High-Intensity Exercise Decreases IP6K1 Muscle Content and Improves Insulin Sensitivity (S_I²*) in Glucose-Intolerant Individuals

Abstract

Context

Insulin resistance (IR) in skeletal muscle contributes to whole body hyperglycemia and the secondary complications associated with type 2 diabetes. Inositol hexakisphosphate kinase-1 (IP6K1) may inhibit insulin-stimulated glucose transport in this tissue type.

Objective

Muscle and plasma IP6K1 were correlated with two-compartment models of glucose control in insulin-resistant hyperinsulinemic individuals. Muscle IP6K1 was also compared after two different exercise trials.

Design

Nine prediabetic [hemoglobin A1c; 6.1% (0.2%)] patients were recruited to take part in a resting control, a continuous exercise (90% of lactate threshold), and a high-intensity exercise trial (6 30-second sprints). Muscle biopsies were drawn before and after each 60-minute trial. A labeled ([6,6²H₂]glucose) intravenous glucose tolerance test was performed immediately after the second muscle sample.

Results

Fasting muscle IP6K1 content did not correlate with insulin sensitivity (S_I²*) (P = 0.961). High-intensity exercise reduced IP6K1 muscle protein and messenger RNA expression (P = 0.001). There was no effect on protein IP6K1 content after continuous exercise. Akt³⁰⁸ phosphorylation of was significantly greater after high-intensity exercise. Intermittent exercise reduced hepatic glucose production after the same trial. The same intervention also increased S_I²*, and this effect was significantly greater compared with the effect of continuous exercise improvements. Our in vitro experiment demonstrated that the chemical inhibition of IP6K1 increased insulin signaling in C2C12 myotubes.

Conclusions

The in vivo and in vitro approaches used in the current study suggest that a decrease in muscle IP6K1 may be linked to whole body increases in S_I²*. In addition, high-intensity exercise reduces hepatic glucose production in insulin-resistant individuals.

Type 2 diabetes (T2D) is a multifactorial metabolic disease characterized by defects in insulin sensitivity (S_I²*), glucose effectiveness (S_G²*), β-cell function, and endogenous glucose production (1). Although the evidence is not conclusive, insulin resistance (IR) seems to result from a decrease in the ability of insulin receptor substrate (IRS) to activate downstream insulin-signaling kinases (2). A reduction in the serine/threonine protein kinase Akt phosphorylation is a known characteristic of IR and T2D (3) and is an important protein in the insulin-signaling cascade.

Insulin-stimulated glucose uptake involves insulin receptor autophosphorylation, tyrosine phosphorylation of IRS, and the subsequent activation of phosphatidylinositol 3-kinase. The downstream target of phosphatidylinositol 3-kinase, Akt, is then activated via the phosphorylation of Thr³⁰⁸ and Ser⁴⁷³ by phosphatidylinositol 3,4,5-triphosphate-dependent protein kinase 1 (PDK1) and PDK2, respectively (4). Akt contains an N-terminal pleckstrin homology (PH) domain, allowing the binding of phosphatidylinositol-3,4,5-triphosphate (PIP3) and the subsequent membrane translocation and activation of Akt (5). Upon activation, Akt is responsible for the phosphorylation of the 160-kDa Akt substrate (AS160), GLUT4 translocation, glucose uptake (6), phosphorylation and inhibition of GSK3β, and glycogen synthesis (7, 8), making Akt a potential target in the treatment of T2D.

Diphosphoinositol polyphosphates, also known as inositol pyrophosphates, are a family of water-soluble inositol phosphates (9). The inositol hexakisphosphate (IP6) kinase 1 (IP6K1) produces a pyrophosphate group at the fifth position of IP6 to generate an inositol pyrophosphate, diphosphoinositol pentakisphosphate (IP7) (10). Production of IP7 results in its binding to the PH domain of Akt/protein kinase B, preventing its translocation to the cell membrane and reducing its subsequent phosphorylation by PDK1. The evidence for this effect comes from the finding that IP7 does not prevent PDK1 phosphorylation of Akt^Thr308 lacking a PH domain (11). The consequence is a potential reduction in insulin (Akt)-stimulated glucose uptake in muscle and adipose tissue (12). Chakraborty et al. (11) showed that IP6K1 knockout (KO) mice demonstrate augmented Akt activity and increased glucose transport rates in skeletal muscle. Key to the current work, recent in vivo data suggest a role of IP6K1 in IR, with IP6K1 KO mice displaying normal glycemic control despite low circulating plasma insulin (13). In addition, the pharmacological inhibition of IP6K1 increases Akt signaling in mouse embryonic fibroblasts while suppressing IP7 synthesis (14). The increased availability of IP6K1 and its product, IP7, is thought to be stimulated by insulin. The competition of IP7 with PIP3 for binding at the PH domain of Akt may represent a negative feedback mechanism whereby hyperinsulinemia eventually decreases Akt activity, indirectly shown through the presence of decreased insulin action despite augmented insulin secretion in prediabetic states (15) and an increase in IP6K1 activity and reduced p-Akt in rodents treated with insulin (11).

