In theory, the real meeting starts in 40 minutes with Steve Austad's session. Before that we have 20 minutes of "getting to know each other", which we seem to have already done, reasonably well -- I think everybody now knows everybody else -- and then we also have 20 minutes scheduled for a description of how this meeting links up with a meeting that Greg organized about 18 months ago, in UCLA, and which was really, for me, the inspiration for organizing this one. There were certain things that I wanted to do differently than Greg, but there were definitely certain things which I felt that Greg's meeting pioneered and should be taken forward, and so I'm delighted that he and I have been able to collaborate in getting this meeting off the ground.

Unless anyone wants to get another cup of coffee before we get going, I suppose there's no reason not to start on that aspect of the meeting right now.

Gregory Stock: Aubrey thinks he's going to be able to decipher all these transcripts very easily, because he knows everyone, and knows all their voices, but my experience has been that it's very difficult to do that even if you know people. So I think it would be helpful if people would say their name before they speak; then it will be relatively easy to decipher, and once we get into that pattern I think it'll be easy to maintain. And if you're saying something outrageous, just say "de Grey."

Aubrey de Grey: Absolutely!

Gregory Stock The idea is that it would be nice to put together a web site for this that will have at least a rendition of the transcript and anything that comes out of it. I think that would be very useful for some other people.

Judith Campisi: Will it be on your site?

Gregory Stock: Yes, we will make it a sequence with the other meeting and maybe there'll be a series of these every year and a half or so. The goal of the meeting we had in 1999 was to come up with a series of milestones on the path toward developing clinical interventions to retard aging. Judy Campisi and Chris were there and no doubt recall how challenging it was to come up with milestones; a lot of work went into it beforehand. People submitted preliminary milestones and their ideas of the major accomplishments that would move the field toward the possibility of clinical intervention. There were about 100 of those, and eventually at the meeting, we came up with maybe 8 or 9 where there was reasonable agreement that they were meaningful. A big conflict developed on what level of detail we should use to specify them, because there was -- let me say who the other people were at that meeting, there was Judy Campisi, Steven Austad, Chris Heward, Tuck Finch, George Martin, Michael Rose, Cynthia Kenyon, Richard Miller, Huber Warner, James Nelson, Jan Vijg and myself -- some people felt that it should be very high-level, that that's the way science should be done. They didn't want to get into specific details, because that wouldn't prove to be very interesting, because research might evolve by unforeseen tracks that would be much more fruitful than those we imagined. And other people, in particular Richard Miller, wanted there to be more specificity. That came out in some of the goals, where we specified a particular percentage increase in lifespan, or actually even specified what tests would be involved. Eventually, though, the milestones that we came up with were these. I'll just read them, they're on our web site and you can also get a copy):

- develop a DNA chip that would effectively assay aging in the mouse, and do it at the molecular level, convincingly correlating broad patterns of RNA expression with aging (the idea being that this was fundamental to doing research, that it would speed up the cycle time of research that would otherwise depend upon lifespan measurements).

- the second was to find human genes, to identify an allele or a cluster of alleles that substantively contributed to unusually successful human aging. Obviously, that would be very useful in moving us towards clinical interventions of some sort.

- Another was to do various kinds of engineering that would be a mouse enhancement, to construct a genetically heterogeneous mouse stock with an least a doubling of normal lifespan.

- Another was to do transgenic engineering, where you would transfer a gene or a cluster of genes from a long-lived species to a short-live one and significantly increase lifespan.

- Another was to emulate caloric restriction, where you somehow determine the mechanisms of the age-retarding effects of caloric restriction by genetically or pharmacologically emulating it.

These are all very big challenges; there were also three goals to clarify some of the theories of aging -- telomere shortening, mitochondrial damage, and oxidative damage. These were just basically to reduce -- in any mammal -- the rate of telomere shortening and the rate of cellular replicative senescence sufficiently that the resulting change or lack of change in lifespan would somehow resolve that controversy. There was a similar one with mitochondrial damage -- to reduce the rate of mitochondrial damage sufficiently that the resulting change or lack of change would put to rest the concept of whether that's really involved with organismic aging; and the same with oxidative damage.

These were meant to be goals that would be accomplished in 3-10 years, it was thought to be reasonable that they would happen and have a very meaningful impact. I think that the goals for this meeting today (and Aubrey will talk a little bit more about that) are both more ambitious -- in the idea of trying to actually reverse aging or to achieve negligible senescence -- and then they're less, because they're very specific engineering strategies targeted at subsystems. They might not necessarily have an impact on the overall length an organism can survive. I'm hoping that by the end of the meeting we can move toward some estimates of low and high timeframes for any goals that we think are realistic, some probabilities of their success, what's realistic and what isn’t, and maybe some rankings of relative importance for the goals. And if these in fact are a set of engineering efforts, as Aubrey had mentioned in his initial write-up, there should be sub goals that are present as well, and when we get there I'm sure Aubrey will talk a little bit more about that.

So that's what happened in Los Angeles a year and a half ago, we ended up with some very specific goals. The idea was also to come up with a very large prize -- a series of prizes actually, kind of Nobel-like in nature -- and I continue to try to raise that funding. I've had a few nibbles, but it hasn't developed yet. In fact, there was some conflict at the meeting in LA as to whether prizes of this sort were even desirable. I thought there would be a uniform "Yeah, this'd be great!"… especially since the people at the meeting were the ones who would likely get the money. I thought that 1/4 million dollar prizes or such would be desirable not only for researchers but also for elevating the profile of efforts to retard aging. I think that Aubrey’s idea of having a paper in Science, or at least a rendition of what happened here, and the very notion of reversing aging, if it is actually realistic, is something that may have similar effects, raising the profile of what is possible and separating out the hype and the ideas of what is not possible.

Aubrey de Grey: Before I start talking about my vision of today, perhaps it would be appropriate to spend a few minutes for the rest of you to comment on what Greg's said, on your own views about his meeting, and so on -- as Greg said, there's only two of you here who were actually there, but let's not restrict it to you two -- in terms of what you have to say about the idea of what Greg did.

Judith Campisi: I think the goals of the meeting were a little bit different. It was really to identify how will we know when we're on the right track, and what are the milestones on the way, that will tell us we're on the right track, that we're getting someplace. I think it's a requirement that we try to limit ourselves to the mouse at this time. It really becomes very complicated when you start getting into delivery and how do you engineer a human, and that's much too premature. I think at the moment that I see this meeting as kind of a subset, focusing a little bit more on the practicalities, because once you can do it in the mouse, then you can worry about how to do it in humans. But can we do it in the mouse? We still don't know.

Aubrey de Grey: Yes: in response to that, I wouldn't *quite* say that I want to restrict the meeting to discussion of mice and only mice, but certainly to organisms that we can manipulate by existing techniques, as opposed to organisms which we can only manipulate by very experimental techniques at this point, in particular gene therapy. Certainly many of the things we'll be discussing today are going to be based on transgenic interventions, which are of course infinitely easier in model organisms, especially mice, but not exclusively mice, than they are in humans. The other reason which I think makes sense as a reason not to focus too much on gene therapy is that there are plenty of people focusing on the practicalities of improving gene therapy delivery techniques, and so on, who aren't interested in aging at all, and what they achieve in the enormous number of different applications where gene therapy may be valuable will be just as applicable to us anyway. So I'm rather optimistic that by the time we've solved these other rather specific problems that concern us, in model organisms, that there will have been very substantial progress on the areas that we're really not going to be focusing on today.

Gregory Stock: I would like to agree with Judith that although the focus was on humans at the last meeting, and I think that's ultimately where our motivations are, that the actual goals were in model organisms, generally -- I think if you emulate caloric restriction you're going to do it in mice. So I think that actually that distinction is maybe more explicit in this case, but I think it's very clear that you're always going to be working with model organisms.

Julie Anderson: Also there already exist several transgenic models, at least for age-related diseases, that are currently being used to look at therapies for these age-related diseases. It's already happening, using animal models.

Christopher Heward: I think the significant thing about these meetings, and this one in particular, is that this, to my knowledge, is the first time that a group of serious scientists have gotten together and discussed the possibility of actually reversing the aging process. I think that's an important thing, because there's a tremendous interest in this field in the lay community. A lot of people would like to reverse the aging process, or at least stop it. There is a huge lack of understanding about aging in our society. Scientists have historically avoided entering into this type of discussion for the purpose of really establishing what is real and what isn't in the lay press. So, the more this type of meeting occurs, the more information about what is scientifically possible will be conveyed to the world and the less opportunity there will be for charlatans to step into this area and make claims that just aren't correct. I think we have a real problem with that right now. So I think this is a very important event and I'm glad to attend it.

Roger McCarter: If I can just add one thing to what Greg said: There is an eccentric millionaire from Houston who's already publicized a one million dollar prize for doing essentially what you're talking about. He calls it "curing the aging disease," and I think that reversing the aging process will do the job.

Aubrey de Grey: What's his name?

Roger McCarter: I knew you were going to say that, Aubrey, and I do have it -- he announced that, in fact, at the Pan-American Congress of Gerontology in San Antonio last year. And he has, of course, a very personal interest in this, since the man is a retired oilman, and he's about 75 I think, and is really angry that he's 75 and going downhill rapidly. So it's out there, and I'll get you his name.

Gregory Stock: Do you remember what the criteria were, how the decision would be made?

Roger McCarter: No I don't. I put this in the same category that we've just been talking about, that there's a lot of, sort of, flim-flam stuff going on out there, and I quite agree that it's very good that these words are being said aloud here where they can be discussed.

Bruce Ames: Maybe we should use the words "delaying aging" -- a little more modest than reversing aging or curing aging.

Gregory Stock: Aubrey has actually suggested that he thinks it's possible to essentially reverse, because I've had that discussion with Aubrey, and there's certainly a big difference between delaying or retarding, and reversing -- returning to some earlier levels of functionality. I think this brings up the issue that it's sort of a two-edged sword, because by using these kinds of words, basically it just adds fuel to the flames of all the ridiculous stuff that's out in the world, of selling nostrums and such.

Aubrey de Grey: In a few minutes' time I'm going to be putting forward the somewhat unorthodox view that reversing aging may actually turn out to be *easier* than dramatically retarding it. But I've got my own place for that in the argument. So, you may be a little bit surprised by what I have to say in ten minutes' time or so. But I don't entirely agree that it's more modest to delay aging than to reverse it.

Gregory Stock: I would say that logically that seems flawed to me, in that I don't see how it's possible to reverse aging without retarding it.

Aubrey de Grey: Right -- wait and see! OK: Andrzej.

Andrzej Bartke: I think whatever I say will be permanently marked by my accent. You mentioned the use of genetic DNA markers as a way of monitoring the aging process. I have some comments on microarrays and aging, I don't know if it's appropriate now?

Aubrey de Grey If you want to summarize?

Andrzej Bartke: It's not very long. I think that right now there is a sort of mixture of enthusiasm for this new approach, and some fairly harsh criticism of this approach -- you attended a meeting with Rich Miller, so I think you were exposed to this -- and I think we'll have to wait until some consensus emerges. Some of the work that was published on the use of cDNA arrays to monitor what happens with aging -- it's of course very exciting, the stuff that Prolla and Weindruch have published -- it's beautiful work and it certainly deserves the publicity it's received. But at the same time, as Rich Miller has communicated on numerous occasions, in most of these studies there's really a lack of acceptable statistical approach to data. And if you look at the methods section you find that the people compared two animals with three, or three with three, and they classify difference in the means of >100% or >200% as a biological difference which was detected. And of course in any other area of physiology this type of idea would be totally unacceptable. Everyone would ask the obvious question, what is the significance? And in all of these studies, that actually has not been tested. And some modeling has suggested that you can get such -- and greater -- effects by chance alone, if you make hundreds and thousands of comparisons. So even though these data look, in many studies, very exciting, and I think that the comparison of young and old, and old versus old calorie-restricted, that Lee and his colleagues did, really reassures that we're studying something real, at the same time, I think that all of this data right now probably consists of an unknown mix of real data and false positives. And I think that while it's obvious that it would be a very desirable goal to have a cDNA test to compare an aging cell with a young cell, I think that the practicality of it and how easy it might be, and how close it might be, is really yet to be determined. And I think one of the issues that may be resolved fairly soon is whether it's more productive to look at a huge number of genes, or to look at some genes for which there is a logical reason to look at them. And then the statistical problems become much easier if you don't deal with so many comparisons. So this is a very tough area which hasn't yet found its own accepted methodology.

Gregory Stock: I totally agree with you. These were 3-10 year timeframe goals, and it was actually to develop a bioassay, using microarrays -- that would be very desirable. But it's not easy at all. I would agree that most of the data that's out there at this point is not very good, and it isn't going to be good until the cost comes way down so that it'll be possible to do the kinds of statistics and the kinds of experiments that'll be necessary.

Andrzej Bartke: An example: there was a very nice paper in Molecular Endocrinology by Feng and colleagues looking at effects of triiodothyronine on gene expression in the liver of hypothyroid mice. The way they did the experiment, it was obviously a very drastic hypothyroidism, and replacement is a good way to find genes that respond. And they published, I believe it was a list of 55 genes that went up or down by the criteria that they used. And I compared their list to the list that Rich Miller generated in the tissue that we sent him, of the livers of dwarf mice, which is a congenitally hypothyroid animal, and he did a comparison of different ages, and had again a list of somewhere between 50 and 100 genes depending where you make the cutoff, and I compared the two lists, and there was *one* *gene* that appeared on both lists. And to make things worse, the effect was in the same direction, where it really should be opposite. So I think this kind of shows that we're all excited about microarrays but I think honestly we still don't quite know what they may tell us.

Aubrey de Grey: Unless there are more comments to Greg, I think I'd like to start on my vision of how today might be, and those aspects of this meeting which will be rather similar to Greg's and those aspects which will be different. The aspects, of course, which are most similar are, first of all that it's purely a roundtable format, with very little in terms of presentation (only a minority presentation component), and of course it's a similar-sized meeting, only about 10 people. The other thing that distinguishes this meeting from most of the ones that we go to is that the speculation content will be extremely high, which is always something I like -- I think it's important to be ambitious and to aim high and think what we could do, which sometimes we don't end up having much opportunity to do in more traditional formats. That has led me to a number of departures from what Greg did. The first question is who's here and who isn't here. First of all I decided that it would be a good idea not to bring in anybody who was explicitly involved in actually paying for what we do, because I think that we often shoot ourselves in the foot a little bit by thinking too soon about how to persuade people to give us the money to actually do the experiment, whereas if we spent a little longer thinking blue-sky technology and blue-sky experiments, before deciding how to make the case for those experiments, maybe we would come up with different experiments that we would choose to try to make the case for, and we might not be unable to make that case, later on, so I think it's valuable to have a little bit of a standing-back from the funding questions.

The second thing I've done is, you may notice, that everybody here is a mammal -- there are no invertebrates here, let alone prokaryotes -- and I think it was quite important to focus the meeting on a particular type of aging. Some people feel it's a little bit comical to say that cerevisiae ages at all; there's a well-known, well-established definition of what aging is in Saccharomyces cerevisiae, but at the level of cell biology it doesn't look very much like aging in anything else, and it's still extremely debatable, despite the trenchant efforts of Lenny Guarente to prove otherwise, that the actual *molecular* basis of aging in cerevisiae has all that much to do with aging in higher metazoans. So I thought it would be valuable to try to get away from questions about the definition of aging, and similarities or contrasts between widely divergent species, by focusing and having participation only from people who work on mammals. Another thing that came up in what Greg said earlier, and what Judy said, was that I felt it would be important to focus on the basic research in model organisms, more than on delivery of such technology to humans. So while I'm very pleased to see Chris here, who's of course very much engaged on the medical side of things, and who may help to bring us down to earth a little bit in a session that we've scheduled right near the end of the meeting, nevertheless I think it's important not to let that aspect of the problem outweigh and overwhelm the more speculative aspects that we may discuss.

There aren't really any genomicists here either. I'm not really terribly excited about microarrays -- I think that the extent that they can tell us how to *reverse* aging, in particular, which is of course the focus of today's meeting, is very limited. I've argued rather forcefully in the past that, for example, trying to restore the patterns of gene expression in an old individual that exists in a young individual of a given species may actually be the wrong thing to do and may actually be harmful to the organism, because most of the changes that are detected -- even insofar as they really are detected, bearing in mind what Andrzej just said -- most of the changes that are detected in a microarray study will typically be coordinated changes that the typical cell in an organ is performing, as an age-related change, and because it's coordinated it seems to me that most such changes are likely to be compensations for probably non-genetic changes such as accumulation of mutations, and accumulation of junk like lipofuscin and so on -- accumulation of contamination that the body has to make the best of. So I'm rather of the view that the things we learn from microarrays may actually turn out to be worse than useless, to be actually distractions from the real processes that we need to intervene in if we want to do something about aging.

But within those constraints, I've done my best to have as wide a breadth of expertise as possible, and certainly had we had Steve and Tony here I think we would have been able to say that virtually every aspect of human aging was represented, one way or another, which was certainly what I was hoping to do. It's pretty difficult to do that in a small meeting like this, but I'm delighted that that's basically been close to coming about.

So now I want to spend a few minutes talking about why I wanted to focus on reversing aging, rather than delaying or postponing aging. First of all, I suppose my main reason is to come back to what Chris and Roger were saying a few minutes ago, which is that I'm not in favor of aging. I consider it basically a barbaric, uncivilized phenomenon that shouldn't really be tolerated in polite society, and so I would like to do something about it. Now there's a variety of opinions on that -- there may be a variety of opinions in this room on that -- there certainly is within the community, as to whether aging is a good thing for society and so on, but the fact is that aging is unpopular in the world at large. People aren't really in favor of it; if we could provide them with a way to avoid it they would take it, by and large; and we as experts in our field -- and as publicly-funded people, in particular -- have a strong responsibility, it seems to me, to accept that that is a legitimate aspiration. It's an aspiration that's very widely held, and we ought to be telling people the best we can about how they might actually achieve it. And this doesn't seem to me, from my relatively brief (compared to most of you) experience of this field, to be something that we talk about nearly enough. I personally don't think, in view of the dramatic advances -- and rate of advance -- that we are making in this field, that it's any longer defensible for us, especially, to remain silent on the question of whether we could actually, in due course, do something to reverse the decline that people are experiencing as they get older.

Reversing is also important because, let's face it, people aged 60 or 70, who are doing all right, are not in anything like the shape they were 30 years previously. Sure, they'd be happy not to get any worse before they die, but they'd be much happier still to be restored to some sort of full degree of functionality that they enjoyed in years gone by. So I think that, again, this is something that is widely seen as highly desirable, despite the acknowledgement, of course, by anyone who thinks about it at all, that there would be enormous social upheaval and enormous problems in adjusting to such a different world as would exist with a dramatic increase in healthspan. There is, nevertheless, the desire to see this happen.

I said I was going to talk about why reversing aging might be easier than retarding aging. First of all I want to comment on what Greg said about this being logically impossible. Of course Greg is right in one sense -- that reversal is the limit of retardation. What I really meant was: therapies that are *aimed* at dramatically retarding aging, let's say at increasing human healthspan by a factor of two or more, are an uphill battle. If they're only aimed at doing that, and they're not aimed at eliminating or reducing the accumulation of damage that aging comprises, then it's going to get steadily harder and harder to continue that reversal. I think it's no surprise that longer-lived organisms have proved much harder to extend the lifespan of -- even by a small amount, like 10-20% -- than shorter-lived organisms. We've seen, for example, a very important result published about a year ago now by the Italian group of Pelicci, where knocking-out of just one gene was able to extend the lifespan of a strain of mice by 30%. Very important work, but it's been widely, not exactly criticized, but dismissed in terms of its generality, on the basis that the strain of mice that was used in the experiment was extremely short-lived in the first place -- it only had a maximum lifespan of about two years. And this is undoubtedly a major difficulty. I'm an optimist on these things, so I have a small wager with George Martin that when it's done on C57Bl/6 there will actually be up to a 20% increase in lifespan, but I wasn't willing to put any more than $5 on it... So I think that we have to remember that shorter-lived species are much easier to extend the lifespan of than longer-lived ones, and this may be a big message to us that retarding aging is rather hard.

I think that the reason *why* retarding aging (by which, I have to say again, I mean approaches to postponing aging which are conceptually only aimed at retardation) -- the reason why they are unlikely to be very much more successful than they've already been, in my view, is because they aim at targets that are very intimately linked with normal metabolism. So for example getting rid of superoxide, which is of course the originating free radical, a side-effect of respiration, has had very important successes recently: it was found that you could increase C. elegans lifespan by 30 or 40% by a small molecule mimetic of superoxide dismutase. Very important result, undoubtedly, but in mice -- a somewhat longer-lived organism than C. elegans -- if you actually *eliminate* the *natural* superoxide dismutase in the cytosol, nothing happens (to speak of) -- the lifespan of the mice is essentially unreduced. That's not true for the mitochondrial isoform of superoxide dismutase, but it is, nevertheless, true for the cytosolic one. This shows that, really, mammals are dead complicated. And if you intervene in a process that's very closely and intimately linked with something that we have to do, such as respiration, or transcription or whatever, or if you do something even more ambitious like actually trying to reduce the rate of production of superoxide in the first place, which is what a number of people have argued is really what you need to do in order to extend lifespan much, then really you're fiddling with something that cells have evolved to make the best of, for a long time. Because they know they're making this stuff. Superoxide, for example, is not a particularly reactive substance, and it's just about usable as a signaling molecule, and certainly some other free radicals, such as nitric oxide, are now known to have extremely widespread signaling roles, but superoxide almost certainly has some as well. So this is an example of how if one intervenes in something that's very intimately linked with metabolism, then one might actually have difficulty avoiding deleterious side-effects that would offset the beneficial effects that one had been thinking about when one designed the therapy.

Now, I've used this phrase "closely linked to metabolism." I want to try to explain what I mean by contrast with the sorts of things we need to do if we were to want to reverse aging. It seems to me that really the most influential aspects of aging are the ones which we can call "one-way" changes. So, mutations -- changes to the DNA sequence, in the mitochondrial DNA or the nuclear DNA -- is an example of a one-way change, in the sense that once the mutation, the actual base-pair substitution or whatever, has occurred, there's no way of telling that it wasn't like that originally, from the cell's point of view. The cell can't chemically distinguish it from what it had before. So there's no way back naturally -- there's no way the cell can do anything about that. But that's the end point of a very long chain of molecular events. Let's say it was oxidative damage. So you've got production of superoxide; the superoxide reacts with a metal and you've got a more reactive radical, like hydroxyl radical; that reacts with, let's say, a lipid, causing a bunch of lipid peroxidation chain reactions; maybe eventually there's a reaction with a nucleic acid, causing an adduct like 8OH-deoxyguanine for example, one of the numerous oxidative lesions that have been found in DNA; subsequently, if that damage isn't repaired, there may be an error in DNA replication, causing a mispairing of a base, because of the oxidative lesion; and eventually you may actually get a double-stranded base-pair change, which is the final point. But every single step before then has the capacity for reversal. And not just the capacity: there is actually stuff in there that's *doing* that reversal. So what we have in our cells, taking a little bit more forward this example of DNA damage, is we have a whole bunch of precursors of mutations, which are at some sort of steady-state equilibrium between the rate at which damage is occurring and the rate at which it's being repaired enzymatically. And however much we bear down on the forward rate and bear up on the backward rate of those reversible processes, those equilibria, there's still going to be a gradual accumulation of the end-point, which our cells can't do anything about on their own. So that's what retardation of aging is about, whereas reversal would be to actually go in and get rid of the mutation one way or another. And I'll be talking at some length in the slot that's "really" me, at about 3:00, to do with how we might do this in mitochondrial DNA, and I'll also be presenting some very new data from the group of Steve Zullo at NIH, showing that the approach that I've been promulgating for quite some time is extremely feasible. The point is that getting rid of mutations may actually be easier as an anti-aging therapy because it is so far removed from normal metabolism (let's say, transcription that allows an oxidative adduct to form in the first place, when it might not form if the DNA was wrapped up in histones). It's much further away in terms of a causal chain of events.

Probably an even more explicit example is the sort of junk that builds up in our cells, especially in lysosomes. Lipofuscin is the stuff that's talked about most often. This is very clearly the end product of an extremely long and complex series of lipid peroxidation chain reactions and other related reactions. Now, lipofuscin in itself has been rather dismissed, popularly, as having much of a causative role in aging, because it looks pretty inert, plus it's wrapped up in these organelles, lysosomes, so it really ought not to do people any harm. But as has been recently pointed out (and also experimentally somewhat supported) by Ulf Brunk's work in Sweden, this may be an oversimplification. The actual functionality of lysosomes may be impaired by their lipofuscin load. I've recently become interested in other things that accumulate in lysosomes of other cells, in particular in macrophages in the artery, and also in the retina, which may have much more obvious links with age-related pathology but may also be related simply to the accumulation of stuff, in lysosomes, that can't be degraded. If we were able to reverse the accumulation of this stuff in cells, then we would be reversing the end-product of these reaction that, until that last step, are basically reversible things. And it seems to me that not only is this large distance from metabolism (in terms of the length of the chain of events) something that helps us potentially to avoid side-effects, it's also likely to feed back into the earlier events. Ultimately, any network of metabolic and biochemical changes which comprises only equilibria between forward and backward rates of processes (let's say the backward rates are all enzymatic, generally they are) -- any system that's made entirely of reversible reactions is going to be, itself, at equilibrium. And equilibrium is basically what we call negligible senescence. So my view is that if we were able to bear down, simultaneously, on all of the really one-way processes that are the culminations of these complicated chains of events, then we would actually be in a state in which the steady-state levels of all these intermediates (oxidative lesions, lipid peroxides, things like that) would actually not be increasing. There are some exceptions to that, certainly: there are some areas of aging which are undoubtedly what we might call programmed. So the most explicit example is probably the loss of follicles in the ovary, which undoubtedly has hormonal consequences, which certainly have very pleiotropic downstream consequences. But that, it seems to me, is really quite an exception, quite a special case, and most of the other aspects of the network that we see as aging, the causal network that we see, are driven much more influentially by the end-products, by the accumulated events, the loss-of-function mutations, the accumulation of junk -- much more by them than by the reversible events.

OK, well that was a bit radical; I'll ask people to comment shortly. I just want to say one more thing, which is that it may sound as though what I'm saying is that we know enough now, we can get on with implementing intervention. I'm not saying that at all, certainly not. There's an enormous amount to be found out, obviously, still, about what the aging process is. But ultimately if we accept that a lot of people want us to do something about it, then we have to think not simply as basic scientists but also, simultaneously, as engineers. And if one is thinking as an engineer and thinking about how to solve a really hard problem, then sure, one starts exploring the problem, and trying to understand it in more and more detail. But eventually one has to make a transition to starting to design what to do about it, at ever-increasing levels of sophistication of that design, the design of those interventions. Now of course the transition between the exploratory phase and the design phase is by no means a sharp transition, and one may spend a long time essentially flipping between those two modes of thought. I think they are very different modes of thought. I've often felt that I'm psychologically much more of an engineer than a basic scientist; I don't think many of the people around this table would say the same, but that's fine, really. I just think that we have a sort of duty to think in both ways, most of the time, these days. And the last thing I want to say is that it's very valid to argue from a pessimistic standpoint, that even if we were able to intervene in all the aspects of aging that we know about at the moment, and to reverse them, then there would still be plenty of subtler aspects of aging that would catch up with us, pretty quickly, so we wouldn't really have progressed very far. I think that may be true. But I also think that, going back to basic scientist psychology, the best way to find out more about those subtle processes is to unmask them: in other words, to try to eliminate the more overt processes that we understand in quite a lot of detail already, and thereby to let us see more of the detail and more of the clarity of the processes that we can as yet only study rather poorly because they're so heavily obscured by these other ones.

OK, that's it. I expect I've probably said a lot of nice reactionary things; please comment.

Bruce Ames: In 60 million years of evolution we've gone from a 2-3 year lifespan mouse or rat to an 80-year-plus lifespan human, so clearly evolution can do this in a relatively short time.

Aubrey de Grey: I think that's a very important point, and in fact this might be a good point to bring up the work of our absent friend Steve Austad, who of course studied a few years ago a population of opossums in an island off Virginia, and found that even though they had only been isolated in an environment that was free of predators for about 5000 years, equivalent to about 4000 generations, nevertheless they had managed to evolve an increased lifespan of about 70% longer than the highly predated opossums on the mainland. This is an area where I do actually feel that genomics might be able to tell us a few things, because when you have organisms that are *so* closely related, that have diverged only a few thousand years ago, the sequence divergence will be of course really rather limited, and one may indeed be able to establish certain aspects of how the increased lifespan has come about by a straightforward genomic approach that doesn't involve prior hypotheses.

PT: To address your question of reversing versus prolonging, I think there are already some examples that it's possible to reverse. If we consider atherosclerosis as a generally universal, progressive deleterious process, one used also, in the past, to say it was irreversible. But today we know that we can, with drugs etcetera, that some of the atherosclerotic lesions can be reversed. So that's a very good example, and this reversion is associated with the transition of the cardiovascular diseases' mortality since 1970. So it is not only a real fact but a fact which has important consequences. So that would be a case of reversing aging. Of course, we don't know if we can consider atherosclerotic lesions as a marker of aging, but if we do, then we have the possibility of reversing, which would be an important example.

Judith Campisi: I have a comment: We're doing that, even as we speak. We've seen that in the last 50 years we've eliminated certain infectious diseases, for example. I do want to point out, though that 60-70 million years is a very long time, and a 75-year-old guy is not going to be happy! But I understand the point: the point is there are probably relatively few genes, and certainly in the past that wasn't know to be true, because 5000 years is a relatively short time. Although, remember, what happened in evolution is selection for existing alleles. It's unlikely that there were new genes that evolved in 5000 years, but there was selection for alleles that additively gave rise to the, again, rather modest increase in lifespan.

Aubrey de Grey: You make an excellent point. And in fact a number of the things that I'm going to be talking about in the later sessions today are precisely the optimist's answer to the challenge that, well, evolution's only been able to manage this rather modest factor in this rather long amount of time, how on earth do you think you're going to do any better? And I've always been rather a fan of Crick's famous comment "Evolution is cleverer than you are", but the fact is that evolution has had fewer tools. We have the capacity genuinely to introduce new genes, not simply new alleles, into the genome of a model organism, really rather easily. A number of the interventions I'm going to be talking about will actually involve exactly that. So I feel that the comparison with evolution has a lot of value, but it can also be very misleading if we take it too literally.

Roger McCarter: However I think we do have to bear in mind your starting point for this discussion, and that is the question of metabolism. We consist of a series of millions of intertwined reactions, and you cannot perturb one.

Aubrey de Grey: Precisely.

Roger McCarter: You perturb everybody, and there is a readjustment. And depending on how radical the readjustment is, the organism does survive, and may survive better, or it may not survive. And so I think, as a structural engineer, you have to in fact do the experiment to make sure that the structure is not, in the end result, weakened as a consequence of your changing one strut, if you like.

Aubrey de Grey: I completely agree.

PT: There is now a whole new area of genetic epidemiology. We know that the genetic diseases which are produced by only one or two genes are very few, and there is this idea of creating genetic profiles of genetic susceptibility, on which, then, the environmental factors will produce the disease or not. And that would be very useful, because if you know what your susceptibility is then you would specifically identify factors that you have to avoid, and those that eventually you have to favor instead. So that's another way of attacking the problem, which I think would be very useful.

Gregory Stock: I'd like to make a comment about what you said about the goals, and basically I think that's a very good one, because the idea of being able to reverse or control aging some day is something that is aspired to by large numbers of people. And in fact would be, as I'll discuss later, much less disruptive in a lot of ways, socially, than simply retarding aging and slowing it down. So I think the idea of discussing these things is very important. The notion that you're essentially doing re-engineering, and whether evolutionary timeframes are meaningful, to me that isn't very relevant actually. Because if you say "well, we can't do that in 50 years, or in 25 years, but extend to 500 years or 1000 years", there are still dramatic differences between that and evolutionary timeframes. You're still re-engineering. So to say that it's a hard problem, and that it might take much longer than we think, and that it's going to be very complex is one thing, but I don't think that the thousands or millions of years that are involved in even modest changes evolutionarily have much relevance for this process. It's really different, it's a design process you're talking about, with very different tools.

Judith Campisi: That was the point I was trying to make, that what happens over evolution is selection for pre-existing things in the gene pools, and the slow emergence of those phenotypes. And that's *not* what we're talking about here.

Andrzej Bartke: Also, in defense of short-living animal models, I think that if short life in a particular strain or species is due to a definable disease, and life-extension is by eliminating or reducing that disease, then of course the data does not have wide applicability. But if the mechanism relates to general mechanisms of aging, then it would probably apply just as well to a longer-living strain of the same species, or to a different species.

Aubrey de Grey: Yes. I think I would agree with you that it's much *more* likely to apply, but I wouldn't agree that it's *likely* to apply. I think that at least it remains a very open question the extent to which a given intervention that extends lifespan, for example, of a 2-year-lifespan mouse will extend the lifespan of a 3-year-lifespan mouse. Perhaps Roger could actually summarize the situation in CR: Do you know whether there is a tendency for very short-lived strains of mice to have their lifespan extended by a higher proportion by CR than long-lived strains do?

