1. Cell Biology

Microtubule-mediated GLUT4 trafficking is disrupted in insulin-resistant skeletal muscle

  1. Jonas R Knudsen  Is a corresponding author
  2. Kaspar W Persson
  3. Carlos Henriquez-Olguin
  4. Zhencheng Li
  5. Nicolas Di Leo
  6. Sofie A Hesselager
  7. Steffen H Raun
  8. Janne R Hingst
  9. Raphaël Trouillon
  10. Martin Wohlwend
  11. Jørgen FP Wojtaszewski
  12. Martin AM Gijs
  13. Thomas Elbenhardt Jensen  Is a corresponding author
  1. August Krogh Section for Molecular Physiology, Department of Nutrition, Exercise and Sports, Faculty of Science, University of Copenhagen, Denmark
  2. Heart and Skeletal Muscle Biology, Global Drug Discovery, Novo Nordisk, Denmark
  3. Exercise Science Laboratory, Faculty of Medicine, Universidad Finis Terrae, Chile
  4. College of Physical Education, Chongqing University, China
  5. Department of Biomedical Sciences, Faculty of Health and Medical Sciences, University of Copenhagen, Denmark
  6. Clinical Drug Development, Novo Nordisk, Denmark
  7. Department of Electrical Engineering, Polytechnique Montréal, Canada
  8. Laboratory of Integrative Systems Physiology, Institute of Bioengineering, École Polytechnique Fédérale de Lausanne, Switzerland
  9. Microsystems Laboratory 2, Institute of Electrical and Micro Engineering, École Polytechnique Fédérale de Lausanne, Switzerland
https://doi.org/10.7554/eLife.83338
Share this article
Cite this article
  1. Jonas R Knudsen
  2. Kaspar W Persson
  3. Carlos Henriquez-Olguin
  4. Zhencheng Li
  5. Nicolas Di Leo
  6. Sofie A Hesselager
  7. Steffen H Raun
  8. Janne R Hingst
  9. Raphaël Trouillon
  10. Martin Wohlwend
  11. Jørgen FP Wojtaszewski
  12. Martin AM Gijs
  13. Thomas Elbenhardt Jensen
(2023)
Microtubule-mediated GLUT4 trafficking is disrupted in insulin-resistant skeletal muscle
eLife 12:e83338.
https://doi.org/10.7554/eLife.83338
  2. Download BibTeX
  3. Download .RIS

Abstract

Microtubules serve as tracks for long-range intracellular trafficking of glucose transporter 4 (GLUT4), but the role of this process in skeletal muscle and insulin resistance is unclear. Here, we used fixed and live-cell imaging to study microtubule-based GLUT4 trafficking in human and mouse muscle fibers and L6 rat muscle cells. We found GLUT4 localized on the microtubules in mouse and human muscle fibers. Pharmacological microtubule disruption using Nocodazole (Noco) prevented long-range GLUT4 trafficking and depleted GLUT4-enriched structures at microtubule nucleation sites in a fully reversible manner. Using a perifused muscle-on-a-chip system to enable real-time glucose uptake measurements in isolated mouse skeletal muscle fibers, we observed that Noco maximally disrupted the microtubule network after 5 min without affecting insulin-stimulated glucose uptake. In contrast, a 2-hr Noco treatment markedly decreased insulin responsiveness of glucose uptake. Insulin resistance in mouse muscle fibers induced either in vitro by C2 ceramides or in vivo by diet-induced obesity, impaired microtubule-based GLUT4 trafficking. Transient knockdown of the microtubule motor protein kinesin-1 protein KIF5B in L6 muscle cells reduced insulin-stimulated GLUT4 translocation while pharmacological kinesin-1 inhibition in incubated mouse muscles strongly impaired insulin-stimulated glucose uptake. Thus, in adult skeletal muscle fibers, the microtubule network is essential for intramyocellular GLUT4 movement, likely functioning to maintain an insulin-responsive cell surface recruitable GLUT4 pool via kinesin-1-mediated trafficking.

Editor's evaluation

This manuscript reveals localization of Glut4 glucose transporters at microtubules in mouse and human muscle fibers and shows that disruption of microtubules or a kinesin-1 motor alters Glut4 trafficking. Evidence is also provided supporting the idea that insulin resistance disrupts Glut4 dynamics at microtubules. Overall, these studies provide compelling evidence that Glut4 and its regulation by insulin involves Glut4 movements that require microtubule function.

https://doi.org/10.7554/eLife.83338.sa0

Introduction

Skeletal muscle is quantitatively the largest site of glucose disposal, a process facilitated by insulin and contraction-responsive translocation and insertion of glucose transporter 4 (GLUT4) into the surface membrane of muscle fibers (Jaldin-Fincati et al., 2017; Klip et al., 2019). Insulin-resistant human and rodent muscle exhibit impaired insulin-stimulated GLUT4 translocation (Zierath et al., 1996; King et al., 1992; Etgen et al., 1997; Garvey et al., 1998) and muscle-specific deletion of GLUT4 is sufficient to cause systemic insulin resistance and glucose intolerance (Zisman et al., 2000). However, the details of GLUT4 regulation – particularly in adult skeletal muscle – and the causes of skeletal muscle insulin resistance remain unclear. In L6 myoblasts and 3T3-L1 adipocytes, insulin resistance not only decreases insulin-stimulated GLUT4 recruitment to the surface membrane, but also affects the distribution of GLUT4 between intracellular compartments (Foley and Klip, 2014; Xiong et al., 2010). This suggests that disturbed intracellular sorting of GLUT4 contributes to peripheral insulin resistance.

Motor protein-mediated trafficking on the microtubule cytoskeleton is well established to allow long-range transport of a diverse assortment of molecules and to position intracellular organelles and membrane structures in various cell types (de Forges et al., 2012). For GLUT4, long-range microtubule-dependent GLUT4 movement beneath the plasma membrane has been observed in adipocyte cell culture (Lizunov et al., 2005) and similar long-range movement was also seen in adult rodent skeletal muscle (Lizunov et al., 2012). A requirement for microtubule-based protein trafficking is supported by several observations in cultured cells. Microtubule disruption dispersed perinuclear GLUT4 in 3T3-L1 adipocytes (Guilherme et al., 2000; Fletcher et al., 2000) as well as L6 myoblasts (Foley and Klip, 2014) and impaired GLUT4 membrane insertion in some (Foley and Klip, 2014; Fletcher et al., 2000; Chen et al., 2008; Emoto et al., 2001) but not all studies (Molero et al., 2001; Shigematsu et al., 2002). Neither the requirement of microtubules for intracellular GLUT4 positioning and trafficking nor the influence of insulin resistance on microtubules and/or microtubule-based GLUT4 trafficking have been investigated in adult skeletal muscle fibers.

Therefore, we presently characterized various aspects of microtubule-based GLUT4 trafficking in predominantly adult human and mouse skeletal muscle. Our findings suggest that an intact microtubule network is required for KIF5B-mediated intracellular GLUT4 movement and maintaining insulin-responsive glucose uptake, and that impaired microtubule-based GLUT4 trafficking is a feature of skeletal muscle insulin resistance.

Results

GLUT4 was enriched at microtubule nucleation sites and traveled on microtubule filaments in adult mouse and human muscle

To study the involvement of microtubules in GLUT4 trafficking, we first used structured illumination microscopy to image the subsarcolemmal (up to 4 µm into the muscle fiber) microtubule network and GLUT4 in mouse and human skeletal muscle at super-resolution. Due to amenability for live fiber isolation we used flexor digitorum brevis (FDB), a muscle consisting predominantly of type IIa and IIx fibers (Tarpey et al., 2018), from mice and vastus lateralis, a highly mixed muscle (Staron, 1991; Horwath et al., 2021), from humans. In both mouse (Figure 1A) and human (Figure 1B) muscle, we observed GLUT4 to be localized on microtubule filaments and to be enriched at microtubule filament intersections, previously identified as microtubule nucleation sites (Oddoux et al., 2013). Next, to study GLUT4 movement in live muscle fibers, we overexpressed GLUT4-7myc-GFP (GLUT4-GFP) (Bogan et al., 2001) alone or together with mCherry-Tubulin in mouse FDB muscle fibers (Figure 1C). GLUT4-GFP was localized in the same pattern as endogenous GLUT4 and observed along the microtubule network, including on the more stable subpopulation (Bulinski and Gundersen, 1991) of detyrosinated microtubules (Figure 1—figure supplement 1A) implicated in trafficking of lipid droplets, mitochondria, and autophagosomes in other cells types (Mohan et al., 2019; Herms et al., 2015), as well as on mCherry-Tubulin-labeled microtubules (Figure 1—figure supplement 1B). Live-imaging revealed long-range lateral directional movement of GLUT4 along filamentous tracks (Figure 1—figure supplement 1C), corresponding to mCherry-Tubulin containing microtubule filaments (Figure 1D and Figure 1—video 1). The GLUT4 structures occasionally exhibited long tubular morphology (>2 µm) but were mostly minor tubular structures or spheres (size varying from ~0.4 μm2 down to the unresolvable) observed to undergo budding and fusion events on the microtubule tracks (Figure 1—figure supplement 1D, E). Live-imaging, including fluorescence recovery after photobleaching experiments, revealed particularly dynamic and bidirectional movement at the microtubule nucleation sites (Figure 1—figure supplement 1F–H and Figure 1—video 2). Collectively, a portion of GLUT4 localized to microtubule nucleation sites and on microtubule filaments in adult mouse and human skeletal muscle. Furthermore, GLUT4 underwent continuous movement, budding and fusion along the microtubule tracks in live mouse skeletal muscle.

Figure 1 with 3 supplements see all
Download asset Open asset
Glucose transporter 4 (GLUT4) was enriched at microtubule nucleation sites and traveled on microtubule filaments in mouse and human muscle.

Structured illumination microscopy (SIM) in mouse flexor digitorum brevis (FDB) muscle (A) and human vastus lateralis muscle (B) of endogenous α-tubulin and GLUT4 (left panel) and 3D reconstruction of GLUT4 (green) and α-tubulin (white) (right panel). Arrows indicate GLUT4 at microtubule nucleation sites, arrowheads mark GLUT4 vesicles along the microtubule filaments. (C) Overview of workflow for live-imaging of fluorescently conjugated proteins in adult mouse FDB muscle fibers. (D) Live-imaging of FDB-expressing GLUT4-GFP and mCherry-Tubulin. Yellow projection of mCherry-Tubulin outlines the microtubule filaments (top panel left). Movement of GLUT4-GFP was visualized by color-coded projection (first image cyan, last image red, top panel right). The merged projection (bottom), demonstrated movement of GLUT4-GFP along the mCherry-Tubulin containing microtubule filaments indicated by color-coded projections on top of the microtubule filaments. The movement of GLUT4-GFP is shown in Figure 1—video 1. (A, B) Images are representative of >5 fibers from ≥3 different mice in A + D and 3 different fibers from 3 different subjects in B. Scale bar = 5 µm (A, B, D) and 2 µm (inserts in B, D).

