fact, now we need to then bring this back to the clinician, and make sure we find the other patients like this patient, and get these patients on these drugs.

Ken Shulman 10:45
Ten seconds to scan that spiraling ladder of sugars and phosphates and bases that compose the human genome. Ten seconds to scan the 3.2 billion base pairs, those are the rungs on that spiraling ladder, the ones that determine whether we'll have blue eyes or brown. Ten seconds to do that, and cross reference those 3.2 billion base pairs with 1000s of possible mutations, mutations that might be driving cancer. That is one heck of a haystack,

Eli Van Allen, MD 11:14
What we would typically have to do is manually go through all of them one by one, and try to figure out which things are important. In fact, that was sort of the standard of care of the time, people were having molecular tumor boards where they would take these data and put it basically into a spreadsheet, project a spreadsheet into a room and a bunch of people would stare at it, hoping that if you were lucky, somebody in the room recognized, oh, that thing and then would act upon it. I'm obviously being a little loose on it, but that's basically what was happening.

Ken Shulman 11:45
The algorithm Van Allen and his lab mates built took a problem that in the old days might have taken days or weeks to solve and figured it out in the time it takes an Olympic sprinter to run 100 meters. But that's not all it did.

Eli Van Allen, MD 11:58
It would do 99% of the heavy lifting for you and find the actionable alterations, find which ones of them might be more important than others, find which one of them might actually have other relationships to other events in the genome. And then also, critically, present this information in a way that a treating provider could actually understand.

Ken Shulman 12:20
The heavy lifting, sifting through those 3.2 billion base pairs, locating mutations and linking those to cancer, that may be the most impressive part of the operation. But Van Allen Elon says the last part, making the findings accessible to doctors in the clinic, may be the most important.

Eli Van Allen, MD 12:39
Because we failed that last step, we could have the most sophisticated algorithms in the world but it won't matter because the person making the decision won't understand what they're looking at. And will just, you know, be inclined not to use it, or even worse, make the wrong decision.

Ken Shulman 13:00
Back in 2010, and even a few years after that, people like Eli Van Allen, doctors who wanted to build a bridge between machine learning and medicine, sometimes had a tough time finding a place to land. It's not like that today. Lots of labs at Dana-Farber and elsewhere have data scientists and AI specialists working side by side with geneticists, biologists, and physicians. But acceptance doesn't always mean assimilation, there were still a times failure to communicate.

Jackson Nyman 13:39
Half of us were just pumping out data. And then the other half of us were getting set data and doing analysis on it. And so it was a lot of like, people that spoke one language but not the other, so to speak.

Ken Shulman 13:51
That's Jackson Nyman, a systems biologist from Harvard, who works in the Van Allen lab. He just defended his doctoral thesis in March.

Jackson Nyman 13:59
I personally wanted to exist in that space, where I could kind of speak both languages and try to make the best out of both of them or start to like, facilitate some of that interaction.

Ken Shulman 14:13
Nyman started his PhD in 2017, with Van Allen as his thesis advisor. The project? To see if it might be possible to use imaging and machine learning to predict whether a specific cancer patient would respond to a specific therapy. The specific therapy here was immunotherapy, the PD-1 blockade, and the specific cancer was kidney cancer.

Jackson Nyman 14:36
This is a cancer where it's actually really common to use immunotherapy even in the kind of first sign or initial treatment setting at this point. So, there's a big need to try to understand which patients respond and why or rather which don't respond well.

Ken Shulman 14:53
We've already heard about using AI to sift through genetic sequences through the haystack of 3.2billion base pairs. What Nyman was trying to do here was different. Pathologists often use imaging -- pictures of tumors and tumor sections -- to diagnose individual cancers and prescribe treatment, Nyman wanted to use AI to scan those and pinpoint physical features that might be correlated with drug response or drug resistance. Essentially, he wanted to train the machine to home in on details in a picture that could tell doctors whether immunotherapy would work or not work. Now, this wasn't just a handful of images. There were hundreds of thousands of them at very high resolution. It's hard to even guess how long a team of pathologists might take to leaf through and analyze that volume of pixels.

Jackson Nyman 15:49
But if we just run it through some of our AI models that can recapitulate some of these same things, all of a sudden, we have this really detailed map for hundreds or thousands of patients that is just directly ready for downstream analysis, I can start to see like: Are there properties that are shared between patients from this lens? Are there certain combinations of features that might be important and things that might be kind of hard to see, if you didn't quantify it with a computer.

Ken Shulman 16:20
It's important to note that Nyman wasn't working in real time with real patients. The outcomes here were already known, the AI team was working backwards to see if there was a pattern or patterns that connected physical features of the tumor with drug response or drug resistance.

Jackson Nyman 16:37
Sometimes you have a patient tumor where the whole thing is basically consistent. It's one pattern versus the other. But then, other times, and this is something that's kind of, it's been described before, but hasn't been quantified, you get this co-occurrence. And so, for our particular work, this was actually really important because when we looked at cases that had this co-occurrence, it appears to correlate with different clinical outcomes actually.

Ken Shulman 17:12
When we talk about artificial intelligence, about training machines to spot microscopic aberrations in tumor cells, we're also talking about neural networks. That's the artificial intelligence, a network of digital circuitry, layers upon layers of it, that we can train to tell us if something's a cat, and even if the cat is alive or dead or both. But here's the creepy thing. Not only do these neural networks learn, they also correct themselves. If they make a bad call, they'll tweak their connections all on their own. It's this last part, this computer, heal thyself, that Nyman and many in AI find troubling, because these machines don't always stay on mission. Philosophers may debate whether these neural nets actually have a mind. But if they do, sometimes they seem to have a mind of their own.

Jackson Nyman 18:13
It might just be learning something else that's not actually related to a cancer, it might be something really mundane. These models that we're using, neural networks in general, are only as smart as what we give it. So in a lot of the cases, if there's a shortcut to be taken, they will take it because these models have way too much modeling power behind them. It's a double-edged sword. They're really good when they work.

Ken Shulman 18:40
Because they are really good when they work, neural nets and other forms of AI have found their way into almost every corner of our lives from spellcheck to smartphones to self-driving cars.

Gregory Abel, MD, MPH 18:55
And artificial intelligence is becoming used more and more in oncology. But there certainly are some ethical issues with its deployment.

Ken Shulman 19:02
That's Gregory Abel. He's a doctor at Dana-Farber and co-chair of the Ethics Advisory Committee there. That committee serves as a sounding board for Dana-Farber officers and physicians and staff, and for patients and their families. For anyone who needs support navigating the difficult decisions that cancer and cancer care so often raise. Among his many other duties, Abel works with Eli Van Alen and several other colleagues to sketch out an ethical framework for AI deployment at Dana-Farber. One of the first features, one of the first concerns in that sketch is equity. The concern that AI will mostly benefit wealthy or well-connected patients, or that AI will be used as a cost-effective substitute for face time with real physicians, as in I'm sorry, the doctor won't see you now. But here's my friend, Robo Doc.

Gregory Abel, MD, MPH 19:55
Like many innovations in oncology, they tend to appear first for people who are well resourced. So if we find really good AI, and it's being used in oncology, it may take a while to get to patients that are more disadvantaged. On the other hand, if we start to use AI, instead of clinical encounters and our clinical and judgment, we may find that the most resource patients are the ones that still have that patient-physician relationship, where there's less AI in the room. So it's sort of a double-edged sword.

Ken Shulman 20:29
Equity in AI doesn't just mean an equitable distribution of care. There's an equity problem in datasets as well. These datasets don't always represent all groups. There can be gaps, data from certain ethnicities, genders or social classes might be absent. And if there are gaps, a patient from one of those underrepresented groups may not get the most precise diagnosis or prognosis, even if the AI follows every rule and protocol. It's called data absenteeism, or in some circles, data chauvinism. AI didn't create this problem, but it can exacerbate it.

Gregory Abel, MD, MPH 21:14
Another piece of my work is to try and make sure that we have people of color in trials, that we make sure that we address the data absenteeism, even data chauvinism, that we see in datasets. Sometimes we'll say, Well, if it's a mutation, it doesn't matter. If it's in a person of color, or white person, that mutation will act in this way. But there are so many other factors that could also affect the environment of that same mutation. To me, you need to have people of all groups hopefully in your dataset that you are basing your AI on.

Ken Shulman 21:54
Deploying AI in oncology also heats up the tug of war between transparency, that free flow of information that brings oxygen to science, and patient privacy. Patients willingly share their personal information and medical histories with their own treatment team. But they may not want to share that same information with the team on another floor, let alone clinical trial partners halfway around the world,

Gregory Abel, MD, MPH 22:20
Transparency can be limited by privacy concerns, because the more transparent we are, than we may be revealing what patients are actually, you know, in the database that the AI was based on. And that's not okay. So, with some rare cancers, just knowing the rare cancer, the age, and maybe the sex of a person and where they live, you could probably figure out who it is. And so that's a really important point. On the other hand, I do think that patients want to help, you know, there's always been altruism in medicine. It's why a lot of patients with advanced disease participate in trials. And so, if they by having their data in a database that AI can use, I think a lot of patients want that and are interested in that.

Ken Shulman 23:03
Transparency in AI and in cancer care also means a free flow of information between doctor and patient, to give patients the information they need to make informed choices about their care, to stay in the driver's seat. With AI, that isn't always easy. Because AI can be opaque. Even data scientists sometimes can't explain how this mechanical brain, how this black box reaches its conclusions. Abel says doctors just need to be honest,

Gregory Abel, MD, MPH 23:35
We have to help patients to understand the usefulness of this, the inexplicability of some of it, how even the best AI models sometimes fail. Every person is unique. Even if the AI says something that's going to happen, it may not happen. And that's where I think we have a real role to play as oncologists in terms of really packaging the AI and explaining it, even if we can't understand exactly how it's working, some of the limitations, and also some of the wonderful things that can result from it.

Ken Shulman 24:09
And there are many wonderful things that can come with it. Some that we can imagine and some that we can't. Abel even thinks doctors might one day use AI to help make ethical decisions in deciding for example, how to parcel out a drug, a leukemia drug, for example, when that drug is in short supply. Do you give it to the sickest patients? To the youngest patients?

Gregory Abel, MD, MPH 24:32
And I could see how that might be something that artificial intelligence could do. Or we could actually turn it back on itself and say, Okay, here's two ethical approaches to something. Let's use an AI model to help us figure out, using both approaches which we feel are both valid, actually ends up saving more lives. And then you'd say okay, we're not sure which is the best approach ethically they're both defensible, but this one actually the data shows through this AI model might actually be a little bit better in terms of saving lives. So we do that one.

Ken Shulman 25:19
Artificial intelligence, machine learning, big data, neural nets. These are all components of a system that has changed our lives, and will undoubtedly change them even more in years to come. They'll change oncology as well, just as profoundly. Doctors like Gregory Abel and Eli Van Allen, want to remind us that in the beginning, and in the end, these things are tools, new tools with capabilities and potential, we're just beginning to solve. And like all tools, they work best and do the least harm in the hands of thoughtful humans. Thoughtful humans, Van Allen says, from many different disciplines,

Eli Van Allen, MD 25:58
Certainly our lab and really there's a few other labs in our environment that are excited about this. But it's not just us. It's actually a larger initiative and a larger, larger effort across the entire campus. Having ethicists involved, having implementation scientists involved, having health services researchers involved, but bringing them into some of these new spaces, sort of as part of a larger community is actually something we're now doing across the board.

Ken Shulman 26:36
This is our final episode of Unraveled Season Two. We've told the story of research that drives clinical trials, clinical trials that test and refine new therapies, and new therapies are just as often existing therapies used in new ways that help patients live longer and healthier lives. It's a story of momentum, a Mobius loop of science and clinic and care, now animated by a new brand of intelligence that promises to step up that momentum. I can't wait to see how all that works out. But until then, I'm Ken Shulman, and this is Unraveled, a Dana-Farber Cancer Institute podcast.

Season 1

Season 1 Trailer

Season 1, Episode 1: The Wolf in Sheep's Clothing

The parable of the wolf in sheep's clothing reminds us that things aren't always as they seem: bad guys can dress up as good guys to do their bad guy things. There's a wolf in sheep's clothing story in cancer research, complete with a crafty predator, a clever disguise, and a visionary team of researchers who found a way to nab the wolf.

Season 1, Episode 1 Transcript

Ken Shulman 00:06
I bet you've heard the parable of the wolf in sheep's clothing. The sharp-toothed predator who slips on the sheepskin, sneaks past the shepherd, and helps himself to a tasty lamb lunch. The tale reminds us that things aren't always as they seem, that bad guys can dress up as good guys to do their bad guy things. Well, there's a wolf in sheep's clothing story in cancer research with a crafty predator, a clever disguise, and a visionary team of researchers that finally found a way to nab the wolf. I'm Ken Shulman and this is Unraveled, a Dana-Farber Cancer Institute podcast. The human immune system is an amazing thing, an intricate choreography of signals and cells that defends us from so many threats, viruses, bacteria, and parasites, for starters,

Gordon Freeman, PhD 01:11
And the thinking for over 100 years it'd be if you could just get the immune system to attack the cancer. That would be wonderful. And so lots of ways were tried. But none of those worked.

Ken Shulman 01:24
Immunologist Gordon Freeman came to Dana-Farber in the early 1980s. Born in Fort Worth, Texas, he travelled north to do undergraduate and graduate work at Harvard. Among other things, he studied T cells, an agile and aggressive strike force, the frontline defenders in our immune system,

Gordon Freeman, PhD 01:42
They're the major orchestrator of the immune fight. And over the last 30 years, we've studied the signals that expand the numbers of T cells, and discovered they're also signals that reduce the number of T cells,

Ken Shulman 02:00
Signals that expand the number of T cells, that's easy to understand. Say you've got an infection, your immune system calls up the troops for a rapid response. There's strength in numbers, but then there are signals that reduce the number of T cells. What's that about? Well, there is a point in downsizing the strike force, and it's a good one. Sure, we want to ramp up our T cells when we're under attack. But we also want those T cells to stand down once the fight’s over. Otherwise, they could run wild and start attacking healthy tissue and cells, they could stage a coup. Fortunately, our immune system also has several layers of stand-down signals, molecular checkpoints that tell our T cells to disarm. In most cases, the system works, but not in all cases. In his lab, Freeman found one reason why. And here's where the wolf story comes in. Sometimes, those stand-down signals, the checkpoints that tell T cells they can drop their weapons, get crossed, stolen, actually stolen by cancer cells. Freeman discovered that certain cancer cells steal that stand-down signal and use it to slip past T cells so they can spread disease. Cancer cells had learned how to slip on sheepskin and fool our defenses, just like the wolf learned to fool the shepherd. “What if”, Freeman started asking, “what if we could get those T cells to see the cancer cells as wolves again?”

Barry Nelson 03:44
No one wants to deal with cancer. No one wants to hear that word to be directed towards their life.

Ken Shulman 03:50
In 2011, Barry Nelson was diagnosed with stage three lung cancer. Nelson has a family history of cancer. The disease took his mother, aunt and grandmother, when he was diagnosed, he prayed to God that he wouldn't die in vain.

Barry Nelson 04:07
I felt like you know, God heard my prayer. Because I wanted, even if I was going to die, I just wanted what I was going through to help somebody else. So I felt like, look at this. It's going to help somebody else. You know,

Ken Shulman 04:19
Nelson did multiple rounds of chemotherapy and radiation with devastating side effects. Nothing worked. His doctors in Boston told him to put his affairs in order.

Barry Nelson 04:30
You know, the senior oncologist comes into the room. He said, this is a waste of time. You're going to die. You need to accept that you're going to die.

Ken Shulman 04:38
But Barry Nelson didn't want to accept it. He talked to his primary care physician who made an appointment for him at Dana-Farber with Dr. Christopher Lathan.

Barry Nelson 04:48
He listened to everything I had to say. And it was a wonderful feeling because after I explained to him what happened, he said, first of all, that'll never happen to you here. Secondly, we have plenty of tools in our toolbox. We're gonna fight for you just as hard as you fight for you. And that's what I wanted to hear.

Ken Shulman 05:05
And it turned out that Dr. Lathan had to try almost every tool in that toolbox, including another drug that when partnered with chemo had extended the lives of some lung cancer patients. But that combination just made Nelson feel worse,

Barry Nelson 05:21
Terrible, terrible side effects for me, I lost my equilibrium for like six months, I had no balance. I couldn't hardly walk or stand up, broke out in all kinds of blisters and all kinds of ... it was nasty. It was ugly.

Ken Shulman 05:34
It took a while for Nelson to get his strength back. When he did. Dr. Lathan told him there was another tool if he was interested. It was a new drug and immunotherapy that was just starting clinical trials at Dana-Farber. The drug was called nivolumab. If you know a thing or two about science, you'll know that major discoveries don't really happen overnight. Every breakthrough is the result of hundreds of small steps taken over months, and years and decades. And even for the best of scientists, most of those steps are missteps. in cancer research, very few lab discoveries make it as far as clinical trials, even fewer go the distance and become drugs. A researcher has to kiss a lot of frogs to find a prince,

Barrett Rollins, MD, PhD 06:28
You got to be sufficiently motivated to take all those hits and to still come back year after year, day after day to do the research.

Ken Shulman 06:38
That's Barrett Rollins, Chief Scientific Officer Emeritus.

Barrett Rollins, MD, PhD 06:42
Yeah, we get seduced a lot of times into thinking that the only thing that matters are positive results that moves you down a pathway. But once you become a scientist, and once you start working on things, negative results, something that doesn't work is just as valuable as something positive because it closes off one of those other pathways that you'd otherwise waste time going down.

Ken Shulman 07:05
Rollins and Freeman both started as researchers at Dana-Farber in the early 80s. It was a different scientific world. sequencing the human genome was still decades away. immunotherapy, the idea that the body's own immune system could be transformed into a cancer fighting force loomed like a sunset, beautiful, beguiling and out of reach. In the meantime, doctors at Dana-Farber and elsewhere used chemical agents to poison tumors and patients, or radiation treatments to shrink them. Let's be clear, these weren't the Dark Ages, every year saw improved cancer therapies with fewer and lighter side effects and better outcomes. But those improvements were mostly incremental, then came Gleevec.

Barrett Rollins, MD, PhD 07:54
It's impossible to overstate the importance of the Gleevec discovery.

Ken Shulman 08:00
Gleevec, a revolutionary therapy for chronic myeloid leukemia or CML. CML is an aggressive blood cancer whose average survival rate used to be about five years. Gleevec changed that, more importantly, Gleevec changed the way we look at cancer. In one trial in 1998 98% of CML patients who received Gleevec were still in remission five years later, doctors hesitate to use the word cured in cancer. But moving from a death sentence to five years of living cancer free sure looked like a cure. It also looked like a miracle. But it wasn't just these results that made Gleevec a game changer. It was the method. And this was no overnight miracle. It started in the 1950s when scientists began to unravel the biological mechanisms that generate cancer.

