35th Annual Killian Award Lecture—H. Robert Horvitz (2007)

Search transcript...

LERMAN: Ladies and gentlemen, colleagues, visitors, guests, welcome. My name is Steve Lerman. I'm serving as chair of the faculty, and one of the great pleasures of being the chair of the faculty is the honor of introducing the recipient of the James Killian Award for Faculty Achievement. This award comes with the privilege of, of course, addressing, giving a Killian lecture. Today's lecture will finish around, more or less, 5:30. If time allows, there will be question and answer. And I just want to remind everybody there's a reception at the end of today's session.

Today it's my privilege to introduce our colleague Professor H. Robert Horvitz, who is the recipient of the Killian award for 2006-2007. As noted in the citation for this award, which is a resolution to the faculty, a copy of which you have, Bob Horvitz is a true pioneer in the study of a simple but remarkable roundworm, C. elegans. During this worm's development, this extraordinary organism develops precisely 1,090 cells of which 131 are eliminated through a process of programmed cell death. Through a series of wonderfully crafted elegant experiments Bob and his research group made, and I'll quote the citation, "seminal discoveries in the area of cell differentiation, programmed cell death, and nerve cell function. "

These discoveries proved to be crucial to our understanding of the development of all multicellular organisms. They've shed incredible insight into processes as diverse as aging, neural development, and behavior. Among the great number of national and international awards and honors that have been bestowed upon him, Bob was given the Nobel Prize in physiology or medicine in 2002.

As is the custom, the website for the Nobel Prize includes an autobiography, which in Bob's case is a wonderful window into the background behind his amazing career. His description of his undergraduate years as a mathematics and economics major here at MIT include a detailed description of his uniquely MIT interests in diverse topics, such as studying and applying probability theory in pursuit of blackjack and poker winnings, writing programs for early generations of IBM computers, and wiring the panels for accounting machines. Among his accomplishments listed in again, this publicly available autobiography, so I'm not telling tales out of school among his accomplishments include being ejected from a Las Vegas casino for having spotted a stacked deck of cards.

His interest in biology started after taking courses in the field during his senior year at MIT. After graduating from MIT, he went to Harvard and received his MA and Ph.D from there. Following a post-doc at Cambridge University, he returned to MIT in 1978 to join our faculty.

It is often said that we know a great deal about an individual from the company he keeps. Bob's autobiography notes that he had four official research supervisors during his student and post-doc years. These were Robert Solow, James Watson, Wally Gilbert, and Sydney Brenner. And, as Bob notes and those of you who keep track of such things, all four of these people who are influential in his career and his development as a scholar, researcher, teacher, and, of course, Bob himself are Nobel Laureates.

It's our great fortune that Bob's entire 27-year career as a faculty member has been spent right here at MIT. He has taught literally thousands of students and been a mentor to an entire generation of graduate students and post-docs who have been in his lab. Many of these former students now populate the top research institutions not just here in the United States but in the world.

Please join me in welcoming this outstanding teacher, scientist, mentor, and scholar to the podium. Bob. Just to get away a bit of business, this is the citation, the official award, the Killian Award, which I'm pleased to hand you. And with that, the floor is yours.

HORVITZ: Well, I have to say thank you for the kind introduction. You had warned me that I didn't know indeed what to expect. That was accurate, but the stories also are accurate.

I want to start by thanking everybody for coming today. It's an enormous honor to be chosen by my colleagues for the James R. Killian Faculty Achievement Award. And I thank, in particular, the selection committee as well as the many members of the MIT community who have supported me over the years and helped me in so many different ways.

I knew Jim Killian. Jim Killian was a great man and he was a great leader at MIT for many years. When I was an undergraduate here, Jim Killian was the chairman of the MIT Corporation. I talked with Jim on numerous occasions, and I had the privilege of being at a number of events that he hosted at his apartment at 100 Memorial Drive. And I think that just the fact that as an undergraduate I got to know the chairman of the MIT Corporation, says something fundamental and something wonderful about Jim Killian.

The Killian Award is special to me in another way, and that is the previous honorees. This is the 35th year of the Killian Award. One of the first awardees was my undergraduate thesis advisor, as you've just heard, Bob Solow. Bob I held in awe when I was an undergraduate, and I must say I continue to hold in awe today.

Other Killian Award recipients include enormously accomplished close colleagues and good friends in a variety of departments. And I am very honored to be following in their footsteps, and I must say they are tough acts to follow.

So you've heard a little bit about my history. I'll try to run through it from my vantage point skipping some of the details that you've heard alluded to. I came to MIT in 1964 as a freshman. To put that into context, at that point Paul Gray was a 32-year-old associate professor and he was then or was soon to become the chairman of the Freshman Advisory Council.

I graduated, as you heard, with degrees in mathematics. For those of you who need the translation, that's course 18. And also in economics, that's 14. I moved upriver to another Cambridge University for my Ph.D studies, and that's where I began my serious inquiry into biology. And at Harvard, I had two advisors again, you've heard. One is James Watson, who's obviously known for the discovery of the DNA double helix with Francis Crick and also Wally Gilbert, who is known for the development of methods to sequence DNA and also for the founding of Biogen, which amongst others he did with Phil Sharp, a good friend and colleague here.

