Thanks largely to the power of genetic analysis in model organisms such as Caenorhabditis elegans (a nematode), Drosophila melanogaster (a fruit fly), and the laboratory mouse, major advances have been made in the elucidation of what can be termed "public" modulations of intrinsic biological aging—that is to say, commonalities of gene actions across widely diverse phyla that explain, in part, the plasticity of processes of aging. There are hints that at least one such conserved pathway may be operative in our own species. These observations, together with related research on other biochemical pathways, a long history of research on the beneficial effects of dietary restriction (most recently including an initial report of its beneficial effects on healthspan and lifespan in a primate) (Fig. 71-1), and spectacular advances in genomics raise the possibility that we may one day be able to delay the times of onset and decrease the rates of progression of aging processes. Such interventions have the potential to extend the healthspans and, therefore, the functional lifespans of a large proportion of our population. This new knowledge, however, is still very distant from clinical translation. Many remain skeptical of the relevance of these experimental findings. Moreover, we need much more information on the pathophysiology of aging, especially in the invertebrate models that have provided most of our new knowledge concerning genetic modulations of lifespan. We will also require more detailed information on the impact of longevity enhancements upon what can be described as the "terminal decline" of the life course, the stage of life in humans responsible for protracted morbidity, frailty, and the consequent loss of the ability to live independently. These terminal declines account for a very substantial proportion of all health care costs. Finally, the promising new knowledge needs fuller discussions by ethicists, economists, sociologists, and political scientists, among others, as to the impacts upon society of any large-scale clinical translations.
Definitions of Aging: Senescent Phenotypes
Mammalian gerontologists usually define aging in terms of the gradual, insidious, and progressive declines in structure and function (involving molecules, cells, tissues, organs, and organisms) that begin to unfold after the achievement of sexual maturity. These declines affect the germ line as well as the soma. For large populations of individuals, exponential declines in the probability of survival are observed. The aging organism is less successful in its reaction to injury and has increasing difficulties in maintaining physiological homeostasis. The organism, therefore, becomes increasingly vulnerable to a wide range of environmental perturbations.
Biological aging is the major risk factor for essentially all of the major geriatric disorders, including dementias of the Alzheimer type (DAT), Parkinson's disease, age-related macular degeneration, ocular cataracts, presbycusis, all forms of arteriosclerosis, type 2 diabetes mellitus, congestive heart failure, sarcopenia, osteoporosis, osteoarthritis, degenerative intervertebral disk disease, immunosenescence, benign prostatic hyperplasia, and most forms of cancer. These and many other disorders (Table 71-1) can be referred to as "senescent phenotypes."
Table 71-1 Alterations in Proliferative Homeostasis in the Tissues of Aging Humans |Favorite Table|Download (.pdf)
Table 71-1 Alterations in Proliferative Homeostasis in the Tissues of Aging Humans
|Integument||epidermal atrophy, "liver spots", seborrheic keratoses, basal cell ca, squamous cell carcinoma, graying and loss of hair, eccrine sweat gland atrophy, apocrine sweat gland hyperplasia, stasis dermatitis, regional subcutaneous atrophy, hyperplasia|
|Sensory||lacrimal gland atrophy, corneal degenerations, ocular cataracts, age-related macular degeneration, presbycusis, olfactory loss|
|Musculoskeletal||sarcopenia, "fatty infiltration" of muscle, osteoarthritis, osteoporosis|
|Hematopoietic/Immune||anemias, myelodysplastic syndromes, leukemia, lymphoma, monoclonal gammopathy and multiple myeloma, autoimmune disorders (e.g., atrophic gastritis and polycythemia vera), immunosenescence (accelerated in AIDS)|
|CNS||reactive gliosis, dural and meningeal fibrosis|
|Cardiovascular||atherosclerosis, arteriolosclerosis, myocardial interstitial fibrosis|
|Pulmonary||interstitial fibrosis, emphysema|
|Renal||glomerulosclerosis, interstitial fibrosis|
|Male Reproductive||benign prostatic hyperplasia (smooth muscle and glands), adenocarcinoma of prostate, testicular atrophy|
|Female Reproductive||ovarian atrophy and theca cell hyperplasia, endometrial atrophy and hyperplasia, endometrial carcinoma, smooth muscle atrophy, leiomyomas of uterus|
|Endocrine||parenchymal atrophy with interstitial fibrosis, cell type-specific hyperplasias, adenomas|
|GI||mucosal and smooth muscle atrophy, hyperplastic polyps, adenomas, adenocarcinomas of colon and rectum|
The Classical Evolutionary Biological Theory of Why We Age
A compelling theory as to why aging occurs has been developed by a series of contributions by evolutionary biologists, beginning with JBS Haldane. Haldane wondered why certain late-onset disorders, such as Huntington's disease, seemed to be so prevalent in England—perhaps of the order of one per thousand instead of what might have been expected from germ-line mutation rates—perhaps one per million. He concluded that this was because the disease had largely escaped the force of natural selection, because the commonest forms of the disease did not manifest until after reproduction had ceased. In age-structured populations (i.e., populations consisting of individuals with a wide range of ages), most of the reproduction is carried out by the younger cohorts. This is because, historically, few individuals living in the wild escaped the effects of infections, predation, nutritional deprivations, and accidents to achieve old age. As such, even late-acting good alleles will have only minor contributions to the gene pools of subsequent generations. Peter Medawar extended this idea, arguing that there are numerous such constitutional mutations, an idea that has come to be known as the "mutation accumulation" theory of aging. (These are mutations that one is born with, not somatic mutations.) By this argument, most of us are likely to have been born with some special vulnerability to a late-life disorder or disorders. A second major theory, known as "antagonistic pleiotropy," was developed by George C. Williams. He argued that there are likely to be many genetic alleles that were selected because of their enhancement of reproductive fitness early in the life course, but that have negative effects late in life, when the force of natural selection will have greatly diminished. A more general conceptualization of tradeoffs between reproduction and lifespan (the "disposable soma" theory) was developed by TBL Kirkwood. The evolutionary theory was quantified by William D. Hamilton and elaborated on by Brian Charlesworth and Michael Rose.
