β-Actin shows limited mobility and is required only for supraphysiological insulin-stimulated glucose transport in young adult soleus muscle

Abstract

Studies in skeletal muscle cell cultures suggest that the cortical actin cytoskeleton is a major requirement for insulin-stimulated glucose transport, implicating the β-actin isoform, which in many cell types is the main actin isoform. However, it is not clear that β-actin plays such a role in mature skeletal muscle. Neither dependency of glucose transport on β-actin nor actin reorganization upon glucose transport have been tested in mature muscle. To investigate the role of β-actin in fully differentiated muscle, we performed a detailed characterization of wild type and muscle-specific β-actin knockout (KO) mice. The effects of the β-actin KO were subtle; however, we confirmed the previously reported decline in running performance of β-actin KO mice compared with wild type during repeated maximal running tests. We also found insulin-stimulated glucose transport into incubated muscles reduced in soleus but not in extensor digitorum longus muscle of young adult mice. Contraction-stimulated glucose transport trended toward the same pattern, but the glucose transport phenotype disappeared in soleus muscles from mature adult mice. No genotype-related differences were found in body composition or glucose tolerance or by indirect calorimetry measurements. To evaluate β-actin mobility in mature muscle, we electroporated green fluorescent protein (GFP)-β-actin into flexor digitorum brevis muscle fibers and measured fluorescence recovery after photobleaching. GFP-β-actin showed limited unstimulated mobility and no changes after insulin stimulation. In conclusion, β-actin is not required for glucose transport regulation in mature mouse muscle under the majority of the tested conditions. Thus, our work reveals fundamental differences in the role of the cortical β-actin cytoskeleton in mature muscle compared with cell culture.

INTRODUCTION

In muscle and fat cells, the dynamic insulin-stimulated reorganization of the cortical actin cytoskeleton is a well-described crucial step in the insulin-stimulated translocation of vesicles containing glucose transporter 4 (GLUT4) to facilitate glucose transport (6, 9). The mechanistic involvement of actin remodeling in GLUT4 translocation has been proposed at multiple steps, including vesicle tethering, movement, and fusion, in addition to localization of signaling proteins (6, 9). A disturbed regulation of the actin cytoskeleton may furthermore be a unifying theme underlying the reduced insulin-stimulated GLUT4 translocation in muscle cells made insulin resistant by various methods (27, 77). Accordingly, muscle insulin resistance correlates with changes in actin regulating signaling by proteins such as Rac1, PAK1, and ROCK1 and/or with altered actin morphology in a number of transgenic mouse models, in cell studies (18, 68, 69, 78, 79, 81, 84), and in type 2 diabetic humans (12, 68). The inhibition of the PtsIns3 phosphate-5-kinase PIKFyve, suggested to regulate actin morphology and GLUT4 translocation in cell culture (63), also reduces insulin- and contraction-stimulated glucose transport when inhibited in mature rodent muscle (25, 36). Furthermore, pharmacological depolymerization of the actin cytoskeleton by latrunculin B reduces insulin- and contraction-stimulated glucose transport into mouse muscles (7, 69, 70). Finally, a study in isolated mature feline cardiomyocytes overexpressing green fluorescent protein (GFP)-tagged β-actin used fluorescence recovery after photobleaching (FRAP) imaging to conclude that the mobility of β-actin is increased following insulin stimulation (1). Taken together, these studies suggest that the requirement and function for actin is conserved between muscle cell culture and fully differentiated skeletal muscle. However, global disruption of the actin cytoskeleton using latrunculin B is a very crude tool to investigate the involvement of actin in glucose transport regulation and cannot establish the exact actin isoform involved. No studies, to our knowledge, have investigated isoform-specific actin dynamics in fully differentiated skeletal muscle.

The mammalian actin family of proteins encompasses four muscle-specific isoforms (α_smooth, γ_smooth, α_cardiac and α_skeletal) and the two ubiquitously expressed cytoplasmic β- and γ-actin isoforms (8, 31). The skeletal muscle actin pool is dominated by the highly expressed α_skeletal isoform that forms the thin filaments of the contractile apparatus compared with much lower expression of the cytoplasmic β- and γ-actins. The latter two actin isoforms are 99% homologous, differing by only 4 amino acids in their NH₂ termini. However, although isolated β- and γ-actin can produce hybrid actin filaments in vitro (5, 41), suggesting that they may substitute for one another, the in vivo functions of β- and γ-actin are likely nonredundant, since 1) evidence suggests that β- and γ-actin localize to distinct compartments and form distinct actin structures within various cell types (8, 31), and 2) the muscle-specific knockout (KO) of either isoform produces a clear phenotype (57, 65). From the literature, most evidence in different cell types, including the muscle cell lines C2 myoblasts and L6 myoblasts and myotubes, points toward β-actin being the main constituent of the cortical actin cytoskeleton previously linked to GLUT4 translocation (22, 23, 30, 35, 64, 77). To directly investigate whether skeletal muscle β-actin regulates whole body metabolism and specifically muscle glucose transport in mice, we obtained and compared in vivo and ex vivo metabolic parameters in wild-type and muscle-specific β-actin KO mice.

METHODS

Animals.

The β-actin floxed mice were a kind gift from James M. Ervasti (Dept. of Biochemistry, Molecular Biology, and Biophysics, Univ., of Minnesota, Minneapolis, MN) (54). The skeletal muscle-specific KO mice were littermates from breeding of hemizygous human α-skeletal actin (HSA)-Cre and β-actin^flox/flox mice (57). To control for the progressive myopathy that develops in the β-actin KO mice (57), a young adult (8–13 wk old) and mature adult (18–23 wk old) cohort of mice was investigated for each sex. The liver kinase B1 (LKB1) muscle-specific KO mice were generated as previously described (75). Results from male and female mice were pooled unless noted, since they showed comparable phenotypes. All mice used at the University of Copenhagen were housed according to the Danish experimental animal legislation, while the wild-type C57BL/6 mice (Jackson Laboratories) used for GFP-actin transfection and confocal microscopy in National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS) were housed according to the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Likewise, all animal experiments were approved by the Danish National Animal Experimental Inspectorate or by the NIAMS Animal Care and Use Committee and conducted in agreement with the Declaration of Helsinki.

