Special Reviews

The Endocrine Regulation of Aging by Insulin-like Signals

See allHide authors and affiliations

Science  28 Feb 2003:
Vol. 299, Issue 5611, pp. 1346-1351
DOI: 10.1126/science.1081447


Reduced signaling of insulin-like peptides increases the life-span of nematodes, flies, and rodents. In the nematode and the fly, secondary hormones downstream of insulin-like signaling appear to regulate aging. In mammals, the order in which the hormones act is unresolved because insulin, insulin-like growth factor–1, growth hormone, and thyroid hormones are interdependent. In all species examined to date, endocrine manipulations can slow aging without concurrent costs in reproduction, but with inevitable increases in stress resistance. Despite the similarities among mammals and invertebrates in insulin-like peptides and their signal cascade, more research is needed to determine whether these signals control aging in the same way in all the species by the same mechanism.

Dozens of genes extend adult longevity. Remarkably, many of these genes are involved with hormonal signals, and both these genes and their endocrine systems are conserved among eukaryotes. Thus, insulin-like peptides, insulin-like growth factor (IGF), lipophilic signaling molecules, and sterols are all candidate effectors of aging in organisms as diverse as the nematode Caenorhabditis elegans, the flyDrosophila melanogaster, and the mouse Mus musculus. Suppression of these hormones or their receptors can increase life-span and delay age-dependent functional decline. This regulation is likely to be adaptive because, at least among invertebrates, these hormones regulate the organism's capacity to survive during states of reduced metabolism coupled with high stress resistance and arrested development. Mutations that increase life-span through hormones are thought to initiate elements of this survival program independent of the appropriate environmental cues. Because mechanisms for survival often oppose the progress of aging, they can illuminate the cellular and molecular causes of senescence. The insulin/IGF system and its associated downstream hormones are likely to prove particularly instructive.

Genetics of C. elegans Dauer and Aging

The genetic dissection of aging rapidly advanced when strains of C. elegans with mutations in the dauer formation (Daf) pathway were found to have unusually long lives (1). Dauer diapause is a nonfeeding, stress-resistant larval state evolved for endurance and dispersal under adverse conditions. Animals with weak alleles of Daf-constitutive mutants in the genes age-1 and daf-2(2) can bypass dauer and become prodigiously long-lived adults in a manner dependent on the gene daf-16(1, 3, 4) (Fig. 1A). The products of these genes and others revealed insulin/IGF-like signal transduction as a central regulator of dauer and aging (5). The C. elegansgenome contains 37 insulin-like ligands that are mainly expressed in neurons, but are also found in intestine, muscle, epidermis, and gonad (5, 6). Ligand bound to DAF-2 insulin-like receptor signals through a kinase cascade to phosphorylate forkhead transcription factor DAF-16, which is thus sequestered in the cytoplasm. In this state, adults directly reach reproductive maturity and age rapidly. Without activation of the pathway, DAF-16 promotes transcription and induces diapause and exceptional longevity (Fig. 1A).

Figure 1

Models for endocrine circuits of aging regulation. Environmental cues and diet are integrated by the central nervous system (CNS), which activates the reproductive, pro-aging mode through endocrine signals. In addition, hormones provide feedback and information on physiological state to the CNS/neuroendocrine centers. In each model, pathways of activated reproductive mode are shown with solid lines; broken lines represent down-regulated pathways; arrows denote agonists, and bars denote antagonists. Elements of the insulin/IGF signal transduction pathway are represented in red; other hormones and neuropeptides are represented in blue. “?” denotes hypothetical or speculative elements. (A) Nematode. Environmental cues influence production of insulin-like peptides from insulin-producing cells (IPC), mainly sensory neurons. DAF-2/Insulin receptor responds in target tissues, including neuroectoderm, and produces putative secondary endocrines that promote the reproductive mode. In the gonadal pathway, signals from the germ line cells (GC) inhibit production of a life-extending steroid produced by DAF-9, potentially synthesized in somatic gonad (SG) tissue. Secondary endocrines could inhibit life-extending activities of DAF-16 and DAF-12 nuclear hormone receptor in multiple target or endocrine tissues. (B) Drosophila/Insects. Cells of the pars intercebralis produce insulin-like peptides. In direct or indirect response to insulin signaling, the corpora allata produces juvenile hormone (JH). JH promotes vitellogenesis and suppresses response to external stress (21). Insulin-like signals acting at the gonad stimulate egg development and 20-hydroxy-ecdysone (20HE) synthesis. Insulin-like peptides also act upon the fat body, a somatic tissue with intermediary metabolism and nutrient-storage functions. (C) Rodent. Food intake alters neuronal input to the hypothalamus to promote release of gonadotropins [luteinizing hormone (LH) and follicle-stimulating hormone (FSH)], thyroid-stimulating hormone (TSH), and growth hormone (GH)] from the pituitary. These hormones cause increased output of steroid sex hormones, thyroid hormones, and insulin-like growth factor 1 (IGF-1). Release of insulin from the pancreatic beta cells is increased by food intake. Combined actions of IGF-1, insulin, gonadal, and thyroid hormones promote growth and reproduction at the expense of anti-aging mechanisms.

