The Critical Role of Metabolic Pathways in Aging

Aging is arguably the most universal contributor to the etiologies of metabolic decline and related diseases, including type 2 diabetes mellitus (T2DM), cardiovascular disease (CVD), and stroke (1). Insulin resistance (IR) represents a major component of metabolic syndrome (MS) and is commonly observed in older adults (2). Major impairments include unrestrained hepatic gluconeogenesis, adipose lipogenesis, and defective glycogen synthesis and glucose uptake in skeletal muscle. Abdominal obesity, which is commonly observed with aging, is a major contributor to IR and MS (3). Many older individuals are abdominally obese, despite a normal BMI, a factor that can decrease the utility of BMI as a predictor of T2DM in older persons. Aging is also associated with an increase in proinflammatory cytokines, which are known to interfere with insulin action. These cytokines are derived from both the age-associated accrual of visceral fat (VF) and secretion of proinflammatory cytokines by increasing numbers of senescent cells (4). Collectively, these age-related alterations in metabolism and body fat distribution are active participants in a vicious cycle that can accelerate the aging process and the onset of disease.

In humans, increased adiposity typically develops between the third and seventh decades of life and may increase, decrease, or remain unchanged thereafter. Computational tomography scans in men and women reveal that as age increases, subcutaneous fat (SF) decreases, whereas VF, which is the sum of fat depots inside the abdominal cavity, increases (3). VF accumulation is associated with IR and development of T2DM. In prospective studies, increased VF is an independent risk factor for coronary artery disease, stroke, and death (3). Aged rodents also develop increased fat mass, with a disproportionate increase in VF compared with SF mass, demonstrating similarity to human aging (5).

In humans and rodents, VF and SF are biologically distinct in terms of gene expression and secretory profiles of adipokines and inflammatory markers, including leptin, tumor necrosis factor-α, interleukin-6, and plasminogen activation inhibitor-1, all of which can contribute to the pathogenesis of IR and its age-associated chronic conditions (6). Furthermore, expression of adipokines from adipose tissue is regulated by nutrients, and these responses are exaggerated with aging (7). Another potential complication of VF accrual is the augmented release of free fatty acids that can reach the liver via portal circulation and interfere with hepatic insulin action. Surgical removal of VF in rodents has been shown to restore insulin sensitivity, improve lipid profiles, decrease hepatic triglycerides, and prolong life span (8). Interestingly, selective clearance of p16Ink4a-expressing cells in mice, many of which were found in adipose tissue, delayed the onset of age-related pathologies (9). Thus, the ability of VF removal to mitigate the cytokine secretory capacity of adipocytes and senescent cells in adipose tissue with aging may have contributed to the extended longevity observed in these rodents (8).

Another metabolic regulator derived from adipose tissue and linked to aging is the hormone adiponectin. As opposed to other fat-derived cytokines that oppose insulin action, adiponectin is an insulin sensitizer with anti-inflammatory properties and a potent activator of AMP-activated protein kinase (AMPK) (10). Adiponectin is paradoxically increased in lean individuals or in response to caloric restriction (CR), and high circulating levels are found in several long-lived mouse mutants (11). Furthermore, adiponectin concentrations are increased in families of centenarians, and, in these studies, a novel del/del polymorphism in its 3′-untranslated region was associated with increased adiponectin concentrations (12).

Sarcopenia represents another unfavorable phenotypic change observed with aging in humans. Skeletal muscle loss is a major contributor to the frailty syndrome of aging and can lead to reduced mobility and increased disability among older adults (13). Sarcopenia has also been linked to a reduction in energy expenditure and IR (14). The causes of sarcopenia are multifactorial, but the proinflammatory state associated with aging and obesity appears to be one key contributor (15). IR is believed to be another contributor to the decline in muscle quantity and quality, as it has been linked to reduced skeletal muscle strength, reduced protein synthesis rates, and accelerated skeletal muscle loss in humans (16). Thus, IR may be both a cause and consequence of sarcopenia, placing it in a vicious cycle of skeletal muscle loss and metabolic dysfunction.

