This question—what is aging?—is posed not as an invitation to semantic quibbling, but to initiate reexamination of facts so familiar that they are seldom examined. A case history of an individual who has mild arthritis, some loss of hearing acuity, some evidence of incipient cataract, loss of muscle mass and strength, a progressive decline in capacity for aerobic exercise, troubles with learning and remembering, and an increased vulnerability to infectious illness would lead any physician to assume that the individual described is a man or woman of 60 or more years of age. But the list of signs and symptoms refers with equal accuracy to a 20-year-old horse, or a 10-year-old dog, or a 2-year-old mouse. The specific list of deficits and impairments shifts a bit from person to person, and from species to species, but it is extremely rare to find an 80-year-old person, or 3-year-old mouse, or 14-year-old dog that has avoided all of these age-associated problems. The aging process is synchronized, in that it is common to see all of these difficulties in older people, horses, dogs, and mice, but rare to encounter any of them in young adults just past puberty. This synchrony, though entirely familiar to schoolchildren, physicians, and scientists alike, is the central challenge in biological gerontology: how is it that such a process affects so many cells, tissues, organs and systems at a rate that varies, even among mammals, over a 100-fold range from the shortest lived shrews to the longest lived whales? Structural features of shrews, mice, dogs, and people are remarkably similar at scales from the arrangement of DNA and histones in the nucleus, to the architecture of the kidney, heart, and thymus, to the role of the central nervous system (CNS) and endocrine systems in regulating responses to heat and cold, hunger and thirst, infection, and predators. Why, then, in molecular terms, will the eye, kidney, immune system, brain, and joints of a mouse last only 2 to 3 years under optimal conditions, while the same cells and organs and systems persevere for 50 or more years in people, and longer still in some species of whales?
The definition proposed at the beginning of this chapter—aging as a process for turning young adults into distinctly less healthy old ones—is straightforward enough to appear simple-minded, but in practice draws some prominent distinctions. In this view, aging is not a disease: Diseases are certainly among the most salient consequences of aging, but aging produces many changes not classified as diseases, and many diseases also affect young people. Similarly, lifespan and mortality risks are influenced by many factors besides aging. Thus, evidence that a gene or diet or public health measure has altered life expectancy, upwards or downwards, does not imply that the effects have been achieved by an effect on aging. In the context of the whole organism definition of aging used in this chapter, it is hard to interpret the meaning of changes that occur as individual cells “age” in tissue culture. While cell culture studies can provide valuable information of great relevance to ideas about aging, the two processes are likely to be fundamentally different.
From a biological standpoint, a critical distinction is the difference between aging and development. Development creates a healthy young adult from a fertilized egg, and is strongly molded by natural selection. Genetic mutations that impair development, creating a slower falcon, a near-sighted chipmunk, or a chimpanzee uninterested in social cues, are rapidly weeded from the gene pool. But the force of natural selection diminishes dramatically at ages that are seldom reached. Mice, for example, typically live only 6 months or so in the wild, before they succumb to predation, starvation, or other natural hazard. There is strong selective pressure against mutations that cause cataracts in the first few months of mouse life, but little or no pressure against genotypes that postpone cataract formation for 2 years. Mice protected, in a laboratory setting, against predation, starvation, and other risks typically do develop cataracts in their second or third year. In wolves, however, a genotype that delayed cataracts for only 2 years would be a disaster; natural selection favors wolf genes that preserve lens transparency for a decade or more. A similar process, working on our ancestors in environments where survival to 15 was common, but survival to 55 distinctly uncommon, has filled our own genome with alleles that postpone cancer, osteoarthritis, coronary disease, Alzheimer's, presbycusis, cataracts, sarcopenia, immune senescence, and many other familiar maladies, for about 50 to 60 years. Thus, although aging and development seem, superficially, to be similar processes in that both lead to changes in form and function, they are different in a fundamental and critical way: Development is molded directly by the forces of Darwinian selection and the changes of aging are the consequence of the failure of these selective processes to preserve function at ages seldom reached by individuals in any given species.
Aging as a Coordinated, Malleable Process
The definition of aging as a process that turns young adults into old ones conflicts with a view of aging as instead a collection of processes, some that lead to arterial disease, some that affect endocrine function, some that impair cognition or cause neoplastic transformation, etc. Because each of these ailments is itself the outcome of a complex interaction among many factors, including genes, diet, accidents, viruses, toxins, antibiotics, and physicians, and because each of these diseases, and many others, seems an inextricable part of aging, it has seemed implausible to regard aging as less complex than its (apparent) constituents. Considering aging as a process, rather than a collection of complex processes, has thus seemed to be an oversimplification.