GLUT4 trafficking and localization required intact microtubules

Next, we tested if microtubule-based GLUT4 trafficking was insulin responsive and dependent on an intact microtubule network. Insulin (30 nM) stimulation increased insulin signaling at the level of Akt Thr308 as expected (Figure 2—figure supplement 1A), and the microtubule depolymerizing compound Nocodazole (Noco) (13 µM) significantly reduced both the total and the Noco-resistant (Khawaja et al., 1988) detyrosinated pool of polymerized microtubules by ~90% and ~50%, respectively (Figure 2—figure supplement 1B, C). We did not observe any significant increase in GLUT4 movement on microtubules upon insulin stimulation but GLUT4 movement was completely prevented by microtubule depolymerization (Figure 2A, B, Figure 2—figure supplement 1D, Figure 2—video 1). Having established that GLUT4 trafficking was dependent on the microtubule network, we next tested if microtubule disruption affected the overall GLUT4 localization and distribution between different compartments. For quantification, we divided the GLUT4 structures into size categories corresponding to (1) large structures at the microtubule nucleation sites (>4 µm2), (2) intermediate endomembrane structures (0.4–4 µm2) (Gruenberg, 2001; Huotari and Helenius, 2011), and (3) the smallest resolvable endomembrane structures (<0.4 µm2) including presumably insulin-responsive GLUT4 storage vesicles (GSVs) (Figure 2—figure supplement 1E). Microtubule disruption by Noco (13 µM) drained the GLUT4 structures at the microtubule nucleation sites and reduced the amount of the smallest structures, while causing an increase in the intermediate structures (Figure 2C, D). These changes were reversed within 9 hr after removal of Noco (Figure 2C, D). Within the smallest category, there was a shift toward fewer but larger area GLUT4 membrane structures (Figure 2—figure supplement 1F). The total number and area of GLUT4 structures did not differ between conditions (Figure 2—figure supplement 1G). In a previous study in L6 myoblasts, microtubule disruption prevented pre-internalized GLUT4 from reaching a Syntaxin6-positive perinuclear subcompartment involved in GSV biogenesis and from undergoing insulin-responsive exocytosis (Foley and Klip, 2014). We therefore tested in adult muscle, if microtubule disruption similarly prevented accumulation in a perinuclear Syntaxin6-positive subcompartment. However, we observed a limited and Noco-insensitive (in mice) co-localization of Syntaxin6 with either endogenous GLUT4 in human and mouse skeletal muscle, or fluorescent GLUT4-EOS (Lizunov et al., 2013) in mouse skeletal muscle (Figure 2E, F, Figure 2—figure supplement 1H). Altogether, our data demonstrate that GLUT4 trafficking and distribution is disrupted by pharmacological microtubule network depolymerization in a fully reversible manner.

Figure 2 with 2 supplements see all
Download asset Open asset
Glucose transporter 4 (GLUT4) trafficking and localization was dependent on an intact microtubule network.

(A) Representative time-lapse traces of GLUT4-GFP vesicle tracking in muscle fibers ± insulin (INS, 30 nM) for 15–30 min with or without microtubule network disruption by Nocodazole (Noco, 13 µM) for 4 hr prior to insulin. The dynamics of GLUT4-GFP in the different conditions are also exemplified in Figure 2—video 1. (B) Quantified microtubule-based GLUT4 trafficking. (C) Representative images of muscle fibers ± pre-treatment with Noco 13 µM, for 15 hr, followed by recovery in Noco-free medium for 9 hr. (D) Quantification of GLUT4 distribution between the microtubule nucleation sites (structures sized >4 µm2), intermediate-sized structures (0.4–4 µm2) and the smallest resolvable structures (<0.4 µm2) in fibers treated as in C. Compartment identification is described in Figure 2—figure supplement 1E. (E) GLUT4 and Syntaxin6 (Stx6) in muscle fiber from human vastus lateralis muscle. (F) GLUT4-Stx6 overlap in perinuclear region of mouse flexor digitorum brevis muscle fibers in Dimethylsulfoxide (DMSO) medium with and without Noco (13 µM) treatment. For A, B, n ≥ 14 muscle fibers from 5 different mice. For C, D, n = 9–11 muscle fibers from 3 different mice. For E, n = 3 subjects. Data are presented as mean with individual data points. ***p < 0.001 different from basal, ###p < 0.001 different from INS, ##p < 0.01 different from Noco recovery. ¤¤¤p < 0.001 analysis of variance (ANOVA) effect. Scale bar = 5 µm (A–C) and 2 µm (E).

Figure 2—source data 1

Data used for quantification of GLUT4-Stx6 overlap in perinuclear region of mouse flexor digitorum brevis muscle fibers in DMSO medium with and without Noco (13 µM) treatment.

https://cdn.elifesciences.org/articles/83338/elife-83338-fig2-data1-v2.zip

Prolonged, but not short-term, microtubule disruption blocked insulin-induced muscle glucose uptake

To investigate the requirement of microtubule-based GLUT4 trafficking and localization for insulin-induced muscle glucose uptake, we assessed muscle glucose uptake ± insulin and ± microtubule disruption in isolated incubated intact mouse soleus and extensor digitorum longus (EDL) muscles. When mouse soleus and EDL muscles were incubated ex vivo ± insulin and ± Noco (13 µM) for 15 min and up to 2 hr, an interaction between insulin and Noco was observed and the insulin-induced glucose uptake was gradually impaired over time and completely disrupted after 2 hr in both muscles (Figure 3A). The increase by insulin stimulation was significantly impaired after 40 min in soleus and 2 hr in EDL muscle (Figure 3B). Insulin-stimulated phospho-signaling via Akt and TBC1D4 was unaffected by Noco treatment (Figure 3—figure supplement 1A–C).

Figure 3 with 1 supplement see all
Download asset Open asset
Time-dependent effect of microtubule disruption on insulin-induced muscle glucose uptake.

(A) 2-Deoxyglucose (2-DG) transport in basal and insulin-stimulated mouse soleus and extensor digitorum longus (EDL) muscles pretreated with Nocodazole (Noco, 13 µM) for the indicated time. (B) Insulin-stimulated 2-DG transport (insulin minus basal) from muscles shown in A. (C) Experimental setup for muscle-on-a-chip system with glucose sensor. (D) Microtubules imaged with α-tubulin in glucose transporter 4 (GLUT4)-GFP-expressing mouse flexor digitorum brevis (FDB) fibers treated ± Noco (13 µM) for 5 min or 2 hr. (E) 180-s measurements of glucose concentration in perifusate from basal and insulin-treated FDB muscle fibers in muscle chips pre-incubated with DMSO, Noco (13 µM, 5 min or 2 hr) or colchicine (25 µM, 2 hr). (F) Insulin-stimulated glucose uptake into FDB muscle fibers in muscle chips calculated from the last 20 s of the concentration curves in E. (G) Representative GLUT4 images from isolated mouse FDB muscle fibers treated ± Noco (13 µM) for 5 min or 2 hr. (H) Quantification of GLUT4 in large, intermediate- and small-sized membrane structures in FDB fibers treated ± Noco (13 µM) for 5 min or 2 hr. The membrane compartment division by size is shown in Figure 2E. For A, B, n = 6–7 muscles from 6 to 7 mice. For D, G, H, n = 8–10 muscle fibers from 3 different mice. For E, F, n ≥ 3 muscle chips from 3 to 4 mice. Data are presented as mean with individual data points. Paired observations from the same mouse are indicated by a connecting line. */**/***p < 0.05/0.01/0.001 different from basal/DMSO, #/##/###p < 0.05/0.01/0.001 different from DMSO. ¤/¤¤/¤¤¤p < 0.05/0.01/0.001 analysis of variance (ANOVA) effect. Scale bar = 5 µm.

Figure 3—source data 1

Data used for quantification of 2-DG transport and glucose clearance and uptake in Figure 3A, B, E, F, polymerized microtubules in Figure 3D and glucose transporter 4 (GLUT4) localization in Figure 3H.

https://cdn.elifesciences.org/articles/83338/elife-83338-fig3-data1-v2.zip

To understand the temporal resolution of microtubule network disruption and its effect on insulin-induced glucose uptake, we investigated glucose uptake adult isolated single fibers in real time using a custom-made perifused organ-on-chip system (Gowers et al., 2015; Trouillon and Gijs, 2016) featuring a glucose-sensing electrode for glucose uptake measurements (Figure 3C). In brief, this chip continuously measures glucose concentration in perifusate post muscle fiber exposure, allowing the estimation of glucose uptake over time. We confirmed the ability of the chip to measure glucose uptake in skeletal muscle fibers (Figure 3—figure supplement 1D–F). Specifically, the chip measured glucose uptake in isolated FDB fibers at ~5 µM glucose concentration sensitivity (defined as a registered fluctuation of thrice the standard deviation [SD] of the baseline measurements) and a temporal resolution of <4 s (Figure 3—figure supplement 1G, H). Noco (13 µM) caused complete microtubule disruption in FDB fibers within 5 min similar to a 2-hr treatment (Figure 3D). Interestingly, acute microtubule disruption (5 min Noco) affected neither basal (Figure 3—figure supplement 1I) nor insulin-induced muscle glucose uptake, whereas 2 hr treatment by Noco or colchicine (25 µM), another microtubule network disrupter, completely blocked insulin-induced muscle glucose uptake (Figure 3E, F). Notably, GLUT4-containing large membrane structures corresponding mainly to microtubule nucleation sites were already reduced after 5 min of Noco exposure, whereas accumulation of GLUT4 in intermediate- and small-sized membrane structures was only observed after 2 hr of Noco exposure (Figure 3G, H). Thus, an intact microtubule network is not required for the immediate insulin-induced GLUT4 translocation response in adult skeletal muscle fibers. However, prolonged disruption of the microtubule network causes a more pronounced missorting of GLUT4 and renders skeletal muscle unresponsive to insulin.

KIF5B-containing kinesin-1 motor proteins regulate muscle GLUT4 trafficking

Next we investigated which motor protein(s) mediate microtubule-dependent GLUT4 trafficking in skeletal muscle. The kinesin-1 heavy chain protein KIF5B has been implicated in GLUT4 trafficking in adipocytes (Semiz et al., 2003; Habtemichael et al., 2018). Thus, we studied the effect of kinesore, a small molecule modulator which both inhibits kinesin-1 interaction with specific cargo adaptors but also stimulates Kinesin motor function (Randall et al., 2017), in incubated soleus and EDL muscles and differentiated primary human myotubes as well as the effect of Kif5b short hairpin (sh) RNA knockdown and kinesore in L6 skeletal muscle cells overexpressing exofacially tagged GLUT4 (Kishi et al., 1998; Wang et al., 1998; Figure 4A). In both soleus and EDL muscle, 2 hr of kinesore (50 μM) exposure strongly impaired insulin-stimulated glucose uptake (Figure 4B) without affecting p-Akt Ser473 and slightly increased basal and insulin-stimulated p-TBC1D4 Thr642 (Figure 4—figure supplement 1A). AMPK is an insulin-independent stimulator of GLUT4 translocation which may indirectly stimulate TBC1D4 Thr642 phosphorylation (Kjøbsted et al., 2015). Notably, phosphorylation of AMPK and downstream ACC2 were stimulated by kinesore in both basal and insulin-stimulated soleus and EDL muscles (Figure 4—figure supplement 1B). In primary human myotubes differentiated for 7 days kinesore (50 μM) and Noco (13 μM) reduced the glucose uptake and completely blocked the insulin response (Figure 4C). In L6 myoblasts, we lowered KIF5B expression using shRNA by ~70% in L6 myoblasts (Figure 4—figure supplement 1C). This did not affect GLUT4 expression (Figure 4—figure supplement 1D) but impaired insulin-stimulated GLUT4 translocation (Figure 4D). Unlike the inhibitory effect in incubated mouse muscle and primary human myotubes, kinesore-stimulated GLUT4 translocation and glucose uptake additively to insulin in L6 muscle cells, and modestly potentiated insulin-stimulated Akt Thr308 without affecting AMPK signaling (Figure 4—figure supplement 1E–J). Collectively, although kinesore surprisingly had a stimulatory and seemingly MT-independent effect on GLUT4 translocation in L6 muscle cells, our shRNA data in L6 myoblasts and kinesore data in adult muscle and primary human myotubes support the requirement of KIF5B-containing kinesin-1 motor proteins in GLUT4 translocation in skeletal muscle.

Figure 4 with 1 supplement see all
Download asset Open asset
Kinesin-1 containing KIF5B-regulated glucose transporter 4 (GLUT4) localization and translocation.