DAF-16 also is a key regulator of heat and oxidative stress resistance, fat storage, developmental arrest, fertility, and metabolism (5, 7, 8). DAF-16 becomes localized to the nucleus in response to various stresses; this response is attenuated by insulin/IGF signaling (9). Elevated stress resistance combined with down-regulated central metabolism and reproduction may be coordinated physiological states associated with slow aging (Fig. 2). Down-regulation ofdaf-2 specifically during adulthood extends longevity, and the impact of daf-2 signaling upon adult life-span and stress resistance is independent of larval dauer and of adult reproduction (10). Thus, survival and reproductive traits coordinated by insulin signaling can be decoupled to postpone aging.

Figure 2

Alternative life-history modes (73) coordinated by hormones. The reproductive mode is activated when conditions are favorable to reproduce. Intermediary and central metabolism are high to support the energetic demand, and stress response mechanisms are dampened because they interfere with reproduction. Consequently, the reproductive mode is pro-aging. The alternative adaptive mode is quiescence, corresponding to dauer, diapause, and perhaps hibernation. In quiescence, the systems of reproduction, together with hormones to promote an enduring, resistant, slow-aging somatic state, coordinately alter metabolism and stress responsiveness. The organism increases somatic survival capacity to withstand poor environments until conditions favor reproduction.

Because functional DAF-16 is usually required to extend longevity, its transcriptional profile is critical to understand how the insulin/IGF system modulates aging. In vertebrates, DAF-16 homologs up-regulate genes that promote DNA-damage repair (for example, GADD45a) and oxidative protection (MnSOD), and down-regulate genes promoting cell cycle progression (D type cyclins) (11, 12).

In C. elegans, insulin/IGF as well as transforming growth factor–β (TGF-β) and serotonin signaling converge on the orphan nuclear hormone receptor DAF-12 (13). In dauer formation, daf-12 functions distal to these primary signals. In aging, daf-12 has complex modulator effects;daf-12–null mutants enhance the longevity of somedaf-2 alleles, but weakly suppress that of others (4, 8). The molecular basis of these interactions may arise from mutual interaction of the two signaling pathways at protein-protein or transcriptional levels.

Endocrine Analysis of C. elegans Aging

Hormonal signals downstream of daf-2 may regulate dauer and aging. The loss of a daf-2+ extrachromosomal duplication from neuroectoderm of animals with adaf-2 mutant background extends adult life-span (14). Similarly, ectopic expression of adaf-2+ transgene in the nervous system of adaf-2 mutant can bypass diapause and restore normal adult longevity (15). These data imply that dauer and life-span are controlled by the expression of daf-2 in only a few cell types, demonstrating that a secondary hormone must act at a distance to regulate aging. Given the placement of daf-12 downstream of insulin signaling in the regulation of dauer, one simple model posits that this secondary hormone is a ligand for the DAF-12 nuclear receptor (Fig. 1A).