My next move was to another Cambridge. This time, some people tell me the real Cambridge, in England. And it was there I began as a post-doctoral fellow with Sydney Brenner the studies of the microscopic worm C. elegans, which I will describe a bit of to you today.

I returned again, as you heard, to MIT in 1978, and I have been here ever since. Now before planning this lecture, I tried to figure out exactly what my instructions were, and as far as I could tell the only instructions that I have is to communicate to the MIT community something about what I've been doing for the past nearly 30 years.

I think of myself as having worn four hats. The biggest hat by far is my MIT hat. And here, I've split my time among teaching, research, and trying to help out both departmentally and institutionally. And I must say over this period I have had many interesting experiences. I can say parenthetically that when my mother uses the word interesting it has a variety of connotations some of which, no doubt, would be applicable here.

The experience that probably has had the most visible consequence was my serving as the chair of the building committee for the current biology building, the Koch Building, Building 68. And in this process I learned about architecture and architects, building design and building codes, and also about some of my colleagues' most private thoughts about other colleagues. I helped define adjacencies for the new building, who wanted to be or not to be near whom.

My second hat is from MGH, the Massachusetts General Hospital. About 20 years ago, I became involved in a program of teaching MDs at MGH. And before long, I had appointments in medicine and neurology and a white coat and a badge that said Ph.D. And I also had an active collaboration developed with neurologist Bob Brown. That's not the ex-MIT provost, it's the MGH neurologist. A wonderful man and we have been working since then on the genetics of Lou Gehrig's disease, ALS, amyotrophic lateral sclerosis. That collaboration led to the identification of the first ALS gene and it continues to this day.

My third hat is commercial. I have founded four biotechnology companies, and I've been involved in a total of 12 biotech, pharmaceutical, or related commercial companies.

My fourth hat involves science policy and science advocacy. I have been very concerned about and actively involved in federal science policy and the federal support of scientific research, and I'm going to return to this topic explicitly at the end of this lecture.

Now for me, these different involvements have been stimulating, complimentary, and synergistic. I've learned how to lead a basic research laboratory, how to help with medical research, and how to advise companies and members of the government about how and why to use basic and clinical research to drive new knowledge and to drive novel practical applications.

Now today what I'm going to do mostly is focus on my basic research here at MIT. And, as I alluded to, at the end, I'm going to turn back to the issue of the relationship between basic research and national science policy.

Okay, so I'm going to start with the science. I've entitled this lecture Worms, Life, and Death, and I'm going to discuss some of our efforts concerning basic problems in development and biology and, in particular, problems in the field of developmental biology.

Now there's not going to be time for me to describe many of the efforts of outstanding young scientists who have been in my laboratory and are there now. And I'm not even going to touch upon our studies, for example, of aging, of chromatin remodelling, and even of neural biology, which has been a longstanding major interest. Instead, I will focus on a few specific aspects of animal development.

Animal development begins with a single cell, the fertilized egg, which divides to make two cells. These two divide again and again generating large numbers of cells. For example, in us, as many as 10 to the 13th cells. Then each of these cells must decide what to do and take on specific characteristics becoming, for example, a nerve cell or a muscle cell or a skin cell.

And the first problem that I'm going to discuss is what we've learned about the generation of cell diversity. How are different kinds of cells generated during this process? The second problem, which I'll touch on only very briefly, involves the fact that during development, cells communicate. They communicate within cells and cells communicate with each other. And I'll discuss aspects of intra- and intercellular signaling.

Then I'll spend most of my time on another feature of animal development. Quite remarkably, many of the cells that are generated as animals develop do not survive but instead die. This naturally occurring cell death, which is often referred to as programmed cell death will be the focus of most of my lecture.

So that introduces two-thirds of the title, life and death. What about worms?

This is a worm. The microscopic nematode, or roundworm, Caenorhabditis elegans. This animal was introduced to modern biology as an experimental organism by Sydney Brenner, who is at the Medical Research Council Laboratory or Laboratory of Molecular Biology in Cambridge, England. And as you've heard, it was with Sydney I was a post-doctoral fellow.

Now C. Elegans is a remarkably simple animal. For example, whereas a human or any other mammal has as I said about 10 to the 13th cells and an insect has something like 10 to the fifth to 10 to the seventh, depending upon the insect, C. elegans has a total of 959 cells.

About 25 years ago, John Sulston, also at the MRC LMB in Cambridge, England, with a little help from his friends, myself included, John determined the complete pattern of cell divisions of C. elegans. The so-called cell lineage, which shows what happens as a single cell divides to make two, each of those two divide again. The y-axis is time and this cell lineage diagram depicts the developmental origin of every cell in the animal.

Now John Sulston is a great man, and for those of you who have not had the privilege of knowing him, the next slide shows his photograph. John is right here, the one on the left. Unfortunately, the photograph is incorrectly labeled, because in this photo he is called the Reverend Gabriel Burdett.

According to the accompanying article, the Reverend Burdett is a proponent of a new theory. It's called the theory of intelligent falling. It's an alternative to the theory of gravity. And the Reverend Burdett is quoted, as you can see here, as saying that "things fall not because they are acted upon by some gravitational force but rather because a higher intelligence is pushing them down."

