Elevated NF-κB Activation Is Conserved in Human Myocytes Cultured From Obese Type 2 Diabetic Patients and Attenuated by AMP-Activated Protein Kinase

This study was performed in accordance with the Helsinki II Declaration and approved by the regional research ethics committee (H-KF-01–141/04). Skeletal muscle biopsies from the vastus lateralis were obtained by the percutaneous needle biopsy method from 24 men and women. Impaired glucose tolerance (IGT) and T2D were diagnosed by an oral glucose tolerance test according to World Health Organization Classification of Diabetes Mellitus (The Decode Study Group, 2001). T2D subjects were all on antidiabetic medications, three of whom were prescribed metformin. None of the volunteers had any history of infections, fever, or inflammatory diseases. Obese volunteers were defined as having a BMI >30 kg/m²; nonobese control subjects had a BMI <30 kg/m². Volunteers were divided, based on the preceding measurements, into nonobese with normal glucose tolerance (Non-Ob), obese with normal glucose tolerance (Ob-NGT), obese with IGT (Ob-IGT), and obese with T2D (Ob-T2D) groups. All volunteers gave their written consent before participation.

F10 nutrient mixture (HAM), Dulbecco’s modified Eagle’s medium (DMEM) low and high glucose, FBS, horse serum (HS), Fungizone antimycotic, and penicillin/streptomycin (P/S) were all from Invitrogen (Taastrup, Denmark). A769662 was purchased from Tocris Biosciences (Bristol, U.K.), and insulin (Actrapid) was from Novo Nordisk (Bagsværd, Denmark). Complete (mini) protein phosphatase inhibitor tablets were purchased from Boehringer-Roche Diagnostics (Copenhagen, Denmark), and protein protease inhibitor I and II and streptavidin–horseradish peroxidase were from Sigma-Aldrich (Brøndby, Denmark). ATP[³²P] and 2-deoxy-D-[³H]glucose were from Perkin Elmer Life Sciences (Copenhagen, Denmark). Phospho–inhibitor of κB kinase (IKK)-α/β (Ser^180/181), anti–NF-κB-p65, phospho-Akt/protein kinase B (PKB) (Ser⁴⁷³), anti–β-tubulin, phospho-ACC (Ser⁷⁹), phospho–stress-activated protein kinase/Jun NH₂-terminal kinase (SAPK/JNK) (Thr¹⁸³/Tyr¹⁸⁵), and anti-Akt/PKB antibodies were from Cell Signaling Technology (Boston, MA). Anti–inhibitor of κB (IκB)-α and anti–AMPK-α1/α2 were from Santa Cruz Biotechnology (Delaware, CA); AMPK-α2 antibody used to immunoprecipitate AMPK was a gift from Professor Hardie (Dundee, Scotland, U.K.). NF-κB-p65 DNA binding activity ELISA kit came from Active Motif (Carlsbad, CA). ELISA kit for TNF-α and IL-6 was purchased from Meso Scale Discovery (Gaithersburg, MD).

Muscle precursor cells were isolated from vastus lateralis muscle biopsies as previously described (17). After removal of fat and connective tissue, the biopsy was digested in a 10 mL buffer containing trypsin and collagenase II for 15 min. To minimize fibroblast contamination, cells were preseeded in a culture dish for 3 h in F10/HAM, 20% FBS, 1% PS, and 1% Fungizone antimycotic. Unattached cells were then removed and seeded into a culture flask, precoated with Matrigel (BD Biosciences, San Jose, CA). After 4 days of incubation, the cell culture medium was changed and then every 2nd day thereafter. Cell cultures were expanded and then seeded on Matrigel-coated plates for differentiation. Next, 100% confluent cells were incubated with DMEM containing 1 g/L glucose, 10% FBS, and 1% P/S to allow cells to align; after 2 days, the medium was changed to DMEM containing 4.5 g/L glucose, 2% HS, and 1% P/S. Culture medium was changed to DMEM containing 1 g/L glucose, 2% HS, and 1% P/S 2 h prior to treatment with A769662. For all experiments, cells were used at day 6 or 7 of differentiation at either passage 5 or 6. All four cell groups were used on the same day of differentiation and at the same passage number for each experimental dataset.

