Calorie restriction-induced insulin sensitivity is mediated by adipose mTORC2 and not required for lifespan extension

Adipose mTORC2 is required for CR to improve insulin sensitivity

We next tested our hypothesis that mice lacking adipose mTORC2 would be resistant to the metabolic benefits of CR by placing 10-week-old WT and AQ-RKO mice of both sexes on either ad libitum (AL) or CR diets. After 8 weeks, which is sufficient for CR to promote glucose tolerance and insulin sensitivity in C57BL/6J mice (Solon-Biet et al., 2015), we assessed the effects of CR on metabolic phenotypes. Both AQ-RKO mice and their WT littermates fed a CR diet had improved glucose tolerance compared to their AL fed counterparts (Figures 2A and ​and2B).2B). However, the improvement of insulin sensitivity induced by CR feeding that we observed in both male and female WT mice was completely blocked by deletion of adipose Rictor (Figures 2C and ​and2D,2D, Figures S1A and S1B). In accordance with these results, we observed increased phosphorylation of AKT T308 and S473 in the adipose tissue of WT CR-fed mice relative to their WT AL-fed littermates following either acute stimulation by insulin or prolonged stimulation by refeeding; while in the absence of adipose Rictor, CR had no effect on the phosphorylation of AKT in adipose tissue in response to these stimuli (Figures S1C–S1F and S2A–B). In contrast, we observed no effect of genotype on AKT phosphorylation in skeletal muscle (Figures S2C and S2D).

Open in a separate window

Figure 2.

Adipose mTORC2 is required for the CR-induced increase in insulin sensitivity.

(A and B) Glucose tolerance test of male (A) and female (B) WT and AQ-RKO mice after 9 weeks on indicated dietary regimens (n=5–14 mice/group). (C and D) Insulin tolerance test of male (C) and female (D) WT and AQ-RKO mice after 10 weeks on indicated dietary regimens (n=9–19 mice/group). (E and F) Plasma insulin in male (E) and female (F) mice after an overnight fast followed by refeeding for 3 hours (n=5–8/group). (G) Pyruvate tolerance test of male WT and AQ-RKO mice after 11 weeks on indicated dietary regimens (n=14–19/group). (A-G) Statistics for the overall effects of diet, genotype (GT), and the interaction represent the p-value from a two-way ANOVA; * = p < 0.05, from a Holm-Sidak’s post-test examining the effect of parameters identified as significant in the two-way ANOVA. Error bars represent SEM.

Consistent with the improved glucose tolerance of both genotypes of mice on a CR diet, and previous reports of fasting hyperinsulinemia in mice lacking adipose Rictor (Cybulski et al., 2009; Tang et al., 2016), we observed hyperinsulinemia following refeeding in AQ-RKO mice of both sexes fed either AL or CR; curiously, CR potentiated the hyperinsulinemia we observed in AL-fed mice (Figures 2E and ​and2F).2F). Consistent with the above results, we found that CR improved pyruvate tolerance equally well in both WT and AQ-RKO mice, indicating effective suppression of hepatic gluconeogenesis by insulin (Figure 2G). In accordance with these results, we observed no effect of genotype on AKT phosphorylation in the liver (Figures S2E and S2F).

Loss of adipose Rictor impairs the metabolic adaptation to a CR diet

The effects of CR on insulin sensitivity likely result in part from reduced adiposity. In addition, CR substantially alters whole-body lipid metabolism, in part by reprogramming WAT; CR upregulates lipogenesis in WAT upon refeeding (Bruss et al., 2010) which may contribute to the beneficial effects of CR on systemic insulin sensitivity. We observed that the loss of adipose mTORC2 did not substantially impact the effect of CR on adiposity; CR was equally effective at preventing the accretion of fat mass in both WT and AQ-RKO mice (Figures 3A and ​and3B).3B). The body weight and lean mass of WT and AQ-RKO mice also responded similarly to CR (Figures 3C–3F). Thus, CR was equally effective at reducing adiposity in WT and AQ-RKO mice (Figures S3G and S3H).