(A) Schematic overview of L6 muscle cell system to assess GLUT4 surface content. (B) 2-Deoxyglucose (2-DG) transport in basal and insulin-stimulated mouse soleus and extensor digitorum longus (EDL) muscles pretreated with kinesore (50 µM) for 2 hr. (C) Deoxyglucose (2-DG) transport in basal and insulin-stimulated primary human myotubes pretreated ± kinesore (50 µM) or Noco (13 µM) for 2 hr. (D) Exofacial GLUT4 signal in serum starved (4 hr) basal and insulin-stimulated (100 nM, 15 min) L6 myoblasts (left) and insulin response (insulin minus basal, right) in GLUT4 surface content. L6 myoblasts were transfected with short hairpin scramble RNA (shScramble) or shRNA targeting Kif5b 72 hr prior to the experiment. Analysis of variance (ANOVA) main effect of insulin (¤¤¤) and shKif5b (¤¤¤) and interaction (¤). (B) n = 8 muscles in each group, lines indicate muscles from same mouse. (C) Each data point represents the average of 3 replicates and originate from at least 3 independent experiments. Data are presented as mean with individual data points. */**/***p < 0.01/0.001 effect of insulin. ##/###p < 0.01/0.001 different from DMSO/Scramble. ¤/¤¤/¤¤¤p < 0.05/0.01/0.001 ANOVA effect.

Figure 4—source data 1

Data used for quantification of glucose transporter 4 (GLUT4) localization and GLUT4 surface content in Figure 4.

Data used for quantification of 2-DG transport and GLUT4 translocation in Figure 4B–D.

https://cdn.elifesciences.org/articles/83338/elife-83338-fig4-data1-v2.zip

Insulin resistance induced by C2 ceramide and high-fat diet impaired microtubule-based GLUT4 trafficking

Having established an essential role for the microtubule network in GLUT4 trafficking and muscle glucose uptake, we proceeded to test if microtubule-based GLUT4 trafficking was impaired in insulin-resistant states. We induced insulin resistance in adult mouse skeletal muscle both in vitro using short-chain C2 ceramide and in vivo using diet-induced obesity (Figure 5A). In vitro, treatment of isolated FDB muscle fibers with C2 ceramide (50 µM) impaired Akt Thr308 phosphorylation (Figure 5—figure supplement 1A) and in insulin-stimulated fibers markedly reduced microtubule-based GLUT4 trafficking defined as the number of moving GLUT4 structures (Figure 5B) and the total microtubule-based traveling of GLUT4 structures (Figure 5—figure supplement 1B). In vivo, mice fed a 60% high-fat diet (HFD) for 10 weeks exhibited impaired tolerance to insulin and glucose as well as reduced insulin-stimulated phosphorylation of Akt Thr308 and Akt substrate TBC1D4 Thr642 in isolated FDB fibers (Figure 5—figure supplement 1C–E), confirming whole-body and skeletal muscle insulin resistance. Similar to C2 ceramide-treated fibers, HFD-exposed FDB muscle fibers exhibited impaired microtubule-based GLUT4 trafficking (Figure 5C, Figure 5—figure supplement 1F). This prompted us to investigate whether the microtubule polymerization was itself insulin responsive and/or affected by insulin resistance. To test this, we transfected mouse FDB muscle fibers with the microtubule plus-end-binding protein EB3-GFP, which binds the tip of growing microtubules via its calponin homology domain and has previously been used for live-cell characterization of microtubule polymerization (Stepanova et al., 2003). As previously reported (Oddoux et al., 2013), EB3-GFP transfection allowed visualization of growing microtubules as a dynamic comet tail-like appearance, an effect completely prevented by the microtubule stabilizer taxol (10 µM) (Figure 5—figure supplement 1G, Figure 5—video 1). In our datasets, we analyzed the microtubule polymerization frequency (by counting EB3-GFP puncta; Komarova et al., 2009), the average polymerization distance, the total polymerization distance and the polymerization directionality following C2 ceramide exposure or HFD. Insulin tended (p = 0.095) to increase the number of polymerizing microtubules by an average of 28% compared to basal fibers while C2 ceramide treatment reduced the amount of polymerizing microtubules significantly and taxol almost abolished microtubule polymerization (Figure 5D, E). C2 ceramide treatment also reduced the total polymerization distance while the average polymerization was unaffected (Figure 5—figure supplement 1H).

Figure 5 with 2 supplements see all
Download asset Open asset
Insulin resistance impairs microtubule-based glucose transporter 4 (GLUT4) trafficking.

(A) Overview of in vitro and in vivo insulin resistance models used. (B) Quantified microtubule-based GLUT4 trafficking in basal, insulin (INS, 30 nM) and insulin + C2 ceramide (C2) (INS + C2, 30 nM + 50 µM) treated flexor digitorum brevis (FDB) muscle fibers. (C) Quantified microtubule-based GLUT4 trafficking in basal or INS (30 nM) treated FDB fibers from chow or high-fat diet (HFD) fed mice. (D) Representative images of polymerizing microtubules in EB3-GFP-expressing FDB muscle fibers treated ± C2 (50 µM), paclitaxel (Taxol, 10 µM) for 2 hr prior to 15–30 min of INS (30 nM) stimulation. Red circles highlight microtubule tip-bound EB3-GFP. (E) Quantification of polymerizing microtubules based on EB3-GFP in FDB fibers treated as in D. (F) Quantification of polymerizing microtubules based on EB3-GFP in FDB fibers isolated from chow or 60% HFD fed mice and treated ± INS (30 nM) for 15–30 min. For B–F, n ≥ 13 muscle fibers from 3 to 4 mice. Taxol-treated muscle fibers were only used as a control and not included in the statistical analysis. NA = not statistically analysed. Data are presented as mean with individual data points. #/##/###p < 0.05/0.01/0.001 different from INS (B) or different from corresponding group in chow fed mice (C) or control fibers (E). ¤/¤¤/¤¤¤p < 0.05/0.01/0.001 main effect (ME) of diet/C2.

Figure 5—source data 1

Data used for quantification of glucose transporter 4 (GLUT4) trafficking and microtubule polymerization in Figure 5B, C, E, F.

https://cdn.elifesciences.org/articles/83338/elife-83338-fig5-data1-v2.zip

In contrast to the effect of C2 ceramide on microtubule dynamics, HFD-induced insulin resistance was not associated with alterations in the number of polymerizing microtubules, average polymerization distance or total polymerization distance (Figure 5F, Figure 5—figure supplement 1I). The polymerization directionality was also not affected by HFD (Figure 5—figure supplement 1J). Altogether, different models of insulin resistance impaired microtubule-based GLUT4 trafficking in adult muscle fibers, suggesting a role in adult skeletal muscle insulin resistance. In contrast, defective microtubule polymerization was observed with C2 ceramide but not with the presumably more physiologically relevant HFD insulin resistance model.

Discussion

In the present study, we provide translational evidence in adult human and mouse skeletal muscle, showing that the microtubule network is crucial for long-range directional GLUT4 trafficking via motor proteins, likely including KIF5B. Microtubule polymerization in isolated mouse muscle fibers could be abolished pharmacologically within minutes without affecting insulin-stimulated glucose uptake whereas longer pharmacological inhibition progressively caused GLUT4 mislocalization and lowered insulin responsiveness. These data are consistent with a model where microtubules are required for correct intramyocellular GLUT4 compartmentalization but not the ultimate insulin-stimulated GLUT4 translocation to the cell surface from these compartments. Importantly, microtubule-based GLUT4 movement was impaired in two classical mouse insulin resistance models, short-chain ceramide treatment and diet-induced obesity. These data implicate dysregulation of microtubule-mediated GLUT4 trafficking and their localization in the etiology of adult skeletal muscle insulin resistance.

What may cause the reduced number of GLUT4 moving on microtubules in insulin-resistant muscle? Based on cell-culture studies, GLUT4 is packaged into specialized GSVs, which upon insulin stimulation can undergo exocytosis (Fazakerley et al., 2022). This exocytosis may include an intermediate step involving release and heterotypic fusion of GSVs with transferrin receptor positive endosomes observable using super-resolution quantum dot single GLUT4 particle tracking in isolated mouse soleus and EDL muscle fibers (Hatakeyama and Kanzaki, 2017). Such a model is consistent with previous reports of insulin-stimulated release of GSVs into an endosomal recycling pool in 3T3-L1 adipocytes (Coster et al., 2004; Govers et al., 2004). In 3T3-L1 adipocytes, GSVs are synthesized via the trans-Golgi network and the Endoplasmic reticulum (ER)–Golgi intermediate compartment (Gould et al., 2020) by mechanisms involving TBC1D4 (Bruno et al., 2016; Brumfield et al., 2021; Sano et al., 2007). We note that TBC1D4 phosphorylation on multiple sites is often decreased in insulin-resistant skeletal muscle of humans (Pehmøller et al., 2012; Vind et al., 2011) and rodents (Castorena et al., 2014; Li et al., 2019). Conversely, TBC1D4 phosphorylation is increased after insulin-sensitizing exercise/contraction (Pehmøller et al., 2012; Funai et al., 2009; Arias et al., 2007; Treebak et al., 2009; Kjøbsted et al., 2019; Kjøbsted et al., 2017) which might augment the insulin-responsive GLUT4 pool (Geiger et al., 2006; Knudsen et al., 2020c; Bradley et al., 2014). The presently observed decreased GLUT4 movement on microtubules could reflect impaired availability of GLUT4 for trafficking, i.e. impaired TBC1D4-dependent GSV biosynthesis and/or heterotypic fusion of GSVs with transferrin receptor positive endosomes. Another possibility is that recruitment of GLUT4 onto microtubules, i.e. release of tethered GSVs and/or motor protein interaction, is impaired. In 3T3-L1 adipocytes, insulin stimulated the frequency but not movement speed of GLUT4 on microtubules, an effect blocked by a dominant-negative cargo-binding kinesin light chain 1 (KLC1) which impairs KIF5B function (Semiz et al., 2003). Also, KLC1 was proposed to regulate GLUT4 translocation in 3T3-L1 adipocytes (Semiz et al., 2003) and L6 myoblasts via insulin-regulated interaction with Double C2 domain B (Zhang et al., 2019). These data are consistent with our observations of KIF5B-dependent GLUT4 translocation in insulin-stimulated L6 muscle cells. KIF5B was proposed to regulate GSV trafficking via insulin-stimulated binding to TUGUL, a cleavage product of the intracellular GSV tethering protein TUG (Habtemichael et al., 2018; Bogan et al., 2003). Multiple studies by Jonathan Bogan’s group linked TUG to skeletal muscle insulin-stimulated GLUT4 traffic and glucose uptake (Belman et al., 2015; Löffler et al., 2013). Insulin-stimulated TUG cleavage and expression of the putative protease Usp25m were reduced in rat adipose tissue after a 3-day high-fat/high-sugar diet-induced insulin resistance (Habtemichael et al., 2018), suggesting possible dysregulation in insulin resistance. Apart from KIF5B and TUG, GLUT4 was proposed to utilize other motor proteins to move on microtubules in 3T3-L1 adipocytes, i.e. dynein via Rab5 (Huang et al., 2001), and the kinesin-2 family member KIF3A via AXIN and the ADP-ribosylase tankyrase 2 (Guo et al., 2012; Imamura et al., 2003). Whether any of these mechanisms regulate GLUT4 traffic in skeletal muscle should be tested further.

Noco treatment potently reduced the number of large GLUT4-containing structures, while causing the number of medium-sized GLUT4-containing structures to increase within 2 hr of treatment. We speculate whether the medium-sized GLUT4 structures are so-called Golgi mini-stacks, small immature membrane compartments of Golgi origin incapable of protein secretion, reported to accumulate in microtubule disrupted cells (Fourriere et al., 2016). Alternatively, the membrane structures accumulated in Noco-treated fibers might be internalized GLUT4 in endosomal structures unable to reach microtubule nucleation sites via retrograde transport. Future studies should characterize the nature of these membrane structures by e.g. measuring different compartment-specific protein markers.

