INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

VASCULAR ENDOTHELIAL GROWTH FACTOR (VEGF) is a potent endothelial-specific mitogen (18). It is an ~46-kDa heparin-binding dimeric glycoprotein that binds to two high-affinity receptors with tyrosine kinase domains mainly located on vascular endothelial cells (8, 10, 14, 18). Four molecular species of the human VEGF family have been identified, generated by alternative splicing of the same mRNA, leading to peptides of 206, 189, 165, and 121 amino acids (13). All four are secreted by cells, but the first two remain mainly cell associated. It is a potent angiogenic factor that has been implicated in physiological angiogenesis as well as in the neovascularization of solid tumors (10). Neovascularization of the atherosclerotic plaque has been correlated with the severity of atherosclerosis as well as the development of intraplaque hemorrhage (2, 9). However, the mechanism and stimuli for neovascularization in the atherosclerotic plaque are not very clear.

VEGF is induced by hypoxia as well as by growth factors such as angiotensin II (ANG II) and platelet-derived growth factor (PDGF) in vascular smooth muscle cells (VSMC) (35, 42, 43). Besides its angiogenic properties, VEGF is also a potent vascular permeablity factor and stimulates monocyte migration through endothelial cells (4, 5). These events are key steps in the initiation and progression of atherosclerosis mediated by the concerted action of multiple growth factor cytokines and lipids (30). Thus VEGF is an attractive candidate for the pathological neovascularization and endothelial permeability observed in atherosclerosis.

Cardiovascular disease is clearly increased by the diabetic state. However, the precise mechanisms by which hyperglycemia can affect vascular dysfunction and atherosclerosis remain unclear. Recent in vivo and in vitro studies have implicated VEGF in the development of retinal angiogenic diseases such as proliferative diabetic retinopathy (1, 19, 20). A very recent report indicated that VEGF expression was upregulated by high glucose (HG) concentration in retinal epithelial cells (34). Evidence suggests that VSMC can express VEGF (35, 43). However, no studies have previously evaluated the effects of HG concentrations on VEGF from VSMC. Therefore, we have examined the regulation of VEGF expression in VSMC by HG culture conditions to simulate the diabetic state. We have also evaluated whether HG culture can modulate the effects of ANG II on VEGF expression.

Products of the 12-lipoxygenase (12-LO) pathway of arachidonic acid metabolism have been shown to mediate the hypertrophic effects of ANG II and also have direct growth-promoting effects in VSMC (25). These growth effects are enhanced under chronic HG culture conditions (25). Furthermore, ANG II as well as HG could increase 12-LO activity and expression in VSMC (26). 12-LO products also possess angiogenic properties (17, 32), suggesting that they could potentially modulate VEGF activity or expression during the hyperglycemic state. We have therefore also examined whether 12-LO products such as 12-hydroxyeicosatetraenoic acid (12-HETE) can increase VEGF expression in VSMC.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Culture of VSMC. Human aortic VSMC (HSMC) used in these studies were an immortalized adult human VSMC cell line, which was a generous gift from Dr. J. McDougall (Fred Hutchinson Cancer Center, Seattle, WA). This immortalized human VSMC has been shown to have typical VSMC characteristics (6) and, furthermore, ANG II could increase intracellular calcium and also lead to cellular hypertrophy in these cells via the AT_1a receptor (27). Porcine VSMC (PSMC) were cultured from porcine aortic smooth muscle explants and used from passages 2-6 as described earlier (26). The PSMC and HSMC were maintained in Dulbecco's modified Eagle's (DME) normal glucose (NG, 5.5 mM) medium containing 10% fetal calf serum (FCS). For HG experiments, the cells were cultured for at least two passages in DME HG (HG, 25 mM) containing 10% FCS.