Genotyping.

(Muscle 5–10 mg) was lysed overnight at 55°C in 100 µl of DirectPCR Lysis Reagent (Tail) (250-101-T, Nordic BioSite) with freshly added 0.2 mg/ml proteinase K (Roche Diagnostics). The lysate was centrifuged at 1,250 g for 5 min at room temperature, and supernatant was kept for DNA extraction. The released DNA was diluted fivefold in dilution buffer (10 mM TRIS, 1 mM EDTA, pH 8.0) and incubated for 15 min at 95°C. Five microliters of this dilution was then amplified in a 25-μl SYBR Green polymerase chain reaction (PCR) containing 1× Quantitect SYBR Green Master Mix (Qiagen) and 200 nM of each primer (gDNA: TCCAGCACGTGGGTCTTAGAGG and CTGGCAGGTAGGCTCAGCAGGT, β-actin flox: TCTCAGCCACGCCCTTTCTCATA and TGGGAGAACGGCAGAAGAAAGA, liver kinase B1 (LKB1) flox: GGCATGGCCTCACAGGGACTT and CGGCAGTACCGGCATAACCAAG). The amplification was monitored in real time using the MX3005P real-time PCR machine (Stratagene). All reactions were performed in triplicates, and the mean C_T was used for calculations. Flox abundance was calculated as 2^(C_{T Flox} − C_{T gDNA}) and normalized to the geometric average of the Cre-negative mice.

Immunoblotting.

Lysate protein concentrations were determined by using BSA standards (Pierce) and bicinchoninic acid assay reagents (Pierce). Total protein and phosphorylation levels of relevant proteins were determined by standard immunoblotting techniques, loading equal amounts of protein. The primary antibodies used were phospho- (p)-Akt Ser⁴⁷³ (9271S), p-Akt Thr³⁰⁸ (9275S), p-AMP-activated protein kinase (AMPK) Thr¹⁷² (2535S), p-acetyl-CoA carboxylase 2 (ACC2) Ser²¹² (3661S), p-eukaryotic elongation factor 2 (eEF2) Thr⁵⁶ (2331S), p-p38 mitogen-activated protein kinase (MAPK) Thr¹⁸⁰/Tyr¹⁸² (9211S), p- Unc-51-like autophagy-activating kinase 1 (ULK1) Ser⁵⁵⁵ (5859S), hexokinase II (2867S) (all from Cell Signaling Technology), Rac1 (05-389, Millipore), GLUT4 (Pierce), β-actin (AC-15, Santa Cruz Biotechnology), dystrophin (ab15277, Abcam), LKB1 (Ley37D/G6, Santa Cruz Biotechnology), sarco/endoplasmic reticulum Ca²⁺-ATPase 1 (SERCA1; A-6, Santa Cruz Biotechnology), and SERCA2 (F-1, Santa Cruz Biotechnology). Membranes were blocked for 1 h at room temperature in TBS-Tween 20 (YBST) containing either 2% skimmed milk or 3% BSA and incubated overnight at 4°C with primary antibody. The membranes were incubated in the corresponding horseradish peroxidase-conjugated secondary antibody for 45 min at room temperature and washed in TBST before visualization of the proteins (ChemiDoc MP Imaging System, Bio-Rad).

GFP-β-actin transfection and fiber isolation.

The p-EGFP-actin plasmid, encoding a fusion protein consisting of the red-shifted, human codon-optimized variant of EGFP and the gene encoding human cytoplasmic β-actin, was a kind gift from A. Klip (no. 6116-1, Clontech Laboratories). For transfections, mouse flexor digitorum brevis (FDB) muscles were electroporated in vivo according to a previously described protocol (15), with a few modifications. Briefly, mice were given a subcutaneous injection of buprenorphine-HCl (0.05 mg/kg) as analgesic and anesthetized using 4% isoflurane. Ten microliters of 0.36 mg/ml hyaluronidase (H3884, Sigma) was injected through the skin at the heel into the footpad. After 1 h, 20 µl of 0.25 mg/ml p-EGFP-actin plasmid was injected in the same place. Six pulses of 20 ms were delivered at 1 Hz with the voltage amplitude adjusted to yield an electric field of ~75 V/cm using acupuncture needles (0.20 × 25 mm, Tai Chi, Lhasa OMS) connected to an ECM 830 BTX electroporator (BTX Harvard Apparatus). After 5–7 days, the animals were euthanized with CO₂, and individual FDB fibers were isolated and cultured as described previously (44). After overnight culturing, fibers were kept in either the basal state or stimulated with 60 nM insulin for 30 min before imaging, unless otherwise stated.

GFP-β-actin transfection and antibody staining of cultured C2 myoblasts.

C2 myoblasts were kept in DMEM containing 25 mM glucose and 1 mM sodium pyruvate supplemented with 20% fetal bovine serum. The cells were transfected with the pEGFP-actin plasmid by using FuGENE HD transfection reagent (Promega) according to manufacturer’s protocol. For fixation, cells were incubated for 30 min in PBS containing 4% formaldehyde (FA) and 0.1% Triton X-100 followed by 5-min incubation in methanol. Cells where blocked in 1% goat serum and 2% BSA for 2 h, followed by 2-h incubation in anti-β-actin (AC-15, Ambion, Life Technologies). After three washes in PBS, the cells were incubated with anti-mouse Alexa 568 (Life Technologies) for 2 h; 0.04% saponin was used throughout blocking and antibody incubation.

Imaging and image analysis.