Aging is associated with progressive loss in mitochondrial function in various tissues, including skeletal muscle (17). Mitochondria are the major source of reactive oxygen species (ROS) generation, which can lead to oxidative damage of macromolecules, including nuclear and mitochondrial DNA (mtDNA) (17). This has led many to embrace the mitochondrial theory of aging, which posits that mitochondrial dysfunction is a fundamental cause of cellular aging and senescence. Experimentally, a mouse mutant with deficient mtDNA repair capacity is characterized by substantial mtDNA damage and displays some features of accelerated aging, including impaired mitochondrial function and sarcopenia (18). However, the contribution of mtDNA mutations to mitochondrial decline and aging is unknown, as they typically comprise only a small fraction of the mitochondrial genome. Furthermore, no mouse mutant with increased mitochondrial biogenesis resulting in a longevity phenotype has been reported, and most mouse mutants with altered production of ROS or protection against oxidative stress have no significant life span alterations (19).

The relationships among mitochondria, glucose homeostasis, and disease have been another area of intense investigation. How impaired mitochondrial function may impinge on insulin signaling is complex but is thought to involve reduced or incomplete β-oxidation of fatty acid substrates in metabolically active tissues such as liver and skeletal muscle (20). Investigations have reported an association between IR and impaired glucose tolerance with decreased mitochondrial oxidative activity and ATP synthesis in elderly and type 2 diabetic subjects (21). Specifically, a study comparing mitochondrial function between healthy, lean, elderly volunteers and matched lean, healthy young subjects found a 40% reduction in mitochondrial activity and oxidative phosphorylation in the elderly subjects with IR (22). Similarly, young insulin-resistant offspring of diabetic parents had reduced mitochondrial activity and increased fat content in skeletal muscle compared with insulin-sensitive control subjects (23).

These and other data have led to the belief that metabolic decline can be reversed by increasing mitochondrial biogenesis in tissues. Certainly, exercise can potently stimulate mitochondrial biogenesis and activity in skeletal muscle and improve insulin action (24). However, East Indians have increased mitochondrial biogenesis in muscle, as compared with whites, yet they are more insulin resistant (25). Likewise, another study concluded that reductions in insulin sensitivity with aging were explained by increases in adiposity rather than advanced chronological age or decreased muscle mitochondrial function (26). Furthermore, it was noted that a higher mitochondrial ATP production capacity was observed in the men, yet the women were more insulin sensitive. Moreover, mice that overexpress the transcription factor peroxisome proliferator-activated receptor-γ coactivator-1α in muscle become severely insulin resistant in spite of increased mitochondrial content (27). Thus, the data do not affirm using mitochondrial content as a proxy of its function nor do they support stimulating mitochondrial biogenesis as a strategy to reverse metabolic dysfunction.

Age-related declines in various hormones have been linked to the aging process, prompting human studies investigating the potential for hormone-replacement strategies to modulate features of aging. The Women’s Health Initiative (WHI) was a landmark study that hypothesized that estrogen plus progesterone replacement would modulate several key age-related outcomes in older women (28). The WHI was stopped early, however, because hormone replacement increased the risk of CVD, cognitive decline, and breast cancer. This largely failed experiment raises the possibility that interactions among hormones that are important for reproduction in young females are different in postmenopausal women. These women are characterized biologically by increases in inflammatory cytokines, senescent cells, and breakdown in repair of cells, and estrogen may interact in a harmful way with this type of environment. Correction of hormonal deficiencies in WHI participants was also limited to ovarian steroids, (i.e., there was no replacement of growth hormone [GH]/insulin-like growth factor-1 [IGF-1], other hormones, or important circulating peptides to more youthful levels). Moreover, the antagonistic pleiotropy hypothesis of aging suggests that some pathways that are evolutionarily favorable for facilitation of development and reproduction can become unfavorable with aging. For example, the pubertal surge in GH/IGF-1 is necessary to promote growth and maturation during adolescence. However, sustained elevations in this axis throughout life could increase the risk for some cancers.

