Cold-Induced Thermogenesis Depends on ATGL-Mediated Lipolysis in Cardiac Muscle, but Not Brown Adipose Tissue

Abstract

Fatty acids (FAs) activate and fuel UCP1-mediated non-shivering thermogenesis (NST) in brown adipose tissue (BAT). Release of FAs from intracellular fat stores by adipose triglyceride lipase (ATGL) is considered a key step in NST. Accordingly, the severe cold intolerance of global ATGL knockout (AKO) mice has been attributed to defective BAT lipolysis. Here we show that this conclusion is incorrect. We demonstrate that although the BAT-specific loss of ATGL impairs BAT lipolysis and alters BAT morphology, it does not compromise the β₃-adrenergic thermogenic response or cold-induced NST. Instead, NST depends on nutrient supply or lipolysis in white adipose tissue during fasting, suggesting that circulating energy substrates are sufficient to fuel NST. Cold intolerance in AKO mice is not caused by BAT dysfunction as previously suspected but by severe cardiomyopathy. We conclude that functional NST requires adequate substrate supply and cardiac function, but does not depend on ATGL-mediated lipolysis in BAT.

Keywords: adipose triglyceride lipase; brown adipose tissue; cold; heart; lipolysis; thermogenesis; white adipose tissue.

Graphical abstract

Figure 1

Global ATGL Deficiency Causes Cold-Induced Hypothermia and BAT Hypertrophy, but Does Not Alter Mitochondrial Function in BAT (A) TG hydrolase activities of BAT infranatants in the absence or presence of HSL inhibitor (HSLi) from mice housed at 22°C–23°C (n = 3). (B) Body temperature in ad libitum-fed mice upon acute cold exposure at 5°C for 3–6 hr (n = 6). (C) BAT weight (n = 7). (D) Histology of BAT. Scale bar, 100 μm. (E) Relative mRNA expression of classical brown fat genes upon acute cold exposure (n = 6). (F) UCP-1 immunoblot of isolated BAT mitochondria upon acute cold exposure. (G) Total DNA content in whole BAT depots (n = 6). (H) DNA content in isolated brown adipocytes (adi) and stroma-vascular fraction (SVF; n = 3). (I) Total protein content in isolated mitochondria from whole BAT depots (n = 10). (J) Relative mtDNA content in BAT assessed by qPCR and calculated from copy numbers of the mtDNA-encoded MtCO1 gene and the nuclear DNA-encoded Ndufv1 gene (n = 6). (K) Representative transmission electron micrographs from BAT. Scale bar, 0.5 μm. (L) Oxygen consumption rates (OCRs) in BAT homogenates (hom) using pyruvate (pyr), glycerol-3-P (G3P) in the absence or presence of rotenone (R, G3P/R), and guanosine 5′-diphosphate (GDP). OCRs were calculated for whole BAT depots (n = 6–7). Analyses were performed in male mice, except for (H), which used female mice, aged 9–10 weeks. Data are presented as means ± SD. Statistical significance was evaluated by unpaired two-tailed Student’s t test. ^∗p < 0.05, ^∗∗p < 0.01, and ^∗∗∗p < 0.001. See also Figure S1.

Figure 2

ATGL in BAT Is Not Essential for Thermogenesis during Acute Cold (A) ATGL immunoblot of BAT homogenates (n = 4; representative of at least two independent analyses). (B) Ex vivo lipolysis in BAT under basal and isoproterenol (ISO)-stimulated conditions (n = 4). (C) TG hydrolase activities in BAT infranatants in the absence or presence of ATGL inhibitor Atglistatin (ATGLi) from mice exposed to 5°C for 6 hr (n = 5). (D) BAT weight (n = 11) and tissue gross morphology (inset). (E) Histology of BAT. Scale bar, 100 μm. (F) Body temperature in ad libitum-fed mice during acute cold exposure at 5°C (n ≥ 5). (G) DNA content in brown adipocytes (adi) and stroma-vascular fraction (SVF; n = 3). (H) UCP-1 immunoblot of isolated BAT mitochondria upon acute cold exposure. (I) Oxygen consumption rates (OCRs) in BAT homogenates (hom) using pyruvate (pyr), glycerol-3-P (G3P) in the absence and presence of rotenone (G3P/R), and guanosine 5′-diphosphate (GDP). OCRs were calculated for whole BAT depots (n = 6). Analyses were performed in male mice, except for (G), which used female mice, aged 9–11 weeks and 4 weeks upon tamoxifen administration. Data are presented as means ± SD. Statistical significance was evaluated by unpaired two-tailed Student’s t test or two-way ANOVA with Bonferroni post hoc tests. ^∗p < 0.05, ^∗∗p < 0.01, and ^∗∗∗p < 0.001 versus ATGL^flox/flox control and ^§p < 0.05 versus basal. See also Figure S2.