This article was published in the Onion, a political satire newspaper which many of you know from Madison, Wisconsin. And if it is, of course, a commentary on intelligent design. The sad part is the number of people who took this article seriously.

Now in case you're wondering what it is John was actually doing-- it is a real photograph-- when the photograph was taken. the answer comes from the original article, which shows very simply that he was announcing the sequence of the human genome.

Okay, so back to John's, are the worm's cell lineage. So this lineage generates the 959 cells that are found in the adult animal. One long-term goal of our laboratory at MIT has been to understand the mechanisms responsible for generating this pattern of cell divisions and the diversity of cells that are generated by this lineage.

And to address this problem, we've used genetics. Identify mutant animals that are abnormal in cell lineage. Mutants are abnormal in genes, and so by doing this we could analyze the mutants to identify and then characterize genes that control cell lineage.

Over the years, our laboratory has isolated and characterized a large number of such mutants. As of last week, we had 5,113 mutants isolated in the lab and many of them affect the worm cell lineage. I'll tell you about a few.

There are mutants that perturb developmental timing of specific aspects of this lineage. We call such mutants heterochronic, and the first heterochronic mutant, in fact, I characterize with John Sulston and Marty Chalfie, who's now at Columbia University. The heterochronic genes were then characterized in some detail by two post-doctoral Fellows, Victor Ambros and Gary Ruvkun, first in my laboratory and then over the years since in their own laboratories.

Heterochronic mutations are generally of two classes as exemplified here. Very simply put, if in normal development we see a series of events one, two, three, four, in a normal wild type animal. In one class of mutant, retarded mutants, normally early events occur later. So instead of going one, two, three, four, this particular one goes one, one, two, two. So you get a normally one-stage event happening when two-stage events should happen. It's retarded. Things go too slowly.

By contrast, there are precocious mutants in which things happen earlier than they should. So in this particular mutant, instead of going one, two, three, four, it goes two, three, four. So two-stage events happen when first-stage events should happen.

These heterochronic mutants defined a series of genes that turned out to be pivotal in controlling developmental timing. Victor's and Gary's laboratories discovered that two of these genes do not encode the standard gene products. Most genes encode proteins. These genes did not. They encoded tiny 21- or 22-base pair long RNAs.

These RNAs were the founding members of a family of RNAs now called microRNAs, and they've proved to be widespread in biology. C. elegans has about a hundred of these, and other organisms, including insects, plants, and mammals have at least as many. And they have proved to be major regulators of gene activity in essentially all multicellular organisms. MicroRNAs are now the subject of intensive study in a variety of laboratories, including a number of laboratories here at MIT.

Now others of our cell lineage mutants block the generation of cell diversity at specific cell divisions. For example, look back at the cell lineage and you see something that looks very complicated. But we can simplify it to that. What this diagram indicates is that we can regard every cell in the cell lineage as having a fate. The fate may be to differentiate into a particular cell type, a muscle cell, an intestinal cell, a nerve cell of a particular class. Or the fate may be to divide in a particular pattern and generate a particular complement of descendant cells.

In this context, at each cell division a cell with one fate A can be said to generate two daughters with fates B and C where, in general, B and C differ both from each other and from the fate of the mother cell. Some of the mutants we found cause sister cells to be identical to each other or daughter cells instead of becoming different, to be like their mothers.

Since these mutants were defective in genes we could conclude that those genes defined by these mutants drove either what makes sisters different or what makes daughters different from others. In other words, drove aspects of the generation of cell diversity.

We characterized many of these genes in a variety of ways and one of the most interesting and, I think, most important findings to emerge is that most of the genes we identified have counterparts in other organisms, including humans. For example, the worm cell lineage genes we identified include founding members of well known gene families, for the biologists in the audience, Powell, Lim, we could go on with names came out of studies of worm cell lineage in this way. In addition, by studying groups of genes with related effects on the same aspect of the worm cell lineage, we identified not only single genes that are responsible for specific developmental steps but, in fact, complete gene pathways.

For example, we studied a set of genes involved in organ development, specifically the development of the vulva of the C. elegans hermaphrodite. The vulva defines the opening between the uterus and the external environment and the animal uses it for egg laying. And there are two classes of mutants that we identified.

First of all, if you look here at a normal animal, the vulva is located here. Here's an animal that is vulvaless, it has no vulva. And what happens in this case, is that eggs cannot be released. And I then have to do this quick aside and tell you a little bit more about C. elegans' biology.

C. elegans has two sexes. One is male, the other is not. The second sex is a hermaphrodite. The hermaphrodite produces both eggs and sperm and it is internally self-fertilizing. The hermaphrodite can mate with males, but she can reproduce all by herself. In other words, the hermaphrodite is entirely self-sufficient.

In a vulvaless hermaphrodite, eggs are fertilized internally, they develop in utero, and they hatch. And unfortunately for the animal that was both mother and father to them, what they then do is devour their parent. So this is actually 50 little worms in the shell of what was their parent. Those are vulvaless mutants.

Multivulva mutants have a multiple number of vulva-like structures. Many of the genes defined by such vulvaless and multivulva mutants proved to be involved in the cell interactions that drive the development of the vulva. And we studied such genes in detail, as did a number of people after they left our lab, most notably ex-graduate student Paul Sternberg, who went to Cal Tech.

