Dr. Aubrey de Grey is a legend in the longevity field who has been steadfastly promoting the idea of life extension since well before it became mainstream. While with SENS Research Foundation, de Grey made significant contributions to geroscience, and at Longevity Summit Dublin last year, he announced the creation of his new brainchild, Longevity Escape Velocity Foundation (LEVF).
Now, the first major and long-awaited LEVF-funded project is being launched: Robust Mouse Rejuvenation (RMR). This is envisioned as a rolling research program aiming to increase both the mean and maximum lifespan of mice by at least 12 months with various combination therapies started late in life. For the first study, four therapies have been chosen: rapamycin, a senolytic, hematopoietic stem cell transplantation (HSCT), and telomerase expression. A groundbreaking experiment by any measure, RMR got us excited, and we reached out to Aubrey to discuss both RMR and LEVF in depth.
The following interview has Arkadi asking questions in bold and Aubrey de Grey answering in normal font.
We are obviously very excited about LEVF’s RMR project. Could you walk our readers through its design and goals?
I believe we’ll have two outcomes. One of them scientific, and the other more, if you like, rhetorical. We want to get mice to live a lot longer than they do now: at least a year longer, starting the treatment or treatments only after middle age. The idea is that this will appeal more directly to people who care, vote, pay taxes, and make donations than if you do early-onset interventions. So, I decided to put numbers on this, to have a milestone that clearly says this is where we want to get to. We believe this will be a sufficiently dramatic result.
What does “sufficiently” mean? The audience that I still care about the most is my prominent colleagues in the community: people who talk to the general public and do a lot of media appearances. We all know each other very well, we’re all good friends. So, I have a very accurate idea of how impressive a result needs to be in order to stimulate these people to say things publicly that they wouldn’t previously have said.
First, we have to take a strain of mice that are inherently healthy and long-lived. Of course, the standard strain that’s normally used are C57Bl/6 (“Black 6”) mice. We start late, at 18 months of age. That’s roughly one year less than the average lifespan of this strain. The idea is to double that one year, i.e. to give an average lifespan of three and a half years rather than two and a half. We also want to have at least a year of increase in maximum lifespan. As I’m sure you know, maximum lifespan is normally defined not as the lifespan of the last surviving mouse but rather the average lifespan of the last 10%.
This is important because many interventions can increase mean lifespan but not maximum. The interpretation that most people, including me, put on that kind of result is that the intervention is hitting some but not all the mechanisms of aging, and those that it’s not hitting are still proceeding unabated.
Meaning, we have compression of mortality instead of an increase in maximum lifespan?
Of course, we must be very careful with what we mean by compression of mortality, because that’s at a population level. We don’t know if any given individual mouse actually had a shorter time when they were going downhill health-wise. So, in this study we absolutely want to increase maximum as well as average lifespan.
There are various other things that we’re throwing into the mix. First, we’re using combination therapies. We take the view that rejuvenation therapies, those that repair damage instead of just slowing down the creation of damage, are inherently more partial in how much they do across the board of different types of damage. We figured out that we might get much more than the sum of the parts by putting multiple of these things in at the same time.
The second thing, again, coming back to rejuvenation, is that we do recognize that there are knock-on effects. In other words, you may be fixing one particular type of damage, but having done that has a beneficial effect on the rest of metabolism and somehow slows down other things as well. We believe that is the basic reason why we see in the literature a small but non-trivial number of reports of increase in maximum lifespan, as well as mean, by rejuvenation-type interventions. So, we believe that those are the most promising ones to start with, and we want to combine them.
In our current, initial, study (this is the first round of what we believe will be a rolling research program) we are taking four interventions, and one of them is not of the type I’ve described: rapamycin. You can think of it as a non-rejuvenation control if you like.
But it’s not just a control, because we want to know how well it synergizes with bona fide rejuvenation therapies that repair damage. Indeed, the entire design of the study is based on my analysis of what will give us the most information, per dollar spent, about the synergies between different treatments. In this study, we have a thousand mice in total, 500 of each sex, and we have split those mice into 10 groups of each sex, so 50 mice per group, per sex.