Figure 3

ATGL-Mediated Lipolysis in WAT Is Essential to Fuel Thermogenesis during Fasting (A and C) Body temperature in iBAKO (A) and AAKO (C) mice during fasting at 5°C at indicated time points. (B and D) Blood glucose, plasma FA, glycerol, and TG levels upon fasting at 5°C for 6 hr in iBAKO mice (B) (n = 5–8) and for 2 hr in AAKO mice (D) (n = 7). Analyses were performed in female iBAKO and ATGL^flox/flox mice aged 40 weeks and 8 weeks upon tamoxifen administration and AAKO mice of mixed sex aged 10–16 weeks. Data are presented as means ± SD. Statistical significance was evaluated by unpaired two-tailed Student’s t test. ^∗p < 0.05, ^∗∗p < 0.01, and ^∗∗∗p < 0.001. See also Figure S3.

Figure 4

Impaired Heart Function Due to Cardiac ATGL Deficiency Causes Hypothermia (A and D) Kaplan-Meier plots depicting mice maintaining euthermia (body temperature ≥35°C) during acute cold exposure at 5°C at indicated time points from AKO (A) and iHAKO mice (D). (B, E, and H) Left ventricular (LV) mass from AKO (B), iHAKO (E), and AKO/cTg mice (H). (C, F, and I) Ejection fraction from AKO (C), iHAKO (F), and AKO/cTg mice (I). (G) Body temperature of male AKO/cTg mice during acute cold at 5°C for 3–6 hr (n ≥ 8). Male AKO mice were analyzed at the age of 6 and 10 weeks (n ≥ 6). iHAKO mice of mixed sex were studied at the age of 12 months and 6 and 10 weeks upon tamoxifen administration (postTAM) (n ≥ 6). Male AKO/cTg mice aged 5–6 months were used to study LV mass and ejection fraction (n = 5). Data are presented as means ± SD, except for Kaplan-Meier plot showing median. Statistical significance was evaluated by Gehan-Breslow-Wilcoxon test for Kaplan-Meier plots or two-way ANOVA and Bonferroni post hoc tests. ^∗∗p < 0.01 and ^∗∗∗p < 0.001 versus control and ^###p < 0.001 versus 6-week-old respective groups. See also Figures S4 and S5.

Figure 5

ATGL-Deficient Mice Survive upon Cold Acclimation and Exhibit Normal Brown Adipocyte Recruitment (A and B) Time course (A) and mean (B) core body temperature (CBT) of ad libitum-fed AKO/cTg mice during cold acclimation at 5°C for 3 weeks using implantable telemetry transmitters (n = 5). (C) UCP-1 immunoblot of BAT homogenates. (D) Total protein content from whole BAT depots (n = 3–5). (E) Calculated total UCP-1 protein amount per BAT of mice housed at 22°C (n = 3) or 5°C for 3 weeks (n = 5). Analyses were performed in ad libitum-fed male AKO/cTg mice aged 10–12 weeks. Data are presented as means ± SD. Statistical significance was evaluated by unpaired two-tailed Student’s t test. ^∗p < 0.05 and ^∗∗∗p < 0.001. See also Figure S6.

Figure 6

Adaptive NST Is Intact in ATGL-Deficient BAT (A–D) Studies in mice acclimatized to thermoneutrality (28°C–30°C). (E–H) Studies in mice acclimatized to cold (5°C). (A and E) Time course of metabolic rates (VO₂) during basal and upon CL administration upon acclimation to thermoneutrality (A) and to cold (E) (n = 5–9). (B and F) Mean metabolic rates during basal and upon CL administration upon acclimation to thermoneutrality (B) and to cold (F) (n = 5–9). (C and G) Plasma FA levels upon 15 min of CL injection upon acclimation to thermoneutrality (C) and to cold (G) (n = 5–9). (D and H) Body temperature upon 60 min of CL injection upon acclimation to thermoneutrality (D) and to cold (H) (n = 4–6). For analyses upon thermoneutrality, male iBAKO, AAKO, and ATGL^flox/flox control mice aged 25–30 weeks were acclimatized for 3 weeks to thermoneutrality at 28°C–30°C (A–D). For analyses upon cold acclimation, male iBAKO, AAKO, and ATGL^flox/flox control mice aged 8–12 weeks were acclimatized to cold at 5°C for 3 weeks (E–H). For iBAKO mice, Atgl gene deletion was induced 1 week prior to acclimation to thermoneutrality or cold. During cold acclimation, iBAKO and ATGL^flox/flox control mice received tamoxifen twice per week to maintain Atgl gene deletion. Analyses in iBAKO mice were performed 4 weeks upon tamoxifen administration. Independent of prior acclimation temperatures, metabolic rates (VO₂) were analyzed at 28°C–30°C in pentobarbital-anesthetized mice using indirect calorimetry (Phenomaster, TSE). Data are presented as means ± SD, except for (A) and (E) showing means. Statistical significance was evaluated by unpaired two-tailed Student’s t test or two-way ANOVA and Bonferroni post hoc tests. ^∗p < 0.05, ^∗∗p < 0.01, and ^∗∗∗p < 0.001 versus control and ^§p < 0.05 and ^§§§p < 0.001 versus basal. See also Figure S7.