And these studies together defined a pathway known as the Ras pathway. This is a fundamental and arguably the best understood signal transduction pathway in biology. It's basic to animal development, and it's also key in processes that generate human cancers.

I note that this key gene, Ras, was discovered as a viral cancer gene, an oncogene, by Ed Skolnick, who's now here at the Broad Institute. And it was identified as the first cellular oncogene by Bob Weinberg, who's been here at MIT for many years.

Now as we were beginning our genetic studies of the C. elegans cell lineage, one aspect of this lineage particularly caught our attention. As you heard from the introduction, in addition to the 959 cells that are generated and found in the adult, there are 131 cells that are generated and not found in the adult. They are not in the adult because they die. Because these deaths are an aspect of the normal so-called program of development, we refer to them as programmed cell deaths.

Before I turn to our studies of programmed cell death, I want to introduce you a bit to the topic more generally. Programmed cell death occurs as a normal aspect of the development of every animal that has been studied. The phrase is also used to refer to deaths that employ the same mechanisms as these naturally occurring deaths, and I'll give you some examples.

During metamorphosis from a tadpole to a frog, the tadpole loses its tail. The cells die by programmed cell death. In us and many other organisms, in utero we have interdigital cellular regions of webbing, which are sculpted out by the process of programmed cell death.

Birds are much the same. Chickens have feet without webbing. They've lost the webbed region by programmed cell death. Ducks, on the other hand, have webbed feet and that programmed cell death does not occur.

I should say for people here, you may recognize some of these birds. This goose lives by the Charles River just across the way, and this pigeon was found by the food trucks right by the MIT biology building. These two I don't know personally.

Now programmed cell death appears to be a major and universal aspect of animal development. Nonetheless, until not many years ago, biologists when they thought about cell death, and I should probably say biologists if they thought about cell death because most did not, simply considered cell death to be what happens to cells when cells were unhappy. If they became too damaged to live, they died. We know today that often this is not the case and instead, cell death can be an active process on the part of cells to die, and that specific genes can act in cells to make those cells die.

So we can say that there is a biology of cell death every bit as much as there is a biology of other basic processes like cell division, cell migration, and cell differentiation. And not surprisingly, where there is a biology there can be a pathology. If normal biology goes wrong in humans, then something is going to lead to disease and programmed cell death is no exception.

And there are many disorders, and this is just a short list, that have now been associated with abnormalities and programmed cell death. For example, the neurodegenerative disorders Alzheimer's, Parkinson's, Huntington's, ALS, all involve neuronal cell death. In each case, specific classes of nerve cells die, which leads to the particular clinical features of each of the neurologic disorders.

One hypothesis, but I want to underscore that word, one hypothesis today is that in at least some of these disorders, what is going on is, in essence, ectopic programmed cell death. So that nerve cells are dying by the process, the mechanism of programmed cell death but at the wrong time, or the process is being expressed in the wrong cells.

That's known to be the case for certain disorders. Certain retinal degenerations are clearly of this class, and for the other disorders it's still an open question. Many other human disorders involve too much cell death, and in many cases, the cell death has been shown to be ectopic programmed death. In other words, using those same mechanisms.

Conversely, some disorders involve too little cell death. For example, consider cancer, which we normally think about as being uncontrolled cell division. That's true, but it's more complicated than that. The number of cells in our tissues is defined by an equilibrium of two opposing processes. Cell addition by cell division and cell depletion by programmed cell death.

The number of cells can go up either by amplifying the rate of cell division or by decreasing the rate of cell death. And certain cancers, in fact, are fundamentally caused by too little programmed cell death. Follicular lymphoma is one striking example of such a cancer.

Furthermore, most and possibly all cancers involve a down regulation of programmed cell death and the major cancer treatments chemotherapy, radiation therapy, work by activating the process of programmed cell death. And it's for these reasons that the identification of the genes and the proteins that function in programmed cell death have provided new targets for possible intervention in diseases as diverse as neurodegenerative disorders, AIDS, cancer, and autoimmune disease.

And what I'm going to tell you about now is some of the work from my lab that has identified such genes and proteins. We've taken, as I described before for cell lineage, a genetic approach. We've identified mutants that are abnormal in the process of cell death and use that to define a four-step genetic pathway.

The first step is that of deciding. Every cell has to decide, am I going to live or am I going to die by programmed cell death? The second step is that of cell killing. The third step is one of engulfment whereby dying cells are engulfed by neighboring cells and thereby removed from the body. And the fourth step is one of degradation of the cellular debris of the cell that was dying. If this sounds too formalistic, you can remember the four steps in a very simple way.

What I'm going to focus on first is the killing step. Two graduate students, Hilary Ellis and Chand Desai, discovered two genes that are required for programmed cell death. They isolated mutant animals in which essentially all 131 cells that normally die, instead survive. Each mutant was abnormal in a single gene. They analyzed the genes and they named them CED-3 and CED-4 where CED, C-E-D, stands for cell death abnormal.

