Centenarians have outlived any exceptionally long-lived invertebrate model by ~100 years and any comparable rodent model by ~30-fold. Human subjects in general and centenarians in particular outlive their nearest relatives, the common chimpanzee, by many decades [50
]. This happy state of affairs is clearly the result of evolutionary changes in our constitutional genomes [51
]. Since the structure of our proteins and those of chimpanzees is very nearly identical [52
], our enhanced life spans are probably related primarily to regulatory RNA species, an area of scholarship that has just begun [6
] and has not yet been applied to the study of the evolution of varying longevities.
Our task here, however, is to review progress towards the elucidation of genetic factors that contribute to exceptional longevities of individual members of Homo sapiens.
The life expectancies of centenarians at birth are nearly double that of most members of their birth cohort and, on average, have surpassed current life expectancy by 22 years. Environmental and stochastic contributions to human life span likely play important roles in the determination of such exceptionally long survivals, as inferred from the twin studies discussed below. Familial aggregates of exceptional longevity do not rule out major environmental factors that are the result of cultural inheritance (e.g., lifestyles, nutrition); such factors could explain, in part, why the progeny of long-lived members of the Framingham study exhibit advantageous cardiovascular risk profiles in middle age [53
]. Nevertheless, evidence consistent with a significant heritable component of exceptional longevity is impressive. Parents of centenarians (born in ~1870) were shown to have approximately nine times the odds of living to the tenth decade as compared to controls [54
]. Siblings of centenarians were shown to have up to an ~18-fold increase in the chance of achieving a similar age [55
]. Such data have raised the possibility that some specific genetic modulators of aging in humans can be identified using such populations, and that conserved pathways for exceptional longevity might thus be validated. Exceptional longevity is obviously coupled with exceptional resistance to diseases that lead to earlier mortalities. We do not have the required biomarkers, however, to clearly disentangle the two phenomena. The research suggested in the last section of this essay may eventually lead to such markers, however.
During the last decade, centenarian populations (New England American, Mormon, Ashkenazi Jewish, Islandic, Okinawan, Japanese, Italian, Irish, and Dutch, among others [54
] have been used for association studies to search for candidate longevity genes or pathways. Particularly striking examples have included PON1
], cytokines, enzymatic antioxidants such as superoxide dismutases [74
], and elements of lipid metabolism [76
]. Some significant differences have been noted between younger cohorts and centenarians in the prevalence of specific genotypes and sometimes in their associated protein activities. These interesting observations, however, have suffered and will continue to suffer from several limitations. In addition to the usual problems and pitfalls of association studies, particularly as we enter the new age of whole genome scans [78
], there is the special problem of the identification of appropriate controls for a cohort of exceptionally long-lived individuals. One innovative approach has been an experimental design based upon a genetic analysis of the progeny of centenarians, giving the opportunity for matched spousal controls [79
The New England Centenarian Study recruited long-lived sib-ships for a genome-wide scan of linkage to exceptional longevity. A region on Chromosome 4 was implicated [80
]. By high density SNP analysis an exonic genotype in microsomal transfer protein was thought to be the locus associated with the exceptional longevity [81
]. The original finding could not be replicated in independent populations [82
]. Such validation is crucial because of the considerable rates of false positives. While it is possible that the role for this gene in longevity may only be significant in certain populations, the most likely explanation for the original linkage was population stratification. The ethnic mix within the long-lived and younger control populations was likely to have differed [83
]. Nevertheless, it would be helpful to evaluate other allelic variants in the same gene or in other related genes. In any case, this early study emphasizes the need to establish additional phenotypes associated with the polymorphism. Although microsomal transfer protein cannot be directly measured, evidence for a role in lipoprotein characteristics or a relationship to age-related diseases would have been helpful in support of a protective role. The population stratification problem can be ameliorated by the selection of better-defined populations, as was done for the case of the Dutch study cited above [83
Studies performed at the Albert Einstein College of Medicine were based upon populations of Ashkenazi Jews and the following considerations [84
]. First, exceptional longevity is obviously a rare phenotype (~1/10,000 individuals live to the age of 94–110). Second, it is also apparent that, for any given cohort, genotypes associated with comparatively early mortality are “weeded out,” while a subset of genotypes are associated with survival. Given large cohorts representing each decade of the life span, one can examine whether those who continue to survive exhibit biologically distinctive phenotypes and genotypes as compared to those of younger cohorts. Thus, the relative prevalence of favorable “longevity” genotypes within the population can be expected to rise monotonically rather than abruptly or intermittently over the life course. Because the genotypes of survivors are “selected,” the greater the attribution of a genotype to longevity, the greater is the divergence from Hardy–Weinberg equilibrium among the elderly. Using this strategy, the Einstein group recruited significant numbers of Ashkenazi Jews of all ages, including ~400 individuals between ages 95–110. Significant increases within aging cohorts were observed for three genotypes from among hundreds of candidate genotypes (selected because of their relevance for lipoprotein phenotypes) that were tested in unrelated populations consisting of individuals between ages 50–110 years () [79
]. These genotypes were: (1) the CETP
gene codon 405 isoleucine to valine variant (CETP VV);
(2) the apolipoprotein C-3 (APOC-3
) gene codon A (−641) C variant (APOC-3 CC)
; and (3) a deletion at +2019 in the adiponectin (ADIPOQ)
gene. The enrichment of the CETP
genotype is supported by evidence from two independent populations [77
Visual Presentation of the Frequency Trends of Favorable Genotypes with Exceptional Longevity
While a significant overrepresentation of a single genotype among nonagenarians and centenarians operationally defines a candidate gene, several other criteria should be fulfilled before considering it to be an important longevity assurance gene ().
