19: Bret Weinstein - The Prediction and the DISC: Difference between revisions

'''EAric:''' Yeah.

'''Eric:''' Okay. I want you to talk to me about that story, and because I lived with you, I know that it happened, and I know that it got buried, and I know that it's part of what I'm calling the Distributed Idea Suppression Complex because, quite frankly, you were not the only person who was a part of the story, and the story had to die because it said something, which is that the power of your theory was sufficient to predict, from first principles, a manifestly observed and surprising result, within molecular biology, from pure evolutionary principles.

'''Bret:''' Evolutionary biology has long been biased in the direction of abstraction. Rather than thinking about mechanism, that is to say we deal in the phenomenology of things. We talk about gross patterns that we see in nature rather than talking about the fine detail of what drives them. That has been changing in recent decades, but it has a long history, and it comes from a very mundane place. That mundane place is that we just haven't had the tools to look, for example, inside of cells and we haven't been able to read genomes. You know, we could have been able to read a gene here and there at great expense, but the ability to peer into genomes is pretty new. The ability to peer into these molecular pathways is pretty new. So anyway, there's a historical bias in evolutionary biology against mechanism and in the direction of phenomenology. I have never been particularly fond of that bias. I have always been interested in [https://en.wikipedia.org/wiki/Mechanism_(biology) mechanism]. I'm interested in the phenomenology too, but I've always kept my foot in the door with respect to mechanism. And as an undergraduate, I took lots of mechanism classes. I took a development class at the time, developmental biology was in my opinion, a bit stuck. It is now unstuck in a very dramatic way. But anyway, I took a developmental biology class. I took some or immunobiology. And anyway, I was armed with these things in an environment in evolutionary biology where most people were not, most people were in the phenomenology. And one day I happened to be in a seminar. Dick Alexander was running a seminar for graduate students, and a student was there who was very out of place. He was studying cancer, and he, on a lark, decided to take an evolution seminar that looked good to him in the catalog, and it wasn't right for him. And he gave a talk at some point, and his talk was on his work with cancer and frankly, because all the other people in the room were evolutionarily oriented, nobody was really tracking what he was saying. But what he said struck me like a bolt of lightning. He said that in the realm of cancer research, people were looking at telomeres, which are these repetitive sequences at the ends of chromosomes. And they were toying with the possibility that the fact that these telomeres shorten every time a cell divides, that that is providing a resistance to tumor formation. Very straightforward—counter counts down, and that would prevent—

'''Bret:''' Yup. It was the discovery of Leonard Hayflick, who basically overturned the prior wisdom about cells, which was that they would grow indefinitely as long as you kept feeding them and making an environment that was conducive to division. So I don't exactly know why that result had been misunderstood at first. Maybe somebody had a cancerous cell line and so they got the wrong idea and it just spread, but Hayflick checked it and it turned out to be false. It turned out there was a number of cell divisions that healthy cells would go through, and then they'd stop. The mechanism was not obvious to Hayflick, but later it became clearer and clearer that the mechanism was these sequences at the ends of chromosomes which shorten each time the cell divides. And the implication was that, potentially, this was a cause of what we call [https://en.wikipedia.org/wiki/Senescence “senescence”]. What in common parlance would often be called “aging”, the tendency to grow feeble and inefficient with age. If your cells are each in a cell line and that line has a fixed number of times that it can replace itself before it has to stop, then some point your repair program starts to fail. And that repair program, failing across the body, looks like what you would expect aging—aging follows the pattern you would expect if cell lines one-by-one stopped being able to replace themselves. So—  

'''Bret:''' And the [https://en.wikipedia.org/wiki/Germline germline] can't because if it did, your lineage would go extinct as a result of small—  

'''Bret:''' Fitness enhancing traits at some costs late in life, then they will tend to be accumulated by selection. And the reason for that is because, well, there are two ways to think of it, really. If a negative trait occurs very late in life, then a large number of individuals who have the gene for that trait will not live long enough to experience the harm. So if it came bound to a positive thing early in life and you're dead before the late life harm accrues, you got away with it. Right? So William's point was, he was building on earlier work of Medawar, but let's skip that for the moment.

His point was, because of tradeoffs, you will have lots of traits that are good early and bad late. Selection sees the early traits much more clearly than it sees the late traits, and it prioritizes them because of the discounting that arises because so many individuals aren't around to experience the late-life harm, and if they are around experienced the late-life harm, a lot of their reproduction is behind them anyway. So they count less. Selection counts more early in life. And this timer starts at the moment of first reproduction, the usual moment of first reproduction for your species. So this was a beautiful hypothesis, and it was beautifully articulated with many predictions, which is the way really good work is done. And we knew, at the point that I was entering graduate school, we knew that the hypothesis was right. It was a theory.