Northern blotting to detect VEGF mRNA. Nearly confluent cells were placed in low-serum medium (NG) or HG containing 0.2% bovine serum albumin (BSA) and 0.4% FCS for 24 h before an experiment. This low-serum medium was then freshly replaced, and the cells were preincubated for 30 min and then treated with agonists for 4 h. RNA was extracted from the cells using RNA-STAT 60 (Tel Test "B", Friendswood, TX). Total RNA (20 µg) from each of the samples was size fractionated on 1.2% agarose gels containing 0.5% formaldehyde. The denatured RNA was then transferred to positively charged nylon membranes and hybridized to the specific ³²P-labeled VEGF cDNA probe. The plasmid containing a 930-base pair fragment of the human VEGF cDNA was a gift from Genentech (San Francisco, CA). The cDNA probe was radiolabeled with [-³²P]dCTP using random primer labeling system. Hybridization was performed in 50% formamide containing 4× saline-sodium phosphate-EDTA buffer (SSPE), 5× Denhardt's solution, 3% sodium dodecyl sulfate (SDS), and 0.5 mg/ml salmon sperm DNA at 42°C overnight. Washing conditions were 2× in SSPE-0.1% SDS at room temperature for 15 min, 1× in SSPE-0.1% SDS at 37°C for 15 min, and 0.5× in SSPE-0.1% SDS at 53°C for 15 min. Blots were exposed to Kodak film for 24 h. To control for RNA quantity and loading efficiency, blots were stripped and reprobed with a labeled glyceraldehyde 3-phosphate dehydrogenase (GAPDH) probe. The autoradiograms were analyzed by densitometry to quantitate the ratio of VEGF to GAPDH mRNA and statistically analyzed by analysis of variance using the INSTAT software (Graph Pad, San Diego, CA).

Immunoblotting to detect VEGF protein. Serum-starved VSMC in 100-mm dishes were placed in fresh DME NG or HG medium containing 0.2% BSA and 0.4% FCS. Cells were preincubated for 30 min and then treated with agonists for 4 h. Cell pellets were lysed in lysis buffer containing 1% Triton X-100, 0.1% SDS, 1 mM sodium metabisulfite, 1 mM ethylene glycol-bis(-aminoethyl ether)-N,N,N',N'-tetraacetic acid, 0.05 mM leupeptin, 5 µg/ml aprotinin, and 0.2 mM phenylmethylsulfonyl fluoride in PBS. The cell lysates were centrifuged and the cytosolic fractions (equal amounts of protein) were subjected to electrophoresis on 8% SDS gels. Protein was transferred to PVDF membranes using a semidry blotting apparatus (Hoeffer Instruments, San Francisco, CA). Nonspecific binding was blocked by incubating the membranes in a blocking buffer provided by Tropix (Bedford, MA). The blots were then incubated with a polyclonal antibody to a peptide derived from the human VEGF sequence (1:200 dilution, catalog no. sc-152, Santa Cruz Biotechnology, Santa Cruz, CA). Washed blots were then incubated with an alkaline phosphatase-labeled second antibody at a dilution of 1:30,000. Detection was by a chemiluminescent technique using the Western light detection system (Tropix). Western blots were quantitated using a computerized densitometer (SCISCAN 5000, US Biochemical), and the values were expressed as arbitrary optical density units or fold-over control.

Blocking experiments to determine the specificity of the VEGF immunoreactive bands were performed as follows: the VEGF polyclonal antibody (10 µl, 1 µg) was incubated with the VEGF antigen 20-amino acid peptide (60 µl, 12 µg; catalog no. sc-152P, Santa Cruz Biotechnology) or with 12 µl (12 µg) BSA in a final volume of 300 µl PBS for 2.5 h at room temperature with gentle shaking. The blocked and free antibodies were then diluted to 2.0 ml (1:200 final) and used to probe the blots containing the VSMC cytosol samples.

Quantitation of secreted VEGF levels by enzyme immunoassay. Preconfluent HSMC in NG or HG culture medium were serum depleted for 24 h by placing in NG or HG DME medium containing 0.2% BSA and 0.4% FCS. The cell monolayers were then rinsed with PBS, and the above-mentioned depletion medium was freshly replaced. The cells were incubated for 3 h, and then the supernatant-conditioned medium from the dishes was stored at 70°C for the measurement of secreted VEGF. VEGF in these samples was quantitated by a specific enzyme-linked immunosorbent assay (ELISA) kit (catalog no. DVE00) obtained from R&D systems (Minneapolis, MN). This assay recognizes both natural and recombinant human VEGF in cell culture supernates, serum, and plasma samples. No significant cross-reactivity or interference was observed with any of the 94 different kinds of cytokines, chemokines, and growth factors tested. All values obtained from the ELISA were corrected for the protein content in each dish.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Effect of HG culture on VEGF mRNA expression in immortalized HSMC. Figure 1 depicts a Northern blot run with total RNA from human VSMC that were cultured for three passages under NG (5.5 mM) or under HG (25 mM) conditions. Results of hybridization of the blot with a specific labeled VEGF cDNA probe are seen in Fig. 1A, top panel, and with a GAPDH probe in the bottom panel. VEGF mRNA is seen with a strong band of ~3.6 kilobase. All three samples cultured in HG showed distinctly higher levels of VEGF mRNA than those in NG samples. The bar graph in Fig. 1B shows the densitometric quantitation of the results from such multiple experiments (n = 6) and indicates that VEGF mRNA was 2.5-fold higher in cells cultured in HG compared with those in NG (P < 0.001). This effect was specific for glucose because similar cultures in mannitol, which was used as a control for osmolality, did not have any significant effect on VEGF expression (results not shown).

