Posted 18 January 1999
Interviewed by Brian J. Cummings
ARF: What is the primary hypothesis that guides your laboratory?
PC: The primary hypothesis that guides my laboratory shifts from time to time as we and other people get more data; but the primary hypothesis in the lab now is that a cell in AD is sick for decades before the cell actually dies and that there are a lot of changes in gene expression over the course of these decades which have consequences for the function of the neuron. And we think the evidence suggests that the changes in gene expression have at their core the re-expression of genes that were originally utilized in the cell cycle when these cells were dividing. And because these cells are no longer capable of dividing, these early sequence of events cannot be carried through in a coherent way, so the cell goes down an alternate pathway which is harmful for the function of the cell.
ARF: Okay. Would you then say by consequence, that even in a transitional or very early case of AD, that there are a larger number of cells that are dysfunctional than we were traditionally recognizing?
PC: I think there is no question that there are a larger number of cells that are not functioning normally than have been traditionally recognized. It is very clear that the traditional marker of NFTs is something that occurs farther along the course of the cascade and that has been obvious ever since the work of, say Randy Nixon, showing that the expression of cathepsin D was present in so many cells in AD - far more cells than those in which you could see NFTs. And now we are seeing cells in which there is immunoreactivity to various phospho-tau antibodies in the absence of NFTs. There is also the work of Peter Davies, where some of his antibodies expose very early markers. So yes, I think there is evidence that cells are doing strange things well before the "traditional" markers of pathology. I think the evidence is overwhelming.
ARF: And what do you think is the event that accounts for the initial symptoms of AD - the beginning of the disease. Would you say that AD is simply pathological aging that gets worse over time, or is there an initial event?
PC: I think that there is no question that AD is not normal aging and from my biased point of view, one of the most convincing pieces of evidence for that is the work that we did with Mark West showing that the pattern of nerve cell loss is different in normal aging and in AD. In other words, in normal aging there is nerve cell loss in some of the hippocampal subfields, but not in CA1 whereas in AD there is nerve cell loss in CA1 (contrary to one weird report not too long ago). So, I haven't answered your question completely...
ARF: Well no, but its a good start. How about, what is the initial event?
PC: The first cause ...
ARF: Do we know what the first thing is? Ten years from now will we say "Oh yeah….
PC: Now obviously we don't know now what the first cause is and the speculation that I would make is that there are multiple first causes which then funnel into a more restrictive number of secondary causes which then eventually funnel into a final common pathway which we call AD pathology.
ARF: I am going to take a side track to that because I'm curious. Are you saying then that plaques and tangles are linked? Are they two separate things both the result of a final pathway, or does one lead to the other?
PC: Well, first of all, I think that if one leads to the other, I think that Aß leads to NFTs, not the reverse. But that is not to say that I really believe that this is so. There is data out there that shows that the two are not well correlated, that you have one without the other. So, that certainly suggests that they are not inevitably linked, but that still doesn't mean that, for example, if Aß is present that it cannot in some way lead to the formation of NFTs. And there are data out that suggests that this may be so; I'm thinking of work showing that Aß fragments of APP will stimulate some kinases that have been shown to phosphorylate. Some of these data may be somewhat controversial, but the data are still there. There are a variety of different kinds of data that show that Aß can lead to some of the events, at least, that produce paired helical filaments. But nobody has yet produced NFTs in an experimental preparation.
ARF: How strong a role do you think genetics plays in AD? Maybe as a follow-up to your answer, would you say that genetics will eventually explain all cases of AD once we understand about all the genes involved… say the year 2010?
PC: In the sense of what you get from your mother and father, no it clearly won't. To me, one of the most telling pieces of evidence in support of that statement is the study of identical twins where one twin gets AD and the other one doesn't. Of course, it might be that the twin that didn't get it will get it if he lives long enough, that could be said about anybody. The point is that these are people with the same genes and one gets it and the other doesn't - period.
ARF: Earlier, you said that AD is not normal aging. You mean to say that if everyone lived long enough, not everyone will get this disease?
PC: No, presumably not. The epidemiological data suggests that the rate of increase and prevalence starts to fall once you get into old age.
ARF: So we all aren't destined to get AD. Outline for us the steps, or the progression the disease takes.
