Scientist Stories: Shirley Tilghman, Epigenetics & Cloning the First Mammalian Gene

Good evening. I'm Jerry Rubin, the director of the Janelia Farm Research Campus, and on behalf of all of us at Janelia and in Howard Hughes Medical Institute, I'd like to welcome you to tonight's presentation in our Dialogues of Discovery series. I've had the pleasure of knowing tonight's speaker, Dr. Shirley Tilghman, for over 40 years, since we were both beginning scientists in the early 1970s. And I've had the pleasure of watching her career develop, first as a scientist, and then, which you'll hear about, her work which she began in the late 1970s studying the problem of imprinting. And epigenetics, which she's going to talk about this evening. She went on in around the early 2000s to become the first female president of Princeton University, a job she held for 12 years. She's now gone back to being a faculty member at Princeton. She has many honors, many accomplishments. They're all listed in this thing, which you can read at your leisure. So I don't want to take more of her time. I'd like to just welcome Shirley up to the podium. Thank you.

So good evening, everyone. Okay, we'll try that one more time. Good evening, everyone. All right, all right, now we're together. Now we're together. It is really a pleasure to be with you this evening and have an opportunity to give a Discovery Lecture here at the Howard Hughes Medical Institute. One of the things that Jerry did not say in his introduction is that I owe the Howard Hughes Medical Institute an enormous debt for my scientific career. Because I became an investigator of the Institute in 1988, one of the early investigators, and receiving the kind of support that the Howard Hughes Institute provides to its investigators meant that I could really take risks as a scientist. If you do not take risks as a scientist, it's very, very difficult to accomplish anything. They gave me freedom and I have always been enormously grateful for the support that I received from Howard Hughes. I am here in part as a way to say thank you. To the Institute. I am also here to tell you all about epigenetics. And it's a subject that has been on my mind this semester because I just completed a course in epigenetics that I taught this spring to 15 bright-eyed and bushy-tailed freshmen at Princeton where I took them through the field of epigenetics from Soup to Nuts. They have just handed in their final paper, which I thought I was going to grade on the train coming down until, unfortunately, because of the tragedy in Philadelphia, there is no train to Washington at the moment. But, so epigenetics has been very, very much on my mind. So, what is epigenetics? If you had asked me that question 20 years ago, I am not sure that I would have been able to tell you that there is a branch of genetics called epigenetics. People didn't use that term. It wasn't a common way to describe phenomena in biology, but today it is now one of the most actively researched areas of genetics. And let me start with the definition, which is it is a study of heritable changes in gene expression or cellular phenotype caused by mechanism other than underlying changes in DNA and DNA sequence. So we all, of course, know that mutations or alterations in the sequence of DNA can change the consequences to an organism. That's what a lot of genetic disease is about. This is what Various kinds of species of plants and so on are all about underlying changes in DNA. But in fact, all of the phenomena that you see on this slide are caused by something other than changes to the structure of DNA or the sequence of DNA. So here, I'm a gardener. So this petunia is very, very close to my heart. So what do I mean by that? And let's just take the petunia as an example. So every single cell in this petal, in these petals of petunia, contain exactly the same DNA sequence. And yet clearly the cells that are in the middle of the petal are expressing a gene that produces a very dark purple pigment. But the cells around the outside are not producing that pigment because they are not expressing that gene. They have the gene. The gene is completely what we call wild type, meaning it's perfectly capable of producing a pigment. And yet here it is not and here it is. And this is a heritable change in the sense that all of these cells in the middle, if they were to replicate and replicate and replicate, they would always produce purple pigment. If the cells on the outside were to reproduce, they would never produce the pigment. The question is, okay, if it's not a difference in DNA, what is the difference? All of the other examples on this slide are similarly examples. Maybe here's one of the most dramatic ones. These two ants have exactly the same DNA. This isn't male/female. This isn't where there's actually a chromosomal difference. These two ants have identical DNA. So you could say the blueprint that leads to the production of these ants is virtually identical. And yet not only are they phenotypically— and phenotypically means how they look— different between the major and the minor, but they have different behaviors. You know, one is an army ant and one is a forager. So how do you get two very different looking and behaving ants out of exactly the same DNA, underlying DNA sequence? And I could go through lots of other examples on this slide, but hopefully you're getting the picture of what we mean by epigenetics. Now, I said that this was not a field that existed until roughly 20 years ago. It is now a field that you could read about in lots of sort of publications. Time Magazine, "How the First 9 Months Shape the Rest of Your Life." Here's another one from Time, "Why Your DNA Isn't Your Destiny." The one that sends me into fits of fury is the New Scientist. Can you imagine a magazine that actually purports to report on science has a cover that says Darwin was wrong? I'm here to tell you folks Darwin was not wrong. The New Scientist is wrong. Nevertheless, what all of these articles that are in these magazines, but all are about this new field of epigenetics and what it means for human health. So I'm going to try and give you a little bit of a sense of that as we go forward. So let's compare genetics and epigenetics. So, and I'm going to start where genetics began, of course, with