Barrett Rollins, MD, PhD 08:56
I think everybody understands that the problem in cancer is that cells grow and they're not supposed to. I mean, it doesn't happen just because there's food and water around. It turns out that it's a process that's controlled by genes. In every cancer that we have looked at, we know that the reason that cancer has happened is because there's been some alteration, some damage, mutation to the genes that regulate cell growth.

Ken Shulman 09:21
All of our cells have these genes. They're basically a molecular switch that tells our cells to be fruitful and multiply. It's how we grow. Think of those switches as hitting the gas pedal. In 1960, researchers in Philadelphia discovered that several CML patients had an abnormally short 22nd chromosome. A decade later, another group saw that this chromosome featured another abnormality, a hybrid gene. Two genes that shouldn't be touching were fused together. It was this hybrid gene this abnormality that sent out signals for cancer cells to grow. and grow and grow. It was a genetic switch jammed in the on position. So now think of your car being stuck in drive with the gas pedal glued to the floor. Doctors treated CML the way they treated most cancers―with chemo. But in the 1990s, a few researchers started to think they could do better. They thought they could develop a therapy that would shut down that hybrid gene. To turn the growth switch back to off. One of these researchers was Dana-Farber's Tom Roberts. Another was Brian Drucker, an Oregon based scientist who trained at Dana-Farber. It was a shift in strategy, instead of blindly attacking the cancer cells, these scientists wanted to target the very reason that these cells were cancerous. inhibiting that protein might keep these cells from behaving like cancer cells. That would take the pedal off the metal, and hopefully, spare healthy cells and tissues.

Barrett Rollins, MD, PhD 10:59
It completely changed the way you look at that leukemia, you go from something that was eventually a clearly fatal and pretty miserable death to, you know, patients have been on this drug for, you know, over 10 years and are doing extremely well. And now we know that if they fail that we have another drug behind it. But for the science of cancer therapeutics, it's a game changer because it proved the concept. It was a proof of concept experiment that you could make a drug targeted against a mutation that would have major clinical impact.

Ken Shulman 11:31
For researchers at Dana-Farber. Gleevec was proof of another perhaps even more significant concept, that there were other game-changing therapies out there, new strategies to fight cancer waiting to be discovered.

Barrett Rollins, MD, PhD 11:44
The confidence it gave researchers, especially in our setting, is that basic science can solve problems. Basic science can tell you which way to go in order to make new therapies. And so we had this long history of doubling down year after year saying that cancer cells are so different from normal cells, that our immune system ought to be able to detect them and reject them. And there were a lot of people who had faith that the immune system would be able to do this. But we had to figure out why. And so Gordon Freeman was one of these people.

Historic Radio Announcer 12:21
Until two days ago, that sound had never been heard on this earth.

Gordon Freeman, PhD 12:25
When I was in high school, Sputnik had just been launched.

Ken Shulman 12:29
That's Gordon Freeman again, he was in school in Fort Worth, Texas in 1957, when the Soviet Union launched the world's first manmade satellite into space, Sputnik. That's Russian for “little Voyager”.

Gordon Freeman, PhD 12:42
So there was a tremendous American concern that the Russians were ahead of us. So Congress funded scientific research all the way down to the elementary and high school level, because we needed to get better at science.

Ken Shulman 12:59
Thanks to the space race, Freeman got hooked on science early, first in his high school lab, and then over the summer at a special program at the University of Texas. That was where he learned he could play in the big leagues. He left the Lone Star State for New England and Harvard, where he got his bachelor's and doctorate degrees. Thanks to his first postdoc at Dana-Farber, Freeman chose to focus on immunology and cancer,

Gordon Freeman, PhD 13:25
Because I thought that it was a field really ripe for discovery. And that it could be applied to so many diseases, that if you could unleash it on cancer, you could have a really novel fight.

Ken Shulman 13:45
When he started, Freeman didn't quite know how the body's immune system could be trained to seek and destroy tumors. He just knew there was a need, because standard treatments like chemotherapy weren't always getting the job done.

Gordon Freeman, PhD 13:58
You hit the cancer, the chemo wipes out 99.99%. But you got one cell left, which then starts to grow again, and the chemo doesn't work on that cell. So what's different about the immune system is it can learn and change and evolve and adapt as the cancer does. Once you stop the tumor from evading the immune response,

Ken Shulman 14:25
And immunotherapy, the use of the body's own defenses to fight cancer, offered another advantage over traditional cancer care. Fewer side effects like hair loss, nausea, fatigue, decreased ability to fight infection, there are more but you get the picture. And it usually isn't a pretty one,

Gordon Freeman, PhD 14:46
Because we're not trying to directly kill the tumor cells. We're trying to activate the immune system to have the immune system kill the tumor cells. Chemotherapy is basically trying to poison cells.

Ken Shulman 15:01
Freeman was looking for a better way to make our T cells smarter, stronger, more effective against tumors. But why do T cells need our help? Why can't they beat cancer on their own? Well, it turns out, they usually can,

Gordon Freeman, PhD 15:18
At the beginning of any disease, you just have a few T cells, maybe just 100 that could attack that specific disease. So the critical thing, when you recognize a disease is those small number of T cells start dividing really fast. And a week later, they've gone from 100 T cells to millions of T cells. Now millions of T cells are an effective fighting force.

Ken Shulman 15:47
Effective enough to defeat most cancers, especially at the onset.

Gordon Freeman, PhD 15:51
Early on in the beginning of cancer, you get one or two or 10 cancer cells. It turns out, the immune system eliminates most of those at the one or two or 10-cell stage. It's only occasionally that those 10 cells become 100 cells. By the time you have cancer, which has become a medical problem. The immune response has been trying to fight that cancer for 10 or 20 years.

Ken Shulman 16:20
And after those 10 or 20 years, the T cells sometimes give up. In the early 1990s, a research lab in Kyoto, Japan, identified a protein on the surface of T cells, a protein that appeared when the T cells died. They named that protein PD-1. PD for programmed death. Then the team started manipulating that surface protein PD-1, turning it on and off in mice. And they noticed when they turned it off, the mice developed autoimmune disorders. It turned out they chose in a bad name for their protein. PD-1 wasn't a T cell kill switch, it was a safety switch. And when the Kyoto team disabled it, the T cells kept going and going, attacking everything in sight. In 1999, Gordon Freeman decided to dig a little deeper.

Gordon Freeman, PhD 17:14
So first, we discovered the B7 molecules, the ones that expand the number of T cells,

Ken Shulman 17:21
The battle cry rallying the troops. Let's take a quick timeout for a crash course in the human immune system. When a threat like a virus or an infection hits our bodies, tiny messengers called antigen presenting cells deliver samples of that threat to our T cells. Think of Paul Revere galloping through the countryside crying, “The virus is coming, the virus is coming!” These antigen presenting cells, the messengers show up with two signals. The first signal tells the T cell who the intruder is. The second signal, the B7 molecules Freeman just mentioned, tell the T cells to gear up for an attack. Now, when those dangerous signalers, the B7 molecules meet up with T cells, they bind with receptors on the surface of the T cells. This molecular handshake switches the T cell into attack mode and tells it what arms it will need to fight the enemy. The handshake also tells the T cell to multiply... fast. But there's even more. Remember, the T cells need to be disarmed once they've done their job, or they'll stay in attack mode. So, when the B7 molecule and its receptor on T cells shake hands, the B7 molecule also tells the T cell to express PD-1 on its surface, PD-1, the safety switch. It's like installing a set of disc brakes on a Ferrari. Even the fastest car needs to stop, especially the fastest car. Gordon Freeman knew all about this molecular handshake. And he asked a basic science question. Were there other molecules similar to B7 molecules that might shake hands and cause T cells to stand down to slam on the brakes? It was a good time to ask that question. Scientist around the world had nearly completed mapping the human genome.

Gordon Freeman, PhD 19:26
And then we looked for other new molecules as the human genome began to be sequenced and all these 20,000 different genes in your body were identified. We asked what looked like the B7 molecules. And we discovered little snippets of information that looked like a B7 molecule and we worked it up and found the full molecule.

Ken Shulman 19:52
Again, basic science. You ask a question. That question leads to another question and If you're lucky, to another,

Gordon Freeman, PhD 20:02
So we isolate that structure and make test tube amount of it. And then we ask, what will it interact to? What will it bind to? What does it glue to

Ken Shulman 20:17
Freeman and his team called their new molecule 292. It looked like a B7 molecule. Now they wanted to know whether it would act like a B7 molecule, whether it would bind, shake hands, with the same things that B7 molecules bind to, like receptors on T cells. Perhaps the new molecule would bind with the PD-1 receptor. Perhaps that handshake would tell T cells to stand down. While questions like these still drive science, technology often provides the wheels. For this experiment, Freeman and his team did something called flow cytometry. It's a high-tech cellular scanner. Researchers suspend cells in a single drop of water, mix them with dye and pass them beneath a laser beam at 500 cells per second. The machinery is complex, but the concept is simple.

Gordon Freeman, PhD 21:11
You know how you go to the rock shop and shine UV light on a rock and it shines back, green or red. Same idea. That laser beam hits a cell for a second and that cell shines back green. And we say okay, it's there, it's bound.

Ken Shulman 21:30
The question was whether molecule 292 would bind with PD-1, the safety switch. So, they prepared two batches of molecule 292 mixed with T cells. In the first batch, the T cells expressed PD-1, in the second they didn't.

Gordon Freeman, PhD 21:46
We found it doesn't bind to the cell without PD-1, but it bound to the cell with PD-1

Ken Shulman 21:53
Freeman's team had discovered a pathway. A few key steps in the complex ballet that is the human immune system. In this pathway, molecule 292 binds with PD-1. It was this bond, this handshake, that told the T cells to disarm once the job was done. PD-1 is a safety switch. But it can't shut down a T cell on its own. It needs to bind to shake hands. In order to hit the brakes. Freeman's team gave molecule 292 a new name, PD-L1, the brakes on the Ferrari, a military checkpoint. The PD-1, PD-L1 checkpoint protects us from a host of autoimmune diseases. And in most cases, the checkpoint system works. But as Freeman studied his new molecule, he found a bug in the system.

Gordon Freeman, PhD 22:48
We found that the PD-L1 molecule was expressed on cancer cells, and that we'd already discovered that it turned off the immune response.

Ken Shulman 23:00
So cancer cells also had PD-L1.

Gordon Freeman, PhD 23:03
So seeing this molecule on cancer cells said this might be a way that cancer cells would turn off the immune response. PD-L1 binds to PD-1, and it causes PD-1 to signal into the T cell and turn off.

Ken Shulman 23:23
Cancer cells stealing signals, cancer cells flying PD-L1 flags, so our T cells can't see them for what they are. The wolf in sheep's clothing. In 2000, after about one year of research, Freeman and his colleagues published their findings in the Journal of Experimental medicine. Freeman and his lab mates had co-authored scores of papers. But this one felt different.

Gordon Freeman, PhD 23:49
We did know it was major, because it was an immune inhibitory molecule that was on cancer cells. So we knew this had, you know, more potential than most

Ken Shulman 24:03
Potential. That's an understatement. Freeman had discovered a molecule on cancer cells that disarmed T cells, our frontline defenders, he discovered a mechanism that cancer used to confuse the body's cancer fighters. But no matter how promising, it was still just a lab discovery.

Gordon Freeman, PhD 24:21
Scientists have a lot of ideas that look good in the test tube, or look good as ideas. And what you've got to do is see it works and it's safe. I think the real confirmation is it works in people.

Ken Shulman 24:44
If cancer science is indeed a race, it's definitely a relay race. Scientists like Gordon Freeman run the first leg, then hand the baton to translational specialists, who ask if that lab discovery might be used with cancer patients The baton then goes to drug companies who develop a therapy or drug. In the final lap, clinicians test that drug on patients. If no one drops the baton, if everything goes right, the FDA approved the drug at the finish line. The relay race that Freeman started produced a drug called nivolumab. It's a checkpoint inhibitor, an antibody produced in the lab that keeps T cells from shutting down. How? Nivolumab binds with the PD-1 receptor on T cells with the safety switch. Nivolumab, the antibody, literally forms a blockade around the PD-1 receptor. Try as it may, the PD-L1 protein on cancer cells can't bind with the PD-1 receptor. It can't signal the T cell to shut down. From behind the blockade, the T cell sees cancer as the threat it is and attacks it. Basically, this drug allows the T cells to do their jobs. It removes any distractions, any bad messaging. Nivolumab was now in clinical trials. The anchor leg in the relay, which brings us back to Barry Nelson, the stage three lung cancer patient who had been told that his time was up. Now his doctors at Dana-Farber asked if he wanted to participate in a clinical trial for nivolumab. They told him the drug had shown promise in patients with melanoma and Hodgkin's lymphoma. But they'd never tested it on lung cancer. To some patients. It might have sounded like a Hail Mary. To Nelson, it sounded like another chance. And it felt like one too, right off.

Barry Nelson 26:57
So, when I got the infusion, I knew something was working. I didn't. I didn't feel like I normally felt when I was getting the standard chemotherapy. So, I went home, I didn't have to go to bed immediately. I don't know I just felt something different. It didn't feel like I normally felt. I didn't feel sick. I didn't feel like... my energy was returning. So, the long and short of it is that after three months, I mean after three treatments. We did a scan my tumors had shrunk 25%. 25%! And so that was amazing. They hadn't seen anything like that before.

Ken Shulman 27:35
Nelson continued with the nivolumab trial, and his health continued to improve. Soon he was well enough to ride his bicycle to Dana-Farber for his treatments.

Barry Nelson 27:45
And while I was waiting, you know, people in the waiting room, they would be whispering. I didn't know what they were talking about. So I was in one of my research nurses, “Did you ride your bike today?” And I said, “Yeah.” she says, “Oh, you know, everybody's talking about you riding the bicycle.”

Ken Shulman 28:05
The nurse asked Nelson another question. Would he like to meet the scientists who developed the treatment that had saved his life? A few months later, Gordon Freeman came to one of Nelson's treatments. Freeman looked at his scans. He asked Nelson how he was feeling, whether there were any side effects.

Barry Nelson 28:25
For me, it was almost like meeting God. Because this is that was he was the one he and his team were the ones that, you know, doing something to change my life and other people's life. You know.

Gordon Freeman, PhD 28:39
It was a real thrill to meet Barry Nelson. Barry had really advanced lung cancer. He was told by his physician to make his peace with the world. And Barry is a fighter and a researcher. And he found a doctor at the Dana-Farber Cancer Institute, who really said we have clinical trials. We have new players in the game.

Ken Shulman 29:10
In 2017, two years after it received breakthrough status from the FDA nivolumab was approved for treatment for a host of other cancers, including bladder, kidney, and melanoma. For a story that began at Dana-Farber, with Gordon Freeman poking around with T cells, it sounds like a very happy ending. But for Freeman, it's only the beginning.

Gordon Freeman, PhD 29:33
The immunotherapies have really opened the door. Again, they've been approved for 21 different types of cancer. But an average response rate is 20 to 30%, which means that it's not working for 70% of the people who take it. Also, it works in a lot of cancer types, but not every cancer type. And so we want to get immunotherapy to work in those cancers.

Ken Shulman 30:05
More than anyone, Freeman understands that checkpoint inhibitors like the one he helped pioneer aren't the cure for cancer. They're part of the cure. But they're an important part. A new way to wage the war, teaching our frontline defenders not to fall for one of cancer’s dirtiest tricks.

Gordon Freeman, PhD 30:24
I think checkpoints are the foundation because you've got to stop the tumor from evading the immune response. I'm not sure every scientist would agree with me. And I'm sure there will be a number of new therapies which use completely something else. But many, many of the new combinations are going to be the PD-L1, PD-1 drugs, plus something else.

Ken Shulman 30:55
Gleevec. Nivolumab. Overnight sensations that were anything but overnight. Dana-Farber, Chief Scientific Officer Barrett Rollins wants us to keep that in mind.

Barrett Rollins, MD, PhD 31:06
You know, if there's anything negative at all to say about these two stories, it's that it will tend to give certain people the idea that, you know, we should be seeing these kinds of successes year after year after year, and you just can't. There has to be a certain amount of a certain sort of, you know, critical mass of research being done for the sake of research. If we don't support basic undirected research without thinking that, “Oh, yes, this project is going to lead to the next big blockbuster drug”, we're not going to get to the next big blockbuster drug.

Ken Shulman 31:40
That said, Rollins believes the culture of Dana-Farber ups the odds of hitting the next jackpot in cancer therapy. Next time, the Nobel Prize comes calling.

Bill Kaelin, MD 32:04
First of all, the real prize was participating in the discovery and seeing the result and understanding something that had never been understood before and just marveling at the beauty of the mechanism that nature had arrived upon.

Ken Shulman 32:20
I'm Ken Shulman and this is Unraveled, a Dana-Farber Cancer Institute podcast.

Season 1, Episode 2: Behind the Science of the 2019 Nobel Prize

It's hard to overstate how important oxygen is to life on earth. Almost every living thing on the planet needs oxygen to convert fuel into energy. We can't survive without it, not even for a handful of minutes. Fortunately, our bodies know this, and have developed several rapid response systems to keep us going when oxygen runs low.

A few cancers have found a bug in the system: a way to sound a false alarm and make the body think it's low on oxygen when it's not. These cancers hijack the body's response to feed hungry tumors. It's a complicated and fascinating process, and it earned one Dana-Farber doctor the biggest prize of all.

Editor's note: A few months after the release of this episode, David Livingston, MD, died at the age of 80. He was a mentor to many and a giant in the field of cancer research. Read more about his life and work.

Season 1, Episode 2 Transcript

Ken Shulman 00:06
It's hard to overstate just how important oxygen is to life on Earth. Almost every living thing on the planet needs it to convert fuel into energy. We can't survive without it, not even for a handful of minutes. Fortunately, our bodies know this. And they've developed several rapid response systems to keep us going when oxygen runs low. It's yet another miracle of biology. But this miracle isn't foolproof. A few cancers have found a bug in the system. They found a way to sound a false alarm to make the body think it's low on oxygen when it's not. And then to hijack the body's response to feed hungry tumors. It's a complicated process and a fascinating one, and it earned one Dana-Farber doctor the biggest prize of all.

Nobel Assembly Emcee 00:56
On behalf of the Nobel Assembly at Karolinska Institute, it is my great privilege to convey to you our warmest congratulations. I now ask you to step forward to receive your Nobel Prize from the hands of His Majesty the King.

