Abstract

Activation of nuclear factor-κB (NF-κB) occurs by dissociation from IκB after serine or tyrosine phosphorylation of IκBα, but the way of NF-κB activation by high glucose has not been defined. High glucose is known to activate NF-κB via protein kinase C and reactive oxygen species (ROS). In this study, we investigated how high glucose activates NF-κB for CC chemokine ligand 2 production in cultured human glomerular endothelial cells. High glucose increased nuclear translocation of p65 and also increased NF-κB DNA binding activity. High glucose-induced NF-κB activation occurred without degradation of IκBα. In agreement with this, there was no increase in serine phosphorylation of IκBα, while tyrosine phosphorylation of IκBα was increased by high glucose. High glucose increased the generation of ROS, whereas both α-lipoic acid and N-acetylcysteine scavenged the ROS and decreased high glucose-induced tyrosine phosphorylation of IκBα, nuclear translocation of p65, and NF-κB DNA binding activity. Protein kinase C pseudosubstrate inhibited high glucose-induced ROS production, tyrosine phosphorylation of IκBα, and nuclear translocation of p65. Both BAY 61-3606, a specific inhibitor of Syk protein-tyrosine kinase, and small interfering RNA directed against Syk inhibited high glucose-induced tyrosine phosphorylation of IκBα as well as p65 nuclear translocation. High glucose increased tyrosine phosphorylation of Syk, while it was inhibited by α-lipoic acid and protein kinase C pseudosubstrate. In summary, high glucose-induced NF-κB activation occurred not by serine phosphorylation of IκBα. Our data suggest that ROS-mediated tyrosine phosphorylation of IκBα is the mechanism for high glucose-induced NF-κB activation, and Syk may play a role in tyrosine phosphorylation of IκBα.

  • CC chemokine ligand 2
  • protein kinase C
  • reactive oxygen species

the transcriptional factor nuclear factor-κB (NF-κB) is implicated in the development of diabetic nephropathy by mediating high glucose-induced cytokine production. High glucose-induced CC chemokine ligand 2 (CCL2; monocyte chemoattractant protein-1) production is such an example. In the studies of cultured glomerular mesangial cells (8, 11), high glucose was shown to induce CCL2 production, in which NF-κB was the responsible transcriptional factor (8). Secreted CCL2 in vivo may cause macrophage recruitment and activation, resulting in diabetic renal injury (30).

The major form of NF-κB is composed of a dimer of p50 and p65 subunits and is sequestered in the cytoplasm through its tight association with specific inhibitory proteins (IκB) (7). Thus NF-κB activation requires dissociation from IκB. This occurs by serine or tyrosine phosphorylation of IκBα. Serine phosphorylation leads to the degradation of IκBα by the ubiquitin-proteasome complex (4, 25). In contrast, tyrosine phosphorylation activates NF-κB without degradation of IκBα (3, 12). After dissociation from IκB, NF-κB translocates to the nucleus and binds to a specific sequence in DNA, which in turn results in gene transcription. The way of NF-κB activation was reported to be different according to the stimuli. Tumor necrosis factor-α (TNF-α) activates NF-κB through phosphorylation of serines 32 and 36 of IκBα (20), while nerve growth factor activates NF-κB through tyrosine phosphorylation of IκBα (3). So far, how high glucose activates NF-κB has not been determined.

In rat mesangial cells, high glucose-induced NF-κB activation was mediated by protein kinase C (PKC) and reactive oxygen species (ROS) (8). α-Lipoic acid (α-LPA) is endogenously synthesized from ocatanoic acid, which is derived from acetyl-CoA (23). It has powerful antioxidant effects by direct radical scavenging and metal chelating, and by increasing intracellular glutathione (23). Dietary supplementation of α-LPA was shown to prevent or ameliorate diabetic nephropathy in an animal model (2, 18). It also prevented aggravation of proteinuria in a study of diabetic patients (21).

Glomerular endothelial cells are in direct contact with high glucose in the blood of diabetic patients. In the present study of cultured human glomerular endothelial cells (HGECs), we evaluated the effects of α-LPA on high glucose-induced activation of NF-κB and CCL2 production, and then investigated how high glucose activates NF-κB.

MATERIALS AND METHODS

Cell culture.

HGECs were isolated from the normal portion of the kidney tissues resected due to renal cell carcinoma after informed consent was obtained, as described previously (24). In brief, the renal cortex was minced and passed sequentially through 250-, 200-, and 150-μm sieves and suspended in media composed of Waymouth MB 752/1 (GIBCO BRL, Grand Island, NY) supplemented with 20% FCS, fetuin (100 μg/ml, GIBCO BRL), glutamine (2 mM, GIBCO BRL), and antibiotics. After centrifugation, the pellet was treated with collagenase type III (Worthington Biomedical), and filtered sequentially through 100- and 50-μm sieves. Glomerular segments retained by the 50-μm filter were collected and plated onto fibronectin-coated 60-mm culture dishes (Corning Life Sciences, Acton, MA) in culture medium containing heparin (100 μg/ml, Sigma, St. Louis, MO) and endothelial cell growth factor (200 μg/ml, Sigma). After ∼2 wk, the cells were transferred to a T-25 flask (Corning Life Sciences) coated with fibronectin. To prevent epithelial cell contamination, the cells were treated with puromycin (10 μg/ml, Sigma) for 24 h.