GH secretion markedly decreases with age, resulting in a concomitant decline in IGF-1 concentrations. Low peripheral IGF-1 levels in humans are associated with increased risk of numerous conditions including T2DM, CVD, sarcopenia, osteoporosis, and frailty (29). Similar to sex hormones, the adverse outcomes linked to declining GH/IGF-1 with age have been used as a rationale for GH treatment in the elderly, particularly in so-called “antiaging” clinics. However, chronological age is not an approved indication for GH therapy, and in the U.S., off-label use of GH as an antiaging therapy is specifically disallowed because of its serious side effects (30). Side effects of GH include elevated circulating IGF-1 levels, which is an established risk factor for many types of cancer. Furthermore, attenuated insulin/IGF-1 signaling (IIS) in invertebrates and rodents has led to extended life span (31,32), and functional IGF-1R mutations in humans have been linked to exceptional longevity (33).

Finally, the rate of living theory, which suggests that longevity is negatively related to metabolic rate, has linked hormones regulating energy metabolism with aging. Thyroid hormones are major regulators of energy expenditure. Experimentally induced hypothyroidism in young rats has resulted in extended life span (34), whereas inducing hyperthyroidism reduces longevity (35). Importantly, several mutant mice with exceptional longevity have decreased or nearly absent thyroid function (36). Hypothyroidism may modulate life span by lowering metabolic rate, core body temperature, and oxygen consumption, thereby reducing generation of ROS and associated oxidative damage. Interestingly, subclinical hypothyroidism, (i.e., elevated plasma thyrotropin with lower thyroid hormone concentrations) was associated with reduced mortality in women, is inherited in families with exceptional longevity, and has been linked to a polymorphism in the thyroid-stimulating hormone receptor (37). These studies suggest that thyroid replacement may be unnecessary and potentially even harmful in elderly subjects without clinical hypothyroidism.

A controversy exists over the physiologic role of IR in aging. In humans, IR accompanied by compensatory hyperinsulinemia has clearly been implicated as a risk factor for multiple age-related diseases (1). In support of this observation, improved longevity and multiple features of delayed aging have been described in mice, rats, and other mammals in which insulin sensitivity is increased by genetic mutations or CR (38). Paradoxically, a decrease (by genetic modulation) in the expression of key IIS pathway intermediates in model organisms is associated with life span extension (32). One explanation for this inconsistency is that there are substantial differences between invertebrate and mammalian aging. However, life span extension has now been described in several mouse mutants that are in fact insulin resistant (31). Thus, there is evidence that the role of attenuated IIS signaling in the mechanisms of aging and longevity is not limited to lower taxa, but is evolutionarily conserved.

Clinically, decreasing IR with drugs is a major strategy to relieve demands on pancreatic β-cells. However, it may remove a very important layer of protection from other tissues. Specifically, IR might actually serve as an adaptive mechanism in some tissues by preventing excess uptake of nutrients by cells. Indeed, differential IR of various organs can direct nutrient flux to adipose tissue, a factor that may have evolutionary benefits in regulating nutrient storage. Moreover, IR has been proposed as an antioxidant defense mechanism (39). Enhanced stress defenses may contribute to increased longevity of invertebrates with reduced IIS and may help explain the extension of longevity observed in some mutant mice with IR. Therefore, pharmacological reduction of IR in humans (without accompanying lifestyle modifications) may help minimize some hazards associated with T2DM, but this may occur at the expense of predisposing individuals to complications associated with excess nutrient flux and impaired stress defenses.

One of the most robust observations in the biology of aging is the ability of CR to delay or prevent a range of age-related processes and significantly extend life span (40). As a result, the CR paradigm has served as a valuable research model in the laboratory for uncovering possible modulators of life span in both simple and complex organisms. In mammals, the CR phenotype includes prevention of some of the potentially harmful changes that are typical of aging, such as increased adiposity and VF, or impaired hepatic and peripheral insulin action (8,41).

Biological characteristics of animals exposed to CR include numerous other changes in the transcriptome, metabolome, and proteome, as well as increases in stress hormones (corticosterone or cortisol depending on the species) (40). CR produces declines in insulin (and glucose), thyroid hormone, reproductive hormones, and GH/IGF-1 levels. Some of these effects represent preservation or restoration to levels typical of younger individuals (such as lower insulin and glucose concentrations), whereas, paradoxically, other changes resemble aging (such as low GH/IGF-1). Major hormonal and other functional changes thought to be important to the life-prolonging action of CR are illustrated in Fig. 1.