However, research suggesting a role of IP6K1 inhibition as a future target of IR has been limited to in vitro and animal work. The current study aimed to address this limitation by measuring IP6K1 muscle content in hyperinsulinemic insulin-resistant humans.

Insulin-stimulated glucose uptake increases 2 to 72 hours after exercise (16–18), with the amount of muscle mass an important determinant to this response, so exercise involving a larger muscle mass is preferable. Higher-intensity exercise recruits a larger proportion of muscle and a greater number of type 2 glycolytic muscle fibers compared with moderate-intensity activity (19), which may offer a greater sink for glucose disposal. Recently a form of high-intensity interval training has been shown to improve S_I²* (20–22). However, data on the other metabolic defects associated with T2D, including hepatic glucose production, β-cell function, and glucose effectiveness, remain sparse. Here we assessed the effects of high-intensity exercise on two-compartment models of insulin sensitivity, glucose effectiveness, and hepatic glucose production.

Therefore, the first aim of this study was to characterize this insulin signaling pathway (Akt-IP6K1) in prediabetic humans. Second, using a stimulus known to improve insulin sensitivity (muscle contraction), we aimed to examine whether IP6K1 could be manipulated after muscle contraction in glucose-intolerant individuals and to evaluate the effects of different types of exercise stimuli on IP6K1 muscle content. The C2C12 skeletal muscle cell line was also used to investigate the role of N2-(m-Trifluorobenzyl), N6-(p-nitrobenzyl)purine (TNP) treatment on Akt-AS160 signaling.

Participants and Methods

In vivo study

Nine sedentary glucose-intolerant individuals (7 male and 2 female) were recruited for this investigation. Subjects’ clinical characteristics were age 47 (±3) years, body mass index 32.0 (±2.4) kg/m², body fat percentage 39.0% (±4.4%); hemoglobin A1c (HbA1c) 6.1% (±0.2%), and homeostasis model of insulin resistance (HOMA_IR) 3.3 (±0.8). Each participant was informed of the study purpose, experimental procedures, and all potential risks before providing written consent to participate. Ethical approval was granted by the local university ethics committee (11, 12, 23) and conformed to Declaration of Helsinki for the use of human participants in research. Exclusion criteria included diabetes-related complications (i.e., neuropathy, peripheral vascular and cardiovascular disease), current smoking, or treatment with insulin or any other pharmaceutical intervention. HbA1c values of >5.7% and <6.4% were used to define individuals in a prediabetic state (23).

Experimental protocol

Participants were required to attend our laboratory on four occasions, each separated by 7 to 14 days. During a preliminary visit percentage of body fat was estimated with a Bodpod (Life Measurement, Inc., Concord, CA) as previously described (24). Venous blood samples were drawn for the determination of HbA1c (Axis-Shields, Dundee, UK). Fasting blood glucose and plasma insulin concentrations were measured for the determination of HOMA_IR [fasting insulin (µU/mL) × fasting glucose (mmol/L)/22.5] and homeostasis model of β-cell function [HOMA_β-Cell; 20 × fasting insulin (µU/mL)/fasting glucose – 3.5 (mmol/L)] (25). During this preliminary visit, individual lactate threshold values were obtained as previously described (26) via a cycle ergometer (Lode Corival; Groningen, the Netherlands).