However, two lines of evidence support the merits of the view of aging as a unitary process, with its own (still ill-defined) physiological and molecular basis, which underlies and tends to synchronize the multiple changes seen in older individuals. The first of these discoveries was caloric restriction: The observation that rodents allowed to eat only 60% of the amount of food they would voluntarily consume would live 40% longer than controls permitted free access to food. This observation, first made by McCay in the 1930s, has now been repeated in more than a dozen species in scores of laboratories, and ongoing studies in rhesus monkeys have now produced preliminary, but highly suggestive, evidence that similar benefits may accrue in our own order of mammals. The key point is not merely that lifespan is extended, but that nearly all of the consequences of aging are coordinately delayed. Caloric restriction delays changes in cells that proliferate continuously (such as gut epithelial cells), cells that can be triggered to proliferate when called upon (such as lymphocytes), and those that never proliferate (such as most neurons), as well as on tissues that are extracellular or acellular (such as lens tissue and extracellular collagen fibrils). It delays aspects of aging characterized by excess proliferation, such as neoplasias, and those characterized by failure to proliferate (such as immune senescence). It delays or decelerates age change at the tissue level (such as degradation of articular cartilage) and those involving complex interplay among multiple cells and tissues (such as loss of cognitive function and endocrine control circuits). Because caloric restriction alters, in parallel, so vast an array of age-associated changes, it seems inescapable that these many changes, distinct as they are, must be in some measure timed, i.e., synchronized by a mechanism altered by caloric intake.
A second, more recent, set of experiments leads to the same inference. In 1996, Bartke and his colleagues showed that Ames dwarf mice, in which a developmental defect in the pituitary impairs production of growth hormone (GH), thyrotropin, and prolactin, had an increase of more than 40% in both mean and maximal lifespan compared to littermates with the normal allele at the same locus. Since then studies of this mutant, and the closely similar Snell dwarf, have documented delay in kidney pathology, arthritis, cancer, immune senescence, collagen cross-linking, cataracts, and cognitive decline, making a strong case that these genetic changes in endocrine levels do indeed modulate the aging process as a whole, with consequent delay in a very wide range of age-synchronized pathology. Since 1996, mouse researchers have documented increased lifespan in at least nine other mutations, of which five others, like the Ames and Snell mutations, lead to lower levels of or responses to GH and/or its mediator insulin-like growth factor I (IGF-I).
These observations, on calorically restricted (CR) rodents and now also in mutant mice, justify a sea change in thinking about aging and its relationship to disease. The new framework includes two key tenets: (1) the aging process, despite the complexity of its many effects, can usefully be considered as a single, coordinating mechanism and (2) the rate at which aging progresses can be decelerated in mammals as well as in other taxa. From this perspective, a fundamental challenge to biogerontologists is to develop and test models of how age-dependent changes are themselves coregulated, during middle age, to produce old people or old mice or old worms. Studies of the aged, as opposed to studies of aging, are relevant to this challenge only insofar as they are exploited to generate or test ideas about coregulation and synchrony, rather than ideas about disease-specific pathogenesis. The key resource for meeting the challenge is not comparisons between young and old donors, but rather comparisons between young or middle-aged adults who are known to be aging at different rates. Fortunately, the same experiments that have documented the malleability of aging rate have done so by producing sets of animals that do indeed age at normal or slower-than-normal paces. With luck, future studies may produce at least tentative answers to the two key questions in biogerontology: How does aging produce the signs and symptoms of aging and what controls the rate of this process in mammals?
Key Themes from Studies of Invertebrate Models
Four main ideas have emerged, in the period 1990–2008, from studies of genetically convenient invertebrate models, the nematode worm Caenorhabditis elegans and the fruit fly Drosophila melanogaster. (1) Single gene mutations—lots of them—can extend lifespan in worms and flies (and mice). Most mutations discovered so far are “loss of function,” in which eliminating or crippling the gene product leads to slower aging and lifespan extension. (2) The genes whose elimination slows aging are typically those used by the normal animal to notice and respond to environmental poverty; they are selected, by evolution, to permit the animal to take on alternate forms or functions to deal with the relative absence of nutritional conditions optimal for rapid growth and reproduction. Many of these genes are familiar to physicians and biological scientists: they encode insulin, insulin-like growth factors, possibly thyroid hormones, and other regulators of fuel usage and metabolic fluxes, as well as the intracellular proteins that change cell properties in response to these hormones. (3) The mutations that extend lifespan in worms and flies (and perhaps in mice) render the animals more resistant to lethal injury, such as that resulting from heavy metals, ultraviolet irradiation, heat, oxidizing agents, and chemicals that damage DNA. It thus seems plausible that augmentations of stress resistance are the mechanism by which these mutant genes slow aging. (4) This three-way association, connecting aging to stress to signals about nutrition, has very deep evolutionary roots and can be noted (albeit with species-specific nuances) in yeast, flies, worms, and mice. The suggestion here is that the pathways tying nutrition to stress to aging rate evolved extremely early in the eukaryotic lineage, in an ancestor predating the branch points between yeast, flies, worms, and vertebrates. The key implication is that investigations into the cell biology of these linked processes in conveniently short-lived organisms may provide valuable insights into aging and disease in humans.
