Skeletal muscle insulin resistance is induced by 4-hydroxy-2-hexenal, a by-product of n-3 fatty acid peroxidation

Reagents

4-HNE and 4-HHE were synthesised as previously described [24]. Insulin (Actrapid, 100 U/ml) was from Novo Nordisk (La Défense, France) and ECL reagent and 2-deoxy-d-2,6-[³H]glucose were from GE Healthcare (Orsay, France). 1,2-Dithiole-3-thione (D3T) was from Interchim (Montlucon, France). All other reagents were from Sigma-Aldrich (Saint Quentin Fallavier, France).

Antibodies

Antibodies to Ser473-Akt, total Akt, p85 and total IRS1 (dilution 1:1000) were from Cell Signaling (Leiden, the Netherlands). Secondary antibodies (dilution 1:10,000) were from Sigma-Aldrich (Heidelberg, Germany). Anti-HHE-Michael adducts antibodies were from Cosmobio (Tokyo, Japan). Antibodies for insulin signalling were validated using insulin as a positive control. Anti-HHE antibody specificity was tested against protein lysates incubated with 4-HHE or 4-HNE.

Participants

Animal experiments

All animal experiments were performed under authorisation no. 69-266-0501 (INSA-Lyon, DDPP-SV, Direction Départementale de la Protection des Populations - Services Vétérinaires du Rhône) according to the guidelines laid down by the French Ministry of Agriculture (no. 2013-118) and the European Union Council Directive for the protection of animals (2010/63UE).

The rats used in all experiments were raised in an air-conditioned room with a controlled environment of 21 ± 0.5°C and 60–70% humidity, under a 12 h light–dark cycle (light on from 07:00 to 19:00 hours) with free access to food (2016C, 12.6 kJ/g; Harlan, Gannat, France) and water. Rats were housed together and randomised into groups on the day of the experiment using random numbers.

Zucker diabetic fatty rats

Euglycaemic–hyperinsulinaemic clamps

Clamps were performed in 3-month-old anaesthetised male Wistar rats [25] purchased from Janvier (Le-Genest-Saint-Isle, France). Rats were fasted overnight, anaesthetised with sodium pentobarbital (35 mg/kg i.p.; Sanofi Santé Animale, Centravet, Lapalisse, France) and chlorpromazine (5 mg/kg, i.p., Largactil; Sanofi-Aventis, Paris, France) and then implanted with indwelling catheters (PE-20; Phymep, Paris, France) in the left carotid artery and left and right jugular veins. After catheterisation, four rats were infused with 4-HHE (10 mg/kg, 0.1 ml) and four rats were infused with the vehicle (DMSO, 0.1 ml) in the left jugular vein. A standard 2 h clamp [25] was conducted using a primed and continuous infusion of human recombinant insulin (Actrapid) at a rate of 6 mU kg⁻¹ min⁻¹ coupled with a variable infusion of 25% (wt/vol.) glucose to maintain blood glucose concentration at approximately 6 mmol/l. Blood glucose was measured every 5 min using a glucometer (Accu-Check Performa). Plasma insulin concentration was determined at the end of the clamp by EIA assay (SpiBio, Montigny le Bretonneux, France).

In vivo insulin stimulation and insulin signalling in skeletal muscle

Male Wistar rats, fasted for 7 h prior to the experiments, were anaesthetised with sodium pentobarbital (35 mg/kg i.p., Sanofi Santé Animale, Centravet) and chlorpromazine (5 mg/kg i.p., Largactil; Sanofi-Aventis). Body temperature was maintained at 37°C with a homeostatic blanket (Harvard apparatus, les Ullis, France). A tracheotomy (PE-240; Phymep, Paris, France) was performed to facilitate breathing. A catheter (PE-20) was inserted into the left jugular vein for 4-HHE infusion. Rats were either infused with 4-HHE (10 mg/kg in 0.1 ml DMSO, n = 5) or 0.1 ml of DMSO as a control (n = 4). One hour after infusion, rats received an intravenous injection of either saline (154 mmol/l NaCl) or insulin (0.75 U/kg body weight). Thirty minutes after insulin injection, rats were killed with an intravenous overdose of sodium pentobarbital (120 mg/kg), blood was collected by heart puncture into a heparinised syringe and gastrocnemius muscle was rapidly dissected out, blotted dry and snap-frozen in liquid nitrogen.