'''Bret:''' The Antagonistic Pleiotropy Hypothesis for senescence. We knew that it was right because it predicted so many phenomenon in nature that we could readily go out and measure. And this is again where the phenomenology versus mechanism comes out.  

'''Bret:''' We know that creatures that are poisonous or have a shell that protects them or can fly away from danger, are disproportionately long-lived for their size. Small creatures tend to live shorter lives than large creatures. But if you can fly, then you're off the line of the other creatures of your size. So for example, their small bats who have been recovered after 30 years in the wild. So creatures that have special protections have disproportionate longevity. This matches William's hypothesis, because it is their ability to fly away from danger that makes the likelihood of their experiencing late-life costs go up.  

'''Bret:''' So selection sees their late life more easily than it sees a small Creek.

'''Eric:''' I just want to say something. This is a podcast. It's an unusual podcast and we can talk science and I'm thrilled, but we always have our colleagues in our minds when we're talking to a general audience and the colleagues are always in a “gotcha” mode. Well, you forgot about this. You didn't mention that. I'm even interjecting little bits because I want to make sure that you're immunized from all the bullshit that the academics, so I just want to make a general statement, which is we can come back and get into any level of specificity that somebody wants to, if they want to take you down, I don't care. What I'd love to do is to tell the story with enough punch that people understand what happens.  

'''Eric:''' Does that mean, to be a gene, it has to be protein encoding?  

'''Bret:''' Yeah. Anyway, I knew this assertively, I was well familiar with William's paper. At the point that I saw this talk on cancer and I knew already about the question of senescence, everything came together. This was obviously the answer, where the missing pleiotropy was. Well, the missing pleiotropy had to do with a telomere, which wasn't exactly a gene. It was genetic, it was DNA, but it wasn't a gene, but it was perfectly capable of producing exactly the effects that we see in senescence across the body, tissue—

'''Eric:''' So I'm not even sure that you fully said it. I want to make sure that I'm even clear on it and I'm going to, I think I'm right, but correct me if I'm wrong. What you're saying is, “What if the Hayflick limit is a protection against dying from immortality at a cytological level”, that some cell gets a dream of immortality that it shouldn't have because, let's say, it's a somatic cell, and it says, “Okay, I just want to keep dividing and dividing and dividing”. Nature knows how to do this, and that immortality, which sounds good at first, is actually called cancer. And so in computer science we would say, okay, you've introduced a recursion limit into a while loop or a for loop to make sure that you don't have a resource leak, which is what a tumor is.

'''Bret:''' Yeah, so let me say it this way. If you have a damage to a tissue cut on your arm or something, the cells on both sides of that cut suddenly become aware that there is a problem, a gap, because the can't hear a neighbor on one side of them and their natural reaction is to start growing into the gap until they can hear a neighbor which is the sign to stop. If you imagine that something like that is occurring in every tissue, or almost every tissue, the problem is that that means that every tissue in your body for which that story is about right, is in danger of having damage from radiation or whatever, turn it deaf to its neighbors. A single cell that has turned deaf to its neighbors will suddenly start replicating, and if it is deaf to its neighbors, then there's no message that it's going to hear that's going to tell it to stop. So that thing, imagine any cell in your body just taking off and growing and growing and growing—

'''Eric:''' Okay, this is terrifying. What you're saying to me is, is that if I'm comprise of let's say 30 trillion cells and I view them as each let's say subroutines, any subroutine that is not denucleated, right? Like this wouldn't happen in the in the lens of your eye because the nucleus has been removed, but any other reasonable cell is potentially your assassin, because it's mitosis process might completely go rogue.

'''Bret:''' And so the rather elegant and very simple idea is that there would be a hard limit so that any cell that had become damaged, so it started down this path would just simply run into the number of cell divisions it was allowed in a lifetime and it would stop.  

'''Eric:''' So like, the moles on my face that some of my less couth commenters loved to talk about—

'''Eric:''' Are effectively attempts to kill me that may have stopped. And that the perimeter where they stop is where the Hayflick limit took over and said, “This cell line must die so that the patient will live”?

'''Bret:''' Yeah. The name I gave him was “prototumor” and the idea is a prototumor is a patch of cells arrested at their Hayflick limit. Because they had become unregulated. If you go to the dermatologist and you say, what do I look for? You know, they tell you certain things to look for. So a round patch of cells that suddenly becomes irregular in shape. Well that's what would happen if you took one of those cells and gave it a second mutation and it started growing again.  

