Endogenous oxidative stress, but not ER stress, induces hypoxia-independent VEGF120 release through PI3K-dependent pathways in 3T3-L1 adipocytes

Authors


  • Disclosure: The authors declared no conflict of interest.

  • Funding agencies: This study was supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Science and Culture (to H. I.) and by a grant from the Japan Private School Promotion Foundation.

Correspondence: Hitoshi Ishida (ishida@ks.kyorin-u.ac.jp)

Abstract

Objective

Expressions of vascular endothelial growth factor (VEGF) are increased in obese adipocytes and is secreted from obese adipose tissue through hypoxia-independent pathways. Therefore, we investigated the hypoxia-independent mechanism underlying increased expression and release of VEGF in obese adipocytes.

Design and Methods

We compared signal transduction pathways regulating VEGF with those regulating monocyte chemoattractant protein-1 (MCP-1), which is increased in obese adipocytes, in an in vitro model of artificially hypertrophied 3T3-L1 adipocytes preloaded with palmitate, without the influence of hypoxia.

Results

Palmitate-preloaded cells exhibited significantly enhanced oxidative stress (P < 0.01) and showed increased VEGF120 and MCP-1 release (P < 0.01, respectively), while endoplasmic reticulum (ER) stress was not induced. Increased VEGF120 release was significantly decreased with PI3K inhibitor LY294002 (P < 0.01). In addition, antioxidant N-acetyl-cysteine (NAC) markedly diminished not only VEGF120 secretion (P < 0.01) but also augmented Akt phosphorylation on Ser473 (P < 0.01). In contrast, increased MCP-1 release was suppressed with JNK inhibitor SP600125 and p38 MAPK inhibitor SB203580 (P < 0.01).

Conclusions

VEGF120 release from hypertrophied adipocytes can be enhanced through PI3K pathways activated by oxidative stress but not by ER stress, suggesting that VEGF120 secretion is regulated through oxidative stress-dependent pathways distinct from those involved in MCP-1 release through either JNK or p38 MAPK activation.

Introduction

Chronic low-grade inflammation in obese adipose tissue has been shown to have a crucial role in development of systemic insulin resistance, a pathogenic factor in type 2 diabetes [1, 2]. In addition, many adipocytokines such as monocyte chemoattractant protein-1 (MCP-1), tumor necrosis factor-α (TNF-α), resistin, and retinol-binding protein 4 (RBP4) have been found to be involved in insulin resistance in various cells [3]. The proinflammatory adipocytokines are highly important in the pathophysiology of systemic insulin resistance. Moreover, angiogenesis is an essential event in expansion of adipose tissue mass [7]. It has been shown recently that vascular endothelial growth factor (VEGF) mRNA expression is increased in obese adipocytes [8, 9] and that administration of anti-VEGF antibody to obese diabetic db/db mice attenuates not only angiogenesis but also adipogenesis to inhibit macrophage infiltration into adipose tissue [10], indicating that VEGF participates in both adipogenesis and macrophages infiltration in adipose tissue. Thus, de novo angiogenesis is crucial for coincident adipogenesis in obesity, and VEGF can act as a key proinflammatory adipocytokine linking adipogenesis to local low-grade inflammation.

VEGF is a potent growth factor that induces migration and proliferation of vascular endothelial cells [11]. VEGF also reinforces vascular permeability and modulates thrombogenicity [12]. The VEGF family includes VEGF-A, −B, −C, and −D, and placental growth factor [13]; VEGF-A has the core role in VEGF function [13, 14]. Three VEGF-A protein isoforms of 188, 164, and 120 amino acids, VEGF188, VEGF164, and VEGF120, respectively, are created by alternative splicing in mouse, although VEGF-A is encoded by a single gene [15, 16]; the primary importance of these isoforms is to possess or not possess a heparin-binding domain, which adheres to extracellular matrix and cell surface [13, 17, 18]. Although VEGF164 is the most abundant isoform, secretion from cells cannot be readily analyzed for possession of a heparin-binding domain. In this study, we investigated VEGF120, the major isoform after VEGF164, which lacks a heparin-binding domain.