Endothelial cells were identified by immunohistochemical staining with rabbit antihuman factor VIII antibody (Dako, Glostrup, Denmark). The experiments were performed using cells between passages 4 and 8.

Before each experiment, cells were rested for 16 h in DMEM (GIBCO BRL) containing 2% FCS and 5 mM glucose. The effects of high glucose and α-LPA were evaluated by placing the cells in DMEM containing 2% FCS, 5 mM glucose, 30 mM glucose, or 30 mM glucose plus α-LPA (50 μM, Sigma). This dose of α-LPA was selected on the basis of published data and our preliminary study. In a study of human aortic endothelial cells (34), α-LPA was used at 50–500 μM for 48 h without alteration of cell viability. In our preliminary study, α-LPA was effective and not toxic at 50 μM. In a separate experiment, N-acetylcysteine (5 mM, Sigma) was used in place of α-LPA to evaluate whether another antioxidant has similar effects.

CCL2 ELISA.

Serum-starved cultured HGECs were incubated with 5 mM glucose, 30 mM glucose, or 5 mM glucose plus 25 mM mannitol for 6, 24, or 72 h. Mannitol (Sigma) was used as an osmotic control for glucose. The supernatants were collected, and CCL2 released in cell culture supernatant was quantified by sandwich ELISA (Quantikine human MCP-1, R&D Systems, Abingdon, UK) according to the manufacturer's instructions. Each experiment was performed in duplicate.

Northern blot analysis of CCL2 mRNA.

The cDNA probe specific for human CCL2 was made by RT-PCR, as described previously (16). Serum-starved HGECs grown to confluence in 100-mm fibronectin-coated culture dishes (Corning Life Sciences) were left untreated or pretreated with α-LPA, N-acetylcysteine, PKC inhibitors, or NF-κB activation inhibitors for 30 min and then incubated with glucose (5 or 30 mM) or PMA (EMD Chemicals, Darmstadt, Germany). PKC inhibitors included myristoylated PKC[19–27] pseudosubstrate (Myr-Phe-Ala-Arg-Lys-Gly-Ala-Leu-Arg-Gln-OH, Biosource, Camarillo, CA), calphostin C (EMD Chemicals), and GF109203X (Alexis Biochemicals, San Diego, CA). NF-κB inhibitors included NF-κB decoy oligodeoxynucleotides (ODNs), NF-κB activation inhibitor I(6-amino-4-4-phenoxyphenylethylaminoquinazoline) (EMD Chemicals) and NF-κB activation inhibitor II, 4-methyl-N1-(3-phenylpropyl)benzene-1,2-diamine (EMD Chemicals). At the indicated time points, total cellular RNA was extracted with a Tri reagent kit (Molecular Research Center, Cincinnati, OH). Equal amount of RNA (10 μg/lane) was electrophoresed through 1% agarose, 2.2 M formaldehyde denaturing gel in MOPS buffer, followed by capillary transfer to nylon membranes. The membranes were stained with methylene blue to check the integrity, uniformity of loading, and transfer of RNA. After being fixed at 80°C for 2 h, the membranes were hybridized with 32P-labeled cDNA probe for CCL2 and 100 μg/ml salmon sperm DNA overnight at 65°C. The 28S and 18S ribosomal RNA bands of the membranes stained with methylene blue were used to control the RNA loading and transfer efficiency.

Preparation of decoy ODNs.

The phosphorothioate double-stranded ODNs against the NF-κB binding site and the mismatched ODNs were prepared by Bioneer (Daejeon, Korea), based on the literature (29). The sequences of ODNs are as follows: NF-κB decoy ODN, 5′-AGTTGAGGGGACTTTCCCAGGC-3′, and mismatched NF-κB decoy ODN, 5′-AGTTGAGGCGACTTTCCCAGGC-3′. The double-stranded ODNs were prepared from complementary single-stranded phosphorothiolate-bonded oligonucleotides.

Transfection of decoy ODNs or small interfering RNA.

Transfection of ODNs was performed to cells grown in 100-mm dish to 80% confluence. DNA was precomplexed with PLUS reagents (Life Technologies, Rockville, MD) at room temperature for 15 min and then mixed and incubated with diluted lipofectamine reagent (Life Technologies) for 15 min at room temperature. Thereafter, the DNA-PLUS Lipofectamine reagent complex was added to each dish containing fresh medium and incubated at 37°C at 5% CO2. After 6-h incubation, complete medium with serum for normal growth was added. After a 2-day transfection period, the cells were subjected to the experiment for Northern blotting.

Transfections of small interfering RNA (siRNA) directed against Syk (Syk-siRNA) or a nonspecific, scrambled, control siRNA (control-siRNA; Santa Cruz Biotechnology) were also performed using Lipofectamine reagent. After transfection of 60 pmol siRNA using 10 μl of Lipofectamine 2000, the cells were maintained with complete medium for 3 days and then serum-starved for 24 h and subjected to the experiment for Western blotting.

Western blot analysis.