Perhaps the best indication that the field of biogerontology has finally matured as a science is the fact that its most cherished theory, what can now be termed as the "classical" evolutionary theory, has undergone several challenges. First, demographers have noted that at the extremes of old age for organisms as diverse as roundworms, fruit flies, med flies, and humans, rates of declines in the force of natural selection diminish. One response to this important challenge (the "cocoon" hypothesis) is that these declines may simply be related to the virtual cessation of locomotor activities at extreme ages. When flies cease flying and worms cease moving, fewer opportunities for serious injuries may result. These "plateaus" in the rates of mortality at very advanced ages are much less striking for people; they might also be related, in part, to diminished motility, as well as to secular trends in the development of central heating, air conditioning, and immunizations. A second challenge comes from geneticists who have discovered that, to our great surprise, many single-gene mutations can substantially increase the lifespans of Baker's yeast, nematodes, and fruit flies This issue has been most systematically explored in C. elegans; a meta-analysis of an initial set of unbiased, genome-wide RNAi screens ("knockdowns," but not "knockouts" of gene expression) has tabulated hundreds of single-gene loci capable of enhancing lifespan when they are dialed down. These genes fall into a finite number of pathways and, moreover, many of them fit within the context of antagonistic pleiotropic mechanisms of aging. For example, hypomorphic mutations in genes within the most famous of these pathways, the insulin-like growth factor (IGF)/insulin signaling pathway, may be reporting on an evolutionarily conserved diapause—nature's way of taking time out from the business of development and reproduction during hard times, such as severe shortages of food. The gene actions associated with such diapauses (many of which are still unexplored at the biochemical genetic level) understandably result in enhanced resistance to various stresses. A large number of these C. elegans "longevity genes" converge upon a single transcription factor, daf2, or influence mitochondrial function. Moreover, it is possible that many such genes may be specific for this highly inbred laboratory model. In general, these investigations support, rather than refute, the evolutionary theory of aging.
A third challenge to the theory comes mainly from anthropologists and economists and emphasizes intergenerational transfers of resources. As an oversimplification, this idea is often referred to as "the grandmother hypothesis." Older members of population groups have survived a number of threats to their existence and can pass along to their younger family members information about successful avoidance or adaptation. Field experiments with prides of lions and olive baboons have failed to support this hypothesis. Moreover, while many grandparents are now contributing to the reproductive fitness of their children and grandchildren via transfer of resources, current evidence, while still incomplete, indicates that such elders were exceedingly rare within populations of our remote ancestors, when the evolution of species-specific gene actions would have evolved. Any such favorable alleles not expressed until those late stages of life would, therefore, have been vastly diluted by the alleles of their progeny.
A final challenge to evolutionary theory comes from a reexamination of the assumptions made by Hamilton in his influential 1966 paper. Baudisch, using different assumptions, demonstrates that, under some conditions, the force of natural selection can increase during aging. For example, species of rockfish that continue to grow beyond sexual maturity are much more likely to become predators rather than prey; as such, the force of natural selection would, indeed, become stronger, not weaker, as they age.
The message is that aging is non-adaptive. It did not evolve via a program of determinative gene actions that are designed to lead to the death of aging organisms because it is good for the species. Moreover, evolution has taught us that lifespans and their associated healthspans are plastic, thus providing a rationale for the potential effectiveness of future interventions.