Plasma 4-HHE and 4-HNE measurement

Five hundred microlitres of plasma were used to measure 4-HHE and 4-HNE by GC–MS as previously described [26].

Cell culture

Rat L6 muscle cells were obtained from the American Type Culture Collection (LC/GC, Molsheim, France) and grown as described previously [15]. Cells were starved of serum for 4 h before treatments, which were performed in serum-free Minimum Essential Medium Eagle Alpha Modification (Sigma-Aldrich, Saint Quentin Fallavier, France). Viability and cell death were determined by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) reduction (Cell Proliferation Kit I; Roche), lactate dehydrogenase release (In vitro toxicology assay kit; Sigma-Aldrich) and caspase 3 activity (Caspase 3 Assay Kit; Sigma-Aldrich). Mycoplasma contamination was regularly tested by PCR.

2-Deoxy-d-[³H]glucose uptake

Cells were treated with 4-HHE and incubated for 20 min with 100 nmol/l insulin or 20 μmol/l cytochalasin B. Glucose uptake was initiated by the addition of 2-deoxy-d-[³H]glucose (747 GBq/mmol) to a final concentration of 0.1 mmol/l for 5 min at 37°C. Uptake was terminated by three washes in ice-cold PBS and solubilisation in 0.1% (wt/vol.) SDS. Tritium was detected by liquid scintillation counting and results normalised by protein concentration measured using the Bradford assay. Non-specific uptake measured in presence of cytochalasin B was subtracted from each determination.

Spectrophotometric 2,4-dinitrophenylhydrazine assay

Carbonyl groups on proteins were determined using 2,4-dinitrophenylhydrazine (DNPH) as previously described [27]. Carbonyl content was determined from the absorbance at a wavelength of 370 nm using a molar absorption coefficient of 22,000 l mol⁻¹ cm⁻¹ and normalised to the protein concentration measured at 280 nm.

Protein extraction, immunoprecipitation and immunoblotting

After incubation with 4-HHE, cells were washed with PBS and proteins were extracted as described previously [15]. For IRS1/p85 detection, whole-cell lysates were immunoprecipitated using antibodies against IRS1. Proteins were boiled in Laemmli buffer, separated by SDS-PAGE and transferred onto nitrocellulose membrane. For dot blotting, 20 μg of proteins from whole-cell lysate were loaded directly onto the nitrocellulose membrane. Following saturation with 5% (wt/vol.) BSA, membranes were probed with antibodies and processed for chemiluminescence (ECL plus; GE Healthcare, Coutaboeuf, France). Quantification was performed using the ImageJ software (National Institutes of Health, Bethesda, USA).

Immunofluorescence

L6 cells grown on coverslips were incubated with 4-HHE, fixed for 30 min with 3% (wt/vol.) paraformaldehyde and quenched with 50 mmol/l NH₄Cl for 5 min. After PBS washing, cells were permeabilised with 50 μg/ml digitonin for 10 min and blocked with 0.1% (wt/vol.) BSA for 30 min. The cells were incubated with the first antibodies and then washed and labelled with fluorescent secondary antibodies. The stained cells mounted in Mowiol (Sigma Aldrich, Saint Quentin Fallavier, France) were examined under a Zeiss LSM 510 confocal microscope (Zeiss, Marly le Roy, France) equipped with 63W oil objective and fluorescence was quantified using Image J software v1.56.

Glutathione assay

Reduced glutathione (GSH) was measured using a commercially available kit from BioVision (Clinisciences, Montrouge, France) and normalised to the protein concentration measured using the Bradford assay.

Statistics

Experimenters were not blind to group assignment. Data were analysed using Graphpad Prism (Graphpad Software, La Jolla, CA, USA) and presented as means ± SEM. Results were compared by one-way or two-way ANOVA followed when appropriate by post hoc Fisher PLSD tests. Simple comparisons were performed with Student’s t test using Welch’s correction for variance in homogeneity whenever needed. Differences were considered significant at the p < 0.05 level.