Because CR modulates a remarkable number of biologic systems, attempts to individually isolate the important factor(s) have largely been unsuccessful (except for lowering IIS). For example, reducing insulin and glucose by overexpression of a glucose transporter did not alter life span in transgenic mice (42). Thus, it is possible, and not surprising, that the concurrent interplay of several mechanisms is necessary to elicit the longevity-promoting effects of CR. However, as discussed below, directly altering the activities of three key nutrient sensors—AMPK, SIRT1, and mammalian target of rapamycin (mTOR)—has been reported to impact life span.

Most CR studies were conducted in rats and mice, but CR experiments in other mammalian species as well as taxonomically distant organisms suggest the universality of its impact on life span (43). However, it is worth noting that the effects of CR on longevity in mice can differ widely depending on genetic background (44). In nonhuman primates, CR decreases the risk of T2DM and other age-related diseases (43), but whether it extends life span in these animals is still open to debate. Importantly, many effects of CR discovered in animal studies have been reproduced in middle-aged humans, although suppression of IGF-1 is dependent on reducing protein rather than caloric intake (45). However, whether CR is capable of achieving life span extension in humans in addition to reducing disease-specific mortality risk remains to be resolved.

An important consideration regarding CR in humans involves weighing the possible benefits of CR against known quality of life concerns. Individuals that subject themselves to long-term, rigorous CR often experience issues with low bone density and muscle mass while complaining of feeling hungry, lethargic, and cold (46). Likewise, CR as a late-life intervention could prove counterproductive in frail, sarcopenic individuals in whom adequate caloric and protein intake are essential to maintaining bone and muscle mass. However, some of these concerns can be overcome by using a more moderate CR regimen with an exercise program. Such a strategy has been shown to promote weight loss and improve several health indices in humans while simultaneously preserving or improving bone mineral density, lean mass, strength, and aerobic capacity (47). Finally, extrapolating from rodent CR longevity studies, which often are initiated in very young animals, the benefits of a rigorous CR regimen for extending life may be very small given that many humans often do not initiate CR until adulthood (46).

The GH/IGF-1 axis declines with aging, yet smaller individuals within a species usually live longer (including ponies and small dogs). Mechanistically, life span extension has been achieved by several single-gene mutations in the IIS pathway in yeast, nematodes, and fruit flies (31,32). Life span extension has also been demonstrated in mutant mice with reduced function of the somatotropic axis, including Ames and Snell dwarf mice and mice lacking the GH receptor, all of which have decreased plasma IGF-1 concentrations (31,32). Although some of these mutant mice lack other hormones, their extended longevity appears to be due primarily to GH deficiency, as restoration of GH levels in Ames dwarf mice reverted their longevity to that of nonmutant controls (48).

Both animal and human studies have also linked reduced IGF-1 levels/signaling per se with reduced risk of many cancers as well as improved longevity. For instance, IGF-1 serum concentrations at 6 months of age in genetically heterogeneous mice were inversely associated with their subsequent life span (49), a conclusion supported by low serum IGF-1 concentrations in long-lived CR rodents (31,32). In addition, functional mutations have been identified in the human IGF-1R gene that result in altered IGF-1 signaling and are more common in centenarians than in younger control subjects (33). Paradoxically, low IGF-1 concentrations in humans have been associated with increased risk for CVD, stroke, T2DM, and osteoporosis. However, Ecuadorian individuals with mutations in the GH receptor (GHR) gene, which led to severe GHR and circulating IGF-1 deficiencies (the syndrome of Laron dwarfism), had reduced risk for T2DM, presumably due to the absence of the anti-insulinemic action of GH (50). These individuals also had lower incidence of cancers, but they did not live any longer than control subjects (50). Collectively, these results suggest that optimizing the IGF-1 axis to promote healthy aging in humans is more complex than originally appreciated and will require a greater understanding of its array of interactions and tissue specificity in order to strike the right balance throughout the life course.

In mammals, the sirtuin family of proteins is a seven-member group (SIRT1–7) of highly conserved, nicotinamide adenine dinucleotide-dependent protein deacetylases that function in the regulation of various biological processes. Mammalian SIRT1 shares the closest similarity to Sir2, a protein that was first shown to play an essential role in the control of yeast replicative life span (51). In mammals, SIRT1 appears to play an important role in controlling glucose metabolism, insulin action, fat storage, and nutrient sensing (52). Furthermore, SIRT1 deacetylates the inflammatory regulator nuclear factor-κB (NF-κB) (53), which may be a key player in IR and the MS.

