Estrogen, Insulin, and Dietary Signals Cooperatively Regulate Longevity Signals to Enhance Resistance to Oxidative Stress in Mice*

To investigate the biological significance of a longevity mutation found in daf-2 of Caenorhabditis elegans, we generated a homologous murine model by replacing Pro-1195 of insulin receptors with Leu using a targeted knock-in strategy. Homozygous mice died in the neonatal stage from diabetic ketoacidosis, whereas heterozygous mice showed the suppressed kinase activity of the insulin receptor but grew normally without spontaneously developing diabetes during adulthood. We examined heterozygous insulin receptor mutant mice for longevity phenotypes. Under 80% oxygen, mutant female mice survived 33.3% longer than wild-type female mice, whereas mutant male mice survived 18.2% longer than wild-type male mice. These results suggested that mutant mice acquired more resistance to oxidative stress, but the benefit of the longevity mutation was more pronounced in females than males. Manganese superoxide dismutase activity in mutant mice was significantly upregulated, suggesting that the suppressed insulin signaling leads to an enhanced antioxidant defense. To analyze the molecular basis of the gender difference, we administered estrogen to mutant mice. It was found that the survival of mice under 80% oxygen was extended when they were administered estradiol. In contrast, mutant and wild-type female mice showed shortened survivals when their ovaries were removed. The influence of estrogen is remarkable in mutant mice compared with wild-type mice, suggesting that estrogen modulates insulin signaling in mutant mice. Furthermore, we showed additional extension of survival under oxidative conditions when their diet was restricted. Collectively, we show that three distinct signals; insulin, estrogen, and dietary signals work in independent and cooperative ways to enhance the resistance to oxidative stress in mice.

To investigate the biological significance of a longevity mutation found in daf-2 of Caenorhabditis elegans, we generated a homologous murine model by replacing Pro-1195 of insulin receptors with Leu using a targeted knock-in strategy. Homozygous mice died in the neonatal stage from diabetic ketoacidosis, whereas heterozygous mice showed the suppressed kinase activity of the insulin receptor but grew normally without spontaneously developing diabetes during adulthood. We examined heterozygous insulin receptor mutant mice for longevity phenotypes. Under 80% oxygen, mutant female mice survived 33.3% longer than wild-type female mice, whereas mutant male mice survived 18.2% longer than wild-type male mice. These results suggested that mutant mice acquired more resistance to oxidative stress, but the benefit of the longevity mutation was more pronounced in females than males. Manganese superoxide dismutase activity in mutant mice was significantly upregulated, suggesting that the suppressed insulin signaling leads to an enhanced antioxidant defense. To analyze the molecular basis of the gender difference, we administered estrogen to mutant mice. It was found that the survival of mice under 80% oxygen was extended when they were administered estradiol. In contrast, mutant and wild-type female mice showed shortened survivals when their ovaries were removed. The influence of estrogen is remarkable in mutant mice compared with wild-type mice, suggesting that estrogen modulates insulin signaling in mutant mice. Furthermore, we showed additional extension of survival under oxidative conditions when their diet was restricted. Collectively, we show that three distinct signals; insulin, estrogen, and dietary signals work in independent and cooperative ways to enhance the resistance to oxidative stress in mice.
Recent experimental evidence suggests that oxidative stress is a principle cause of aging (1). Biochemical data further support the hypothesis that the oxidative damage accumulated in macromolecules such as DNAs, lipids, and proteins leads to the physiological decline in aging tissues. In a genetic analysis of invertebrate models, longevity mutants have been identified that show extension of lifespan associated with an enhanced resistance to oxidative stress (2). In Caenorhabditis elegans, a mutation of the daf-2 gene that encodes an insulin/insulin-like growth factor type 1 (IGF-1) 1 receptor significantly enhanced resistance to oxidative stress and extended lifespan (3). These phenotypes are regulated by the insulin/IGF-1 signaling pathway from DAF-2 to DAF-16, a homologue of the hepatocyte nuclear factor-3/forkhead transcription factor (4,5). Furthermore, expression of sod3 encoding manganese superoxide dismutase (MnSOD) is up-regulated in daf-2 mutants (5). In insulin-like signaling mutants of flies, it is suggested that the increased expression of SOD genes confers resistance to oxidative stress and the extension of lifespan (6). Thus, the regulation of SOD levels is critical for the determination of lifespan, which is modified by the insulin/IGF-1 signaling pathway. This signaling pathway has evidently been conserved from C. elegans to Drosophila (5,6).
Holzenberger et al. (7) reported that a heterozygous deficiency for the IGF-1 receptor in mice enhances resistance to oxidative stress and extends lifespan. Because in IGF-1 receptor deficient mice heterozygous females exhibited a greater increase in lifespan than males, the beneficial effect of reduced IGF-1 signaling is also modulated by gender difference. However, it is unclear whether insulin signaling controls tolerance to oxidative stress and the extension of lifespan in mammals or is modulated by gender difference. In humans, the life expectancy of women is generally longer than that of men. For example, in Japan, with the highest life expectancy in the world, life expectancy was 78.36 years for men and 85.33 years for women in 2003. However, in laboratory animal models, the relationship of gender to longevity is variable, with lifespan being dependent on other factors like breeding and diet (8). Borras et al. (9) paid attention to the gender difference in rats, reporting that the expression and enzyme activity levels of MnSOD as well as glutathione peroxidase are higher in females than males. Ho et al. (10) reported that human MnSOD transgenic mice showed a significant increase in survival when exposed to 90% oxygen. Furthermore, in pulmonary epithelial cells of mice the overexpression of human MnSOD confers protection against exposure to hyperoxygen (11). The regulation of antioxidant enzymes including MnSOD plays an important role in the determination of lifespan, suggesting that the gender difference in antioxidant enzymes explains the difference in life expectancy between females and males.