Classes of Gene Action that Modulate Rates of Aging
If we are ever to intelligently intervene in one or more aging processes, we must be thoroughly familiar with the nature of the underlying gene actions. We must also be aware of the influences of idiosyncratic constitutional mutations, genetic polymorphisms, gene-gene interactions, environmental agents, gene-environmental interactions, and stochastic events. All of these complexities support what most experienced clinicians have learned in the course of their practices—namely that no two patients (even identical twins) age in precisely the same way; they share some commonalities but also have unique subsets of structural and functional impairments. Enhancements in the prevention and treatment of geriatric disorders will, therefore, require a more sophisticated and comprehensive understanding of the cellular, molecular, and integrative physiological underpinnings of these intra-specific variations in the patterns of aging.
The evolutionary biological theory of why we age provides clues as to how we age (i.e., the nature of the underlying gene actions). We can, in fact, outline twelve distinct classes of gene actions suggested by the evolutionary theory.
Class One: Good Alleles with Good Effects Early and Late: Longevity Assurance Genes
There are many examples of such genes. For instance, about 150 distinct human genetic loci have already been identified for the repair of DNA. Table 71-2 gives an example of the many different types of genetic loci involved in the oxidative-damage theory of aging. According to that venerable theory, various sources of oxidative damage to macromolecules, notably those produced as byproducts of the oxidative metabolism of mitochondria, are the primary causes of intrinsic biological aging. There is now mounting evidence, however, that this theory is an insufficient explanation for variations in longevity. Nevertheless, increasing evidence also suggests that the modulation of oxidative damage is important for major aspects of healthspan. A cogent example is the evidence that the genetic engineering of mice to provide high concentrations of a human cDNA for catalase directed to the mitochondria greatly ameliorates a form of congestive heart failure responsible for many geriatric hospital admissions. By the criteria of anatomic pathology and echocardiography, the features in aging mice are quite comparable to those observed in aging human subjects. This finding gives support to the proposition that even potentially good novel gene actions can also escape the force of natural selection. Nature has targeted catalase to peroxisomes, but it does not supplement mitochondrial protection by also directing some optimal amount of catalase to the mitochondria to decrease the steady state levels of a dangerous by-product of mitochondrial metabolism, peroxide (H2O2). In the presence of iron, H2O2 results in the synthesis of the highly reactive hydroxyl radical, with resulting damage to all classes of macromolecules in its immediate vicinity.
Table 71-2 Categories of Gene Action Relevant to the Oxidative Damage Theory of Aging |Favorite Table|Download (.pdf)
Table 71-2 Categories of Gene Action Relevant to the Oxidative Damage Theory of Aging
Structural and regulatory genes modulating genesis of free radicals
Examples: Cytochrome C Oxidase; P450 family
Structural and regulatory genes for scavenger enzymes
Examples: SOD-1,2,3; Catalase; γ Glutamyl cysteine synthetase
Genes regulating flux of nonenzymatic free radical scavengers
Examples: uric acid synthetic enzymes
Genes regulating target copy number
Examples: Genetic regulation of mitochondrial DNA replication, fusion, fission
Genes specifying target structure
Examples: Structural genes for chromatin proteins and membrane lipoproteins
Structural and regulatory genes for repair of target macromolecules
Examples: Specification of machinery for reversal, repair, tolerance of DNA damage
Genes specifying the orderly replacement of effete cells
Examples: Genes modulating DNA replication and cell cycle progression, apoptosis, growth factors, growth factor receptors, and stem cell biology
The last category (VII) of gene actions listed in Table 71-2 is relevant to a very important aspect of mammalian aging—the maintenance of proliferative homeostasis. As mammals age, there is a puzzling juxtaposition of both atrophy and inappropriate hyperplasia, often side by side. This can involve multiple tissues and is associated with numerous geriatric pathologies (Table 71-1). The hyperplasias may act as tumor promoters, leading to benign and malignantneoplasms.
Tissue atrophy may be attributable, in part, to the accumulating effects of cell apoptosis and necrosis from various causes and the failure of stem cells to compensate for the gradual attenuation of the replicative potentials of somatic cells, a process known as replicative senescence or the "Hayflick limit." It remains controversial as to whether or not an in vivo phenomenon exists that reflects the in vitro phenomenon of replicative senescence. The dominant (but not the only) mechanism responsible for replicative senescence is the loss of telomere repeat units from the ends of chromosomes. The germ line, many stem cells, and most cancers are protected from the erosion of telomeres by an enzyme known as telomerase, but this enzyme is absent from most somatic cells. Scientists shared a 2009 Nobel Prize for research in this field, the first such prize to recognize modern research in biogerontology. Cells that have exited the cell cycle typically do not undergo necrosis or apoptosis. They have a variety of phenotypes that characterize them as senescent; most notably a senescence-associated secretory phenotype (SASP) that has important consequences for regional pathology. In brief, such cells secrete a range of pro-inflammatory cytokines, metallothioneins, and mitogens. While such cells may be few in number within a given tissue, they can have a "field effect," thus driving the proliferation of neighboring epithelial cells, altering the connective tissue matrix, and contributing to sustained chronic inflammation. The latter is receiving increasing attention as a major concomitant of intrinsic biological aging.