Molecular Leads for Further Study
Studies of the invertebrate model systems have also called attention to a range of intracellular pathways that influence longevity, perhaps through regulation of resistance to a variety of forms of lethal and metabolic stress. Components of these pathways are now under scrutiny to see if they also affect aging and disease in rodents. Many of the antiaging mutations in worms, for example, act by increasing the actions of a DNA-binding factor called Daf-16, whose targets include genes that modulate resistance to oxidative damage, DNA repair, and other modes of cellular defense. The mammalian equivalent to Daf-16, a member of the FoxO group of transcription factors, has already been implicated in control of cell death in human and rodent cells, and studies of the role of FoxO family members in aging are under way. A second family of proteins, the sirtuins, was initially implicated in lifespan regulation in studies of the budding yeast Saccharomyces cerevisiae, but then shown also to be able to increase lifespan in nematode worms. There is some evidence that members of the sirtuin family may contribute to the improved health and longevity of mice on a calorie-restricted diet, and interest in these proteins has been spurred by observations that median lifespan of mice on a diet high in saturated fat can be increased significantly by resveratrol, whose many biochemical effects include stimulation of sirtuin activity. Work with invertebrates has also brought new attention to the possible role of TOR, the “target of rapamycin,” in control of lifespan and responses to dietary interventions. TOR plays a major role, conserved throughout the evolutionary tree, in regulating protein translation rates in response to both external cellular stress and the available supply of amino acids. TOR inhibition by rapamycin is a mainstay of clinical immunosuppressive therapy after organ transplantation, and TOR inhibitors also show promise as antineoplastic agents. Other molecular circuits, including those triggered by the p66 (shc) protein, the tumor-suppressor p53, and the stress-responsive kinase JNK, were initially discovered through work with rodents and now are also being examined in the context of invertebrate models for aging. Each of these biochemical pathways has powerful and still only partially defined links to hormonal signals, neoplastic transformation, stem cell homeostasis, and the balance between cell growth and cell death, so their elucidation is likely to provide both rationale and direction for translational work aimed at slowing the aging process and retarding age-related disease and dysfunction.
Delayed Aging in Mice and Rats
Mice or rats fed approximately 30% or 40% less food than they would ordinarily consume typically live up to 40% longer than freely fed animals. These CR animals stay healthy and active, with good physical, sensory, and cognitive function, at ages at which most of the controls have already died. The intervention extends lifespan if initiated at very young ages (e.g., at weaning), or when started early in adulthood (e.g., at 6 months, in a species where puberty occurs at 2 months and median survival is about 28 months). Whether CR diets extend lifespan when initiated in animals already older than half the median lifespan is controversial, with some early studies suggesting little or no response, and some more recent experiments leading to a positive result. Lifespan can be extended using CR diets of widely varying composition, and there is strong evidence that total caloric intake, rather than proportions of specific fuel sources (carbohydrates, fats, proteins), is responsible for the beneficial effects. CR diets do not, except in the first few months after their imposition, alter metabolic rate or oxygen consumption per gram of lean body mass because CR rodents lose weight, or if adolescent fail to gain weight, so that lean body mass matches calorie supply once equilibrium is established. Although CR rodents are less obese than control animals, it seems unlikely that the effect of CR diets on aging is caused entirely by avoidance of obesity, in part because CR diets extend lifespan in mice genetically engineered to lack leptin (ob/ob mice). When placed on a CR diet, ob/ob mice are both longer lived and more obese than normal mice with normal caloric intake. There is some evidence that CR diets can delay at least some aspects of aging, for example, changes in immune function, in nonobese rhesus monkeys, but definitive evidence, based on lifespan data, is not yet available.
Mice and rats on CR diets show delay or deceleration in a very wide range of age-dependent processes, including neoplastic and nonneoplastic diseases, changes in structure and function of nearly all tissues and organs evaluated, endocrine and neural control circuits, and ability to adapt to metabolic, infectious, and cardiovascular challenges. Mouse stocks that have been engineered for vulnerability to specific lethal diseases, such as models of lupus or early neoplasia, also tend to live longer when placed on a CR diet. Intensive study of CR rodents in the past 20 years has suggested many ideas about the mechanism of its effects, including ideas about altered levels and responses to glucocorticoids and/or insulin, increases in stress-resistance pathways including resistance to oxidative injury, diminished inflammatory responses, changes in stem cell self-renewal, and many others, each plausible, none of them at this point more than plausible. Further work on the early and midterm effects of CR diets is among the most attractive avenues for testing basic ideas about aging in mammals.