Each deathless mutant lacked the activity of either CED-3 or CED-4 So that revealed that the activities of these genes are needed for cells to die by programmed cell death. In short, that these are killer genes This finding was key because it was this finding that demonstrated that programmed cell death requires the function of specific genes and hence that programmed cell death is an active biologic process.

Graduate student Junying Yuan then showed that CED-3 and CED-4 both act within the cells that die and that established that, to at least to this extent, programmed cell death is a process of cellular suicide. Junying and another graduate student, Shai Shaham, molecularly characterized CED-3 and found that it looked like a protein that had been identified by two pharmaceutical companies interested in human inflammatory disease. This protein called ICE, interleukin-1 beta converting enzyme, if you want the details, converts the pro-form of a cytokine, a signaling molecule, into the active molecule.

Junying then, in her own lab, showed that if she expressed the worm CED-3 gene in a mammalian cell, she could cause the mammalian cell to undergo programmed cell death. These two proteins, CED-3 and ICE, proved to be the founding members of a protein family today referred to as caspases.

Caspases are proteases, enzymes that cleave other enzymes. This finding and others of our subsequent studies demonstrated that CED-3, the killer gene, functions in cells in C. elegans by acting as a cysteine protease. This was the first molecular mechanism for programmed cell death.

Given this observation, we also suggested, particularly given that ICE was in human cells, we suggested that CED-3 ICE-like proteases were likely to be involved in programmed cell death in mammals, including humans. And, in fact, today we know of about a dozen different mammalian caspases, many of which have been shown to function in programmed cell death.

Junying Yuan also characterized molecularly the gene CED-4 and five years later, work in Dallas, Texas identified a human CED-4 counterpart called Apaf-1 and showed that it, like CED-4, acts to promote programmed cell death. So CED-3 and CED-4 have conserved functions between worms and humans.

Then another graduate student, Ron Ellis, discovered yet another gene involved in programmed cell death. We call this gene CED-9. CED-9 is different from CED-3 and CED-4. CED-3 and CED-4 are killers. CED-9 is a protector. It protects cells against programmed cell death.

Graduate student Michael Hengartner characterized CED-9 and discovered that it looks like a protein that is a human cancer gene. A protein called BCL-2, which stands for B-cell lymphoma and is causally implicated in follicular lymphoma.

BCL-2, like CED-9, protects cells against programmed cell death and the way in which BCL-2 causes cancer is by protecting cells that should commit suicide from doing so, allowing the number of B-cells to then proliferate.

Soon thereafter, at Stanford University, David [? Vo ?] working with Stuart Kim and Irv Weissman, demonstrated that the human BCL-2 gene, if put into C. elegans works. It will protect against programmed cell death. And Michael Hengartner, in my lab, confirmed this result and took it a step further.

What he showed was that if you took the human gene and put it in a worm that lacked CED-9 function, the human gene would substitute for CED-9. What that said is not only are single genes similar in structure and capable of functioning in analogous ways, the human gene could interface with that endogenous pathway in a worm, and therefore CED-9 and BCL-2 must act in similar molecular genetic pathways.

By analyzing various genetic interactions, graduate student Shai Shaham helped define this pathway. And what he found was what's shown here. CED-3 kills, CED-4 kills by promoting the activity of CED-3 and CED-9 protects by preventing CED-4 from promoting the activity of CED-3. And given the human counterparts, we propose that the human proteins were likely to act in a similar pathway, which indeed has proved to be the case.

Now these human counterparts define potential therapeutic targets. For example, consider a disease in which there is too much programmed cell death, be it retinal degeneration or a heart attack. If we could inhibit a killer gene, a caspace--

Oops, I'm sorry. I have skipped a little bit. I have to tell you one more bit of the story first. I back up. We'll get to therapy in a second.

The pathway. A key question, CED-9, if it's active, cells live. If it's inactive cells die. What controls CED-9? Barbara Conradt, a post-doc, obtained the answer to this. And the answer was EGL-1. EGL-1, like CED-3 and CED-4, is a killer, but unlike CED-3 and CED-4, it acts upstream of, not downstream of CED-9. So EGL-1 kills by preventing CED-9 from preventing CED-4 four from activating CED-3.

OK now, given the four, what about therapeutics? And again, as I was starting to say, if we could inhibit a killer, we should be able to prevent the pathological process of programmed cell death. So, for example, inhibit a caspase and cause cells that, in disease, are dying by programmed cell death, to then survive in a variety of diseases, such as the ones indicated here.

Conversely, for diseases in which there is too little programmed cell death, for example, in certain cancers as I've described, if we could activate the cell death pathway by inhibiting an inhibitor of cell death, as CED-9/BCL-2-like protein, we should be able to cause cells to die that have otherwise escaped death.

So with these thoughts in mind, our identification of the genes and proteins that function in programmed cell death has provided new targets for possible intervention in a wide variety of human disorders. And I should say that therapeutics against the two targets that I've just described, caspases and CED-9/BCL-2 proteins, have been developed by a biotechnology company that I founded and clinical trials are in progress for both of these targets. For caspases they are in phase three. For BCL-2 antagonists they are in phase one. And both of these clinical trials are in the hands of major pharmaceutical companies at present.

Okay, so now let's return to the core pathway. In this pathway, EGL-1 is key. If EGL-1 is active, cells die. If EGL-1 is inactive, cells live. So now we have to go a step back. What controls EGL-1?