All imaging experiments were carried out in the Light Imaging Section of NIAMS (Bethesda, MD). Confocal images were collected using a ×63 1.4 NA oil immersion objective lens or a ×10 0.45 NA objective lens on a LSM 780 confocal microscope (Zeiss) driven by Zen 2011. For live-imaging, fibers were imaged in DMEM with 5% horse serum while kept at 5% CO₂ in air and 37°C by a Pecon Laboratory-Tek S1 heat stage and an objective heater. For measurements of intensity accumulation, regions of interest (ROIs) were drawn around striations before insulin stimulation, and the signals in the ROIs were quantified relative to the total signal before and 10 min after insulin stimulation. For FRAP sequences, a ROI surrounding the cortical cytoskeleton in close proximity to the nucleus was manually drawn. Fibers were bleached for 10 iterations with 100% laser power and a ×63 objective, which yielded bleaching of 50–80% of the fluorescent molecules in the ROI. Some samples were fixed before imaging as a negative control by incubation with 2% FA in PBS for 20 min at room temperature.

Raw intensity values from FRAP sequences were exported and processed in the MATLAB plug-in easyFRAP (60). Images were exported in 8-bit TIF format, cropped, and linearly adjusted using Photoshop CC 2015 (Adobe).

MRI scan and weight.

Mice were weighed and MRI scanned (Echo MRi, Echo Medical Systems) to determine the body composition.

Glucose tolerance tests.

Prior to an intraperitoneal glucose tolerance test (IPGTT, 2 g/kg body wt d-glucose) mice were fasted for 5 h from 8:00 AM. Blood glucose concentration was measured in duplicate from the tail vein at time points 0, 20, 40, 60, and 90 min. A 50-µl blood sample was collected after 20 min for determination of the plasma insulin levels. Plasma insulin levels were determined using the Mouse Ultrasensitive Insulin ELISA kit (ALPCO Diagnostics).

Maximal running tests.

Mice were acclimated to the treadmill three times (11 min at 0.12 m/s, 60s at 0.38 m/s) (Treadmill TSE Systems) before the maximal running tests. The test started at 0.17 m/s for 300 s with 0% incline, followed by a continuing increase (0.15 m/s) in running speed every 60 s. The individual maximal running capacity was determined as the point where the mouse failed to run and stayed on the shocker for at least 5 s. All running tests were carried out as blinded experiments.

Indirect calorimetry.

Mice were singly housed and acclimated for 9 days and then singly housed in a 16-chamber indirect calorimetry system (PhenoMaster, TSE Systems) for 9 days. Baseline (72 h), fasting (24 h), refeed (72 h), and high-fat diet (48 h) conditions were included. Oxygen consumption rate (V̇o₂, ml·h⁻¹·kg⁻¹), carbon dioxide production rate (V̇co₂, ml·h⁻¹·kg⁻¹), respiratory exchange ratio (RER), total activity (beam breaks), food, and water intake were measured simultaneously for each mouse.

Muscle incubations.

Soleus and extensor digitorum longus (EDL) muscles were excised from anesthetized mice (6 mg pentobarbital + Lidocaine/100 g body wt ip) and suspended in the incubation chambers (Multi Myograph System, Danish Myo-Technology) containing 30°C Krebs-Ringer-Henseleit (KRH) buffer with 8 mM mannitol and 2 mM pyruvate, as previously described (28). For insulin stimulation, muscles were preincubated for 30 min followed by 20 min of insulin stimulation. For submaximal insulin stimulation, soleus and EDL muscles were incubated with 1.8 and 3 nM insulin (Actrapid, Novo Nordisk), respectively, whereas maximal stimulation was 60 nM insulin for both muscles (40). For 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR) experiments, the muscles were preincubated for 30 min and subsequently stimulated with 4 mM AICAR for 40 min. Contractions were induced by electrical stimulation every 15 s with 2-s trains of 0.2-ms pulses delivered at 100 Hz (~30 V) for 15 min. Passive stretch was induced by manually stretching the muscles to a tension of 100 mN for 5 min followed by adjustment to 130 mN for 10 min. Prior to contraction and passive stretch the muscles were pre-incubated for 45 min.

2-Deoxyglucose transport.

2-Deoxyglucose (2-DG) transport was measured for the last 10 min during all incubation protocols using ³H and ¹⁴C radioactive tracers, as described previously (28).

Muscle analyses.

Immediately after ex vivo incubation, the muscles were washed in ice-cold KRH medium and snap-frozen in liquid nitrogen. Tissue was homogenized for 1 min at 30 Hz using steel beads and a Tissue Lyser II bead-mill (Qiagen) in ice cold lysis B buffer (0.05 M Tris Base, pH 7.4, 0.15 M NaCl, 1 mM EDTA, and EGTA, 0.05 M NaF, 5 mM sodium pyrophosphate, 2 mM sodium orthovanadate, 1 mM dithiothreitol, and benzamidine (BZM), 0.5% protease inhibitor cocktail (P8340, Sigma-Aldrich), and 1% NP-40. Supernatants were collected after end-over-end rotation for 30 min at 4°C and centrifugation (18,327 g) for 20 min at 4°C.

Statistical analyses.

Results are presented as means ± SE. Statistical testing was performed using unpaired two-tailed t-test and one- or two-way analysis of variance (ANOVA). Repeated-measures ANOVA was used for data with repeated sampling in the same animal, including the glucose tolerance tests as well as the indirect calorimetry data. Tukey’s post hoc test was used to identify differences if significant (P < 0.05) ANOVA effects were found. The statistical analyses were carried out in SigmaPlot v. 11.0.

RESULTS

Validation of the muscle-specific β-actin KO mouse model.