Gregor Mendel, the Austrian monk in the 19th century, and his his famous peas. I think everyone who has ever taken a biology course has suffered through the explanation of this experiment where Mendel took a pigmented pea plant and he crossed it to an unpigmented pea plant. In the first generation, all of the progeny were pigmented, were green pigmented. The reason now we know from molecular biology studies is because the green plant was producing a pigment. The yellow plant had the gene, but the gene was mutated. So it didn't produce a pigment. And when you cross these two, what you get are heterozygotes that produce the pigment because they are inheriting the active gene from the green plant and 100% of the progeny are green. And if you then go and do what we call an intercross, where you take two plants like this and you cross them together, you get the famous outcome where 25% of the plants contain two active green pigment producing genes. 50% get a green one and a yellow one from each parent. And there's two ways you can do that, obviously. And a quarter of the plants inherit the mutated yellow gene. And this is classic recessive inheritance. And it's how Mendel began to work out the laws of inheritance. So I'm going to show you an epigenetic effect in a different plant. And that plant is corn. So you start off with the same experiment, only in this case because it is maize, The pigmented plant is actually purple. So the leaves are actually purple. And that's because this plant has two very active copies of the pigment-producing gene. And we cross it to a green plant, which is green because it's missing that purple pigment. And what is your expectation if Mendel was correct? What should the first generation look like between this and this? All purple. But look what happens. They're all green. They're all green. So somehow or other, here you have a case where you're not getting any of this pigment-producing— this gene that is active in the parent is suddenly silent in the offspring. What's even more mysterious is that if you take this now green plant and you cross it back to the purple plant, sort of saying, well, let's get the purple back. We want to get that purple pigmentation back. If you cross these two, what you will get numerically is a quarter of the plants will be heterozygous, but three-quarters of them will only have purple genes. But look what happens. 100% of them are green. So even, even in this plant where you know you have two perfectly wild type, normal, active, potentially active copies of this gene, you get no pigment produced at all. And you can do this cross for 100 generations. You can do it over and over and over and you will never recover the expression of that purple gene. So when this was first being described by scientists, people just thought it was black magic. But in fact, we now know why this happens, and it's an epigenetic phenomenon. So this happens in humans as well. Let me give you a comparison with human genetics. So here's a Here's a classic Mendelian disease gene called cystic fibrosis, which I think everybody knows about. And the classic Mendelian behavior of this gene is if you have a carrier that is carrying a wild-type copy— a mutant copy, which is the dark blue, and a normal copy, that individual is a carrier. The disease is not being expressed. You cross it to a female here who is completely wild type, and it will only produce carriers because it takes two mutant copies to produce the disease. This is true whether the carrier is a female or whether the carrier is a male. You'll only produce carriers. But what happens when a carrier mates with a carrier? That is when you finally begin to see the disease manifesting itself because you have inherited mutant from both mother and father and you have the disease, the child with the disease. So that's classic Mendelian inheritance. So what does an epigenetic disease look like? So here's an epigenetic disease. This is a disease called Angelman syndrome. It's about 1 in 15,000 births in the United States. Initially, it looks pretty much like what we just saw with cystic fibrosis. Here's a carrier, completely normal. A male mates with a wild-type female and you produce a carrier male over here and you produce a carrier female over here. If this carrier male mates with a wild-type female, you get another You get a couple of carriers and some wild type who inherit the wild type copies from both parents. That looks like just classical Mendelian genetics. But look what happens over here— oops, sorry. Look what happens over here when the carrier is a female. When the carrier is a female and mates to a wild type male, All of the heterozygous offspring who inherit this mutation from their mother are affected and have the disease. So this is a disease that you can only inherit from your mother. You cannot inherit this disease from your father. And it is not an X chromosome linked disease. This is on an autosome. So why? The question, the epigenetic question is What can possibly explain the fact that you can only inherit this disease from your mother? It's an epigenetic effect. So that raises the question of what is really going on in these two instances that I've told you about that can explain these effects. Oh, and this is just one other example. These are identical twins, meaning their DNA is virtually identical. Yet this toddler is unaffected, is wild type. Her sister here has a disease called Beckwith-Wiedemann syndrome, which is a slightly more common disease than Angelman syndrome, about 1 in 14,000 births in the United States. So it's not DNA. They're genetically virtually identical to one another. Yet this twin is suffering from a whole slew of phenotypes. You can tell immediately somatic overgrowth. This is a much larger twin. Organomegaly is this— her internal organs are too big. Macroglossia means the tongue is larger than can fit into the mouth. She is susceptible to embryonic tumors. This is a serious condition, but it is the underlying cause is not mutations in DNA, it's an epigenetic effect. So what causes epigenetic effects? And the answer is really that the effects are caused by the way in which DNA is packaged in the nucleus. So if we were to take the DNA in all of your chromosomes and we lay them out end to end, just string them out, you know, DNA