'''Bret:''' Is adaptive to protect you from cancer—

'''Eric:''' K, so there's a little bit of a mind bender because what you're telling me is that I've got to avoid immortality, which can kill me, and that the solution to not dying is death.  

'''Eric:''' It's a very elegant thing. And now the problem is, is that there's all this weird attended complexity that you had to deal with.  

'''Eric:''' So it was like stem cells versus germ versus ...

'''Bret:''' Yes, maybe even literally on occasion. But the question was, I began to wonder if there was something wrong with the idea that mice had long telomeres. Sometimes, like in Hayflick's case it turned out that a bunch of people were copying some wrong result, so it seemed like a lot of people had seen it, but only one had. And I checked, was it true, that there was some, that everybody was parroting one study that said mice had long telomeres?  

'''Bret:''' It turns out lots of people had tested it. Mice have long telomeres like 10 times the length of human telomeres. It just didn't fit. So finally, it occurred to me that it was possible that what was going on—I discovered something in trying to figure out what they meant by “mice”. Right? There's a lot of species of mice, but all the mice that we use in the lab, with rare exception, are from one genus, and often from a particular target species.  

'''Bret:''' Mus musculus, which is the common one. What shocked me was that it turned out all the mus musculus that were being used in labs across the country, and in many cases, farther afield than that were coming from one place, which I had no idea. There was one—

'''Bret:''' The [https://en.wikipedia.org/wiki/Jackson_Laboratory JAX Lab] in Bar Harbor Maine, right? They seemed to be the source of everybody's mice. And so it began to be—it was a possibility I could not shut down in my mind, that there was something about what was going on at the JAX Lab that had resulted in the mice that were being sent out to all these other labs—

'''Eric:''' Is it that they were representative animals—

'''Bret:''' Right, these are a model organism. People were just using mice because mice were a convenient mammal, but they're all coming from one place, and it began to occur to me that that one place was not just a source of mice in the sense that we might think it, it was actually a selective environment that was impacting those mice. And when I dug deeper, it turned out that the mice had all, they were descendants of a long lineage that had lived in captivity under conditions at the [https://en.wikipedia.org/wiki/Jackson_Laboratory JAX Lab]. And at some point I realized that the most likely thing going on was that there was something about this environment that had wildly elongated the telomeres of these mice. And that was simultaneously an unbelievable idea, but the only one I could think of that made sense of everything I had seen. And so—

'''Eric:''' Well, it's unbelievable because the consequences, I mean, look, I have not even heard whether anyone has said, “Yeah, we did that, we screwed that up.” But it is, like, your favorite model organism for mammalian trials being screwed up by a central facility. Because also there's this weird thing where medical people very often stop taking into account evolutionary theory because they treat that as “Well, that's that class I took in college or the beginning of graduate school.”

'''Bret:''' [https://en.wikipedia.org/wiki/Elizabeth_Blackburn Elizabeth Blackburn]. Exactly. She was her student and they shared the Nobel prize with [https://en.wikipedia.org/wiki/Jack_W._Szostak Szostak]. In any case, her work seemed good to me. I called her up, cold, you know, I went into the insect division office and I sat down at the phone. I called her, I said, Carol, you don't know me. I'm a graduate student at Michigan. I'm an evolutionary biologist. I'm racking my brains trying to understand something. Can you tell me, is it possible that mice don't have ultra long telomeres? That it's only laboratory mice that do? And she said, huh, that's really interesting. I'm pretty sure that mice have long telomeres universally. But it is odd that if you order mus spretus instead of mus musculus and you order from European suppliers, the lengths are very different than what you get if you order mus musculus from the JAX Lab. I said, Whoa.

And she said, yeah, that's really interesting. And then she said, I can't remember if it was the same phone call or if we had a second phone call, but she said she was gonna put her student, her graduate student, [https://biology.mit.edu/profile/michael-t-hemann/ Mike Hemann], who I think is now at MIT, on the project. And he was going to do a little work to figure out whether there was anything to this. And Mike did some work. They sourced some different strains of mice that were, they were actually not wild mice. Wild mice would have been the right test, but she couldn't get wild mice for obvious reasons.  

'''Eric:''' I'm sorry, this is like as close to a who'd done it Discovery J'accuse— the mice, you know, I remember, you were over the moon.  

'''Bret:''' I still am! I still can look at this email and it is the moment at which I realized, A, there's no way I'm kidding myself about how well I understand this.