What are those ten groups? Of course, there’s a control group that gets nothing, and there’s also a group that gets all four of our interventions, but there are also four groups that get exactly one intervention, so that we have some kind of baseline. And then we have four groups that get three out of the four. This is very important. The purpose of this is to determine whether there are antagonistic interactions between things, which is possible.
With this, we believe that we’ll basically get all the possible information about those interactions. We believe that we would not gain appreciably more information if we also did the six more groups where you do exactly two out of the four. We will be able to extrapolate with very high confidence what their results would be. That saves us a lot of money.
What was your rationale for choosing the interventions? You made some intriguing choices here.
As I said, we did want to have rapamycin in the mix as a non-rejuvenation control. It’s a calorie restriction mimetic and probably the most effective one out there. The real question is, why the other three?
There were two main reasons. One was that there were already studies by others showing increase in mean and maximum lifespan in long-lived strains of mice when starting late in life. So, by and large, we are reproducing the protocols that these other groups used over the years.
The other big criterion is that the interventions should, as far as we can say, be targeting different types of damage. As I mentioned, we are drawing on the idea that there will be knock-on effects from one type of damage onto others.
Still, would it be fair to say that among those four interventions, we only have robust life extension data in mice for rapamycin?
Among the calorie restriction mimetics, I would agree with you. But remember, we are restricting ourselves to late onset interventions, and there are not a lot of those. One of the most horrifying things in the history of the field was the fact that a decade ago, when the NIH had this enormous piece of good luck, when they accidentally did this study starting at twenty months of age rather than four months, which was their original intention…
Yes, with rapamycin.
They got this fantastic result that was as good as calorie restriction itself, starting at that same kind of age. It’s so obvious that what they should have done was to go back and revise the guidelines for further studies to focus more on late-onset studies.
Of course, I understand why that didn’t happen – because people in academia are constantly fixated on what are diplomatically called “positive results”, on getting their stuff published in those high-profile journals. That means doing stuff that’s more likely to work, even if it’s less informative, which is insane. But it’s the way of the world. It’s a good example of the reason why I chose the direction I did 20 years ago and decided to lead a series of independent non-profits that were funded by philanthropy rather than by grants.
Not only might late-onset studies be more informative, but using pre-aged mice also allows you to drastically shorten the whole process, right?
Yes. Also, we are planning on putting out interim data all the time, very frequently, once a week or once every two weeks, because we’ve got two things going for us. The first one is, as you say, starting late means that the overall experiment will take two years rather than four years. But the other one, which is just as important, is that because we are mostly doing rejuvenation therapies that remove damage rather than just slowing down the creation of new damage, we have a good chance of seeing they’re working from the divergence of survival curves quite quickly, like after only six months.
Did you consider dietary restriction of any kind as one of the interventions?
We thought about it. At this point, I believe that the data on rapamycin is strong enough that it doesn’t matter which of the two you do. And, honestly, there’s just less labor involved in putting rapamycin in the chow. It was a more straightforward way to go.
With such a lofty goal at hand, would you like to make some predictions about the results? For instance, which interventions or combinations are more likely to succeed?
Definitely not. Let’s be clear: I do not actually have lofty expectations for this first experiment. We’ve been saying from the beginning that this is a rolling research program, and our top priority is, as soon as we get this one kicked off, we’re going to design the next one, and to bring in the money, which is about three million dollars for each round.
So, no, I have no idea what we’re going to get with this one, but I’m hoping that we’ll be able to do subsequent rounds more than once a year – maybe every nine months or so – because we don’t need to wait for the results of the first one to decide how to do the second one. We’re also incorporating masses of information from the community, from literature, and we already have a plenty good list of things that we’d like to try in the next round.
Have you decided on what senolytic will be used?
