Feed-forward signaling of TNF-α and NF-κB via IKK-β pathway contributes to insulin resistance and coronary arteriolar dysfunction in type 2 diabetic mice

Abstract

We hypothesized that the interaction between tumor necrosis factor-α (TNF-α)/nuclear factor-κB (NF-κB) via the activation of IKK-β may amplify one another, resulting in the evolution of vascular disease and insulin resistance associated with diabetes. To test this hypothesis, endothelium-dependent (ACh) and -independent (sodium nitroprusside) vasodilation of isolated, pressurized coronary arterioles from mLepr^db (heterozygote, normal), Lepr^db (homozygote, diabetic), and Lepr^db mice null for TNF-α (db^TNF−/db^TNF−) were examined. Although the dilation of vessels to sodium nitroprusside was not different between Lepr^db and mLepr^db mice, the dilation to ACh was reduced in Lepr^db mice. The NF-κB antagonist MG-132 or the IKK-β inhibitor sodium salicylate (NaSal) partially restored nitric oxide-mediated endothelium-dependent coronary arteriolar dilation in Lepr^db mice, but the responses in mLepr^db mice were unaffected. The protein expression of IKK-α and IKK-β were higher in Lepr^db than in mLepr^db mice; the expression of IKK-β, but not the expression of IKK-α, was attenuated by MG-132, the antioxidant apocynin, or the genetic deletion of TNF-α in diabetic mice. Lepr^db mice showed an increased insulin resistance, but NaSal improved insulin sensitivity. The protein expression of TNF-α and NF-κB and the protein modification of phosphorylated (p)-IKK-β and p-JNK were greater in Lepr^db mice, but NaSal attenuated TNF-α, NF-κB, p-IKK-β, and p-JNK in Lepr^db mice. The ratio of p-insulin receptor substrate (IRS)-1 at Ser307 to IRS-1 was elevated in Lepr^db compared with mLepr^db mice; both NaSal and the JNK inhibitor SP-600125 reduced the p-IRS-1-to-IRS-1 ratio in Lepr^db mice. MG-132 or the neutralization of TNF-α reduced superoxide production in Lepr^db mice. In conclusion, our results indicate that the interaction between NF-κB and TNF-α signaling induces the activation of IKK-β and amplifies oxidative stress, leading to endothelial dysfunction in type 2 diabetes.

the prevalence of obesity and diabetes has led to a significant increase in complications associated with cardiovascular disease. Many of these complications can be linked to a dysregulation of the body's immune system (16, 33). Tumor necrosis factor-α (TNF-α) is a proinflammatory cytokine implicated in cardiovascular diseases (14, 34). We previously found that an increased TNF-α expression induces the production of reactive oxygen species (ROS), leading to endothelial dysfunction in type 2 diabetes (15). Importantly, TNF-α activates the transcription nuclear factor-κB (NF-κB), which regulates the expression of genes involved in inflammation, oxidative stress, and endothelial dysfunction (15). TNF-α also initiates signaling cascades via the inhibition of the NF-κB (IκB) kinase (IKK) complex, which contains IKK-α and IKK-β. The inhibitory protein IκB-α is phosphorylated, ubiquitinated, and degraded by proteasomes, releasing NF-κB to translocate into the nucleus. Under normal physiological conditions, the inflammatory response is terminated by binding NF-κB with the inhibitory protein IκB (24, 36). An investigation of the TNF-α-mediated NF-κB pathway is needed to identify specific proinflammatory agents in type 2 diabetes.

Salicylates, including aspirin and sodium salicylate (NaSal), are common nonsteroidal anti-inflammatory drugs, which have antiplatelet and anti-inflammatory effects. Aspirin inhibits prostaglandin production, but NaSal has a prostaglandin-independent effect. NaSal inhibits the activity of IKK-β, which is required for the activation of NF-κB. IKK-β, the molecular target of NaSal, is associated with the rise of insulin resistance (7, 43). Recent work suggests that a specific antagonism of the NF-κB inflammatory pathway through the inhibition of IKK-β reduces acute myocardial damage following ischemia-reperfusion injury (26) and that insulin action plays a critical role in reducing myocardial infarct size by increasing cardiac myocyte metabolism (5, 8, 29). We previously found that TNF-α-activated Jun NH₂-terminal kinase (JNK), which mediates O₂⁻ production and impairs endothelium-dependent, nitric oxide (NO)-mediated vasodilation in coronary arterioles (45). NaSal prevents TNF-α-induced JNK activation from blocking insulin signaling via a serine phosphorylation (18). Therefore, we evaluated 1) whether reciprocal interactions between TNF-α and NF-κB via the activation of IKK-β accentuate the evolution of vascular disease in type 2 diabetes and 2) whether NaSal restores endothelial dysfunction in coronary arterioles and insulin resistance by inhibiting IKK-β/NF-κB activity, thereby preventing JNK activation in type 2 diabetic mice.