In an extensive series of studies, we've provided some answers to this question. It turns out that EGL-1 is not controlled globally. All of these cells are globally involved in programmed cell death. But rather, EGL-1 is controlled on a cell-by-cell basis. Each cell decides should EGL-1 be active or not. If active, it dies. If not, it lives.

And more specifically what we found, is that in many cases the function of the EGL-1 gene is regulated at the level of the transcription of the EGL-1 messenger RNA on a cell-by-cell basis. Different sets of so-called transcription factors determine in different cells whether EGL-1 is copied into an RNA and then subsequently made into the EGL-1 protein.

And I'm going to describe our conclusions from a few of these studies. First, we'll think about some nerve cells called the NSM sisters. The NSMs are particular neurons located in the head of the animal here. And the NSM sister cells normally die. So they're not present in a normal animal, wild type animal. But if these cells survive, you can find them located as indicated here. There are two of them. One on each side, right and left.

The deaths of the NSM sisters were studied by graduate students Ron Ellis and Mark Metzstein and also by post-doc Barbara Conradt, first in our lab and then Barbara in their own lab thereafter. You've seen pictures of Ron and Barbara. You haven't seen Mark who is here.

What these studies have indicated is that a set of transcription factors controls EGL-1 RNA synthesis in the NSM sisters. And interestingly, these factors have human counterparts. The human counterparts are involved in programmed cell death, and they have been implicated in human cancer in particular in ALL, acute lymphoblastic leukemia.

The fact that these human cancer genes drive cancerous growth because of their role in programmed cell death was discovered as a direct consequence of our identification of these worm genes as homologues of human cancer genes.

Now let me consider a second example, the life versus death decision of another set of nerve cells, cells called the HSNs. These nerve cells are sexually dimorphic. They're present in hermaphrodites but not males. In hermaphrodites, they drive egg laying. Males don't lay eggs. They don't have HSN neurons. And what happens is in both sexes these nerve cells are generated but in males they undergo programmed cell death.

So this is a sexually dimorphic programmed cell death. What controls this? The answer is very similar to what I've just described. Post-doc Barbara Conradt identified a set of transcription factors, a different set of transcription factors from those I've just alluded to, that regulate EGL-1 synthesis in the HSNs and they have counterparts that act in humans in programmed cell death, also known as apoptosis, and in cancer.

Specifically the worm transcription factor TRA-1 is a member of the glioblastoma family of proteins. One member of which, GLI-1 is overexpressed in most, if not all, basal cell carcinomas. The counterpart of another of these worm transcription factors, EOR-1, has been implicated in acute progranulocytic leukemia and it is thought to promote cancer by affecting apoptosis, programmed cell death.

From these studies and from a series of other studies, studies that have been done by a number of people in the lab-- Hillel Schwartz, Brendan Galvin, are graduate students currently in that lab. Takashi Hirose, Hirose is a post-doc in the lab and Barbara Conradt and Scott Cameron, both were post-docs who have now left the lab. From their studies and the studies I've just described, we've come to an overall picture that is depicted here.

The cell-type-specific control of whether particular cells live or die is determined by whether or not the killer gene EGL-1 is expressed in that cell. And we have identified, in some detail, precisely how this decision is made in each of a variety of different cell types.

In addition, as I've already noted, mammalian counterparts of the worm genes that control whether particular cells live or die, mammalian counterparts are, at least in some cases, involved in cancer.

Now interestingly, we've recently found that such genes can be involved in other disorders. For example, graduate student Hillel Schwartz has been analyzing the life versus death decision of another set of nerve cells, nerve cells called the CEMs. These, like the HSNs, are sexually dimorphic, but this time it's opposite. These cells are generated in both sexes and they live in males and die in hermaphrodites. And in fact, in males they become sensory nerve cells that the male uses to chemically sense the presence of the opposite sex, the hermaphrodite.

Now Hillel has identified and characterized a gene called CEH-30. And it turns out that CEH-30, protects these nerve cells from dying in males. So that when CEH-30 is active, which it is in males, these neurons survive, and in hermaphrodites, CEH-30 is turned off and these cells die.

Like the other cell-type-specific regulators for programmed cell death I've told you about, CEH-30 encodes a transcription factor, this is for a particular class known as a homeodomain protein, and obviously this raises the question, what about mammalian counterparts?

It turns out that a mutation in a mammalian counterpart of CEH-30 causes a disorder in a mouse counterpart. That's the only one that's been analyzed so far. And the disorder is not cancer, it's deafness. There is a loss of hearing caused by the progressive degeneration of the hair cells in the ear.

Now how might this be? Well, if we think about what we know about CEH-30, CEH-30 protects particular nerve cells from dying, and when it's turned off, those nerve cells die. In this mouse, the CEH-30 counterpart is inactivated by mutation and hair cells are lost.

Our hypothesis, very simply, is that the functioning of this mammalian gene is very comparable and by losing protection of the hair cells, those hair cells die and then deafness results. And in fact, based upon this finding, studies in the lab that's been working on this particular mouse gene seem to confirm that that indeed is the case.