Yes, we just decided on it, so it’s going to be exclusive for you. We’re going to use conjugated navitoclax. As you probably know, navitoclax has a reputation as a reasonably good senolytic. However, it’s not very specific. But Manuel Serrano had this extraordinarily simple and brilliant idea based on the fact that most senescent cells have a high expression of beta-galactosidase.
You can encapsulate your navitoclax, or any other drug for that matter, in galactose. If it goes into a regular cell, then nothing will happen, while if it ends up in a senescent cell, the galactose will be broken down, the navitoclax will be liberated and will probably kill the cell. It’s just brilliant.
So, he published a bit on that a few years ago, but the manufacturing of this encapsulation is finnicky and hard to reproduce. A few years ago, he and some of his colleagues in Spain decided to try a variation on the theme: conjugated navitoclax. Conjugated means that you actually make it into a pro-drug. You covalently attach galactose to the molecule in a location that makes the molecule not work. But because it’s galactose, if the cell is producing beta-galactosidase, galactose will be cleaved off in senescent cells and only in senescent cells, and, lo and behold, you get the same result. This turned out to be a lot more reproducible.
With mTERT, you will be using a protocol that was only used once in a small proof-of-concept study, including the rather unconventional intranasal method of delivery. How comfortable are you with that?
It was a small study, true, but the results were very impressive, and it was done by very good people. George Church would not have put his name on the paper otherwise. Obviously, you have to weigh all these things.
What will you be measuring?
We’re going to measure all sorts of stuff in addition to lifespan. We will focus heavily on function with tests such as the rotarod, so that we have good information on healthspan. We’ll be doing that in different ways. First, we’ll have a bunch of non-invasive things that measure agility, visual acuity, physical appearance, including alopecia and kyphosis (the bending of the spine). These are well-established measures of biological age.
In addition, we will be sacrificing some mice at various periods during the study and asking what condition they’re in. On top of that, we will be looking at mice that die naturally during the experiment and figuring out what they died of. So, we’re really covering all the bases.
The innovation that I introduced, and I don’t think this has ever been done before (I wonder why because it’s a bit obvious) is that rather than choosing our cull points at particular chronological age intervals, we look at the survival curve of each group. Say, we wait until 20% of them have died, and then we kill a few and ask what state they’re in. This is, to my mind, obviously an improvement on the way things are normally done, because there’s no point in measuring two different things if you know in advance that they are highly correlated.
We’re basically factoring out that correlation with lifespan. If you’re comparing one group with another group, you may be looking at them at different ages, but if it’s the same survival point, you expect them to be roughly equally healthy, and you can find out whether some of them are healthier in one way and less healthy in another way. We feel we’ll get much more information that way. This was my innovation, but it’s been very well received so far. I’m pretty happy with this.
If you do the culling at different chronological points for different groups, could this complicate intergroup analysis?
Not at all. It’s going to reveal things that would normally not be so well-revealed. If you do it at chronological age, and the therapy is working, then the mice will be on average biologically younger. But you knew that already because you saw how many of them have died, right? So, you’re actually learning less because things are already tightly correlated. This method factors out the correlation with lifespan and therefore gets you more information, a better signal to noise ratio, so to speak.
Say you achieve this goal of 30-40% life extension in mice. How do you think that would translate to humans?
Our goal here is twofold, as I said earlier. We would love to identify therapies that will translate to humans, and we certainly would predict that damage repair therapies, meaning rejuvenation therapies, will in general translate more directly, more effectively across species than therapies of the kind that I call “messing with metabolism”. That’s because damage is more similar across species, across mammals anyway, whereas metabolism has a lot of differences.
We hope that will happen, but whether or not it happens, we will certainly achieve the rhetorical goal. We will achieve the goal of people saying that Aubrey de Grey was right all along (albeit maybe not in those words!), and we’re within striking distance of achieving this with humans. That’s very important because this will make Oprah Winfrey start saying that this is coming, and that’s “game over”. The following day, it will become impossible to get elected unless you promise to put proper money into this. Those two goals are of equal importance.
RMR is not a new idea. What happened that finally made it possible?