On experimental days, volunteers reported to the laboratory at ∼08:30, having fasted for 12 hours, abstained from caffeine and alcohol for 24 hours, and abstained from exhaustive exercise for 72 hours. An 18-gauge cannula was positioned into a dorsal hand vein to allow frequent sampling of arterialized blood, with a thermoregulated hot box (∼60°C) (27). A second 18-gauge cannula was placed into a prominent contralateral antecubital vein for administration of labeled glucose. In a randomized fashion, subjects completed a resting control trial (rest) of 60 minutes of passive sitting, continuous exercise at an intensity equal to 90% lactate threshold for 60 minutes (continuous) (cycle ergometer, Lode Corival), and high-intensity intermittent exercise (6 × 30-second Wingates) (intermittent) (Monark 894 E Weight Ergometer; Monarch, Vansbro, Sweden). The 30-second sprints were interspersed with 9.5 minutes of passive recovery in the intermittent trial. Each trial lasted 60 minutes. Muscle biopsies were drawn under local anesthesia from the vastus lateralis with the conchotome method (28) at baseline (0 minutes) and immediately after (60 minutes) trials, with volunteers in a supine position. Immediately after the posttreatment muscle biopsy, a 4-hour labeled intravenous glucose tolerance test was administered (28.4 mg/kg [6,6²H₂]glucose and ∼250 mg/kg unlabeled glucose), prepared under sterile conditions. Thereafter, frequent arterialized (∼5 mL) blood samples were drawn over the ensuing 240 minutes, as previously described (26). The concentration of circulating glucose was measured in whole blood with a YSI 2300 (STAT; Yellow Springs, OH), and spun separated (4°C, 10 minutes, 5000 rpm) plasma was frozen and later analyzed for plasma insulin, endogenous glucose, and isotope-enriched [6,6²H₂]glucose concentrations.

Blood analysis

Glucose-enriched plasma samples were deproteinized in ethanol (99%), with the resulting supernatants centrifuged to dryness. Hydroxylamine hydrochloride (100 μL, 0.18 M pyridine) was then added before a 60-minute incubation at 70°C, after which Bis(trimethyl)trifluoroacetamide 1% trimethyl-chlorosilane (99%) (Sigma-Aldrich, Exeter, UK) was added before a further incubation (45 minutes at 70°C). Samples were then analyzed for glucose derivatives of 319 (unlabeled glucose; trace) and 321 ([6,6²H₂]glucose; tracer) by gas chromatography–mass spectrometry. Endogenous glucose concentration was measured in whole blood (YSI 2300; STAT) and plasma insulin via a commercially available enzyme-linked immunosorbent assay (DRG Diagnostics, UK).

Plasma insulin, endogenous glucose concentrations, and [6,6²H₂]glucose-enriched values were used to model the metabolic indices: insulin sensitivity (S_I²*), glucose effectiveness (S_G²*), and hepatic glucose production (HGP) as described previously (29, 30) (SAAMII Institute, Seattle, WA) S_I²* explains the effects of insulin on glucose disposal rates, and S_G²* quantifies the effects of glucose to cause its own transport via mass action at basal insulin concentrations.

Muscle analysis

Immediately after collection, muscle samples were washed in ice-cold saline, with visible fat removed before being frozen in liquid nitrogen and transferred to −80°C until analysis. Muscle tissue homogenates were used for Western blot protein analysis. Protein content of the homogenates was quantified via Lowry’s method (Bio-Rad DC protein assay) with 20 μg of total protein separated with 7.5% precast polyacrylamide gels before being transferred via semidry method to nitrocellulose membranes (Bio-Rad).

Immunoblotting

Membranes were blocked in 5% bovine serum albumin (BSA) (1 hour), and polyclonal antibodies pAkt308, pAkt473, total Akt, AS160, pAS160 on Thr642 (Cell Signaling) and IP6K1 (Abcam) were incubated overnight at 1:1000 in 5% BSA at 4°C. Membranes were then washed and incubated with anti-rabbit secondary antibody (Cell Signaling; 1:10,000) in 0.5% to 5% BSA. Membranes were quantified with an Odyssey^® Fc Imaging System (LI-COR). Blots were normalized to total protein (31) because this method shows greater sensitivity than “housekeeping” proteins.