To study the role of β-actin in skeletal muscle glucose metabolism, we obtained and bred a previously characterized muscle-specific β-actin KO mouse model (57). This mouse model was originally reported to show an ~50% reduction in β-actin, with no compensatory changes in other actin isoforms (57). Myonuclei constitute approximately one-half of the total nuclei in adult mouse muscle (80); hence, the ~50% reduction of the floxed allele DNA in the quadriceps, soleus, and EDL muscles of the β-actin KO compared with wild-type mice verified the muscle-specific excision of the β-actin gene (Fig. 1A). Possibly due to the low expression of β-actin in myofibers vs. nonmuscle cells in skeletal muscle tissue (76), it was not possible to show decreased β-actin protein expression in the muscle-specific KO mice (data not shown). To verify that an ~50% reduction in the floxed allele would be expected from a complete muscle-specific KO, floxed allele excision was also measured in the muscle-specific LKB1 KO mice, together with protein expression of LKB1 in quadriceps, soleus, and EDL muscles. A similar ~50% reduction of the floxed allele was observed in the muscles of the LKB1 KO compared with wild-type mice (Fig. 1B), supporting that myonuclei constitute about one-half of the total nuclei in mouse skeletal muscle. LKB1 protein expression was almost completely abolished in the muscles of the LKB1 KO mice, showing that an ~50% reduction in floxed allele is consistent with a complete KO in myofibers (Fig. 1C). Furthermore, the previously reported reduction in dystrophin protein expression in the quadriceps muscles of the β-actin KO mice (57) was confirmed (Fig. 1D). The KO validation was limited to excision of the floxed allele and could not be confirmed by reduced β-actin protein expression. However, these data strongly support that efficient muscle-specific excision of β-actin in the KO mice had occurred as intended, but genotyping at protein level is not possible due to the low expression of β-actin in muscle fibers compared with nonmuscle cells.

Fig. 1.

Verification of the β-actin knockout (KO) genotype. A: quantification of the flox allele in quadriceps, soleus, and extensor digitorum longus (EDL) muscles of wild-type (WT) and β-actin KO mice; n = 7–9 mice. All values are shown as means ± SE; ***P < 0.001. B: quantification of the flox allele in quadriceps, soleus, and EDL muscles of WT and liver kinase B1 (LKB1) KO mice; n = 4–6 mice. All values are shown as means ± SE; **P = 0.019, ***P ≤ 0.001. C: quantification of LKB1 protein expression in quadriceps, soleus, and EDL muscle lysates of WT and LKB1 KO mice; n = 5–7. All values are shown as means ± SE; ***P < 0.001. D: quantification of dystrophin protein expression in soleus, tibialis anterior (TA), quadriceps, and EDL muscle lysates of young adult WT and β-actin KO mice; n = 3–4. All values are shown as means ± SE; **P = 0.003. E: representative blots of quantified proteins in C and D. Coomassie staining is used as loading control. R.E., relative expression to WT; A.U., arbitrary units.

Young adult β-actin KO mice exhibit no overt in vivo phenotype except during repeated maximal running.

An extensive in vivo characterization of wild-type and muscle-specific β-actin KO mice was undertaken. As previously reported (57), a greater decline in maximal running performance was found in the β-actin KO mice compared with wild-type mice when repeatedly challenged with maximal running tests on consecutive days (Fig. 3A). No genotype-related differences were found for other in vivo parameters, including body composition, blood glucose, insulin levels (Table 1), substrate utilization (RER; Fig. 2), and glucose tolerance (Fig. 3, B–E). Neither did total activity, food, or water intake differ between the genotypes (data not shown). This shows that the muscle-specific β-actin KO mice are largely indistinguishable from their wild-type littermates at whole body level but do exhibit an impaired ability to recover from repeated exercise bouts.

Fig. 3.

No differences in glucose tolerance, but β-actin KO mice display decreased running performance. A: young adult female mice were subjected to a maximal running test each day for 3 days. Maximal running speed is denoted as stop speed (m/s; n = 8–10). Values\shown are means ± SE; ***P ≤ 0.001 within WT; ###P ≤ 0.001, ##P = 0.016 within KO; $P = 0.005, genotype difference. Young adult (B) and mature adult (C) male mice were fasted for 5 h and injected with 2 g/kg body wt d-glucose; n = 10. Blood glucose was measured after 0, 20, 40, 60, 90, and 120 min. Young adult (D) and mature adult (E) female mice were subjected to a similar protocol, n = 9–10. Area under the curve (AUC) of blood glucose is calculated from 1 to 120 min and shown as means ± SE. Young adult mice, 8–13 wk old; mature adult mice, 18–23 wk old.

Table 1.

No differences in selected in vivo characteristics

	Wild Type		β-Actin KO
	Young Adult (8–13 wk old)	Mature Adult (18–23 wk old)	Young Adult (8–13 wk old)	Mature Adult (18–23 wk old)
Male
Body weight, g	27.3 ± 0.61	33.0 ± 0.98	27.0 ± 0.46	33.1 ± 0.49
MRI - fat mass, g	2.42 ± 0.16	4.13 ± 0.56	2.16 ± 0.15	3.75 ± 0.36
MRI - lean mass, g	23.8 ± 0.52	28.1 ± 0.52	23.7 ± 0.51	28.4 ± 0.46
Blood glucose, mmol/l	15.5 ± 1.24	17.0 ± 1.47	17.1 ± 0.86	16.9 ± 0.82
Insulin, ng/ml	0.59 ± 0.05	1.31 ± 0.29	0.60 ± 0.05	1.25 ± 0.17
Max running test, stop speed, m/s	0.59 ± 0.02	0.56 ± 0.01	0.60 ± 0.02	0.57 ± 0.02
Female
Body weight, g	20.1 ± 0.30	28.3 ± 0.97	20.3 ± 0.24	27.6 ± 0.58
MRI - fat mass, g	2.06 ± 0.12	5.33 ± 0.59	1.97 ± 0.11	5.01 ± 0.62
MRI - lean mass, g	17.1 ± 0.27	22.3 ± 0.43	16.9 ± 0.28	22.0 ± 0.27
Blood glucose, mmol/l	15.5 ± 1.72	15.6 ± 1.11	13.9 ± 0.76	17.8 ± 1.52
Insulin, ng/ml	0.45 ± 0.01	0.71 ± 0.08	0.55 ± 0.04	0.73 ± 0.05
Max running test, stop speed, m/s	0.55 ± 0.01	0.56 ± 0.02	0.54 ± 0.02	0.56 ± 0.02

All values are shown as means ± SE. Mice were injected with 2 g/kg body wt d-glucose for the intraperitoneal glucose tolerancew test. Blood glucose values are measured 20 min post-glucose injection together with a 50-µl blood sample to measure insulin levels. Young adult mice: 8–13 wk old, mature adult mice: 18–23 wk old; n = 9–10 mice.