4-HHE accumulates in plasma from individuals with type 2 diabetes

The concentration of free 4-HHE, measured using GC–MS, was significantly increased in plasma from individuals with type 2 diabetes compared with plasma from healthy volunteers (33 vs 14 nmol/l, p < 0.001) (Fig. 1a). Diabetic individuals also exhibited a significant increase (sevenfold) in 4-HHE Michael adducts on plasma proteins (Fig. 1b). In contrast, 4-HNE levels did not differ between diabetic and healthy individuals; neither did the plasma concentrations of n-3 and n-6 fatty acids (data not shown). Normalisation of 4-HHE and 4-HNE to the concentration of their respective n-3 and n-6 precursors showed that one out of 20,000 n-3 PUFA were oxidised in the form of 4-HHE, while only one out of 100,000 n-6 PUFA were oxidised in the form of 4-HNE (Fig. 1c). Significant correlations with 4-HHE plasma concentrations were found for blood glucose levels (Fig. 1d) and for BMI, HbA_1c, triacylglycerol and HDL-cholesterol (Table 3).

Open image in new window

Table 3

 Univariate correlations with 4-HHE plasma concentrations in participants

Variable	r	p value
Age, years	0.420	0.017
BMI, kg/m2	0.692	<0.001
SBP, mmHg	0.014	0.942
DBP, mmHg	−0.300	0.113
Fasting glucose, mmol/l	0.695	<0.001
HbA1c (mmol/mol)	0.750	<0.001
Triacylglycerol, mmol/l	0.588	0.003
Total cholesterol, mmol/l	−0.302	0.184
HDL-cholesterol, mmol/l	−0.549	0.007
4-HNE, nmol/l	0.267	0.139
Total n-3 PUFA, mmol/l	0.643	<0.001
4-HHE Michael adducts, AU	0.805	<0.001

AU, arbitrary units; DBP, diastolic BP; SBP, systolic BP

4-HHE and 4-HNE in plasma from ZDF rats

A plasma concentration of 40 nmol/l of free 4-HHE was measured in lean ZDF rats and a fourfold increase (p < 0.001) was observed in obese ZDF rats (Fig. 1e). The plasma concentration of 4-HNE was increased by 80% in ZDF rats (p < 0.001). As in humans, a significant correlation was found between the concentration of 4-HHE and blood glucose levels (Fig. 1f).

4-HHE infusion induces insulin resistance in rats

To demonstrate that 4-HHE plays a causal role in insulin resistance, insulin sensitivity was measured using euglycaemic–hyperinsulinaemic clamps in Wistar rats given an intravenous bolus of 4-HHE (10 mg/kg). The injection induced a transient increase in plasma concentration of 4-HHE, which reached 140 nmol/l 30 min after injection, while plasma 4-HNE concentration remained unaffected (Fig. 2a). The clamp was applied at a blood glucose level of approximately 6 mmol/l (Fig. 2b,c). The glucose infusion rate necessary to maintain euglycaemia was significantly lower in the 4-HHE infused rats than in the DMSO-infused rats (Fig. 2d,e), indicating whole-body insulin resistance. In another set of experiments, rats were infused with 4-HHE for 2 h and then received a single injection of insulin for 20 min. Under these conditions, the infusion of 4-HHE prevented the decreased in blood glucose level induced by insulin (Fig. 2f) and abolished insulin-induced phosphorylation of Akt in skeletal muscle (Fig. 2g,h).

Open image in new window

Impaired insulin-stimulated glucose uptake in response to 4-HHE

In lean animals, 80–90% of infused glucose is taken up by skeletal muscle [28] so the clamp technique primarily reflects skeletal muscle insulin sensitivity. We therefore went on to characterise 4-HHE-induced insulin resistance in skeletal muscle cells. In rat L6 cells, 4-HHE completely blunted insulin-induced 2-deoxy-d-[³H]glucose uptake (Fig. 3a), with maximal inhibition observed starting at 10 μmol/l (Fig. 3b), lower than the concentration of 4-HNE previously reported to impair insulin response in adipocytes and muscle cells [14, 15]. Insulin-induced glucose uptake was similarly blunted in L6 cells exposed to 4-HNE, but with a maximal inhibition obtained at a concentration of 50 μmol/l. Significant inhibition of insulin-induced glucose uptake by 4-HHE was observed from 10 min until 4 h of treatment (Fig. 3c). In cells treated for 30 min with up to 100 μmol/l 4-HHE, MTT reduction, lactate dehydrogenase (LDH) release and caspase-3 activity remained unaffected, demonstrating no deleterious effects on cell viability at times and concentrations used (Table 4).