Real-time quantitative polymerase chain reaction

Total RNA from muscle biopsy samples (20 mg) was extracted with the RNeasy Plus Mini Kit (Qiagen, Hilden, Germany) according to the standard manufacturer’s protocol. Cell lysates were homogenized with Qiashredder spin columns (Qiagen). The concentration and purity of extracted RNA were measured at 260 nm by spectrophotometry with a NanoDrop 1000 Spectrophotometer (Thermo Scientific, Wilmington, DE). Extracted RNA samples were stored at −80°C. For relative quantification of messenger RNA (mRNA), total RNA was reverse transcribed to complementary DNA (cDNA) with the QuantiTect Reverse Transcription Kit (Qiagen) according to the manufacturer’s instructions. In brief, ≤1 µg RNA was reverse transcribed to cDNA in a final volume of 20 µL using oligo (dT)₁₅ primers (0.5 µg per reaction). Each quantitative real-time polymerase chain reaction (PCR) mixture (20 µL) contained 1 µL of RT product (cDNA transcribed from 1 µg total RNA), and PCR was performed with Rotor-Gene SYBR Green PCR Kit (Qiagen) according to the manufacturer’s instructions. The mixture was initially incubated at 95°C for 5 minutes, followed by 40 cycles of 95°C for 15 seconds, 52°C for 15 seconds, and 72°C for 30 seconds. PCRs were carried out on a Rotor-Gene Q (Qiagen) in triplicate. Samples were normalized relative to the mRNA level of glyceraldehyde 3-phosphate dehydrogenase. For each subject, all samples were simultaneously analyzed in one assay run. Measurements of the relative distribution of each target gene were performed for each subject; a cycle threshold (C_T) value was obtained by subtracting glyceraldehyde 3-phosphate dehydrogenase C_T values from respective target C_T values. The expression of each target was then evaluated by the Rotor-Gene Q Software version 2.3 (Qiagen).

In vitro experiment

Myoblasts from the muscle-derived mouse C2C12 cell line (ATCC #CRL-1772) were grown in growth media of Dulbecco’s modified Eagle medium (DMEM; Gibco #22A320), supplemented with 10% fetal bovine serum, penicillin (50 U/mL), and streptomycin (50 U/mL) in a standard manner (37°C, 5% CO₂, 100% humidity) until ∼80% confluent. Cells were trypsinized and seeded for experimental conditions in standard 6-well plate. To induce differentiation into myotubes, confluent cells were washed in Dulbecco’s phosphate-buffered saline and incubated in differentiation media of DMEM with 2% equine serum, penicillin (50 U/mL), and streptomycin (50 U/mL) for 96 hours, with differentiation media changed every 24 hours before experimental conditions were applied.

After 96 hours for myotube formation, cultures were incubated under control, hyperglycemic, and hyperinsulinemic conditions with or without TNP, a pan-IP6 kinase inhibitor. For hyperinsulinemic treatment, myotubes were incubated in 100 nM insulin (Sigma) for 24 hours in serum-free DMEM containing 5 mM glucose (32), with control cells incubated in DMEM containing 5 mM glucose for the same time period. For hyperglycemia, C2C12 cells were treated in serum-free DMEM containing 30 mM d-glucose for 24 hours (33). Each treatment was performed with or without TNP at 10 μM (34). At the 24-hour point, total protein was extracted. Cells were aspirated and washed on ice with ice-cold phosphate-buffered saline before 400 µL lysis buffer with protease inhibitor (1:100) (Cell Signaling) was added. After 20 minutes of incubation, cells were scraped into 1.5-mL Eppendorf tubes and spun (6 minutes, 6000 rpm), and supernatant was removed and aliquoted for protein quantification (Lowry and Bio-Rad DC protein assay) and later analysis of proteins of interest by Western blot (described earlier).

Statistics

The areas under the curve for both glucose and insulin were calculated with the trapezoidal rule. Differences over time and between conditions were evaluated by two-way repeated-measures analysis of variance. Tukey post hoc tests were used when statistical significance was found. Linear regression analyses were carried out to test for significance where appropriate. All statistical tests were carried out in SPSS (version 15; IBM). Data are expressed as mean (standard error). Statistical significance was set at P < 0.05.

Results

Correlation analysis

IP6K1 has been shown in cell culture and animal models to be implicated in reduced glucose control. One of the key aims of this research was to assess whether muscle and plasma IP6K1 correlated with whole body measures of glucose control. Figure 1A and 1B show correlation analysis between plasma IP6K1 concentration and two-compartment measures of peripheral glucose control. These data have been combined with those of previously published work to include a range of insulin-resistant individuals including patients with T2D (26). Neither S_I²* (r = 0.402; P = 0.055) nor S_G²* (r = 0.151; P = 0.281) showed a significant relationship with plasma IP6K1. Baseline measures HbA1c (r = 0.357; P = 0.080), fasting blood glucose (r = 0.232; P = 0.185), fasting insulin (r = 0.365; P = 0.075), body fat percentage (r = 0.028; P = 0.457), and HOMA_β-Cell (r = 0.006; P = 0.491) were also correlated with plasma IP6K1, with only HOMA_IR demonstrating a significant relationship (r = 0.429; P = 0.043) with this measure. A full set of correlation data is displayed in Table 1. No relationship was noted between muscle IP6K1 protein content and S_I²* (r = 0.019; P = 0.961). This comparison was for the current data set (Fig. 1C) as muscle tissue was not collected in our earlier work (26).

