H. Robert Horvitz ’68 - 2002 Nobel Laureate Lecture in Physiology or Medicine

MODERATOR: Thank you all for joining us in the privilege of being present today to celebrate this wonderful occasion in the life of science, and of MIT, and certainly of our friend and colleague Bob Horvitz.

There is no event that more symbolizes and reminds us of why we are all here than when one of our colleagues wins the highest possible accolade for the science that he or she has done. The things that we do at MIT, the things that people, especially like Bob, do in basic research, in my view, is at the absolute highest level of human accomplishment. And it is truly a privilege to count Bob among our friends and colleagues, and to recognize how he has advanced the state of human knowledge. And how that knowledge ultimately will improve health for humankind.

So thank you all for being here. And I would like now to ask our colleague, Phil Sharpe, to introduce Bob. Phil.

SHARPE: It's a great honor to be asked to introduce Robert Horvitz. And I want to begin by congratulating Bob for his outstanding research and work here at MIT as a colleague. Bob's research, as you will hear, is fundamental to all life processes. Fundamental to how a multicellular organism is formed, fundamental to how our brain is formed and works, and fundamental to the processes related to disease such as cancer. And therefore, as you would expect, Bob has many homes. He is a professor in the Department of Biology, he is an investigator in the McGovern Institute for Brain Research, and he is a member of the Center for Cancer Research here at MIT.

Beyond the borders of MIT, Bob has also other homes. He is a investigator in the Howard Hughes Medical Institute, and has been supported by Howard Hughes for many years. And from the day he came to MIT, almost, he has been supported by the National Institute of Health. So Bob has many friends and many homes.

Now, it has been said that the best predictor of whether you're going to receive a Nobel Prize, statistically, is having your office within 50 yards of another Nobel Prize winner. In Bob's case, you want to interpret that as mentors. Now Bob was an undergrad at MIT, where he majored in mathematics and economics, and his advisor here was Bob Solow, who, if you don't know, has a Nobel in economics.

He went to Harvard, and his advisor at Harvard, where he got his PhD in biology was Jim Watson and Wally Gilbert. And that's three for three.

And then he went to Cambridge to do a post-doc. And there, things were getting a little slow for him in keeping this perfect record. And his mentor at Cambridge was Sidney Brenner and John Sulston. Sidney was his post-doctoral mentor, John is who he worked with, and they share this honor with him this year. And that makes it 5 for five for Bob. So if he wants to be a post-doc in somebody's lab around here, I'm sure others would accept him.

Bob has been at MIT for many years, since I think '76, or something like that. He not only has done outstanding research, but he has been a wonderful colleague. He teaches. He teaches remarkably well, as you will learn, and enthusiastically. And he has trained some of the best scientists this country has. Every major university in this country has someone who has been trained, either as a graduate student or a post-doctoral fellow, in Bob Horvitz's lab. He is the professional constant. So it's a great pleasure to introduce Bob today. Let's all give him around of applause.

HORVITZ: Thank you, Chuck and Phil, for the introduction. And thank all of you [LAUGHING] including my wife Martha, for coming here today.

This is a very exciting time for me and I can think of nothing better than to share this time with my MIT friends. And I'd like to start with a few specific statements of thanks to these friends. First and foremost, to my lab. I thank the students and postdocs who have been members of our research group, because, as I'm sure most people in this audience know, it's really their efforts, their insights, and their accomplishments, that have been recognized.

I thank my colleagues in the Department of Biology. These colleagues have provided a spectacularly supportive environment, and have fostered my growth as an independent scientist. And their help, their friendship, remains a major force behind my scientific efforts. I particularly thank Boris Magasanik, who hired me. And also my other ex- and current department heads, Jean Brown, Morie Fox, Richard Hines, Phil Sharpe, and Bob Sour, for their support over many years. I thank Bob Sour and Graham Walker, with whom I grew up at MIT. We entered as faculty members at very much the same time. And I thank Salva Luria and David Baltimore, both of whom made me think about science, but also made me think beyond science, to the responsibilities of a scientist in the country and the world.

Now Phil, when I first came to MIT, it was 1964. I'm kind of homegrown. I came in 1964 as a freshman. At that time, Paul Gray was a 32-year-old associate professor, and if I recall correctly, he was then, or soon thereafter, chair of the freshman advisory council. The associate dean of students was Fred Facet. And I remember Fred very distinctly from day one, because as a freshman, I sat in Kresge auditorium as he addressed us. And what he said was, look to your left, look to your right. One of you won't be here when graduation time comes.