In the present study we generated a homologous murine model similar to the daf-2 mutant (Ir P1195L/wt ) mouse in order to investigate the biological significance of longevity mutations found in the daf-2 gene of C. elegans. We demonstrated that the mutant mice acquired an enhanced resistance to oxidative stress. Furthermore, we also revealed that the influence of gender difference and dietary restriction (DR) on the acquired resistance involved defective insulin signaling.

Generation of Insulin Receptor (IR) Mutant
Mice-The 129 mouse genomic library in FIXII (Stratagene, La Jolla, CA) was screened with rat IR cDNA containing exon 19 -21 as a probe. Three overlapping clones covered exon 15-22 of the mouse IR gene. The 1.1-kb fragment containing exon 20 of the gene was PCR-amplified with two SpeIanchored primers (5Ј-GGA CTA GTA GCA TTG AGA ACT GGA-3Ј and 5Ј-GGA CTA GTA CCT GAG TTC AAT GCC A-3Ј). The SpeI-restricted fragment carrying exon 20 was mutagenized with a 19-bp mutagenic oligonucleotide (5Ј-GGA TGT CAC TCG AGT CCC T-3Ј) using the pALTER system (Promega, Madison, WI). The introduced mutation, P1195L, was confirmed present by sequencing. The 2.6-kb short 3Ј homologous fragment was amplified with a XhoI-anchored primer (5Ј-CCG CTC GAG ATA GAG ACT ATT GTA CG-3Ј) and an ApaI/NotIanchored primer (5Ј-ATG GGC CCG CGG CCG CTG TGA ACA TAC CTC TG-3Ј). XhoI/ApaI-restricted short 3Ј homologous fragments were inserted into a XhoI/ApaI-restricted pBluescript II SK (pBSII SK, Stratagene) vector. The 1.3-kb MC1neo cassette flanked by the loxp sequence was PCR-amplified from a pMC1neo-lox vector (12) with a SalI-anchored primer (5Ј-CGC GTC GAC ATA ACT TCG TAT AAT G-3Ј) and a SalI/EcoRI-anchored primer (5Ј-CGC GTC GAC GAA TTC ATC GAT ACC GGC GAC ATA-3Ј). The SalI-restricted MC1neo-loxp cassette was inserted into a XhoI-restricted pBSII SK vector containing the short 3Ј homologous fragment. The SmaI/SalI-restricted fragment including a mutagenized exon 20 was inserted into a SmaI/SalI-restricted pBSII SK vector containing a short 3Ј homologous fragment and a MC1neo-loxp cassette. The 7.3-kb-long 5Ј homologous fragment was PCR-amplified with a NotI-anchored primer (5Ј-ATT TGC GGC CGC TGG CTA ACA ACT GAC TC-3Ј) and an SmaI-anchored primer (5Ј-TCC CCC GGG TTT CTT GAG ACG AGC-3Ј). The NotI/SmaI-restricted long 5Ј homologous fragment was inserted into a NotI/SmaI-restricted pMC1DT-A (Oriental Yeast, Tokyo, Japan). The resulting construct encompassing the fragment carrying a modified exon 20, the MC1neo-loxp cassette, and the short 3Ј homologous fragment was restricted with NotI, blunted, and inserted into SmaI-restricted pMC1DT-A containing the long 5Ј homologous fragment. The vector was then linearized with NotI and used for the electroporation of embryonic stem cells. Genomic DNAs from 240 G418-resistant embryonic stem clones were screened for homologous recombination by Southern blotting. One embryonic stem clone with the expected homologous recombination was used for generating chimeric mice by the aggregation method as described (13). The chimeric mice were cross-bred with C57BL/6CrSlc (Nihon SLC), and the germline transmission was confirmed by PCR amplification with the primers 5Ј-GCA TGT ATG TGG ACA CT-3Ј and 5Ј-GTG GAG GTC ATG GTT GAG CA-3Ј in agouti offspring. To analyze the sequence of the mutated allele in IR mutant mice, we prepared IR transcripts containing exons 19 -21 from livers of homozygous embryos by reverse transcription (RT)-PCR using as primers 5Ј-GCG AAT TCA ATA ACC CAG GCC GCC CT-3Ј and 5Ј-AGT CTC TCT GGA CAG TTA TCA-3Ј, as described below. We confirmed the presence of the introduced mutation, P1195L, and the absence of additional mutations in exon 20 of the transcripts by sequencing. To delete the neomycin resistance gene from the germline, we cross-bred heterozygous mice with CAG-Cre mice (kindly provided by Dr. Miyazaki, Osaka University) (14).
Animals were housed in specific pathogen-free facilities on a 12-h light/dark cycle (0800 on, 2000 off) and were fed a standard rodent chow (mouse diet CRF-1, Oriental Yeast, Tokyo) and water ad libitum (AL). All protocols for animal use and experiments were reviewed and approved by the Animal Care Committee of the Tokyo Metropolitan Institute of Gerontology.
Southern Blotting and PCR Analysis-For the knock-in allele, genomic DNA was digested with EcoRV, subjected to electrophoresis on a 0.7% agarose gel, and then transferred to a Hybond-N ϩ nylon membrane (Amersham Biosciences). Probes were labeled with [␣-32 P]dCTP (6000 Ci/ mmol, Amersham Biosciences) using the Megaprime DNA labeling system (Amersham Biosciences). Membranes were hybridized with a HincII/ HincII fragment as a 3Ј probe, and the fragment was visualized using a Bioimage analyzer BAS-1000 (Fuji Film, Tokyo, Japan). To differentiate wild-type and mutated alleles, we amplified the genomic fragment harboring exons 20 and 21 by PCR with a sense primer (P1, 5Ј-GGA ATG ACA AGG GAC ATC TA-3Ј) and antisense primer (P2, 5Ј-CAC CTG TTC ATT AGA CAG GCC-3Ј). PCR products were then digested with EcoRI to separate wild-type and mutated alleles.