PC: Before I do that (and I don't know if this is something that is going to come up later), something I do want to say is that about 100 years ago Alzheimer's described the senile plaques and the NFTs of AD and I think that to a great extent the field has been like drunk looking for keys under a lamp because that's where the light is. So here are two very obvious things that are going on in the AD brains and I think that there has been an over-emphasis on those things, to then be ignoring other kinds of things that may be going on in the AD brain. I think that it is important to understand what is behind the formation of senile plaques and NFTs, but I think that to adopt the position that these are central to the disease is counter-productive.
ARF: Okay, so let's say that you were going to recommend to the NIA a new program announcement. With the specific idea in mind that we have too many plaque and tangle people -- myself included -- what direction would you look in and suggest to young people to go off in?
PC: NIH review groups do not like fishing expeditions, but I think that at this point we do need a fishing expedition, and the technology is there for it. And basically the nature of the expedition, in my mind, would be to look at profiles of expression in individual neurons as a function of disease progression. And by disease progression I don't mean disease progression in the total brain or the whole individual, I mean disease progression in the neuron. And one of the things that's required are markers of disease progression. We don't really have a complete set of those, although we do have some that do indicate different disease states. But maybe we don't need a marker for every step along the way because once we have the complete profile of gene expression, then we should be able to reconstruct the progression of events. That would be a real problem for an informatics analysis, but I think that it's something that's plausible. Not only does it get us away from the NFT and the senile plaque, I also think that it could lead us back to the first cause of the disease because knowing what genes are turned on very early - extremely early -- in the course of the disease and which are turned off, then we can start to think about what is turning these genes on and off. That's one aspect of the "Gedanken" experiment. One of the other aspects of the "Gedanken" experiment ("Gedanken": to think about an experiment. It's what the German nuclear physicists used to do - the theoretical physicists...they would never do the experiments).
ARF: That's exactly what we are looking from this interview process, good "thought experiments" or "Gedanken" experiments. So one suggestion is going on a fishing expedition for genes whose regulation changes very early in the disease?
PC: Another aspect is that basically so much of the work that we do in AD comes from brains from people that have died. What we need to do is to be able to look at the brains of people while they are alive in far more detail than we can now. We need to have imaging methods that allow us to look at the chemistry, molecular biology, and morphology of individual neurons in a living person, without opening the head, preferably.
ARF: At what stage in this pathway of primary causes (secondary causes, tertiary causes, etc..) do you think intervention is going to be most successful? The obvious answer is to design interventions for those first causes. But since we don't know what those first causes are, how should we guide therapeutic research?
PC: Well, in terms of our current state of knowledge, I think that the basic principle is therapeutic intervention as early as possible. Basically the goal currently is therapeutic intervention before somebody becomes symptomatic -- before somebody goes to the doctor's office and before he says that he thinks something is wrong. Basically, intervention while the person is still functioning at the normal level and then halting the progression of the disease as strongly and efficiently as possible, so that they live out the rest of their life span without ever exhibiting any symptoms.
ARF: Is this partly from an idea that once you have it, there is no way of reversing the changes?
PC: Well, I hesitate to make that assumption. I think that if that kind of statement is based more or less on the state of knowledge that you can anticipate a few years down the road when you may have a reasonable bio-marker and you may have more effective treatments, but not treatments that are getting at the basic mechanism -- the treatments like the cholinergic agents and so on. Clearly not basic, but are helpful. So, in terms of reversal, I think that reversal is absolutely possible if detection is early, before the cell is committed to die, before it has formed insoluble junk inside of it. To reverse the process is totally feasible and what I can imagine would be that once we have defined the genes that play an important role very early in the cascade, then it would be possible to deliver gene therapy.
The kind of vision that I have about gene therapy is that, first of all, I think that AD is a systemic disease. Peripheral cells are affected, as are neurons. So, I think the molecules or the genes that are likely to be found as important in the early steps of the disease are going to be genes that are expressed throughout the body. And are we going to want to mess with those genes? Are we going to want to mess with protein kinase C, for example, on a global scale? No, I don't think so. I think that we are going to want to do is to deliver the gene therapy to defined regions and be able to turn genes on and to turn it off under our control. And the technology for this exists already. Under various external controls, it is now possible in experimental animals to deliver genes to specific brain regions, turn them on when you wish, turn them off when you wish, and so on.
ARF: But now we need to know what gene we want to turn on or off.
PC: Right, we need to know what genes we want to turn on and off in AD at a very early state.
ARF: How would you respond to someone saying that AD is primarily environmental, and so that there are a certain set of genes that will be turned on if a neuron is under stress. And if you turn them on enough they will go down some pathway that results in cell death. But that pathway is a side effect in AD. If you look at all of these genes you'll end up following a lot of false leads. How do you follow the leads that are AD specific? Isn't there that risk of studying irrelevant genes versus those that are causal to AD?
