October 15, 2013

The notion that telomeres play a central role in both age-related disease and aging itself is generally misunderstood and is often criticized without an actual understanding of either disease or telomeres, yet there is a growing sense of the obvious about the role of telomeres in human aging and disease. More and more people – […]

Telomere Misconceptions

The notion that telomeres play a central role in both age-related disease and aging itself is generally misunderstood and is often criticized without an actual understanding of either disease or telomeres, yet there is a growing sense of the obvious about the role of telomeres in human aging and disease. More and more people – in the street, in the clinic, or in the lab – have the working assumption that telomeres somehow define aging and that the relationship is obvious, which is certainly an overstatement.

We are living through the reality of Schopenhauer’s ironic remark that all truth has three stages: first ridicule, then violent opposition, and finally the belief that it was completely self-evident and it was obvious all along. Nevertheless, even those who assume the importance of telomeres to aging – even those with research or clinical backgrounds – often share one of the many common misconceptions.

Misconception #1: it’s telomere length that matters

The first inaccuracy is that telomere length defines age, which is almost true, but entirely wrong at the same time. The critical measure is not what you have, but what you lost. Putting it more accurately, the length of the telomeres has nothing to do with age, rather it’s the change in telomere length that determines cell aging. The fact that mice have long telomeres is often cited as evidence against telomeres as determining aging in an organism, yet the absolute length is irrelevant. If, however, you watch the change in telomere lengths from birth to senility – in mice as well as other organisms – the correlation is clear and unambiguous. The key to aging is not the telomere length per se, but the way in which a shortening telomere causes changes in gene expression: as the telomere shortens, gene expression changes, DNA repair and molecular recycling (e.g., protein production and breakdown) rates both slow significantly. The result is a gradual increase in the percent of errors in both DNA and biologically critical molecules, for example SOD, elastase, and others. These errors have a profound effect on the cell as a whole, but also undercut mitochondrial function: more free radicals are produced, more escape, fewer are scavenged, and less damage gets repaired when it occurs. Aging occurs not because you have short telomeres, but because your shortening telomeres have caused your cells to stop dealing with the day-to-day damage at an effective rate. For those of you who want a longer discussion and formula describing molecular turnover within aging cells, I refer you to my first book, Reversing Human Aging, but the quick formula is this:

Assume a constant rate of cell damage (here 1%/unit time) and that anabolism  equals catabolism (i.e., there is a constant pool size). Then if the percentage of damaged molecules equals  and the percentage of molecules produced per unit-time equals M, then the formula defining the accrued damage is:      

 = 1 + [ (100-M)/100]


if  M = 50% in a young cell with a high turnover rate, then  = 2% (very little damage)

if M = 2% in an old cell with a low turnover rate, then  = 50% (critical damage)

Misconception #2: chromosomes unravel, cells die, and that’s aging

A related inaccuracy is common to almost all the news stories that try to explain telomeres to the general public: “telomeres shorten, the DNA unravels, and the cells die.” Chromosome simply don’t unravel in normal aging. To the contrary, cell aging begins long prior to the telomere’s putative disappearance. Only in the most extreme cases (such as the F5 generation of telomere knockout mice) do cells ever “lose all their telomeres”. It simply doesn’t happen in normal organisms as they age. The most important change, alluded to above, is not that telomeres are lost (or that the chromosome is at risk), but rather that the shortening telomere has caused a changing pattern of gene expression, usually referred to as the telomere position effect (TPE). Not only do a great many genes near the telomeres change expression as the telomeres shorten, but these genes go on to change the expression of more distant genes as well, including critical changes in the master epigenetic regulators, such as histone methyltransferases and DNA methyltransferases.  This effect occurs long before growth arrest occurs, long before you “run out of telomeres”, and long before you get “frayed chromosomes”. In real life, your chromosomes are actually in pretty good shape even if you live to be 120 and they only time they actually fray is during decomposition. The result of shortening telomeres is an aging cell, but not because the chromosomes unravel. Not at all. The chromosomes are fine, thank you, but the pattern of expression has changed remarkably and to the detriment of cell function. Cell aging is caused by a “continuous spreading of telomeric heterochromatin” which can cause either a decrease in gene expression or an increase in gene expression, depending on the gene in question One example of this effect occurs in Facioscapulohumeral Muscular Dystrophy (FSHD) where the gene is only 25-50 kb from the telomere, although much more distant effects have been documented as well.