Cells were rinsed once in ice-cold PBS and lysed in 20 mmol/L Tris, pH 7.5, 150 mmol/L NaCl, 1 mmol/L EGTA, 1 mmol/L EDTA, 0.1% Triton X-100, protease inhibitor (1 tablet/10 mL), and 1% phosphatase inhibitor cocktail. Whole cell lysates were centrifuged (12,000g at 4°C for 10 min), and supernatants were removed for storage at −80°C until required.

Protein concentrations were determined using the Bradford reagent (18). Whole cell lysates (20 μg) were subjected to SDS-PAGE using Invitrogen 4–12% precast gels and I-blot dry transfer machine according to the manufacturer’s instructions. Polyvinylidene fluoride membranes were probed with primary antibodies raised against the protein of interest as indicated in the figure legends. Detection of primary antibodies was performed using appropriate peroxidize-conjugated IgG and protein signals visualized using FEMTO-enhanced chemiluminescence and Bio-Rad Chemidoc XRS imager (Bio-Rad, Hercules, CA). Quantification of immunoblots was done using ImageJ (National Institutes of Health, Bethesda, MD, http://rsb.info.nih.gov/ij).

Measurement of NF-κB-p65 DNA binding activity was carried out on 10-μg whole cell lysate using Active Motif ELISA kit according to the manufacturer’s instructions.

Levels of TNF-α and interleukin (IL)-6 were measured 4 h post changing of cell culture media (100 μL) using Meso Scale Discovery proinflammatory ELISA kit according to the manufacturer’s instructions.

Myocytes were serum starved in DMEM containing 1 g/L glucose for 2 h prior to assaying glucose uptake and incubated with reagents for times and at concentrations indicated in the figure legends. Cells were washed twice with HEPES-buffered saline (140 mmol/L NaCl, 20 mmol/L HEPES, 5 mmol/L KCl, 2.5 mmol/L MgSO₄, and 1 mmol/L CaCl₂, pH 7.4). Glucose uptake was assayed by incubation with 10 μmol/L 2-deoxy-D-[³H]glucose (1 μCi/mL) for 10 min. Nonspecific tracer binding was determined by quantifying cell-associated radioactivity in the presence of 10 μmol/L cytochalasin B. Medium was aspirated before washing adherent cells twice with 0.9% ice-cold NaCl. Cells were subsequently lysed in 50 mmol/L NaOH, and radioactivity was quantified using a Perkin Elmer Tri-Carb 2810TR scintillation counter. Protein concentration was determined using the Bradford reagent.

Protein G-sepharose beads were washed once in PBS and incubated with anti–AMPK-α2 for 2 h at 4°C on an orbital platform shaker. Bead-antibody mix was then incubated with 500 μg of whole lysate protein for 2 h at 4°C before washing. The immunoprecipitates were washed twice in immunoprecipitation buffer (50 mmol/L Tris, 150 mmol/L NaCl, 50 mmol/L NaF, 5 mmol/L NaPPi, 1 mmol/L EDTA, 1 mmol/L EGTA, 1 mmol/L dithiothreitol, 0.1 mmol/L benzamidine, 0.1 mmol/L phenylmethylsulfonyl fluoride, 5 μg/mL soybean trypsin inhibitor, and 1% triton, pH 7.25) containing 0.5 mol/L NaCl, once in immunoprecipitation buffer containing 0.15 mol/L NaCl, and once in HEPES assay buffer (50 mmol/L HEPES, 1 mmol/L dithiothreitol, and 0.02% Brij-35, pH 7.0).

