SIRT7 suppresses energy expenditure and thermogenesis by regulating brown adipose tissue functions in mice

Abstract

Brown adipose tissue plays a central role in the regulation of the energy balance by expending energy to produce heat. NAD⁺-dependent deacylase sirtuins have widely been recognized as positive regulators of brown adipose tissue thermogenesis. However, here we reveal that SIRT7, one of seven mammalian sirtuins, suppresses energy expenditure and thermogenesis by regulating brown adipose tissue functions. Whole-body and brown adipose tissue-specific Sirt7 knockout mice have higher body temperature and energy expenditure. SIRT7 deficiency increases the protein level of UCP1, a key regulator of brown adipose tissue thermogenesis. Mechanistically, we found that SIRT7 deacetylates insulin-like growth factor 2 mRNA-binding protein 2, an RNA-binding protein that inhibits the translation of Ucp1 mRNA, thereby enhancing its inhibitory action on Ucp1. Furthermore, SIRT7 attenuates the expression of batokine genes, such as fibroblast growth factor 21. In conclusion, we propose that SIRT7 serves as an energy-saving factor by suppressing brown adipose tissue functions.

Fig. 1. Sirt7 KO mice display elevated energy expenditure and body temperature under normal conditions.

a, b Body weight (a) and percent epiWAT and iBAT weight calculated relative to body weight (b) in 15-week-old male WT and Sirt7 KO mice. p = 0.0286 (epiWAT), p = 0.0288 (iBAT). c–j Data of indirect calorimetry experiments from 15-week-old male WT and Sirt7 KO mice. VO₂ rates (c), average VO₂ (d), average VCO₂ (e), energy expenditure (f), RER (g), locomotor activity (h), water intake (i), and food intake (j). p = 0.0021 in (c); p = 0.0023 (Dark), p = 0.0082 (Light), p = 0.0021 (Total) in (d); p = 0.0255 (Dark), p = 0.0101 (Light), p = 0.0106 (Total) in (e); p = 0.0063 (Dark), p = 0.0049 (Light), p = 0.0026 (Total) in (f); p = 0.0296 in (j). k Representative H&E-stained sections of iBAT from 15-week-old male WT and Sirt7 KO mice (left panel, scale bar = 50 μm), and quantification of the lipid area (n = 3 independent animals per group) (right panel). p = 0.0347. l Oscillation of body temperature in 10-week-old male WT and Sirt7 KO mice. p = 0.0013. Data are presented as means ± SEM. All numbers (n) are biologically independent samples. Two-way ANOVA with Bonferroni’s multiple comparisons test (c, l); two-tailed Student’s t-test (a, b, d–k). *p < 0.05; **p < 0.01. Source data are provided as a Source Data file.

Fig. 2. Sirt7 KO mice exhibit excessive energy expenditure and thermogenesis in the hypometabolic state.

a–c Data of indirect calorimetry experiments and body temperature from 2-year-old male WT and Sirt7 KO mice (n = 5 independent animals per group). Data from young mice (from Fig. 1d, f, l) are shown for a comparison of their elevated rate with that of the 2-year-old mice. Average VO₂ (a), energy expenditure (b), and body temperature at zeitgeber time (ZT) 10 (c). p = 0.0021 (15 weeks), p = 0.0260 (2 years), p = 0.0087 (WT 15 weeks vs. 2 years) in (a); p = 0.0026 (15 weeks), p = 0.0283 (2 years), p = 0.0170 (WT 15 weeks vs. 2 years) in (b); p = 0.0025 (15 weeks), p = 0.0143 (2 years), p = 0.0002 (WT 15 weeks vs. 2 years) in (c). d–f Data of indirect calorimetry experiments and body temperature during artificially-induced daily torpor in 10-week-old male WT and Sirt7 KO mice. Representative pattern of energy expenditure (d) and the average energy expenditure (e) from 12 h (ZT14) after initiation of fasting (ZT2). Body temperature after 24 h fasting (f). p = 1.4E–05 in (e); p = 0.0406 in (f). Data are presented as means ± SEM. All numbers (n) are biologically independent samples. Two-way ANOVA with Bonferroni’s multiple comparisons test (e); two-tailed Student’s t-test (a–d, f). *p < 0.05; **p < 0.01. Source data are provided as a Source Data file.

Fig. 3. SIRT7 suppresses energy expenditure and thermogenesis via multiple pathways.