As noted above, Category VII is also highly relevant to the emerging field of regenerative medicine. A possible generalization emerging from research on the interface of stem cell biology with the biology of aging is that a dominant factor in the decline of the success in mobilizing stem cells for the response to injury resides in the microenvironment of stem cells. This has been most effectively demonstrated in the repair of injury to skeletal muscle, in which satellite stem cells are mobilized quite effectively for repair in young mice but not in old mice. A clever parabiotic experiment provided compelling evidence for a non-cellular factor circulating from young to old mice that was capable of markedly ameliorating the deficient repair in the old mice. The satellite stem cell model has also demonstrated a switch, during adult myogenesis, from the Notch pathway—required for stem cell proliferation—to the Wnt pathway—required for effective differentiation. As is the case with so many signal transduction pathways, there is cross-talk between these two pathways. The situation is more complex, however, with evidence of an important role for a member of the transforming growth factor beta family of cytokines. There may also be a role for Klotho, a transmembrane protein with a structure suggestive of certain glycosidases. In any case, the clinical implications of this field of research are clear: as a much less complex and much safer alternative to the transplantation of stem cells, including those derived from the patient's own cells, clinicians might one day be able to inject small molecular weight compounds to "wake up" the patient's endogenous stem cells.
Class Two: Bad Alleles with Late-Life Penetrance: Idiosyncratic Constitutional Mutations
There is a national debate concerning health care reform, including the need for legislation for the protection of children with pre-existing genetic disorders. It seems likely, however, that we all have pre-existing conditions—namely gene actions that will lead to variable times of onset and variable severities of late-life disorders. Among these gene actions are individually rare but numerically numerous mutations that do not reach some phenotypic level of expression until middle age or beyond, when the effects will have escaped the force of natural selection. A prototypic example is Huntington's disease, one of a number of triplet repeat diseases; those unfortunate enough to have been born with the requisite number of CAG repeats (coding for a run of polyglutamines) in the affected locus will develop the disease after the peak of reproduction. More cogent examples, perhaps, come from rare, but pathogenetically informative, autosomal dominant mutations at three distinct loci resulting in "early" onset DAT. ("Early" here means younger than age 60; most patients are in late middle age when they're diagnosed with these forms of Alzheimer's disease.) To get some estimate of the genetic load of mutations that were known to lead to late-onset dementias of various types, the following is a systematic analysis of several editions of McKusick's Mendelian Inheritance in Man. In the 1975 edition, some 55 loci were identified from among 2336 listed at that time. Thirty years ago, the conventional wisdom was that there were ~100,000 protein-coding genes in the human genome. When the results of full genome sequencing were about to appear, geneticists participated in a contest (GeneSweep) to see who got closest to the sequencing results; the low bidder (~26,000 genes) won the event. Assuming that number is, in fact, correct (some new estimates suggest somewhat lower or higher figures), we can conclude that about 2.4% of these protein-coding genes, or a total of 624, have the potential to modulate one's susceptibility to dementing disorders of late life. That has to be an absolutely lower limit of how we can get in trouble with cognition as we age; however, as there are thousands of functionally distinct splice variants and thousands of DNA sequences that code for various families of RNA molecules (the "dark matter" of DNA), most of which seem likely to have a role in the regulation of gene expression. We are also learning about so-called "moonlighting proteins"—single proteins with two distinct functions.
Class Three: Bad Alleles Early with Good Effects Late: Paradoxical Antagonistic Pleiotropy
In theory, such genes can persist in a population either because of a comparatively recent founder effect, or because the allele participates in a balanced polymorphism. Elevated frequencies of the 4G allele and of the homozygous 4G4G genotype were found in centenarians and were associated with high levels of plasminogen activator inhibitor-1. Such high levels are predictive of recurring myocardial infarction in young men. It is, therefore, paradoxical that these high levels are associated with extreme longevity. Like all such studies with centenarians, however, one requires independent confirmations and more sophisticated controls, such as the use of centenarian progeny and their spouses.