It was recently found that VEGF can be secreted despite hypoxia in adipose tissue in obese patients [19]. Furthermore, hypoxia is known to strongly upregulate VEGF expression [20]. It has also been demonstrated that the O2 level in adipose tissue of obese subjects in vivo is not as low as 1%, which is used in cell culture experiments in vitro, and which fails to induce a counterregulatory response during hypoxia [21]. Thus, it has been assumed that the hypoxia level in obese adipose tissue in vivo is not low enough to induce VEGF expression and release. However, enhanced VEGF mRNA expression in obese adipocytes has also been demonstrated [8, 9], but the hypoxia-independent mechanism regulating VEGF120 release from hypertrophied adipocytes remains to be clarified. Exploration of VEGF120 secretion exclusive of influence by either the mild hypoxia induced in obese adipose tissue [22, 23] or macrophages infiltration into the tissues in vivo is problematical. We utilized our in vitro system previously reported [24] to evaluate VEGF120 secretion from adipocytes, in which mature 3T3-L1 adipocytes are preloaded with high palmitate resulting in artificial hypertrophy. In this system, the secretory mechanisms can be analyzed without the effects of either hypoxia or recruited macrophages.

Like VEGF, MCP-1 is a crucial adipocytokine in local low-grade inflammation. MCP-1 has a large chemotactic effect in monocytes, predominantly yielded by macrophages and vascular endothelial cells [25]. MCP-1 from obese adipose tissue accelerates macrophage infiltration into the tissue and leads to insulin resistance under obese conditions [26]. In comparison against lean mice, obese mice have increased levels of MCP-1 within their adipose tissue and plasma [4]. It has been proposed that MCP-1 expression is enhanced by oxidative stress in adipose tissue under obese conditions [29]. In addition, we previously found that MCP-1 secretion can be increased via activated JNK and nuclear factor-κB (NF-κB) pathways in our hypertrophied adipocytes model [24]. This suggests comparison of the intracellular signal transduction pathways involved in VEGF120 release from hypertrophied adipocytes with those in MCP-1 to elucidate the regulation of local inflammation in obese adipose tissue.

Methods

Reagents

Palmitate was purchased from Wako (Osaka, Japan). N-acetyl-cysteine (NAC), hydrogen peroxide (H2O2), and Wortmannin were obtained from Sigma (St. Louis, MO), and SP600125 from A. G. Scientific (San Diego, CA). PD98059, SB203580, and LY294002 were from Calbiochem (La Jolla, CA), murine recombinant MCP-1 was from Abcam (Tokyo, Japan), and murine recombinant VEGF was from Bender Medsystems (Burlingame, CA). Antibodies against VEGF120 and MCP-1 were obtained from R&D Systems (Minneapolis, MN), antibodies against C/EBP homologous protein (CHOP) and Akt1/2/3 were from Santa Cruz Biotechnology (Santa Cruz, CA), and antibody against Akt1/2/3 (phosphor-Ser473) was from Assay Designs (Ann Arbor, MI). Antiglucose-regulated protein of 78 kDa (GRP78) and β-actin antibodies from Sigma were used.

Preparation and treatment of 3T3-L1 adipocytes

3T3-L1 cells were obtained from the cell bank of the Japanese Collection of Research Bioresources (Tokyo, Japan). Cells were seeded and fed every 2 days in DMEM containing 25 mmol/l glucose supplemented with 50 U/ml penicillin, 50 μg/ml streptomycin, 100 mmol/l MEN sodium pyruvate, and 10% FCS. Cells were grown under 5% CO2 at 37°C. At confluence, differentiation was induced by addition of medium containing 500 μmol/l isobutylmethylxanthine (Sigma) and 250 nmol/l dexamethasone (Sigma) and 1.7 μmol/l insulin. After 48 h, the mixture was replaced with fresh medium. The medium was then changed every 2 days until the cells were used for experiments. On day 14 after the induction of adipocyte differentiation, 0.3 mmol/l palmitate was added to the culture medium (the maximum concentration without inducing cytotoxicity) to prepare the artificially hypertrophied mature adipocytes in vitro, as we have previously reported [24]. At 8 h after the addition of palmitate, VEGF120 release and MCP-1 release were quantified by immunoblotting. At 48 h after addition, triglyceride contents of the adipocytes were extracted and measured with or without palmitate loading for 48 h, and the amount of intracellular or secreted VEGF120 and MCP-1 and intracellular CHOP and GRP78 were quantified by immunoblotting, and VEGF-A and MCP-1 mRNA were measured by quantitative real-time RT-PCR. Lactate dehydrogenase (LDH) levels in culture medium were determined by a commercially available ELISA kit (Roche, Mannheim, Germany) according to the manufacturer's protocol to ascertain that cytotoxicity was not induced.

Fourteen days after induction of differentiation, the fully differentiated mature 3T3-L1 adipocytes were incubated with 0.3 mmol/l palmitate for 24 h. Cells were treated with either 0.5 mmol/l NAC, 50 μmol/l LY294002, 20 μmol/l PD98059, 10 μmol/l SP600125, 10 μmol/l SB203580, 10nmol/l Wortmannin, 100 ng/ml MCP-1, 50 ng/ml VEGF, or vehicle alone for another 24 h, still under 0.3 mmol/l palmitate. Efficacies of PD98059 and SB203580 were preliminarily identified by Surveyor IC EIA kits (R&D Systems) to detect phosphorylated ERK and p38 MAPK levels, respectively, and efficacy of SP600125 was by TiterZyme EIA kit (Assay Designs) to detect phosphorylated JNK level.