AMPK-α2 immunoprecipitates were used to measure AMPK activity assayed in vitro by the phosphorylation of a synthetic peptide HMRSAMSGLHLVKRR (the “SAMS” peptide) in the presence of a saturating concentration of AMP (0.2 mmol/L). Kinase activity was assessed by a 15 min incubation at 30°C with reaction buffer ([³²P]ATP, 5 mmol/L MgCl₂, 200 μmol/L AMP, plus or minus 200 µmol/L SAMS peptide) on a shaking platform. The reaction was terminated by spotting 30 μL of the reaction mixture onto P81 phosphocellulose paper and washing in 1% phosphoric acid. Radioactivity was quantified using Perkin Elmer Tri-Carb 2810TR scintillation counter.

For multiple comparisons, statistical analysis was performed using one-way or two-way ANOVA with Bonferroni corrections. For data that were normalized to basal (i.e., fold changes), statistical analysis was performed using a one-sample t test. Data analysis was performed using GraphPad Prism (La Jolla, CA) software and considered statistically significant at P < 0.05.

The clinical characteristics of the subjects used in this study are shown in Table 1. All subjects were age-matched (54 ± 1 years) from a mixture of males and females. All obese subjects were matched for BMI (34.2 ± 0.9), which was statistically higher than Non-Ob control subjects (P < 0.05) (Table 1). We have previously shown that plasma markers of inflammation (IL-6, TNF-α, and C-reactive protein) were significantly upregulated in both obese and diabetic subjects compared with healthy nonobese control subjects (19,20). However, inflammation in skeletal muscle isolated from these subjects was not investigated. Here we show that skeletal muscle from Ob-T2D patients has significantly lower protein expression of IκB-α (∼70% lower) when compared with skeletal muscle from Non-Ob subjects (Fig. 1A and B). It is important that this lower expression of IκB-α in Ob-T2D muscle was sufficient to significantly increase NF-κB (p65 subunit) DNA binding activity (Fig. 1C). There was no significant difference in NF-κB-p65 subunit DNA binding activity between skeletal muscle from Non-Ob and either Ob-NGT or Ob-IGT muscle.

We wanted to assess whether this increased inflammation was conserved in our system of cultured muscle cells isolated from the same subjects. To assess this, we measured the phosphorylation of IKK-α/β and the associated loss in its downstream target, IκB-α, as a readout for NF-κB activation. We show that myocytes from Ob-T2D have threefold higher IKK-α/β phosphorylation and concomitant loss in IκB-α protein abundance of 45% (Fig. 2A and B) compared with Non-Ob myocytes. Importantly, we found that the level of IκB-α protein expression in skeletal muscle tissue and muscle precursor cells from the same individual had a significant correlation (P = 0.0034) with an r² value of 0.5 (Fig. 2C). To assess whether this loss in IκB-α protein was sufficient to upregulate NF-κB activity, the ability of the p65 subunit of NF-κB to bind to DNA was measured. Figure 2D shows that Ob-T2D myocytes have a significantly higher level of p65 DNA binding compared with Non-Ob myocytes. It is important that Ob-NGT myocytes had comparable levels of NF-κB signaling with Non-Ob control subjects, whereas Ob-IGT myocytes seemed to have an intermediate level of NF-κB signaling (between that of Non-Ob and Ob-T2D). To fully elucidate the level of inflammation in these cells, the concentrations of TNF-α and IL-6 in cell culture media were measured. TNF-α concentrations were significantly higher in Ob-T2D compared with Non-Ob control subject concentrations (Fig. 2E). Of interest, concentrations of IL-6 in cell culture media from Ob-IGT and Ob-T2D was significantly lower than that in Non-Ob cell media (Fig. 2F). This latter finding supports the idea that IL-6 exerts an inhibitory effect on TNF-α in muscle (21).