Rats fed a diet containing much reduced levels of the essential amino acid methionine were shown, in 1993, to live 40% longer than rats, on a standard diet, and more recent work has shown a similar, though smaller, effect on mice. These animals are not calorically restricted—they eat more calories, per gram of lean body mass, than control rats, and rats pair-fed normal food at levels that match the total caloric intake of a methionine-restricted (MR) rat show a much smaller degree of lifespan extension. MR diets could, in principle, affect aging by causing changes in protein translation or rate of protein turnover, by changes in DNA methylation (which depends upon metabolites of methionine), by alterations in levels or distribution of the antioxidant glutathione (also a metabolite of methionine), by changes in hormone levels (MR mice show low levels of insulin, glucose, and insulin-like growth factor 1, IGF-I), or by induction, through hormesis, of augmented stress response pathways at the cellular level. Some of these ideas could be tested using diets that restrict levels of other amino acids, and there is older, fragmentary evidence that such diets, too, may induce lifespan extension in rodents. MR is thus the second confirmed method for extending lifespan in mammals and comparisons of similarities and differences between CR and MR rodents are likely to prove highly informative.
Single-Gene Mutations that Extend Mouse Lifespan
The initial reports in the 1980s and 1990s that mutations of single genes in worms, and then in flies, could produce dramatic increases in lifespan were the strongest support, along with the CR data, for viewing aging as a unitary process that could be decelerated. The report in 1996 by Bartke and his colleagues that the lifespan of mice could be extended more than 40% by mutation of a gene required for pituitary development opened the door to new genetic models for study of aging in mammals. This Ames dwarf gene (Prop1) leads to an endocrine syndrome featuring low levels of growth hormone (GH), IGF-I, thyrotropins and the thyroxines, and prolactin. The observation that longevity in mice could be improved by reduction of hormones in the insulin/IGF family, i.e., the same family of signals implicated in lifespan extension in worms and (later) flies, provided a key foundation for exploiting comparative cell biology in biogerontology research. Subsequent work then showed that similar degrees of lifespan extension could be seen in Snell dwarf mice, whose endocrine defects are very similar to those of Ames dwarfs, and also in mice that lack the GH receptor (GHR-KO mice) (Figure 1-1). This last observation, together with documented lifespan extension in mice lacking GH-releasing hormone receptor (GHRHR) and in mice with diminished expression of the IGF-I receptor (IGF1R heterozygotes), suggested strongly that a common factor, i.e., low production of or response to IGF-I, is a major cause of lifespan extension in all five models, although it is certainly possible that other factors, such as altered insulin sensitivity, thyroid tonicity, adipokine levels, etc., may contribute to antiaging effects in some of these mouse models. Male and female mice are affected by each of these mutations except that the effect of the IGF1R mutation seems stronger in females than in males. Mice in which tissue levels of IGF-I are reduced by genetic manipulation of a protease that controls local concentrations of IGF-I binding proteins are also long-lived, again consistent with models in which abnormally low IGF-I levels cause lifespan extension in mice. Studies of Snell and Ames dwarf mice have shown that the exceptional longevity of these mice is accompanied by a delay or deceleration of age-dependent changes in T lymphocytes, skin collagen, renal pathology, lens opacity, cognitive function, and neoplastic progression; taken with the lifespan data, these observations suggest strongly that these mutations, like the CR diet, act to slow the aging process itself.
(A) A young adult Snell dwarf mouse, with a littermate control (on the vehicle). (B) Survival curves for Snell dwarf (dw/dw) mice and littermate controls. (C) Glomerular basement pathology scores, at terminal necropsy, for Snell dwarf (N = 40) and control (N = 46) mice. Higher scores indicate a greater degree of kidney pathology. Despite living 40% longer, a higher percent of Snell dwarfs had 0 scores at necropsy and a smaller percent had scores of 2 or 3. (D) Cataract scores determined by slit lamp examination in Snell dwarf and littermate control mice at 18 and 24 months of age. p < 0.001 for the difference between dwarfs and controls at each age. (B) Data from Flurkey K, Papaconstantinou J, Miller RA, Harrison DE. Lifespan extension and delayed immune and collagen aging in mutant mice with defects in growth hormone production. Proc Natl Acad Sci U S A. 2001;98(12):6736–6741. (C and D) Data from Vergara M, Smith-Wheelock M, Harper JM, Sigler R, Miller RA. Hormone-treated Snell dwarf mice regain fertility but remain long-lived and disease resistant. J Gerontol Biol Sci. 2004;59:1244–1250.