In other experimental series, the differentiated 3T3-L1 adipocytes were treated with 3, 100, and 300 μmol/l H2O2 or vehicle alone for 24 h, respectively. Both VEGF120 and MCP-1 released into the culture medium were then quantified by immunoblotting.

Immunoblotting

Cultured 3T3-L1 adipocytes were lysed in SDS sample buffer containing 50 mmol/l NaF, sonicated, and centrifuged. Resulting supernatants were boiled in the presence of 50 mmol/l dithiothreitol. To measure secreted proteins, the cultured medium of the cells was also boiled in SDS sample buffer with 50 mmol/l dithiothreitol. Boiled samples were subjected to SDS-PAGE and transferred onto polyvinylidene difluoride membranes (Bio Craft, Tokyo, Japan). Membranes were incubated with primary antibodies as described in Reagents section, and thereafter with horseradish peroxidase-conjugated secondary antibody. Protein bands were visualized with chemiluminescence reagents according to the manufacturer's protocol (Amersham, Little Chalfont, Buckinghamshire, UK). Bands were scanned and analyzed with NIH Image software [24]. Protein band intensities under basal conditions were set as 100% for normalization purposes. VEGF120, MCP-1, CHOP, GRP78, Akt1/2/3, and phosphorylated Akt1/2/3 proteins were confirmed by observations that the strong bands were found at 20.0, 13.8, 30.0, 78.0, 60.0, and 60.0 kDa with their corresponding antibodies, respectively.

Quantitative real-time RT-PCR

Total RNA was extracted from 3T3-L1 adipocytes using the RNAqueous®-4PCR kit (Ambion, Austin, TX) according to the manufacturer's instructions. Quantitative real-time RT-PCR was conducted using the 7300 real-time RCR system (Applied Biosystems, Foster City, CA). The following primers and probes were ordered from Applied Biosystems: VEGF-A (Mm03015192_m1) and MCP-1 (Mm00441242_m1). The mRNA signal was normalized over 18S rRNA signal. A mean value of triplicates was used for relative mRNA level.

Measurement of intracellular triglyceride contents

At 48 h after the addition of 0.3 mmol/l palmitate, the cultured 3T3-L1 adipocytes with or without palmitate loading were washed with PBS three times. Intracellular triglycerides were extracted with isopropanol and measured using Triglyceride E-test Wako (Osaka, Japan) according to the manufacturer's protocol. The resultant concentrations were adjusted to the intracellular total protein contents.

Hydroperoxides measurement

Fully differentiated 3T3-L1 adipocytes were incubated with 0.3 mmol/l palmitate for 24 h and then treated with either 0.5 mmol/l NAC or vehicle alone for another 24 h, still under 0.3 mmol/l palmitate. Supernatants were removed and the cells were washed three times with PBS. Cells were lysed in 0.5 mmol/l Tris-HCl (pH 7.4), 1.5 mmol/l NaCl, 2.5% deoxycholic acid, and 10% Nonidet P-40. Lysates were centrifuged for 10 min at 15,000g and 4°C; supernatants were assayed for intracellular endogenous hydroperoxides by the Free Radical Elective Evaluator system (Diacron, Grosseto, Italy) according to the manufacturer's protocol. Hydroperoxide units of Carratelli units were adjusted to intracellular total protein contents.

Statistical analysis

Statistical analysis was performed by unpaired t-test or by analysis of variance. Results are expressed as mean ± SEM and P < 0.05 was considered significant.

Results

Effects of palmitate preloading on triglyceride contents of mature 3T3-L1 adipocytes and increased VEGF120 expression in palmitate-preloaded adipocytes

Preloading cells with 0.3 mmol/l saturated fatty acid palmitate for 48 h resulted in a significant, 1.4-fold increase in intracellular triglyceride content (P < 0.01; Figure 1A). Measurement of LDH levels in the culture medium showed no significant changes in LDH cells incubated with vehicle alone or 0.3 mmol/l palmitate (data not shown), indicating that the concentrations used were not cytotoxic.

Figure 1.

