Cytosolic ROS production by NADPH oxidase 2 regulates muscle glucose uptake during exercise

Abstract

Reactive oxygen species (ROS) act as intracellular compartmentalized second messengers, mediating metabolic stress-adaptation. In skeletal muscle fibers, ROS have been suggested to stimulate glucose transporter 4 (GLUT4)-dependent glucose transport during artificially evoked contraction ex vivo, but whether myocellular ROS production is stimulated by in vivo exercise to control metabolism is unclear. Here, we combined exercise in humans and mice with fluorescent dyes, genetically-encoded biosensors, and NADPH oxidase 2 (NOX2) loss-of-function models to demonstrate that NOX2 is the main source of cytosolic ROS during moderate-intensity exercise in skeletal muscle. Furthermore, two NOX2 loss-of-function mouse models lacking either p47phox or Rac1 presented striking phenotypic similarities, including greatly reduced exercise-stimulated glucose uptake and GLUT4 translocation. These findings indicate that NOX2 is a major myocellular ROS source, regulating glucose transport capacity during moderate-intensity exercise.

Fig. 1

Moderate-intensity exercise causes a pro-oxidative shift in human and murine muscle. a 2′,7′-dichlorodihydrofluorescein diacetate (DCFH) oxidation (DCF, n = 3)) and b exercise signaling in human vastus lateralis before and after a bout of moderate-intensity cycling (30 min, 65% maximal power output) in young-healthy volunteers (n = 3). c Representative images and quantification of the exercise-stimulated (20 min, 65% maximal running speed) DCFH oxidation in tibialis anterior muscle from WT and ncf1* mice (n = 5 per group). d Exercise signaling in quadriceps muscle from WT and ncf1* mice (n = 6 resting and n = 8 exercising group). e Total proteins levels of p38 MAPK, ERK 1/2, Total alpha2 AMPK, Total ACC, and Coomassie staining as loading control. For a paired t-test was performed for statistical analysis, * denotes p < 0.05 compared to resting condition. For c, d a two-way ANOVA was performed to test for effects of exercise (Exer) genotype (Geno), and interaction (Int), followed by Tukey’s post hoc test with correction for multiple comparisons. *^, **^, ***  Denotes p < 0.05,  p < 0.01,  and p < 0.001, respectively, for main effects/interactions. ^#, ### Denotes p < 0.05 and p < 0.001 compared to the WT group. Individual values and mean ± standard error of the mean (SEM) are shown. Scale Bar = 50 µm. For b western blots for total proteins are shown in Supplementary Fig. 1, uncropped blot and quantifications are shown in the Source data file. For c–e, source data are provided in the Source Data file

Fig. 2

NOX2 is a major ROS source during exercise in skeletal muscle. Subcellularly targeted redox-sensitive GFP2 (roGFP2) were electroporated in ncf1* and p47roGFP in inducible muscle-specific Rac1 mice. Representative image and quantification of a Mito-roGFP2-orp1 (WT, n = 4 and ncf1* = 3), b cyto-roGFP2-Orp1 in Flexor Digitorum Brevis fibers (WT, n = 3–4 and ncf1* = 3) and c p47roGFP oxidation in tibialis anterior muscle in WT and Rac1 imKO mice (WT, n = 4/8 for rest/exercise, Rac1 imKO, n = 3/8 for rest/exercise groups). Two-way ANOVA was performed to test for effects of exercise (Exer) genotype (Geno), and interaction (Int), followed by Tukey’s post hoc test  with correction for multiple comparisons. * Denotes p < 0.05 for main effects/interaction. #, ### Denotes p < 0.05 and p < 0.001 compared to the WT group. Individual values and mean ± SEM are shown. Scale Bar = 10 µm (a and b)/30 µm (for c). For a–c, source data and p values are provided in the Source Data file

Fig. 3

Exercise-stimulated glucose uptake is reduced in ncf1* mice. Different parameters were compared between p47phox-mutated mice (ncf1*) and wild-type (WT) mice. a Respiratory exchange ratio (RER) during light and dark cycles and b habitual activity (n = 8 per group). c–e In vivo running-induced (20 min, 65% maximal running speed) 2-deoxy-glucose uptake (2DG) uptake in quadriceps (Quad) soleus and tibialis anterior (TA) muscles from WT and ncf1* mice (n = 12–16). f Maximal running speed (n = 15 per group), g blood glucose concentration after rest/exercise (n = 8–12), h NAD⁺, i NADPH levels in quadriceps muscle (n = 6–8 for rest/exercise groups), and j Plasma lactate concentration (n = 6–8 for rest/exercise groups). Unpaired t-test f and two-way ANOVA were performed to test the effects of exercise (Exer), genotype (Geno), and interaction (Int), followed by a Tukey’s post hoc test with  correction for multiple comparisons. *^, **^, **** Denotes p < 0.05, p < 0.01, p < 0.0001, respectively, for main effects/interaction. ^{#, ##} Denotes p < 0.05 and p < 0.01, respectively compared to the WT group. Individual values and mean ± SEM are shown. For all Items, source data and p values are included in the Source Data file

Fig. 4

Ncf1* mice show similar muscle size, fiber-type composition, mitochondrial content and capillary density compared to WT group. a Fiber size-related parameters in WT and ncf1* mice were determined using wheat germ agglutinin (WGA) staining. b Myofiber cross-sectional area, and c minimum (min.) Ferret diameter. d Quantification and e representative merged image of fiber-type staining where Type I fibers (blue), Type IIa (green), Type IIb (red), Type IIx (non-stained) in WT (n = 6), and ncf1* (n = 5) TA muscles. Subunits of mitochondrial oxidative phosphorylation complexes were determined in f quadriceps and g soleus muscle lysates (n = 14 per group). h Capillary density was estimated by PECAM immunostaining in tibialis anterior sections. Two-way ANOVAs was performed to test for effects of exercise (Exer), genotype (Geno), and interaction (Int), followed by Tukey’s post hoc test with correction for multiple comparisons. Individual values and mean ± SEM are shown. Scale Bar = 100 µm. For b, c, d, and h, source data and p values are provided in the Source Data file. For f and g, uncropped blots are provided in the Source Data file

Fig. 5

Exercise-stimulated GLUT4 translocation requires NOX2 activity. a the GLUT4-myc-GFP construct was electroporated into tibialis anterior muscles of both WT and ncf1* mice. Non-permeabilized muscle fibers from exercised (20 min, 65% maximal running speed) or resting mice were stained with anti-myc antibody and imaged by confocal microscopy (n = 4 mice, a minimum of 10 fibers/muscle). Total endogenous GLUT4 and HK II were determined by western blot in the following muscles b quadriceps, c soleus, and d TA muscles in WT and ncf1* mice (n = 14). Unpaired t-test (b–d) and two-way ANOVA were performed to test for effects of exercise (Exer), genotype (Geno), and interaction (Int), followed by a Tukey’s post hoc testwith correction for multiple comparisons. * Denotes p < 0.05 for main effect. # Denotes p < 0.05 compared to the WT group. Individual values and mean ± SEM are shown. Scale bar = 20 µm. For a, source data and p values are provided in the Source Data file. For b–d, uncropped blots are provided in the Source Data file