Abstract
The Drosophila melanogaster gene insulin-like receptor (InR) is homologous to mammalian insulin receptors as well as to Caenorhabditis elegans daf-2, a signal transducer regulating worm dauer formation and adult longevity. We describe a heteroallelic, hypomorphic genotype of mutantInR, which yields dwarf females with up to an 85% extension of adult longevity and dwarf males with reduced late age-specific mortality. Treatment of the long-lived InR dwarfs with a juvenile hormone analog restores life expectancy toward that of wild-type controls. We conclude that juvenile hormone deficiency, which results from InR signal pathway mutation, is sufficient to extend life-span, and that in flies, insulin-like ligands nonautonomously mediate aging through retardation of growth or activation of specific endocrine tissue.
Molecular similarity between fly InR and worm daf-2 suggests that mutants ofInR in flies should affect adult life-span, as do mutants ofdaf-2 in worms. InR and daf-2 are members of the insulin receptor family with homology to mammalian insulin and insulin-like growth fac tor–1 (IGF-1) receptors (1, 2). Worms carrying temperature-sensitive mutations indaf-2 form dauers at high temperature, but at lower temperature develop directly into adults with extended longevity and resistance to exogenous stress (3, 4). Genotypes homozygous for mutant InR have been reported to be lethal (5, 6), but several heteroallelic combinations ofInR alleles produce viable, dwarf adults that are slow to develop: InREC34 /InRE19 and InRGC25 /InRE19 [reported in (5)], andInRE19 /InRp5545 (Table 1). In addition,InRE19 /InRE19 was found to be viable and dwarf once crossed into a new isogenic background (7). Dwarf females eclose with extremely immature ovaries, and the egg chambers of young adults remain previtellogenic (Fig. 1, A through C).
Ovaries and ovarioles of wild-type andInR dwarf flies, stained with rhodamine-phalloidine and oligreen. Bars in (A) through (D), 50 μm. (A) Wild-type ovariole prepared 4 hours after eclosion. Germarium and stage 1-3 egg chambers are present. (B)InRE19 /InRE19 ovary 24 hours after eclosion. All ovarioles are immature. (C)InRE19 /InRE19 ovariole 11 days after eclosion. The degree of egg chamber maturation resembles that of newly eclosed wild-type flies [compare to (A)]. (D) A representative ovariole of dwarfInRE19 /InRE19 11 days after a single topical application of methoprene (1 μg) at 1 day after eclosion. The gradient of maturation is typical for normal, virgin females: a mature stage 14 egg is silhouetted on the right (arrow), while previtellogenic stage 1-6 egg chambers are distal [stages described in (35)]. (E) Effect of methoprene treatment on vitellogenesis in InR mutant females. Single applications of methoprene in 0.1 μl acetone solvent were made upon the abdomen of anesthetized flies at 1 day after eclosion (control, acetone only). Females were dissected at 11 days after eclosion (10 to 22 flies per group) and scored as vitellogenic if any ovariole was at or beyond stage 8. For continuous exposure, methoprene was volatilized in glass culture bottles for 10 days (four applications of 0.01 μg methoprene per bottle).
Phenotypes of InR dwarf and nondwarf genotypes relative to the coisogeneic wild-type control (mean ± SEM).
Measurement of INR kinase activity (8) indicated that theInRp5545 and InRE19 alleles both confer loss of INR function (Fig. 2). Basal activity of heterozygotes +/InRE19 and+/InRp5545 was 45% of that of the wild type. Insulin stimulation increased kinase activity of INR from +/+ and +/InRE19 flies by 60%, but only by 26% from +/InRp5545 flies (P < 0.05, for insulin stimulation of+/InRE19 versus+/InRp5545 ). Basal kinase activity of INR from InRE19/InRE19 and fromInRp5545/InRE19 flies was 14 and 11% of that of the wild type, respectively; neither were stimulated by insulin. InRp5545 is a P-element insertion in exon-1 (5, 6); the molecular lesion ofInRE19 has yet to be identified, but it does not appear to occur in the known coding region of the gene (9).
Kinase activity of INR from membranes prepared from adult heads of defined genotypes. Membranes were solubilized and INR autophosphorylated in the absence or presence of insulin (1 μM), and immunoprecipitated with Ab dp1040. Each mean ± SEM represents data from three to six independent experiments for each genotype, each performed in triplicate.