Open image in new window

Table 4

 Effect of 4-HHE and 4-HNE on rat L6 muscle cell viability

Aldehyde concentration	Cell viability (% of control)	LDH activity (% of total)	Caspase 3 activity (% of control)
4-HHE
10 μmol/l	108 ± 2	0.9 ± 0.2	116 ± 2
25 μmol/l	101 ± 3	1.5 ± 0.8	117 ± 6
50 μmol/l	107 ± 4	0.9 ± 0.2	116 ± 12
75 μmol/l	120 ± 5	1.8 ± 0.7	119 ± 14
100 μmol/l	125 ± 1	1.1 ± 0.3	121 ± 12
4-HNE
10 μmol/l	105 ± 5	2.1 ± 0.4	115 ± 16
25 μmol/l	105 ± 3	2.6 ± 0.6	122 ± 15
50 μmol/l	108 ± 5	2.4 ± 0.9	116 ± 13
75 μmol/l	109 ± 3	2.4 ± 0.9	119 ± 19
100 μmol/l	106 ± 2	4.1 ± 1.3	140 ± 20

Data are means±SEM from four independent experiments

L6 muscle cells were treated for 30 min with increasing doses of 4-HHE or 4-HNE in serum-free conditions. Cell viability was measured by MTT assay and expressed as a percentage of untreated control. LDH activity in the supernatant fraction was expressed as a percentage of the maximal level of LDH activity determined after total cell lysis. Caspase-3 activity was expressed as a percentage of the specific activity present in vehicle-treated cells

No significant effects were found

Impaired insulin signalling induced by 4-HHE in rat L6 muscle cells

The insulin signalling cascade leading to glucose uptake can be affected at various stages: GLUT4 translocation; protein kinase B (PKB)/Akt and/or upstream at the insulin receptor and IRS. In response to 4-HHE or 4-HNE, insulin-induced phosphorylation of PKB/Akt was reduced by 50% (Fig. 4a). 4-HHE was effective at inhibiting insulin-induced Akt phosphorylation at concentrations ranging from 10 to 100 μmol/l (Fig. 4b). Similarly, exposure to 4-HHE significantly impaired insulin-induced IRS1 phosphorylation (Fig. 4c), in parallel with a marked reduction in the level of p85 protein co-immunoprecipitated with IRS1 (Fig. 4d).

Open image in new window

4-HHE generates protein adducts in rat L6 muscle cells

Many deleterious effects of aldehydes have been attributed to their ability to form covalent adducts on biomolecules [5]. Indeed, incubation of L6 muscle cells for 30 min with 4-HHE resulted in a dose-dependent increase in carbonyl content (Fig. 5a) and increased formation of 4-HHE Michael adducts on proteins (Fig. 5b). Western blot analysis revealed the presence of 4-HHE adducts on many proteins from 30 to 100 kDa (Fig. 5c) and confocal microscopy showed that the formation of adducts was diffuse in the cell cytoplasm but was excluded from the nuclei (Fig. 5d) without any specific co-localisation with actin cytoskeleton or the Golgi apparatus (data not shown). These data demonstrate that 10 μmol/l 4-HHE is sufficient to induce carbonylation of cytoplasmic proteins in muscle cells.

Open image in new window

Elevation of glutathione pools reverses 4-HHE-impaired glucose uptake

Reduced glutathione (GSH) is one of the major aldehyde detoxification mechanisms [29]. Treatment with increasing concentrations of 4-HHE indeed resulted in a sharp dose-dependent decrease in GSH content in rat L6 muscle cells (Fig. 6a). This decrease was reversed by pre-treatment with D3T (100 μmol/l, 24 h), which doubled the intracellular GSH content and counteracted the decrease induced by 4-HHE (Fig. 6b). D3T also reversed the 4-HHE-induced increase in total carbonyl content (Fig. 6c) and protein adducts in the cells (Fig. 6d). Pre-treatment with D3T increased the basal glucose uptake and effectively prevented 4-HHE-induced insulin resistance (Fig. 6e). Similarly, pre-treatment of L6 muscle cells with N-acetyl-cysteine (NAC) or the scavenger of lipid peroxidation by-products aminoguanidine (AGD) also attenuated Michael adduct formation (Fig. 6f) and prevented the impaired insulin-induced glucose uptake (Fig. 6g) induced by 4-HHE. Altogether, these data confirm that adduct formation is a major player in 4-HHE-induced insulin resistance.

Open image in new window