MATERIALS AND METHODS

Animals.

The procedures followed were approved by and in accordance with the guidelines of the Laboratory Animal Care and Use Committee at Texas A&M University. Heterozygote controls (mLepr^db), homozygote type 2 diabetic (Lepr^db), and Lepr^db mice null for TNF (db^TNF−/db^TNF−) were purchased from Jackson Laboratory and maintained on a normal rodent chow diet. mLepr^db mice are normal in body weight, blood glucose, and plasma insulin. Our studies used 12–16-wk-old, 15–25-g mLepr^db, 25–50-g Lepr^db, and db^TNF−/db^TNF− mice of either sex. We used the same strain (C57BL/6J) of mLepr^db and db^TNF−/db^TNF− mice to match the backgrounds of Lepr^db mice, type 2 diabetic models that are characterized by leptin resistance and obesity. The cross (db^TNF−/db^TNF−) of Lepr^db with TNF knockout is heterozygous for Lepr^db and homozygous for TNF knockout (TNF⁻/⁻). The db^TNF−/db^TNF− mice show the phenotypes of hyperglycemia and obesity that are consistent with diabetes and the penetrance of the leptin receptor mutation. The obese mice from the second round of breeding of the Lepr^db and TNF⁻/⁻ mice were used in experimentation.

Measurement of blood parameters: blood glucose, cholesterol level, insulin level, and homeostasis model assessment-insulin resistance.

Blood was obtained from the vena cava after anesthesia with pentobarbital sodium (50 mg/kg ip) and the exposure of the vein. Whole blood samples were spun at 3,000 rpm for 10 min, and serum was stored at −80°C until analysis. We used a OneTouch Ultramini glucometer (LifeScan) for measuring blood glucose in mLepr^db, Lepr^db, and Lepr^db mice treated with NaSal at the same time (8:00 am–10:00 am) throughout the protocol. The serum cholesterol level was measured with the Cholesterol/Cholesteryl Ester Quantitation Kit (Biovision), and insulin was measured with the use of a commercial kit, insulin (Mouse) Ultrasensitive EIA (ALPCO Diagnostics) using spectrophotometry (Multiskan MCC, Fisher Scientific). Insulin resistance (IR) was determined by utilizing the homeostasis model assessment (HOMA) by using the following formula: HOMA-IR = {[nonfasting glucose (in mmol/l)] × [nonfasting insulin (in mU/l)]}/22.5.

Measurement of superoxide by electron paramagnetic resonance spectroscopy.

Superoxide (O₂⁻) quantitation from the electron paramagnetic resonance (EPR) spectra was determined in the homogenates of 4–6 isolated coronary arterioles by a double integration of the peaks, with reference to a standard curve from horseradish peroxidase generation of the anion from a standard solutions of H₂O₂, using p-acetamidophenol as the cosubstrate (32, 45), and then normalized to the mean value of mLepr^db mice.

Treatment with TNF-α neutralization.

The neutralizing antibody to TNF (21) is 2E2 monoclonal antibody (2E2 MAb, 94021402, NCI Biological Resources Branch). At 12–16 wk of age, all mice received anti-TNF (2E2 MA, 0.625 mg·ml⁻¹·kg⁻¹·day⁻¹ ip for 3 days); the dosage was based on our estimates of TNF-α expression, which was estimated to be in the low nanogram or picogram range. This amount was able to neutralize 10–100 fold more than the estimated levels of TNF-α.

Treatment with MG-132, apocynin, NaSal, or SP-600125.

MG-132 tripeptide [Z-Leu-Leu-Leu-aldehyde (Sigma) dissolved in dimethyl sulfoxide (DMSO)] is a peptide aldehyde proteasome inhibitor. Ultimately preventing IκB-α degradation by the proteasome restricts NF-κB activity in vitro (13, 23, 24). Lepr^db mice were injected (10 mg·kg⁻¹·day⁻¹ ip for 3 days) (25). The oxidative stress was assessed by treating mice with the NAD(P)H oxidase inhibitor apocynin (100 mg·kg⁻¹·day⁻¹ ip for 3 days) (28). We administered NaSal (6 mg/ml in drinking water for 7–10 days, pH 7; Calbiochem) to Lepr^db and mLepr^db mice to determine whether the blocking of IKK-β affects their blood glucose level. To determine whether the blockage of JNK affects insulin receptor substrate-1 (IRS-1) activity, we also administered the JNK inhibitor SP-600125 (10 mg·kg⁻¹·day⁻¹ ip in 0.05 ml DMSO for 7 days; CalBiochem) to Lepr^db mice (3).

Functional assessment of isolated coronary arterioles.

