2000 LECTURE SERIES

Is Alzheimer's Our Reward for Living Longer?

Dr. Dennis Selkoe
Center for Neurologic Diseases
Harvard Medical School
February 22, 2000

I have a story to tell you and I think it's a true story. But you know, in science you never know for sure whether it's true or not and that's sort of the excitement of science. And if it's not true, it'll be indeed highly embarrassing. But I think that we'll show whether it's true or not in all of you. And that is, we're going to prove the point by trying to treat this dread disease that we call Alzheimer's. So indeed, as Dr. Bryant mentioned, I've chosen a provocative title, "Is Alzheimer's Disease Our Reward for Living So Long?" And I hope to give you a positive answer, not positive yes, but rather positive in the sense that you'd like it to be positive, which is no. And that's what I hope I'll be able to explain for you today.

Now, in trying to help Macbeth and so many others with the misery that they experience from brain diseases, those of us who practice what might be called applied neurobiology are in a wonderful position. Because we get to do fascinating scientific experiments that are using the same tools that the most basic scientists among us use. But we have the chance to gain insight into major brain diseases. And that's what the Alzheimer's story is about, a wonderful joining together of basic science and the application to human suffering. And indeed, as a neurologist, that's what turns me on, so this is a great privilege to speak with you about this. It's what I love to do and it's also a very nice time of year to leave Boston. I wouldn't have come if it had been June. That's not true. I find that people are interested in this subject for different reasons. There are a lot of people who come to a lecture like this because they really want to figure out what's going on with Alzheimer's. It's a curiosity. It interests them. It intrigues them. They recognize it as a major problem. And they just would like to understand a little bit about the biology and how we're going to solve it. And if you came for that reason, I hope not to disappoint you. But there are other reasons. Some people come because they recognize that this is such a common disease that, God forbid they might themselves develop it. And they wonder about what I'm going to say at a very personal level. And that's understandable. And then there are a third group of individuals, many of them young, who are in the audience, who come because they're already convinced that their elders have the disease. And many times they sit in the back of an auditorium and they don't say much, but they're thinking all the time, what am I'm going to do about my mentor's disease? Maybe I'll gain some insights today. So for all of you in the audience who have those different reasons for being here, I'm glad you came and again, I hope not to disappoint you.

So if you're not a scientist and I know that many of you are not, I'm going to make an effort to explain the Alzheimer's story as it's unfolded, sort of historically as it came about. And I'm going to try to do it without getting too lost in biochemical detail. Now, some of you are scientists so I'm going to also mention a few things that appear esoteric, or arcane. But I'll ask you to stay with it, because actually even those arcane details, one can explain without too many polysyllables. And I'm going to make that effort to do that. Now, what's wonderful about the modern study of brain science is that we are perhaps the only creature on the face of the globe that has the chance to examine our own consciousness, if you will. And primates and human primates in particular, can sort of think about how it was that this remarkable structure, the human brain, arose out of the evolutionary primordial swamp and dominated the planet, as clearly it does right now, for better or worse. So Alzheimer's provides a window into how we human beings use our most remarkable tissue or organ, and that is the human brain. And at the same time, it provides a window into how that amazing machine self destructs, which is sadly what it does in people who develop Alzheimer's.

The story starts in many ways on a bench, on a boat, on a lake, in a province of Germany. And that is on the [inaudible] where four men, these four men were out on a weekend excursion, on a boat, on the [inaudible] and with a little bit of poetic license I can tell you that they were talking at some point during the afternoon about the quite interesting discovery made by this man. And so, I'm introducing to you Alois Alzheimer, a Bavarian, looking a little bit Prussian I think, but he actually was a Bavarian. And by the way, his house is now immortalized, it's been purchased where he was born in 1869 and now it's a place where people can visit if they are concerned or interested in the history of this problem. But Alzheimer himself, who described the syndrome that bears his name, did not have the gall, so to speak, to apply his own name to it. And rather, we have to all be respectful of and dutiful towards our mentors. Here's his mentor in the light suit, Emil Kraepelin. And Kraepelin was the one in his second or third edition of his textbook of psychiatry at the turn of the century who decided to apply Alzheimer's name to the syndrome he described. So I remind, Burt Seng for example who is in the audience, always be good to your mentors. I was a mentor of his. And you never can tell when it's going to turn out that they're going to do something good for you. And if there are any other of my mentees out there, I remind them of the same. Now, the gentleman here we don't know much about. That's Dr. Gaup, he wrote some biographical words about Alzheimer a few years after this slide was made, this picture was taken. Because Alzheimer died prematurely of kidney disease just a few years after he made his now renowned discovery. And this man, for those of you in science, is renowned. His name is Franz Nissl. And Nissl is a name we recognize in neuroscience because he was one of the great students of the nervous system and the neuron in particular in those early days. So this is how Alzheimer's disorder began. And I'm tipping my hat here to psychiatry, because these four individuals were all psychiatrists and all working together in Munich.

Now, if we jump almost a century ahead and say what is Alzheimer's disease the year 2000? It is, of course, the most common cause of dementia in essentially all developed nations, and it's rapidly rising even in less developed nations. As such, it affects 3 to 4 million Americans and perhaps 15 to 20 million worldwide. As everything that I will say tonight, don't hold me to it precisely. We don't know a lot of these things and we have to infer as best we can, but it is a very common disorder. And wherever you look for it in a population you find it. It occurs in all races and ethnic groups and Americans of European, or Asian, or African descent have very similar prevalence of Alzheimer's disease. And it is a very expensive disease, which as you know if you've faced this tragedy in your personal life, is largely spent for the care of individuals for whom we have no good treatment. These numbers from the Society for Neuroscience are now a bit out of date and we don't need any larger number to convince us that this is a public health problem we have to do something about.