Open in a separate window

Figure 3.

Loss of adipose mTORC2 does not affect the response of body composition to CR.

Fat mass, weight, lean mass, and adiposity of male (A, C, E, and G) and female (B, D, F, and H) WT and AQ-RKO mice fed either AL or CR diets starting at 10 weeks of age (n=12–16/group; Statistics for the overall effects of diet, genotype (GT), and the interaction represent the p-value from a repeated measures two-way ANOVA conducted separately for each genotype). Error bars represent SEM.

We next examined the role of mTORC2 in the upregulation of lipogenesis by CR following refeeding. WT and AQ-RKO mice were placed in metabolic chambers, allowing us to assess substrate utilization by examining the respiratory exchange ratio (RER). RER is calculated using the ratio of O₂ consumed and CO₂ produced, and approaches 1.0 when carbohydrates are being primarily utilized for energy production and approaches 0.7 when lipids are the predominant energy sources (Hasek et al., 2010). An RER > 1.0 reflects the utilization of carbohydrates for active de novo lipogenesis (Bruss et al., 2010). Consistent with previous reports that CR-fed mice engage in rapid lipogenesis following refeeding and sustain themselves the rest of the day via the utilization of these stored lipids (Bruss et al., 2010), we observed that RER rapidly rose above 1.0 in WT CR mice following refeeding, and approximately 6 hours later fell below 0.8 (Figures 4A and ​and4B).4B). In contrast, AQ-RKO CR-fed mice of both sexes had a blunted lipogenic response to refeeding, and AQ-RKO CR male mice had an intermediate RER for the majority of the dark cycle, suggesting that these animals relied on a mix of carbohydrate and lipid stores (Figures 4A and ​and4B4B).

Open in a separate window

Figure 4.

Loss of adipose mTORC2 attenuates CR-mediated reprogramming of adipose lipid metabolism.

(A and B) Respiratory exchange ratio (RER) of male (A) and female (B) WT and AQ-RKO mice after 7-week AL or CR regimens (n = 4–12/group; statistics for the overall effects of genotype (GT), time and the interaction represent the p-value from a repeated measures two-way ANOVA conducted on WT CR and AQ-RKO CR mice; * = p < 0.05, from a Holm-Sidak’s post-test; dark (shaded) and light phases were analyzed separately). (C and D) Expression of Chrebp-β and related lipogenic genes in the iWAT of male (C) and female (D) WT and AQ-RKO mice after 12 weeks on the indicated diets (n = 6/group; statistics for the overall effects of diet, genotype (GT), and the interaction represent the p-value from a two-way ANOVA for each gene; * = p < 0.05, from Holm-Sidak’s post-tests conducted for the effect of diet and genotype). (E and F) Plasma triglycerides (E) and free fatty acids (F) in WT and AQ-RKO mice after 12 weeks on the indicated diets (n=5–8/group; statistics for the overall effects of diet, genotype (GT), and the interaction represent the p-value from a two-way ANOVA conducted separately for each sex; * = p < 0.05, from Holm-Sidak’s post-tests conducted for the effect of diet and genotype). Error bars represent SEM.

To verify our model that AQ-RKO mice fed a CR diet had a reduced ability to induce lipogenesis following refeeding, we sacrificed mice 3 hours after refeeding and examined lipogenic gene expression in inguinal WAT (iWAT). In agreement with our RER data, we observed that CR robustly induced the expression of many genes required for de novo lipogenesis, including ATP-citrate lyase (Acly), acetyl-CoA carboxylase (Acc), fatty acid synthase (Fasn), and elongation of very long chain fatty acids 6 (Elovl6), in WT mice (Figures 4C and ​and4D).4D). However, CR had a substantially weaker impact on the expression of these lipogenic genes in AQ-RKO mice (Figures 4C and ​and4D4D).