Our glucose uptake time-course data in perifused mouse FDB fibers suggest that microtubules are required to build and maintain the pool of insulin-responsive GLUT4 vesicles near the cell surface observed in previous studies (Lizunov et al., 2012), and that disruption of this process may contribute to insulin-resistant GLUT4 translocation in adult skeletal muscle. If true, then one would predict GLUT4 mislocalization to be observable in unstimulated insulin-resistant skeletal muscle. Consistent with this prediction, Garvey et al. measured GLUT4 content in different sucrose density fractions in human vastus lateralis muscle biopsies and found a basal redistribution of GLUT4 to a denser fraction in type 2 diabetic subjects compared to control (Garvey et al., 1998). Similar observations of altered intracellular GLUT4 distribution were made using a different fractionation protocol in subcutaneous adipose biopsies from type 2 diabetic subjects compared to control (Maianu et al., 2001), implying that GLUT4 mislocalization is a shared hallmark of insulin-resistant muscle and adipose cells. Application of higher-resolution microscopy to the study of basal and insulin-stimulated GLUT4 localization in adult human and rodent insulin-resistant skeletal muscle could help to resolve the relative distribution of GLUT4 between specific membrane compartments.

A previous study reported that the microtubule disrupting drug colchicine (25 µM for up to 8 hr) had no effect on insulin and contraction-stimulated glucose uptake, whereas Noco (3–83 µM for 30–60 min) potently inhibited glucose uptake in ex vivo incubated adult rat muscle and 33 µM Noco added for <5 min directly inhibited GLUT4 transporter activity into sarcolemma derived giant vesicles (Huang et al., 2001). Regarding colchicine, these divergent findings mirror previous studies in 3T3-L1 adipocytes where some support the absent effect of colchicine on insulin-stimulated cell surface GLUT4 translocation/glucose uptake despite strong microtubule disruption (Molero et al., 2001; Huang et al., 2001) whereas others found colchicine to inhibit insulin-stimulated GLUT4 translocation (Fletcher et al., 2000; Emoto et al., 2001). The reason for this variation between studies is not readily apparent since the models and colchicine treatment protocols overlap. Regarding Noco, a direct inhibitory effect of 33 µM Noco added for >2 min on GLUT4 activity was suggested by previous 3T3-L1 adipocyte studies (Molero et al., 2001; Shigematsu et al., 2002; Huang et al., 2005), However, Molero et al. also observed no inhibition of insulin-stimulated glucose uptake using 2 µM Noco for 1 hr despite complete disruption of microtubules (Molero et al., 2001), suggesting that Noco is useful to study the microtubule-glucose uptake connectivity at low concentrations. Since Noco was shown to inhibit GLUT4 activity within 2 min and we observed no effect of 13 µM Noco in isolated muscle fibers on basal glucose uptake or insulin-stimulated glucose uptake after 5 min at a concentration causing maximal inhibition of microtubule polymerization, we find it unlikely that Noco at the concentration used had major effects on GLUT4 activity.

We were unable to detect a significant effect of insulin on the number of GLUT4 moving on microtubules in the individual experiments performed in the current study. This contrasts the previously reported observation that rodent skeletal muscle expressing GFP-tagged GLUT4 had increased recovery of fluorescence after photo bleaching when stimulated by insulin, suggesting that insulin increased the overall GLUT4 movement (Lauritzen et al., 2008). Meanwhile, in primary and 3T3-L1 adipocyte cell culture total internal reflection fluorescence (TIRF) microscopy studies, insulin increased the number of GLUT4 halting and docking beneath the plasma membrane prior to insulin-stimulated insertion (Lizunov et al., 2005; Gonzalez et al., 2006; Lizunov et al., 2009; Bai et al., 2007). TIRF imaging of mouse muscle fibers expressing HA-GLUT4-GFP suggested a similar GLUT4 halting and fusion-promoting effect of insulin in skeletal muscle (Lizunov et al., 2012). Regarding microtubule dynamics, previous studies reported that insulin increased microtubule polymerization and/or density in 3T3-L1 adipocytes (Parker et al., 2019; Olson et al., 2003; Dawicki-McKenna et al., 2012) and in L6 skeletal muscle cells (Liu et al., 2013) but insulin also decelerated CLASP2-positive MT polymerization in 3T3-L1 adipocytes (Parker et al., 2019). In adult muscle, we did not detect any significant effect of insulin on the assessed parameters. Given that insulin may differentially increase or decrease the mobility of subpopulations of GLUT4 and microtubules, it seems likely that our relatively crude analyses of total subsarcolemmal GLUT4 movement could mask larger or even opposite behaviors of specific GLUT4 subpopulations.

The actin cytoskeleton is known to exhibit extensive interactions with microtubules (Pimm and Henty-Ridilla, 2021). Importantly, the actin cytoskeleton and associated regulators and binding proteins were suggested in L6 muscle cells and adult mouse muscle to be required for skeletal muscle GLUT4-dependent glucose transport (Sylow et al., 2013a; Sylow et al., 2013b; JeBailey et al., 2007; Tong et al., 2001; Brozinick et al., 2004; Ueda et al., 2010; Toyoda et al., 2011; Kee et al., 2015), likely by promoting the final steps of GLUT4 docking and fusion (Zaid et al., 2008). In adult muscle, the cytoplasmic beta and gamma-actin isoforms are lowly expressed compared to cultured cells and their individual KO in mouse muscle does not affect insulin-stimulated glucose uptake, making the exact actin isoform and mechanistic involvement of actin unclear (Madsen et al., 2018). Nevertheless, multiple studies support a role of actin in regulating adult muscle GLUT4-dependent glucose uptake. Hence, the interconnectivity between the actin and microtubule cytoskeletons in mediating GLUT4 trafficking should be investigated.

In conclusion, we presently demonstrated that GLUT4 is present on microtubules in adult mouse and human skeletal muscle and that acute microtubule disruption causes intramyocellular GLUT4 redistribution and eventually decreases insulin responsiveness of glucose transport. Decreased microtubule-dependent GLUT4 movement was observed in in vitro and in vivo mouse insulin resistance models, suggesting that disturbed microtubule-based GLUT4 trafficking is a feature of insulin resistance in adult skeletal muscle.

Materials and methods

Sample obtaining

Request a detailed protocol

Human muscle samples are tissue from m. vastus lateralis from young healthy men fasted for 6–7 hr. Further details on the subjects and tissue processing are described in previously published studies (Knudsen et al., 2020c; Steenberg et al., 2020). The study was approved by the Regional Ethics Committee for Copenhagen (H-6-2014-038; Copenhagen, Denmark) and complied with the guidelines of the 2013 Declaration of Helsinki. Written informed consent was obtained from all participants prior to entering the study.

Mouse muscle samples were from 10- to 16-week-old C57BL/6 mice. All animal experiments were approved by the Danish Animal Experimental Inspectorate or by the local animal experimentation committee of the Canton de Vaud under license 2890 and complied with the European Union legislation as outlined by the European Directive 2010/63/EU. The current work adheres to the standards outlined in the ARRIVE reporting guidelines. Male C57BL/6NTac mice, 16 weeks old, were used for the experiments including HFD fed mice. The mice were fed a 60% HFD or a control standard rodent chow diet ad libitum. For the rest of the experiments, the mice were female C57BL/6JRj aged 10–16 weeks fed ad libitum with a standard rodent chow diet. All mice were exposed to a 12 hr:12 hr light–dark cycle.

Gene transfer and fiber isolation

Request a detailed protocol

FDB muscles were electroporated in vivo similar to DiFranco et al., 2009 and isolated as previously described (Knudsen et al., 2019). The following plasmids were used: pB-GLUT4-7myc-GFP plasmid (a gift from Jonathan Bogan, Addgene plasmid #52872); p-mCherry-Tubulin-C1 (a gift from Kristien Zaal); HA-GLUT4-EOS (originally from the Zimmerberg laboratory, Lizunov et al., 2013, was a gift from Timothy McGraw); and p-EB3-GFP-N1 (originally from the Akhmanova laboratory, Stepanova et al., 2003, was a gift from Evelyn Ralston).

Fiber culturing and drug treatments

Request a detailed protocol

Experiments with isolated fibers were performed the day after isolation. For prolonged (15 hr) nocodazole treatment, nocodazole (M1404, Merck) was added for a final concentration of 4 μg/ml at this step. When palmitic acid treatment is indicated, this was added for a final concentration of 0.5 mM at this step as well. Palmitic acid was dissolved to a 200 mM solution in 1:1 ethanol and α-minimal essential medium (MEM), from which a 16× solution containing 100 mg/ml fatty acid-free bovine serum albumin (BSA) was made. Non-treated fibers were treated with BSA without palmitic acid. When indicated C2 ceramide (50 μM) (860502, Avanti), Paclitaxel (10 μM) (T7402, Merck), kinesore (6664, Tocris), or colchicine (25 μM) (C9754, Sigma) was added 2 hr prior to imaging/lysing whereas nocodazole (13 μM) was added 4 hr prior unless otherwise mentioned. For signaling analyses, 30 nM insulin (Actrapid, Novo Nordisk A/S) was added 15 min prior to lysing, for microscopic analyses 30 nM insulin was added 15–30 min prior to imaging. For fixation fibers were incubated with 4% paraformaldehyde (Electron Microscopy Sciences) in phosphate-buffered saline (PBS) for 20 min.

Cell culturing and experiments

Request a detailed protocol

Mycoplasma-free L6 rat myoblasts expressing myc-tagged GLUT4 (Wang et al., 1998) were maintained in α-MEM (12561056, Gibco) supplemented with 10% fetal bovine serum and 1% pen/strep antibiotic in a humidified atmosphere containing 5% CO2 and 95% air at 37°C. Differentiation to myotubes were achieved by lowering the serum concentration to 2% for 7 days. Visual myotube formation was used as authentication. For specific knockdown of KIF5B, shRNA constructs containing a 19-nucleotide (GGACAGATGAAGTATAAAT) sequence derived from mouse Kif5b mRNA (Zhao et al., 2020; a gift from Dr. Kwok-On Lai, City University of Hong Kong) were used using JetPRIME (Polyplus) according to the manufacturer’s protocol. As control shRNA with the sequence CCTAAGGTTAAGTCGCCCTCGCTCGAGCGAGGGCGACTTAACCTTAGG (Addgene plasmid # 1864) were used. Three days after initial transfection, the experiments were conducted as described in the figure legends. For GLUT4 translocation assessment, cell surface GLUT4myc was detected using a colorimetric assay (Wijesekara et al., 2006). Drug treatments were performed as described in the figure legends.

Human primary myoblasts (SK111, Cook MyoSite) were maintained and passaged in myotonic basal media (MB-2222, Cook MyoSite) supplemented with 10% myotonic basal media (MB-3333, Cook MyoSite) and 1% P/S antibiotic in a humidified atmosphere containing 5% CO2 and 95% air at 37°C. For differentiation and 2-DG transport measurements myoblasts were seeded at 50,000 cells/cm2 in collagen coated 96-well Cytostar-T scintillation plates (Perkin Elmer). After 2 days differentiation was initiated by switching to MEM (Gibco, 41090-028) with 2% horse serum, 10 µM 1-Dimethylethyl Ester (DAPT, Stemcell Technologies) and 1 µM Dabrafenib (Stemcell Technologies). Visual myotube formation was used as authentication. After 7 days of differentiation, myotubes were starved in PBS with magnesium and calcium and 10% MEM in 3.5 hr with DMSO/kinesore/Noco included the last 2 hr. Next, cells were incubated 15 min ± insulin (100 nM) before 14C-labeled 2-DG were added and tracer accumulation were measured for 3 hr.

Glucose uptake measurements

Request a detailed protocol

2-DG transport into Soleus and EDL muscles were assessed as described before (Knudsen et al., 2020b).

2-DG transport into L6 cells was measured by washing cells in PBS containing 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) and incubating them in PBS + HEPES containing 2-[3H] deoxyglucose for 5 min before cell harvest in lysis buffer. Tracer accumulation was then measured by liquid scintillation counting.