is a very long polymer. There is simply no way that that DNA could fit inside the nucleus of a single cell. No way. So what nature must do with all of that DNA is it must find a way to package it and condense it. And it uses 3 different strategies in order to do that. So here's Here's the DNA down here at the very bottom. The classic double helix discovered by Watson and Crick. So the first thing that happens is that DNA gets wound around a group of proteins called nucleosomes. And it gets wound around so that the distance from one nucleosome to another is about 140 bases of DNA. And it gets wound around again. And this is an electron micrograph where you can actually see see these nucleosomes. They are often referred to as beads on a string. But that's just the first level of condensation where we're trying to really package the DNA so it will fit inside the nucleus. So once you have this structure, it then gets sort of wrapped around itself again forming a slightly more dense structure. That gets condensed even further, even further, until finally at the very top here you have the famous picture of a metaphase chromosome which will fit inside a nucleus along with the other 22 chromosome pairs. And here's just an electron micrograph so that you can see what one of these metaphase chromosomes actually looks like. And you can see how tightly condensed The DNA is in here. So DNA is not sort of floating around in the nucleus. It's tightly, tightly condensed. So let's look at the proteins that actually do this condensation. So here are our beads on a string. And within each one of these beads are actually 8 proteins. 2 copies of each one of these proteins. Each one is called a histone. 2A, 2B, H4, and H3. And these are the basic proteins that the DNA is wound around. Now, you'll notice that each one of these histones has tails that kind of float out into space. They're not actually in that condensed particle. And the tails turn out to be the really important part of this story, this epigenetic story. Now, when I was knowing Jerry 40 years ago, and when I was in graduate school and starting out as an investigator, everybody thought histones were the most boring proteins on the planet. That all they were were scaffold proteins and no one— oops. Let me see if I can do— oh, somebody is— You're going to help me do that. All right. Thank you. So nobody thought these were interesting proteins. It turns out they're incredibly interesting proteins. So what happens to these tails? Well, these tails get modified with chemicals, with methyl groups, with acetyl groups, with phosphate groups. But they get modified in different ways depending upon whether the gene is going to be on, in which case all of these modifications lead the beads on a string to kind of spread out. If they get modified in a different way, although the groups are the same, but how they get modified is different, then what that leads is to very dense condensation of these nucleosomes.. And the consequence of a gene in this kind of a state is that the gene is off. So it's a way for the cell to structurally organize the genes it wants to express from the genes that it wants to silence. So these histone— this is called the histone code, and it is one level at which genes are regulated on and off. But there's another mechanism that is equally important, and that is DNA can actually be chemically modified itself, not just the histone tails, but the DNA itself. And the most common modification on one of the 4 bases of DNA— this is cytosine in DNA— is that this carbon sitting here in the 6 rings the 6-ring structure, this is the 5th position, the carbon 5 position, the DNA gets methylated and the methyl group gets added onto it. So what are the consequences for gene expression of that modification? Well, when these Cs near a gene are unmethylated, the gene tends to be on. But if these C's get methyl groups attached to them, then the gene cannot be expressed. The gene is therefore off. So that's a second way that you can modify the DNA without changing its underlying sequence, but modify it so that you can turn a gene on or turn a gene off. And when I was a student and when Jerry was a student. Those are the two things we knew about gene regulation. We've learned a third level at which genes can be regulated on and off, and that is epigenetic effects that are caused by both long and short non-coding RNAs. And the famous example of this is derived from the tabby cat. So how many in the audience have either tortoiseshell or tabby cats? Yes. Good. They're all female. They're all female because this phenotype is being caused by a phenomenon that only happens in females that have two X chromosomes. So what happens very early in development in female mammals is that one of these two X chromosomes is inactivated. And the entire chromosome is inactivated for all of the progeny of the cell that makes the decision. And what's fascinating is that the decision is random. So roughly half the time you inactivate the chromosome inherited from your mother, and half the time you inactivate the chromosome that you inherit from your father. If you look in a cell and look at the two X chromosomes, you can really see that the active X is very long and open and expressing the roughly 1,000 genes that are on the X chromosome. The other X is this tight little nubbin of a chromosome called a Barr body. For many years scientists tried to understand how can you activate a whole chromosome? Mechanistically? And in the 1990s, a gene was discovered on the X chromosome called XIST, for X inactivation site transcript. And what was realized very early is that this gene gets activated on only one of the two X chromosomes in a female. Then the RNA spreads all up and down that chromosome and is required to turn that X chromosome that expresses the gene into this little nubbin of a chromosome where all the genes are completely silent. If you're missing that RNA, you cannot inactivate as a chromosome. You can actually see that here. This is a fluorescence microscope of a single cell. This is a the nucleus of the cell, and the blue you're looking at is DNA. And here is this exist RNA covering the inactive X chromosome, and you can see that the DNA of that chromosome is that little yellow dot there, and it's completely missing from the