De novo lipogenesis is regulated by several transcriptional factors including carbohydrate-responsive element-binding proteins (ChREBPs) (Uyeda and Repa, 2006) and sterol regulatory element-binding proteins (SREBPs) (Horton et al., 2002). In agreement with a recent report implicating adipose mTORC2 in the control of ChREBP-β-dependent de novo lipogenesis (Tang et al., 2016), we observed that CR induced Chrebp-β strongly in the iWAT of CR fed WT mice, but to a much lesser degree in mice lacking adipose mTORC2 (Figures 4C and ​and4D).4D). Other transcriptional factors important for lipogenesis in adipose tissue such as Chrebp-a and Srebp1-c were also downregulated in AL-fed AQ-RKO mice; however, deletion of adipose Rictor attenuated the ability of CR to upregulate these transcriptional factors only in females (Figure S3A). While adipose tissue is the major site of lipogenesis in CR-fed mice (Bruss et al., 2010), the liver is also an important site for de novo lipogenesis. To assess possible indirect effects of adipose Rictor deletion on hepatic lipogenesis, we examined the expression of a panel of key lipogenic genes in this tissue. Although CR robustly increased the expression of almost of all these genes, we observed no effect of genotype on lipogenic gene expression on either AL or CR diets (Figure S3B).

CR activates thermogenesis as a mechanism to defend body temperature (Fabbiano et al., 2016). We found that WT males, but not WT females, fed a CR diet had increased expression of uncoupled protein 1 (Ucp1) mRNA as well as increased expression of UCP1 protein in the iWAT; however, neither male nor female AQ-RKO mice had increased expression of UCP1 at either the mRNA or protein level in response to CR (Figures S3C–S3F). The expression of two other key thermogenic genes, cell death-inducing DNA fragmentation factor alpha-like effector A (Cidea) and dlongation of very long chain fatty acids protein 3 (Elovl3), were significantly upregulated by CR in the iWAT of WT mice of both sexes, but were not induced by CR in AQ-RKO mice (Figures S3C and S3D). Despite these significant differences at the molecular level, the morphology of iWAT from WT and AQ-RKO mice on either diet was very similar (Figure S4A). Furthermore, CR similarly remodeled the BAT of both WT and AQ-RKO mice (Figure S4B) at the histological level. As in the iWAT, the expression of UCP1 was increased by CR in the BAT of WT male mice, but not in AQ-RKO mice (Figure S4C).

We reasoned that as adipose Rictor deletion impacts the effects of CR on both substrate utilization and adipose lipogenesis, AQ-RKO mice might have altered levels of circulating lipids following refeeding. We observed that loss of adipose Rictor increased circulating levels of triglycerides in both males and females, irrespective of diet, and CR significantly reduced the plasma triglycerides levels of both WT and AQ-RKO males (Figure 4E). Plasma free fatty acids were significantly decreased by CR in female WT mice but not in female AQ-RKO mice; in males, plasma free fatty acids were not significantly affected by genotype, but CR diet showed a strong trend to decrease plasma free fatty acids (Figure 4F). Furthermore, female AQ-RKO mice placed on a CR diet had significantly higher plasma cholesterol than WT CR-fed females (Figure S4D). Finally, we assessed lipid deposition in the liver and muscle by histology; we found that CR promoted a similar reduction in lipid content in the liver of both WT and AQ-RKO mice, and that no group had significant lipid deposition in the muscle (Figures S4E and S4F).

Longitudinal assessment of the role of adipose mTORC2 in body composition and glucose homeostasis

The importance of adipose mTORC2 in the metabolic response to CR suggested to us that adipose mTORC2 might play a key role in regulating the many health benefits of long-term CR. To test this, we initiated a survival study in which 10-week-old WT and AQ-RKO mice were placed on either AL or CR diets; we then longitudinally assessed the effects of CR on weight and body composition, glycemic control, healthspan and lifespan.