Serum-starved HGECs were left untreated or pretreated with α-LPA, N-acetylcysteine, genistein (EMD Chemicals), herbimycin A (EMD Chemicals), PP2 (EMD Chemicals), or BAY 61-3606 (EMD Chemicals) for 30 min and then incubated with glucose, TNF-α (R&D Systems), or PMA. For the experiment requiring inhibition of proteasome, the cells were pretreated with ALLN (calpain inhibitor I, Sigma). Cytosolic fractions were separated using pipettes after centrifugation. The remaining nuclear fractions were lysated again with lysis buffer. After repeated freezing and thawing with liquid nitrogen and a water bath set to 37°C, the nuclear lysates were incubated at 4°C for 20 min with a rocking platform. Finally, nuclear protein was obtained after centrifugation. Equal amounts of protein extracts were separated by 12% SDS-PAGE and transferred to an Immobilon-P transfer membrane (Millipore, Bedford, MA). The membrane was probed with a rabbit IgG antibody directed against human p65 or IκBα or an anti-phospho-IκBα (Ser-32) antibody (Santa Cruz Biotechnology). Bands were visualized using horseradish peroxidase-conjugated anti-rabbit IgG (Santa Cruz Biotechnology) and an enhanced chemiluminescence agent (Amersham) according to the manufacturer's instruction.

Immunoprecipitation.

Equal amounts (100 μg) of whole cell lysates were immunoprecipitated by addition of antibody to IκBα or Syk (Cell Signaling Technology, Danvers, MA). Immune complexes were recovered by addition of protein A-agarose (Roche Applied Science, Indianapolis, IN) and analyzed by Western blotting with an anti-phosphotyrosine (4G10) antibody (Upstate USA, Chicago, IL).

EMSA.

HGECs were left untreated or pretreated with α-LPA or N-acetylcysteine for 30 min and then incubated with 5 or 30 mM glucose. At the end of indicated incubation periods, nuclear extract was prepared. NF-κB consensus oligonucleotides (5′-AGTTGAGGGGACTTTCAGGA-3′) (Promega, Madison, WI) were end-labeled with [γ-32P]ATP (50 μCi at 3,000 Ci/mmol, New England Nuclear, Boston, MA) with T4 polynucleotide kinase. Nuclear protein (10 μg) was allowed to react with 32P-labeled NF-κB oligonucleotides (1.75 pmol) in 25 μl of binding buffer. Nucleoprotein-oligonucleotide complexes were resolved by electrophoresis on a 6% nondenaturing polyacrylamide gel. Thereafter, the gel was dried and DNA-protein complexes were localized by autoradiography for 6∼12 h. For the competitive assay, 100 times excess unlabeled NF-κB oligonucleotides were added in the reaction mixture before addition of the 32P-labeled NF-κB oligonucleotides. In the supershift assay, the nuclear protein was incubated with anti-p65 antibody (Santa Cruz Biotechnology) before the binding reaction.

Measurement of ROS.

The production of intracellular ROS was evaluated with the probe 5-(and-6)-chloromethyl-2′-7′-dichlorodihydrofluorescein diacetate (CM-H2DCF-DA; Molecular Probes, Eugene, OR). This reagent enters into cells and reacts with ROS to generate the fluorescent product 2′-7′-dichlorofluorescein. HGECs grown in 24-well plates were serum-starved for 16 h and then were left untreated or pretreated with α-LPA (50 μM), N-acetylcysteine, or PKC inhibitors for 30 min and further incubated with 5 or 30 mM glucose, TNF-α, or PMA. Thereafter, the cells were incubated with 5 μM H2DCF-DA for 10 min at 37°C. After removal of the media and washing of the cells, the expression of intracellular ROS was visualized with a LEICA DM-IRE2 inverted microscope (Leica Microsystems, Wetzlar, Germany) equipped with the Leica TCS-SP2 confocal system (excitation 488 nm, emission 520 nm).

Statistical analysis.

Data are presented as means ± SE, with n representing the number of different experiments. Comparisons of the values among groups were performed by ANOVA and Scheffé's test. P < 0.05 was considered statistically significant.

RESULTS

High glucose-induced CCL2 production: inhibition by α-LPA.

The morphology of HGECs isolated and used in this study is shown in Fig. 1. Virtually all of the cells were positive for factor VIII in the immunohistochemical staining. The cells were exposed to 5, 30, or 5 mM glucose plus 25 mM mannitol as an osmotic control for 6, 24, or 72 h. The supernatants were collected and assessed for CCL2 production. High glucose significantly increased CCL2 production compared with control glucose or mannitol (582 ± 39 vs. 904 ± 81 vs. 490 ± 39 pg·ml−1·104 cells−1 at 24 h, P < 0.05; 1,995 ± 148 vs. 2,738 ± 183 vs. 1,595 ± 141 pg·ml−1·104 cells−1 at 72 h, P < 0.05, n = 6, each n is in duplicate) (Fig. 2A). α-LPA inhibited basal (1,227 ± 35 vs. 388 ± 23 pg·ml−1·104 cells−1, P < 0.05, n = 6) or high (2,856 ± 95 vs. 554 ± 37 pg·ml−1·104 cells−1, P < 0.05, n = 6) glucose-induced CCL2 production (Fig. 2B). In agreement with this, high glucose increased CCL2 mRNA, while α-LPA downregulated it (Fig. 3).

Fig. 1.

Identification of human glomerular endothelial cells (HGECs). A: phase-contrast photomicrograph of HGECs (×100). B: in immunohistochemical staining, virtually all of the cells were positive for factor VIII (×200).

Fig. 2.