Today, I will not announce a cure or effective treatment. But I will point clearly to that at the end of my lecture, so I encourage you to stay and stay awake. Because at the end, I will say my views of how we're going to treat Alzheimer's and where all this science for the last several decades is going. Now, let's go over a few definitions so that we're all starting at the same point. Now, I apologize to those of you who care for Alzheimer patients on a daily basis in this is not interesting to you. Senile dementia, which the general public often calls senility, is a progressive mental failure after the age of 65. And I actually learned, as I've learned many things from Jim McGaugh today, why it is that we choose 65. Well, we choose 65 because the Social Security Administration says that's when we're supposed to retire. And Jim taught me today that the reason that they say that is because they took that from Bismarck. So Bismarck apparently, at that time in Germany, wanted to define an age as Jim, I'm just telling Jim's anecdote now, at which most people would be dead. And in Germany in the 1870's, indeed most folks would be dead by 65. So Bismarck said fine, that's the year that we'll say they can get retirement benefits from the, do I have that right, Jim? And interestingly enough, Bismarck's selection, so there's a dramatic influence to this topic tonight, Alzheimer was a Bavarian, Bismarck was a Prussian. And it turns out that we still use the term after 65 for senile dementia and before 65 for pre-senile dementia. But as you just heard me say, that's an artificial construct and actually Alzheimer's is a continuum that can begin tragically as early as the 30's or 40's, that's exceedingly rare. And then it just keeps on going with age. As such, it's the most common of certainly well over 20 causes of dementing syndrome, Alzheimer's. And these cases of dementia are perhaps 60 to 70 percent Alzheimer's disease. So there are many other dementias and Alzheimer's is the most common one. It's not a new disorder and I like to say sometimes that perhaps, again Shakespeare who knew all things, was thinking of Alzheimer's when, in "As You Like It", he had Jacques' soliloquy talking about the seven ages of man. And the concept that as we get older, we return to an infantile state. So one can read into Shakespeare what one wishes. And I choose to read into it that Shakespeare was familiar with the syndrome that now is referred to as senile dementia or Alzheimer's disease. We simply didn't diagnose this many times in the past. And where I grew up in the Midwest, we spoke of hardening of the arteries.

Now, the disease begins insidiously and usually with decreased memory and some confusion. And it progresses slowly. And all families who deal with Alzheimer's will tell one that they can't really say when the disease began. And that's helpful to the clinician, because anything that has a defined starting point is unlikely to be Alzheimer's. Sometimes you hear that a bang on the head or some kind of head trauma initiated the process. That may be formally possible, but usually that's just something that brings out the underlying condition. And we don't think of Alzheimer's as actually starting at one moment in time. It shortens life expectancy by some 30 to 70 percent. And it is something that ultimately is fatal. People always ask, how can Alzheimer's kill a person? And it does it like so many brain diseases by weakening many, many systems in the body to the point that the patient develops an ancillary illness, like pneumonia, or a fall with a hip fracture, or a urinary infection. And that is then, sadly, the final event.

Now, this statistic is worrisome. By 2010, the United States will have arguably more than 40 million people over 65 and some 6 million over 85. And so, so many of us, as the question in my title suggests, will be rewarded by entering the age where Alzheimer's becomes highly prevalent. But we're going to do something about that. Let's see if I can go forward here. Here we go.

Now, in many populations when Alzheimer's has been studied, it seems to have a linear relationship to age. So here's age, this is a semi-log plot. Here are new cases of Alzheimer's disease. So this is actual incidence. Very few places in the world can calculate incidence of a complex, chronic disease. But one place it can is Rochester, Minnesota. And you know why that is. That's where the Mayo Clinic is and they're about the business of figuring out how these things develop. And they put together this slide from earlier studies in other populations and their own study, and showed that the number one risk factor for Alzheimer's is time on the planet. The longer you've been around, the more likely it is to occur. And it doesn't really plateau or become ascentodic like this line indicates. That's an artifact of this particular Stockholm study. It just keeps on climbing. And I see individuals in their 90s who have this disorder. As a result, this kind of prevalence data, per 100,000 men, women, and children in the United States population, is probably an underestimate and will soon be definitely an underestimate. So Alzheimer's with 6 per 1,000, any time you can quote prevalence in per 1,000 you have a common disorder, is 3 times as common, roughly, as Parkinson's disease and many times more common than other well known diseases. And this is of course, why people who have Alzheimer in the family got together and created the Alzheimer Association, which does a lot to publicize the enormous public health burden. Not to say that we don't want to study all of these diseases and indeed, to say in a positive sense that understanding Alzheimer's may well help us understand disorders like Huntington's. And I think there's good, scientific defense for that statement.

Now, I'm going to stop for just a moment and make a point about clinical aspects. I am a neurologist and I see patients with the disease, but today, my mission is to talk to you about the remarkable science that has led us forward in the Alzheimer area. I will say, however, that Alzheimer's looks like a remarkably insidious process. You can't say when it starts. It just keeps on going. Very few people plateau and stop getting worse. A few of them do, but after a while they do again. And it's very symmetrical. That is, it affects both sides of the brain. It affects the memory centers importantly, but then other things like language and finding your way around your daughter's house, even a familiar area, or driving to a place that you don't go so often. Those things creep into the picture. And the clinician has to diagnose the disease based on that story, on how Alzheimer's evolves. We do that with the help of imaging procedures, where we look for the damage that Alzheimer inflicts in the brain. But even the imaging procedures are second to the clinical diagnosis. We can usually call it Alzheimer's, based on what it looks like in the patient. There is little that looks exactly like Alzheimer's. Now, for example, by positron emission tomography or PET scanning, which is a remarkably powerful method that I know nothing about really and don't work with at all. One can image the brain and then project those images in pseudo color and we can just say that low metabolism is blue and high metabolism is red, arbitrarily. And we can see that in Alzheimer's disease there is sort of a symmetrical left and right hypo-metabolism, a blue color here compared to the normal brain. And here you see that again. There's some hypo-metabolism frontally and especially in this part of the brain, in the posterior, parietal, temporal, occipital region. And this is a signature of Alzheimer's. While such PET scans aren't 100 percent specific, they're helpful. And there are also less expensive methods than this that help the clinician figure out this indeed is Alzheimer's.

By the way, the other things I'm going to tell you about Alzheimer's disease, that are very interesting scientifically, cannot be seen in the PET scan. So the plaques and tangles that we'll speak about are not visible to the PET scan. We have no way of imaging those. And it would be great if someone figures out how to image the actual lesions of the disease in vivo. We don't have that yet. Well, that's what I'm going to say today about the clinical picture. And I'm mostly going to speak for the rest of the time on these three questions. So if you try to get your arms around the incredibly complex study of Alzheimer's disease, you have too many competing voices. You have possibly viruses or preons or aluminum, estrogen, a million different possible players in this. And I've tended to organize this in my own mind by asking which of the three basic questions about Alzheimer biology do these different disparate studies address? Do they speak about what the cause of Alzheimer's is? Do they say that regardless of the cause, we want to understand the mechanism of brain cell death? Or do the studies speak to what types of brain cells are dying? So in simple terms, we want to know cause, mechanism, and effect. And any one study may give us clues to more than one of these questions.