As we anticipated, both WT and AQ-RKO mice of either sex fed a CR diet weighed substantially less than AL-fed mice throughout their life (Figures S5A and S5B), and were leaner with reduced levels of both fat and lean mass (Figures 5A–5D). However, we did observe significant differences in the body composition of AL-fed mice with age; in particular, male AQ-RKO AL-fed mice had significantly less fat mass than their WT littermates, particularly at 12 and 18 months of age (Figure 5A). Both male and female AL-fed AQ-RKO mice had greater lean mass throughout life than their AL-fed WT littermates, an effect that was statistically significant in females from 6–18 months of age (Figures 5C and ​and5D).5D). The overall effect of Rictor loss was reduced adiposity over the course of the lifespan for AL-fed males, but not for AL-fed females (Figures 5E and ​and5F5F).

Open in a separate window

Figure 5.

Longitudinal assessment of body composition of WT and AQ-RKO mice fed AL or CR diet.

(A-F) Longitudinal body composition analysis of WT and AQ-RKO mice. (A and B) Fat mass of male (A) and female (B) mice. (C and D) Lean mass of male (C) and female (D) mice. (A-D) Numbers vary month-by-month; maximum n = 24–39/group. (E and F) Percentage of body fat mass and lean mass of male (E) and female (F) mice at 2, 12 and 24 months of age (n = 13–28 mice/group). Statistics for the overall effects of genotype (GT), age, and the interaction represent the p-value from a repeated measures two-way ANOVA conducted for AL fed mice; * = p < 0.05, from Holm-Sidak’s post-test comparing WT AL and AQ-RKO AL fed mice at each age. Error bars represent SEM.

During the course of the study, we assessed glucose homeostasis as the mice aged. In agreement with our initial findings, and studies conducted by Tang and colleagues in young mice (Tang et al., 2016) we found that AQ-RKO mice fed an AL diet had normal glucose tolerance even at advanced ages; indeed, there was a trend towards improved glucose tolerance of 8 month old AQ-RKO male mice relative to the their WT littermates (Figures S5C–S5F). This glucose tolerance was observed despite the lifelong insulin resistance of AQ-RKO mice of both sexes relative to their WT littermates (Figures 6A–6D). Consistent with these results and previous reports (Tang et al., 2016), we observed increased fasting hyperinsulinemia and hypersecretion of insulin in response to glucose stimulation in AL-fed AQ-RKO mice (Figures S5G and S5H).

Open in a separate window

Figure 6.

Deletion of adipose Rictor blocks calorie restriction from improving organismal insulin sensitivity.

Insulin tolerance tests were performed on (A and B) 9-month-old male (A) and female (B) WT and AQ-RKO mice as well as (C and D) 25 month-old male (C) and female (D) WT and AQ-RKO mice. n = 6–13/group; statistics for the overall effects of diet, genotype (GT), and the interaction represent the p-value from a two-way ANOVA for each measurement; * = p < 0.05, Tukey post-test. Error bars represent SEM.

As we observed in our short-term study, while WT mice placed on a CR diet showed significantly improved insulin sensitivity at 9 months of age relative to WT AL-fed mice, a CR diet did not improve the insulin sensitivity of AQ-RKO mice at this age (Figures 6A and ​and6B,6B, Figures S6A and S6B). At 2 years of age, although CR improved the insulin sensitivity of AQ-RKO mice relative to AL-fed mice of the same genotype, CR-fed AQ-RKO mice remained significantly more insulin resistant than CR-fed WT mice (Figures 6C and ​and6D,6D, Figures S6C and S6D). Critically, at no age were AQ-RKO mice fed a CR diet more insulin sensitive than age-matched WT AL-fed mice.