In addition to the six mutations that lead to lower levels of IGF-I signals, there are now five other mutations that have been shown, in each case in a single unreplicated report, to extend lifespan in mice. Table 1-1 presents a summary of these 11 published mutations that extend maximal lifespan in mice. Transgenic overexpression of urokinase-type plasminogen activator in the brain has been reported to produce a significant extension of lifespan, perhaps by suppression of appetite with consequent mimicry of a CR diet. Mice in which the insulin receptor has been inactivated specifically in adipose tissue show an 18% lifespan increase, consistent with the idea that altered insulin sensitivity or adipokine levels might play a role in aging rate in this species. Transgenic mice overexpressing the klotho protein, a cofactor for fibroblast growth factor (FGF) signals whose absence leads to early death by elevation of vitamin D levels, have been reported to live 19% to 31% longer than controls, perhaps reflecting the ability of klotho to block insulin or IGF-I signals. A mutation that inactivates the 66 kDa splice variant of the Shc protein, involved in the pathway to programmed cell death after exposure to hydrogen peroxide or ultraviolet light, also extends longevity (by about 28%), as does transgenic overexpression, in mitochondria, of catalase, an enzyme involved in detoxification of hydrogen peroxide (a 20% lifespan increase). These last two reports, if confirmed in other laboratories, may give new insights into the connections linking aging and late-life diseases to agents that damage DNA or induce oxidative injury at intracellular sites.
Table 1-1 Mouse Mutants That Improve Longevity ||Download (.pdf)
Table 1-1 Mouse Mutants That Improve Longevity
EFFECT ON LIFESPAN (%)
Ames dwarf (df), Prop1
Low GH, IGF-I, thyrotropin, thyroxine, prolactin
Snell dwarf (dw), Pit
Low GH, IGF-I, thyrotropin, thyroxine, prolactin
GHRHR (lit, little)
Low GH, IGF-I
Survival increase on low-fat diet only
Heterozygous mice; significant only in females only
Pregnancy-associated plasma protein A, PAPP-A
Protease for IGF-I binding proteins
Impairs insulin and IGF-I signals
Insulin receptor; FIRKO
Receptor diminished in adipose cells only
Lower apoptosis after UV, H2O2
Urokinase-type plasminogen activator, uPA
Transgenic, expressed in brain; may suppress appetite.
Catalase transgenic, MCAT
Overexpression in mitochondria only
More work is needed to determine to what extent these mutations influence common pathways, to see whether they do or do not influence aging through the same mechanism. It is possible, for example, that some of these mutations diminish risk of cancer, a common cause of death in many inbred mouse strains, without effect on any other age-dependent trait, while others may modulate the effects of aging on multiple organ systems, thus diminishing mortality risk from both neoplastic and nonneoplastic diseases. There are hints, for example, that CR diets and the Ames dwarf gene may affect longevity by different mechanisms. As shown in Figure 1-2, caloric restriction extends lifespan in normal as well as in dwarf mice, and the Ames mutation extends lifespan on both CR and control diets, showing that the effects are at least partly additive. The survival curves in Figure 1-2 also suggest that the Ames dwarf mutation may be affecting the age at which deaths begin to occur (“delay” of aging), and that the CR diet may affect, instead, the rate at which deaths occur once mortality risks become detectable (“deceleration” of aging). Dwarf and CR mice also differ dramatically in adiposity (CR mice are extremely lean; dwarf mice have average or above-average proportions of fat), and in resistance to liver toxicity induced by acetaminophen poisoning. Analysis of the mechanisms by which each of the mutations shown in Table 1-1 affects lifespan and age-dependent traits will help to sort out questions about whether initial mortality rate and mortality rate doubling time are under separate physiological control.
Survival curves for genetically normal (wild-type [WT]) or Ames dwarf (df) mice on caloric-restricted diet (CR) or on unrestricted ad libitum (AL) food intake. Each symbol represents a mouse dying at the indicated age. The dwarf mice live longer than the WT mice on either diet, and the CR diet extends the life span of mice of either genotype. Reproduced with permission from Bartke A, Wright JC, Mattison JA, Ingram DK, Miller RA, Roth, GS. Extending the lifespan of long-lived mice. Nature. 2001;414:412.