The Stages Needed in Order to Support the Association of a Genotype with Longevity
The first step by which the functionality of the genotype can be studied is by determining the serum and plasma levels of the coded protein, if it is secreted and circulating. For example, for each of the genotypes in (CETP
, and ADIPOQ),
appropriate alterations in plasma levels have been demonstrated [76
]. Detailed information on the approach to the choice of controls typically used for these studies is given in [87
]. They included spouses or other nonrelated age-matched pairs for the progeny of centenarians. The genomes of the latter can be expected to be enriched with alleles for unusual longevity. Indeed, these offspring were shown to be healthier than age-matched controls and had more favorable lipid profiles.
A second step in establishing functionality is the identification of an intermediate phenotype. For example, alleles at CETP
differentially modulate lipoprotein characteristics. These effects may vary with age and should therefore be examined in cohorts of varying ages. The functional value of a genotype can also be assessed directly by functional studies of the mutant in a cellular system in vitro.
Such studies may underestimate or overestimate the real physiological importance of the relevant gene action in vivo, however. If the gene encodes a protease, then studies of its specificity, tissue distribution, and regulation are called for. If the gene encodes a cell-surface receptor, then studies of the biochemistry of the receptor should be done. For example, the Einstein group recently identified novel mutations in the IGF-1 receptor of three centenarians. AKT phosphorylation was assessed in lymphoblastoid cell lines from these subjects and controls, both basal and induced levels (stimulating with IGF-1). A marked decrease in phosphorylation was observed in cells from the centenarians with the mutant IGF-1 receptors [89
As noted in the Introduction, the enhancement of life spans in model organisms via single gene mutations raises the question of whether allelic variations at this pathway in human subjects might impact intrinsic biological aging within all tissues and thus lead to substantial increases in life span. Among the many billions of human beings who have lived since the time of recorded history, it is unlikely that a spontaneous mutant of this type, say leading to a doubling of the usual human life span, would have been missed. One could not rule out gene actions of this type, however, that contribute to the generation of rare, relatively healthy centenarians. This possibility is supported by at least three lines of evidence. First, as in the case of CETP VV,
the protection provided by certain genotypes can extend beyond a known role in a disease entity associated with its ascertainment (in this example, cardiovascular disease). The CETP VV
genotype is also associated with enhanced insulin sensitivity and lower risk for hypertension, the metabolic syndrome, and diabetes [87
]. Moreover, the CETP VV
genotype protects against age-related cognitive decline and Alzheimer's disease [91
], although the role of particular haplotypes at that region may interact with polymorphic alleles at the APOE
A second example is a 2.5-fold increase, among centenarians, in the prevalence of an apoC III promoter variant. This variant is associated with significant declines in plasma levels of apoC III and a phenotype of large lipoprotein particles. There is also significantly less hypertension among subjects homozygous for this variant. The most striking data, however, were obtained from a retrospective study of a cohort of subjects bearing this variant. They live significantly longer; in fact, subjects <95 years old with this genotype lived on average over four years longer than those who were not homozygous for the variant. This is indeed a very large impact upon life span when one considers the conclusions of demographers, who have noted that the elimination of ischemic heart disease, a disorder that was responsible for 25.73% of all deaths in 1985, would increase life expectancy at birth by only 3.0 years for females and 3.55 years for males [93
A third argument suggesting that some longevity genes are not merely disease specific is the marked conservation of some of these loci. For example, apoC III is under the control of FOXO-1,
a transcription factor homologous to the DAF16
gene of Caenorhabditis elegans. DAF16
is a key regulator of a downstream suite of genes that are thought to protect the organism from macromolecular damage and thus enhance life span [94
]. Some centenarians have novel functional mutations in the IGF-1 receptor, as noted above. There is evidence that the homologue in mice regulates life span and resistance to oxidative stress [95
]. Thus, while the impact of variants at the CETP
locus upon age-related diseases and longevity may be a special feature of the biology of humans, there is also evidence that the fruits of research on the genetic modulation of the life spans of worms, flies, and mice may in fact be applicable to our species.
Some favorable “longevity genotypes” may act to buffer the deleterious effects of genes that lead to age-related diseases. As a result, the frequencies of deleterious genotypes may, paradoxically, be increased among individuals with extreme life spans. This may explain why the cholesteryl ester transfer protein (CETP-VV) genotype appears to exhibit an additional advantageous effect—the neutralization of the deleterious effects of the lipoprotein(a) (Lp(a)) gene [84
]. Such buffering effects cannot be ascribed to genetic linkage. For the example just cited, those loci are in fact on separate chromosomes. More generally, however, it is clear that one can define two distinct populations, each bearing the disease susceptibility allele in question, but only one of which exhibits the putative buffering effect.