Electrochemical glucose sensing – Inspired by Trouillon et al., 2017, we fabricated a microfluidic chip system using standard soft lithographic techniques. The chip was divided into a tissue chamber and a glucose-sensing chamber connected by tubing. Both chambers were molded based on SU-8 master wafers. The tissue chamber consisted of two identical units each containing three layers of poly(dimethylsiloxane) (PDMS) (Figure 6A). The glucose-sensing chamber was a single PDMS layer bonded to a glass slide by air plasma (Figure 6B). Fluid connection was achieved by punching holes using biopsy punchers and inserting tubes in the portholes (Figure 6C). A customized electrode was fabricated by threading a working electrode (platinum wire, Ø 51 µm, Teflon coated) and a reference electrode (silver wire, Ø 75 µm, Teflon coated) in the lumen of an 18 G syringe needle and embedding them in fluid epoxy. A counter electrode was attached to the metal of the needle by silver paste (Figure 6D). On the experimental day, the electrode was carefully polished using fine sand paper and aluminum oxide slurry (0.05 µm particles) and a layer of chloride was deposited on the electrode by immersing it in 3 M KCl and exposing it to six current steps consisting of −20 µA for 1 s followed by 20 µA for 9 s. The working electrode was cleaned electrochemically in 0.1 M H2SO4 by running 10 cyclic voltammogram cycles. Next, to form an exclusion membrane on the sensor, a layer of poly-(m-phenylenediamine) (m-PD) was electropolymerized on to the working electrode by applying 20 s of 0.0 V, 5 min at 0.7 V and at 0.0 V. Finally the sensor was modified by addition of glucose oxidase by embedding a PBS solution consisting of glucose oxidase (60 mg ml−1), BSA (30 mg ml−1), and poly(ethylene glycol) diglycidyl ether (60 mg ml−1) and 2% glycerol on top of the electrode via 2 hr incubation at 50°C.

Figure 6
Download asset Open asset
Overview of muscle chip for glucose sensing.

(A) Overview of the different poly(dimethylsiloxane) (PDMS) layers for the tissue chamber unit. Scale bar = 5 mm. (B) Microfluidic system for the glucose-sensing chamber. The electrode was placed in the center of the system in the punched hole. (C) Overview of the muscle chip system showing the various layers of the tissue chamber as well as the connection to the glucose-sensing chamber. (D) Schematic drawing and picture of the customized electrode based on glucose oxidase fabricated to sense glucose. RE = reference electrode, CE = counter electrode, WE = working electrode.

In parallel, FDB fibers were isolated as described but cultured on 4-mm paper patches (filter 114, Whatman) by diluting the fiber solution 5:1 in extracellular matrix gel and adding 50 µl to each patch. Two hours prior to experiment, the fibers were starved from serum and glucose. Just before experiment start, the microfluidic system was assembled. First, the electrode was inserted into the detection chamber and calibrated by perifusing solutions with known glucose concentrations through the system. Next, the fiber-containing paper patch was inserted into the tissue chamber and perifused with serum-free Dulbecco´s minimal essential medium (DMEM) containing 4 mM glucose for 3–5 min. Then, the glucose concentration in the perifusate was monitored during basal and insulin-stimulated conditions and ∆glucose was calculated as previously described (Trouillon et al., 2017). Colchicine and Noco treatment for 2 hr was achieved by pre-incubating the paper patches containing fibers in colchicine and Noco and keeping the drugs in the perifusate at all times after the assembly of the tissue chamber. Noco treatment for 5 min was achieved by switching the perifusate to one containing Noco 5 min prior to insulin stimulation.

Western blotting

Request a detailed protocol

Samples were lysed in lysis buffer (50 mM Tris base, 150 mM NaCl, 1 mM Ethylenediaminetetraacetic acid (EDTA), 1 mM ethylene glycol-bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid (EGTA), 50 mM sodium fluoride, 5 mM pyrophosphate, 2 mM sodium orthovanadate, 1 mM dithiothreitol, 1 mM benzamidine, 0.5% protease inhibitor cocktail [Sigma P8340], 20% NP-40, pH 7.4) before processing as previously described (Knudsen et al., 2020a). The following antibodies were used: phospho (p)-Akt Thr308 (9275, CST), Akt (9272, CST), p-TBC1D4 Thr642 (4288, CST), or TBC1D4 (ab189890, Abcam). Coomassie staining was used as a loading control.

Immunolabeling

Request a detailed protocol

Human fiber bundles were teased into individual fibers and transferred to wells in a 24-well plate containing PBS using fine forceps. FDB fibers were similarly incubated in PBS. Fibers were washed 3 × 10 min in PBS and incubated in blocking buffer 1% bovine serum albumin (Merck), 5% goat serum (16210-064, Gibco), 0.1% Na Azide (247-852, Merck), 0.04% Saponin (27534-187, VWR) for 1 hr. The muscle fibers were then incubated in blocking buffer containing primary antibodies overnight at 4°. The next day, the fibers were washed 3 × 10 min in PBS containing 0.04% saponin and incubated in blocking buffer with Alexa 488 anti-rabbit or Alexa 568 anti-rabbit or anti mouse (Invitrogen) for 2 hr. Finally, the fibers were washed 3 × 10 min in PBS and mounted on glass slides in Vectashield (H-1000, Vector Laboratories) or imaged directly from the glass bottom dish. The following antibodies were used, raised in rabbit: GLUT4 (PA5-23052, Invitrogen), detyrosinated α-tubulin (AB48389, Abcam), Syntaxin6 (110 062, Synaptic Systems), or in mouse: GLUT4 (MAB8654, R&D Systems), α-tubulin (T9026, Merck).

Image acquisition and processing

Imaging was performed using the following systems: Zeiss 710, 780, 900, 980 or Elyra PS.1. Confocal imaging was performed using a Zeiss Plan-Apochromat ×63 1.4 NA objective. Laser source was an Argon laser emitting excitation light at 488 nm (25 mW) and helium neon laser emitting excitation light at 543 nm (1.2 mW), assembled by Zeiss. Emission light was collected by PMTs with matching beam splitters by Zeiss. The different channels were acquired sequentially. All live-imaging was performed in an integrated incubator at 37° in 5% CO2 and the fibers were kept in α-MEM containing drug/hormone as described. Specific imaging details for time series are provided in figure legends. Time series with 1 image per second for 60 s were obtained for mCherry-Tubulin and GLUT4-GFP dual color imaging. EB3-GFP time series were 1 image per second for 30 or 60 s. GLUT4-GFP imaging were time series of 30–300 s with 0.1–1 image per second. At all times, pixel size was kept at ≤90 × 90 nm. The pixel dwell time was 1.27 µs. To visualize dynamics, color-coded projections were generated as described in figure legends. In these projections, moving objects appear rainbow colored whereas static objects appear white.

Structured Illumination Microscopy (SIM) was performed using an Elyra PS.1 system (Zeiss), with a Zeiss Plan-Apochromat ×63 1.4 NA objective and a ×1.6 tube lens. The system was driven by Zen Black 2.3 SP1 from Zeiss which automatically assigns a diffraction pattern for each used wavelength (namely 28 µm for 488 nm, and 34 µm for 561 nm). Laser source was diode lasers, emitting excitation light at 488 nm (200 mW) and 561 nm (200 mW), assembled by Zeiss. Emission light was collected by a PCO.edge 5.5 sCMOS camera (PCO, Kelheim, Germany) with matching beam splitters by Zeiss. The different channels were acquired sequentially. Settings for image collection aimed at obtaining a dynamic range >5000 grayscale levels, and bleaching was assessed to be <20% of the dynamic range across the imaging sequences. 3D stacks were acquired at 100 nm steps by using a PI E-655 Z-piezo drive by Physik Instrumente (Karlsruhe, Germany).

Unless otherwise noted, images shown are single frames. For visualization purposes only, some images were cropped and contrast or levels were linearly adjusted. Images were processed using ImageJ (Tong et al., 2001) and Adobe Photoshop 21.2.1 (Adobe).

3D-SIM image reconstruction

Request a detailed protocol

The structured illumination image processing module built in Zen Black 2.3 was used for reconstruction, keeping a theoretical PSF and the same noise filter parameter (−5.5) for all the processed images. The resulting super-resolution images were kept in raw scaling and were baseline shifted. Super-resolution image quality was assessed by applying FFT to the reconstructed images, compared to FFT of the widefield acquisition. System performance was assessed by using an Argolight SIM patterned standard sample (Argolight, Pessac, France), obtaining resolutions ~120 nm consistently, and PSF was evaluated using 100 nm TetraSpeck beads from Zeiss.

Live-imaging videos

Request a detailed protocol

Representative videos were generated from the live-imaging time series at 10 frames per second. For GLUT4-GFP and mCherry-Tubulin dual color time series (Figure 1—video 1), the video was generated by merging the two channels with the GLUT4-GFP channel being green and the tubulin channel being magenta. To facilitate visualization, the single-color time series (Figure 1—video 2, Figure 2—video 1, Figure 5—video 1) were generated by removing every other frame and switching the colors from green to red between remaining frames.

Depletion and relocalization analysis

Request a detailed protocol

Using the particle analysis tool in ImageJ (Schneider et al., 2012), GLUT4-GFP structures (sized >0.02 µm2, circularity between 0 and 1) from background and threshold adjusted 8-bit images were identified and structure areas were determined. From this, small structures <0.04 µm2 were counted and related to the total number of structures as a reference for overexpression. These structures were sized from the smallest resolvable and up to ~225 nm in diameter. We analyzed this fraction since insulin induces membrane insertion of small 50–70 nm GSVs (Martin et al., 1997) which are expected to be part of this fraction. For relocalization analysis, structures were identified as for the depletion analysis. The individual structures were allocated into one of the following three groups: large structures (>4 µm2), medium-sized structures (between 0.4 and 4 µm2), and small structures (<0.4 µm2).

Polymerization rate analysis

Request a detailed protocol

Via calponin homology domains, EB3 proteins interact with tubulin at the microtubule tip and can thus be used to label polymerizing microtubules (Stepanova et al., 2003). On 8-bit threshold-adjusted images, the tip of polymerizing microtubules were identified as a 0.08–0.2 µm2 region with a circularity between 0.2 and 1 with accumulated EB3-GFP signal. Based on these criteria, the average number of polymerizing microtubules was calculated per image in a 30- or 60-s time series with an image every second.

Tracking analysis was performed on 8-bit images using the ImageJ plug in TrackMate. Threshold was adjusted in TrackMate and tracking settings were adjusted to maximize fitting of the automated tracking. The following tracking settings were used, for GLUT4-GFP: LoG detector, 0.5 µm blob diameter estimate, LAP tracker, 1.5 µm for gap-closing and max. distance, and a max. frame gap of 1. Settings were similar for EB3-GFP, except 2 µm was used as gap-closing distance, max. distance was 1 µm and max. frame gap was 2. Tracks with a <1.5 µm displacement was not considered microtubule-based GLUT4 movement or microtubule polymerization and excluded from further analysis.

Statistical analyses

Request a detailed protocol

Results are shown as mean, mean with individual values or mean ± SD. Statistical testing was performed using t-test or one- or two-way analysis of variance (ANOVA as described in the figure legends. Tukey’s post hoc test was performed following ANOVA). The significance level was set at p < 0.05.

Data availability

All data generated or analyzed during this study are included in the manuscript and supporting files.

References

    1. Fletcher LM
    2. Welsh GI
    3. Oatey PB
    4. Tavaré JM
    (2000)
    Role for the microtubule cytoskeleton in GLUT4 vesicle trafficking and in the regulation of insulin-stimulated glucose uptake
    The Biochemical Journal 352 Pt 2:267–276.

Decision letter

  1. Michael Czech
    Reviewing Editor; University of Massachusetts Medical School, United States
  2. Anna Akhmanova
    Senior Editor; Utrecht University, Netherlands
  3. Amira Klip
    Reviewer; Hospital for Sick Children, Canada

Our editorial process produces two outputs: (i) public reviews designed to be posted alongside the preprint for the benefit of readers; (ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.

