|Home | About | Journals | Submit | Contact Us | Français|
Multiple biological and environmental factors impact the life span of an organism. The endocrine system is a highly integrated physiological system in mammals that regulates metabolism, growth, reproduction, and response to stress, among other functions. As such, this pervasive entity has a major influence on aging and longevity. The growth hormone, insulin-like growth factor-1 and insulin pathways have been at the forefront of hormonal control of aging research in the last few years. Other hormones, including those from the thyroid and reproductive system have also been studied in terms of life span regulation. The relevance of these hormones to human longevity remains to be established, however the evidence from other species including yeast, nematodes, and flies suggest that evolutionarily well-conserved mechanisms are at play and the endocrine system is a key determinant.
More than 40 years ago, Everitt and coworkers (1965; Olsen and Everitt, 1965) removed the pituitary gland (hypophysectomized) of rats and noted a retardation of the aging process in tail tendon collagen fibers and a delayed onset of proteinuria. Building upon these observations and those of others, Everitt postulated that the pituitary gland secreted both anti-aging and life-maintaining factors. Integrating this information along with decades of data showing that food restriction in animals inhibited the secretion of most pituitary hormones (Mulinos and Pomerantz, 1940; Everitt and Porter, 1976) in addition to the delayed onset of age-related disease and life span extension, he performed experiments to determine the role of hormones on aging and longevity. The thought was that the anti-aging effects of calorie restriction (CR) may be due to a reduction in a pituitary aging factor (Everitt et al., 1980). Specific results from these studies included evidence that both hypophysectomy at 70 days of age and CR begun at 70 days markedly retarded collagen aging and abolished the rise in protein excretion with age. Hypophysectomy performed in middle-aged rats (400 days) also retarded collagen aging however, CR started at this same age had no effect. Proteinuria also progressively declined following both hypophysectomy performed at or CR started at 400 days. Furthermore, CR and hypophysectomy had similar effects on delaying the onset of several different age-related pathologies such as total tumor incidence (endocrine and non-endocrine), hindlimb paralysis, aortic wall thickening, and cardiac and kidney enlargement. These early studies also revealed that adrenal corticosteroids were necessary for viability, therefore subsequent hypophysectomy studies included the replacement of this hormone. Using this model system, Everitt and several collaborators showed that pituitary hormones played a major role in life span regulation.
The endocrine system is highly integrated in mammals. Therefore, life long alterations in one system almost certainly result in secondary endocrine effects in other systems. A case in point: the life long reduction in plasma insulin-like growth factor-1 (IGF-) levels found in growth hormone receptor/binding protein (GHR/BP) knockout mice results in reduced circulating insulin and glucose, decreased thyroid hormones, altered reproductive capacity, hyperprolactinemia and life span extension. This animal and other longevous mammalian phenotypes will be discussed in the context of their endocrine environment and the relationships to longevity and hormone levels. There is an abundance of literature on hormones and aging in other species including nematodes and flies, with most, if not all, directly supporting that found in mammals. Together, this interspecies information provides very strong evolutionary evidence implicating the endocrine system as a major regulator of aging and life span. Potential mechanisms of life span extension in mammals involve multiple systems including metabolism, growth, reproduction, stress resistance and repair. The endocrine system exerts significant control over metabolism and reproductive function and may also modulate antioxidant defense and stress resistance in specific tissues. These areas represent physiological mechanisms held in high regard in the aging research community as primary factors affecting longevity. The relevant literature in mammals regarding major hormones and longevity will be reviewed along with their interrelationships and significance to human longevity. A list of some of the animals discussed in this review and their phenotypic characteristics can be found in Table 1.
The endocrine pathways that have received the most attention in the aging field over the past several years are GH, IGF-1 and insulin. These interrelated pathways have been shown to have a significant place in longevity assurance in multiple species. Potential mechanisms of longevity in mammals causally related to GH and IGF-I include body size, metabolism, reproduction, and resistance to stress. Both a deficiency in GH and GH resistance (due to nonfunctional GH receptor activity) are strongly implicated in life span extension, several examples of which are presented in Table 1. In rodents, Ames and Snell dwarf mutants lack circulating GH, along with prolactin and thyrotropin and exhibit significant extensions (42-70%) in life span (Brown-Borg et al., 1996; Flurkey et al., 2001). The GH deficiency itself has been shown to play a remarkable role in delaying aging via enhanced antioxidative defense, enhanced stress resistance, reduced tumor burden and associated increases in insulin sensitivity (reviewed in Bartke and Brown-Borg, 2004). The GHR/BP KO or Laron mice exhibit GH resistance with high levels of GH due to a mutated receptor preventing both GH actions and an inactive negative feedback loop. These long-living animals share characteristics common to the GH deficient animals including small body size, enhanced antioxidant defenses, stress resistance and insulin sensitivity (Zhou et al., 1996; Coschigano et al., 1999; Hauck et al., 2001). The IGF-1 receptor +/− mice are considered partially IGF-1 resistant and also enjoy life span extension, enhanced antioxidant defenses and improved insulin sensitivity when compared to normal, wild type mice (Holzenberger et al., 2003). The hormone, Klotho, has been recently shown to inhibit IGF-1 signaling by binding to a cell-surface Klotho receptor and inhibiting FOXO phosphorylation, promoting FOXO nuclear translocation. Mice created to overexpress the klotho protein are long-lived (19-25%, female and males, respectively; Kuruso et al., 2005). These animals are insulin/IGF-1 resistant but are normal in size.