Ken Shulman 01:33
I'm Ken Shulman and this is Unraveled a Dana-Farber Cancer Institute podcast. On October 7, 2019, the Royal Swedish Academy of Sciences announced that Dana-Farber’s Dr. William Kaelin, was one of three physician researchers who'd won that year's Nobel Prize in Medicine.

Bill Kaelin, MD 02:03
I was asked the night before the phone call whether I was going to be able to sleep but I said I'd be able to sleep because I think the chance of winning is no better than 1%. But because it's 1%, I should probably leave my ringer on my cell phone. And indeed, my cell phone did ring at 4:40 a.m. to tell me I'd won the prize.

Ken Shulman 02:22
Now, it's always great to dream. If you're a kid who loves soccer, you dream about winning the World Cup. If you're a kid who loves music, you dream about sold out stadiums. And if you are a kid who loves science,

Bill Kaelin, MD 02:33
I had the great fortune of growing up in the 60s. And of course, that was the post Sputnik era. And I also had heard about the work of Jonas Salk, and I remember getting the polio vaccine. And so I grew up thinking about scientific heroes, and then heard about the Nobel Prize. And, you know, I knew it was out there. But of course, it was unimaginable that it would ever happen to me,

Ken Shulman 02:58
Bill Kaelin had lots of dreams as a kid and as a young man, not all of them came true. At least not at first,

Bill Kaelin, MD 03:05
I dreamt as a young boy of going to Harvard. And I was rejected from Harvard, I wound up going to Duke. I applied to Harvard Medical School, was again rejected, but I was fortunate enough to go to Duke medical school. I applied to Harvard for my internship and residency, was again rejected but had the good fortune of going to Johns Hopkins. And then I finally made it in 1987 as a medical oncology fellow and came to the Dana-Farber.

Ken Shulman 03:33
Today, Kaelin is the Sidney Farber professor of medicine at Harvard Medical School. But in 1987, he was a young postdoc at Dana-Farber. He needed 18 months of laboratory work, so he could be board certified in oncology. So, he started knocking on doors. One of those doors belonged to David Livingston, a physician scientist, who'd supervise some of Kaelin's clinical work at Dana-Farber. It was a very good fit. Livingston was impressed from the start.

David Livingston, MD 04:01
And he asked more questions then than I asked him. But what was amazing to me, was it this chap who'd never worked on the kind of problems that my colleagues and I are working on in the lab at that time, asked questions that were apt. And then he began to ask me what we were working on, that you know, might be of interest to someone, someone like him. I mean, this man is modest.

Ken Shulman 04:33
There was lots of work in the Livingston lab that interested Bill Kaelin, but one thing in particular, something called the RB gene, RB. That's short for retinal blastoma. And retinal blastoma is a rare eye cancer that mostly affects children. Physicians had already observed that in some families, this eye cancer was passed down from generation to generation. In the 1980s, researchers identified the gene involved in this transmission, they call it the RB gene. Several scientists had observed that when the RB gene was damaged, the eye cancer seemed to flourish. What Livingston and others wanted to know, was why, what was the link between the defective RB gene and this particular form of eye cancer?

David Livingston, MD 05:21
What Bill Kaelin came in wanting to work on was the structure of the retinoblastoma gene, with the idea that maybe from understanding the structure, and how the structure worked, it might be possible to gain some insight into how the retinoblastoma gene work. And that's exactly what he set out to do. And that's exactly what he accomplished and did it elegantly.

Ken Shulman 05:52
We need a break for clarification here. It's true that Kaelin and Livingston were interested in the RB gene. But the structure that Kaelin charted wasn't the gene. It was the protein that the gene produced. Because it was this protein, the RB protein that did the heavy lifting, it disposed of cells that might otherwise develop into cancer.

Bill Kaelin, MD 06:13
Scientists, as you know, are awful about their jargon. But genes are really just blueprints, typically for making proteins. And so, when scientists say they're studying genes, they really mean they're studying the proteins that are made with those instructions.

Ken Shulman 06:25
Kaelin looked hard at the protein that was made from the instructions in the RB gene. And he saw that the structure of that protein resembled the sort of pocket,

David Livingston, MD 06:34
He wanted to know how the pocket that he discovered, and the retinoblastoma gene works. So right away, I know we're talking about the motherlode here, because you're talking about a gene whose product stops cancer. So, the RB gene, which produces a protein that helps stop cancer, Kaelin discovered that a part of the structure of this RB protein, a part called the pocket, helps it bind with the protein that causes the childhood eye cancer. When this bond occurs, the cancer-causing protein is almost completely neutralized, disarmed with Livingston's help. Kaelin showed that it was the structure of the RB protein, the pocket, that helped it suppress tumor cells. It was one of the most powerful experiments I'd seen, designed and acted and controlled because Bill is famous for controlling his experiments, that is testing the validity of every result in a way that's impeccable, that it virtually took my breath away. He had in one experiment demonstrated how the retinoblastoma pocket actually works generically, and even in some detail.

Ken Shulman 07:57
That's a hell of a thing for a postdoctoral fellow to do. And it wasn't the only accomplishment he had made. He made a number of them in my lab in 1992. After five years with Livingston, Kaelin got his own lab. Now he needed a project, something timely, something compelling, something that matched his experience and could also break new ground. It didn't take him long to find it. In 1993, the National Cancer Institute isolated a gene called VHL. That's V H L for von Hippel Lindau. VHL was a tumor suppressor gene. Check. Kaelin had studied tumor suppressor genes with Livingston, and he was familiar with VHL. He was particularly interested in two clinical features that often appeared in patients who had tumors linked to VHL disease. These patients tended to produce new blood vessels, and they tended to produce additional red blood cells. Another check

Bill Kaelin, MD 09:00
We say that these people have so-called von Hippel Lindau disease. So, one thing is they developed several different types of tumors, one of which is kidney cancer, which is one of the 10 most common cancers in the developed world. And I thought, everything else being equal, why not work on a common cancer as opposed to an uncommon cancer, it seemed at the time, a lot of our progress related to cancers that were fascinating, but really, numerically quite rare. So, this seemed like an opportunity to really tackle one of the top 10 cancers, the assumption being that if we understood the VHL gene better, we would understand more what makes a kidney cancer tick.

Ken Shulman 09:34
So, a tumor suppressor gene, a disease he was familiar with, and a chance to help large numbers of patients, those three checkboxes would already have been enough, but there was more.

Bill Kaelin, MD 09:47
The tumors these patients develop are notoriously rich in blood vessels, we would say they're highly angiogenic and there was a lot of excitement in the early 90s trying to treat cancers by blocking their ability to obtain blood vessels, and so I thought by studying the VHL gene, we would learn something about the molecular control of angiogenesis

Ken Shulman 10:08
Angiogenesis, the production of blood vessels. For scientists in search of a problem, VHL seemed to have it all. The clincher, though, came when Kaelin recalled how some patients with VHL disease produce too many red blood cells.

Bill Kaelin, MD 10:22
And what angiogenesis and red blood cell production have in common is that they would normally be activated or induced if a tissue, for example, or your body wasn't getting enough oxygen. And so that was really the clue that if we could learn more about the VHL gene, we would also learn perhaps how the cells and tissues in your body sense and respond to changes in oxygen, which wasn't really understood at that time.

Ken Shulman 10:47
So, here's the mission, we know there's a link between the defective VHL gene and certain types of cancer. We also know that some of these cancers cause patients to produce additional blood vessels and additional red blood cells. What we don't know is how we get from A to Z. That's what Bill Kaelin wanted to discover. The story had a beginning, and it had an end. But the all-important middle was missing.

Bill Kaelin, MD 11:15
So, we set out to understand what the VHL protein actually did, what it did biochemically, how it did its work. And why when it was defective, you develop these tumors.

Ken Shulman 11:27
So where to start, you design experiments that can reveal, at least in part, how that protein works. So Kaelin and his colleagues cultivated cell lines derived from kidney cancers in a plastic laboratory dish, these cells carried a mutant or defective version of the VHL gene. When these cells were injected into laboratory mice, the mice developed tumors. Then, working with the same cell lines Kaelin injected the normal VHL gene. When these cells with the normal VHL gene were injected into laboratory mice, the mice did not develop tumors. Mice who got cells with a defective gene got cancer. Mice who got cells with the normal version of the gene didn't. This was visible proof that VHL did, in fact, suppress tumors.

Bill Kaelin, MD 12:20
But perhaps the more important discovery we made is once we now had kidney cancer cells that did or did not have an intact version of the VHL gene, we could grow them under low oxygen or high oxidant. And then we could ask them how they responded. And in particular, we can measure certain molecular distress signals that you and I would normally produce if we were starved of oxygen.

Ken Shulman 12:48
It sounds a little complicated, but it's really classic science. You've just proven in laboratory conditions that VHL suppresses tumors. Now you want to draw a connection between the effective form of that gene and the production of excess blood vessels and red blood cells. You know the connection exists, because you've seen it in patients. You just can't explain it. Not yet. So, what do you do? You take two groups of cells from the same cell line you used before. These cells are identical except for a single factor. One of them has the VHL gene. And the second one doesn't. You grow each group in a low oxygen environment, and then you grow them in a high oxygen environment.

Bill Kaelin, MD 13:34
And as we had guessed, the cells lacking the VHL gene, they couldn't sense oxygen, they behaved as though they weren't getting enough oxygen, 24/7, and they produce high levels of these distress signals that normal cells would only make if they weren't under low oxygen.

Ken Shulman 13:54
Okay, here's where we are now. And where Kaelin was; we know some patients with a defective VHL gene developed cancer. And we know that some of those VHL patients produce excess blood vessels and excess red blood cells. Those blood vessels and blood cells funnel additional oxygen to the tumor and fuel its growth. Now, Kaelin has discovered that cells lacking the VHL gene can't sense oxygen. They think they're starving, even when they're not. And because they think they're starving, they send out an SOS distress signals calling for more oxygen, even when there's plenty of it. So, the defective VHL gene hacks the cellular alarm system and tricks the body into feeding those tumors,

Bill Kaelin, MD 14:49
Certain cancers -- and kidney cancer would probably be at the top of the list -- hijack this oxygen-sensing mechanism for their own evil purposes and in particular, to trick the body into providing oxygen to the tumor. And so, amongst the things that the tumor does is it stimulates new blood vessels to be formed by the host to provide it with a blood supply.

Ken Shulman 15:12
By the late 1990s, Bill Kaelin had answered many of his initial questions - he'd gotten from A to K in the story he was writing. He'd observed that some VHL tumors trick the body into sending out an SOS distress signal that stimulated the growth of new blood vessels. What he still didn't know was how it all worked. There was still more of the story to write, and he wasn't writing it alone. Researchers at other institutions were also working in parallel on hypoxia in conditions of low oxygen. These researchers included Greg Semenza at Johns Hopkins and Sir Peter Ratcliffe at Oxford. Those parallel lines soon converged.

Bill Kaelin, MD 15:55
We got a big break in 1999, when Sir Peter Ratcliffe showed that cells lacking the VHL protein couldn't destroy a protein in the cell that scientists refer to as hypoxia inducible factor or HIF for short, and HIF is sort of a master regulator of those distress signals we talked about a moment ago.

Ken Shulman 16:17
Here's another character in the story. HIF, hypoxia inducible factor. It's a protein complex that governs much of the body's reaction to low oxygen. In many cases, most cases HIF is beneficial. HIF helps Alpine climbers ramp up their red blood cells at altitude. HIF helps stroke and heart attack victims grow new blood vessels to supply the damaged brain or heart. In many cases, most cases, HIF is only present when there isn't enough oxygen. So, what switches HIF on and off? How does HIF know when to call for reinforcements? With further experiments, Kaelin showed that when oxygen levels were normal, VHL gets rid of HIF. The functioning VHL protein binds directly to HIF, and it targets HIF for destruction. Let's break that down. When oxygen is present, VHL acts like an advanced scout for waste disposal. It surveys the neighborhood, find cells left on the curb and puts a tag on them to make sure they go straight to the scrap heap with HIF tagged and towed it away. There's no distress signal, no false alarm.

Bill Kaelin, MD 17:33
So that was incredibly exciting. And now we're in the year 2000. But it didn't answer the question: well, how does the VHL protein know, if you will, whether oxidant is or is not present, and hence whether it should or should not bind to HIF and mark it for destruction? And so that was the puzzle that both our laboratory and Sir Peter Ratcliffe's laboratory set out to solve.

Ken Shulman 17:54
How does it know? That was the question, the Nobel-Prize-winning question, or so it turned out? Neither Ratcliffe nor Kaelin could know it at the time. They were just doing what scientists do, pose a question, test it under laboratory conditions, examine the results, and pose new questions. The newest question: how the VHL protein, which helps to suppress tumors, knows whether or not to tag HIF. HIF, that's the command center for the body's response to hypoxia. In other words, how does the VHL protein decide that the supplemental oxygen system isn't needed and should be carted away? Both laboratories, one in Boston, the other in Oxford, set out to write the next chapter.

Bill Kaelin, MD 18:41
And what we both showed simultaneously, working independently, was in the presence of oxygen, a little chemical flag, which scientists would refer to as pro little hydroxylation, gets added to the HIF. And that serves as the signal for VHL to bind. And this turned out to be extremely satisfying because in this little flag, there actually is an oxygen atom. So even I could understand how this would be linked to oxygen availability. Now it turns out to be a little bit more nuanced than that as often is the case in biology. But to first approximation, this was a remarkably simple and elegant way for cells to know whether they were getting enough oxygen, because they actually use the oxygen to make the flag that then determines whether HIF will be destroyed or not.

Ken Shulman 19:28
It's almost a failsafe system. When oxygen levels are normal, the system uses that oxygen to print a chemical tail to pin on HIF. It's like a label you might print and stick on an old TV you want collected. VHL reads the label and takes HIF to the dump. But when oxygen is low or absent, this system can't print the label because it needs an oxygen atom to make the label. VHL leaves HIF on the curb and HIF sounds the hypoxia alarm, all good, except when it's a false alarm. And when the VHL protein can't sense oxygen, when it can't read the label, the false alarm goes off. Kaelin knew he'd found the missing piece to the puzzle. Now, he wondered whether Peter Ratcliffe and Oxford had also found it. The two had what Kaelin describes as a gentlemen's agreement. If both laboratories were close to a solution, they would publish their papers together.

Bill Kaelin, MD 20:36
And so when we understood the nature of the chemical flag, I called up Peter Ratcliffe, to see where they stood. And they described this as sort of like the dance of the seven veils, because I didn't want to give away the answer if he was nowhere close to the answer, in part because I didn't want to pre-empt him from having the joy of making the discovery himself. So, we sort of started to use another metaphor, flipping over cards until it was pretty clear. We both had the same answer, and then we agreed to co-publish.

Ken Shulman 21:09
Kaelin and Ratcliffe published their papers together in 2001. Subsequent experiments demonstrated that HIF was a major culprit in kidney cancer. So the next logical step seemed to be to work towards a drug that would inhibit HIF to stamp out the problem at the source. Unfortunately, at the time, drug manufacturers didn't think they could make a molecule that would bind with HIF. It didn't have the right nooks and crannies to make it an attractive target.

Bill Kaelin, MD 21:39
But fortunately, we knew that one of those genes regulated by HIF was vascular endothelial growth factor or VEGF, for short.

Ken Shulman 21:48
VEGF. It's a growth factor, and it's regulated by HIF. VEGF induces the growth of new blood vessels

Bill Kaelin, MD 21:56
And certainly, VEGF has become famous in terms of being produced by tumors, especially kidney tumors as a means of inducing new blood vessels. And so we and others work with companies to test VEGF inhibitors in various cancers, including especially kidney cancer, and we now have seven VEGF inhibitors approved for the treatment of kidney cancer, as we had hoped, and predicted.

Ken Shulman 22:23
Unlike HIF, VEGF turned out to be a very attractive target. And drug makers have become more agile since Kaelin first discovered the oxygen-sensing mechanisms in VHL and HIF. A few years ago, a small Texas biotech company came out with a molecule that can inhibit HIF. That drug has shown great promise in phase two and phase three trials for kidney cancer and von Hippel Lindau disease. In addition to drugs that inhibit HIF, Kaelin's research has also led to drugs that stimulate HIF. Those drugs can help patients recover from heart attack or stroke, or counteract the effects of anemia. It's a very impressive yield for some very impressive research. But it's not over. Kaelin thinks that one day it may be possible to go even further up the chain to correct VHL disease at its source, at the VHL gene. Scientists have already treated some diseases at the genetic level by altering a person's DNA to prevent or treat that disease. But it's a lot trickier with cancer, in order for the genetic treatment to work, close to 100% of the tumor cells have to take up the corrected gene.

Bill Kaelin, MD 23:37
Because if it's not 100%, you can be sure that the cells that didn't take up, the gene will quickly outgrow those cells that did take up the gene. So I think they'll have to be some further advances in gene therapy and gene delivery technologies. So we may get there, it certainly would be the more elegant solution just actually correct the genetic defect itself, rather than using drugs that sort of deal with the consequences of that defective gene. But we're not quite there yet.

Ken Shulman 24:15
On the night of October 9, 2019, the night before that year's Nobel Prize in Medicine was to be announced. Bill Kaelin says he thought he had a 1% chance of winning. He did, however, keep his cell phone ringer on.

Bill Kaelin, MD 24:29
When we understood the accident-sensing mechanism in the year 2008. You know, I was aware of the fact that this was the kind of discovery that might put you in the conversation for major prizes. But I think to my credit, I did a pretty good job ignoring that as best I could because you know, I always tell young people, the real prize was participating in the discovery and seeing the result and understanding something that had never been understood before and just marveling at the beauty of the mechanism that nature had arrived upon.

Ken Shulman 25:05
But despite his best efforts, he couldn't always ignore the call of the Nobel.

Bill Kaelin, MD 25:11
Over the years, we did win several of the so-called pre-Nobel prizes, which made it even harder to ignore. So even though I did try to ignore it, when October was looming every year, it was in the back of my mind.

Ken Shulman 25:23
Scientists like to joke about two conditions related to the Nobel Prize, pre-Nobelitis and post-Nobelitis. In pre-Nobelitis, you become obsessed with winning the Nobel Prize, your work and your life both suffer. In post-Nobelitis, the Nobel Prize winner becomes a celebrity and succumbs to the lore of fame and fortune. Again, your work and your life both suffer.

Bill Kaelin, MD 25:49
Unfortunately, there are people who you know when they win the Nobel Prize, it becomes sort of the brass ring and then they sort of jump off the rails and either go on the banquet circuit or, or become experts on things they’re not really experts in

Ken Shulman 26:04
Kaelin says he's done his best to ensure that the Nobel Prize has changed his life and his work for the better. There are still questions to be answered. Still patients in the waiting room.