Decision letter after peer review:

Thank you for submitting your article "Microtubule-mediated GLUT4 trafficking is disrupted in insulin resistant skeletal muscle" for consideration by eLife. Your article has been reviewed by 2 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Anna Akhmanova as the Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Amira Klip (Reviewer #3).

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential revisions:

1) This issue of microtubules being required for proper localization of Glut4 into an insulin sensitive compartment is viewed as important, and the data are supportive. However, there are further experiments that should be performed to test or strengthen this point (see Reviewer 2 comments on first weakness, and Reviewer 3 comments on Figure 2 in particular).

2) The Kinesore data do not appear to support the authors' hypothesis (see Reviewer 2 comments in particular). Please address this issue in a substantive way, with additional data or commentary and appropriate interpretations.

Reviewer #1 (Recommendations for the authors):

Perhaps a better way to probe for the effects of insulin on GLUT4 dynamics would be to use the FRAP assay they employed to measure anterograde and retrograde trafficking in basal condition (FigS1G-H). This assay may be a more robust means of capturing the effects of insulin on GLUT4 dynamics; doing so would greatly strengthen the manuscript.

The manuscript would benefit from more detailed discussion of their results and perhaps editing of the amount of data presented. Perhaps too much narrative has been removed in trying to aware of the overall the length of the text. There are too many instances where data are presented as an observation but not fully explored. As an example, the statement on line 94-95 is too vague/descriptive as are the data in Figure 1SD. In my view, focusing the manuscript by removing extraneous data and more clearly describing the findings and how they advance our understanding of GLUT4 in muscle will enhance the impact of the work.

Reviewer #2 (Recommendations for the authors):

Overall, there are a number of technical questions (including better analysis of the colocalization of GLUT4 with Syntaxin 6) and throughout there is marked need to include the control of Noco treatment in absence of C2 or insulin.

Figure 1:

– A and B, nice parallel between mouse and human morphology and microtubule structural similarities. Please comment on fiber types – it appears that the general structure is not dependent on fiber type, but future phenotypes may be explained by fiber type, so it would be useful to introduce this variable early on, or at least in the discussion. Also, please clarify that FDB is used because its amenability for fiber isolation, and discuss how it compares to the intact muscles used.

Figure 1 – —figure supplement 1 – G and H = Y axis states "normalized" – please specify normalized to what, total fluorescence intensity?

Figure 2:

– Data describing GLUT4 dependence on microtubules for traffic is sound, but the dependence of insulin-stimulated GLUT4 traffic on intact microtubules is lacking the essential control of the effect of Noco in the basal state (A-C).

– While in aggregate the results support the concept that GLUT4 traffic along microtubules is required for proper GLUT4 sortin, but not for acute insulin mobilization of GLUT4 from GSV. However, testing the effect of insulin in all conditions would go a long way to clarify this concept, if possible.

– Is the effect of Noco on the insulin response (minute) seen also with other stimuli? AICAR? Contraction? It may be that this is work for a future study, and could be commented so.

– Please consider to stimulate with insulin to mobilize GLUT4, then disrupt microtubules and perform a re-stimulation with insulin – in this way both acute (1 stimulation) and prolonged (re-stimulation) insulin-regulated GLUT4 traffic along microtubules could be examined

– B-C, Are the data in C the results from 4 experiments, as performed in B? It seems somewhat misleading if the variability within a single experiment is as shown in B, but when pooled in C the effect is now apparent when there was no significant insulin effect in any single experiment. If C is indeed an average of 4 independent experiments, why is only a single experiment shown in B?

– Please clarify the sources of data for B and C, and include error bars for both B and C. Also specify in C, normalized to what?

– D-E, 15 hours Noco, then 9 hours recovery – why these hours? If previous experiment used 4 hours, why are 15 hours required to disrupt microtubules in this experiment?

o Is there a reason why this experiment was not also conducted in the presence of insulin? Similarly, in F-G, would this localization change in presence of insulin?

– Please indicate "mouse" in G

– In supplement 1 to this figure:

– B-C, Normalized to what?

– D , Why was basal+Noco not tested here but in E-G insulin is not tested?

– D, Noco treatment back to 4 hours but in F-H treatment is 15 hours please state the rationale for treatment times

– It appears the Pearson's R value for Syntaxin 6 and GLUT4 was applied to the entire fiber, or is it the Entire image? Fields? In any case, it would be highly useful to segregate colocalization according to depots of GLUT4: at nucleation site clusters, intermediate structures and small structures. This could be far more informative than the overall which will likely muddle any potential colocalization. Additionally, since these are overexpressed proteins, it is important to calculate the colocalization of one protein within the other's pool and vice versa.

Figure 3:

– A-B, glucose uptake phenotype also nicely demonstrated for FDB in C-F. Please suggest explanation for why this insulin effect was not apparent in Figure 2?

– In figure 2, could endogenous GLUT4 be detected in static images to examine localization relative to microtubules in soleus or EDL muscles? This could provide some context as to fiber type specific dependence on microtubules for insulin-mediated GLUT4 traffic.

– Noco treatment only 2 hours, compared to 4-15 hours when measuring GLUT4 traffic , what is the rationale for this difference?

– Please provide potential explanations for why for microtubules are important for the acute insulin action on on 2-DG uptake but not GLUT4 traffic. This is a key question that must be addressed.

– H, What is different from Figure 2E – in 3H the number of small structures is elevated with 2 hours Noco but in 1E they are reduced. Does the presence of insulin alter this distribution?

– G-H, What does GLUT4 localization look like in presence of insulin +/- Noco?

Figure 4

– Nice demonstration of Kinesin-1 dependence. In C, It would be useful to see this performed with insulin +/- Noco and present in Figure 2

Figure 5:

– 5B and supplement 1B, why is there no basal + C2? If the conclusion is that C2 inhibits insulin stimulated vesicle movement, it is essential to show that this is not happening in the basal state

Discussion:

– The finding that more GLUT4 'vesicles' appear in intermediate and small structures along the microtubule in response to Noco is fascinating. Given that those pools increase, they are unlikely to be the ones responding to insulin, unless they no longer engage with the elements that cause mobilization (whether Kinesin-1 or other). On the other hand, Kinesore causes GLUT4 dispersion into small clusters and promotes translocation to the membrane even additive to insulin's action. Please reconcile these observations.

– The dispersion of GLUT4 upon activation of Kinesin-1 via Kinesore is also very interesting, but requires further assessment of specificity towards Kinesin-1.

– One cannot call the 'puncta' vesicles, these are likely aggregates unless the point spread function is calculated for each item and found to be individual

– Throughout the study there is no mention of the other fiber and motors that contribute to GLUT4 traffic in muscle. In the past the authors have not detected insulin-dependent remodeling of transfected actin, but this does not rule out a dependency of the positioning of the GLUT4 pools on actin cytoskeletal elements. -

– Key and related to the above is testing whether Noco for 2 h alters any of the known elements of the Rac1 signaling pathway or the proteins regulating actin filaments. There are studies in the literature showing interconnectivity between the microtubule and microfilament networks.

https://doi.org/10.7554/eLife.83338.sa1

Author response

Essential revisions:

1) This issue of microtubules being required for proper localization of Glut4 into an insulin sensitive compartment is viewed as important, and the data are supportive. However, there are further experiments that should be performed to test or strengthen this point (see Reviewer 2 comments on first weakness, and Reviewer 3 comments on Figure 2 in particular).

2) The Kinesore data do not appear to support the authors' hypothesis (see Reviewer 2 comments in particular). Please address this issue in a substantive way, with additional data or commentary and appropriate interpretations.

Thank you for the comprehensive and constructive feedback on our work. Based on the comments from the reviewers and the reviewing editor, and with particular focus on the above points, we have performed additional experiments to better understand the connection between insulin-regulated GLUT4 localization and to characterize the effect of kinesore in adult muscle in more detail. Our new findings are integrated into an updated and improved version of the manuscript. We address the reviewer comments point by point below.

Reviewer #1 (Recommendations for the authors):

Perhaps a better way to probe for the effects of insulin on GLUT4 dynamics would be to use the FRAP assay they employed to measure anterograde and retrograde trafficking in basal condition (FigS1G-H). This assay may be a more robust means of capturing the effects of insulin on GLUT4 dynamics; doing so would greatly strengthen the manuscript.

Thank you for this proposal. We agree that FRAP experiment would be a good idea and may be a more robust means of capturing an insulin effect. Unfortunately, the first author of this paper is no longer employed in the lab and thus it is currently not feasible to perform this experiment.

Worth mentioning, we previously attempted to isolate an insulin effect on anterograde kinesin-dependent trafficking by using the cytoplasmic dynein inhibitor Ciliobrevin D. However, we could also not detect any effect of insulin using this approach.

The manuscript would benefit from more detailed discussion of their results and perhaps editing of the amount of data presented. Perhaps too much narrative has been removed in trying to aware of the overall the length of the text. There are too many instances where data are presented as an observation but not fully explored. As an example, the statement on line 94-95 is too vague/descriptive as are the data in Figure 1SD. In my view, focusing the manuscript by removing extraneous data and more clearly describing the findings and how they advance our understanding of GLUT4 in muscle will enhance the impact of the work.

We have modified the manuscript to more clearly focus on the data supporting our main conclusions. However, we also believe that publishing descriptive data and odd observations rather than just a polished storyline increases the opportunities for others to interpret and build on the work. Thus, we still include descriptive/observational data in this updated version of the manuscript.

Reviewer #2 (Recommendations for the authors):

Overall, there are a number of technical questions (including better analysis of the colocalization of GLUT4 with Syntaxin 6) and throughout there is marked need to include the control of Noco treatment in absence of C2 or insulin.

Figure 1:

– A and B, nice parallel between mouse and human morphology and microtubule structural similarities. Please comment on fiber types – it appears that the general structure is not dependent on fiber type, but future phenotypes may be explained by fiber type, so it would be useful to introduce this variable early on, or at least in the discussion. Also, please clarify that FDB is used because its amenability for fiber isolation, and discuss how it compares to the intact muscles used.

Figure 1 – —figure supplement 1 – G and H = Y axis states "normalized" – please specify normalized to what, total fluorescence intensity?

Thanks. The rationale behind choice of muscle and their fiber type distribution has been addressed (line 78-81). Normalization was min-max normalization, which is now specified in the legend.

Figure 2:

– Data describing GLUT4 dependence on microtubules for traffic is sound, but the dependence of insulin-stimulated GLUT4 traffic on intact microtubules is lacking the essential control of the effect of Noco in the basal state (A-C).

The microtubule disruption by nocodazole in the insulin-stimulated state completely blocked all microtubule-mediated GLUT4 movement. This is clear from figure 2 A&B and particularly from the uploaded video showing no microtubule-mediated GLUT4 trafficking following 2h of microtubule disruption. Thus, a further reduction in GLUT4 movement without insulin stimulation is not possible. For this reason, we believe the proposed Noco treated basal fiber control to be redundant.

– While in aggregate the results support the concept that GLUT4 traffic along microtubules is required for proper GLUT4 sortin, but not for acute insulin mobilization of GLUT4 from GSV. However, testing the effect of insulin in all conditions would go a long way to clarify this concept, if possible.

For practical reasons, we are limited in the experiments we can currently perform in the lab. Thus, additional GLUT4-GFP electroporation and fiber isolation experiments are not possible to perform within a reasonable time-frame.

As a proxy measure of GLUT4 mobilization from GSV, we tested whether insulin would reduce the number of small vesicles (<0.4 µm2) in our live imaging experiments shown in figure 2A-B. However, we could not detect any significant insulin regulation by this approach.

– Is the effect of Noco on the insulin response (minute) seen also with other stimuli? AICAR? Contraction? It may be that this is work for a future study, and could be commented so.