Growth hormone is a pituitary-derived hormone that has both somatic and metabolic functions. The somatic actions of GH are primarily mediated via the stimulation of liver IGF-1 secretion and its subsequent actions on bone, cartilage and muscle. Somatic growth is driven by components of this pathway; therefore, body size differences are obvious in many GH/IGF mammalian mutants. Those mutants with reduced signaling of this pathway are significantly smaller than wild-type controls with the range of differences directly related to the degree of suppression of the IGF-I pathway, in particular. IGF-I receptor knockdown mice are slightly smaller (8% in males) than wild-type mice while Ames and Snell dwarf mice are 66% smaller (one third the size of wild-type; Bartke, 1964; Holzenberger et al., 2003). The GHR/BP knockout mice are 40% the size of wild type siblings as adults (Coschigano et al., 1999). Furthermore, the fact that small body size is strongly associated with longevity also holds true for domestic dogs (Patronek et al., 1997) and humans (Samaras et al., 2003). The relationship between small body size and longevity is further supported by reports showing that mice selected for reduced body sizes live longer (Eklund and Bradford, 1977; Roberts, 1961; Rollo, 2002; Miller et al., 2002) and calorie restriction also reduces growth and adult body size (Duffy et al., 1990). Therefore, growth negatively influences life span in mammals (Rollo, 2002) and likely other species.
The metabolic activity of GH includes but is not limited to actions on glucose regulation, lipolysis and oxidative metabolism. The actions of GH on glucose regulation are well known. This hormone is considered a diabetogenic factor in that it opposes the actions of insulin. GH elevates plasma glucose concentrations by stimulating gluconeogenesis and glycogenolysis and inhibiting glucose uptake at the tissue level. Humans and genetically engineered mice with elevated plasma GH levels exhibit hyperinsulinemia, hyperglycemia, and/or insulin resistance (Balbis et al., 1996; Dominici et al., 1998). More than 50% of humans with supraphysiological GH levels become diabetic and develop micro- and macrovascular complications associated with hyperglycemia. In sharp contrast, GH-deficient and GH-resistant animals have low glucose, low blood insulin and increased insulin sensitivity (Borg et al., 1995; Dominici et al., 2000, 2003; Hauck et al., 2001). Administration of a single dose of insulin induced a larger decrease in blood glucose levels in Ames and GHR/BP KO mice than in corresponding wild-type mice (Coschigano et al., 1999). However, disposing of glucose was reduced in both types of mice, indicating a decreased ability to increase insulin secretion in response to an acute glucose challenge. The heightened insulin responsiveness in these animals may be due to the elevated levels of liver insulin receptors (IR; protein) in GHR/BP KO mice and higher levels of IR, IRS-1 and IRS-2 (downstream effectors of the insulin receptor) in liver tissues of Ames mice (Dominici et al., 2000; 2002). This evidence is supported by decreased numbers of pancreatic islets in the Ames mouse (Parsons et al., 1995). Thus, two different mechanisms may be involved. Caloric restriction, long known to increase life span in rodents (McCay et al., 1935; Weindruch and Walford, 1982), flies (Chapman et al., 1996), and likely primates (Lane et al., 1996) also reduces insulin levels in the blood. Despite normal blood glucose concentrations and clearance, liver specific IGF-I deficient mice (Yakar et al., 1999) are hyperinsulinemic and exhibit increased insulin resistance due to GH hypersecretion and β-cell hyperplasia. Therefore, these mice do not exhibit life span extension likely because signaling of this pathway is not truly reduced (Yakar et al., 2001; Yu et al., 2003). Mechanistically, reducing circulating glucose delays aging by decreasing the accumulation and detrimental processes associated with glycation end products (Reiser, 1998; Baynes and Monnier, 1989) and reducing metabolism (less fuel) and associated metabolic ROS generation.
As mentioned above, reduced insulin secretion and enhanced insulin sensitivity are common to long-living mutant mice. Insulin sensitivity declines with age (Facchini et al., 2001) and is specifically related to visceral fat stores; sensitivity increases with reduction in visceral fat (Barzilai et al., 1998; 1999). The FIRKO mouse lacks expression of the insulin receptor specifically in fat tissue. These insulin receptor knockout animals have significantly reduced fat mass, are protected against age-related obesity and decreased insulin sensitivity. The decrease in fat mass occurs in the absence of reduced food intake. In addition, these mice do not develop diabetes or glucose intolerance and, importantly, live 18% longer than wild-type control mice (Bluher et al., 2002; 2003). Studies of centenarians have shown that greatly enhanced insulin sensitivity is strongly correlated with longevity in this unique population (Paolisso et al., 1996; 2001; Kojima et al., 2004; van Heemst et al., 2005).
Reproductive competence is significantly affected by GH and IGF signaling pathways. The GHR/BP KO mice exhibit delayed puberty and most animals are fertile. However, the Ames and Snell dwarf mice have significant delays in sexual maturation; the females are infertile while males are considered subfertile (Bartke, 2000) although the degree of gonadal function is dependent on the background strain. The IGF-I receptor knockdown mice exhibit normal puberty and fertility although the males have not been examined thoroughly. Concomitant with lower plasma IGF-I concentrations, calorie-restricted rodents exhibit delayed puberty, reduced litter size, and lower fecundity (Weindruch and Sohal, 1997; Meites, 1993; Holehan and Merry, 1985). The effects of CR on reproductive function result from alterations in neuroendocrine factors including IGF-I. The klotho overexpressing mice also display reduced fecundity supporting evidence indicating that GH and IGF-1 affect reproductive capacity (Kuruso et al., 2005).