View larger version (34K): [in this window] [in a new window]   Fig. 1.   Effect of high glucose (HG) culture on vascular endothelial growth factor (VEGF) mRNA expression in an immortalized human vascular smooth muscle cell line (HSMC). Serum-depleted HSMC growing for at least two passages in normal glucose (NG, 5.5 mM) or HG (25 mM) medium were incubated for 4 h, and then RNA was extracted from cells. Equal amounts of RNA were subjected to Northern blotting. A: representative Northern blot probed with a specific VEGF cDNA probe (top). Blot was then stripped and probed with a glyceraldehyde 3-phosphate dehydrogenase (GAPDH) probe (bottom). B: densitometric quantitation of VEGF band from 6 experiments after correction for intensity of internal control GAPDH mRNA. * P < 0.001 vs. NG.

Regulation of VEGF mRNA expression in HSMC by ANG II and 12-HETE. We next compared the effects of ANG II on VEGF mRNA expression in the HSMC cell line cultured under NG versus HG conditions. Figure 2 shows that, as seen earlier, basal VEGF mRNA expression was higher in the cells cultured in HG versus NG. Interestingly, under NG culture conditions, a 4-h treatment with ANG II (10⁷ M) did not have a significant stimulatory effect on VEGF mRNA expression (Fig. 2A, top). In contrast, ANG II treatment led to a significant twofold increase in VEGF mRNA above basal expression under HG conditions as also seen in Fig. 2B, which is obtained by densitometric quantitation of the bands from multiple experiments and corrected for the expression of the internal control, GAPDH mRNA expression, shown in the Northern blot in Fig. 2A (bottom).

View larger version (30K): [in this window] [in a new window]   Fig. 2.   Regulation of VEGF mRNA in HSMC by angiotensin II (ANG II) and 12-hydroxyeicosatetraenoic acid (12-HETE). Serum-depleted HSMC in NG or HG were treated for 4 h with ANG II or 12-HETE (107 M each). Total RNA from the cells were subjected to Northern blotting to evaluate VEGF mRNA expression. A: representative Northern blot. B: densitometric quantitation of VEGF band from 4 experiments after correction for GAPDH mRNA expression. * P < 0.001 vs. control.

Increasing evidence suggests that 12-LO products of arachidonate metabolism can mediate the effects of ANG II and also have angiogenic properties. We therefore examined the effects of the 12-LO product 12-HETE on VEGF mRNA expression in HSMC. Figure 2 shows that treatment of the cells with 12-HETE (10⁷ M; Biomol, Plymouth Meeting, PA) for 4 h led to a significant twofold stimulatory effect under both NG as well as HG culture conditions.

Regulation of VEGF mRNA expression in PSMC by ANG II and 12-HETE. To determine whether the responses observed in the immortalized HSMC cell line could be demonstrated in primary cultures of nontransformed VSMC, we also examined whether ANG II and 12-HETE could regulate VEGF expression in primary cultures of PSMC. Figure 3A shows a representative Northern blot depicting the effects of treating PSMC cultured in NG or HG with ANG II or 12-HETE (10⁷ M each) for 4 h on VEGF mRNA expression. The bar graph in Fig. 3B shows the densitometric quantitation of the results. It is seen that, similar to the results seen in HSMC, basal VEGF mRNA expression was nearly twofold greater in the cells cultured in HG (solid bar in Fig. 3B). ANG II treatment again had no clear effect in the cells cultured in NG but led to nearly a twofold increase in VEGF mRNA expression in HG. With regard to the effect of 12-HETE, it is again seen that it increased VEGF mRNA to a similar extent (twofold) in both cells cultured in NG as well as HG.

View larger version (31K): [in this window] [in a new window]   Fig. 3.   Regulation of VEGF mRNA expression in porcine VSMC (PSMC) by HG (25 mM), ANG II, and 12-HETE. Serum-depleted PSMC in NG or HG medium were treated for 4 h with ANG II or 12-HETE (107 M each) and then processed for VEGF mRNA quantitation as described in MATERIALS AND METHODS. A: representative Northern blot. B: densitometric quantitation of blot shown. Results shown are representative of 2 separate experiments.

