Insulin-like growth factor 1 receptor regulates hypothermia during calorie restriction

Significance

Energy homeostasis is fundamental for the survival of living organisms and contributes to their health, longevity, and aging. When food resources are scarce, and during experimental calorie restriction, endothermic animals can lower their core body temperature. Here, we found that this response is regulated by the insulin-like growth factor 1 receptor. This demonstrates that the three main factors affecting aging and longevity (calorie restriction, reduction of the insulin-like growth factor 1 signaling, and lowered temperature) are components of the same pathway that modulates energy homeostasis. The finding also identifies body temperature reduction as a common determinant of the effects of both calorie restriction and reduced insulin-like growth factor 1 receptor signaling.

Abstract

When food resources are scarce, endothermic animals can lower core body temperature (T_b). This phenomenon is believed to be part of an adaptive mechanism that may have evolved to conserve energy until more food becomes available. Here, we found in the mouse that the insulin-like growth factor 1 receptor (IGF-1R) controls this response in the central nervous system. Pharmacological or genetic inhibition of IGF-1R enhanced the reduction of temperature and of energy expenditure during calorie restriction. Full blockade of IGF-1R affected female and male mice similarly. In contrast, genetic IGF-1R dosage was effective only in females, where it also induced transient and estrus-specific hypothermia in animals fed ad libitum. These effects were regulated in the brain, as only central, not peripheral, pharmacological activation of IGF-1R prevented hypothermia during calorie restriction. Targeted IGF-1R knockout selectively in forebrain neurons revealed that IGF signaling also modulates calorie restriction-dependent T_b regulation in regions rostral of the canonical hypothalamic nuclei involved in controlling body temperature. In aggregate, these data identify central IGF-1R as a mediator of the integration of nutrient and temperature homeostasis. They also show that calorie restriction, IGF-1R signaling, and body temperature, three of the main regulators of metabolism, aging, and longevity, are components of the same pathway.

Energy homeostasis is fundamental for the survival of living organisms and is one of the main regulators of health, aging, and longevity. Endothermic animals have the ability to respond to calorie restriction (CR) by reducing their core body temperature (T_b; reviewed in ref. 1). This physiological response is regarded as part of an adaptive mechanism that may have evolved to conserve energy when nutrient availability is limited, and it is a remarkable example of the integration of nutrient and temperature homeostasis. The mechanisms that regulate it are not fully understood.

T_b is regulated centrally, primarily by the hypothalamic preoptic area (POA). This region contains temperature-sensitive neurons that are pivotal in sensing and regulating T_b (2–4). Among them, the warm-sensitive neurons (WSN) exert a tonic inhibition, directly or indirectly through the dorsomedial hypothalamus, on the raphe pallidus that can activate spinal sympathetic and somatic motor circuits to drive thermogenesis and heat dissipation (5). WSN are typically investigated for their role in regulating fever or response to peripheral and local changes in temperature. However, electrophysiological studies and molecular characterization demonstrated that these specialized cells also respond to nutrient signals (2, 6–10). Nutrient homeostasis is likewise regulated centrally and primarily in the hypothalamus (11). There, the paraventricular (PVN), the arcuate (ARC), and the lateral hypothalamic (LH) nuclei act as main modulators of feeding and satiety, and some of their specific neuropeptides can mediate hyperthermia during feeding (12).

We recently demonstrated that POA-WSN neurons express the insulin and the insulin-like growth factor 1 (IGF1) receptors (insulin receptor and IGF-1R), and that local administration of insulin or IGF1 in the POA induced a transient elevation of T_b (6, 9, 13). CR reduces T_b and lowers the level of insulin and IGF1 in the circulation and in the hypothalamus, where IGF1 can be synthesized in the POA, the PVN, the LH, and the supraoptic nucleus (14–21). Furthermore, the long-lived Ames, Snell, and growth hormone receptor null mice all have reduced IGF1 production and lowered T_b (22, 23). Taken together, these data suggested the possibility that the insulin/IGF1 signaling may contribute to the regulation of T_b during CR.

Here, we tested this hypothesis in the mouse by pharmacological inhibition and activation of IGF-1R, as well as its genetic disruption or dosage.

Results

Pharmacological Inhibition of IGF-1R Reduced T_b.