An additional example of GH action on reproduction can be taken from GH transgenic mice. These mice have high circulating GH levels and live half as long (12 months) as wild type siblings (Steger et al., 1993). Female GH transgenic mice exhibit early puberty and increased ovulation rates; however, fertility is reduced and the reproductive life span is severely shortened when compared to wild-type mice (degree of suppression appears to depend on level of circulating GH; Naar et al., 1991; Bartke et al., 1994; Cecim et al., 1995). Male mice with elevated GH levels are fertile, but exhibit alterations in sexual behavior, reproductive pituitary hormone secretion, and a reduced reproductive life span (Chandrashekar et al., 1988; Bartke et al., 1994; Meliska et al., 1997). High levels of GH and IGF-1 are also associated with accelerated aging and tumor promotion (Burroughs et al., 1999; Yu and Rohan, 2000; Swerdlow et al., 2002).
Resistance to oxidative stress is another factor significantly affected by altered signaling of the GH, IGF-1 and insulin pathways and closely regarded in the aging field as a major player in life span determination. The free radical theory of aging proposes that endogenously generated reactive oxygen species (ROS) cause aging via damage to DNA, proteins, and lipids (Harman, 1956; 1988). The effects of GH and IGF-I on oxidative metabolism and oxidative damage have been documented in numerous reports. GH is an anabolic factor that increases cellular metabolism. Increased metabolic activity (glucose oxidation and oxygen consumption) leads to increased oxidative phosphorylation and increased production of ROS as byproducts of metabolism. Rollo and coworkers (1996) showed that GH overexpression increased superoxide radicals and oxidative damage to membrane lipids (lipid peroxidation). Perhaps adaptively, tissues from mice with elevated plasma GH exhibit significantly reduced levels of antioxidative enzymes including manganese superoxide dismutase (MnSOD), copper-zinc SOD, catalase, and glutathione peroxidase (GPX; Brown-Borg et al., 1999; Brown-Borg and Rakoczy, 2000; Hauck and Bartke, 2001). In addition, direct effects of GH and IGF-I in vitro strongly support the in vivo data showing that these two hormones directly down-regulate catalase activity, GPX protein and activity and MnSOD protein in hepatocytes from normal mice (Brown-Borg et al., 2002). Other protein hormones tested (luteinizing hormone and prolactin) had little if any effect on antioxidant enzymes, indicating specificity of GH and IGF-I (Brown-Borg et al., 2002).
In stark contrast to suppressed antioxidative defense induced by GH, there are several studies demonstrating enhanced defense capacity of this system when GH is deficient. Ames dwarf mice exhibit elevated catalase, GPX and SOD levels in liver, kidney, heart, and hypothalamic tissues (Brown-Borg et al., 1999, Brown-Borg and Rakoczy, 2002; Hauck and Bartke, 2000; Brown-Borg, unpublished data). Muscle tissue GPX activity is preserved in dwarf mice at several ages following both acute and chronic exercise while that from wild-type mice declines with age (Romanick et al., 2004). Interestingly, GH replacement in dwarf mice down-regulates catalase, GPX, and MnSOD proteins and activities in both young and adult animals (Brown-Borg et al., 2003). Metallothionein and glutathione both exhibit ROS scavenging abilities, and levels of these are significantly increased in multiple tissues from the dwarf mouse (Meyer et al., 2003; Brown-Borg et al., 2001). The amino acid methionine, whose metabolic pathway feeds cysteine residues into the GSH pathway, is also highly upregulated in the Ames mouse (Brown-Borg et al., 2005; Uthus and Brown-Borg, 2003; 2006).
In general, circulating GH/IGF-I status is negatively correlated with antioxidative capacity; high levels of plasma GH suppress this defense mechanism while the absence or low levels of GH/IGF-I enhance the ability of an organism to counter oxidants. The signal transduction pathways used by GH, IGF-I, and insulin overlap and influence redox-regulated transcription factors (Meyer et al., 1994). Insulin resistance in diabetes and obesity is improved by antioxidants (Blair et a., 1999; Hansen et al., 1999; Trosh et al., 1999). This point is further illustrated in p66shc mice with disrupted IGF-I/insulin signaling. p66shc is an adaptor protein that is crucial to IGF-I/insulin receptor signaling. Disruption of this gene results in mice with enhanced resistance to oxidants and extended life span when compared to wild-type control mice (Migliaccio et al., 1999). As mentioned earlier, the IGF1R knockdown mice are also resistant to oxidative stress (Holzenberger et al., 2003) as are mice that overexpress the protein, klotho (Kuruso et al., 2005).
Mitochondrial oxidant production (liver H2O2) is significantly lower in dwarf mice, possibly indicating decreased metabolic activity in the absence of thyroid hormone and GH (Brown-Borg et al., 2001). Consequent to reduced ROS and elevated antioxidants, dwarf mice exhibit lower nuclear DNA, mitochondrial DNA, and protein oxidative damage in several tissues (Brown-Borg et al., 2001; Sanz et al., 2002). Functionally, these GH-deficient mice out-survive their GH sufficient counterparts following administration of the systemic oxidative stressor paraquat (Bartke et al., 2000). The IGF-I receptor knockdown mice challenged with paraquat also lived significantly longer than mice with normal levels of IGF-I receptors (wild-type; Holzenberger et al., 2003).