Effect of α-lipoic acid (α-LPA) on high glucose-induced CC chemokine ligand 2 (CCL2) production. Cultured HGECs were serum starved for 16 h and incubated with 5, 30, or 5 mM glucose plus 25 mM mannitol for the indicated times (A), or with 5 or 30 mM glucose in the presence or absence of α-LPA (50 μM) for 72 h (B). The supernatants were collected, and CCL2 was measured by ELISA. (n = 6). *P < 0.05 compared with 5 mM glucose. **P < 0.05 compared with 30 mM glucose.

Fig. 3.

Effect of α-LPA on high glucose-induced CCL2 mRNA expression. Serum-starved HGECs were left untreated or pretreated with α-LPA (50 μM) for 30 min and then incubated with 5 or 30 mM glucose for 12, 24, or 72 h. Total RNA was extracted, and CCL2 mRNA level was measured by Northern hybridization. The 28S and 18S ribosomal RNA bands on the methylene blue-stained membrane were used to control the RNA loading. The result shown is representative of 4 independent experiments.

NF-κB in high glucose-induced CCL2 mRNA expression and effect of α-LPA.

In the following experiments, we evaluated whether NF-κB mediates high glucose-induced CCL2 gene transcription and the effect of α-LPA on it. Both NF-κB activation inhibitor I (50 nM) and II (30 μM) inhibited high glucose-induced CCL2 mRNA expression (Fig. 4A). In addition, transfection of NF-κB decoy ODNs also significantly inhibited high glucose-induced CCL2 mRNA expression, while mismatched NF-κB decoy ODNs had no effect (Fig. 4B). In the EMSA, high glucose increased DNA binding activity of NF-κB, whereas α-LPA downregulated high glucose-induced DNA binding activity of NF-κB (Fig. 5).

Fig. 4.

Downregulation of high glucose-induced CCL2 mRNA expression by NF-κB inhibitors and NF-κB decoy oligodeoxynucleotides. A: serum-starved HGECs were left untreated or pretreated with NF-κB inhibitor I (50 nM) or II (30 μM) for 30 min and then incubated with 5 or 30 mM glucose for 72 h. B: cultured HGECs were transfected with NF-κB decoy oligodeoxynucleotides or mismatched oligodeoxynucleotides. After serum starvation, the cells were incubated with high glucose (30 mM) for 72 h. Total RNA was extracted, and CCL2 mRNA level was measured by Northern hybridization. The 28S and 18S ribosomal RNA bands on the methylene blue-stained membrane were used to control the RNA loading. The results shown are representative of 3 independent experiments.

Fig. 5.

Effect of α-LPA on high glucose-induced NF-κB DNA binding activity. Serum-starved HGECs were left untreated or pretreated with α-LPA (50 μM) for 30 min and then incubated with 5 or 30 mM glucose for 15, 30, or 60 min. The nuclear extracts were assayed for the ability to bind 32P-labeled NF-κB oligonucleotides by EMSA (left). To determine the specificity of the band, 100 times excess unlabeled NF-κB oligonucleotides were added in the reaction mixture before addition of the 32P-labeled NF-κB oligonucleotides. In lane 3, the nuclear protein was incubated with anti-p65 antibody before the binding reaction (right). The results shown are representative of 4 independent experiments.

High glucose-induced p65 nuclear translocation, cytosolic IκBα degradation, and serine or tyrosine phosphorylation of IκBα: effect of α-LPA.

To evaluate the dynamics of IκBα and p65 protein, the cells were left untreated or pretreated with α-LPA (50 μM) for 30 min and then incubated with 5 or 30 mM glucose for 15, 30, or 60 min, and the cytosolic and nuclear extracts from the cells were evaluated for IκBα and p65, respectively. High glucose increased nuclear translocation of p65 (Fig. 6A, top), while it was inhibited by α-LPA. However, degradation of cytoplasmic IκBα was not induced by high glucose (Fig. 6A, bottom). α-LPA also had no effect on the amount of cytoplasmic IκBα. TNF-α is a well-known activator of NF-κB. As a positive control, the cells were treated with TNF-α (10 ng/ml), and cytoplasmic IκBα was measured. Cytoplasmic IκBα began to undergo serine phosphorylation as early as 3 min after stimulation with TNF-α and disappeared, and then reappeared at 60 min (Fig. 6, BD). Thus TNF-α had a very rapid effect on the serine phosphorylation of IκBα that then faded over time. In contrast, either serine phosphorylation or degradation of IκBα was not induced by high glucose (Fig. 6, B and D).

Fig. 6.