To relate the inflammatory status with glucose tolerance in the cultured myocytes, it was important not only to refer to the subjects from whom cells were isolated (as shown in Table 1) but also to confirm that impaired insulin sensitivity was retained in cultured myocytes. To ascertain this, insulin-stimulated glucose uptake and PKB/Akt phosphorylation of myocytes were measured. Both Non-Ob and Ob-NGT myocytes had comparable levels of basal and insulin-stimulated glucose uptake, with a significant increase in glucose uptake (approximately twofold) in response to insulin (Fig. 3A). It is important that both Ob-IGT and Ob-T2D myocytes had no increase in glucose uptake when challenged with insulin. This lack of insulin sensitivity in these cells was reflected in the loss of insulin-stimulated PKB/Akt Ser 473 phosphorylation (Fig. 3B and C). Hence, the in vivo glucose tolerance status of the donors is reflected in cultured myocytes.

Obesity had no effect on the level of AMPK-α1/α2 protein expressed in the cells. However, in the Ob-T2D myocytes, there was a significant increase in the total protein amount of downstream AMPK target, ACC (Fig. 4A and B). It is interesting that obesity was associated with a trend of increased basal AMPK-α2 activity, independent of glucose tolerance of the cells. Figure 4C shows that in obese myocytes, there is a threefold increase in basal AMPK activity compared with that observed in Non-Ob myocytes. Treatment of cultured human myocytes with A769662 for 4 h significantly increased AMPK-α2 activity to equal extents in all cell groups (Fig. 5A). Of interest, AMPK activation in Ob-IGT and Ob-T2D myocytes led to a significantly higher level of ACC phosphorylation compared with Non-Ob and Ob-NGT control subjects. The latter observation cannot be fully accounted for by increased protein expression levels of ACC (Fig. 5C and D). A769662 treatment did not alter the protein expression of AMPK-α1/α2.

It has been reported that AMPK is able to antagonize the effects of a number of proinflammatory markers in a variety of cell lines (12–15). However, its role as a modulator of inflammation has not been well established in human skeletal muscle. Therefore, we assessed the effect of pharmacological activation of AMPK using the specific AMPK activator, A769662, on NF-κB signaling in Non-Ob and obese myocytes. It is important that activation of AMPK using A769662 attenuated IKK phosphorylation to a level that was not significantly different from Non-Ob control subject levels. The loss of IκB-α protein expression was also attenuated in cells from Ob-T2D (Fig. 6A and B). In line with this, A769662 treatment of Ob-T2D cells attenuated the level of NF-κB-p65 DNA binding activity to that comparable with Non-Ob myocytes (Fig. 6C). Because levels of TNF-α were robustly increased in the cell culture media from basal Ob-T2D myocytes, we measured concentrations of this cytokine after 4 h of AMPK activation. A769662 suppressed TNF-α in the cell culture media to a level that was not significantly different from that measured in media from Non-Ob cells (Fig. 6D). Although A769662 did not significantly increase IL-6 concentrations in Ob-IGT and Ob-T2D media, there was a trend to increase IL-6 concentrations after A769662 treatment in these groups (Fig. 6E).

A769662 treatment had no additive effect on insulin-stimulated glucose uptake or PKB/Akt phosphorylation in any of the cell groups tested (Fig. 7A and B). As shown in Fig. 2, Ob-IGT and Ob-T2D myocytes show no response to insulin as measured by glucose uptake and PKB/Akt phosphorylation. It should be noted, however, that there was a trend toward increased basal glucose uptake in Ob-T2D myocytes (P = 0.065). Of interest, 4 h of A769662 treatment, which was sufficient to attenuate inflammation, did not increase insulin sensitivity of Ob-IGT or Ob-T2D myotubes (Fig. 6A). This finding was also confirmed at the level of PKB/Akt phosphorylation, which remained suppressed in Ob-IGT and Ob-T2D myocytes compared with Non-Ob and Ob-NGT myocytes (Fig. 7B).