Reduced GH/IGF and insulin signaling is also associated with increased resistance to other stressors including heat shock, UV, and gamma irradiation. Overall, the reported studies strongly support the notion that GH, IGF-I and insulin signaling pathways are intimately involved in the modulation of oxidative stress. The suppressive effect of GH on multiple components of the antioxidant system and consequent oxidative damage may be one mechanistic reason that levels of this hormone decline with aging.
GH-deficient dwarf mice and rats resist cancer development following administration of chemical carcinogen (Bielschowsky and Bielschowsky, 1961, Ramsey et al., 2002) and exhibit reduced growth of transplanted tumors (Rennels et al., 1965). Spontaneous tumor incidence is delayed and the severity reduced in the hypopituitary dwarf mice (Ikeno et al., 2003; Flurkey et al., 2001). Moreover, tumor growth in IGF-I deficient mice is reduced relative to control mice (Yang, 1996). Cancer incidence is also lessened in calorie-restricted rodents (Weindruch and Walford, 1982). In addition, GH/IGF-I deficient mice develop significantly less osteoarthritis than wild-type mice (Silberberg, 1972). It has been postulated that stress resistance is coordinately increased (heat shock proteins, antioxidants, detoxification systems, metal chelators, and repair systems) and this up regulation results in multi-stress resistance to different stressors (Rollo, 2002; Jazwinski, 1996; Martin et al., 1996). Work by Murakami and coworkers (2003) exemplifies this idea well in showing that fibroblasts from the skin of long-living Snell dwarf mice are resistant to multiple forms of cellular stress including heat, paraquat, H2O2, UV light, and the toxic metal cadmium. This data is supported by similar work in the Ames mouse (Salmon et al., 2005).
Fewer reports describe the actions of insulin itself on oxidative stress. Chronic hyperglycemia and hyperinsulinemia are known pro-oxidant factors (Mezzetti et al., 1996; Guigliano et al., 1996; Paolisso and Guigliano, 1996). Oxidative stress is associated with impaired insulin signaling (Kanety et al., 1998). Studies have also shown that insulin-mediated glucose uptake is improved with vitamin E administration along with lower plasma free radical concentrations (Paolisso et al., 1994; 1998). Healthy centenarians have been shown to exhibit reduced oxidative stress and a preserved insulin-mediated glucose uptake response (Paolisso et al., 1998). In vitro, insulin inhibits catalase synthesis in rat liver thus lowering catalase activity (Xu and Badr, 1999) and catalase and MnSOD in rat brainstem (Radijicic et al., 1997).
Related data more directly associates lower insulin levels with longevity. Evans and Meyer (1992) showed that in rats life-long feeding of the insulin-lowering drug, chromium picolinate, increased median life span by 36%. Daily injection of ad libitum fed mice with an insulin-sensitizing biguanide (phenformin) resulted in a 23% increase in life span over that of control animals (Dilman and Anisimov, 1980). Thus, a reduction in GH, IGF-1 and insulin signaling results in increased insulin sensitivity and reduced oxidative stress, likely significant factors in achieving long health and life spans.
There is evidence of a phenomenon of longevity in humans similar to that of Ames mice. Individuals with mutations in the PROP-1 gene, the same gene mutated in the Ames dwarf, also display characteristics of hypopituitarism. These individuals did not receive hormone replacement and were shown to reach very advanced ages, exceeding the average life expectancy of the general population (Krzisnik et al., 1999). In addition, the employment history and educational performance of these human dwarfs indicated no significant cognitive issues. On the other hand, there is evidence of reduced life span in a pedigree of human dwarfs with untreated isolated GH deficiency (Besson et al., 2003) possibly due to increased adiposity leading to insulin resistance in these GH deficienct subjects. Other studies have shown that GH deficiency is associated with serious functional deficits and increased risk of premature death due to cardiovascular complications (Sacca et al., 1994; Bates et al., 1996). Mutations similar to Pit-1 and Prop-1 have been reported in humans and those children affected require thyroid hormone and GH treatment (Tatsumi et al., 1992; Pfaffle et al., 1996; Fluck et al., 1998).
We also know that in the case of GH excess in both humans (acromegaly) and mice (transgene expression), life span is severely reduced (Bengtsson et al., 1988; Orme et al., 1998; Wolf et al., 1993; Cecim et al., 1994; Bartke et a., 1998). Edema, arthralgia, symptoms of carpal tunnel syndrome and insulin resistance are among the detriments experienced by individuals administered GH (Blackman et al., 2002). In GH transgenic mice, signs of premature aging are abundant and early death is related to glomerulonephritis and glomerulosclerosis in addition the development of mammary and liver tumors (Wanke et al., 1992; Steger et al., 1993; Yang et al., 1993). IGF-1 transgenic mice do not experience such severe pathological changes as GH transgenic mice (Doi et al., 1988).
The major concern of GH therapy in elderly humans focuses on the adverse effects of this hormone. Short-term studies using GH or IGF-1 in healthy elderly individuals have shown modest improvement in body composition (Rudman et al., 1990). Other studies have confirmed these findings but also suggest that the benefits are neither clear nor consistent (Papadakis et al., 1996) as no parallel functional improvements (strength, endurance, mood, and mental status; Lange et al., 2002; Blackman et al., 2002). Long-term studies are lacking and are necessary to determine whether this hormone has therapeutic potential. This is true for most hormone replacement-type studies.
