Rhonda: What do you think about some of the...so, there's consumer-available epigenetic, you know, aging clocks out there, like, what are your thoughts? Like, are they accurate? I mean, what's...and, of course, I know you're an advisor for Elysium, which makes one of them. So...

Dr. Levine: Yeah, no, this is an important question. And I think, you know, the hard thing with epigenetic age is, again, this is something people can't...you don't know the answer, right? So, you can take one of these tests and you get a number back and there's no way for the consumer to verify whether that's, number one, correct, I mean, if there is a correct answer, or if it's even meaningful. And I think, basically, there are two things that I think are really important when you're talking about epigenetic tests being used by individuals or even clinically. This comes back to the reliability of these tests. So, what I mean by that is, if I were to take the exact same test twice on the same day, will I get the same answer? And unfortunately, what we found is actually, if you use the original epigenetic clocks, you do not get the same, you get wildly different answers. They're actually highly unreliable and very noisy. So, we've taken blood samples. You can split them, like, the same sample, run it twice, and you can get upwards of eight-years difference in your epigenetic age, using traditional clocks.

So, actually, a few years earlier, and this actually came out from my work with Elysium because they saw this in their data first and then we went back and looked at it more. Because I thought, "Oh, this is the end of epigenetic clocks." So, if this is what's actually happening, they can't be useful for anything, unless you have thousands of people. And so, we actually developed a statistical method that completely removes all this technical noise, and I won't go into the math for people on the podcast, but basically, we can get this down to you can split the sample and now you're getting only about one-year difference at max. Most people are predicted exactly the same age on their two tests.

And so, what I would say...so, Elysium, as you mentioned, I'm no longer an advisor for them because I'm doing other stuff with Altos but I was an advisor for them, and they actually did care a lot about this reliability thing, this is why I was helping them to try and sort this out. So, at least for their tests they felt like, if someone took it twice, they would get the same answer. You know, assuming you're taking it twice within a short period of time. But most tests, I would say, on the market are using the old methods that are highly unreliable. And I don't know that but I would suggest to consumers to, you know, find out if there is data on the reliability of tests.

The second thing that's really important for epigenetic tests is something that statisticians call construct validity, which is just this idea of biological age is, what we call, latent. You can't see it, you can't truly measure it. It's not CRP where I know I'm trying to measure something very specific. So, it's can we try and approximate something that's not really measurable? And then the way we evaluate if we did that well is does it predict or track with things we would expect? So again, with epigenetic tests, it should predict things like mortality risk after you adjust out the chronological age. So, being higher/lower than your chronological age should be very predictive of mortality risk or disease risk. And people who are using these first-generation clocks, the ones trained to predict chronological age, are not as good at that.

So yes, there's a lot of tests on the market but I think it's really important to make sure you're using ones trained more like the second-generation clock, so, things like GrimAge or PhenoAge, but again, using these methods that make them more kind of reliable and take out the noise.

Rhonda: Are there that many epigenetic clocks that are consumer-available now, are there?

Dr. Levine: I think, I mean, I actually don't know the number but I constantly see different companies launching epigenetic age tests. There's at least, I would say, probably close to 10 on the market.

Rhonda: Wow. Is the one that was developed by Elysium, when you were advising for them, that was more for biological age, that was the...

Dr. Levine: Yeah. So, it was similar to the PhenoAge one but with this additional statistical method that removes the technical noise. So, that one we've shown is predictive of mortality above and beyond chronological age. And actually Elysium I licensing these systems' measures, so, I think they'll be putting those out too where you can get kind of approximation of aging in different systems and get your kind of ageotype kind of thing.

Rhonda: My questions kind of got many layers to it and it has to do with like, if a person is, you know, trying to measure their epigenetic age and they want to do a lifestyle intervention and then measure it again...so, there was a very very small study, an extremely small study, it was published by a gal that reached out to me, her name was Kara Fitzgerald. And she and her colleagues had taken a small sample of people and they underwent like an extreme dietary change where, you know, they were eating a lot of leafy greens, cruciferous vegetables, blueberries, but they also were eating animal meat, liver, like, organ meat, and eggs. No refined sugar. Meditation, exercise, probiotics. I mean it was just a kitchen sink, it was a lot.

And this was like...I think it was pretty short treatment, like it wasn't like six months or anything. I can't remember off the top of my head, maybe two months or something like that or three months, I don't know. But their epigenetic age, according to the clock they used, I think it was maybe the original...

Dr. Levine: [inaudible 01:09:18] the original.

Rhonda: So, I guess you kind of answered the question. But it had reversed by like 3 years or something. So, I guess the question is, you know, like, can you track interventions accurately? Can you, you know, make assumptions based on it, or what are your thoughts? Like...

Dr. Levine: Yeah, and I'm not saying this to speak badly about any study but I mean the great thing about that study is they made their data public. And we were actually able to go back in and show that the entire effect was noise. So, actually, once you do the statistical method that removes the noise, there was actually no effect of the intervention. And, of course, you know, these methods weren't available originally and they were using an original clock like they thought they should.

But I think this shows us that we need to be a little bit careful how we interpret the results from these things and the epigenetic clocks are powerful tools to kind of...I think they are good at giving you an idea of your health status, same as if I go into my doctor's office for, you know, a kind of lab test and do a metabolic panel or whatever. They can give you a good idea of your status. But when we think of how they're applied to interventions, I think we need to be careful that we don't take any change at just face value, we need to really think about, you know, "If I measure it again, was that sustained?" like things we talked about earlier. Because I would argue that we haven't proven that a change in the epigenetic clock is a change truly in your biological aging process. And like we said, you can have changes in kind of inflammatory markers that are acute and not truly capturing the aging process but that'll change what you see in terms of your clock.

So, I think epigenetic clocks are really helpful for like a wake-up call. So, you know, a lot of young people are kind of going about their life, we don't know, like we said, till you see these dysfunctional things start happening, if we're doing well or not. And I think they can be used to kind of inform. And as we move forward and develop better more sophisticated ones and we look at actually linking changes in epigenetic aging to changes in other kind of health parameters, they will be good for kind of people testing interventions, either personally or in clinical studies. But I think right now we're in the really early days and these are really high-potential tools but we're not at the point yet where you can just, you know, take it and believe everything that you get back.

Rhonda: Right. So, you're better off also doing all the classical biomarker tests too and measure lots of things?

Dr. Levine: Yeah.

Rhonda: It's always good to have more data. Before we wrap this up, two questions. What are you most excited about with respect to the aging field in general, like what's being researched right now, what's coming out of the pipeline? And then what are you most excited about coming out of your lab? Or I guess they could be the same answer.

Dr. Levine: Yeah. No, they're related. So, the thing in the aging field, again, is this kind of transient or partial reprogramming thing. And just, again, not necessarily for an intervention but just figuring out what it is. To me, it's just so magical that you can change the state of a cell. And what does that mean for the cell, and does it now function better, and do populations of cells work better together, and how does that happen? Can you figure out even better ways to do this? You know, we've used kind of the Yamanaka factors but there could be, you know, tons of other ways that you could actually change this and just not knowing that that's possible. So, that kind of basic science I'm the most excited about.

And then, in my lab, I'm just really excited to figure out what the epigenetic clocks are. We have no idea, we apply these measures to a bunch of different things, we have no idea what drives these changes or, functionally, what they even mean. So, why is your epigenetic clock related to your mortality risk? It's like what is the pathway that links those two things? And I think that'll keep us busy for a really long time, that's something I'm excited to kind of start working on.