Now, as often happens in biology and medical research, we tended to start with the third question first. That is, the effect. What kinds of brain cells are dying? Were they located in the brain? What is their connectivity? How do they speak to each other? And very importantly, for anyone who cares about how the brain works, how do they communicate? Which means how do they release transmitters? And what transmitters are those? So here's a very nice cartoon of the business end of a neuron in many ways. And that is, how a neuron communicates with the next neuron. The top half of this cartoon is one ending of a neuron and the bottom half is another neuron. And the two are kissing, as you can see here. And they kiss all the time in the brain, via the synapse. So there's a very tiny synaptic cleft here that's in between the two neurons. It's a real space, although it almost looks virtual, it's so skinny. And what happens there, as many of you are aware, is that this neuron can signal to the next one by releasing little packets of chemicals, which are message molecules or neurotransmitters. So this was the first thing that became exciting in Alzheimer research. What kind of cell is releasing what kind of transmitter that's dying? And the answer came back neurons that synthesize, that make the following message molecules are known to be lost in Alzheimer's. And there's a long risk. Now, there is going to be a quiz about all this at the end. I don't see many of you taking notes but, Bert, are you taking notes? You ought to be, because Bert, you have to know this stuff. He was in the lab for about a year and a half so he knows this all cold. So it turns out that acetylcholine is the first one discovered and that was why I listed the year 1976, because that was sort of the first neurochemical handle we got on the disease. But now, there are many other neurotransmitter, neuromodular substances. So Alzheimer's is a quintessential, multi-transmitter disease, that is more than one message molecule.

Now, what's a good example of a mono-transmitter disease? Perhaps Parkinson's. Many of you know that we treat Parkinson's with L-dopa. And dopa indeed is the precursor of dopamine, which could actually be on this list of Alzheimer portabations as well. So Parkinson's is quite successfully treated by replacing the missing dopa. In the case of Alzheimer's that has been tried. And of course, the only two drugs now used in the United States officially are drugs that try to replace or beef up acetylcholine. Those are the drugs called Cognex and Aricept. And Aricept is the one that we mostly use in practice and it is a cholinesterase inhibitor. It prevents the breakdown of acetylcholine in that little tiny synaptic cleft that I showed you in the last picture. But as you hear me unfold the story, the reason that Aricept doesn't work powerfully is because there's too much else going on. As you can see, Aricept would only address the first of these, not the rest of these transmitter alterations. So we have to get to questions two and one. That is, mechanism and cause. And I'd like you leave the auditorium today having some concept of at least my biases about what causes Alzheimer's. To get to that question, a number of us, over the years, have studied the ultimate defining characteristics of Alzheimer's. By which I mean the detailed phenotype. Phenotype is what you see. Phenotype is blue eyes, or brown eyes, or brown hair, or blond hair, the external picture. And that can include what you see in the brain itself. And genotype, of course, is what might lead to that, at least on a genetic basis. We're going to talk about that tonight.

So defining characteristics include the two lesions Alzheimer described back in 1906, analytic plaques, neurofibrillary tangles, but beyond that many other things are wrong. There's a profound glial disturbance with astrocytosis and microgliosis. These are the supporting players. The neuron is the guy who always takes the Oscar for leading man or leading woman in the drama of brain function. But astrocytes and microglia are also very important and they're very unhappy about the Alzheimer process. And are sort of inflamed. There's a selective loss of the main players, the neurons that are the ones that have the synapses that communicate. And beyond the loss of neurons, there's a profound alteration of synapses, which is exactly what you would expect. So you don't need me to tell you this, synaptic loss means that people can't communicate thoughts. And of course, that's what our patients with Alzheimer's cannot do. And downstream of that synaptic loss are again, the multiple message molecule deficits, acetylcholine and all the others that I've mentioned. So what many of us have been trying to do in the last couple of decades, 15 years particularly, is understand how all this arises. And I'm going to give you my view of that.

This is what Alzheimer saw. In 1906, he had the chance to look at the brain of a woman who died in her early 50s with a great paranoia. She believed her husband was cheating on her. And she was assured that there was no evidence subjectively for this. We've heard that kind of story before. But in point of fact, objectively speaking, everything that one knew about this woman was that her paranoia was misplaced. And she also had a gradually progressive memory loss. This is true. This is from Alzheimer's description. And when the doctor had the chance to look in the patient's brain, he saw, as I did in a patient of mine whom I followed during life, two amyloid plaques, shown here in the amygdala, which is an important structure for behavior in a number of different ways. And adjacent to the two amyloid plaques, a number of neurofibrillary tangles. Now, amyloid plaques are spherical masses. And I'll show you how many there are in the patient's brain. It's frightening how many there are. Here at high power you're just looking at two in my patient's amygdala. And the amyloid is the brownish golden material in the center, and its surrounded by these squiggly lines, which are abnormal axons and dendrites. Those are the guys that kiss, that make the synapses. So looking at this you know that probably something is not good about the communication of the axons and dendrites that are surrounding this amyloid plaque. Adjacent are tangles, these are silver positive masses of filaments that occupy much of the cytoplasm of these neurons. But notice that my patient when he died, after nine years of disease, had a number of neurons that looked normal by this particular stain. And there is nothing that I can tell you is definitely wrong about these. I wish I could tell you why that is. We still don't understand why some neurons get in trouble and adjacent ones are still all right. But this is why many Alzheimer patients can still do a lot of things. They can swing a golf club. They can play some tennis. It's not a good time to learn how to play those sports, because you lose the nuances of the game. But you can still do things, because many, many neurons are functional, even at the late stages of the disease.

