The Role of Ca²⁺ Influx for Insulin-Mediated Glucose Uptake in Skeletal Muscle

Adult male mice (NMRI or C57BL/6, weight ∼30 g) or rats (Wistar, weight ∼100 g) were housed at room temperature and fed ad libitum. In some experiments, male obese leptin-deficient ob/ob mice (45–55 g) were used. Animals were killed by rapid neck disarticulation and muscles removed. Isolated extensor digitorum longus (EDL; mainly fast-twitch fibers), flexor digitorum brevis (FDB; mainly fast-twitch fibers), and soleus (mainly slow-twitch fibers) muscles were used for measurements of 2-deoxyglucose (2-DG) uptake and protein phosphorylation. Intact single FDB fibers were used in experiments with fluorescent indicators measuring Ca²⁺ influx and 2-DG uptake. The single fibers were obtained either by microdissection (20) or by enzymatic digestion (21). All experiments on living cells were performed in solutions bubbled with 95% O₂/5% CO₂, with the exception of hypoxia experiments (see below). When used, insulin was applied at a supramaximal concentration of 60 or 120 nmol/l (10 or 20 mU/ml). All experiments were approved by the Stockholm North local ethical committee.

We used two highly sensitive methods to measure Ca²⁺ influx that have been frequently used to study nonselective cation channel activity: Mn²⁺ quenching (22) and Ba²⁺ influx (17). In Mn²⁺-quenching experiments, single FDB fibers were isolated by microdissection and mounted in a stimulation chamber continuously superfused by standard Tyrode solution (no glucose) (20). The fiber was microinjected with the fluorescent Ca²⁺ indicator fura-2 (Invitrogen) and then excited at 360 ± 5 nm (the isosbestic wavelength of fura-2 under our experimental conditions), while measuring the fluorescence light emitted at 495 ± 5 nm. MnCl₂ (100 μmol/l) was first added to the bath solution to obtain the rate of basal quenching. We then applied insulin and/or 2-APB. The rate of Mn²⁺ quenching was measured during periods where the rate of decrease of the fluorescent light was stable (i.e., ignoring transient changes induced by changes of the bath solution). Measurements were not performed until 2 min after application of insulin, thus allowing time for insulin to act. A brief tetanic contraction was produced before and after the period of exposures to ensure that the fiber was fully viable; fibers that showed a marked decrease (>10%) in tetanic force were discarded.

For Ba²⁺ influx experiments, FDB fibers were enzymatically isolated and cultured for ≥20 h (21). The fibers were loaded with Calcium Green-1 (10 μmol/l; Invitrogen) for 1 h at room temperature. For imaging, we used a BioRad MRC 1024 confocal unit with a krypton/argon mixed-gas laser run at 15 mW (BioRad Microscopy Division, Hertfordshire, U.K.) attached to a Nikon Diaphot 200 inverted microscope. A Nikon Plan Apo 20× objective lens was used. The cells were excited at 488 nm, and the emitted light was collected through a 522-nm narrow-band filter simultaneously from six to eight cells at 10-s intervals. To measure the baseline fluorescence, cells were first superfused with normal, glucose-free Tyrode solution for 5 min. Cells were then exposed to the Ba²⁺ solution, where 1 mmol/l Ca²⁺ was replaced by 1 mmol/l Ba²⁺, for 20 min. Data were normalized to the Calcium Green-1 fluorescence at the onset of Ba²⁺ exposure, which was set to 1.0. When used, insulin, OAG, and/or 2-APB were added to the Ba²⁺ solution. In a subset of experiments, cells were first exposed to insulin for 20 min, and OAG was then applied in the continued presence of insulin. Viability of the cells was tested before and after Ba²⁺ experiments with brief electrical stimulation. We only used data from cells that contracted and displayed normal [Ca²⁺]_i transients in response to stimulation.

