Extension of Life-Span by Loss of CHICO, a Drosophila Insulin Receptor Substrate Protein

See allHide authors and affiliations

Science  06 Apr 2001:
Vol. 292, Issue 5514, pp. 104-106
DOI: 10.1126/science.1057991


The Drosophila melanogaster gene chicoencodes an insulin receptor substrate that functions in an insulin/insulin-like growth factor (IGF) signaling pathway. In the nematode Caenorhabditis elegans, insulin/IGF signaling regulates adult longevity. We found that mutation of chicoextends fruit fly median life-span by up to 48% in homozygotes and 36% in heterozygotes. Extension of life-span was not a result of impaired oogenesis in chico females, nor was it consistently correlated with increased stress resistance. The dwarf phenotype ofchico homozygotes was also unnecessary for extension of life-span. The role of insulin/IGF signaling in regulating animal aging is therefore evolutionarily conserved.

Mutations that extend life-span illuminate the molecular mechanisms underlying aging and longevity. InCaenorhabditis elegans, mutation of the genesdaf-2 and age-1, which encode components of an insulin/IGF signaling (IIS) pathway, enhances stress resistance and increases adult life-span by up to 200% (1). This pathway also controls the formation of dauer larvae, which are developmentally arrested, stress resistant, long-lived, and produced in response to crowding and reduced food (2). Potentially, insulin/IGF mutants could be long-lived by virtue of expression of dauer longevity in the adult, in which case the extension of adult life-span by these mutations could be a peculiarity of C. elegans. We examined whether the role of IIS in aging has been evolutionarily conserved and therefore might also operate in humans.

In the fruit fly Drosophila melanogaster, the insulin/IGF receptor INR, the insulin receptor substrate CHICO, the phosphatidylinositol 3-kinase (PI3K) Dp110/p60, and the PI3K target protein kinase B (PKB, also known as DAkt1) form a signaling pathway that regulates growth and size (3–7). We examined the effects on aging of hypomorphic mutations in Inr(equivalent to daf-2) and PKB, and null mutations in chico and the catalytic (Dp110, equivalent toage-1) and adapter (p60) PI3K subunits (8). All mutants were tested as heterozygotes. We also tested chico1 (3) andPKB3 (9) homozygotes andInrGC25 /InrE19 transheterozygotes, which form viable dwarf adults. The remaining mutations were homozygous lethal.

Most mutants tested had normal or significantly decreased life-span (10). For example, PKB3 homozygotes andInrGC25 /InrE19 flies were short-lived. By contrast, chico1 extended life-span (Fig. 1). Homozygouschico1 females exhibited an increase of median and maximum life-span of up to 48 and 41%, respectively.chico1 heterozygotes also exhibited increases in median life-span of up to 36 and 13% in females and males, respectively. Homozygous males, however, were slightly short-lived.

Figure 1

Effect of chico1 on life-span. Median life-spans and sample sizes, respectively, for females are as follows. (A) +/+, 53 days andN = 119; +/chico1 , 56 days andN = 128. (B) +/+, 44 days andN = 105; +/chico1 , 60 days andN = 100; chico1 , 65 days andN = 86. (C) +/+, 49 days andN = 61; chico1 , 65 days andN = 67. (D) +/+, 46 days andN = 172; +/chico1 , 54 days andN = 171; chico1 , 58 days andN = 176. Data for males are as follows. (E) +/+, 54 days and N = 136; +/chico1 , 56 days and N =135. (F) +/+, 57 days and N = 104; +/chico1 , 62 days and N = 140. (G) +/+, 59 days and N = 54;chico1 , 50 days and N = 48. Statistical comparisons (log rank test) are as follows for females: +/chico1 versus +/+, P < 0.0001 (A, B, and D); chico1 versus +/chico1 , P = 0.0008 and <0.0001 (B and D, respectively); and chico1 versus +/+, P < 0.0001 (B, C, and D). Statistical comparisons (log rank test) are as follows for males: +/chico1 versus +/+, P < 0.0001 and < 0.0866 (not significant) (E and F, respectively). In all panels, circles representchico1 /chico1 ; squares represent +/chico1 , and triangles represent +/+.

