Transcription Factor EB Controls Metabolic Flexibility during Exercise

Abstract

The transcription factor EB (TFEB) is an essential component of lysosomal biogenesis and autophagy for the adaptive response to food deprivation. To address the physiological function of TFEB in skeletal muscle, we have used muscle-specific gain- and loss-of-function approaches. Here, we show that TFEB controls metabolic flexibility in muscle during exercise and that this action is independent of peroxisome proliferator-activated receptor-γ coactivator1α (PGC1α). Indeed, TFEB translocates into the myonuclei during physical activity and regulates glucose uptake and glycogen content by controlling expression of glucose transporters, glycolytic enzymes, and pathways related to glucose homeostasis. In addition, TFEB induces the expression of genes involved in mitochondrial biogenesis, fatty acid oxidation, and oxidative phosphorylation. This coordinated action optimizes mitochondrial substrate utilization, thus enhancing ATP production and exercise capacity. These findings identify TFEB as a critical mediator of the beneficial effects of exercise on metabolism.

Keywords: PGC1alpha; TFEB; autophagy; diabetes; exercise; glucose; insulin; metabolic flexibility; mitochondria; mitochondrial fusion.

Graphical abstract

Figure 1

TFEB Induces Mitochondrial Biogenesis (A) Genes involved in lipid and glucose metabolism that were affected by TFEB expression are shown in colored circles. The upregulated (red circles) and downregulated genes (green circles) are shown in TFEB-overexpressing and TFEB KO, respectively. The genes were divided into functional sub-categories. (B–D) EM analysis of skeletal muscles from WT mice transfected with AAV2.1-GFP (B) and AAV2.1-TFEB (C), and from TFEB KO mice (D). Normal mitochondria are indicated by black arrows (B and C), while abnormal mitochondria in TFEB KO muscle (D) are indicated by empty arrows. Accumulation of glycogen is shown by white arrows in (C). The scale bars represent 500 nm. (E and F) Morphometrical analyses of mitochondrial number (E) and size (F) in TFEB-overexpressing and TFEB KO muscles compared to controls. Error bars represent mean ± SE, n = 3; ^∗∗p < 0.01. (G) mtDNA analysis of muscles infected with AAV2.1-TFEB and from TFEB KO compared to WT mice. Quantitative real-time PCR of mtDNA copy numbers. Data are shown as mean ± SE, n = 3; ^∗∗p < 0.01. (H) Quantification of abnormal mitochondria in TFEB KO gastrocnemius (GCN) muscles compared to WT muscles. Data are expressed as percentage of abnormal mitochondria on total mitochondria of 16 pictures (for detailed description of the criteria, see Experimental Procedures). Error bars represent mean ± SE; ^∗p < 0.05.

Figure 2

TFEB Regulates Mitochondrial Function (A) Quantitative real-time PCR of genes related to mitochondrial biogenesis and function in muscles of WT (blue bar), TFEB-overexpressing (red bar), and TFEB KO (green bar) mice. Data were normalized for GAPDH and expressed as fold induction relative to the WT. Data are expressed as mean ± SE, n = 3; ^∗p < 0.05, ^∗∗p < 0.01. (B) ChIP analysis in muscles from TFEB-overexpressing and WT mice. The CLEAR elements of NRF1 and NRF2 promoters are depicted as red boxes. TSS (transcriptional start site) indicates the first codon. The histograms show the amount of immunoprecipitated DNA as detected by quantitative real-time PCR assay. Data represent mean ± SE of three independent experiments; ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001. (C) Western blots of mitochondrial respiratory chain proteins. Blot is representative of four independent experiments. (D) Densitometric quantification of blots shown in (B). Data are shown as means ± SE (n = 4); ^∗p < 0.05. (E) Mitochondrial respiratory chain activity in muscles of WT, AAV2.1-TFEB transfected, and TFEB KO mice. Data are expressed as a percentage relative to the WT. Data are expressed as mean ± SE, n = 4; ^∗p < 0.05, ^∗∗p < 0.01. (F) H&E, succinate dehydrogenase (SDH), and cytochrome oxidase (COX) staining of AAV2.1-GFP, AAV2.1-TFEB transfected, and muscles from TFEB KO mice. The scale bars represent 100 μm. Magnification, 20×. (G) ATP production rate per gram of tissue from TFEB-overexpressing and TFEB KO mice compared to controls. Data are shown as means ± SE (n = 5); ^∗∗p < 0.01, ^∗∗∗p < 0.001. (H) Mitochondrial membrane potential of myofibers isolated from FDB muscles of WT and TFEB KO mice. Where indicated, 6 μM oligomycin or 4 μM FCCP were added. Each trace represents the tetramethylrhodamine methyl ester (TMRM) fluorescence of a single fiber. The fraction of myofibers with depolarizing mitochondria is indicated for each condition. (I) Densitometric quantification of the carbonylated proteins extracted from WT and TFEB KO mice. Values are mean ± SEM (n = 4; ^∗p < 0.05). A representative immunoblot for carbonylated proteins is shown in Figure S3C.

