mTORC2 sustains thermogenesis via Akt-induced glucose uptake and glycolysis in brown adipose tissue

Abstract

Activation of non-shivering thermogenesis (NST) in brown adipose tissue (BAT) has been proposed as an anti-obesity treatment. Moreover, cold-induced glucose uptake could normalize blood glucose levels in insulin-resistant patients. It is therefore important to identify novel regulators of NST and cold-induced glucose uptake. Mammalian target of rapamycin complex 2 (mTORC2) mediates insulin-stimulated glucose uptake in metabolic tissues, but its role in NST is unknown. We show that mTORC2 is activated in brown adipocytes upon β-adrenergic stimulation. Furthermore, mice lacking mTORC2 specifically in adipose tissue (AdRiKO mice) are hypothermic, display increased sensitivity to cold, and show impaired cold-induced glucose uptake and glycolysis. Restoration of glucose uptake in BAT by overexpression of hexokinase II or activated Akt2 was sufficient to increase body temperature and improve cold tolerance in AdRiKO mice. Thus, mTORC2 in BAT mediates temperature homeostasis via regulation of cold-induced glucose uptake. Our findings demonstrate the importance of glucose metabolism in temperature regulation.

Keywords: brown adipose tissue; glucose uptake; mTORC2; thermogenesis.

Figure 1. NE activates mTORC2 in vitro via cAMP, PI3K, and Epac1

1. Immunoblot analysis of BAT cells stimulated with norepinephrine (NE) for the indicated proteins.

2. Immunoblot analysis of BAT cells stimulated with NE for 5 min in the presence of rapamycin (Rapa), Torin, or wortmannin (Wrtm) for the indicated proteins.

3. Immunoblot analysis of BAT cells stimulated with 8‐Br‐cAMP for 5 min in the presence of Rapa, Torin, or Wrtm for the indicated proteins.

4. Immunoblot analysis of BAT cells stimulated with NE for 5 min in the presence of Wrtm, H89, or ESI‐09 for the indicated proteins.

5. Immunoblot analysis of BAT cells stimulated with 8‐Br‐cAMP for 5 min in the presence of Wrtm, H89, or ESI‐09 for the indicated proteins.

Data information: All experiments were performed in triplicates, and a representative replicate is presented.

Figure EV1. AdRiKO mice do not display alterations in body weight, plasma IGF‐1 and locomotor activity.

1. Immunoblot analysis of BAT and sWAT of AdRiKO and control mice housed at 22°C for the indicated proteins.

2. Body weight of AdRiKO and control mice housed at 22°C [n = 14 (control), n = 12 (AdRiKO)].

3. Body composition of AdRiKO and control mice housed at 22°C (n = 18/group).

4. Plasma IGF‐1 levels in AdRiKO and control mice housed at 22°C [n = 11 (control), n = 9 (AdRiKO)].

5. Quantification of Akt‐pS473 band intensity relative to total Akt band intensity shown in Fig 2B (n = 3/group).

6. Quantification of Akt‐pS473 and mTOR‐pS2481 band intensity relative to total Akt or total mTOR band intensity shown in Fig 2C (n = 6/group).

7. Immunoblot analysis of sWAT of AdRiKO and control mice housed at 22 or 4°C for 2 h for the indicated proteins (n = 6/group, each lane represents a mix of 3 mice).

8. Locomotor activity of AdRiKO and control mice housed at 22°C (n = 13/group).

9. Body temperature loss of AdRiKO and control mice upon cold exposure with ad libitum access to food [n = 11 (control), n = 10 (AdRiKO)].

10. Cold‐induced shivering of AdRiKO and control mice housed at 4°C for 4 h (n = 6/group).

Data information: Data represent mean ± SEM. Statistically significant differences between AdRiKO and control mice were determined with unpaired Student's t‐test and indicated with asterisks (*P < 0.05; **P < 0.01; ***P < 0.001). Statistically significant differences between temperatures or treatments were determined with unpaired Student's t‐test and are indicated with a number sign (^# P < 0.05; ^## P < 0.01). The exact P‐value for each significant difference can be found in Appendix Table S2.