Figure 1.

Correlation analysis between plasma IP6K1, S_G²* (r = 0.151; P = 0.281), S_I²* (r = 0.402; P = 0.055), and HOMA_IR (r = 0.429; P = 0.043). (A–C) Additional data from a previously published article (Mackenzie et al., 2011) and the current data set (n = 17) are included. Correlation analysis includes both patients with prediabetes and patients with T2D. (D) Correlation analysis for muscle IP6K1 protein content (n = 9) with S_I²* from the current data only.

Table 1.

Correlation Analysis Between Plasma IP6K1 and Measures of Glycemic Control

	Patients With Prediabetes	Patients With T2D	Combined
SI2*	r = 0.033 (0.932)	r = 0.733 (0.025)*	r = 0.402 (0.055)
SG2*	r = 0.247 (0.521)	r = 0.472 (0.200)	r = 0.151 (0.281)
HbA1c	r = 0.194 (0.595)	r = 0.310 (0.493)	r = 0.357 (0.080)
Fasting blood glucose	r = 0.119 (0.760)	r = 0.043 (0.913)	r = 0.232 (0.185)
Plasma fasting insulin	r = 0.179 (0.645)	r = 0.340 (0.370)	r = 0.365 (0.075)
Body fat %	r = 0.086 (0.825)	r = 0.061 (0.870)	r = 0.028 (0.457)
HOMAβ-Cell	r = 0.171 (0.661)	r = 0.261 (0.498)	r = 0.006 (0.491)
HOMAIR	r = 0.194 (0.617)	r = 0.311 (0.415)	r = 0.429 (0.043)*

Values are means (standard error of the mean).

Exercise intervention

Immediately after the labeled intravenous glucose load, the area under the curve for glucose was significantly lower after intermittent (P = 0.008) and continuous (P = 0.016) exercise treatments when compared with the resting control trial. No difference was noted for the area under the curve for glucose between exercise treatments (P = 0.084) [Fig. 2(D)]. Despite a trend for being lower after treatment, neither exercise condition affected the area under the curve for insulin [Fig. 2(E); P = 0.421]. Endogenous glucose, labeled glucose, and insulin concentrations were modeled to determine two-compartment measures of S_I²*, S_G²*, and HGP. Both exercise conditions demonstrated a significant increase in S_I²* over the control trial (P < 0.001). S_I²* was also significantly higher after high-intensity intermittent exercise when compared with traditional moderate-intensity exercise [Fig. 2(A); P < 0.01]. Despite being higher after both exercise conditions, S_G²* was not found to be statistically different from resting control [Fig. 2(B); P = 0.561].

Figure 2.

(A) S_I²*, (B) S_G²*, and (C) HGP in response to a control, continuous, and intermittent exercise trials. The integrated area under the curve for (D) arterialized blood glucose and (E) plasma insulin after labeled intravenous glucose loads 4 hours after trials. *Denotes significant difference between resting control (P < 0.01). †Denotes difference between continuous and intermittent exercise (P < 0.001).

Skeletal muscle signaling

Comparisons between treatments in human skeletal muscle samples were made as fold changes from fasting control. IP6K1 was significantly lower immediately after intermittent exercise compared with fasted samples (P = 0.001), with no difference noted for the same comparisons for continuous exercise (P = 0.337; Fig. 3A). Phosphorylation of Akt at serine 308 was elevated for intermittent exercise (P = 0.003) above fasted values, with no difference for the same comparison with the continuous treatment (P = 0.175; Fig. 3B). There was no difference between treatments for pAkt⁴⁷³ (P = 0.200; Fig. 3C). The downstream target for Akt, AS160, was significantly increased after both intermittent (P = 0.012) and continuous exercise (P = 0.041; Fig. 3D). Intramuscular IP6K1 mRNA expression decreased significantly after both exercise treatments (P < 0.01) with continuous exercise lower than the intermittent protocol (P < 0.05) (Fig. 4A). Akt and GLUT4 expression was significantly higher after intermittent and continuous exercise when compared with fasting samples (P < 0.01; Fig. 4B and 4C). PDK1 mRNA expression was significantly greater for the intermittent exercise treatment only (P < 0.01).

