| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||
a Gerontology Division, Beth Israel Deaconess Medical Center, Harvard Division on Aging, Boston, Massachusetts
Thomas Perls, Beth Israel Deaconess Medical Center, Gerontology Division, Rabb 417, 330 Brookline Ave., Boston, MA 02215 E-mail: thomas_perls{at}hms.harvard.edu.
Decision Editor: John E. Morley, MB, BCh
IN 1825, Benjamin Gompertz proposed that mortality rate increased exponentially with age. Since then, most researchers have accepted that this rule indeed applies to the age range around the average life expectancies of many species, humans included. However, at extreme old age an exception to the rule exists. At very old age, mortality begins to decelerate in species such as medflies (1), Caenorhabditis elegans (2), and humans (3). Why does mortality decelerate? Most likely it is because frailer individuals drop out of the population, leaving behind a more robust cohort that continues to survive. Because these frail individuals drop out of the population, the distribution of certain genotypes and other survival-related attributes in a cohort changes with older and older age (4). This selecting-out process is termed demographic selection.
The effect of demographic selection is exemplified by the drop out with extreme age of the apolipoprotein E 
-4 allele (5). Rebeck and colleagues noted the frequency of the -4 allele to decrease markedly with advancing age, even among nonagenarians and centenarians with dementia (6). One of its counterparts, the -2 allele, becomes more frequent with advanced age. Consistent with these observations are the findings reported by Frisoni and colleagues in this issue (7). Presumably the drop out at earlier age of the -4 allele is because of its association with "premature" mortality secondary to Alzheimer's disease and heart disease.
A similar trend exists in the case of the apolipoprotein B locus, for which investigators from the Italian Centenarian Study comparing 143 centenarians with younger controls demonstrated an association between specific multiallelic polymorphisms and extreme longevity (8). In another study, nonagenarian subjects had an extremely low frequency of HLA-DRw9 and an increased frequency of DR1. A high frequency of DRw9 and a low frequency of DR1 are associated with autoimmune or immune-deficiency diseases, which can cause premature mortality (9). Tanaka and colleagues demonstrated single nucleotide substitutions in three mitochondrial genes that were present in the majority of a small centenarian sample but rare in the general population (10). This study requires verification in other populations given its small and homogeneous sample. Franceschi and colleagues, analyzing mitochondrial DNA haplogroups from 212 healthy centenarians and 275 younger controls, found that in male centenarians from Northern Italy, the J haplogroup was ten times more frequent than in controls; that is, 23% versus 2%, with p = .005 (personal communication, October, 1999).
Such findings suggest that centenarians are ideal subjects for the discovery of other polymorphisms (or lack of polymorphisms) associated with a survival advantage. Furthermore, the apolipoprotein E -4 allele serves as an example of a polymorphism with an influence powerful enough to have a noticeable effect upon survival in the general population and across various ethnic lines. Wachter recently addressed the question of how many other such polymorphisms might exist. He argued that, based upon the significant differences in mortality risk between people in their 90s versus those aged greater than 105, "there could be a relatively small number of genes—hundreds, not tens of thousands—which one has to not have in order to survive ad extrema." Of course, having the right polymorphisms may be as important as lacking the wrong ones (11).
The importance of lacking the wrong polymorphisms has been elegantly demonstrated by Tomita-Mitchell and colleagues, who used American mortality records to calculate the expected decrease in single nucleotide polymorphisms coding for premature mortality in newborns compared with centenarians (12). Thus, though Frisoni and colleagues propose that the increase in -2 allelic frequency with extreme age nominates this allotype as a longevity assurance gene, it is also plausible that it is the lack of the -4 allele that matters. Of course, another alternative might exist, in which it is not the apolipoprotein E allotype that matters, but rather a polymorphism of some other gene that maps along with it.
Finally, in general, the term "longevity assurance gene" misleads one to believe that such a gene ensures longevity. As the saying goes, the only for-sure things in this world are death and taxes. Certainly a host of environmental and other genetic factors can influence a person's mortality risk in such a way as to negate the effect of any such assurance. Thus, the term longevity enabling or predisposing gene is probably more appropriate.
 |
Nature Versus Nurture |
|---|
 Top Nature Versus Nurture References |
|---|
A number of findings strongly argue against natural selection's indifference to aging and argue for selection pressure to develop longevity enabling genes. Principally, experiments in yeast (17)(18), in the nematode C. elegans (19)(20)(21)(22)(23)(24)(25)(26), in Drosophila, and in mice suggest that a few alleles can exert a powerful influence on life span. Specific mutations in C. elegans (daf-2, daf-15, daf-23, age-1, and clk-1) increase the nematode's life span up to three to five times (19)(20)(21)(22)(23)(24). The daf genes and age-1 regulate the formation of the hibernating dauer state in C. elegans. The dauer state is an alternative larval stage that allows the C. elegans worm to survive periods of low food availability. Well-fed worms live for about 3 weeks, but dauer larvae can live more than 2 months. The human homolog of daf-2 is the insulin/IGF-1 receptor (27). The clk-1 gene controls cell-cycle duration and adult rhythmic behaviors, and when altered in the laboratory it slows metabolism, also markedly increasing life span (28).
In Drosophila the mth (methuselah) mutant strain survives 35% longer than wild types (29). Overexpression of Cu/Zn super oxide dismutase (SOD) can increase the maximum life span of transgenic Drosophila up to 48% (30). SOD plays a critical role in lower organisms and humans in the scavenging of oxygen radicals (31). Likewise, longevity-associated daf mutations are associated with an increased resistance to reactive oxygen species.