Although there is no consistent evidence that SIRT1 content and/or activity declines with aging, activating this pathway might still be beneficial in preventing some manifestations of aging. Resveratrol is a plant-derived polyphenol that may activate SIRT1 (54) and has antioxidant, anti-inflammatory, and antitumorigenic properties. It has been suggested that through activation of SIRT1, resveratrol may function as a CR mimetic. Resveratrol treatment has been shown to increase life span in several organisms, including, most notably, high-fat–fed mice, which had improved insulin sensitivity, mitochondrial function, and survival (55). More recently, treatment of obese mice with SRT1720, a synthetic activator of SIRT1, resulted in similar improvements in survival as were observed in resveratrol-treated mice (56). These studies make a case that both natural and synthetic SIRT1 activators can improve health and survival, though the belief that these effects are SIRT1-dependent remains controversial.

Recent studies have shed further light on the emerging interplay among SIRT1, dietary, and hormonal perturbations. For example, one study found that the effects of CR on SIRT1 expression in the central nervous system of mice may be responsible for CR-induced suppression of GH secretion (57), thus linking sirtuins to nutrient-sensing and hormonal changes with aging. However, the dogma that CR uniformly increases SIRT1 has been challenged, as CR decreased SIRT1 activity in liver (58). This suggests that targeting SIRT1 to delay aging may be more complex than had been originally anticipated.

AMPK has been described as the energy gauge of the cell due to its exquisite sensitivity to changes in intracellular AMP levels or external cues such as nutrients and hormonal signals (59). AMPK plays a critical role in regulating whole-body energy balance with effects that are tissue-specific. For example, in the hypothalamus, AMPK activation in agouti-related peptide or pro-opiomelanocortin neurons stimulates food intake, whereas activation of AMPK in muscle promotes glucose transport, fatty acid oxidation, and mitochondrial biogenesis (59).

Given that AMPK is activated by various interventions such as exercise or CR and regulates many apparent beneficial adaptations, studies have attempted to determine whether AMPK activators can serve as exercise or CR mimetics to improve health and function. In sedentary mice, it was reported that 4 weeks of treatment with the AMPK agonist, AICAR, substantially enhanced running endurance (60). However, it is unknown whether AICAR is capable of reproducing these effects in humans or will prove safe at the necessary dosage. Likewise, the antidiabetic drug metformin, which is widely believed to be an AMPK activator, increased life span in yeast and mice (61), but not in rats (62). Furthermore, there are no data to support a role for metformin in human aging, but it is well-tolerated in patients with T2DM and has recently emerged as an anticancer agent (63). Finally, although the mechanism(s) by which AMPK activation modulates aging and disease risk is unclear, its ability to inhibit mTOR signaling may be critically involved.

The mTOR signaling pathway is highly conserved and integrates energy and growth factor signaling to cell growth and basic cellular processes such as RNA translation, stress resistance, and autophagy (64). In addition, mTOR is closely linked to components of IIS pathways and plays a pivotal role in energy metabolism and glucose homeostasis (65). Aberrant activation of mTOR signaling has been linked to several age-related diseases such as T2DM, cancer, Alzheimer disease, Parkinson disease, and CVD, leading to studies on the role of this pathway in metabolism, aging, and life span.

Inhibition of the mTOR signaling pathway by genetic or pharmacological intervention extends life span in invertebrates, including yeast, nematodes, and fruit flies (64). Likewise, inhibition of the downstream effector of mTOR, S6K, increases life span in worms and flies (66) and protects against diet-induced obesity and enhances insulin sensitivity in mice (67). Pharmacologic inhibition of mTOR (with rapamycin) was also shown to extend median and maximal life span in mice (68).