Administration of the selective IGF-1R inhibitor picropodophyllin (PPP) to ad libitum (AL)-fed Igf1r^+/+ mice diminished T_b in a dose-dependent manner (Fig. 1). PPP and vehicle were administered at time 0 at four different doses (2–40 mg/kg i.p.). Consistent with the short half-life of PPP (2–4 h) (24), the hypothermic effects were transient and animals resumed pretreatment values of T_b within 5 h of injection. The hypothermic profile was similar in both female and male mice for concentrations of PPP of 20 and 40 mg/kg, where temperature reached the lowest value of 35 °C and 31 °C, respectively. No significant (see SI Appendix, SI Statistical Analysis for details) effects were observed in either male or female mice at a dose of 2 and 10 mg/kg. Similar data were obtained with a distinct IGF-1R inhibitor, NVP-ADW742 (SI Appendix, Fig. S1).

Fig. 1.

Pharmacological inhibition of IGF-1R reduces T_b. (A) Temperature profile of female (F) and (B) male (M) Igf1r^+/+ mice after i.p. injection of different doses of the IGF-1R-specific inhibitor PPP. Injections were performed at t(0). n = 6 WT females and 5 WT males per condition. Arrows indicate time of injection. P < 0.0001 between vehicle and dose of PPP. ^§Vehicle vs. 20 mg/kg; ^#Vehicle vs. 40 mg/kg.

Genetic Deletion of IGF-1R Enhanced the Hypothermic Response to CR.

We next measured the effects of deletion of IGF-1R on T_b. Since constitutive full inactivation of IGF-1R signaling is lethal at birth, genetic deletion was induced in adult mice using Ubc-CreER^T2 transgene and homozygous IGF-1R^flox knock-in (UBIKOR) (25–27). Circadian T_b profiles of AL-fed UBIKOR and control mice were similar for both sexes (SI Appendix, Fig. S2 A and B). As could be expected (8), females showed slightly higher T_b compared with males (P < 0.05; RM-ANOVA for genotype; SI Appendix, Fig. S2 A and B). On 50% CR, female and male control mice lowered T_b similarly, although not identically, as females showed slightly lower values at some times of the day (SI Appendix, Fig. S2C; P < 0.05). Importantly, UBIKOR mice showed a significantly larger hypothermic response to CR than did control mice (Fig. 2; P < 0.001 for males and females). This response was identical in both sexes, as UBIKOR male and female mice showed overlapping circadian profiles and dropped their temperature 5–6 °C below that of controls, reaching a lowest T_b value of 26 °C (P < 0.001) (Fig. 2 A and B and SI Appendix, Fig. S2D). The possibility that the observed differences may be a result of changes in calories ingested was ruled out by demonstrating that UBIKOR and control mice consumed the same amount of food. All animals also showed equal calorie intake response on cessation of CR (SI Appendix, Fig. S3).

Fig. 2.

T_b in UBIKOR mice and controls. (A) Female and (B) male circadian temperature profiles showing that hypothermic response to 50% CR was more pronounced in mice with ubiquitous ablation of IGF-1R (UBIKOR) compared with control mice. (Lower) AL profiles. UBIKOR mice carry a Ubc-CreER^T2 transgene and homozygous IGF-1R^flox knock-in, controls are Ubc-CreER^T2 with wild-type IGF-1R alleles. Arrowheads indicate the 7 AM 50% CR meal. *P < 0.05; **P < 0.01; ***P < 0.001. n = 5 UBIKOR males, 6 control males, 7 UBIKOR females, 6 control females. Note that when we tested at intermediate points during the early daylight period (SI Appendix, Fig. S4; 10 AM), we found significantly low T_b also in AL-fed UBIKOR mice, similar to the findings in Igf1r^+/− females.

Genetic Dosage of IGF-1R Confers Sex- and Estrus-Specific Reduction of T_b in Mice Fed AL.