High glucose increases tyrosine phosphorylation of IκBα and activates NF-κB without degradation of IκBα: inhibition by α-LPA. A: effects of α-LPA on high glucose-induced p65 nuclear translocation and IκBα degradation. Serum-starved HGECs were left untreated or pretreated with α-LPA (50 μM) for 30 min and then incubated with 5 or 30 mM glucose for 15, 30, or 60 min. After the cells were harvested and lysed, the cytosolic and nuclear lysates were obtained separately. The nuclear proteins were subjected to immunoblotting using an anti-p65 antibody (top), while the cytosolic proteins were subjected to immunoblotting using an anti-IκBα antibody (bottom). The result shown is representative of 4 independent experiments. B: effects of high glucose and TNF-α on IκBα degradation. HGECs were treated with 5, 30, or 5 mM glucose plus TNF-α (10 ng/ml) for 5, 15, or 30 min, and cytosolic extracts from the cells were analyzed by immunoblot using an antibody to IκBα. The result shown is representative of 4 independent experiments. C and D: effect of high glucose or TNF-α on serine phosphorylation of IκBα. Serum-starved HGECs were incubated with 5 or 30 mM glucose or TNF-α (10 ng/ml) for 3, 10, 15, 30, or 60 min. The whole cell lysates were immunoblotted with an anti-phosphospecific IκBα (Ser-32) antibody. The membranes were stripped and reprobed with anti-IκBα or an anti-actin antibody. The result shown is representative of 4 independent experiments. E: high glucose-induced tyrosine phosphorylation of IκBα: inhibition by α-LPA. Serum-starved HGECs were left untreated or pretreated with α-LPA (50 μM) for 30 min and then incubated with 5 or 30 mM glucose for 15, 30, or 60 min. The whole cell lysates were immunoprecipitated with an antibody to IκBα. The immunoprecipitated IκBα was subjected to immunoblotting using an anti-phosphotyrosine antibody. The result shown is representative of 4 independent experiments. F: TNF-α-induced tyrosine phosphorylation of IκBα. Serum-starved HGECs were pretreated with 100 μg/ml of ALLN for 1 h and then further treated with or without TNF-α (10 ng/ml). The whole cell lysates were immunoprecipitated with an antibody to IκBα and then subjected to immunoblotting using an anti-phosphotyrosine antibody. The result shown is representative of 4 independent experiments.

Because tyrosine phosphorylation of IκBα can also activate NF-κB without degradation of IκBα, we measured tyrosine phosphorylation of IκBα. For this, IκBα in whole cell lysate was immunoprecipitated with an antibody to IκBα and then fractionated on 12% SDS-PAGE. Western blot analysis was performed using an anti-phosphotyrosine antibody. As shown in Fig. 6E, high glucose increased tyrosine phosphorylation of IκBα, while α-LPA downregulated it.

Serine-phosphorylated IκBα undergoes degradation by the ubiquitin-proteasome complex, while ALLN, an inhibitor of proteasome, inhibits the proteolysis of IκBα. To determine whether TNF-α induces tyrosine phosphorylation of IκBα, the cells were pretreated with 100 μg/ml of ALLN for 1 h and then further treated with or without TNF-α. Thereafter, whole cell lysates were subjected to immunoprecipitation with an antibody to IκBα and then immunoblotting with an anti-phosphotyrosine antibody. As shown in Fig. 6F, tyrosine phosphorylation of IκBα also was induced by TNF-α.

ROS in high glucose-induced CCL2 production.

To evaluate whether the inhibitory effect of α-LPA is related to antioxidant activity, we investigated whether high glucose increases ROS in HGECs and the effect of α-LPA on it. As shown in Fig. 7, intracellular ROS were increased by high glucose. Both basal and high glucose-induced ROS were scavenged by α-LPA. N-acetylcysteine, another antioxidant, also scavenged basal and high glucose-induced ROS. Similarly to α-LPA, N-acetylcysteine downregulated high glucose-induced tyrosine phosphorylation of IκBα, nuclear translocation of p65, DNA binding activity of NF-κB, CCL2 mRNA expression, and the production of CCL2 protein (3,321 ± 198 vs. 2,553 ± 219 pg·ml−1·104 cells−1, P < 0.05, n = 6) (Fig. 8). TNF-α also increased the generation of ROS (Fig. 7).

Fig. 7.

Generation of reactive oxygen species (ROS) by HGECs treated with high glucose or TNF-α, and the scavenging effect of α-LPA and N-acetylcysteine. Serum-starved HGECs were left untreated or pretreated with α-LPA (50 μM) or N-acetylcysteine (5 mM) for 30 min and then incubated with 5 or 30 mM for 60 min (left). In another experiment, the cells were treated with or without TNF-α (10 ng/ml) for 30 min (right). Thereafter, the cells were loaded with 5 μM CM-H2DCF-DA and incubated for 10 min at 37°C. After removal of the media, the cells were washed and the expression of intracellular ROS was visualized with confocal microscopy (×200).

Fig. 8.

Effects of N-acetylcysteine on high glucose-induced tyrosine phosphorylation of IκBα, NF-κB activation, and the production of CCL2. A: N-acetylcysteine inhibits high glucose-induced CCR2 protein production. Serum-starved HGECs were left untreated or pretreated with N-acetylcysteine (5 mM) for 30 min and then incubated with 5 or 30 mM glucose for 72 h. The supernatants were collected, and CCL2 was measured by ELISA. (n = 6). *P < 0.05 compared with 5 mM glucose. **P < 0.05 compared with 30 mM glucose. B: N-acetylcysteine downregulates high glucose-induced CCL2 mRNA expression. Serum-starved HGECs were left untreated or pretreated with N-acetylcysteine (5 mM) for 30 min and then incubated with 5 or 30 mM glucose for 72 h. Total RNA was extracted, and CCL2 mRNA level was measured by Northern hybridization. The 28S and 18S ribosomal RNA bands on the methylene blue-stained membrane were used to control the RNA loading. The result shown is representative of 4 independent experiments. C: N-acetylcysteine inhibits high glucose-induced NF-κB DNA binding activity. Serum-starved HGECs were left untreated or pretreated with N-acetylcysteine (5 mM) for 30 min and then incubated with 5 or 30 mM glucose for 15, 30, or 60 min. The nuclear extracts were assayed for the ability to bind 32P-labeled NF-κB oligonucleotides by EMSA. The result shown is representative of 4 independent experiments. D: N-acetylcysteine inhibits high glucose-induced p65 nuclear translocation. Serum-starved HGECs were left untreated or pretreated with N-acetylcysteine (5 mM) for 30 min and then incubated with 5 or 30 mM glucose for 15, 30, or 60 min. The nuclear lysates were subjected to immunoblotting using an anti-p65 antibody. The result shown is representative of 4 independent experiments. E: N-acetylcysteine inhibits high glucose-induced tyrosine phosphorylation of IκBα. Serum-starved HGECs were left untreated or pretreated with N-acetylcysteine (5 mM) for 30 min and then incubated with 5 or 30 mM glucose for 15, 30, or 60 min. The whole cell lysates were immunoprecipitated with antibody to IκBα and then subjected to immunoblotting using an anti-phosphotyrosine antibody. The result shown is representative of 4 independent experiments.