Tyrosine Kinase Assay of IR-Livers from day1 neonates were solubilized in 1% Triton X-100, 50 mM HEPES (pH 7.6), 150 mM NaCl, 10 g/ml leupeptin, 10 g/ml pepstatin, and 1 mM phenylmethylsulfonyl fluoride. The insoluble material was separated by ultracentrifugation at 100,000 ϫ g for 1 h at 4°C. The supernatant was applied to a wheat germ lectin-agarose column pre-equilibrated with buffer A containing 0.1% Triton X-100, 50 mM HEPES (pH 7.6), 150 mM NaCl, and protease inhibitors. The column was extensively washed using buffer A, and the bound glycoproteins, such as IR, were eluted in buffer A containing 0.3 M N-acetylglucosamine. Aliquots of eluate were incubated in the absence or presence of 0.1 nM insulin (Eli Lilly) for 1 h at room temperature. Thereafter, phosphorylation was initiated by the addition of 0.1 mCi of [␥-32 P]ATP in the presence of 0.1 nM ATP. After 20 min at room temperature, the reaction was stopped by the addition of stopping solution containing 0.1 M NaF, 4 mM NaVO 3 , 1 mM EDTA, and 1 mM sodium pyrophosphate. Samples were immunoprecipitated with anti-IR antibody (kindly provided by Dr. Ebina, Tokushima University) (15) and protein A-Sepharose CL-4B (Amersham Biosciences). Labeled phosphoproteins were separated on sodium dodecyl sulfate, 10% polyacrylamide gels and detected by autoradiography.
Measurement of Protein Kinase B Phosphorylation-Mice were injected intravenously with 0.75 units kg Ϫ1 insulin or equal volume of vehicle. All tissues were collected in liquid nitrogen 5 min after the injection. Frozen tissues were homogenized with 6 (for muscle) or 10 (for liver and kidney) times the volume of the tissue in homogenization buffer (1% Triton X, 50 mM HEPES, pH 7.4, 100 mM sodium pyrophosphate, 100 mM NaF, 10 mM EDTA, 10 mM NaVO 3 , and 2 mM phenylmethylsulfonyl fluoride). Phosphorylation at Ser-473 of protein kinase B in tissue extracts was analyzed with anti-phosphoprotein kinase B Ser-473 and anti-protein kinase B antibodies (Cell Signaling) as described (16).
Oxidative Stress-Twenty (4-month-old) mice were continuously exposed to 80% oxygen for 14 days in a chamber (50 ϫ 30 ϫ 30 cm). The concentration of fractional inspired O 2 in the chamber was monitored with an oxygen analyzer (G-101-Y, Iijima Products, Gamagouri, Japan) and maintained with a constant flow of 80% oxygen gas (3 liters/min). Exposure was continuous for the time indicated except for a few minutes when the chamber was opened for housekeeping purposes weekly. The mice were fed food and water ad libitum and kept on a 12-h dark-light cycle at 25°C. We checked the survival of mice every 12 h. As chemical oxidative stress, the mice (4-month-old) were intraperitoneally injected with paraquat (Sigma) prepared in phosphate-buffered saline at a dose of 70 mg/kg of body weight. We checked the survival of mice every 2 h until 120 h after the injection.
Analytical Procedures-Blood glucose levels were determined in fed mice (3, 6, and 12 months old) using an automatic monitor, Gluco-card (Hoechst Marion Roussel). Serum insulin levels were measured in 16week-old mice. Serum obtained from fed mice was analyzed for insulin using a mouse insulin enzyme-linked immunosorbent assay kit (Shibayagi, Japan). In glucose tolerance tests fasting (4-month-old) mice were injected intraperitoneally with 2 g/kg body weight of 20% D-glucose. Blood glucose values were determined in whole blood obtained from the tail at 0, 15, 30, 60, and 120 min after the glucose injection. In insulin tolerance tests, 4-month-old mice were injected intraperitoneally with 1 units/kg of body weight of insulin. Blood glucose values were measured at 0, 15, 30, and 60 min after the injection.
SOD Activity Assay-Livers were homogenized in phosphate-buffered saline, pH 7.4, containing 2 mM EDTA, 2 mM EGTA, 2 mM phenylmethylsulfonyl fluoride, and 4 g/ml leupeptin, then sonicated 10 times using the handy sonicator. The lysates were spun at 15,000 ϫ g for 30 min to remove cell debris. SOD activity was measured by a competitive assay as described previously (17). The amount of protein that inhibits the tetrazolium dye WST-1 (Dojindo, Kumamoto, Japan) reduction to 50% of maximum is defined as 1 unit of SOD activity. Enzymatic activity was expressed in units per mg of protein. To measure MnSOD activity, KCN was added to tissue lysates at a final concentration of 1 mM to inhibit the cupper/zinc (Cu/Zn) SOD activity.
Ovariectomy (OVX) and Estradiol (E2) Administration-At 12 weeks of age, female mice were ovariectomized or sham-operated via a midline incision under anesthesia with pentobarbital (40 mg/kg, intraperitoneal injection). After 4 weeks, treated mice were exposed to 80% oxygen or SOD activity, and gene expression levels were analyzed. Male (12-weekold) mice and OVX female (16-week-old) mice were administered a weekly subcutaneous injection of 17␤-estradiol (20 mg/kg per week; Sigma) in corn oil (Ajinomoto, Japan). After 4 weeks of treatment, the male and female mice were exposed to 80% oxygen. E2-treated male mice were analyzed for SOD activity and gene expression.