Prolactin is a pituitary-derived multifunctional hormone that plays species specific modulatory roles in many physiological systems. Prolactin secretion is under tonic inhibitory control of the hypothalamus by dopamine. Much of our knowledge regarding prolactin has been derived from animals that lack this hormone or its receptor due to spontaneous mutations (Ames, Snell dwarfs) or genetic engineering (prolactin gene knockout, prolactin receptor knockout). In addition to its well-documented effects in the reproductive system in target tissues such as the mammary gland and prostate, prolactin has been shown to promote β-cell growth and insulin production in the pancreas, to alter immune function and to affect metabolic homeostasis including glucose tolerance and lipolysis (Neville et al., 2002; Sorenson and Brelje, 1997; Sorenson et al., 1987; Costello and Franklin, 1982; Fielder and Talamantes, 1987; Houseknecht et al., 1996). However, the physiological effects can be species specific, transient, and sometimes present only in specific stages of life (prepubertal versus adult). Prolactin-receptor mediated signaling is not likely involved in regulating insulin sensitivity in adult mice as both hypoprolactinemic Ames mice and hyperprolactinemic GHR/BP KO mice exhibit enhanced responses to insulin (reviewed in Bartke 2005). In agreement, a recent report utilizing prolactin deficient animals indicates that a lack of prolactin has negligible metabolic effects in mice (LaPensee et al., 2006).
Little information exists regarding the role of prolactin in longevity. One study using NZB/NZW mice showed that prolactin administration shortened the life span of these mice whereas treatment with a prolactin antagonist (bromocriptine) prolonged the life span (McMurray et al., 1991; Walker et al., 1998). NZB/NZW mice exhibit many aspects of severe systemic lupus erythematosus and have non-characteristic short life spans even in the control animals. These treatments were initiated in this line of mice as a relatively strong association between hyperprolactinemia and the incidence of systemic lupus erythematosus in humans had been discovered (Walker et al., 2001). Therefore, the changes in life span using prolactin in this study are not truly reflective of the hormone longevity potential. A study by Flurkey and coworkers (2001) replaced prolactin with pituitary transplantation in Snell dwarf mice (which are prolactin, growth hormone and thyrotropin-deficient) and reported that fertility was restored and the enhanced longevity of the strain was maintained. A similar study by Bartke and coworkers using Ames dwarf mice found that prolactin replacement (via pituitary transplantation) somewhat shortened the life span of these long-living mutants (Bartke, personal communication).
Early work by Bartke (1965) established that prolactin is absolutely required for luteal function, implantation and maintenance of pregnancy in mice. Relationships between the level of gonadal function and longevity exist. However, it is unlikely that the gonad is a major player in the case of dwarf mice as Snell and Ames mice with different reproductive capacities (fertile or sterile) exhibit similar average and maximal life spans (Bartke, 2000). Finally, only a few reports describe the direct effects of prolactin on antioxidative defense. A positive correlation between prolactin and mammary gland SOD levels was reported (Bolzan et al., 1997), and SOD expression was induced by prolactin in the rat corpus luteum (Sugino et al., 1998). In vitro data in mouse hepatocytes indicated no effect of prolactin on the expression of several antioxidative enzymes (Brown-Borg et al., 2002). Considering these and other data, hyperprolactinemia may accelerate autoimmune disorders by suppressing immune function thus leading to increased mortality in a subset of immunocompromised patients. However, neither hyper- nor hypoprolactinemia by itself has been reported to have a significant effect on longevity in healthy mammals.
Thyroid hormones are secreted by the thyroid gland under the control of the pituitary hormone, thyroid stimulating hormone (TSH) which, in turn, is stimulated by hypothalamic secretion of thyroid releasing hormone. The two active thyroid hormone molecules are L-thyroxine (3, 5, 3', 5' tetraidothyronine; T4), the major circulating hormone, and 3, 5, 3' triiodothyronine, also known as T3. T4 exerts its effects as a prime regulator of intermediary metabolism in virtually every tissue, and its action is mediated by binding and activation of the nuclear receptors TR (NR1A1) and TR (NR1A2) (Weinberger et al., 1986). It accelerates the rate at which cells oxidize fuel, leading to increased thermogenesis and up-regulates enzymes and cytochromes involved in mitochondrial function (Weisner et al., 1992; Izquierdo et al., 1990), proteins involved in glucose transport and metabolism (Weinstein et al., 1990; Hoppner and Seitz, 1989; Wall et al., 1989) and in fatty acid metabolism (Swiercynski et al., 1991; Stapleton et al., 1990).
The most direct evidence for a role of thyroid hormones in regulating longevity is derived from studies by Ooka and coworkers (1983; 1986). These investigators induced hypothyroidism in young rats resulting in T4 levels at ~2/3 of normal for up to 20 months of age. These animals lived four months longer than euthyroid control rats. This life span modification is relatively modest in comparison to the life extension documented in GH-deficient hypothyroid dwarf mice. Subsequent studies using hyperthyroid rats produced a three-month life span reduction relative to euthyroid control animals (Ooka and Shinkai, 1986).
In Ames and Snell dwarf mice, characteristics of hypothyroidism are prominent. These mice retain infantile facial features in addition to small body size throughout life reflecting deficiencies of both GH and thyroid hormones. Thyroxine (T4) levels in Ames mice are below the detectability limits of radioimmunoassay (0.4 ng/ml; Brown-Borg et al., 1995). Another very long-living mammal, the naked mole rat (known to live more than 25 years), also has very low levels of circulating T4 (0.004 mg/dl; Buffenstein, 2005).