Roger McCarter: No -- in general I think the extension is proportional to the degree of restriction, and I guess that's one reason why people who work in the area feel that it's affecting really fundamental aging processes, whatever they are, because it does seem to have very similar effects all over the place. You know that I once plotted all of the available data on the degree of restriction versus the extension of lifespan, and it's fairly linear.

Aubrey de Grey: But for a given degree of restriction, is there any study comparing a long-lived strain with a short lived strain, and seeing which one had its lifespan extended by a greater proportion?

Roger McCarter: Not to my knowledge.

Judith Campisi: There is Steven Austad's data, and it's really too bad Steven's not here, suggesting that it may not work in wild strains, which he also predicts would have a longer lifespan than the laboratory strains.

Roger McCarter: Yes, but Aubrey's question was are there already existing studies that demonstrate that, and I agree that Steve's is very relevant, but we don't have the controls.

Judith Campisi: Yes, that's correct. And it needs to be repeated.

Roger McCarter: But it's a very remarkable result, no question.

Andrzej Bartke: But caloric restriction is being applied to primates, and I think the data for the squirrel monkey are already quite convincing that it extends life, and the data for rhesus are suggestive enough that it would be really bizarre if it didn't. The people don't yet have the longevity data, the physiological data are very convincing.

Aubrey de Grey: Yeah -- I mean, certainly the physiological data and the biochemical data look as though everything is happening the same way -- marvelous. But, I'm a good deal more hesitant than you to say that the present situation, even for the squirrel monkeys, is all that persuasive, simply because we don't know what happens to the later part of the survival curve as a result of what happens to the early part. We don't know how much aging is simply being rectangularised, as opposed to postponed, at this point. But we have to go with what we have. Any other comments?

Roger McCarter: Just, perhaps, the effects of exercise -- to address one of the comments that Bruce made -- in terms of reversal. If you look at the function of a given physiological system, let's say the pulmonary system, the pulmonary system of master athletes ages at the same rate as the pulmonary system of sedentary individuals. However, at any given age, the level of function of the pulmonary system is much higher in athletes than in the sedentary individual. So it appears there that the aging rate has not been affected. However, should you choose to start exercising, you'll bump yourself up to the point that you had when you were 10-15 years younger. So one interpretation of that is that it's not a retardation but a reversal.

Aubrey de Grey: Absolutely, and I definitely want to be discussing exercise at some length in your session.

Gregory Stock: That brings up a very important issue -- what do we mean when we're talking about aging? And I don't want to get into a mess --

Aubrey de Grey: And nor do I!

Gregory Stock: No, but if we talk about reversing and you go back and restore function to a more youthful time, and then in ten years you stop exercising and go up to where you would have been at that age, there's no reversal going on there at all.

Aubrey de Grey: That's not true -- I feel that there's a simple answer to that. There is reversal going on, but it's only reversal of some components of aging. Reversal of some components, while other components of aging are continuing to progress unabated, results in a somewhat artificial situation in which, physiologically, phenotypically, at the gross level, one may see, let's say, increase in ambulatory activity or fitness or whatever, but since the other things are still going on, they're going to catch up with us. So absolutely I agree that reversal of aging cannot have what we would like to think of it as having, in other words a significant impact on healthspan, unless it is across-the-board, and applied simultaneously in all aspects of aging. And the reason why these things that look like reversal in some ways are not reversal in other ways is because they're only partial.

Judith Campisi: Do you think we could find something -- some things?

Aubrey de Grey: A large panel of things, yes.

OK, we're starting to run a little bit late; maybe I should try to move on to talking about negligible senescence in more detail. We'll of course have a pretty flexible schedule, so we can come back and not necessarily stick precisely to the schedule right through the day. Well, this was supposed to be Steve's session, but I'm going to make the best of it: I'm going to get various people who know a bit about Steve's recent work, especially, to chip in and talk about it. I think it's worthwhile first of all to remind ourselves what negligible senescence is. It was defined over ten years ago by Tuck Finch as a statistically undetectable increase in risk of death with age -- in other words, when you're N years old, your risk of dying before you get to N+1 is basically the same, for any N starting at adulthood. This is very relevant to the idea of reversing aging, because one *can* think of negligible senescence as "starting late" -- for example, we might age until the age of 70 and then find some way to keep us in a 70-year-old physiological state. That would conform, in a formal sense, to Tuck's definition of negligible senescence. But not to the spirit of it, I don't think. I think realistically what has to be meant by negligible senescence is maintenance of the risk of death that one had at the beginning of adulthood.

Now, in terms of whether negligible senescence really exists, since it is after all defined statistically as the *failure* to detect an increase in the susceptibility to life-threatening challenges, the question of whether it exists is very hard. In addressing it, it's really absolutely necessary, I think, to start with the fantastic study that came out a couple of years ago from Daniel Martinez, on hydra, which absolutely transformed the situation with regard to what we actually see in real organisms. hydra reproduce by budding off a new, small hydra off the side; they do this roughly once every other day; and if you try to keep hydra alive without being very careful about the conditions, they don't tend to live all that long. But it was suspected that most of the mortality of hydra was essentially extrinsic, age-independent mortality, and Martinez succeeded in identifying conditions for growing hydra in which the hydra were sufficiently happy that the extrinsic mortality was pretty much eliminated. And he kept these hydra alive, and budding away every other day, for four years, before some accident with his fishtank. Now, that is a ratio of lifespan to rate of reproduction, and also to age of onset of reproduction, that is totally off the scale compared to any other organism known.

Judith Campisi: I'm not sure that this is a productive discussion, for the following reason -- and this came out of the Symposium on Slow-Aging organisms that Tuck hosted. I'm not convinced that it is useful to discuss these examples -- hydra, plants, yeast, for example -- in the same way we discuss mammals, and the reason is the totipotency of the cells within the organism.

Aubrey de Grey: You took the words out of my mouth. The hydra example is *superficially* mind-blowing. It was discussed in somewhat modest terms in the actual paper, in straightforward negligible senescence terms, as a failure to detect an increase in the propensity to die with age -- no increase in the *rate* of death -- but the actual result was far, far stronger. It was no death. There was, like, one or two hydra died, in the entire experiment. Which meant that he really had completely eliminated the extrinsic mortality, to speak of, and therefore demonstrated far more strongly than anyone else had that there was no increase in intrinsic mortality.

However, as Judy so rightly says, there's a difference between hydra and us -- a very big difference -- which is that every cell in a hydra is turned over. Cells in hydra don't live very long, any more than cells in plants live very long. 3000-year-old bristlecone pines, as Hayflick pointed out some time ago, don't have any cells in them that are more than about 30 years old, that are actually still alive. So it's extremely dubious to make comparisons between plants and humans. It's slightly less dubious to make comparisons between hydra and humans, because after all hydra are animals, but they're only just animals; and I'm actually not terribly in favor of making too strong comparisons between lower invertebrates, such as C. elegans and Drosophila, and mammals, because the body plan, in terms of what cells do and how cells can be turned over, is dramatically different. In the case of C. elegans and Drosophila the difference is really the opposite difference, in the sense that essentially every cell in those organisms is postmitotic, whereas we have a rather wide spectrum of mitotic competence of our different cell types. So yes, I think it's a very important question to address whether negligible senescence really does exist in organisms like us, and I feel that the situation is still extremely open. The best-known examples of vertebrates which are, certainly extremely long-lived, and appear to exhibit very slow senescence if senescence at all, even they are somewhat questionable because of their environment. Negligible senescence has never been claimed to exist in any warm-blooded organism -- mammal or bird. It's been claimed to exist in certain amphibians, and the most spectacular examples are in fish. Rockfish are the ones which are catching most people's attention, because they've been found now to live, if I'm not mistaken, occasionally at least, up to over 200 years. But the enormous problem with these rockfish is that they live in an environment that is *incredibly* cold, and also incredibly low in oxygen. So in terms of their deviation from the legendary "rate of living" generalisation, they may not actually be terribly much better than we are -- if *any* better than we are. They may not actually have much better life energy potential than humans, or certainly not than birds, who certainly outstrip humans by some way.

Judith Campisi: One thing that came out of this symposium which I was very impressed with was that I do now believe in the dating methods that can determine the age of a rockfish, and other organisms. I was very skeptical, you know, they grow throughout life and this is a very big rockfish so it must be old, I said no, I'm sorry! But now there are very serious scientists who are tagging these fish and injecting them with tetracycline, so the way they age a fish is there is a bone in the head that grows rings, very much like a tree, does it every year, and what these scientists are doing is injecting with tetracycline so they light up a ring. And then they come back, seven years later and they ask, are there seven rings? Is this really an accurate way of determining the approximate life of the fish? I was pretty impressed that there's a series of probably pretty good estimates of these fish that are 150-200 years. But what we're really missing, which is still completely anecdotal, is "I caught a big rockfish and it looked pretty healthy when I cut it open". You know, no pathology, no real objective measure that they have really not undergone some aging process. It doesn't mean to say it doesn't happen, but the there's no data out there yet.

Gregory Stock: That's still pretty good, the fact that there exists a 200-year-old rockfish.

Judith Campisi: 150 -- but there are 150-year-old elephants. Nature has given us -- species have evolved characteristic lifespans, but the question is, is there negligible senescence?

Julie Anderson: And in this cold environment with low oxygen...

Aubrey de Grey: This is the real problem, yes. So in the absence of data, which Judy so rightly brings attention to, I'm always keen to try and do thought experiments and try and decide, well, is this likely to exist? Is it plausible that negligibly-senescing vertebrates could exist? And if I had to come up with a single strongest argument that makes me think that negligible senescence probably *doesn't* exist, it relates to cancer. Because it seems to me that, whereas all other aspects of aging -- the end-product aspects of aging that I've spoken about, like accumulation of mutations and accumulation of lipofuscin and so on -- all these things could conceptually be overcome by different physiology, different cell biology than we have, like for example better cell turnover, better induction of apoptosis in appropriate cells, things like that. The thing about cancer is that it's a war between us and them, us being the organism and them being the cells --

Judith Campisi: Actually it's a war between us and us!

Aubrey de Grey: Yes, absolutely -- it's natural selection at the cellular level, within the organism, and eventually the cell is going to outsmart the organism, it seems to me. So eventually, pretty much however strong the defenses are, that an organism with our sort of body plan has, we're going to be outsmarted, and cancer's going to find a way to get away. I've been pretty impressed over the years, though I know Bruce hasn't, with the low tendency of sharks to get cancer, that's been reported. Bruce rightly points out that the epidemiology of sharks is decidedly lacking, or at least limited, but the anecdotal evidence is at least as good as the anecdotal evidence for negligible senescence. So maybe it's worth considering the possibility that we might substantially be able to improve the defenses that we have against cancer by emulation of organisms like that. However, I think that an approach which really eliminated our risk of death from cancer would have to be a good deal more high-tech, shall we say, and a good deal less in the way of copying what organisms that are already very similar to us do. And Bruce and I will be discussing a number of those ideas in his slot. What do people think about that? Does that sound to you like a persuasive argument that negligible senescence is really a mirage?

Andrzej Bartke: If you present eliminating age-related mutations as a feasible and realistic goal, then why would mutations relating to cancer not be in the same category?

Aubrey de Grey: This is a very good question. I don't know, at this point, whether any mutations that are involved in loss of function other than cancer are involved in determining our lifespan. It seems to me that cancer is certainly a life-threatening disease which is based around nuclear mutations, but I'm not sure whether any other life-threatening disease is based around nuclear mutations.

Bruce Ames: If you look at cancer versus age of the animal, it goes up at something like the fourth or fifth power of age. So mice live two years, rats live three years, and so on. And so clearly evolution is more able to defend against cancer.

Judith Campisi: I think that this is really an important point that it hasn't been eliminated. I think this is really something that we need to think about -- what is the difference that separates the species?

Bruce Ames: Part of it is metabolic rate.

Judith Campisi: But humans are different from mice - maybe they can repair DNA damage better.

Bruce Ames: So anyway, evolution is handling this.

Gregory Stock: Your argument that negligible senescence couldn't exist because eventually the cells are going to win out, that doesn't seem very convincing, because it's just sort of hand-waving. There are clearly mechanisms for delaying the time when cancer is able to win out, so there's no reason to think that one couldn't do that indefinitely. There may not be any motivation -- it may not be desirable.

Judith Campisi: Greg, I've thought about this problem a lot, and I hate to say it but I agree with Aubrey in this very depressing thought. I kind of agree with him that cancer is going to be the hard wall, at least within this modern era of aging research. I don't want to talk about fish because I really think that it's a different biology. When you talk about mammals I think cancer is a very hard wall. If you think about other aspects of aging, these are degenerative changes in aging, and we have at least in our minds a conceptual framework for replenishing degenerating tissue -- stem cells, reversing AGE cross-links, this sort of thing. Conceptually, we do not know how to stop a cell that is undergoing somatic mutation in the nucleus, and prevent that cell from growing inappropriately, which is the evolution, the selective pressure for that cell to win out. We don't have an idea yet, in biology, for how to do it. We can think about the immune system and killing them off as they arise, we can think about telomerase, but stopping them from occurring is going to be a hard wall.

Gregory Stock: But at what point do you detect that wall?

Bruce Ames: So clearly, every year life expectancy gets longer, so we're moving up that curve, but on the other hand people understand a lot about prevention -- it's going to be a series of curves that just take more time.

Judith Campisi: But that's true for everything. I mean if we as humans got cancer in two years, you would never see us because we're not sexually mature. So evolution of course has designed strategies for everything, including the degenerative diseases -- everything's time of onset has been increased throughout evolution. I'm not talking about whether we would *die* of cancer -- that's something where we're not sure what the future holds. But *getting* cancer: I know pathologists, who have told me, at autopsy, old people have cancer. They may not even know it, but they've got it. Every male probably has the seeds of prostate cancer. We're going to get it. So in terms of eliminating it I'm not optimistic, at least within this era of the biology of aging.

Aubrey de Grey: I want to explore briefly something that Andrzej just brought up -- that it may be a little bit doubtful whether we should mark out nuclear mutations leading to cancer as distinct from other nuclear mutations. Now I agree that a lot of the logic that underlies what Judy and I have been saying also applies to mutations that don't lead to cancer -- mutations that lead to some sort of loss of function, and doesn't necessarily induce proliferation. It seems to me that some of those types of accumulation of mutations could in theory be a very important aspect of aging, albeit almost certainly one that needs a lot of unmasking, largely because of the fact that many of our cell types remain quiescent for a long time, with a lot of genes not being significantly expressed, or at least not expressed in a manner that's necessary for the cell to do its job. And then some signal happens, let's say the cell next door dies, or whatever, such that suddenly genes have to spring into action which have not had to be in action before. Now this poses a very strong conceptual difficulty and challenge to eliminating such mutations, because if a mutation gives rise to a phenotype then at least you've got a handle. You've got something to eliminate. Whereas if you have a population of cells which are quietly accumulating mutations in genes that they're not using, and then they are required to do something, and they can't do it, then the situation is altogether different, and might lead to pathology as well. Now I've thought about this a little bit, and explored it a little bit, and I actually remain uncertain whether we can yet say that any aspect of aging as we currently know it does revolve around nuclear mutations other than ones involved in cancer. Can anyone do better than me on that?

Judith Campisi: I agree. I would go so far as to say that there's absolutely no evidence that somatic mutations that lead to cell dysfunction have a role in aging. First of all if you think about the numbers, 10% of the genome is expressed -- even whatever a cell has to do, it only needs 10%. These genes are in a vast sea of DNA in which virtually no point mutation is going to lead to a problem. And the chances of your liver, say, accumulating a critical number of mutations, you can't imagine it, it would be ridiculously low. Plus we have evidence from Jan Vijg's work, which says that if you look at how many somatic mutations accumulate in the mouse over the course of its lifetime, it's up by a factor of two or three at most. We're not accumulating hundreds of dramatic mutations in silent loci -- so this was a neutral reporter integrated into the genome of the mouse. It's negligible. But what I would say, that I think we haven't paid enough attention to, is the cellular response to particular types of damage. Not necessarily what gene the damage happens to, but one double-strand break, and wham! That cell is not the cell it was before the double-strand break. And I'll try to address that a little bit in my talk. I think that has a much greater impact than exactly where the mutation is. So going and revisiting somatic mutations that don't lead to cancer, I agree.

Aubrey de Grey: It's another hard wall, but it's a hard wall that's way out of sight.

Judith Campisi: And it may have such a negligible effect that it's not worth dealing with.

Roger McCarter: Bruce, could we have a comment on the mitochondrial DNA, because I'm intrigued by the whole argument.

Bruce Ames: Mitochondria are turning over, so the lysosome is eating up 10% of them per day, the mitochondria are putting out a little more superoxide, or that's the supposition, nobody's proved that, but that's what people think, but your lysosomes which are doing this turnover get clogged up, and then the whole system's going out, so the oxidants from mitochondria are going up with age, but I think it's not just the genes.

Roger McCarter: But how about the idea that if you take a very old individual and exercise them, their muscles increase their mitochondrial number perfectly well? So they retain the capacity for increasing the oxidative capacity of the muscle fibers, into very old age. So it does seem to me that if the body needs more healthy mitochondria, it simply makes them.

Bruce Ames: Nobody knows whether there are stem mitochondria. Maybe there are stem mitochondria the way there are stem cells, those mitochondria just lolling around not doing anything serious except dividing once in a while, and the other mitochondria are doing all the work.

Aubrey de Grey: Yes, Bruce and I have discussed this a little bit and tried to think of a way of testing it; we didn't get very far.

Gregory Stock: Did you say it's 10% a *day*?

Bruce Ames: That's the number I have.

Aubrey de Grey: Yes -- in rats it's about that. There's very preliminary data on humans, just in muscle, which leads to a number of perhaps 2-3% a day, so the same order.

Roger McCarter: Could you also comment on the idea -- I know that CP Lee did some high voltage electron microscopy on muscle several years ago, and found that all the mitochondria in the muscle fibers that she looked at were connected. And so the idea was that you've got one giant mitochondrion, that actually buds. And this is 20 years back, and I've never seen anything about that since that time.

Aubrey de Grey: I can make a few comments on that. There are studies which identify these highly reticulated mitochondria, two aspects of which are extremely important and which cast doubt on whether this really means very much. One is the EM study of mitochondria (taken out of muscle, again), which has been done, I think the most important study has been done by Vladimir Skulachev, in Moscow. And he was able to show pretty clearly that, in most cases at least, what looked at the light microscope level like reticulated mitochondria were actually mitochondria that were perfectly separate from each other but were abutting, so that you could actually see four membranes, inner-outer-outer-inner, stuck together. And Skulachev found this incredibly often, this structure, and he postulated that what's actually being done is that there is some sort of gap-junction-like thing forming between the mitochondria, some sort of channel. He proposed that the physiological role of this might be, in long cells like muscle fibers, to transmit electrochemical potential from an area with high nutrient density to an area that needed ATP. Because you could just transmit, electrically, the chemiosmotic potential along the mitochondria, and along a chain of mitochondria, so long as the protons could get through. But nothing else has to get through. So it doesn't really say that the mitochondria are fused in the sense that would allow travel of proteins. Another thing that's very important to bear in mind as a caveat on these studies is that people have looked in vitro at a number of different cell types and found, again, very fused and reticulated mitochondria, and in this case there's genetic evidence to suggest that quite a lot of fusion really is occurring. But, it's been very clearly demonstrated that fusion is induced -- wholly induced -- by oxidative stress. And of course all of these experiments have been done in 20% oxygen, which is not what mitochondria see in the body -- they see about 5% or less, probably a lot less actually, in muscle. So it's almost certain that any measure of the extent of real fusion in mitochondria that is seen in vitro is completely wrong to be extrapolated to the in vivo situation.

Julie Anderson: Are there any other tissues that this has been done in, besides muscle?

Aubrey de Grey: Muscle is the only one that's been looked at in that much detail.

-- coffee break --

Roger McCarter: Aubrey asked me to talk about muscle, and in particular aging in muscle. And I wanted to start off by not talking very quickly, because I've decided to stretch my 5-minute limit and go to 7 minutes, with the acquiescence of the chair! I thought I should start by saying what is sarcopenia, because I get this question regularly, just to clarify in your minds what we're talking about, and then the consequences, and then its reversal. When Aubrey asked me to participate in this meeting, and I really want to thank him for this invitation, what I thought he wanted me to do was to use my 25 years' experience of dietary restriction to discuss exactly the sort of thing that Bruce was talking about, how do you retard aging? Because I think we have some handle on that, on how to do it. But that's exactly what Aubrey did not want me to talk about, so this is where I would have started -- I wanted to show you how aging is indeed retarded by dietary restriction. Here is the mass of the leg muscle, the gastrocnemius in rats, as a function of age eating ad libitum, and you see here is a classic example of sarcopenia -- loss of muscle mass with age. And how this loss of muscle mass is retarded, or delayed in time, should you choose to eat 40% less. So we have a state of delaying aging, if you like, by dietary restriction. On the other hand we have to put up with a very much smaller muscle if you eat 40% less, and I think this is clearly an example of retarded aging, not reversed aging. So I took my mission very seriously and said, well, can we truly reverse aging in terms of taking the ad libitum-fed gastrocnemius, is there a possibility in fact that we could take a rat at this stage, or even at this stage, and do something which would build its muscle mass back to here. So that's the challenge that I took, and I hope that will be satisfactory from your point of view -- that that is a satisfactory definition, in terms of actually trying to reverse aging. And the more I got into the literature, the more it became obvious to me that there are such exciting things going on in this area that it is indeed, theoretically at least, possible to do that.

Now, I just want to point out one other thing. The definition of sarcopenia really relates to loss of muscle mass, that's how it is classically defined. But it's become very clear that that's not good enough. When people talk about sarcopenia today, they really mean not just restoring the mass of the muscle but the *quality* of the muscle. What they mean by the quality of the muscle is that muscles, as they get older, accumulate fat; they accumulate connective tissue. So you don't want just to get the muscle *mass* back up, you want to get rid of the fat, you want to get rid of the connective tissue. And one other thing which is extremely important: it's become clear from studies on mobility that muscle mass is not the only, or in fact even the most, important thing for enabling old people to get out of chairs and to move around. In fact the most important thing is the power that the muscle can exert. Power, just to get back to first-year physics, is force times speed. So in terms of getting muscle mass back, it's not sufficient to build up the mass which gives you enough strength or force. We have to get back the characteristics of the muscle, which enabled it to contract quickly. So the challenge that I took was, what devices are available to enable us to take a low-weight muscle and not only build up its mass, but to change the quality of the muscle in terms of restoring the power that the muscle had when it was way back here. And that's really what I would like to talk about. Now, let me get back to the question of the causes of this, and simply focus on this drawing here. The loss of mass of any particular muscle is the consequence of two things. There are a number of muscle fibers, and each individual muscle fiber has a certain diameter. And animals seem to be different from human beings in this respect. When we lose muscle mass, we lose numbers of muscle fibers (primarily). There is also a decrease in the diameter or the cross-sectional area, a loss of contractile protein, of many of the muscle fibers which are present. But the major effect in human beings, at least identified so far, and these are early days, because there are not really good data from different muscles -- much of the evidence that we have on sarcopenia comes from leg muscles, the vastus lateralis and triceps brachiae. But the evidence available today indicates that loss of muscle fiber number is the important problem. So we have a bit of a problem there, because it's very difficult to restore fiber number. What we can do, of course is to take the existing numbers which are present in old muscle and build up the diameter, to restore the appropriate amount of contractile material, and then change the character of that contractile material to increase the power of the muscle, so that the quality of the muscle is restored.

Gregory Stock: The loss of fibers, does that occur in a relatively linear fashion across the muscle?

Roger McCarter: Yes. Again, there are not very good measurements on this, but the supposition is that most of the loss is a consequence of the numbers, not the cross-sectional area. And the idea is that when you've lost 20% of your muscle mass you've in fact lost almost 20% of the numbers of muscle fibers.

Gregory Stock: So if your muscle mass decreases because of adverse environmental effects -- you're starved or whatever -- that that would be fiber loss?

Roger McCarter: I'm going to get into that, as a matter of fact. There is another thing that happens in human muscle, and it's not particularly well identified, there's a lot of controversy in this area, and that is that the one really good study done in this area shows that there's a preferential loss in what are called type 2 muscle fibers, and these are muscle fibers which are rapidly contracting, and they're also much more fatigable. And this is where the power story comes in, because if you're preferentially losing the quickly contracting ones, then of course you're losing power, so you wind up with what you've got left, which is the slowly contracting fibers. So then the question becomes, if you've got an old muscle, composed mainly of slowly contracting muscle fibers, what can you do to change the quality of those muscle fibers to make them rapidly contractible? Now, here's what makes muscle so exciting. No matter what the age of the muscle fiber, it is heavily plastic -- it's highly plastic. Muscle fibers respond to the demands put on them. There's very good evidence that no matter what the age of the muscle fiber, it will respond appropriately to the demand that is put upon it. So muscle fibers retain this plasticity into very old age, which makes them a very exciting tissue in terms of time, again, and retardation of aging.

We've done a lot of measurements in rats and mice, as they age. They get slower, they move around less, as they get older, and this is probably one of the causes of sarcopenia. However, there's a great deal of evidence that the major cause is the loss of nerve-muscle contact. With aging, there is a dropping-out of motor neurons from the spinal column, so there are fewer motor neurons, and a preferential loss of the rapidly-contracting motor neurons, the type 2 motor neurons. And the motor neurons which are left, in fact, increase in size. So a motor neuron in a young person may have 1000 muscle fibers attached to it, the same motor neuron in an older person may have 2000 muscle fibers attached to it, so there's an increase in size of the motor neuron, or motor unit, because there are fewer motor neurons in the muscle. Another popular mechanism for the loss of muscle mass is the gradual accumulation of sustained damage, the old wear-and-tear theory -- that muscle, as a consequence of time and usage, gradually accumulates damage which is not repaired, and that damage accumulates to the point where eventually an apoptotic cycle is initiated and the muscle fiber is destroyed. And finally there's one which I think Andrzej may address in a bit more detail: certainly the hormonal profile of the older person changes dramatically with age, so that the growth hormone level, which has a trophic influence on muscle, declines with age, and even more importantly, there is an 80% drop in the circulating levels of IGF-1, insulin-like growth factor 1, which has a major trophic effect on muscle. And we're going to get back to this in a moment, because that is where the reversal strategies come in.

Now let me just briefly mention the consequences of sarcopenia. They are severe. Probably one of the least desirable, which we can all relate to, is increase in fat mass. Your lean body mass goes down with lost muscle; your fat mass goes up, because you're still eating just as much but you're not moving around as much. So there's an increase in fat content, a decrease in lean body mass, about 50% over 50 years, and very importantly for all of us -- men and women -- the consequence of sarcopenia is a decrease in bone density. One of the causes of loss of bone density with age is decreased mobility associated with loss of muscle mass. From the system point of view, the cardiovascular system, when there's a decrease in mobility, is compromised; the pulmonary system is compromised; the immune system is up-regulated by exercise, if you exercise less the immune system goes down; from a whole body point of view, obviously your whole body weight declines because you've got less muscle mass. The BMI goes down; there is decreased glucose tolerance because you've got less muscle mass; a decreased thermoregulatory capability; and a decrease in maximum oxygen consumption. So there are severe consequences of sarcopenia, so reversing it is a very very important job.

So let me move on now to the reversal strategies, which is really where I'd like to spend my last three or four minutes. Can we do something about it? Indeed we can. This is what we need, close up. And I think it's pretty obvious when considering muscle, that there are three essential things that we need. Firstly, we need to be able to increase muscle protein synthesis. And I was astonished by the increases in protein synthesis which can be accomplished today. And I'm going to give you a strategy for doing that, but let me just say that in one of these techniques, which has been done very carefully by Goldspink, at University College Medical School in London, he's calculated that by doing this particular exercise, the muscle nuclei are induced to increase their rate of synthesis so that they are producing 90000 new molecules of myosin every minute. So muscle mass, as a consequence of this, is increased by 30-40% within the course of three days, as a result of doing this particular exercise. Now this is pretty darned dramatic, because sarcopenia results in a loss of muscle mass of about 30% over a period of 30-40 years, and this man can induce this in mice, at least, to the tune of 30-40% in the course of 3-4 *days* using a simple exercise strategy. So, we need to increase protein synthesis.

Now, here's something that's come up quite recently and is quite important. Increase in protein synthesis is not good enough. The reason is that in new muscle power, there is a limitation imposed by the number of cell nuclei. People have looked at the muscle fibers from well-trained athletes, and the number of nuclei per unit volume, the ratio of nucleus to cytoplasm, is fixed. It's as if each nucleus is capable of controlling a given volume of cytoplasm and that's it. So, you can synthesize all the protein you want, but once you exceed your critical ratio of nucleus to cytoplasm, it's no good. So it's very clear that you can't just increase protein synthesis, because you will slowly hit that wall, where the number of nuclei is simply not enough, so you have to import new nuclei from somewhere. And of course this is where the satellite cells, or myoblast stem cells, come in. Muscle fibers have, under the basal lamina but sitting outside the cell membrane, a population of undifferentiated cells called satellite cells or myoblasts. In response to inflammation, let's say we have an inflammatory insult in the muscle: here is a muscle fiber that's been damaged, and we have satellite cells sitting underneath the basal lamina but outside the muscle membrane, which are mobilized in response to stimuli. Now they're mobilized in response to exercise, particular types of exercise, as well as damage, and these satellite cells, under the direction of myogenin, a transcription factor, are activated and liberated, and then they differentiate in response to another transcription factor, and fill in the gap, and you have an intact muscle fiber again. Now you can see the critical nature of satellite cells -- you have to have new nuclei. So where do they come from? Previous results indicated that there was an age-related loss of satellite cells. More recent data suggest that that is not so, so I think we're up in the air at the moment as to whether satellite cell number declines or does not decline. But what *isn't* controversial is that the responsiveness of the satellite cells which *are* present declines as a function of age. So the responsiveness of the satellite cells to trophic hormones does decline. But this fellow Goldspink has taken satellite cells out of elderly animals and re-engineered them in vitro and then reinserted them. So you do away with the immune problem by using autologous myoblast transfer. And you take the cells out, grow them up in vitro, re-engineer them genetically and then put them back.

Aubrey de Grey: What sort of re-engineering?

Roger McCarter: Well, the re-engineering can take many aspects. He's actually proposing quite a revolutionary thing. He's proposing to use muscle as an artificial endocrine organ to, in fact, replace whatever deficiencies you have. And the re-engineering is such that the expression of a chemical in which. let's say, you are deficient, can be switched on and off using different drugs -- rapamycin, or different antibiotics. So it's a very exciting developments, where you can get around the immune problem by using autologous myoblast transfer to repair not only whole muscle fibers where it's been damaged, but take older muscles and introduce these myoblasts to do it.

Gregory Stock: How do they get these myoblasts back?

Roger McCarter: By simple injection. It's injected into the muscle. The amazing thing about these satellite cells, and again this is something that I was unaware of: you would think, being located under the basal lamina, that their possibility of movement would be very very limited. It turns out, people have tagged satellite cells and found that satellite cells from a neighboring muscle -- not just a neighboring muscle fiber but a neighboring muscle -- somehow, presumably via some chemotactic influence, will find the damaged area. So satellite cells are clearly mobile, and they can be induced to move around from one part of a tissue to another.

Judith Campisi: Helen Blau has done this with her cells, and has shown that you can get sustained hormone secretion from the muscle. Myoblasts are highly migratory.

Roger McCarter: Yes. I found this very exciting, because I was unaware of this stuff.

OK, so let's talk reversal strategies. The first reversal strategy is exercise, and I think this is the thing that your wife has been trying to get you to do for a long time, and all of us have a certain sense of guilt about, because we feel we should be doing more of it; what's the evidence? There are three types of exercise. Aerobic exercise, in fact, as we all know, the sort of typical running-type thing, does a lot for your cardiovascular system; it doesn't do a lot to your muscles. However, even for aerobic exercise, it's one strategy to increase muscle mass a small amount. The one that we've all heard about, which is so boring, is lifting weights. There's no question that that has a major trophic effect on the muscle. But how much? Remember, we're trying to build up 30-40% of muscle mass. Resistance-, strength-type exercise, even in compulsive "type A" individuals, will get you about 10% of your muscle back. Not really in the ballpark we want -- it's good, but it's not *that* good. So resistance exercise is good; now, here's where Goldspink and crowd come back in. What they have done is to pioneer the use of a different type of exercise. It's been known for a very long time that one of the economies of movement -- for instance you're in a tree and you're lowering yourself from the tree -- is that you have active muscles which are stretched while they're active. Our muscles work in pairs. So every time a muscle shortens under tension, its rate of shortening is controlled by another muscle which is active while it's being lengthened. It's been known by people who study muscle energetics, for a very long time, that active lengthening involves almost no energy cost. The rate of ATP hydrolysis is almost switched off in active lengthening, because the molecular motors are prevented from going through their power stroke, so that the hydrolysis of ATP does not occur. So that active lengthening involves almost no usage, or very little usage, of metabolic energy. But, it's also been known for a very long time that this also creates enormous amounts of damage. There's a huge free radical flux associated with active lengthening of skeletal muscles. What these people have found out is that in the process of active lengthening exercises, not only is there great free radical flux, but the genetic machinery is switched on to synthesize a new type of growth hormone. It is a splice variant of the usual IGF-1 which is circulating round the body from the liver. And this particular splice variant of IGF-1, which they have termed MGF, or mechano-growth factor, they have found that this particular variant of IGF is released from the muscle -- it's synthesized by the muscle and released from the muscle -- and it comes back to the muscle by influencing the growth of the muscle very dramatically. It is also picked up by the nerve fibers innervating that muscle and transported back to the central nervous system. So this is very exciting, because it provides a link which muscle physiologists have always known has to exist, but which wasn't clear: a trophic influence or feedback mechanism, from an active muscle, to the central nervous system. So MGF is very new, it's very exciting, and it offers great possibilities, because of course now that they've cloned it, they are in a position to synthesize this, or engineer vectors, which can then be engineered into the myoblasts to be able to produce great quantities of this IGF-1 variant. And this variant is the chemical which apparently stimulated muscle growth dramatically. I mentioned previously that this would increase muscle growth by 30-40% in three days.