Class Four: Bad Alleles Early and Late: Segmental Progeroid Syndromes
These conditions can be defined as genetic disorders that mimic, to various extents, many, but not all, senescent phenotypes found in the general population. Prototypic examples are the Werner syndrome (WS) ("Progeria of the Adult") and the Hutchinson-Gilford syndrome (HGPS) ("Progeria of Childhood"). WS results from homozygosity for null mutations at the WRN locus, which codes for a member of the RecQ family of helicases. To "do business" with DNA, one must first unwind the double helix. The WRN protein appears to have several functions in DNA replication, transcription, recombination, and repair. Telomeres are particularly favored substrates. Almost all HGPS patients suffer from the same C-terminal mutation in the Lamin A gene (LMNA), which codes for an intermediate filament that lines the nuclear membrane. The mutation results in the preferential use of a cryptic splice site, resulting in the deletion of a sequence of 50 amino acids. The abnormal gene product (progerin) no longer acts as a substrate for the enzymatic removal of a post-translational modification (farnesylation). This contributes to structurally distorted nuclei and abnormalities in gene expression and has, therefore, led to clinical trials employing farnesylation inhibitors. Other factors are likely to contribute to the pathology, however. Many segmental progeroid syndromes are characterized by genomic instability. For the case of WS, there is a 10–100 fold increase in mutation rates; large deletions are particularly common. There is a striking limitation of the replicative lifespans of somatic cells from such patients, probably related to unrepaired damage of a major product of oxidative damage, 8-oxo-2′-deoxyguanosine, at telomeres. Deficiency of the WRN helicase also results in pro-inflammatory gene actions characteristic of normative aging.
For the case of HGPS, in addition to the genomic instability related to the abnormal structures of nuclei, there are likely to be important aberrations in the regulation of gene expression. Evidence that the study of HGPS can provide insight into normative aging comes from the discovery that small, potentially pathogenetically relevant amounts of progerin can be found in cells from normal individuals and that the effects of this abnormal protein may be particularly relevant for stem cells.
Class Five: Good Alleles Early with Bad Effects Late: Antagonistic Pleiotropy
This category of gene action may contribute to three of our most devastating geriatric disorders—cancer, atherosclerosis, and DAT. The evidence that it plays a role in the pathogenesis of cancer comes from research referred to above, largely from the laboratory of Judith Campisi. The underlying hypothesis, as noted above, is that the repression of telomerase in somatic cells and alternative modalities of expediting exit from the mitotic cell cycle (such as DNA damage and oncogenic stimuli), evolved because they were adaptive for young, actively reproducing organisms, where they act as tumor suppressors. Later in life, however, the accumulation of replicative senescent cells may act in tumor promotion, both via mitogenic effects upon neighboring epithelial cells and via degradative actions upon the associated matrix, which may enhance local invasion by neoplastic cells. The arguments for roles in atherosclerosis and in Alzheimer's disease are much more speculative but worthy of more research. For the case of atherosclerosis, an argument can be made that macrophages play a primary pathogenetic role, particularly in the presence of functionally impaired, aged endothelial cells. The primary functions of macrophages are to engulf and destroy pathogens, for which purpose they employ a range of receptors including promiscuous "flypaper" receptors that can also recognize oxidized lipoproteins, which are of relevance to the pathogenesis of atherosclerosis. Given the exposure of modern humans to high-fat diets, late-life deleterious effects of the phagocytosis of oxidized lipoproteins may be an unfortunate tradeoff. Evidence for such a tradeoff has come from experiments with a mouse model of atherosclerosis (APOE knockout mice on a high-fat Western diet). When these mice were crossed with mice deficient in a class-A scavenger receptor (MSR-A), there was the expected amelioration of atherosclerosis. These hybrid mice, however, were shown to be highly susceptible to infection with Listeria monocytogenes or herpes simplex virus.
The evidence for a role of antagonistic pleiotropic gene action in the pathogenesis of DAT remains modest but is of considerable interest. Most such discussions revolve around polymorphic alleles of the APOE gene. The ancestral allele found within primates is the notorious epsilon 4 allele, the single-most important genetic susceptibility factor for common "sporadic" forms of DAT. In populations within the developed world, this has become a minor allele, however. Arguments have been developed that this allele had been (and still is, in some parts of the world) a major allele of Homo sapiens because it is under selection for its putative protection against a variety of infectious agents, either because of its enhancement of the immune/inflammatory response or because of its less efficient delivery of lipids to the membranes of infectious agents such as Trypanosoma brucei, which must obtain its lipids from its infected host. Yet another antagonistic pleiotropic hypothesis for the existence of DAT in our species comes from studies of polymorphic forms of a locus coding for an adaptor protein of importance in the metabolism of the beta amyloid precursor protein (APP), widely considered to be of central importance in the pathogenesis of DAT. The gene coding for this protein is formally designated as the APBB1 (Amyloid Beta A4 Precursor Protein-Binding, Family B, Member 1), but is widely referred to as FE65. Two laboratories have provided evidence that polymorphic alleles of this locus modulate susceptibility to what has been termed "very-late-onset Alzheimer's disease," that is to say, dementias with onsets after the peak ages of onset of the sporadic late-onset forms of Alzheimer's disease that are associated with the epsilon 4 allele of APOE (an association that peaks between 65 and 75 years of age). The FE65 polymorphism was shown to be independent of the impact of the APOE polymorphism. Its minor allele, the dominant allele found in other mammals and primates (species resistant to DAT) was shown to bind with much less avidity to APP. This is likely to significantly alter the modulation of functions of that protein, potentially including the role of the APP/FE65/TIP60 complex in transcription. The authors suggested that the new allele emerged as part of a suite of gene actions to enhance cognitive functions, but that this came with deleterious effects of APP metabolism in late life.