Our findings indicate that skeletal muscle precursor cells that are grown and differentiated into myocytes in culture, despite multiple passages, retain both an obesity/T2D-associated inflammatory phenotype and insulin resistant phenotype seen in vivo. This finding provides evidence for justifying the use of muscle precursor cultures as tools to define the role of increased inflammation in obesity and T2D in humans ex vivo. It is interesting that no increase in inflammation was found in myocytes from Ob-NGT subjects when compared with Non-Ob control subjects. Inflammation was, however, significantly increased in Ob-T2D myocytes when compared with both Non-Ob and Ob-NGT cells (Fig. 1C). Similar findings have also been shown in adipose tissue, where adipose tissue from moderately obese insulin-resistant individuals expressed significantly higher mRNA levels of IL-8, IL-6, and monocyte chemoattractant protein–1 when compared with levels in moderately obese insulin-sensitive tissue (22). One hypothesis that can be drawn from these data is that hyperglycemia (associated with insulin resistance) is a major driving force behind inflammation seen in Ob-IGT and Ob-T2D. In support of this, it has been shown that in healthy individuals, clamping glucose at hyperglycemic levels (20 mmol/L) leads to a significant increase in plasma TNF-α, IL-18, and IL-6 within 2 h (23). A subsequent study shows that a glucose meal (75 g) alone resulted in increased plasma inflammatory and oxidative stress markers, including vascular cell adhesion molecule–1 and intracellular adhesion molecule–1 (24). It is possible that duration of obesity, partitioning of adiposity, or stage of T2D the volunteers were at when biopsies were taken could affect how obesity inflammation and insulin sensitivity interact in these cells. The analysis of these parameters in the human subjects is still ongoing and will be necessary for future understanding of the directionality of the inflammation–insulin resistance relationship.

AMPK is a known target for the treatment of diabetes; however, recently it has been suggested that AMPK confers an anti-inflammatory response (although the mechanisms by which this occurs are currently unclear). Metformin, a widely prescribed insulin-sensitizing drug, and AICAR are both well-established (nonspecific) AMPK activators that have been shown to inhibit NF-κB activity in endothelial cells (13,25). Because of the potential health benefits of anti-inflammatory drugs for both metabolic syndrome and other chronic inflammatory diseases, it is important to establish this role of AMPK in humans. The thienopyridone A769662 directly activates AMPK, mimicking the effects of AMP; however, it has no effect on the binding of AMP to the Bateman domains of the γ-subunit (26). Instead, it has been shown that A769662 activates AMPK by a mechanism involving the β-subunit carbohydrate binding domain in addition to other sites on the γ-subunit, although it directly activates only AMPK-β1–containing complexes (in cell-free assays) (27). It is important, therefore, to note that human skeletal muscle has been shown to express the AMPK-β2 subunit predominantly, with very little β1 expression (28). Despite this, AMPK was significantly activated in all myocytes treated with A769662. We have attributed this to the 50% increased expression of the β1 subunit (with no effect on β2 expression) in cultured myocytes when compared with skeletal muscle tissue (Supplementary Fig. 1).

Having confirmed that A769662 significantly activated AMPK in human skeletal myocytes, we found that the increased inflammation associated with Ob-T2D myocytes could be attenuated by this direct, pharmacological activation of AMPK. More specific, AMPK activation suppressed NF-κB signaling (IKK phosphorylation, IκB-α, and p65 DNA binding) leading to decreased secretion of the inflammatory cytokine TNF-α. However, pharmacological activation of AMPK, although comparable in all cells, failed to attenuate obesity-associated insulin resistance in either Ob-IGT or Ob-T2D myocytes. Similar findings were recently published by Green et al. (16) in rat L6 myotubes, where inflammation induced by palmitate could be attenuated by AMPK activation; however, this failed to increase the insulin sensitivity of these cells. This latter finding would suggest that increased NF-κB signaling and increased TNF-α concentrations in obesity are not the main causative factors in the development of insulin resistance. Yet it is possible that other factors that promote insulin resistance may not be regulated by AMPK activation, for example, the mitogen-activated protein kinase (MAPK) family. Both JNK and p38 MAPK have been implicated in the pathogenesis of insulin resistance in muscle (29). It has been shown that AMPK activation in L6 muscle cells attenuates palmitate-associated extracellular signal–related kinase phosphorylation but fails to attenuate either palmitate-induced JNK or p38 phosphorylation under the same conditions (16). Further investigation into the causes and potential role of MAPKs in insulin resistance in muscle precursor cells and the affect of AMPK activation on this will be investigated in our future work to fully elucidate the interplay between inflammation and insulin resistance in obese human muscle precursor cells. It is possible that our findings might reflect the complex process by which insulin resistance is developed and that once established in the muscle precursor cells, it probably requires more than just reversing a few contributing factors.