Figure 3.

(A) Muscle protein content of IP6K1, (B) phosphorylation of Akt at Thr³⁰⁸, (C) Akt at Ser⁴⁷³, (D) AS160 at Thr642, and (E) representative blots (n = 9). *P < 0.05 vs. fasting (before exercise).

Figure 4.

Intramuscular mRNA expression of (A) IP6K1, (B) Akt, (C) GLUT 4, and (D) PDK1. Values are expressed as fold change from fasting before exercise and after each trial as mean (standard error) (n = 9). *Different from fast, P < 0.01. †Different from continuous exercise, P < 0.05. ‡Different from rest, P < 0.05.

Insulin signaling in C2C12 cells

C2C12 skeletal muscle cells were treated with both insulin and glucose with or without TNP (10 μM) to assess the effects of the stated treatment on IP6K1. Insulin treatment increased IP6K1 protein content over the control treatment (P = 0.010). Insulin plus TNP treatment was not different from control (P = 0.647) but was significantly lower than in the insulin condition (P = 0.008). Glucose treatment increased IP6K1 (P = 0.007), which was lowered with the addition of TNP (Glu + TNP; P = 0.008). Phosphorylation of Akt at 308 was significantly lower with insulin when compared with the control condition (P = 0.041) and elevated in the insulin plus TNP over the insulin-only treatment (P = 0.030). Twenty-four hours of insulin treatment with and without TNP increased phosphorylated/total protein Akt⁴⁷³ in the C2C12 skeletal muscle cells (P < 0.05). In addition, phosphorylated/total protein Akt⁴⁷³ was significantly higher in the insulin-only trial when compared with insulin plus TNP (P < 0.05). AS160, one of the last proximal steps in glucose transport in skeletal muscle, was elevated in both the insulin and insulin plus TNP treatments (P < 0.05). The same target was significantly reduced with the addition of glucose plus TNP to the treatment media (Fig. 5D; P = 0.039).

Figure 5.

(A) Muscle protein content of IP6K1, (B) phosphorylation of Akt at Thr³⁰⁸, (C) Akt at Ser⁴⁷³, (D) AS160 at Thr642, and (E) representative blots for C2C12 treatments. *P < 0.05 vs. control. ^αP < 0.05 vs. Ins+TNP and †P < 0.05 vs. Ins. Data are mean ± standard error (n = 4). (F) PathScan^® Akt Signaling Antibody Fluorescent read (700). *P < 0.05 vs. control. †P < 0.05 vs. IGF. ^αP < 0.05 vs. Ins. ^γP < 0.05 vs. Ins+TNP. Data are mean ± standard error (n = 4).

Discussion

Previous work has shown that IP6K1 KO mice display both normal glycemic control and low circulating plasma insulin (13). In addition, IP6K1 gene–deleted mouse embryonic fibroblasts demonstrate increased Akt phosphorylation and glucose transport rates (11, 13). Though important, these basic and reductionists approaches lack complete translational relevance. Here we show that IP6K1 protein content in insulin-resistant skeletal muscle does not correlate with whole body measures of glucose control. Conversely, exercise decreased IP6K1 protein content in human skeletal muscle, and the exercise treatment that caused the greatest improvements in S_I²* also caused the greatest decrease in muscle IP6K1 content. A supplemental aim of this work was to investigate IP6K1 roles in insulin signaling in skeletal muscle. Our work showed that the chemical inhibition of IP6K1 in vitro increased phosphorylation of Akt at both Ser⁴⁷³ and Thr³⁰⁸ in the skeletal muscle C2C12 cell line. Whereas insulin increased AS160^Thr642, IP6K1 inhibition had no additive effect on this important target in insulin-stimulated glucose uptake, suggesting that IP6K1 may not interfere with AS160 activity despite increasing its upstream activator pAkt.

The production of IP7, stimulated in part by insulin, is known to compete with PIP3 at the PH domain of Akt, inhibiting subsequent translocation and phosphorylation of Akt by PDK1. This process may represent a negative feedback mechanism whereby hyperinsulinemia eventually decreases Akt activity (11). Given this notion, we hypothesized that muscle IP6K1 protein content would correlate with two-compartment modules of (S_I²*), yet despite a negative relationship, this correlation was not found to be significant (r = 0.019; P = 0.961). Interestingly, plasma IP6K1 demonstrated the strongest relationship with S_I²*, although it was not significant (r = 0.389; P = 0.110). This finding suggests that muscle and plasma IP6K1 are not key mediators in the development of whole body IR, despite being seemingly important in an in vitro model of IR.