2-(N-[7-nitrobenz-2-oxa-1,3-diazol-4-yl]amino)-2-deoxyglucose (2-NBDG; Invitrogen) has been used to assess glucose uptake in single vascular smooth muscle cells (23), and we adapted this method to single FDB fibers. Enzymatically digested FDB cells were cultured for ∼20 h and then imaged using a BioRad MRC 1024 confocal microscope (see above) with illumination at 488 nm, while measuring the emitted light at >515 nm. Fibers were superfused with a glucose-free (2 mmol/l pyruvate) Tyrode solution containing 50 μmol/l 2-NBDG for 15 min and then washed for 20 min. Fibers were then again exposed to 50 μmol/l 2-NBDG for 15 min followed by 20 min of washing. When used, insulin, 2-APB, and/or OAG were present during the last 15 min of the first washing period and during the second exposure to 2-NBDG. In some experiments, the [Ca²⁺] in the Tyrode solution was decreased to ∼1 nmol/l by decreasing the added CaCl₂ to 10 μmol/l, increasing MgCl₂ to 4.5 mmol/l, and adding 2 mmol/l EGTA. Up to six fibers were followed in each experiment, and only fibers that at the end of the experiment gave a robust contraction in response to electrical stimulation were included. Three confocal images were used to assess 2-NBDG uptake: immediately before the first exposure to 2-NBDG (a), after the first washing period (b), and after the second washing period (c). For each muscle fiber, the increase in fluorescence during the second 2-NBDG exposure period (ΔF_c-b) was divided by the increase in fluorescence during the first exposure (ΔF_b-a). In each set of experiments, the mean ΔF_c-b/ΔF_b-a of fibers in one dish, with insulin added on its own, was set to 100% and compared with the ΔF_c-b/ΔF_b-a of fibers not exposed to insulin (basal), exposed to insulin in low [Ca²⁺], or exposed to insulin plus 2-APB or OAG.

2-DG uptake in whole muscles was measured essentially as described elsewhere (24). Briefly, muscles were incubated at 35°C in a shaking water bath in Krebs bicarbonate buffer containing 2 mmol/l pyruvate (no glucose) with or without OAG or an inhibitor of Ca²⁺ influx, either 2-APB or cis-N-(2-phenylcyclopentyl)azacyclotridec-1-en-2-amine (MDL) (25). When used, insulin was applied 30 min before addition of 2-DG (1 mmol/l; 1 mCi per mmol/l) and inulin (1 μCi/ml medium). Muscles were blotted and frozen in liquid nitrogen 20 min after addition of 2-DG. Muscles were then weighed, digested in NaOH at 70°C, cooled, and centrifuged, and aliquots of the supernatants were added to scintillation cocktail and counted for ³H (2-DG) and ¹⁴C (inulin).

Hypoxia-mediated 2-DG uptake with or without 2-APB or OAG was studied by exposing muscles to 95% N₂/5% CO₂, instead of the normal 95% O₂/5% CO₂, for 80 min with 2-DG present during the last 20 min. The effect of 2-APB or OAG on contraction-mediated 2-DG uptake was studied in EDL muscles. The muscles were stimulated in standard Tyrode solution with two tetanic contractions (50 Hz, 100 ms) per second for 10 min, which results in a decrease in the glycogen store to <20% of the rested value (26). Muscles were transferred to the shaking water bath immediately after the cessation of tetanic contractions and incubated for 40 min with or without 100 μmol/l 2-APB or 30 μmol/l OAG. 2-DG uptake was measured during the last 20 min.

For protein kinase B analyses, whole EDL muscles were incubated at 37°C in standard Tyrode solution with or without insulin, 2-APB, or wortmannin and then frozen in liquid nitrogen. Frozen muscles were homogenized in lysis buffer consisting of (in mmol/l): 20 Hepes (pH 7.6), 150 NaCl, 5 EDTA, 25 KF, and 1 Na₃VO_4, as well as protease inhibitor cocktail (Roche), 20% glycerol, and 0.5% Triton X-100. Lysates were cleared by centrifugation for 5 min at 1,000g. The protein content was determined using the Bradford assay (Bio-Rad). Phosphorylated protein kinase B (PKB) and total PKB were measured using the PathScan Elisa Kit (Cell Signaling 7160 and 7170, respectively) following the supplier’s test procedures.