In vivo tests

For glucose and pyruvate tolerance tests, food was withheld overnight from ad libitum fed animals for 16 hours, and mice were then injected intraperitoneally with glucose (1 g/kg) or pyruvate (2 g/kg) as desribed previously (Yu et al., 2018). The insulin tolerance test desribed in Figure 1 was performed by fasting mice at 5 pm and injecting 0.75 U/kg of insulin intraperitoneally around 9 am the next morning. As CR-fed WT mice are extremely insulin sensitive and this procedure leads to extremely low blood glucose values during an ITT, all other insulin tolerance tests (except for the ones performed in 25-month-old mice, Figures 6C, ​,6D,6D, S6C, S6D) were performed by fasting mice at 6 am in the morning and then injecting 0.5 U/kg of insulin (0.5U/kg) intraperitoneally at 10 am. ITTs in 25 month-old mice were performed using 0.75 U/kg as the response of all groups of mice to 0.5 U/kg was limited. In all ITTs, WT CR-fed animals that experienced hypoglycemia and needed to be rescued by adminsitration of glucose were not included in the presented results. A Bayer Contour blood glucose meter and test strips were used to measure blood glucose. For body composition analysis, lean and fat mass were measured using an EchoMRI 3-in-1 Body Composition Analyzer (EchoMRI™, Houston, TX, USA) according to the manufacturer’s procedures. To assess RER, we utilized an Oxymax/CLAMS metabolic chamber system (Columbus Instruments, Columbus, OH, USA).

Tissue harvest

Mice used for the short-term study were sacrificed in either in the fasted state, refed state or insulin-stimulated state after 12 weeks of feeding AL or CR. Mice sacrificed in the fasted state were fasted overnight starting at 5 pm and then sacrificed between 9 and 11 am the next morning. Mice sacrificed in the refed state were fasted at 5 pm for overnight and food was provided to both AL and CR mice at 8 am the next morning; the mice were then sacrificed approximately 3 hours later. Mice sacrificed in the insulin-stimulated state were fasted overnight for approximately 16 hours starting at 5 pm and the next morning were then injected I.P. with 0.75 U/kg of insulin 15 min prior to sacrifice.

Plasma hormone and lipid measurement

Plasma IGF1 and insulin were quantified using a mouse IGF-1 ELISA kit (80574) and ultra-sensitive mouse insulin ELISA kit (90080), respectively, from Crystal Chem (Elk Grove Village, IL, USA). Plasma triglyceride and free fatty acid were assessed using triglyceride colorimetric assay kit (10010303) and free fatty acid fluorometric assay kit (700310) from Cayman Chemical (Cayman Chemical, Ann Arbor, MI, USA), respectively. Plasma cholesterol was assessed using the stanbio cholesterol liquicolor kit (1010225) from Stanbio Laboratory (Boerne, TX, USA).

Quantitative PCR

RNA was extracted from liver or adipose tissue using Trireagent (Sigma-Aldrich, St. Louis, MO, USA). Briefly, the concentration and purity of RNA were determined by absorbance at 260/280 nm using Nanodrop (Thermo Fisher, Waltham, MA, USA), and 1 μg of RNA was used to generate cDNA using Superscript III (Invitrogen, Carlsbad, CA, USA). Oligo dT primers and primers for real-time PCR were obtained from Integrated DNA Technologies (Coralville, IA, USA); primer sequences are listed in Table S1. Reactions were run on an Applied Biosystems StepOne Plus machine (Applied Biosystems, Foster City, CA, USA) with Sybr Green PCR Master Mix (Invitrogen, Carlsbad, CA, USA) as described previously (Cummings et al., 2018). β-actin was used to normalize the results from gene-specific reactions.