Figure 3

Exercise Induces TFEB Expression and Nuclear Localization in PGC1α^−/− Muscle (A) TFEB mRNA levels of PGC1α^−/− muscles infected with AAV2.9 control virus (light gray) or AAV2.9-TFEB (dark gray). Data were compared with TFEB mRNA level of WT mice (white). The levels of TFEB were measured in sedentary condition and post-exercise as indicated. Error bars represent mean ± SE for n = 3; ^∗p < 0.05, ^∗∗∗p < 0.001. (B) TFEB immunohistochemical analysis of PGC1α^−/− muscles from sedentary and exercised mice. GCN muscle cryosections were immunostained by using anti-TFEB antibody. Control means endogenous TFEB in muscles infected with AAV2.9 control virus. TFEB means the transgene TFEB in muscles infected with AAV2.9-TFEB. Arrows indicate exercise-induced TFEB nuclear localization. The scale bars represent 50 μm. (C) Analysis of genes related to mitochondrial biogenesis in PGC1α^−/− skeletal muscle infected with AAV2.9 control or with AAV2.9-TFEB virus. The mRNA levels were measured before and after acute exercise, as indicated. Data are shown as mean ± SE, n = 3; ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001. (D and E) Muscle fatigue and physical performance are improved by TFEB expression in PGC1α-deficient muscle. (D) The mice were systemically injected with AAV2.9-TFEB or AAV2.9 control virus. After 3 weeks from the virus injection, mice were subjected to run until exhaustion, as described in the Experimental Procedures. Data are shown as mean ± SE, n = 6; ^∗p < 0.05. (E) Soleus muscles were transected by intramuscular injection of AAV2.1-TFEB or AAV2.1 control virus, which resulted in 100% transfection efficacy. Index for fatigue means tetanic muscle force generated after 100 s of stimulation divided by the maximal tetanic tension produced during the fatigue protocol. Data are shown as mean ± SE, n = 4; ^∗p < 0.05, ^∗∗∗p < 0.001.

Figure 4

Exercise Induces TFEB Nuclear Translocation (A) GCN tissue sections of TFEB KO and WT mice were immunostained with anti-TFEB antibody and counter-stained with hematotoxylin as indicated. Arrows indicate nuclear localization of endogenous TFEB in WT mice after an acute bout of exercise. The scale bars represent 50 μm. (B) Endogenous TFEB was immunostained in muscle cryosections of sedentary, 4 days mild exercised, and 7 weeks intense trained WT mice. Arrows indicate TFEB nuclear localization. (C) Expression analysis of genes related to mitochondrial biogenesis and mitochondrial respiratory chain in WT skeletal muscle before and after acute exercise. Data are shown as mean ± SE, n = 3; ^∗p < 0.05, ^∗∗p < 0.01.