The techniques for the identification and the isolation of coronary microvessels were previously described in detail (20, 45). Briefly, coronary arterioles (40–100 μm in diameter) from murine hearts were carefully dissected for in vitro study. Acetylcholine (ACh) is an endothelium-dependent vasodilator, and sodium nitroprusside (SNP) is endothelium-independent vasodilator. Flow-induced vasodilation is NO mediated, endothelial dependent, but agonist independent. To determine whether NF-κB or IKK-β played a role in endothelial injury in type 2 diabetes, ACh (0.1 nmol/l to 10 μmol/l)-, SNP (0.1 nmol/l to 10 μmol/l)-, and flow (4–60 cmH₂O)-induced vasodilation were assessed in the coronary arterioles in mLepr^db, Lepr^db, and Lepr^db mice treated with MG-132 or NaSal. Flow was established by the production of a pressure drop across the vessel and linearly related to the pressure drop (ΔP).

Insulin tolerance test to nonfasting mice.

The tail was nicked with a fresh razor blade by a horizontal cut of the tip, and the OneTouch Ultramini glucometer was used to measure baseline blood glucose. A total of 0.75 U/kg body wt of diluted insulin from porcine pancreas (Sigma) was injected into the intraperitoneal cavity. Blood samples were taken for glucose determinations at 0, 30, 60, and 120 min later. The blood was sampled from the tail of each mouse by gently massaging a small drop of blood onto the glucometer strip.

Protein expression by Western blot analyses.

Protein concentrations were assessed using a BCA Protein Assay Kit (Pierce), and equal amounts of protein (15 μg) were separated by SDS-PAGE and transferred to nitrocellulose (Hybond, Amersham) or polyvinylidene difluoride membranes (Pierce Biotechnology). IKK-α (Santa Cruz), IKK-β (Abcam), phosphorylated (p)-IKK-β (Cell Signal), IκB-α (Santa Cruz), p-IκB-α (Santa Cruz), NF-κB (Santa Cruz), TNF-α (Santa Cruz), N-Tyr (Abcam, an indicator for peroxynitrite-mediated tissue injury), JNK, p-JNK (Abcam), IRS-1, and p-IRS-1 (Millipore) antibodies were used to assess the protein expression and protein modification in mLepr^db, Lepr^db, and Lepr^db mice treated with anti-TNF (0.625 mg·ml⁻¹·kg⁻¹·day⁻¹ ip for 3 days), apocynin (100 mg·kg⁻¹·day⁻¹ ip for 3 days), NF-κB antagonist MG-132 (10 mg·kg⁻¹·day⁻¹ ip for 3 days), NaSal (6 mg/ml in drinking water), or SP-600125 (10 mg·kg⁻¹·day⁻¹ ip for 7 days). Our results were presented as fold changes normalized to the internal control (β-actin), compared with normal control mice (defined as 100 for control). Signals were visualized by enhanced chemiluminescence (Santa Cruz), scanned with a Fuji LAS3000 densitometer, and quantified by Multigauge software (Fuji film). The relative amounts of protein expression were quantified and normalized to those of the corresponding internal reference and β-actin and then normalized to corresponding mLepr^db control, which were set to a value of 1.0.

Data analysis.

At the end of each experiment, the vessel was relaxed with 100 μmol/l SNP to obtain its maximal diameter at 60 cmH₂O intraluminal pressure (45). All diameter changes to pharmacological agonists were normalized to the control diameter and expressed as a percentage of control diameters. All data are presented as means ± SE, except the protein expression that is presented as means ± SD. Statistical comparisons of vasomotor responses under various treatments were performed with two-way ANOVA, and intergroup differences were tested with Bonferonni inequality. Significance was accepted at P < 0.05.

RESULTS

Body weight, abdominal girth, serum concentration of glucose, cholesterol, insulin, and insulin resistance.

Body weight and serum parameters were measured at 12–16 wk for the different strains of mice (Table 1). Lepr^db mice were treated with NaSal for 7–10 days, and serum was collected from nonfasting mice in the morning. Nonfasting measurements were required since the Lepr^db murine model was intolerant to a fasting protocol.

Table 1.

Baseline serum parameters

	mLeprdb	Leprdb	mLeprdb + NaSal	Leprdb + NaSal
Body weight, g	25.82±1.5	50.53±3.17*	24.1±0.6	47.81±4.7*
Abdominal girth, cm	8.05±0.5	11.9±0.4*	8.3±0.4	11.75±0.6*
Glucose (nonfasting), mg/dl	139±17.26	442±52.38*	122.25±16	234.5±57.08†
Cholesterol level, μg/μl	1.12±0.23	1.8±0.4*	1.24±0.32	1.8±0.25*
Insulin (nonfasting), ng/ml	2.234±0.73	2.58±0.52	2.8±0.45	2.8±0.24
HOMA-IR	18.3±6.1	67.5±12.3*	20.3±4.4	38.9±9.2†