As Table 1-1 suggests, the evidence from mice indicates that diminution of IGF-I levels, either in early life, adult life, or perhaps both, can create mice that are long-lived compared to controls. The association has also been noted using experimental designs in which individual mice are tested for IGF-I levels as young adults: those mice with the lowest IGF-I levels were found to live significantly longer than those with higher IGF-I concentrations. Similarly, mouse stocks bred by selection for slow early life growth trajectory are found to be smaller than controls and also longer-lived. There is also highly suggestive evidence for a similar relationship between IGF-I, body size, and lifespan in dogs and horses. For dogs, several studies have shown greater longevity in small-size breeds than in large breeds, and a strong relationship between body size and life expectancy among mixed-breed (mongrel) dogs as well (Figure 1-3). Anecdotal and limited published data suggest that pony breeds of horses are also substantially longer lived than horses of full-sized breeds. The relationship between body size and life expectancy in humans is complicated by the strong effects of socioeconomic status on both endpoints: wealthier people tend to be both taller and longer-lived than poor people. On the whole, tall stature is associated with lower mortality risks from cardiovascular diseases, which are a major cause of mortality in developed countries. In contrast, a remarkably consistent set of studies (Table 1-2) show that tall stature is associated with significantly higher mortality risk from a wide range of neoplastic diseases. There is also limited data that centenarians, on average, were shorter as mature adults than those who do not attain centenarian status. Testing the idea that short midlife stature is associated with delayed or decelerated aging in humans will depend on measuring a wide range of age-dependent traits, rather than merely lifespan, on large populations of middle-aged people.
Longevity, size, and insulin-like growth factor I (IGF-I) levels among breeds of purebred dogs. Left panel shows mean breed lifespan as a function of mean breed weight for each of 16 breeds of dogs; three breeds are indicated by arrows. Data from Li Y, Deeb B, Pendergrass W, Wolf N. Cellular proliferative capacity and life span in small and large dogs. J Gerontol A Biol Sci Med Sci. 1996;51(6):B403–B408. Right panel shows mean plasma IGF-I levels as a function of body mass in eight breeds of purebred dogs. For details and citations see Miller and Austad. Growth and aging: why do big dogs die young? In: Masoro EJ, Austad, SN, eds. Handbook of the Biology of Aging. 6th ed., New York, New York: Academic Press; 2006.
Table 1-2 Population-Based Association Between Short-Stature and Lower Mortality Risk for Multiple Neoplastic Diseases ||Download (.pdf)
Table 1-2 Population-Based Association Between Short-Stature and Lower Mortality Risk for Multiple Neoplastic Diseases
TALL PEOPLE DO BETTER
SHORT PEOPLE DO BETTER
15 000 Scots
Colorectal, prostate, hematopoietic cancers
12 000 NHANES (United States); men
Cancers; 40–60% effect; adjusted for race, smoking, income
Breast and colorectal cancer
22 000 U.S. male physicians
Cancer; adjusted for age, BMI, exercise, smoking
570 000 Norwegian women
400 000 American women
Breast cancer, postmenopause
1.1 million Norwegians
England and Wales (by county)
Ischemic heart disease
Breast, prostate, ovarian cancer
Stress Resistance and Aging
Mutations that extend lifespan in invertebrates typically render the animals resistant to multiple forms of lethal injury, whether the threat comes from oxidative agents, heat, heavy metals, or irradiation. Indeed, this stress resistance seems likely to represent the mechanism by which these mutations delay the aging process. Thus presumably much of the cellular and extracellular pathology that produces dysfunction and increases mortality risk in older animals is held in abeyance by the same, poorly defined, defenses that permit nematodes and flies to survive when exposed to external stress in an experimental setting. Genetic dissection of the relevant pathways has shown, surprisingly, that in normal, nonmutant worms, the levels of stress resistance, and thus resistance to aging, are actively diminished by specific DNA-binding transcription factors. These factors, whose human homologs are members of the FoxO family, are retained by evolutionary pressures because they provide reproductive advantages in the natural environment, in which animals must be able to quickly take advantage of transient access to nutrients. Genetic inactivation of these FoxO pathways in the laboratory produces mutant animals that are not ideally suited for natural conditions, but which are resistant to many kinds of stress and which age more slowly than normal. Studies of gene expression patterns in the long-lived mutant worms have shown that the FoxO proteins can trigger transcription of over 100 genes that together protect against many different forms of cellular damage. The list includes enzymes that destroy free radicals, heat shock proteins, and other chaperones that guard against misfolded proteins, proteins that protect against infection, and chelating agents that bind toxic metal ions, among others.
The connection between induction of these stress-resistance pathways and late-life diseases has been shown by two sets of informative experiments. In the first, genetically identical worms were exposed to a brief, nonlethal heat stress, and physically separated into those that showed a strong response of chaperone proteins and those that did not. Worms with the most robust response to transient stress were found to be longer lived than those with lower stress responses. A second approach involved worms bearing genetic variants that cause aggregation of proteins and neurodegeneration (Huntington's disease) in people; neurological dysfunction in these worms can be delayed, and in some cases prevented entirely, by augmentation of the FoxO-dependent stress-resistance pathways. Similarly, age-dependent increases in susceptibility to stress-induced cardiac arrhythmias in Drosophila can be significantly postponed by activation of FoxO-dependent protective pathways.