X chromosome that is active. So here you have an RNA that can actually silence silence a whole chromosome. This was a revelation when it was first discovered. Okay. So having told you about these epigenetic mechanisms, I now want to apply it to one, what I think is really fascinating epigenetic phenomenon called genomic imprinting, which is a mechanism I studied for many years as a Hughes investigator. So here we have the normal state of affairs for a normal garden variety gene. You inherit a copy from your mother, it's on. You inherit a copy from your father and it's on. Both copies are on. That would be the cystic fibrosis gene if both of these copies are active. Imprinted genes, are different. Imprinted genes are only expressed when you inherit it from one parent. The other parent's copy, although perfectly normal, is completely silent for your entire life. So in this example, this is an example where the copy you inherit from your mother is very open in terms of chromatin. There's no methylation up on this chromosome, and so this gene is on in the cells that normally express the gene. But look what the state of the father's gene. The father's copy of this gene is covered with highly condensed nucleosomes. The chromosome along here is highly, highly condensed. The gene is methylated. Remember methylation was a signal for silencing. And the gene is completely silent. So that's an imprinted gene. We call that an imprinted gene. And we know that the mammalian gene home has anywhere from roughly 100 to 200 genes that behave in this way, which is a pretty high number when you consider that the total number of genes that we have is about 22,000. So this is a fairly substantial fraction of the genome that is being devoted to these very unusual genes that are only inherited when you— only expressed when you inherit it from one parent. And we know these genes are important, and I'll show you why. Oh, so— sorry. So let's go back to Angelman syndrome and see if we can now understand Angelman syndrome based on what I just told you about imprinted genes. So here we have a carrier. And Angelman is a gene— it is caused by mutations in a gene that are only expressed when you inherit them from mothers. So how is this carrier normal? Because he inherited a normal copy from his mother, so he's completely normal, and the mutation came in from his father, but it doesn't matter that his father's copy is mutant because it's never going to be expressed anyway. So as long as he has a good copy from his mother, he is normal and therefore only a carrier of the disease. Now, he can pass to his offspring that mutant copy, but it will always be in a silent state. So it doesn't matter. So as long as he continues to pass this along, he will only produce Carriers, not affected. But look what happens when the carrier is a female. Now she has inherited a wild-type copy from her mother up here. So she's just a carrier. She's fine. But when she passes this mutant copy on to her children, that mutant copy now becomes a maternally inherited gene.. And it's mutant, it doesn't work, and therefore she is completely missing this gene product. And so that's why you can only inherit this disease from your mother because it's only your mother's copy that is ever active. So we know, for example, that these are very, very important genes. Because if you create a mammalian embryo through nuclear transplantation here that has only maternal genes, therefore will be missing all of those roughly 100 genes that are only expressed when you inherit them from your father, you're going to be missing all of those genes and these embryos fail to develop. And likewise, if you have two male genomes, you start life instead of with a mother and a father, two male genomes, you're going to be missing the expression here of all the maternally expressed genes because there's no maternal genome here. Then there is also a failure of development. So these genes are important to development, and yet we've done something very unusual, which we have created a system circumstance where you only have one copy of the gene instead of having the protection of two copies of the gene. So when does this happen? When do these chromosomes get differentiated from one another? When do these genes get marked so that once in an embryo, that the embryo, every cell can tell which chromosome, which gene came from mother, which came from father? So it turns out that the only time you can really do this is before the chromosomes form in the embryo. So if you look at an early embryo, sort of the cells that are going to become either eggs or sperm depending upon whether you're a male or a female, there are no marks on these chromosomes. These chromosomes have kind of been wiped clean, if you want to think about it that way. Then if this is going to become a female, she begins to make eggs, oocytes, which are premature eggs. She puts onto these imprinted genes marks. We think the mark is DNA methylation. DNA methylation marks that signal that this is a gene here that is going to be silent when we inherit it from mothers. And that same region in sperm is going to be completely free of marks. Likewise, if this early embryo becomes a male, then what you're going to get— trying to keep this pointer from failing on us. If this if you're producing sperm, then different marks will be placed on this chromosome. Again, we think the marks are paternal-specific DNA methylation that are going to be missing in those regions of the chromosome in oocytes. And when these two chromosomes finally come together in the zygote, which is the early embryo, one chromosome is marked to silence maternal genes, the other chromosome chromosome is marked to silence paternal genes. So it really is happening before fertilization is when this phenomenon really begins to kick in. Now, what I want to spend a little bit of time on is the great paradox that genomic imprinting presents to those of us who've been studying it for many years. And that is that Imprinting results in offspring that have only a single active copy of what I just showed you are essential genes. Genes that if they are missing, you completely fail to develop as an embryo. These should have been selected against in evolution because they put the organism at significant risk. Now, one of the reasons