a Real-time qPCR analysis of BAT-related genes in iBAT of 15-week-old male WT and Sirt7 KO mice. p = 0.0288 (Ebf2), p = 0.0098 (Dio2), p = 0.0003 (Ucp1), p = 0.0206 (Elovl3), p = 3.4E–06 (Clstn3b), p = 0.0053 (S100b), p = 0.0296 (Bmp8b). b T3 level in iBAT and serum of 13-week-old male WT and Sirt7 KO mice. p = 0.0192. c Western blot analysis of UCP1 and OXPHOS complexes I–IV in iBAT of 15-week-old male WT and Sirt7 KO mice (left panel), and quantification of the UCP1 protein bands relative to GAPDH control (right panel). p = 0.0347. d Western blot analysis of tyrosine hydroxylase (TH) in iBAT of 15-week-old male WT and Sirt7 KO mice (left panel), and quantification of the TH protein bands relative to RNA Polymerase II (Pol II) control (right panel). p = 0.0005. e Real-time qPCR analysis of batokine genes in iBAT of 15-week-old male WT and Sirt7 KO mice. p = 0.0109 (Fgf21), p = 0.0016 (Nrg4). f–h Data of indirect calorimetry experiments and body temperature from 20-week-old male WT and Sirt7 KO mice at thermoneutrality. VO₂ rates (f), energy expenditure (g), and body temperature at ZT10 (h). p = 4.5E–02 (Dark) in (f); p = 0.0386 (Dark), p = 0.0473 (Total) in (g); p = 0.0439 in (h). Data are presented as means ± SEM. All numbers (n) are biologically independent samples. Two-way ANOVA with Bonferroni’s multiple comparisons test (f); two-tailed Student’s t-test (a–e, g–h). *p < 0.05, **p < 0.01, ***p < 0.001. Source data are provided as a Source Data file.

Fig. 4. SIRT7 deficiency in adipose tissue elevates whole-body energy expenditure and body temperature.

a, b Data of indirect calorimetry experiments and body temperature from 12-week-old male Adipoq-Cre control and Sirt7 AdKO mice. VO₂ rates (a) and energy expenditure (b). p = 3.7E–29 in (a); p = 0.0011 (Dark), p = 0.0030 (Light), p = 0.0006 (Total) in (b). c Oscillation of body temperature in 20-week-old male Adipoq-Cre control and Sirt7 AdKO mice. p = 0.0025. d Real-time qPCR analysis of BAT-related genes in iBAT of 12-week-old male Adipoq-Cre control and Sirt7 AdKO mice. p = 0.0089 (Pparg), p = 0.0239 (Ppargc1a), p = 0.0026 (Ebf2), p = 0.0005 (Dio2), p = 0.0010 (Cox3), p = 0.0006 (Cox8b), p = 5.0E–06 (Clstn3b), p = 0.0001 (S100b), p = 0.0002 (Bmp8b), p = 0.0003 (Fgf21), p = 0.0008 (Nrg4). e T3 level in iBAT of 11-week-old male Adipoq-Cre control and Sirt7 AdKO mice. p = 0.0420. f Western blot analysis of UCP1 in iBAT of 12-week-old male Adipoq-Cre control and Sirt7 AdKO mice (left panel). Quantification of the UCP1 protein bands relative to GAPDH control is shown on the right side. p = 0.0019. Data are presented as means ± SEM. All numbers (n) are biologically independent samples. Two-way ANOVA with Bonferroni’s multiple comparisons test (a, c); two-tailed Student’s t-test (b, d–f). *p < 0.05, **p < 0.01, ***p < 0.001. Source data are provided as a Source Data file.

Fig. 5. Brown adipocytic SIRT7 suppresses energy expenditure and thermogenesis in vivo.

a, b Data of indirect calorimetry experiments and body temperature from 12-week-old male Ucp1-Cre control and Sirt7 BAdKO mice. VO₂ rates (a) and energy expenditure (b). p = 9.1E–36 in (a); p = 1.8E–04 (Dark), p = 3.8E–05 (Light), p = 3.6E–05 (Total) in (b). c Oscillation of body temperature in 20-week-old Ucp1-Cre control and Sirt7 BAdKO mice. p = 0.0399. d Real-time qPCR analysis of BAT-related genes in iBAT of 12-week-old male Ucp1-Cre control and Sirt7 BAdKO mice. p = 0.0245 (Prdm16), p = 0.0046 (Cebpb), p = 0.0009 (Fgf21), p = 0.0056 (Nrg4). e Western blot analysis of UCP1 in iBAT of 12-week-old male Ucp1-Cre control and Sirt7 BAdKO mice (left panel). Quantification of the UCP1 protein bands relative to GAPDH control is shown on the right side. p = 0.0192. f, g Data of indirect calorimetry experiments and body temperature from 20-week-old male Ucp1-Cre control and Sirt7 BAdKO mice at thermoneutrality. Energy expenditure (f) and body temperature at ZT10 (g). Data are presented as means ± SEM. All numbers (n) are biologically independent samples. Two-way ANOVA with Bonferroni’s multiple comparisons test (a, c); two-tailed Student’s t-test (b, d–g). *p < 0.05, **p < 0.01, ***p < 0.001. Source data are provided as a Source Data file.