PC: No, I think that the term genes includes all of the thousands of genes that are expressed in a cell and they can be turned on by what you have gotten from your mother or father or they can be turned on by environmental factors, or by anything. When I talk about genes being expressed or not expressed, it can be from any variety of things. I don't think that is going to take us down a garden path... However, I think it could unless our methods of data analysis are well developed, because we're going to have tremendous amounts of information. Think of it. We're going to be able to get a complete profile of all the genes that are expressed by a single neuron. So, we're talking about ten thousand genes that are expressed by one neuron and we're going to be looking at multiple neurons, within one brain, within one brain region of that one brain. We're going to be looking at other brain regions, we're going to be looking multiple brains, we'll be looking at multiple disease states. That is a huge data matrix, and of course if you're not looking at that in some kind of rational, reasonable way, you can go down all sorts of garden paths. Some of the challenges in the future will be not just to get all of these data, but to develop methods to look at it in a way that makes sense.
ARF: And one of the consequences might be that if we study the genes that are turned on as cells are degenerating, maybe we'll cure AD, maybe we'll be able to cure other degenerative diseases as well. As we learn about the damage that stress does to an individual neuron and what genes are altered, we'll be able to cure more than one disease?
PC: Oh absolutely. I think that this approach is a model for dealing with all kinds of diseases. I would also caution that the expression of the message is a far cry from the delivery of the active protein to its site of action. Because once you have the message it has to be translated, it's probably got a pro-protein, it has to be broken up, or it has to be phosphorylated or glycosylated. It has to be transported and delivered to where it is going to do its job. We don't have the technology for looking at proteins in the same huge arrays now than we do for looking at messages. And we don't have the technology for the localization of these proteins. The message's localization is less important (although it has some importance) but with proteins it is vitally important. But we don't have the technology for doing that at a large scale -- and we need that technology.
ARF: If you could design the perfect experiment, or in your view of 5 to 10 years from now, if we could do A, B, and C, what would be the experiment that would really help us understand AD, and what would A, B and C be?
PC: Forget 5 or 10 years from now.
ARF: Okay, we'll leave a specific time frame out. In the future, what is the ideal experiment to do if there were no limitations in our technology, in our data analysis capabilities, in our funding, in man-power required to do it, etc… What would the perfect experiment be?
PC: My perfect experiment would be to be able to follow the expression and processing of messages and proteins in compartments of individual cells in the living person with known cognitive status..
ARF: We're going to need a lot to accomplish that. Maybe molecular biology is coming along fast, but in the protein and imaging arena we're going to need similar leaps and bounds.
PC: Yes, that's true.
ARF: What if you were wrong -- the primary focus of the lab is wrong -- what if that was not the right focus? What is the most nagging criticism people raise against your work? And is there an experiment you are going to do in the future or could do to address this criticism? How are you going to answer your critics?
PC: Well, in terms of the work that we're doing now on message profiling in single cells, one of the most frequent criticisms that we usually get is that it's a fishing expedition. What if we're wrong? I think that this is the kind of work in which it is not possible to be wrong. I think that information is going to come out of it no matter what. And I think that it will be useful information. It may not be the kind of information we want and the kinds of things that I said about cell cycle genes may be entirely wrong, but this approach is going to uncover something else.
ARF: Something else that is going to be useful to understanding AD.
ARF: What would be your advice to graduate students starting out in field of neurobiology? Or maybe post-docs? What would you like to pass on to them?
PC: My advice to graduate students or post-docs is that you have to know molecular biology, you have to know protein chemistry. You must never forget that you have to know neurobiology. It seems as though there are too many people who do molecular biology -- who are very good molecular biologists -- who don't know the neurobiology and therefore do the wrong experiments. And they don't know the neuropathology and then they do the wrong experiments. And similarly, I have seen people who know the neurobiology and forget the molecular biology.
ARF: So, do you think the length of time to get a Ph.D. should increase because there is so much more to know?
PC: There is always more to know but I am not in favor of increasing the time it takes to get a Ph.D. I think that the bottom line is that you have to be continuously learning. I think that the length of time one spends as a post-doc clearly is increasing, whereas 3 years used to be standard and now 5 years isn't at all unusual. Forgetting about these overt markers of where you are, I think that you have always got to be learning.