Both reduced core body temperature and metabolic rates are consistent with hypothyroidism. Ames mice were observed to exhibit core body temperatures 1.5°C lower than wild type mice (Hunter et al., 1999) and Hauck and coworkers (2001) also reported reduced body temperatures in GHR/BP KO mice. The role of body temperature in life span regulation is not entirely clear, however, a recent report by Conti and coworkers (2006) showed a 15% extension in life span in transgenic mice that overexpressed UCP2 in hypocretin neurons, effectively lowering core body temperature by 0.3-0.5°C. Unfortunately, thyroid hormone levels were not reported in this study. These data are interesting however, somewhat confounded by the reported short mean life span of wild type animals (18.5 and 23 months, females and males, respectively).
Ames and Snell dwarf mutant mice live 40-70% longer than mice with normal thyroid hormone levels. Short-term (11 weeks) replacement of T4 and GH in dwarf mice had no effect on longevity (Vergara et al., 2004). However, life-long thyroxine administration in dwarf mice shortened life span to 83% of that of control dwarf mice (Vergara et al., 2004) yet the treated dwarf mice still live significantly longer than normal, wild type mice. Early life T4 (and IGF-1) levels have been shown to be predictive of life span in mice (Harper et al., 2003). An important aspect to keep in mind is that these dwarf mice not only live longer than wild type mice but also exhibit characteristics of delayed or decelerated aging of their connective tissue, immune system, joint pathology and in the occurrence fatal neoplastic disease (Silberberg, 1972; Flurkey et al., 2001; Ikeno et al., 2003; Vergara et al., 2004).
The reduced oxygen consumption and reduced core body temperature of dwarf mice suggest that metabolic rate is also reduced in these long living mutants. Metabolic rate was shown to be significantly reduced in Snell dwarf mice (Gruneberg, 1952). The reduced metabolism of dwarf mice may contribute to their long life span by decreasing reactive oxygen species generation leading to less oxidative damage to cellular components and tissues. Indeed, studies using liver and muscle mitochondria from Ames mice reflect lower H2O2 production (Romanick et al., 2004). However, these animals consume more food per gram of body weight suggesting that the opposite is true (Mattison et al., 2000). There is an abundance of literature regarding the role of thyroid hormone on metabolism, and particularly, oxygen metabolism that has been reviewed elsewhere. This discussion will be limited to a few examples.
The relationship of thyroid status to oxygen consumption and oxygen radical production by mitochondria and oxidative DNA damage have been studied experimentally in rats rendered hyper- and hypothyroid (Lopez-Torres et al., 2000; Asayama et al., 1987; and many others). Low thyroid hormone levels were associated with reduced ROS generation, less oxidative damage and inconsistent changes in antioxidative defense enzyme levels. The opposite findings were observed in hyperthyroid animals. In general, it appears that hyperthyroidism increases mitochondrial production of ROS in a variety of cells and tissues, an effect variably associated with a compensatory increase in antioxidant defenses (Asayama et al., 1987; Mano et al., 1995; Asayama and Kato, 1990; Tapia et al., 1999; Yavuz et al., 2004).
Regarding stress resistance, it has been shown that activation of CAR decreases serum T4 levels in mice (Maglich et al., 2004). The orphan nuclear receptor CAR (NR1I3) has been characterized as a central component in the coordinate response to xenobiotic and endobiotic stress. In this study, it was demonstrated that CAR plays a pivotal function in energy homeostasis via thyroid hormone regulation. This evidence suggests an additional hormonal mechanism regulating overall stress resistance in long-living animals and possibly humans.
In humans, a deficiency of thyrotropin resulting in low T4 and T3 levels produces a range of outcomes depending on the age of the individual. Young adults with subclinical hypothyroidism can exhibit weakness, fatigue, peripheral edema, weight gain, cold intolerance, depression, memory problems, cardiac dysfunction and an increased rate of cardiovascular events (Surks et al., 2004) as thyroid hormone deficiency can affect virtually all body functions. However, in elderly patients, no consistent associations between depression, memory loss or other symptoms and subclinical hypothyroidism have been observed (Meneilly, 2005). While this syndrome is extremely common in the elderly population, many older adults receive treatment without any randomized controlled trials that support this therapy (Surks et al., 2004). Based on studies of CR, and its effects on reductions in metabolic rate and longevity, administration of thyroid hormones could theoretically accelerate aging (Meneilly, 2005). In very old patients (>85 years), the biological response to hormone replacement may differ from that of younger populations, therefore hormone therapy in this group of individuals is called into question. With longevity as a concern, Gussekloo and coworkers (2004) found that in a cohort of 85 year olds, decreased levels of free thyroxine were strongly associated with longer life span, in agreement with an earlier report (Parle et al., 2001). These data contrast a study in middle-aged males with subclinical hypothyroidism showing higher rates of mortality from all causes (Imaizumi et al., 2004). These observations result in a lack of consensus regarding whether or not to treat these patients because of limited evidence justifying therapy (Col et al., 2004).