DAF-9, a cytochrome P450 related to steroidogenic and fatty acid hydroxylases, provides an important link in this endocrine circuit (Fig. 1A) (16, 17). Nutritional signals acting through insulin/IGF and TGF-β appear to stimulate production of a DAF-12 ligand by DAF-9, which promotes reproduction and permits rapid senescence. In contrast, under adverse environments,daf-9 is down-regulated and dauer is induced.daf-9 is expressed in a limited set of tissues—two neurons, hypodermis, and spermatheca—consistent with an endocrine function. Some mutants of daf-9 produce adults with life-spans at least 25% longer than that of the wild type (16, 17). However, becausedaf-9 mutants further enhance daf-2 longevity and act strictly through daf-12, DAF-9 likely works in parallel to insulin/IGF for life-span regulation.

Endocrine influences on C. elegans aging are demonstrated by the fact that ablation of germ line–derived cells results in a 60% life-span increase (18). The pro-aging, reproductive state (Fig. 2) maintained by insulin/IGF and nuclear hormone receptors likely involves positive feedback from the gonad because longevity induced by germ line ablation requires functionaldaf-16, daf-9, and daf-12(16, 18). Conceivably, the germ line induces hormones that sustain the reproductive state, which either directly or indirectly accelerate aging. Removal of both the germ line and the somatic gonad was seen to have no net effect on life-span, suggesting that germ line and somatic gonad act antagonistically to inhibit and promote longevity (Fig. 1A).

Primary and Secondary Hormones of Insect Aging

Thirty years ago the sesquiterpenoid juvenile hormone (JH), best known for its role in larval development, was shown to influence insect adult longevity (19). In adult insects, JH promotes vitellogenesis (20) and inhibits adult diapause, an enduring, somatic state with arrested reproductive development (Fig. 2). Diapause is induced in a variety of insects when the source of JH is removed; under these conditions adult longevity is increased more than 100% (19). Drosophila also overwinter as diapause adults, and in this state stress resistance is increased and the rate of aging is retarded (21). Mutation of the gene encoding insulin-like receptor (InR) (22) and of the gene encoding insulin-receptor substrate (chico) (23) of D. melanogaster increases adult survival (24–26). Extended life-span by these mutants is likely to be secondary to JH deficiency because JH synthesis is negligible in InR dwarfs and treatment with an analog of this hormone initiates vitellogenesis and restores normal longevity.

JH may be the secondary pro-aging signal downstream of insulin/IGF (Fig. 1B). The pars intercebralis are the major insulin-producing cells of adult D. melanogaster; insulin-like protein is released into the protocerebrum, at the corpora cardiaca, and into the hemolymph (27). Because insect JH synthesis is controlled by neuropeptides (28), insulin/IGF may modulate neuropeptidogenic tissues, or the insulin-like peptide may act directly to stimulate JH synthesis. Through any of these channels, inhibition of insulin ligand or receptor could reduce JH synthesis.

The steroid hormone ecdysone is a second candidate regulator of insect aging (Fig. 1B). Ecdysone, like JH, is best known for its influence on development and on yolk production in the adult (29). In the adult, the active form of ecdysone, 20-OH-ecdysone, is made largely in egg follicle cells. Transported to the fat body, 20-OH-ecdysone binds to the ecdysone receptor (EcR), which forms a heterodimeric receptor complex with the retinoid X receptor Ultraspiracle (USP). USP may use JH as ligand (30), and both the molecular chaperones Hsc70 and Hsp90 and the histone deacetylases Sin3A/Rpd3 are thought to interact with EcR/USP (31, 32). Overexpression of chaperones extends longevity in D. melanogaster and inC. elegans (33, 34), and mutation of the gene encoding Rpd3 increases the life-span of yeast and of D. melanogaster (35, 36). In mosquito, insulin acts directly at the ovary, where it regulates ecdysone synthesis (37). In D. melanogaster, insulin regulates the rate of stem cell proliferation (38), and ovaries of InR-mutant adults produce little ecdysone (39). Ecdysone also can inhibit JH synthesis (40). Ecdysone, therefore, may be a gonad-derived signal through which insulin can affect insect aging, and recent data show that D. melanogaster deficient for ecdysone synthesis or with mutants of EcR have increased longevity accompanied by stress resistance and surprisingly normal fecundity (41).