PKC in high glucose-induced CCL2 production.

With regard to the involvement of PKC in high glucose-induced CCL2 production, HGECs were left untreated or pretreated with myristoylated PKC[19–27] pseudosubstrate (20 μM), calphostin C (1 μM), or GF109203X (10 μM) for 30 min and then incubated with 5 or 30 mM glucose for 72 h, and the CCL2 mRNA level was measured. As shown in Fig. 9A, all the PKC inhibitors downregulated high glucose-induced CCL2 mRNA expression. In addition, myristoylated PKC[19–27] pseudosubstrate inhibited high glucose-induced tyrosine phosphorylation of IκBα and nuclear translocation of p65 (Fig. 9, B and C).

Fig. 9.

PKC in high glucose-induced NF-κB activation. A: serum-starved HGECs were left untreated or pretreated with myristoylated PKC[19–27] pseudosubstrate (20 μM), calphostin C (1 μM), or GF109203X (10 μM) for 30 min and then incubated with 5 or 30 mM glucose for 72 h. Total RNA was extracted, and CCL2 mRNA level was measured by Northern hybridization. The 28S and 18S ribosomal RNA bands on the methylene blue-stained membrane were used to control the RNA loading. The result shown is representative of 4 independent experiments. B: serum-starved HGECs were left untreated or pretreated with myristoylated PKC[19–27] pseudosubstrate (20 μM) for 30 min and then incubated with 5 or 30 mM glucose for 15, 30, or 60 min. The nuclear lysates were subjected to immunoblotting using an anti-p65 antibody. C: in similar experiments, the whole cell lysates were prepared and immunoprecipitated with an antibody to IκBα and then subjected to immunoblotting using an anti-phosphotyrosine antibody. The results shown are representative of 4 independent experiments. D: serum-starved HGECs were left untreated or pretreated with myristoylated PKC[19–27] pseudosubstrate (20 μM), calphostin C (1 μM), or GF109203X (10 μM) for 30 min and then incubated with 5 or 30 mM glucose for 1 h (left). In another experiment, the cells were treated with or without PMA (100 nM) for 30 min (right). Thereafter, the cells were loaded with 5 μM CM-H2DCF-DA and incubated for 10 min at 37°C. After removal of the media, the cells were washed and the expression of intracellular ROS was visualized with confocal microscopy (×200).

Next, we evaluated the effect of PKC inhibition on the generation of ROS. High glucose-induced production of ROS was downregulated by all the PKC inhibitors (Fig. 9D).

Finally, we evaluated whether PMA, an activator of PKC, has a mechanism of NF-κB activation similar to that by high glucose. PMA increased nuclear translocation of p65 and CCL2 mRNA expression (Fig. 10, A and B). After treatment with PMA, cytoplasmic IκBα underwent degradation and was nearly completely degraded at 60 min, and then recovered later (Fig. 10B). PMA also increased the generation of ROS (Fig. 9D). To prevent degradation of IκBα, the cells were pretreated with 100 μg/ml of ALLN for 1 h and then further treated with or without PMA. Thereafter, whole cell lysates were subjected to immunoprecipitation with an antibody to IκBα and then immunoblotting with an anti-phosphotyrosine antibody. As shown in Fig. 10C, tyrosine phosphorylation of IκBα also was transiently induced by PMA.

Fig. 10.

PMA-induced NF-κB activation and CCL2 mRNA expression. A: serum-starved HGECs were treated with different doses of PMA for 72 h. Total RNA was extracted, and CCL2 mRNA level was measured by Northern hybridization. The 28S and 18S ribosomal RNA bands on the methylene blue-stained membrane were used to control the RNA loading. B: serum-starved HGECs were incubated with 100 nM PMA for 0, 15, 30, 60, or 120 min. After the cells were harvested and lysed, the cytosolic and nuclear lysates were obtained separately. The nuclear proteins were subjected to immunoblotting using an anti-p65 antibody (top), while the cytosolic proteins were subjected to immunoblotting using an anti-IκBα antibody (bottom). C: in similar experiments, the whole cell lysates were prepared and immunoprecipitated with an antibody to IκBα. The immunoprecipitated IκBα was subjected to immunoblotting using an anti-phosphotyrosine antibody. All the results shown are representative of 3 independent experiments.

Syk protein tyrosine kinase in high glucose-induced NF-κB activation.