Class Six: Bad Alleles Early and Late: Nuclear and Mitochondrial Somatic Mutations
The accumulation of somatic mutations in the tissues of aging mammals has been well-documented and can be surprisingly high. For example, non-leaky mutations (i.e., severe loss of function mutations) have been shown to rise exponentially in the renal tubular epithelium of human kidneys, reaching levels—by about age 80—between 10–3 and 10–4. If one assumes that, for each such severe mutation, there are of the order of ten "leaky" mutations (i.e., mutations with diminished function), one can conclude that the levels of mutations at the single locus that was investigated could approach one in a hundred cells at advanced ages. In aging mice, the rates of increase and the types of nuclear somatic mutations were found to vary substantially from tissue to tissue. The frequencies of chromosomal mutations from the kidneys of aged F1 hybrid mice were found to be as high as one in three cells, but these remarkable results were likely related, in part, to transient exposures of the cells to ambient oxygen, which is now known to be particularly cytotoxic for murine cells. The late Howard J. Curtis, in an early example of the use of comparative gerontology, demonstrated that the levels of carbon tetrachloride-induced chromosomal mutations in mammals were inversely related to their lifespan potentials.
About 1500 genes (encoded by nuclei and mitochondrial DNA) contribute to the functioning of mitochondria. These targets clearly also contribute to the load of somatic mutations during aging and age-related diseases. The frequency of somatic mutation in mitochondrial DNA is ~500–1000-fold higher than it is in nuclear DNA. The tissues of individuals may, in fact, provide unique and dynamic mosaics of patterns of these mutations, serving as a sort of forensic fingerprint. While we have noted above that there is a waning of support for the idea that lifespan is limited by oxidative damage related to mitochondrial metabolism, arguments are still garnered for a primary role in aging processes. We will require more research on the characterization of the specific types of mitochondrial mutations that might be important actors in normative aging, because certain mutations may be more likely to out-compete wild type molecules within cells. Perhaps mutations with rearrangements of the mitochondrial genome produce more than one origin of replication and may lead to dominance ("homoplasty") of deleterious mutant molecules and thus the death of cells bearing that type of mutation.
In contrast to the present debate on the role of mitochondrial mutations and mitochondrial dysfunctions in the genesis of normative cellular aging, support for the roles of mitochondrial dysfunction in common geriatric diseases is becoming stronger. It is noted above that the evidence for an important role of mitochondrial dysfunction in a common form of geriatric congestive heart failure. The evidence of important roles of dysfunctional mitochondria in the genesis of Parkinson's disease is now quite compelling, given the roles of mutations at the DJ-1 and PINK1 loci. There is also great interest in observations of marked increases in the prevalence of mutations in the mitochondrial control region in the brains of patients with DAT. Given the key role of mitochondria in the control of apoptosis, aberrations in this pathway can skew the balance of cell proliferation and cell death and, therefore, modulate carcinogenesis; there is, in fact, growing interest in targeting mitochondria as treatments for cancer. Mitochondrial deletions appear to be key events in the genesis of sarcopenia. Mitochondrial mutations or dysfunctions may also be involved in the pathogenesis of atherosclerosis. Finally, a novel hypothesis has been proposed implicating changes that occur in utero as influences on diseases of aging. Epidemiologic data show an association of small birth weight with the increased risks of developing type 2 diabetes, metabolic syndrome, and cardiovascular disease in adulthood. Leduc and Levy propose that placental mitochondrial dysfunction is present in cases of placental insufficiency and may be a critical influence on the fetus leading to atherosclerosis in later life.
Epigenetic Shifts in Gene Expression
The next six classes of gene action (half of the total) can all be grouped under this heading. The term "epigenetic" refers to covalent chemical alterations in the expressions of DNA that are "on top of" the DNA—that is to say, in contrast to mutations or polymorphisms, they do not change the primary sequence or arrangement of nucleotides. These alterations produce the striking specificities of gene expression that define the numerous functionally distinct mammalian cell types and are essential components in the reactions to various injuries. These chemical changes are of two broad types. One type involves methylations of cytosines, typically at "islands" of CpG runs within domains, such as promoters and enhancers, that regulate gene expression. Such methylations are associated with gene silencing. A second general type involves alterations of specific amino acids within the histone proteins that coat the DNA, such as acetylations (associated with gene activation), deacetylations (associated with gene silencing), phosphorylations, methylations, ubiquitinylations, and ADP ribosylations. Methods for molecular epigenetic analysis are being applied at the single cell level, a key technical development for gerontology, because there are many shifts in the population heterogeneity of tissues during aging. Such studies are of interest in the assessment of yet another antagonistic pleiotropic mechanism of aging (Class Twelve).
