48. T. J. Imig, H. O. Adrian, Brain Res. 138, 241 (1977).49. R. A. Galuske, W. Schlote, H. Bratzke, W. Singer,Science 289, 1946 (2000).50. S. L. Juliano, P. J. Hand, B. L. Whitsel, J. Neurophysiol.46, 1260 (1981).51. C. N. Woolsey, E. M. Walz, in Cortical SensoryOrganization, Multiple Auditory Areas, C.N. Wool-sey, Ed. (Humana, Totawa, NJ, 1982) vol. 3, pp.231–256.52. J. Kaas, C. E. Collins, Curr. Opin. Neurobiol. 11, 498(2001).53. H. Kosaki, T. Hashikawa, J. He, E. G. Jones, J. Comp.Neurol. 386, 304 (1997).54. We thank S. Ghaznavi and H. J. Alitto for theirassistance with the animals, and L. Ungerleider [Na-tional Institute of Mental Health (NIMH) and NIH]and J. Rauschecker (Georgetown University) for theircritical reading of the manuscript and helpful sugges-tions. Supported by NIMH-IRP, NIH, and the U.S.Department of Health and Human Services.Supporting Online Materialwww.sciencemag.org/cgi/content/full/299/5606/568/DC1Materials and MethodsFig. S130 September 2002; accepted 4 December 2002Extended Longevity in MiceLacking the Insulin Receptor inAdipose TissueMatthias Blu¨her,1Barbara B. Kahn,2C. Ronald Kahn1*Caloric restriction has been shown to increase longevity in organisms rangingfrom yeast to mammals. In some organisms, this has been associated with adecreased fat mass and alterations in insulin/insulin-like growth factor 1 (IGF-1)pathways. To further explore these associations with enhanced longevity, westudied mice with a fat-specific insulin receptor knockout (FIRKO). These an-imals have reduced fat mass and are protected against age-related obesity andits subsequent metabolic abnormalities, although their food intake is nor-mal. Both male and female FIRKO mice were found to have an increase inmean life-span of ⬃134 days (18%), with parallel increases in median andmaximum life-spans. Thus, a reduction of fat mass without caloric restric-tion can be associated with increased longevity in mice, possibly througheffects on insulin signaling.Longevity is dependent on many factors in-cluding genetics (1, 2), hormonal and growthfactor signaling (3, 4), body weight (5), bodyfat content, and environmental factors (4, 6).Food restriction is the most potent environ-mental variable and has been shown to in-crease longevity in diverse organisms (6).The effect of restricted feeding on life-spanhas been studied in rodents for more than 60years (7–10), and although some studies havesuggested that reduced food intake is moreimportant than adiposity (8, 9), it is difficultto separate the beneficial effect of caloricrestriction from that of leanness and the var-ious biochemical correlates of leanness.To investigate this question, we evaluatedthe life-span of the fat-specific insulin recep-tor knockout (FIRKO) mouse. These animalswere derived by crossing insulin receptor IR(lox/⫹) mice, in which exon 4 of the insulinreceptor is flanked by loxP sites (11), with IR(lox/⫹) mice that also express the Cre recom-binase under the control of the aP2 promoter/enhancer (12). This breeding strategy alsogenerated three littermate control groups:wild-type (WT), IR (lox/lox), and aP2-Cremice, which were indistinguishable with re-gard to physiologic and metabolic parametersand have the same mixed genetic backgroundas the FIRKO mice. For the aging experi-ments, 250 animals were housed under thesame conditions in a virus-free facility on a12-hour light/dark cycle and were given astandard rodent feed and water ad libitum.Growth curves were normal in male andfemale FIRKO mice from birth to 8 weeks ofage. Starting at 3 months of age, FIRKO micemaintained 15 to 25% lower body weightsand a 50 to 70% reduction in fat massthroughout life (Fig. 1A). The reduction inadiposity was estimated by perigonadal fatpad weight but was apparent in all fat depotsand was also reflected by a reduction of⬃25% in total-body triglyceride content (13).FIRKO mice are healthy, lack any of themetabolic abnormalities associated with lipo-dystrophy, and are protected against age-re-lated deterioration in glucose tolerance,which is observed in all control strains (13).FIRKO mice maintained low body fat, de-spite normal food intake (Fig. 1B). Indeed,because FIRKO mice were leaner, the foodintake of FIRKO mice expressed per gram ofbody weight actually exceeded that of con-trols by an average of 55% (Fig. 1C).The median life-span of most laboratory1Joslin Diabetes Center and Department of Medicine,Harvard Medical School, One Joslin Place, Boston, MA,02215 USA.2Department of Medicine, Beth IsraelDeaconess Medical Center and Harvard MedicalSchool, Boston, MA, 02215 USA.*To whom correspondence should be addressed. E-mail: [email protected] 3 6 9 12 15 18 21 24 27 30 333600.511.522.533.56 12182430360501001502006 1218243036% Fat pad mass/Body weightAge (months)Age (months)Food intake per body weight(mg/g)Age (months)Food intake (g/mouse/day)ABCFig. 1. Age-related changes in fat pad mass per grambody weight (A), food intake (B), and food intakeper gram body weight in FIRKO mice (C) (initialcohort, n ⫽ 40) and control littermates (initialcohort, n ⫽ 20 per genotype). (A) We determinedfat pad mass in animals killed at each age by directlyweighing perigonadal fat pads. There was no differ-ence in body weight or fat pad mass among thethree control genotypes [WT, IR (lox/lox), and aP2-Cre]. After the mice reached the age of 3 months, the differences in fat pad mass per gram of bodyweight were significant for all data points between FIRKO mice (diamonds) and all three controls(squares) (P ⬍ 0.05). (B) In mice caged singly, we determined food intake (gram per mouse per day)daily over 5 days by using at least five FIRKO (white bars) and four control mice (black bars) pergenotype (n ⫽ 12). Data of the control genotypes [WT, IR (lox/lox), and aP2-Cre] are plottedtogether in the black bars, because there were no differences in daily food intake among them.(C) Food intake per gram body weight—calculated from the food intake and body weightdata—was significantly increased in FIRKO mice (white bars) as compared with controls (blackbars) (P ⬍ 0.05).R EPORTS24 JANUARY 2003 VOL 299 SCIENCE www.sciencemag.org572mice, such as BALB/c-nu (5) or C57BL/6J(14), is 30 months. Consistent with this, wefound that 45 to 54% of the mice in thecontrol groups lived to 30 months of age. Bycontrast, about 80% of the FIRKO mice werealive at this age (Fig. 2A). The increasedsurvival of FIRKO mice was seen in bothmales and