The phosphorylation state of the mitogen-activated protein kinase extracellular signal–related kinase (ERK) 1/2 was measured with Western blot and phospho-specific antibodies (New England Biolabs) as described elsewhere (27).

Data are presented as means ± SE. Statistical analyses of two groups were performed as paired or unpaired t tests, as appropriate. We used one-way ANOVA for analyses of more than two groups and one-way repeated-measures ANOVA for repeated measurements in one group. When ANOVA yielded a significant difference, either the Student-Newman-Keuls or Dunnett’s post hoc tests were used. P values <0.05 were considered statistically significant.

We used two methods frequently employed to detect Ca²⁺ fluxes over the cell membrane: Mn²⁺-quenching of fura-2 fluorescence (22) and Ba²⁺ influx measured with the visible-light Ca²⁺ indicator Calcium Green-1 (17). A major attraction with these methods is that they are very sensitive and can detect even small changes in the Ca²⁺ influx.

The basal rate of Mn²⁺ quenching was 0.70 ± 0.06%/min (n = 11). Figure 1A shows a representative Mn²⁺ quenching experiment with application of insulin followed by 2-APB. Insulin increased the rate of quenching in four of five cells tested (Fig. 1B). The subsequent application of 2-APB (100 μmol/l) resulted in a significant reduction of the rate of quenching (P < 0.01). Surprisingly, removal of 2-APB in the continued presence of insulin gave a rebound effect with a rate of quenching that was ∼30% higher than basal (P < 0.01).

Application of insulin caused a significant increase in the Ba²⁺ influx (P < 0.001) (Fig. 1C and D). Application of insulin and 2-APB caused a transient increase in the Ba²⁺ influx, but at the end of the 20-min incubation period, the relative fluorescence was significantly lower than with insulin alone (Fig. 1D).

Application of OAG (30 μmol/l) caused a significant increase (P < 0.001) in the Ba²⁺ influx with a magnitude similar to that obtained with insulin (Fig. 1C). However, the rate of increase was markedly lower with OAG than with insulin; the time to reach half the maximum signal was ∼12 and 6 min, respectively. Application of OAG plus 2-APB caused a transient increase in the relative fluorescence (as was the case with insulin plus 2-APB), but at the end of the incubation period, the OAG-induced increase in Ba²⁺ influx was abolished in the presence of 2-APB (Fig. 1D).

We also studied the combined effect of insulin plus OAG on Ba²⁺ influx (Fig. 1E). When OAG was added in the continued presence of insulin, there was an ∼40% additional increase in the signal as compared with the increase induced by insulin alone (P < 0.001). Noteworthy, the time to reach half of the maximum OAG-induced increase in the presence of insulin was ∼5 min, which is markedly faster than when OAG was added on its own (∼12 min).

Application of 2-APB caused a marked decrease of the rate Mn²⁺ quenching under basal conditions (i.e., no insulin present) (Fig. 2A), and mean data show that 2-APB reduced the rate of quenching to ∼30% of the control rate (P < 0.001) (Fig. 2B). Intriguingly, there was a marked increase in the rate of quenching when insulin was added in the presence of 2-APB. In fact, the effect of insulin on the rate of quenching was much larger in the presence than in the absence of 2-APB (Figs. 1 and 2). Nevertheless, in the presence of insulin plus 2-APB, the rate of quenching was decreased to 60–70% of the control rate with either insulin or 2-APB added first.