Figure 2. NE and cold activate mTORC2 in vivo

1. Immunoblot analysis of BAT from control mice treated with either norepinephrine (NE) or vehicle for 30 min for the indicated proteins (n = 3/group).

2. Immunoblot analysis of BAT from AdRiKO and control mice treated with either norepinephrine (NE) or vehicle for 30 min for the indicated proteins (n = 3/group).

3. Immunoblot analysis of BAT from AdRiKO and control mice housed at either 22 or 4°C for 2 h for the indicated proteins (n = 6/group, each lane represents a mix of 3 mice).

4. Body temperature of AdRiKO and control mice housed at 22°C [n = 11 (control), n = 9 (AdRiKO)].

5. Body temperature of AdRiKO and control mice housed at 30°C for 2 weeks (n = 8/group).

6. Body temperature loss upon cold exposure of AdRiKO and control mice [n = 20 (control), n = 17 (AdRiKO)].

Data information: Data represent mean ± SEM. Statistically significant differences between AdRiKO and control mice were determined with unpaired Student's t‐test and are indicated with asterisks (*P < 0.05; **P < 0.01; ***P < 0.001). The exact P‐value for each significant difference can be found in Appendix Table S2.

Figure EV2. mTORC2 in adipose tissue does not affect BAT and sWAT weight.

1. BAT weight of AdRiKO and control mice housed at 22 or 4°C for 8 h (n = 6/group). Data represent mean ± SEM.

2. sWAT weight of AdRiKO and control mice housed at 22 or 4°C for 8 h (n = 6/group). Data represent mean ± SEM.

Figure 3. mTORC2 in adipose tissue is not required for cold‐induced lipid droplet mobilization

1. Representative H&E staining of sWAT sections from AdRiKO and control mice (n = 5/group).

2. Non‐esterified fatty acids (NEFAs) in plasma of AdRiKO and control mice (n = 6/group).

3. Glycerol in plasma of AdRiKO and control mice (n = 6/group).

4. Representative H&E staining of BAT sections from AdRiKO and control mice (n = 5/group).

5. Triglycerides (TGs) in BAT of AdRiKO and control mice housed at 22 or 4°C for 8 h (n = 6/group).

6. NEFAs in BAT of AdRiKO and control mice (n = 6/group).

Data information: Data represent mean ± SEM. Statistically significant differences between AdRiKO and control mice were determined with unpaired Student's t‐test and are indicated with asterisks (**P < 0.01; ***P < 0.001). Statistically significant differences between temperatures were determined with unpaired Student's t‐test and are indicated with a number sign (^# P < 0.05; ^## P < 0.01; ^### P < 0.001). The exact P‐value for each significant difference can be found in Appendix Table S2.

Figure 4. mTORC2 in adipose tissue is not required for cold‐induced mitochondrial uncoupling and β‐oxidation

1. mRNA levels of the indicated genes in BAT of AdRiKO and control mice housed at 22 or at 4°C for 8 h (n = 6/group).

2. Immunoblot analysis of BAT from AdRiKO and control mice housed at 22 or at 4°C for 8 h for the indicated proteins (n = 6/group, each lane represents a mix of 3 mice).

3. mRNA levels of the indicated genes in BAT of AdRiKO and control mice housed at 22 or at 4°C for 8 h (n = 6).

4. Immunoblot analysis of BAT from AdRiKO and control mice housed at 22 or at 4°C for 8 h for the indicated proteins (n = 6/group, each lane represents a mix of 3 mice).

5. Mitochondrial DNA content of BAT from AdRiKO and control mice housed at 22 or at 4°C for 8 h (n = 6/group).

6. Representative electron micrographs of BAT from AdRiKO and control mice housed at 22 or at 4°C for 4 h (n = 3/group).

7. Oxygen consumption rate (OCR) of BAT explants from AdRiKO and control mice housed at 22 or at 4°C for 4 h (n = 7/group).

