Data Mining Reveals Human Aging Clocks in Blood Proteins

Human plasma protein levels change with age and can act as predictive aging clock

The ticking clock

Even if you’re generally healthy, once you reach a certain age, slowly but surely you’ll begin to notice the deterioration of various bodily functions and abilities.

No matter how hard we try, eventually we all fall to the inexorable influence of father time. Joints begin to complain, aches and pains become a part of daily life, health markers during our routine blood checks start to act wonky…

One by one — or all together — our body’s systems decide it’s been enough.

Aging affects all parts of the body, it is a systemic process, a multi-factorial issue. Multi-factorial also implies that a lot of data has to be gathered in orde to makes sense of the mechanistic details of what is actually happening.

Machine learning and data mining are particularly suited to deal with such issues. In fact, in a previous post, we looked at implementing machine learning in aging research. In that post, one of the possibilities we discussed involved the identification of biomarkers, certain ‘signals’ that could tell us our chronological age, as well as our potential lifespan.

Not much later, a study in mice appeared where machine learning was used to develop two ‘aging clocks’. One of those clocks was a strong predictor of chronological age, and the other one accurately predicted lifespan and the effect of potential lifespan-extending interventions. But, as noted in the post about that study, mice are not human (not entirely anyway).

Human proteins as aging clocks

In humans, it’s already known that blood protein levels change with age. Quite reliably, even:

Waves of changes in the proteome [the full set of proteins] in the fourth, seventh and eighth decades of life reflected distinct biological pathways and revealed differential associations with the genome and proteome of age-related diseases and phenotypic traits.

Now, a new study ups the ante.

Based on a dataset of blood plasma proteins of over 4,200 people, the researchers first tested 529 proteins that are known to change their expression level with age.

They found that at least 64 of these proteins can regulate lifespan in animal models, and 9 of them are known to significantly extend lifespan in mice or fish when manipulated.

Next, the scientists performed machine learning analyses (on a dataset of over 3300 people) to find out which protein combinations strongly correlated with aging — so-called aging clocks.

They found at least twelve. The strongest one incorporated information on no less than 491 protein levels. This clock correlated very strongly with age, and the median error was around 2 years (a bit lower in the training set, a bit higher in the real dataset). That’s pretty darn good.

Two interesting observations were made as well:

• Physically active individuals were consistently ‘aged’ younger by the clocks. We can’t stop the clock from ticking, but physical activity does seem to slow it down a little bit.
• A lot of the most predictive proteins play a role in the immune system and signal transduction.

That’s all intriguing, but why would we want such ‘aging clocks’? They may tells us how old we are, but they can’t really predict when we’ll pass away, right? In the words of the authors:

Given that it is not realistic to perform life span studies in humans, a prominent appeal of aging clocks is their potential ability to accelerate anti‐aging clinical trials…

With the ultimate objective of improving human health span in mind, we sought to better understand proteomic aging clocks and to identify high‐quality protein targets that exhibit anti‐aging clinical potential.

What is especially noteworthy in this quote is the focus on healthspan rather than lifespan. It is, in other words, about living healthier for as long as possible rather than about simply living as long as possible. Sounds similar, but the difference matters.

Of course, that is not to say that the health- and lifespan cannot be correlated. Perhaps living healthier for longer also makes you live longer overall. Seems probable.

Of course n° 2, we already know some things that can help us with living healthier for longer: eat well, move a lot (gently), and pick the right parents. (Naturally, in case of illness or genetic conditions, these things will only get you so far…)

Of course n° 3, even if we find ways to (drastically) extend health- and/or lifespan, there will be difficult questions to address about overpopulation, retirement, resource use, and the availability of life extension technology.

Tick tock…