Times have changed and some of us are still here. In this room I took 601, Introductory Electrical Engineering from Amar Bose. Most of you probably know him from speakers, radios, and headsets. He knows electrical engineering. As Phil said, I graduated with degrees in, well he didn't quite say it this way, 18 and 14. Mathematics and economics. And, as Phil also noted, I had the wonderful opportunity to work with Bob Solow for my undergraduate thesis in economics.

The time then was in the late 60s. And I wanted to do something different. And that something turned out to be biology. So what I would like to do now is to turn to that biology. The 2002 Nobel Prize in physiology or medicine has been awarded to Sidney Brenner, John Sulston, and me, for a set of interrelated studies that focus around discoveries concerning a phenomenon known as programmed cell death. Why should anyone care about dying cells? What does it mean that certain deaths are programmed?

First, let me say that it is not notable that cells die. This has been known for many years by biologists. In fact, the real problem has been that in order to study cells one has to keep them alive. When they die that's not interesting, that's a problem. What is notable is that, in many cases, when cells die, the cell death is a manifestation of a normal process. More specifically, what's happening is that cells die not because they've been damaged beyond life, or in some terrible environment, but rather because they have activated an endogenous program of cellular suicide.

Now what about the term programmed cell death? To me, programmed cell death is synonymous with naturally occurring cell death. And let me explain what I mean by that. During the development of an animal from a single cell, from a fertilized egg, many things happen. First, the egg divides. Over and over and over again, generating a large number of cells. For example, in us, we don't know exactly what that number is, but it's probably on the order of 10 to the 13th. 10 million million. And then each of these cells must differentiate. They must take on specific characteristics and become certain types of cells, like a nerve cell, or a muscle cell, or a skin cell. And then all of these cells must interact as groups to form proper structures, an arm, or a liver, or whatever, and proper interactions, proper interconnections, as in the most complicated of organs, the brain.

These processes of cell division, cell differentiation, and morphogenesis constitute basic events of development, and they define basic problems in developmental biology. But now, in addition to these processes, there's another process that appears to be universal among developing animals. And that is the process of cell death.

Quite remarkably, many of the cells that are generated as animals develop do not survive, but instead die. And it's this naturally occurring cell death that is often referred to as programmed cell death, because it is an aspect of the natural quote "program" of development. Now, as biologists have learned more about programmed cell death, the term has been extended to other deaths that use the same mechanisms as those that occur normally during development. So, certain cell deaths that biologists study in vitro, In the test tube or on Petri plates, are programmed cell deaths. And in some diseases what one sees are programmed cell deaths.

Now, that cell death occurs during the normal course of animal development has been known for many years by developmental biologists. For example, the regression of the tail of a tadpole, as a tadpole becomes a frog, involves the programed deaths of essentially all of the cell types in the tadpole tail. Similarly, the formation of digits, for us fingers and toes, that formation involves the elimination of interdigital regions that are cellular, in us, in utero. And these regions are removed by programmed cell death.

We can see this exemplified in birds. If you think about birds, think about bird feet, some are webbed, and some are not. And here we see some friends, MIT friends. This is from the food court, this pigeon. There near all the trucks, near building 68. This is Charles River Goose. Webbed, not webbed. Programmed cell death, no programmed cell death, in the foot.

Programmed cell death is not a minor event. For example, approximately 95% of the thymus sites, blood cells, in our immune system, 95% that are generated die by programmed cell death. In areas of the developing mammalian brain, as many as 85% of the nerve cells die by programmed cell death. Yet, despite this very pervasive phenomenon, it was ignored for many years by biologists. Why is that?

Well, I think there are a number of reasons. One, I already alluded to. But another has to do with timing, with kinetics. And that is, that this process of cell death is quite rapid. So if one takes a snapshot, by looking at a sample of tissue, and says, how many cells are in the throes of death? The answer is very few. It's only if you look over time. And that's something that one must know to look for in order to assess it.

To me, perhaps the most interesting and fundamental aspect of programmed cell death is simply what I've said already. And that is that there is a biology of cell death. And I think the major finding in the field of cell death is the fact that cell death is a normal biological process, and it is an active process on the part of cells that die.

The key studies that led to this conclusion demonstrated that specific genes must function for programmed cell death to occur. And many of these studies were performed in my laboratory here at MIT. And these studies indicated that there is a biology of cell death, every bit as much as there is a biology of other basic cellular processes, like cell division, cell migration, cell differentiation.

