Regulation of Vascular Endothelial Growth Factor Expression by Advanced Glycation End Products*

,

zymatic reactions between glucose and free amino reactive group of proteins and lipids (2). AGEs and their intermediates have been implicated in pathophysiological dysfunction associated with the vascular complications of diabetes mellitus, such as retinopathy (3).
Diabetic retinopathy remains a leading cause of blindness in the Western countries. Retinopathy is characterized by a progressive alteration of retinal microvasculature, increased vasopermeability, and the pathologic intraocular neovascularization leading to loss of vision (4). It has been proposed that an angiogenic factor released from the retina may stimulate this neovascularization. Vascular endothelial growth factor (VEGF) is such an angiogenic factor and has been linked to the development of diabetic retinopathy. Indeed, it has been observed that intraocular VEGF levels are increased in diabetic patients and that a correlation exists between the levels of glycated proteins and the development of retinopathy (5)(6)(7). Alternative splicing of VEGF mRNA leads to the formation of five distinct isoforms of 121, 145, 165, 189, and 206 amino acids (8). VEGF 165 is the most predominant form of VEGF. VEGF expression is mainly regulated by tissue oxygen content (9 -11) but also by growth factors and cytokines, including plateletderived growth factor, epidermal growth factor, insulin, insulin-like growth factor-I, tumor necrosis factor ␣, and transforming growth factor ␤ (12)(13)(14)(15)(16)(17). Hypoxia stimulates VEGF expression, through an increase in gene transcription, a regulation at a translational level, and stabilization of the mRNA (18 -20). Hypoxia regulates VEGF gene transcription by activating the transcription factor, hypoxia inducible factor-1 (HIF-1) (21,22). HIF-1 is a basic-helix-loop-helix transcription factor, which is composed of two subunits, HIF-1␣ and HIF-1␤. HIF-1␤, also known as the arylhydrocarbon nuclear translocator, is constitutively expressed, whereas HIF-1␣ expression is increased upon hypoxia. In the absence of adequate signals (hypoxia or growth factor stimulation), HIF-1␣ is rapidly ubiquitinated by the von Hippel-Lindau tumor suppressor E3 ligase complex, and subjected to proteasomal degradation (23). Under hypoxic conditions or after stimulation with growth factors, HIF-1␣ is not degraded and accumulates to form an active complex with HIF-1␤. Mechanisms by which cells respond to hypoxia and activate HIF-1 are not fully understood (21). It has been shown that ERK activates HIF-1 by promoting the phosphorylation of HIF-1␣ (24,25) and that PI3K-dependent pathways are involved in HIF-1 activation and VEGF expression (26 -28).
AGEs stimulate VEGF expression in epithelial cells and in vascular smooth muscle cells (29,30). In vivo elimination of AGEs from the circulation occurs by macrophage scavenger receptor-mediated endocytosis in liver endothelial and Kupffer * This work was supported in part by funds from INSERM, Association pour la Recherche contre le Cancer (grant 5492). University of Nice-Sophia Antipolis and by grants from the European Community (QLG1-CT-1999-00674) and Aventis Pharma Deutschland GmbH (Frankfurt, Germany). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ Recipient of a fellowship from the Ministère de l'Enseignement Supérieur et de la Recherche (France).