Judith Campisi: This is the engineered, or the exercised muscle?

Roger McCarter: The exercise will do it, but even if you don't do the exercise, which creates a lot of damage, this re-engineered increased level will do the job, because this is what gets the job done in terms of stimulating protein synthesis.

So this is where this is. I think it's very exciting. And in terms of gene therapy there is another group, Lee Sweeney's group at the University of Pennsylvania, which is also doing some really excellent work. Instead of using myoblast transfer, Sweeney's group is using viral vectors to get re-engineered DNA into muscle cells, and they're in fact using recombinant adeno-associated virus, and having some pretty good results. The drawback to their particular technique is it increases muscle size by about 10-15%, not in the same ballpark as Goldspink and his 30-40%.

OK, so we've got a means of increasing muscle mass. Remember, that's not good enough. We have to not only increase muscle mass, we have to increase muscle power. We have to change the characteristics of the muscle fibers. It's very well accepted at this stage that exercise -- resistance-type exercise -- will slowly, slowly effect a change in the characteristics of the muscle. So that by varying the type of exercise that you have, when you now have your bigger muscles, you can then institute a program of regular exercise to change the character of the muscle from mainly slow myosin to up-regulate to a fast myosin.

And I would never have done this had Aubrey not forced this upon me, and in retrospect I was very pleased, because the more I read, the more I got excited, seeing that there is in fact a confluence of different areas here, coming together, where we may not satisfy Greg's definition of a reversal of aging, but at least from the muscle physiologist's point of view I think theoretically without question, the possibility exists so that a 70-year-old human being at least has the theoretical possibility of recovering the muscle composition that he or she had in their 20s or 30s by using this type of technology.

Aubrey de Grey: Thank you very much. I must say that it excites me very much as well, and I'm delighted that you found out all about this stuff.

OK, I have a couple of questions before anyone else gets a chance. You mentioned that the cross-sectional area of fibers is more easily increased than their number, in reversing muscle loss; is there a relationship between the fiber cross-sectional area and the unit cross-sectional area -- in other words, are big fibers fundamentally less efficient than small fibers?

Roger McCarter: No, I don't think so. At least in terms of the measurements, the normalization is all to cross-sectional area, and provided you normalize appropriately the numbers are consistent across a large number of different muscles.

Aubrey de Grey: OK. The second thing I wanted to ask is: something I had no idea until you mentioned, though I evidently should have known, because it seems not to be new stuff, is that the size of the motor unit, in other words the number of fibers attached to a given motor nerve, increases in fast muscle with age, while the number of actual fast muscle fibers is decreasing -- have I got that right?

Roger McCarter: No you didn't; let me restate it. The number of motor units decreases with age. But the size of each motor unit increases with age.

Aubrey de Grey: Right, that's what I was saying.

Roger McCarter: The fast and slow business is a little bit controversial, but I think there's no controversy about the dropping-out of fast or type 2 motor units. There's a lot of controversy as to the characteristics of the muscle fibers.

Aubrey de Grey: They might change into slow?

Roger McCarter: Exactly.

Aubrey de Grey: OK, so the question I really have, then is: what that seems to imply is that the rate of loss of fast motor neurons is the greatest change, percentage-wise - it's faster than everything else.

Roger McCarter: Yes.

Aubrey de Grey: Now, is there any sign of THAT being reversed by anything?

Roger McCarter: There is. In fact, somebody mentioned -- I think Judy you mentioned -- Helen Blau's work. One of Helen Blau's former postdocs, Charlotte Peterson, set up a really good group in Arkansas, and with Esther Dupont-Versteegden, has been doing a lot of really excellent work on spinal cord transection. And they have just put out a number of studies indicating that taking fetal spinal cord extracts and injecting into the spinal column of Sprague-Dawley rats who have had their spinal cord transected greatly increases muscle mass, in response to IGF-1 and in response to passive exercise

Aubrey de Grey: Helen Blau was one of those people?

Roger McCarter: No, Charlotte Peterson and Dupont.

Aubrey de Grey Thank you very much. My last question was about the ability of exercise of the right sort to increase muscle mass: of course, some muscles are easier to do that to than others, I guess. How much easier? How creative does one have to be?

Roger McCarter: I think that's a very good point, because obviously there are some muscles which are difficult to exercise. So the idea would be that you could presumably have little stimulators. Let's say you wanted to exercise your neck muscles: put little stimulators on and at least let them perform isometric contractions.

Aubrey de Grey: Give yourself cramp?

Roger McCarter: Right, you could be in real trouble. So the trouble with that is that the mechano-growth factor or IGF-1 variant that's put out by muscle is put out in response to work. In other words, it's not just enough to develop a force. You must have a stretching and you must have a change of length of the muscle.

Julie Anderson: It's got to be mobile?

Roger McCarter: Exactly. You can't just be developing a tension. And that's a bit of a shame, because each one of us could be doing isometrics as we sit here, and do ourselves some good, but apparently that's not going to result in the secretion of this autocrine IGF-1. So that there has to be a change of length of the muscle, preferably an increase of length of the muscle. So that the muscle is actually stretched while it's active. And that is a very very strong stimulus to the secretion of this growth factor. And I see in the future that in fact there will be whole exercise programs devoted to active lengthening of muscles, much more than has ever been done before. Physical therapists have been using this as a modality of treatment for years, under the name of pliometric contractions, and it's only now starting to be appreciated that this could be a major factor in changing the sarcopenia of old age.

Judith Campisi: What types of exercise do this?

Roger McCarter: Well, you've got to be stretching an active muscle. So one way is to lower yourself -- so, right, I'm hanging on and I lower myself, I'm stretching my muscles while they're active. Imagine yourself in a tree and you lower yourself to the ground from the tree.

Judith Campisi: That'll do your arms. If I'm an old lady and I'm losing leg muscle?

Roger McCarter: Yes, that's a different problem, because the character of that muscle is a little bit different. So I'm not sure how to address that one, but that's just because I haven't thought about it -- the smart muscle physiologists will find an exercise, I know, for that. In terms of your legs, it's been known since 1850 that it was much more fatiguing, but not much more energy-consuming, to run downhill than to run uphill. You use a lot more energy running uphill, but the movement is much more fatiguing to go downhill.

Judith Campisi: I like it -- running downhill -- that's my kind of exercise!

Roger McCarter: Well, you know the interesting thing is that it's not tied into the energy consumption. In fact you're switching off the energy consumption. But let me just add that one of the interesting experiments that was done on medical students in the time of the 1850s, when you could do anything on medical students, was to get them to run up the steps (and this was done with French medical students), run upstairs, and stop at the top of the steps, and then go backwards, at the same rate, down the steps, and their oxygen consumption was measured at the bottom of the steps and the top of the steps. And this was the first indication that muscle metabolism really switched off in these lengthening active contractions, because obviously it consumes far less oxygen when going backwards. But people felt very fatigued at the bottom.

HA: I think the decrease in the number of fibers should be synchronized with the number of mitochondria -- we were discussing the network of mitochondria. Treatment by undifferentiated myoblasts with presumably healthy mitochondria may be part of the mode of action, because if the mitochondria in the fiber are damaged, maybe they are signaling somehow that these fibers are damaged. Is this possible?

Bruce Ames: I want to introduce Hani Atamna, a postdoc in the lab.

Roger McCarter: Hani, I think this is an area which is ripe for experimentation, because I have never seen anything addressing the point that you raise. The fibers which are created as a consequence of this type of activity are Type 1 muscle fibers, which are slow, oxidative, and have a high mitochondrial population. But the origin of the mitochondria or the rate of increase in the mitochondria in response to this type of exercise, as far as I know nobody has even looked at it. So it would be something very interesting to study.

PT: How are the beneficial effects of MGF dissected? Because we have also IGF and growth hormone in there,

Roger McCarter: That's an excellent point. These have been done mainly in acute studies. There is one study in which at least the expression was monitored over a period of eight months. So the possibility for long-term monitoring is there, and in fact in this regard Sweeney's group recently had a very exciting paper in which they have developed a method of non-invasive monitoring using magnetic resonance spectroscopy. So it would be possible in vivo, using this particular reporter gene that they've developed, to monitor the expression of the gene over a long period of time. So all of the little bits are coming into place now, to do this in a very useful way.

PD: Is anything known about the chemotactic agent which drives the satellite cells to begin to proliferate? It seems like that would be a great modality, and not as invasive as taking out satellite cells, growing them and putting them back.

Roger McCarter: Well of course you do want to engineer them. And you do want to expand them, too, because the satellite cells which you're going to extract from the old muscle are few in number, possibly, and unresponsive -- they're lost a great deal of their responsivity. So using the engineering techniques you could re-engineer the satellite cells to be much more responsive, and then you put them back in greater numbers. And I think this is part of the beauty of this technique.

Andrzej Bartke: In terms of the recruitment of these cells and the repopulating of the muscle which then might improve the size: If I'm not mistaken, body-builders believe that you need to damage the muscle -- they call it "tearing" -- by very severe exercise, and then you get more increase in size. So that would be, I think, already an applied aspect.

Roger McCarter: Absolutely. And in fact it goes along with what we know about muscle, that you need to have heavy loads under controlled changes of length, and that induces this autocrine factor. And in fact it's a paracrine factor, in the sense that the chemical not only goes back on the same muscle, but it goes to other tissues, and it's taken up by the nerves, for instance.

Aubrey de Grey: I hate to bring this discussion to a close, but we are beginning to run rather late; I'm sure that we'll be able to continue talking about these things over lunch.

Judith Campisi: I have a question that maybe you can address. I saw a stem cell study in which somebody made the claim that mesenchymal stem cells can be modified to differentiate into myoblasts under exercise. Have you see this?

Roger McCarter: No, I hadn't heard that.

Judith Campisi: So this was presented as an explanation for the apparent controversy as to whether old muscles really lack myoblasts or not, and the issue is how much exercise the individual has done -- under exercise you could recruit stem cells to then differentiate into myoblasts. It makes me wonder if maybe that's a positive feedback.

Roger McCarter: Yes. And in fact Goldspink's group has addressed that. Because they have found a feedback between bone, for instance -- there is the same MGF, thought to be a paracrine, endocrine factor in bone. So that provides the feedback; again, you know there's been a lot of controversy as to what the feedback is between muscle and bone. It becomes very clear that this MGF is a major candidate. And at least some of the additional studies support that.

Aubrey de Grey: OK, last question, how do you spell "Goldspink"?

Roger McCarter: Goldspink.

Judith Campisi: Where is he?

Roger McCarter: He's at the University College Medical School in London. And I have several of his reprints here if you'd like to take them.

Aubrey de Grey: OK -- Julie.

Julie Anderson: So Aubrey has asked me to cover neurodegenerative diseases of aging, which are basically where you have the loss of lots of cells in a particular brain area, and depending on the area of the brain that has the cell loss, that gives you the symptomology you see. We see in Parkinson's disease loss in the midbrain, loss of voluntary muscle movement, for example. There are actually several good genetic models for most of these diseases -- they're not perfect models, we can talk about that. People are using transgenics to create models of these diseases, and then use these to look at various treatments for reversing the disease. There are pharmacological treatments: metal chelators; antioxidants; different enzyme inhibitors that may inhibit a basic process that seems to underlie all of these diseases, such as caspase induction; more may be particular enzymatic reactions that may be specific to a particular disease, be it Alzheimer's disease or Parkinson's disease, so there are a lot of different pharmacological treatments; dietary types of treatments are being tried; also reversal of the disease by repopulation of the cells that are lost in the neurodegenerative disease. Aubrey had also suggested a discussion about the reversal of these diseases by destruction -- in each of these diseases you see an extracellular or an intracellular deposit or aggregation of proteins, and a lot of the therapy in these various diseases is towards removal of this aggregation in these animal models. However, whether those aggregates are actually causal or are just a tombstone of cell loss, or actually may be beneficial, is a big area of controversy in the field, so I want to talk a little bit about that. We've seen publications on these various controversies, that loss of the aggregates may be something that would be problematic in terms of reversal of the disease.

So in terms of these various neurodegenerative diseases, in the past 5, 10, 15 years, many of the genes responsible for these diseases have been identified -- genes for Huntington's disease, the early-onset forms of Alzheimer's disease, Parkinson's disease; for late-onset Alzheimer's disease, presenilin-2; apoE4 is a risk factor for age of onset for Alzheimer's disease; IDE is a recent one that's come out of the Tanzi lab; and in terms of early-onset forms of Parkinson's, parkin, which has recently been indicated to be a deubiquitinating enzyme, again addressing this involvement of aggregates in disease; alpha-synuclein, which is found in some early-onset forms of Parkinson's disease, and just on the aggregates, Lewy bodies are in Parkinson's disease. In terms of ALS there are also familial forms of ALS, accounting for about 10% of the familial forms of amyotrophic lateral sclerosis, and about 20% of those familial forms are due to a mutation in the copper-zinc superoxide dismutase. Again, another commonality amongst all of these diseases is the involvement of oxidative stress and mitochondrial dysfunction, so in terms of thinking of therapy for these diseases one might want to address them, or reversing the sort of damage that you get due to this oxidative stress and mitochondrial dysfunction. So because of the availability of these different genes in these familial forms, which seem to have the same symptomology although earlier on, that these have been used to make various animal models of these diseases. As I pointed out, these are generally incomplete models: for example, in the APP model of Alzheimer's disease, these animals do show the beta-amyloid plaques that you see in Alzheimer's disease; you do see tau phosphorylation like you see in Alzheimer's disease; however you don't see the formation of neuritic tangles of the disease; cell loss is more variable in the APP transgenic; it's not really a true, complete model of Alzheimer's disease, but it's sort of the best thing that we have at the moment, and at least in terms of addressing that portion of the disease they are useful and helpful. So these models have allowed us to examine disease progression, as well as looking at preclinical testing of novel therapeutics. I just want to make a point that in terms of current therapies in neurodegenerative diseases, most of these involve slowing the disease progression after the onset of symptoms -- someone has symptoms and you try to slow the disease progression at that point. Most of the therapies, unfortunately, are things that can delay the symptoms but not the course of the disease. In terms of Parkinson's disease the main therapy that's used clinically is L-dopa; this is so that you get conversion of L-dopa to dopamine in the remaining cells in the Parkinsonian patients, making those cells (if you will) work harder; but however it doesn't do anything to prevent the cell loss that you see in the disease. So many of the therapies that are currently being used are going to affect symptomology but not the actual etiology of the disease. So really what we want to develop are treatments that are going to delay or stop even the initiation of disease progression. And so could some of these models be used in combination with some prescreening techniques? There are various prescreening techniques that are getting better and better in terms of being able to look at individuals and assess, before symptoms are seen, the problems of the disease, which could be used to prescribe therapeutics before the development of symptoms. The primary goal of developing these therapeutics would be therapies that have a long-term effect and limited side-effects, which I think is true in any of these diseases.

So one of the big areas of excitement recently, in at least Alzheimer's disease, has been immunotherapy, this is work out of Florida, where they've immunized some of the APP mice with synthetic 1-42 Abeta, and they found that that reduced plaque formation in these animals and found that it decreased the memory loss that we see (by Morris water maze). The problem is that in making antibodies against this Abeta form, are you then affecting the normal function of the protein? In general with a lot of these therapies you have to think about this, that there probably is some normal function of the protein, and that the therapy that you're using is going to affect the normal function of the protein. If you actually look at the data from this kind of work, the normal mice that are being immunized actually seem to do slightly worse than the normal mice that aren't being immunized. So it's something to be aware of. Another broad sort of immunotherapy is the use of non-steroidal anti-inflammatory agents: inflammation also seems to be a major contributor to all of these diseases, via infiltration of microglia and macrophages producing, let's say, oxidative stress, superoxide and so on, that can actually exacerbate the disease.

Another big area of excitement in the field is the use of various metal chelators. In terms of these diseases there does seem to be metal deposition. Again you have to be careful in terms of the concept that metals are important for biological functions, so there's probably some dosage issues in terms of using this kind of therapy. I was at a meeting in Padua recently where Ashley Bush from Harvard presented some data on using some copper chelators in one of these APP models of Alzheimer's disease, and showed that the use of these copper chelators that were being added to the drinking water actually prevented the plaque formation that you see in these APP transgenics, and it improved the memory function in terms of the water maze. There are some early studies done by Gurney's lab, in some of the ALS transgenic models, feeding them copper chelators, and they found that in those models it seems to delay disease onset and extend survival in these animals.

In terms of this issue of copper chelators, both copper and zinc cause aggregation of Abeta. Zinc is oxidant-neutral; copper actually, the transition between copper(I) and copper(II) can produce oxidative stress, and there are some indications that the ratio of copper to zinc is important in terms of Alzheimer's disease, so if you're able to chelate up copper or affect this ratio of copper to zinc, that may be important in terms of slowing it. Abeta seems to act as an oxidant, so if you get Abeta aggregation you get hydrogen peroxide production. And there's some indication, again from Ashley's lab, and Rudi Tanzi's lab, that perhaps the normal function of APP is to act as a SOD-like molecule. And so then this ratio would be important, and if it's out of whack, that it may be acting as a pro-oxidant. We have some data from our lab looking at transgenic expression of an iron chelator, in an MPTP model of Parkinson's disease, that suggests that -- I mean, it's known in Parkinson's disease that in these patients if you look at the basal ganglia of a Parkinsonian patient versus an age-matched control there's an increase in levels of free iron in Parkinsonian patients, so a lot of people talked for a long time about the use of iron chelators as a therapy for Parkinson's disease. We now have this transgenic model and we were able to show that expressing an iron chelator actually considerably delays the neuronal cell loss that you see in this MPTP model of Parkinson's disease. So we've expressed the iron-binding protein ferritin in the substantia nigra of these animals, and there are several ferritin-type chelators being developed, which affect the ratio of iron(II) to iron(III). Again with the idea that of course iron is important in terms of neuronal function, and so probably dosage issues may be important.

HA: When you measure the increased iron, at advanced ages in the disease, it might be that this is present earlier but you don't see it.

Julie Anderson: We're looking at autopsy samples, and yes, it's probably more common in the advanced stages and a contributor. Another aspect of work in our lab is looking at the early loss of glutathione, because actually a thing you see in Parkinson's disease is this loss of glutathione levels, and we have an indication in a paper that just came out in JBC, that if you decrease glutathione you're affecting the function of mitochondrial Complex I, which seems to be through thiol oxidation, and again in terms of oxidative stress in all of these diseases, a lot of these particular susceptibilities of certain proteins may be due to thiol oxidations. That may be something that we want to go into some more.

In terms of antioxidant therapies, a lot of the early studies have been done were with vitamin E. There were some studies done, again in the ALS transgenic mice, showing that vitamin E or selenium, which increases glutathione peroxidase function, delayed the symptoms, but not an increase in survival. And vitamin E has been used in clinical studies, both in Parkinson's disease and in Alzheimer's disease. And it's really shown relatively little efficacy in terms of preventing the progression of the disease. The problem is vitamin E is not catalytic, so maybe we want to be looking at compounds which are small molecular weight compounds that can easily cross the blood-brain barrier, that are going to be catalytic. The EUK compounds are compounds that were mentioned previously that act as SOD/catalase mimetics, that are catalytic, that are low molecular weight; they're been used in tissue cultures with Abeta and shown to prevent the toxicity of the disease. They've also been shown in not very thorough studies done by Eukarion, that they seem to prevent the degenerative cell loss seen in the MPTP model. So we may want to be exploring some of these. Also there's NOS inhibition. Nitric oxide is another oxidant that seems to be involved in the etiology of these diseases, and some people have suggested the use of NOS inhibitors in these transgenics, which might be protective in the MPTP model of Parkinson's disease. They inject animals with an inhibitor of Complex II, you see a very similar type of etiology of the disease as you see in Huntington's disease, loss of striatal neurons, and this suggests that there may be involvement of Complex II in that disease. This is a very common toxic model of HD. CoQ is an electron donor and acceptor of Complex II and III, and Beal has done a lot of work looking at these various animal models and found that Q seems to be protective against SOD/ALS transgenics, the mouse model of HD, and the MPTP model of Parkinson's disease. They've done some small clinical studies that suggest that overall, CoQ seems to delay the course of the disease. And I put this under antioxidants though of course it's mitochondrial function: creatine, which serves to buffer ATP stores as phosphocreatine, and this appears to be protective in the transgenic mouse? model of HD. Creatine kinase is oxidatively sensitive; it's been shown to be inactivated by beta-amyloid, and actually it's found to be oxidized in Alzheimer's brain. So again this is one of these enzymes that's a thiol enzyme, and it seems to be the oxidation of the thiols that is important. So again, maybe we want to be talking about glutathione or glutathione-like compounds. And as we look, many of the mitochondrial enzymes that seem to be affected in this disease are enzymes that are oxidatively sensitive.

In terms of treatment with enzyme inhibitors, there seems to be in all of these diseases a problem of apoptosis and caspases. We did some studies crossing ALS transgenics to a dominant negative form of caspase 1, and this seems to slow the disease progression. The problem is that in ALS you don't really see the morphological signs of apoptosis, so it's not clear. Caspases get touted a lot as being involved in apoptosis, but they're also involved in inflammatory events, so this may be an anti-inflammatory mechanism. But in terms of all these other neurodegenerative diseases, caspases have been shown to be involved in APP processing in Alzheimer's disease, involved in Parkinson's disease -- there was a study that showed that caspase 3 is present in Parkinsonian brain, and we have a transgenic model where we've expressed a caspase inhibitor in neurons and shown that it's resistant to MPTP toxicity, and we're actually going back into the Parkinsonian brain and delineating the pathway for caspase involvement, so that we can be sure that what we see in the animal model we actually see Parkinson's disease itself. Caspases also seem to be involved in processing of huntingtin protein, in cleavage of the huntingtin protein.

Of course in terms of Alzheimer's disease there's a lot of interest in terms of finding inhibitors that will inhibit the Abeta secretases, to prevent accumulation of Abeta. It's been pointed out that actually presenilins seem to be involved in early development at the stage where cells are making decisions between going neuronal or becoming epidermal, which involves Notch/Delta local inhibition; part of the action of the Notch receptor involves cleavage of Notch by presenilin, so if you give a presenilin inhibitor it's going to affect Notch; of course if you're talking about giving treatment in Alzheimer's disease, which is very late in life as opposed to something that happens early in life, but again there's the idea that this protein has a normal function, and you want to develop therapy that's ideally not affecting the normal function. In Parkinson's disease, MAO-B inhibitors, the deprenyl studies done suggest that deprenyl as an MAO-B inhibitor can delay the course of the disease. It's somewhat controversial as to whether this is at the symptomatic effects or actually the cell loss of the disease. Also, deprenyl seems to act as an antioxidant, so it's not clear that the effect is really inhibition of MAO-B as opposed to its antioxidant effects. However there were studies recently done by Shih using knockouts which assayed the MPTP toxicity of these, so there does seem to be an involvement of MAO in the progression of the disease.

In terms of cell repopulation, of course in all of these diseases you're losing a neuronal subtype.

Judith Campisi: My understanding is that that's not the case -- that the cell bodies are there but the synapses have failed. Unlike Parkinson's.

Julie Anderson: Actually it's somewhat controversial, but we actually do have a substantial loss of cell bodies. In some of the models, what we're seeing *is* just a loss of synapses, but at the end stages there is actual loss of cells. In Parkinson's as well, the first thing that's affected is the synaptic loss. Which, in a way, if you think about the functionality of neurons, it's not the cell body, it's more the axon, and a lot of the action end of the cell is the axon connections.

Judith Campisi: Has anyone tried to get synapses to regrow? It seems much less draconian than dumping cells into your brain.

Julie Anderson: So just looking at replacing the synapses? I think a lot of the other types of therapy we're talking about are therapies towards before you even see degeneration of synapses -- trying to slow the actual processes that are causing cell damage. Of course the most dramatic thing you see in terms of loss of cell function is the actual loss of the cell. So a lot of these therapies are looking towards effects like oxidative stress; even in terms of apoptosis, looking at inhibitors of caspases, which a lot of people are doing, it's such a downstream event that if you're inhibiting caspases then the cell's so far gone that you're going to die anyway. But some studies of caspase inhibitors have shown that they maintain functionality of cells. But yeah, my bias is that we want to look at and try to prevent the events that are actually causing the damage, rather than cell replacement. But that's what Aubrey asked me to talk about! So in terms of repopulation of cells, there has been some success in the use of human fetal tissue in replacing the cells that have been lost. The problem is to obtain sufficient tissue. That might be overcome by isolating a combination of human neuronal stem cell populations, but the problem is that they apparently undergo very rapid replicative senescence after only seven or eight population doublings.

Gregory Stock: These are very specific diseases. If you look at just the general loss of the level of neuronal processing that's going on in aging, what's going on there?

Julie Anderson: Well, I was asked specifically to address neurodegenerative diseases, but I can address normal aging in this context, to say that I think that a lot of these processes that occur in neurodegenerative diseases -- be it mitochondrial dysfunction, increase in reactive oxygen species, which may be accelerated in particular brain regions in these various diseases -- are processes that happen in normal aging as well, and have been postulated to play a role in neuronal cell loss, or not even neuronal cell loss but loss of neuronal function in normal aging. Only some of the specific therapies that I'm talking about are specific to specific diseases; therapies like use of antioxidants, use of apoptosis inhibitors -- apoptosis is something else that occurs during normal aging and in normal cells. Increasing mitochondrial function through various dietary agents -- those are all things that may have an effect in normal aging as well as in these diseases.

Neurons are postmitotic; you're not going to replace them after cell loss barring these types of therapies, so there is this rewiring, and so in terms of regaining function you might also want to think about different types of factors that are aiding in terms of rewiring the system around the cell loss -- this cell's making a connection to this neuron, and you lose this neuron, this neuron rewires and makes a connection with another -- what happens to that process, which is a process that occurs in early neuronal development, and there certainly are a number of people who are interested in how that can be recapitulated with aging, how one can reverse age-related decline by rewiring the system, as well as preventing cell loss or cell repopulation.

Also, what you want is the cells to differentiate into particular cell types, so in terms of making these types of cultures in Parkinson's disease, it looks as if when they continue to culture these cells, what they really wanted to replace were the dopaminergic cells, and these cells seemed to be losing a lot of dopaminergic characteristics, so are there ways in which you could maintain the particular neuronal subtype that you wanted to replace, such as for example engineering, forcing the cells to make a particular dopaminergic transcription factor so that you're not losing TH cells. There have been, also, studies using heterologous cells which have been used in Parkinson's disease; the most promising seems to be carotid body autotransplant, where they've transplanted cells very high in dopamine content into MPTP-treated monkeys. There were only a couple of monkeys in the study but it seemed to be effective in terms of MPTP tolerance of these animals.

And I just wanted to end by talking about this issue of removal of extra- or intracellular aggregates. The question as to whether these aggregates are causal as opposed to protective remains; there was a paper by Hayden where they actually created a Huntington's model by putting the whole Huntington's gene, the human Huntington's gene mutation into mice in YACs, and this actually shows the correct phenotype and the same striatal neurodegeneration that you see in HD, along with these nuclear inclusions; however they found that the nuclear inclusions weren't important -- only the transport of this N-terminal fragment of the HD protein from the cytosol to the nucleus was required. So in fact there's a big controversy in the field now as to whether these aggregates are protective. I think when you're looking at neurodegenerative disease the easiest thing to see is the pathology, where you've got this increase in aggregates, but we don't know -- maybe you have the aggregate formation to take out the toxic molecule. For example in Alzheimer's disease, the fibrillar Abeta plaque deposition doesn't correlate with the disease -- the *soluble* Abeta does. There's some suggestion that maybe you have this aberrant fragment that acts as an oxidant, producing hydrogen peroxide, and if you're aggregating it and taking it out, it's not able to perform that function. So this is something that we have to be careful of in terms of designing therapies. In terms of Parkinson's disease, it may indeed be that the aggregates are something that relate to the pathology -- if you express the neuronal PD mutant form of alpha-synuclein in flies or in mice, you get Lewy body formation and also loss of neuronal function. However, again I'll point out that these models are not perfect models, and you don't see the loss of dopaminergic cells. You do see degeneration of the dopaminergic synapses that you see early on in PD, but you don't see this 80-90% loss of dopaminergic neurons that you see in Parkinson's disease. Again, these are not perfect models.

Aubrey de Grey: Thanks very much. OK, I have a couple of questions to kick things off. First of all, really following up on Greg's question, you've covered in detail a number of the neurodegenerative diseases, and you pointed out the various distinctions between them, really defined by their different names and so on, but of course they do have commonalities. And you made the point that these commonalities may be things that accumulate at a slower rate in normal aging, in people who don't die of these things. What can be said at this point about the presence -- well, OK, let's start with aggregates: what can be said about the presence of these things in people who die at very old age but not ostensibly of any neurodegenerative cause? Are they seen?

Julie Anderson: I know, for example, when we've looked at Alzheimer's patients versus age-matched controls -- and this is part of the argument against amyloid deposition as being actually causing the disease -- is that you have people who have no sign of the disease but have much higher Abeta plaque.

Aubrey de Grey: And is that true of the tangles as well as for plaques?

Julie Anderson: I'm not sure. As I said, I think it's because just visually you're looking mainly at the plaques.

Aubrey de Grey: The other thing I wanted to ask, by way of getting a handle on hypotheses for what might be causing and what might not be, is coming back to what you said about Huntington's disease having these inclusions in the nucleus. Now I would say that something can be predicted to be very likely to be causal if clearly the cell is trying not to let it do anything. So in particular if it's lysosomal, which I understand Lewy bodies are -- is that right?

Julie Anderson: They're cytoplasmic -- again there's even some debate in the field as to whether these aggregates are being primarily causal because you see ubiquitin in almost all of these aggregates, be it Abeta, plaques, Lewy bodies -- people have done histological studies and see ubiquitin there. So the commonality amongst all of these diseases in the end is that you're getting these large protein aggregations and somehow the ubiquitin system is messed up and it's not able to take care of them and you get a build-up of aggregates in the cell. But it's not even clear, because people have done histological studies, as opposed to looking to see whether that ubiquitin is free ubiquitin or bound -- I think only in Alzheimer's disease have they actually looked to see whether you see ubiquitin in proteins as opposed to free. Part of the work that we're doing in Parkinson's disease is it's been shown that if you deplete thiol levels in tissue culture paradigms, because the E1 and E2 ligase enzymes that are involved in transferring ubiquitin to proteins are thiol-sensitive, you actually prevent the binding of ubiquitin to damaged proteins by lowering glutathione levels.

Aubrey de Grey: OK, would it be a fair summary to say that really in all of these diseases it's very unclear whether the aggregates (whether they're intracellular or extracellular) are causal or are not, but, by contrast, it *is* clear in all cases that cell loss is the fundamental cause of the loss of function? Well, synapse loss but also subsequent cell loss?

Judith Campisi: I got Fred Gage to talk about these things, and he's quite adamant that except for the very very late stages of Alzheimer's, when the patient practically can't breathe, there is no loss, or minimal loss, of cells. So there's merit in putting growth factors back in the brain, getting them to grow better synapses from the cell body.

Bruce Ames: There might be loss of mitochondria.

Julie Anderson: The issue is cell function.

PT: When you speak of cells you must be very careful, because neurons might be lost, but the glial cells increased, so in terms of the total number of cells you have more.

Judith Campisi: You're right; he's concerned about inflammation, but even with neurons, he claims that the neurons are there but they're dysfunctional because their synapses are gone.

Julie Anderson: So maybe we don't want to go to, like, cell repopulation in AD, but it is cell function that declines.

Aubrey de Grey: Judy, can I ask you to elaborate on this? From what you do remember about what he said, what success did he have in regeneration of synapses?

Judith Campisi: Very dramatic effects but very local. He really has done it in one region of the brain.

Julie Anderson: There have been clinical trials done with NGF, and they haven't been very promising, but it may be they didn't go to the right place.

Bruce Ames: How do they do it?

Judith Campisi: Oh, they go in with a needle, deliver it to a local area, and they mark that area and look at the synaptic outgrowth of those cell bodies.

Christopher Heward: So they basically infuse NGF in very locally?

Judith Campisi: Yes.

Julie Anderson: This has been a very big area of research in the last 10-15 years.

Gregory Stock: So given what you're saying, this idea of infusing in the cell bodies seems to be ...

Julie Anderson: Dependent on the disease.

Judith Campisi: Exactly. For Parkinson's in humans, you do need to replace them.

Julie Anderson: And there has been some success in Parkinson's disease. There you're losing 80 or 90%.

PT: I think first of all we have to remember that we have a rather huge redundancy of cell number. So even if we lose many of them that should not be the cause of the decline in function. The loss of neurons alone, even if it is severe, will not be sufficient to explain the loss.