8. Maximal respiration (VO ₂ max) of AdRiKO and control mice housed at 22 or at 4°C for 8 h [n = 9 (control 22°C), n = 7 (AdRiKO 22°C), n = 8 (control 4°C), n = 8 (AdRiKO 4°C)].

9. Respiration (VO ₂) of AdRiKO and control mice upon cold exposure (n = 8/group).

Data information: Data represent mean ± SEM. Statistically significant differences between AdRiKO and control mice were determined with unpaired Student's t‐test and are indicated with asterisks (*P < 0.05; ***P < 0.001). Statistically significant differences between temperatures were determined with unpaired Student's t‐test and are indicated with a number sign (^# P < 0.05; ^## P < 0.01; ^### P < 0.001). The exact P‐value for each significant difference can be found in Appendix Table S2.

Figure 5. mTORC2 in adipose tissue is required for cold‐induced glucose uptake and glycolysis

1. 2‐deoxyglucose‐6‐phosphate (2DG6P) accumulation in BAT of AdRiKO and control mice housed at 22 or at 4°C for 4 h (n = 6/group).

2. Extracellular acidification rate (ECAR) of BAT explants from AdRiKO and control mice housed at 22 or at 4°C for 4 h (n = 7/group).

3. Immunoblot analysis of BAT from AdRiKO and control mice housed at 22 or at 4°C for 8 h for the indicated proteins (n = 6/group, each lane represents a mix of 3 mice).

4. Immunoblot analysis of isolated plasma membranes from BAT of AdRiKO and control mice housed at 22 or at 4°C for 8 h for the indicated proteins (n = 6/group, each lane represents a mix of 3 mice).

5. Immunoblot analysis of mitochondrial and cytosolic fractions from BAT of AdRiKO and control mice housed at 22 or at 4°C for 4 h for the indicated proteins (n = 6/group, each lane represents a mix of 3 mice).

6. Cytosolic hexokinase activity in BAT of AdRiKO and control mice housed at 22 or at 4°C for 4 h [n = 5 (control 22°C), n = 5 (AdRiKO 22°C), n = 7 (control 4°C), n = 7 (AdRiKO 4°C)].

7. Mitochondrial hexokinase activity in BAT of AdRiKO and control mice housed at 22 or at 4°C for 4 h [n = 5 (control 22°C), n = 5 (AdRiKO 22°C), n = 7 (control 4°C), n = 7 (AdRiKO 4°C)].

Data information: Data represent mean ± SEM. Statistically significant differences between AdRiKO and control mice were determined with unpaired Student's t‐test and are indicated with asterisks (*P < 0.05). Statistically significant differences between temperatures were determined with unpaired Student's t‐test and are indicated with a number sign (^# P < 0.05; ^## P < 0.01; ^### P < 0.001). The exact P‐value for each significant difference can be found in Appendix Table S2.

Figure EV3. Cold exposure decreases blood glucose and circulating insulin.

1. Blood glucose of AdRiKO and control mice housed at 22 or 4°C for 8 h (n = 7/group).

2. Plasma insulin of AdRiKO and control mice housed at 22 or 4°C for 8 h (n = 6/group).

3. Quantification of raptor‐pS792 and ACC‐pS79 band intensity relative to total raptor or total ACC band intensity shown in Fig 5C (n = 6/group).

Data information: Data represent mean ± SEM. Statistically significant differences between AdRiKO and control mice were determined with unpaired Student's t‐test and are indicated with asterisks (*P < 0.05; **P < 0.01). Statistically significant differences between temperatures were determined with unpaired Student's t‐test and are indicated with a number sign (^# P < 0.05; ^## P < 0.01; ^### P < 0.001). The exact P‐value for each significant difference can be found in Appendix Table S2.