Fig. 2.

β-Actin KO mice display no whole body metabolic phenotype. Respiratory exchange ratio (RER) during baseline (72 h), fasting (24 h), refeed (72 h), and high-fat diet (HFD; 48 h) conditions were measured in young adult (8–13 wk old; A–D) female and mature adult (18–23 wk old; E–H) male mice; n = 8. Measurements were obtained in both young and mature adult male and female mice, showing the same pattern in both sexes with no genotype-related differences. Gray bars indicate dark hours. Values shown are means ± SE.

Young adult β-actin KO mice display decreased maximal insulin-stimulated glucose transport into soleus but not EDL muscle.

The slow-twitch oxidative soleus and the fast-twitch glycolytic EDL muscles were investigated ex vivo. In our current experiments, young adult β-actin KO mice showed a sex-independent 40% reduction in maximal insulin-stimulated glucose transport in soleus (Fig. 4A). Surprisingly, submaximal insulin-stimulated glucose transport was not reduced in the absence of β-actin (Fig. 4A), nor were there differences in insulin-stimulated glucose transport into EDL (Fig. 4A). In terms of signaling, we found no differences in insulin-related Akt phosphorylation (Fig. 4, B–C). Likewise, we found no differences in the expression of hexokinase II, GLUT4, the actin cytoskeleton-regulating GTPase Rac1, or the Ca²⁺ handling proteins SERCA1 and SERCA2 (Fig. 4, D–I). We conclude that β-actin is necessary to attain maximal rates of insulin-stimulated glucose transport in slow-twitch muscle of young adult mice, and this is likely not related to defects in Ca²⁺ handling.

Fig. 4.

2-Deoxyglucoes (2-DG) transport and insulin signaling during submaximal and maximal insulin stimulation in young adult mice. A: insulin-stimulated 2-DG transport during submaximal (soleus 1.8 nM, EDL 3 nM) and maximal (60 nM) insulin concentrations in soleus and EDL muscles from young adult (8–13 wk old) WT and β-actin KO mice. Akt Ser⁴⁷³ (B) and Akt Thr³⁰⁸ (C) phosphorylation in soleus and EDL muscles from young adult WT and β-actin KO mice, as well as total protein expression of hexokinase II (D), Rac1 (E), GLUT4 (F), sarco/endoplasmic reticulum Ca²⁺-ATPase (SERCA1; G) and SERCA2 (H) in soleus and EDL muscles from WT and β-actin KO mice; n = 10–15. I: representative blots of quantified proteins. Coomassie staining was used as loading control. All values shown are means ± SE; *P < 0.05, ***P < 0.001 vs. basal; #P < 0.05, ##P < 0.01, ###P < 0.001 max vs. submax insulin; $$$P < 0.001 genotype difference; £P < 0.05, £££P < 0.001 soleus vs. EDL.

Maximal insulin-stimulated glucose transport is not reduced in mature adult β-actin KO mice.

According to the original report, adult mice, 12–24 wk old, begin to show signs of myopathy with 5–10% centronucleated fibers in weight-bearing muscles such as quadriceps and triceps, whereas non-weight-bearing muscles including tibialis anterior, EDL, and diaphragm are much less affected (57). Since soleus is a weight-bearing postural muscle, we hypothesized that allowing the mice to age would further aggravate the reduction in insulin-stimulated glucose transport observed in young adult mice. Surprisingly, the difference in maximal insulin-stimulated glucose transport in soleus from young adult mice was absent in the mature adult cohort (Fig. 5A) and we found neither β-actin KO effect on insulin-stimulated glucose transport in EDL muscles (Fig. 5A), nor β-actin KO effect on cell signaling (Fig. 5, B–D). Loss of β-actin might be more critical in young adult muscles because of the ongoing growth and maturation, higher β-actin abundance (data not shown), and/or the higher habitual activity level (data not shown).

Fig. 5.

2-DG transport and insulin signaling during submaximal and maximal insulin stimulation in mature adult mice. A: insulin-stimulated 2-DG transport during submaximal (soleus 1.8 nM, EDL 3 nM) and maximal (60 nM) insulin concentrations in soleus and EDL muscles from mature adult (18–23 wk old) WT and β-actin KO mice. Akt Ser⁴⁷³ (B) and Akt Thr³⁰⁸ (C) phosphorylation in soleus and EDL muscles from mature adult WT and β-actin KO mice; n = 12 for submaximal insulin and n = 19–21 for maximal insulin stimulation. D: representative blots of Akt Ser⁴⁷³ and Akt Thr³⁰⁸ protein phosphorylation in soleus and EDL muscles of mature adult WT and β-actin KO mice. Coomassie staining is used as loading control. All values shown are means ± SE; ***P < 0.001 vs. basal; ###P < 0.001 max vs. submax insulin.

Ex vivo 2-DG glucose transport in young adult β-actin KO mice trends downward following electrically stimulated contractions but not with passive stretch or AICAR stimulation.