Julie Anderson: People talk about it as if, so, you've got some minimal number of neurons, say, in the substantia nigra, that you need for function, and in a normal individual you have loss but you don't fall below this biological threshold, but in PD, with an accelerated loss, you do. And again, cell loss is the end state, the last thing that you look at, and maybe we should be more concerned with the sublethal damage that is happening where the cell's still there but it's sick and not functioning.

Aubrey de Grey: I think I would like to comment on this, actually: it seems to me that is another case where loss of connectivity may be more important. I don't know whether we can say that there's anything like so much redundancy in the connectivity of the synaptic connections between cells as there is in the actual number of cells. Maybe, I mean, even in PD you know they go down to say 15% of young people, whereas in normal substantia nigra it's like 80% of young people. Maybe the threshold is much higher in terms of connectivity.

PT: It is important also to remember, at least from the work of McCabe (?), who has done lots of work on them, that if you grow neuronal stem cells, only 2% become neurons. All the others become glial cells. So even if you make large amounts of them, the production of neurons is very limited -- extremely limited. On the other hand, just producing more glial cells might be a good idea, because glial cells might rearrange or reorganize the microenvironment of the neurons and make them divide better.

Julie Anderson: And they're also stated to be a sink for a lot of the things that are causing damage, like oxidative stress.

PT: So that would be an indirect benefit. But just the replacement of the lost neurons?

Julie Anderson: I'm not a big proponent of that.

Aubrey de Grey: OK, I think we're going to have to go to lunch.

---

Andrzej Bartke: ... models for identifying possible means of intervention into the aging process. And whether we consider these potential interventions as delaying or reversing or both, my personal bias is that the most effective therapeutic interventions would prove to be prevention rather than treatment. That's just my personal opinion. So the two models that I want to talk about are the genetic dwarf mouse and the growth hormone receptor knockout. And I will say very little about the dwarf mouse, this animal shown here, because I think some of you may have heard this story already: these are mutants at the Prop1 (prophet of Pit1) locus; they have congenital problems with development of the anterior pituitary, and consequently they have congenital deficiency of three of the pituitary hormones -- growth hormone, prolactin and TSH. And there are two mutants with essentially identical phenotype, they are both spontaneous mutants, one at the Prop1 locus, one at the Pit1 locus, and the only piece of data that I want to show for these mutants is the survival plot. This particular survival is from the data we obtained in the prop1 mutant, which was recalculated and plotted by Rich Miller, and here is data for the Snell dwarf, the other mutant, that's on a different genetic background of animals which have somewhat longer lifespan and also more uniform genetic composition while those are not, and here is a graph of the Snell dwarfs, looks like a Xerox copy of this one. So the very obvious characteristic of these mutants is that they live much longer than normal, and it seems to be a true delay of aging, an increase of both average and maximum lifespan. Also we looked at various parameters, the most interesting probably are the learning and memory, and they also support the notion that this is truly delayed aging.

Now the other animal that I wanted to mention is the growth hormone receptor knockout, the animal produced by John Kopchick at the University of Ohio, and these animals do not express growth hormone receptor, the growth hormone-binding protein gene, and this was shown by RT-PCR, by Western blot for the protein. They do not have a total absence of growth hormone binding. It's somewhat unclear whether this is evidence for some other yet unknown growth hormone receptor, or perhaps it's an assay problem. They do not have the serum growth hormone-binding protein; depending on the assay they have either very low or no detectable IGF-1 in blood, that's a very important feature of these animals. They have elevated growth hormone because of absence of negative feedback, and they are somewhat smaller, roughly normal, at birth, but they are only about half of normal size adults. And these are data from the study of John Kopchick published in the July issue of the Journal of Endocrinology, showing average lifespan of males and females; the data are based on a rather limited number of animals but the difference compared to normals in the homozygotes or heterozygotes versus normal is highly significant. I think there's no doubt that these animals have a major extension of lifespan. We are now doing a study on a large number of animals and it's already obvious that we'll have a big difference.

Now, some of the physiological characteristics of these creatures. Here's some data on learning and memory: this is the maze test, and the knockouts do significantly better than normals of the same age. We also looked at the different age groups and the normals show age-related decline in learning and memory and the knockouts have much smaller or no significant change, at least over the period we've done it. This particular slide shows animals of the same age, and here they were tested at several intervals and the knockouts in each case were numerically better, and after three weeks significantly better than normals.

Another characteristic of these animals is extremely low insulin. With two different commercial assay systems for insulin, in fact, you cannot get a detectable reading from either fasted or fed knockout mouse. And with these extremely low insulin levels, they maintain also significantly reduced glucose. Here are data collected in the morning and in the afternoon, from animals that have constant access to food, and at both intervals the knockouts were significantly lower. This is the growth hormone receptor knockout, the glucose was lower than normal. We were very excited about the availability of these animals because we thought that we would be able to separate the effects of growth hormone deficiency from the effects of hypothyroidism, which is a prominent feature of genetically dwarf mice, so we wanted to demonstrate what we thought would be very obvious, that these animals have normal thyroid function. But they don't. They have a mild but significant reduction in the levels of both thyroxine and triiodothyronine, which presumably is the secondary consequence of low IGF-1.

We also looked at the body core temperature, and it was sort of marginally reduced, on the average about 0.4 degrees Centigrade, significant at some times of the day and not at others. We were also interested in levels of glucocorticoids, because of the evidence from studies of caloric restriction that this perhaps is the most constant and physiologically important feature of calorically restricted animals is elevated glucocorticoids. So we half expected to find the same thing here and we did not, by and large. This shows normal animals of both sexes sampled in the morning and in the afternoon, and we used a method where we thought they really had no chance to have the stress response before we collected the blood, and in each instance, normals and knockouts of both sexes, we saw the expected increase in the afternoon versus the morning, but the only difference that we found between normals and knockouts was for the afternoon levels in males. So out of the four comparisons, knockouts were high only in one. So clearly if there is elevation of glucocorticoids then it's not major.

Roger McCarter: Just a question on that: have you done longevity, male versus female?

Andrzej Bartke: Yes -- they were improved, both of them.

Roger McCarter: But there was no difference between male and female?

Andrzej Bartke: There is not enough data yet.

Roger McCarter: It really addresses the importance of the glucocorticoids in this effect.

Judith Campisi: Actually, I thought from the slide you showed that females actually were slightly lowered?

Andrzej Bartke: I believe so. But, you know, there were seven and a dozen animals. Also it depends a little bit if you compare them to the homozygous normals, or to all normals combined, so I think it's a bit early to say.

So what I was trying to do here was to compare the caloric restriction with the Ames dwarf and with the growth hormone receptor knockout, and I think that these comparisons allow at least some tentative conclusions. Obviously there are lots of similarities. All of these animals are smaller than the corresponding normals. All of them have very low IGF-1. (Obviously the two are connected.) All of them have low insulin; all of them have low glucose. Presumably all of them have some reduction in body fat -- there are no data for knockouts yet, we hope to have the data very soon, initial data recently obtained using a DEXA machine by Hyman using the animals we sent him. Dwarfs have a small but significant reduction in body fat when they were middle-aged or older. Young ones did not have this but the middle-aged did. Somewhat to our surprise -- we always assumed somehow that there was no difference in body fat.

Roger McCarter: There's one that you don't have there which is the temperature. Because in the mice, calorically restricted --

Andrzej Bartke: Next slide!

And there are also some differences: food intake in the calorically restricted animals is of course reduced by definition at the beginning of the experiment. Once their weight stabilizes they have to eat about the same per gram body weight as normals. The dwarfs and the growth hormone receptor knockouts in fact eat more. And the interpretation of this is not very straightforward. There's a huge difference in body size, but also a difference in body composition. Their brain size is not reduced nearly as much as body size, so they have lots of metabolically active tissue. They have less lipids so they have somewhat less of less active tissue. There's also a possibility that they are not as effective at absorbing food, so my hunch would be that this increased food intake per gram body weight doesn't necessarily mean that they have higher metabolism, in fact their metabolism might be lowered. Also their challenge for maintaining body temperature is different because they are so small.

Other comparisons shown here: all of these animals are hypothyroid, but there are big differences in degree. Ames or Snell dwarfs are severely hypothyroid, the others mildly. Body temperature is reduced by caloric restriction; severely reduced in the dwarf, about 1.5 degrees Centigrade; a very marginal effect in the knockout. Glucocorticoid data were discussed; prolactin -- when I made this slide I really didn't know if there was enough data on prolactin, but I think it's generally reduced in caloric restriction. It's severely reduced, essentially absent, in the dwarfs; in the knockout males it's elevated. And fertility is severely suppressed in the dwarfs, and somewhat suppressed by both caloric restriction and the growth hormone receptor knockout, but these are both animals, both sexes.

So I think what these data do is they point to some things which seem like a common theme, and therefore are presumably involved in it. There are others which vary, and I think also what at least I think I can see in the data is that the animals achieve this prolonged longevity by different combinations of the same mechanisms. That some of them may have a very pronounced reduction in insulin but less effect on body temperature or thyroid function, and others have a quantitatively different combination of the same mechanisms. And we thought that it might be instructive to look at the interaction of caloric restriction and the genetic life-extension. And the next two slides will be a progress report on an experiment I've been doing almost three years, on the calorically restricted dwarf mice. And what I plotted here were the result of an analysis of variance for their age at death or age of live animals at the time we looked at the data, which was a few weeks ago. And I added here the difference of ratios of animals still alive. So what this slide shows is that caloric restriction in this strain of mice extends life, not a very surprising finding; dwarfs live longer than normals; but looking at the number of survivors at this particular time it seems that the dwarfs benefit from caloric restriction. At the start of this experiment we half expected this would kill them -- you know, these small hypothyroid, frail animals. In fact, not only do they do fine, but I think they do better than ad libitum-fed dwarfs. And the data for the females are even more interesting: they show again that caloric restriction extends life in normals; they show also that female dwarfs live much much longer than normals, a difference of almost a year; in addition the data show that ad libitum-fed dwarfs live longer than calorically restricted normals. So at least 30% caloric restriction is not as effective as the dwarf. And again we have a significant difference in the percentage of survivors at this point in the experiment. So what we would like to do is to use this kind of experiment and look at various physiological parameters, look with this effect on normal aging, and in this way try to identify potential interventions which are the common themes: which are the potential mechanisms and which mechanisms could be utilized. And what I envisage at this point is one day, perhaps, treatments (this may be better for example if better drugs for insulin resistance were available), to include them in the treatment just in the way that blood pressure-lowering drugs or cholesterol-lowering drugs are now used in the population, to consider this a risk of short life and maybe pick up other parameters as targets for therapy.

And on the last slide I just want to bring up an issue -- do these little mice have anything to do with humans. And no one knows enough at this point to answer this question with any degree of assurance. But there is this bit of data which suggests that at least it might be relevant. This is a paper which was recently published on a few pedigrees from Croatia, of people who have a mutation at the Prop1 locus, the same locus which is mutated in the Ames dwarf. The phenotype of these people is somewhat more severe than the dwarf mice: they are growth hormone deficient, they are severely hypothyroid; they're also hypogonadal. And here are the data on age at death of all of these individuals, who are apparently not treated, so they did not have thyroid replacement -- where they live I guess it was kind of warm -- and this treatment was never instituted, and some of these people were interviewed, so there is data on them. And they complained that they were always cold and sleepy. But they lived to a rather impressive age, one of them living to 91, and in this particular population the average age was about 55. So these data, I think, bring up another issue, I don't know if we have time to discuss this, that some features of extended life may not be everything we would hope for. That maybe there are some incurred costs in programming for long life, and the phenotype of these individuals is certainly not something that any one of us would wish to have that kind of clinical feature for someone we know. But the fact is that they did quite well. By the way, these people were not cretins -- with this untreated severe hypothyroidism, they went to school and they held white-collar jobs. So they apparently managed quite well and they lived to a very old age. So they obviously represent an extreme, as I think the Ames dwarfs do, but they point, I think, to a connection between reduced growth, reduced body size, perhaps reduced metabolism, and long life. And this relationship, I think, becomes more and more real, and I think it's being taken more and more seriously as more data accumulate.

So this is really all that I have formally; I did not get into growth hormone and androgen regimens as sort of reversal of some aspects of aging; I obviously would be happy to try to answer questions.

Aubrey de Grey: Right, let's start there. It seems to me that the popular perception, at least, of how this work relates to growth hormone supplementation is that the evidence is quite compelling, taken as a whole, that growth hormone is bad for us if we have too much of it early in life, and indeed through young adulthood -- maybe it shortens life as a result -- but that it has a reasonably general therapeutic effect later in life, when we're suffering cell loss and loss of cell mass and so on. I wonder whether it's possible, or perhaps already in the pipeline, to do much more direct tests of that hypothesis, by developing mice with inducible growth hormone receptor genes, for example, or inducible Pit1 wild-type transgenes in a mutant background, so as to be able to vary the level of growth hormone in these animals constitutively, without having to worry about injections every day or anything like that, on an age-related basis. It seems to me that that could be a very aggressive and very direct test of the age-related benefits versus deleterious effects of growth hormone. Is anything ongoing?

Andrzej Bartke: Well, several experiments of this type were proposed in various applications, including some of the ones we put in. I know that there is a study ongoing with replacement of growth hormone early in life in dwarfs, which is addressing a different question. There are plans to use transgenic growth hormone expression in these type of animals under control of a promoter which you would be able to modulate by drugs. Of the data which are available now, I think there isn't an awful lot. There is abundant information about the transgenic mice overexpressing growth hormone to a considerable extent throughout the lifespan, which would be clinically the counterpart of gigantism. And these mice have reduced lifespan -- this is a major effect. Very major -- some of these mice live less than half of normal lifespan. Which I personally found very striking, because when they are young they look like Supermouse, they are slick, big, healthy, beautiful animals -- they certainly don't look like animals that there's something wrong with. And then as early as six months of age, they begin to gray and thin and get scoliosis and start looking unhealthy.

Aubrey de Grey: So it sounds as though even taking into account the work that you say is being planned, that there really isn't a sufficiently comprehensive effort to explore the modulation of growth hormone.

Andrzej Bartke: There was a recent paper (abstract, I don't remember) by Kalu from San Antonio, where he injected both rats and mice for prolonged periods of time with growth hormone. And I think the question he was asking was really do you do any harm by giving growth hormone? And his answer was no. That was with the doses he used, which he considered reasonable doses, and he started sort of in mid-life, if I remember this well. And he elevated IGF-1 in these creatures, and apparently it did nothing to either lifespan or age-related pathologies. So his answer was that treatment of middle-aged to elderly rodents with growth hormone, chronic treatment, doesn't seem to have any negative effects.

Aubrey de Grey: And that's even taking into account the presumably negative effects of being injected every day. So I guess that counts as a preliminary support of that hypothesis.

I did want to talk rather more globally about the endocrine system in general. Of course the problem is that growth hormone naturally, in normal people and normal animals, declines with age, and many other hormones do the same. Now, I mentioned the loss of follicles in the ovary at the beginning of the day as the one of the very few things in aging which we can really unambiguously say is a program that has a systemic effect on the levels of certain hormones, which has very pleiotropic other effects. Of course telomere shortening is another example of something that *may* have systemic effects, but the actual pathway from the primary thing to any pathology is of course much more tendentious and preliminary at this point, whereas in the case of the changes of hormone levels resulting from the menopause, and indeed in males from loss of androgens, the situation is much more direct. Now, we didn't talk about bone when we talked about muscle, but of course exercise has an enormous amount of beneficial effect on loss of bone mass, and loss of bone density during aging. But of course so does replacement of estrogen and replacement of androgens. So it seems to me that we need to have as accurate as possible a global picture of what we might achieve by stabilizing the levels of all hormones in the body indefinitely.

Andrzej Bartke: Well, I think there are several people here who know more about it than I do, so please correct me if I overgeneralize. But I think that maybe it would be fair to say that the current evidence indicates that you definitely can improve muscle mass, bone mineral density and body composition in terms of lean to fat, with particular emphasis on frontal obesity which we think is very important, that you can achieve these objectives by sex steroid and/or growth hormone replacement. I think these studies all also indicate that you can get positive interactions, augmentation of this effect, by exercise, and that in many cases people are able to get a similar effect with exercise alone. There is also a concern for demonstrated and presumed negative effects of these therapies. I think these effects are kind of obvious -- cardiovascular complications of steroids, cancer-inducing effect of steroids, insulin resistance of growth hormone and what not.

Aubrey de Grey: So let's go into, to interrupt you a little bit, to go into the deleterious effects and what is known about how they come about. One thing that of course always comes up with the very primitive systems that we have for maintaining and supplementing hormone levels right now is the fact that it's a bolus that goes in, once a day, it's not the same sort of gradual -- diurnally varying maybe but nevertheless relatively gradual -- release that happens naturally. Is it believed that the very sudden spike of hormone supplementation that happens with current treatments is largely responsible for those deleterious effects that you listed a moment ago?

Andrzej Bartke: I don't think so, and also I think this problem is really technically solvable, because already on the market there are two methods of androgen replacement which reproduce reasonably faithfully the normal pattern, there's the patch and the gel. And I think, if I'm not mistaken, they're using the growth hormone-releasing peptide. You can get something not too far from normal growth hormone release. So this is already manageable.

Aubrey de Grey: OK -- and these deleterious effects that you mention, have they been assayed in such systems, such slow-release systems?

Andrzej Bartke: I think not extensively. No -- no, many of them were, they were also tested with the more conventional methods of application, and, you know, these studies are ongoing in various human populations and in very large studies, and in general the reports indicate no clinically significant problems, no significant changes, within the normal range, things of that sort. There were some negative effects with growth hormone. These are actually quite long tests, including some people dropping off studies because of these uncontrollable side-effects. But many of these things are sort of in the eyes of the beholder: some people using growth hormone clinically consider weight gain and water retention as one of the significant side-effects, other people using growth hormone clinically say that dehydration is a symptom of growth hormone deficiency, so of course you correct it when you give growth hormone and it's not a side-effect but an expected normalization of the parameter. So you have this kind of divergence of opinion. But I think that the consensus seems to be that for both androgens and growth hormone if you keep the doses low you're probably not doing any harm.

Judith Campisi: What about if you're castrated? Because we know if you castrate males they don't get prostate cancer, so that would indicate that the normal fluctuation in androgens is bad.

Andrzej Bartke: Well, this is something that of course is always mentioned in terms of androgen replacement, and it's considered a strong contraindication if the disease exists. As I understand it there is still no consensus on what induces prostate cancer in men. Obviously androgens increase it -- there's no question about it, the data that you mentioned that it does not develop in castrated or hypogonadal people. And everything points out that it's testosterone, but I don't think anybody has ever demonstrated that testosterone is the real culprit.

Christopher Heward: In fact, in the case of prostate cancer, the current thinking is that it may actually be estrogen, or estrogen derivatives, that are responsible, rather than testosterone.

Judith Campisi: May it maintain the cancer?

Andrzej Bartke: Oh, absolutely. There is also, by the way, an increasing amount of evidence that IGF-1 may be a major risk factor in breast cancer. And this particular type of cancer, and other types of cancer, are on the list of the potential negative effects of growth hormone. But again this is reasoned from the epidemiological data rather than demonstrated.

Aubrey de Grey: Well, that really brings me back to your mice. Mice tend to die of cancer quite a lot, and your mice live a long time, from which I conclude that your mice are rather resistant to getting cancer. Is that the correct interpretation?

Andrzej Bartke: We have a rather large study of histopathology of dwarf mice dying of natural causes, but for various reasons that have to do with money or funding of our collaborators, things like that, we don't have very much data. Of the data that we do have, it seems that the most reasonable interpretation was that the dwarf mice definitely develop cancer. Which was a new finding, by the way, because there was some literature on Snell dwarf mice, showing that they do *not* get tumors, that they are resistant to induction of tumors by chemical carcinogens, and that they do not support well growth of transplanted tumors -- there were older papers of this type on Snell dwarfs. So we really thought that dwarfs may be tumor-free or nearly free-free, but they clearly are not. So that's one thing that we know for sure. It seems that the type of tumors and the overall incidence is similar to what we see in other mice. We assume that they develop at later age, but really from the data we have, because we didn't do cross-sectional studies, from the data we have we could not distinguish the possibility that they just grow extremely slowly. It's something we obviously need to do, but this would have to depend on funding.

Bruce Ames: The health food stores are selling all these DHEA pills, and old people seem to be popping them. Do you thing this is bad?

Andrzej Bartke: We don't work in this area; in the mouse models that I was discussing, or in any other mouse models, it's sort of difficult to evaluate as well, because DHEA is not a major product of the mouse. The levels of DHEA and DHEA-S are extremely low. So it's probably not a very good model. From what I understand about the human use, I think there's lots more opinions than good data on the benefits. There's some literature that's difficult to ignore, from very good people, but there is also negative. Some people who work in this area think that most of what's sold on the market probably doesn't contain enough DHEA or DHEA-S to be biologically active, and that it may be just as well, because the effect when you take it, primarily it's a precursor for sex steroids and most people acknowledge that unsupervised use of sex steroids may be not a very good idea.

Christopher Heward: I can speak to that issue a little bit. It's, obviously, something that we encounter at Kronos quite frequently. There's been an ongoing debate at Kronos regarding that very issue, between some of the physicians, Mitch Harman and me.First of all, with respect to the products that are available in the marketplace, there is tremendous variation. When I was at Emerald Laboratories, we tested a number of DHEA products and found them to contain anything from zero DHEA to 100% of what was represented on the label. So that's a problem. If you have a good product, however, you can deliver DHEA orally. It does pass though the intestinal mucosa and it does measurably raise blood levels of DHEA sulphate. So, if, in fact, DHEA sulphate is something that we need in order to live healthy youthful lives, then we can definitely raise blood levels by giving supplements. The debate is still whether or not DHEA is an end hormone in itself. If you look closely at the literature, and Mitch Harman and I have looked pretty closely, it is difficult to find an effect of DHEA that can not be explained by conversion of DHEA to estrogen or testosterone. There's really not much there. Moreover, if it really is a hormone, one would expect that somebody would have isolated a DHEA receptor, and that's not been done. There are no DHEA receptors that we know of. So DHEA is probably just as the textbooks have always said, a biosynthetic intermediate on the pathway between cholesterol and the sex steroids.

Bruce Ames: They're doing a big study at UCSF, of which the first part has just finished, on serum DHEA levels, and we're getting a lot of samples and we're supposed to be doing some assays; the last time I talked to the PI he hadn't broken all the codes yet.

Christopher Heward: Another thing that we have to be very careful of, in our patients that do take DHEA, is that we don't just monitor DHEA sulphate levels. We also have to monitor testosterone levels, dihydrotestosterone levels and estrogen levels, because one can very easily become abnormal in one of these derivative steroids with normal or only slightly elevated DHEA levels. That's a concern.

Aubrey de Grey: One last thing I'd like to mention in this slot, before we go to Judy, is a paper I came across, some very recent work from England, from Richard Aspinall's lab, in which not IGF-1 this time but IGF-7, was found to be able to reverse the involution of the thymus during aging, which is probably one of the most dramatic losses of cells of a given organ that happens during aging. This can be very dramatically, very substantially reversed by IGF-7 treatment. It seems to me that if we were to want to reverse age-related changes in hormone levels in general, then one thing that we'd definitely need to look at would be to make sure that in any glands for which the cell loss was the major cause of the lowering of hormone levels, to find similar ways of reversing that cell loss. So one thing I'd just like to ask Andrzej and others is, which glands in particular appear to undergo greatest cell loss, and in particular in which cases has that been linked to lowering of hormone levels with age? Is the pituitary a lot smaller in older people, for example? Not counting the cancers!

Andrzej Bartke: Well, all the hormones we're talking about, sex steroids, growth hormone, IGF-1 they're all mitogens. So it would make sense that they would work partially by cell depletion in their target. Cell loss: you mentioned the thymus of course.

Judith Campisi: Peter ??? has looked at the adrenal in this regard, and it's very clear that his data show that it's not cell loss but loss of function.

Aubrey de Grey: Yes, that's what I wanted to ask about, exactly. So, what is driving the loss of function, if it's not cell loss? Why are the cells getting worse at producing the stuff?

Judith Campisi: Actually what happens in the adrenal gland is a little hard to relate to culture studies and in vitro studies. But what Peter has shown is that the cells undergo rapid senescence in culture, and it's faster in older animals, but the phenotype of the senescent cells is not that they stop producing adrenal steroids but they start producing altered steroids.

Aubrey de Grey: So you'd see this as an example of senescent gene expression, then?

Judith Campisi: Yes. But as I recall, there's no evidence for shrinkage.

Andrzej Bartke: For the ??? cells there is evidence of reduced responsive to stimuli with age, at least in rats. Rather like in humans. And I don't think there is a major decrease in number, but rather in function.

Aubrey de Grey: OK. I'm afraid we're going to have to move on, fascinating though this is; Judy, your turn.

Judith Campisi: I may change my talk as we go on, in the light of the discussion. Aubrey asked me to talk about cell senescence. So I'm not going to describe to you very much about the phenomenon, it was discovered 40 years ago, and it's been experimentally proven that its driven by DNA replication. It's not driven by time. So the first thing I want to tell you is that, at least in my mind -- I don't know how pervasive this is in the field, but certainly in my mind, that there's been a major paradigm shift in the way we think about senescence. The initial emphasis has always been on cell division. So when Hayflick said this may have something to do with aging, everybody was talking about the fact that you've got tissues that can no longer replicate, and that causes aging. I really think we need to get away from that idea. And the reason is, two major bodies of evidence, not only in my lab, but in 20 or 30 labs working in concert, have now clarified a lot about what the phenotype is. So the first is, a whole lot more happens than that cells don't divide. So for stromal cell, or T cells, probably for endothelial cells but not necessarily all cell types, there is an accompanying resistance to apoptosis. It doesn't mean the cells don't die, but for example if you take T cells, the natural inducer of death is Fas ligand, after they've undergone senescence and you apply Fas ligand, they don't want to die, though if you give very very high doses, then you can get them to die. So the cells become resistant to apoptotic death. The second thing, which I thought was most important, is the phenotype of altered function. So in the case of neocortical cells, we see this altered profile. In the case of fibroblasts they go from being a matrix-producing cell to being a matrix-degrading cell, pouring out collagenase. Endothelial cells develop a very ??? phenotype, senescent endothelial cells are pouring out ???. So what we don't know, at the molecular level, is how these three phenotypes are coupled. We've very good reason to believe that the phenotype involves global changes of gene expression, rather than genes here and there being turned on at the gene level, but this is still a wide-open field. And coming back to microarrays, Aubrey, this is one example where the microarrays will be extremely helpful as we begin to tinker with the chromatin, to understand how the chromatin is really adapting.

So that's the first paradigm shift, is that this has a lot more to do with cell phenotype than cell division. The second paradigm shift is that we know from some very early work, that replicative senescence, what Hayflick first described, is driven by telomere shortening. There's no question about that now. But now we also know that you don't need telomere shortening to acquire the phenotype. Direct DNA damage, particularly double-strand DNA breaks, will turn a normal cell into this phenotype. Similarly, some well-known oncogenes -- take a proto-oncogene, Ras, everyone knows Ras is an oncogene, take a normal cell, activate Ras and you get a tumor cell; it doesn't work unless that cell had a pre-existing mutation. So normal cells, that have all their checkpoints intact, will respond to oncogenic stimuli by senescing. Not by apoptosis. So what this tells us -- what all these things have in common -- is they can cause cancer. Short telomeres predispose towards genetic instability, which is a prescription for cancer. Obviously oncogenes predispose to cancer, and so does this. So what this tells us -- tells me -- is that nature gave us what we're now calling the senescence response, as a first line of defense against cancer. I really think that there are very good reasons to believe that, just as nature gave us apoptosis as a second line of defense against cancer, the first line of defense in a normal cell is to senesce. So if that's the case, that really says that this process, not just cell senescence, but now I'm going to call it cellular senescence, induced by any of these methods, evolved to protect us from cancer. And I think there's lots of good reasons to believe that. The most powerful evidence we have is the mouse model: single gene knockouts, whose cellular phenotype is failure to senesce. The organismal phenotype is early death through cancer, multiple cancers.

Aubrey de Grey: What genes?

Judith Campisi: Two genes: p53 and p16. With knockouts of either of those genes, the cell no longer senesces.

-- tape change -- And this is what Steve Austad would have told us, if he were here, which is that all of our genes evolved under a scenario in which lifespan was very short. Probably all our ancestors until recently didn't live much more than 25 years. So any allele that had no effect until age 50 would never have been weeded out of the population -- there were simply no survivors against which the force of natural selection could act. The point was, if there are processes that have unselected adverse phenotypes, they will be maintained, because there's no selection against them. So my proposal now is that cellular senescence is an example of what the evolutionary biologists call antagonistic pleiotropy. I think there's a huge amount of evidence that it's good for us when we're young -- it does protect us against cancer -- and I'll show a little bit of data that says it's bad for us when we're old.

So I would argue that this is the selected phenotype. If you have irreversible growth arrest, that cell will never become a tumor. So picture a damaged epithelial cell or a fibroblast that's senescent -- you'll never get a tumor from that cell. Like postmitotic cells, this cell cannot become a tumor. But we can suggest that these are the unselected, deleterious phenotypes that evolution has never bothered to clean up. And if we stick around, and live as long as we have been in recent times, for a million years or so, probably evolution will get around to fixing this.

Aubrey de Grey: That's a big "if"...

Judith Campisi: Right -- if we don't kill ourselves! I should point out that we have escaped predation, and the only predators we haven't escaped are ourselves, we're our biggest predator.

OK: I think many of you know that starting around five years ago we discovered a marker that correlates very well with the senescent phenotype -- not necessarily replicative senescence, but the senescent phenotype -- and we could show that cells with this marker accumulate in tissue with age. So I think we can say safely that senescent cells accumulate with age. So why might this contribute to aging? We think it's because, in every cell type that I'm aware of, when the cells undergo senescence they develop an altered function, a secretory phenotype. This is just gleaned from the literature on stromal cells: they make matrix proteases; fibroblasts and epithelial cells in particular, but also others, are pouring out proinflammatory cytokines, etc.

So very simply, the idea is that with age, you begin to accumulate senescent stromal cells, senescent epithelial cells, that are producing all these things that disrupt the tissue, leading to loss of function. Now this is a tough model to test, and I'll tell you how we think we can test it, but it's not going to be easy.

But I want to take you back now to our earlier discussion about mutation load. Even in relatively young organisms, there are now, in quite a few papers, data showing that apparently normal tissue has an unbelievably high load of mutations. If you look at normal breast tissue, there's an amazing amount. Other people have shown that there's a p53 mutation in apparently normal tissue, with very sensitive assays. I think this tells us that throughout our life, there are more mutations than we think, even early in life. But most of us don't develop cancer until much later in life -- cancer rates go up after about 50. We've thought about the fact that these senescent cells may be destroying their microenvironment, and we thought that maybe one of the reasons we develop cancer in the later part of our life is that after senescent cells begin to accumulate, the tissue environment can now encourage the growth of whatever potentially tumorous cells are present. So I don't need to remind you that cancer rates go up exponentially with age. There's no age on this graph, and that's because if this is a mouse, this is a year, and if it's a human, this is 50 years. And we don't understand what it is that makes the difference. But in all mammalian species, to my knowledge, it is in the latter third of the lifespan that you get cancer.

Now, cancer biologists have been telling us for years that this is mutation accumulation -- you need five hits, six hits or seven hits to get a cancer, and that takes time. And I'm not saying that mutations are not important in cancer -- they are preconditions for cancer. However, we now know that if you look at normal tissue, you can always find four or five mutations in critical oncogenes. We also now know that if you look at a benign tumor, as against a malignant breast cancer, a cancer that's going to kill you, and ask how many mutations there are, the answer is 30, 40, 50 -- and there's no difference. We cannot tell the difference between a benign carcinoma and breast cancer by simply looking at the total mutations: there is a huge mutational load in even relatively benign tissue.

So I think we need to invoke something that synergizes with mutation accumulation. And I also want to remind you that a great deal of cell biology has told us that you can take a very messed-up tumor and get that cell to behave pretty normally, by serum deprivation. So this is an example. These are normal and abnormal human breast cell lines, and if you passage these cells in he absence of growth factors, you get tumor cells that no longer form nice colonies. And these cells kill the animal. If you inject these cells into an animal the animal dies, and if you inject these other cells that animal does not die. Now what we have been able to show is that if you take an anti-integrin antibody, and trick the cell into thinking it has a normal microenvironment, they now form nice normal colonies, and they no longer kill the animal. This cell has probably 50 different mutations. So the idea, then, is that the problem is mutation accumulation coupled to aging of the cellular environment. If this is the case, then we should be able to show some effect of senescent cells on precancerous cells. I'll just show you a couple of slides: these are three different types of endothelial cells. None of them are normal, they've all got p53 mutations, but here, they do not form tumorous tissue. Put them on a lawn of presenescent fibroblasts, they behave like tumor cells. This effect can be shown in vivo. So this is a tumor experiment where we put the cells in the absence of any fibroblasts and we got no tumors, with presenescent fibroblasts the tumors are present, we have some tumor growth, smaller tumors, and with the senescent cells we see many many more tumors and larger ones.

Gregory Stock: What was the axis on the bottom, the time axis?