To determine the tyrosine kinase responsible for high glucose-induced tyrosine phosphorylation of IκBα, the cells were left untreated or pretreated with broad-spectrum inhibitors of tyrosine kinases [genistein (50 μM), herbimycin A (1 μM)], [a specific inhibitor of Syk (BAY 61-3606; 0.1, 1, or 10 μM)] or an inhibitor of the Src family of tyrosine kinases (PP2; 1, 5, or 10 μM) for 30 min and then stimulated with high glucose. The concentrations of the inhibitors were selected based on published studies (5, 13, 31). Both genistein and herbimycin A prevented high glucose-induced nuclear translocation of p65. The inhibitory effect of PP2 on high glucose-induced nuclear translocation of p65 was small. In contrast, BAY 61-3606 almost completely inhibited high glucose-induced tyrosine phosphorylation of IκBα as well as p65 nuclear translocation (Fig. 11, A and B). To see the effects of specific suppression of Syk protein, we transfected Syk-siRNA or control-siRNA into the cells and then treated them with 5 or 30 mM glucose for 30 min. As shown in Fig. 11, CE, Syk-siRNA effectively suppressed the expression of Syk protein and inhibited high glucose-induced tyrosine phosphorylation of IκBα and p65 nuclear translocation.

Fig. 11.

Syk protein-tyrosine kinase in high glucose-induced NF-κB activation. A: serum-starved HGECs were left untreated or pretreated with genistein, herbimycin A, PP2, or BAY 61-3606 for 30 min and then incubated with 5 or 30 mM glucose for 30 min. The nuclear lysates were subjected to immunoblotting using an anti-p65 antibody. B: cells were left untreated or pretreated with BAY 61-3606 (1 μM) for 30 min and then incubated with 5 or 30 mM glucose for 30 min. The whole cell lysates were immunoprecipitated with an antibody to IκBα and then subjected to immunoblot analysis using an anti-phosphotyrosine antibody. CE: cells were transfected with control small interfering RNA (siRNA) or siRNA directed against Syk and then treated with 5 or 30 mM glucose for 30 min. The nuclear lysates were analyzed by Western blotting using an anti-p65 antibody. In similar experiments, the whole cell lysates were immunoprecipitated with an antibody to IκBα and then subjected to immunoblot using an anti-phosphotyrosine antibody. The whole cell lysates were also subjected to immunoblot using anti-Syk antibody. F: serum-starved HGECs were left untreated or pretreated with α-LPA (50 μM) or myristoylated PKC[19–27] pseudosubstrate (20 μM) for 30 min and then incubated with 5 or 30 mM glucose for 10 min. The whole cell lysates were immunoprecipitated with an anti-Syk antibody and then subjected to immunoblot analysis using an anti-phosphotyrosine antibody. All the results shown are representative of 3 independent experiments.

Activation of Syk requires phosphorylation of tyrosines in the activation loop of the Syk kinase domain (33). To determine whether high glucose induces tyrosine phosphorylation of Syk protein, the cells treated with 5 or 30 mM glucose for 10 min were immunoprecipitated with an anti-Syk antibody and then subjected to immunoblot analysis using an anti-phosphotyrosine antibody. As shown in Fig. 11F, high glucose induced tyrosine phosphorylation of Syk protein. Pretreatment of the cells with α-LPA or myristoylated PKC[19–27] pseudosubstrate inhibited high glucose-induced tyrosine phosphorylation of Syk protein.

DISCUSSION

High glucose stimulated CCL2 production in cultured HGECs, as in murine mesangial cells. High glucose increased nuclear translocations of p65, a component of NF-κB, and increased DNA binding activity of NF-κB, while either treatment of the cells with NF-κB activation inhibitors or transfection of NF-κB decoy ODNs downregulated high glucose-induced CCL2 mRNA expression. These findings suggest that NF-κB mediates high glucose-induced CCL2 production.

A new finding in this study was that the activation of NF-κB following exposure of HGECs to high glucose occurred without degradation of IκBα. Phosphorylation of IκBα at the serine residue also did not increase after exposure to high glucose. In contrast, phosphorylation of IκBα at the tyrosine residue was increased by high glucose, which explains the activation of NF-κB without degradation of IκBα. In contrast, in the cells treated with TNF-α, IκBα was rapidly phosphorylated at the serine residue, and degradation of it followed. Therefore, the way of NF-κB activation by high glucose was different from that by TNF-α. However, serine phosphorylation was not the only mechanism of NF-κB activation by TNF-α because blockade of IκBα degradation by pretreatment of the cells with ALLN revealed that tyrosine phosphorylation of IκBα was also induced by TNF-α.

Both PKC(s) and oxygen free radicals, which are known to be activated or increased after exposure to high glucose, are activators of NF-κB (22, 27). In the present study, high glucose stimulated HGECs to generate ROS. α-LPA, an antioxidant, scavenged the ROS and downregulated high glucose-induced tyrosine phosphorylation of IκBα, activation of NF-κB, and CCL2 production. N-acetylcysteine, another antioxidant, also downregulated high glucose-induced tyrosine phosphorylation of IκBα, nuclear translocation of p65, DNA binding activity of NF-κB, and CCL2 production, suggesting that generation of ROS leads to the tyrosine phosphorylation of IκBα. In the case of TNF-α, it induced tyrosine phosphorylation as well as serine phosphorylation of IκBα. It may be because TNF-α also stimulates the production of ROS.