‡ Both authors contributed equally to this work. ¶ To whom correspondence should be addressed: Tel.: 33-4-93-81-54-47; Fax: 33-4-93-81-54-32; E-mail: peraldis@unice.fr. 1 The abbreviations used are: AGE, advanced glycation end; Alb-AGE, albumin AGE; ARPE, retinal epithelial cells; ERK, extracellular signal-related kinase; MAPK, mitogen-activated protein kinase; MEK, MAPK/ERK; JNK, Jun amino-terminal kinase; PKB, protein kinase B; PI3K, phosphatidylinositol 3-kinase; VEGF, vascular endothelial growth factor; HIF, hypoxia inducible factor; SAPK, stress-activated protein kinase; CREB, cAMP-response element-binding protein; PBS, phosphate-buffered saline; DTT, dithiothreitol; PMSF, phenylmethyl-cells (31,32). However, during diabetes the increased concentration of AGEs result in their binding to AGE-binding proteins. Several AGE-binding proteins have been identified, including p60 homologous to OST-48, p90 homologous to 80K-H (a protein kinase C substrate), galectin-3, and RAGE (33)(34)(35). However, the signaling pathways mediated by the activation of these AGE-binding proteins are poorly understood. Cell treatment with glycated proteins leads to generation of reactive oxygen species, activation of ERK1/2, and activation of the transcription factor NF-B (36,37). The signaling pathways activated by AGEs and implicated in VEGF expression remain to be defined. In this report, we found that AGEs stimulate VEGF mRNA expression through an increase in HIF-1␣ accumulation and activation of HIF-1. Moreover, AGEs-induced VEGF expression and HIF-1␣ accumulation are dependent on ERK. RNA Isolation and Northern Blot Analysis-TRIzol reagent (Life Technologies, Inc.) was used to extract total cellular RNA from tissues or confluent cells grown in 100-mm tissue culture plates according to the manufacturer's instructions. Cells were serum-deprived overnight in medium containing 0.2% (w/v) bovine serum albumin and cells were pretreated or not with inhibitors for 30 min and stimulated for indicated times. RNA was extracted, and 10 g of total RNA was denatured in formamide and formaldehyde and separated by electrophoresis in formaldehyde-containing agarose gels. RNA was transferred to Hybond-N membranes (Amersham Biosciences AB, Uppsala, Sweden) and cross-linked to the membrane by UV radiation. Probes were labeled with [␥-32 P]dCTP by random priming using the Rediprime kit (Amersham Biosciences, Inc.) and purified with the Probequant kit (Amersham Biosciences, Inc.). Hybridizations were performed at 42°C in NorthernMax hybridization buffer (Ambion, Inc., Austin, TX). Membranes were washed in 1ϫ SSC, 0.5% SDS, and radioactivity was quantitated using a Storm 840 PhosphorImager, Molecular Dynamics.
Nuclear Extract Preparation-Nuclear extracts were prepared as previously described (38). Serum-starved cells were treated with ligands, chilled to 4°C, and washed with ice-cold PBS. Cells were scraped into 5 ml of PBS and pelleted by centrifugation at 1500 rpm for 10 min at 4°C. Cell pellets were washed with four packed cell volumes of buffer A (10 mM Tris-HCl, pH 7.6, 1.5 mM MgCl 2 , 10 mM KCl, supplemented with 2 mM DTT, 0.4 mM PMSF, 2 g/ml leupeptin, 2 g/ml aprotinin, 2 g/ml pepstatin, and 1 mM Na 3 VO 4 ), resuspended in four packed cell volumes of buffer A, and incubated on ice for 10 min. Cells were lysed by 20 strokes in a glass Dounce homogenizer with a type B pestle. Nuclei were pelleted at 3000 rpm for 10 min and resuspended in three packed nuclear volumes of buffer C (0.42 M KCl, 20 mM Tris-HCl, pH 7.8, 20% (v/v) glycerol, 1.5 mM MgCl 2 ) supplemented with 2 mM DTT, 0.4 mM PMSF, 2 g/ml leupeptin, 2 g/ml aprotinin, 2 g/ml pepstatin, and 1 mM Na 3 VO 4 . Nuclear proteins were extracted by stirring at 4°C for 30 min. After centrifugation at 13,500 rpm for 30 min, the supernatant was dialyzed against buffer Z-100 (25 mM Tris-HCl, pH 7.6, 0.2 mM EDTA, 20% (v/v) glycerol, 2 mM DTT, 0.4 mM PMSF, 1 mM Na 3 VO 4 , 100 mM KCl) at 4°C. The dialysate was clarified by centrifugation at 13,500 rpm for 30 min at 4°C and designated as crude nuclear extract. The nuclear extracts were aliquoted, frozen in liquid N 2 , and stored at Ϫ80°C. Protein concentration was determined by a Bio-Rad assay using bovine serum albumin as standard.
Binding reactions were performed as described previously (38). Reactions contained 10 g of nuclear extract and 0.1 g of denatured salmon sperm DNA (Sigma) in 10 mM Tris-HCl, pH 7.5, 50 mM KCl, 50 mM NaCl, 1 mM MgCl 2 , 1 mM EDTA, 5 mM DTT, and 5% (v/v) glycerol. After preincubation for 5 min at room temperature, probe (2.5 ϫ 10 5 cpm) was added and the incubation was continued for an additional 15 min, after which the reaction mixtures were loaded onto 5% nondenaturing polyacrylamide gel. Electrophoresis was performed at 185 V in 0.25ϫ TBE (22.25 mM Tris-HCl, 22.25 mM boric acid, and 1.25 mM EDTA) at 4°C. Gels were vacuum-dried, and radioactivity was determined on a Storm 840. Competitor DNAs were preincubated with nuclear extract and salmon sperm DNA for 5 min prior to addition of labeled probe.