Judith Campisi: These are just individual animals. We had 15 animals here and we got five tumors, and so on. But yes, we have time course data and we can show that it's time-dependent. Actually these tumors have a much greater tendency to shrink. Our guess is that the presenescent cells are undergoing apoptosis and they die and the senescent cells are resistant so they're sticking around longer.

So cellular senescence may affect tumorigenesis. One quick point about this stimulation: we can account for 80% of the phenotype by secreted materials. This is simply putting epithelial cells on medium that is secreted by senescent or presenescent fibroblasts, and we see a big effect.

So this gives us some ideas about how we might treat tumors. Before I do that, let me get back to the telomere hypothesis, as the idea of what causes senescence, and address this issue of the role of telomere shortening. So I want to go on record that there is no telomere hypothesis. It really is a misnomer. I'm not saying that telomeres and telomerase are unimportant, and I'm not saying that therapies directed against telomerase do not hold great promise for the treatment of cancer -- that is really really valuable. But there is no telomere hypothesis of aging. The hypothesis is that cellular senescence contributes to aging, telomere shortening being but one of many routes to that phenotype. So is Dolly older because her telomeres are shorter? It's a completely irrelevant question! Because what happens is when the telomeres reach a certain critical length, which is actually quite short, that triggers cellular senescence. But independent of that are other mechanisms for producing the phenotype: background damage, oxidative damage to mitochondria, and so on. So the hypothesis really is that cellular senescence contributes to aging, and I'm the first to admit it's an as yet untested hypothesis. But again, I really want to point out if I had, especially an early cancer, and I could take a pill to inhibit telomerase, I would take it. I think there's really great promise for anti-telomerase therapies for treating cancer. It'll meet problems -- we now know that cells can often maintain telomeres without telomerase -- but as a first line of defense, in combination with other therapies, it has great promise.

I'm going to finish up by telling you what we've learned from the telomerase knockout mice. So mice are a little bit different from humans in that they tend to be much more promiscuous with expression of telomerase. But nevertheless, it's not true that all mouse tissues express telomerase. If you knock out the gene for telomerase, for the first four generations there's no phenotype. That's not unexpected, because we know that mouse telomeres are much longer than human telomeres. And therefore it takes several generations to get them down to human size, where we might expect to see a phenotype. But in the fifth or sixth generation we do see cellular senescent cells in highly replicative tissues. The first thing you might expect is a lowered cancer rate. Not true. What happens is that they do get cancer. How much cancer they get, of course, depends on the genetic background: in a p53-/- background they get a lot more cancer, in a p16-/- background they get less cancer. But nonetheless, they all get cancer. And the reason is because the telomeres have shortened, and that induces genetic instability, and that gives rise to cancer. All these cancers are telomerase negative. So what this tells us is that the process of cellular senescence is still active -- it's not telomerase-dependent. So I think what we've learned from all of this is that there's been much too much focus on telomeres. The focus on telomeres and telomerase in cancer is very appropriate, but when we look at the aging process it's really part of a much larger process of cellular senescence. So what might we do? Realizing that this was selected for preventing cancer, we don't want to get rid of cellular senescence -- we don't want to create animals whose cells don't senesce, because they'd die of cancer. So I can think of two things, and the first is something that will only work in mice. We know that there are certain genes that are turned on when cells undergo senescence. We can take the promoters of those genes -- we've already done this with one, the collagenase gene. We've identified the region of the promoter that drives the senescent gene expression, and you can hook it up to an apoptosis inducer. What you really want is to create an animal where senescent cells die. You don't what them sticking around, and you certainly don't want them not to senesce. So that's one approach. The other approach is to inhibit the secretory phenotype of these cells. And so we recently started a collaboration with ??? in Chicago, which has developed some dominant negative proteins that are reported to inhibit secretion. And our prediction would be, then, if we could inhibit the secretory phenotype, we may be able to lower the unselected deleterious effects of senescence. But again these are all in process and we don't know if they're going to work. But these are the kinds of strategies I can imagine. With regard to translation to humans, the second approach obviously is something we could think about working with. The first approach -- until we get much more liberal, we're not going to make transgenic humans!

Aubrey de Grey: Thanks very much Judy, that was marvelous. Since you ended up talking about the telomerase knockout mice, and about what they can teach us, I wanted to ask a slightly complicated question about that, bearing on one thing that you said and also a couple of pieces of information that you didn't mention. You mentioned that p53 and p16 mutations appear to eliminate replicative senescence in mouse cells.

Judith Campisi: And senescence induced by oncogenic stimuli.

Aubrey de Grey: So now: 20 years ago now, it was published that mouse cells in vitro have a slow growth phase, at a respectable number of doublings commensurate with mouse lifespan, something like 10 or 15, whereas now if you take, for example, embryonic stem cells, which don't exhibit that slow growth phase, and you put them in a telomerase-negative environment, as I understand it they go on for about 450 generations, and they don't actually exhibit replicative senescence -- rather, they undergo apoptosis.

Judith Campisi: What happens is they undergo crisis, which is a very unstable situation. But you do get survivors that come out of that.

Aubrey de Grey: Rather like human cells. Whereas telomerase-positive but normal cells, like fibroblasts from mice, it's a rather poor form of senescence but nevertheless something that looks very much like replicative senescence, at a very small number of doublings. Now, I'm not aware of any work on the relevance either of telomerase, or of p53 or p16, to that process, so I wonder what you can tell us about that.

Judith Campisi: I can tell you quite a lot. The first thing is that people have taken fibroblasts from p53 knockout mice and p16 knockout mice, and the statement that I made that the cells don't undergo replicative senescence is just the experiment you describe. So we know that that happens in fibroblasts. The second thing is that we're now growing -- first of all, to undergo 10-15 doublings is really rare. They undergo more like 7-10 doublings. We've now taken mouse embryonic fibroblasts and grown them in 3% oxygen. So you know from Saito's work that if you do that to human fibroblasts you get a 20-30% increase in replicative lifespan. If you do that to mouse embryo fibroblasts, we're getting 2-3 *fold* increase. So the mouse cells are much more sensitive to oxidative stress. I think the reason for this, which we touched on earlier, is that mouse cells and human cells have different repair phenotypes. They simply don't repair as much of the damage that the DNA is incurring. And even at 3% oxygen, they show the same senescent phenotype. So the bottom line to your question is: I do not believe that mouse cells acquire the identical phenotype as human cells. There's something completely different that we don't understand.

Aubrey de Grey: So, another question, closely related: You gave a summary of why it wouldn't be expected that there would be much influence on lifespan in the telomerase knockout mice until you've got down to 4-5 generations of inbreeding. Of course that very much begs the question of, well, what are they dying of in the wild-type, but I think you've answered that by saying we must focus on these other, non-telomere-dependent triggers of p53- and p16-dependent replicative senescence methods. Are you saying, really, that normal mice die at least partly from replicative senescence, but not at all from telomere-dependent replicative senescence?

Judith Campisi: First of all, normal mice of course die of starvation or foxes or whatever. But in the laboratory, I would say that there's a good chance that a contributing factor is cellular senescence.

Gregory Stock: You said that cellular senescence was an example of preventing cancer. I was wondering if you would mention some of the possible mechanisms for that, because it would seem to me that you were also saying that cancer can't develop when there is a normal cellular environment -- that a disruption of that environment is required for tumor development. It seemed to me you were also suggesting that the presence of senescent cells is instrumental in creating the abnormal environment that generates the cancer, so it's not obvious to me what the advantage is of senescence.

Judith Campisi: So the advantage is, when you're young, and along comes a carcinogenic agent that causes a double-strand break, this is a potentially catastrophic condition. In the absence of normal checkpoint control, the absence of the mitotic checkpoints, cells have the potential to repair that damage. And the primary way that they do it is by a process called non-homologous end-joining. If there's a sister chromatid around it'll try to do it by homologous recombination, which is very faithful and error-free, but most of the time cells are in other parts of the cell cycle, and then they do it by non-homologous end-joining. It's very dangerous, and it often gives translocations. So that is what normal cells are accustomed to doing. Now, if your normal checkpoints are intact, that's fine -- you join the DNA, probably as a translocation -- and you want that cell sitting there, functioning, but you don't want it to divide. And that's what senescence achieves.

Gregory Stock: But the functioning is very abnormal in the senescent cell.

Judith Campisi: That's the unselected adverse effect. So when you're young, it's simply preventing the proliferation of such damaged cells. The only downside is late in the lifespan. It's only the time that cells with this abnormal phenotype need to accumulate in tissue. And that synergizes with mutation accumulation. Imagine two types of cells, one with a mutation in p53 and one where there's not.

Gregory Stock: You're saying it's the checkpoints that are requiring the normal environment?

Judith Campisi: That's correct. And so again, probably you have to develop a mutation in the checkpoint and also with age you have mutation accumulation, then you have a problem where you can no longer senesce, but you've got the microenvironment keeping you in check, but meanwhile the neighbor might have senesced and be affecting the microenvironment.

Gregory Stock: So, what do you think would happen if you do as you said and place a cell death gene after a senescence-dependent promoter: then what's going to happen?

Judith Campisi: I would predict that in a normal mouse, where there's telomerase, that the incidence of cancer would be reduced.

Aubrey de Grey: How's it going, this work? How far have you got with it?

Judith Campisi: We have the element, we have it hooked up to a reporter gene, and we have one line of transgenic mice. But the mouse has silencer mutations, so we're going back now to recreate the mouse. So we're going to do it first with a reporter, I think.

Gregory Stock: Wouldn't you also predict, since the microenvironment would be uncorrupted, that you would also have reduced aging?

Judith Campisi: That's a much harder hypothesis to test. And if it were true, I would be dancing on the table, but at this stage...

Aubrey de Grey: I was going to ask a question very much along those lines. You explained the very localized, paracrine hypothesis of cancer formation due to senescent cells very nearby a cell that's already activated or initiated. But of course that works in a situation where you have just very occasional cells undergoing senescent gene expression, on the sort of scale that you've demonstrated with your assays -- one in 10,000 or something like that. Something that we talked about in Andrzej's slot, to do with for example the change of hormone levels from the adrenal gland, would only (it seems to me) be explainable in such terms if the proportion of senescent gene expressing cells was vastly higher. Do you have any evidence plus or minus on that point?

Judith Campisi: None. Very little evidence. All the evidence we do have suggests that even in very old tissue the fraction of senescent cells never gets to be very high. So this is why, of course I would be delighted if by killing off senescent cells you could actually preserve tissue function, but I don't think that's the most realistic expectation at the moment. The greater expectation is that we would reduce cancer rates.

Gregory Stock: But if that's the case, aren't you saying that the senescent phenomenon really has very little to do with aging?

Judith Campisi: It's very possible that it has very little to do with the generalized decline in tissue function that is recognized as aging. I think it might be most important for cancer.

Gregory Stock: Just one final thing: it seems to me that there was a third approach -- if, in fact, the senescent phenotype has any markers on the cell surface that are unique, then it might be possible to develop ways of attacking them and actually going in and cleaning them out.

Judith Campisi: The only surface marker I know of is that Jim Smith, many years ago, thought he had such a marker, but to my knowledge that's never been confirmed.

Bruce Ames: What's the reason for cells senescing rather than apoptosing?

Judith Campisi: The theory of antagonistic pleiotropy, which gives rise to unselected phenotypes, is the reason. If evolution was thinking about how to keep us healthy when we're 80, then yes, evolution would have designed a way to do that. But since that phenotype is a late-acting one, it's not been selected away. But let me go on to a fundamental question, which is where does this phenotype come from? Does the fibroblast, for example, create a secretory phenotype? And the answer is no. In every instance that I'm aware of, the phenotype that's indicative of senescence still comes from the rules that are laid out when we're young for that cell type.

Aubrey de Grey: Which makes your death gene promoter experiment a very difficult experiment to do -- you wouldn't want it to be killing lots of cells when they were trying to heal wounds, for example.

Judith Campisi: You wouldn't want it to be killing cells during development either, right.

Roger McCarter: I'm still not clear what switches it on, though. What switches on when they get to be senescent? Why does that secretory phenotype occur? Is it uncontrolled, or the genes are not nearly as well regulated as they were before?

Judith Campisi: The question you're asking is what do a double-strand break in DNA, oncogenic Ras, or short telomeres have in common, that they converge on this common set of genes that are now over-expressed? The short answer is, of course, we have no idea. The longer answer is that all three stimuli have, at least in theory, the potential to reorganize chromatin. And that's why I think we should not be looking at transcription factors for collagenase, and so on, but rather genes that unravel heterochromatin, which in turn then allows gene expression changes, within the boundaries of the differentiated phenotype. Senescent T cells don't express the same genes as senescent fibroblasts.

-- coffee --

Bruce Ames: The remarkable thing about this is that cancer goes up at this power of age, but of course we have to understand the aging process. So I was interested in oxidation for a long time, because the thing about this, is that evolutionary biologists don't think cancer and heart disease are big factors in evolution -- mostly people were dying before they were 40, and when life was much shorter the degenerative diseases of aging weren't so important. So one of the trade-offs that we need to think about is related to metabolic rate. So we started when the radiation biologists had done the chemistry, and at one point 20 years ago we had this vision, ah, maybe we can look in urine and see what had been repaired out of the DNA. So if you get a lesion in the DNA, and the cell divides, you've got a mutation, but if you repair it out, by nucleotide excision repair or base excision repair, a very general system that takes a patch out of the DNA, you don't. So we started looking for oxidized bases that the radiation biologists had thought of. And we found them, and we tried to estimate how many hits per rat cell per day, and we came up with a number like 100,000. So every cell in the body is getting 100,000 oxidative hits to the DNA every day. But that's a somewhat soft number, because you can't control for cell turnover, and you don't know how much is coming from the mitochondria which are turning over all the time. But in any case the number was large, and that kind of said that oxidative damage was likely to be important. But what got me interested in micronutrients was thinking about the prevention of cancer. Because all the top epidemiologists think about 1/3 of cancer is due to smoking, about another third is due to diet, and then 20% or so is from infection.

So the interesting part seemed like the diet, because that's what people don't know about. One thing everybody agrees on -- there are 200 epidemiological studies that have looked at it -- is the importance of fruits and vegetables: the quarter of the population that eat the fewest fruits and vegetables have double the cancer rate of the quarter that eat the most. That's for every type of cancer. So that seemed pretty interesting, and then the question is what's in fruits and vegetables? It's known that 90% of lung cancer is due to smoking, but eat a lot of fruits and vegetables and you cut your lung cancer rate in half.

Christopher Heward: A quick question. If you eat a lot of fruits and vegetables, you probably don't eat as much meat. Is it clear that it's the presence of fruits and vegetables rather than the absence of meat?

Bruce Ames: That's called a confounder. There's an epidemiologists' joke that Miami is a weird place, where everyone is born Hispanic and dies Jewish. So in epidemiology you have to know is it being causative, so people have thought a lot about that, and the consensus is that it's the presence of the fruits and vegetables. Now for example smoking, where 90% of lung cancer is due to smoking, you have the chemicals, every kind of studies agrees on that, but yet, the smokers who eat their fruits and vegetables have half the cancer rate of smokers who don't. Lots of studies have shown this big interaction between diet and smoking. We showed that if you don't take your vitamin C, you're oxidizing your DNA. So that in human intervention studies, where you take ten people and treat them like rats, everything is controlled in the diet, when the vitamin C goes down, the DNA damage goes up. That shows in sperm damage. And what smoking does is deplete the vitamin C, so smokers have to have a much better diet in order to keep their vitamin C up.

So anyway, thinking about that got me interested in micronutrients, and the thing that really got me all excited about micronutrients was folic acid, and I'll just say a few words about that. By the way, DNA damage rates are going up with age. If you look at chromosome breaks, anybody's measure just goes up. So we're not keeping up. So why is folic acid interesting? Well, a guy named Jim MacGregor has been looking at micronuclei. So, when you make a red blood cell, you extrude a nucleus. So if you take mice and stain red blood cells for DNA, there shouldn't be any DNA, yet one in 2000 red blood cells stain for DNA. And what it turns out is that they had a broken chromosome which got left behind and surrounded by a membrane, and you have a little quarter or half chromosome left behind. And if you stain with a DNA reagent you see nice little bodies.

And what MacGregor stumbled on is that folic acid deficiency breaks chromosomes. It looks just like radiation. So he came to my lab and spent a sabbatical, and studied hundreds of mice, and then one of my graduate students took up the problem and figured out why. So folic acid is kind of the hot part of this. Homocysteine goes to methionine with methyl tetrahydrofolate and vitamin B12, so those two vitamins are necessary to methylate homocysteine to methionine. And you're constantly recycling here, and if you don't have enough of this or this, then homocysteine accumulates, and that gives you heart disease. So homocysteine really damages endothelial cells. A guy named McCully did an autopsy on a kid who had a genetic disease where he accumulated homocysteine, and at autopsy his blood vessels were like an 80-year-old man with severe atherosclerosis, and he'd never seen that in a young person. So the light went on in his head, and he started reading the literature, and he found there's a different genetic disease that causes homocysteine accumulation, and when one of those kids died and they did an autopsy, they saw the same thing. So he wrote some papers saying homocysteine gives you atherosclerosis. And nobody was interested, he never got tenure at Harvard, but it turns out to be true, and now there are a lot of papers that homocysteine really damages endothelial cells.

Now what we got interested in is why folic acid deficiency is breaking chromosomes. And that's because of a different mechanism. dUMP goes to dTMP with methylenetetrahydrofolate. This is a different chemical. And the reason is, it's working in exactly the same way as radiation. What radiation does is give you an occasional oxidative damage. And then base excision repair repairs it out, and you make an AP site, and then you get a nick, and then it's mended So that's how it works, and so there are efficient enzymes that can do base excision repair, glycosylases that recognize oxidized bases like uracil and take them out. But what radiation does is that it occasionally gives you a cluster of them. Occasionally you hit both strands at once. And that's bad, because then you get a double nick, and the chromosome falls apart. So the radiation biologists said that the dangerous part of radiation is the double-strand breaks. And since radiation does two at once, it's kind of a one-hit process. And what we found is that folic acid deficiency puts 4 million uracils in the cell, and with the occasional oxidative damage on the other strand you get a chromosome break. And what shook me up is when I looked at what percent of the population have a level of folate intake that causes chromosome breaks, it's 10% of the US population. Folic acid is named from the word folia, like foliage, so you get folic acid from green vegetables, spinach, things like that.

So that got me all interested in micronutrients. And now folic acid deficiency is associated with colon cancer, and others. So this list is getting longer, and I think it's likely to apply to all kinds of cancer. And now what we've shown is that B12 deficiency traps everything in this pool, and uracil goes up in people and in rats, doing the same thing. For B12 we've shown it in humans, and B6 deficiency too. So these three vitamins now end up breaking chromosomes in the same way as radiation. And when you look up what percent of the population has below half the RDA of vitamin B6, B12, and folate, it's huge. So folate was 10%; now they supplement flour with folate so it's going down. B12 is 4%, but it's 14% in the elderly, and old people should probably takes higher levels of vitamin B12. So any of those deficiencies works like radiation. But it turns out that there are 40 micronutrients you need in your diet. And when you look up any one of them, you find say 5% of the population is low, or 20% is low -- there are large numbers of people. Then, vitamin D, which may protect against cancer, all the Northern tier of the United States isn't getting enough vitamin D unless they're drinking a lot of fortified milk, because you need the ultraviolet light to make vitamin D in your skin, and dark-skinned people and most Northern people, like Brits, who are starved of sunshine, don't make much. I think that's why the Swedes have such a light skin, because you're selecting for light skin in Northern climates and selecting for dark skin in Southern ones. Anyway, so there's a long list of vitamin deficiencies that I think we should be looking at. And one that we're working on a lot is iron. So it's known that too *much* iron is bad, but we've been collaborating with ??? and Patrick Walter in my lab, and that's the biggest micronutrient deficiency in the world, because most of the women and children in poor countries are iron-deficient. And what the epidemiology says is that iron deficiency and zinc deficiency gives kids who don't do well in school, and if you supplement young kids then these kids they do better in school. With many things it seems better, but the evidence with iron and zinc is really strong. Even in the US, guess what percent of the population is really low in iron? This is women of menstruating age. 19% of the menstruating women. Even in Berkeley students, some high percentage of them are getting too little iron. So women need iron, and men are probably mostly getting too much iron. So you put iron in the flour, you're hurting the men and helping the women. So that's why they sell vitamin pills in two kinds, one for the women and one for the men, and the men's have very low iron. Anyway, what Patrick Walter in my lab has now shown is that when you're iron-deficient, it fouls up your mitochondria: you're breaking your mitochondrial DNA. Maybe it's the iron-sulphur clusters aren't passing electrons so efficiently, but whatever it is, you're getting extra oxidants, and you get a lot of oxidative damage. So in this whole idea of tuning up metabolism, getting everybody up to snuff, iron is important. Zinc deficiency is too. People say "cancer" and they think of toxins; they're not thinking of what would be happening in the diet. Zinc is in 300 enzymes with zinc fingers, including p53 and copper-zinc SOD, and enzymes involved in oxidative DNA repair, and 18% of the population are walking around with low zinc. So you just know it's going to foul up your enzymes. So, my colleague Keen at Davis showed that oxidative damage to DNA goes up when you're low in zinc. And there are other similar things. Your immune system goes out when you don't have enough iron and zinc; so there are a lot of things that go wrong with you if you're low in a micronutrient. Which is kind of what you'd expect.

Roger McCarter: We're all living longer and longer and longer. Does this mean that people earlier on were even worse off?

Bruce Ames: Oh, much worse. After my wife and I got married we went to England on sabbatical, and she's Italian, and this was in January, in Cambridge England, and there was nothing -- you could buy potatoes, and cabbage, and that was it! That was before the age of freezers. So she learned to cook potatoes in fifteen ways. The refrigerated boxcar and pesticides really lowered the price of vegetables; refrigerated boxcars brought oranges to Minnesota, and so on.

The really nice thing about micronutrients is that it costs less than a penny a day to make a multivitamin. They sell at between two and three cents. So for nine bucks you can get a jar that'll last you a year's supply of vitamins. So it costs nothing. And a quarter of the population are taking a multivitamin, but it's the rich eating the good diets that are taking them as insurance, when the people who really need them are the poor.

Aubrey de Grey: So, are there any tricks one can play like with the folate, to sneak multivitamin supplements into people's food that they're eating anyway?

Bruce Ames: There are various ways. One is you supplement, but in the long run I think we're better off with pills, because when you put iron in the food you differentially affect men and women; old people probably need different things than young people; and eventually they're going to have five different types of pills for ten different profiles, depending on your age or your sex or your genetic constitution. I can tell you an anecdote: I gave a seminar at Hershey medical school, and Hershey was the man that founded the Hershey chocolate company. He founded splendid medical schools and orphanages, he was very philanthropic. And I got a tour of the Hershey chocolate factory. One of my former students was chief toxicologist at Hershey Chocolate. So I said why don't you give me a tour, and she said: "You know, five years ago when the first rumors started coming out about folic acid deficiency for women giving this horrible birth defect, I said, well, it doesn't cost anything, we may as well add it to chocolate. So I talked my company into putting folic acid into our chocolate. And we were about to do it, when someone in the company said you'd better fly to Washington and check with the FDA. Why would they care, we said, but he said "just go and check." So she flew to Washington, made her pitch, and the FDA said no, and they never did it. The government has a rule that you can't put anything good into anything that's considered bad, like chocolate or cola. Anything that's poor to eat is bad, and therefore you can't put vitamins in it.

Aubrey de Grey: A bit of a Catch-22, really!

Bruce Ames: Yes. So anyway, for the government to do anything it takes a while. So right now we're going to have a pilot project to see if we can get vitamin pills into a poor black area and a poor Hispanic area, they're the two minorities that are really eating a terrible diet. We're working, based on what we find with human cells in culture, leaving out each micronutrient and figuring out the best ways to measure DNA damage and then going to small human intervention studies. Human intervention studies are very powerful -- it's not epidemiology, everything's controlled, it's like rats. So you just vary the micronutrient. And there's a guy up in the state of Washington who published a paper about B6. He had nine subjects on defined diets, he had them on the RDA for vitamin B6, and a third the RDA, and then half the RDA. So I hope to get at some of the DNA in his freezer, and we'll assay uracil, and if we see a significant increase in uracil we've nailed it down to vitamin B6. So intervention studies are powerful, so I think we'll doing an end-point of DNA damage, which everybody accepts is important, and then running through all the micronutrients and see which ones are important. But in a way it's easier to just give people a pill, and you don't have to measure all these things, but of course one has to do both.

Just one last thing. People in nutrition are freaked out. They think you should be getting all your vitamins and everything from a good normal diet. But they've been trying for 30 years to get poor people -- and a lot of rich people too -- to eat better, with not much success. So I think you should sell it as insurance: take your multivitamin as insurance, and eat five portions of fruits and vegetables a day too.

Aubrey de Grey: I would like to know much more about these epidemiological studies, because I feel that your strongest chance of turning the nutritionists around -- and turning government around, for that matter -- is to be able to provide genuine data on significantly increased delay or postponement of certain pathologies, and possibly extension of lifespan as well.

Bruce Ames: That's coming. Let me say something about this thing. There's a polymorphism in the enzyme for converting this pool of methylene tetrahydrofolate to methyltetrahydrofolate. It's done by an enzyme called methylenetetrahydrofolate reductase. 40% of Brits and Northern people are defective in that enzyme. What that does is it makes this pool smaller, which puts you at more risk of heart disease, and makes this pool larger, which puts you less at risk of chromosome breaks.

Aubrey de Grey: And do we see that? -- do we see a correlation between levels of that enzyme and death from those two causes?

Bruce Ames: Yes. They've done this on colon cancer and heart disease, and what ??? Smith did is he went to Britain and looked at AOL and AMM, two types of leukemia, and the AOL patients were low for this polymorphism, which suggests that in fact what was causing the AOL was insufficient folate.

Aubrey de Grey: OK: so in a way, this sort of dietary intervention counts as retardation of aging, and in a way it counts as reversal, because of course what one is doing is reversing the gradual increase in the steady-state level of, for example, uracil in DNA -- reducing it to a more youthful level - by suitable supplementation.

Bruce Ames: Yes. So in a way, what we want to think about is that when you drive your car, the amount of oil you need in your car is not the amount to get out of the garage, it's the amount for the lifetime of the engine. And the same thing for all of these: we want to know what the optimum level of each of these is. And we might as well get the whole world supplemented, because it's so cheap to bring the whole world up to snuff, that's what I want to do.

Aubrey de Grey: So, time for some speculation. What's the upper limit on this, in your view? How much could we further increase average lifespan in America by the absolute ideal dietary supplementation, everyone taking a multi-vitamin of your choice?

Bruce Ames: Well, it depends on how much of this fruit and vegetable effect -- that's double, between two and three times more cancer in the bottom quarter. And this is the proportions of other ones. We're not talking about big numbers, but that's where you're getting the vitamins and minerals. Now, you get some zinc from meat, also vitamin B12; vegetarians tend to be very low in vitamin B12, so it's all about a balanced diet. There may be other things we can supplement, antioxidants, things like that, but I think a good part is the vitamins and minerals. So I think this could be a big effect on everything from brain function to immune function, etc.

Aubrey de Grey: OK, back to these epidemiological studies you said you were doing in the Hispanic and black community in Oakland.

Bruce Ames: All we're doing is can we get compliance -- can we talk poor people into taking control of their own lives and taking a pill.

Roger McCarter: How do you know they're taking the pills every day?

Bruce Ames: Well, you can measure ascorbate in urine, or things like that. It's not an easy problem. In the third world, in poor countries, they supplement food and the kids do better at school, everything's better, but here, the poor are doing themselves in by eating all this high glycemic carbohydrate. High glycemic carbohydrate gives you insulin resistance and type II diabetes, and that's really bad. That may do in your mitochondria just as badly as iron. So how do you get people to not eat so much? I actually have a theory about this. If you take a rat and starve it of iron, or starve it of zinc or some essential nutrient, they start to eat more. Maybe there's a feedback to satiety, and they're eating to try to find that missing ingredient, because nature would rather have you fat and fertile than thin and non-fertile. So anyway, whether that's true I don't know, but we're going to try and test that. But one thing is that high glycemic carbohydrate does seem to be bad. A lot of the top people are saying now that we should get rid of the potatoes and the bread and the rice. So that's a lot of work to do to motivate people, but I think it's an easier problem than getting them to change their diet.

Gregory Stock: Aubrey was asking you what the larger impact might be. You also said that for majorities of the population, certainly in the developed world, probably a quarter of the population is somewhere deficient.

Bruce Ames: I can say it's the bottom third that's really in trouble, and the middle third maybe you'd do a little bit of good, and the top third are probably doing fine.

Gregory Stock: So the effect would probably be relatively minimal in terms of changing overall aging demographics or anything like that?

Aubrey de Grey: Well, the way that most people talk about aging is with regard to *maximum* lifespan, of course, and I think that may not necessarily be very fair. I mean, certainly what Bruce is talking about mainly concerns increase in the *average* lifespan, with, as you say, significantly less scope in increasing maximum lifespan because the top third of us are already doing most things pretty much right.

Bruce Ames: So if we can get people to stop smoking, stop eating this high glycemic carbohydrate, and take a multivitamin, I think we're going to move everything up to where the top people are. The Seventh Day Adventists have half the cancer rate and live two or three years longer, and they eat a very good diet, they don't drink, they don't smoke, but they die of boredom.

Roger McCarter: Is it possible that this could be self-defeating? (I'm just trying to play devil's advocate here.) For instance, is the beta polymerase gene regulated by demand, in the sense that if there are fewer breaks, would the repair processes be down-regulated? Is it inducible?

Bruce Ames: Yes. All our defense enzymes are inducible. So we're buffered.

Roger McCarter: So if you down-regulate the damage, wouldn't you automatically down-regulate the protective mechanisms? And so actually the real thing is the balance between damage and repair, and that may not be altered. Is that reasonable or not?

Bruce Ames: It's a good question -- I mean, should we all be getting a little radiation that would induce our antioxidant defenses, and all these things? I think we could argue about this.

Aubrey de Grey: Well, we got to a quarter to four before anybody said "hormesis"! Thanks very much Bruce, that was very interesting, and no doubt we'll come back to many of those points later on.

So: one of the reasons why I haven't been too strict about people's timing is because I'm on for this session and the next session and I really *will* take only five minutes, or thereabouts. OK, so this session is going to be basically on mitochondria. I'm going to talk about mitochondrial DNA, and the role it may have, and in particular how to do something about the accumulation of mutations in the mitochondrial DNA, and Bruce is going to say a few words about non-DNA aspects of mitochondrial damage. So this is a just very brief summary slide of stuff that I'm sure all of you know very well: that the idea that free radicals matter during aging has been around for a very long time, and Denham Harman suggested in 1972 that mitochondria might have the main role in this, might be the main mediators of aging due to free radicals, by virtue of being the main site of production and therefore the main site of free radical damage. There was a big problem with this, which was pointed out by Alex Comfort in 1974, and that is that even in a postmitotic cell, the mitochondria are constantly being recycled, and therefore naturally one would expect that any mitochondria which suffer a mutational event will be eliminated and replaced by proliferation of non-mutant mitochondria and therefore one would have a steady-state level of mitochondrial mutant load. This challenge was very robust and very important, and it didn't get nearly enough airtime as far as I can tell -- during the next 20 years or so, actually. Various theories were come up with to explain how mitochondria might accumulate mutations within cells, particularly positive feedback loops involving mutant mitochondria making more free radicals than wild-type mitochondria and therefore causing other mutations in nearby mitochondria. These theories were really not terribly well thought out, as far as I can tell, because for example they certainly never took into account the possibility of selective advantage, preferential elimination of the mutant mitochondria, that Comfort had suggested. And really nothing changed until 1993, when Josef Muller-Hocker showed that, for whatever reason, Comfort was wrong -- that in fact it's the *mutant* mitochondria that have a selective advantage in non-dividing cells. They get clonally expanded. And this was, among other things, a complete refutation of the idea that there is a positive feedback loop causing a cascade of independent mutations: what this showed is that in fact you have a rather rare mutational event, followed by a selection process that amplifies copies of that original mutation, to take over the cell at the expense of the wild-type mitochondrial DNA, and you *don't* get significant subsequent occurrence of further mutations in the same cell.

So that was all very well, but the next question was, does it really matter?

Gregory Stock: Why is there a selective advantage, before you go on?

Aubrey de Grey: A fair question. Nobody had a clue for a long time. People put forward a number of rather interesting hypotheses that didn't really fit the facts, such as for example that there are mutations that cause large deletions of the mitochondrial DNA, so maybe they have a replicative advantage by virtue of being smaller and therefore easier to replicate more quickly. It turns out that this doesn't work, because first of all mitochondrial DNA replication isn't nearly frequent enough for that to be limiting, and secondly because point mutations are also now well known to be expanded.

Bruce Ames: Mitochondria are putting out oxidants, and if what the lysosome detects is malondialdehyde on the surface, you would expect at least that class to be taken out.

Aubrey de Grey: Right -- so I put forward a mechanism about 3-4 years ago now, which essentially says that the reason why mitochondria with a mutation have a selective advantage is actually because they put out *less* free radicals than wild-type mitochondria, and therefore damage their own membranes more slowly, and therefore are less frequently engulfed by the lysosomal phagocytosis machinery.

Judith Campisi: Why would they put out less?