Pituitary Endocrine Deficiency in Mammals

Of the half-dozen genetic models that retard murine aging, four involve deficiency of pituitary endocrine action. The mutationsProp1 df (42) andPit1 dw impede pituitary production of growth hormone (GH), thyroid stimulating hormone (TSH), and prolactin; reduce growth rate and adult body size; and increase adult life-span by 40 to 60% (43, 44). Small adults with similar improvement in longevity are also produced by a knockout of growth hormone receptor (GHR-KO) (45). Expressed throughout life, these mutations produce many secondary alterations in endocrine systems. Without GH, the synthesis of circulating IGF-1 is suppressed, as is plasma insulin as a result of enhanced sensitivity in the liver combined with altered pancreatic islet development (46). Thyroid function is reduced inProp1 df and Pit1 dwmutants deficient for TSH (45); GHR-KO mice are mildly hypothyroid, presumably as a result of impaired development (47). The challenge is to identify which of these hormones regulate aging and at which stage of life.

In invertebrates, reduced insulin/IGF signaling increases longevity, but it remains unclear whether or how reduction in GH and IGF-1 directly affects aging in rodents (Fig. 1C). In addition to its impact on IGF-1, GH influences somatic metabolism—for instance, by inducing adipocyte lipolysis. IGF-1 itself may affect aging in both beneficial and detrimental ways. In rodents as in humans, levels of GH and IGF-1 decline with adult age. Short-term GH supplementation in aged adults restores some aspects of body composition and cognition (48, 49). Thus, the withdrawal of GH and IGF-1 has been suggested to be a cause of senescence rather than a condition that retards aging. On the other hand, because it stimulates metabolism and cell growth, GH may hasten tissue pathology. Indeed, chronic treatment of GH-deficient dwarf rats with GH increased tumor incidence in response to a carcinogen (50). High IGF-1 titers in young wild-type animals may produce a trade-off between current benefits to reproduction and later costs in senescence (51).

Powerful evidence for the direct role of IGF-1 signaling in the control of mammalian aging was provided by mice mutant for the IFG-1 receptorIgf1r (52): Igf1r +/−females, but not males, live 33% longer than wild-type controls. These mutants exhibit minimal reduction in growth with no alterations in the age of sexual maturation, fertility, metabolism, food intake, or temperature. Life extension is associated with increased tolerance of oxidative stress and reduced phosphorylation of Shc, a gene previously implicated in the control of longevity and stress resistance in mice (53).

Mouse longevity is also increased 18% by fat-specific disruption of the insulin receptor gene (54). These mice have normal caloric intake yet retain leanness and glucose tolerance with age. Multiple intriguing changes in adipocytes underlie these effects, including elevated plasma leptin relative to adipose tissue mass, reduced lipolysis, and polarization of adipocytes into populations with altered expression of fatty acid synthase (55). Thus, insulin at adipose tissue may affect aging through impacts on neural-targeted hormones as well as through regulation of intermediary metabolism.

Reduced thyroid hormone synthesis also is characteristic of pituitary mutants with retarded aging (Fig. 1C). Because of their low thyroid function, dwarf mice exhibit traits hypothesized to slow senescence, including hypothermia and delayed maturation. Hypothyroidism induced by treatment of neonatal rats modestly extends adult life-span (56) although reduced IGF-1 may still play a role because thyroid hormone stimulates GH during rodent development. Induced hypothyroidism in rats reduces free-radical leak from heart mitochondria, decreases free-radical production, and reduces oxidatively damaged DNA (57). Intriguingly, the target of thyroid hormone, thyroid receptor, is a member of the nuclear receptor family (58) that includes C. elegansDAF-12 and D. melanogaster EcR. The potential for thyroid hormones to affect aging remains to be thoroughly explored.

Reduced caloric intake increases rodent longevity, and similar effects may occur in nonhuman primates and humans (59). Caloric restriction (CR) may extend life-span because availability of “energetic substrate” is directly reduced at each cell, perhaps increasing the efficiency of central metabolism and lowering the rate of free-radical generation, or activating histone deacetylases that increase transcriptional silencing (60, 61). Alternatively, CR may indirectly mediate aging (62, 63). Signals from current nutritional state and body composition, such as leptin and resistin, are integrated centrally through the hypothalamus, which in turn modulates food intake, endocrine function, and metabolic homeostasis (64). Through hypothalamic patterning, long-term CR may adjust endocrine levels to sustain a somatic endurance program (Fig. 2). In this context, suppression of IGF-1 and insulin during CR may be functionally necessary.

