Circadian lipid synthesis in brown fat maintains murine body temperature during chronic cold

Abstract

Ambient temperature influences the molecular clock and lipid metabolism, but the impact of chronic cold exposure on circadian lipid metabolism in thermogenic brown adipose tissue (BAT) has not been studied. Here we show that during chronic cold exposure (1 wk at 4 °C), genes controlling de novo lipogenesis (DNL) including Srebp1, the master transcriptional regulator of DNL, acquired high-amplitude circadian rhythms in thermogenic BAT. These conditions activated mechanistic target of rapamycin 1 (mTORC1), an inducer of Srebp1 expression, and engaged circadian transcriptional repressors REV-ERBα and β as rhythmic regulators of Srebp1 in BAT. SREBP was required in BAT for the thermogenic response to norepinephrine, and depletion of SREBP prevented maintenance of body temperature both during circadian cycles as well as during fasting of chronically cold mice. By contrast, deletion of REV-ERBα and β in BAT allowed mice to maintain their body temperature in chronic cold. Thus, the environmental challenge of prolonged noncircadian exposure to cold temperature induces circadian induction of SREBP1 that drives fuel synthesis in BAT and is necessary to maintain circadian body temperature during chronic cold exposure. The requirement for BAT fatty acid synthesis has broad implications for adaptation to cold.

Keywords: body temperature; brown adipose tissue; circadian; thermogenesis.

Fig. 1.

Lipolytic, fatty acid oxidation (FAO), and de novo lipogenesis (DNL) genes acquire high-amplitude circadian rhythms in BAT of mice exposed to 4 °C for 1 wk. (A) Lipolytic and FAO circadian gene expression in BAT from WT mice housed 1 wk at 29 °C or 4 °C. (B) DNL circadian gene expression in BAT from WT mice housed 1 wk at 29 °C or 4 °C. Results were compared by 2-way ANOVA (n = 5 to 6 per group). Statistical significance of rhythmicity was determined using JTK_CYCLE (56). *P < 0.5, **P < 0.1, ***P < 0.001.

Fig. 2.

Srebp1 acquires high-amplitude circadian rhythms in thermogenic BAT after chronic cold exposure. (A) Srebp1a and Srebp1c circadian gene expression in BAT from WT mice housed 1 wk at 29 °C or 4 °C. (B) Circadian expression of full-length (Fl-SREBP1) and nuclear-cleaved SREBP1c (N-SREBP1) protein in BAT from WT mice housed 1 wk at 29 °C or 4 °C. Liver SCAP-KO sample is used as negative control. The results were compared by 2-way ANOVA (n = 3 to 6 per group). Statistical significance of rhythmicity was determined using JTK_CYCLE (56). *P < 0.5, **P < 0.1, ***P < 0.001.

Fig. 3.

The circadian nuclear receptor REV-ERB regulates SREBP/DNL gene expression in BAT. (A) Circadian expression of REV-ERBα after 1 wk at 29 °C or 1 wk at 4 °C (n = 3 per condition). (B) Circadian expression of Bmal1 in BAT from WT mice housed 1 wk at 29 °C or 4 °C. The results were compared by 2-way ANOVA (n = 5 to 6 per group). (C) Endogenous HA-epitope–tagged REV-ERBα protein model. (D) ChIP-seq of BAT from HA-REV-ERBα mice, using HA antibody, comparing mice housed at thermoneutrality or 1 wk at 4 °C. Scatterplots of the HA-REV-ERBα cistrome in BAT at 29 °C (y axis) and 4 °C (x axis). The sites selective at 4 °C are plotted in blue, and the sites common at 29 °C and 4 °C are plotted in black. The cutoff for differentially regulated sites was a fold-change of >4 (n = 4 per group). (E) Gene ontology analysis of HA-REV-ERBα ChIP-seq peaks selective to BAT after 1 wk at 4 °C. (F) Genome browser view of HA-REV-ERBα ChIP-seq peaks in mice housed at 29 °C (red) and 4 °C (blue) near DNL genes. (G) Rev-erbα and Rev-erbβ gene expression in BAT from control (Cre-) and Rev-erbα/β DKO mice housed 1 wk at 4 °C (n = 3). (H) Canonical REV-ERB target genes expression in BAT from control (Cre-) and REV-ERB/β DKO mice housed 1 wk at 4 °C (n = 3). (I) DNL gene expression in BAT from control (Cre-) and REV-ERBα/β DKO mice housed 1 wk at 4 °C at ZT10 (n = 7 to 9 per group). In G–I, the results were compared by unpaired t test. *P < 0.5, **P < 0.1, ***P < 0.001.