The effects that gene dosage of IGF-1R had on T_b were evaluated in mice carrying heterozygous null mutation (Igf1r^+/−) and their wild-type littermates (Igf1r^+/+), previously described to have reduced IGF-1R signaling and increased female life span (28). Continuous radiotelemetric recording revealed that AL-fed female but not male Igf1r^+/− mice displayed transient hypothermia (Fig. 3 A and B). Specifically, during the first part of the light cycle, female Igf1r^+/− mice reduced T_b to an average of 33 °C for a period lasting up to 2 h before resuming a profile that was similar to that of Igf1r^+/+ female mice. This hypothermic profile of female Igf1r^+/− mice differed in duration and/or amplitude, depending on the estrous phase, indicative of a possible hormonal contribution to its regulation (Fig. 3C). The transient hypothermic effect during AL feeding was also evidenced in UBIKOR mice when we screened their T_b specifically during the first part of the light cycle (SI Appendix, Fig. S4).

Fig. 3.

AL-fed female but not male Igf1r^+/− mice display transient hypothermia. (A) Temperature profile of female (F) and (B) male (M) Igf1r^+/− and Igf1r^+/+ mice fed AL during a period of 4 d of continuous recording. D, dark phase; L, light phase. (C) Detailed temperature profile of the areas showing hypothermia in A. The profile for females was adjusted to each day of estrous cycle, as indicated. Two-way RM-ANOVA followed by Bonferroni’s multiple comparisons correction test was performed to analyze differences in T_b recorded every 5 min. n = 5 animals per group. *P < 0.05 between genotypes.

The possibility these effects could be a result of inherently impaired capacity of thermogenesis, of differences in calorie intake or altered locomotor activity, were ruled out by demonstrating that fever response, food intake, and locomotor activity did not differ across genotypes and sexes (SI Appendix, Figs. S5–S7).

Female Mice with Genetic Dosage of IGF-1R Have Enhanced Hypothermic Response to CR.

Next, we compared the T_b profile of Igf1r^+/− and Igf1r^+/+ mice during CR. In our CR paradigm, animals were fed 100% of AL diet for 2 d, followed by 75% of the AL diet for 4 d and 50% thereafter. Female and male Igf1r^+/+ mice showed similar temperature profiles during CR, with circadian T_b reduction observed by day 4 of the 75% diet and gradually progressing to reach a lower average point of 33 °C by the fifth day of 50% CR (SI Appendix, Fig. S8). As observed in UBIKOR mice, on returning the CR animals to an AL diet, Igf1r^+/− and Igf1r^+/+ of both sexes engaged similarly in increased feeding and promptly resumed their original T_b profile within 24–48 h (SI Appendix, Fig. S9). However, female Igf1r^+/− mice displayed a faster and stronger hypothermia to CR than Igf1r^+/+ animals, which was evident, although not significant, starting on the second day of 50% CR, when average T_b reached 33 °C. By the fifth day of 50% CR, T_b of female Igf1r^+/− mice was as low as 23 °C compared with the 33 °C of female Igf1r^+/+ mice (F_1,8 = 7.861; P = 0.0231; Bonferroni’s P < 0.05) (Fig. 4A). In contrast, no significant (SI Appendix, SI Statistical Analysis for details) differences in T_b response to CR were observed between Igf1r^+/− and Igf1r^+/+ male mice (Fig. 4B). The possibility that the faster and larger hypothermic response could depend on differences in body size, on the amount of adipose tissue, or on its possible early exhaustion was ruled out by analyzing differences of T_b, using body weight as a covariate (SI Appendix, Fig. S10 and Table S1), with EchoMRI analysis of body composition showing no difference in the amount of fat across sexes and genotypes under AL or CR regimen (SI Appendix, Fig. S11 and Table S2).

Fig. 4.

Female (F), but not male (M), Igf1r^+/− mice display stronger hypothermic response to CR than their wild-type littermates. (A) Temperature profile of F and (B) M Igf1r^+/− and Igf1r^+/+ mice during a period of 19 d of CR, as indicated. n = females, 4 Igf1r^+/+ and 6 Igf1r^+/−; males, 5 Igf1r^+/+ and 7 Igf1r^+/−. *P < 0.05 between genotypes.