The effect of 2-APB on basal Ca²⁺ influx was also tested with the Ba²⁺ influx method. Measurements performed during the normal 20-min measuring period without adding Ba²⁺ to the solution showed a significantly (P < 0.05) lower fluorescent signal (∼0.9 of the control) than when Ba²⁺ was added (∼1.0 of the control) (Fig. 2C); this decrease in the Calcium Green-1 signal without Ba²⁺ might represent leakage of the dye out of the cell, or some bleaching of the dye, due to repeated exposures to the laser light, which would be counteracted by influx of Ba²⁺ when the ion was present. When Ba²⁺ was added together with 2-APB, the fluorescent signal had decreased down to the level observed without Ba²⁺ at the end of the 20-min exposure period (Fig. 2C). Thus, both the Mn²⁺-quenching and the Ba²⁺-influx experiments indicate that there is a basal Ca²⁺ influx that can be inhibited by 2-APB.

Previous studies on isolated whole muscles have not shown any effect of decreasing bath [Ca²⁺] on insulin-mediated glucose uptake (9,10). In accordance, we did not detect any effect on insulin-mediated 2-DG uptake in EDL muscles exposed to a Tyrode solution with 4 mmol/l EGTA and no Ca²⁺ added (113.5 ± 4.8 vs. 103.5 ± 6.3 μmol/l per min in control; n = 4). However, in a normally polarized skeletal muscle cell, the extracellular [Ca²⁺] concentration has to be <1 nmol/l before the inward driving force for Ca²⁺ is abolished, and in isolated FDB fibers, the bath [Ca²⁺] has to be decreased to 1 nmol/l to significantly reduce the insulin-mediated increase in [Ca²⁺]_mem (14). Thus, it is doubtful whether the extracellular [Ca²⁺] in isolated whole muscles can be decreased to levels low enough to affect insulin-induced Ca²⁺ uptake and, hence, glucose transport. We therefore used fluorescent 2-DG (2-NBDG) to assess glucose uptake in single FDB fibers. Under control conditions, ΔF_c-b/ΔF_b-a after insulin (60 nmol/l) exposure (0.96 ± 0.08; n = 27) was about twice as large as that without insulin (0.50 ± 0.05; n = 21; P < 0.001), which demonstrates that the isolated cells respond to insulin with increased glucose uptake. The insulin-mediated increase in 2-NBDG fluorescence was markedly reduced when cells were exposed to low [Ca²⁺] (∼1 nmol/l) (Fig. 3A and B), and mean data show that ∼40% of the normal insulin response was lost in low [Ca²⁺] (Fig. 3C).

The application of 2-APB (100 μmol/l) significantly decreased the insulin-induced increase in 2-NBDG uptake by ∼60% (Fig. 3C). Insulin plus OAG increased the Ca²⁺ influx above that observed by insulin alone (Fig. 1E), which suggests that application of both agents would result in an increased glucose uptake. Accordingly, exposure to insulin plus OAG (30 μmol/l) gave a 2-NBDG fluorescence that was significantly larger than that of insulin alone (Fig. 3C). We also measured the 2-DG uptake in whole EDL muscles exposed to insulin plus OAG and found a significantly larger uptake (98.7 ± 5.7 μmol/l per min) compared with insulin alone (79.0 ± 3.0 μmol/l per min) (P < 0.01; n = 8).

The uptake of 2-NBDG was also tested in single FDB fibers of obese ob/ob mice. These mice are insulin resistant with an insulin-mediated 2-DG uptake in EDL muscles that is ∼30% lower than in wild type (28). Accordingly, the insulin response in single ob/ob FDB fibers was small (Fig. 3D), and while insulin caused an approximate twofold increase in 2-NBDG uptake in wild-type cells, the increase was only ∼1.2-fold in ob/ob fibers. Exposure to insulin plus OAG resulted in a significantly increased 2-NBDG uptake compared with insulin alone in ob/ob fibers as well.