Now, this is a high power view. Let's look at a lower power view that is back off. And you see really a large number in my patient's amygdala. And if you kept on looking, you'd see millions and millions of these spherical deposits of amyloid beta protein, as we now call the protein. And peppered among them are these neurofibrillary tangles. So a big issue for so many of us in the field was whether tracking where these began would be helpful. And a number of people said, those are tombstones. Those are late in the disease and they're unlikely to lead to early events. And I can tell you that that, by and large, has not turned out to be true. Tracking how these began has, in my view, led us to the etiology of Alzheimer's disease in many ways. Now, the plaque is complicated. And here's another amyloid plaque. The brown material is this amyloid protein that I'm going to tell you more about in a minute. And these are angry looking microglial cells. Microglia are inflammatory cells and they are normally in your brain, but they don't like foreign things, they're xenophobic, so to speak. And when they see something that is not to their liking, they sometimes go right to it and almost want to eat it up. It turns out that in Alzheimer's disease, as best we know, and this is again a simplification, there's little evidence that the microglia trying to respond to the presence of amyloid is good news. In fact, they get stirred up and they probably release a number of mediators, chemicals that don't help. And here's another cell, the astrocyte, which is this fibrillary structure. A bit player so to speak, but probably I think much more important than a bit player. And also causing trouble when it reacts to the microglial rich amyloid plaque. So in short, the last two sides have said that in essentially every patient in the world with Alzheimer's disease, there are plaques of protein called amyloid protein that have angry looking microglial cells, reactive astrocytes, supporting cells, and damaged nerve endings, synapses.

Next slide. I can do that. Now, I'm going to shorten an awful lot of biochemistry by simply saying that in the 1980's a number of labs, and my lab participated in this, learned that neurofibrillary tangles, which are inside the neuron, the black structures you saw, are composed of a particular protein, a microtubular associated protein called tau, the Greek letter tau. And I'll speak about that again in a moment. And the amyloid plaques, which are outside of the cells in the space in between neurons are composed of 40 and 42 residue, which means amino acid amyloid proteins, or A-beta.

Now, what did I just say? What I said was that biochemists figured out what these lesions are made of. And they hope that by understanding that, they'd go backwards and find out what genes made those proteins. So what does that mean? Well, this is a good time to introduce this concept. Many of you know, and I apologize for those of you who do know. But many of you also don't, or don't feel comfortable with the notion that there is material called DNA, which is the substance that genes are made of. Which is basically, as you know the starting point of life and dictates an awful lot about what we are, along with virtually everything. DNA can be transcribed into RNA, which is a very similar molecule chemically, but is slightly different and that's what DNA has to do to go on to allow a protein to be made. And then RNA gets converted to a protein. Here I've made a dotted line, because it's not like the RNA directly turns into the protein, but rather RNA is a code, a blueprint by which the cell assembles the building blocks for proteins, amino acids, to build a protein. So I introduced the concept, because I'm going to be talking about genes. And you can feel comfortable with this, because basically, a gene is a piece of DNA that has a blueprint, which is transcribed into RNA, and then the RNA tells the cell how to assemble a protein. So in some ways, there can be an amyloid gene, if you will, and an amyloid protein, a protein that arises as a result of that blueprint.

Now, in that regard, we got interested in the tau, which is a protein normally found in your brain and mine, that makes tangles in the brain. And we looked at the brains of people with tangles. And we stained the brain section with an antibody, the tau. And you can see some dark staining nerve cells here that are chock full of tau. And some lightly staining ones, which we call ghost tangles, where the cell has given up the ghost. It's not healthy to have this much tau in one place, in an insoluble form. And the cell has apparently died and left behind this tau material as a cast. But here you're not looking at Alzheimer's. This is a much more rare disorder called PSP, Progressive Supernuclear Palsy. And I show you this to make the point that there are about a dozen human diseases in which you see tangles like this made up of this tau protein, that is not Alzheimer's disease. In these cases, there is no amyloid. There are no plaques. And the patient doesn't really have the picture, clinically, of this insidious, progressive dementia that looks like Alzheimer's. When I recognized more and more that this was what neuropathologists were teaching me, I felt that it was unlikely the tau or tangles would be a primary event in the disease. And so I put my own studies of tau aside. Although it's a very interesting molecule and many people have continued to study it.

And now in 1998, a very interesting development occurred. Some families were found on the face of the earth that have mutations in tau. That is they were born with a defective tau protein. And they get enormous numbers of tangles like this. Indeed, this very condition, I think might turn out to be a problem with tau primarily. The fact that these rare cases of dementia, due to a tau alteration, never led to amyloid plaques allowed us to use genetics to answer a question. Probably the tangles came after amyloid in Alzheimer's disease, because you couldn't get it the other way around. If you had a really bad tau disease, where people inherited a tau mutation, they didn't go on to get the amyloid plaques. And yet, the amyloid plaques were in all of the Alzheimer's patients. So starting with tau couldn't really explain how amyloid arose in the disease. And I think that's largely been borne out. In contrast, the amyloid deposits have led us, I think, much closer to the cause of Alzheimer's.

Here's a brain of a huge number of amyloid deposits, with many, many of them occupying much of the area of the brain. And people like Brian Cummings here at UC Irvine and many others, have done a beautiful analysis of how it's bad news to have this much amyloid beta protein in the brain. And that this does correlate with how impaired the patient is during life. Now, a wonderful clue to the importance of this image that is, having a ton of amyloid plaques in the brain, came from sort of an accident of nature, if you will, and that was Down's Syndrome. So many of you may be aware that in Down's Syndrome the particular picture or phenotype the patient has is due to having an extra copy of chromosome number 21. So here's a nice picture of what all of us have in every cell of our body. We have 46 chromosomes, but if one has Down's Syndrome one has 47 and it's number 21 that has an extra copy. Now, it turns out that patients with Down's Syndrome almost always develop the lesions of Alzheimer's, the plaques and the tangles. So here was a powerful clue. Something about being born with an extra copy of chromosome 21 led to a picture that resembled Alzheimer's in the brain. And indeed, the patients often show some signs of behavioral impairment. Let me give you an example. Here is the brain of a Down's patient, perhaps around 18 years of age, and there are wispy deposits of this amyloid protein, the same 40 to 42 amino acid protein that I'm going to re-introduce to you in just a moment. And if you look immediately next to this brain section, which was stained with an antibody that recognizes the amyloid plaque, if you will, and we'll show that in the next slide. Just six microns away, a very small distance, we stained with an antibody the tau that recognizes the tangles and there are no tangles here. So in an 18-year old Down's brain there's just the amyloid deposit and not the other player in Alzheimer's disease. Now, if we stain a section on the other side of that last slide, just six microns, which is a very small distance away, and look for the angry microglial cells, the inflammatory cells, there's nothing wrong. So these three slides signify that in a Down's patient in teenage years, the amyloid is there, but there is little evidence of reaction to it. So when we understood this more clearly, either the amyloid was absolutely not worthy of further study, because it was there, but it didn't do anything, or exactly the opposite. The amyloid was there first and then the other lesions of Alzheimer's occurred afterwards. The reason we could say that that was likely is because all Down's patients, if they live long enough, develop the picture of Alzheimer's in their brain. And so if they die in an accident at age 18 you can answer the question how, in a way, does the Alzheimer process begin? And the answer from Down's seems to be, it seems to begin with amyloid deposition, as least as far as what we can see under a microscope.