Values are means ± SD; n = 8 mice. Cholesterol level was higher in Lepr^db (homozygote, diabetic) and Lepr^db mice treated with sodium salicylate (NaSal) than in mLepr^db (heterozygote, normal) mice. Glucose concentration was higher in Lepr^db, but the treatment of NaSal attenuated blood glucose in Lepr^db mice vs. mLepr^db mice. Abdominal girth and body weight were higher in Lepr^db and Lepr^db mice treated with NaSal vs. mLepr^db on the day of surgery, but weight loss in diabetic mice was 2.3 ± 2.5 mg in Lepr^db mice treated with NaSal in drinking water (6 mg/ml for 7–10 days). Nonfasting insulin level was identical in Lepr^db and mLepr^db mice before and after treatment with NaSal; however, homeostasis model assessment-insulin resistance (HOMA-IR) was higher in diabetic Lepr^db mice vs. control, and NaSal significantly attenuated HOMA-IR in Lepr^db mice.

^*P < 0.05 vs. mLepr^db control mice;

^†P < 0.05 vs. Lepr^db.

Roles of IKK-α and IKK-β in type 2 diabetes.

To evaluate the roles of IKK-α and IKK-β in diabetes, Lepr^db mice were treated with anti-TNF, apocynin, and MG-132, which plays a role in the selective proteolysis of IκB-α (13, 23, 24). The protein expression of IKK-α and IKK-β was increased in Lepr^db mice (Fig. 1). IKK-β expression was significantly attenuated in Lepr^db mice treated with MG-132 or apocynin, or in db^TNF−/db^TNF− mice versus Lepr^db mice. However, the alteration of IKK-α expression was not observed in Lepr^db mice treated with a neutralizing antibody to TNF-α or in db^TNF−/db^TNF− mice. These findings suggest a link among IKK-β, NF-κB, oxidative stress, and TNF-α in type 2 diabetic mice.

Fig. 1.

Western blot analysis of protein expression of IKK subunit from heart extract. A: protein expression of IKK-α was slightly higher in Lepr^db (homozygote, diabetic) than in mLepr^db (heterozygote, normal) mice. Lepr^db mice treated with anti-TNF, apocynin, MG-132, and the genetic deletion of TNF-α in diabetic mice (db^TNF−/db^TNF−) did not show significantly changes in IKK-α expression. B: protein expression of IKK-β was higher in Lepr^db than in mLepr^db mice. db^TNF−/db^TNF− mice and Lepr^db treatment with MG-132 or antioxidant apocynin significantly decreased IKK-β expression in Lepr^db mice. Data represent means ± SD; n = 4 separate experiments. *P < 0.05 vs. mLepr^db; #P < 0.05 vs. Lepr^db.

Activation of IKK-β, NF-κB, and degradation of IκB-α induced by TNF-α.

The phosphorylation of IKK-β activates IκB-α to decrease the NF-κB inhibitory action of IκB-α. In Lepr^db mice, the protein modification of p-IKK-β was higher, but the protein modification of p-IKK-β was attenuated by anti-TNF, MG-132, apocynin, or the genetic deletion of TNF-α in diabetic mice (Fig. 2). In Lepr^db mice, the protein expression of IκB-α was lower versus mLepr^db mice (Fig. 3A). However, the protein modification of p-IκB-α was higher in Lepr^db versus mLepr^db mice. The protein modification of p-IκB-α was attenuated in db^TNF−/db^TNF− and Lepr^db mice treated with anti-TNF, apocynin, or MG-132 versus mLepr^db mice (Fig. 3B). In Lepr^db mice, the protein expression of NF-κB was higher, but apocynin, MG-132, or the genetic deletion of TNF-α presented a reduced protein expression of NF-κB (Fig. 3C).

Fig. 2.

Protein modification of phosphorylated (p)-IKK-β (A) and ratio of p-IKK-β to IKK-β (B) was higher in Lepr^db vs. mLepr^db mice heart extract. Anti-TNF, apocynin, NF-κB inhibitor MG-132, and genetic deletion of TNF-α attenuated phosphorylation of IKK-β in Lepr^db mice. Data represent means ± SD; n = 4 separate experiments. *P < 0.05 vs. mLepr^db; #P < 0.05 vs. Lepr^db.

Fig. 3.

Protein expression of IκB-α (A) was higher in mLepr than in Lepr mice heart extract. In contrast, p-IκB-α (B) was higher in Lepr^db vs. mLepr^db mice. Anti-TNF, antioxidant apocynin, NF-κB inhibitor MG-132, and genetic deletion of TNF-α in Lepr^db mice attenuated the phosphorylation of IκB-α, thereby reducing the protein expression of NF-κB (C) as well. Data represent means ± SD; n = 4 separate experiments. *P < 0.05 vs. mLepr^db; #P < 0.05 vs. Lepr^db.