Studies of the relationship of stress resistance to aging in mammals are underway, but suggestive data have begun to emerge. Both CR diets and at least some of the long-lived endocrine mutant stocks show elevated levels of enzymes with antioxidant action, heavy metal chelators, and intracellular chaperone proteins, as well as have lower levels of oxidative damage to DNA, proteins, and lipids. Cells grown, in tissue culture, from long-lived Snell and Ames dwarf mutant mice, or from mice lacking GH receptor, are resistant to lethal injury caused by cadmium, peroxide, heat, a DNA alkylating agent (MMS), ultraviolet light, and paraquat (which induces mitochondrial damage by free radical generation). Mice prepared by CR or MR diets are resistant to liver damage induced by the oxidative hepatotoxin acetaminophen, and long-lived mutant mice are somewhat more resistant to death induced by paraquat injection. Stress resistance also seems to play a role in evolution of long-lived species in that cells from long-lived rodents and other mammals are resistant in culture to several forms of oxidative and nonoxidative damage. This work provides initial support for models that attribute variations in aging rate to differences in stress resistance pathways, but many questions remain unanswered at this early stage. Figure 1-4 shows representative results for resistance to stress in long-lived mutant worms and cells from long-lived mutant mice, as well as data on resistance of CR and MR mice to an oxidative hepatotoxin. Figure 1-5 presents data from two studies showing stress resistance in culture of cells from longer-lived mammalian species.
Association of stress resistance to longevity. (A) Resistance to heat shock (thermotolerance) is higher in mutant worms that have extended longevity. Each point is the mean for one mutant strain. (B) Resistance of skin-derived fibroblasts to lethal hydrogen peroxide concentrations is higher in cells from Ames dwarf (df/df) mice compared to controls. Each symbol shows an individual mouse. (C) Hydrogen peroxide (H2O2) resistance of skin-derived fibroblasts from long-lived growth hormone receptor knock out (GHRKO) mice. LD50 is the amount of peroxide that kills 50% of the cells. (D) Resistance of calorie-restricted (Cal-R) mice to liver injury induced by injection of acetaminophen (APAP); serum LDH levels indicate the level of damage to hepatocytes at varying intervals after a single injection (unpublished data of Chang and Miller). (A) Reproduced with permission from Gems D, Sutton AJ, Sundermeyer ML, et al. Two pleiotropic classes of daf-2 mutation affect larval arrest, adult behavior, reproduction and longevity in Caenorhabditis elegans. Genetics. 1998;150:129–155. (B and C) Reproduced with permission from Salmon AB, Murakami S, Bartke A, Kopchick J, Yasumura K, Miller RA. Fibroblast cell lines from young adult mice of long-lived mutant strains are resistant to multiple forms of stress. Am J Physiol Endocrinol Metab. 2005;289(1):E23–E29.
Stress resistance in fibroblasts from animals of long-lived species. Left panel: Resistance of cultured skin fibroblasts to hydrogen peroxide as a function of species maximum lifespan in years. Species, from left to right, are hamster, rat, rabbit, sheep, pig, cow, and human. Reproduced with permission from Kapahi P, Boulton ME, Kirkwood TB. Positive correlation between mammalian life span and cellular resistance to stress. Free Radic Biol Med. 1999;26(5–6):495–500. Right panel: Resistance of cultured skin fibroblasts to cadmium as a function of species maximum lifespan in years. Species, left to right, are laboratory mouse, wild-caught mouse, rat, red squirrel, white-footed mouse, deer mouse, fox squirrel, porcupine, beaver, and little brown bat. LD50 is the amount of hydrogen peroxide or cadmium that kills 50% of the cells. Reproduced with permission from Harper JM, Salmon AB, Leiser SF, Galecki AT, Miller RA. Skin-derived fibroblasts from long-lived species are resistant to some, but not all, lethal stresses and to the mitochondrial inhibitor rotenone. Aging Cell. 2007;6:1–13.
Genetic Approaches to Analysis of Aging in Humans
Attempts to find genetic variations that influence aging in humans have been plagued with practical and conceptual problems in addition to the obvious difficulty that selection of mating partners is not amenable to experimental control. For one thing, heritability calculations show that only about 15% to 20% of the variation in lifespan among humans can be attributed to genetic factors. Furthermore, an unknown but potentially large fraction of this genetic variation probably reflects genetic variants that influence susceptibility to diseases of childhood, infectious agents, and specific common illnesses of old age. For example, genetic variants that cause Huntington's disease or type 1 diabetes or which triple the normal risk of myocardial infarction by the age of 50 years would all contribute to the measured heritability of lifespan, but do so by altering mortality risks from a specific form of illness rather than by alteration of aging with its effects on multiple late-life traits. Thus genetic variants that mold lifespan by effects on aging per se, if they exist at all, are likely to influence only a small fraction of variation (perhaps 5%) in how long people live.