While body composition did not explain altered regulation of T_b, we could clearly demonstrate, using indirect calorimetry, that CR had marked effects on energy expenditure (SI Appendix, Fig. S12). Energy expenditure during CR was significantly lower in Igf1r^+/− than in Igf1r^+/+ mice (F_1,7 = 8.131; P = 0.025) (SI Appendix, Fig. S12 A and B). Consistent with what we observed for T_b, these differences were found only in female mice. The analysis also showed that the respiratory exchange ratio was similar across genotypes, indicating that Igf1r^+/− and Igf1r^+/+ mice did not differ in the way they used carbohydrates (SI Appendix, Fig. S12 C and D).

Central IGF1 Receptors Mediate the Hypothermic Response to CR.

Next, we performed a set of experiments to gain insights on the site at which IGF-1R regulates T_b. CR reduced the level of IGF1 in the circulation, as well as in the hypothalamus: serum IGF1 (AL, 600 ± 32 vs. CR, 240 ± 66 ng/mL; n = 11 AL, 7 CR; t(16) = 8.492; P < 0.0001) and hypothalamic level of IGF1 transcript [AL, 4.14 ± 2.30 vs. CR, 1.29 ± 0.37 AU; n = 12 AL, 12 CR; t(22) = 3.218; P = 0.004]. Thus, we first tested the effects of systemic or local administration of recombinant IGF1 on T_b in animals on CR or AL diets. We found that IGF1 fully prevented the hypothermic response to 50% CR when administered in the POA (Fig. 5 A and B). These effects were larger [10 ng: F_1,6 = 9.893 (P = 0.019); 100 ng: F_1,6 = 19.52 (P = 0.004); Bonferroni’s P < 0.01] in females compared with males, ruling out differences in body weight [t(9) = 0.214; P = 0.834]. In contrast, no significant effects [females: F_2,12 = 0.048 (P = 0.952); males: F_2,10 = 0.590 (P = 0.572)] were observed when IGF1 was injected intraperitoneally (Fig. 5 C and D). These data indicate that IGF1 acted centrally and not peripherally to regulate T_b. They also identify the POA as being a main site of action. This is consistent with our previous finding, that IGF1 had hyperthermic effects when applied to the POA, but not when injected into the dorsomedial hypothalamus or the raphe pallidus, the two other regions recognized to regulate T_b (13). The possibility that LH, PVN, and ARC, the three major hypothalamic nuclei that regulate feeding, could be sites for the effects of IGF1 on T_b was also assessed. Stereotactic application of 10 ng IGF1, previously shown to induce hyperthermia when injected in the POA (13), had no effects in the LH or the PVN and caused only a minor and short-lasting increase of T_b in the ARC of female, but not male, mice (SI Appendix, Fig. S13).

Fig. 5.

Central, but not peripheral, administration of IGF1 blocked T_b reduction during CR. Temperature profile of female (F) and male (M) Igf1r^+/+ mice on 50% CR diet after (A and B) POA or (C and D) i.p. injection of recombinant IGF1 or vehicle at t(0). Arrows indicate time of injection. n for POA injection: females, 4 vehicle, 3 IGF1 10 ng and 4 IGF1 100 ng; males, 4 vehicle, 5 IGF1 10 ng and 5 IGF1 100 ng. n for i.p. injection: females, 5 per treatment; males, 4 vehicle, 5 IGF1 2 μg and 4 IGF1 20 μg. **P < 0.05 between vehicle vs. 10 or 100 ng IGF1.

Finally, to further dissect IGF action on CR-induced T_b regulation in the central nervous system, we generated mice with IGF-1R knockout selectively in prosencephalic neurons (inIGF1RKO), but not in the periphery nor caudally to the POA/hypothalamus (Fig. 6 A–C), and submitted them to CR (SI Appendix, Fig. S14). We found that the targeted inhibition of IGF signaling in neurons located upstream of POA also produced significant hypothermia in CR. Surprisingly, this was the case in males (Fig. 6 D and E; P < 0.01), where the maximum difference in hypothermia was −3.6 °C, thus accounting for almost half the effect previously evidenced in ubiquitous IGF-1R KO (−8.3 °C; Fig. 2B). In contrast, T_b in female inIGF1RKO mice was not different from controls (Fig. 6 F and G). Collectively, these results suggested sex-dimorphic effects of IGF signaling depended not only on gene dosage but also on the neuronal target, the latter also including noncanonical thermoregulatory regions.

Fig. 6.