It is worth noting the small sample size in the current work and the fact that this study focused on skeletal muscle in isolation. Therefore, future work should examine the role of IP6K1 in adipose tissue as well as other insulin-sensitive tissue. In addition, a major limitation of these data is the assumption that muscle protein signaling in the vastus lateralis reflects insulin signaling in other muscle groups and insulin-sensitive tissue. Data published elsewhere report that glucose uptake is different for different muscle types (35). However, our data did show a significant relationship between HOMA_IR and plasma IP6K1, with the former known to correlate with one-compartment modules of insulin sensitivity (36) and validated against the euglycemic-hyperinsulinemic clamp technique (37), considered the gold standard assessment of insulin sensitivity and secretion (25, 38). Taken together, these findings suggest that HOMA_IR is a useful measure in the assessment of glucose control and that circulating IP6K1 may be implicated as an available predictor of IR.

In skeletal muscle, insulin-mediated IRS activation causes the downstream phosphorylation of Akt and AS160 to facilitate translocation of GLUT4 proteins to the plasma membrane, where they fuse, leading to increased glucose uptake into the cell (39). Phosphorylation of AS160^Thr642 was also elevated after both exercise conditions, suggesting that exercise can increase S_I²* and pAS160^Thr642 while also decreasing muscle IP6K1. Our data do not allow us to speculate on a possible mechanism linking IP6K1 and AS160. Yet it is likely that any relief of the inhibitory effects of IP6K1 on Akt (11) would probably result in an increase in pAS160^Thr642, particularly in a postexercise muscle cellular environment. Chemical inhibition of IP6K1 with TNP in vitro elevated Akt³⁰⁸ and Akt⁴⁷³ phosphorylation yet had no additive effect over insulin on pAS160^Thr642, suggesting that IP6K1 may play a part in Akt activity, but this may not be sufficient to change downstream related targets.

Contrary to our hypothesis, IP6K1 protein content was not correlated with whole body measures of S_I²* or S_G²*. At least from the current data set, it appears that exercise can decrease muscle IP6K1 content in the acute period (∼1.5 hours) (40) after muscle contraction and that the same stimulus improved S_I²*. We note two findings in our exercise data: high-intensity exercise had a greater effect on two-compartment models of insulin sensitivity when compared with the lower-intensity exercise at 90% of lactate threshold, and high-intensity exercise decreased HGP in the 4 hours after exercise. Continuous moderate-intensity exercise offered clear improvements in S_I²* but showed no change in HGP. These data suggest that greater improvements in glucose control can be obtained with high-intensity exercise over more traditional forms in prediabetic patients. The cellular mechanisms explaining improvements in insulin sensitivity in response to exercise have been well documented and reviewed elsewhere (41, 42). The mechanisms by which muscle contraction influences other insulin-sensitive tissue remains a key question. Elevated HGP is a major contributing factor to hyperglycemia in T2D (43) owing to hepatic IR. Increased glucose rate of appearance (Ra) is a normal and well-documented response to exercise, and higher-intensity exercise is met with a greater glucose Ra over moderate-intensity exercise (44–46). The increase in glucose Ra during exercise is a product of increased hepatic adenosine 5′-monophosphate, elevated adenosine 5′-monophosphate–activated protein kinase (AMPK) levels (47), and increased hepatic glucagon delivery (48) and sensitivity (49). Indeed, increased glucagon and reduced insulin (50–52) are key contributory factors of HGP during exercise. The rise in glucagon causes a decrease in hepatic glycogenolysis and gluconeogenesis, while a reduction in insulin secretion also causes hepatic glycogenolysis (51). Thus postexercise hepatic Ra is likely to be downregulated after higher-intensity exercise due to a reduction in hepatic insulin requirements (53), resynthesis of liver glycogen (54), and an exercise-induced increase in hepatic AMPK (55) and IRS-2 (56). AMPK inhibits phosphoenolpyruvate carboxykinase and glucose-6-phosphatase (57), both of which are key enzymes responsible for reducing gluconeogenesis while upregulation of IRS-2 is associated with improved hepatic sensitivity to insulin (58). Data show that postexercise ingestion of ¹³C-glucose increased hepatic glycogen resynthesis by 0.7 mg/kg/min over a 4-hour period in humans (52, 59).