Formal analyses of the genetics of human aging have so far relied mostly on candidate gene approaches, in which the investigator evaluates long-lived and control populations for variations at one or a small number of genetic loci, selected on theoretical grounds as most likely to be involved in aging or disease processes. The alternate approach, whole genome screening of large populations for association of longevity to hundreds of thousands of genetic variants, is also beginning to produce initial results at an accelerating pace. Although most of the published studies, and those in progress, focus on nuclear genes, there are tantalizing suggestions from studies of large families that a surprisingly high proportion of inherited effects on lifespan may come through the maternal line alone, suggesting that variations in mitochondrial gene sequences (which are inherited almost entirely from mothers to both sons and daughters) may also influence life expectancy in humans.
A major problem with all these approaches, from the perspective of biological gerontology, is the lack of a defensible phenotype: a measure of aging better than lifespan. There are now several dozen reports of candidate loci at which particular alleles are overrepresented among centenarians or near-centenarians, and advocates of this strategy hope that among this collection are some loci that control aging rate. But skeptics note that alleles that increase risk of cardiac disease, or Alzheimer's disease, or stroke, or various common forms of cancer, or severe osteoporosis are likely to have contributed to disease and death before the age of 90 or 100 years, and thus to have been eliminated or greatly reduced among the very old. Thus it should be assumed that a collection of genetic loci whose frequency discriminates very old people from others of the same birth cohort will include many genes with influence over common forms of lethal illnesses rather than genes that modulate aging per se. This problem is not one that can be solved by technological innovation or larger numbers of tested subjects; it requires development of a phenotype that provides more information about health in old age than merely a record of the age at death. For example, a genetic allele that identified, among 70-year-old people, those most likely to have excellent eyesight and hearing, no history of cancer, angina, diabetes, or arthritis, above-average responses to vaccination, and retention of baseline levels of cognition and muscle strength would be a much stronger candidate for an authentic “antiaging” gene than one that predicted survival to the age of 100 years.
Models of Accelerated Aging
There are a small number of rare, inherited, diseases, of which Werner's syndrome and Hutchinson-Gilford syndrome are the most celebrated, that have been mooted as possible examples of “accelerated” aging. Some of the physical features and symptoms of these diseases do resemble, at least superficially, some of the changes that typically affect older people, including in particular changes in skin and connective tissues. Hutchinson-Gilford syndrome, sometimes called “progeria,” is now known to be caused by mutations in the gene for Lamin A, a component of the nuclear membrane. Werner's syndrome patients usually have mutations in an enzyme (“WRN”) that has activity as a DNA helicase (unwinding coiled DNA) and as an endonuclease. Patients with Hutchinson-Gilford syndrome typically survive to their early teens, and Werner's syndrome patients frequently survive to their mid-forties, about 10 years after the age of typical diagnosis.
However, it is highly debatable whether either of these diseases provides strong clues about the molecular or cellular basis for age-related changes in normal individuals. Werner's patients do resemble elderly people in some ways: they frequently suffer from cataracts and premature graying of the hair, and by their early thirties often develop osteoporosis, diabetes, and atherosclerosis. On the other hand, many features of normal aging are not seen in Werner's patients and many features of Werner's syndrome are not seen in normal old individuals. Werner's syndrome patients, for example, do not show signs of Alzheimer's disease or other amyloidoses, hypertension, or immune failure. Mesenchymal tumors, which are rare in normal people, are about 100-fold more frequent in Werner's patients, but the epithelial and hematopoietic tumors characteristic of normal aging are not seen in Werner's syndrome patients. Furthermore, Werner's patients exhibit many features that are not seen in normal aging, including subcutaneous calcifications, altered fat distribution, vocal changes, flat feet, malleolar ulcerations, high levels of urinary hyaluronic acid, and a number of other idiosyncrasies not seen commonly in elderly individuals. Mice with mutations in the WRN gene live a normal lifespan and do not show signs of premature senescence. It seems possible that investigation of the pathogenesis of Werner's syndrome may provide key clues to the mechanisms of age-related diseases. But it seems at least equally plausible that the WRN mutation, perhaps through alteration of cells responsible for connective tissue maintenance, induces multiorgan failure through processes quite distinct from the changes that impair some of the same organs in normal aging.