Dehydroepiandrosterone (DHEA) is a steroid molecule synthesized in large quantities, (15-30 mg/day) mainly by the adrenal cortex and found in relatively (compared with other plasma steroids) high concentrations in the blood of humans and a few non-human primates. Most DHEA circulates as the sulfate conjugate, DHEAS. Lower primates have relatively little circulating DHEA, and other mammals secrete almost none. Both DHEA and DHEAS are biotransformed into biologically active estrogens and androgens in tissues. It is thought that over 90% of estrogen in postmenopausal women and 30% of total androgens in men are derived from this peripheral conversion (Labrie et al., 1995). DHEA levels decline with age from puberty to senescence (2-4%/year) and do so more dramatically than any other circulating steroid measured with the exception of estrogen during menopause (Orentreich et al., 1992). These observations give rise to speculation that DHEA secretion in large amounts may be a unique adaptation contributing to the relative longevity of the human species, and that the age-related decline in DHEA contributes to senescence. However, there has been difficulty in trying to correlate endogenous human DHEA levels with age-related changes in body composition or function (including immune function, insulin sensitivity and cognitive function), generally weak or non-significant relationships have been identified (Straub et al., 1998; Berr et al., 1996; Maccario et al., 1999; Abbasi et al., 1998; Abassi et al., 1998; Kostka et al., 2000; Haffner et al., 1994; Phillips, 1996; Kalmijn et al., 1998; Moffat et al., 2000). DHEA treatment of older men and women has not been convincingly demonstrated to have significant clinical benefits, with the exception of a study by Morales et al. (1994) showing improvement in self-reported physical and psychological well-being in post-menopausal women. DHEA may be a reliable biomarker of aging specifically in men (Roth et al., 2002) and a potential predictor of mortality (Mazat et al., 2001). Hornsby (1997) makes the case that the most likely function for DHEA is as a precursor for conversion to potent androgens, which mediate adrenarchy, a sexual signaling mechanism occurring just before puberty in higher primates. The shrinkage of the DHEA-secreting tissue zone of the adrenal cortex and the decrease in DHEA levels could be considered as the post-pubertal involution of an organ whose function has been fulfilled and is no longer necessary, rather than as a phenomenon of aging or senescence.
Many studies in animals suggest that DHEA is a multifunctional hormone with many beneficial effects. The rodent studies are problematic in these animals have little or no detectable plasma DHEA, therefore these studies are considered pharmacological experiments, introducing a hormone into a naïve environment. There is little clinical use of DHEA with the exception in the case of pathologic adrenal failure. DHEA replacement is associated with improvements in these patients (Arlt et al., 1999).
One mechanism by which this hormone may be involved is through its effects on oxidative defense. However, DHEA has been demonstrated to behave both as an antioxidant and a pro-oxidant, depending on the conditions of the experiment (Mooradian, 1993; Gallo et al., 1999). Both in vitro and in vivo, DHEA has been shown to protect against lipid peroxidation, cell death and toxicities induced by H2O2, carbon tetrachloride, copper, and hyperglycemia (Gallo et al., 1999; Bastianetto et al., 1999; Whitcomb and Schwartz, 1985; Rom and Harkin, 1991; Brignardello et al., 2000; Boccuzzi et al., 1997; Aragno et al., 1993; 1994; 1997). In rats subjected to repeated immobilization stress, DHEA administration partly reversed stress-induced inhibition of body weight gain, increases in adrenal weight and glucocorticoid receptor levels, and decreased lipid peroxidation, suggesting that DHEA may act an anti-stress hormone by reducing free radical generation (Hu et al., 2000). DHEA treatment both in vitro (Mohan and Cleary, 1989; McIntosh et al., 1993) and in vivo (Mohan and Cleary, 1991) lowers respiratory rates of mitochondria isolated from rat adrenals, heart, kidneys, brain and brown adipose tissue, suggestive of potential oxidant antagonist activity. However, there are also studies suggesting that DHEA has pro-oxidant properties (Swiercynski et al., 1997; McIntosh et al., 1993; Goldfarb et al., 1994; 1996).
At present, there is not enough evidence of anti-aging activity of DHEA to recommend as a supplement (Allolio and Arlt, 2002). Overall, it is not clear whether DHEA plays a role in regulating life span or is merely a biomarker of aging.
The discussion regarding testosterone and estrogen will be limited. First, the data on testosterone and longevity is somewhat minimal while data on estrogens' potential involvement with longevity is quite the opposite. Plasma testosterone concentrations in men decline in an age-related manner, decreasing 1.6% per year (Feldman et al., 2002). The challenge for clinicians is the non-specific nature of symptoms associated with adult androgen deficiency and the risks and benefits of testosterone replacement (IOM report, 2003). Direct longevity effects of testosterone are limited mostly to castration studies showing that a modest but significant extension of life span is observed in rats while gonadectomy increases the life span of domestic dogs (Drori and Folman, 1976; Michal, 1999). An increase on life span via castration of human males is improbable although one study in institutionalized mentally retarded adults suggests otherwise (Nieschlag et al., 1993; Wilson and Roehrborn, 1999; Hamilton and Mestler, 1969).
There is a rich literature regarding both the effects of estrogen on factors affecting life span and estrogen replacement in postmenopausal women that is beyond the scope of this review. It is clear that in most species of mammals, females live longer than males. Several theories have been postulated to define the biological mechanism for gender differences in longevity. Perhaps the potential mechanism that has received the strongest support is that females produce fewer free radicals when compared to males and estrogens upregulate antioxidative defenses resulting in less oxidative stress and damage accumulation (Vina et al., 2005). In addition, longevity associated genes are increased following estrogen treatment (Vina et al., 2005). Epidemiological studies have shown that estrogen therapy is associated with numerous health benefits including a reduction in risk for cardiovascular disease, lower incidence of osteoporosis and associated bone fractures, decreased risk for neurodegenerative diseases, increased cognitive function and reduced risk of cataract (reviewed in Singh et al., 2006).