DR-At 16 weeks of age we measured food intake per day in male and female mice and randomly assigned the mice to two groups, AL and DR. Four mice were housed in every cage (20 ϫ 10 ϫ 20 cm). The mice in the AL group were fed freely, whereas those in the DR group were fed a restricted diet on Monday, Wednesday, and Friday. First, DR mice were fed a 10%-restricted diet for a week, and then the restricted diet was gradually reduced by 65% for 3 weeks. Finally, DR consisted of 65% of food intake in AL mice. We adjusted the amount of restricted diet by measuring food intake in AL mice every 10 weeks. We housed DR mice until the age of 60 weeks.

RESULTS
The Generation of Modified IR Mutant Mice-To elucidate the biological significance of the longevity mutation found in the daf-2 mutant of C. elegans, we generated a homologous murine model by replacing Pro-1195 of IR with Leu by targeted knock-in of the genomic IR gene in mice. As shown in Fig. 1A, we first isolated an ES cell carrying a mutation homologous to daf-2 (e1391) (3), i.e. Leu-1195 in exon 20 of the IR genomic gene by homologous recombination with the target vector. Germline transmission was confirmed in the F1 offspring of the chimeric and C57BLACK/6 mice by Southern blotting. As shown in Fig. 1B, the blot showed that the 3Ј probe detected a 6.2-kb polymorphic EcoRV allele in wild-type (Ir wt/wt ) as well as heterozygous (Ir P1195L/wt ) mice, whereas the probe detected a 7.5-kb mutant EcoRV allele in heterozygous and homozygous (Ir P1195L/P1195L ) mice. To delete the neomycin resistance gene from the germline, we cross-bred Ir P1195L/wt mice with CAG-Cre mice (14). We successfully deleted the neomycin resistance gene from the germline by cross-breeding CAG-Cre mice as schematized in the bottom of Fig. 1A. To detect the wild-type and mutant alleles, we amplified the genomic fragment harboring exon 20 and 21 by PCR with a sense primer (P1) and an antisense primer (P2). The PCR products were then differentiated based on the EcoRI site that is specifically introduced with Open triangles represent loxp sequences. After the cross-breeding with Cre-expressing transgenic mice, the neo cassette was popped out from the genome (mutated allele). The probe for Southern blotting is shown as a black box. The arrows indicate primers P1 and P2 for PCR. RI, EcoRI; RV, EcoRV. B, Southern blot analysis of EcoRV-digested liver DNAs in embryos. The knock-in allele yielded a 7.5-kbp EcoRV fragment, whereas the wild-type allele yielded a 6.2kbp EcoRV fragment. C, PCR-based genotyping of Ir wt/wt , Ir P1195L/wt , and Ir P1195L/ P1195L embryos. Products amplified with primers P1 and P2 were digested by EcoRI to identify wild-type and mutated alleles. D, insulin-induced IR phosphorylation in liver of IR mutant mice. E, insulin-induced protein kinase B (PKB) phosphorylation in liver, kidney, and muscle of IR mutant mice. the knock-in allele (Fig. 1C). A genotypic analysis of the offspring showed that the frequency of wild-type, heterozygous and homozygous mice was 31:61:0 at the 4-week-old stage. Further observation, however, showed that some mice were identified as homozygous at 1 day old, which apparently showed that the growth retardation was complicated by the diabetic ketoacidosis. 2 The result suggested that the homozygous mice failed to thrive due to diabetic ketoacidosis in the neonatal stage, which is consistent with the phenotypes previously reported in IR-deficient mice (18,19).
To investigate the autophosphorylation activity of IR in IR mutant mice, insulin was added to extracts from the neonatal liver of the mice. As shown in Fig. 1D, we did not detect any insulin-induced autophosphorylation activity in the Ir P1195L/P1195L mouse. On the other hand, the Ir P1195L/wt mouse did show the insulin-induced autophosphorylation although it was less than 20% that of the Ir wt/wt mouse. Furthermore, to investigate whether the downstream activity of IR is suppressed by the mutation of the IR gene, we monitored protein kinase B phosphorylation, a more distal event in IR signaling (Fig. 1E). As with IR autophosphorylation in Ir P1195L/wt mice, insulin-induced protein kinase B phosphorylation was suppressed in liver, kidney, and muscle of Ir P1195L/wt mice compared with Ir wt/wt mice (Fig. 1E). These results indicated that the tyrosine kinase activity of IR from heterozygous (Ir P1195L/wt ) mice was severely suppressed in a dominant negative manner, suggesting that IR mutant mice provide an excellent model for the analysis of longevity phenotypes induced by suppressed insulin signaling compared with heterozygous IR-deficient mice with a haploinsufficiency of insulin signal.
IR Mutant Mice Showed Enhanced Resistance to Oxidative Stress-To investigate whether IR mutant mice acquired enhanced resistance to oxidative stress, the mice at 4 months of age were exposed to 80% oxygen in a chamber. We then compared the survival of IR mutant mice to that of wild-type mice (Fig. 2). IR mutant females survived 33.3% longer than female wild-type mice, i.e. IR mutant mice had a 50% survival rate at 8 days compared with 6 days for wild-type mice ( Fig. 2A). On the other hand IR male mutant mice survived 18.2% longer than wild-type male mice, i.e. IR mutant mice had a 50% survival rate at 6.5 days compared with 5.5 days for wild-type mice (Fig. 2B). These results suggested that IR mutant mice are more resistant to oxidative stress, but the benefit of the mutation was more pronounced in female than male mice.
Next, we intraperitoneally injected paraquat, a chemical oxidant, into female and male mice to evaluate the defense system of the mutant mice against chemically induced oxidative stress. As shown in Fig. 2C, both IR mutant and wild-type female mice began to die ϳ48 h after the administration of paraquat. IR mutant female mice, however, survived significantly longer than wild-type mice in stages beyond 72 h. IR mutant male mice also survived longer than wild-type mice, but the beneficial effects in male mutant mice were less pronounced in comparison to those in female mutant mice (Fig.  2D). These results suggested that IR mutant mice of both genders survived longer than wild-type mice in oxidative conditions. It is also of note that there is a gender difference in the resistance to oxidative stress in IR mutant mice. This data suggested that the insulin signaling may be modulated by sex hormones (i.e. either by the beneficial effects of estrogen or by the deleterious effects of testosterone).