Application of 2-APB resulted in a dose-dependent decrease in insulin-mediated 2-DG uptake in mouse fast-twitch EDL and slow-twitch soleus muscles (Fig. 4A and B). In whole FDB muscles, application of 2-APB (100 μmol/l) approximately halved the 2-DG uptake in the presence of insulin (n = 4, P < 0.01). The inhibitory effect of 2-APB on insulin-mediated 2-DG uptake was not restricted to mouse muscle, since similar results were obtained in rat EDL and soleus muscles (Fig. 4C). Thus, 100 μmol/l 2-APB decreased insulin-mediated 2-DG uptake by ∼50% in fast- and slow-twitch muscles of mouse and rat. We also tested the effect of MDL, another inhibitor of nonselective cation channels that is structurally different to 2-APB (25). Application of MDL (100 μmol/l) resulted in a marked decrease in insulin-mediated 2-DG uptake in both EDL and soleus muscles (P < 0.001) (Fig. 4D).

Measurements of insulin-mediated 2-DG uptake were also performed in EDL muscles that had been exposed to 2-APB for 30 min and the drug then removed in the continued presence of insulin (i.e., a situation where Ca²⁺ influx was markedly increased) (Fig. 1B). The results showed a significant (∼30%; P < 0.01) increase in 2-DG uptake after 2-APB removal compared with insulin alone (Fig. 4E). To determine whether the insulin effect could also be potentiated in insulin-resistant muscles, we repeated the same experiment in EDL muscles from ob/ob mice. Indeed, the 2-APB rebound effect was also present in the insulin-resistant ob/ob muscles (∼15% increase; P < 0.05) (Fig. 4E).

Basal 2-DG uptake in EDL and soleus muscles was not significantly affected by 2-APB, MDL, or OAG (Fig. 5A). Hypoxia or repeated contractions are known to increase glucose uptake in skeletal muscle by pathways that differ from that of insulin (29). Hypoxia (induced by exposing muscles to N₂) resulted in about a twofold increase in 2-DG uptake that was not affected by 2-APB or OAG (Fig. 5B). Repeated tetanic contractions for 10 min increased 2-DG uptake about threefold in EDL muscles, and this increase was not affected by 2-APB or OAG (Fig. 5C). We added 2-APB or OAG after the series of contractions because skeletal muscles have been shown to take up Ca²⁺ during repeated contractions (30), and we wanted to avoid potential problems with differences in sarcolemmal Ca²⁺ fluxes during the contraction period. It might then be argued that 2-APB and OAG can only exert their effects if they are present during the induction of increased glucose uptake. However, control experiments, in which EDL and soleus muscles were exposed to insulin for 30 min before adding 2-APB (100 μmol/l) and measuring 2-DG uptake, showed ∼35% lower uptake in the presence than in the absence of 2-APB (P < 0.001). Thus, while 2-APB and OAG significantly affected insulin-mediated Ca²⁺ fluxes and glucose uptake, they had no significant effect on other modes of glucose uptake.

Tetanic force and [Ca²⁺]_i were measured in isolated FDB fibers before and immediately after or during exposure to 2-APB. Records of tetanic [Ca²⁺]_i from a representative experiment are shown in Fig. 5D. Mean data showed that tetanic force was 100 ± 5% and [Ca²⁺]_i 99 ± 10% of that before 2-APB exposure (n = 6). Tetanic force was also measured in whole muscles exposed to 2-APB (100 μmol/l). After 30 min exposure, tetanic force was increased by 14 ± 4% in soleus (n = 5, P < 0.05) and 7 ± 5% in EDL muscles (n = 4, P > 0.05) and the relaxation speed slightly decreased. Thus, 2-APB had no adverse effect on action potential–mediated sarcoplasmic reticulum Ca²⁺ release and force production.

Insulin increased PKB (Akt) phosphorylation, and this increase was not affected by 2-APB (Fig. 6A). The insulin-induced phosphorylation of PKB was absent in the presence of wortmannin (0.5 μmol/l), which inhibits phosphoinositide 3-kinase; this phosphorylation is an early step in the insulin signaling cascade (4). 2-APB and wortmannin, on their own, had no effect on PKB phosphorylation, nor did any of the treatments significantly affect the total PKB expression (data not shown). Insulin significantly (P < 0.05) increased ERK1/2 phosphorylation in both the absence and presence of 2-APB (Fig. 6B). These findings demonstrate that 2-APB interferes neither with early signaling steps in the pathway terminating with glucose uptake nor with the branch of the insulin signaling cascade that results in mitogen-activated protein kinase activation (31).