So we can summarize this particular section by saying that our studies to date of worm genes that control whether individual cells live or die have led us to genes with mammalian counterparts involved in such diverse disorders as cancer and deafness. And what we suspect is that this is just the tip of the iceberg.

OK, so let me now share with you a not quite up-to-date summary of the overall pathway of programmed cell death. Luckily for you I've discussed only a few of these genes, but the message is much the same. Most and possibly all of the genes here have human counterparts and I believe that many of these counterparts are likely to define novel therapeutic targets for a broad variety of human disorders.

Now I want to transition. I want to put the discoveries I've discussed into a somewhat broader context. The work I've described involved absolutely basic research. When I began, neither the generality nor the application of our efforts was at all clear.

C. elegans, as Susan and I were just discussing, was an obscure organism. Genetic studies are often highly abstract. I did not target any disease. And in fact, I didn't have any idea if what we found would be relevant to any organism other than this microscopic roundworm. Nonetheless, our studies established mechanisms that appear to be universal, and our findings may well help provide the basis for new treatments for a broad variety of human diseases.

And I think there are two important messages here. First, basic research, discovery based research, may lead not only to intellectually stimulating findings, and I believe that that, in and of itself, is exceedingly important, but also will lead to insights of major practical import. Basic research is the driver of scientific knowledge.

Secondly, extensive true basic research must be supported outside of the private sector by governments and by foundations because only such organizations can act based upon the fact that discoveries from basic research will benefit humanity but in ways that can't be predicted, in ways that can't possibly constitute a business plan in the private sector.

So what does this mean? For biomedical research, it means the NIH, the National Institutes of Health. The NIH has been, and should continue to be, the leading supporter of biomedical research in the country and in the world. Without NIH funding, the work I've described, as well as many, many other studies with at least as great even greater impact and import, simply could not have been done.

And with that comment in mind, I want to convey my deep concern about the future of government-funded basic science in this country. For example, if we look at this slide, what we're looking at here is the rate of change in the NIH budget. And the light blue line shows the percent changes in the budget since the 1994. The black line shows the rate of general inflation. The dark blue line shows the federal government's [INAUDIBLE] of the rate of biomedical inflation which is higher than the general rate of inflation. And the red line shows the changes in the budget request from the President.

1999 to 2003, this period here, is the period of the doubling of the NIH budget. Now there's been a recent study by the Medical Research Council in Britain of the rate of biomedical inflation. The consequence of this study is to show, and many people in this country already knew it, that this value that's shown in this dark blue line, what's called BRDPI. Everything in the federal government has an abbreviation. BRDPI is the Biomedical Research and Development Price Index. POTUS is the President of the United States, in case you didn't know that, OK.

BRDPI drastically underestimates the real current rate of biomedical inflation. Given the inflation rate and the costs that all of us in the field know about, the increased cost of doing genomic studies, animal maintenance, and so on and so on, it has been estimated that a 6% to 8% annual increase in the NIH budget is needed simply to maintain the current level of biomedical research, and that a 10% to 12% percent increase is needed to propel biomedical research in the ways that are necessary to take advantage of current knowledge and opportunities.

Instead, look what's happened. 2004 and 2005, the NIH budget has increased 2% to 3%. 2006 there was a decrease. 2007, a very small increase. And the President's proposal for 2008 is flat. For the five years from 2004 to 2008, the average increase is about 1%, drastically below the 6% to 12% that is needed.

The next slide shows another way to look at some of these data. I plotted these data on a log scale to reflect growth rates, and the graph shows two things. First, it shows very clearly what the NIH doubling did. The doubling was from here to here and what it did was it brought the level of the NIH budget back to the level of the long-term growth of the NIH prior to that. It was simply a restoration. In fact, if one goes all the way back to 1938, the level of growth was higher than is seen for this period.

The second thing is if you look at the consequence of what's happened in the last most recent years, the upshot is that we are now at a lower level compared to the classical growth rate than ever before in history since the NIH was founded.

The consequences are actually worse than these graphs indicate. NIH supports research with multi-year commitments. Most of each year's budget goes to fund commitments from prior years. In a time of diminished real dollar budgets that means that few, if any, new efforts can be funded, and the consequence-- and this is now unfortunately our current history-- the consequence has been unprecedented low success rates for new and renewal grant applications, the curtailment of long-term successful projects that have come up for renewal, and the irreversible drain of talented and highly-trained scientists from the biomedical community.

OK, so what is needed? Well ex-NIH director has stated very clearly what is needed. The operative phrase is a soft landing. And what that means is, if you're going at a rate of increase that's high and you have multi-year commitments and you want to go to a lower rate, you don't do it by going flat, which is a disaster. You do it by bringing things down more slowly. That's what is needed.

Now this picture concerning the NIH is bad, but the broader picture concerning the federal funding of science and technology in the US is even worse. A recent blue-ribbon panel examined the current state of US science and technology. This panel was chaired by Norm Augustine, the ex-CEO of Lockheed, and it had many respected and experienced members, including our own Chuck Vest. The report of this group is entitled "Rising Above the Gathering Storm," and reading it is eye-opening, and I would say, beyond distressing.

The data about the current trajectory of science and technology in the US is positively frightening. For example, the committee showed data indicating that the federal funds to support research and development as a function of the national Gross Domestic Product has been falling essentially continually since the time of the space program.