Role of oxidative stress and NF-κB activation in type 2 diabetes-induced vascular dysfunction.

The vasodilation to the endothelium-dependent vasodilator ACh was significantly impaired in Lepr^db versus mLepr^db mice (Fig. 4A), but NF-κB antagonist MG-132 partially restored NO-mediated coronary arteriolar dilation in Lepr^db mice. The EPR results showed that O₂⁻ production was elevated and that MG-132 decreased O₂⁻ production in Lepr^db versus mLepr^db mice (Fig. 4B).

Fig. 4.

A: Lepr^db mice impaired but MG-132 treatment significantly restored ACh-induced endothelium-dependent vasodilation in Lepr^db mice. Data represent means ± SE; n = 4 mice. B: O₂⁻ production from isolated coronary arterioles was higher in Lepr^db mice vs. mLepr^db mice, and MG-132 attenuated O₂⁻ production in Lepr^db mice. Values were normalized to the mean value of mLepr^db. Data represent means ± SD; n = 5 separate experiments. *P < 0.05 vs. mLepr^db; #P < 0.05 vs. Lepr^db.

Effect of NaSal on improvement of insulin sensitivity in the heart.

Blood glucose, body weight, abdominal girth, lipid levels, insulin levels, and HOMA-IR were higher in Lepr^db versus mLepr^db mice. There were no significant differences in body weight, abdominal girth, lipid level, and insulin level before and after treatment with NaSal (See Table 1). Lower values of glucose concentration and HOMA-IR were found in Lepr^db mice treated with NaSal. The average body weight loss of Lepr^db mice was 2.3 ± 2.5 mg after 7 days administration with 6 mg/ml NaSal in drinking water. Insulin tolerance testing (Fig. 5) showed a significant difference in the glucose clearance rate at 120 min after insulin administration in Lepr^db mice treated with NaSal compared with Lepr^db mice.

Fig. 5.

Insulin was injected (0.75 U/kg ip) and blood samples were taken for glucose determinations 0, 30, 60, and 120 min later. A: insulin tolerance test (ITT) showed that the glucose clearance rate at 120 min after insulin administration in Lepr^db mice treated with sodium salicylate (NaSal) was significantly higher compared with Lepr^db mice. The value was normalized to the basal glucose level. B: actual glucose level. Each value represents the mean ± SE from 6–8 mice. *P < 0.05 vs. mLepr^db; #P < 0.05 vs. Lepr^db.

Role of NaSal in vascular dysfunction in type 2 diabetes.

To establish the role of IKK-β in the signaling pathway, we studied whether the blockade of IKK-β improves the endothelium-dependent vasodilation in Lepr^db mice. Functional results (Fig. 6) showed that NaSal partially restored the ACh- and flow-induced NO-mediated vasodilation in Lepr^db mice. SNP-induced vasodilation was equivalent in mLepr^db and Lepr^db mice, indicating that the function of smooth muscle was preserved.

Fig. 6.

A: sodium nitroprusside-induced dilation in coronary arterioles was identical between mLepr^db and Lepr^db mice. NaSal did not affect sodium nitroprusside-induced vasodilation in Lepr^db mice. B: ACh-induced vasodilation in isolated mice coronary arterioles was blunted in Lepr^db vs. mLepr^db mice, and NaSal partially restored this dilation in Lepr^db mice. C: NaSal partially restored flow-induced vasodilation in Lepr^db mice. Data represent means ± SE; n = 5 mice. *P < 0.05 vs. mLepr^db; #P < 0.05 vs. Lepr^db. ΔP, change of pressure.

Insulin resistance enhanced activity of JNK and IKK-β in type 2 diabetes.

Western blot analysis (Fig. 7) showed that the protein expression of TNF-α and NF-κB and the protein modification of p-IKK-β and p-JNK were greater in Lepr^db mice, but NaSal attenuated the protein expression of TNF-α and NF-κB and the protein modification of the phosphorylation of IKK-β and the phosphorylation of JNK in Lepr^db mice.

Fig. 7.

Protein modification of p-IKK-β (A) and p-JNK (B) was greater in Lepr^db mice vs. mLepr^db mice heart extract, but NaSal decreased p-IKK-β and p-JNK without affecting the expression of IKK-β and JNK. Protein expression of TNF-α (C) and NF-κB (D) was greater in Lepr^db mice vs. mLepr^db mice heart extract, but NaSal decreased TNF-α and NF-κB. Data represent means ± SD; n = 5 separate experiments. *P < 0.05 vs. mLepr^db; #P < 0.05 vs. Lepr^db.