Exercise signaling to glucose transport is distinct from the proximal insulin signaling cascade and is known to recruit a different subset of GLUT4 vesicles (55). To evaluate whether contraction-stimulated glucose transport was impaired in young adult β-actin KO mice, we measured electrically induced contraction-stimulated glucose transport ex vivo. Similar to the insulin data, we observed reduced contraction-stimulated glucose transport in soleus but not in EDL muscles. However, the reduction was more variable than for insulin stimulation and did not reach statistical significance (Fig. 6A). The increases in glucose transport by the AMPK activator AICAR or by passive stretching of the muscles showed no changes related to β-actin (Figs. 7A and 8A). Finally, no genotype-related differences were observed in signaling pathways known to be involved in contraction-, AICAR-, or stretch-stimulated glucose transport (p-AMPK Thr¹⁷², p-ACC Ser²¹², p-TBC1D1 Ser²³¹, p-eEF2 Thr⁵⁶, p-p38MAPK Thr¹⁸⁰/Tyr¹⁸²; Figs. 6, B–F; 7, B–E; 8, B–D). Taken together, the ex vivo analyses indicate that the majority of the glucose transport regulation in adult skeletal muscle occurs independently of β-actin, which is in great contrast to studies in cultured muscle cells.

Fig. 6.

2-DG transport during electrically stimulated contraction (CTXN) and contraction-mediated signaling in young adult mice. A: contraction-stimulated 2-DG transport during electrical stimulation in soleus and EDL muscles from young adult (8–13 wk old) WT and β-actin KO mice. AMP-activated protein kinase(AMPK) Thr¹⁷² (B), acetyl-CoA carboxylase (ACC) Ser²¹² (C), TBC1 domain family member 1 (TBC1D1) Ser²³¹ (D), and eukaryotic elongation factor 2 (eEF2) Thr⁵⁶ (E) phosphorylation in soleus and EDL muscles from young adult WT and β-actin KO mice. F: representative blots of AMPK Thr¹⁷², ACC Ser²¹², TBC1D1 Ser²³¹, and eEF2 Thr⁵⁶ protein phosphorylation in soleus and EDL muscles of young adult WT and β-actin KO mice; n = 28–29. Coomassie staining is used as loading control. All values shown are means ± SE; ***P < 0.001 vs. basal; $$P < 0.01 genotype difference.

Fig. 7.

2-DG transport during 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR) stimulation and AICAR-mediated signaling in young adult mice. A: AICAR-stimulated (4 mM) 2-DG transport in soleus and EDL muscles from young adult (8–13 wk old) WT and β-actin KO mice. AMPK Thr¹⁷² (B), ACC Ser²¹² (C), and TBC1D1 Ser²³¹ (D) phosphorylation in soleus and EDL muscles from young adult WT and β-actin KO mice; n = 12. E: representative blots of AMPK Thr¹⁷², ACC Ser²¹², and TBC1D1 Ser²³¹ protein phosphorylation in soleus and EDL muscles of young adult WT and β-actin KO mice. Coomassie staining is used as loading control. All values shown are means ± SE; *P < 0.05, **P < 0.01, ***P < 0.001 vs. basal.

Fig. 8.

2-DG transport during passive stretch stimulation and stretch-mediated signaling in young adult mice. A: stretch-stimulated (130 mN) 2-DG transport in soleus and EDL muscles from young adult (8–13 wk old) WT and β-actin KO mice. p38 mitogen-activated protein kinase (MAPK) Thr¹⁸⁰/Tyr¹⁸² (B), and eEF2 Thr⁵⁶ (C) phosphorylation in soleus and EDL muscles from young adult WT and β-actin KO mice. No increase in eEF2 phosphorylation indicates that the muscles were not stretched excessively, which otherwise might lead to Ca²⁺ influx through the plasma membrane; n = 16. D: representative blots of p-p38 MAPK Thr¹⁸⁰/Tyr¹⁸⁰ and p-eEF2 Thr⁵⁶ protein phosphorylation in soleus and EDL muscles of young adult WT and β-actin KO mice. Coomassie staining is used as loading control. All values shown are mean ± SE; **P < 0.01, ***P < 0.001 vs. basal.

β-Actin mobility does not increase upon insulin stimulation.

The insulin-stimulated actin remodeling necessary for GLUT4 translocation is thought to be associated with increased dynamics and continuous movement of the actin cytoskeleton in muscle cells in culture (10). To test whether insulin stimulation increased actin motility in mature mouse skeletal muscle, we performed a FRAP analysis of nonstimulated and insulin-stimulated motility of electroporated GFP-β-actin in single, live FDB muscle fibers. The only published similar experiment was done in isolated feline cardiac fibers and found increased β-actin half-time of recovery with insulin stimulation (1). The increase in dynamics of GFP-β-actin in insulin-stimulated cardiomyocytes was observed in regions including the z-disks (1). However, the mobile fraction of the GFP-β-actin was not stated, making it difficult to interpret these observations. We validated our construct by transfecting C2 myoblasts and verifying that GFP showed an actin stress fiber-like morphology and colocalized with β-actin antibody staining (data not shown). We also verified that the electroporated FDB fibers responded as expected in regard to Akt phosphorylation by insulin (Fig. 9A) and that a portion of the fibers from a FDB muscle expressed detectable GFP-β-actin (Fig. 9B). In general, GFP-β-actin was expressed in a striated pattern (Fig. 9C) that colocalized with the z-disk protein α-actinin (data not shown). This expression pattern is similar to what has been reported for β-actin in cardiomyocytes (1) and for GFP-γ-actin in FDB muscle fibers (52). We also found GFP-β-actin in an additional band between the z-disks, corresponding to the A-band region, and around nuclei (Fig. 9C). No changes were observed in the intensity of the β-actin signal in striations upon insulin stimulation (Fig. 9D). When FRAP analyses were applied to this region in the muscle fibers, we could not detect any mobility of GFP-β-actin or changes with insulin (data not shown). Since it is the cortical fraction of β-actin that is considered involved in glucose transport (9), FRAP analyses were performed on GFP-β-actin localized in the superficial part of the muscle fibers in a well-defined region surrounding the nuclei (data not shown). It is likely that the achieved level of GFP-β-actin expression is insufficient to cause its incorporation into the muscle contractile filaments, as was the case for highly overexpressed γ-actin (26), but, just in case, the cortical region surrounding the nuclei was chosen for analysis to avoid potentially immobile β-actin in the contractile filaments. In these regions, the GFP-β-actin in live muscle fibers displayed greater mobility than that in control muscle fibers fixed with formaldehyde (Fig. 9E). However, insulin stimulation changed neither the mobile fraction nor the half-time of recovery of GFP-β-actin (Fig. 9, E and F). Although the FRAP analysis cannot exclude actin reorganization within the bleached region, or assess whether β-actin mobility equates to actin reorganization, the limited to null FRAP recovery of GFP-β-actin in insulin-stimulated as well as unstimulated muscle fibers, indicates that insulin stimulation does not elicit a major change in β-actin mobility in adult skeletal muscle.