Intracellular triglyceride contents and intracellular VEGF120 and MCP-1 in 48-h palmitate-preloaded 3T3-L1 adipocytes. Mature 3T3-L1 adipocytes were preloaded with 0.3 mmol/l palmitate [▪] or ethanol vehicle alone [□] for 48 h. A: Triglyceride contents in cells were measured, and the concentration was adjusted to intracellular total protein contents. Results are means ± SE (n = 6). **P < 0.01 compared to corresponding control cells. B and C: The mRNA levels of VEGF-A (B) and MCP-1 (C) were measured by quantitative real-time RT-PCR. The mRNA signal for each gene was normalized over 18S rRNA signal. Results are means ± SE (n = 4). **P < 0.01 compared to vehicle. D and E: Intracellular VEGF120 (D) and MCP-1 (E) were analyzed by quantitative immunoblots. β-Actin was assessed as an internal control. B and C, top: representative picture of immunoblotting that was quantified. Results are means ± SE (n = 4). **P < 0.01 compared to vehicle.

Immunoblotting results showed that intracellular VEGF120 levels were increased significantly by 1.4-fold in 48-h palmitate-preloaded mature adipocytes relative to control cells (P < 0.01; Figure 1D). MCP-1 contents in the cells also exhibited a marked increase of 1.7-fold (Figure 1E), corresponding with our previous study [24]. In addition, mRNA levels of VEGF-A including VEGF120 and MCP-1 were also increased by 2.3-fold; and 1.6-fold, respectively (P < 0.01; Figure 1B,C, respectively). These results show that the increased VEGF120 and MCP-1 expression occurred at transcriptional level.

Increased secretions of VEGF120 and MCP-1 from 48 h, but not 8 h, in palmitate-preloaded mature adipocytes

Secreted VEGF120 and MCP-1 levels were significantly increased in 48-h palmitate-preloaded adipocytes relative to control cells: VEGF120 was increased 1.4-fold and MCP-1 was increased 1.7-fold (P < 0.01; Figure 2A,B, respectively).

Figure 2.

Secreted VEGF120 and MCP-1 from 3T3-L1 adipocytes preloaded with palmitate for 8 or 48 h. Mature 3T3-L1 adipocytes were preloaded with 0.3 mmol/l palmitate [▪] or ethanol vehicle alone [□] for 8 (C and D) or 48 h (A and B). A–D: VEGF120 release (A and C) and MCP-1 (B and D) were analyzed by quantitative immunoblots. β-Actin was assessed as an internal control. A–D, top: representative pictures of immunoblotting that were quantified. Results are means ± SE (n = 4). **P < 0.01 compared to vehicle.

No significant effect of palminate was elicited by preloading for 8 h on either VEGF120 release or MCP-1 release from adipocytes (Figure 2C,D), suggesting that the enhanced VEGF120 and MCP-1 after palmitate preloading is specifically derived from indirect effects of palmitate-induced adipocyte hypertrophy.

Induction of oxidative stress, but not endoplasmic reticulum stress, in hypertrophied mature adipocytes

The intracellular concentration of endogenous hydroperoxides in nonpreloaded relative to preloaded cells was used as a marker for endogenous oxidative stress. Intracellular hydroperoxides were increased significantly in palmitate-preloaded adipocytes by 2.7-fold relative to nonpreloaded cells (4.93 ± 0.12 and 1.85 ± 0.09 Carratelli units/μg protein in preloaded and control cells, respectively: P < 0.01; Figure 3A). The antioxidant agent NAC (0.5 mmol/l) clearly attenuated the elevated content of hydroperoxides in palmitate-preloaded cells by 55% relative to preloaded cells receiving vehicle alone (2.21 ± 0.09 Carratelli units/μg protein: P < 0.01; Figure 3A).

Figure 3.

Intracellular hydroperoxides, CHOP, and GRP78 content in 48-h palmitate-preloaded adipocytes. Mature 3T3-L1 adipocytes were preloaded with 0.3 mmol/l palmitate [▪] or ethanol vehicle alone [□] for 48 h. A: At 24 h after the addition of palmitate, 3T3-L1 adipocytes were further treated with 0.5 mmol/l NAC (an antioxidant agent) or vehicle (dimethyl sulfoxide) alone for an additional 24 h. Hydroperoxides content was assayed by the Free Radical Elective Evaluator system. Results are means ± SE (n = 4). **P < 0.01 compared to corresponding control cells. B and C: Intracellular CHOP (B) and GRP78 (C) protein content were quantified by immunoblot analysis. β-Actin was assessed as an internal control. B and C, top: representative pictures of immunoblotting that were quantified. Results are means ± SE (n = 4). NS, no significant difference compared to vehicle.

In contrast, immunoblotting results showed no significant change in the amount of CHOP and GRP78 contents in hypertrophied mature adipocytes, indicating the lack of endoplasmic reticulum (ER) stress in these adipocytes (Figure 3B,C).