The phosphorylation of IRS-1 on the serine site blocks insulin signal transduction in diabetes. The protein expression of IRS-1 (Fig. 8A) was significantly reduced in Lepr^db, Lepr^db treated with NaSal, and Lepr^db treated with JNK inhibitor SP-600125 versus mLepr^db mice. The protein modification of p-IRS-1 (Ser307) was unaffected (Fig. 8B) in Lepr^db, Lepr^db treated with NaSal, and Lepr^db mice treated with JNK inhibitor SP-600125 versus mLepr^db mice. However, the ratio of p-IRS-1 to IRS-1 was significantly higher in Lepr^db. Lepr^db mice treated with NaSal and Lepr^db mice treated with JNK inhibitor SP-600125 showed decreased p-IRS-1/IRS-1.

Fig. 8.

A: total insulin receptor substrate-1 (IRS-1) protein expression was decreased in Lepr^db vs. mLepr^db mice heart extract; NaSal and SP-600125 (SP) treatment did not increase IRS-1 expression. B: p-IRS-1 (Ser307) protein modification was elevated in Lepr^db compared with mLepr^db control mice heart extract, and both NaSal and SP-600125 treatment appeared to reduce p-IRS-1 protein modification although there was no significant difference between groups. C: ratio of p-IRS-1 to total IRS-1 (p-IRS-1/IRS-1) was significantly higher in Lepr^db vs. mLepr^db heart extract. NaSal and SP-600125 treatment reduced p-IRS-1/IRS-1. Data represent means ± SD; n = 8 mice. *P < 0.05 vs. mLepr^db; #P < 0.05 vs. Lepr^db.

DISCUSSION

Our results suggest that the interaction of NF-κB and TNF-α signaling induces the activation of IKK-β in insulin-resistant Lepr^db mice heart, which amplifies oxidative stress, leading to endothelial dysfunction in the coronary arterioles of Lepr^db mice. Our findings support the concept that a feed-forward signaling of TNF-α and NF-κB via the IKK-β pathway induces insulin resistance and coronary arteriolar dysfunction in type 2 diabetes based on the following observations: NF-κB antagonist MG-132, or IKK-β inhibitor NaSal, restored NO-mediated endothelium-dependent coronary arteriolar dilation in Lepr^db mice, but the responses in mLepr^db mice were unaffected. The protein expression of IKK-α and IKK-β were higher in Lepr^db than in mLepr^db mice; the expression of IKK-β, but not the expression of IKK-α, was attenuated by MG-132, antioxidant apocynin, or the genetic deletion of TNF-α in diabetic mice. Insulin resistance was increased in Lepr^db mice, and NaSal improved insulin sensitivity. The protein expression of TNF-α, NF-κB, and protein modification of p-IKK-β and p-JNK were greater in Lepr^db mice, but NaSal attenuated TNF-α, NF-κB, p-IKK-β, and p-JNK in Lepr^db mice. The p-IRS-1-to-IRS-1 ratio was elevated in Lepr^db compared with mLepr^db mice. Both NaSal and the JNK inhibitor SP-600125 reduced the p-IRS-1-to-IRS-1 ratio in Lepr^db mice. MG-132 (Fig. 4B) or anti-TNF-α reduced O₂⁻ production in Lepr^db mice. NF-κB induces TNF-α signaling to accentuate oxidative stress and endothelial dysfunction via an IKK-β-dependent mechanism, which may be associated with inflammatory and insulin signaling pathways in type 2 diabetes. Our findings are consistent with NF-κB involvement, by interacting with TNF-α, thereby inducing IKK-β and oxidative stress, in endothelial dysfunction in type 2 diabetes. The present molecular results further support our previous physiological observations that the increases in TNF-α expression induce the production of ROS, leading to endothelial dysfunction in type 2 diabetes (15).

Roles of interaction of TNF-α and NF-κB signaling in type 2 diabetes.

The inducible transcription factor NF-κB regulates the expression of genes encoding oxidants, cytokines, chemokines, and adhesion molecules, which are associated with inflammation and activated by the gene products of NF-κB, e.g., a feed-forward interaction (12, 15, 19, 30). These proinflammatory agents play a detrimental role in the vascular pathology, and TNF-α initiates signaling cascades that converge on the IKK complex via the NF-κB signaling pathway. TNF-α is an inducer of IKK-β and IκB-α in the classical NF-κB pathway. Thus TNF-α initiates signaling cascades, predominantly acting through IKK-β (6, 22). TNF-α induces oxidative stress by activating NAD(P)H oxidase, which is the primary source of O₂⁻ in vascular tissue. O₂⁻ is dismutased into oxygen and H₂O₂ by O₂⁻ dismutase (SOD) in mitochondria. Mitochondria-derived H₂O₂ diffuses through the cellular membrane and then is involved in the regulation of endothelial NF-κB by activating the IKK complex (10, 11, 37). In this study, we examined whether a feed-forward interaction between TNF-α and NF-κB, via the IKK-β pathway, contributes to the evolution of vascular disease and insulin resistance in type 2 diabetes. Our Western blot analysis results showed that MG-132, apocynin, or the genetic deletion of TNF-α attenuated the expression of IKK-β but not the expression of IKK-α. This points to a linkage among IKK-β, NF-κB, oxidative stress, and TNF-α in contributing to type 2 diabetes.