Fig. 9.

A fraction of GFP-β-actin is mobile but not affected by insulin stimulation. A: Akt Thr³⁰⁸ phosphorylation in overnight-cultered flexor digitorum brevis (FDB) single fibers with and without 30 min of prior insulin stimulation (60 nM). B: overview of live single fibers isolated from a p-EGFP-actin-transfected FDB muscle. Bar = 1 mm. C: GFP-β-actin structure in successfully transfected live single fiber. Top: characteristic striated pattern, which was observed throughout the core of the fibers. Bottom: GFP-β-actin in the region surrounding a nucleus (arrows). Bar = 5 µm. D: GFP-β-actin intensity accumulation in striations relative to total GFP-β-actin intensity in live single fibers before and 10 min after insulin stimulation (60 nM); n = 3. E: average fluorescence recovery curve after photobleaching (FRAP) of FDB fibers with and without 30 min of prior insulin stimulation (60 nM); n = 41–43 from 3 independent experiments. As a negative control, fibers (n = 17) were fixed before the FRAP experiment. F: mean time of half recovery and mobile fraction from recovery curves fitted using a single-term exponential equation, as described in Ref. 60; n = 41–43 from 3 independent experiments. All values shown are means ± SE.

DISCUSSION

β-Actin has been proposed as the major actin isoform required for GLUT4 translocation in muscle and fat cells (6, 77). However, here, we demonstrate that β-actin is unlikely to play a major role in the GLUT4 translocation mechanism in mature mouse muscle, since glucose transport stimulation in β-actin KO muscle was indistinguishable from that in wild type under all conditions except maximal insulin stimulation in young adult soleus muscles. Increasing the age of the mice from 8–13 wk to 18–23 wk caused the glucose transport phenotype in the soleus muscle of the β-actin KO mice to disappear, suggesting that it might relate to muscle growth/maturation or be observable only at very high glucose transport rates in young mice. Furthermore, our FRAP analysis of overexpressed GFP-β-actin suggests that the β-actin cytoskeleton is largely static and unresponsive to insulin in adult mouse muscle. The imaging analyses were performed in FDB muscle fibers, whereas the ex vivo characterization was done in soleus and EDL muscles. Although the EDL and FDB muscles are both characterized by high proportions of type II fibers (3), and we consider a genotype-related difference in FDB unlikely, based on our EDL incubation data, we cannot exclude potential muscle type-specific differences.

Studies using the actin depolymerizing agent latrunculin B suggest the requirement of the actin cytoskeleton in insulin-, contraction-, and stretch-stimulated glucose transport (7, 68, 69, 71). In addition, a number of studies have suggested that the function and regulation of the cortical actin cytoskeleton are conserved between muscle cell culture and mature muscle. Our group (68–70) and others (24, 79) demonstrated that the small Rho family GTPase Rac1 is activated by insulin, passive stretch, and contraction/exercise in rodents and humans and is necessary for insulin-, stretch-, contraction-, and exercise-stimulated GLUT4 translocation and glucose uptake in mouse muscle (68, 69, 71–73). This effect was nonadditive to that of latrunculin B, suggesting that the effects of Rac1 require an intact actin cytoskeleton (68, 70). A role for Rac1 is also supported by reduced insulin-stimulated glucose transport and GLUT4 translocation in mice lacking the Rac1-activated kinase PAK1 (78). This is consistent with the proposed involvement of Rac1 signaling in insulin-stimulated cortical actin remodeling and GLUT4 translocation in muscle cell culture (34). These data indicate that Rac1 signaling regulates actin dynamics and consequently glucose transport in mature muscle. There are several possible explanations for the discrepancy between muscle-specific β-actin KO mice and Rac1 KO mice. First, although it seems unlikely, based on the literature (8, 16, 31, 57, 65), we cannot rule out that γ-actin partly compensated for the absence of β-actin, as also suggested recently by others (43, 53). One study found γ-actin to be the main constituent of the cortical actin mesh in various rat and pig fibroblast cell cultures (16), and another study showed γ-actin, but not β-actin, staining in mechanically peeled sarcolemma from mouse muscle fibers (61). Furthermore, genetic manipulations to remove or increase expression of tropomyosin 3.1 (Tpm3.1) in skeletal muscle correlate with glucose uptake during glucose and insulin tolerance tests, and Tpm3.1 interestingly defines a subset of γ-actin filaments located at the sarcolemma and T-tubules (32). Since latrunculin B disrupts all actin isoforms, it might be speculated that the previously reported Rac1-dependent reduction in glucose transport by this compound occurs via either γ-actin or both β- and γ-actin. Second, Rac1 and PAK1 might control GLUT4 translocation and glucose transport through non-actin-dependent mechanisms. Rac1 is a subunit of the NADPH oxidase 2 (NOX2) complex, and reactive oxygen species (ROS) can influence GLUT4 translocation and glucose transport both positively and negatively (2, 13, 21, 29, 39, 62, 67). NOX2 is known to interact with the actin cytoskeleton, and latrunculin B and other actin depolymerizing agents were found to increase basal or stimulated NOX2 activity in different nonmuscle cell types (4, 11). Since NOX2 is the major source of ROS during in vitro contraction and stretch (33, 82) and may signal to insulin-stimulated muscle GLUT4 translocation (13) and contribute to insulin resistance (14, 66), the exact roles of actin and NOX2 and their potential cross-talk downstream of Rac1 in regulating glucose transport deserves further study. If the absence of actin isoforms in skeletal muscle increases basal or stimulated NOX2 activation, it might have contributed to the present insulin-resistant phenotype in soleus muscles from young mice. Furthermore, if the absence of actin isoforms in skeletal muscle increases basal or stimulated NOX2 activation, it might also contribute to the progressive myopathy in mice lacking β- and/or γ-actin (57, 58, 65).