OK, so now, I've been talking about the biology of cell death. And if you think about it, any biology can go wrong. And if it does so, in us, that can lead to disease. And programmed cell death is no exception to that. It appears that some, and maybe a very substantial number, of human diseases are consequences of disruptions in the control of programmed cell death. For example, the major clinical features of many neurologic disorders, the neurodegenerative diseases include Alzheimer's, Huntington's, Parkinson's, ALS, other disorders, stroke, traumatic brain injury. In all of these cases, there are cells that die that shouldn't.

For example, in Lou Gehrig's disease, ALS, amyotrophic lateral sclerosis, specific nerve cell types, the so-called motor neurons, the cells that drive the contractions, contractions of our muscles. They die in the brain spinal cord, leading to paralysis, and eventually to death. What causes these cells to die in ALS and other types of nerve cells to die and other neurodegenerative diseases? Let me say, very specifically, we don't know. But, one hypothesis is that at least some of the cell deaths in these disorders are ectopic programmed cell death. In other words, the cells die by using that same mechanism that's used developmentally.

There is a range of data that support this to varying degrees in these various disorders, and in at least one neurodegenerative disorder, retinal degeneration, I would say the data are very strong. And if you ask me to guess, I would guess that this process is involved in others of these disorders as well.

So these are disorders in which there are too much cell death. But there also are disorders of the opposite sort, in which there is too little programmed cell death. So that, for example, cancer, we generally think about cancer as a loss of control over cell division. So too many cells are generated. And that's a very appropriate way to think about cancer, but it's not the only source of cancer, ultimately.

If you think about the number of cells in our tissues, it turns out that this number is defined by two opposing processes, the process of cell division, which adds cells, and the process of programmed cell death, which takes them away. That number can increase, either because of too much cell division, or because of too little cell death. And certain cancers, the one studied in the most detail is follicular lymphoma, in fact are cancers that ultimately are caused because there is too little cell death.

So what I noted earlier is that one of the major reasons for concluding that cell death is an active biological process is the finding that there are specific genes involved. And what I'm going to do now is to tell you about such genes, based upon the work that we've done here in our laboratory. And we've discovered such genes not by studying cell deaths in humans, or human disease. Humans are complicated, and not ideal experimental organisms for many purposes. But rather what we studied is a microscopic roundworm, or nematode, caenorhabditis elegans, or c. elegans

Now, if you ask about this year's Nobel Prize, the major player is this little worm, c. elegans Because, all three of us, Sydney Brenner, John Sulston, and myself, are being honored because of studies of this little worm.

This is Sydney, often referred to as the father of the worm. Sydney is a brilliant scientist. Sydney helped pioneer the field of molecular genetics. He had a major role in the elucidation of the genetic code, he was a discoverer of messenger RNA, and he has countless other major discoveries. Sydney was drawn to c. elegans in part, because it's very simple. We now know that there are only 959 cells in the c. elegans adult. And this can contrast with the 10th to the 13th and in humans, or something in the hundreds of thousands to millions in insects. 959.

He was also drawn to c. elegans because of features of its genetics, features that made it exceptionally well suited for genetic studies. Geneticists are concerned with rare individuals, events that occur one in a million. And so, it's nice to be able to study a million individuals, or 10 million, or 100 million, or what have you. And with some organisms, that's very hard. With c. elegans, because it's only one millimeter in length, and very easy to grow, on a single Petri dish we can grow 10,000 individuals, and many more than that in a liquid culture.

So the cellular simplicity and appropriateness for genetic studies. The other thing I should note is that c. elegans grows very rapidly. So geneticists are also interested in generations. You go from one generation, to the next, to the next. C. elegans has a three day generation time. From the time an egg is produced to the time that egg is adult producing its own eggs is three days. Many generations, lots of animals, simple cellular anatomy, Sydney said worms are good. Let's use them for studies, genetic studies, of developmental biology and neurobiology. And he pioneered such analyses.

John Sulston, the third of this year's recipients, and I both worked in Sydney's lab at the Medical Research Council Laboratory in England, of molecular biology. Cambridge, England. Early 1970s, early to mid 1970s. I was a post-doctoral fellow. I was supported by the Muscular Dystrophy Association, and as Phil noted, I came from the Harvard lab of Jim Watson, Wally Gilbert, and also Claus Weber. And there, I had studied the bacteria phage T4, and what it does to its host RNA polymerase. An interest that Phil has continued with in a broader sense.

John was a staff scientist at the MRC. He had received his PhD in organic chemistry from the University of Cambridge and then went to La Jolla and worked with Leslie Orgel on essentially biochemical studies of the origin of life. And came back to work with Sydney to focus on the neurochemistry of c. elegans.