Our results showed that the protein expression of IκB-α was lower. However, the protein expression of p-IκB-α was higher in Lepr^db mice compared with mLepr^db mice. In db^TNF−/db^TNF− and Lepr^db mice treated with anti-TNF, apocynin, or MG-132, the protein expression of p-IκB-α was attenuated. In Lepr^db mice, the protein expression of NF-κB was higher, but apocynin, MG-132, or db^TNF−/db^TNF− mice had a reduced protein expression of NF-κB, indicating that TNF-α, oxidative stress, and IκB-α degradation may increase the activity of NF-κB. NF-κB contributes to the endothelial inflammation in ROS-dependent activation of the IKK complex (11), and this supports the concept that TNF-α induces oxidative stress by activating NAD(P)H oxidase, which then activates NF-κB.

Our functional results showed that ACh-induced vasodilation was impaired in Lepr^db mice versus mLepr^db mice, but NF-κB antagonist MG-132 partially restored ACh-induced, NO-mediated coronary arteriolar dilation in Lepr^db mice. EPR results showed that O₂⁻ production was elevated and that MG-132 decreased O₂⁻ production in Lepr^db mice, which indicates that the blocking of NF-κB transcriptional action markedly attenuated O₂⁻ production in isolated coronary arterioles in Lepr^db mice. This supports the concept that IKK-β is a putative activator of NF-κB, which is thought to alter subsequent gene expression in this pathway (12, 19, 30).

Role of IKK-β in vascular dysfunction in type 2 diabetes.

NaSal is an IKK-β inhibitor that interrupts the cascade of molecular events leading to inflammation. NaSal inhibited the expression of inducible NO synthase, which causes endothelial dysfunction and inflammation in cardiac fibroblasts (40) and TNF-α production, which is dependent on NF-κB activity, in lipopolysaccharide-stimulated macrophage (39). NaSal may interfere with NF-κB activity through a mitogen-activated protein kinase-dependent process (39).

Yuan et al. (43) showed that reduced signaling through the IKK-β pathway inhibition by NaSal in obese mice is accompanied by improved insulin sensitivity. Recent studies noted that NF-κB blockade by inhibiting IKK-β activity decreased myocardial injury and preserved cardiac function following ischemia-reperfusion (26), which supports the concept that the inflammatory response participates in the development of heart failure (42). Our results show that p-IKK-β increases in diabetes, but the blockade of NF-κB and oxidative stress and the genetic deletion of TNF-α attenuated p-IKK-β, suggesting that the phosphorylation of IKK-β increases in diabetic mice by the activation of oxidative stress and TNF-α.

The pathogenesis of endothelial dysfunction in diabetes is initially related to a decrease in NO synthesis or the inactivation of NO due to an increased endothelial production of ROS (31). An underlying mechanism proposed for TNF-α-induced endothelial dysfunction is that TNF-α signaling leads to oxidative stress via NAD(P)H activation, which in turn may lead to reduced NO bioactivity (15, 28). We identified a link between vascular dysfunction and insulin resistance: the treatment with NaSal in coronary microcirculation resulted in an improvement of endothelial dysfunction in diabetic Lepr^db mice. The NO donor SNP induced an identical vasodilation in Lepr^db and mLepr^db mice, but ACh-induced vasodilation was impaired in isolated coronary arterioles in diabetic Lepr^db mice, which indicates that endothelial-dependent and NO-mediated vasodilation is impaired in diabetes. Most importantly, NF-κB antagonist MG-132 or IKK-β inhibitor NaSal partially restored NO-mediated coronary arteriolar dilation in type 2 diabetes. These results further support our hypothesis that TNF-α and NF-κB signaling via an IKK-β-dependent mechanism plays a key role in endothelial dysfunction in diabetes.

Blood glucose, body weight, abdominal girth, lipid level, insulin level, and HOMA-IR were higher in Lepr^db versus mLepr^db mice. However, the lower values of glucose concentration and HOMA-insulin resistance found in Lepr^db mice treated with NaSal indicate that NaSal improved insulin sensitivity in Lepr^db mice (Table 1). MG-132 or anti-TNF antibody did not change blood glucose or body weight. Insulin tolerance testing (Fig. 5) shows that there was a significant difference in the glucose clearance rate at 120 min after insulin administration in Lepr^db mice treated with NaSal compared with Lepr^db mice. Our results indicate that hyperglycemia and insulin resistance in Lepr^db mice were amended by a short-term treatment with NaSal via a downregulation of IKK-β and NF-κB signaling. An increased phosphorylation of IKK-β may also contribute to the development or progression of vascular disease and hyperglycemia in type 2 diabetes.