Television personalities notwithstanding, your chromosomes don’t come apart, but your changing pattern of gene expression does result in your body’s “coming apart” as you acquire age-related disease.

Misconception #3: there are “aging genes”

This brings us to a third inaccuracy, quite common to the research world: the idea that “genes cause aging”. Putting it succinctly, there are no aging genes, only aging patterns of gene expression. Aging is not genetic, it’s epigenetic. Just as the difference between your nose and your toe is not a matter of having different genes, but having the same genes with a different pattern of gene expression, so too the difference between your body at age 5 and at age 95 is not different genes, but a different pattern of gene expression. There are no more “aging genes” than there are “nose genes”. It’s quite true that aging and noses are both dependent upon genes, but they are defined by the pattern of how those genes are expressed, not by specific genes for a body part or for aging. You might better say that ALL genes are aging genes: the epigenetic pattern of aging results from changes throughout the chromosomes, not from a few “aging genes”. It’s a gestalt, not a few genes.

The same misconception occurs in regard to age-related diseases. One example is that of Alzheimer’s dementia in which genetic predilections do occur, but even apoE4 is not an “Alzheimer’s gene”: it’s a normal gene, although associated with an increase in the risk of Alzheimer’s as you age, although not always. Also, thinking of Alzheimer’s as a “genetic” disease merely looks at a single, static set of parameters – your genes – and ignores the epigenetic changes that occur dynamically over time. The result is that genome-wide association studies (“GWAS”) identify a great many gene associations with neurodegenerative diseases, but these genes explain only a tiny portion of the risk of actually getting the disease, as occurs in the case of apoE4. Alzheimer’s disease, coronary artery disease, stroke, osteoarthritis and the entire gamut of age-related diseases are not “genetic diseases” but are more appropriately and usefully thought of as “epigenetic diseases”. Only when we recognize the subtle – but critical – difference can we begin to prevent and cure these diseases effectively.

Misconception #4: telomere length(s) predict length of life

This misconception arises not because there isn’t an element (a strong element) of truth in the statement, but because the statement implies several wrong assumptions, so it ends up being not only wrong, but essentially meaningless.

Let’s look at the truth first. To do so, we’ll make several naïve (and usually unrealistic) assumptions:

  1. You are only going to die of one particular age-related disease, such as coronary artery disease, and no other disease whatsoever, no matter what you do to yourself.
  2. The only predictor of your having a heart attack is the state of your coronary arteries and the only predictor of your coronary disease is the telomere lengths of your vascular endothelial cells (rather than lifelong diet, etc.).
  3. Your coronary artery disease is gradual and accumulative – and the rate of shortening is invariant despite any short term changes in diet, exercise, or other risk factors (such as smoking, blood pressure, or cholesterol to name just three risk factors) until some predictable “tipping point” occurs when your endothelial cell telomeres reach a certain precise length and then occlusion inevitably and instantly occurs.
  4. Finally, despite the inevitable variation that will be present in those telomere lengths, there is a particular length measurement – whether it’s the shortest telomere or the mean telomere – that is the precise and only predictor of that tipping point.

You see how unrealistic this set of assumptions can be, but if these were all true (i.e., never) then we might reliably predict your actual lifespan based on your telomere lengths. Of course – and perhaps this is the most unrealistic idea of all – we would have to do biopsies of your coronary arteries to measure the important (i.e., predictive) cells that cause disease, something that very few patients (in fact, very few physicians of those patients) are willing to have done to them. Blood sample, yes; coronary biopsies, not me!

The reality is that most of us have multiple age-related diseases in progress in multiple tissues and organs, any of which might finally kill us. Predicting exactly which organ is most likely to kill us is strictly a statistical bet, not a reliable prediction. Even if we knew (how could we?) that you would definitely die of a coronary artery occlusion and not die of any other problem, such events are not merely probabilistic, but they are actually stochastic: they are affected by random events such as small clots or transient inflammation that are not strictly speaking a local problem of the coronary arteries at all. Rather they involve other tissues and organs, as well as transient whole-body events, such as infections. As a result of such stochastic events (and other factors, such as dietary changes, medications, behavioral risks, etc.) the risks do not climb in a gradual curve, but rise and fall in spurts and unpredictable jumps that defy mathematical precision – in fact they defy mathematical imprecision for that matter. Moreover, since there are innumerable cells in your coronary arteries, each with its own set of telomeres, and since each endothelial cell has 23 chromosomes, each with four telomeres, which telomere in which endothelial cell would we use for predicting lifespan? Or do we have to remove every single cell and measure the telomeres? Finally, we can’t actually (at least practically) sample the endothelial cells on real patients in the first place.