The executive summary, this is the cover of "Rising Above the Gathering Storm," the executive summary of the Rising Storm stated, amongst other things, "The committee is deeply concerned that the scientific and technical building blocks of our economic leadership are eroding. We fear the abruptness with which a change in science and-- with which a lead in science and technology can be lost and the difficulty of recovering a lead once lost-- if indeed, it can be recovered at all. "

The committee has detailed and highly specific recommendations. One of them is exceedingly straightforward. Increase the federal investment in long-term basic research 10% annually over the next seven years. I could not agree more.

Most of us in this room are involved in cutting edge work in the fields of science and engineering and I think it is vitally important that we make our voices and are arguments heard. We must strongly encourage our government to commit the appropriate resources that will allow our country and the world to benefit from the science and technology that is so vital to our future.

Okay, so with that, I want to finish with, I would say, maybe my most important slides, which are the acknowledgements, acknowledging and thanking the members of my laboratory who have been responsible for these efforts over the years.

These are the current members of my laboratory, and if you want to reach any of them, here is the contact information. The next slide shows a number of people who had been in the labs over the years at a reunion we had a few years ago, many, many very now close friends in this picture.

And finally, and more comprehensively, I want to thank all of the members and visitors to my laboratory from undergraduates to sabbatical visitors ready for retirement who have been with me here at MIT since I came more than 29 years ago. And I'll stop with that and thank you very much.


LERMAN: I think it may be time for just a couple of questions. I know some people may have to depart, but do you have a moment?

HORVITZ: Absolutely.

LERMAN: So if you want just take questions from the floor, that would be great.

HORVITZ: I've been told the most important thing I'm supposed to tell you is that there's a reception out there with food and drink. So I see no questions. One question.

The question is how optimistic am I that the funding will be increased? And the key word there is the timescale, when? And what it's going to take and what will happen in the interim? So I believe that it will be increased as eyes get opened, but I think it's a very difficult sell. Particularly today, with the federal budget in the trouble that it's in, it's very hard to make the arguments.

It looks like this year is better than last year for a variety of reasons with respect to, at least, the budget for NIH. There have been some increases that have been made to other federal budgets, including the National Science Foundation, but when you look at the total dollars that have been put into these increases, even though the National Science Foundation has been increased in a reasonable way percent wise, the number of dollars is piddling.

If you want to read a document that really spells out in a deep, thoughtful way what today's issues are, read the 1949 report from Vannevar Bush, in which he proposed to the federal government why basic research is so vital to our future. And in fact, he was pivotal in making the federal government change its ways and really dedicate the resource that's led to the revolution that has occurred in the interim. We need today a new Vannevar Bush, because otherwise, it's not going to happen in a timely enough way to really save us.

AUDIENCE: Could you say something about the too little cell death connection to auto-immune disease and what progress, if any, there is in that direction?

HORVITZ: So in autoimmune disease what's basically happening in some cases, is that cells that should be destroyed because they're capable of recognizing self, our own bodies, and they are destroyed by programmed cell death are not destroyed and so they survive to recognize our own bodies. So that's the basic biology. Apologies to the immunologists for the very summary statement.

And the thought is that by intervening in that process and in fact, in a variety of other diseases that are related, you should be able to do something in principle. The step from in principle to actually deriving and developing proper therapeutics is a big one.

AUDIENCE: Has your math major been useful? And how did you switch?

HORVITZ: Has my math major been useful? Well, I can tell you its utility came into play even yesterday as I got an email from a lobbyist in Washington who said, explain to me what a log graph is. Now sophisticated the math wasn't, but I think I explained what the log graph was and why it was the appropriate way to look at data of this sort.

In short, I think the mathematics has been useful as a training in rigorous thinking more than in the mathematical methodologies themselves, but I must say that there are times when the math is useful, but I can't think of any time that Calculus on Manifolds has come into biology in a straightforward way.

How did I switch? Basically as an experiment. I knew very little about biology. I knew I wasn't a mathematician. That was very clear to me, and so I thought I would do something a little different.


HORVITZ: That is, I think, actually a very complicated question to define in its entirety. The genetic code basically defines the relationship between DNA, where the information is stored, and proteins, which do the majority of the work in cells and organisms. Okay, that then becomes read out in a variety of ways, including in development and including in the cell lineage. In many contexts, that DNA leads to a varied set of defined responses. In some contexts, it leads to defined responses within a limited repertoire that can then be affected, for example, by the environment.

So in the case that I was talking about there, basically what's going on is defined by the genes and the way they are translated into the workhorse proteins. The A to B to C are the same letters perhaps, two of them, as is used in the genetic code but it's being used in a fundamentally different way. It doesn't mean the same thing here.

AUDIENCE: Once the EGL-1 protein is unleashed and the cell commits to dying, is that an absolute commitment, or is there any way that it can be reversed?

HORVITZ: EGL-1 basically triggers a cascade of events. And so you could stop events downstream of EGL-1 and nothing will happen. And that's been demonstrated very clearly. The question as to what is the last irreversible step, is one about which we only know a little so for me to answer it would take too long.

Okay, let's do the refreshments and again, I thank you all very much.