Effect of CR on T_b in neuron-specific IGF-1R knockout (inIGF1RKO). We used Tam-inducible CaMKIIα-CreER^T2 to inactivate floxed IGF-1R in adult mice. (A and B) Micrographs of sagittal and (C) coronal brain sections from adult Tam-induced CaMKIIα-CreER^T2;tdTomato reporter (52) mice. (A) KO specificity: red fluorescence reveals efficient recombination in cortex (Cx), olfactory bulb (OB), hippocampus (Hp), and caudate putamen (CPu). White arrowhead indicates bregma. (B and C) Absence of significant Cre-loxP recombination in POA. (B) Magnification from A showing no recombination in lateral preoptic area (LPO) and lateral hypothalamic area (LH). (C) Coronal view at bregma +0.25 demonstrating absence of recombination in medial preoptic area (MPA) and LPO. Diffuse red in MPA and LPO corresponds to afferents from CaMKIIα-positive regions. Other hypothalamic areas known to play a role in temperature regulation and showing no recombination in adjacent sections: AHC, anterior hypothalamic area, central part; AHP, anterior hypothalamic area, posterior part; VMH, ventromedial hypothalamic nucleus. Abbreviations: aca, anterior commissure; Cb, cerebellum; HDB, nucleus of the horizontal limb of the diagonal band; Hth, hypothalamus; LDB, lateral nucleus of the diagonal band; Mb, midbrain; Me, medulla; OCh, optic chiasm; Pn, pons; Th, thalamus; Tu, olfactory tubercle; VP, ventral pallidum. (D–G) Five cycles of T_b recorded during the last days of 50% CR. Male inIGF1RKO (D and E) showed consistent hypothermic episodes (RM-ANOVA for time, P < 0.001; for genotype, P < 0.05) in contrast to females (F and G) (RM-ANOVA for time, P < 0.001; for genotype, P = 0.29, followed by post hoc Bonferroni’s correction test). Arrowheads indicate the 50% CR meal at 7 AM. (E and G) Circadian profiles from D and F aligned on Zeitgeber time (dark-to-light transition at 8 AM). Significant CR-induced hypothermia occurred at 8 AM in inIGF1RKO males (RM-ANOVA for genotype P < 0.05, followed by Bonferroni’s correction), but not in females. On refeeding, hypothermia disappeared (T_b at 8 AM 37.3 ± 0.1 in control females vs. 37.4 ± 0.1 °C in inIGF1RKO females; P = 0.62; control males, 36.7 ± 0.1 vs. inIGF1RKO males, 36.9 ± 0.1 °C, P = 0.23; Mann-Whitney U test). Data are mean ± SEM. Males: inIGF1RKO, n = 5, and controls, n = 5; females: inIGF1RKO, n = 8, and controls, n = 8. *P < 0.05; **P < 0.01; ***P < 0.001.

Discussion

The data presented here demonstrate that IGF-1R regulates the hypothermic response to CR, and identify this receptor as one of the mediators of the integration of nutrient and temperature homeostasis.

The involvement of IGF-1R was demonstrated by a combination of its pharmacological inhibition or activation and by the data obtained with three different and complementary animal models: UBIKOR, Igf1r^+/−, and inIGF1RKO mice representing ubiquitous adult deletion, reduced expression, or prosencephalic neuron-specific ablation of IGF-1R, respectively. Pharmacology and the genetic deletion of IGF-1R in UBIKOR and inIGF1RKO mice were obtained in adult animals, eliminating the possibility that the phenotype observed could be the result of developmental defects. The possibility that the effects observed could be due to differences in body weight or body composition was also ruled out. Using these experimental approaches, we found that IGF-1R regulated T_b differently in males and females, depending on both the site of action and genetic dosage.

Pharmacological manipulation showed that the effects by which IGF-1R effectively influenced T_b are not peripheral, but central. These experiments also identify the POA as one of the regions in which IGF-1R regulated T_b. This is consistent with our previous demonstration that insulin and IGF1 induced hyperthermia by direct inhibition of POA WSN that regulate temperature homeostasis, and expressed the insulin receptor and the IGF-1R (6, 9, 13). In those studies, we also showed that the dorsomedial hypothalamus and the raphe pallidus, the two other canonical thermoregulatory regions and the sites of POA neuronal projections, were insensitive to the action of insulin or IGF1. Here we also found that IGF1 did not elevate T_b when administered in the LH or the PVN, and had minor effects in the ARC, although only in female mice. In addition, results from prosencephalic neuron-specific inIGF1RKO mice suggested that the implication of IGF signaling in T_b control is not limited to the POA, and that the forebrain neuronal circuitry may play a hitherto unknown role in modulating hypothermic responses.