Link between inflammation-induced insulin resistance and vascular dysfunction.

The heart is an insulin-responsive organ, and disorders of insulin action, such as diabetes and obesity, can have profound effects on cardiac performance (1). Coronary artery disease causes acute myocardial infarction, which is associated with significant morbidity and mortality in patients with diabetes (9). The inflammation-induced insulin resistance, but genetic disruption of NF-κB and JNK signaling pathways, has been shown to improve insulin resistance (33). Human type 2 diabetes is currently characterized by defects in both insulin action and insulin secretion, which lack the focus needed to identify a single molecular abnormality underlying these features, if indeed one exists. IRS proteins may be involved in type 2 diabetes (41) and tyrosine phosphorylation activates insulin IRS-1, which further leads to the translocation of glucose transporters (GLUT4) to the cell surface (38). IKK-β may affect the glucose metabolism by altering NF-κB transcriptional action to express GLUT4 (4). JNK is a main regulatory molecule that contributes to coronary arteriolar dysfunction and insulin resistance (44, 35). Whereas IRS-1 tyrosine phosphorylation by Akt induces insulin signal transduction (18), serine phosphorylation by JNK can downregulate or inactivate IRS-1. The impaired activation of Akt and the enhanced activation of JNK by TNF-α are correlated with insulin resistance (2, 17, 35). Although NaSal had no obvious effect on the activation of IKK-β or JNK, the reduced signaling by inhibiting IKK-β in obese mice is accompanied by an improved insulin sensitivity (43). Consistent with the findings described above, the inhibition of IKK-β or JNK activity by NaSal significantly reduced the blood glucose level and augmented insulin signaling in cardiac tissue. Our results showed that the protein expression of TNF-α and NF-κB were higher in diabetic mice but that NaSal attenuated the protein expression of TNF-α and NF-κB. Specifically, NaSal attenuated the phosphorylation of IKK-β and JNK without altering the protein expression of IKK-β and JNK. This suggests that NaSal interrupts the phosphorylation of IKK-β and the phosphorylation of JNK. TNF-α activates the IKK-β (Fig. 1B) and JNK signaling pathways (44), and, conversely, the expression of the TNF-α gene is regulated by NF-κB through the activation of IKK-β (Fig. 7D). NaSal attenuates TNF-α; however, the complete role of NaSal in this process remains to be elucidated.

To further elucidate the signaling pathway of defective insulin action in the cardiac tissue of Lepr^db mice, we tested whether IRS-1 serine phosphorylation is attenuated by SP-600125 and NaSal. p-IRS-1 protein modification was elevated in Lepr^db compared with mLepr^db mice; both NaSal and JNK inhibitor SP-600125 reduced the p-IRS-1 protein modification in Lepr^db mice (Fig. 8). Cardiac myocyte metabolism occurs through paracrine mechanisms by activating insulin receptors in the heart (5), thereby decreasing myocardial infarct size and increasing cardiac contractility (8, 29). Because insulin signaling influences numerous functions within the heart, an altered insulin action can have many significant indirect effects on the heart that may lead to vascular dysfunction. Our results indicate that TNF-α, the phosphorylation of JNK and IKK-β, and NF-κB contribute to insulin resistance and vascular dysfunction putatively via the activation of IKK-β in type 2 diabetes. NaSal and JNK inhibitor SP-600125 sensitize insulin signaling by preventing serine phosphorylation on the IRS-1 of IKK-β and JNK. This indicates that NaSal inhibits the phosphorylation of IKK-β and JNK in insulin-resistant diabetic murine hearts.

In summary, our results demonstrate that TNF-α is correlated with IKK-β to induce NF-κB activation. IKK-β inhibitor NaSal or NF-κB antagonist MG-132 restored ACh-induced endothelium-dependent vasodilation in coronary arterioles in diabetes. Diabetic mice treated with NaSal show a decrease in glucose level and a reversal of insulin resistance. Both IKK-β and JNK activation are increased in Lepr^db mice, and NaSal attenuates their activation. Overall, our molecular and functional results show that IKK-β is a crucial component of the biochemical pathway responsible for vascular dysfunction, inflammation, and insulin resistance in type 2 diabetic murine hearts. The blockade of IKK-β activity not only preserves coronary arteriolar vasodilation but also prevents insulin resistance in type 2 diabetic mice. Although the current study confirms that NaSal amends whole body insulin sensitivity and demonstrates that IKK-β and NF-κB signaling is associated with insulin resistance and inflammation in diabetic murine hearts, further validation is needed to show how a direct insulin signaling in coronary arterioles affects vessel function.

GRANTS

This study was supported by an American Heart Association Scientist Development Grant 110350047A; a Pfizer Atorvastatin Research Award 2004-37; and National Heart, Lung, and Blood Institute Grants RO1-HL-077566 and RO1-HL-085119 (to C. Zhang).