So all three of us were interested in the worm, and in fact all three of us were really very much interested and maintain a strong interest in the nervous system.

This is the worm. The worm comes in two sexes. The lower one is a male, the top one is not. It is a hermaphrodite. It's essentially a female whose gonad, for a brief period of time, forgets and makes sperm. That has various experimental advantages, but I won't go into that right now. This image shows photographs taken by John Sulston and diagrams that I drew.

When I arrived in Cambridge, that Cambridge, John had just discovered that by observing living c. elegans, using a Nomarski optics under the light microscope, he could see individual cells, and follow them as they divided. And here's an image from a paper John and I did. If you look at either of these cells, p9 nine and p10, here's a time course, each of them undergoes the process of mitosis, dividing to make two daughter cells.

Watching life. Well, this was pretty amazing. You could follow these cells, and then follow what their descendants did, and they would go on and divide. And so you could follow a cell, and its daughters, and its daughters, and so on and so forth. And for example, these two cells, P9 and P10, they went on and divided as shown here, to generate ultimately six descendants each, in a very short order of time.

So in talking about it, John and I decided that it should be possible to determine every cell division that occurred, at least during larval development after hatching, in this animal. And with the added efforts of Judith Kimball, who was a graduate student with David Hirsch in Boulder, and she was interested in gonadal development, we did exactly that. John and I published the complete description of the non-gonadal post embryonic cell lineages of c. elegans in 1977. And two years later, Judith Kimball and David Hirsch published the gonadal lineages, thereby bringing the post embryonic lineage to completion.

John then closed the loop. He studied the egg, and followed that first fertilized cell, fertilized egg cell, as it divided and divided, ultimately to make the larva. Now this was hard. Let me just tell you that this is, if you picture a bowl of grapes with about 550 grapes, and your assignment is to watch all 550 as somebody is standing there shaking the bowl, that's essentially what John did. So my part was easy. John did the hard part, and he did it completely.

And that led to this, which is the complete cell lineage of the c. elegans hermaphrodite. We also have the male, looks a little bit different. Starting with a single self-fertilized egg, divides to make two, divides again, so on and so forth, y-axis is y. And that's the developmental origin of all 959 nuclei that are found in the adult.

This pattern, and the corresponding pattern of the fates of these cells, is essentially invariant from animal to animal. There is enormous biologic information in the cell lineage, but I won't go into that today, except to say that the knowledge of the cell lineage has made possible the analysis of the mechanisms responsible for the development of this animal at the level of resolution of single cells. Now, in addition to the 959 cells that are found in the adult, this cell lineage generates another 131 cells, not found in the adult. The reason? They undergo programmed cell death.

These cell deaths occur at different points in the lineage. Here we look at the development of the feeding organ, the pharynx, and basically you find that the cell deaths occur at different times, different points in the lineage, and with nearest relatives varying quite substantially over the sister cells. When you look at these deaths, what you see, again using no Nomarski optics, is most obvious here, this flat disk like appearance is a cell in the throes of death, and later that cell goes away, no sign of it can be seen.

Looking at the deaths with an electron microscope, they have specific ultra structural features, most obvious here is the electron density of this dying cell. There are other features as well, which told us that these deaths looked like a process that had been defined in pathological specimens from human patients of cells that were dying and were called, the process was called apoptosis. So programmed cell death in c. elegans is similar to mammalian apoptosis, which I should add is pronounced in any of a variety of ways.

So what we decided, then, was that since cells die in a specific morphological, through a specific morphological set of events, at different times, in different places in the lineage, that we could think about programmed cell death as a cell fate. No different from any other fate, like differentiating into a gut cell, or a serotonergic nerve cell, or whatever it happens to be. And if that's true, then we postulated there ought to be genes that control that fate, that control the decision should cells live or die, and that having made the decision to die, should carry out that process of death.

And so what we did was look for, identify, and characterize such genes. And those studies led to the definition of a genetic pathway, as outlined here. This pathway consists of essentially four sequential steps. The first step is a decision step. Every cell in the body has to decide, I will live or I will die by killing myself. The second step is the killing step itself. That's what generates the corpse. The third step, basically, is getting rid of the body. It's a process of engulfment whereby the dead cell is removed from the animal. And the fourth step is eliminating the debris of that cellular corpse.

Let me briefly tell you about these steps. Some years ago Hilary Ellis, shown here, and Chan Desai, shown here, both as graduate students in the lab, identified two genes with functions needed for cells to die by programmed cell death. These genes are called CED-3 CED for cell death, abnormal. CED-3 and CED-4. And if either of these genes is inactivated, essentially all of the cells that should die instead live. And it was this finding that said that programmed cell death is an active, biologic process.