Aubrey de Grey: Right, good question. The whole idea that they might put out more free radicals was proposed in 1990 by analogy with experiments that had been done mainly in the 70s, using antibiotics. If you take certain antibiotics and throw them at intact mitochondria with a genetically intact respiratory chain, you can certainly get a very dramatic increase in the rate of free radical production. And these antibiotics do certainly block electron transport along the respiratory chain. So, by analogy, it was reasoned that maybe a mutation that blocks the electron transport chain would have the same effect. But of course that depends on the mutation. If you have a mutation that doesn't actually even allow *assembly* of the mitochondrial electron transport chain -- you could for example eliminate a tRNA and therefore you don't make any of the mitochondrially-encoded proteins -- then you wouldn't expect the same effect at all. And in fact it's very remarkable that this was overlooked by the community in general, because the original paper that first formally outlined the vicious cycle hypothesis (of mutations causing a higher production of superoxide) did actually have this caveat very explicitly in it, and pointed out that it had already been shown that Complex III, for example, doesn't actually assemble at all in the absence of cytochrome b, which happens to be the only mitochondrially-encoded subunit. That's very obvious indeed now that we have the crystal structure of Complex III, because cytochrome b is the only truly transmembrane component, so it's got to be there before you can do anything. Complex I is a very great deal more confused at this point, because it's a much more complicated enzyme and we don't have the structure, but it does seem very likely that it won't assemble right if you don't have any of the seven mitochondrially-encoded subunits.

Bruce Ames: There's another thing that could explain having clusters of particular mutations, and that is if you have stem mitochondria. So if you have a stem mitochondrion, and you hit the stem cell, then that takes over. So that, to me, would be a more plausible explanation.

Aubrey de Grey: Perhaps. Certainly it's been shown in yeast -- which is of course the first place in which clonal amplification of mitochondrial mutations was shown, as the phenomenon of suppressive petites, found in the 50s by Ephrussi, -- that in general, petite mutations do not have a replicative advantage. They have clonal expansion, but it's not driven by faster DNA replication of the mutant species. So that argues for slower degradation instead of more rapid replication. But again, as you say, stem mitochondria are an idea that could potentially give a much more clonal result.

But I don't want to talk much more about that, because I don't really care how this happens. What I care about is, first of all, does it actually matter, and secondly, whether or not it matters, what can we do about it? So the question of whether it matters or not at all is very profound, because it turns out that very very few cells are taken over in this way. If you examine tissue histochemically, for example, or immunocytochemically, to see whether there is activity or presence of the mitochondrially-encoded subunits of the respiratory chain, you find that in more or less all tissues, the level of mitochondrial mutations is down well below 1% -- 0.1%, or something like that. And you simply can't ascribe bioenergetic deficiency to that sort of level of mutations. One has to come up with something that involves some sort of active toxicity. One has to propose mechanisms whereby these cells which are taken over by mitochondrial mutations are actually actively poisoning their neighborhood -- or poisoning the circulation, and therefore the entire body. And so I put forward a theory a few years ago now which essentially said exactly this, that these cells which have become completely anaerobic, due to these mutations, survive by some biochemical acrobatics that involves essentially recycling of NADH, which would normally be recycled at Complex I, instead recycling it at the plasma membrane. There's now known to be an enzyme which seems to be present in all cells -- very suspicious, how ubiquitous it is -- that can oxidize cytosolic NADH and give the electrons to something outside the cell, such as ferricyanide, which can't enter the cell. This has been shown in vitro a number of times now. And this mechanism is potentially able to allow cells not only to maintain glycolysis, but also to maintain even the TCA cycle, without running out of anything, because it can get rid of these electrons, and thereby recycle NADH to NAD+ for reuse in intermediate metabolism, even in the complete absence of a respiratory chain. And there are a few other things that have to be done in order to make this work, but it all seems to be reasonably plausible. Certainly a number of preliminary tests of this hypothesis have now been undertaken and it does seem to be holding up so far.

Now, the idea is that these electrons, when you have these cells that are bristling with electrons, because they've got to get rid of them very fast in order to survive and maintain intermediate metabolism, these electrons are not picked up by any nice antioxidant technique, because there are not appropriate antioxidants around in the extracellular medium -- for example, almost all the ascorbate is in the reduced state, where it can't take electrons -- and therefore it is quite possible that molecular oxygen in the extracellular environment is an acceptor of these electrons, on the surface of these anaerobic cells, with the result that superoxide is in danger of being made in the extracellular space. And there are many areas in the extracellular space in which the level of extracellular superoxide dismutase is very low, in particular in muscle. So there is the potential for this superoxide to do nasty things: if it gets near transition metals, for example, it can initiate peroxidation chain reactions in circulating material such as lipoproteins. That therefore gives the possibility not only for transmission of oxidative stress from an anaerobic cell to neighboring cells: there is also the opportunity for dramatic amplification of that toxicity. It means that only a small number of cells, which are really effectively throwing out electrons, can actually initiate very profoundly toxic lipid peroxidation chain reactions which are themselves amplifying the problem. Because when the circulating material, LDL for example, is affected by this, it gets imported into other cells, including into mitochondrially healthy cells, which is of course most of our cells, where it can do harm. So the theory I put forward for this effect was that this happens at an increasing rate as we get older and as we accumulate more and more of these anaerobic cells. Even though they still remain at a very small proportion of the total number of cells. With the result that the circulating material becomes more and more contaminated, and therefore the average degree of contamination of the stuff that normal cells are taking in -- LDL in particular, because that's what they need for their cholesterol supply -- becomes more. And so cells that are perfectly healthy at the level of the mitochondrial DNA will nevertheless be experiencing greater oxidative stress.

Really I'm more interested in testing this in the most direct way possible -- namely by stopping it. It would be very straightforward to avoid this problem if we didn't have mitochondrial DNA in the first place. And it's clear that evolution has done its best -- done really rather well -- in eliminating mitochondrial DNA, because at the time of the original endosymbiotic event mitochondria certainly had upwards of 1000 genes, and they've only got 13 protein-coding genes left. And in some plants, they've got even less than that. The question is, could we complete the process? And a few ways have been put forward to try to do something about that. The first two ideas up here are really alternatives to that, which have been explored in a number of labs. The idea of getting rid of these cells that are overtaken by mitochondrial DNA mutations can be addressed in all of these three ways. The top way is basically, before they've been completely taken over, while there is still some wild-type mitochondrial DNA left, try to inhibit the replication of the mutant mitochondrial species and thereby get the wild-type mitochondrial DNA to take over again. It's not really applicable to aging in the form that it's been looked at so far, because the way it's been looked at is to supply a sequence that is complementary to a mutant species and therefore get rid of that mutant species. But unfortunately of course we get different mutant species in different cells, so it would require too many constructs. The other possibility that's been looked at quite hard is to get new copies of the mitochondrial DNA into mitochondria, to complement the mutations, of any loss of function that happened. And that would be fine, except for a number of problems. First of all, that's an awful lot of DNA. We have a lot of mitochondria, and getting a copy of the mitochondrial DNA -- or of any bits that were necessary -- into all the mitochondria in a cell, or even into a respectable proportion of them, that's a hell of a lot of delivery. And that's rather an important problem, because a certain amount of delivery might be enough to rescue oxidative phosphorylation in the cell, but then you have a situation similar to what you had before the mutant mitochondrial DNA had taken over the cell in the first place. The successfully transfected mitochondria would be at a selective disadvantage, and they'll be eliminated as fast as they're transfected. So I don't think that's going to work either. I'm much keener on *this* idea, completing the process that evolution has left incomplete, namely transferring -- or rather putting in copies of -- the protein-coding mitochondrial genes into the nucleus, suitably modified so that they can work, so that the proteins will go back into mitochondria and be assembled into the oxidative phosphorylation enzyme complexes, just as they are when they're constructed on mitochondrial ribosomes. And this is something which I've been pushing for a little while. I certainly wasn't the first person to think of it: it was first suggested in 1990 in a brief review in Cell by Eric Lander and Harvey Lodish.

There are many reasons why it might not work. (This is a graphical summary of the previous slide, basically showing the various different ways that one might go about fixing the effects of mitochondrial mutations, by the three pathways that I explained before.) There are various reasons why this has been supposed not to work. Some of them are really not thought out at all: for example people said, well, hang on, there are mutations that would affect the tRNAs and the ribosomal RNAs that are encoded in the mitochondrial DNA, and this of course completely overlooks the fact that if you're getting all the proteins in then the tRNAs and rRNAs are completely redundant, they have no function whatsoever over and above the synthesis of these proteins that's not going to be necessary any more. I won't go through all the other reasons, because they basically don't really work -- they're mainly based on the perceived requirement for a reason why evolution hasn't finished the job.

Bruce Ames: How big is yeast mitochondrial DNA compared to human?

Aubrey de Grey: It's about five times as big, but it's only got about the same number of genes.

Bruce Ames: So it can't be very easy for evolution to get rid of these things. Maybe you want them. If you take out the bad mitochondria with lysosomes, maybe that's a good thing.

Aubrey de Grey: This has been said, yes. But we don't need to suggest such things, because first of all is the fundamental problem that the mitochondrial DNA in yeast and in animals has a different genetic code than in the nucleus. It didn't have a different genetic code at the time of the endosymbiotic event, but after nearly all the protein-coding genes had been transferred, it became less traumatic to have drift in the genetic code, and that's what happened. There are now four differences between the human mitochondrial and nuclear genetic codes, with the result that of course if a chunk of DNA encoding a protein were transferred from the mitochondrial DNA into the nucleus, even if it did happen to land in an appropriate place such that it were attached to a signal sequence resulting in the encoded protein's being taken back into the mitochondria, it still wouldn't have the slightest chance of working because it would have the wrong amino acid sequence. In fact what would actually happen is that it would be truncated, because there are differences in the stop codons. So that's the only reason one needs. And in fact one can support this idea by looking in plants, in which the genetic code is *not* different -- it's the same in plants in the mitochondrion and in the nucleus -- and sure enough, there are cases where plants have succeeded in moving genes into the nucleus which we have failed to move. Now they haven't succeeded all the way, so there's clearly more to it than that. But all of these reasons I've written up here are not the answer, this is the answer. Hydrophobicity. The proteins that are still encoded in the mitochondrial DNA are viciously hydrophobic, and getting these proteins through the mitochondrial membrane, from the cytosol into the matrix, requires that the proteins be completely unfolded, which of course hydrophobic proteins don't like to do. So it turned out to be very easy to get a very very short protein, a 48-amino-acid protein from cerevisiae, expressed in this way, expressed from a nuclear transgene and imported into the mitochondria, and this was done in 1986. And only two years later it was shown that this was actually able to achieve phenotypic rescue of a mutation in the endogenous mitochondrial copy of that gene. So that really worked, and everyone was quite hopeful, and in 1991 they got one further gene in, which was only 76 amino acids, still very short. Since then there's been blindingly little success, except that as Katrina may know there is a paper in review in Science right now, which has succeeded in mammalian cells -- in Chinese hamster ovary cells, in fact -- in getting ATP6, which is a perfectly respectable-sized protein, into cells, and indeed demonstrated phenotypic rescue and growth in the presence of a non-functional mitochondrial copy. So there is no real progress yet, this is very much the first time that any respectable-sized protein has been expressed in this way in mammalian cells. And I very much expect the floodgates to open now, because people will be able to see that it is feasible.

Whether or not all proteins can be got into the mitochondria in this way remains to be seen. Some of them are probably a good deal harder than the one that's just been successfully got in. But the point is that, small or large as the effects of mitochondrial mutations on aging may be, there is now a realistic way of simply eliminating the effects of mitochondrial mutations *completely*, in a model organism like a mouse for example, simply by creating transgenic nuclear copies of these 13 -- a very small number -- 13 mitochondrially-encoded proteins. So there are 13 problems. They're hard problems, they'll need a lot of work, but, you know, it's something that we shouldn't be scared of any more. That's really all I have to say.

Judith Campisi: Is there evidence for anaerobic cells in tissue?

Aubrey de Grey: Absolutely. Sorry -- this is the growth curves from the paper that's in review in Science, so it's really true, the top two are controls in the absence of the thing that makes the mitochondrial copy not work, and the growth curve that tails off horizontally is the one where the nucleus has not been transfected and the one that gets away is the one where transfection has occurred. OK yes, to your question, is there evidence, yes, there's exceptionally good evidence. The paper I mentioned, from Muller-Hocker in 1993, gave absolutely conclusive evidence. What Muller-Hocker showed was, he took slices of muscle, and first of all he did histochemistry, and showed that there were individual muscle fibers that were completely inactive for cytochrome oxidase, which is partly encoded by the mitochondrial DNA. And they were just individual fibers, so the next-door fiber would be perfectly normal. And that looked really convincing. But what he did in 1993 was something even better than that: in situ hybridization experiments. So in other words, take small fragments of the mitochondrial DNA, use them as probes to a sample of muscle for example, and if you can see no signal with your probe, that means that pretty much all the mitochondrial DNA molecules within that cell, that muscle fiber, lack that segment of the mitochondrial DNA. So this was proof not only that these cells are completely devoid of wild-type mitochondrial DNA, it also shows that they are filled up with copies of a molecule that misses the same segment, which is almost certain to be a clonal expansion of one mutation.

Gregory Stock: I don't see how this qualifies as anything relating to reversing of aging. It seems to me it would have to be a germline intervention.

Aubrey de Grey: Not at all. Supposing we had gene therapy in humans, and we were able to get these transgenes into the nucleus of all our cells.

Gregory Stock: That was the supposition that I was having a problem with.

Aubrey de Grey: OK, so let's look at the situation in mice. Supposing we arranged to have all these allotopically expressed genes in the nucleus of mice, introduced by germline transformation, but suppose that we had them under some inducible promoter, so that they were actually silent until old age. And let's suppose that it's true that mitochondrial DNA mutations really matter in aging. Then, if we do the induction, the mouse is getting old, and looking old, it ought to restore the OXPHOS function (oxidative phosphorylation function) in those cells that have lost it, and by that means it would eliminate the spewing of electrons out into the extracellular space, that's going on, and therefore it should eliminate the toxicity of these cells.

Gregory Stock: So you're saying you'd artificially allow the spewing-out to occur, and then turn on these genes which there wouldn't be any reason not to have turned on all the time?

Aubrey de Grey: Not in a mouse, no, but of course if we want to do anything eventually when we do get gene therapy, we'll want to do it on people who haven't got these things, we want to be able to introduce these things into adults.

Gregory Stock: So you're assuming you would have a relatively robust -- to say the least -- gene therapy.

Aubrey de Grey: Oh yes, certainly. The whole thing, in terms of human application, entirely relies on very robust gene therapy, much more robust than we reliably have at the moment. But actually you make a very good point, because it seems to me that a lot of people are concerned about the time it takes to show even a modest increase in maximum lifespan in mice -- two or three years -- and what a relatively small impact it has on people's perceptions of what might actually be possible in humans in the fullness of time. There could be a real boost to the public perception of the importance of this work, of this sort of work -- anti-aging work in general, especially anti-aging work involving transgenics. Maybe there's a lot of scope in making these conditionally-expressed transgenes, and turning them on only in middle to late ages, in mice that have been allowed to age normally because they've had natural gene expression until that time. Because that would be a much more dramatic demonstration of what we could really describe as reversal of aging. One could expect that one would take a group of mice out to where their subsequent life expectancy was only let's say six months, on average, and a maximum life expectancy of a year, from the time that one switched on the transgene, and one would be able to see, therefore, in two years, a doubling of lifespan relative to controls. So this would be a much more dramatic result. Only if it works, of course, but for things that *did* work it would be, I think, rather a good thing. The problem is it's quite hard work to make inducible transgenes in the first place.

PD: The final steps of haem synthesis and iron-sulphur formation are inside the mitochondria. Is there any evidence that you would need to have some subunits translated inside the mitochondria to have proper insertion of the iron-sulphur clusters?

Aubrey de Grey: As far as I can tell there's no evidence for that at all. And in fact there's good reason to believe that that's not the case, because most, or at least a fair proportion, of the nuclear-coded subunits of the respiratory chain complexes are imported in a rather indirect sort of way that involves import right through into the matrix followed by semi-export out into the membrane. So in other words there is a period during which these proteins are in exactly the same compartment that they would be when they're mitochondrially-encoded. So similarly, if we use that pathway to take a transgene from the cytoplasm, then we'd have a good chance.

Julie Anderson: Aubrey, maybe I missed something, but so what was the trick in terms of getting, in this paper you were talking about, this sizeable protein imported from the cytoplasm?

Aubrey de Grey: OK, it's not really clear what these people have done that people didn't do before. You have to put a leader sequence on the front, to direct it, but of course the leader sequence has to work in some way. So, people have tried with this same protein, to get it in with other leader sequences, and they haven't got anything like this. A very important aspect of this, which I should draw your attention to, is that the transfected line, the one that gets away and does really well eventually, actually starts very slowly, and eventually catches up, whereas the line that hasn't been transfected is all right early on. I think that that tells us that these transgenic proteins are actually quite often getting pretty badly stuck on the way through the membranes, and thereby actually inhibiting general protein import of mitochondria. But eventually just enough of this stuff is getting through, and is able to complete successful import, to be able to restore OXPHOS function such that they overwhelm that problem.

Bruce Ames: I'll be the wet blanket. This seems too much pie in the sky to me, because we've had an awfully long time in evolution having our mitochondria adapting to what they want. Why are certain genes working in the mitochondria? There's so much regulatory circuits, so it just doesn't seem anything that's at all conceivable within the next five hundred years. There are other problems that seem more pressing.

Aubrey de Grey: Well, let me address that, because basically what you're saying is "Cells are really clever. And if we try to tell a cell what's good for it, it's going to have strong opinions on the matter and it's going to tell us that we're wrong." That's essentially what you're saying. And I agree with that, completely. But what I'm saying is that this is actually a fine example of *not* trying to tell the cell what's good for it, but rather, giving it a more robust way to do what it wants to do anyway.

Bruce Ames: Well, I think one could try and get particular genes in, and see if one works, try another; I'm just not very optimistic.

Aubrey de Grey: Thirteen is a very small number, though.

Judith Campisi: Let me add a cellular dimension into all this molecular biology. You're saying there's like a field effect, where one anaerobic cell may create -- the critical lynchpin to making all of this worthwhile is the idea that there's one anaerobic cell in a field and it has a field effect, a toxic effect, on the surrounding tissue. What is the evidence? Is there evidence that this is occurring?

Aubrey de Grey: Really only very indirect evidence. I mean, when I put this idea forward, it was only February of 1998, it was completely off the wall, no one had ever thought in this way at all, so there wasn't even really anything that could be considered as preliminary data in favor of it. What we can say now is that there does seem to be evidence that the enzymatic activity that exports electrons is expressed at a higher level in older people. So that does provisionally indicate that this is, at least broadly speaking, the way that these cells are staying alive. The best evidence that we can say that these cells are having an effect on circulating material, and thereby having what you called a field effect, is really negative -- by elimination. Atherosclerosis researchers have had a big problem for a long time, starting in the mid-80s when oxidized LDL was implicated in the biogenesis of atherosclerosis -- in the etiology of atherosclerosis -- a big problem, which is that if you take endothelial cells in vitro and you incubate LDL with them, in a medium that has anything like physiological amounts of vitamin E and other antioxidants, nothing happens. There is no oxidation of that material. Whereas in vivo, as we see, it happens. So people have suggested that maybe once the LDL gets underneath the endothelial cells, into the arterial intima, maybe there's a depletion of some of the important components of the antioxidant arsenal, such as albumin, and maybe that allows LDL oxidation to proceed even though it doesn't proceed in the artery. But now that's been tested quite well as well, because you can extract plasma from that space, and do the experiment with that, and again, nothing happens. So one needs to postulate a microenvironment in which something happens to LDL that allows it to become atherogenic. And nobody has found one yet. So this is my candidate microenvironment.

Bruce Ames: Bringing back stem mitochondria: wouldn't stem mitochondria be another possible explanation for why you get mutations accumulating?

Aubrey de Grey: Sure, they might be. I really don't give a damn how these cells come into being. I care about fixing it.

Judith Campisi: So it seems to me, Aubrey, that this might be practical if you take Attardi's system, right, and you would predict that if you would introduce mitochondria into those cells that have mutations in DNA, and then coculture those cells with LDL or with other cells, you should be able to see damage?

Aubrey de Grey: You're right, this is the sort of idea that --

Judith Campisi: It's a grant proposal that you could write and give to somebody!

Aubrey de Grey: Yes, well that was one of the suggestion I made in my first paper on this, actually, was that one ought to be able to find realistic culture conditions, in which cells would be induced to -- rho-0 cells, ones without mitochondrial DNA -- would be induced to behave appropriately. The difficulty is that it seems that -- well, one thing that does seem clear is that rho-0 cells in culture, which are rapidly dividing, do not seem to do the same sort of thing. First of all they're vastly more glycolytic.

Judith Campisi: That's true of cells in culture in general.

Aubrey de Grey: Right. So, it's tricky to make them do these biochemical tricks in the first place. One thing that was only found out about six years ago was that you don't necessarily have to give rho-0 cells an electron acceptor that can get into the cell, be reduced and come out, the way it's usually done with pyruvate: you can use things like ferricyanide that are cell-impermeant. And this was the first real proof that a plasma membrane electron export system can keep cells alive when they haven't got oxidative phosphorylation. But yes, if one could produce very high levels of this enzyme in vitro, one ought to be able to incubate LDL with those cells and ask what happens.

Gregory Stock: Or another way would be to try to figure out a way to selectively kill those cells.

Aubrey de Grey: Right -- we'd love to do that. Exactly what Judy's trying to do with senescent cells. It's much harder here, I'm afraid. The reason it's much harder is because the cells that are most predominantly affected are muscle, and the worst thing is, which I should have mentioned earlier on, is that it's not the whole fiber that goes down. Because mitochondria are immobile in muscle, the result is that -- or this is how it looks, from the mitochondrial mutations -- is that individual nuclei, or perhaps small numbers of nuclei, maintain individual fiefdoms of the mitochondrial population, so that the selective advantage exists, but it causes a very slow spreading of the mutant species along the fiber. It's been shown very recently and very convincingly by Judd Aiken's group in Wisconsin that the fiber segments that become anaerobic are longer in older individuals than in younger ones, indicating that these fibers are surviving for in the region of a lifetime (and are getting gradually more and more taken over); but then the problem with killing these cells is that if one were to take the obvious approach and find a way of inducing apoptosis in a muscle fiber that had this problem, one would probably end up killing the whole fiber, not just the bit that was anaerobic. And if you do the numbers and you work out how long a segment is as a proportion of the total fiber, and how many times you see one of these anaerobic segments in a transverse section across muscle, then the number you come up with is that there's actually a hefty proportion of fibers, maybe even a quarter or a half, that are affected.

Gregory Stock: What about other tissues than muscle?

Aubrey de Grey: Other tissues haven't been nearly so well studied. There is certainly good evidence from histochemistry that expansion of these mitochondrial species occurs, and cells get into this state, but there's by no means such good evidence that they do anything unusual with their biochemistry. One of the in vivo assays which have been done that show that something really strange happens in muscle is that in the areas which are cytochrome oxidase negative, there's very substantial *up*-regulation of succinate dehydrogenase, which is one of the enzymes of the TCA cycle. And this is one of the first things that led me to suspect that maybe the TCA cycle is being maintained, very paradoxically, in muscle. But this, again, has not been reproduced in vitro.

Gregory Stock: Are you going to go and do the lysosomal stuff?

Aubrey de Grey: Yes, that'll be the next session.

Gregory Stock: I would suggest that you actually do that; then we have a break and wrap up.

AdG: Yes. First of all I want to give Bruce at least a couple of minutes to talk about non-DNA damage to mitochondria.

Bruce Ames: So, six years ago, two of my senior postdocs and I wrote a review on mitochondrial decay and aging -- Mark Shigenaga and Tory Hagen were the postdocs. If you take young rats and old rats and compare the mitochondria, we've been doing that for a long time, so in the review we put together all the evidence suggesting that mitochondrial decay is important in aging. I think the results were reasonably convincing, to us anyway. But experimentally there's a problem, because you can look at a tissue that's not turning over all the time, like the heart or the brain, or the liver, so if you take a liver out of an old rat versus a young rat, the mitochondria are very fragile in the old rats. Tory Hagen in my lab came up with an experimental way to get around this problem. Basically what he does is perfuse the rats, and isolate hepatocytes, and they're not fragile, so you get the same yield from the old rats and you can look at mitochondrial function in the whole hepatocyte. And what he found is that if you take the old rat and compare it to the young rat, the membrane potential is less. The cardiolipin (that's a lipid you need for the mitochondrial enzymes) is down, oxidative respiration is way down, and oxidant leakage is much higher. So we can measure all those things in old mitochondria compared to young mitochondria.

And then he figured out how to reverse it. So he got onto this because the first thing we tried, we were reading the work of Paradies and Gadaleta -- there's a mitochondrial group in Bari, Italy, and they worked with an Italian company that makes carnitine, the acetyl ester of it. Carnitine is a transported that brings fatty acids into the mitochondria for fuel. It's imported by acylcarnitine transferase. And that enzyme is known to go down with age. And the Italians fed acetyl carnitine to old rats and measured mitochondrial transcription. So we fed our old rats acetyl carnitine in the drinking water of these old rats for a month, and the membrane potential came back, but oxidant leakage was no different -- if anything it was worse. So we seemed to have solved a lot of the problems, and the first thing we found that solved the oxidant problem was PBN. PBN is a spin trap, and Carney and Floyd had spent years using gerbils, and had lowered the oxidized protein in gerbils down to the level of young animals. And they formed a company in San Francisco called Centaur, got 100 analogues of PBN and try to get some sort of treatment for Alzheimer's and so on. So, when we tried feeding PBN to these rats, it lowered the level of oxidants. So ALCAR plus PBN solved all the problems. So that was our first step, and since then -- I'll get on to lipoic acid in a minute, I'll just say one more thing about PBN. One of my former postdocs showed that when you give PBN to human primary cells in culture, it made them take more generations before they senesce. And Hani, who's here, was repeating that experiment and he found that somehow it seemed a lot more active than it should be, more active than she had described in the first paper. And it was a year-old solution of PBN, so he made up a fresh solution and compared the two, and the old solution was much more active. So he tracked it down to a breakdown product of PBN, a hydroxylamine, and it's 100 times more active than PBN in a tissue culture system and it's terrific in old animals -- whatever PBN does, this is better. So we think PBN is just a prodrug. And he tried ten hydroxylamine derivatives and they all worked, and they're very non-toxic molecules.

So that's that; but we dropped PBN and switched to lipoic acid, because you can buy lipoic acid in the health food store. And as you know it's the coenzyme for pyruvate dehydrogenase and ketoglutarate dehydrogenase. So Lester Packer in our department has been using lipoic acid for years and says it's a wonderful antioxidant. You actually feed the oxidized form. So we decided now we'd try lipoic acid, and it works beautifully.

So I'll also discuss why I think these things are working. What we measure is ambulatory activity. You have a video camera spying on the rats at night, young rats are frisky and old rats are very lethargic, and either acetyl carnitine or lipoate makes them a lot friskier. They sell acetyl carnitine in Italy as a pick-me-up pill -- you can just buy it in any drugstore. And they sell lipoic acid as a diabetes drug all over the world. We also looked at the Morris water maze, which is a test of spatial mapping, and again either lipoic acid or acetyl carnitine seems to make old rats a little better. So that's what we've been able to measure now. Glutathione goes down with aging, so we're asking what are the effects on that. Ascorbate goes down with age, and again it's better.

Now, why does this work? And I've got a big elaborate theory, which I think is likely to be right, but at least it's a theory.

Christopher Heward: Can I ask one question? I remember you mentioning at a meeting recently that you were going to be doing the longevity experiments at that time or were getting ready to do them. Were there any preliminary experiments that didn't come out?

Bruce Ames: No. We've done dose-response stuff, and we're ready to do it. We've had trouble getting the lipoic acid from this company in Germany, and we had trouble getting the rats from NIA, and there've been various pitfalls, but it's due to start.

So: Lipoic acid sold in a health food store is a synthetic mixture, a racemic mixture. And R is the natural form and S is some unnatural one. R is reduced in mitochondria, S is reduced in the cytoplasm. And in our hands R works and S doesn't. So why does this work? I'd just like briefly to tell you about this review I'm writing, which gets thicker and thicker. This review started 20 years ago when I was teaching a course in bacterial genetics, and I had the students isolate mutants that would grow on a complex medium and not on minimal medium, and each one had a different set of mutants and they had to track down what gene was mutated. A little exercise in detective work in biochemistry. A very common class of mutants are what I call Km mutants. And we published one 20 years ago. Threonine goes to alpha-ketoglutarate goes on to isoleucine. So we convert the amino acid threonine to the amino acid isoleucine. If you mutate this gene, you have to add isoleucine to the medium, because the bacteria then can't make isoleucine. But this is a B6 enzyme. About a quarter or a third of the mutants turn out to grow either on B6 or on isoleucine. So the interpretation is that we make a mutation in which you affect the structure or function, sometimes you're affecting the Km, and you can get the activity back by raising the level of your cofactor. And the protein chemists say that these should be very common. So I looked in the literature and found a few cases and I decided to write a review, but I got side-tracked. And I got back to it again with aging, because I think the same thing is happening in aging. You're oxidizing proteins, and what happens is you deform the structure and then you lose activity; you can get some of it back if it's a Km mutant. So if the acylcarnitine transferase doesn't work very well you can get some of that activity back by adding more substrate. So that says that a lot of the decay might be reversible by higher levels of certain micronutrients.

Aubrey de Grey: But hang on. You're suggesting that the Km of various proteins, especially in the mitochondria, is diminished during aging not through a mutational event but through oxidation. But then these proteins are being recycled all the time. So new copies that haven't been oxidized --

Bruce Ames: But the level of oxidized protein goes up with age in mitochondria. And membranes are getting stiffer with age.

Aubrey de Grey: So what's driving it? If new proteins are being made all the time?

Bruce Ames: What's driving it is, we can measure that more and more oxidants are coming out of the mitochondria. Now, why? Is it DNA mutation, or is it as Ulf Brunk says, the lysosomes getting clogged up with oxidative material and iron and that's getting fouled up and the whole cleanup system is degenerating? I suspect it's an interaction between these, it's not just one. But what's clear is you make more oxidants with aging, you have a higher level of oxidized proteins, a higher level of oxidized DNA and a higher level of lipid aldehydes.

Aubrey de Grey: I think we're all happy that the steady-state level of oxidative damage is a lot higher, and I really like your proposal that a large part of why this is bad for us is reduction in Km. But I think it's really important, in trying to decide how to do something about that, to do more than simply take things back the very little amount that one could do by increasing the amount of substrate of the various things. So if one wants to do that then one has to go further back to how this is going on. You've worked on hepatocytes and you've found this very pronounced effect; hepatocytes don't accumulate anything in their lysosomes, because they get recycled too fast.

Bruce Ames: Well, now Tory's doing the heart and we're doing brain. And in this review I found 50 cases, 50 human diseases where a third of the people get better when you feed them high pyridoxamine, or in other cases they go blind and a third or a quarter of the patients get better if you feed them high thiamin or riboflavin. There's just a long list of B vitamins that if you feed high levels, they get better in all these different diseases. People have shown it's Km in certain cases, and it all makes sense. And if you look at aging, people have shown that in Complex I and Complex IV the Km for substrates goes down with age. With all kinds of enzymes the activity goes down with age, yet the enzymes are there, it's probably that they're being oxidized.

Aubrey de Grey: OK, thanks. Let's break for five or ten minutes.

....

Aubrey de Grey: ... which has a profound physiological effect in relation to the age-related accumulation of protein-protein cross-links in the extracellular medium. In a variety of different models it's now been established that this happens. The first one that was looked at in detail was rat tendon, which gets steadily stiffer during aging, and this was reversed. That was obviously a symptom that doesn't kill the rat, in general, but it was something that could easily be tested. But now there's plenty of good data on heart function in dogs, and so on, it's really an amazing thing. And the reason it's possible to identify a small molecule which can have this profound physiological effect, with as far as we can see no side-effects to speak of whatever, is because the molecular nature of most of the cross-links that are formed in the extracellular medium, as a result of glucose-induced glycation damage, these cross-links are rather unusual. They have this sort of structure -- you have a protein, another protein, and you end up with a short carbon chain across, with two double-bonded oxygens on consecutive carbons, a diketone moiety. And these molecules can come along and essentially cleave, here, and most of the agent gets attached to one side, and something else gets attached to the other side, so the result is you have hydrolysis of the bond. And it doesn't look much like it did before glucose came along in the first place, but that seems to be all right, because you end up with no cross-linking and therefore the physical properties of the material that was undergoing this random cross-linking in the first place are very substantially restored as a result of this. And I'm extremely optimistic about the therapeutic value of this compound. They've had a good deal of trouble actually turning it into something commercially available, mainly in respect of getting investment from big pharma in order to get it through the various stages of trials, Phase III trials have yet to be secured. But I understand that the problems they've had are entirely based on the fact that there is perceived to be doubt about the robustness of the patent, as opposed to any doubts about the pharmacological efficacy.

Gregory Stock: What's the name of the chemical?

Aubrey de Grey: It's called ALT-711. Yes, the problem is that the concept behind the chemistry is rather simple, so it's perceived that one could design chemicals which are outside the boundary of the patent but nevertheless work in essentially the same way, and this is a problem with getting pharmaceutical companies to invest the appropriate amount of money. But they're gradually getting enough money to get the thing through; Phase II trials are happening in Europe right now, and maybe it'll happen.