Now, what I just told you about how amyloid can come real early is mirrored in your dogs and cats. So it turns out, and we did some of the work that showed this in collaboration with Linda Cork and Don Price and others at Johns Hopkins, that a number of species accumulate the same amyloid beta protein that makes up the millions of plaques in the Alzheimer patient's brain. And again, when this became apparent, people like myself felt that it was very unlikely the amyloid would be just a disease related thing, because we couldn't say all these animals when they got older had Alzheimer's disease. They didn't necessarily. So it looked like just the aging process in many mammals would bring this cerebral beta amyloidosis as we call it, that is the deposition of these little deposits of a beta. Let me tell you that polar bears are hell to work with. This is one of the things that, and I know that some of the people who have reviewed my grants are in the audience, and this is why our cage costs are so high, you know. And so actually, I've learned that it's better to deal with dogs and cats, polar bears are just, and they're ungrateful too, they're not, they always want to be in the water. But it turns out that all the animals in green here have what look like bona fide amyloid beta protein deposits as they age. And indeed humans are right at the top of the list. We all get a little bit of amyloid beta in our brain as we age. Unfortunately, the mouse and the rat do not. Their cage costs are much less, although they're still expensive. And they have to be engineered to get this pathology or treated in some way, and indeed, that's been done by many scientists. And I'll show an example of that in just a moment.

Now, that let's me point strongly to this guy. This is the actual amyloid beta protein. And scientists like to write out what a protein looks like by a single letter code. So every one of these letters is a different amino acid. And I've learned this well enough now that I can read them all off. But don't ask me to do that. I can do it though. And it turns out that this 42 residue, amyloid beta peptide as such is about 4,000 molecular weight. Every amino acid is 100, so 40 times 100 is 4,000. We call that 4,000 daltons, or 4 kilodaltons. So you're learning a little bit of scientific lingo here. The peptide or protein that we see here is very water hating. That's what all this means. You can always translate scientific words into regular speak. And hydrophobic means that because this has some water hating amino acids in this part of the molecule, it sort of crashes out of solution. And indeed we've showed already in Ô86 and 1987 with Dan Kirshner that these amyloid peptides, if you make them synthetically on a machine, can crash out of solution and form little plaques in a test tube, basically.

Now, in 1990 or so, Carl Cotman and separately Bruce Yankter, began to examine what these aggregates of A-beta do. And Carl's lab in particular, with Christian Pike and others, made a very important observation. And that is that this peptide is trouble, but when it forms insoluble aggregates and you put those aggregates on the neuron, they're neurotoxic. Neurotoxic just meaning the neuron hates them. It gets sick from them. So the fact is that Carl pointed out to us that the A-beta itself isn't enough, it has to be in an aggregated, clumped form. And then it is trouble. And that is a robust observation that has stood the test of time and is one way to model the badness that amyloid build-up brings with it. Now, it's also true that A-beta is too small to be made directly from DNA, to RNA, to protein, in that scheme I showed you. It needs to be made from a high molecular weight precursor, as a lot of little peptides are, like insulin and other things, they're made from pro-proteins that are bigger. And I will introduce to you this molecule, amyloid precursor protein.

This is the structure of the precursor molecule, APP, and the egg shaped region here is the amyloid beta peptide, that 42 amino acid fragment that had all the letters there, that thing is this thing, just the egg part. And as you can see, it's toward the tail end of a larger protein, which we call amyloid parent protein or APP, quite simply. And by 1991, when I drew the original version of this cartoon that Scientific American then made more pretty, we envisioned that there was normal enzymatic processing. There was a cut in the amyloid beta region, cutting that egg in half. And since this is the only part that makes those millions of little deposits in the Down's patient's brain or in the aged monkey's brain, this was good news, because you have to have the intact A-beta to make those plaques, or so it seemed. And this would be against Alzheimer's. But we imagined at that time, we didn't know it for sure, that there was an alternate enzymatic processing, by which we mean there's a way of cutting this full length protein here and here, at the beginning and end of the egg. You pop the egg out, the egg now is floating around in the extra cellular space of this neuron. This is the cell membrane here where APP sits. And now you have the beginning of badness, the A-beta could now build up. Now, I'm going to show you this scientifically in another cartoon. Just imagine seeing the same thing that I just showed you but in something that we scientists feel more comfortable with. It's not as clear and colorful as this. It's a schematic diagram that looks like this. See now, right away, you know, we scientists can [inaudible], because now it looks more complicated than I think it was before. And that's how we convince people that we know a lot, because we just make it look a little more complicated. So this is the cartoon that I usually show. Here's APP, a long molecule. The red is now the egg region. It's called amyloid beta or A-beta. And there are a couple of ways to cut up APP. You can cut right at the yellow dot by an enzyme called alpha-secretase that leaves a stub, everything from the yellow dot to the end here, called C83.

C83 comes from alpha-secretase cutting this big parent protein. And then another enzyme called gamma-secretase cuts C83 and pops out a little peptide that we first discovered called P3. And here's the even more interesting reaction, if you don't get this cutting, you can cut with an enzyme called beta-secretase right here, right at the beginning of the red box. And that's at this position, leaving a 99 amino acid piece left here. And that can be cut by the same enzyme gamma-secretase, making A-beta. So I'll ask you to remember that C99 is the name of the protein that gives rise to A-beta. And you should be interested in this, because you're doing it all the time.