Bill Kaelin, MD 26:15
There's too much work to do. And so, I allowed myself to enjoy it. There was a magical wonderful week in Stockholm, and I'm going to try to judiciously try to use my Nobel Prize to increase awareness in science and supportive science.

David Livingston, MD 26:35
Bill Kaelin wasn't the only Dana-Farber to leave his phone ringer on when he went to bed. The night before the Nobel Prize in Medicine was announced. David Livingston, Kaelin's Dana-Farber colleague and former mentor kept his phone on when he went to bed. And early the next morning, before the sun rose over Boston, I was not surprised to receive a call from Bill rather early in the morning to let me know that he won the Nobel Prize in Medicine. I was utterly thrilled. I was just taken with the notion that justice, scientific justice, had been done. A great scientist had won the greatest honor in all biological and medical science. I was utterly thrilled.

Ken Shulman 27:31
It's said, there's no greater joy for a teacher than to see a student succeed. And it's hard to imagine a greater level of success for a mentor, than to see a protege win the profession’s most important prize. In our next episode, we look at the culture of mentorship at Dana-Farber.

Julie Losman, MD 27:55
I have to say that the postdocs who work in my lab have generated really beautiful data. And I hope that I've conveyed to them that they themselves have produced their own Da Vinci's.

Ken Shulman 28:08
I'm Ken Shulman and this is Unraveled, a Dana-Farber Cancer Institute podcast.

Season 1, Episode 3: A Culture of Mentorship

What do cancer researchers and 15th century Florentine masters have in common? For one, a culture of mentorship. Both come of age in a culture where knowledge is transmitted from master to pupil. In this episode, we take a closer look at that culture and the crucial role mentorship plays at Dana-Farber, with Nobel laureate Dr. William Kaelin.

Season 1, Episode 3 Transcript

Julie Losman, MD 00:07
I do remember the very first time that I got the results of the experiment. I brought it to Bill and I showed it to him and I was very…I don't think we even were supposed to meet. I was so excited to show it to him that I sort of popped into his office. And he looks at that and he said, “Julie, this is your Da Vinci.”

Ken Shulman 00:26
That's Julie Losman, an assistant professor at Dana-Farber. The Bill she's talking about is Bill Kaelin, her former mentor at Dana-Farber and the winner of the 2019 Nobel Prize in Medicine, and Da Vinci. That's Leonardo da Vinci. Painter of the enigmatic portrait known today as the Mona Lisa. So, what do cancer researchers and 15th century Florentine masters have in common? Well, for one, a culture of mentorship, both came of age in a culture where knowledge is transmitted from Master to pupil. In this episode, we'll take a closer look at that culture, and at the crucial role mentorship plays at Dana- Farber. I'm Ken Shulman, and this is Unraveled, a Dana-Farber Cancer Institute podcast. If institutional memory could take human form, at Dana-Farber, it would look a lot like David Livingston. He's professor of genetics and medicine at Harvard Medical School. He's won many major awards during his nearly 50 years at Dana-Farber. He's also served as mentor to dozens of post-doctoral fellows, young scientists who've honed their craft in this lab, young scientists including Bill Kaelin. Livingston says, mentorship is a big part of the job.

David Livingston, MD 02:00
Which becomes for me and the vast majority of my colleagues, a dedicated pleasure of interacting as a mentor, for a young person who is working on difficult but very important problems. Then there is nothing like watching a young person grow, watch them sprout their mythical wings, and begin to think on their own, and begin to ask questions that simply take your breath away.

Ken Shulman 02:33
Livingston caught the science bug early on in life at 11 years old, when a physician cousin gave him a subscription to The New England Journal of Medicine.

David Livingston, MD 02:42
Well, I tell ya, whatever it was, I couldn't understand the words, naturally. I didn't know any anatomy even. But I could look at the pictures. When the New England Journal came out, there was a case report from the Massachusetts General Hospital that I would read. It was like a magnet. Hmm. I couldn't wait to read one. Even in hard, even when the weather was crummy, even when there was something fun to be done.

Ken Shulman 03:08
The pull of science and medicine, the magnet, Drew Livingston on to medical school, and then in 1971, to the National Cancer Institute, where he worked as a research fellow. At the time, the cancer research field was still largely a mystery. It wasn't exactly a blank slate, but it was pretty close.

David Livingston, MD 03:28
Almost everything we didn't know. We knew something about cancer cells, we knew that they grew in ways that normal cells didn't, that they had capacities that normal cells didn't, like to cause disease. And we knew that cancer cells came in various, I would say, they wore different clothing. No two cancer cells necessarily wore the same clothing. They didn't look the same, nor did they replicate at the same rates.

Ken Shulman 03:58
Scientists did know about DNA in the early 70s. But they hadn't come close to decoding it. They hadn't sequenced the human genome.

David Livingston, MD 04:06
And we had very little knowledge of how genes were built. Except we were pretty sure they were built with DNA. We didn't know much about the punctuation in genes, not nothing, but almost nothing. It was just being worked out in those days. And what made them cancer cells was a total mystery.

Ken Shulman 04:29
Livingston came to Dana-Farber in 1973, when it was still called the Children's Cancer Research Foundation. He immediately opened his own lab, and he began studying a virus found in both monkeys and humans. The virus is called SV40. SV for simian virus, and it's tiny with only a handful of genes.

David Livingston, MD 04:52
I knew that this virus when it was injected into a rodent, typically a hamster or a mouse, or a rat that lacked a normal immune system... could not grow. But what it could do was live long enough to make a tumor. And that was amazing.

Ken Shulman 05:13
The SV40 virus had another amazing thing to reveal. Researchers found that its cancer-causing agent was located in a single gene. And that single gene contained the code for a single protein. And when that single protein was added to cell cultures in a laboratory, it formed micro-tumors.

David Livingston, MD 05:33
That was almost a matter of addiction for those of us who entered the field, because it meant an enormous leap had taken place in cancer science that gave us an inkling, an insight, and tools that made it possible to study cancer for the first time, much more simply than ever before. Intoxicating discoveries don't come in big bushels. You know, this series of observations that others, many others observed in the world, and I was privileged enough to be one of them, was intoxicating to all of us.

Ken Shulman 06:10
That sense of wonder, the sense of intoxication grew as Livingston and the post-docs in his lab began to unravel the mystery that was cancer. He soon found a promising area to explore: tumor suppressor genes. These are genes that single out potentially cancerous cells and neutralize them. One day in 1987 Livingston got a call from Bill Kaelin. Kaelin told him he'd been working in another Dana-Farber lab, but that lab had shut down. Now he, Kaelin, was orphaned. He needed another lab where he could finish his postdoctoral fellowship. Livingston agreed to meet with him.

David Livingston, MD 06:53
I thought he was one of the tallest people I'd ever seen. I come from a family of short people. And this large, tall, very gracious, well-spoken young man came on and knocked and came in. And I thought to myself, “Boy, Americans are making them tall”. It's great. It's absolutely wonderful. And he was obviously extremely intelligent, quite gracious, and he asked great questions.

Ken Shulman 07:21
Kaelin was also taken with Livingston, with his infectious enthusiasm for science. And there were some interesting things going on in Livingston's lab. They’d just begun studying a tumor-suppressor gene, known as the RB gene. Children born with a defective copy of this gene develop a rare eye tumor called retinal blastoma. It's where the RB gene takes its name. Kaelin was eager to dive into a project with such a direct link to cancer, especially with Livingston as his mentor.

Bill Kaelin, MD 07:52
But what never changes in science is the ability to sort of see a good question, ask a good question, to have sort of the nose or the instincts to decide where you're going to work, what problems you're going to tackle. And so, I think the first job of a mentor is to kind of imprint upon you that sort of scientific taste and scientific intuition on you know, what might be a good thing to work on what might be a fertile area to start working on it. So, I certainly learned that from David, he has incredible scientific taste. David also taught me, you know, the art of experimental design, how to design really effective penetrating experiments that would get to the heart of a question and will hopefully give you a relatively unambiguous answer. So, you know, I say that, you know, thinking never goes out of style. I mean, David really helped me to think like a scientist.

Ken Shulman 08:43
Thinking like scientists. The pair found a project for Kaelin. Genes, as we know, are molecular blueprints, they contain the instructions for making proteins. The RB gene that Livingston's lab was looking into contains the instructions for making the RB protein. And it's that RB protein that actually suppresses tumors. What Kaelin set out to do was understand how the RB protein worked, what biochemical properties that this protein has, that enabled it to block cancer. It was an ambitious project. And Livingston says Kaelin absolutely nailed it.

David Livingston, MD 09:26
And even in some detail. That's a hell of a thing for a postdoctoral fellow to do. And it wasn't the only accomplishment he had made. He made a number of them in my lab, and others had too. So that was a happy moment. That was a good moment. That was a self-satisfying moment. And the entire lab was thrilled. And of course, Bill was thrilled. It was a great discovery.

Ken Shulman 09:52
Kaelin is quick to share much of the credit for this early triumph with his mentor. He says Livingston constantly challenged him to ask more daring questions, and to design even more bomb-proof experiments that could answer those questions beyond any doubt. But Livingston wasn't just a hard-driving coach.

Bill Kaelin, MD 10:12
So, David was just a fantastic cheerleader. psychotherapist, at times, I know myself, and I'm sure this is true for most young people doing science, there are days when it's all very frustrating. And you're wondering whether you're just a square peg going into a round hole. And David was very, very good at giving me perspective, encouraging me, keeping me moving forward. So, I'll always be very grateful for that. Again, in addition to asking great questions, David was a fantastic cheerleader, no matter what you discovered, David made you think you were going to Stockholm, which is sort of ironic that I eventually did go to Stockholm.

Ken Shulman 10:52
Livingston doesn't remember it quite that way.

David Livingston, MD 10:55
Well, in those days, the only thing I knew about Stockholm was that some of my relatives were born not far away. That was about it. I mean, I knew all a lot of the myth and, and I knew that the Nobel Prize had its home there with the royal family and on and on, but that really wasn't something that crossed my mind in a serious way, until a year ago.

Ken Shulman 11:18
What he does remember is the thrill he felt when he first saw the results of Caitlin's work, the thrill and the pride that only a mentor can know when the student paints his first masterpiece, his first Da Vinci.

David Livingston, MD 11:31
To encounter someone with a mind that's golden, that can approach a really complicated problem in medical science, really complicated, you know, it's largely shrouded in endlessly deep mystery. And crack it open in one experiment. Excuse me? You see that on the fingers of one hand in your lifetime.

Ken Shulman 12:00
Many of Livingston's former postdocs have gone on to make major contributions in cancer research and treatment. And many, including Kaelin are now mentors to one or more generations of researchers.

David Livingston, MD 12:12
Now reaching the success that Dr. Kaelin did is of course, not a daily event. But may I say that Bill Kaelin has a laboratory now and for many years now, and he has also earned the right to be viewed as a great mentor. No question. He has a right to be viewed as a great scientist, unequivocally. He also has, in my view, the right to be viewed as a great mentor. And as some of us would say, a mensch in the truest order of the word.

Ken Shulman 12:54
In 1992, after five years in David Livingston lab, Bill Kaelin opened his own laboratory, he began what would become his life's work, a study of the body's oxygen-sensing system. That study would yield fascinating information about cancer and metabolism and would eventually lead to the Nobel Prize. As head of his own laboratory, Kaelin also became a mentor, and helped to shape a new generation of researchers at Dana- Farber, a generation that impressed him right off the bat.

Bill Kaelin, MD 13:27
A lot of young people are doing everything they can to come to Boston, and they're interested in cancer and to come to the Dana-Farber to do their work. So, I like to say I think the Dana-Farber and Harvard generally, you know, they've set up this sort of nice, positive feedback loop where they have, you know, great young people beating on their door. And that's what makes them a great institution. And because our great institution, they have great young people beating on their door. So, it's almost self-fulfilling. It's a little bit like Silicon Valley and computer science, I think we really attract some of the best and brightest from around the country, and around the world. So, you start with that, I think as being part of the secret sauce.

Ken Shulman 14:08
Kaelin often says that David Livingston made him into the scientist he is today. He also credits Livingston with making him an effective mentor. From Livingston, he learned that one of the first and most important things a mentor can do is help a mentee decide what question they want to answer.

Bill Kaelin, MD 14:27
What problem am I going to work on? Because there's an opportunity cost to everything you do. So, I think that the first decision of what I'm going to work on turns out to be quite critical. In fact, getting back to David Livingston, he used to say that the most important thing for a young scientist is just plant them in a fertile area and water them once in a while.

Ken Shulman 14:45
Kaelin also learned that a good mentor needs to help a mentee stay on mission, to focus on the initial question and avoid distractions and dead ends. It's a problem that plagues today's postdocs in particular, postdocs who have tools that Kaelin and his generation could only dream about.

Bill Kaelin, MD 15:03
You know, there's always some shiny new technology or toy that it's easy to kind of focus on. I want to learn how to do this approach. I want to learn this technology. And so, one of the other things I learned from David Livingston is that every good experiment actually starts with the question, you know, what question are you trying to answer? And if you start with the question, you can figure out what technologies you might need to answer the question. I think with some of these shinier and newer and in some cases, sexier technologies, there's a temptation to start with the technology and try to figure out what questions you can answer with that technology.

Ken Shulman 15:40
In 2001, almost 10 years after he opened his own lab, Kaelin published a blockbuster paper, the paper described in exquisite detail the mechanisms the human body uses during hypoxia, during times when oxygen is low. This and additional work on oxygen sensing and metabolism would eventually earn him the Nobel Prize in Medicine. Most postdocs came to Kaelin's lab to work on those specific problems, or on closely related ones. Enter Julie Losman.

Julie Losman, MD 16:13
When I started at Dana-Farber, I was reasonably sure that I wanted to study, research-wise I wanted to study leukemia. There's really good treatments for a lot of patients, but there's still a huge unmet need. So, it had that resonance for me.

Ken Shulman 16:25
Julie Losman started at Dana-Farber in 2006. As a resident at Johns Hopkins, she'd worked on leukemia. The biology of the disease fascinated her. So, it followed that at Dana-Farber, she joined a lab that studied leukemia, but that lab closed a year and a half later. Orphaned, Losman started looking for a new lab where she could work. She met with several scientists, including Bill Kaelin. He liked her immediately and her story was touchingly familiar.

Bill Kaelin, MD 16:55
I actually did not start out working with David Livingston. I started out working with a young faculty member named Shelly Bernstein. I was Shelly Bernstein's first postdoctoral fellow. And it turned out I was also his last postdoctoral fellow because four or five months after I worked with him, he called me into his office and told me he was shutting down his laboratory and he was going to go into clinical practice. And likewise, Julie, started out working with a wonderful scientist named Gary Gilliland. But then Gary decided to leave academia to go work at Merck. And so, Julie arrived at my doorstep an orphan. And it made me think back of my time, looking for laboratories and landing with David Livingston. And Julie is just a superstar.

Ken Shulman 17:37
But sympathy and shared history weren't enough of a reason to work together. They needed a plan.

Julie Losman, MD 17:43
Bill doesn't study leukemia, and I definitely did not want to switch out of the leukemia field. So, I actually ended up meeting with Bill early on in 2009. And then sort of mid-2009, making the decision to join his lab, because basically, what he told me was look, as long as you're studying an important question, as long as you're sitting an interesting question, and as long as you're studying a question that I feel that I can bring some knowledge and some insight to, you can work on whatever you want in my lab.

Ken Shulman 18:14
The pair came up with a suitable project shaped by their shared expertise. The work went well. And then things got even better. Another postdoc in Kaelin's lab was studying a genetic mutation that just been identified in brain tumors. The mutation was called IDH. IDH was of great interest to Kalin because it plays an important role in metabolism. Then there was an extraordinary stroke of luck. Researchers at Washington University found that same IDH mutation in leukemia.

Julie Losman, MD 18:49
This is the perfect storm - these things are not supposed to happen in nature just where all of a sudden, you know, this enzyme involved in metabolism that Bill and the entire lab was very interested in, had been working on in brain tumors, and all of a sudden I joined the lab wanting to study leukemia, and this mutation pops up in leukemia.

Ken Shulman 19:07
Kaelin asked Losman to investigate the role that this specific enzyme, the mutant IDH, played in leukemia. It started off as a side project. Within a few weeks, it took over her work life, the perfect storm morphed into the perfect project. In its natural state, the IDH gene is a fundamental part of the cycle that generates energy for cells. Losman knew that the mutant form of IDH changes a cell's metabolism. It changes the way a cell produces energy. The mutant IDH also changes the by-products the cell emits during metabolism. These byproducts are called metabolites.

Julie Losman, MD 19:50
And so instead of basically converting one metabolite to another, it produces a completely new metabolite, this metabolite called 2-hydroxygluterate or 2HG. This metabolite normally is found at very, very low levels in cells. Cells kind of get rid of it because it's considered to be somewhat toxic. But in these IDH mutant cells, you have these massively high concentrations of this metabolite.

Ken Shulman 20:16
The mutant IDH gene causes cells to produce a metabolite called 2HG. And this metabolite 2HG, looks a lot like another metabolite that Kaelin's lab knew well. That other metabolite was part of the system that sounds the alarm when oxygen is low. You might remember that Kaelin had already shown how cancer games the oxygen-detection system, how it tricks that system into sounding a false alarm, and how that false alarm triggers a series of mechanisms that push tumor growth at the expense of healthy cells. Losman and everyone in Kaelin's lab expected 2HG to do the same thing, just like its chemical cousin did. They expected it to also trick the body into thinking that oxygen was scarce. That trick would help hijack the system and feed tumors. This was how many cancers grew. Many, but not all. Just after Losman joined the Kaelin Lab, another postdoc discovered that the false low oxygen alarm factors that drove so many cancers did not drive brain cancer. In fact, they stifled it.

Julie Losman, MD 21:31
And so this was all really interesting and very paradigm shifting. One of the questions was, is this all true, likewise true in leukemia? And so that's how I started working on this in the context of leukemia.

Ken Shulman 21:44
So Losman designed her initial experiment, she took a cell line, she removed the cytokine. That's the growth factor that cells need to grow. Then she introduced the mutant IDH. If the cells resumed growth, it would mean that mutant IDH caused cancer. But the initial experiments failed. The cells couldn't tolerate the mutant IDH. So, she found another cell line, a human leukemia cell line called TF1.

Julie Losman, MD 22:16
These cells could tolerate it. And when I put mutant IDH into the cells, and then withdrew cytokine, the cells grew.

Ken Shulman 22:25
Losman would repeat the same experiment more than 50 times during her five years in Kaelin's lab, the results were always the same. Mutant IDH was a driver in leukemia. And the lab work was exhilarating. With this model in hand, she was able to show that the Leukemia was indeed driven by the metabolite called 2HG. She deciphered how 2HG allows leukemia cells to grow.