The present study shows that changes in Ca²⁺ influx have significant effects on insulin-mediated glucose uptake in skeletal muscle. By manipulating the Ca²⁺ influx, we could both decrease (in the presence of 2-APB, MDL, or low extracellular [Ca²⁺]) and increase (after removal of 2-APB or in the presence of OAG) the insulin response. On the other hand, a solitary change in the Ca²⁺ influx is not enough to alter glucose transport because decreased (with 2-APB) or increased (with OAG) Ca²⁺ influx had no effect on 2-DG uptake in the absence of insulin. This means that insulin activates additional factors that are required for stimulation of glucose uptake. The insulin-induced activation of these additional factors is probably mediated via the PKB/Akt and/or CAP-cbl-TC10 pathways (2,3), because 2-APB and OAG had no effect on hypoxia- or contraction-mediated glucose uptake, which are mediated via signaling pathways different from those of insulin (29).

The present results show a complex interaction between basal, insulin- and OAG-mediated Ca²⁺ influx (Figs. 1 and 2), which indicates the involvement of several different ion channels or a complex regulation of one set of channels. Earlier studies suggested the involvement of voltage-activated l-type Ca²⁺ channels in insulin action because pharmacological inhibition of these channels reduced insulin-mediated increases in glucose uptake in rat muscle (32) and [Ca²⁺]_mem in mouse FDB fibers (14). However, l-type Ca²⁺ channels are normally activated upon depolarization, and insulin causes a hyperpolarization of the sarcolemma (14,33). Also, there was no simple relationship between Ca²⁺ uptake and glucose transport in the presence of Ca²⁺ channel blockers (32). Thus, the present results show an important modulating role of Ca²⁺ influx in insulin-mediated glucose uptake, but the identity of the ion channels involved remains uncertain.

The changes in Ca²⁺ influx induced by insulin, 2-APB, and OAG were rather small and would have limited effect on [Ca²⁺]_i (14,15), but still, they had a major impact on insulin-mediated glucose uptake. This implicates a model with strict spatial control of these Ca²⁺ fluxes with participating Ca²⁺ channels being located in close vicinity to other components required for glucose transport. The situation would resemble that in neurons where the diverse response to activation of different Ca²⁺ channels seems to be caused by physically associated signaling molecules that sense local Ca²⁺ increases at the mouth of the channels (34). The Ca²⁺-binding protein calmodulin has a central role in this process (35). Interestingly, pharmacological inhibition of calmodulin/calmodulin kinase decreases insulin-mediated glucose uptake (24,36).

Ca²⁺ influx appears to act on later steps in the insulin signaling pathway, since 2-APB did not affect PKB or ERK1/2 phosphorylation (Fig. 6). Several steps in the docking and fusion of GLUT4 transporters into the plasma membrane are potentially Ca²⁺/calmodulin sensitive (37); in other words, this process may be similar to the Ca²⁺-dependent docking of synaptic vesicles in nerve terminals (38). Furthermore, results from an in vitro assay for GLUT-4 translocation show that the association of GLUT-4 with the plasma membrane is Ca²⁺ dependent (39).

In conclusion, Ca²⁺ influx is an important modulator of insulin-mediated glucose uptake. The present results indicate that therapeutic interventions that increase Ca²⁺ influx in association with insulin exposure may be used to improve glucose uptake in insulin-resistant conditions. As a proof of principle, we show that increased Ca²⁺ influx is accompanied by increased insulin-mediated glucose uptake in insulin-resistant muscles.

This study was supported by the Swedish Research Council, the Biovitrum Partner Fund, the Swedish Diabetes Foundation, the Swedish National Center for Sports Research, and funds at the Karolinska Institutet.