What we showed in 1992 was that all of this happens throughout life that this chopping up of the parent protein happens normally. And that therefore, all of us are potentially at risk for Alzheimer's, because we normally make A-beta. And it's circulating in your blood, it's found in your spinal fluid, and in your brain, which does not mean that you'll all get Alzheimer's disease. And even if we don't develop a treatment soon, not everybody who has this necessarily develops the disease. But I'd just ask you to remember that C83 cut by gamma-secretase yields P3. C99 cut by a gamma-secretase yields a beta. These are enzymatic reactions that cut and liberate products. The rest of this protein, by the way, doesn't seem to play near as important a role in Alzheimer's disease, in my view at least, as the A-beta box itself. And for those of you who are knowledgeable about proteins, of course APP has one so called TM, or transmembrane domain, this is like the membrane of a cell, if we extended these dotted lines, and APP sits inside the membrane.

Now, let's get back to real life. I'm sorry for that digression. And what does that have to do with what causes Alzheimer's? So question one, what's the cause of Alzheimer's? And sadly we can't help it, genes have a lot to do with it. And so I always caution you to choose your parents wisely, because once you choose them you're stuck with them. And they pass on many different features that lead to your ability to age or not to age properly. And what we now know about Alzheimer's, which is partly why you're here tonight, is that there are several genetic alterations that produce Alzheimer's disease. Chromosome 21, it's the APP gene, the amyloid precursor protein gene. And the onset of the disease is early, namely in the 50's, really quite early for Alzheimer's. And I mention here again that three copies of chromosome 21, which is Down's Syndrome, leads to very early amyloid beta protein deposition, very early plaque formation in the brain, as I told you already.

Now, this is very exciting news. When John Harding and Alison Goate discovered the first mutation, because APP is a gene that is on chromosome 21. And so we had an ah-ha experience. If you are born with 3 copies of chromosome 21 in Down's Syndrome, you got an extra copy of APP and you make too much A-beta throughout your life. And that's been proven. Babies with Down's Syndrome have two or three fold more A-beta production than they normally should. And that's probably why they go on to get a picture of Alzheimer's disease.

Now, there are three other genes in brief, a protein called Apo E4, which is a cholesterol transport protein, was discovered by Alan Rosas and colleagues on chromosome 19, to lead to late Alzheimer's, that is the usual age, in the 60's and 70's. And in 1995, several laboratories, especially Peter Hislap at Toronto showed that presenilin one and a similar gene called presenilin two can cause early onset Alzheimer's in the 40's and 50's. And again, I think you can all understand this because this is a pre-senile form of Alzheimer's disease.

Now, importantly there are other genetic loci, that is other genes that are likely. And I think each year and each month perhaps we'll see new genes identified. And what I'm about to show you has to be done with those new genes as well. But we've done it for these four genes. We've asked the question what is the genotype to phenotype relationship? In other words, how do these genes do their dirty work? And what we've found is for example, that when there are mutations in APP, the gene that makes the amyloid protein on chromosome number 21, they're all found, and here's the APP cartoon again. Here the A-beta is in green. Here's that single transmembrane domain of APP. And here's the again, the single letter amino acid code of this part. And the bad guys are in blue in this cartoon. These are mutant amino acids. That is, there is a DNA mutation, the DNA makes a faulty RNA, the faulty RNA instructs the cell to make a faulty protein. And the mutations are all right where beta-secretase cuts APP, or right after where alpha-secretase cuts APP, or right after where gamma-secretase cuts APP. So my lab and Steve Younkins' lab figured a lot of this out in cultured cells and this was a very helpful experience, because the APP mutations were all clustered right where you might think they would be, right where they could help make too much A-beta. And that's what we think they do. And Steve has nicely shown this in living patients with these very mutations. So that's one genetic cause of Alzheimer's. It seems to work by cranking out too much A-beta protein, the little peptide. But that's a rare cause.

Here's a more common cause, Apo E4. And Alan Rosas and his colleagues published that there is increased amyloid beta peptide again, in the brain cortex as a consequence of the Apo E genotype. Now, what this means in simple language is that if you inherit the Apo E4 gene, and about 15 to 20 percent of the folks in this room have at least one copy of Apo E4, you have an increased likelihood of getting Alzheimer's in your 60s and 70s. And if you've got an Apo E4 gene from mom and one from dad, we call that being homozygous, then you have an even greater likelihood of getting it.

How does it work? It doesn't produce too much amyloid. It seems to work by not clearing away the amyloid. There are two ways to fill a bathtub. You can turn up the spigot or you can block the drain. And it appears that Apo E4 is somehow acting by blocking the drain, by preventing the normal removal of this little A-beta from the brain.

Now, finally there are the presenilin genes, and I'll tell you a little bit about them and that's sort of towards the end of my story. The presenilin genes are really very interesting. Here's a picture, a cartoon of presenilin one and it has eight transmembrane domains. In other words, it winds its way in and out of the membrane of the cell eight times. And the red dots are a few of the mutations in presenilin that can cause familial forms of Alzheimer's. Now, you probably know that in any one family, one mutation is more than enough. You only expect a single mutation in a particular patient would cause the disease. These dots are few in number here because there are actually 75 now in presenilin one. And the yellow dots are examples of the two or three in presenilin two, which is the sister of presenilin one, so to speak.

So what are these guys doing? This is not APP. This is not the molecule that gives rise to A-beta. But it turns out when you make a mouse that has a mutation in presenilin one, like this mutation right here. Here's what happens. The amount of A-beta 42 and that's shown here on the side, you can't see it here, yeah. We can make that a little smaller. The amount of A-beta 42 in that mouse's brain goes up. This is work that we did with George Carlson and Peter Hislap and Peter Subert and there's more A-beta 42, which is the form of A-beta that has two extra amino acids and is definitely bad news. Normal mice have this level. Mice that have a mutant presenilin have this level. And by the way, the wild type or normal presenilin molecule put into a mouse doesn't change the A-beta production. So what I've just told you, in short, is that all the arrows point up. The genes that predispose you to getting Alzheimer's disease all have a credible relationship to what we call the beta amyloid phenotype, what you see in the patient's brain. And even you can see it in the patient's blood. They increase A-beta by production. This one, this one, and this one increase the production. And the Apo E4 doesn't increase the production, but it increases the density of A-beta in the brain apparently by preventing its removal. Now, the last little bit of science I want to share with you before I sort of try to put all of this in perspective and answer a few more of your questions is, how does the A-beta get made in the cell? And this is a very interesting, scientific issue. Here is a neuron, a nerve cell in a culture dish, and it gives you a little flavor for how we do these experiments. And in green is the amyloid parent protein, the large protein, the stick that's going to give rise to A-beta. And in red and yellow is another protein called Map two that is a dendritic marker. And we've looked very closely at neurons, in terms of how they move APP around. And in the next slide I'll show you an example of how we do that. We can grow cells in a remarkable little device called a campanode chamber that a scientist at Harvard invented in 1977. You can put the cells, the neurons in the living state, in the middle chamber of this three-chamber system. And you can nurture the cells in the middle chamber and watch their little terminals grow right out into the lateral chamber. Here's this chamber in cross section. The cell bodies are in the middle and the wires of the neuron go out to the sides. And just to give you an example of how exciting and beautifully done neurobiology research is, you can actually study what happens to APP. And we did that. APP comes down the nerve ending, to the end, and then travels back up. And as it travels back up, it goes all the way out to the surface of the cell body.