MnSOD Plays a Role in the Acquisition of Resistance to Oxidative Stress-Because MnSOD and Cu/ZnSOD play an important role in the intracellular defense against reactive oxygen species (ROS), we measured SOD activities in the liver of IR mutant mice. As shown in Fig. 3A, the MnSOD activity of IR mutant female mice was significantly up-regulated by 39.9% compared with that of wild-type female mice (484.9 Ϯ 96.1 units/mg of protein for IR mutant mice versus 346.5 Ϯ 51.7 units/mg of protein for wild-type mice, p Ͻ 0.05). As for male mutant mice, MnSOD activity was up-regulated by 22.9% in comparison with that of wild-type male mice, but the up-regulation was less significant than that found in wild-type female mice (Fig. 3B). In contrast, we did not detect any pronounced difference in Cu/ZnSOD activity between IR mutant and wildtype mice of either sex (Fig. 3, C and D). To analyze the molecular basis for the up-regulation of MnSOD, mRNAs for MnSOD and Cu/ZnSOD were analyzed by RT-PCR (Fig. 3E). We found that the transcription of MnSOD was up-regulated in liver, brain, kidney, and heart of both female and male mutant mice, whereas that of the Cu/ZnSOD and glyceraldehyde-3phosphate dehydrogenase genes was not up-regulated in mutant or wild-type mice (Fig. 3E). Furthermore, to investigate the contribution of other antioxidant enzymes to resistance to oxidative stress, we measured activities of glutathione peroxi-

FIG. 2. IR mutant mice at 4 months of age show resistance to oxidative stress.
A and B, oxidative stress was induced by exposure to 80% oxygen. Kaplan-Meier analysis showed a significant increase in survival of Ir P1195L/wt compared with Ir wt/wt mice. Female mutants exhibited increased stress resistance (p Ͻ 0.005, g-Wilcoxon test, at least 16 mice per genotype), whereas male mutants showed a significant increase (p Ͻ 0.05, g-Wilcoxon test, at least 15 mice per genotype). C and D, paraquat was injected. Ir P1195L/wt female mice survived longer than wild-type mice beyond 72 h. IR mutant male mice also survived longer than wild-type mice (not significant, g-Wilcoxon test, n ϭ 10 in each group). dase and catalase. However, we did not detect any significant differences between IR mutant and wild-type mice (data not shown). These results suggested that the transcriptional upregulation of MnSOD is specific to the enhanced antioxidant system in IR mutant mice. In C. elegans and Drosophila melanogaster, the up-regulated SODs are reported to degrade the oxygen radicals that cause oxidative stress in IR/ IGF-1 receptor mutants (5,6,20). Kops et al. (21) reported that MnSOD was up-regulated upon treatment with insulin in cultured 3T3-L6 cells. Here we demonstrated for the first time that the expression and the activity of MnSOD were up-regulated by the mutation introduced into the kinase domain of the IR gene in mice. Transgenic flies, which overexpressed Cu/ZnSOD in motor neurons, survived 40% longer than control fruit flies, whereas transgenic flies that overexpressed MnSOD in motor neurons survived 30% longer than control flies (22,23). The up-regulation of MnSOD may confer an extension of lifespan in Drosophila. The regulation of MnSOD by insulin signaling was shown to be conserved between invertebrates and mammals, suggesting that the antioxidant system against ROS plays an important role in the determination of animal lifespan in both invertebrates and vertebrates.
Estrogen Signaling Enhances Tolerance to Oxidative Stress in IR Mutant Mice-The MnSOD activity of female mice was significantly up-regulated in comparison with that of male mice (Fig. 3, A and B). Strehlow et al. (24) reported that the gene expression of MnSOD and extracellular SOD in ovariectomized mice was down-regulated to 40 -60% that of the control level and was recovered through estrogen replacement therapy (24). To investigate whether estrogen signaling confers resistance to oxidative stress and up-regulates MnSOD expression in IR mutant mice, we performed OVXs on female mice. For male mice, on the other hand, we administered E2 for 4 weeks.
First, we investigated the oxidative tolerance under hyperoxic conditions and the activity of MnSOD in the liver of OVX mice. As shown in Fig. 4, A and B, we compared the survival rate of OVX mice with that of non-OVX mice in both the IR mutant mice and wild-type strains. OVX mutant mice died significantly earlier than non-OVX mutant mice, i.e. OVX mutant mice had a 50% survival rate at 6.5 days compared with 12 days for non-OVX mutant mice (Fig. 4A). OVX wild-type mice died earlier than non-OVX wild-type mice, i.e. OVX wild-type mice had a 50% survival rate at 5.5 days compared with 7 days for non-OVX wild-type mice (Fig. 4B). The shortening of the survival rate was more pronounced in IR mutant mice than the wild-type mice, when both were ovariectomized. To clarify the molecular mechanisms of shortened survival in OVX mice, we investigated the gene expression and enzyme activity of MnSOD in the liver of mutant mice. As shown in Fig. 5A, the MnSOD activity of IR mutant mice was significantly downregulated upon OVX (484.9 Ϯ 96.1 units/mg of protein for mutant mice versus 387.3 Ϯ 88.9 units/mg of protein for OVX mutant mice, p Ͻ 0.05). The down-regulation of MnSOD activity was less pronounced in wild-type OVX mice. In contrast, we did not detect any difference in Cu/ZnSOD activity upon OVX (Fig. 5C). The transcription of MnSOD was significantly downregulated in OVX mice (Fig. 5E).
