Wednesday, July 8, 2015

An age-old surprise....or not

Yesterday's Big Story of the microsecond on the BBC's web page was the profound discovery by scientists that people age at different rates. This in itself is something that has been so obvious that it hardly merits a sentence, much less a whole story.  But it is a news report of an actual science paper in the journal PNAS.  There are some interesting aspects that are worth discussing, but why is this paper in a major journal?

According to the paper, "Already, before midlife, individuals who were aging more rapidly were less physically able, showed cognitive decline and brain aging, self-reported worse health, and looked older." As a result, we've got a new 'science', ''geroscience' coined as mentioned by the authors, but this seems to be much more about marketing than actual science, because we already have fields known as gerontology, the biology of aging, and so on.

Still, an important health-related objective is to anticipate disease before it has occurred so that it can be prevented.  If signs of aging can be detected early, and there are interventions that may slow it--or, even just if the person could be advised of his/her state so s/he can live to the fullest--then this will be progress.  It is, of course, a main idea behind the notion of 'precision genomic medicine'.

This paper reports on a study of 38 year olds who had had some baseline measures of basic health taken at age 26, so it's a report of change over 12 years.  Measures of cardiovascular, metabolic, immune and other health-status were taken.  It isn't clearly stated in the paper that we can find, but these individuals were presumably all 'healthy' at the time of the study.  The paper shows that if typical health-related criteria are used, there is a normal (bell-shaped) curve of biological age among these 38 year olds, as shown here:


Anyone knows, or should, that biological traits vary among individuals.  Most traits vary with age as experience, wear-and-tear, and so on occur.  Age-related variation can be because of genetic and/or environmental/lifestyle variation.  The normal distribution is so robust a general pattern of variation that its shape is no surprise either, and also not particularly informative per se, but it does reflect the fact that we vary in almost everything.

The different biological systems showed regular increases in average age-related measures over this 12-year period, as this figure, also from the paper, shows:

Again, there is absolutely no surprise in seeing that obvious age-related traits change with age on average.  But the authors also found that individuals who seemed biologically 'older' at age 38 had experienced greater age-related change in their systems over the 12-year period:

How risk factors affect risk
The notion that genetic or environmental risk factors raise risk is only part of the story.  Risk factors don't just change your risk.  Typically they change the 'hazard function', that is, the change of risk with time, that is, with age.  Exposure per dose-duration usually accelerates risk, often exponentially.  In other words, the more exposure the faster risk increases--the faster you 'age' with respect to when you may get the disease.

This is why, for example, those with high risk genetic factors usually are found initially by study of early onset cases of a given disorder.  Biological effects are ongoing and their effects, say tissue damage or somatic mutation, accumulate with exposure level and duration.

These generalizations have been known for many decades (I say this because even in the 1970's I was working on the biology and genetics of aging and these things were in the textbooks as well as research literature.  So it is not any sort of surprise to see the reported effects, and in that sense they should not have been treated as if they were new.  In essence, gradual risk increase with age is aging.)

This is not to say the results are wrong in any way, nor that following a cohort in this way is not useful.  The past literature dealt with multiple 'competing' causes in various ways, usually as if they were independent.  But it was of course known that if you were exposed to multiple risk factors then multiple causes would accelerate in risk and so on--as this paper reports.

Since we all vary in our genomes and in our lifestyle habits, our various systems will 'age' in ways that respond to those individual exposures.  Some factors seem to affect multiple tissues and their associated diseases (telomere shortening, for example, is argued by some to accelerate many different risks).  If we could identify the causes of these differences, we could in principle ameliorate them for the fast-agers. Then, of course, the 'geroscience' community would demand big studies to intervene in everyone, because even the slower agers could be made to age more slowly still.

We must also expect that when these measures are subjected to genomewide mapping---how could there not be a demand for funds to do this?--the traits will largely turn out to be genetically complex and rarely due to tractably simple genotypic variation.  Similarly, environmental/lifestyle variation will be shown to be important, if not the main factor, and we know how difficult such things are to change, or even to identify.

The authors clearly point out several limitations to this study, and do stress environmental factors that may be responsible and ascertainable.  They argue that animal models (such as mice, who are short-lived among other things) don't provide the kind of direct individual evidence we would like.  We already know of some specific genetic factors that accelerate what otherwise are age-related traits (people with Down Syndrome and other progerias show these symptoms, including bone, skin, and muscular effects), but it is debated whether these should properly be viewed as generic aging-related rather than simply tissue-specific effects of a given gene.  The 'pace of aging' in the above figure represents an average or total of the various measures.  So one needs to discern whether fast-agers in the PNAS study population were aging fast in all measures or just one or a few (and if the latter, were those measures related).

As with many things of this nature, a point we regularly note, relevant environmental factors may, if studies are lucky, be identified in retrospective studies like this one, but since we cannot predict lifestyles of the future, the studies may be interesting but often of unknowable relevance to people who are now young relative to their health futures.

1 comment: said...

There were three areas I expected to see covered that weren’t addressed in this study:

- Where were the links back to all of the relevant measurements and predictions made when these subjects were ages 3, 5, 7..? Other studies of these same subjects made such links, but only cognitive testing was linked back in this study. Were the researchers really trying to pretend here in 2015 that these dramatic later-life physical measurements were not effects of earlier-life causes?
- Where were the psychological measurements? Are we to believe that the subjects’ states of mind had no relationships to their biomarkers?
- I didn’t see any effort to use newer measures such as the degree of epigenetic DNA methylation. I’d expect that these subjects’ historical tissue samples were available. The reviewer certainly was familiar with newer biomarkers.