Volume 23, Issue 3, Pages 441-453 (March 2016) FGF21 Lowers Plasma Triglycerides by Accelerating Lipoprotein Catabolism in White and Brown Adipose Tissues  Christian Schlein, Saswata Talukdar, Markus Heine, Alexander W. Fischer, Lucia M. Krott, Stefan K. Nilsson, Martin B. Brenner, Joerg Heeren, Ludger Scheja  Cell Metabolism  Volume 23, Issue 3, Pages 441-453 (March 2016) DOI: 10.1016/j.cmet.2016.01.006 Copyright © 2016 Elsevier Inc. Terms and Conditions

Cell Metabolism 2016 23, 441-453DOI: (10.1016/j.cmet.2016.01.006) Copyright © 2016 Elsevier Inc. Terms and Conditions

Figure 1 FGF21 Reduces VLDL-TG Secretion and Liver Lipid Content by Diminishing Plasma NEFAs (A–F) In a VLDL secretion experiment, C57BL/6J mice (n = 7 per group) were injected with either saline or FGF21, followed by a gavage of saline or glucose (2 g/kg). This resulted in experimental groups receiving neither FGF21 nor glucose (−Glc-mock), FGF21 but no glucose (−Glc-FGF21), no FGF21 but glucose (+Glc-mock), or both FGF21 and glucose (+Glc-FGF21). Subsequently, mice were injected with tyloxapol to determine hepatic production of VLDLs. (A) Schedule of the experiment. (B and C) Change of plasma cholesterol (Chol) (B) and plasma TGs (C) after tyloxapol injection. (D) Plasma NEFAs. (E and F) TGs (E) and Chol (F) per milligram of protein of total liver lysate. (G–I) Chronic versus acute FGF21 dosing experiment (n = 8 per group). Mice were subcutaneously injected for 10 days with either saline (mock) or FGF21 (chronic FGF21) or for 9 days with saline and then with FGF21 on the final day (acute FGF21). Plasma NEFAs (G), liver TGs (H), and liver Chol (I) were determined in blood and tissues harvested 4 hr after final injection and food withdrawal. (J) VLDL secretion experiment (according to A) performed in DIO mice (n = 5–6 per group) treated acutely with FGF21. Mean ± SEM, Student’s t test, ∗p < 0.05 mock versus FGF21, +p < 0.05 −Glc-mock versus +Glc-mock. Cell Metabolism 2016 23, 441-453DOI: (10.1016/j.cmet.2016.01.006) Copyright © 2016 Elsevier Inc. Terms and Conditions

Figure 2 FGF21 Stimulates Fatty Acid Uptake in WAT (A and B) Hepatic DNL in C57BL/6J mice assessed by gene expression after acute or chronic FGF21 treatment (same mice as in Figures 1G–1I) (A) and by 3H2O incorporation into liver lipids after acute FGF21 treatment with or without glucose gavage (n = 4 mice per group) (B). (C) VLDL secretion from primary hepatocytes preloaded with 3H-glycerol, in the presence or absence of oleate. Secretion of 3H-glycerol lipids incorporated into lipoproteins was determined in supernatants collected 2 hr after treatment with insulin (50 nmol/l) and/or FGF21 (250 nmol/l). Results are presented as mean ± SEM of three independent experiments. (D) Phosphorylation of hormone-sensitive lipase at Ser563 (pHSL) in ingWAT of fasted C57BL/6J mice after acute FGF21 treatment determined by western blot analysis. (E) Effect of insulin (50 nmol/l), FGF21 (250 nmol/l), and the beta-3 adrenergic agonist CL316,243 (50 nmol/l) on NEFA release from primary white adipocytes determined in triplicates. (F) Organ uptake of 14C-labeled oleate 10 min after intravenous injection in C57BL/6J mice (n = 6 per group). Glucose gavage and FGF21 administration were performed as described in Figure 1. Mean ± SEM, Student’s t test, ∗p < 0.05. Cell Metabolism 2016 23, 441-453DOI: (10.1016/j.cmet.2016.01.006) Copyright © 2016 Elsevier Inc. Terms and Conditions