Two things. First, we’ve got enough money to do the experiment (it’s really expensive). Second, we have the substrate, the interventions that have already shown individual efficacy in other people’s hands. We didn’t have that until a couple of years ago for any damage-repair treatment.
Two years ago, at SENS Research Foundation, Alexandra Stolzing and I took a first step in this direction, combining just two interventions – a senolytic and a stem cell therapy. Since they fired both of us, it’s not gone so well. But the idea was just about that – only just. And now we can throw in a couple of other things, and it’s time to put our pedal to the metal on this.
Tell me about your new foundation, LEV (Longevity Escape Velocity) Foundation. By the way, it’s a bold name that tells me you haven’t abandoned the idea of reaching escape velocity.
Ha! Well, people don’t donate to Aubrey de Grey because they want the work they’re supporting to be timid.
The key thing is that we have moved on enormously over the past decade. Certainly, in the past 20 years since I started talking about rejuvenation. If you remember my journal, Rejuvenation Research, that I started back in 2004, on three separate occasions the publishers asked me to change the name of the journal “because people think it’s about cosmetics, and it’s harming the circulation”. And I had to fight back hard.
It was only about four years ago that I obviously won that battle. I actually wrote an editorial called “Rejuvenation Reclaimed”. That’s when we started getting actual conferences, and companies from the very top labs, using the word properly rather than in the way it had been used.
Although this battle is won, we still have to convince people that this is worth doing. And many people are deep-seated defenders of aging. They believe that we shouldn’t or can’t do anything about aging. That’s why they persist in going on with this completely unscientific nonsense about compression of morbidity, which is never going to happen to a significant degree, because the healthier you are, the longer you are going to live, irrespective of how long ago you were born.
It’s still, “Yes, we might be able to extend lifespan a bit by this kind of approach, but people are still going to get sick and die”. Of course, this is what’s going to happen if people get the first generation of rejuvenation therapies that add 20 to 30 years of extra life, and then we stop developing new therapies – but anyone who thinks we would stop has rather a lot of the history of technology to explain away.
To me, the time frame that I place on reaching longevity escape velocity, on getting those first 20 years or so – that’s the speculative part. But the idea that there will be any chance of failure to maintain longevity escape velocity once we get to it is completely crazy. And I don’t understand why people don’t get that. Of course, I do understand why people pretend not to get it.
So, now I’m doing what I always do, placing myself at the tip of the spear and taking the bullets, opening the doors that other people can walk through afterwards. And, yes, the name of the foundation is very much emblematic of that.
Another part of it is that I have a long-standing and public interest in cryonics. I believe that if we’re going to save so many lives in due course by rejuvenation therapies, we have the duty to save as many lives as we can of people who are not going to make the cut. So, we’ve also put quite substantial money into a couple of startups in the cryonics space.
How is LEV different from what we’ve seen before and how do you see its role in the longevity space?
How is it different from SENS Research Foundation? I’m actually in control. Also, SENS Research Foundation was created in 2009, when I didn’t know anything about how to run an organization or even how to tell whether it was being run.
I’ve learned a few things over the years, some of them the hard way. First of all, governance. The board of directors of SRF was chosen by my co-founder Mike Kope, who was a CEO with me, whereas the board of directors of this foundation were chosen by me with one thing in mind above all – a gold-plated track record of respect for donor intent, which is exactly what was flouted at SRF. Of course, in addition to that, they have a wide spectrum of skills you need on a board of directors. So, I’m very happy about how that’s going.
Also, just being leaner and meaner. Any organization that’s been around for a decade accumulates baggage. It becomes harder to make decisions, more bureaucracy. If I look back at the ghastly stuff that happened in 2021, I’m thinking that I probably haven’t lost much time. I did initially, but by virtue of having my own organization, I’m moving a lot faster than I would have been able to in SRF. I’m steadily resaving all the lives that SRF’s directors caused to be lost.
I understand that for the new foundation, advocacy and outreach are an important part. We at Lifespan.io obviously can relate, so please tell me more about it.