we have— we're diploid, we have two copies of each gene, is to protect us as it does in the case of cystic fibrosis, for example. And yet we have this mechanism that seems to throw away that tremendous insurance policy. And so the question is why? Why would mammals in particular do this? So the first clue to the answer to that question came when people began to ask which organisms have imprinting. Fairly quickly, we realized that the species with imprinting include eutherian mammals—that's us—and our nearest relatives: rodents, cows, pigs, horses, what we think of as sort of the group of mammals that we all know very well, as well as marsupials, including things like kangaroos and opossums. Hopefully everybody remembers how marsupials reproduce. They are internally fertilized. They spend the first half of gestation in the uterus just as with us. But about halfway through gestation, they're born. The one thing that is developed are claws. They crawl the way up the mother's tummy into the pouch. And spend the second half of what we would consider gestation in the pouch drinking mother's milk, lactating. So all of these animals have this phenomenon, this genomic imprinting phenomenon. The species that don't have it are a much bigger group, including monotremes. Monotremes are mammals. It's sort of the most primitive mammal like the platypus. And the echidna. They lay eggs. They're the only mammal that lays eggs. The reason they're included in the mammalian genus is because they lactate. Once born, monotremes actually lactate. That's sort of an unusual variation on the themes up here. And as for all other vertebrates, you know, birds, amphibians, lizards, on and on and on, none of them display this genetic phenomenon. So that's the first clue. Why mammals? Why no other vertebrates? And I think to think about this, you have to really think about what's unique about mammalian development. The first is the fact that fertilization is happening internally. Here is an oviduct. Of a mammal. And here is an egg that has just been fertilized. It travels as it begins to divide down the oviduct and into the uterus where it does something that is just stunning in a biological sense. It burrows, literally burrows into the wall of the uterus. It becomes actually completely enclosed initially in the wall of the uterus. And of course, at that point, it stops sets up a common circulation with the mother so that the embryo developing here inside this yolk sac, which it has produced, feeds itself through the placenta. So if you want to think about it this way, the mother and the embryo are in constant communication over nutrient allocation. Another way I used to say was that the embryo, a mammalian embryo, is a parasite of its mother. And those of you with teenagers know this doesn't end anytime soon. So that makes it very unique, this way in which we reproduce. Now the second clue really came when we began to identify imprinted genes and we began to ask, Well, what do these genes do? What's their function in biology? And although this— what I'm just showing on this slide is a bit of an overgeneralization. In biology there are never any absolute rules. But in general, what was realized is that mothers silence growth-promoting genes, and so as a consequence, the genes that mothers do express tend to slow down the growth of their offspring. Now you can think that might actually make a lot of sense given that the embryo is a parasite of the mother. And if that embryo overgrows, then not only does the embryo die because it cannot be delivered in the wild, but the mother dies as well. When that happens. So it's a catastrophic event for a mother to have an overgrown embryo. And the genes that the father silences tend to be genes that silence growth-reducing genes. And his genes, the genes he expresses, are the genes that are trying to actually increase the size of their— of his offspring. And again, you could rationalize that By saying that it is in the interest of the paternal genome to have his offspring be as strong as possible, be as large as possible, so that he can send his genes into the next generation. So these two observations led a brilliant evolutionary biologist named David Haig at Harvard to propose an explanation for imprinting that I think has held up very well over time. It basically takes these two clues and synthesizes them. What does he say? He says that in polygamous species, and all mammals are polygamous, there are no such thing as a monogamous mammal. I looked. I tried to find one. Not personally, you understand. Among species is what I meant. The paternal genomes that will be selected in a population are those that increase the likelihood that offspring will be able to garner maternal nutrients during gestation, even if that is at the expense of other offspring that the mother will have. And the reason is because of polygamy, because he will not be the father of those future offspring. Offspring. So he has no investment in her future well-being or her future offspring, only in the offspring that are— that he shares with her at that point. Now maternal genomes that will be selected in such a population are those that favor equal allocation of maternal resources among all offspring irrespective of paternity. So this is It's basically a parental conflict between mothers and fathers over the optimal growth of offspring, the vast majority that they will not share going forward. And this theory has led to the idea that imprinting really is a genetics arms race that functions very much like the arms races we all knew about, at least those of us over a certain age. During the Cold War. And it works its way through this. If we're looking across evolutionary time, you begin with a rodent, and of course the first mammals were rodents. So you begin with a rodent where there is a balance between the maternal genome and the paternal genome leading to sort of an optimal size for this rodent, this little mouse. But because Darwin was right, there will be variation in the population. And one of the variations that could arise in the population is that you have a paternal genome that has silenced a gene that normally suppresses growth. And therefore, in the offspring of such a mating between a female and a paternal genome, will be larger than normal offspring. And