Then Junying Yuan, as a graduate student, performed so-called genetic mosaic analyzes, asking the question in what cells must these genes act in order for death to occur? And the answer was in those cells that are going to die. Conclusion, programmed cell death is, at least to this extent, a process of cellular suicide. Junying and another graduate student, Shai Shaham, then worked together to clone and molecularly characterize these genes. And what they found is that first these two killer genes, CED-3 and CED-4, first CED-3, looks similar to a human protein known as ICE, interleukin 1 beta converting enzyme, a protein that had been purified by two pharmaceutical companies interested in human inflammatory disease.

And this information told us something about CED-3, because they had shown that ICE acted as a protease, an enzyme that cleaves other proteins, suggesting to us that CED-3 kills by acting as a protease. And [? Ding Sho ?] in the lab confirmed that. And also suggesting, that since there was a human counterpart, that ICE or other CED-3 ICE-like proteins may well play a comparable role in the process of programmed cell death apoptosis in humans. And there has now been much evidence supporting that.

The second gene, CED-4, also has a human counterpart, identified through biochemical studies done by [? Shel Don ?] [? Wong ?] in Dallas, Texas. Biochemical studies in which he was assaying in a test tube the process of programmed cell death, and found an activity that activated this process. And it turned out to look like CED-4, which we had already described.

The third c. elegans gene that plays a key role in this core killing part of programmed cell death is known as CED-9. And CED-9 is different. Because whereas CED-3 and CED-4 are killers, CED-9 is a protector. It keeps cells alive. CED-9 was discovered by graduate student Ron Ellis, and characterized molecularly by graduate student Michael Hengartner. And what Michael found was that CED-9 also looked like a human protein, this one that had been isolated because of its involvement in cancer, a protein known as BCL-2, were BCL stands for b-cell lymphoma. BCL-2 is proto-oncogene, causally responsible for follicular lymphoma. It's what I referred to earlier.

OK, so we have CED-9, CED-4, and CED-3. And Shai Shaham analyzed through genetics the interactions amongst these genes, and defined the core genetic pathway for programmed cell death, as shown here, where it CED-3 kills, CED-4 kills by causing CED-3 to kill, and CED-9 protects, by preventing CED-4 from killing, causing CED-3 to kill. Given the mammalian counterparts, we postulated a pathway as shown here. With BCL-2, the Apaf-1 CED-4-like factor, and these proteases, that are today known as caspases. So this is the key core pathway. What else?

Barbara Conrad, a post-doc in the lab, analyzed another gene, egl-1 a gene that we had actually studied and discovered in the course of studies of behavior. But it turned out the behavioral abnormality was reflecting abnormalities in programmed cell death. And what Barbara discovered was that egl-1 just like CED-3 and CED-4 is a killer. We won't go through data but you can see egl-1, like CED-3 and CED-4 here. And the genetic studies she did showed that although it was a killer, it was different from CED-3 and CED-4, because instead of acting, were CED-3 and CED-4 act downstream of the protector CED-9, it acts upstream of CED-9.

She cloned and characterized this gene and, whoops. First pathway. Pathway, egl-1 upstream. So egl-1 kills by preventing CED-9 from preventing CED-4 from activating CED-3. Right? Pathway. What is it? Barbara cloned it. And it turned out to encode a gene related to BCL-2, the Proto-oncogene. And it turns out that there is a family of these BCL-2-like genes in us. And egl-1 looks like another one of these. Some of these genes, like the BCL-2 protect against cell death. Others, like egl-1, cause cell death. So egl-1 is upstream, now, of CED-9 in this pathway.

And it was known at that point that BCL-2 members could physically interact with each other. So Barbara asked the question what about CED-9 and egl-1, and she found that they interact physically. And Furthermore, in studies we did, and that were also done in many other laboratories, it became clear that said CED-9 and CED-4 also interact physically, leading us now to a molecular pathway, as indicated here, where we say egl-1 is a direct negative regulator of CED-9. So egl-1 prevents CED-9 nine from acting by physically interacting with it. And then CED-9 prevents CED-4 from acting by physically interacting with it. And that's what the genetic and molecular studies indicate.

What is it well one does? We think the answer to that is it changes the location in the cell of CED-4. And the way in which we discovered that was through a series of experiments done in part by Barbara Conrad, whom I've already shown you a picture of, and also by Brad Hirsch, the graduate student in the lab, Zheng Zhou, a post-doc in the lab, and Fangli Chen, who was an undergraduate student in the lab. And what they did was to use antibodies against these various proteins involved in cell death, and ask the question, where are the proteins?