RESULTS
Glycated Albumin Increases VEGF mRNA Expression-To study the regulation of VEGF mRNA expression, mice were injected with saline solution alone (control) or with insulin, albumin (Alb), or glycated albumin (Alb-AGE). After the indicated times, eyes were removed from the animals, and RNA was extracted and analyzed by Northern blotting using a VEGF 165 cDNA probe. RNA loading and integrity were verified by a Northern blot using 18 S as a probe, and expression of VEGF mRNA was normalized as the ratio of VEGF mRNA over 18 S (Fig. 1).
We observed that both insulin and Alb-AGE induced a 2-fold increase in the level of VEGF mRNA in eyes of mice compared with control injection. Albumin alone had no effect on VEGF expression. Because Lu et al. (29) have shown that in retina the VEGF mRNA level is increased in the ganglion, inner nuclear layer, the retinal pigment epithelial, and the choroids, we have chosen here to study the effect of AGEs on VEGF mRNA expression in retinal epithelial cells.
Glycated Albumin Stimulates VEGF mRNA and Protein Expression in Retinal Epithelial Cells-Human retinal epithelial cells (ARPE-19) were treated for 6 h with Alb-AGE, insulin, or with both molecules, and RNA was extracted. RNA was analyzed by Northern blotting for VEGF mRNA expression ( Fig.  2A). As observed, Alb-AGE and insulin induced a 2-and 3-fold increase in VEGF mRNA expression, respectively. Co-treatment with Alb-AGE and insulin had an additive effect, because together they induce a 5-fold increase in VEGF mRNA expression.
As shown in Fig. 2B, we verified that the increase in VEGF mRNA expression was associated to an increase in the protein levels. ARPE-19 cells were treated with Alb-AGE or with insulin, and whole cell lysates were analyzed by Western blotting using antibody to VEGF. Both Alb-AGE and insulin stimulated the expression of VEGF isoforms (VEGF 121 , VEGF 165 , VEGF 189 ) in ARPE-19 cells. Signaling pathways involved in insulin-induced VEGF expression have been studied (17). In contrast, the regulation of VEGF mRNA expression in response to glycated proteins is poorly documented. For this reason, we have focused our work on the signaling pathways activated by AGEs and implicated in VEGF expression.

Signaling Pathways Implicated in Alb-AGE-induced VEGF mRNA Expression-Because stimulation of cells with Alb-AGE
has been shown to produce reactive oxygen species (36), we investigated whether antioxidants modulate Alb-AGE-stimulated VEGF expression. ARPE-19 cells were treated with Alb-AGE in the absence or presence of antioxidants, N-acetyl-Lcysteine (NAC) or pyrrolidinedithiocarbamate (PDTC). RNA was extracted and analyzed by Northern blotting using a VEGF 165 cDNA probe. As shown in Fig. 3A, we observed that the level of VEGF mRNA detected after Alb-AGE treatment is increased by the use of NAC and is not modified by PDTC treatment. The observation that NAC increases VEGF mRNA level could be due to its ability to activate HIF-1 (40).
Then, we investigated the role of ERK and PI3K on Alb-AGEinduced VEGF mRNA expression. ARPE-19 cells were treated with Alb-AGE in the absence or presence of specific inhibitors for MEK, U0126, or for PI3K, wortmannin. RNA was extracted and Northern blot analysis was performed using a VEGF 165 cDNA probe (Fig. 3B). Alb-AGE stimulated VEGF mRNA expression in ARPE-19 cells (5-fold increase). Inhibition of MAPK activation by U0126 blocked the Alb-AGE-induced VEGF mRNA expression, whereas inhibition of PI3K by wortmannin did not seem to affect the ability of Alb-AGE to induce VEGF mRNA expression.
In conclusion, our data show that Alb-AGE stimulates VEGF expression through an ERK-dependent pathway.
Glycated Albumin Activates MAPKs but Not PI3K/PKB-We investigated the signaling pathways activated by Alb-AGE. First, we studied the ability of Alb-AGE to activate the MAPK family members, extracellular signal-related kinases (ERK), Jun amino-terminal kinases (JNK), and p38 MAPK. ARPE-19 cells were stimulated with Alb-AGE for 10 or 20 min, and whole cell lysates were analyzed by Western blotting using antibodies to phosphorylated forms of JNK, ERK, or p38 MAPK (Fig. 4, A-C). Alb-AGE activated JNK1 and JNK2. Maximal activation of ERK1 and ERK2 by Alb-AGE was reached within 10 min of stimulation. However, we were unable to detect any activation of p38 MAPK after Alb-AGE treatment of ARPE-19 cells.