IGF1 can reach the brain from the periphery through neuronal transport or can be produced locally, and CR lowers both its peripheral and central levels (29, 30). The IGF-1R inhibitors used here can cross the blood–brain barrier and act peripherally and centrally. However, i.p. administration of recombinant IGF1 failed to induce hyperthermia (13) or alter the hypothermic response of CR. Thus, the pharmacokinetics by which peripheral IGF1 can reach the brain may not be compatible with its role in influencing T_b. Instead, it is more likely that the system responds to the fluctuation of local IGF1 that can be produced by several nuclei, including those involved in the regulation of feeding (PVN, LH, and supraoptic nucleus) or of T_b (POA) (18).

Comparing the T_b responses to CR observed in UBIKOR and Igf1r^+/− mice indicated that if blocking IGF-1R signaling affected both sexes equally, its reduction was effective only in females. Female Igf1r^+/− mice also had a transient but significant hypothermic phase when fed AL that was not seen in males. Furthermore, although the increase in T_b caused by IGF1 injection in the ARC was modest, it was only observed in females. These observations are consistent with studies on growth hormone receptor KO mice in which lowered IGF1 was associated with a larger T_b reduction in females (22, 23). Thus, female mice appear to be more susceptible to IGF-1R-dependent hypothermia and may respond to lowering of IGF-1R stronger and/or faster than males. The exact mechanisms accounting for the sex-specific differences remain to be elucidated, but they may be central and include the interplay between IGF-1R and estrogen receptor signaling that was reported at least in the POA (31–33). More surprising was the finding that the inIGF1RKO mice deficient for IGF-1R in the prosencephalic brain (but not in the POA and not in other hypothalamic nuclei) still showed enhanced hypothermic response to CR that was unique to the males. These findings could extend the current view that the POA contains the most frontal structures integrating and controlling thermoregulation in the central nervous system (34). Several areas in POA, for instance, ventrolateral preoptic nucleus (35), receive afferents from forebrain circuitry; namely, through the lateral septal nucleus, but also directly from cortical areas. Using a fluorescent tdTomato reporter transgene to trace IGF-1R ablation, we confirmed experimentally that resident neurons in the POA receive abundant afferents from the forebrain, but are not receptor-deficient themselves in inIGF1RKO mice (SI Appendix, Fig. S15). Together with the known sensitivity of forebrain neurons to estrogens and other sex steroids, this potentially connects effects of IGF-1R knockout in prosencephalic neurons with dimorphism of thermoregulation, as observed in this study. These results added complexity to the sex-dimorphic effect of central IGF signaling on T_b, requiring further investigations. In aggregate, the findings presented here support the idea that central hypothermic effects of IGF-1R inhibition can be ascribed to neurons in hypothalamus, and POA in particular, but not exclusively so.

IGF-1R participates in several biological functions regulating growth, metabolism, longevity, and aging (28, 36). In poikilotherms, IGF-1R also regulates thermotolerance (37). The data collected here demonstrate that IGF-1R also contributes to the regulation of energy homeostasis. Lowering of IGF-1R signaling during CR reduced both T_b and energy expenditure. This is similar to what is observed in dieting and opposed to what occurs in cachexia, in which reduced calorie intake is coupled with an elevation of T_b and energy expenditure. An important difference between these two conditions is that while dieting, as well as CR, is characterized by hunger, appetite is lost in cachexia. Thus, determining how the nuclei and the molecules that regulate appetite affect IGF-1R signaling may provide important information on the mechanisms by which they can influence energy homeostasis.