And what they found first was that CED-9 is localized to mitochondria. OK, what we're looking at here is a developing c. elegans embryo. This panel shows where mitochondria are. And basically, CED-9 and mitochondria are in the same location. So CED-9 is in mitochondria.

They further showed that CED-4 is also in mitochondria. And remember, I've told you CED-9 and CED-4 physically interact. So the question, then, was is either required for the localization to mitochondria of the other? And that we could ask very simply, because we had mutants that were lacking, either CED-4 CED-9.

So in the absence of CED-9, was CED-4 still mitochondrial? In the absence of CED-4, was CED-9 nine still mitochondrial? Again, without going through details, the answer was, CED-4 localization depends upon CED-9. And in the absence of CED-9, you see a very different localization pattern. Here is CED-4 normally, here is CED-4 without CED-9. And where it is, is instead of mitochondria, it's located around the nucleus, in a perinuclear way.

A series of studies that went from here led us to the model that's shown here. And that is that this egl-1 protein, this killer protein, what it does, is it interacts with CED-9 on the mitochondria, displacing, or at least causing the release of CED-4, and the translocation of CED-4 to the nucleus, allowing CED-4 to be active, and thereby activate CED-3 and cause programmed cell death. OK? So that's the basic model. That is our core pathway for programmed cell death.

What I want to do now, is to turn quickly to the next step in this pathway, the step of engulfment. When a cell dies by programmed cell death, it is swallowed and destroyed by another cell. We have analyzed that process in much the same way as we've analyzed the killing process. I won't tell you how we got there, but here's where we got.

There are two parallel and partially redundant pathways that lead to the engulfment of cell corpses. There are three genes that act in the order CED-716 in one pathway, and four genes that have been identified that act in the order 2, 5, 12 together, and 10 in the second pathway. These two pathways appear to be involved in the communication between the dying cell and the cell that is swallowing it. Signal transduction pathways.

In one case, what we believe is going on is that the dying cell uses CED-7, which encodes of protein of a family known as ABC transporters, to basically put on its surface a signal that says come and get me. The engulfing cell has a receptor that recognizes that signal, and then transmits that signal via another protein and a whole pathway, which has yet to be determined.

And parallel to this process, there's another process of intercellular communication. Again, much of it is not yet known. We think the corpse talks to the engulfing cell, and signals, now via a protein complex, that is responsible for activating a protein that is known to be involved in cytoskeletal re-organization. That means changes in cell shape. When one cell swallows another, it changes its shape, and this is how we think it does it.

OK, now I'm not going to talk about the work that led to this model, or the details of this model, but rather use it to introduce another question. So I've been talking about programmed cell death as cellular suicide. But we've been suspicious for some time about this process of engulfment. If you think about it, if you were on the scene and you watched something die, and as that death was occurring somebody else was right there surrounding that individual, you might be suspicious. Now, in fact programmed cell deaths in c. elegans, the engulfment process starts even before the cell division that generates the cell that's going to die is completed.

So we had been suspicious for some time that maybe cell death isn't entirely suicide. Maybe there's a component of murder. But we couldn't get enough evidence to convict. Peter Reddien, a graduate student in the lab, obtained the evidence. So he started with the background knowledge that is a mutant defective and engulfment, programmed cell death still occurs.

So the first order statement is OK, engulfment's not needed for programmed cell death. But Peter said no, I want to go a little further. I want to really address this in a more rigorous way. And what he did, was he took advantage of an assay that had been developed in the lab by postdoc Scott Cameron. And this was an assay that allowed us to look very easily for cells that should die, but instead live. And if you think about it, that's not always the easiest thing to do. Because you've got an animal with an awful lot, 959 doesn't sound like a lot, but if you're going to look for one extra, you better know how to do that in a very precise way. And Scott developed a way to do this.

What he did was, he took advantage of a gene called lin-11, characterized by Gwen Actin when she was a graduate student in the lab, and which she showed was expressed in a set of cells derived from those p cells. Remember, I showed you those p cell divisions and p cell lineage right in the beginning? Well there aren't just two of them, there are actually more than that. There are 12 p cells and a friend called w. And these cells all divide to generate six descendants, except for w, which does only five. And the corresponding descendants of each of these lineages differentiate in the same way, in general, most of them becoming nerve cells for particular types.