Exogenous H2O2 increases VEGF120 release and MCP-1 release from differentiated mature adipocytes; treatment of NAC reduces their release from hypertrophied adipocytes

As endogenous oxidative stress was induced in palmitate-preloaded 3T3-L1 adipocytes, we examined the effects of exogenous H2O2 on VEGF120 release and MCP-1 release from differentiated, nonpreloaded adipocytes. Differentiated 3T3-L1 adipocytes were exposed to various concentrations of H2O2 or vehicle alone for 24 h. Treatment with 30, 100, and 300 μmol/l H2O2 dose dependently increased VEGF120 release by 1.2-, 1.4-, and 1.6-fold relative to control cells receiving vehicle and 100 or 300 μmol/l H2O2 exposure (P < 0.01; Figure 4A). Similarly, MCP-1 release induced by 30, 100, and 300 μmol/l increased H2O2 by 1.8-, 1.9-, and 1.9-fold, respectively, in all the cases showing significant differences (P < 0.01; Figure 4C).

Figure 4.

Exogenous H2O2 increases VEGF120 release and MCP-1 from differentiated adipocytes; NAC reduces their release from 3T3-L1 adipocytes. A and C: Mature 3T3-L1 adipocytes were treated with 30, 100, and 300 μmol/l H2O2 [▪] or sterile distilled water vehicle alone [□] for 24 h. VEGF120 release (A) and MCP-1 (C) were then quantified by immunoblot analysis. B and D: 3T3-L1 adipocytes were preloaded with 0.3 mmol/l palmitate [▪] or ethanol vehicle alone [□] for 48 h. At 24 h after the addition of palmitate, adipocytes were further treated with 0.5 mmol/l NAC or vehicle (dimethyl sulfoxide) alone for an additional 24 h. VEGF120 secretion (B) and MCP-1 (D) were assayed by quantitative immunoblots. β-Actin served as an internal control. A–D, top: representative pictures of immunoblotting that were quantified. Results are means ± SE (n = 4). **P < 0.01 compared to the corresponding controls.

However, addition of NAC (0.5 mmol/l) to palmitate-preloaded adipocytes suppressed increased VEGF120 secretion by 21% compared to preloaded cells with vehicle alone, and inhibited increased MCP-1 secretion by 44% (P < 0.01; Figure 4B,D, respectively). NAC alone had no effect on VEGF120 release or MCP-1 in nonpreloaded adipocytes (Figure 4B,D).

To the contrary, we have previously reported that increased adiponectin release with palmitate was not inhibited by NAC, but that H2O2 significantly decreased adiponectin release [24]. Thus, the elevated release of VEGF120 and MCP-1 from hypertrophied adipocytes via endogenous oxidative stress may be specific to these proteins.

Involvement of phosphatidylinositol 3-kinase pathways in VEGF120 release from hypertrophied mature adipocytes

We investigated the pathway involved in VEGF120 release from palmitate-preloaded adipocytes. Akt phosphorylation on Ser473 was examined by immunoblotting with a phospho-specific antibody. Although Akt phosphorylation on Ser473 was markedly enhanced 1.4-fold (P < 0.01) in palmitate-preloaded adipocytes, LY294002 (50 μmol/l), an inhibitor of phosphatidylinositol 3-kinase (PI3K), reduced Akt phosphorylation by 32% (P < 0.01; Figure 5A). As shown in Figure 5B, treatment of 0.5 mmol/l NAC to eliminate the enhanced endogenous oxidative stress attenuated the augmented phospho-Akt by 19% compared with the preloaded cells receiving vehicle alone (P < 0.01). There were no significant differences in effect using NAC on the nonpreloaded adipocytes, but slightly diminished Akt phosphorylation was observed (Figure 5B). Moreover, the treatment with LY294002 significantly decreased VEGF120 secretion by 23% relative to the preloaded cells (P < 0.01), but LY294002 alone showed no significant effects on VEGF120 secretion from nonpreloaded cells (Figure 5C). In addition, 10 nmol/l Wortmannin, another PI3K inhibitor, was used as a control. The treatment with it also repressed the increased VEGF120 secretion by 21% relative to the palmitate-preloaded cells (P < 0.01; Figure 5D). On the other hand, MCP-1 release was completely unaffected by PI3K inhibitor LY 294002 (50 μmol/l) (Figure 5E), suggesting that the intracellular mechanism of enhanced MCP-1 secretion from hypertrophied adipocytes is independent of PI3K activation.

Figure 5.