To date, two tyrosine kinases, c-Src and Syk, were reported to be involved in tyrosine phosphorylation of IκBα, depending on the nature of stimuli or cell types. In an epithelial cell line (HeLa cells), c-Src was responsible for the redox-mediated NF-κB activation following hypoxia-reoxygenation or pervanadate treatment (5). c-Src was also involved in IL-1β-induced tyrosine phosphorylation of IκBα in a mouse mesangial cell line (13), silica-induced tyrosine phosphorylation of IκBα in RAW 264.7 macrophages (14), and TNF-α-induced tyrosine phosphorylation of IκBα in marrow macrophages (1). In contrast, Syk mediated hydrogen peroxide-induced tyrosine phosphorylation of IκBα in a leukemic cell line (KBM-5 cells) (28) and in rat small intestine cells (35). In the present study, genistein and herbimycin A, both of which are broad-spectrum inhibitors of tyrosine kinases, downregulated high glucose-induced p65 nuclear translocation, further supporting that tyrosine kinase(s) is involved in the signal pathway. Of the specific inhibitors of tyrosine kinases, PP2, the selective inhibitor of the Src family of tyrosine kinases (9, 10), had only a small inhibitory effect on high glucose-induced p65 nuclear translocation even at high dose, whereas BAY 61-3606, a highly selective inhibitor of Syk with no inhibitory effect against Btk, Fyn, Itk, Lyn, and Src (31), almost completely inhibited high glucose-induced tyrosine phosphorylation of IκBα and p65 nuclear translocation. Consistent with it, Syk-siRNA also inhibited high glucose-induced tyrosine phosphorylation of IκBα and p65 nuclear translocation. In addition, activation of Syk, as assessed by tyrosine phosphorylation of Syk, was also induced by high glucose. Thus these data suggest that Syk may play a role in tyrosine phosphorylation of IκBα in HGECs treated with high glucose.

It is well known that high glucose activates PKC(s) in a variety of cell types. In the present study of HGECs, high glucose-induced CCL2 gene expression was downregulated by PKC inhibitors and PKC[19–27] pseudosubstrate, an inhibitor of PKCα, -βI, and -βII, suggesting that PKC(s) is a mediator in signal transduction. PKC[19–27] pseudosubstrate inhibited high glucose-induced tyrosine phosphorylation of IκBα and NF-κB activation. Both PKC inhibitors and PKC[19–27] pseudosubstrate inhibited high glucose-induced generation of ROS. Therefore, activation of PKC(s) preceded the generation of ROS. This is in agreement with the study of rat mesangial cells by Ha et al. (8). In addition, high glucose-induced tyrosine phosphorylation of Syk was suppressed by α-LPA and PKC[19–27] pseudosubstrate. Taken together, it may be summarized as follows. High glucose activates PKC. PKC increases the generation of ROS, which in turn activate Syk. Syk phosphorylates IκBα at the tyrosine residue and thereby activates NF-κB.

As with high glucose, activation of PKCs by treatment of the cells with PMA also resulted in the generation of ROS, tyrosine phosphorylation of IκBα, and NF-κB activation. Additionally, PMA caused degradation of IκBα, which suggests that it has other additional mechanisms of NF-κB activation that high glucose does not have. PKCs consist of at least 12 isozymes and can be divided into 3 groups according to the requirement of DAG and Ca2+ for enzyme activity: conventional (α, βI, βII, and γ); novel (δ, ε, η, θ, μ); and atypical PKCs (λ, ι, ζ) (19). PMA is known to activate all the isozymes of both conventional and novel PKCs (19). In contrast, high glucose was shown to activate NF-κB via PKCβI (15) or -βII (17). Thus there is a possibility that the way of activation of NF-κB could be different depending on the isozymes of activated PKCs.

In the animal models of diabetic nephropathy (2, 18), dietary supplementation of α-LPA was shown to prevent the development of glomerulosclerosis, tubulointerstitial fibrosis, and proteinuria. Diabetic glomerulosclerosis is characterized by increased glomerular extracellular matrix that is mainly synthesized by mesangial cells. In the pathogenesis of diabetic nephropathy, as in immune-complex glomerulonephritis, the infiltration of monocytes/macrophages into the glomeruli is known to play an important role in glomerular injury by provoking signals to mesangial cells to produce extracellular matrix. In streptozotocin-induced diabetic rats (26, 32), macrophages were recruited within the glomeruli at the very early phase of hyperglycemia. This glomerular macrophage infiltration was associated with the subsequent mesangial matrix expansion. Macrophage recruitment was preceded by an increased glomerular expression of CCL2. In the kidney biopsy specimens obtained from patients with non-insulin-dependent diabetes mellitus (6), macrophages were demonstrated in the lesion of glomerulosclerosis, suggesting that macrophages may contribute to irreversible structural damage. Thus downregulation by α-LPA of high glucose-induced activation of NF-κB and CCL2 production may in part explain the beneficial effects of it on diabetic nephropathy.

In conclusion, activation of NF-κB by TNF-α in HGECs occurred through phosphorylation of IκBα at both serine and tyrosine residues. With serine phosphorylation, IκBα underwent degradation. In contrast, high glucose-induced NF-κB activation occurred without serine phosphorylation of IκBα and without degradation of IκBα. Our data suggest that ROS-mediated tyrosine phosphorylation of IκBα is the mechanism for high glucose-induced NF-κB activation and Syk may play a role in tyrosine phosphorylation of IκBα.

GRANTS

This study was supported by Grant 2203-070 from the Asan Institute for Life Sciences (Seoul, Korea) and the research fund of the University of Ulsan 2000.

Footnotes

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

REFERENCES

View Abstract