But, there's a difference. Six of these cells derive from p-3 through 8, survive to become motor neurons known as vc neurons. But their counterparts from the other cell lineages in the front and back of the animal undergo programmed cell death. Now, if we block programmed cell death, for example, with a mutation in the CED-3 killer gene, then all of these cells survive, and they take on the characteristics of these nerve cells. So, we then could use a reporter, expressed in these nerve cells, and basically use a reporter that's fluorescent, that uses the green fluorescent protein, characterized by Marty Chalfie at Columbia University. And ask if we ever see fluorescent cells in the front or back of the animal, where normally there would be none, in a mutant defective in engulfment.

And I wouldn't tell you this if the answer weren't yes. So here we see a wild type animal. And you see green spots. You also see expression in another cell type here. And a CED-3 mutant, no cell death, lots of green spots. And here's a CED-7 mutant, defective in that signal on the corpse. Says come and get me. And what do you see? You see a cell up here, where in the normal animal there is never one. So you block engulfment, and you get some survival.

Peter quantified it, showed it was real, and showed that it was true for all of the engulfment genes, leading to the conclusion that the engulfment process is necessary for total cell death to occur.

Now Peter also used no Nomarski optics to watch this process in living worms. And what he found was something quite striking. So here is a cartoon version of what I showed you already as a photograph, series of photographs, process of programmed cell death. In a CED-3 mutant this doesn't occur at all. In a mutant defective in engulfment what happens is this process begins. There's no engulfing cell that's active and swallowing it. The process begins, but then it can regress. And in fact, it can be episodic. So that a cell starts to die, comes back, looks normal, starts to die again, comes back, looks normal. Back and forth, and generally dies. But occasionally doesn't.

In the absence of engulfment, it looks like cells are kind of poised on a knife edge, biased towards death, usually die, but not always. Now I think this has implications for biology, and also possibly for therapy. Implicit in what I've told you already about the involvement of programmed cell death, in a variety of diseases, is the fact that these genes define potential targets for intervention. If you block the process of cell death, in some disorders that would be good. If you cause the process of cell death, in some disorders that would be good too.

Now, obviously we have to be careful. And as with all biology, genes and targets aren't enough. To know either the biology or the therapeutics, you have to really do the biological studies. But there are ways of pursuing that. And that, in fact, is being done. From these studies I want to add the comment that maybe the engulfment process involves another biologic phenomena and another set of targets. Because in many of these disorders, it's likely that in fact cells are poised on such a knife edge. And maybe we can push them one way or the other.

OK, now, having said that, I'd like to finish basically by telling you a little bit of some of the newest. And that is that Peter used this assay that I just told you about here, oh and I should say, this was the conclusion, engulfment is responsible. He used this assay, now very sensitive assay, looking for extra green fluorescent spots to try to find new genes that might have a partial role, like the in engulfment genes, in the cell killing process.

And what he found, again, by using this green fluorescent protein assay, and looking for mutations that would tweak the number of living or dying cells, what he found, for example, in a so-called enhancer screen, looking to make the effects of a week CED-3 mutations stronger, making a partial killer into a stronger killer, he found mutations, which we call n3376 3380, the 3376th and 80th mutation characterized in the lab. If these two mutations define two genes involved in the killing process that we didn't know were involved in the killing process.

The first of them encodes a novel protein that has a motif that is suggestive of interactions with other macromolecules, and as I'll tell you in a moment, we think it's actually interaction with DNA. The gene is called mcd-1, for modifier of cell death. And the second gene, much to our surprise, was an old friend. The second gene, the second mutation, turned out to be an allele of a gene that we called dpl-1. From studies we had done in what I thought was a totally unrelated project of an oncogene pathway involved in cell signaling in c. elegans development. Oncogene, the RAS oncogene. An oncogene discovered here by Bob Weinberg some years ago.

Now, dp proteins are known in mammals, they affect cell proliferation, and they also affect cell survival, and cell death. But nobody knows how. They act by interacting with other proteins, like e2-f. And they interact with Rb proteins, retinoblastoma, another gene that's been characterized by Bob Weinberg and by Tyler Jacques here at MIT.

So, we had already analyzed the roles of dp protein, and e2-f protein, and an Rb protein in the c. elegans RAS pathway. We had all the reagents. And Peter said, well, are they also involved in this process of cell death? And again, don't worry about the details, just worry about the title. The answer is yes, dp, e-2f, and Rb in the worm are all involved in programmed cell death.

What they do, we don't know. But given what we do know about what they do in other contexts, what we postulate is they act by controlling transcription, RNA synthesis, gene expression. We know that they act downstream of the protector gene CED-9. And we have two basic hypotheses: either that they are regulating the expression of the killer gene CED-3 and/or CED-4, or that they interface in this pathway in a different way, possibly with genes involved in the cell cycle, something we know that these genes can do. And that's something that we are currently exploring.