Fig. 6. SIRT7 cell-autonomously suppresses mitochondrial respiration via a reduction in the UCP1 protein level in brown adipocytes.

a Oil red O staining of differentiated primary brown adipocytes from WT and Sirt7 KO mice (left panel, scale bar = 50 μm), and quantification of the stained oil red O (right panel). b Real-time qPCR analysis of BAT-related genes in the cells described in (a). c Western blot analysis of UCP1 in the cells described in (a) (left panel), and quantification of the UCP1 protein bands relative to GAPDH control (right panel). p = 0.0307. d Evaluation of the mtDNA copy number in differentiated primary brown adipocytes from WT and Sirt7 KO mice by determining the ratio of mtDNA to nDNA. e, f Mitochondrial stress test in the cells described in (a). Time course OCR (e) and quantification of mitochondrial respiration (f). p = 0.0247 (basal), p = 0.0239 (maximal). Data are presented as means ± SEM. All numbers (n) are biologically independent samples. *p < 0.05 by two-tailed Student’s t-test. Source data are provided as a Source Data file.

Fig. 7. SIRT7 suppresses energy expenditure via IMP2 in brown adipocytes.

a Halo-SIRT7 pull-down assay with extracts from brown adipocytes. Eluted proteins were resolved by SDS-PAGE, followed by silver staining. The gel lanes were cut into 26 pieces and the proteins within each gel piece were analyzed by mass spectrometry. b Halo-SIRT7 pull-down assay performed with lysates from HEK293T cells overexpressing 3×HA-IMP2. Overexpressed IMP2 was detected by WB with an anti-IMP2 antibody. c Co-IP assay detecting the interaction between FLAG-SIRT7 and 3×HA-IMP2 in HEK293T cells. d, e The effect of Imp2 deficiency on mitochondrial respiration in differentiated primary brown adipocytes from WT and Sirt7 KO mice. SVF cells were infected with the indicated recombinant AAV and differentiated for 9 days. Time course OCR (d) and quantification of mitochondrial respiration (e). p = 0.0020 (WT + Ctrl AAV vs. WT + Imp2 KO AAV), p = 0.0005 (WT + Ctrl AAV vs. Sirt7 KO + Ctrl AAV) in basal respiration; p = 0.0039 (WT + Ctrl AAV vs. WT + Imp2 KO AAV), p = 0.0031 (WT + Ctrl AAV vs. Sirt7 KO + Ctrl AAV) in maximal respiration; p = 0.0099 (WT + Ctrl AAV vs. WT + Imp2 KO AAV), p = 0.0348 (WT + Ctrl AAV vs. Sirt7 KO + Ctrl AAV) in uncoupled respiration. f, g Western blot (f) and real-time qPCR (g) analysis of Ucp1 in the cells described in (d). n = 4 independent samples per group in (g). WB western blotting, IP immunoprecipitation, N.S. not significant. Data are presented as means ± SEM. All numbers (n) are biologically independent samples. The screening experiment (a) were performed one time. *p < 0.05, **p < 0.01 by two-tailed Student’s t-test. Source data are provided as a Source Data file.

Fig. 8. Deacetylation of IMP2 by SIRT7 attenuates the translation of Ucp1 mRNA.

a The effect of SIRT7 overexpression on IMP2 acetylation. HEK293T cells were transfected with the indicated expression plasmid, and IMP2 acetylation was assessed by IP and WB (left panel). Quantification of the acetylated IMP2 relative to IMP2 (n = 3) (right panel). p = 0.0174. b The effect of Sirt7 deficiency on IMP2 acetylation. Protein lysates of iBAT were subjected to IP, after which the acetylated IMP2 was detected by WB (left panel). Quantification of the acetylated IMP2 relative to IMP2 (n = 3) (right panel). p = 0.0170. c Mapping of the region of the interaction between IMP2 and SIRT7. Halo-SIRT7 pull-down assay with lysates of HEK293T cells expressing the indicated IMP2 deletion mutants fused with GAL4DBD. See also Supplementary Fig. 6a. d The effect of SIRT7 overexpression on IMP2^K438R and IMP2^K439R acetylation. HEK293T cells were transfected with the indicated expression plasmid, and IMP2 acetylation was assessed by IP and WB. Quantification of the acetylated IMP2 is shown in Supplementary Fig. 6b. e Binding between in vitro-transcribed long Ucp1 3′-UTR and IMP2 (IMP2^WT, IMP2^K438R, IMP2^K438Q) (n = 4 independent samples per group). p = 0.0401 (IMP2^WT vs. IMP2^K438Q), p = 1.7E-05 (IMP2^K438R vs. IMP2^K438R). f The effect of IMP2 mutant on the translation of luciferase-Ucp1 UTR fusion mRNA. luciferase-Ucp1 UTR mRNA was transcribed in vitro and used in an insect cell-free translation system with recombinant protein in E. coli for IMP2^WT, IMP2^K438R, and IMP2^K438Q (n = 3 independent samples per group). Translation was estimated by the gain in luciferase activity after incubation for 5 h at 25 °C. p = 0.0016 (Control vs. IMP2^WT), p = 0.0106 (Control vs. IMP2^K438R). g Proposed model for the suppression of brown fat thermogenesis through attenuation of Ucp1 mRNA translation via SIRT7-dependent deacetylation of IMP2. See Discussion for details. WB western blotting, IP immunoprecipitation, N.S. not significant. Data are presented as means ± SEM. All numbers (n) are biologically independent samples. *p < 0.05, **p < 0.01 by two-tailed Student’s t-test. Source data are provided as a Source Data file.