To determine whether Alb-AGE activates the PI3K-dependent cascade, we measured the phosphorylation of a downstream effector of PI3K, PKB (Fig. 4D). ARPE-19 cells were treated with Alb-AGE or insulin, and whole cell lysates were analyzed by Western blotting using antibodies to PKB phosphorylated on serine residue 473 or to total PKB. We observed that insulin induced a rapid and strong phosphorylation and activation of PKB, whereas no phosphorylation of PKB could be

FIG. 2. Glycated albumin stimulates VEGF mRNA and protein expression in ARPE-19 cells. A, ARPE-19 cells were treated with
Alb-AGE (100 g/ml), insulin (100 nM), or both, for 6 h. RNA was extracted and analyzed by Northern blotting using a VEGF 165 cDNA as probe. B, ARPE-19 cells were treated with Alb-AGE (100 g/ml) or insulin (100 nM) for 24 h. Cell lysates were analyzed by immunoblotting using antibody to VEGF. Experiments shown are representative of experiments performed two to three times with identical results. detected after Alb-AGE treatment. Using a PI3K assay, we did not detect activation of PI3K in response to glycated albumin (data not shown).
Glycated Albumin Stimulates HIF-1␣ Accumulation-To study whether glycated proteins stimulate VEGF expression through HIF-1 activation, we first investigated whether Alb-AGE induced an accumulation of HIF-1␣. ARPE-19 cells were treated with Alb-AGE or with a known inducer of HIF-1␣, CoCl 2 , for 4 h, and cell lysates or nuclear extracts were analyzed by Western blotting using an antibody to HIF-1␣ (Fig. 5,  A and B). Expression of HIF-1␣ was normalized using a Western blot with antibodies to Shc or to CREB (cAMP-response element-binding protein). Expression of the HIF-1␣ protein was increased upon Alb-AGE stimulation (2.5-fold) in both total cell lysates and in the nucleus. As expected, CoCl 2 treatment of cells led to a greater accumulation of HIF-1␣ in ARPE-19 cells (10-fold stimulation).
Glycated Albumin Stimulates HIF-1 DNA Binding Activity-Because Alb-AGE was found to stimulate HIF-1␣ accumulation, we determined whether this accumulation was correlated with an activation of HIF-1 DNA binding activity. ARPE-19 cells were treated in the absence or presence of Alb-AGE, and nuclear extracts were isolated. Double-strand oligonucleotides containing the hypoxia response element site present in eryth-ropoietin promoter were used to measure the ability of HIF-1 to bind to a specific sequence (Fig. 5C). We observed that Alb-AGE induced an increase in the DNA binding activity of HIF-1. This binding is specific, because it is abolished in the presence of an excess of unlabeled oligonucleotides.
Accumulation of HIF-1␣ in Response to Alb-AGE Is Dependent on ERK-Because Alb-AGE stimulated VEGF mRNA expression through an ERK-dependent pathway (Fig. 3), we determined whether ERKs are implicated in the induction of HIF-1␣ in response to AGEs. ARPE-19 cells were pretreated with wortmannin or U0126 and then stimulated with Alb-AGE. Whole cell lysates were analyzed by Western blotting using antibodies to HIF-1␣ or to Shc (Fig. 6). As a positive control, we used CoCl 2 to stimulate HIF-1␣ accumulation. Alb-AGE induced a 3-fold increase in HIF-1␣ expression, which was not affected by pretreatment with the PI3K inhibitor, wortmannin. Inhibition of ERK activation by U0126 completely inhibited the expression of HIF-1␣ in response to Alb-AGE. In conclusion, we found that Alb-AGE stimulated HIF-1␣ expression through an ERK-dependent pathway. DISCUSSION In the present study, we showed that Alb-AGE stimulates VEGF mRNA expression in an ERK-dependent pathway. Moreover, Alb-AGE activates the transcription factor, HIF-1. To the best of our knowledge, this is the first report showing that AGEs stimulate VEGF expression and HIF-1 DNA binding activity.
AGEs stimulate expression of VEGF mRNA and protein in ARPE-19. This is not due to an increase in the stability of the messenger (data not shown) and results from an increase in the transcription of the gene. We found that AGEs stimulate VEGF expression through an ERK-dependent pathway, but not through a PI3K-dependent pathway. Moreover, we have observed that the antioxidant, NAC, did not inhibit VEGF mRNA expression in ARPE-19 but rather increased VEGF mRNA level. It has been reported that NAC appears to increase both the elevation of glutathione and reduction of the oxidized form of glutathione, and then, induces an increase in HIF-1␣ nuclear abundance and activity (40). These observations could explain the fact that NAC increases VEGF mRNA level. PDTC, like NAC, induces HIF-1␣ nuclear translocation but fails to activate DNA binding activity (40), explaining the observation that PDTC did not modify the level of VEGF mRNA. However, our results are in contrast with a previous report, which shows that antioxidants, N,NЈ-dimethylthiourea and NAC, inhibit AGEsinduced VEGF mRNA expression in bovine smooth muscle cells (29).