Class Seven: Good Alleles Downregulated Early for Good Reasons: Adaptive Silencing
Given our evolutionary biological premise that nature does not really "care" much about the impact of gene actions that are going on late in the life course (i.e., when the force of natural selection is very weak), any switches in the degrees of gene expression that are initiated at some earlier stage of life because of their adaptive nature could, in principle, have a life of their own and could, therefore, continue to be downregulated or upregulated to a degree that could eventually become deleterious. The following are two examples of such downregulations, one observed in lab mice and the other in people. The first example occurs at the period of life when the somatic growth of mice is dramatically slowed and when resources are switched toward the business of reproduction. Beginning at around that time, a subset of genes that code for the synthesis of ribosomal proteins is silenced. Consequently, the rates of protein synthesis and protein turnover decline in most tissues. This leads to the accumulation of post-translationally modified proteins, including alterations that lead to diminished functions of those proteins, a process likely to contribute to senescent phenotypes. The second example involves the silencing, probably beginning with human puberty, of the estrogen receptor in human colonic mucosa, including regions of the colon that are particularly susceptible to the development of adenocarcinoma. This silencing continues throughout the adult lifespan and, given other lines of evidence for a role of this locus in the regulation of gene expression, likely plays a role in the development of colon cancer the elderly. Both examples involve gene silencing via methylations of CpG dinucleotides.
Class Eight: Good Alleles Upregulated Early for Good Reasons: Adaptive Expression
Here is the counterpoint for the mechanism mentioned above. An example is the enhanced expression of androgenic loci after sexual maturation, a likely contributor to the eventual emergence of benign prostatic hyperplasia later in life.
Class Nine: Good Alleles Inappropriately Upregulated in Late Life: Non-Adaptive Loss of Silencing
The best example of this mechanism comes from research on aging laboratory mice. To cite only one of several examples, a locus (Atp7a) on the X-inactivated chromosome (a normal dosage compensation event) increases expression in the spleen with increasing age. While the links to specific pathophysiological effects of this and other aberrant upregulations of genes remains to be elucidated, it seems likely that that there will be contributions to senescent phenotypes, given large number of such alterations.
Class Ten: Good Alleles Inappropriately Downregulated in Late Life: Non-Adaptive Epigenetic Loss of Expression
An example comes from studies of monozygotic human twins. Both gains and losses of expression were seen in a large number of loci. While it is possible that a number of these alterations were adaptive responses to age-related changes in physiology ("sageing"), given the very large number of these epigenetic alterations in gene expression, many are likely to have resulted in pathophysiological effects.
Class Eleven: Good Alleles for Females, Bad Effects for Males (and Vice-Versa): Sex-Based Antagonistic Pleiotropy
It would seem likely that evolution will have optimized gene actions somewhat differently in males versus females to enhance the fitness for these behaviorally, morphologically, and physiologically distinct organisms. Thus, alleles that evolve because of optimization for females may not be optimal for males, and vice versa—a sort of evolutionary battle of the sexes. In fruit flies, a large proportion of the alleles that have been optimized for male fertilization are detrimental for the fecundity of females, and vice versa. Arguments have also been made that, given the exclusive inheritance of mitochondria via the female germ line, mitochondrial structure, and function, including numerous interactions with the nuclear genome, may not work as well in males as they do in females. These effects can translate into differential longevities and patterns of late life dysfunction.
Class Twelve: Good or Bad Alleles Early with Bad Effects Late: Epigenetic Gambling and Epigenetic Drift
The histograms of Fig. 71-2 demonstrate a phenomenon that has puzzled gerontologists for generations. Why is it that, despite every effort to control genetics and environment, there are still marked variations in the lifespans of cohorts of experimental animals ranging from worms to mice? By far the best job of controlling for both genetics and environment has been achieved in experiments with C. elegans (Fig. 71-2). These organisms are hermaphrodites and, therefore, every diploid locus is driven to homozygosity—in other words, populations of these worms essentially consist of identical twins. They also can be grown in axenic media (devoid of bacteria) in temperature-controlled suspension cultures, including those with magnetic stirring rods, so that each individual worm "sees" the same food and waste and has equal opportunities for contact with other worms. Nevertheless, lifespans among such genetically defined organisms show considerable variability. This intrinsic variation is sufficient to produce a remarkable degree of overlap in the distributions of lifespans among wild type controls and a long-lived mutant population. These variations in lifespan are not heritable. Therefore, when considering the relative contributions of nature, nurture, and chance to intraspecific variations in lifespan, chance seems to be the dominating factor. Phenotypic evidence for stochastic variations is seen in the differences in the rates of development of aberrations in the ultrastructure of the skeletal muscles of aging worms. Some of these stochastic factors could involve somatic mutations—possibly in mitochondria—but their probable frequencies seem insufficient to explain these striking variations. Therefore, a theory suggests that, based upon the proposition that stochastic variations in gene expression develop in cohorts of all organisms, they enhance survival of the population. In a given environment, some worms may have developed a "lucky" set of gene expressions, while the patterns of gene expression may be quite nonadaptive in some of their identical twins. These individual outcomes could be changed in different environments. It is speculated that such "epigenetic gambling" may have evolved before the development of meiosis as a mechanism to ensure the survival of the species under unpredictable environments. Once initiated, "epigenetic drift" would ensue, eventually leading to departures from physiological homeostasis and senescent phenotypes. Published evidence indicates that epigenetic drift occurs within families of isogenic single cells.