However, estrogen replacement remains controversial as a means to reduce age-related disease. The series of papers published based on data from the Women's Health Initiative, a very large clinical study evaluating the effects of hormone replacement on various physiological factors suggests that both estrogen and progesterone replacement have adverse effects in postmenopausal women and that caution should be taken in prescribing these hormones long term (Roussow et al., 2002; Cauley et al., 2003; Wassertheil-Smoller et al., 2003; multiple other reports are available). This particular clinical study is not with its' own controversy as steroid formulation, dosage, age of administration and other factors significantly affected the results obtained. Plus, promotion of estrogen as a supplement to increase health span would appeal to only one half of the population due to the feminizing effects of this hormone. Studies evaluating the role of estrogens in aging and age-related disease will certainly continue to be explored.
The sex steroids were long thought to be responsible for the differences in longevity between females and males, however, another possibility exists. In close proximity to the sex steroids lie the hormones that regulate synthesis and secretion of estrogen and testosterone, namely the gonadotropins. Luteinizing hormone (LH) and follicle stimulating hormone (FSH) act in concert to regulate sex steroid biosynthesis, secretion and gonadal function. As the sex steroids decline with aging (dramatically at menopause in females), both LH and FSH rise. Bowen and coworkers (2000) have put forth the postulate that the gonadotropins play a significant role in the incidence and progression of age-related diseases, specifically Alzheimer's disease. Thus, it has been suggested that the gonadotropins are key players in hormonal modulation of aging in all sexually reproductive organisms (Bowen and Atwood, 2003). These hormones would mechanistically act via antagonistic pleiotropy – promoting growth and development early in life to advance reproduction but becoming dysregulated and driving senescence once reproductive life span has ended. The theory of antagonistic pleiotropy was first put forth by Williams (1957) and further analyzed in a recent review by Leroi and coauthors (2005). Williams proposed that aging occurred because of a decline in the force of natural selection late in life, and that alleles with positive effects on fitness early in life also had deleterious effects late in life. In agreement with that proposal, the LH transgenic mouse exhibits pathologies in multiple diverse systems (Mann et al., 2003). Furthermore, a collection of elegant experiments provide additional support for a role of LH in the pathogenesis of Alzheimer's disease (Bowen et al., 204; Webber et al., 2006; 2007). Further exploration of these endocrine hormones and their relationship to aging, age-related disease and longevity is warranted.
The work of Everitt, together with years of calorie restriction studies, provided key evidence that the endocrine system was involved in aging and longevity. No one knew at the time that this system would contribute so widely to the various physiological processes involved in aging. This research has been recently repeated in mice where 15-21% extensions in life span were achieved following hypophysectomy at one and nine months, respectively (Powers et al., 2006). An important outcome of the current study demonstrated that removal of pituitary hormones in normal adult mice still extends life span. Thus, these experiments indicate that the maturational abnormalities in long living, hormone deficient dwarf animals is not prerequisite for long life.
There are multiple hormones produced and secreted by the pituitary gland. The natural age-related decline in plasma GH levels (14% per decade during adult human life) and the concomitant decrease in IGF-I that occurs in mammals is likely a protective mechanism to decrease metabolic activity and cellular division (Iranmanesh et al., 1991). High levels of either of these hormones throughout life could play a contributory role in pathological changes associated with aging such as increased oxidative damage and cancer. Enhanced insulin sensitivity is a physiological mechanism that appears to prevail in almost every mammalian example of longevity. Decreasing IGF-I and insulin signaling without a concomitant decline (or increase) in GH concentrations may provide a balance of hormones that would maintain muscle and bone mass yet enhance stress resistance and prevent many of the physical signs of aging. In addition, this type of treatment may result in decreased incidence of age-related diseases such as cancer, diabetes and cardiovascular disease. Reduced activity of the somatotropic axis may be key in the quest to slow or delay aging and promote life extension. It has been shown that polymorphic variants of the IGF-I receptor and PI3K genes affect plasma IGF-I levels and human longevity, suggesting that life span control is evolutionarily conserved (Bonafe et al., 2003).
Both GH and thyroid hormones have well documented effects on growth. In fact, hypothyroid rats have significantly reduced levels of pituitary GH as T4 is required for GH synthesis and secretion suggesting perhaps that the GH is more significant with regard to longevity of dwarf animals. In addition, the hypothyroidism cannot explain the remarkable longevity of Ames and Snell mice as the GHR/BP KO mice also live significantly longer than wild type animals in the presence of only a small reduction in thyroid hormone levels (Hauck et al., 2001). Moreover, T4 replacement in subclinical hypothyroid aged individuals is a therapy that remains controversial.
Hormones are also important modulators and mediators of oxidative stress. Hormonally active molecules function both in protecting cells against oxidative stress and in generating oxygen free radical damage or sensitizing cells to such damage. These actions may occur via hormone receptor interactions, but are, in some cases, due to redox characteristics of the molecules themselves and are mediated via non-receptor dependent mechanisms.
One commonality shared by many of these hormones is that many are known to affect mitochondrial transcription including estrogen, insulin, T4, ACTH and others (Berdanier et al., 2006). Therefore, many of these endocrine components likely have direct effects on mitochondrial function and efficiency, certainly affecting free radical generation and likely oxidative damage. Alterations in hormone levels between individuals could then result in differences in rates of damage accumulation and eventually life span.
Are hormones regulating longevity in mammals? The answer to this question is probably, yes, but the degree of influence is hormone- and species-specific. Further exploration of the endocrine system will likely uncover new pathways and mechanisms that lead to potential therapeutic interventions to delay aging, treat aging-related disorders and extend life span in humans.
The author apologizes to those whose work could not be cited because of space limitations or inadvertent omission. The research from our laboratory reported in this review was supported by the NIA.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.