Involvement of phosphatidylinositol 3-kinase pathways on VEGF120 secretion from 3T3-L1 adipocytes. Mature 3T3-L1 adipocytes were preloaded with 0.3 mmol/l palmitate [▪] or ethanol vehicle alone [□] for 48 h. A, C, D, and E: At 24 h after the addition of palmitate, 3T3-L1 adipocytes were further treated with 50 μmol/l LY294002 (an inhibitor of PI3K) and 10 nmol/l Wortmannin (an inhibitor of PI3K), or vehicle (dimethyl sulfoxide) alone for an additional 24 h. Akt phosphorylation on Ser473 and total Akt protein with LY294002 were shown in A by immunoblotting. VEGF120 release with Wortmannin (D) was analyzed by quantitative immunoblots. VEGF120 (C) and MCP-1 (E) secretion with LY294002 was assayed by quantitative immunoblots. β-Actin served as an internal control. B: At 24 h after the addition of palmitate, adipocytes were further treated with 0.5 mmol/l NAC or vehicle (dimethyl sulfoxide) alone for an additional 24 h. Akt phosphorylation on Ser473 was then quantified by immunoblot analysis. Phospho-Akt was normalized by total Akt protein. A–E, top: representative pictures of immunoblotting that were quantified. Results are means ± SE (n = 4). **P < 0.01 compared to the corresponding controls.

MCP-1 release from preloaded adipocytes is increased through JNK and p38 MAPK pathways

We examined other pathways involved in the regulation of VEGF120 secretion from 3T3-L1 adipocytes preloaded with palmitate. Treatment with any agent of PD98059 (20 μmol/l), an inhibitor of mitogen-activated protein kinase kinase/extracellular signal regulated protein kinase kinase 1/2 (MEK1/2), SP600125 (10 μmol/l), an inhibitor of c-Jun NH2-terminal protein kinase (JNK), and SB203580 (10 μmol/l), an inhibitor of p38 mitogen-activated protein kinase (MAPK), did not attenuate the VEGF120 secretion seen in palmitate-preloaded cells (Figure 6A). These results indicate that MAPK pathways participate little if at all in VEGF120 release from hypertrophied mature adipocytes.

Figure 6.

MCP-1 release from palmitate-preloaded adipocytes is enhanced through JNK and p38 MAPK pathways, but not MAPK pathways were unrelated to VEGF120 secretion. Mature 3T3-L1 adipocytes were preloaded with 0.3 mmol/l palmitate [▪] or ethanol vehicle alone [□] for 48 h. A and B: At 24 h after the addition of palmitate, 3T3-L1 adipocytes were further treated with 20 μmol/l PD98059 (an inhibitor of MEK1/2), 10 μmol/l SP600125 (an inhibitor of JNK), and 10 μmol/l SB203580 (an inhibitor of MAPK) or vehicle (dimethyl sulfoxide) alone for an additional 24 h. VEGF120 secretion (A) and MCP-1 (B) were then quantified by immunoblotting, with β-actin as an internal control. A and B, top: representative pictures of immunoblotting that were quantified. Results are means ± SE (n = 4). NS, no significant difference compared to the corresponding controls.

We also investigated whether these pathways are involved in MCP-1 release. The treatment with p38 MAPK inhibitor SB203580 (10 μmol/l) as well as JNK inhibitor SP600125 (10 μmol/l), the effects of which we have previously investigated [24], clearly reduced the increased MCP-1 secretion from preloaded cells (P < 0.01), whereas no significant effect was seen with MEK inhibitor PD98059 (20 μmol/l) (Figure 6B).

Exogenous MCP-1 does not affect endogenous VEGF120 release from adipocytes or vice versa

To identify the effect of exogenous MCP-1 on endogenous VEGF120 release, 3T3-L1 adipocytes were stimulated with exogenous MCP-1 (100 ng/ml). As shown in Figure 7A, the increased VEGF120 release from palmitate-preloaded cells did not exhibit any significant changes by treatment with MCP-1. In addition, VEGF120 secretion from the control cells receiving vehicle alone remained unaffected by exogenous MCP-1 as well.

Figure 7.

Effects of exogenous MCP-1 on endogenous VEGF secretion and vice versa from 3T3-L1 adipocytes. Mature 3T3-L1 adipocytes were preloaded with 0.3 mmol/l palmitate [▪] or ethanol vehicle alone [□] for 48 h. A: At 24 h after the addition of palmitate, 3T3-L1 adipocytes were further treated with 100 ng/ml MCP-1 or vehicle (sterile distilled water) alone for an additional 24 h. VEGF120 secretion was analyzed by quantitative immunoblots. B: At 24 h after the addition of palmitate, adipocytes were further treated with 50 ng/ml VEGF or vehicle (sterile distilled water) alone for an additional 24 h. MCP-1 secretion was assayed by quantitative immunoblots. β-Actin was assessed as an internal control. A and B, top: representative pictures of immunoblotting that were quantified. Results are means ± SE (n = 4). NS, no significant difference compared to the corresponding controls.