To evaluate whether the previously reported progressive myopathy in β-actin KO mice would affect glucose transport, we performed our measurements in two cohorts of mice, young adult and mature adult. The mature adult cohort was chosen because all muscles measured in the original report at 6 mo (52) showed a significant, albeit highly variable, genotype difference in centronucleated fibers, indicative of muscle regeneration. In contrast to our hypothesis that muscle damage-associated inflammation in the aged β-actin KO mice would impair insulin-stimulated glucose transport, we observed a partial reduction in maximal insulin-stimulated glucose transport in the soleus muscles of the young adult β-actin KO mice, which was absent in the mature adult mice. Even in young mice, this phenotype was observed only with maximal insulin stimulation and only in soleus muscle. We speculate that this phenotype relates to the interaction of the load-bearing postural soleus muscle with the rapid postnatal muscle growth phase. For instance, an ~14-fold increase in mouse EDL fiber volume occurs between weeks 1 and 8 post partum (83). Another recent study compared the effect of satellite cell depletion on synergist ablation-induced hypertrophy of the plantaris muscle in 8- and 16-wk-old mice and observed a strong inhibitory effect in young but not mature adult mice (42). Whether these two phenotypes are related and whether the involvement of β-actin is direct or indirect are presently unknown. Another possibility is that a genotype difference in glucose transport is evident only at very high glucose transport rates in very insulin-sensitive young mice. However, the fold increase in glucose transport did not differ between the two age groups, making this an unlikely explanation.

Despite the known downregulation of β- and γ-actin during muscle cell differentiation (20, 37, 51, 61, 76), and despite previous reports of exclusive β-actin localization to the neuromuscular junction (19, 38), what might the function of β-actin then be if present throughout mature skeletal muscle? Apart from a role in vesicle movement beyond those containing GLUT4, it may play a role in glycogen resynthesis. After in situ contraction-mediated glycogen depletion in rabbit tibialis anterior muscle, β-actin was reported to colocalize with glycogen synthase and phosphorylase in the spherical glycogen synthesis-initiating structures termed glycosomes (56). Furthermore, β-actin expression in membrane-enriched fractions correlated with so-called transversal stiffness (i.e., nonlongitudinal force transmission perpendicular to the length of the fibers) measured using atomic force microscopy in rodent soleus muscles subjected to unloading by hindlimb suspension or space flight (45–49). Interestingly, actinin-4 showed a similar decrease in membranes with unloading as β-actin (45, 47) and was proposed to link GLUT4 translocation to the cytoskeleton in L6 muscle cells (17, 74). Finally, the actin cytoskeleton positions many organelles, and β-actin might participate in this process, too. To give a few examples from nonmuscle cell types, the actin cytoskeleton positions mitochondria, perhaps through mitochondrial myosin 19 isoform (59) and nuclei, or through A-type lamins and the nuclear envelope proteins Nesprin and Sun (50). The above-mentioned clearly highlights the significance of using the muscle-specific β-actin KO mice in future studies to unravel the role of β-actin in mature skeletal muscle.

In summary, the present study examined the contribution of β-actin to glucose transport regulation in adult mouse muscles. The β-actin KO mice showed a surprisingly mild phenotype, implying that the β-actin cytoskeleton might behave differently and is not as crucial for GLUT4 translocation in adult muscle as indicated by previous cell culture studies. Instead, the closely related γ-actin isoform might be the main actin isoform involved. Further studies are required to scrutinize which actin isoform is involved in the actin cytoskeleton rearrangements facilitating GLUT4 translocation and glucose transport in adult muscle.

GRANTS

The study was supported by funding from a Novo Nordisk Foundation Excellence Grant (no. 15182) to T. E. Jensen, Danish Diabetes Academy PhD stipends supported by the Novo Nordisk Foundation to A. B. Madsen and J. R. Knudsen, and Intramural Research Program of the National Institute of Arthritis and Musculoskeletal and Skin Diseases to E. Ralston.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

T.E.J. conceived and designed research; A.B.M., J.R.K., C.H.-O., Y.A., K.J.Z., L.S., P.S., and T.E.J. performed experiments; A.B.M., J.R.K., C.H.-O., Y.A., K.J.Z., L.S., and P.S. analyzed data; A.B.M., J.R.K., K.J.Z., P.S., E.R., and T.E.J. interpreted results of experiments; A.B.M., J.R.K., and C.H.-O. prepared figures; A.B.M., E.R., and T.E.J. drafted manuscript; A.B.M., J.R.K., C.H.-O., Y.A., K.J.Z., L.S., P.S., E.R., and T.E.J. edited and revised manuscript; A.B.M., J.R.K., C.H.-O., Y.A., K.J.Z., L.S., P.S., E.R., and T.E.J. approved final version of manuscript.