Our findings suggest that obesity, independent of insulin sensitivity, is associated with a significant increase in basal AMPK-α2 activity. This latter finding fits with findings from other laboratories that show a higher level of basal AMPK activity in obese human muscle and in rat muscle cells treated with saturated fatty acids (30,31). Because AMPK activation is important in regulating fatty acid oxidation and energy balance in the cell, increased basal AMPK activity may be a reflection of the cell’s attempt to compensate for the obesity-driven oversupply of fatty acids, and accumulation of these in skeletal muscle. In line with this, it has been reported that intracellular accumulation of fatty acids increases mitochondrial oxidative capacity, which leads to the impairment of fuel selection from fat to carbohydrate (32). It is therefore likely that increased basal AMPK activity in obese myocytes reflects increased “impaired” fatty acid oxidation. This impaired fatty acid oxidation could subsequently lead to increased levels of reactive oxygen species (ROS) and the promotion of insulin resistance (33,34). It is important that ROS levels have been shown to be increased in human skeletal muscle from obese insulin-resistant subjects (34). It is also important to note that Ob-T2D myocytes show an increased sensitivity to AMPK activation in terms of ACC phosphorylation. This increase in sensitivity cannot be explained simply by the increased abundance of ACC protein and is not attributable to higher AMPK activity or, indeed, the expression of α-AMPK subunits in these cells. It would be interesting, therefore, to investigate whether this higher level of ACC phosphorylation in response to AMPK activation translates into a higher level of functional fatty acid oxidation or whether this increase further reflects the impaired fatty acid oxidation associated with the oversupply of free fatty acids in obesity.

In conclusion, our findings provide solid evidence that muscle precursor cells that are grown ex vivo in culture are useful tools to investigate the relationship between inflammation and insulin resistance in skeletal muscle. It is also likely that this system retains other disease phenotypes that have not been investigated here. Even though obesity is associated with inflammation in vivo, we demonstrate that obesity is not enough to establish an inflammatory phenotype in muscle precursor cells. In contrast, the increased inflammatory response observed in the muscle precursor cells derived from subjects with T2D possibly reflects the irreversible status of the disease. In addition, we have confirmed the anti-inflammatory role for AMPK in human skeletal muscle, verifying its potential therapeutic role in inflammatory-associated morbidity and mortality associated with metabolic syndrome. However, it must be noted that attenuation of NF-κB activity by AMPK in muscle may not be sufficient to increase or restore insulin sensitivity of the cells alone, and further investigation into how other inflammatory mediators (including JNK and ROS) may promote insulin resistance under these conditions should be investigated in future research using this primary cell model.

The Centre of Inflammation and Metabolism (CIM) is supported by a grant from the Danish National Research Foundation (02-512-55). This work was funded by Marie Curie Inter-European fellowship (FP7) awarded to C.J.G.

No potential conflicts of interest relevant to this article were reported.

C.J.G. researched data and wrote the manuscript. M.P. screened subjects, carried out biopsies on human volunteers, and reviewed and edited the manuscript. B.K.P. and C.S. contributed to discussion and reviewed and edited the manuscript.

The authors thank Prof. Grahame Hardie, University of Dundee, Dundee, Scotland, for the AMPK-α2 antibodies and his laboratory for their expertise during this project. The authors also thank Dr. Thomas Solomon, CIM, Copenhagen, Denmark, for his help with the preparation of the manuscript.