for a period of time, that will advantage that genome in the population, and you will— because the larger size will be beneficial. But eventually that catches up with you because females this size have fewer offspring. And so, in fact, fecundity goes down. And so what will happen as variation in the population is again sampled a maternal genome will arise that can then balance out the paternal genome and you will go back to a normal-sized offspring until it happens again and you get the balance again. And so over a period of time, each one of these is the generation of another imprinted gene. And it ratchets up just like the arms race that the normal arms race that we all know about, ratchets this up. And the consequence is, by the time you get to where we are today, you have 100 to 200 genes that are being regulated by this very, very unusual phenomenon. So I want to leave you with just a few samples of the impact of this genetic mechanism that is in all of us on human health. So the first one that I want to tell you about is that we are now learning that this— these genomic imprinted genes are playing a very significant role in the development of some cancers. And that will not surprise you because cancer is a disease of overgrowth. And these are genes that that regulate growth. So for example, here you might have a growth-promoting gene that is normally only expressed from the father. The mother's copy, for example, is completely silent. But through a mistake during early development, the maternal gene is— these epigenetic marks are lost. And the consequence is that the maternal gene turns on. Now you go from having one copy of a growth-promoting gene to now you have two copies of this growth-promoting gene. And that is the mechanism that underlies Beckwith-Wiedemann syndrome. So if you look at children with Beckwith-Wiedemann syndrome, they have misregulated their maternal gene, and as a consequence, They are overgrown, and because the gene that is overgrown is a gene that is involved in fetal growth, you see the development of cancers in these children at a substantially increased rate over the normal population. We now know that this regulation is involved in a lot of human cancers that we see where there's not mutation, as we were all taught in the early days that cancer is a disease of mutation. Cancer is also a disease of the misregulation of imprinted genes. So these genes are very important for that reason. Here's another implication. I told you that these marks are put onto the chromosomes during the development of eggs and sperm. And that when they come together, the difference between the chromosomes is then inherited from cell generation to cell generation. The one time we know that these marks are unstable is very, very early in development. And that is when people are generating embryonic stem cells, like these embryonic stem cells, or they are cloning animals, like Dolly. This was the first cloned animal. So anytime you take a fertilized egg out of the body and you put it in culture, you risk the possibility that these imprints will be lost. They're unstable at this stage of development. And so, for example, the major phenotype of a cloned animal that is not born is somatic overgrowth. So it's misregulated these imprinted genes. So it's making— imprinting is making it much more challenging to be sure that these manipulations that are being done at the time of early development are actually not misregulating these crucial growth-regulating genes. Here's another really fascinating epigenetic example that— where imprinting has been implicated. This was a terrible, I mean, a tragic human experiment that happened in the Netherlands in about a 4-month period between 1944 and 1945 where the Germans basically starved the Dutch population. And the consequence you can see for the phenotype of children that were born during this very terrible period in terms of their weight, their length, their head circumference. All of these things were affected if they were conceived during this period of time when virtually everybody in the population was starving. But what studies have now shown is that these effects, which have been— where imprinting has been implicated in these effects are having long-time implications that are being seen in adults of these— the hunger winter. And obesity is one of them, as well as other disorders including schizophrenia. So some of these disruptions in early development— this is caused not by manipulation of eggs, but by starvation. Have long-term detrimental effects in the future. And the last place where epigenetic effects and particularly involving imprinting are beginning to be identified are in the genesis of obesity. So I think we all know that the environment has a profound effect on the likelihood of developing obesity. Genetic variation, we now know, is a big part of that story. But the third part of the story that is now beginning to be recognized is that there are epigenetic mechanisms, sort of misregulation of methylation, misregulation of histone modifications, that are also contributing to the obesity. And one of the reasons why people are paying so much attention to this mechanism right now is because there are some studies from both animals as well as humans that are suggesting that there is a transgenerational effect here. In other words, that obese women are much more likely to have obese offspring who will have obese offspring who have obese offspring. And so that becomes a really serious public policy and public health issue if in fact these studies, which are very preliminary at the moment, turn out to be the case. So let me just end by thanking you all for coming this evening to learn about epigenetics. And if— and this cartoon is pretty much what I'm going to say in the answer to questions. So thank you all for coming. So I think there is a microphone that's going to be going around. We have a question over here. Yeah. Can you tell me, I'm an IVF nurse and I know that there's some mitochondrial kind of transplant happening in Canada right now where they're taking the mitochondria of one woman's eggs, a young woman's eggs, and putting them with the DNA of an older woman. Can you tell me how epigenetics would impact that? Yeah, and I think it comes back to the manipulation of the eggs. So