At SRF, it was already important for me to do advocacy and outreach, but what we actually did as an organization was very little. SRF’s outreach consisted mostly of fundraising, and other aspects were basically me going out and doing it on my own.
Here, we’re doing it a bit differently. We’re emphasizing it more – because we can. We’ve moved to the point where the conversation in the wider world, including the corridors of power, is a lot more sophisticated than it was. Lots of people have genuinely got the message that aging is a medical problem, and we might be able to fix it very soon.
At the moment, we’ve been funding two groups. One of them, the Alliance for Longevity Initiatives, is focused on Capitol Hill. They are interfacing very energetically with members of Congress to get legislation changed, and this can work. Just a week ago, there was a change to the rules governing the FDA, which allows them to approve drugs based on animal-free testing, with things such as organoids, and this was partly due to our work.
In parallel with that, because elected representatives care about having their finger on the public pulse, we’re funding a group called the Healthspan Action Coalition, which is led by the most amazing people, starting with Bernie Siegel and Melissa King. Bernie for nearly 20 years has run by far the premier networking event in regenerative medicine. Melissa was the first executive director of CIRM – California Institute for Regenerative Medicine. She also led the campaign to have that institute refunded at 5 billion dollars at the last round of California elections. They’re both amazing, and they’ve recruited more people.
I describe HAC as the antidote to the AARP. They want to speak to the same audience that AARP does, the older generation, but they want to speak in a language of hope rather than the language of fatalism.
Speaking of fatalism: as a veteran in the field, what is your mood overall? Do you feel despair or exhaustion sometimes, or are you mostly optimistic, or maybe both?
I have the great good fortune that, first, I have a lot of intrinsic energy and fight in me. It takes a hell of a lot to slow me down, to make me despondent or anything. Just as importantly, I have people around me with just as much fighting spirit as I have. No, I don’t ever feel significantly despondent. And, of course, looking at the data, you can’t deny that progress has been made.
I don’t like to praise myself too much, but I feel it’s pretty much unarguable that I’ve made a significant contribution over the years, and there are many more contributions that I’m in a position to make. I’m not going to just rest on my laurels and spend my time doing math in a hot tub.
What are the most promising directions in geroscience today? Specifically, but not limited to, I’d like to hear your thoughts about cellular reprogramming.
The fact that damage repair has become the dominant school of thought in terms of doing something about aging changes everything. It means that there’s no longer a kind of running battle between people who favor different theories of aging, whatever the hell that ever meant. At this point, everyone knows that a lot of things are going on at the same time, and they’re only weakly interacting with each other, and that a comprehensive approach is going to involve multiple different interventions applied to the same people at the same time. That’s huge. That means I don’t have to justify that to people anymore. But that also means that one can’t point to any particular direction that people are taking and say: this is the dominant, the most promising one.
As to partial reprogramming: obviously, there’s a huge amount of money in it now. It’s the main thing for Altos, for Retro, for New Limit (or so we think). This means, if it works, we will know that pretty soon. But if you ask me based on what’s known today, what my expectation is, I think that we’re probably going to need different ways to do it. Simply titrating the amount of Yamanaka factors that we express, whether by using mRNA, or by having inducible promoters or whatever… these things will stop the mouse or the human from getting teratomas, but I don’t think they’re going to stop them from getting regular cancers, simply because the body of even people your age or my age is already chock-full of cells that have spontaneously accumulated most of the mutations needed to become cancerous.
With partial reprogramming, you’re taking cells like those, whose “cytostatic cage”, so to say, meaning, their network of mechanisms that stops them from dividing inappropriately, is already impaired, and you’re whacking that cage with a sledgehammer. The occasional cell here and there is not going to survive that, it will be knocked into being cancerous. Of course, you’ll never notice that by looking at cell culture. What’s worse, you also can’t tell this by looking at mice, because they don’t live long enough. The cancers that kill mice start getting going really early on. So, this is my problem with partial reprogramming.