Figure EV4. Intra‐BAT injection of AAV targets genes of interest specifically to BAT.

1. Representative immunostainings for RFP of BAT from control mice infected with either AAV8‐RFP or AAV8‐empty (n = 4/group).

2. RFP mRNA expression in BAT, liver, quadriceps, and WAT of control mice infected with either AAV8‐RFP or AAV8‐empty (n = 4/group).

3. Plasma insulin of AdRiKO and control mice infected with either AAV8‐Akt2^S474D or AAV8‐empty housed at 4°C for 4 h [n = 7 (control AAV8‐null), n = 6 (AdRiKO AAV8‐null), n = 6 (control AAV8‐Akt^S474D), n = 6 (AdRiKO AAV8‐Akt^S474D)].

Data information: Data represent mean ± SEM.

Figure 6. Restoration of glucose uptake or Akt signaling suppresses the thermogenic defect in AdRiKO mice

1. HKII mRNA expression level in BAT of AdRiKO and control mice infected with either AAV9‐HKII or AAV9‐empty (n = 8/group).

2. Cold‐induced 2‐deoxyglucose‐6‐phosphate (2DG6P) accumulation in BAT of AdRiKO and control mice infected with either AAV9‐HKII or AAV9‐empty housed at 4°C for 4 h (n = 8/group).

3. Body temperature of AdRiKO and control mice infected with either AAV9‐HKII or AAV9‐empty housed at 22°C (n = 8/group).

4. Body temperature upon cold exposure of AdRiKO and control mice infected with either AAV9‐HKII or AAV9‐empty. The left panel represents body temperature after each hour of cold exposure, while the right panel represents body temperature as a bar graph for the 3‐h cold exposure time point (n = 8/group). a: significant difference between AdRiKO and control mice infected with AAV9‐empty; b: significant difference between AdRiKO and control mice infected with AAV9‐HKII; d: significant difference between AdRiKO mice infected with AAV9‐empty and AAV9‐HKII.

5. Immunoblot analysis of BAT from AdRiKO and control mice infected with either AAV8‐Akt2^S474D or AAV8‐empty (n = 6/group, each lane represents a mix of 3 mice).

6. Body temperature of AdRiKO and control mice infected with either AAV8‐Akt2^S474D or AAV8‐empty housed at 22°C (n = 11/group).

7. Body temperature upon cold exposure of AdRiKO and control mice infected with either AAV8‐Akt2^S474D or AAV8‐empty. The left panel represents body temperature after each hour of cold exposure, while the right panel represents body temperature as a bar graph for the 3‐h cold exposure time point (n = 11/group). a: significant difference between AdRiKO and control mice infected with AAV8‐empty; b: significant difference between AdRiKO and control mice infected with AAV8‐Akt2^S474D; d: significant difference between AdRiKO mice infected with AAV8‐empty and AAV8‐Akt2^S474D.

8. Cold‐induced 2‐deoxyglucose‐6‐phosphate (2DG6P) accumulation in BAT of AdRiKO and control mice infected with either AAV8‐Akt2^S474D or AAV8‐empty housed at 4°C for 4 h [n = 7 (control AAV8‐null), n = 6 (AdRiKO AAV8‐null), n = 6 (control AAV8‐Akt^S474D), n = 6 (AdRiKO AAV8‐Akt^S474D)].

Data information: Data represent mean ± SEM. Statistically significant differences between AdRiKO and control mice were determined with unpaired Student's t‐test and are indicated with asterisks (*P < 0.05; **P < 0.01). Statistically significant differences between viruses were determined with unpaired Student's t‐test and are indicated with a number sign (^# P < 0.05; ^## P < 0.01; ^### P < 0.001). The exact P‐value for each significant difference can be found in Appendix Table S2.

Figure 7. mTORC 2 in BAT is activated by adrenergic stimulation and mediates temperature homeostasis via regulation of cold‐induced glucose uptake and glycolysis