Immunoblotting

Tissue samples from liver and muscle were lysed in cold RIPA buffer supplemented with phosphatase inhibitor and protease inhibitor cocktail tablets (Thermo Fisher Scientific, Waltham, MA, USA) as previously described (Baar et al., 2016) using a FastPrep 24 (M.P. Biomedicals, Santa Ana, CA, USA) with bead-beating tubes (16466–042) from (VWR, Radnor, PA, USA) and zirconium ceramic oxide bulk beads (15340159) from (Thermo Fisher Scientific, Waltham, MA, USA). Protein lysates were then centrifuged at 13,300 rpm for 10 min and the supernatant was collected. For WAT lysates, protein lysates were then frozen overnight at −80°C, thawed, and the aqueous non-lipids were collected using a needle and syringe; this supplemental step to remove lipids was repeated twice as previously described (Baar et al., 2016). Protein concentration was determined by Bradford (Pierce Biotechnology, Waltham, MA, USA). 20 μg protein was separated by SDS–PAGE (sodium dodecyl sulfate–polyacrylamide gel electrophoresis) on 8%, 10%, or 16% resolving gels (Thermo Fisher Scientific, Waltham, MA, USA) and transferred to PVDF membrane (EMD Millipore, Burlington, MA, USA). Rictor (2140), AKT (9272), HSP90 (4877), and pS473-AKT (4060) were purchased from Cell Signaling Technologies. Imaging was performed using a GE ImageQuant LAS 4000 imaging station (GE Healthcare, Chicago, IL, USA). Quantification was performed by densitometry using NIH ImageJ software.

Lifespan and necropsy

All mice for the lifespan study were bred by crossing mice homozygous for a conditional allele of Rictor (Rictor^L/L) with mice homozygous for a conditional allele of Rictor and hemizygous for Adipoq-Cre (Adipoq-Cre; Rictor^L/L). Mice were co-housed (grouped by sex and dietary regimen) in SPF housing. Mice were inspected daily and were euthanized for humane reasons if moribund, if the mice developed other specified problems (e.g. excessive tumor burden), or upon the recommendation of the facility veterinarian. Mice found dead were noted at each daily inspection and saved in a refrigerator for necropsy, during which the presence or absence of cancer was recorded. Mice that died during the adaptation to a CR diet, or as a result of the conduct of their cagemates, were excluded from the lifespan analysis. Mice that died within a one week period following the conduct of an in vivo procedure, as a result of husbandry errors, or due to anal or uterine prolapse prior to 20 months of age were included and censored as of the date of death. A complete list of mice for the lifespan study is in Table S2.

We would like to thank all members of the Lamming lab, particularly Dr. Cara Green for critical reading of the manuscript, and our colleagues at UW-Madison and beyond for their thoughtful input. We thank A. Broman of the UW-Madison Department of Biostatistics & Medical Informatics for assistance with analysis of lifespan data and R, and Drs. Christine Patterson and Dr. Evan Rosen for the gift of Adipoq-Cre mice and advice on genotyping. We thank Dr. Dena Cohen for advice regarding the implementation and phenotyping of CR mice and continual support and encouragement.

FUNDING

The work was supported in part by the NIH/National Institute on Aging (AG041765, AG050135, AG051974, AG056771, and AG062328 to D.W.L.), by a Glenn Foundation Award for Research in the Biological Mechanisms of Aging to D.W.L., and startup funds from the University of Wisconsin-Madison School of Medicine and Public Health and Department of Medicine to D.W.L. This research was conducted while D.W.L. was an AFAR Research Grant recipient from the American Federation for Aging Research. D.Y. was supported in part by a fellowship from the American Heart Association (17PRE33410983). N.E.C. was supported by a training grant from the UW Institute on Aging (NIA T32 AG000213). The UW Carbone Cancer Center (UWCCC) Experimental Pathology Laboratory is supported by UWCCC Support Grant P30 CA014520 from the NIH National Cancer Institute. The project was supported by the Clinical and Translational Science Award (CTSA) program, through the NIH National Center for Advancing Translational Sciences (NCATS), grant UL1TR002373. The Lamming laboratory is supported in part by the U.S. Department of Veterans Affairs (I01-BX004031), and this work was supported using facilities and resources from the William S. Middleton Memorial Veterans Hospital. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. This work does not represent the views of the Department of Veterans Affairs or the United States Government.

DECLARATION OF INTERESTS

D.W.L has received funding from, and is a scientific advisory board member of, Aeonian Pharmaceuticals, which seeks to develop novel, selective mTOR inhibitors for the treatment of various diseases.