In another important finding relating to oxidative damage, Migliaccio and colleagues recently identified a mutation in the mouse SHC gene ( p66shc) that is associated with a 30% longer life span (32). The protein product of this gene responds to reactive oxygen species. Ablation of the gene reduces the apoptotic response to oxidative damage induced by hydrogen peroxide or ultraviolet irradiation. Mice with the mutation are also less susceptible to the effects of paraquat, which causes rapid and lethal oxidative damage in wild-type animals. These and other findings in mammals and lower organisms support an important role for oxidative damage in aging and its modulation by specific genes (33).
Several human studies also suggest a significant genetic component to the ability to achieve exceptional old age. McLearn and colleagues noted an increased heritability of cognitive functional status at older ages, suggesting a demogaphic selection for genetic polymorphisms associated with both increased risk for survival and decreased risk for dementia (34). From the New England Centenarian Study, a review of 102 pedigrees of centenarian probands compared with 77 pedigrees of a similar birth cohort who died at age 73 (average life expectancy for that group who also survived beyond age 20) found that the siblings of centenarians had a greater than four times risk of living to age 92 years (35). A statistically nonsignificant trend was observed in which the relative risk climbed to 10 times to achieve age 95 and 15 times to achieve age 100. As the study continues to collect more pedigrees from both groups, we anticipate statistically confirming this trend. Do these statistics indicate that it is worthwhile to conduct a centenarian sibling-pair study? Certainly if we believed that centenarianism was a monogenic trait, a linkage study would be an obvious pursuit. However, a complex trait is more likely, making such a project more risky. Another line of evidence makes taking such a risk more compelling...
In the process of enrolling centenarians, the New England Centenarian Study has discovered several families highly clustered for extreme longevity. Mathematical analyses using cohort life tables indicate that assuming a homogeneous population, these sibships with multiple centenarian and nonagenarian members would occur just by chance at a much lower frequency than could be encountered in the world today. A number of observations including the fact that some siblings were reared in different parts of the world and spouses lived to average life expectancy make it likely that there is an important genetic influence contributing to such clustering (36)(37). But how important? Clearly the population is not homogeneous and the sibships must have important environmental factors in common. Arguably, environment must still play the predominant role in survival even to extreme old age. However, a gene that could explain even 7–10% of the variation in life expectancy at very old ages would be of keen interest. Such a degree of variation might be expected from a gene found by a linkage study to be statistically linked to the phenotype of exceptional longevity.
The search for longevity enabling genes might therefore be a worthwhile pursuit. Perhaps the most conservative approach is determining frequencies of polymorphisms or point mutations of candidate genes among centenarians compared with appropriate controls, as performed by Frisoni and colleagues (7). Several groups have improved upon this approach by combining it with survival analyses (12)(38). An association study among members of specific families demonstrating striking clustering is tempting but unlikely to succeed for a multigenic trait. A linkage study of sibling pairs or sibships should be feasible and worthwhile if enough sibships can be found and enrolled.
Pursuing a long and involved, as well as expensive, sibling pair study begs the question of what the utility of finding a gene and polymorphism common to centenarians would be. Rather than fitting the myth of the older you get the sicker you get, centenarians typically achieve their age by living 90–95% of their lives in very good health, thus escaping, or at the least markedly delaying, diseases normally associated with aging (39). Discovering genes that could impart such an advantage should yield a great deal of understanding of how the aging process increases susceptibility to diseases associated with aging and how this susceptibility might be modulated.
The compression of morbidity and functional impairment toward the end of life observed among centenarians is supportive of Jim Fries's compression of morbidity hypothesis (40). Fries proposed that as the limit of human life span is approached, the onset and duration of lethal diseases associated with aging must be compressed toward the end of life. However, this brings up an important point about another work presented in this issue by Arai and colleagues (41). In their study, centenarians were studied for levels of insulin-like growth factor-1. An inherent problem in studying levels of practically anything in centenarians is that the majority of these subjects are in the morbidity phase of their lives with an annual mortality rate between 30% and 50%. Except in the case of truly healthy centenarians, these studies do not reflect levels conducive to living to 100, but rather levels associated with the last year or two of life. Thus, findings such as low IGF-1 levels among demented subjects could be explained by the subject's frailty and have no association with longevity or dementia.
An interesting finding in the Arai study that also appears in a number of other centenarian studies but that has not been discussed is that the men tend to be better off than the women, both in terms of physical and cognitive function. This would at first seem paradoxical since women seem so much better able to achieve extreme old age. In the United States and other industrialized countries, women comprise the vast majority of centenarians at a prevalence rate of about 85%. One explanation may be that compared to women, men have to be in particularly good condition to achieve extreme old age. Those who are not, die before reaching the centenarian mark. These observations may represent a demographic crossover in terms of functional status in which women are better off than men at younger old age, but then men, though much fewer in number, become better off at extreme old age (4).
Much time, effort, and expense have been spent in the search for biomarkers of aging. However, confounding, environmental factors, and heterogeneity are just some of the factors that have made such a search so difficult. In the midst of the molecular genetics revolution that we are all experiencing, the prospect of finding genetic correlates of longevity and of decreased or increased susceptibility to diseases associated with aging appear to be promising. The gene coding for apolipoprotein E is one promising candidate, though the associations of certain polymorphisms with diseases such as Alzhiemer's disease and vascular disease are still unclear. As with potential biomarkers, investigators will have to be most cautious inferring an association between specific genotypes and phenotypes such as exceptional longevity.
|
Acknowledgments |
|---|
Received July 11, 2000
Accepted July 12, 2000
|
References |
|---|
|
TopNature Versus NurtureReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  H. T. Blumenthal Milestone or Genomania? The Relevance of the Human Genome Project to Biological Aging and the Age-Related Diseases J. Gerontol. A Biol. Sci. Med. Sci., September 1, 2001; 56(9): M529 - 537. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