This is of course an interesting and important question. We have not studied this. In the new experiments performed in our revision process, we observed a marked increase in AMPK activation (p-AMPK and p-ACC) following kinesore treatment, both in basal and insulin stimulated EDL and soleus muscle. Despite this AMPK activation, we observed no increase in glucose uptake in the basal state and impaired insulin-stimulated glucose uptake. The prediction from these data would be that the effect of Noco/kinesore would also be seen with AICAR/AMPK activation and contractions.

– Please consider to stimulate with insulin to mobilize GLUT4, then disrupt microtubules and perform a re-stimulation with insulin – in this way both acute (1 stimulation) and prolonged (re-stimulation) insulin-regulated GLUT4 traffic along microtubules could be examined

We stimulated L6 cells with insulin (100nm, 15 min) before 2h ± Noco or Colchicine and then performed re-stimulation with insulin or kept cells basal. There was a clear main effect of both Noco and Colchicine when cells were pre-treated with insulin. Overall, there was no clear effect of microtubule disruption on insulin-stimulated GLUT4 translocation when cells were pre-stimulated with insulin before microtubule disruption. From these data, it seems that, in the L6 cells overexpressing GLUT4, a residual insulin-responsive GLUT4 pool remains after microtubule disruption, even with prior GLUT4 mobilization by insulin.

– B-C, Are the data in C the results from 4 experiments, as performed in B? It seems somewhat misleading if the variability within a single experiment is as shown in B, but when pooled in C the effect is now apparent when there was no significant insulin effect in any single experiment. If C is indeed an average of 4 independent experiments, why is only a single experiment shown in B?

Yes, data are from 4 experiments performed as in B. Each data point set reflects the average effect size from each experiment. In addition to the experiment shown in B, the data in C also originate from the basal and insulin control groups in figure 5.

Because this particular figure panel was highlighted as confusing by two reviewers and since we could not detect an effect of insulin in single experiments, we decided to remove the panel from the manuscript. We discuss several possibilities for the lack of significant insulin effect on GLUT4 movement in individual experiments in the Discussion section (lines 342 to 361).

– Please clarify the sources of data for B and C, and include error bars for both B and C. Also specify in C, normalized to what?

See reply above.

– D-E, 15 hours Noco, then 9 hours recovery – why these hours? If previous experiment used 4 hours, why are 15 hours required to disrupt microtubules in this experiment?

For practical reasons, we choose to do an overnight treatment with nocodazole, then image the fibers in the morning, allow recovery throughout the working day and re-image the fibers again in the afternoon. This amounted to 15h Noco treatment and 9h recovery. As seen from the representative live-imaging of the experiment in 2A-B, the changes induced by microtubule disruption had occurred already after 4h. Hence, we do not believe that a prolonged 15h disruption would be required for our localization observations. However, as also noted by the reviewer in a later comment, the number of small GLUT4 membrane structures was higher after 2h of nocodazole treatment compared to the control condition while it was lower after 15h. We have not investigated this time-dependent effect further.

– Is there a reason why this experiment was not also conducted in the presence of insulin? Similarly, in F-G, would this localization change in presence of insulin?

The research question we asked in this experiment was whether microtubule disruption would eventually result in GLUT4 mis/re-localization and whether this was reversible.

Whether insulin stimulation would affect mis/re-localized GLUT4 is an interesting question, but would in our opinion have been better addressed by an isolated experiment.

For F-G, our hypothesis was that insulin would reduce GLUT4 Syntaxin6 co-localization. Since the correlation between GLUT4 and Syntaxin6 was low already in the basal state and unaffected by nocodazole we did not pursue this hypothesis further.

– Please indicate "mouse" in G

Corrected.

– In supplement 1 to this figure:

– B-C, Normalized to what?

We normalized the individual intensity to the average intensity across the full data set.

– D , Why was basal+Noco not tested here but in E-G insulin is not tested?

The reasoning for the different treatments is similar in figure 2 and in the supplement to figure 2.

– D, Noco treatment back to 4 hours but in F-H treatment is 15 hours please state the rationale for treatment times

The treatment times were different for practical reasons outlined above.

– It appears the Pearson's R value for Syntaxin 6 and GLUT4 was applied to the entire fiber, or is it the Entire image? Fields? In any case, it would be highly useful to segregate colocalization according to depots of GLUT4: at nucleation site clusters, intermediate structures and small structures. This could be far more informative than the overall which will likely muddle any potential colocalization. Additionally, since these are overexpressed proteins, it is important to calculate the colocalization of one protein within the other's pool and vice versa.

Thank you for pointing this out. We have reanalysed the data specifically analysing the perinuclear region and using Manders´ correlation to estimate the fraction of GLUT4 in Syntaxin 6 positive regions. We did not see any changes in overlap by Noco treatment (0.269 vs. 0.273, p=0.96).

Figure 3:

– A-B, glucose uptake phenotype also nicely demonstrated for FDB in C-F. Please suggest explanation for why this insulin effect was not apparent in Figure 2?

Please see our response to weakness 1 by reviewer #2.

– In figure 2, could endogenous GLUT4 be detected in static images to examine localization relative to microtubules in soleus or EDL muscles? This could provide some context as to fiber type specific dependence on microtubules for insulin-mediated GLUT4 traffic.

Indeed, this would be possible to do. However, our ex vivo incubation glucose uptake data suggest no major differences between EDL and Soleus. For this reason, we would expect no major fiber type differences and did not prioritize these analyses.

– Noco treatment only 2 hours, compared to 4-15 hours when measuring GLUT4 traffic , what is the rationale for this difference?

Our ex vivo preparations are difficult to keep viable for extended time periods. Since we hypothesized that the immediate insulin-induced glucose uptake would be preserved whereas longer duration microtubule disruption would be inhibitory, we decided on a long ex vivo incubation time at which the muscles would still be viable, i.e. 2h.

We judged that the additional information from re-performing the isolated FDB fiber experiments conducted in figure 1 and 2 using only 2h Noco would be limited and thus decided not to perform these experiments.

– Please provide potential explanations for why for microtubules are important for the acute insulin action on on 2-DG uptake but not GLUT4 traffic. This is a key question that must be addressed.

We are unsure whether there has been a mistake in posing this comment as we see the opposite, i.e. that microtubules are important for GLUT4 traffic but not acute glucose uptake. Our working hypothesis is that the muscle fiber retains a GLUT4 reservoir that is capable of reaching the surface membrane even in the absence of microtubules, likely from a pre-tethered pool as proposed by Samuel Cushman’s group (PMID: 22297303). This reservoir may allow the muscle fiber to retain a normal initial insulin response on glucose uptake. We discuss this hypothesis in the discussion (lines 308-321).

If this question instead relates to the fact that insulin-stimulated glucose uptake is impaired after 2h of microtubule disruption despite no effect of insulin on microtubule-mediated GLUT4 trafficking, then our response is the following:

In addition to the 2-DG uptake measurements ±Noco, we have now demonstrated that Kinesore treatment (2h, 50uM) also impairs insulin-stimulated 2-DG uptake in adult skeletal muscle, overall supporting the requirement of microtubule-mediated transport for 2-DG uptake. Additionally we have clearly demonstrated that the microtubules are important for adult muscle GLUT4 traffic. As pointed out, we have not demonstrated that the microtubules are important for the acute action of insulin on GLUT4 traffic in muscle fibers. We hypothesize that a fraction of the insulin-regulated GLUT4 trafficking is dependent on intact microtubules, but we currently do not have sensitive enough tools to fully test this hypothesis as discussed in the current version of the manuscript (lines 341 to 361).

– H, What is different from Figure 2E – in 3H the number of small structures is elevated with 2 hours Noco but in 1E they are reduced. Does the presence of insulin alter this distribution?

We have not investigated whether insulin would alter this distribution, nor have we investigated why there is a difference between the two time points. It would be really interesting to explore these observations further, especially since the small structures would presumably include insulin-responsive vesicles. However, as also pointed out by reviewer #2, the data would be hard to interpret without identification by compartment specific markers. Since this would require a lot of optimization to set up, we find this to lie beyond the scope of the current study.

– G-H, What does GLUT4 localization look like in presence of insulin +/- Noco?

We quantified the number of small GLUT4 structures following insulin stimulation +/- Noco in the live-imaging experiment in Figure 2, but we did not observe significant changes (see figure 4 in this rebuttal). These data were not included in this manuscript.

We believe that the lack of effect is due to lack of sensitivity in the image analyses. Supporting this notion, we previously demonstrated, that insulin-stimulated GLUT4 translocation can occur without discernible visual GLUT4 localization changes in mouse muscle (PMID: 30710396), indicating that higher resolution microscopy and more specific image analyses approaches may be a requirement.

Figure 4

– Nice demonstration of Kinesin-1 dependence. In C, It would be useful to see this performed with insulin +/- Noco and present in Figure 2

During the revision work, we observed that kinesore impairs insulin-stimulated glucose uptake in incubated mouse muscle and primary human myotubes. This has made us unsure of the interpretation of the data presented in figure 4C and, also considering also reviewer 2´s comments about too many underexplored observations, we thus chose to remove it from this version of the manuscript.

Figure 5:

– 5B and supplement 1B, why is there no basal + C2? If the conclusion is that C2 inhibits insulin stimulated vesicle movement, it is essential to show that this is not happening in the basal state

We retrospectively agree that this is an essential control to interpret the insulin-effect. We therefore acknowledged this limitation by changing the wording in the manuscript to reflect that we cannot conclude whether C2 inhibits the effect of insulin (lines 218-222).

Discussion:

– The finding that more GLUT4 'vesicles' appear in intermediate and small structures along the microtubule in response to Noco is fascinating. Given that those pools increase, they are unlikely to be the ones responding to insulin, unless they no longer engage with the elements that cause mobilization (whether Kinesin-1 or other). On the other hand, Kinesore causes GLUT4 dispersion into small clusters and promotes translocation to the membrane even additive to insulin's action. Please reconcile these observations.

We included some speculative discussion on the identity of the structures (lines 311-318). Future studies should aim to identify the nature of these structures.

– The dispersion of GLUT4 upon activation of Kinesin-1 via Kinesore is also very interesting, but requires further assessment of specificity towards Kinesin-1.

We agree that the dispersion of GLUT4 upon kinesore treatment is very interesting but requires further assessment. We are currently uncertain of the interpretation of the dispersion data and considering also reviewer 2´s comments about too many underexplored observations, we decided to remove the data from this manuscript.

– One cannot call the 'puncta' vesicles, these are likely aggregates unless the point spread function is calculated for each item and found to be individual

We have rephrased to “structures” throughout the manuscript.

– Throughout the study there is no mention of the other fiber and motors that contribute to GLUT4 traffic in muscle. In the past the authors have not detected insulin-dependent remodeling of transfected actin, but this does not rule out a dependency of the positioning of the GLUT4 pools on actin cytoskeletal elements. -

– Key and related to the above is testing whether Noco for 2 h alters any of the known elements of the Rac1 signaling pathway or the proteins regulating actin filaments. There are studies in the literature showing interconnectivity between the microtubule and microfilament networks.