Our data also demonstrate that reduction of IGF-1R signaling is one of the mechanisms by which CR reduces T_b. This is particularly meaningful when considering the mechanisms of aging, since CR, lowering of IGF-1R signaling, and reduction of T_b all promote longevity and retard aging independently (25, 38–42). Temperature reduction prolongs life span in ectothermic as well as endothermic animals (38, 40, 43, 44). It is now becoming clear these effects may be mediated not only by thermodynamic effects but also by the activation of specific molecular pathways that respond to hypothermia (45–47). CR promotes longevity across species (reviewed in refs. 41 and 42) and lowers T_b in endothermic animals (48–50). The finding that lowering T_b alone could prolong the life span of mice fed AL demonstrated that one of the mechanisms by which CR promotes longevity is by reducing T_b (38). IGF-1R signaling, first found to regulate longevity in Caenorhabditis elegans, also affects life span in endothermic animals (28, 40, 51). Since several groups reported that CR lowered IGF1 (15–17), the reduction of IGF-1R signaling is considered one of the mechanisms by which CR prolonged life span. Here, we provide an important addition to this paradigm by showing that suppression of IGF-1R signaling exacerbated CR-induced T_b reduction. This demonstrates that CR, IGF-1R signaling, and T_b, the three main determinants of health, longevity, and aging, are components of the same pathway that regulates energy homeostasis. Our findings also identify lowered T_b as a common determinant of both CR and IGF1 signaling. Consistent with the role of T_b in longevity, only female Igf1r^+/− mice showed an increased life span (28, 36). We speculate that one of the reasons only female Igf1r^+/− mice are long lived may be the fact that they display hypothermic profiles even when fed AL. With this respect, it will be interesting to investigate whether differences in T_b depending on the IGF-1R signaling can also be found in other long-lived rodent models, in which increased longevity is unique or larger in females.

Materials and Methods

Mice and Husbandry.

All procedures were approved by the TSRI IACUC and by Comité d’éthique en expérimentation animale (Paris, #02214.02), respectively. UBIKOR mice were hemizygous for Ubc-CreER^T2 transgene (26) and homozygous for IGF-1Rflox knock-in mutation. Control mice were hemizygous for Ubc-CreER^T2 and carried wild-type IGF-1R alleles. UBIKOR and control mice were on C57BL/6 genetic background. Ubiquitous IGF-1R inactivation was induced at age 6 mo by administrating i.p. tamoxifen (Tam; Sigma; 84 µg/d/g body weight for 5 d) to all mice, as described (25). inIGF1RKO (CaMKIIα-CreER^T2;IGF-1R^flox/flox) mice were generated as described (25), and gene knockout induced at 3 mo of age by Tam administration. CR was performed and T_b analyzed at 12 mo. To visualize Cre-loxP recombination, we used the tdTomato reporter transgene (52). Mice were singly caged during CR, T_b measurements, and refeeding at 22 ± 1 °C room temperature and 12:12 h light:dark cycle (lights on 8:00 AM). Animals were fed AL with irradiated LASQCdiet Rod18 (18.9% protein, 53.1% carbohydrates, 5.3% fat, 3.9% crude fiber, and 7.0% ash; LASvendi) or restricted during the CR experiments.

Igf1r^+/− and Igf1r^+/+ littermates were generated by crossing heterozygous males with 129/SvPas wild-type females in a 129/SvPas genetic background (28). Animals were singly caged in standard 12:12 h light:dark (lights on 9:00 AM; Zeitgeber time 0 or ZT0) at a controlled ambient temperature of 22 ± 0.5 °C. Water was available AL during CR experiments. Food (Lab Diet 5053 Irradiated Pico Lab containing 3.41 kcal/g, 20.0% protein, 52.9% carbohydrates, 10.6% fat, 4.7% crude fiber, and 6.1% ash) was provided AL or restricted during the CR experiments, as specified.

Treatments.