But for the rest of this session I want to talk not so much about extracellular problems based on accumulation of undegradable and generally deleterious cross-linked substances, but about the same thing that happens *within* cells. I've begun to get really interested in lysosomes. They seem to get rather sidelined and ignored a great deal -- hardly anyone seems to take any notice of them. But they're incredibly important for us, because they do all the hard parts of recycling of things. We have a number of other processes, of course, which are involved heavily in recycling of proteins and so on, particularly the proteasome, which has been getting a lot of interest recently, but ultimately the lysosomes are the sort of back-stop. They're the things that can degrade and recycle pretty much anything that was unsuccessfully recycled by more specialist machinery in the cell. And unfortunately, even lysosomes can't do absolutely everything. It's been discussed in a number of contexts that perhaps the junk that accumulates within lysosomes due to its undegradability in postmitotic cells may have something to do with aging. This stuff's called lipofuscin or age pigment; it's fluorescent, if you shine blue light at it glows red. But really lipofuscin is not by any means the most highly implicated lysosomal material in age-related pathology. Something that I'm very interested in is what happens in macrophages in the artery. We talked briefly about atherosclerosis in other ways earlier on during the day. Various steps in the chain of event that lead to atherosclerosis have been addressed as potential therapeutic avenues; one, of course, is to reduce the steady-state level of oxidation of low-density lipoprotein, which seems to be a very clear causative factor in the accumulation of fatty streaks and the subsequent atherosclerotic lesion. And most of the other interventions into atherosclerosis are much later on in the chain of events, so for example trying to diminish the tendency of vascular smooth muscle cells to proliferate and increase the size of the lesion, and then later on than that, trying to stop the lesions from bursting and spewing out nasty chunks of cholesterol that get stuck in your brain and give you strokes for example. Also trying to open up the cross-sectional area of an artery that's affected and become almost blocked due to growth of the lesion. These very late-acting therapies are really pretty short-termist and they don't work for very long, they have a very large rate of recurrence of problems such as restenosis. One thing that hasn't been looked at all is trying to stop the fatty streaks, the macrophages that become the beginning of the atherosclerotic lesion, to stop them forming.

We know that macrophages invade the space underneath the endothelial cells, the intimal space, of arteries -- on purpose, in some sense, in a physiological way, in order to get rid of contaminated lipoproteins that are stuck there. And they need to do this very well. But these contaminated lipoproteins contain some very unlikely and unusual cross-linked material, oxidized cholesterol in particular, and oxidized cholesteryl ester, and these materials seem to be very refractory, or at least seem to a significant extent to be very refractory, to degradation within lysosomes. So that's what happens in the arterial intima: the macrophages proceed perfectly well breaking down the LDL that they're taking up, except that a small proportion of that LDL is not broken down but accumulates in lysosomes. And it doesn't look anything like lipofuscin, this stuff -- it's a sort of toothpaste- textured white material, it's not fluorescent or anything -- but that doesn't matter, the main thing is that macrophages can't break it down. And so eventually their lysosomes just end up filling up the cell, and the macrophage gives up the ghost and becomes toxic. And that's the start of the fatty streak, and thence the atherosclerotic lesion. So if macrophages were any better than they are now at breaking down the stuff that they're taking up, then this would happen much more slowly -- potentially *very* much more slowly. So this is, in other words, a rather similar problem to the problem of accumulation of age pigment or lipofuscin in postmitotic cells. One area which has already been mentioned today and accepted for a long time as being quite clearly a lysosomal event is macular degeneration, which is the accumulation in lysosomes of the pigmented epithelium in the retina of essentially photodamaged rhodopsin, which is recycled at a phenomenal rate because the damage that rhodopsin undergoes as a result of exposure to light is very rapid. And so the lysosomes in the pigmented epithelium, Bruch's membrane, accumulate this particular protein, rhodopsin, in damaged form, at a really phenomenal rate, and their lysosomes get very much taken over, with the result that the cell starts to lose function in a very clear way. It's clear that this is a primary causative mechanism of age-related macular degeneration. So there are really a lot of areas in which lysosomes can be not quite complete, and where the very good, but not quite good enough, ability of lysosomes to break stuff down, heterogeneous stuff, is clearly implicated in important age-related pathology. It's been quite difficult to demonstrate in vivo that lipofuscin per se matters, but it has been explored in vitro by Ulf Brunk, as Bruce mentioned not long ago. Lysosomes are made to accumulate junk at a much greater than physological rate in vitro, in cultured cardiac myocytes, which are the cells that Ulf has used, and it's clear that the cells with lipofuscin- loaded lysosomes are a lot less happy than they were when the lysosomes did not contain any age pigment. And so the problem is the function of the lysosome, which is impaired in its functionality as a result of this. One thing I did mention has been shown in microglia, which are of course the macrophage of the brain, effectively: they have a tendency to try to take up, and to some extent succeed in taking up, Abeta plaques, which have been more often talked about as being extracellular. Julie mentioned earlier on today that maybe Alzheimer's plaques are potentially beneficial in retarding damage caused by the primary defect in Alzheimer's disease, but insofar as the stuff gets taken up by microglia, that may indicate that that's not correct.

So the question is, what might we be able to do about it? And no one to my knowledge has really worked very much on this, on trying to identify a way to actually reduce and reverse the accumulation of junk in lysosomes. One problem is that there are differences in age-related pathology in the retina and in the arterial situation: clearly the nature of the stuff is very specific for the cell. In postmitotic cells such as muscle and motor neurons, there's pretty good evidence that a large source of what's taken up, or at least of what's taken up and then fails to be degraded, is mitochondrial in origin.

Bruce Ames: Would it make sense to have this system unless the lysosomes are taking out the more damaged mitochondria?

Aubrey de Grey: Absolutely. And of course the model that I put forward in '97 was that that's exactly what lysosomes are doing and it's how mitochondrial turnover is driven, except that the ones that are more damaged are actually the wild-type ones not the mutant ones. The mitochondria that are damaged are those that produce more free radicals. In a given cell. Of course that doesn't mean to say that's what's happening in a young cell versus an old cell, because that's the wrong comparison to make.

So anyway: the question is what to do about it? And it does look pretty ambitious, but a year or so ago I came up with a very crazy idea. If Bruce thinks it'll take 500 years to get 13 proteins into the nucleus, then I wonder what he's going think of this idea. This is not my question, this is a question posed by a collaborator of mine. The fact is, graveyards don't fluoresce. They're not full of lipofuscin, even though they're full of people. So something -- something in the soil, presumably -- is better at breaking down lipofuscin than we are. And in fact that's not so surprising once one looks at the literature and talks to specialists about this. Luckily, by a very great stroke of luck, I happen to have a colleague in my department who works on Rhodococcus and other proteobacteria, the Nocardioforms, and these are very common in the soil bacteria, and they break down absolutely everything. It's completely incredible what they can break down. The commercial application that's been explored by this guy and his colleagues in my department, for a few years now, is that they can break down toluene, or even trinitrotoluene. So they have commercial applications in detoxifying contaminated areas of land for housing developments, things like that. It's completely extraordinary. So, I mean, the question of course, the implication of the question, is that maybe these same bacteria have the potential to break down lipofuscin, and of course not only lipofuscin but the other types of undegradable material that we talked about earlier in these other cell types.

Bruce Ames: Trees and fungi have all this cross-linked material, and they're broken down.

Aubrey de Grey: Absolutely. So in other words the diversity of very elementary organisms is completely amazing. I should point out that it is diversity -- it's not just that you can get a particular bacterium, a particular strain of bacteria, which can break down absolutely everything. What you get is an enormous range of functionality, of hydrolytic diversity, in the soil. So that whatever you need to break down, there's going to be something that can do it, but there's also going to be masses of things that can't do it.

Bruce Ames: There's a problem with metals, though: so one is cross-linked proteins, things like that, the other is are they pulling out the iron that we need.

Aubrey de Grey: Sure -- but young lysosomes have no trouble releasing, in a controlled way, the iron that they have taken up and extracted from proteins that they're breaking down, for example. So eventually they get sequestered within lipofuscin, which is potentially problematic. That's what the authors propose. So yes, that's a very important point. If the proteins and lipids and so on, that are so heavily cross-linked that none of the hydrolytic enzymes in our arsenal can break them down, if those things could be broken down, then we might not have a problem with the accumulation of this sort of junk in all these various cell types. And that might have extremely pleiotropic beneficial effects in reversing age-related pathology. So the question then is, the obvious idea is, well, if particular bacteria seem to be able to do this, then could we identify the enzymes that have this functionality? And then we could make transgenes and target the proteins to lysosomes and see what happens.

Of course there are potential pitfalls in this. The first pitfall is that we might end up with these proteins eating the lysosomes, from the inside, that's the most obvious one. But actually it seems quite unlikely that that would be a real problem, because lysosomes already have a bit of a difficulty in not being eaten from the inside, which they succeed in avoiding very effectively, and the way they do it is by having extremely specialized membranes. The proteins in lysosomal membranes are fantastically glycosylated, which is clearly protecting them from lysosomal hydrolytic activity within the lysosome. The membrane itself has one or two very unusual lysophopspholipids that have their fatty acid tails in an extraordinary orientation. So there are very specialized things that are specifically designed not to be attackable by any of the enzymes that the lysosome contains, thereby to maintain lysosomal integrity. And the point, of course, is that when one isolates a single enzyme from, let's say, a bacterium, then it's going to have pretty specific activity. It's not going to be any less specific if you get it from a bacterium than if you get it from a mammal. So it should be very straightforward indeed to find enzymes that can break down the molecular structures that are prevalent in lipofuscin or oxysterols or whatever, without having any activity against the lysosomal membrane, and therefore it may turn out not to be much of a problem after all.

Judith Campisi: Aubrey, what do we know about the chemistry of these lysosomal gels and proteins? Maybe we'd need thirty enzymes to deal with lipofuscin alone, and more for the other things.

Aubrey de Grey: Well, if you needed in the region of thirty enzymes I might be worried. I'm inclined to suspect that you'd need in the region of half a dozen.

Judith Campisi: Why?

Aubrey de Grey: Because the stuff is very heterogeneous, this is in the nature of lipid peroxidation and protein oxidation that you're bound to get pretty much random cross-links, but you don't need to break all of them in order to disaggregate these granules that accumulate in these various cell types. Basically you need to break down enough of the knottedness of something to allow access to molecular structures that we've already got enzymes for. So my completely subjective thinking, because of course it would have to be tried, my hunch is that at the moment we have a hell of a lot of different things to break down, and we probably have only about thirty hydrolases, even including all the proteases, lipases, whatever, in our own lysosomes, so I'd be pretty optimistic that adding half a dozen more would make a big difference to the rate of accumulation of these sort of things.

Judith Campisi: Why not take the approach that Alteon have taken?

Aubrey de Grey: Using small molecules? Basically because the difference between lipofuscin and the structures that ALT-711 is good at, there's a single particular structure that seems to account for maybe 80% of the cross- links in the extracellular medium, that indeed has been susceptible to single agents. But -- there have been reports of small molecules that can disaggregate lipofuscin, but those reports have never stood up. There's a chemical called centrophenoxine, which has been talked about for 30-35 years with respect to an ability to disaggregate lipofuscin, and careful studies, including studies by Ulf, in fact, have basically shown that this is not the case.

Bruce Ames: So the short term approach, though, not thinking of 100 years or however long it takes, is to look at polymorphisms. Practically every gene has polymorphisms. So you look at the people who don't accumulate lipofuscin as much, and maybe those genes that they have are slightly better, and you slowly track down, among all the people who are living to 100, what's special about them, and then you can try and get back ...

Aubrey de Grey: Let me explain what the real problem is with that. It turns out that a very large proportion of the variability in the population is not at the level of the accumulation of the junk, but downstream of that. Strokes and atherosclerosis are a fine example. Atherosclerosis obviously, some people are prone to it and some people aren't. It's very very clear that there's a lot of variation in the population. But, it turns out that that's not a variation in the tendency to accumulate fatty streaks -- fatty material in the arterial intima. Everyone has fatty streaks in their aorta by the age of 10 or 15. Extraordinary -- absolutely everybody. The difference in terms of susceptibility to atherosclerosis is in the tendency of the vascular smooth muscle cells in the vicinity of a fatty streak to begin to proliferate, to surround the macrophages, and in some people's view to be neoplastic in fact. And of course the lesions become problematic.

Bruce Ames: But still, there should be some variability in people's lysosomes, so we could find out who has better lysosomes.

Aubrey de Grey: With the exception of diseases that are known to very markedly accelerate the accumulation of junk, lysosomal storage diseases, with the exception of those, no one has so far reported any variability in the population on that point. But I completely agree with you that it's very desirable to look for it -- there could be something there.

Now, here are some pictures. These are bacteria that are, we reckon, actually growing on this stuff. You can see occasional ones that appear to have taken the stuff up and not be growing on it, so we presume they can take it up but they can't break it down, that's the variability I was talking about. This obviously is comically preliminary data, but it definitely suggests to us that these things are actually growing, because these bacteria aren't being given anything else to grow on. So the idea is that not only are they breaking the stuff down, they're actually being able to extract energy from it. That's no surprise, because it's bound to be extremely energy-rich, with lots of double bonds of the appropriate sort. So this is the sort of approach we're taking. It'd take a bit of funding to actually get anywhere with this idea, but ...

Gregory Stock: The comment that you made to Bruce's suggestion about finding polymorphisms that have better ability to break down: for something like macular degeneration, is that the same thing, where it's not the accumulation that matters but the response to it?

Aubrey de Grey: Good point. I think that in macular degeneration, there may actually be scope for doing what Bruce said, because I think there is a substantial variability in the primary defect.

Gregory Stock: It would certainly be very interesting to take that whole range.

Judith Campisi: Yes -- some people don't get it, but most people do.

Bruce Ames: There could be dietary influences on that too, so it's complicated. But still.

Christopher Heward: I'd like to comment on ALT-711. Tony Cerami came by Kronos about two weeks ago and brought us up to speed on what's going on with Alteon and that molecule. It turns out the company may be in financial trouble and they have no ongoing source of funding. The clinical trials are underway, but they were poorly designed. The end-point they used in the clinical trials was blood pressure due to hardening of the aorta. The problem with that is there are already better drugs for treating blood pressure. The FDA doesn't consider hardening of the aorta a disease. So the reason the financial community isn't red hot on Alteon is that ALT-711 is not likely to be approved. The management team has taken this product down the wrong path. Even Tony has abandoned the company and has gone his own way. He has developed another class of molecules that do exactly the same thing and he says they're better than ALT-711. So AGE-Breaker technology is on the horizon, but I am afraid it's not going to become available in the very short-term.

Gregory Stock: Was your impression the same, that it's not really protectable, therefore that is a major problem for the funding of fairly expensive clinical trials?

Christopher Heward: No, that wasn't the impression that Tony gave us.

Aubrey de Grey: Yes, I got that from one of the chemists who made the negative decision at a big pharma company, he was on the other side of the argument.

Judith Campisi: But exactly, if Tony came up with a different compound that does the same thing...

Christopher Heward: Tony claims to have developed a completely different class of compounds. He says it's totally different chemistry, different structures. He says that patent protection is not an issue.

Judith Campisi: Aubrey, I remember Vince Monnier some time ago isolated or had found a bacteria that degraded these cross-links. And so he was trying to make a transgenic mouse that expressed this, so to do the same thing as you're suggesting. I would definitely give him a call and ask why he abandoned it, of why that never got anywhere. This was several years ago.

Aubrey de Grey: Do you remember that he did actually isolate the gene in question?

Judith Campisi: I believe the bacterial gene had been isolated -- I would definitely give him a call. There are sometimes, as you know, big problems in expressing bacterial genes in clones, so it's not a theoretical variable, it's a practical problem. But you could write the grant!

So I thought what you were going to talk about, a little bit, was this idea that Bruce kind of alluded to that if there is an oxidized lipid on the surface of a damaged mitochondrion, that the lysosome might preferentially take that one, and then the totally clean mitochondria are allowed to overpopulate: so wouldn't another strategy be to see if there was some way you could target different mitochondria?

Aubrey de Grey: Sort of. It's a problem that really comes down to what I was saying right at the beginning of the day, about things that are close to metabolism and things that aren't. The hypothesis that I put forward about how clonal expansion occurs, how mutant mitochondria gain this proliferative advantage, is a hypothesis that also -- in the same breath, by exactly the same mechanism -- explains why and how mitochondrial turnover happens in the first place, in the absence of any mutation. So basically you need some sort of way to maintain the normal, wild-type mitochondria at a steady-state level of lipid and protein damage, and the simple way to do that is to have turnover of mitochondria -- dilution of damage, effectively -- by destroying some mitochondria and replicating others, which of course entails the incorporation of pristine lipids and proteins into the mitochondrion to grow it so it can divide. And that's a dilution of the damage, so it keeps the damage at steady state. And this is again, in the hypothesis I put forward, driven by the preferential destruction of mitochondria that have gone critical and become too damaged to work.

Bruce Ames: The LDL receptor takes out any LDL molecule that has a malondialdehyde on the surface, which is a marker of oxidation, so, I mean, it's not implausible that that could be the same mechanism where the lysosome takes out the mitochondria.

Aubrey de Grey: Right.

Well, it says here that I'm supposed to do some sort of summing up of what we've done so far. This is going to be fun, isn't it? Maybe I'll get Greg to start, and talk about how he sees today as having followed on from February 1999.

Gregory Stock: OK. Well, to me there are a number of questions when we start to look at what's been done today. I really like the tone, the approach of looking at ways in which you can manipulate systems and come at them in rather new and creative ways to try and enhance function in one way or another. So, I think the kinds of questions that need to be asked in the advertisement of this as reversing aging, potentially, and if the goal of promoting them as engineering projects is to be met, is which of these projects, which of these sorts of possibilities, are more important and which are less important. Which are more possible, more realistic, and which are less realistic? If you have a difference between possibly decades and 500 years, that's a very different vision of the feasibility.

Bruce Ames: 500 years was a bit hyperbolic.

Aubrey de Grey: No weaseling out of it!

Bruce Ames: One can't see 20 years in the future in science. But it's still not around the corner.

Gregory Stock: But I think that your relative vision of the likelihood of that being successful would be lower, if only minimally, than Aubrey's.

Aubrey de Grey: I got my $5 wager in...

Gregory Stock: I think it would be interesting to take these and get some thoughts about which are the most likely, which are the most plausible, and what the measures of success really would be. Could they be used in the adult or could they only be used through germline interventions? I'll mention that a little bit later when we talk about the social implications. Also, how much of this is sort of baby-boomer fantasy that we really will be able to have major major impacts on aging, to the extent of reversing aging. Certainly some of these that we've discussed don't cross my threshold of what I would evaluate as really being able to reverse aging, in the sense that if you say "reversing aging", the image is of someone who is older, looking, acting and behaving like a younger person.

Aubrey de Grey: Which of these things do not come into that category, in your opinion, of the things that we've talked about today?

Gregory Stock: I think that the things that we've discussed in terms of mental functioning weren't really directed at that, in that direction. I don't see restoration of youthful mental activity.

Bruce Ames: I'm somewhat optimistic about that, that that may be part of general mitochondrial and other kinds of decay.

Aubrey de Grey: Yes. I'm actually very surprised that you chose that one, because of all the things that I've learned today, the thing that most impressed me on the optimism side of things was what came out about the role of loss of connectivity, loss of synaptic connections, in the brain, as opposed to loss of neurons themselves, and the potential for reversing that loss of connectivity as being a great deal more accessible and feasible than reversing the loss of neurons themselves.

Judith Campisi: It may not work for Parkinson's, for example, where there is a lot of loss, but it might work elsewhere.

Julie Anderson: Depending on where you intervene.

Bruce Ames: In the neurons, the mitochondria are way out in the tip; maybe that's what's hurting the synapses.

Gregory Stock: That really points out the differences in the way we might view the discussions that have gone on. I think it would be very useful to have everybody give feedback in terms of which would be the most plausible, in terms of making serious progress within what you had mentioned, a ten-year timeframe. I think that if you want to stick to that, that would be fine, or a 10 to 20 year timeframe.

Aubrey de Grey: Yes, I think that even once one gets out to talking about 20 years from now, one is talking so subjectively in terms of what's going to happen in the meantime that it's virtually meaningless, so I'd very much prefer to talk in terms of what we perceive as conceivable within ten years.

Judith Campisi: Are we talking not in humans?

Aubrey de Grey: Certainly.

Gregory Stock: In mice.

Judith Campisi: So Gage is taking fibroblasts and getting them to produce nerve growth factor and reversing synapse loss in the brain in rats, and synapses form. I mean, he's doing that now. And I think in ten years we'll certainly have that under control.

Julie Anderson: I was asked to direct it towards neurodegenerative disease, but as you said, a lot of the processes that are occurring at an enormous rate in those particular brain areas in these diseases are things that certainly are occurring in aging. So things like antioxidant therapy, addressing increased levels of free iron, all these things, go on in aging, so the therapies that we're thinking about in terms of neurodegenerative diseases may have a role in terms of maintaining mental function in normal aging.

Gregory Stock: Let's list the ones that we think are close, if we could get a ranking of some sort, in terms of what our perceptions are of these general things. The muscle intervention, for instance, you seemed to be very optimistic about being able to at least make inroads into that general problem.

Roger McCarter: If we talk mice, and we talk achievable within ten years, I would be very surprised if the age-related loss of muscle mass and character of muscle is not reversed in ten years. It's certainly achievable.

Aubrey de Grey: Andrzej. (Let's go round the table.)

Andrzej Bartke: Well, I still think that improving our hormones in the context of extending lifespan seems to be very very strong. I mean, we already have one dietary manipulation and three loss-of-function mutations, which improve function. In the case of these mutants, I didn't have time to talk about it but we also have measures of mobility and locomotor performance, to make sure that what we see in hazard avoidance is a real effect on memory, not that the animal is less active -- it improves the general situation. So it seems to be a genuine improvement in mental function. And we can link it to specific mutations, so we can build plausible mechanisms of action. And it occurs in two different models, with very little variation.

Gregory Stock: Could we have your thoughts about what other people's reactions are to the significance of mitochondrial damage, not of the kind that you were mentioning that is relatively easily reversible, but the other kind?

Aubrey de Grey: This whole idea of putting the protein-coding mitochondrial genes into the nucleus was, as I say, developed in the mid-80s, and it was suggested as a potential therapy in 1990. Now I happen to know that a number of very good labs have worked on this over the past decade, without significant success until now, until this paper I mentioned that is currently in review. But it's very reasonable to conclude that the reason they didn't have any success was because they weren't trying very hard. All of these labs were doing this work completely unfunded. Nobody has ever succeeded in getting a grant to work on this problem. They've always been doing it because if it works, it will be a bloody goldmine, but they haven't got the money to invest in it, so they just do it as a sideline, when a postdoc hasn't got anything else to do that day, and when they hit a very minor problem, basically the whole thing gets shelved for another few years, and that's the sort of progress you get with unfunded research, even in the top labs. So now, with a somewhat serendipitous but nevertheless very demonstrable success in one of these 13 genes, in a mammalian cell system, it seems very much to me that the floodgates will open and that it will be very straightforward for good labs to write grants and get them funded to do the same thing in some of the other genes. Furthermore, it will be easy for people to summon up the motivation to persist when they hit the sort of problems that any substantial project always hits. So I'm pretty confident that I'll be collecting my $5!

Gregory Stock: What about replicative senescence and cancer?

Judith Campisi: Cellular senescence. I think certainly we can find ways that we can test the effect on the secretory phenotype of senescent cells. I think we'll test that in mice. Now, how we translate that to humans I don't know, but I think within mice that it's doable.

Gregory Stock: And would you see the possibilities you mentioned, the few possibilities for interventions in adults, as being things you would be able to do in mice, if it turns out to be very critical?

Judith Campisi: Yes. I think we have at least the road map of how to do that. I mean, we're still at the stage of validating the effect in culture, but I think in ten years, yes, we'll be able to test this hypothesis in adult mice.

Aubrey de Grey: With neurodegeneration I guess we have the problem that most of the things that we get, mice don't get unless you throw in some mutation or other, which of course always casts doubt on the fidelity of the model.

Julie Anderson: And as I was saying, most of these models are not perfect models. But we've got pretty good mice at least for segmental parts of the process of neurodegeneration. And these mice are already in place, and actually a lot of this therapy is now being tested. And a lot of the pharmacological evidence or dietary evidence, there are easy things to do to test whether we're able to reverse the effects we see.

Roger McCarter: Julie, are you relatively optimistic about being able to address these?

Julie Anderson: Oh yes, I mean as I said, in terms of models that are in place, a lot of these things are already being tested.

Roger McCarter: So in terms of the models that are available -- forget about application to humans, but at least in the models, you think these things are only a few years off?

Julie Anderson: Oh, absolutely.

Gregory Stock: So this is really interesting. There's a general optimism about making solid progress, I think, within a decadal sort of a timeframe, on all of these.

Aubrey de Grey: With the exception, I mean, I'm certainly not going to claim that the lysosomal side of things is in that timeframe. I think for that, we do have an order of magnitude further to go. I don't mean an order of magnitude in terms of years, but in terms of difficulty of the problems, that's beyond the range where we could accurately be optimistic. The reason why I focused on that, even though it's an extremely ambitious approach, is simply because there doesn't seem to be anything else we can do about getting rid of this stuff that accumulates in lysosomes of various types of cell. People have thought about trying to trigger or engineer exocytosis of the stuff, and eventually excretion of the stuff, but as I say, cells are clever, and cells seem to have chosen very clearly to be unwilling to relinquish this stuff, and they think it's better for us in some way, or for them, to sequester it and not to let it out into the extracellular medium and to be excreted. So I have a feeling that attempting to engineer exocytosis of this stuff isn't going to work, and we do have to find ways to improve the degradative capacity. I can't think of anything -- short of extreme serendipity like finding small molecules that can do for that what ALT-711 did for extracellular glucose-induced cross-links, and of course we can't tell what will happen in terms of serendipity -- I can't think of any rational way to approach this problem other than the way I summarized. And I think that is going to take at least ten years to demonstrate any sort of suggestion of efficacy.

Bruce Ames: Another approach is to look at people who are living pretty long lives, and looking at polymorphisms. What's special about these people? Now, two groups have done that in mitochondria, and the distribution of mitochondrial mutations is not random, not the same as in the general population. I'm certain there are centenarians that have certain types of mitochondrial DNA that are relatively fit.

Aubrey de Grey: Yes, the maternal component of longevity is definitely there, and it's fascinating, and it certainly seems that the obvious interpretation is that it's from mitochondrial DNA. But what then? What do we do with it?

Bruce Ames: I suspect that there are a lot of polymorphisms, so that we'll learn which genes are important.

Aubrey de Grey: But what do we do? How do we exploit that?

Bruce Ames: My own feeling is that the first thing people are going to do is tune up metabolism. Get the diet to include the vitamins and minerals people need to optimize longevity. We'll find out which compounds, maybe old people should be taking high levels of B vitamins and ubiquinone and creatine, acetyl carnitine, there may be a long list of these things. So in the short term I think that's where the advances are. In the long run, the tools are coming fast and there are more people working on aging.

Gregory Stock: Let me ask a larger question. We've looked at a number of very specific challenges, and there's a reasonable level of optimism about making progress, at least in rodent systems. So, say that in ten years, there were enormous progress made in all of these areas, and they could be used in mice … what elements are missing? What would be the effect on those mice, and is it really a second layer of things that are relatively minor now, or are there major things that are not being addressed here today?

Aubrey de Grey: Have we been sufficiently comprehensive today?

Gregory Stock: If the things we've talked about today were actually successful in mouse systems, then what would those mice --

Aubrey de Grey: Die of, and when?

Gregory Stock: Yes, what would they be doing?

Judith Campisi: I think maybe we're not being totally comprehensive. I think one of the elements that we've missed in this discussion is addressing an issue that we've skirted on, which is the species specificity. So I think there are very specific ways that we'll be able to humanize the mouse. And in the past we've already done this with some of the neurodegenerative changes, in the sense of making them susceptible to human symptomology. But I predict we're going to find ways to make the mouse more human in the way that human cells are more resistant, for example, to cancer and types of damage, certain types of oxidation. And in the process, we will extend mouse lifespan.

Aubrey de Grey: But not in a manner that can easily be extrapolated to humans? Even with gene therapy.

Judith Campisi: Except that it will tell us really what the possibilities are in mice, and in mammals. If we limit ourselves to mice, this is something that's very important.

Gregory Stock: So if you limit yourself to mice, and you apply the kinds of things we've talked about here today, then what are these mice going to be like? How long do they live, what do they die of?

Judith Campisi: I think they'll still die of a spectrum of diseases. Because you're talking about general processes. For example, there is a protein in human cells, and in mouse, which recognizes damaged bases. In mice, this is a low level constitutively expressed protein. In humans it's highly regulated. So you might imagine that if you took this protein and made it more human-like in the mouse, the mouse now may be able to repair oxidative damage much better. So you might have longer-lived mice, or at least a longer-lived mouse brain or a longer-lived mouse CNS.

Aubrey de Grey: I'm not clear what you're getting at in terms of using such mice as a stepping-stone towards anti-aging human research. The problem I have is that mice are a good model because they age so quickly. And so the manipulations that would surely teach us about what to do to humans if we could, are ones which do not try to reduce the underlying rate at which damage is occurring, but try to make the mouse resistant to that damage anyway. Because we have slower damage occurring already, so making the mouse suffer that damage more slowly would indeed be predicted to extend the mouse lifespan, but it wouldn't tell us what to do to ourselves.

Judith Campisi: It would tell us what the vulnerable systems are.

Bruce Ames: I think as you learn about how things work, human interventions come up all the time. The main thing is to understand the mechanism, and I think people are going to contribute a lot with these polymorphisms.

Aubrey de Grey: I'm not sure that the main thing *is* understanding mechanisms. This was my point. I think that these end-points of these mechanisms are extremely influential in pathology. The accumulation of mutations is very much an endpoint of a long chain of events. It doesn't really matter whether the mutations arose through 8OHdG or some other intermediate lesion, or through an error of base excision repair or nucleotide excision repair: the fact is, the lesions accumulate. And they're always going to accumulate, so we have to make sure they don't kill us.

Bruce Ames: But once you see which lesions accumulate, if it's oxidative then you see how you manipulate oxidative damage; if nitrogen oxide turns out to make a lot of mutations maybe we can look at that. All these things are going to come out.

Judith Campisi: The information to get, Aubrey, is what strategies work?

Gregory Stock: I think it's amazingly important. If, by doing things that you really understand, you could actually make a significant impact in a controlled way in the healthspan of mice and in their lifespan as well, I think that generalizing that to make further progress and eventually to do various things of benefit to humans would be very likely.

Judith Campisi: You're assuming that what works in the mouse is already perfect in humans, we've optimized it. There's no evidence for that.

Aubrey de Grey: I'm not really assuming that. I think really what I'm trying to say is in the context of the difference between retarding and reversing aging. If we work out how to slow things down, such as if we work out how calorie restriction works, and we develop mimetics, it does look from the primate studies, that we could substantially extend primate lifespan, probably human lifespan, by appropriate intervention, and that if we came up with something that was sufficiently palatable, that people might actually take it, and we could actually do the experiment, though it would be a rather long experiment. But you've still got all the old people, who aren't going to live very long and would like to live longer, and I'm really interested in doing something to reverse the processes that have already accumulated. And that's where I think we can gain only a very limited amount of information from by intervening comparatively.

Judith Campisi: Do you think that the fact that calorie restriction works at virtually any age that it's tried is simply because the additional aging is slowed, or is there reversal?

Aubrey de Grey: Well, let me ask people who are more expert on this.

Roger McCarter: Can I just make a point. Bruce's drawings of the exponential increase of cancer, you can do that with so many different diseases, and if we plotted them all as a fraction of that animal's lifespan, of that species lifespan, they would be superimposable. And I think therefore it lends a lot of credence to exactly the line of argument that if you understood it, if you had mechanism, a mechanistic understanding of the process, whatever the rate of aging was, in that particular animal, you'd have a darn sight better idea of what to do with the species of choice -- let's say human beings -- simply because there seems to be such a commonality in the progression of disease and decay.

Aubrey de Grey: OK, but let's do this CR question first. Are there any aspects of aging which you can think of, which are actually reversed by, let's say, late-onset CR, such that an animal that was put on CR at age two actually are younger than they were in some sense?

Roger McCarter: Well, I think you could make a very good case for doing exactly the drawing that I was suggesting earlier on in terms of the pulmonary system of an exercising athlete. In DR, you can plot any physiological function you like, and at a given age, it's operating at a healthier level than an ad libitum-fed animal.

Aubrey de Grey: But that's CR applied from an early age.

Roger McCarter: But for many systems that would be true if applied in adulthood too. It's just that if you apply it in adulthood, it takes a lot longer for some things, for instance body composition, to adjust.

Judith Campisi: So do things get better in CR or it's a slowing?

Roger McCarter: It's a sudden jump. In other words, the physiological system, within several weeks, essentially gives you a new animal. A new animal wherein all of the different body parameters are adjusted to a new level of homeostasis. Plasma glucose is lower; plasma insulin is lower; body temperature is lower; protein turnover is higher. Immune function is up-regulated. So the animal is essentially operating in a totally different system, and I think it gets back to what we were talking about earlier about metabolism. You can't change one thing, and CR changes everything.