Now, the last step in this pathway, you'll recall, is that of degrading the cellular corpse. And there's one gene in this step that was identified many years ago by John Sulston. It's a gene he called nuc-1, possibly a bad name, because he showed that it controlled the nuclease. And what happens is that this nuclease is involved in degrading the DNA in cell corpses. And if it's not active, the DNA remains. But other aspects of the death process occur, and one is left with these little kinetic nuclei, as shown here. And that's how John found this nuc. He showed that it controlled the activity of a nuclease, but it wasn't clear whether this control was direct or indirect. And it also wasn't obviously at all clear what kind of nuclease it might be. And if you read the cell death literature, there are a large number of nucleases, each of which has been implicated in the process of programmed cell death. In one study or another.

So Gillian Stanfield was a graduate student in the lab, cloned and molecularly characterized nuc-1, and found that indeed it does directly encode a nuclease, and a nuclease of the sort DNase-2. Which made some people very happy, and other people very disappointed. But, we could add, this wasn't all there was to it, and studies that Julian did with graduate student Yi-Chun Wu, she showed that in fact there are at least three nucleases involved in this process, two of which are not identified. The field is very happy.

OK, so we take all of this together, and we end up with today's molecular genetic pathway for programmed cell death. Again, we have the four general steps that I described. I've told you about molecular, genetic, and developmental studies for some, but certainly not all, of these players. But what I will say is that essentially all of these players have human counterparts. And insofar as they've been analyzed, they seem to be involved in the process of programmed cell death.

Some of these have counterparts that are associated with other diseases. For example, these two genes here, ces-1 and ces-2, cell that specification, have human counterparts that are associated with ALL, acute lymphoblastic leukemia, another cancer. And there are other counterparts through most of this pathway. So from this, we conclude that much, and possibly all of this pathway, is conserved amongst organisms as superficially different as a microscopic worm and us.

Now, I should say that if one looks at what we know today about worms, and what we know today about human programmed cell death, mammalian more general, there are apparent differences. We don't know yet if these differences are real biologic differences, or if these differences are in part like the differences of the wise men who expected the elephant, and one looked at the trunk, and one looked at the tail, and one looked at the leg, and had somewhat different impressions. We know we don't know at all, we're looking to find out more.

Let me put this, then, in a broader perspective. I think, if one looks at the major findings, maybe the major finding, in the field of molecular genetics over the last 5 to 10 years, what we as a field would say is that what has really been striking is a phenomenon that I like to refer to as the principle of biologic universality. In other words, despite the diversity of organisms, we share a lot in common. Worms and people, and the fruit fly drosophila, and a yeast cell that's used for baking bread or making beer. There are genes and gene pathways in all of us that are strikingly similar. And we can learn about each from studies of the other. And that, I think, maybe I should say I thought, was the big news.

Until I saw this. The cover of Punch Magazine, from 1882, where clearly the artist knew more than I did, until very recently, in saying, man is but a worm. With that the people, members of my laboratory over the years who have been involved in the study of programmed cell death, current members involved in that project, members of the lab involved in a variety of projects today. And I actually would like to ask the members of my lab, past and present who are here today, please stand up. Because it's you who are being recognized.

[APPLAUSE]

MODERATOR: Keeping to an old tradition, when you give a talk, and you give new data, and you express new ideas, you take questions. Bob has done that, and Bob deserves a couple of questions. We'll take a couple of questions from the audience. Yes.

AUDIENCE: [INAUDIBLE] was talking about, but after hearing him, I was wondering, how close are you to practice this in human genes?

HORVITZ: To-- the question is how close are we to practice this? And the question is, do you mean to use this information for treating disease, or to study this process? To study this process. There are many studies ongoing today in human beings, in other mammals, that are drawing upon this information, and also deriving brand new information in lots of exciting ways. So this is very actively on-going.

AUDIENCE: My question would be, what turns those genes on to start the process [INAUDIBLE]?

HORVITZ: Right, what turns those genes on? OK, that's my next half seminar. So, the pathway that I told you about began with this gene egl-1, the bcl-2-like killer gene. And in fact, we've shown that this gene is regulated at the level of gene expression, RNA synthesis, transcription. We know one of the transcriptional regulators. And basically, it causes this gene to be off in certain cells that should live, and when that regulator is not active, and I can tell you how that happens, but it's a longer story, then this gene is on.