Life-span distributions for individual Caenorhabditis elegans nematodes in isogenic populations of wild-type (green bars) and age-1 (red bars) strains. (Reproduced from TB Kirkwood et al: Mech Ageing Dev 126:439, 2005; with permission.)
The Identification of Signal Transduction Pathways Capable of Modulating Lifespan and Their Potential as Guides to Drug Targets
Earlier in the chapter evidence was cited for the existence of biochemical genetic pathways (or "signal transduction pathways") that, when appropriately modified, could extend the lifespans of several distantly related species—in other words, there is now evidence for the existence of "public" or shared mechanisms of aging. Because of claims that drugs that act in such pathways may enhance lifespans in model organisms (and, by extrapolation, perhaps in people), we need to review these pathways and the evidence for pharmacological interventions in more detail. We must keep in mind, however, that there is likely to be extensive cross-talk among these various pathways. Therefore, the "tweaking" of one pathway may have unanticipated effects on distant types of gene actions, especially given the high degree of genetic polymorphism in our species.
The best-documented such longevity-related signaling pathway is the IGF-insulin signaling pathway. The details of this pathway vary somewhat between nematodes, fruit flies, mice, and men, but a key common element is a downstream transcription factor (daf16 in worms and members of the FOXO family in mammals). In response to the regulation of its phosphorylation state, the transcription factors can enter the nucleus and expedite the transcription of a very large suite of genes with diverse activities that serve to enhance the protection of the soma from macromolecular damage. Downregulation of IGF-insulin signaling, as in the case of upstream hypomorphic mutations, releases inhibition of this transcription factor. Such decreased signaling is associated with enhanced longevity. It has also been shown to protect C. elegans from the types of proteotoxicity associated with a variety of human neurodegenerative disorders. It is apparent, however, that an optimum level of functioning of this pathway must have evolved to enhance reproductive fitness and survival in both good times and bad times, including the prevention of diabetes mellitus. That such optimization is subject to modulations by genetic polymorphisms in human subjects is suggested by initial findings that exceptionally healthy and exceptionally long-lived subjects (centenarians) exhibit an enrichment of variant hypomorphic alleles for the IGF1 receptor. In that study, it was interesting that the progeny of these centenarians carrying these alleles were more likely to have shorter stature. Such short stature has been associated with a specific haplotype for IGF1 and with longer lifespans of breeds of dogs.
A second pathway (actually two related pathways) involves mTOR signaling. The nomenclature derives from the discovery that this protein, a serine/threonine kinase, is a target of rapamycin, a drug in clinical use as an immunosuppressive agent, for the prevention of coronary artery restenosis and for the treatment of malignantneoplasms. The pathways are involved in a plethora of vital functions, including cell proliferation and survival, the recycling of intracellular materials, aspects of metabolism including nutrient sensing, the generation of ribosomes, and the regulation of protein translation. It has been referred to as the conductor of the cell-signaling symphony. Rapamycin enhances the lifespans and healthspans of mice even when administered at ages roughly comparable to those of 60-year-old humans. While there appeared to be no impact of this treatment upon the numbers of neoplasms, the dominant cause of death of these strains of lab mice, it remains to be seen if the increased lifespan was attributable, at least in part, to a slowing of the rate of growth of these neoplasms.
A third pathway of growing interest to biogerontologists involves a family of histone deacetylases known as sirtuins. Early experiments in yeast demonstrated that such enzymes are associated with the silencing of genes. Their functions are intimately related to the metabolism of nicotinamide adenine dinucleotide (NAD), a coenzyme used as an oxidizing or reducing agent in a variety of essential metabolic processes. Although still controversial, an important member of this family is SIRT1, homologues of which enhance the lifespans of model organisms. SIRT1 is a putative target of resveratrol, which is thought to activate the enzyme and, therefore, might enhance lifespan and, presumably healthspan as well. Health benefits of large doses of resveratrol have been noted in obese, diabetic mice with fatty livers. Resveratrol, a component of red wine, has caught the imagination of the general public and the investment dollars of some pharmaceutical companies. The latter are searching for more potent variants of the molecule; phase IIa clinical trials in patients with type 2 diabetes and other disorders are in progress.