Life tables (10) of InR mutant adults were compared to concurrent cohorts of a wild-type coisogenic strain (Fig. 3). Dwarfs ofInREC34 /InRE19 andInRGC25 /InRE19 are short-lived. DwarfInRE19 /InRE19 and nondwarf +/InRp5545 have moderately reduced survival; nondwarf +/InRE19 individuals are normal. In contrast, females ofInRp5545/InRE19 are 85% longer lived than wild-type controls and overall present reduced age-specific mortality. As described in the accompanying paper (11), the life-span of female D. melanogaster is also extended by mutation of the insulin receptor substrate homolog chico. Survivorship among male InR genotypes follows the pattern observed for females. Relative to the wild type,InRp5545/InRE19 males exhibit high mortality as early adults, but because of reduced mortality at late ages, dwarf life expectancy at 10 days is 43% greater than that of controls. It is likely that not all InR alleles increase longevity because the gene is highly pleiotropic, with some alleles producing developmental defects that carry over to the adult stage, which counterbalance positive effects of the allele upon aging.
Survivorship and age-specific mortality ofInR mutant genotypes relative to the wild-type coisogenic control. Survivorship calculated asN x/N 0, whereN x is the number alive beginning each census interval and N 0 is the initial cohort size. Age-specific mortality is smoothed first with a 3-day running average and presented as logarithm of hazard, ln(μx) ≈ ln{–[ln(1 – d x/N x)]}, where d x is the number of deaths. Pairwise comparisons among genotypes for age-specific survival (1 – μx) by log-rank test with significance corrected for multiple tests. (A and B) Among females, relative to the wild type, the age-specific survival ofInREC34 /InRE19 ,InRGC25 /InRE19 ,InRE19 /InRE19 , and+/InRp5545 is reduced (smallest χ2 = 5.22, P < 0.02), survival of+/InRE19 does not differ (χ2 = 0.79, P = 0.37), and survival of InRp5545 /InRE19 is increased (χ2 = 175.1, P < 0.0001). (C and D) Among males, relative to the wild type, age-specific survival ofInREC34 /InRE19 ,InRGC25 /InRE19 ,+/InRp5545 , andInRE19 /InRE19 is reduced (smallest χ2 = 29.8, P < 0.001), survival of +/InRE19 does not differ (χ2 = 4.89, P = 0.03, nonsignificant under multiple tests), and survival ofInRp5545 /InRE19 is increased (χ2 = 21.4, P < 0.0001).
The fact that InR mutants are nonvitellogenic suggests a plausible mechanism for the extended longevity ofInRp5545/InRE19 flies.Drosophila overwinter as adults in a reproductive diapause where egg development is arrested at previtellogenic stages (12,13). In many insects, including Drosophila, reproductive diapause is proximally controlled through down-regulation of juvenile hormone (JH) synthesis by the corpora allata (CA) (14, 15). Ovaries of InR dwarf females morphologically resemble ovaries of diapause wild-type flies, and exogenous application of the JH analog methoprene to dwarf females initiated vitellogenesis (Fig. 1, C and D). Females ofInRE19 /InRE19 responded to a single treatment of methoprene in a dose-dependent manner, but females of InRp5545/InRE19 required continuous exposure to hormone to induce any vitellogenesis (Fig. 1E). Direct assay of adult JH synthesis (16) verified that CA activity was reduced in InR dwarfs to about 23% of the wild-type level (Fig. 4). Because reduced JH synthesis is seen inInRE19 /InRE19 flies, which exhibit normal life-span, as well as in long-livedInRp5545/InRE19 flies, the simple lack of JH may not be enough to extend longevity.
Synthesis rates of juvenile hormone (JH) by incubated corpora allata isolated from 2-day-old wild-type and InR mutant females (mean ± SEM, replicate assays). JH is calculated from sum of counts across fractions identified as JHB3, JHIII, and methyl farsinate. Significant differences occur only between dwarfs and nondwarfs [F(4,14) = 6.57, P < 0.005].
