The landmark studies of Leonard Hayflick and colleagues established that embryonic cells have a finite potential to divide. Using embryonic fibroblasts, Hayflick found that these cells could only double about 50 times when cultured in vitro before they lost the ability to divide further and underwent growth arrest, also called “replicative senescence.”4 Subsequent studies extended this replicative limit on doubling number to a variety of other cells, a phenomenon termed their “Hayflick limit.” Striking correlations have been found between the natural lifespan of different animals and the Hayflick limit of cultured primary cells from those animals. Intriguingly, cells from patients with progerias have a reduced Hayflick limit.
Taken together, these observations suggest a causal relationship between replicative senescence and organismal aging. This inference should be approached with some caution, however. The correlation can disappear when patient samples are corrected for health status.5 With an estimated 236 cells in an average adult, fifty population doublings (250 cells) is more than enough to provide for even the longest recorded human lifespan. Indeed, most human tissues lack detectable telomerase and the cells of many tissues whose degeneration marks the progression of age, such as the central nervous system and the heart, show little or no evidence of telomere erosion.
Are telomeres limiting in replicative senescence? Here again, the scientific literature is replete with paradoxes. There is a correlation between telomere length and replicative senescence in cultured cells.6 On the other hand, mouse cells have a lower Hayflick limit than human cells and the typical lab mouse lives for only two years, yet laboratory mice have an average telomere length three times longer than that of humans. Mice engineered to lack telomerase activity are viable for up to six generations, clearly demonstrating that a normal mouse does not die from telomere insufficiency. Humans have the shortest telomeres, but the longest life spans, of all the primates. In a survey of over 60 mammalian species, the short-lived species tended to have the longest telomeres.7
Our germ line cells, the cells that produce sperm and eggs, express the telomerase enzyme. This insures that each generation inherits chromosomes with an adequate length of telomere repeat sequences to support the many cell divisions between fertilization and adulthood (See Figure 4). We are born with an average of ca. 10,000-20,000 nucleotides of telomeric DNA on the ends of our chromosomes, although the exact amount depends on the chromosome and the individual. Our stem cells, the cells that replace the regular loss of cells that turn over during our lifespan, also express telomerase. It is estimated that, without telomerase, our cells lose about 100 nucleotides from each telomere with each cell cycle.
However, there is evidence that aging is accompanied by the loss of telomeric DNA sequences, suggesting that the level of telomerase activity in somatic stem cells is not sufficient to completely compensate for telomere loss over decades. This possibility was first suggested by the finding that telomeres in human germ line cells are considerably longer than those of somatic cells.8
So if we could induce our somatic cells to express significant amounts of telomerase, would we live longer? While the clinical trials necessary to test this idea in human patients are not yet underway, the evidence based on studies of human cells in culture make this appear possible. Artificially inducing telomerase expression in primary retinal epithelial cells, foreskin fibroblasts, and vascular endothelial cells extends the proliferation of each of these cell types in culture for many cell divisions beyond their Hayflick limit.9 Importantly, these cells also lack markers for malignant transformation, so they are not being turned into cancer cells. Instead, they simply maintain the appearance and behavior of younger cells. If these benefits at the cellular level could be achieved throughout all the cells of our bodies, it would represent rejuvenation beyond that achievable by any cosmetic or surgical intervention.
While correlations have been reported between telomere shortening and age in humans, these correlations often disappear when the epidemiological data are corrected for other variables that also correlate with age. A recent review of the literature suggests that the use of telomere length as a biomarker for aging is premature.10 Longitudinal studies spanning decades of the lives of the same individuals would provide a strong test of the utility of telomere length as an aging biomarker.
At this point, telomere shortening could best be described as a frequent concomitant of aging, much like gray hair or wrinkled skin. Certainly, people with gray hair and wrinkled skin have fewer years of life left, on average, than those with smooth skin and pigmented hair. But nobody would say that gray hair or wrinkled skin cause aging, nor would anyone conclude that we would live longer if we just colored our hair and got a facelift.
The defining feature of cancer is uncontrolled cell proliferation. In effect, cancer cells are immortal. In culture, cancer cells escape the Hayflick limit. Unlike most of our somatic cells, 85-90% of cancer cells express telomerase, and compared to normal somatic cells that do express telomerase, cancer cells achieve much higher telomerase expression levels.
Are tumor cell telomeres the Achilles heel of cancer? Telomerase inhibition can induce senescence in certain cancer cells.11 Mice engineered to lack telomerase activity have a lower incidence of cancer than telomerase-replete strains, although they are not cancer-free.12,13 This contrasts with DKC patients, who have a significantly elevated risk for several forms of cancer. On the other hand, mice engineered to over-express telomerase do have a higher incidence of cancer.14 Thus, drugs that target telomerase would seem to hold great promise for cancer therapy. Importantly, since telomerase is dispensable in most cells, there should be little or no toxicity to the patient, in contrast with current cancer chemotherapies. Indeed, telomerase-targeted anti-tumor drugs (imetelstat, a lipidated nucleic acid that targets the catalytic site of telomerase) and immunotherapies (GV1001, GRNVAC1) are in clinical trials.
Unfortunately, 10-15% of cancers achieve telomere stability through a telomerase-independent mechanism. This mechanism, first described in yeast cells in which telomerase was genetically inactivated, is termed “Alternative Lengthening of Telomeres” (ALT) (See Figure 5).The ALT mechanism involves the use of the cell’s homologous recombination machinery to copy longer telomere sequences onto chromosomes with shorter telomeres. ALT-dependent tumors would be resistant to telomerase-based therapies. While ALT-dependent tumors are relatively rare, ALT-dependent cells could emerge among tumors cells under strong selective pressure from anti-telomerase therapy. There are currently no therapies that target this mechanism.
The case for telomeres as the “fountain of youth” for cancer cells is strong.15 The alternative lengthening of telomeres (ALT) pathway poses a challenge to telomerase-based anti-tumor therapies, but the evidence that targeting of telomeres of cancer cells should have a high therapeutic index is compelling.
Much less compelling is the evidence that telomere length, per se, is limiting for the human lifespan. This could reflect technical limitations in the published studies. For example, telomere lengths are typically estimated using blood samples. If leukocyte telomeres are not representative of telomere erosion in other tissues, important correlations could be missed. Also, in most studies, comparisons are made between average telomere lengths, while it is likely that the shortest telomeres are physiologically relevant for age-dependent senescence.
Recent work has shown a correlation between chronic severe stress and shorter telomeres, at least in white blood cells.16 However, studies by some of the same authors suggest that telomerase activity is enhanced in stressed patients and in stressed rats. Since enhanced telomerase activity would be expected to extend telomeres, the mechanistic connections between acute stress, enhanced telomerase activity, shorter telomeres and long-term health are still unclear. It is safe to say, however, that minimizing acute stress has long-term health benefits.
The most consistent predictor of longevity in every organism tested is caloric restriction.17 Thus, any model for a rate-limiting effect of telomeres on organismal aging must account for this observation. Correlations pointing to telomere length as a potential driver for organismal longevity are balanced by other studies showing a lack of a relationship, as noted above. The weight of evidence at present supports the view that telomere shortening could magnify the morbidity associated with other stressors and the pathophysiology of certain diseases. By this view, telomere-extending therapies would be expected to have only an incremental effect on prolonging life span. Furthermore, evidence from transgenic mice suggests that telomere-extending therapy could increase the risk of cancer. For the foreseeable future, the best hope for our longevity, as with our wardrobe, is to take good care of what we already have.