Second, to investigate the influence of exogenous estrogen on male mice, we administered E2. After exposure to 80% oxygen, we compared the survival of males treated with E2 to that of untreated make of both IR mutant mice and wild-type mice. Interestingly, mutant males administered E2 survived significantly longer than untreated mutant mice (Fig. 4E). The extension of survival was more pronounced in IR mutant mice than wild-type mice, when both were administered estrogen (Fig. 4, E and F). As shown in Fig. 5B, MnSOD activity was up-regulated by 20.6% in IR mutant mice treated with E2. The MnSOD activity of wild-type male mice was also up-regulated by 15.3% upon administration of E2. In contrast, we found no up-regulation of Cu/ZnSOD activity with E2 administration. (Fig. 5D). The transcription of MnSOD was significantly upregulated in male mice administered E2 (Fig. 5F).

FIG. 3. The expression and activities of antioxidant enzymes.
A, MnSOD activity was significantly stronger in Ir P1195L/wt female mice than Ir wt/wt female mice. B, Ir P1195L/wt male mice showed higher levels of MnSOD activity than Ir wt/wt male mice but not significantly so. C and D, Cu/ZnSOD activity was not up-regulated in mutant mice. E, RT-PCR analysis. Expression of the MnSOD gene was stronger in liver, brain, kidney, and heart of Ir P1195L/wt male and female mice than Ir wt/wt mice. Expression of the Cu/ZnSOD gene was not up-regulated in mutant mice. The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene as an internal control is shown. Each bar represents the mean Ϯ S.D. in each group from at least five mice per genotype. *, p Ͻ 0.05; Ir P1195L/wt mice versus Ir wt/wt mice.
Finally, to exclude the possibility that the deficiency of testosterone caused by the feedback regulation of E2 administration is attributable to the phenotype of E2-treated males, we investigated the influence of exogenous estrogen on OVX mice. After 4 weeks of E2 administration, OVX mice were exposed to 80% oxygen. OVX mutant mice treated with E2 survived significantly longer than untreated mutant mice; the E2-treated group had 50% survival at 10.5 days compared with 6.5 days for the untreated group (Fig. 4C). OVX wild-type mice treated with E2 also survived longer than untreated wild-type male mice but not significantly so (Fig. 4D). We demonstrated here that mutant mice administered E2 survived longer with hyperoxygen as well as showed an up-regulation of MnSOD expression. Interestingly, the influence of estrogen was remarkable on IR mutant mice compared with wild-type mice.
The Gender Difference of Glucose Metabolism in IR Mutant Mice-To determine whether the suppression of IR signaling induces an impairment of glucose metabolism, levels of glucose (Fig. 6A) and insulin (Fig. 6B) were analyzed in blood samples. In a fed state, IR mutant mice had almost the same blood glucose concentrations as wild-type mice. However, about 6% of mutant male mice developed hyperglycemia (more than 200 mg/dl) (Fig. 6A). This rate was the same as in mice heterozygous for insulin receptor knock-out (25). In contrast, we did not detect any differences in urinary glucose between mutant and wild-type mice (data not shown). Serum insulin concentrations in the fed state were significantly increased in both female and male mutant mice compared with wild-type mice (Fig. 6B). FIG. 4. Estrogen signaling enhances resistance to oxidative stress in Ir P1195L/wt mice. A and B, OVX mice died earlier than non-OVX mice. The shortening of the survival period was greater in Ir P1195L/wt mice (p Ͻ 0.05, g-Wilcoxon test) than Ir wt/wt mice (not significant, g-Wilcoxon test). C and D, estrogen replacement improved the resistance to oxidative stress in OVX mice. The influence of E2 was remarkable in Ir P1195L/wt female mice (p Ͻ 0.05, g-Wilcoxon test) compared with Ir wt/wt mice (not significant, g-Wilcoxon test) upon OVX. E and F, male mice treated with E2 survived longer than untreated mice. The influence of E2 was remarkable in Ir P1195L/wt male mice (p Ͻ 0.05, g-Wilcoxon test) compared with Ir wt/wt mice (not significant, g-Wilcoxon test).
Although we failed to detect any gender difference in insulin concentrations of wild-type mice, insulin concentrations were significantly increased in IR mutant males (6.40 Ϯ 1.49 ng/ml) compared with IR mutant females (2.96 Ϯ 1.63 ng/ml) (Fig. 6B).
To determine the physiological consequence of mutated IR, glucose and insulin tolerance tests were performed on 4-monthold mice. Both female and male IR mutant mice showed normal fasting blood glucose concentrations as compared with control mice (Fig. 6, C and D). However, IR mutant male mice showed a moderately impaired glucose tolerance as compared with control male mice (Fig. 6D), whereas IR mutant female mice showed normal glucose tolerance (Fig. 6C). In addition, IR mutant females were more sensitive to blood glucose-lowering effects than wild-type mice when insulin was administered (Fig. 6E). In contrast, IR mutant male mice showed resistance to blood glucose-lowering effects in the insulin tolerance test (Fig. 6F). Although IR signaling was suppressed in IR mutant mice, mutant females sustained insulin sensitivity with hyperinsulinemia. This gender difference suggested that sex hormones may modulate the insulin signaling in IR mutant mice.
Then, to investigate whether sex hormones influence glucose metabolism, glucose and insulin tolerance tests were performed on female OVX mice and male mice treated with E2. Upon OVX, the glucose tolerance of IR mutant mice showed no significant changes as compared with that of wild-type mice (Fig. 6G). However, OVX mutant mice showed impaired insulin sensitivity compared with non-OVX mutant mice (Fig. 6, E and  I). Furthermore, E2 administration improved glucose intolerance and insulin resistance in IR mutant male mice (Fig. 6, H and J compared with D and F). In contrast, wild-type mice showed no significant changes upon the administration of E2.