Loss of corpora allata JH accounts for dwarf infertility. Mutation of InR may increase longevity because infertility reduces allocation of metabolic resources to reproduction and frees resources for somatic maintenance (17) or because reduced JH in mutant flies induces specific physiological mechanisms of somatic persistence normally expressed during adult reproductive diapause. Adult D. melanogaster in reproductive diapause age at negligible rates and are stress resistant; these traits are reversed by treatment with methoprene (18). Extended survival is characteristic of adult reproductive diapause in acridid grasshoppers and in the monarch butterfly, and surgical ablation of the corpora allata to eliminate adult JH synthesis induces both diapause and increased longevity [reviewed in (19)]. Consistent with the notion that reduced JH synthesis can directly extend life-span,InR dwarf flies show somatic physiological changes (20) (Table 1): (i) triglycerides are elevated fourfold (F = 32.2, P < 0.001), as observed in diapause D. triauraria(21) and in dwarf D. melanogaster mutant forchico (22), and (ii) Cu/Zn-superoxide dismutase concentration is increased twofold (F = 9.42,P < 0.02), as is characteristic of long-lived mutants of Caenorhabditis elegans (23). Measured in young adults, no difference in mass-specific metabolic rate was detected (Table 1). We suggest that infertility need not be the direct cause of slowed aging in InR mutants; JH may simply control both fertility and life-span.
To test directly whether JH modulates survival inInRp5545/InRE19 female dwarfs, we investigated whether treatment with methoprene restores wild-type longevity to these mutants, even if it does not fully restore fertility (24). In concurrent trials of dwarf and wild-type flies, survival of methoprene-treatedInRp5545/InRE19 females was reduced toward the level observed in coisogenic controls (Fig. 5). This rescue is physiological rather than toxicological because, in wild-type controls, methoprene produced no significant change relative to ethanol-treated flies.
Survival of female dwarf and coisogenic wild-type cohorts treated with methoprene in ethanol, ethanol (EtOH) only, or no solution. Methoprene per application: (A) 0.02 μg, (B) 0.002 μg. Mortality of methoprene-treatedInRp5545 /InRE19 is at least 90% greater [in (A)] and 72% greater [in (B)] than either ethanol treatment alone or untreated dwarfs [Cox-proportional hazard, (A) and (B), respectively: exp(coef.) = 1.898,P < 0.0001; exp(coef.) = 1.721, P< 0.0001]. Mortality of methoprene-treated wild type does not differ significantly in (A) from its ethanol-treated control (χ2 = 0.34, P = 0.56), or in (B) from its ethanol-treated or untreated controls (χ2 = 2.16, P = 0.34).
The InR pathway may alter endocrine function in two ways. Adult CA is derived from neurosecretory tissue of the larval ring gland. Adult dwarf CA may be immature upon metamorphosis as a result of cell autonomous effects of InR upon the development of neuroendocrine cells. A second way InR may alter endocrine function is that JH secretion by CA may be impaired by reduced neuropeptide transmission in the adult brain (25), due to a reduction of INR function in brain areas where it is normally expressed (26).
In C. elegans, the insulin/IGF-1 pathway influences dauer formation, fertility, and aging in part through nonautonomous, secondary signaling; sterility is not required for extended longevity in C. elegans because some long-lived daf-2 are fully fertile (27). For Drosophila, we have shown that InR affects neurosecretory tissue specialized for secretion of juvenile hormone. Therefore, mutations in the insulin signaling pathway in flies autonomously affect cell proliferation, growth, and body size (5, 22), but nonautonomously affect diapause, reproduction, and life-span through effects upon specific neuroendocrine cells. Deficiency in a juvenoid-like hormone signal in worms and in flies may extend longevity because its absence leads to the inappropriate expression of parallel physiological programs normally reserved for dauer or diapause.
This invertebrate model may have parallels with mammalian aging. Ames and Snell mice are mutant for the genes Prop-1 orPit-1, respectively, and are defective for pituitary development [reviewed in (28)]. Consequently, they are deficient in growth hormone, prolactin, and thyroid-stimulating hormone, leading to hypothyroidism and presumably reduced synthesis of thyroxin, a retinoid hormone with potential functional similarity to JH (29). These mice are phenotypically dwarf, mildly obese, and long-lived (28). A remarkably similar phenotype is observed in mice lacking insulin receptor function in the central nervous system or those lacking the chico homolog, IRS-2, in all tissues: increased fat mass and infertility with accompanying neuroendocrine deficiency (30, 31). Although effects on life-span in these mice remain to be determined, the concordance of phenotypes suggests that insulin signaling may be central to a common mechanism that exists across taxa for the neuroendocrine regulation of metabolism and the reproductive state, and their associated consequences upon aging.
↵* To whom correspondence should be addressed. E-mail: Marc_Tatar{at}Brown.edu