Figure 5

TFEB Controls Energy Balance in Skeletal Muscle (A) High-intensity exhaustive exercise. To determine exercise capacity, mice were run on a treadmill. During high-intensity exercise, transgenic mice (red) ran more, while TFEB KO mice (gray) ran half as much as WT mice (white). Data are shown as mean ± SE, n = 10; ^∗∗p < 0.01. (B and C) Energy expenditure (B) and RER (C) were determined during exercise in TFEB KO and WT mice. The mice ran at a fixed speed of 10 m/min and an incline of 20°. The figure shows the mean RER measured at peak oxygen consumption. Data were transformed by Blom’s method to obtain both normally distributed data and normally distributed residual. Two-way ANOVA was used for comparison of RERs in the two groups during the entire resting period, whereas t test was used for comparison of individual time points, n = 8. (D–F) Blood and muscle metabolites before and after high-intensity exercise in WT (white), TFEB transgenic (red), and TFEB KO mice (gray). Data are shown as mean ± SE, n = 8; ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001. (G) Enzymatic quantification of muscle glycogen before and after high-intensity exercise in WT (white), TFEB transgenic (red), and TFEB KO mice (gray). Data are shown as mean ± SE, n = 8; ^∗∗p < 0.01. (H) Periodic acid-Schiff (PAS) staining of cryosections from AAV2.1-GFP, AAV2.1-TFEB transfected, and TFEB KO mice. Inserts show PAS staining after glycogen breakdown. The scale bars represent 100 μm. (I and J) Quantitative analysis of serum (I) and muscle (J) NEFA before and after high-intensity exercise in TFEB KO (gray) and WT mice (white). Data are shown as mean ± SE, n = 8; ^∗p < 0.05, ^∗∗p < 0.01. (K) Enzymatic quantification of B-hydroxybutyrate before and after high-intensity exercise in TFEB KO (gray box) and WT mice (white box). Data are shown as mean ± SE, n = 8; ^∗p < 0.05. (L) TFEB ablation results in a decreased rate of insulin-stimulated GIR. EU clamps were used to assess whole-body insulin sensitivity by determining the GIR required to maintain euglycemia in WT (white) and TFEB KO (gray) mice. Data represent the means ± SE from five to six individual mice per group; ^∗p < 0.05. (M) TFEB deletion results in a reduction of insulin-stimulated glucose uptake in muscle. Insulin-stimulated glucose uptake into muscle tissues was determined by 2-deoxy-d-[1-¹⁴C]glucose injection during the last 35 min of insulin infusion during the EU clamp. These data represent the means ± SE of five to six mice per group; ^∗p < 0.05. WT, white; TFEB KO, gray. (N) The amount of glucose conversion to glycogen in GCN muscles from WT (white) and TFEB KO (gray) mice was determined during the EU clamp by infusion of [3-³H]glucose. Data represent the means ± SE from five to six mice per group; ^∗p < 0.05. (O and P) TFEB KO mice do not show any significant difference in glucose homeostasis of liver or adipose tissue. (O) Glucose uptake into adipose tissue (EPI) and (P) insulin suppression of hepatic glucose output (HGP, liver) were determined during the EU clamp by infusion of 2-deoxy-d-[1-¹⁴C]glucose or [3-³H]glucose infusion into WT mice (white) and TFEB KO (gray) mice. These data represent the means ± SE from five to six mice per group.

Figure 6

TFEB Controls Genes Related to Glucose Metabolism (A and C) Quantitative real-time PCR of glucose uptake and glycogen biosynthesis-related genes in GCN muscles infected with AAV2.1-TFEB (black bar) and in TFEB transgenic muscles (red bar) compared with WT (white bar) muscles. Data were normalized for GAPDH and expressed as fold induction relative to the WT. Data are shown as mean ± SE, n = 3; ^∗p < 0.05, ^∗∗p < 0.01. (B) Western blot of GLUT1 and densitometric quantification. Values are normalized for GAPDH; ^∗p < 0.05. (D) Western blot analysis of total and phosphorylated GYS from extracts of GCN. Representative blot images are shown (left panel). Densitometric quantification is depicted on the right panel. Data are shown as mean ± SE, n = 3; ^∗p < 0.05. (E–G) TFEB regulates glycogen synthesis independently of PGC1α. (E) Quantitative real-time PCR of genes related to glucose uptake and glycogen biosynthesis in PGC1α^−/− and WT mice that were infected with AAV2.1-GFP (white) or AAV2.1-TFEB (black). Data are shown as mean ± SE, n = 3; ^∗p < 0.05, ^∗∗p < 0.01. (F) PAS staining of cryosections from PGC1α^−/− muscles that were transfected by AAV2.1-GFP or AAV2.1-TFEB. The scale bars represent 100 μm. (G) Enzymatic quantification of basal glycogen levels in PGC1α^−/− and WT muscle, infected with AAV2.1-GFP (white) or AAV2.1-TFEB (black).

Figure 7

TFEB Also Controls Glucose Uptake via Regulation of AMPK Signaling Pathways (A) Quantitative real-time PCR of nNOS transcript in WT (white bar), TFEB-overexpressing (black), and TFEB KO muscles (gray). Values are expressed as fold induction compared to controls. Data are shown as mean ± SE, n = 8; ^∗p < 0.05, ^∗∗p < 0.01. (B) Western blot of nNOS and densitometric quantification. Value are normalized for GAPDH; ^∗p < 0.05; ^∗∗∗p < 0.001. (C) ChIP analyses of muscles from TFEB transgenic and WT mice on nNOS, GLUT1, and GLUT4 promoters. The histograms show the amount of immunoprecipitated DNA as detected by quantitative real-time PCR assay. Data represent mean ± SE of three independent experiments; ^∗p < 0.05. (D) Western blot of AMPK, PAMPK, PACC, and PAKT performed on protein extracts from GCN muscles of AAV2.1-TFEB transfected, TFEB KO, or WT mice. Representative images (left) and densitometric quantification (right) are shown; ^∗p < 0.05.