this is called mitochondrial gene replacement. Actually, I had a senior advisee this year who wrote her senior thesis about mitochondrial replacement therapy. Which has been approved in the UK, just recently been approved in the UK to go into clinical trials. So this is— let me just, for people who know less than you do, I suspect, these are diseases caused by mutations in the DNA in mitochondria. So mitochondria are small organelles inside cells. You often have as many as 1,000, 2,000 mitochondria per cell. Cell. And the critical thing is you always inherit the mutations from your mother because you only inherit mitochondria from your mother. So there have been scientists in both the United States, Canada, and the UK working on a transplantation strategy that could take a fertilized egg or an unfertilized egg from a normal mother who does not have mitochondrial disease, take out her nucleus, so now there's only mitochondria in the egg, replace it with a nucleus from the mother so that— and it could be a fertilized nucleus, it could be an unfertilized nucleus. And therefore what you will reconstitute is a fertilized egg that has normal mitochondria and a normal nucleus, and thereby potentially completely cure the disease. Of all of the risk factors that I see in the development of that strategy as a wonderful cure for mitochondrial disease, imprinting is, I think, the biggest risk factor. Because in order to do that experiment, you've got to take out the eggs, you've got to put it in culture, you've got to poke it with a needle a couple of times. And every time you do that, you risk the chance that these imprints will start to be lost. So I think it's a— I think it is the most— as I've understood it, it's the most serious risk. Thank you. Hi. Oh, yeah. Sorry. Oh, hi. I was wondering in regards to the XXX syndrome, how would epigenetics affect and how would the RNA be— how would the RNA affect those chromosomes in the XXX syndrome? So the wonderful thing about X chromosomes is they know how to count. Which is a remarkable thing when you think about it, that chromosomes know how to count. So if— for individuals who, because of what is called nondisjunction during meiosis, inherit 3 X chromosomes instead of 2 X chromosomes. In which case, what happens in those individuals is that 2 of the Xs become inactivated. So there's a counting mechanism in the cell that ensures that you will only ever have 1 active X chromosome. So it actually— the cell itself cures the X chromosome inactivation problem. The reason why there are phenotypic effects with individuals who have triple X is because unlike male sex chromosome inactivation, which is, as far as we can tell, complete— I mean, the entire chromosome is silent— there are genes that escape X inactivation on the human X chromosome. It's leaky, in other words. And so instead of having 2X of those genes, you have 3X of those genes. And that's why there's a phenotype. But they actually have two inactive X chromosomes. So does imprinting explain the liger? Yes. Yes. So that— oh, great question. Yes. So the imprinting explains the liger. So if you cross a lion with a tiger, what you get is a gigantic organism that is terrifying to look at. And it's somatic overgrowth. It's basically somatic overgrowth. And the reason, despite the fact that no experiments have actually been done on that animal, the reason that I think I can say this with almost 100% certainty is because we did experiments maybe 10 years ago, 15 years ago, I guess, with two species of mice that had been separated in evolution about the time that lions and tigers have been separated. And if you cross them in one direction, you get mice that look like ligers in the sense that they're enormous. If you cross them in the other direction, you get dwarfs. And so we went in and did the genetic analysis of what was happening in these two different crosses. And what is happening is when you take species that are too far apart in evolution, they're really speciated, what you get is misregulation of a significant fraction of their imprinted genes. And so the overgrowth are obviously overexpressing the paternally expressed genes, and the dwarfs are overexpressing the normally maternally expressed genes. And the other place you see that is between horses and donkeys. If you do those crosses, you get size effects like this. Yes. So can you say anything about the drug called methylcysteine? I'm missing a The full name of it, methotrexate, or no, um, no, methylcysteine. It's like a supplement that— but I think it's being widely prescribed now because there's considered to be a whole range of things that are connected to this methylation that psychiatrists and others are thinking is— I don't know if this is hokey or if this is science, um, but I know doctors are actually prescribing L-methylcysteine? I am sorry, I am not going to be helpful here. I just don't know enough about it. Does that suggest that maybe it's not very scientific or you just can't say either way really? No. There is an experiment in mice I can tell you about that isn't hokey at all where there is a gene called a GUTI. Which affects coat color, actually, and actually affects obesity. And depending upon whether the mother is fed— I think they feed her S-adenosylmethionine, which is a source of methyl groups, basically. If they feed her that, the gene that normally leads to the yellow coat and the obesity is silenced, and she has normal offspring. If you take that out of the diet, she has offspring that are obese and where there is the yellow coat color. So her diet can affect the methylation of her offspring's genes and affect the phenotype. So the principle has been well established with animal models. What I don't know is anything about the clinical trials that suggest that that would be a sensible thing to do. One of the things to recognize is that there would be no specificity in that experiment, right? Because that methyl donor would have no way of knowing where those extra methyl groups should go or whether they should not go. The fact that in the mice it hits this gene doesn't mean that's the only gene it hits. Probably hitting lots of other genes, but you can see it because the effect of this one gene creates a phenotype. So it's a highly unspecific way to try and affect gene expression.