To confirm that chico1 itself extended life-span, we tested the effect on life-span of pCSR4-chico,a P element containing chico(+). This construct fully rescues the dwarf phenotype of chico1 (3). chico1 was crossed to two stocks containing independent pCSR4-chico insertions (pCSR4-chico 1.1 and 2.3). As a control,chico1 was also crossed to the base stock in which the P element insertions were made. Progeny with either two copies (chico1 heterozygotes with onechico transgene) or one copy (chico1 heterozygotes alone) of chico(+) were compared (11). The rescue construct significantly reduced life-span relative to the +/chico1 control. The median female life-span of 54 days in +/chico1 was reduced to 46 days in +/chico1 ,+/pCSR4-chico 1.1 flies and 52 days in +/chico1 ,+/pCSR4-chico 2.3 flies (P = 0.0002 and 0.0243, respectively). Similar effects were observed in males (10). Thus, mutation ofchico itself increases life-span. Becausechico1 is a null allele, its effect on life-span indicates that the wild-type chico gene acts to accelerate aging.

Of the mutations tested, onlychico1 increased life-span. This may be because the effect of reduced IIS on life-span depends on the degree to which signaling is reduced. Unlike the other null mutations in IIS genes tested, chico1 is not homozygous lethal, presumably because the INR receptor can signal to PI3K directly, as well as indirectly via CHICO (3). Thus,chico1 mutants may be long-lived because of the relatively mild reduction in pathway activity that they bring about. Notably, severe IIS mutations in C. elegans can cause premature mortality in some adults, although the maximum life-span of populations is invariably increased (1). This is probably whyInrGC25 /InrE19 flies are short-lived: Demographic analysis indicates that a reduction in the age-specific mortality rate acceleration occurs, whose effect on survival is masked by an elevated rate of age-independent mortality (12). Furthermore, a different heteroallelicDrosophila Inr mutant to that tested here exhibits an 85% increase in female life-span (13). By contrast, in short-lived PKB3 populations, no reduction in mortality rate acceleration is seen (12). This raises the possibility that a second pathway downstream of chico might regulate aging in Drosophila. Interestingly, CHICO contains potential binding sites for the Drk/Grb2 docking protein, consistent with signaling via Ras/mitogen-activated protein kinase.

We next investigated whether extension of life-span bychico1 was mediated by processes previously shown to affect aging. A reduction in fecundity extends life-span inDrosophila females (14, 15);chico1 heterozygous females have reduced fecundity, and the homozygotes are almost sterile (3, 12). To test whether the increased life-span ofchico1 females was due to reduced fecundity, we examined the interactions between chico1 and the dominant, female-sterile mutant ovoD1 . This mutation blocks oogenesis at stage 4, before vitellogenesis commences (16), and extends female life-span (15). If chico1 extends female life-span by the exact same mechanism asovoD1 , then the three sterile genotypes (chico1 , +/ovoD1 , and +/ovoD1 +/chico1 ) should have similar life-spans and live longer than the subfertile chico1 heterozygotes. In fact, the chico homozygotes lived significantly longer than all other genotypes (Fig. 2). In addition, the partially fertile chico1 heterozygotes lived as long as the sterile flies that were heterozygous for both ovoD1 andchico1 and lived significantly longer than the sterile ovoD1 heterozygotes. The effect of chico1 on female life-span is therefore not a consequence of the same mechanism of reduced fecundity as is produced by ovoD1 . Ifchico1 does extend female life-span through an effect on reproductive effort, the interaction must occur through some process other than oogenesis (for instance yolk protein synthesis) or before stage 4 in oogenesis becauseovoD1 flies are blocked at that stage.

Figure 2

Effects of chico1 andovoD1 on female life-span. Median life-spans and sample sizes, respectively, are as follows: +/+, 44 days andN = 105; +/ovoD1 , 50 days andN = 90; +/chico1 , 60 days andN = 100; +/chico1 +/ovoD1 , 59 days and N = 93; andchico1 , 65 days and N = 86. Statistical comparisons (log rank test) are as follows: +/+ versus +/ovoD1 , P < 0.0001; +/ovoD1 versus +/chico1 +/ovoD1 , or +/chico1 , P < 0.0001; +/chico1 +/ovoD1 versus +/chico1 , no significant difference; andchico1 versus +/chico1 +/ovoD1 or +/chico1 ,P = 0.0008 in each case.

In C. elegans, long-lived IIS mutants are stress resistant and overexpress the antioxidant enzyme superoxide dismutase (SOD) (1). We examined the resistance ofchico1 flies to three stressors, but only one showed any correspondence with life-span (17) (Fig. 3). No resistance to heat stress (37°C) was seen (Fig. 3A). Slight resistance to oxidative stress (methyl viologen) was observed inchico1 heterozygotes but not in homozygotes (Fig. 3B). However, some correspondence between starvation resistance and life-span was seen (Fig. 3C). Increased SOD levels were seen inchico1 homozygotes but not in heterozygotes (17) (Fig. 4). Thus, modulation by IIS of longevity, and of SOD levels, has evidently been conserved betweenC. elegans and Drosophila. Furthermore, effects of this pathway on fertility are widespread (1, 3, 18). However, effects on stress resistance are not well conserved, nor do any of the above associated affects appear to be causal in extending life-span.