We also examined the influence of exogenous VEGF (50 ng/ml) on endogenous MCP-1 secretion from adipocytes. There were no effects on MCP-1 release from either palmitate-preloaded or nonpreloaded cells through VEGF stimulation (Figure 7B).

Discussion

It has been reported that VEGF mRNA expression is enhanced in obese adipocytes [8, 9] and that this is involved in further adipogenesis and macrophages infiltration through induction of angiogenesis in adipose tissue [10]. VEGF is released from obese adipose tissue through hypoxia-independent pathways [19]. On the other hand, MCP-1 expression in adipose tissue leads to macrophage infiltration into these tissues and insulin resistance under obese conditions [26]. Obese mice have increased levels of MCP-1 within their adipose tissue and plasma in comparison against lean mice [4]. In this study, we compared the intracellular signal transduction pathways involved in VEGF120 release from hypertrophied mature adipocytes with those involved in MCP-1 release.

When fully differentiated 3T3-L1 adipocytes are preloaded with palmitate in vitro, intracellular triglyceride contents show significant increases. This augmented triglyceride content can increase the diameter of the cell, reflecting hypertrophy of the adipocytes [30]. Intracellular and secreted VEGF120 of adipocytes preloaded for 48 h with the saturated fatty acid palmitate were markedly increased, in accord with the recent finding that VEGF mRNA expression is enhanced in obese adipocytes in vivo [9]. As palmitate took 48 h, but not 8 h, to induce VEGF120 release and the control H2O2 took 24 h to induce it, our time course data suggest the possibility of an indirect effect. We previously reported that adipocyte hypertrophy interpreted as the source of endogenous ROS was induced by preloading with palmitate for 48 h but not 8 h [24]. Hence, this indirect effect appears to be caused by enhanced endogenous oxidative stress, which might be induced through the activation of NADPH oxidase, with hypertrophy of adipocytes by palmitate. Accordingly, we used 48 h to determine the VEGF120 secretion. In the previous studies, the quantity of VEGF secretion could not be precisely determined because of combination with VEGF isoforms holding a heparin-binding domain. We therefore investigated VEGF120, which lacks a heparin-binding domain. As in our previous study [24], the intracellular concentration of hydroperoxides, an endogenous oxidative stress marker, was found to be increased in these hypertrophied mature adipocytes, suggesting that induction of endogenous oxidative stress was enhanced in these cells. In addition, increased VEGF120 release from palmitate-preloaded adipocytes was diminished by treatment with NAC, a known antioxidant, and exogenous H2O2 stimulation upregulated VEGF120 secretion as in other cells [31, 32], indicating that enhanced endogenous oxidative stress in hypertrophied adipocytes drives the increase in VEGF120 secretion. We also determined whether ER stress, the other major cellular stressor, is enhanced in palmitate-preloaded mature 3T3-L1 adipocytes. Interestingly, there were no changes in CHOP or GRP78 (factors induced by ER stress) intracellular contents by treatment of palmitate, demonstrating that ER stress is not induced in mature adipocytes hypertrophied by palmitate preloading. In this regard, previous studies found ER stress to be brought about by palmitate exposure to cells other than mature adipocytes [33]. The reason for these discrepancies among the various cells in terms of susceptibility to ER stress remains to be elucidated, but some protective mechanisms may be exerted in fully hypertrophied mature adipocytes in our in vitro system without producing local low-grade inflammation derived from infiltrated classical macrophages [34, 35]. These results indicate that oxidative stress strongly affects VEGF120 release, whereas ER stress barely influences the secretion. The contribution of angiotensin II-induced endogenous oxidative stress to intracellular VEGF120 protein expression was previously reported in proximal tubular epithelial cells [36]. In this study, we demonstrate its effect in palmitate-preloaded hypertrophied mature adipocytes. Our findings suggest that reduction of endogenous oxidative stress in adipocytes can induce suppression of de novo angiogenesis in adipose tissue and result in inhibition of coincident adipogenesis in vivo to mitigate increasing obesity.