Figure 3 FGF21 Stimulates Postprandial Disposal of TRLs in WAT and BAT (A–K) An OGFTT using radioactive tracers was performed to determine the effect of acute and chronic FGF21 treatment on plasma parameters (B–E), organ-specific tracer uptake (F–I), and gene expression (J and K). C57BL/6J mice (n = 8 per group) were acutely or chronically (10 days) treated with FGF21. Tracers used in the oral gavage were 14C-triolein and 3H-DOG. (A) Schedule of the experiment. (B) TGs. (C) Cholesterol. (D) Glucose. (E) Insulin in plasma. (F and G) 14C-oleate uptake in disintegrations per minute per organ (F) and fold (G). (H and I) 3H-DOG uptake in disintegrations per minute per organ (H) and fold (I). (J and K) Normalized mRNA expression in ingWAT (J) and in gonWAT (K) in relation to control (mock). See also Figure S1. rpWAT, retroperitoneal WAT; muscle, quadriceps muscle. Mean ± SEM, Student’s t test, ∗p < 0.05 mock versus FGF21. Cell Metabolism 2016 23, 441-453DOI: (10.1016/j.cmet.2016.01.006) Copyright © 2016 Elsevier Inc. Terms and Conditions

Figure 4 FGF21 Stimulates TRL Particle Uptake into Adipose Tissues (A–E) C57BL/6J mice were treated acutely with FGF21. Subsequently, radiolabeled rTRLs were intravenously injected to detect organ-specific fatty acid and particle uptake, respectively (n = 6–7 per group). (A) Schedule of TRL uptake experiment. (B and C) 14C-oleate uptake in disintegrations per minute per organ (B) and fold (C). (D and E) 3H-CE uptake in disintegrations per minute per organ (D) and fold (E). (F) Uptake of fluorescence-labeled QD-TRLs (green) in gonWAT by intravital microscopy. A representative image of n = 3 animals per group is shown. (G) Disposal of 3H-triolein TRLs after 5 min in vivo perfusion following treatment with FGF21 or the beta-3 adrenergic agonist CL316,243. Mean ± SEM, Student’s t test, ∗p < 0.05 mock versus treated. Cell Metabolism 2016 23, 441-453DOI: (10.1016/j.cmet.2016.01.006) Copyright © 2016 Elsevier Inc. Terms and Conditions

Figure 5 FGF21-Stimulated TRL Clearance into WAT Depends on CD36 (A–C) Cd36−/− mice and WT littermates (n = 5 per group) were treated acutely with FGF21. Subsequently, radiolabeled rTRLs were intravenously injected. (A) Schedule. (B) 14C-oleate organ uptake. (C) 3H-CE organ uptake. (D) Plasma membrane (PM) localization of CD36 in primary white adipocytes 1 hr after treatment with PBS (−) or FGF21 (+, 250 nmol/l), determined by cell surface biotinylation, pull-down with streptavidin, and anti-CD36 western blot. Two prominent CD36 isoforms were detected (low glycosylation, 70 kDa; high glycosylation, 90 kDa). (E) PM abundance of CD36 quantified by normalization to actin after densitometric quantification of protein bands from two independent experiments. Mean ± SEM, Student’s t test, ∗p < 0.05 mock versus FGF21, +p < 0.05 Cd36−/− versus equally treated WT group. Cell Metabolism 2016 23, 441-453DOI: (10.1016/j.cmet.2016.01.006) Copyright © 2016 Elsevier Inc. Terms and Conditions

Figure 6 LPL-Dependent TRL Clearance into Adipose Tissue (A–E) An OGFTT using radioactive tracers was performed in Apoa5−/− and in adipocyte-specific LPL-deficient (aLPL−/−) mice. (A) OGFTT schedule. (B and C) 14C-oleate organ uptake (B) and 3H-DOG organ uptake (C) in Apoa5−/− mice and FVB-WT controls (n = 6–7; 2–3 females per group). (D and E) 14C-oleate organ uptake (D) and 3H-DOG organ uptake (E) in aLPL−/− and respective aLPL+/+ control mice (WT) (n = 8; 3–4 females per group). Mean ± SEM, Student’s t test, ∗p < 0.05 mock versus FGF21, +p < 0.05 aLPL−/− versus equally treated WT control. Cell Metabolism 2016 23, 441-453DOI: (10.1016/j.cmet.2016.01.006) Copyright © 2016 Elsevier Inc. Terms and Conditions

Figure 7 FGF21-Induced TRL Disposal Is Mediated by Activated BAT in Obese Mice (A–E) An OGFTT (for schedule, see Figure 6) using radioactive tracers was performed in DIO mice (n = 7 per group) housed at 22°C or at thermoneutrality (30°C). DIO mice were acutely or chronically (10 days) treated with FGF21. (A and B) Plasma glucose during OGFTT. (C) Plasma insulin at 120 min. (D) 14C-oleate organ uptake. (E) 3H-DOG organ uptake. Mean ± SEM, Student’s t test, ∗p < 0.05 mock versus FGF21. Cell Metabolism 2016 23, 441-453DOI: (10.1016/j.cmet.2016.01.006) Copyright © 2016 Elsevier Inc. Terms and Conditions