High-intensity exercise also decreased muscle IP6K1 content while increasing pAkt^Thr308. Historical exercise data consistently show increased Akt expression after exercise because of its key role in protein synthesis (60) and insulin-stimulated glucose uptake (12). Short-term TNP treatment increases pAkt^Thr308, pGSKα^Ser21, and pGSKβ^Ser9 in mice (61), suggesting that inhibition of IP6K1 has the potential to increase the activity of key proteins in the insulin signal cascade. This notion is supported by the current study (Fig. 5F). The mechanism by which high-intensity exercise decreases IP6K1 is currently unexplained. Previous research has suggested that increased intercellular Ca²⁺ levels (62) may interfere with IP6K1-IP7 signaling (63). Muscle contraction requires the depolarization of the sarcolemma, resulting in Ca²⁺ releases from the muscle sarcoplasmic reticulum. Thus, the increase in intracellular Ca²⁺ concentration may be the link between exercise and reduced IP6K1 levels. Yet IP6 has been shown to suppress excitatory neurotransmission in hippocampal neurons by inhibiting the presynaptic Syt1-C2B domain (64). The synaptotagmin 1 (Syt1) is a key Ca²⁺ sensor essential for synaptic membrane fusion. The interaction of IP6K1 and its products on Ca²⁺ actions in skeletal muscle warrants further investigation.

Ghoshal et al. (61) suggested that IP6K1 inhibition in rodents may reduce the inhibitory effects of IP7 on both pAkt and energy expenditure, the latter caused in an AMPK-dependent mechanism. IP6K/5-IP7 inhibits Akt and LKB1-AMPK pathways (65–68), with both pathways known to increase UCP1-mediated thermogenesis (58, 69–72). Exercise also stimulates AMPK during muscle contraction, with elevated pAkt-GSK3 a current picture in a postexercise muscle environment. Thus, IP6K1-mediated regulation of AMPK- and Akt-dependent mechanisms has the potential to upregulate glucose transport and offer the appearance of increased whole body insulin sensitivity. In support of the former point, data from our laboratory show that pAMPK^Thr172 is increased in C2C12 muscle cells in response to insulin and insulinlike growth factor treatment when supplemented with TNP (Fig. 5F). AMPK protein content was not determined in human muscle homogenate from the current study.

In conclusion, muscle IP6K1 did not correlate with insulin sensitivity, as measured by the labeled intravenous glucose tolerance test. However, plasma IP6K1 was related to HOMA_IR in hyperinsulinemic prediabetic humans, suggesting that global IP6K1 and not muscle-bound IP6K1 may be implicated in IR. However, high-intensity exercise did reduce muscle IP6K1 content, and this reduction is met with a significant increase in insulin sensitivity. Here we have shown that TNP inhibits IP6K1 in C2C12 myotubes, but this effect is not accompanied by changes in AS160 phosphorylation. Taken together, these data suggest that muscle IP6K1 may play a part in IR, but they do not provide a complete picture, with other signaling intermediates likely to be involved.

Abbreviations:

 
• AMPK

  adenosine 5′-monophosphate–activated protein kinase

 
• AS160

  160-kDa Akt substrate

 
• BSA

  bovine serum albumin

 
• cDNA

  complementary DNA

 
• C_T

  cycle threshold

 
• DMEM

  Dulbecco’s modified Eagle medium

 
• HbA1c

  hemoglobin A1c

 
• HGP

  hepatic glucose production

 
• HOMA_IR

  homeostasis model of insulin resistance

 
• HOMA_β-Cell

  homeostasis model of β-cell function

 
• IP6

  inositol hexakisphosphate

 
• IP6K1

  inositol hexakisphosphate kinase-1

 
• IP7

  diphosphoinositol pentakisphosphate

 
• IR

  insulin resistance

 
• IRS

  insulin receptor substrate

 
• KO

  knockout

 
• mRNA

  messenger RNA

 
• PCR

  polymerase chain reaction

 
• PDK1

  phosphatidylinositol 3,4,5-triphosphate-dependent protein kinase 1

 
• PH

  pleckstrin homology

 
• PIP3

  phosphatidylinositol-3,4,5-triphosphate

 
• Ra

  rate of appearance

 
• S_G²*

  glucose effectiveness

 
• S_I²*

  insulin sensitivity

 
• T2D

  type 2 diabetes

 
• TNP

  N2-(m-trifluorobenzyl), N6-(p-nitrobenzyl)purine.

Acknowledgments

Disclosure Summary: The authors have nothing to disclose.