We correlated the ability of AGEs to stimulate VEGF expression through an ERK-dependent pathway with their ability to activate MAPK family members. Indeed, in ARPE-19 cells, we found that AGEs activate ERK and JNK but not p38 MAPK. Activation of ERK has been observed in numerous cell types (36,37). However, activation of stress-activated protein kinases (SAPK), JNK and p38, is more controversial. In C6 glioma cells, activation of RAGE, a receptor for AGEs, leads to activation of both JNK and p38 MAPK (41). In contrast, in THP-1 monocytes, AGEs activate only p38 MAPK, without activation of JNK (42). These observations suggest that activation of MAPKs could be cell-specific. However, SAPK are not involved in the activation of the transcription of VEGF mRNA but in its stability (43).
We did not detect activation of PI3K/PKB by AGEs. Such AGEs action has been reported only in Jurkat and in PC12 cells (44). Because AGEs activate neither PI3K nor PKB in ARPE-19 cells, it is possible that PI3K activation is also cell-specific.
AGEs stimulate VEGF expression through the accumulation of HIF-1␣ and the subsequent activation of the transcription factor HIF-1. The induction of HIF-1␣ by AGEs occurs through an ERK-dependent pathway. This is in agreement with results from Agani and Semenza (45), who show that ERKs are involved in HIF-1␣ accumulation in response to mersalyl, an organomercurial compound. ERKs have also been implicated in VEGF expression in response to hypoxia by phosphorylating HIF-1␣ subunit and then leading to an activation of the transcriptional activity of HIF-1 (24,25).
However, HIF-1 expression and activity can be regulated through other mechanisms. Indeed, overexpression of an activated form of PI3K or PKB, or expression of dominant-negative phosphatase and tensin homolog stimulates HIF-1␣ expression in response to hypoxia and induces angiogenesis (26,27). Moreover, epidermal growth factor regulates HIF-1␣ expression through the signaling cascade PI3K/PKB/FRAP (FKBP-rapamycin-associated protein) (28). Additional studies are required to determine whether PI3K-or ERK-dependent signaling pathways modulate the degradation of HIF-1␣ through the proteasome.
Finally, we have observed that, compared with insulin or glycated albumin alone, a co-treatment of ARPE-19 cells with both insulin and glycated albumin has a synergistic action on expression of VEGF mRNA. Insulin induces VEGF expression mainly through a PI3K-dependent pathway (17). Because AGEs stimulate VEGF expression through an ERK-dependent pathway, it is possible that the two polypeptides use distinct signaling pathways, leading to an additive action on VEGF expression. FIG. 5. Glycated albumin stimulates HIF-1␣ accumulation and HIF-1 DNA binding activity. ARPE-19 cells were stimulated with Alb-AGE (100 g/ml) or CoCl 2 (200 M) for 4 h. Whole cell lysates (A) or nuclear extracts (B) were prepared and analyzed by Western blotting using antibodies to HIF-1␣, to Shc, or to CREB. C, ARPE-19 cells were stimulated with Alb-AGE (100 g/ml) for 4 h prior nuclear extracts preparation. 10 g of protein was incubated with radiolabeled oligonucleotide containing the hypoxia response element from the erythropoietin promoter in the absence or presence of excess unlabeled probe (Competitor). Experiments shown are representative of experiments performed three times with identical results.
FIG. 6. Glycated albumin stimulates HIF-1␣ accumulation through an ERK-dependent pathway. ARPE-19 cells were pretreated for 30 min with wortmannin (100 nM) or U0126 (10 M) and stimulated with Alb-AGE (100 g/ml) for 4 h, or were stimulated with CoCl 2 (200 M) for 4 h. In lane 5, U0126 (10 M) was added before Alb-AGE stimulation, and after 2 h of treatment with Alb-AGE. Whole cell lysates were analyzed by Western blotting using antibodies to HIF-1␣ or to Shc. Experiments shown are representative of experiments performed three times with identical results.
In conclusion, to the best of our knowledge, this is the first report showing that AGEs stimulate VEGF expression through an ERK-dependent pathway. Moreover, our results suggest that AGEs-induced expression of VEGF is dependent on the transcription factor HIF-1. Based on these observations, we suggest that blockage of HIF-1 activity by ERK inhibitors could be used as a therapeutic approach to inhibit AGE-induced neovascularization during diabetes.