Julie Losman, MD 22:53
It's really fun to do science that feels like you're going downhill, right? So much of science is going uphill, every once in a while, when you hit on a project where it actually feels like you're running to catch up with the science, it's really, really fun. I do remember the very first time that I got the results of the experiment, I plotted the graph of the growth of the cells that were expressing wild type IDH. So, the non-cancer associated form and the mutant form, and I brought it to Bill and I showed it to him and I was very - I don't think we even were supposed to meet, I was so excited to show it to him that I sort of popped into his office. And he looks at that and he said, “Julie, this is your Da Vinci.”

Ken Shulman 23:30
For Kaelin, Losman's result was as beautiful and timeless as the Mona Lisa. Perhaps even too beautiful. Like everyone else, scientists need to be wary of getting what they wish for. Kaelin remembered an important lesson he learned from his mentor, David Livingston.

Bill Kaelin, MD 23:54
So, David was very good after the initial wave of euphoria, when you thought you had discovered something, quietly suggesting that maybe there were other things you'd still be thinking about, maybe there was one more experiment you should do, one more thing you might have overlooked. And so, I'm very much the same way with my mentees that I'm pretty good at coming up with alternative explanations for their data that will then force them to maybe think about doing an additional experiment or two.

Ken Shulman 24:22
But it didn't take long to see there was nothing ambiguous in Losman's experiments, or her data.

Julie Losman, MD 24:28
And that was one of the you know, the beautiful thing about that experiment, and about that result was that the result was just plain to see, you didn't have to, there was nothing to interpret. He just looked at the result was there and that was really the result on which I built my entire postdoc, which was focused on understanding how this mutation, how this metabolite, how this 2HG promotes cancer specifically in the context of leukemia.

Ken Shulman 24:57
Kaelin knew Losman had made a major breakthrough, but neither of them were done. In science as an art, great work inspires more great work. Losman's Da Vinci asked as many questions as it answered. And the most obvious question was whether you could kill cancer cells if you somehow blocked the 2HG, if you block the toxic metabolite produced by the mutant IDH gene.

Julie Losman, MD 25:23
Now, I think, again, a lot of scientists who are maybe not quite as sophisticated as Bill would say, “Well, of course, right? Of course, you can, you know, you have a mutation, that mutation is important to drive the cancer. Of course, if you inhibit the mutant, you're going to kill cancer.”

Ken Shulman 25:40
But they weren't sure. There are two basic models on how cancer develops. In the first model, a genetic mutation turns on an irregular growth program. And that irregular growth program drives the cancer cell. The second model, a model called the hidden run, also starts with a genetic mutation. But in this model, the genetic mutation sets off a series of changes, one after another. And it's these downstream changes that drive the cancer. The initial mutation, the hit and run mutation does ignite the chain reaction, but it doesn't sustain it. And you can't stop the cancer by targeting the hit and run mutation.

Julie Losman, MD 26:25
And there were reasons to think, very plausible reasons to think, that this hit and run phenomenon could be how mutant IDH functions. Because when you rewire the metabolism of the cell, you have a lot of secondary and tertiary effects that may not be reversible.

Ken Shulman 26:46
Losman spent her final two years in Kaelin's lab on this problem. She wanted to see if you could reverse the transformation from healthy cell to cancer cell by blocking 2HG. Through Kaelin's network, she connected with a local drug company that had just built an inhibitor that would block 2HG.

Julie Losman, MD 27:05
And so, we got our hands on these inhibitors. And I was able to use my assay to interrogate whether these inhibitors actually did reverse transformation. If indeed, leukemias are dependent on this mutant protein. And it turns out that they are.

Ken Shulman 27:21
Losman worked with drug company scientists to translate her research into a therapy for leukemia.

Julie Losman, MD 27:27
And I learned a little bit about, you know, how science and industry is a little different than academia. But really, how you can mesh those two approaches in a way that really brings out the strengths of both It was really incredibly fruitful collaboration.

Ken Shulman 27:41
The drug went into clinical trials. A few years later, the FDA approved it, to treat leukemia.

Julie Losman, MD 27:48
Which, you know, was incredibly gratifying and incredibly rewarding to have spent all that time, you know, digging in the weeds of the molecular biology, where I have to be honest, you know, from on a day-to-day basis, I wasn't really thinking about the translational implications, I was interested in understanding the biology. But once you really understand the biological underpinnings of a system like a cancer cell, you can really start to make headway in terms of targeting it. Bill and I shared a glass of champagne, you know, when the drug really you know, started working clinically because this is really why we do this.

Ken Shulman 28:31
In 2014, after five years working with Bill Kaelin, Losman opened her own lab at Dana-Farber. Over time, her focus has shifted from leukemia to cancer metabolism and epigenetics. In essence, she studies how a cell's metabolism can determine which genes are selected for expression. And like her mentor, Bill Kaelin, and his mentor David Livingston, she offers fertile soil and occasional watering for the newest generation of researchers at Dana-Farber.

Julie Losman, MD 29:02
I have to say that the postdocs who work in my lab have generated really beautiful data. And I hope that I've conveyed to them that they themselves have produced their own Da Vinci's.

Ken Shulman 29:16
She still remembers the thrill she felt after painting her first great work in Kaelin's lab.

Julie Losman, MD 29:23
It's still to this day the most beautiful figure I've ever generated. I mean, I'm hoping to produce a few more. You know, Da Vinci was successful because he had more than one Mona Lisa, equivalents, but there is also something to be said for generating something like that for the first time. But on a personal level, you know, for the first time doing an experiment and just having this just sense of what's in front of me is important, is something that I love to say that it happens once a week. Unfortunately, it doesn't. If you're lucky, it happens maybe once or twice a year. But that was a very, very special moment.

Ken Shulman 30:14
Next time, one of Dana-Farber's youngest and most brilliant minds sheds light on one of biology’s hottest field, epigenetics.

Cigall Kadoch 30:22
Imagine a box of books. If you're packing a bunch of books in a box, and suddenly you want to access a book that's at the very bottom of that box, you're gonna have to do some unpacking. And so, the factors that we study are essentially the unpackers.

Ken Shulman 30:35
I'm Ken Shulman and this is Unraveled, a Dana-Farber Cancer Institute podcast.

Season 1, Episode 4: The Blueprints of Your Cells

There are approximately 20,000 genes in the human genome. 20,000 packets that house the blueprints for every human cell. And every cell contains a complete copy of that genome. It's an incredible feat of bioengineering. But here's the question: If every human cell contains every single human gene, how does the cell know what to do? What process determines whether the cell becomes a pancreas or a patella? And what happens if that process falters - if the right gene is selected but is cast in the wrong role?

Season 1, Episode 4 Transcript

Ken Shulman 00:12
The human genome, the instructions that make our species what it is. Our genome includes about 20,000 genes, 20,000 packets of information made of DNA -- packets that contain the instructions to build every human cell. The genome's a package set. There aren't any partial copies. Each of our 30 trillion cells gets a complete version of the genome, all 20,000 genes. And yes, it's hard to do that math in your head.

Cigall Kadoch 00:47
So, if you can imagine each cell within the human body has essentially two meters of DNA. If you were to take the DNA out of each cell, it would stretch to two meters, and that two meter length of DNA must fit inside the nucleus of a cell, which is less than the size of a pinhead in diameter.

Ken Shulman 01:02
If there's a better job of packing in all of biology, I'd like to see it. But here's the question: If every human cell contains every single human gene, how does the cell know what to do? What process determines whether it becomes a pancreas or a patella? And what happens if that process falters? If the right gene is selected, but cast in the wrong role. I'm Ken Shulman, and this is Unraveled, a Dana-Farber Cancer Institute podcast. Genetics, as the name implies, is the study of genes. Those packets of DNA that contain the instructions to create all living things from yeast cells to fruit flies to humans. Genetics have given us a huge leg up in the fight against cancer. Today, we can trace the source of many cancers back to mutations in specific genes. This opens the door for targeted therapies that can be more precise, and less toxic than traditional cancer therapies. But as usual, there's more to the story here, a lot more. Starting with size. The human genome is microscopic, six feet of DNA, wrapped tightly around tiny protein spools, and then crammed into a space 15 times smaller than the diameter of a human hair. At that level of compression, only a very small subset of our 20,000 genes are on the surface and accessible at any given time.

Cigall Kadoch 02:54
So, to allow that amount of genetic information to be condensed so many times and to be put inside the nucleus of a cell, there must be factors that can control how this DNA, this genetic material is actually packaged within the nucleus. And so, there, if you can pack, you can imagine a box of books if you're packing a bunch of books in a box, and suddenly you want to access a book that's at the very bottom of that box. You're gonna have to do some unpacking.

Ken Shulman 03:21
That's Cigall Kadoch. She studies epigenetics at Dana-Farber. That's right. Epigenetics, the study of the epigenome. The epigenome are the factors that determine which of our genes are expressed, and which of them are muted. The factors that determine whether a given cell becomes a bicep or a brain, whether a page in the instruction booklet is read, scanned, or skipped over.

Cigall Kadoch 03:49
And so the factors that we study are essentially the unpackers. As they unpack tightly coiled DNA, which is largely turned, the genes that are in that region of DNA are largely turned off, and they open up that DNA to be able to be expressed. So the genes that are contained in that stretch of DNA can then be turned on at the appropriate times.

Ken Shulman 04:10
Epigenetics, the study of factors that unpack our DNA, that six-foot strand wrapped around those tiny protein spools. Epigenetic factors jiggle those spools, to loosen a stretch of DNA so it can be read, or they crank the spools to put another section out of eyeshot. It's wonderfully and almost confoundingly complex. Imagine playing a four-dimensional slot machine, where dozens and dozens of charities have to line up in space and time in order for you to have your father's eyes and your mother's nose. And that's only a vague metaphor. For now, just know the epigenome has myriad moving parts, and that it keeps our 20,000 genes in line. Just like the genome, the epigenome occasionally stumbles. And these missteps, these epigenetic mutations can favor or even cause cancer. Epigenetic mutations can turn off genes that cells need to grow healthy and strong. And they can switch on genes that fuel runaway growth. Genes that feed tumors. The part of the epigenome that Cigall Kadoch studies is called a BAF complex. That's BAF -- B A F. The acronym stands for a group of proteins, a protein complex that rolls up and down our DNA, choosing which genes will be selected for expression. How does it work? The BAF complex loosens the spools of DNA, so key genes can be accessed and read at the right time. By doing its job, the BAF complex helps cells and tissues develop normally, and this can help prevent cancer. The BAF complex isn't a loner. There are many similar complexes in the epigenome. But this one, the one that Kadoch studies, punches way above its weight in terms of cancer prevention. In fact, she's linked more than a fifth of all human cancers to defects in the BAF complex to this specific epigenetic miscue,

Cigall Kadoch 06:21
You know, in some cancers, it's 100%. The hallmark feature is a problem in this one complex that we study. And in other cancers, you know, much more common cancers, for example, lung cancer, you know, 12% of lung cancers have a mutation in this complex, you know, 60-70% of ovarian cancers, you know. So to see that if you tally up the mutations and all the various components of this multicomponent regulator these tally to over 20% of human cancers.

Ken Shulman 06:47
The idea of the epigenome isn't all that new. Back in the 1740s, Carl Linnaeus, the Swedish botanist observed changes in flowering plants that he couldn't explain with classical genetics. That was an early glimpse of the epigenome. But it wasn't until the 1970s that scientists began to understand how the system worked. And it would take another two decades before researchers made the first connection between the epigenome and cancer. And even then it was a limited connection. That first connection was made when scientists discovered something called DNA methylation. That's a process that occurs during cell division, where a chemical tail is pinned onto a DNA molecule. The chemical tail quiets the gene. And with that gene silenced, other genes take its place in the growth program. If enough of the growth program is altered, the result can be cancer. DNA methylation was an important discovery. But it was only the tip of the epigenetic iceberg.

Myles Brown 07:52
Over the years, we've come to realize that, you know, things that we thought were relatively simple and straightforward are always more complicated.

Ken Shulman 07:59
That's Myles Brown. He directs the Center for Functional Cancer Epigenetics at Dana-Farber. Brown studies how steroid hormones, hormones like estrogen and androgen drive certain cancers. When he started out, he thought those steroid hormones caused DNA methylation, that they help produce the chemical tail that muffles genes. It turned out he was wrong. Brown learned that the steroid hormones didn't use DNA methylation. Instead, they drive cancer by bonding directly with receptors in tumor cells. That bonding makes tumors grow faster. As Brown and many others learned, the epigenome has lots of tools in its toolkit. There are chemical tails, like the ones in DNA methylation, but there are also enzymes that scientists called writers, enzymes that scribble notes on DNA. And there are other enzymes called readers that, you guessed it, read those notes. There are even enzymes called erasers. If all this sounds complicated, that's because it is and because it has to be.

Myles Brown 09:07
The human genome has only about 20,000 genes. And that's only about three times more than a simple organism like a yeast. But of course, humans are more than three times more complicated than a yeast.

Ken Shulman 09:25
So it's not the number of genes that makes us what we are. It's the way we've evolved over time in using those genes. If there's a system that makes us truly special, it's the epigenome, the software program that runs on our cellular genetic hardware. And it's a big program that requires lots of computing power and lots of memory. And it's a program that occasionally malfunctions. Let's take a moment here. We know that our genome and our epigenome can experience mutations. And we know that in both cases, these mutations can lead to cancer. But the two processes, the way they unfold, are different. A DNA mutation is like a script error, a case of garbled instructions. A mutation in the epigenome is like a total rewrite, when one set of instructions is swapped out for another. Think of a soap factory. A genetic mutation would be a broken mold that makes the bars of soap come out looking like Swiss cheese. An epigenetic error in that same factory would be like someone reprogramming the machine to make it spit out tennis rackets. Perfectly good tennis rackets, mind you, but they won't get your hands clean. But here's the most important thing about the epigenome and cancer. Unlike genetic mutations, which are mostly permanent, mutations in the epigenome can be undone. Which is why researchers like Brown are so interested in the epigenome,

Myles Brown 11:02
The attraction of epigenetic changes, is that they are by their nature reversible. That's their nature is to change program during development, for example. And while not simple, they aren't permanent.

Ken Shulman 11:21
Back to Cigall, Kadoch, and the BAF complex, that's the protein group that fine-tunes our DNA. Kadoch is one of Dana-Farber's youngest lab directors. She can't remember a time when she wasn't interested in science. At six years old, she already had her own microscope. All through her childhood, she went on regular nature walks with her parents, and with a beloved caretaker,

Cigall Kadoch 11:45
I loved kind of going back and forth between reading books, you know, with this caretaker and with my parents, and then going out into nature and trying to understand how much of what was in the textbook or in the book that I was reading was actually true, and how much were there other nuances that perhaps one could only glean from really being out in the fields or outside and I think that's a parallel for research as well. One can learn a certain amount from textbooks, and course you need the fundamentals there. But so many of the answers are still unknown.

Ken Shulman 12:11
Shortly after Kadoch turned 12, her caretaker was diagnosed with end stage breast cancer.

Cigall Kadoch 12:18
And within just short, short three months, she actually passed away. And this had a profound impact on me as a young adolescent, and as a person and it has continued to have an impact on me. All throughout my life. In fact, this was relayed I would say what initiated my interest and ultimately lifelong commitment to cancer research.

Ken Shulman 12:38
When she was in high school, Kadoch shadowed her uncle, a radiation oncologist, as he made his hospital rounds. She watched him treat patients. She saw how some of them got better, and some of them didn't.

Cigall Kadoch 12:51
And I just couldn't find the answer as to why that would be. And you know, I sat, in between patients I would sit in his office and read textbooks, and I tried to read every textbook that was on his shelf, just trying to see how much I could understand this advanced material, which was advanced for me at the time. And a lot of times the answers weren't in there.

Ken Shulman 13:09
Kadoch continued to shadow patients in college. She read everything she could about cancer and its biology. But the answers still weren't coming. And then it dawned on her. She wasn't going to find what she was looking for in a textbook.

Cigall Kadoch 13:25
I was going to have to search for them myself. And I was going to have to start engaging in the research community to figure out what are all the strategies we have to explore - scientific questions, and that's what really got me excited about research.

Ken Shulman 13:39
The BAF complex that Kadoch and her lab study, the group of proteins that, when defective, is implicated in more than 20% of all human cancers belongs to a family of molecular machines called chromatin remodelers. Let's unpack that. Chromatin. That's the fancy name for the weave of DNA and proteins that makes up our chromosomes. Chromatin remodelers, like the BAF complex, do exactly what the name suggests they remodel chromatin. Think of a well-honed team of master artisans, supervising proper cell development and growth. Kadoch was interviewing for a PhD program at Stanford when she first encountered the BAF complex. The admissions committee there arranged for her to meet with a senior faculty member named Gerald Crabtree. Crabtree studied stem cells along with the development of the brain and nervous system. Kadoch, on the other hand, was almost solely interested in cancer biology. To this day, she doesn't know how Crabtree ended up on her list.

Cigall Kadoch 14:46
And he had done some beautiful work in the context of the vertebrate nervous system, and even other tissues and really exploring the roles for chromatin regulatory machinery in development. And I just didn't realize that that was going to be of interest to me but when I heard him describe the work that was going on in the lab, again, focus on the brain and on development of human tissues, I was intrigued by his description of these factors which I now study and our whole lab studies -- these BAF complexes, chromatin regulatory complexes -- that were required for virtually every cell fate transition, for every development of every tissue. And it just seemed that as he was describing this, it's exactly the type of process that goes awry in cancer.

Ken Shulman 15:32
Whether their meeting was by accident or design, Crabtree became her thesis advisor. Kadoch told him she wants to find a link between the BAF complex chromatin remodeler that Crabtree worked on, and cancer.

Cigall Kadoch 15:47
And so he sort of set me loose. And you know, I was surprised and still looking back on that I'm still surprised that he let me do that. There wasn't anybody in the lab studying cancer, he himself wasn't really studying cancer, I happen to be in the cancer biology program. But jumping into the lab, he essentially just showed me a bench, it was just an empty bench, gave me a set of pipettes and just said, knock yourself out, and go, go for it.

Ken Shulman 16:08
The work went well. At the same time, a flurry of papers appeared in major journals all about chromatin remodelers in cancer.

Cigall Kadoch 16:18
Time and time again, I remember actually pushing. I could barely keep up with the printer pushing "Print" to print these papers and run down the hall to show Jerry yet another paper was implicating these genes and mutations in these genes as a major cause or contributor to human cancers.