As for the mechanism of action of estrogen, Pedersen et al. (26) reported that IR binding was increased in adipocytes from rodents treated with estrogen. In that case, IR binding seemed to be due to an increased number of IRs (26). Yu et al. (27) reported that an IR signaling pathway is involved in estrogenmediated retinal neuroprotection. The influence of estrogen is remarkable on IR mutant mice when compared with wild-type mice, suggesting that estrogen partially enhances its effect through insulin signaling.
DR Additively Enhances the Resistance against Oxidative Stress in IR Mutant Mice-Dietary restrictions that facilitate a reduction in body weight, mainly fat mass, are known to prolong lifespan by enhancing resistance against oxidative stress in mammals (28 -30). Bluher et al. reported that fat-specific insulin receptor knock-out mice showed reduced adiposity, suggesting that the reduction of fat mass may be beneficial for the extension of lifespan (31,32). To clarify whether there is a reduction of adipose tissue induced by mutation in the IR gene, we observed whole body and organ weights of IR mutant mice. Both male and female IR mutant mice sustained a significant 10 -15% reduction in body weight compared with wild-type mice (Fig. 6, A and B). The fat mass of IR female mutant mice decreased significantly by 41.3% when compared with wildtype mice (446.6 Ϯ 36.1 versus 761.4 Ϯ 183.0 mg, p Ͻ 0.05), whereas that of IR male mutant mice dropped significantly by 57.5% compared with wild-type mice (522.9 Ϯ 54.7 versus 1230.0 Ϯ 196.0 mg, p Ͻ 0.01). We found no difference in the weights of other organs between IR mutant and wild-type mice (Fig. 6, A and B). To exclude the possibility that IR mutant mice show reduced food intake, we measured the food intake of mice at age of 16 weeks. We did not detect any significance difference in food intake between wild-type and mutant mice (female; 3.37 Ϯ 0.1 versus 3.45 Ϯ 0.18 g/day, male; 3.54 Ϯ 0.11 versus 3.57 Ϯ 0.15 g/day). We demonstrated that IR mutant mice specifically exhibited a reduction in fat mass, suggesting that the phenotype of IR mutant mice is very similar to that of fat-specific insulin receptor knock-out mice (31,32).
To elucidate the influence of DR in IR mutant mice, the diet was restricted to 65% that consumed by the mice fed ad libi- tum. As a result, IR mutant female mice with DR showed a 78.9% reduction of fat mass compared with AL wild-type mice, whereas the mutant males with DR showed an 88.4% reduction (Fig. 7, A and B). Although IR mutant and wild-type mice with DR also showed a reduction in the weights of other tissues, it was not as remarkable as that observed in fat mass. We demonstrated that genetic manipulation further enhances the reduction of fat mass induced by DR.
To clarify the influence of the reduction in fat tissue in IR mutant mice, we exposed IR mutant and wild-type mice to 80% oxygen. Interestingly, IR mutant mice showed a comparable survival rate to wild-type mice with DR (Fig. 7, C and D). The result suggested that genetic factors as well as DR independently play an important role, although we could not rule out cross talk between them. Surprisingly, IR mutant female and male mice both showed an extended lifespan upon DR (Fig. 7,  C and D). Houthoofd et al. (33) reported that lifespan extension via DR is independent of the insulin/IGF-1 signaling pathway in C. elegans. Although the reduction of fat mass induced by DR is more pronounced in mutant males than females, IR mutant female mice survived longer than IR mutant male mice under oxidative conditions. These results indicated that IR mutant female mice showed beneficial effects of DR as well as endogenous estrogen for the resistance to oxidative stress.

MnSOD Is Up-regulated in Vivo in IR Mutant
Mice-Daf-2 is one of the longevity mutants found in C. elegans (3). The causative mutation was found in the gene encoding the insulin-like receptor, suggesting that the daf-2 mutant has a defective insulin-like signaling, which eventually triggers the dauer formation as well as the extension of lifespan in C. elegans (3,4). The mutation of daf-16, a major suppressor mutant for daf-2, was found in the gene encoding the forkhead transcription factor (34,35). Biochemical analyses revealed that DAF-16, or dFOXO of the fly ortholog, played a pivotal role as a transcription factor by regulating the longevity genes downstream of the insulin signaling pathway (34 -36). The signaling pathway triggered by insulin or insulin-like ligands is well conserved among various animal species, including mammals and invertebrates such as D. melanogaster, in which mutants for the insulin receptor, IR, or IRS-1, chico, also exhibit the longevity phenotype (6,20). One of the candidate longevity genes in the downstream of daf-16 is the MnSOD gene (37), whose expression is specifically up-regulated in daf-2 mutants (5). Because biochemical studies show that MnSOD catalyzes the superoxide endogenously generated in the mitochondrial matrix, the balance between the generation of ROS and its degradation capacity in mitochondria may be critical for the determination of lifespan (38), which favors the free radical theory of aging (39,40). Kops et al. (21) recently demonstrated in vitro that FOXO3a, a homologue of DAF-16, actually induced the expression of MnSOD, enhancing the antioxidative defense system in cultured mammalian cells challenged with oxidative stress. In the present study we demonstrated for the first time that MnSOD expression is up-regulated in vivo by defective insulin signaling in IR mutant mice. Because IR mutant mice carry a mutation homologous with daf-2 with respect to the severely suppressed signaling of insulin, the up-regulation of MnSOD suggests that one of the longevity signals is sent to the anti-oxidative defense system, which is well conserved between invertebrates and mammals. Although further analyses are needed for elucidation of the molecular basis for the regulation of MnSOD, we found the transcriptional up-regulation of the Mn-SOD gene in IR mutant mice, suggesting the direct involvement of DAF-16 in the regulation of the mammalian MnSOD gene.