Indeed our data fit well in a working model in which microtubules mediate long range movement of GLUT4 towards the cell periphery where the microtubules could be interconnected with the actin cytoskeleton that could mediate the peripheral trafficking and docking steps in the translocation process. We included a small discussion of the evidence for the involvement of actin in L6 and adult muscle and the possible crosstalk between the two cytoskeleton types in l. 400-410.

https://doi.org/10.7554/eLife.83338.sa2

Article and author information

Author details

  1. Jonas R Knudsen

    1. August Krogh Section for Molecular Physiology, Department of Nutrition, Exercise and Sports, Faculty of Science, University of Copenhagen, Copenhagen, Denmark
    2. Heart and Skeletal Muscle Biology, Global Drug Discovery, Novo Nordisk, Soeborg, Denmark
    Contribution
    Conceptualization, Resources, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Visualization, Methodology, Writing – original draft, Project administration, Writing – review and editing
    For correspondence
    jrk@nexs.ku.dk
    Competing interests
    Affiliated with Novo Nordisk A/S
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-5471-491X

  2. Kaspar W Persson

    August Krogh Section for Molecular Physiology, Department of Nutrition, Exercise and Sports, Faculty of Science, University of Copenhagen, Copenhagen, Denmark
    Contribution
    Validation, Investigation, Writing – review and editing
    Competing interests
    No competing interests declared

  3. Carlos Henriquez-Olguin

    1. August Krogh Section for Molecular Physiology, Department of Nutrition, Exercise and Sports, Faculty of Science, University of Copenhagen, Copenhagen, Denmark
    2. Exercise Science Laboratory, Faculty of Medicine, Universidad Finis Terrae, Santiago, Chile
    Contribution
    Validation, Investigation, Writing – review and editing
    Competing interests
    No competing interests declared

  4. Zhencheng Li

    1. August Krogh Section for Molecular Physiology, Department of Nutrition, Exercise and Sports, Faculty of Science, University of Copenhagen, Copenhagen, Denmark
    2. College of Physical Education, Chongqing University, Chongqing, China
    Contribution
    Investigation, Writing – review and editing
    Competing interests
    No competing interests declared

  5. Nicolas Di Leo

    August Krogh Section for Molecular Physiology, Department of Nutrition, Exercise and Sports, Faculty of Science, University of Copenhagen, Copenhagen, Denmark
    Contribution
    Investigation, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-6268-890X

  6. Sofie A Hesselager

    August Krogh Section for Molecular Physiology, Department of Nutrition, Exercise and Sports, Faculty of Science, University of Copenhagen, Copenhagen, Denmark
    Contribution
    Investigation, Writing – review and editing
    Competing interests
    No competing interests declared

  7. Steffen H Raun

    Department of Biomedical Sciences, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark
    Contribution
    Investigation, Writing – review and editing
    Competing interests
    No competing interests declared

  8. Janne R Hingst

    Clinical Drug Development, Novo Nordisk, Soeborg, Denmark
    Contribution
    Investigation, Writing – review and editing
    Competing interests
    Affiliated with Novo Nordisk A/S

  9. Raphaël Trouillon

    Department of Electrical Engineering, Polytechnique Montréal, Montréal, Canada
    Contribution
    Supervision, Writing – review and editing
    Competing interests
    No competing interests declared

  10. Martin Wohlwend

    Laboratory of Integrative Systems Physiology, Institute of Bioengineering, École Polytechnique Fédérale de Lausanne, Lausanne, Switzerland
    Contribution
    Investigation, Writing – review and editing
    Competing interests
    No competing interests declared

  11. Jørgen FP Wojtaszewski

    August Krogh Section for Molecular Physiology, Department of Nutrition, Exercise and Sports, Faculty of Science, University of Copenhagen, Copenhagen, Denmark
    Contribution
    Investigation, Writing – review and editing
    Competing interests
    has ongoing collaborations with Pfizer inc and Novo Nordisk A/S unrelated to this study

  12. Martin AM Gijs

    Microsystems Laboratory 2, Institute of Electrical and Micro Engineering, École Polytechnique Fédérale de Lausanne, Lausanne, Switzerland
    Contribution
    Supervision, Writing – review and editing
    Competing interests
    No competing interests declared

  13. Thomas Elbenhardt Jensen

    August Krogh Section for Molecular Physiology, Department of Nutrition, Exercise and Sports, Faculty of Science, University of Copenhagen, Copenhagen, Denmark
    Contribution
    Conceptualization, Resources, Supervision, Funding acquisition, Investigation, Writing – original draft, Project administration, Writing – review and editing
    For correspondence
    tejensen@nexs.ku.dk
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-6139-8268

Funding

Novo Nordisk Fonden (15182)

  • Thomas Elbenhardt Jensen

Novo Nordisk Fonden (16OC0023046)

  • Jørgen FP Wojtaszewski

Novo Nordisk Fonden (17SA0031406)

  • Jonas R Knudsen

Lundbeckfonden (R313-2019-643)

  • Thomas Elbenhardt Jensen

Lundbeckfonden (R266-2017-4358)

  • Jørgen FP Wojtaszewski

Sundhed og Sygdom, Det Frie Forskningsråd (FSS8020-00288B)

  • Jørgen FP Wojtaszewski

Sundhed og Sygdom, Det Frie Forskningsråd (#9058-00047B)

  • Jonas R Knudsen

Danish Diabetes Academy (NNF17SA0031406)

  • Carlos Henriquez-Olguin

Independent Research Fund Denmark (#9058-00047B)

  • Carlos Henriquez-Olguin

The funders had no role in study design, data collection, and interpretation, or the decision to submit the work for publication.

Acknowledgements

We acknowledge the Core Facility for Integrated Microscopy, Faculty of Health and Medical Sciences, University of Copenhagen and especially Pablo Hernández-Veras for his guidance with the SIM imaging. We acknowledge Prof. Johan Auwerx (École Polytechnique Fédérale de Lausanne) for help with the mouse work. Finally, we acknowledge Profs. Bente Kiens and Erik A Richter (Department of Nutrition, Exercise and Sports, University of Copenhagen) for the expertise and help with conducting the human study. This study was financed by grants to TEJ (Novo Nordisk Foundation [NNF] Excellence project #15182), Lundbeck Foundation (LF) Ascending Investigator (R313-2019-643), to JFPW (NNF16OC0023046, LF R266-2017-4358 and the Danish Research Medical Council FSS8020-00288B), to JRK (a research grant from the Danish Diabetes Academy [DDA], which is funded by the NNF, NNF17SA0031406 and an International Postdoc grant from the Independent Research Fund Denmark, #9058-00047B), to CHO (Postdoc research grant from DDA, #NNF17SA0031406).

Ethics

The work involving human subjects was approved by the Copenhagen Ethics Committee (H-6-2014-038; Copenhagen, Denmark) and complied with the guidelines of the 2013 Declaration of Helsinki. Informed written consent was obtained from all subjects prior to entering the study.

All animal experiments were approved by the Danish Animal Experimental Inspectorate or by the local animal experimentation committee of the Canton de Vaud under license 2890 and complied with the European Union legislation as outlined by the European Directive 2010/63/EU. The current work adheres to the standards outlined in the ARRIVE reporting guidelines.

Senior Editor

  1. Anna Akhmanova, Utrecht University, Netherlands

Reviewing Editor

  1. Michael Czech, University of Massachusetts Medical School, United States

Reviewer

  1. Amira Klip, Hospital for Sick Children, Canada

Version history

  1. Received: September 8, 2022
  2. Preprint posted: September 22, 2022 (view preprint)
  3. Accepted: April 2, 2023
  4. Accepted Manuscript published: April 19, 2023 (version 1)
  5. Version of Record published: May 10, 2023 (version 2)

Copyright

© 2023, Knudsen et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 1,149
    Page views
  • 255
    Downloads
  • 1
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Jonas R Knudsen
  2. Kaspar W Persson
  3. Carlos Henriquez-Olguin
  4. Zhencheng Li
  5. Nicolas Di Leo
  6. Sofie A Hesselager
  7. Steffen H Raun
  8. Janne R Hingst
  9. Raphaël Trouillon
  10. Martin Wohlwend
  11. Jørgen FP Wojtaszewski
  12. Martin AM Gijs
  13. Thomas Elbenhardt Jensen
(2023)
Microtubule-mediated GLUT4 trafficking is disrupted in insulin-resistant skeletal muscle
eLife 12:e83338.
https://doi.org/10.7554/eLife.83338

Categories and tags

Research organisms

Further reading

    1. Biochemistry and Chemical Biology
    2. Cell Biology

    Fluorescein-based sensors to purify human a-cells for functional and transcriptomic analyses

    Sevim Kahraman, Kimitaka Shibue ... Rohit N Kulkarni
    Tools and Resources

    Pancreatic a-cells secrete glucagon, an insulin counter-regulatory peptide hormone critical for the maintenance of glucose homeostasis. Investigation of the function of human a-cells remains a challenge due to the lack of cost-effective purification methods to isolate high-quality a-cells from islets. Here, we use the reaction-based probe diacetylated Zinpyr1 (DA-ZP1) to introduce a novel and simple method for enriching live a-cells from dissociated human islet cells with ~ 95% purity. The a-cells, confirmed by sorting and immunostaining for glucagon, were cultured up to 10 days to form a-pseudoislets. The a-pseudoislets could be maintained in culture without significant loss of viability, and responded to glucose challenge by secreting appropriate levels of glucagon. RNA-sequencing analyses (RNA-seq) revealed that expression levels of key a-cell identity genes were sustained in culture while some of the genes such as DLK1, GSN, SMIM24 were altered in a-pseudoislets in a time-dependent manner. In conclusion, we report a method to sort human primary a-cells with high purity that can be used for downstream analyses such as functional and transcriptional studies.

    1. Cell Biology
    2. Physics of Living Systems

    Force propagation between epithelial cells depends on active coupling and mechano-structural polarization

    Artur Ruppel, Dennis Wörthmüller ... Martial Balland
    Research Article Updated

    Cell-generated forces play a major role in coordinating the large-scale behavior of cell assemblies, in particular during development, wound healing, and cancer. Mechanical signals propagate faster than biochemical signals, but can have similar effects, especially in epithelial tissues with strong cell–cell adhesion. However, a quantitative description of the transmission chain from force generation in a sender cell, force propagation across cell–cell boundaries, and the concomitant response of receiver cells is missing. For a quantitative analysis of this important situation, here we propose a minimal model system of two epithelial cells on an H-pattern (‘cell doublet’). After optogenetically activating RhoA, a major regulator of cell contractility, in the sender cell, we measure the mechanical response of the receiver cell by traction force and monolayer stress microscopies. In general, we find that the receiver cells show an active response so that the cell doublet forms a coherent unit. However, force propagation and response of the receiver cell also strongly depend on the mechano-structural polarization in the cell assembly, which is controlled by cell–matrix adhesion to the adhesive micropattern. We find that the response of the receiver cell is stronger when the mechano-structural polarization axis is oriented perpendicular to the direction of force propagation, reminiscent of the Poisson effect in passive materials. We finally show that the same effects are at work in small tissues. Our work demonstrates that cellular organization and active mechanical response of a tissue are key to maintain signal strength and lead to the emergence of elasticity, which means that signals are not dissipated like in a viscous system, but can propagate over large distances.

    1. Biochemistry and Chemical Biology
    2. Cell Biology

    Inositol pyrophosphate dynamics reveals control of the yeast phosphate starvation program through 1,5-IP8 and the SPX domain of Pho81

    Valentin Chabert, Geun-Don Kim ... Andreas Mayer
    Research Article

    Eukaryotic cells control inorganic phosphate to balance its role as essential macronutrient with its negative bioenergetic impact on reactions liberating phosphate. Phosphate homeostasis depends on the conserved INPHORS signaling pathway that utilizes inositol pyrophosphates and SPX receptor domains. Since cells synthesize various inositol pyrophosphates and SPX domains bind them promiscuously, it is unclear whether a specific inositol pyrophosphate regulates SPX domains in vivo, or whether multiple inositol pyrophosphates act as a pool. In contrast to previous models, which postulated that phosphate starvation is signaled by increased production of the inositol pyrophosphate 1-IP7, we now show that the levels of all detectable inositol pyrophosphates of yeast, 1-IP7, 5-IP7, and 1,5-IP8, strongly decline upon phosphate starvation. Among these, specifically the decline of 1,5-IP8 triggers the transcriptional phosphate starvation response, the PHO pathway. 1,5-IP8 inactivates the cyclin-dependent kinase inhibitor Pho81 through its SPX domain. This stimulates the cyclin-dependent kinase Pho85-Pho80 to phosphorylate the transcription factor Pho4 and repress the PHO pathway. Combining our results with observations from other systems, we propose a unified model where 1,5-IP8 signals cytosolic phosphate abundance to SPX proteins in fungi, plants, and mammals. Its absence triggers starvation responses.