Ken Shulman 16:33
While still in grad school, Kadoch published two important papers. In the first, she showed how a gene mutation in the BAF complex leads to a rare and hard to treat cancer called synovial sarcoma. In the second paper, she dropped a bomb, the bomb that linked defects in the BAF complex to 20%, or even more of all human cancers.

Cigall Kadoch 16:58
This is a huge, previously entirely unappreciated burden. And it behooves us then as researchers to understand the mechanisms of these large molecular machines, and to be able to find new strategies to impact their activities on the genome.

Ken Shulman 17:16
In science, as in most fields, you need to know how things work before you can fix them. If we're going to invent new strategies to impact the epigenome, we need to fully understand it. We already have some of the details. The DNA methylation, that chemical tail that hits a genes mute button, hormones that bond with cell receptors, and convince them to change their preference and chromatin remodelers, those artisans that fine tune our chromosomes, jiggling or cranking the spools to make sure the right genes are expressed. We've seen that some cancers are driven by genetic mutations by hardware glitches, and that other cancers are fueled by mistakes in the epigenome by software bugs. But sometimes, the line between hardware and software is blurred. Scott Armstrong studies a cancer that hijacks both systems, a pediatric cancer called mixed lineage leukemia, or MLL.

Scott Armstrong 18:20
It's most prominent, most well-known, if you will, in children less than one year of age: infant leukemia. Because if a child is diagnosed with leukemia before the age of one, the likelihood that they have mixed lineage leukemia as defined by a mutation in the mixed lineage leukemia gene is about 70%.

Ken Shulman 18:44
Armstrong is Chairman of the Department of Pediatric Oncology at Dana-Farber. He says doctors can cure many forms of pediatric leukemia with chemotherapy and other standard therapies, but they haven't had the same success treating MLL

Scott Armstrong 19:00
So as a pediatric oncologist where we cure most of our patients with leukemia, these types of leukemia stand out because it's much more difficult to cure them.

Ken Shulman 19:13
Armstrong and his Dana-Farber colleagues were among the first researchers to map out how mutations in the MLL gene induced leukemia. But as the cancer progressed, they noticed that its growth changed. It looked like the initial genetic mutations were now influencing gene selection, determining which genes were active. The lab suspected that something in the epigenome was driving this change in growth in gene activity. So they did a gene activity profile. That's a survey to see whether an aberrant set of genes was activated in MLL tumors. They got an assist from a very unlikely species.

Scott Armstrong 19:56
From, believe it or not, studies in Drosophila, in the fruit fly, where they had already shown us that this similar protein that's found in the fruit fly was controlling those same sets of genes.

Ken Shulman 20:07
Let's unpack that one. The mutant MLL protein was controlling the same set of genes in leukemia patients that a similar protein controlled in fruit flies, genes that wouldn't have been activated in a healthy patient's genes that drove abnormal growth. It turns out that the mutant MLL gene, the gene that causes the cancer, packs a double wallop. First, it causes the cancer, then it hijacks the epigenome, jiggles the DNA spools, to uncoil a set of genes that will sustain the tumors. This cancer, initially caused by a genetic mutation by a script, typo, produces a protein that takes over the cellular software. It gets a foot in the door, and then it rearranges the furniture. It's a very clever and very sinister survival mechanism.

Scott Armstrong 21:02
During normal development, cells are to some extent constantly sampling their environment and deciding which genes should be turned on and off as a result of what the body needs at any given time. And this type of cancer and probably many, if not most types of cancers, take, unfortunately take advantage of that, either in the initial steps of development, or later, when we try to treat them. They use that adaptation process to become resistant to treatments. And we think, even though it's still a little hard to prove, that cancers that have mutations in these genes that influence epigenetics directly, may be better able to do that, have more options, if you will, to adapt. And so, yes, we we know, we're fighting an enemy here that is very adaptable, and that obviously creates problems.

Ken Shulman 21:57
Yet along with the problems, Armstrong and others also see opportunity. Because changes in the epigenome are by definition reversible, they offer an attractive target for therapies. The target is still a few years off, but we've already drawn much closer to it. A number of drugs, many developed by Dana-Farber researchers, are in clinical trials, primarily drugs for blood cancers, like leukemia and lymphoma. The drugs are mostly small molecules that are agile enough to operate within the cell nucleus, that tiny space where epigenetic mischief and magic happen.

Scott Armstrong 22:35
I think there's a lot to be done there. We know the fundamentals now. But I think once we understand that, then we can develop increasingly selective drugs that can turn specific genes on and off. And that's, that's really the goal. If we can do that, we could potentially use this therapy for all kinds of things... beyond cancer.

Ken Shulman 23:00
It's not unusual that research in one area of science spills over into a seemingly unrelated area, especially when you study DNA. In addition to linking the BAF complex to many human cancers, the Kadoch lab has tied defects in the same complex to intellectual disability and autism spectrum disorders.

Cigall Kadoch 23:22
And that was a very exciting set of findings to come from this, in the sense that it tells us a whole new way by which the chromatin remodeling activity, the unpacking activity of this complex, is actually regulated. In some cases, the changes are so subtle, just one change, one single amino acid is sufficient to produce a majorly devastating outcome.

Ken Shulman 23:44
Examining protein units like the BAF complex can be tricky. For one thing, they're large, and often contain 15 or more sub-units. And it's hard to extract them from cells and reduce them to the native state so they can be studied. They tend to break apart during purification. Kadoch struggled with the purification process for several years. Two years ago, when she and her team finally perfected it, they were able to learn much more about the 3D structure and organization of the BAF complex.

Cigall Kadoch 24:18
And that one breakthrough that we had, which was published in 2018 Cell paper that we had, we used that to inform the order of assembly. How do these big macro molecular machines piece together from 11 to 15 sub-units? How do they come together? How are they pieced together? And then how would the disappearance of any one of those 11 to 15 subunits or puzzle pieces, how would that change the overall architecture of that complex? And how would it change its ability to know where it needs to go in the genome, to turn on certain genes, to open DNA accessibility. And that's something that we've been spending quite a lot of time on.

Ken Shulman 24:50
Given the critical role the BAF complex plays in human cancer, Kadoch wants to study it from as many vantage points as possible. Her lab includes experts in Biochemistry, Physics, Structural Biology, Genomics, and even Computational Biology and Machine Learning -- any discipline that might shed new light. Imaging is also extremely important.

Cigall Kadoch 25:13
And more and more, what's going to be critical is understanding how these complexes are structurally arranged with one another, you want to see how all of these are interacting together. You want to be able to achieve a picture, with very high resolution, of the cell itself, of the nucleus, of the stretches of DNA that you're interested in, and be able to examine the various proteins that are engaging there.

Ken Shulman 25:35
Kadoch intends to follow her work as it translates into therapies that can reverse cancer. She knows the foundation of the translation is basic research,

Cigall Kadoch 25:45
The best translation and the best approaches toward therapeutic discovery really come from mechanisms that are well understood. So from our perspective, our key area of focus is to achieve the highest degree of detail, for any given molecular mechanism that we are exploring. You know, we aim to achieve a complete picture of the structure and function of these chromatin remodeling complexes, understanding through the lens of evolution, why we've evolved to have certain components on these complexes that weren't present in yeast. They weren't present in flies. And now here they are present in human cells. Why do we need to add those on during the process evolution to accommodate what new feature of our genome?

Ken Shulman 26:27
The problem, she says, needs to be viewed through multiple lenses, from the wide angle of evolution to an in-your-face close up of BAF machines at work, and genes turning on and off. Kadoch and her lab have already made the first important steps towards translation.

Cigall Kadoch 26:46
And so in our case, there are a number of cancer types that are uniquely dependent on the activity, or the aberrant activity of these chromatin remodeling complexes. And so, with as many of the times we've found a new mechanism, almost always there is a route toward a potential therapeutic opportunity that comes directly from the mechanism. You know, but of course, as you well know, to try to translate that into meaningful therapeutic for our patient, a lot of work needs to be done, a lot of work needs to be done.

Ken Shulman 27:13
So how would an eventual therapeutic work and which parts of the epigenome would make the best targets? Myles Brown thinks it might be possible to recruit the epigenome to ramp up our body's natural defenses to write a software program that could turn our T cells into cancer fighting Navy SEALs. Scott Armstrong sees parallels between his work on childhood leukemia, and Cigall Kadoch's work on the BAF complex. In both cases, he says, cancer is driven by mutant proteins that hijack the machinery. And he'd like to defeat the hijackers.

Scott Armstrong 27:51
And if we can figure out how to reverse that hijacking, meaning get rid of that interaction between the cancer-causing protein and in this case, the BAF complex, then that should reverse the epigenetic process that's driving the cancer. MLL's the same way; it's a different complex that's been hijacked. But it's the same concept that the epigenetic complex that normally does something is now being used to do something wrong.

Ken Shulman 28:20
There are myriad potential applications of the work that Kadoch, Armstrong and Brown are doing on the epigenome. Armstrong thinks we may see some of the first viable epigenetic therapies in pediatric cancers, because pediatric cancers in general, are simpler than adult cancers. Of course, he'd be happy to see a drug that could treat MLL.

Scott Armstrong 28:43
The opportunity to develop drugs that help people is certainly what drives, not just as a physician, but many of us in the field -- physicians and scientists. But the details of exactly how things work in a very complicated system like this is kind of what gets you up every day. And, you know, when talking to younger PhD students or people in training, they'll frequently say, the first time that they discovered something that they knew that they were probably the only person in the world that knew that's how it worked. That's kind of a career "Aha" moment. And you don't lose the excitement around that when somebody shows you something that you wouldn't have thought of, or it's a new idea that you know, that no one's really ever thought of before. That's very exciting.

Ken Shulman 29:45
Next time, science fiction becomes science fact.

Judy Wilkens 29:50
It was like they put a Pac Man in my body, and the Pac Man would go through and they would see a cancer cell, omm, and just eat that cancer cell instead of going through my body and destroying all my good cells.

Ken Shulman 30:03
I'm Ken Shulman and this is Unraveled, a Dana-Farber Cancer Institute podcast.

Season 1, Episode 5: Turning Science Fiction into Fact

It's always fun to think about the future. Flying cars. Cities in the clouds. Colonies on Venus and Mars. OK, we didn't end up living like the Jetsons. But a few of those visions did come true. And Judy Wilkins knows that first-hand. Wilkins is a former patient at Dana-Farber and is the poster child for a bold new therapy where science fiction becomes science fact.

Season 1, Episode 5 Transcript

Ken Shulman 00:05
It's fun to think about the future. Flying cars, cities in the clouds, colonies on Venus and Mars. Okay, so we didn't end up living like the Jetsons. But a few of those visions did come true.

Judy Wilkens 00:23
What they do is they manufactured my cells in petri dishes, and they grow them for a period of approximately two weeks, so that they can use these cells to put back into my body once they have re-energized the cells. And to me, it was just like, total science fiction.

Ken Shulman 00:46
That's Judy Wilkens. Sometimes she goes by another name.

Judy Wilkens 00:50
I mean, we had a funny thing that I was Judy Jetson.

Ken Shulman 00:53
Wilkins is a former patient at Dana-Farber, and she's the poster child for a bold new therapy, where science fiction becomes science fact. I'm Ken Shulman, and this is Unraveled, a Dana-Farber Cancer Institute podcast.

Judy Wilkens 01:28
My name is Judy Haley Wilkins. I live in Salem, Mass, and I own my own hair salon. And I'm a hairdresser.

Ken Shulman 01:38
Judy Wilkens grew up in central Massachusetts listening to vinyl records and watching the Jetsons on TV. She's had her own hair salon for 40 years. In May of 2013, she came down with a staph infection, it wouldn't go away.

Judy Wilkens 01:53
And then after that, I ended up with a lump under my left arm in my armpit area. So, at that point, I went to my primary care physician just to have everything looked at. And at that point, we thought it was breast cancer.

Ken Shulman 02:09
It turns out Wilkens did have cancer, but it wasn't breast cancer. It was lymphoma. She saw an oncologist and he wanted to start treatment right away. Wilkens wasn't sure; she made an appointment at Dana-Farber for a second opinion.

Caron Jacobsen, MD 02:25
She had a slow-growing type of B cell non-Hodgkin lymphoma when I first met her that didn't require immediate therapy and so I watched her for some time.

Ken Shulman 02:34
That's Caron Jacobsen. She treats lymphoma patients at Dana Farber. She told Wilkins, “Many people with lymphoma don't need treatment, at least at first.” For the moment, they just needed to monitor her condition. Wilkins would come in every three months for a checkup and bloodwork.

Caron Jacobsen, MD 02:52
And then after a couple of years of keeping a close eye on her, she developed... the slow-growing lymphoma underwent something that we call transformation, which is where it acquires new genetic changes or mutations that cause it to behave like a fast-growing lymphoma. So, it grows much more quickly.

Ken Shulman 03:09
Wilkins’ slow-growing lymphoma had transformed into a fast-growing aggressive form of cancer that changed the landscape, and the strategy. Jacobson prescribed an immediate course of chemotherapy.

Judy Wilkens 03:23
And that's when the roller coaster ride started.

Ken Shulman 03:29
First, they tried one drug, and when that didn't work, they tried another. We did…

Judy Wilkens 03:33
We did RICE, Rituxan, blah, blah, blah, many, many, many. And every time I would go back for my PET scan, after I would do my chemotherapies, I would still have light up spots.

Ken Shulman 03:46
It was a roller coaster ride. Because Judy Wilkens wasn't fighting just one cancer, she was fighting to the original slow-growing lymphoma and the newer, fast-growing lymphoma.

Caron Jacobsen, MD 03:59
Over the next 18 months or so, she went from one therapy to the next because each time we treated one of her lymphomas, you know, successfully, the other lymphoma wouldn't go away. And then we would treat the other lymphoma.

Ken Shulman 04:13
It was like wrestling a tag team all by yourself.

Caron Jacobsen, MD 04:16
So, we went back and forth. But over the course of a year and a half, we never were able to get her into a complete remission.

Ken Shulman 04:24
Complete remission. Jacobsen needed Wilkens to go into remission so she could have an allogenic stem cell transplant. That's a treatment for patients when chemo hasn't worked, or whose cancer is likely to return even after chemo. The allogenic stem cell transplant is basically a total bone marrow swap. The patient's bone marrow is completely replaced by marrow from a donor. You need to be in temporary remission to do it. Otherwise, any lingering cancer cells can contaminate the new marrow. For Wilkens, the chemo treatments weren't getting the job done. Even so, Jacobson started prepping her patient for the transplant. She explained how they'd use chemo and radiation to kill Wilkens' own bone marrow. Then they deliver the donor cells by infusion. She told Wilkens that after the transplant, her immune system would be wiped out. It would take months for the new marrow to churn out enough white blood cells to protect her from infection. During that time, even a common cold could be disastrous. Wilkens would have to isolate until her new immune system kicked in.

Caron Jacobsen, MD 05:41
And Judy is a self-employed hairdresser. She owns her own hair salon. And after allogeneic stem cell transplant, you need to be out of work for a year because of your immune suppression and risk of infection. And that was not going to be tenable for Judy, so she was really struggling with whether that was something she would be willing to do.

Judy Wilkens 06:01
And one of the main things for me was, I would have to have - I have parrots and I have a cat that I would have to have them removed from my home, which is like having my kids taken away. Sorry. So, for me, it's all of that was just really, really scary. But I knew that if that's what I had to do, I would do it. You know, with Caron following me all the way through and my nurses were fabulous; Kathleen McDermott walked me through everything. So, at that point, I understood that this is what we needed to do in order to get rid of the lymphoma because the chemos weren't working. The chemos were just poison in my body. At that point. I was getting so sick.

Ken Shulman 06:47
Judy Wilkens was ready for the transplant, at least in her mind. Her body had other ideas. Try as she might, Jacobson just couldn't get her into remission. She tried a non-chemotherapy drug, but that didn't work either. Without remission, there was no way they could do the stem cell transplant. The situation looked bleak. And then a ray of light. An alternative to stem cells appeared on the horizon. It was early 2016. There was a clinical trial underway. A trial for patients with the same lymphoma that Judy Wilkens had, the trial was called Zuma One. And the therapy involved was something called CAR T cells. It was a living drug made from T cells, T cells that were taken from the patient, genetically re-engineered in the lab, and then returned to the patient. There was one spot left in the Zuma One study, Wilkins would be patient number 101. But Judy wasn't sure she wanted to participate. There was a lot of risk.

Judy Wilkens 07:59
One of the main things that we talked about were the side effects of these type of things. And one of the side, bad side effects were brain swelling, big one was death, which I didn't want to go down that road at all, but the brain swelling and it's hard for me to remember everything and just dizziness and sick, sick, sick. So, I'd said to Caron, I don't think I want to do this. So, Caron, an incredible person, held on and she kept calling the research people telling them that, you know, I need this slot for this one patient. And I'm on the other hand, I'm sitting at office saying, “You know what? I think I'm going to work through the holidays. Let me get through Christmas.” Now we're in June, May-June at this point. And so maybe I'll try the next trial that you know, you're talking about because I don't want to do this. At that point, I didn't realize that this was would have been the end for me.

Ken Shulman 09:01
There's an imaginary beast you might have seen in a museum or a book on Greek mythology. It's mentioned in The Iliad. Homer describes a monstrous fire-breathing creature with a lion's head a goat's body and a dragon's tail. He calls it a chimera. Of course, the chimera was strictly mythical. There aren't any fossils or footprints. But there is one place where the chimera is real... in biology. No, we're not talking about Frankenstein experiments, at least not yet. Still, there are some real-world chimeras. Single organisms made up of cells with two or more different sets of genes. That's right, a composite organism. Sometimes these chimeras occur naturally. When a fetus for example, absorbs its paternal twin in utero. And sometimes chimeras occur as a side effect of medical treatment. If Judy Wilkens had gone ahead with her stem cell transplant, she would have become a chimera. For the rest of her life, she would have lived with blood cells that had her donor's DNA. Chimeras can also be manufactured in the laboratory. One of these are CAR T cells, Chimeric Antigen Receptor T cells.

Jerome Ritz, MD 10:31
We're essentially changing the characteristics of a normal immune cell. We're putting in new genetic information that that cell does not normally have.

Ken Shulman 10:43
That's Jerome Ritz. He runs the Connell O'Reilly Family's Cell Manipulation Core facility at Dana-Farber.

Jerome Ritz, MD 10:51
So that when it - that cell - is then given back to the patient, that cell can actually target the patient's own tumor cells.