The long-term interaction of CR and Prop1df was explicitly examined when mice with and without the mutation were fed a full or a restricted diet (65). Critically,Prop1df mice maintained under CR lived longer than mice under either treatment did alone, suggesting that diet restriction and pituitary deficiency may affect aging through independent mechanisms despite their similar hormonal profiles. Alternately, CR and the Prop1df mutation could still function through a common pathway, but under test conditions neither alone produce their fullest potential gain in life-span (66).

Insulin/IGF Function in Mammals and Invertebrates

The regulation of aging may be a conserved trait of ecdysozoa and tetrapods; the molecular components of the insulin signaling pathway are homologous, and reduced insulin/IGF activity is commonly associated with extended longevity. From this basis we may begin to identify the molecular and cellular targets of a putative aging regulator, targets that are likely to be fundamental survival mechanisms that evolved to resist age-dependent degeneration. To take this step, however, we must first consider whether the insulin/IGF system has evolved in different genomic environments within the histories of these phyla (67).

The genomes of D. melanogaster and C. eleganseach contain a single locus encoding an insulin/IGF-like receptor (InR or daf-2) and many loci that encode insulin-like ligands (Fig. 3). Seven ligand loci occur in theD. melanogaster genome (68), while at last countC. elegans contains 37 insulin-like genes (6); many of these genes occur within clusters suggestive of recent duplication events. An impressive number of phenotypes are regulated through the single insulin/IGF receptor of fly and worm, including growth, nutrient metabolism, diapause and dauer, reproduction, and aging. Diversification of function despite a single receptor may have evolved from the duplication of ligand loci, which multiplies both the coding and the regulatory regions. Regulatory duplication, in particular, permits function to evolve by altering the tissue and temporal specificity of expression, even when the protein is conserved (69).

Figure 3

Genomic distribution of genes for insulin-like peptides and insulin-like receptors in model organisms; map positions on numbered chromosomes. (A) D. melanogaster: Drosophila insulin-like peptides 1 to 7 (dilp); insulin-like receptor (InR). (B) C. elegans: Insulin-like peptides 1 to 37 (ins); dauer formation-2 (daf-2). (C) Mouse: Insulin peptides 1 and 2 (Ins); insulin-like peptide 3 and 5 ; insulin growth factor 1 and 2 (Igf); insulin receptor (Insr); insulin growth factor receptor (Igfr); insulin-related receptor (Insrr).

In contrast, the genomes of tetrapods contain four candidate receptors—IR, IGF-1R, IGF-2R, and perhaps IRR—each of which likely responds to one primary ligand (70). If diversity in mammalian receptors arose from the tetraploidization of the primitive deuterostome genome (71), there would be four loci for the ligand, receptor, and all elements of their signal transduction. With duplication of entire pathways, the function of the single ancestral insulin/IGF system could evolve independent regulatory outputs with less pleiotropic constraint. In addition, rather than being limited to a specific tissue or ontogenetic stage, ligands could be systemically released without ensuing confusion of physiological signals. The evolution of IGF binding proteins in vertebrates also may have contributed to specialization of circulating insulin-like ligands (72). In these ways, insulin could have evolved primarily for regulation of nutrient metabolism, and IGF-1 for mediation of growth.

Aging-related targets of insulin/IGF may be similar in the ecdysozoa and mammals, but this functional orthology could be obscured by the asymmetric phyletic evolution of the gene family. The genomic structure of fly and worm might have favored localized insulin/IGF action—for instance, at neuroendocrine cells—and readily evolved secondary endocrine regulation of aging. The genomic structure in tetrapods, on the other hand, could have favored broad systemic action of insulin and IGF peptides and thus permitted evolution of cell-autonomous aging regulation if one receptor type and its coordinated transduction pathway specifically controlled systems of somatic survival. Research attuned to the evolutionary history of the insulin/IGF system as well as to its molecular operation is required to resolve these alternatives and to find the aging-related targets regulated by these signals.


View Abstract

Stay Connected to Science

Navigate This Article