'''Bret:''' When I got that email it was 1999? 98? Something like that.

'''Eric:''' By the way, that puts you at about 30. You're at the beginning of your career, and you—in this story, you've just predicted that—

'''Eric:''' Or, one of the great moments in evolutionary theory, which is—and let me just curate this, because I'm not a biologist, but I think I can more or less get this—because it's a breeding protocol that is the alteration in the evolutionary landscape for these laboratory mice, and because it's acting on a non-protein coding region, the adaptation to a change in the breeding protocol can be extremely rapid. It doesn't have to undergo some sort of completely crazy typical Darwinian story about random mutation and some of them being retained and others being rejected.  

'''Bret:''' It's even better than that. The creatures are presumably—so we haven't gotten to what the breeding protocol has to do with this—but the creatures are built in some sense to detect how dangerous their environment is, and to the extent that the level of extrinsic danger changes, their telomeres respond quickly so that they are better adapted to the environment. So, they're built to detect the environment and then what is actually a strict matter of market forces.

'''Bret:''' Environmental insult is more or less absent. What we are doing is imposing an economic rule on breeding so that we can maximize the rate at which we turn mouse chow into mice, which is obviously economically the right thing to do, if you're selling mice to all these labs, you want to produce as many mice as cheaply as possible. So producing as many mice as people—

'''Bret:''' So in order to produce as many mice as cheaply as possible, what you do is you don't breed animals past eight months. They breed faster when they're younger because of senescence. And so you don't breed older mice. You throw them out and you replace them with younger mice who breed faster. What that effectively did was it eliminated the selection against cancer, and it turbocharged the selection in favor of youthful vigor

'''Eric:''' Well let me see if I get this—in general, almost all cancers, like, cancer of the germline happens early in life, but all the other cancer, in general, is much more common later in life.  

'''Bret:''' In addition to telling me that there was something funny about mus spretus, she told me that, consistent with the hypothesis that I was conveying to her, that all mice die of cancer. She said, “If you let them live long enough, and then you do the necropsy, you find cancer of one kind or another”, and that was perfectly consistent because they had these wildly long telomeres and no cancer protection. That would be the prediction of the hypothesis—

'''Eric:''' That’s an extrapolation—it's not really all mice. It's all mice that we see in the lab, which happens to be the mice that are ordered.  

'''Bret:''' Right. She was still speaking from the mindset of somebody who thought that the mice she was getting in the mail representative representative of mice in the wild.  

'''Bret:''' Okay, so let me clear up why the breeding protocol—and I should say, that it is the breeding protocol that is causing this? That part, I would say, is still a hypothesis. It has not been directly tested by anybody, but, what I would say is that many hypotheses were tested in the aftermath of the discovery, that lab mice have bizarrely long telomeres, and wild mice don’t,  and no other hypothesis has stood up to scrutiny. So it is the last hypothesis standing and I'm all but certain that it will turn out to be true.  

'''Bret:''' When you throw them out for breeding purposes at eight months of age, you are increasing the importance of their early life breeding, and you are discounting anything related to their ability to fend off cancer because they don't live long enough in that period of time to get cancers that kill them. And so what has happened, according to this hypothesis, is that the mice that have longer telomeres have driven out the other animals from the colony. The trait of having long telomeres has swept through the colony and the telomeres have been elongated to an absurd degree, creating animals that do all die of cancer. And interestingly enough, another thing that's evident from the literature is that if you look at their tissues, their tissues do not age in the way that a normal mammal’s tissues age, they remain young.  

'''Eric:''' So there's one aspect of aging, but that there's a far darker interpretation of what you've just said. If I'm understanding you—correct me, I’ve never taken a class in biology, but I lived this adventure with you—those tissues have, at a histological level, the level of how cells are organized, the possibility of radical histological repair.  

'''Bret:''' Yes, radical effectively indefinite capacity to repair, which is going to come back in this story in the worst possible way. So—

'''Eric:''' This is like a—I mean, I just forget how great of a—

'''Bret:''' The number one senescence guy at the evolutionary level in the world, and he writes a beautiful recommendation letter for this piece. We're going to send it to Nature. George Williams tells Nature, you need to take this piece very seriously. We send it to Nature and they send it back with one of their absurd form letters that says that “The nature of the article is such that it's probably not—

'''Bret:''' To their readers. And we're, you know, I mean, we had a good laugh about that. You know, it's cancer, it's senescence—

'''Bret:''' It’s the craziest thing, and you know, the cherry on top is that they're turning down George Williams recommendation? Like, how cra— do they know who he is? Like, what? Where?