Now, the next slide will tell you why we bothered showing that. Because we think that while APP travels in the cell, a process we call endosomal recycling. It comes in from the surface of the nerve cell, it gets cut by that enzyme we call a beta-secretase inside an endosome, and then this is the 99 amino acid immediate precursor of A-beta, A-beta being in red here, the putative bad guy. And a second cutting occurs by gamma-secretase. But curiously, that occurs in the membrane, in the fatty substance that makes this little vesicle, or endosome, in the cell. And so beta followed by gamma releases the little A-beta peptide, the egg in my earlier cartoon, which when the endosome comes back to the surface, releases A-beta into the bloodstream, or the spinal fluid, or the extra cellular fluid of the brain. And we're all doing this throughout life.

Now, how is it then that presenilin, which causes Alzheimer's as early as the 20s, extremely rare, but it happens, how does it do its dirty work? We told you already that presenilin, when it has a mutation and it's an eight transmembrane protein. Here it is in blue. Can cause too much A-beta 42 to form. And that's bad news. There are two ideas about how presenilin does its dirty work. One is that it forms a complex with the elusive enzyme called gamma-secretase that again, cuts at the back side of the A-beta box, this is the A-beta box here, and lets it form. Or rather than that, that that is not the case. And rather, presenilin moves the components of this chemical reaction around the cell. It traffics gamma-secretase. So it's a trafficker, a word we use in lay language as well to signify someone who's a bad guy. So we wanted to understand, through the beauties of cell biology, which of these two models is true. Next slide. And we learned, give you one little taste of the science behind this, that we could show APP, which is shown as a black line on this cell, complexed with presenilin. That is, presenilin, the protein that gives the most malignant form of Alzheimer's could be brought down with an antibody and APP came along for the ride. And that's shown right here. And in the next slide, I show you an example of mice that were engineered through the wonders of modern genetic engineering, to lack this presenilin protein. So they do not have the presenilin protein. What happens to them is they die in utero, they have a very abnormal phenotype or picture, including a hemorrhage in the brain. Here's blood in the brain of the little baby mouse in utero. And they have an abnormal body form where they don't make the nice segmented spinal canal that you would normally see. What's also true in these presenilin minus, minus, mice is that they don't make hardly any A-beta. So their A-beta production goes way down, as if presenilin indeed is necessary to make A-beta. But the trafficking of even presenilin itself and of APP isn't changed. So what I've just told you is that our studies in mice like this and in the number of cell lines suggest that this is more likely to be true, that presenilin does its dirty work by actually forming a catalytic complex, an enzymatic complex with APP and gamma-secretase.

Now, we took this a step further, a very clever medicinal chemist, Mike Wolfe, who is a collaborator of mine and now works in our center at Harvard. He looked very closely at the presenilin molecule and here it is again, winding its way eight times through the membrane of a cell. And he spotted two amino acids. And we actually did this on purpose, so he was going to look for one type of amino acid and [inaudible] Shaw looked for another kind. And it turned out that Mike showed two aspartic acids, which are given the letter "D," circled in red here. And for the aficionados in the audience, these two aspartates are right near the site where presenilin itself is cut, which is at the green end. We thought this was very intriguing, because the gamma-secretase enzyme that is so important in Alzheimer's had the properties of a protein that uses aspartates to cut other protein. It's called aspartal protease. And so we did a few experiments that I'll very quickly summarize. We looked at cells that have wild type or normal presenilin and these cells that had one of those aspartates change to another amino acid and the amount of APP, which is this black line you see on this cell was normal, it wasn't changed. But something very striking happened and that is C99 and CD3, which I asked you to remember earlier, which are the immediate substrates of A-beta that make A-beta, they went up dramatically when one of these aspartates or "Ds" was changed to another amino acid. And indeed, when we looked at the amount of A-beta made by cells that had this mutant form of presenilin, here it is down here. This is normal presenilin in cells. Here's how much A-beta is secreted. And here's what happens when you change one of the presenilins from aspartate to alanine, and here if you have both of the presenilins, changed their aspartates to alanine. They don't make A-beta anymore, as if these aspartates are crucial for making A-beta.

So what did I just tell you? What I said was that this molecule we call presenilin winds its way in and out of a membrane of a cell eight times and in the sixth and seventh transmembrane region, it has two aspartates, which we label with a "D." And what we think happens is that these two aspartates allow the presenilin molecule to cut itself, so it cuts itself by an auto protealisis and now it's set to coordinate those two little aspartates with whatever comes along that needs cutting. And now if you add to this cartoon a thing that needs cutting, which is APP, the amyloid precursor protein, here's the last 99 amino acids of APP, C99, and the cutting occurs between the green and the blue parts of APP, liberating this bad guy. This again, is the A-beta, the peptide I've been speaking about all along now. What we think happens here then is that presenilin is an unprecedented enzyme that cuts within the membrane.

Now, this is an unpopular theory, because no enzyme really has ever been seen that looks like that. But as we did this work and this is all right up to date, through the wonders of modern biology, other scientists made a very interesting related discovery. They showed that flies, here's a normal fly right here with its wing, need presenilin, because when presenilin is minus, is absent, the wing of the fly is deformed. And it's deformed in a way that is called a notched wing phenotype. The fly's wing becomes notched in an abnormal way. And that's not all that's wrong. Up here in the brain of the fly, instead of the neat rows of nerve cells, there is a disordered chaotic arrangement of the nerve cells. So people working on flies said if presenilin doesn't work, then you get an abnormal phenotype and it's called a notched wing phenotype. What that is all about is that there's another protein in biology called notch, and that's illustrated here, that also has to be cut.