Most clinical telomere assays are based on easily available cells, such as oral swabs or white blood cell samples. Such samples may be – quite loosely – correlated with the telomere lengths of your coronary endothelial cells, but the correlation is only fair and nowhere near precise enough to be an accurate predictor of your coronary artery disease, let alone predict your lifespan.

There is one other problem in using serum samples, particularly if we do them serially, in an attempt to see “how things are changing with time”. For example, we might measure the telomere lengths in your lymphocytes, then have you change your lifestyle, then measure them once again in order to show that the intervention has “extended your telomeres”. Unfortunately, such measures don’t show what you might think. Even if your telomeres are several thousand base pairs longer after the intervention, that would not show that your telomeres had actually “gotten longer”. Surprising, by true.

Consider an analogy. Imagine that instead of measuring telomere lengths in lymphocytes, you were to measure the age of inhabitants of one particular city block over time. The first time we interview those living on that block, imagine the area is economically depressed and populated largely by elderly retirees with an average age of 75 years. Now, over a twenty year period, we use a “clinical intervention” and we change the “lifestyle” of the city by offering tax breaks for new start-up companies, repair the old streets, increase the police presence, and so forth. As a result, perhaps more young professionals move into the area and those young people have even younger children, resulting in an average age of only 25 years. Have we “reversed aging” in the population? Not at all. Certainly, the average age has dropped fifty years, but not because we turned back the aging clock in the inhabitants (or even any single one of the inhabitants), but because the old inhabitants died and a new population of younger people have moved into the area.

The same process occurs with regard to circulating white cells in your blood. Even if the “average age” of the white cells (i.e., the average telomere length) has improved, it is not because we have relengthened any particular telomere, but because we are sampling a different and newer population of white cells. In short, serial sampling of peripheral white cells is not as reliable as we wish it were. In real life (i.e., in clinical measurements of peripheral white cell telomeres), it’s not as bad a watching an urban neighborhood over a fifty year period, but the assumption that we can relengthen telomeres in peripheral cells is usually based on wishful thinking rather than on an understanding of white cell population dynamics. We can certainly measure peripheral white cell telomere lengths over time, but interpreting differences in telomere lengths is chancy and should be taken with a grain of salt. It rarely means what researcher would like you to believe it means.

Even is all of this weren’t enough to undermine our faith in the value of telomere lengths as a predictor of lifespan, as I pointed out above, the key issue isn’t telomere length in the first place; the key issue is the change in telomere length. It’s the change in length that changes the pattern of gene expression, that underlies age-related pathology in the first place. In short, I might have much longer telomeres than you do and – even if we could overcome all the other issues discussed in this section – I might still age faster than you do and I many die of any number of age-related diseases long before you reach a comfortable and healthy middle age.

If we wanted to give our best guess as to lifespan (or at least a better correlation), we should be measuring the telomeres from every organ of the body, with an emphasis on those tissues that are most likely to show age-related pathology that results in mortality: coronary vascular endothelial cells, microglial cells of the brain, etc. While any reasonable person, faced with the above problems in using telomere lengths to predict aging and lifespan would probably throw up their hands and surrender, the fact is that even serum-based telomere samples are actually somewhat correlated to lifespan.


Misconception #5: Some aging disease can’t be related to telomeres

Almost invariably, someone – almost always someone without any knowledge of human pathology – will argue that “telomeres couldn’t possibly cause heart disease and Alzheimer’s dementia!” This naïve criticism reflects a number of accurate, but irrelevant facts. In the case of coronary artery disease, the argument goes like this:

  1. telomeres only shorten during cell division, so non-dividing cells don’t lose telomere length as you age,
  2. cardiomyocytes almost never divide, so their telomeres certainly won’t shorten as you age, and
  3. so heart attacks can’t possibly be due to telomere shortening!

In the case of heart disease, the summary of this criticism would be as follows:

  • Telomere loss can’t cause heart attacks because heart muscle cells don’t lose telomeres.

Unfortunately, this is almost exactly like saying:

  • Cholesterol can’t cause heart attacks because heart muscle cells don’t accumulate cholesterol.