Ken Shulman 11:00
T cells serve as our frontline defenders. They identify and neutralize countless threats, including in most cases, cancer. But sometimes our frontline defenders get overwhelmed. Some cancers put on disguises and slip past our soldiers. Other cancers find a way to turn off the immune response. And sometimes our T cells, after years of battle, just give up the fight. If this were war, the high command might send these troops for some R&R. Then they’d send them back to boot camp, to retrain them and re-arm them, so they can recognize and defeat an ever-evolving enemy. That's what CAR T cells are all about. It's a relatively new therapy. The FDA first approved it to treat leukemia in 2017. And CAR Ts aren't for everybody or every cancer. But few therapies have shown such spectacular results in trials. Here's how it works. You take the patient's blood, you filter and isolate the T cells and put them in solution. Remember, these T cells are either spent, or unable to spot the enemy, or in some cases, both. The next step is to reactivate the T cells. That's done by adding magnetic particles into the mix. Once the T cells have got their mojo back, you want to make them smart to reset them so they can spot tumor cells and attack them. And that's where the chimera comes in. CAR T cells get an injection of artificial DNA. That's what makes them a chimera. And the injection, it's not done with a microscopic needle. It's done with the help of a virus.

Jerome Ritz, MD 12:55
The next step is we then add a viral vector, it's actually made from HIV, which is a virus that can penetrate into cells.

Ken Shulman 13:11
Now hold on, we're not talking about infecting cancer patients with HIV. The HIV virus here is deactivated, it's just a shell, a vector used to penetrate the nucleus of the T cell. Once it's inside the nucleus, the HIV vector delivers its payload, a batch of artificial DNA manufactured in the lab. Now the T cell is a chimera. It has its own DNA and the artificial DNA. The artificial DNA tells the T cell to sprout a new receptor, a receptor that's programmed to seek out a protein that appears on the surface of tumor cells.

Jerome Ritz, MD 13:53
And so, as the cells are activated, they start dividing. They start replicating and they're expanding. And when they're doing that, they're doing that with the genetic information that is now been added. So, all of the cells that have been genetically modified, now continue to express these new receptors.

Ken Shulman 14:13
Pretty soon you've got an army, and that army is infused back into the patient. Shock troops equipped with the latest in tumor vision, a biochemical view finder that guides them to their target, like a heat-seeking missile.

Jerome Ritz, MD 14:29
You're arming them with the right specificity and now we're able to target tumors that would not normally be susceptible to an immune attack because they have good defenses, or they can't be recognized.

Ken Shulman 14:41
CAR T cells do more than just sniff out trouble. They're incredibly buff, unlike antibodies or chemical therapies, which eventually break down. CAR T cells just keep going and going and going. They're living drugs and that can be a problem. Like many immune cells, T cells secrete substances called cytokines. These are proteins that regulate the immune response, both up and down,

Jerome Ritz, MD 15:11
These cells are alive, and they begin to function autonomously. And if the cells are over activated, they start secreting a whole variety of inflammatory cytokines that cause all kinds of problems, toxicities in the patients.

Ken Shulman 15:31
Caron Jacobsen knew the potential upside of CAR T cell therapy. There was data from trials beginning in 2010. Most of the results were dramatic. But she knew about the potential downside too. Side effects were not uncommon, and sometimes dramatic. One of those was cytokine release syndrome.

Caron Jacobsen, MD 15:53
The T cells themselves then release cytokines which are inflammatory substances into the blood. And those lead to downstream inflammation themselves. And they also activate other downstream inflammatory cells, which lead to a cascade of inflammation. And obviously, that can cause flu-like symptoms at its minimum, and it can cause people to drop their blood pressure, go into shock, even go into respiratory failure at its maximum.

Ken Shulman 16:16
And there were other concerns, the CAR T cells, it seems, were able to penetrate the brain-blood barrier to open a breach for fluids and inflammatory cells to trickle into the brain. Once there, these inflammatory cells could cause additional problems. And

Caron Jacobsen, MD 16:34
We know that there are an abundance of other cells that go into the central nervous system that are releasing some of these cytokines which cause local inflammation in the brain. And so, you know, again, at its minimum, patients can get sort of mild confusion and disorientation and at a maximum, they can get brain swelling, which can be potentially fatal.

Ken Shulman 16:56
Still, Jacobson was convinced that CAR T cell therapy was the right choice for Wilkens. Chemo wasn't working, the stem cell transplant was out. Judy Wilkens was losing her battle with cancer. But for Jacobsen, it wasn't just that her patient seemed to have run out of options. It was the type of cancer they were dealing with. Let's take a step back here. In order to work effectively, CAR T cells need two conditions. First, the protein that the CAR T cell is programmed to find has to appear on every single cancer cell. You don't want to leave any cancer cells behind. Second, that same protein has to be vital for the survival of the cancer cell. Otherwise, the cancer cells can lose the protein, evade capture, and continue to replicate. On the flip side, that same target protein can't appear on healthy cells in any vital organ. CAR T cells aren't programmed to recognize tumors, they zero in on specific target proteins. And if that target protein appears, say, in the lining of the colon, or the lungs, well, it won't be pretty. Jacobson knew that B cell cancers like the one that was killing Judy Wilkens checked all the boxes for CAR T cell therapy because there was an optimal target, a protein called CD19.

Caron Jacobsen, MD 18:30
It's necessary for the B cell survival so it's not easily lost. It’s present on all or most of the B cells. And the only other sort of normal cell counterpart that it's on is the mature B cell and we know people can live without B cells. There are genetic, inherited conditions where people are born without B cells, and we can keep them alive and free of infection by giving them intravenous immunoglobulin.

Ken Shulman 18:56
On August 1, 2016, Judy Wilkens, patient number 101, in the Zuma One study, sat in her hospital bed at the Brigham and Women's Hospital, awaiting her first CAR T cell infusion. One of her clients at the hair salon had made her a pair of antennae to bring with her like a character in some space age cartoon. Kathleen McDermott, the nurse practitioner, dared Wilkens to wear the antennae when the medical team showed up.

Judy Wilkens 19:26
So, I'm sitting in my bed and I'm you know trying to be really relaxed and thinking about all the good things and Kathleen texts me back. She's like, “Make sure you have your antennae on.” So sure enough I put my antennae on and all these doctors walk into the room. They're probably like 17 people in the room, ballpark, with my doctor Kathleen and all my support team but then also all the research people for CAR T cell so they came in the room and they're like "What the heck?” and Kathleen says, "Well, this is Judy Jetson."

Ken Shulman 19:58
The infusion lasted about 15 minutes. For Wilkens, it was like any other chemo treatment, a bag of foggy liquid hanging on a pole dripping into her veins. Except this foggy liquid contained her reactivated T cells. And she didn't feel sick the way she did with chemo. After it was over, she jumped out of bed and asked one of the nurses to take a picture of her with her antennae. A doctor came and ordered her back to bed. She didn't feel sick the next day either.

Judy Wilkens 20:33
I did a ton a ton of reading on it. So, I'm anticipating getting really sick right away. And I remember that, you know, the next day, I'm feeling just fine. Everything is no problems at all. And I talked with Caron and later in the day I had called Kathleen, I said to Kathleen, and I said, "You know, I need to talk to the doctors." She said, "What's the matter?" I said, "I think they may have put the wrong cells in my body.” And she's like, "What?!" And I'm like, I said, "I feel absolutely fine," I said. "There's nothing wrong with me."

Ken Shulman 21:01
Kathleen McDermott and Caron Jacobsen both told her there'd been no mistake. Those were her T cells in the bag, T cells reprogrammed to seek and destroy her B cell cancer. She'd be feeling them. And pretty soon.

Judy Wilkens 21:18
So, we went on to the next day. And I remember, you know, taking my temp and all of a sudden I said, “I'm not feeling good.” And the nurse came in and my temp was up over 100 and I was like, I was so excited because I'm like, oh my god, it's finally working. It's working.

Ken Shulman 21:33
It was working. And it was rough. Judy Wilkens spiked a high fever. Light hurt her eyes. She felt disoriented and struggled to answer questions during her daily cognition tests. It felt like a bad case of the flu. Despite that, every morning, she checked a day off the calendar she'd brought with her. She'd set a target date to go home. Nine days after the infusion. She knew it was optimistic. Jacobson had warned her that most people don't leave the hospital on their target date.

Judy Wilkens 22:08
I remember that my target date, I got up that morning and I took a shower and I put a pretty dress on and did my hair because I didn't, well, I didn't have hair, put my wig on and made some eyebrows. And I remember when the research doctor came in the room that morning, and I have my bags are all packed at the door anything. "What are you doing?" "I'm going home today," I said. " Whoa, I I don't know about that." And then later they, you know talked with Caron and I did go home on my target date. It was phenomenal.

Ken Shulman 22:38
Judy Wilkens went home on time. But she spent the next three weeks in bed, exhausted. Friends came over and lay on the bedspread beside her. She vaguely remembers watching the Summer Olympics from Rio, but she was home surrounded by friends and her parrots and cat. And each day she felt a little stronger. Four weeks after the infusion a friend drove her back to Boston for a PET scan. The scan was clear. Wilkens was in remission. Now she wanted to get back to work.

Judy Wilkens 23:14
So, I did the CAR T cell August 1. And the second week of September, I was actually standing behind my chair again cutting hair. I won't say for a full day. But I was there. My daughter would drive me every day to work so that I could at least be there, and it was incredible, incredible.

Ken Shulman 23:33
By Christmas, Wilkens was working almost full time. She'd gained back some of the weight she'd lost, her hair was growing in, every three months she went back to Dana-Farber for a scan and bloodwork, each time the exams came off the same. Judy Wilkens was cancer free.

Judy Wilkens 23:51
And the best way I described to my clients was, “It was like they put a Pac Man in my body. And the Pac Man would go through, and they would see a cancer cell, and um, and just eat that cancer cell instead of going through my body and destroying all my good cells.”

Ken Shulman 24:10
Wilkins’ Pac Man, the CAR T cells, had a healthy appetite in the Zuma one study. Of the 101 subjects in the clinical trial, 84% responded to the treatment. 59% showed a complete response. The numbers were even better in a follow-up trial. In that study, CAR T cells were given to patients with refractory mantle cell lymphoma. It's an incurable cancer that usually leaves patients with only a few months to live. One year later, 83% of those patients were still alive, and most of them are still in remission. The FDA has approved CAR T cell therapy for several forms of leukemia and lymphoma that don't respond to standard treatment. Trials are underway for a wide array of other cancers, including brain cancer. And researchers are working towards ways to make CAR T cells safer, more specific and less toxic. They'd like to program CAR T cells to target two or three proteins instead of just one, to insert a kill switch into the genetic code to turn off the T cells once they've completed their mission. And in the not-too-distant future, we could even see bespoke CAR T cells, T cells programmed with DNA sequence directly from the patient's tumor. These would seek out the patient's cancer cells and nothing else. For the moment, CAR T cell therapy can be a bumpy ride. Nearly half the subjects in the Zuma One study reported serious adverse effects, although none was fatal. It's these adverse effects that give doctors like Caron Jacobsen pause when prescribing CAR T cells. That and the fact that in most patients, standard treatments work quite well.

Caron Jacobsen, MD 26:10
Our traditional chemotherapy for lymphoma, although it is usually five and a half or six months of therapy, is generally well tolerated and the mortality risk is very, very low. And the long-term toxicities are actually pretty low as well.

Ken Shulman 26:24
There are some obvious advantages to CAR T cells over traditional chemotherapy. The window for potential side effects is far more narrow, three or four weeks compared to five or six months for chemo. But the stakes are also much higher.

Caron Jacobsen, MD 26:40
Of course, anybody would prefer a two-week period of toxicity over a six-month period of toxicity. But you don't want to do that when there may be an increased risk of dying as a complication. So, I think CAR T cells need to get a little bit safer to be a formidable competitor to some of our frontline therapies, at least in lymphoma.

Ken Shulman 26:59
CAR T cells and stem cell transplants are for cases where traditional therapies fail. Both come with warning labels. But CAR T cell therapy has a couple of advantages over transplants. The patient doesn't need to be in remission to receive the therapy. And the patient doesn't have to live in a bubble afterwards. Judy Wilkens would not been lying on her bed with friends at home, nine days after a stem cell transplant, or even nine months.

Caron Jacobsen, MD 27:31
This was not just lifesaving for Judy, because it literally saved her life and cured her lymphoma. But it was lifesaving because it meant that she didn't lose her livelihood and her vitality. So, it's really a remarkable story. And Judy, as I'm sure you know from having spoken to her, is just a remarkable lady.

Ken Shulman 27:50
Judy Wilkens is back at work at her salon, full time. Her parrots and cat are fine. Since her recovery, she's become the number one advocate for CAR T cell therapy. She meets with drug companies and oncology groups to stump for the Pac Man that saved her life. She speaks at medical conventions. She meets with prospective CAR T cell patients as well. She knows she's very fortunate.

Judy Wilkens 28:17
I'm back to 100%. I mean, you know, I'm, I'm not doing... well, I am a little bit older now. But work is great. My life is good. I have a couple of grandkids who I thought I may not ever see them, or you know, watch them grow up. So, for me, it's just God's blessing that I'm here to see these children. Because at one point I didn't think I would be.

Ken Shulman 28:41
This summer, if all goes well, Judy Wilkens will celebrate her fifth year of being cancer free. She visits Jacobson twice a year at Dana- Farber for checkups and a scan. The last time they met, they were both wearing masks because of COVID.

Judy Wilkens 28:58
Last time I saw Caron was just a few weeks ago and we can't hug anymore. You know, we used to dance in the - don't tell her that - we used to dance in the room and just like you know, so proud of me. But you know, with the COVID, you have masks and all that and I remember walking out of our examining room and you know she looked back at me and she said, "Judy Jetson, we're doing great."

Ken Shulman 29:25
This past year, a year like no other, has been particularly challenging at Dana-Farber.

Kevin Haigis, PhD 29:31
We get our experiments done. But we're not swapping ideas with people anymore. We're not sitting around the lunch table talking about why my experiment didn't work and how you can help me because your experiment did work. These are the things that make academic science productive. It's your ability to brainstorm and to connect with other scientists.

Ken Shulman 29:52
In our next episode, we'll learn how the Institute was affected by COVID-19 and how the staff found a way to persevere and contribute. I'm Ken Shulman and this is Unraveled, a Dana-Farber Cancer Institute podcast.

Season 1, Episode 6: Dana-Farber in the Time of COVID

A beauty… that in some cases becomes tragic. It's not the way most people would describe a virus. In the simplest of terms, a virus is a snippet of genetic code. It can't reproduce on its own. In order to replicate, it needs to infect and hijack a living cell.

Scientists can't even agree on whether viruses are alive or dead. What they do know is that every so often, one of them goes, well, viral. And when it does, it can bring everything to a skidding halt.

Well, almost everything. One thing the COVID-19 pandemic didn't stop was cancer. And that meant that everyone at Dana-Farber had to find a way to keep on working and find a way to keep themselves and their patients safe.

Season 1, Episode 6 Transcript

Joseph Sodroski, MD 00:07
There is a beauty there. It's a tragic beauty in some cases. But nonetheless, we have to, I think have a respect, and an awe for what nature can accomplish.

Ken Shulman 00:20
A beauty that in some cases becomes tragic. It's not the way most people would describe a virus. In the simplest of terms, a virus is a snippet of genetic code. It can't reproduce on its own. So, in order to replicate, it needs to infect and hijack a living cell. Scientists can't even agree whether viruses are alive or dead. What they do know is that every so often, one of them goes well, viral. And when it does, this zombie genetic charm bracelet, this relic from a time when life was just emerging on Earth, can bring everything we've built in the meantime, to a skidding halt. Well, almost everything. One thing the COVID pandemic didn't stop was cancer. And that meant that everyone at Dana-Farber had to find a way to keep on working, and they had to find a way to keep themselves and their patients safe. I'm Ken Shulman, and this is Unraveled, a Dana-Farber Cancer Institute podcast. It's been quite a year, a year like no other, at least for those of us born after the 1918 Spanish flu epidemic. It's not that there haven't been viral outbreaks since then. We've seen flare ups of Ebola and Zika. A couple of aggressive flu strains killed several million people in the 50s and 60s. And HIV, the virus that causes AIDS, has taken more than 35 million lives worldwide. But none of these viruses shut down the world the way that SARS Coronavirus has.

Laurie Glimcher, MD 02:11
Nobody expected this. But, you know, thinking retrospectively, it's clear that this was going to happen at some point that there were going to be viruses that we didn't know anything about, and they were going to be transmitted from animals to humans.

Ken Shulman 02:31
That's Laurie Glimcher. She studies the human immune system. Since 2016, she's been president and CEO of Dana-Farber. While she and her colleagues may not have expected SARS Coronavirus, or the COVID epidemic it caused, they were ready for it. They had to be.

Laurie Glimcher, MD 02:50
We had to be vigilant. We had to be ready to help the patients who needed us during a time that was extremely frightening for them, because we have to protect our cancer patients who are immunosuppressed. And we are very familiar with and comfortable with the safety techniques one needs to put into place to keep them safe.

Ken Shulman 03:13
The COVID pandemic arrived in the US in early 2020. Everyone scrambled to adapt and adjust. Businesses in school shutdown, basic goods and necessities suddenly became scarce. More than a few people panicked. But a cancer treatment and research center can't afford to panic.

Laurie Glimcher, MD 03:32
Well, we had to react very quickly because as you know, cancer patients have a suppressed immune system. So COVID poses a special risk to many cancer patients. And we had to very quickly implement changes to ensure that we could safely care for our patients and continue to treat our patients because cancer does not pause just because there's a pandemic out there.

Ken Shulman 03:57
As a first step, Glimcher formed an emergency response team. They made decisions, some of them difficult. They drafted strategy, because they had to stay open. And they had to protect their patients. Glimcher said it was almost like working in wartime.

Laurie Glimcher, MD 04:15
My younger son was in the Marine Corps for four years in Afghanistan and in the Special Forces. So, I had a sense from him of what it's like to be in the middle of a war. I think it's not so different as well from being a leader of a wonderful institution like Dana-Farber, because the leaders have to be calm. They have to be reassuring. And they have to be extremely efficient at harnessing all that was needed to deal with a pandemic like this.

Ken Shulman 04:57
With a pandemic spiking, the Dana- Farber team dug deep to keep the institute up and safely running. They reinforced and expanded existing patient safety protocols. They hustled and scrambled to find more protective equipment. And they succeeded. Patients got their care, including the 20% of Dana-Farber patients involved in clinical trials. It wasn't business as usual. COVID changed that. The team altered many treatment plans. They reduced in-person visits to an absolute minimum. Telemedicine surged from 20 appointments a week in pre-pandemic times, to over 400 appointments a day. Doctors changed drug regimens as well. Whenever possible, they shifted chemotherapy patients from intravenous to oral medications, medications they could take at home. All these adjustments reduced human contact and helped reduce the chance of infection and transmission among patients and staff. Adapting to COVID conditions took a lot of work and concentration. But it also provided a psychological lift during a very difficult time.