We also investigated intracellular signal transduction pathways downstream of endogenous sources of oxidative stress. Increased VEGF120 release from palmitate-preloaded adipocytes was significantly suppressed by treatment with PI3K inhibitor LY294002. Moreover, Akt phosphorylation on Ser473, which occurs downstream of activated PI3K, was clearly enhanced in adipocytes preloaded with palmitate, and addition of NAC that moderates enhanced endogenous oxidative stress markedly diminished the increased Akt phosphorylation. These effects were not observed in the hypertrophied adipocytes treated with MEK1 inhibitor PD98059, JNK inhibitor SP600125, or p38 MAPK inhibitor SB203580. Although ROS is known to be associated with PI3K activation by hypoxia in a very quick manner on MCF-7 breast cancer cells [37], our results indicate otherwise that on hypertrophied 3T3-L1 adipocytes, endogenous oxidative stress activates PI3K/Akt pathways and subsequently augments VEGF120 secretion independent of hypoxia, whereas MAPKs pathways are involved only very little. We also show that increased endogenous oxidative stress in hypertrophied adipocytes activates PI3K/Akt pathways and leads to increased VEGF120 release. It was shown previously that insulin can upregulate VEGF mRNA expression through activation of PI3K in adipocytes [38]. However, other signal transduction pathways may also be involved in augmentation of VEGF120 release in hypertrophied adipocytes with insulin resistance.

In addition to VEGF, MCP-1, one of the major chemokines, is considered an important factor in local low-grade inflammation through infiltrated macrophages in adipose tissue. We previously verified that MCP-1 release is increased via activated JNK and NF-κB pathways in hypertrophied adipocytes [24]. We speculate that MCP-1 secretion from the hypertrophied cells further encourages macrophages infiltration together with VEGF-induced de novo angiogenesis into adipose tissues. Although PI3K pathways are involved only very little if any at all in MCP-1 release, endogenous oxidative stress increased its release in vitro [24], as in the case for VEGF120. In contrast, the activation of the JNK and p38 MAPK pathways, which have little influence on VEGF120 expression, is necessary for MCP-1 release from hypertrophied mature adipocytes, indicating that the intracellular signal transduction involved in endogenous VEGF120 release from palmitate-preloaded adipocytes differs from that involved in endogenous MCP-1 release. Although endogenous secretion was generally enhanced by endogenous oxidative stress, treatment with exogenous MCP-1 had no effects on endogenous VEGF release from hypertrophied mature adipocytes or vice versa. Thus, it is likely that their mutual secretory regulation through paracrine stimulation is negated.

It was reported very recently that three obesity-associated factors, preadipocyte differentiation, insulin, and hypoxia, stimulate VEGF secretion from adipocytes [39]; hypertrophy of mature adipocytes in this study is the novel, fourth factor known to stimulate VEGF release. The increase is caused primarily by endogenous oxidative stress and not by ER stress. Moreover, VEGF120 release can be enhanced through PI3K/Akt pathways activated by endogenous oxidative stress in hypertrophied adipocytes, suggesting that VEGF120 secretion is regulated via oxidative stress-dependent signaling pathways distinct from those of MCP-1 in these cells (Figure 8). Interestingly, hypertrophy of adipocytes itself can induce active de novo angiogenesis in adipose tissue in vivo through enhanced secretion of endogenous VEGF120, which in turn links to coincident adipogenesis in a vicious cycle to further expansion of adipose tissue and increasing obesity. These outcomes suggest a strategy to diminish endogenous oxidative stress in palmitate-induced hypertrophied adipocytes and to subsequently attenuate the activation of PI3K/Akt pathways regulating VEGF120 secretion, resulting in suppression of angiogenesis in adipose tissue and mitigation of further development of obesity and insulin resistance. Inhibition of JNK or p38 MAPK activation might also specifically prevent such macrophage infiltration through suppression of MCP-1 release to ameliorate insulin resistance in adipocytes, in accord with findings in previous in vivo studies [40].

Figure 8.

Schema of putative mechanisms concerning endogenous VEGF and MCP-1-induced low-grade chronic inflammation and insulin resistance in adipose tissue in vivo. Hypertrophy of adipocytes by high concentration of palmitate increases endogenous oxidative stress, permitting endogenous VEGF release to be enhanced through PI3K/Akt pathways activated by oxidative stress; MCP-1 release in these cells nevertheless can be upregulated via activation of distinct signaling of the JNK and p38 MAPK pathways. Moreover, active de novo angiogenesis in adipose tissue in vivo is induced through the enhanced secretion of endogenous VEGF, which in turn links to coincident adipogenesis in a vicious cycle to expansion of adipose tissue with adipocyte hypertrophy (obesity). The putative cooperation of both this angiogenesis and the increased endogenous MCP-1 secretion from hypertrophied adipocytes promotes infiltration of classical macrophages into obese adipose tissue, and subsequently low-grade chronic inflammation occurs, leading to the development of systemic insulin resistance.

Acknowledgments

The authors are grateful to Ms. S. Kitazumi for her secretarial work.

Ancillary