Although systemic knockdown of mTOR signaling in rodents results in longevity, greater fine-tuning will be required before translating to humans. Indeed, disruption of mTOR signaling has yielded diverse, tissue-specific effects in rodents. For instance, specific deletion of raptor, a component of mTOR complex in adipose tissue, protects against diet-induced obesity (69), whereas deletion of raptor in skeletal muscle is deleterious, resulting in a muscular dystrophy phenotype (70). In liver, overexpression of dominant-negative raptor improves insulin sensitivity, whereas inhibition of mTOR in the pancreas decreases insulin production by islets, leading to hypoinsulinemia and glucose intolerance (71). Meanwhile, increased mTOR activity in the brain or activation of S6K in the hypothalamus leads to decreased appetite via modulation of leptin and ciliary neurotrophic factor (72).

Rapamycin use in humans is also limited by its numerous side effects, including hyperglycemia, dyslipidemia, immunosuppression, vasospasm, and renal failure. Furthermore, human skeletal muscle from older adults has been shown to have impaired mTOR complex 1 activation and protein synthesis in response to an acute resistance exercise bout (73). Thus, systemic blockade of the mTOR complex 1 pathway could elicit an aging phenotype in skeletal muscle and accelerate the onset of sarcopenia and frailty. Therefore, further work is needed to optimize and refine approaches in terms of selectivity and safety before it can be contemplated as a strategy to delay the onset of age-related diseases in humans. The specific molecular targets currently believed to be most critical for the effects of rapamycin, resveratrol, and metformin on life span are depicted in Fig. 1.

Because humans age at different rates and die at different ages, with fewer survivors populating the age continuum at each successive decade, we can test if functional genetic variations in candidate genes involved in metabolic and endocrine systems contribute to death or longevity. One of our approaches to identify longevity genes is to look for polymorphisms in candidate genes that are enriched in long-lived humans, as compared with unrelated control subjects who are ∼30 years younger (Fig. 2). Perhaps the strongest example of a longevity gene is FOXO3A, which is one member of a family of transcription factors that mediate insulin action and stress resistance. Remarkably, the association of FOXO3A polymorphisms with human longevity has been replicated by eight independent groups in unrelated cohorts of long-lived subjects (74), a finding that supports evidence from model organisms that first identified FOXO transcription factors as regulators of life span. Another example of a gene linked to healthy aging and longevity is cholesterol ester transfer protein (CETP) gene. Homozygosity in the 405VV variant of CETP is associated with lower concentrations of CETP, higher concentrations of HDL cholesterol, and greater HDL particle size, all associated with protection against CVD and Alzheimer disease (74). A CETP inhibitor is currently in phase 3 industry trials, demonstrating the potential of genetic discoveries for directing drug development. Although some characteristics of exceptionally long-lived people may be unique to this highly selected group, further studies utilizing unbiased approaches and high-throughput platforms are virtually certain to lead to additional breakthroughs in the field.

We have described in this study how key features of metabolic signaling pathways can modulate age-related disease risk and longevity. Novel approaches and unexpected discoveries have revealed an important paradox regarding the role of IIS signaling in aging, which has had an important impact on the field moving forward. Efforts to uncover potential therapeutic targets to delay human aging have resulted in the discoveries that compounds such as rapamycin can meaningfully extend mammalian life span. Furthermore, genetic discovery in long-lived human populations (including epigenetics, which is rapidly emerging as a critical mechanism in transcriptional control) will undoubtedly lead to exciting breakthroughs in the field. Although there is certainly more work needed, substantial progress in our understanding of the interplay between metabolism and aging has made the goal of developing strategies to delay the onset of age-related diseases and improve quality of life realistic and attainable.

N.B.’s work on CR and hormones is supported by grants from the National Institute on Aging (NIA) (R01-AG-618381, P01-AG-021654, and P30-AG-038072 to the Einstein Nathan Shock Center) and the Glenn Foundation for Medical Research. D.M.H. is supported by an NIA-sponsored K99/R00 award (AG-037574). R.H.M. is supported by grants from the NIA (K08-AG-027462, K08-AG-027462-03S1, and R01-AG-035114). Work in A.B.’s laboratory was supported by the NIA (Grants P01-AG-031736 and R21-AG-038850), the Ellison Medical Foundation, and the Glenn Foundation for Medical Research.

No potential conflicts of interest relevant to this article were reported.

N.B., D.M.H., R.H.M., and A.B. researched data and wrote, reviewed, and edited the manuscript. N.B. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

The authors apologize to those whose work pertinent to this topic was not discussed or referenced due to strict format limitations or inadvertent omissions.