So let's look at this diagram. This is presenilin in blue. Here are the two aspartates that I'm calling attention to. And it appears that notch lines up to coordinate with those two aspartates and allow itself to be cut. This is crucial for life. On the other hand, here is that APP, amyloid parent protein. Here is the amyloid beta box, the two aspartates can align with that. And when you look at this, and I believe this is reality, not just imagination, you ask the question is there still a gamma-secretase, an enzyme that cuts and, comes in and cuts here? Or is the blue pre-senilien molecule enough to do the cutting? We have more and more evidence that you don't need to draw in this scissors, the blue is the scissors.

Now, why bother telling you all this? Because it's really a beautiful story. What I'm saying is that notch, a protein necessary for life, had presenilin around to cut it and when you developed as a baby this was very good news. You needed a cut notch to let your neurons mature and make the proper connections. As an artifact of the fact that we all live so darn long nowadays, another protein called APP, that perhaps we wish weren't cut as much, is also cut by this very efficient machine. And so thinking of it this way, Alzheimer's disease is an outcome of living too long after reproductive years. Prior to your reproductive prime, this is very, very helpful. After your reproductive prime, this is probably not so helpful. Although I will debate with you whether there are some good qualities to this happening to APP. But it's a new way to look at Alzheimer's, that Alzheimer's is, if you will, a side effect of good development of the brain and of other tissues throughout the body.

Now, in closing I'll tell you that I only talked about the production of A-beta today, but there are other features of this mystery. What happens to A-beta when it is made? How is it degraded? And how is it aggregated? The aggregation, as Carl and others showed originally, is bad news. And I'll just give you one example of what happens in terms of degradation in this graph. Here's the normal amount of A-beta that exists in a cell, in a cultured cell medium. And here is the amount if you over-express a protein called IDE. IDE is something that degrades insulin, it's insulin degrading enzyme. And we believe this is a whole additional story in the Alzheimer riddle. That insulin degrading enzyme is one of several enzymes that cuts up A-beta, the little box, in a good way and helps remove it. And we predict that molecules like IDE will be the site for genes that cause other forms of Alzheimer's.

Now, before I let you walk out, if you're going to do that, or ask me questions, I want to summarize what I've told and put it together in terms of treatment. But while I have your undivided attention, I want to remind you that the kind of thing that I'm presenting today is an enormous team effort. And this isn't even all the team. In my lab at Harvard, these individuals contributed very importantly to this work. And we did this in collaboration with people like Michael Wolfe, and Peter Subert, Dale Schenk, and others who I have mentioned today. There are former lab members who also contributed, this is the beauty of science that you work as a team and through a lot of trial and error you come up with what you hope will turn out to be answers. So I thank my colleagues for letting me give this talk.

So in conclusion, Alzheimer's disease, I think, is a syndrome that arises from a chronic imbalance between producing A-beta and clearing A-beta away, leading to a gradual accumulation of A-beta 42, which is the particular bad guy. This imbalance can be caused by numerous genetic or environmental alterations.

Now, you're going to ask me in a few minutes what the environmental alterations are and I'm going to tell you we don't know a lot about that, but I'm looking forward to discuss that with you. In cartoon form, what I've tried to tell you today is that A-beta is made by you and me throughout life. Neurons, which are the large cells here in the foreground, but also microglia, the inflammatory scavenger cells, the astrocyte, even blood vessel cells, they're all capable of making A-beta. If you're born with a presenilin mutation, I've told you today that I believe that presenilin is the crucial enzyme that cuts APP and makes more A-beta. And you begin to develop cloud like deposits of A-beta that are like those very early plaques that we showed you in the Down's Syndrome patient at age 18. When you first develop those, the nerve terminals that run through the plaque are okay. But as the plaque gets messier and somewhat larger and adds other bad helper molecules, shown in yellow. Now, at this point, the axons and dendrites are not as happy. And so we believe this transition is a very important one, but it's not the whole story, even though what I've told you here is a simplification. Now, in conclusion, five steps in the Alzheimer disease cascade I'd like you to try to recall. There are faulty genes that can cause this disease. The amyloid protein builds up. It makes microglial cells angry and brings out other inflammatory changes in the brain. That, I think, leads to damage to neurons, including the tangles, made up of that tau protein. And of course, the amyloid build-up may very directly injure the neurons. And this is a part of the black box that clearly needs a lot more work. I think downstream of this process is the loss of synapses, the information switches, and the neurotransmitters like acetylcholine, which is the only thing we're treating right now when we offer our patients Aricept. And then comes the dementia. The environmental factors impinge on this, what I think is the critical cascade of Alzheimer's, we don't know exactly how they impinge on it, but we have some ideas about that.

So my very last slide is what we're doing about it. And here's where I think comes some good news. Maybe you can make that slightly smaller, just a little bit. And there are several specific therapeutic approaches that are entering the clinic right now. You could inhibit the A-beta generating proteases. The enzymes called beta or gamma-secretase. I've given you a provocative opinion that presenilin is the gamma-secretase. And I'd like to see drugs that bind to presenilin and block its activity. I think this would be a way to treat Alzheimer's disease, which is like giving a Staten drug to lower cholesterol so that you won't get that build up of A-beta. And I think this is something that will enter clinical trials, probably still in the year 2000. You could prevent the aggregation of A-beta into the fibrils that make the amyloid plaque. You could interfere with the toxic response of the nerve cell to the amyloid, including this altered form of tau. And you could inhibit the inflammatory process around the plaques, which is the angry looking microglia. All of these are worthy targets.

Now, I can't resist showing you this one last image, because I'm a scientist and we just can't stop. And that is that a friend of mine, Dale Schenk at a company that I was involved in helping to begin, but I had absolutely nothing to do with this work. Dale came up with an idea of treating the amyloid problem in the brain, in this case the brain of a mouse, by vaccinating the mouse with the amyloid protein. So he injected the amyloid protein and after about seven months of treatment, the mouse brain looked like this. The amyloid was gone. So I think Dale has come up with a very provocative and potentially clever idea, that if the amyloid is the be all and end all of Alzheimer's, as people like myself believe, could you vaccinate a person so they make antibodies that clear away the amyloid? And it remains to be seen whether that turns out to be true. Thank you for listening.