IGF-1R-specific inhibitor PPP (Selleckchem) was dissolved in 10% DMSO and administered i.p. at different doses (2, 10, 20, 40 mg/kg) at ZT0, when animals go to sleep. NVP-ADW742 was reconstituted in 25 mM tartaric acid, which was also used as vehicle. NVP-ADW742 and vehicle were administered i.p. at time (0) at three different doses. Recombinant IGF1 (R&D systems) was reconstituted in sterile 0.9% saline. IGF1 was administered 5–6 h before the hypothermic onset, i.p. or into the brain through a previously placed cannula. We used a 7-mm, 26-gauge guide cannula and a 7.25-mm injector, 33-gauge internal cannula. For administration into the brain, the working solution was transferred to artificial cerebrospinal fluid. The guide cannula was implanted at the following stereotactic coordinates: preoptic area, anterior to bregma, 0.40 mm; ventral, 4.75 mm from the skull; injector, 7.25 mm; final ventral coordinate, 5.00 mm. Arcuate nucleus, posterior to bregma, 1.25 mm; ventral, 5.65 mm from the skull; injector, 7.25 mm; final ventral coordinate, 5.90 mm. Lateral hypothalamus, posterior to bregma, 2.06 mm; bilateral from midline, 0.75 mm; ventral, 4.75 mm from the skull; injector, 7.25 mm; final ventral coordinate, 5.00 mm. Paraventricular hypothalamus, posterior to bregma, 3.00 mm; bilateral from midline, 0.25 mm; angle, 25°; ventral, 4.08 mm from the skull; injector, 8.10 mm; final ventral coordinate, 5.18 mm.

Bacterial lipopolysaccharides (0127:B8; Sigma) reconstituted in sterile 0.9% saline were administered i.p., using 20–30 μL per mouse at 100 µg/kg.

CR was carried out providing a percentage of the amount of food consumed during AL diet, which was calculated as an average during a period of 4 d. Animals were given 100% of the AL diet for 2 d to habituate to daily rations of food, followed by 75% of AL for 4 d and 50% thereafter. Food was provided daily at a 1-h window before lights were turned off (ZT12). All animals ate the entire amount of food provided during CR experiments. Since T_b reduction during CR is progressive over several days, possible T_b heterogeneity across experiments reflects the day of CR regimen at which the measurements were carried out.

Temperature Recordings.

T_b was monitored using a rectal temperature probe in UBIKOR and inIGF1RKO, or by radiotelemetry in all other experiments. Rectal measurements were carried out using RET-3 probes and thermo recorder HH506A (Omega) and were completed within 40 s after first moving the cage to avoid recording any thermogenesis due to handling (53). Radiotelemetry was performed using a transmitter (TA10TA-F10; Data Sciences, Inc.) surgically implanted into the peritoneal cavity, as described (38). Animals were allowed a recovery time of 14 d postsurgery and 2 d of habituation to the experimental environment before beginning experiments. Data were recorded by placing a cage containing an animal implanted with a radio transmitter on a receiver plate (RPC-1; DataScience, Inc.). Data collection and offline analysis were performed using the DATAQUEST A.R.T. software (DataScience, Inc.).

Estrous Cycle Determination.

Estrous cycle was determined by histological analysis of vaginal smear by lavage performed in the first 3 h of the dark cycle, using the modified version of the procedure (8).

IGF1 Determination.

Serum levels of IGF1 were determined using the Quantikine ELISA (R&D Systems). Levels of IGF1 transcript were determined by real time RT-PCR, using Taq polymerase enzyme (PowerUP Sybr Master Mix, #A25777; Life Technologies), along with the following primers from Qiagen: RT2 qPCR Primer Assay, Beta Actin (PPM02945B-200), IGF1 (PPM03387F-200).

Body Composition.

Body composition during CR was measured during the first hours of the light period, using the Echo3-in-1 NMR analyzer (Echo Medical).

Energy Expenditure.

Energy expenditure was determined by using a computer-controlled open-circuit system (Oxymax System) that is part of an integrated Comprehensive Lab Animal Monitoring System (Columbus Instruments), as previously described (54).

Statistical Analysis.

GraphPad Prism 7 and SPSS statistics were used for data analysis. Two-way repeated measure ANOVA followed by Bonferroni’s multiple comparison correction was used to analyze difference between genotypes and sexes on T_b, locomotor activity and body composition on multiple collected points during AL, CR regimen, and treatments (lipopolysaccharides, IGF-1R inhibitors, and recombinant IGF1 administration). Energy intake, energy expenditure, and CR were covariate normalized by body weight when stated in the text. Body composition differences within same group were analyzed by paired two-tailed Student’s t test. All statistical analysis is detailed in SI Appendix.