Laurie Glimcher, MD 06:10
Understandably, there was fear but that fear did not just fester, it was translated into keeping themselves safe and our patients safe. That was the focus.

Ken Shulman 06:23
Of course, crisis leadership involves clear vision and timely decisions. But there's more. In a hospital, it also involves taking care of your patients and taking care of the people who take care of those patients.

Laurie Glimcher, MD 06:37
Of course, there were stress and we paid a lot of attention to emotional burnout, to stress, particularly in our patient-facing health care faculty and staff.

Ken Shulman 06:51
Part of that attention involve providing a steady flow of information. Before the pandemic, Dana-Farber staff convened about twice a year in the Jimmy fund Auditorium for a town hall seminar series. With the pandemic, they met weekly, every Wednesday, at noon, the entire Dana-Farber community could log on to the forum to find updates on almost everything COVID, local and national infection rates, progress on vaccines, research into drugs that could help alleviate symptoms caused by the virus. Medical staff from other institutions also logged on.

Laurie Glimcher, MD 07:31
I think it's always important to have the data, to have the information. And that was I think the point of the open forum was: Here's where we are; here's what we're doing about it; and we're here for you. We are here for you. We are here for each other. And we are here for our patients. And you know, if you keep as your central guidepost, taking the best care possible of our patients and each other, then I think you can reduce some of the stress.

Ken Shulman 08:07
There was a lot of stress to reduce, among patients and among doctors and nurses who treated patients.

Paul Richardson, MD 08:13
This is a catastrophic pandemic with a novel virus that is a basically behaving outside of the boundaries of nature. And the reality is that it's transformed the way we view our current healthcare environment and the risks to our patients.

Ken Shulman 08:28
That's Dana-Farber's Paul Richardson. He directs clinical research at the Jerome Lipper Multiple Myeloma Center. Multiple myeloma is a blood cancer. It attacks our plasma cells. These are late-stage white blood cells that secrete powerful antibodies. Plasma cells are a vital element in our immune system. They fight off all manner of threats, bacteria, fungi, and of course, viruses. Multiple myeloma hits patients with a double whammy. It's a highly aggressive cancer and can easily spread through the body, and it knocks out our plasma cells. So, our defenses lag.

Paul Richardson, MD 09:08
And so we see a constellation of risk factors for our patients that make them particularly vulnerable to the SARS COV-2, I should say, and then by virtue of that complications of COVID-19. And in fact, surveys in myeloma patients across the world have shown mortality rates for hospitalized myeloma patients infected with SARS COV-2 to be as high as 60%.

Ken Shulman 09:33
Richardson works both as a clinician and researcher. He's accustomed to seeing critically ill patients, including patients whose immune systems have turned against them following a stem cell transplant. But he says what he saw during COVID was different.

Paul Richardson, MD 09:49
Seeing this disease in its worst form, and seeing the incredible commitment of the nurses, the physicians, and the teams in the ICU's caring for the patients. I mean, we've treated patients in the ICU setting, they're the sickest of the sick. I've never seen actually such incredibly ill patients, even in my transplant experience. And to see the commitment that we have to our patients is just unbelievable. It's been a privilege.

Ken Shulman 10:17
Although they'd never dealt with this specific virus before, Richardson and his team knew what they needed to do. First, protect their patients from becoming infected. And then, once their patients were safe, to repurpose their research to help in the fight against COVID. Protect the patients. That meant shielding them from the virus that causes COVID. Doctors like Richardson know how to do that. They're used to protecting patients whose immune systems are down. But repurpose? How could the lab leverage its work in multiple myeloma to help counter the effects of the pandemic? One thing that lab could leverage was success. The Multiple Myeloma Center is one of Dana-Farber's crown jewels. Richardson and his colleagues have led the way in developing almost all the drugs doctors use today to treat multiple myeloma. The results have been stunning. In 2003, a patient diagnosed with multiple myeloma could hope to live for three years, maybe five. Today, the median survival rate can be 10 to 15 years.

Paul Richardson, MD 11:27
Dana-Farber has been right in the thick of it. And I would think it's fair to say there probably is no comparable track record for another translational group, certainly in the United States, probably globally as well for that matter.

Ken Shulman 11:40
One of the drugs developed at Dana-Farber is called Defibrotide. It's used in stem cell transplants. Transplants are used for many different types of blood cancer, including myeloma. It's a complete replacement of the bone marrow. Transplants can be very effective, but they can also trigger some pretty nasty side effects as the body integrates its new immune system. One of the side effects is damaged to the endothelium. The endothelium forms the inner lining of all our blood vessels and lymph nodes. Endothelial cells control a number of vital functions, including blood vessel integrity, and blood clotting. When these functions are compromised, patients suffer. Some of them even die. Defibrotide helps protect endothelial cells and reverse endothelial damage after the transplant. But stem cell transplants aren't the only thing that can damage the endothelium. As the pandemic spread, doctors began to see endothelial damage in severe COVID cases. That damage and the clots the damage caused, were factors in many COVID deaths.

Paul Richardson, MD 12:54
When we recognized this phenomenon of endothelial damage being such an important part of the pathobiology of COVID-19 launched a comprehensive international project to use a drug we've developed at Dana-Farber called Defibrotide. to specifically target this endothelial complication.

Ken Shulman 13:13
Working with domestic and international partners, the Dana-Farber team coordinated clinical trials for Defibrotide in patients with severe COVID symptoms The early results are promising. Defibrotide appears to be both safe and potentially effective in limiting and perhaps even reversing endothelial damage in COVID patients. Naturally, Richardson reminds us that these are early results.

Paul Richardson, MD 13:39
COVID-19 is such a complex disease and such a complex process, especially in its advanced stages, making the interpretation of data really quite challenging. But having said that, we have very compelling preclinical data. We have an FDA approved drug that's used for the blood vessel injury syndromes of transplant, and does so highly successful in this life saving, and now bringing that same platform with a relatively favorable tolerability profile to SARS COV-2 infection, augmenting and complementing existing treatment strategies, we're very hopeful that will make a real contribution.

Ken Shulman 14:16
All hospitals had to grapple with the fallout of the COVID pandemic. All had to adapt, and quickly. The Dana-Farber team was able to sustain patient care and almost all ongoing clinical trials, even as infection numbers spiked last March. But clinical care is only one side of the equation. There was also the research side. For that side, the leadership team drafted a different strategy.

Kevin Haigis, PhD 14:43
We shut down completely the laboratory operations at our institution. And every institution across the country, as far as I know, shut down completely for six or eight weeks.

Ken Shulman 14:53
That's Kevin Haigis. He took over as Chief Research Officer at Dana-Farber on January 1, 2020 Just in time, he jokes now, to oversee preparations for the pandemic. Haigis and the leadership team elected to shut down research until they got a better handle on the virus and the risk it posed. And until they could secure enough protective equipment to guarantee staff safety.

Kevin Haigis, PhD 15:19
Shutting down was actually not really difficult. Tell people don't come in and we gave them a bit of a runway, we said, you know, you have a week to shut off your current experiments. Make sure all your reagents are frozen down, make sure all your incubators are turned off. But then by whatever Friday, nobody’s allowed to come in anymore. And, you know, stay tuned, we will tell you when you're allowed to come back.

Ken Shulman 15:51
The shutdown lasted from mid-March until mid-May. But that doesn't mean research stopped at Dana-Farber. It shifted, just like patient care. And Dana-Farber scientists persevered with the same sense of urgency as their clinical partners. Cancer is a tenacious foe, and the scientists who fight it are just as hard boiled. They stayed busy working remotely. Researchers working in wet labs, those labs with cells and drugs and DNA, they had to put their experiments on hold. But all other research continued -- computational models, simulations and analyses. Researchers with time on their hands seize the opportunity to learn new skills, especially in bioinformatics, the rapidly growing field where big data is applied to medicine. Others wrote important papers they hadn't been able to make time for. And there was a notable upsurge in grant applications. And everyone stayed in touch.

Kevin Haigis, PhD 16:54
I think that scientists are often obsessed with their work. And they're so obsessed that if they can't come to work, they find ways to do work. And I actually was very proud of the response of, of our research population to adapt that way.

Ken Shulman 17:10
Dana-Farber reopened its research facilities on May 18, 2020. It was among the first Boston institutes to do. Scientists returned to their labs, but in limited numbers. They worked in shifts and stayed at least six feet apart. They wore masks and sometimes gowns. And they didn't linger as they once did. There were few casual contacts or conversations.

Kevin Haigis, PhD 17:35
We get our experiments done. But we're not swapping ideas with people anymore. We're not sitting around the lunch table talking about why my experiment didn't work, and how you can help me because your experiment did work. These are the things that make academic science productive. It's your ability to brainstorm and to connect with other scientists. COVID has completely obliterated that from the academic research environment. You know, we see each other on Zoom now. But these kinds of interactions, as I'm sure you know, they don't work well on Zoom, you don't sit around and shoot the breeze with people on Zoom, because you're on Zoom all day, and you want to be off as soon as you can. And I think that this, if I had to pick one thing, this is going to be a lasting effect of the pandemic. It's the thing that we really missed the most.

Ken Shulman 18:27
Still research move forward, even without the water cooler conversations. And even when COVID infection spiked again, in August 2020.

Kevin Haigis, PhD 18:36
We actually did not even think about shutting down our research operations when the second spike occurred. And that is because we were very confident that we knew how to manage our risk of spreading the infection.

Ken Shulman 18:49
And Haigis doubts they'll have to close their doors again anytime soon. Or even in the future.

Kevin Haigis, PhD 18:54
There's almost certainly another pandemic out there. But I think that we know what we need to do to protect our staff and to allow cancer research to remain ongoing.

Joseph Sodroski, MD 19:06
So, I don't think it's an absolute surprise to me that viruses can become this deadly and become globally prevalent.

Ken Shulman 19:17
That's Joseph Sodroski. He's a virologist on staff at Dana-Farber since 1981.

Joseph Sodroski, MD 19:24
But I think that it is a surprise when any given virus achieves that goal. And the rapidity of the SARS COV-2 pandemic and the deadliness of the virus in just a little over a year of time, I think was unexpected and I don't think anyone could have precisely predicted this.

Ken Shulman 19:53
If you're like me, you may be asking why a Cancer Research Institute has a virologist on staff. It is true that certain cancers are caused by viruses. Hepatitis B and C viruses, for example, can lead to liver cancer. Adult T cell leukemia is caused by a virus. And the human papilloma virus is tied to cervical cancer. But Sodroski says there's more. He says viruses have a lot in common with cancer cells, both in what they want to do, and how they do it.

Joseph Sodroski, MD 20:26
What every virus wants to do is to make more of itself, and then to spread to new hosts. There's a strong analogy between viruses and how they desire, if you will, to replicate and make more of themselves. And the same kind of trend in a cancer cell to replicate itself and to make more of itself. And there is a Darwinian selection that occurs so that the viruses that are more successful, the cancer cells that are more successful at achieving that level of replication, tend to be the ones that survive.

Ken Shulman 21:06
Sodroski has spent decades studying HIV. Specifically, he studies how the HIV virus enters cells to then deliver its genetic payload. Studies of HIV have yielded life-saving drugs to patients suffering from AIDS. Sodroski's work has also helped Dana-Farber scientists develop some of the first viral vectors. These are deactivated viral shells that doctors use as vehicles to deliver vaccines. The vaccine for Ebola, for example, uses a viral vector. So do several COVID-19 vaccines. Work in HIV also helped scientists who were sprinting to develop a vaccine for COVID-19. How? Well, for starters, HIV and SARS Coronavirus, also have a lot in common. HIV is what's called an envelope virus. It's like the virus is wearing a coat.

Joseph Sodroski, MD 22:04
On the surface of that coat are viral spikes. And those spike proteins are what the virus uses to attach to target cells. The spike proteins allow the virus to enter the target cell and get its genome--its genetic information--into the interior of the cell and to begin replicating.

Ken Shulman 22:28
But those spike proteins are also an Achilles heel. Yes, first and foremost, the spikes are weapons, biochemical grappling hooks that help the HIV troops breach the cell wall. But they're also targets, targets for antibodies that can stem and even stop the progression of the virus.

Joseph Sodroski, MD 22:48
Now that we're seeing SARS Coronavirus, it also has spike proteins. Those spike proteins kill the infected cell. And those spike proteins are very good targets for generating antibodies that can block the virus and you know, the spike proteins are the key component of all of the vaccines that are now getting into the clinic for SARS COV-2.

Ken Shulman 23:16
Because scientists had a strong foundation in spike proteins and in other viral knowledge, they were able to develop a vaccine for SARS Coronavirus in record time. Sodroski says there's an important lesson here.

Joseph Sodroski, MD 23:30
I think that the one thing that I've learned as a scientist over the years is that there are many aspects of knowledge that are valuable, and they aren't immediately practical. Sometimes decades later, one sees how important they are and how much of a practical impact they have on our ability to either prevent or treat disease. And what we learned about a human T cell leukemia virus, studying a cancer-causing virus, turned out to be highly relevant when we turn to HIV and the AIDS causing viruses. And what we learned there turned out to be useful in understanding SARS Coronavirus.

Ken Shulman 24:23
Like every crisis, the COVID-19 pandemic has presented challenges and opportunities. There have been many victims. There have been many lessons, some of them hard, some of them golden, and some of them surprising. And there will be consequences, many of them even after the pandemic is tamed. For Dana-Farber, one of these consequences is an expected surge in cancer deaths. President and CEO Laurie Glimcher says people all over the country postponed or canceled routine screenings during the pandemic.

Laurie Glimcher, MD 24:59
The number of people who had colonoscopies, mammograms, PAP smears decreased by about 90%. So, catching cancer at stage one becomes a lot more difficult when you don't have these screenings that pick up early cancers. And it's very dispiriting. The director of the National Cancer Institute, Ned Sharpless has predicted that there'll be a 10% increase in mortality from cancer because of the absence of screening.

Ken Shulman 25:36
Most people don't come to Dana-Farber for routine cancer screenings, they go to primary care. The majority of Dana-Farber, patients, as many as 75%, come to the institute with advanced cases of cancer, cancer that has already spread and is much harder to treat.

Laurie Glimcher, MD 25:54
So you can imagine that if you have a 90% reduction in early detection, by screening, you're going to have an increase in patients that come in, who have cancer that's already spread. And then are hence more likely to die from their cancer.

Ken Shulman 26:11
Reflecting over the past year, she's pleased that how Dana-Farber responded to the crisis, how it continued to treat patients and conduct clinical trials. She's equally pleased the Institute forged ahead with research. The mission, she says, isn't just to treat this year's patients, it's also to treat the patients who may arrive 10 or 20 years down the road. That's why it was imperative to reopen the labs as soon as possible.

Laurie Glimcher, MD 26:38
Because this is an amazing time in cancer research. As you know, the last two decades have been transformative. We are able to treat cancers and in some cases cure cancers that previously were lethal, by using targeted genetics, and by immunotherapy. And I was absolutely determined to not let the pandemic slow our progress towards new discoveries.

Ken Shulman 27:06
Along with testing Dana-Farber's skill and stamina, the COVID epidemic also tested its mettle. Glimcher is particularly proud of the outcome in the clinic and in the lab.

Laurie Glimcher, MD Science takes a long time. Research is something that, you know, most of your experiments fail to begin with, you have to persevere. You have to be passionate and be dedicated, and not be disturbed by failure. You're going to fail a good part of the time as a researcher, but that doesn't mean that you, your energy and your dedication is diminished. You just have to keep at it. You learn that you can never give up and be prepared to fail. And just to keep at it until you succeed, because we have to do something about these cancers that are particularly intransigent, like pancreatic cancer, or glioblastoma, or ovarian cancer. And we're not going to give up. And Dana-Farber, we don't give up.

Ken Shulman 28:16
We hope you've enjoyed this window into the workings of Dana-Farber. I know I have, it's been a fascinating journey, learning about cancers that show up in disguise, and about the drugs that unmask them. Listening in as world class researchers decode the intricate dialect that cells speak with each other. Watching as doctors reset the human immune system to dismantle tumors. And most of all, witnessing the extraordinary commitment that Dana-Farber researchers and doctors and staff show every day. If you want to learn more about research at Dana-Farber, check out the Insight blog at blog.dana-farber.org/insight. I'm Ken Shulman and this is Unraveled, a Dana-Farber Cancer Institute podcast.

Dana-Farber's Cancer Podcasts

Dana-Farber's cancer podcasts cover a wide range of topics, including new research, the latest treatments, and survivorship.

Explore all Podcasts

Season 2

Season 2 Trailer

Season 2, Episode 1: BCL-2 Inhibitors: Driving Cells to Destruction

Season 2, Episode 2: Thalidomide and its Second Act in Multiple Myeloma

Season 2, Episode 3: Stop the Division: CDK-4/6 Inhibitors and the Cell Cycle

Season 2, Episode 4: Making it Personal: Targeted Cancer Therapies

Season 2, Episode 5: With a Little Help From My Friends: Combination Immunotherapy

Season 2, Episode 6: Big Data, AI, and Cancer Research

Season 1

Season 1 Trailer

Season 1, Episode 1: The Wolf in Sheep's Clothing

Season 1, Episode 2: Behind the Science of the 2019 Nobel Prize

Season 1, Episode 3: A Culture of Mentorship

Season 1, Episode 4: The Blueprints of Your Cells

Season 1, Episode 5: Turning Science Fiction into Fact

Season 1, Episode 6: Dana-Farber in the Time of COVID

Dana-Farber's Cancer Podcasts

Stay in touch with us

About Dana-Farber

Locations

Discovery and Insights

Care and Treatment

Health Library

Why Choose Dana-Farber

Unraveled

Season 2

Season 2 Trailer

Season 2, Episode 1: BCL-2 Inhibitors: Driving Cells to Destruction

Season 2, Episode 2: Thalidomide and its Second Act in Multiple Myeloma

Season 2, Episode 3: Stop the Division: CDK-4/6 Inhibitors and the Cell Cycle

Season 2, Episode 4: Making it Personal: Targeted Cancer Therapies

Season 2, Episode 5: With a Little Help From My Friends: Combination Immunotherapy

Season 2, Episode 6: Big Data, AI, and Cancer Research

Season 1

Season 1 Trailer

Season 1, Episode 1: The Wolf in Sheep's Clothing

Season 1, Episode 2: Behind the Science of the 2019 Nobel Prize

Season 1, Episode 3: A Culture of Mentorship

Season 1, Episode 4: The Blueprints of Your Cells

Season 1, Episode 5: Turning Science Fiction into Fact

Season 1, Episode 6: Dana-Farber in the Time of COVID

Dana-Farber's Cancer Podcasts

About Dana-Farber

Locations

Discovery and Insights