Fig. 4.

SREBP function in BAT is necessary to induce DNL during chronic cold exposure. (A) Scap, (B) Srebp1c gene expression, and (C) detection of nuclear-cleaved nSREBP1 protein at ZT16 in BAT from control (ScapFlox/Flox) or SCAP BAT KO (ScapFlox/Flox-Ucp1CreER+) treated with tamoxifen at 6 wk of age and housed for 1 wk at 4 °C beginning at 12 wk of age. Liver SCAP-KO sample was used as negative control in C. The results were compared by unpaired t test (n = 3 to 5 per group). (D) DNL gene expression at ZT4 or ZT16 in BAT from control or SCAP BAT KO mice housed 1 wk at 4 °C. The results were compared by 2-way ANOVA (n = 7 to 12 per group). (E) BAT weight from control or SCAP BAT KO mice housed 1 wk at 4 °C. The results were compared by unpaired t test (n = 22 to 26 per group). (F) Triglycerides (TG) in BAT from control or SCAP BAT KO mice housed 1 wk at 4 °C. Measurements were performed at ZT4. The results were compared by unpaired t test (n = 4 per group). *P < 0.5, **P < 0.1, ***P < 0.001.

Fig. 5.

Mice in chronic cold require SREBP function for maximal thermogenic capacity in response to norepinephrine. Thermogenic capacity of control or SCAP BAT KO mice in response to norepinephrine (NE) after 1 wk at 29 °C or 4 °C measured by oxygen consumption (A–C) and heat production (D–F). Time-course and direct comparison 1 h after NE injection. A, anesthesia. The results were compared by 2-way ANOVA (n = 5 to 7 per group). *P < 0.5, **P < 0.1, ***P < 0.001.

Fig. 6.

Mice in chronic cold require SREBP function in BAT to maintain body temperature during circadian trough and fasting. (A and B) Body temperature of WT mice housed 1 wk at 29 °C or 4 °C. The results were compared by 2-way ANOVA (n = 4 to 5 per group). (C and D) Body temperature of control or SCAP BAT KO mice housed 1 wk at 4 °C. The results were compared by 2-way ANOVA (n = 12 to 14 per group, from 3 independent experiments). (E) Body temperature of control or SCAP BAT KO mice housed 1 wk at 4 °C and fasted from ZT1. The results were compared by 2-way ANOVA (n = 7 per group). (F) Body temperature of control and REV-ERBα/β DKO mice housed 1 wk at 4 °C and fasted from ZT1. The results were compared by 2-way ANOVA (n = 10 per group). *P < 0.5, **P < 0.1, ***P < 0.001.

Fig. 7.

Model depicting the circadian regulation of DNL by chronic cold exposure. In response to chronic cold stimulation, mTORC1 pathway is activated and triggers SREBP activity, as previously described (–30). In parallel, chronic cold shift-advances REV-ERBα expression to the light phase and remodels its cistrome, targeting SREBP and DNL gene induction. Continuous activation of SREBP by mTORC1 combined with its circadian inhibition by REV-ERB results in the circadian expression of SREBP and DNL genes, which maximize thermogenic capacity and maintain body temperature during the light phase and fasting.