Yeah. You mentioned that the reason this developed was because the offspring were developing inside the parent. So I was wondering for certain species of sharks and seahorses where they also develop inside the parent, do you find imprinting as well?

Yeah. It's a great question. A really great question. And I was dying to work on seahorses to see if you could see an imprinting effect in seahorses. It just proved— pragmatically impossible to get these animals, to get them to breed, to get— but it's a great experiment. There's also apparently a bird where the embryo develops, if I have this right, we looked into it a long time ago, inside the father, actually. We thought that would be a really interesting look at. But you're thinking about this exactly correctly. That's the kind of thing to test. Yeah. I understand that when an embryo is fertilized, they know every single day, almost to the hour, what type of development can be expected. You mentioned with Dolly, and my question is really about surrogacy, where a woman is unable to conceive and the desire is to transplant. Could you speak to that a little bit? Yeah, I think we know from, you know, literally how many years now of in vitro fertilization that in vitro fertilization is is an inefficient process, even conducted in the very, very best clinical settings. It is— you know, it's— there's no guarantee at all. And there are many reasons for that. I'm not for a second suggesting that defects in imprinting could explain even, you know, a substantial fraction of them. But what we certainly know is that every time you manipulate a human egg outside the woman's body in a petri dish, the risk is there that the instability of these imprints will be manifested. And therefore, it makes in vitro fertilization more challenging. It makes surrogacy more challenging.

We just know. Yeah. You spoke a minute ago about diet as a possible means of correcting the dysregulation. I wonder if there are some other treatment strategies that are being thought about and if you could speak about those briefly.

Yeah, it's very challenging. It really comes back to the issue that we were talking about a minute ago, which is specificity. So, for example, I'll give you an example of colon cancer. So we now know that colon cancer is a disease that is the result of both mutations in very specific genes and the misregulation of genes that are both growth-promoting and growth-inhibiting. If you want to think about how can we correct the second thing, not the mutation, but the epigenetic effect, Because after all, you've got a perfectly wild-type gene there to work with. The question is how do you target the silencing to that gene and not to other genes in the genome that might be very critical, right? There are— people have ideas about how to do that. We're learning how to do genetic engineering in ways that are both exciting and troubling. But what we don't know is how to deliver those methyl groups or take them away in a specific region of the genome without affecting a lot of the other genome. So that's the challenge right now. There's one. Oh. So you talked about all the different ways that these genes are deactivated, like methylation and imprinting. But you also read often about other genes that are inactivated, like our tail, for example. Is that the same mechanism or is that a different mechanism to inactivate, say, our ability to grow a tail? I mean, that's a complex mechanism because, of course, it involves the disappearance of genes that are required to make a perfectly good tail. And to a certain extent, the changes in gene regulation that are there to take genes that might have been used for one purpose in the past and repurpose them for an organism that doesn't need a tail anymore. So there are a lot of things going on in that case.

You cited a couple different diseases that you can definitely draw the correlation of this taking effect. Do you have other diseases that you have pinpointed that you believe, things like alcoholism or addiction, where things run in a family but, you know, different people can or cannot?

Is there any work being done on that or others? So there, if you do sort of a literature search of right now of virtually any human disease and the word epigenetics, you will find literature of people who are asking the question, is there an epigenetic effect in that particular disease? It's remarkable. I did it for this course and was just stunned. If you go in and actually read those papers, most of them are— have no really compelling evidence. For an epigenetic effect. And— but obesity is a good example of these kind of polygenic diseases, diseases that are the result of multiple genes interacting with the environment and potentially interacting with epigenetic effects. I don't think it is going to be the only one that actually gets nailed down I think it's going to be seen elsewhere as well. But this is so new and the evidence is so poor right now for most of the diseases that are claiming epigenetic effects. The best by far is in cancer. There it's absolutely clear that there are these epigenetic effects largely through the misregulation of these imprinted genes. Yeah. Actually, somewhat on this subject, are there any viruses that employ— there are some viruses which promote cellular reproduction. So I'm going to repeat the question because they're filming this and we didn't wait for the mic. So the question is, are there epigenetic effects using viruses? And— or epigenetics effects that viruses use. So the most compelling example of that is in our genome we have both active and probably dead viruses all over our genome. It's a very substantial fraction, actually, of our genome. And we use DNA methylation to keep them under control. And that methylation is absolutely critical for maintaining the integrity of our genome in the face of these viruses that otherwise would be running amok. So there's no— of that there is no question. I'm trying to think of an example where a virus has taken advantage of epigenetic mechanisms to enhance its own replication. I can't think of an example right now, but that's probably because I just can't think of an example right now.

So since I have the mic, I'm going to use it to ask you all to thank Dr.

Tillman for a wonderful lecture. Thank you.