Estrogen Confers Resistance to Oxidative Stress by Up-regulating the MnSOD Gene-As shown in Fig. 4, the acquired FIG. 6. The gender difference of glucose metabolism in IR mutant mice. A, most Ir P1195L/wt mice exhibited blood glucose concentrations within the range of Ir wt/wt mice. B, insulin concentrations were significantly increased in Ir P1195L/wt male mice as compared with Ir P1195L/wt female mice. C and D, glucose tolerance tests performed on 4-month-old Ir P1195L/wt mice. We did not detect any differences in glucose tolerance between Ir P1195L/wt and Ir wt/wt female mice (C). Ir P1195L/wt male mice showed moderately impaired glucose tolerance compared with Ir wt/wt male mice (D). E and F, insulin tolerance tests were performed on random-fed, 4-month-old Ir P1195L/wt mice. The testing demonstrated that Ir P1195L/wt female mice were more sensitive to insulin than Ir wt/wt female mice (E). Ir P1195L/wt male mice were resistant to insulin (F). Upon OVX, glucose tolerance of Ir P1195L/wt mice showed no significant change as compared with that of Ir wt/wt mice (G). OVX Ir P1195L/wt mice showed an impaired insulin sensitivity compared with non-OVX mice (I). E2 administration improved glucose intolerance and insulin resistance in Ir P1195L/wt male mice (H and J). Each bar represents the mean Ϯ S.D. in each group from at least six mice per genotype. *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.005; Ir P1195L/wt mice versus Ir wt/wt mice. resistance to oxidative stress observed in female IR mutant mice was significantly attenuated when their ovaries were surgically removed. Because the ovary secretes the sex hormones estrogen and progesterone, either or both sex hormones confer resistance to oxidative stress in the mutant mice. To address this endocrinological issue, we administered E2 to male IR mutant mice as well as wild-type littermates. Interestingly, both the mutant and wild-type mice survived longer when they were administered E2, suggesting that estrogen made the male mice resistant to oxidative stress. This result is particularly interesting given that patients with prostate cancer are often prescribed with estradiol for hormone therapy. It is also interesting that female animals including humans generally live longer than males. In this context female hormones may protect the body from oxidative stress during the reproductive period, but the beneficial effects would fade after menopause. These assumptions are highly compatible with the notion that age-associated disease, such as atherosclerosis, osteoporosis, and dementia tend to appear more frequently after menopause in elderly females.
As for the molecular mechanism by which estrogen affects the signaling pathway triggered by insulin, Bailey et al. (41) showed that ovariectomized rodents showed an impaired glucose tolerance with an impaired secretion of insulin, suggesting that sex hormones have a direct pharmacological effect on the endocrinological function of insulin. In humans as well, postmenopausal women develop insulin resistance associated with an increase in body fat. Estrogen hormone replacement therapy improves insulin sensitivity with significant loss of fat tissue (42). Because the pharmacological actions of estrogen depend on the receptor subtypes expressed in each tissue, subtype-specific effects of estrogen should be investigated in subtype-specific estrogen receptor (ER)-deficient mice, ER␣deficient or ER␤-deficient mice (43,44).
DR Confers Distinct Signaling for the Determination of Lifespan-Evidence has accumulated that DR extends the lifespan of animals associated with an enhanced resistance to oxidative stress in various species (28 -30). Mutant mice with growth hormone signaling such as Ames dwarf mice showed extended lifespan compared with wild-type mice when they were raised in an ordinary laboratory environment (45)(46)(47). Interestingly, these mutant mice had an even longer lifespan when raised with a reduced diet (48). The result suggests that the longevity signaling evoked by DR may be in large part independent and distinct from the insulin signaling, although DR itself suppresses glucose metabolism and insulin secretion because of the restricted dietary intake.
Another genetic model, the fat-specific insulin receptor knock-out mouse, which has a tissue-specific defect of insulin signaling in adipose tissue, showed reduced adiposity without a restriction of diet. These mice also showed extended lifespan even when they were raised with an ordinary diet, suggesting that the reduction in fat mass attributed to the longer lifespan (32). It is, thus, speculated that the amount of fat mass may be a major determinant of lifespan irrespective of environmental factors or genetic factors. In the present study, however, we revealed that the amount of fat mass and the resistance to oxidative stress are not necessarily correlated (Fig. 7, B and D), suggesting that not only the reduction of fat mass but other environmental or genetic factors contribute to the acquired resistance to oxidative stress. The idea that DR has an independent and distinct signaling mechanism from insulin is also supported by the report that daf-2 mutants showed additional extension of lifespan in C. elegans when their diet of was restricted (33).
In the present study we demonstrated that in IR mutant mice resistance to oxidative stress was further enhanced by DR or E2. We are still finishing an analysis of the lifespan of IR mutant mice under normoxic conditions. We also expect a gender difference in the lifespan of IR mutant mice based on the analysis in oxygen chambers. Female mutant mice may have an enhanced defense system due to the estrogen secreted by the ovaries during the reproductive period. We demonstrated here that three distinct signals, insulin, estrogen, and dietary signals, work in different and independent ways and together increase resistance to oxidative stress and MnSOD levels in mice. FIG. 7. The influence of DR in IR mutant mice. A and B, comparison of mean body weights (BW) and organ weights between Ir P1195L/wt mice and Ir wt/wt mice. Ir P1195L/wt mice had significantly lower body weights than Ir wt/wt mice. The reduction in fat mainly causes the reduction in body weight. C and D, R additively enhances the resistance against hyperoxygen in Ir P1195L/wt mice. Each bar represents the mean Ϯ S.D. in each group from at least five mice per genotype. *, p Ͻ 0.05; **, p Ͻ 0.01; Ir P1195L/wt mice versus Ir wt/wt mice.