Figure 3

Effect ofchico1 on stress resistance. Histograms show mean survival time. Probability that values are identical to wild type is indicated as follows: *, 0.05 > P > 0.01; **,P < 0.01 (log rank). (A) Heat stress resistance. Sample sizes for females were as follows: 49 (+/+), 50 (+/chico1 ), and 51 (chico1 ) on day 1 and 49 (+/+), 50 (+/chico1 ), and 52 (chico1 ) on day 2. Sample sizes for males were as follows: 50 (+/+), 50 (+/chico1 ), and 50 (chico1 ) on day 1 and 52 (+/+), 51 (+/chico1 ), and 47 (chico1 ) on day 2. (B) Oxidative stress (methyl viologen) resistance. Sample sizes for females were 100 (+/+), 100 (+/chico1 ), and 99 (chico1 ). Sample sizes for males were 99 (+/+), 100 (+/chico1 ), and 69 (chico1 ). (C) Starvation resistance. Sample sizes for females were as follows: from 0 to 6 hours, 150 (+/+), 150 (+/chico1 ), and 150 (chico1 ); from 3 to 4 days, 147 (+/+), 148 (+/chico1 ), and 146 (chico1 ); and from 5 to 6 days, 150 (+/+), 150 (+/chico1 ), and 151 (chico1 ). Sample sizes for males were as follows: from 0 to 6 hours, 150 (+/+), 150 (+/chico1 ), and 101 (chico1 ); from 3 to 4 days, 149 (+/+), 148 (+/chico1 ), and 101 (chico1 ); and from 5 to 6 days, 149 (+/+), 149 (+/chico1 ), and 97 (chico1 ). In all panels, white bars represent +/+, gray bars represent +/chico1 , and black bars representchico1 /chico1 .

Figure 4

Effect of chico1 on total superoxide dismutase (Cu/Zn and Mn SOD) activity. Histograms show mean values with 95% confidence intervals. Probability that the values are identical to that of the wild type is indicated as follows: *, 0.05 > P > 0.01; **, P < 0.01 (paired t test). Replicate assays were conducted at several different times. Sample sizes for females were 6 (+/+), 6 (+/chico1 ), and 4 (chico1 ). Sample sizes for males were 5 (+/+), 6 (+/chico1 ), and 4 (chico1 ). White bars represent +/+, gray bars represent +/chico1 , and black bars representchico1 /chico1 .

Our results raise the question of whether IIS regulates aging in mammals. Whereas both the C. elegans andDrosophila genomes contain a single insulin/IGF receptor, mammals possess distinct receptors for insulin and IGF-I, plus a third insulin receptor–like receptor of unknown function. Potentially, any or all of these receptors may play a role in regulating aging. Caloric restriction (CR), which increases life-span in rodents (19), and possibly primates (20), reduces circulating levels of both insulin and IGF-I (21, 22). In the case of IGF-I, there is further evidence for a role in the control of longevity (22,23). Growth hormone (GH) acts via IGF-I to control mammalian body size, and circulating IGF-I levels correlate with body size in mice, dogs, and humans (23, 24). Furthermore, CR can reduce body size (22). In mice and dogs (and possibly humans), there is a marked negative correlation between body size and longevity (24–26). In addition, long-lived Ames hypopituitary mouse dwarves are deficient in GH and other pituitary hormones and have reduced circulating IGF-I (27). Mutation of the human equivalent of the Ames dwarf gene,Prop-1, also causes dwarfism and, possibly, delayed aging (28). The Laron dwarf mouse, which has no GH receptor and very low IGF-I levels, exhibits life-span increases of up to 55% (29).

Whereas the effects of chico1 on development that result in reduced body size are recessive, its effects on life-span are semidominant. This may reflect the noncatalytic and dosage-dependent nature of the function of CHICO as a docking protein. It has been proposed that reduced body size per se increases life-span in mammals (23). Alternatively, the same genes may independently regulate growth during the preadult period and regulate survival during the adult period. Our data support the latter interpretation because chico1 heterozygotes are long-lived, yet of normal size. Likewise, the effect of CR on aging may be observed in the absence of its effects on body size. Together, our results with fruit flies and recent findings with nematodes and mice suggest that the role of IIS (perhaps IGF-I in mammals) in regulating longevity is evolutionarily conserved throughout the animal kingdom.

  • * To whom correspondence should be addressed. E-mail: david.gems{at}


View Abstract

Navigate This Article