Both statements are not only foolish and misleading; both statements are also  wrong. It’s not the heart muscle cells that underlie the pathology, it’s the coronary arteries, which DO lose telomeres and DO accumulate cholesterol. The fact that cardiomyocytes don’t divide is irrelevant to the pathology of heart disease. People don’t die because cardiomyocytes divide and lose telomeres, they die because the cells lining the coronary arteries divide and lose telomeres. Likewise, people don’t die because heart muscle cells accumulate cholesterol, they die because the coronary arteries accumulate cholesterol.

The same is true with regard to Alzheimer’s dementia. Neurons may not (generally) divide nor their telomeres shorten, but microglial cells divide continually and their telomeres do shorten with age. Microglial telomeres shortening correlates with Alzheimer’s disease and appears to precede the onset of amyloid deposition and other hallmarks of the dementia. Central neuron function is critically dependent upon microglial function in the same sense that cardiomyocyte function is critically dependent upon the function of coronary vascular endothelial cells. In both cases, the neurons and the cardiomyocytes are the “innocent bystanders” in the pathology and in both cases the pathology begins in cells that show telomere shortening with age.

These two major age-related human diseases – atherosclerotic disease and Alzheimer’s dementia – are both examples of “indirect pathology”. The pathology of the end organ – dead myocardium and dead neurons – is indirect and is due to primary pathology that occurs in other cells, in this case the vascular endothelial cells and the microglia.

There are also a host of “direct pathologies”, in which the cells that divide and lose telomeres are also the very  cells that underlie the final pathology. In the case of osteoarthritis, for example, the cells that divide and lose telomere length – the chondrocytes – are the same cells whose loss defines the actual clinical disease – the loss of joint surface. Direct pathology includes osteoarthritis, osteoporosis, immune senescence, skin aging, and a number of other clinical problems. Of course, in most of these diseases, there are interactions with other body systems that may play a significant role. For example, skin aging is due to cell aging of the primary skin cells – keratinocytes and fibroblasts for example – but it is also partially due to the loss of small vessels throughout the skin as the vascular endothelial cells age. The same is almost certainly true of most dementias, which often have a vascular component, whether because of decreased blood flow, micro-infarcts, or micro-hemorrhages. Although pathology is rarely simple, the rough distinction between indirect and direct age-related pathology will prove useful as we move into the realm of clinical intervention using telomere extension.

Misconception #6: Telomeres “cause” aging

Telomeres no more “cause” aging than the engine of your car “causes” you to get to work in the morning. More critically, the use of the word “cause” usually represents sloppy thinking, either because the word is simply wrong or because the underlying processes are so complex and sophisticated that simply referring to “causation” is entirely misleading. If two things are correlated, we should say so, rather than assume causation. If the relationship is actually a cascade of events,  a complex web of processes that really does involve “causation” (rather than just correlation) then we need to be extremely careful in using a naïve word to represent a far more sophisticated reality. Cholesterol, for example, doesn’t actually cause coronary heart disease, but it does a significant role within a complex cascade of pathology. As part of this complexity, remember that some people with low cholesterol have significant coronary disease (e.g., progerics) and other people with critically high cholesterol may not have any measurable coronary disease. Causation is rarely simple.

Finally, causation is often not the issue and discussing causation at all (even carefully), often misses the point. This is particularly true in the case of human disease, where the key question is “what causes this disease”, but rather “can we intervene?” In the clinical world – the world of real patients and their health problems, intervention is the primary concern and causation plays a role only where it contributes to intervention (which it usually does, but again, only secondarily because it helps intervene). The issue of causation is not primary and is relevant only as it support the primary question of intervention. Many people have disagreed with the statement that telomeres “cause aging”, but they seldom understand either the data, the complexity, or the primary issue. Aging is full of correlations that result in misattribution of cause; aging is enormously complex with an interacting web of causation behind every definable pathology. Moreover, the primary clinical issue is intervention rather than causation, which is only a contributory (even if necessary) issue.

In this sense, discussions of causation are bootless exercises in philosophy (except as they further intervention), better reserved to conversations over wine than reasoned discussions of science or medicine. The most common issue with causation per se, however, is that many prominent “theories” of aging aren’t actually testable. One of the critical characteristics of the contention that telomeres play a central role in (not “cause”) aging, is that not only does all current data support this contention, but that we can actually test the theory both in vitro and in vivo. Not only does this result in good science (a non-testable theory isn’t science), but it meets the criteria for good medicine (we can use it to help patients get healthier).

Leave a Reply

Your email address will not be published. Required fields are marked *