Aubrey de Grey: Yes, it's like Walford said, it supports every theory of aging.

Roger McCarter: Yes.

Gregory Stock: And I think you were suggesting the possibility that if you could go into mice, it was going into adult mice and actually reversing many of the components of aging. If you can do that in mice, it seems to me that that's way ahead of where you are now, that's for sure.

Bruce Ames: In a fit of enthusiasm I called up my son in New York, and I said "One of my students seems to be changing old rats into young rats!" And my son said "That's all very well and good, but you let me know when you do the next step and change old people into young rats."

Roger McCarter: Can I just add one final point on this? There does seem to be an age limit on carrying out calorie restriction. There have been very few studies of calorie restriction in old age, and the results have been very conflicting. So there isn't a resolution to how late in life you can do it and still get beneficial effects. In many of the earlier studies the animals died more rapidly when they were restricted in old age. There have been some more recent studies in which the animals lived a bit longer. But it's not clear that restricting somebody's caloric intake when they're older -- let's say the equivalent of 60 or 70 years old -- that that wouldn't be very detrimental.

Judith Campisi: Maybe the effect, if you're stress-sensitive, is bad.

Andrzej Bartke: We don't work on lifespan, but rather on physiological performance. I think that the very obvious and very effective and very safe intervention is exercise. If we do it, a number of things are very well documented: it will reduce fat, it will increase muscle, it will improve insulin sensitivity, it will reduce glucose and reduce insulin. It has been done, it works. And I think it works at any age. It is not dangerous, and it's been tested in humans.

Bruce Ames: And it's slowly spreading through the population.

Aubrey de Grey: While we can accept that exercise, taking up a lot of the right sort of exercise, when you're old will not be likely to significantly increase your lifespan, we can accept that, but I think we can also accept that even if we were able to do all these other things, and we didn't do the things that exercise does, like restoration of muscle mass and restoration of bone density, then again we probably wouldn't massively increase lifespan, because of the enormous homeostatic role that muscle has. So I think that exercise is right up there as a first class member of the family of interventions that would be necessary, in union.

I'd like Chris to take the floor now.

Christopher Heward: I guess my role here is to offer a view from the trenches. At Kronos we have to deal with the practical realities of helping people who are seriously interested in combating aging in their own bodies. They want to do it right now, today. Although I understand why you are reluctant to talk, number one, about humans, and number two, about today, rather than ten years from now, there are practical reasons for doing so. There are a lot of people in the world who are aging now and would like to do something about it now. Kronos' mission, if you will, is to help such people understand what is clinically possible and help them take advantage of it as safely as possible. Our goal is to translate technology that is in the laboratory today, from the laboratory to a clinical environment where we can apply it to our patient population. Rather than giving a talk myself, what I am hoping to accomplish here is twist your arms, all of you, to talk in more specific terms and more short-term terms about what new technologies might be applicable in the clinic in the near future. I can sense already a tremendous reluctance to do that, so a couple of things came to mind that I'll just comment on. In the interest of time, I won't dwell on any one thing. The first is that the approach that was taken here is very similar to the approach that we take in the more practical environment at Kronos. We don't try to solve the aging process. We have identified a lot of degenerative processes, as were identified in setting up this conference, that are associated with aging, and we try to address them individually in each patient as they arise. For example, we conduct a broad spectrum of tests on each patient. We do over 150 biochemical tests. We put them on a treadmill and do an exercise ECG, we measure basal metabolic rate, we measure their fitness level, we measure pulmonary function and reserve capacity. With DEXA we measure body composition and bone mineral density. We also measure highest audible pitch, vibrotactile sensitivity, hand grip strength, static balance, a number of parameters reaction time and cognitive function. We get a real clear picture of the patient's functionality, both biochemically and physiologically. Then, if we find risk factors that we know may lead to a disease state later on in life, such as cardiovascular disease, we suggest interventions immediately. Rather than waiting until the clinical disease actually manifests itself, we try to intervene early, with whatever technology is available and appropriate, whether it's dietary supplements, hormone replacement, drugs, or just diet and exercise. , It doesn't matter what the problem is -- we do whatever we think will positively impact the patient’s health and prevent the further decline towards the disease state. We focus individually on each patient’s unique situation to combat whatever disease processes we find.

Bruce Ames: Are you getting people who tend to eat good diets?

Christopher Heward: Some of them. But generally, to characterize the typical Kronos patient -- they don't smokeand they tend to have fairly good diets, although not always. We have some smokers, but not many. Most people want to live healthier longer, period. They really don't want to exercise or diet. Some people don't even want to give up their beer and tobacco. One of the practical problems that we face in a clinical setting is compliance. Our biggest problem is to come up with interventions that require little or no effort on the part of the patients, so that we can help them in spite of themselves. Major progress in this area will require a lot more basic research.

The thing that's most exciting about this meeting, to me, is the fact that it's being held at all. It's an indication that there is interest and thought being given to this whole idea of solving the aging process in a practical way. None of the things that were talked about here today were addressing the underlying fundamental question of what is the nature of aging itself; they were all focused on the practical consequences of the aging process, which is really what our patient population is interested in. People who come to us don't care about aging, but they do care about the fact that they're out of breath when they run up the stairs, or that they can't play basketball anymore, or that they have aches and pains, or their hair is graying, or their skinis wrinkling. These are issues that we all face as we age, and these are the concerns of our patients at Kronos. I think some of the things that were said here today give me reason for optimism about what kinds of treatments we'll be able to offer patients ten years from now.

Bruce Ames: You're getting the wealthy slice of the population, because they can afford to pay.

Christopher Heward: Exactly. As is the normal course of development with new technologies, this will continue to be true until the technology progresses to the point where is inexpensive enough to be affordable to the masses.

Bruce Ames: There are plenty of people out there, I guess, in that slice.

Christopher Heward: There are. The people that can afford ourprogram, the marketing people say, is the upper 3-5% of the US population. It's expensive. If you do all the testing on a routine basis and you monitor longitudinally the interventions, it costs a lot of money. The value of this approach, and the thing that excites me most about it, is that, after doing this for ten years, we're going to have a lot of data on a large number of patients.

Bruce Ames: You're taking DNA samples too, so that in the future you can find polymorphisms?

Christopher Heward: We're not. We're not taking DNA samples. One step at a time, but we are talking about that. At this point, we do not have the facilities to store samples long-term.

Aubrey de Grey: I have a question about compliance. Do you think that if there were a perception that within, let's say, ten years, supposing there were very dramatic advances in the next five years in gene therapy delivery, for example, such that advances made in other mammals might be perceived to be rather more rapidly translatable to humans than they are perceived now, and then supposing that the advances that we've been talking about had a significant impact on mouse lifespan even if they were under some sort of inducible promoter system such that in general they were only applied to mice late in their lives, such that it looked as though, basically, significant extension of human lifespan was let's say only a decade away. Do you think that compliance would improve, that people would be more willing to do things that they don't like doing, if they thought they were going to be potentially gaining a lot more than let's say ten years of their lives?

Christopher Heward: Well, that's a good question.

Bruce Ames: Ten years is huge.

Aubrey de Grey: I agree -- but not huge enough, I think.

Bruce Ames: Then, the short-term thinkers are out there. There must be pleasure in smoking.

Christopher Heward: It's hard to say; I think it depends on the individual. We have patients who, as I said, clearly know things they can do today that will increase their lifespan significantly, if they would just change their lifestyle in very subtle ways. It depends on how much effort and energy are required. Something I've actually been told is "Don't bother me with the details, just give me a pill." If they can take it without any effort, and they don't have to give up their six-pack every Saturday night, then it’s acceptable.

PD: I think the compliance issue is that if they thought that five years down the road there is going to be this magical pill, they would be even *more* prone to say hey, I'm not going to quit smoking, because I can make it five more years.

Gregory Stock: And there's actually a lot of literature about it. I think that people have switched from using "compliance" to "adherence", because "compliance" sounds like forcing. But there's a lot of literature about what leads to better adherence, and one of them is the motivational factors, which is what you're talking about. But there are many many other ones, including the relationship with the caregiver, the frequency of contact with them, a whole bunch of things, and actually we've been working on this with models and using computer-generated coaches, essentially, where a person gets reminded on a regular basis. It turns out that many of the things that contribute to adherence you can recreate, and we'll be able to very easily do that in the next ten years. So I think that adherence generally is going to go up substantially. Right now, in terms of motivation, you have patients who go for a transplant operation, and even though they know that the organ will be rejected if they don't take their drugs, they don't take them.

Andrzej Bartke: By the way, in your patients, do you see any evidence that if you can produce a tangible improvement of something in the patient, that the patient can see, it's not measured but they feel it, as you say if it bothers them that they get out of breath, if you can do something that makes the symptom go away, does this act as positive, does it improve compliance?

Christopher Heward: Yes. In fact it's particularly strong when they came in specifically aware of a problem and it improves. Then they're very motivated and the feedback is positive. But it's even true for some parameters that people cannot feel. When people just come in generally, they want a health check-up and they want the best possible health check-up, so they come to us. When we do follow-ups, and they can see a change in their blood lipid profile, it can be very motivating. It's not something they can feel, but having the visual data. The report is designed in a way that the patient can see how they compare not only to the population as a whole but also to people of their own age group, and they can gauge their progress very conveniently. Just having the feedback of improvement motivates people. They talk to each other and they compare numbers.One of the things within the patient population that's always a big item is HDL cholesterol.Everyone wants their HDL level up as high as possible. There's a lot of discussion about it among our patients. They compare notes about what you can do about it, the kinds of interventions we use at Kronos, the effect of exercise, and so forth. Having a regular check-up that gives you a number, or a group of numbers, to work towards or to work from, provides a tremendous incentive to stay the course and continue to do what you know will work. Now, that's true for a subset of our population. There are others that are not very compliant at all. We tell them exactly what to do and how they can improve, but it's just like talking to a wall.

Andrzej Bartke: These people who are motivated by the responses, do they get impressed with slowing down the decline, or do they want to see a reversal? We were talking about reversal...

Christopher Heward: Good question. As a scientist, I would consider a 25% reduction in the rate of aging, the rate of decline, dramatic and wonderful. But if a person starts on a program at age 60, and we can promise them a 25% reduction in their rate of decline, they do a quick calculation, and the guy says "Gee, so does that mean I'm going to spend all this money, every year, and you're telling me that by the time I'm 80, I'm going to function like a 75-year-old?" It's not enough!

Aubrey de Grey: That's just what I was trying to say.

Christopher Heward: What they want is to function like a 30-year-old. That's what they're looking for. Although we don't promise that, obviously, because we don't know how to do that, people want an improvement in their numbers, back to youthful levels.

Andrzej Bartke: I was wondering, you know, which of these outcomes is enough to make people stick with it, to actually change their behavior.

Aubrey de Grey: Right. I mean, let's be clear about this. The thing that really we've been trying to do today, as a departure from the meetings that we're familiar with, is not to try to be bound by our preconceptions about what's possible in terms of new therapies. I was very careful not to invite anyone who has anything to do with money here today, so that we could be speculative and not really worry too much about it, and I think it's been very successful. I think that in general we haven't allowed ourselves to be too constrained. But it is important to remember that we have to impress people in order to get adherence -- we have to impress people in order to get money! -- and there is a case that people have tended to aim too low, in order to get funding for projects that they know they can succeed in, rather than to work on higher-risk stuff with much higher reward. Now of course, that is fundamentally a consequence of the nature of the funding process -- it's very hard to get money for high-risk projects, because feasibility is a very large component of the fundability of this work. But if we don't start trying to aim higher, and thinking about what we could do, we won't even try to construct a case for doing the research, that's solid enough to persuade someone to give the money. And especially in aging research as it is at the moment, with a massive increase year-on-year of money that's not public, government-funded money, and therefore is a great deal less constrained by the standard paradigms of what's fundable and what isn't than we're used to, we really have to get out of that rut, that psychological rut, I think, of thinking in terms of what used to be fundable.

Bruce Ames: But as scientists, we're always trying to do something that has a reasonable chance of succeeding in our lifetimes. So it's a matter of judging where you get most bang for your buck.

Aubrey de Grey: From our personal point of view, sure. But from the recipient's point of view, from the general public's point of view, there's not much bang for the buck.

Gregory Stock: I think that this is part of the larger view of how this all fits in socially.

Aubrey de Grey: Yes -- it's a lead-in to you.

Gregory Stock: I'll make some comments, then we can open up the discussion again.

I was trying to consider what the larger implications of reversing aging might be. When we talk about anti-aging, it seems to me that there are three kinds that are really possible and plausible. One is that there will be interventions that work in adults and are essentially a slowing of aging. That's what everybody usually discusses. The second is the focus of this meeting – that we could actually reverse aging, which obviously is what most adults would like to see happen. And the third is that we won't be able to make any meaningful anti-aging interventions except at a germline level, which would not affect us but future generations. I've heard lots of arguments, particularly with germ-line interventions, but with all of them, that it would take a long time after such interventions were developed before we would feel the effects, a long time before there would be a meaningful impact on the population. After all, only a small fraction of the population would be using these sorts of things initially, especially with germline interventions. In anti-aging a common thing to say is "well, if you alter genes at the germ-line or embryo level, it's going to be 80 or 90 years after you do that before you'd have any impact on the population."

Judith Campisi: You wouldn't even know if it worked.

Gregory Stock: Right, but assuming that you had a pretty good idea.

Aubrey de Grey: Even in the "worst-case scenario" where it does work!

Gregory Stock: And even if you have biomarkers, so that you actually know that, it would take a very very long time. If, however, you were able to reverse aging in adults, the effects on the population would be massive and immediate, because you could have rapid large-scale changes in the population by reducing the level of decrepitude. There would be changes in the whole make-up of the population, and I think they would largely be very positive at the individual level. (There would be a huge social adjustment, of course, but I'll get to that in a moment.) And these changes would strongly contrast with the effects of the "successful aging" programs we see today. What is involved in successful aging? Well, it's maintaining functional immunity, maintaining a level of activity, slowing one's deterioration, making it as moderate as possible; that's the vision of it today. If you had a slowing of aging as well, there would be significant impacts. They would, however, be kind of nightmarish to a lot of people, because if we could allow people to live longer but couldn’t retard the debility associated with it, then relatively soon we would begin to end up with a much larger and sicker elderly population than previously, a huge bulge in this demographic group.

Bruce Ames: But that's happening already.

Gregory Stock: And there are quite a few consequences to that. But I think the implications of these things, and their manifestations, are going to be much more rapid than most people realize. I suspect that change may occur practically overnight, because the psychological implications even of retarding aging are so substantial. Any believable progress would reverberate through society and bring about huge funding changes, and this would happen as soon as people realized that anti-aging interventions were really plausible. At present, most people are not comfortable thinking about either retarding aging in a significant fashion, or even attacking the problem, and they certainly don’t think about reversing aging. They think "These are people <those who espouse these goals> basically can't deal with their mortality. They have psychological problems, and they just can't accept it." So, if you could make a convincing case that you might really be able to treat adults, massive effects would arise quickly and spread through the population. Even if you were only able to do interventions to the germ line, very substantive psychological effects would develop within a matter of decades.

Judith Campisi: Could we manage the changes?

Gregory Stock: If you think of the situation where we’re actually able to begin to reverse aging, it seems unlikely that it's all going to come at once. My suspicion, however, is that the biology of aging would become a major focus of medicine at that time. Medicine would take a much bigger percentage of GNP. There would be enormous changes in therapy. The initial impact of the interventions would be governed by whether they were the ones that have relatively high capital cost, or relatively high variable costs. If the interventions are very expensive because of large expenditures on equipment and such, but the actual cost of doing the intervention to large numbers of people were not that high, then it seems to me that there would be huge political pressure to make these things available to everyone. You would not want to exclude all but the wealthier people who are now using supplements and vitamins. After all, what better way to spend our health care money?

Bruce Ames: But it's a knowledge problem too. Vitamin pills are useful but you have to convince them to take them. That may take years.

Aubrey de Grey: Tell them it's going to reverse their aging, and then see whether they take it!

Gregory Stock: It's interesting: some of the things that you've pointed out, the public may easily decide "ah, you know, this actually might work!". I think people doing body-building and such are already using whatever they can. You made comments about some of these quack treatments that actually may work in some ways better than a placebo. I think that if people really felt that there were substantive changes to be made to their level of health and their youthfulness, there would be enormous incentive for people to use them.

If there are very expensive variable costs, because a lot of surgery is involved or a lot of attention needs to be given to individuals or to training on an ongoing basis, then I think you run into problems with possible bottlenecks. In this situation, there might be lots of conflict about how to make this available to more people. One issue here is whether the treatment is a one-time intervention or a long-term one. Is it going to be a daily kind of a treatment? What are the side-effects? How onerous are the procedures? Do they involve lots of pharmaceuticals, and if so, how much are the drugs going to be tested? If you can demonstrate effects in mice, the whole infrastructure of testing and drug clinical trials will have to be altered, because there will be huge social forces pushing these procedures into use. Of course, the way drugs are tested is going to change anyway, because of the advent of genetics and the move from huge blockbuster drugs to highly targeted ones. With increasing numbers of people voluntarily using drugs with many different levels of safety, I think it would behoove us to move in the direction of allowing this and monitoring these people very closely so that we are essentially doing broad clinical human tests that are dynamically tailored on the fly. As we move forward, we're going to need to constantly try to get informational feedback on what we're doing, what's happening to our cells. The best example of this is in in-vitro fertilization and certain treatments for infertility, where all the interesting work that's going on is really using human experimentation, because different people try different things and report their results. You can't do this work in animals because you simply don't have the same fertility problems in these model systems.

Bruce Ames: The whole drug thing, now, is that there are huge hurdles to get past.

Gregory Stock: A lot of people are thinking about that, and there are forces that are building, I think, to seriously modify the whole approval process. Most of the diseases that are being dealt with affect very small numbers of the population, so people are relatively conservative. Something that affects aging, however, which is in everybody's interests to validate rapidly, will accelerate this whole effect.

Bruce Ames: But in health food stores, people are just gobbling up every herb and all these weird things, but there's no follow-up. There ought to be follow-up. I'd like to know what people are taking and why, and what the statistics are, and the placebo effects. Maybe there's some real stuff, and we're throwing away all this experimentation.

Gregory Stock: And there are some people that I've talked to down at UCLA who are doing work on what's called quality engineering, where the idea is to develop test designs where you don't really do a normal clinical trial but actually alter what's being done within the test groups as you're going forward, breaking them apart into subgroups. Because even with regular testing you're having a problem with maintaining groups whenever you show a difference. You often can’t maintain adherence within a group that's only getting a placebo.

Andrzej Bartke: I think it could be argued, coming back to the point you made about DHEA, that experimentation with DHEA and melatonin and so on is already ongoing, uncontrolled, unsupervised, without good data collection. But it's ongoing, and I think it would be nice if the data could become available, even in the negative: it would at least be possible to announce the fact that they are negative.

Gregory Stock: I think some sort of a safety rating system will be the kind of thing that's going to happen, where things are graded toward the level of agents involved, and people would be given much more latitude to make interventions at whatever level of risk they feel is appropriate.

Aubrey de Grey: I find that I completely agree with all of these predictions that you're making about the changes that will occur, and the enthusiasm to put more of the GNP into the appropriate areas of medicine or into research and development and so on. The interesting thing that I'd like to explore is the question of how sharply, and at what point in the progress towards this medicine, one would expect that cusp to happen. Because I think what you're basically saying, which I certainly would agree with, is that essentially there's a sort of denial -- the public are in this sort of denial about aging, they don't want to think about it, they don't want any minor progress to go to their heads, and they sort of belittle in their own minds the prospect of any change, but eventually there will be some critical advance that breaks the camel's back and suddenly shifts opinion to the view that maybe it won't be all that long, and then I would think it will cascade. It seems to me that that may not be a human advance -- that it may be an advance in mice.

Gregory Stock: I would agree.

Aubrey de Grey: If it's a sufficiently dramatic one.

Gregory Stock: I think that what is required -- if you think about that, it's a very hard thing to decide -- but I think basically if you have focused interventions in rodents that clearly extend rodent longevity and lifespan in ways that people could imagine applying to themselves at some point. Caloric restriction is not such a mechanism.

Bruce Ames: If our acetyl carnitine and lipoate extends these rats' lifespan, everyone in the world will be popping them. They won't be waiting for human trials.

Gregory Stock: Just as with drugs -- you really cannot control drugs now. With something of this sort, there's going to be widespread use of these things, followed by a retreat of the regulatory environment. People will say, "Well, this stuff's going to be used anyway, at least what we should be doing is getting data". It is not going to be opposed, there's going to be such massive use of these sorts of things.

Aubrey de Grey: Supposing that dietary manipulations such as the one you're doing at the moment *don't* have a significant effect, but transgenic ones -- or transgenic ones combined with dietary ones -- *do* have a very significant effect. Then, it seems to me that the applicability to people that people see for themselves is absolutely crucial. So this is to repeat the recommendation that I'm thinking in terms of, that pretty much whatever transgenic intervention one would be experimenting with in mice, if it looked as though it was really working, to basically do the same experiment again in a late-onset inducible system, under cre or something like that, so as to demonstrate that the effect can be done in an animal that has been aging perfectly normally up until let's say middle age.

Gregory Stock: I would say that it would even not need that. It's definitely true that that is the case, but what I've thought about is in terms of germ- line interventions that could be made that would significantly extend healthspan and lifespan. What would happen would be that suddenly something that was never viewed as malleable would be popularly viewed as malleable.

Aubrey de Grey: The perception would be that the incremental advance required to develop reversal would be small?

Gregory Stock: Right. People would really want it to happen. A lot of people talk about the opposition to it, but I think that would fade very rapidly. The reason for this is that there's demonstrable desire for youth and healthfulness, which is clear from the amount of money that's spent on all sorts of things that don’t work and that are not very pleasant either. Certainly there would be a generational issue as well. People who are younger will want such interventions, but maybe someone who's very old will say "Well, I've had enough". People do say that, because you feel broken down, and tired.

Bruce Ames: George Bernard Shaw said that "Youth is wasted on the young"!

Aubrey de Grey: That indicates that if he could have had it back, he wouldn't have felt that he'd had a long enough lifespan at all -- it would have been very welcome.

Gregory Stock: I think that eventually we will appreciate that aging and death are not needed to make life meaningful. Also, the people who are more predisposed to embracing these kinds of things would end up being the vital people in society, who drive public opinion. Those who are most interested in it would actually propagate it. I think that when you start talking about the social problems that would be caused by anti-aging interventions, they would be tremendous, in that they would really rip humanity free. There's so much about the way we live our lives and the structure of society that is based on the anchor of human mortality. People mention overpopulation, social security, and such, but I don't think they would be big issues. I think we will easily adjust to those things. What will be the most difficult to deal with will be things such as intergenerational differentials and the whole way that power is handed off from one generation to the next. How would many of the benefits, like property, jobs, and other things that are usually passed on to the next generation be handled: that would be a very big challenge.

Bruce Ames: Old professors never want to retire!

Andrzej Bartke I don't think we need to worry about the consequences too much. Because I think the type of interventions we're talking about today, even with the most optimistic assumptions, in terms of years added to healthspan or years added to lifespan, will probably do less than has already happened in the last 100 years. And societies were not overthrown or revolutionized. Things have changed. You know, people's perception of what's old has changed, people's idea of what's appropriate behavior for a 60- or 70-year-old has changed. But, you know, there wasn't a huge upheaval of values, goals, perception of sense of life.

Gregory Stock: I think you're largely right. But there are going to be a lot of things that we'd be dealing with over a period of decades. It depends how suddenly these things occur, and it depends on the relative timing of what systems we were able to revitalize. Ultimately it would be very challenging,. You could say we've adjusted to things very well as they are, but there's a great deal of social chaos. There's a sense of people not understanding what the trajectory of their life is, what life is all about. There are issues, as well, if you have extended lifespan, of reorienting many of our biological drives.

Aubrey de Grey: There's a very real sense in which we would be different organisms.

Gregory Stock: Right -- but we would have many of the same orientations.

Aubrey de Grey: I'd like, in the context of what Greg's been saying, to ask a slightly introspective question of everybody here. There is a very wide spectrum of publicly stated opinion, on behalf of the specialists in biogerontology, with regard to the desirability of spending money on anti-aging research. Some people are quite happy to come out these days, still, and say that it's basically an illusion that we'd get anywhere at all -- I certainly haven't heard any of that opinion expressed by anyone here today. But a much more popular view, that's certainly I think something that the majority of biogerontologists still tend to say when confronted by the press, is that working on extending maximum healthspan is a distraction from the important business of working on age-related degeneration. I think it's time that that became a less fashionable way for this community to interact with the media. Do people think I'm overstating the case? Do people say that still?

Roger McCarter: I think it's very important to say what we mean by it. I think you raised this right up front this morning. Obviously we don't want to get into definitions of aging, but when we talk about anti-aging, for instance, I find it very easy to view this in terms of getting the muscle mass that I had when I was 20. And I think put in those terms it's less threatening. If you talk about reversing the aging process, people tend to think you're a crank, because in some way you are searching for the fountain of youth. So I think in some ways it's a matter of what we mean by it. And if we can get the meaning of it across, in a way that doesn't sound like we're chasing the same thing that Ponce de Lyon was, 500 years ago, I think that would be important.

Aubrey de Grey: So do you mean talking about it more as engineering than as an arbitrarily distant goal?

Roger McCarter: Yes -- in terms of a very specific realizable goal, which is possible in terms of the technology that we have available to us today. Rather than turning back, almost like reversing time. The thing is I think that's a popular perception. And so we're not talking about reversing time, what we're talking about is returning to a level of function which we have lost.

Gregory Stock: I think it's really interesting that many of the people who talk about making progress in age-related diseases regard trying to retard aging as not a useful goal for society. I see this at the UCLA School of Medicine where I had the first conference. Some of the department chairs were saying that this wasn't a proper focus, not a reasonable goal, because of the social problems. Working on age- related diseases, of course, was OK. To me it's very interesting that we get that reaction. And these people were reasonably old!

Aubrey de Grey: Would you agree with Roger that the reason why they distinguish between these things is because they feel that people who say that don't know what they're talking about, haven't defined what they're talking about?

Gregory Stock: It's interesting -- there was one woman in particular whose view was that if was very threatening for society. People realize that although you might say that it's not going to happen overnight and that we'll adjust to it, it will change virtually everybody. People say that the public doesn’t really want to change the human lifespan, and yet we are quite accepting of the idea of spending fortunes trying to get an extra six months of what is a really very poor quality of life. I've found that the biggest confusion about the idea of retarding aging is understanding that it means a longer period of vitality. Virtually everyone is very supportive of the idea of extending the human healthspan. What is scary is the image of a bunch of incredibly old people living extra time and being cared for in all sorts of ways. When you start talking about increasing lifespan, that's what people think of, because the big issue now is deciding when to stop high-tech interventions for people who have an extraordinarily low quality of life.

Judith Campisi: Another issue which we haven't talked about, which I've run across, is I've given public lectures, and I have been criticized for working on any intervention that would increase the number of people alive and wasting resources, polluting the atmosphere and being totally antisocial. I mean, this is not something that I've heard outside of California, but definitely it exists. So the idea is why do you want to give people another 20 years of life, when all they're going to be doing is increasing garbage, depleting the ozone layer, using up fossil fuels, and so on.

Aubrey de Grey: And what do you find to be the most successful rebuttal of that?

Judith Campisi: I don't know that I've found a successful one! Obviously what I say is that first of all we have a big societal problem in that regard, in that Western society are major consumers, and there's no reason we can't reduce our consumption. There's such a societal problem that the impact of another 20 years of lifespan would be minuscule.

Aubrey de Grey: I was wondering whether the most effective argument was the argument that postponing "frailspan", as you might call it, or reducing frailspan by postponing it, will actually *diminish* the amount resources taken up, even by a greater number of people, because it's the frail who take up the greatest amount of resources (if you include healthcare)?

Judith Campisi: They don't put it in terms of money, but in terms of pollution of the environment.

Gregory Stock: I've found two particular ways of dealing with that. One being that in fact it is the birth rate that is a significant issue in terms of creating larger numbers of people, so that if they really feel that way they should focus their energy on seeing that various methods of birth control are provided to the rest of the world and made accessible. If you cut the death rate, it will not have as much of an impact on the increase of population as a reduction in the birth rate. The second thing is that if people really feel that way, then what they should be doing is try to stop there being any expenditures on addressing diseases that occur late in life. You should certainly stop the war on cancer, stop the war against heart disease, all of the specific diseases where huge amounts of money are spent.

Judith Campisi: I've actually had people say "I would have no problem with extending the lifespan of Pablo Casals, or some great artist, but I don't think it should be done for the general population". This is getting very scary. And it really gets to the point where it's not very logical or very meaningful having discussions. But I think the issue, the argument, remains that birth rates have a much larger impact that if you stopped death, and that's a very good point, that's definitely true. If you cut birth rates by 50% and stopped death rates cold, the population would decline!

Gregory Stock: Plus, they use much more resources. There are incredible resources in getting a person to a functional level. If you can maintain a person where he's actually functioning and contributing it is very effective. After all, you're just getting smart and learning to deal with society and make a contribution when you begin to wind down. So, longer lives are actually a much more efficient way of using resources. And just as some people restrict the size of their own families because of feelings about population issues and such, you know, it starts at home. Tell them to take a pledge not to have any medical care, it'd change their mind really fast!

I really want to congratulate Aubrey for putting this together. The general thrust of it, which I think is very fresh, is a wonderful extension of the original effort that I made last year. I think this event has raised it to another level, and I think having this discussion has been really great.

Aubrey de Grey: Well, thank you very much. And I'd like to express my thanks to everyone for participating -- I think, obviously, it wouldn't have been the same without you! I certainly regard today as a very great success. I'm sure this will be repeated, and I very much hope that everyone here will continue to be involved in future meetings of the same sort. Perhaps we can have them not just organized two months in advance, as a one-day meeting, next time. We can probably have it slightly less crushed in terms of the schedule, that would be nice.

Andrzej Bartke: Would both of you feel that it would make sense to project this kind of meeting every year, every other year?

Aubrey de Grey: Well, there are other meetings happening around the world, and I feel it's important to be inclusive about these things, to make the best of these things. Greg and Chris and I were at a meeting in Manhattan Beach, in southern California, in June, organized by a new potential investment group in anti-aging research: they just got hold of a bunch of people and said are you interested in coming along, and we had a discussion over two days, and it had certain merits. I'll be seeing David Kekich, the guy who ran this thing, a week today, when I go down to southern California, and I'll certainly be telling him about how today went, and he's already organizing something which might be very similar to this or similar to what he organized before, for January -- this coming January, only three months from now. So of course if one ended up with a situation where meetings like this were happening somewhere around the world, organized by some group of people or other, in the region of every three months, then most people aren't going to go to all of them, but it could become a real process.

Judith Campisi: It would become quite boring!

Gregory Stock: Yes, I think the idea of doing something like this every year and a half would be very interesting, to sort of assess progress. I think what would be important would be to be more specific, too. We've picked some realms where progress might be made, but it would be useful to think about what form that progress might take and align it with the milestones we originally created. I’d like to see how that changes over time; it might people's thinking. It seems like your Roger’s thinking was very stimulated by the task that was set out, and it's really important to be able to capture this, so I think that the idea of creating a series that will have some history to it is important.

Aubrey de Grey: In terms of the range of people who participate, I think that, for example, I mentioned to Judy in email a few weeks ago: this meeting, as I say, I made a conscious decision to avoid involving anyone who was involved in funding of research. I definitely don't think that that's something that should apply in general to such meetings. I think that in fact it would be appropriate to have a meeting in the relatively near future which essentially balanced this one in that respect, by being very heavily weighted in favor of people involved in making decisions of both public and philanthropic and commercial funding of anti-aging research. So, you know, it might not be all that boring! Each meeting is likely to end up having very much a distinctive flavor.

Gregory Stock: I think the one thing that is absolutely crucial, and when you mentioned Kekich's meeting, is that these have been very solid, scientifically. Everybody who was invited to the first event was clearly very solid in their research and their perceptions of things, and I think this meeting has been the same. When you're dealing with an area that is as subject to distortion as is research on aging, and where inflations can occur, then if you have just one or two people who are marginal in the way they speak and how they look at things, it will hurt the whole process.

Aubrey de Grey: So for example at this meeting three months ago, Robert Freitas was there, the expert on nanomedicine, certainly looking a very great deal longer than ten years ahead, in terms of the applicability of the area he specializes in. And in fact Kekich has already acknowledged that there was probably too much breadth of long-term-short-termness in his meeting, and he's decided to try to organize two separate meetings, on consecutive weekends I think it's going to be, in January. I certainly told him that this gave him an opportunity to separate the really long-termists from the more short-termists, the more, let's say, biology-oriented people. I will be interested to see whether he decides to take me up on that suggestion.

Gregory Stock: But it's even more than the time horizon, because there can be too much looseness, an inability to make a distinction between what might is possible and what is nonsense. And if you don’t make that distinction, then any outsider looking at it takes the most nonsensical thing and uses them to discredit other possibilities that may seem extravagant as well, but are actually quite plausible. This effort will only be important to the extent that the participation is kept absolutely solid.

Aubrey de Grey: Absolutely. Are we all worn out yet?

OMNES: Yep!