Regulation of VEGF protein expression in VSMC. To determine whether the increase in VEGF mRNA expression by HG and by ANG II was also associated by a corresponding increase in VEGF protein expression, we performed immunoblotting with a specific VEGF antibody on cytosolic fractions from lysates of HSMC cultured in NG or HG and treated with or without ANG II for 4 h. Results are seen in Fig. 4. The blot in Fig. 4A indicates that basal expression of VEGF protein (43-46 kDa) was higher in the cells cultured in HG than those cultured in NG. ANG II did not have any effect on VEGF protein expression in the cells cultured in NG. However, it led to a clear stimulatory effect at 10⁸ and 10⁷ M concentrations in the cells cultured in HG. The bar graph in Fig. 4B shows the densitometric quantitation of the results obtained from multiple experiments and indicates that ANG II had nearly a twofold significant stimulatory effect on VEGF protein expression only in the cells cultured in HG (hatched bar in Fig. 4B). The solid bar in Fig. 4B indicates that HG culture alone led to a significantly greater than twofold increase in VEGF protein expression over NG.

View larger version (34K): [in this window] [in a new window]   Fig. 4.   Regulation of VEGF protein expression in HSMC. Quiescent HSMC in NG or HG medium were treated for 4 h with ANG II (108 or 107 M). Cells were lysed and equal amounts of cytosolic protein were subjected to immunoblotting to detect VEGF protein expression. A: representaive blot. B: cumululative data obtained from densitometric quantitation of bands from 4 experiments. * P < 0.01 vs. HG control; ** P < 0.01 vs. NG control by analysis of variance.

We next examined the effect of a 4-h treatment with the 12-LO product 12-HETE (10⁷ and 10⁸ M) on VEGF protein expression in HSMC. Figure 5A shows a representative immunoblot. Basal VEGF expression in HG was greater than that in NG as seen before. 12-HETE treatment led to a marked increase (over twofold) in VEGF protein expression in the cells cultured in both NG as well as in HG (Fig. 5B). This is similar to the results obtained for the regulation of VEGF mRNA by 12-HETE).

View larger version (29K): [in this window] [in a new window]   Fig. 5.   Effect of 12-HETE on VEGF protein expression in HSMC cultured in NG vs. HG. Quiescent HSMC in NG or HG media were treated for 4 h with 12-HETE (108 or 107 M). Equal amounts of cytosolic fractions were subjected to immunoblotting to detect VEGF protein expression. A: representative immunoblot. B: densitometric quantitation of same blot. Results shown are representative of 2 separate experiments.

To determine whether HG culture can also lead to the upregulation of VEGF protein expression in the PSMC, we performed immunoblotting with the VEGF antibody on lysates from PSMC cells. Figure 6 shows that, similar to the results obtained with the HSMC, basal VEGF protein expression was clearly greater in the cells cultured in HG.

View larger version (72K): [in this window] [in a new window]   Fig. 6.   Effect of HG culture on VEGF protein expression in PSMC. Serum-depleted PSMC growing in NG or HG medium were incubated for 4 h, and then cytosolic fractions were subjected to immunoblotting to detect VEGF protein expression. Results shown are representative of 3 experiments.

We also performed studies to determine the specificity of the bands observed in the Western blots. Figure 7 shows that lysates from PSMC and HSMC showed distinct bands that appear as doublets around 43-46 kDa, which is the reported molecular weight of dimeric VEGF (13) (Fig. 7A). To evaluate the specificity of these bands, we carried out antibody-blocking studies. Figure 7B shows an immunoblot of the same lysates from PSMC and HSMC. However, in this case the blots were probed with VEGF antibody that had been preadsorbed with the peptide antigen. This process led to the complete disappearance of the bands (Fig. 7B), indicating the specificity of the bands for VEGF protein. Under these conditions, we did not observe any bands around 23 kDa, which represents the monomeric form of VEGF, although we have observed such low-molecular-weight bands in some experiments.

View larger version (47K): [in this window] [in a new window]   Fig. 7.   Specific expression of VEGF protein expression in PSMC and HSMC. Cytosol fractions of lysates from PSMC or HSMC were subjected to immunoblotting to detect VEGF protein expression as described in MATERIALS AND METHODS. A: probed with 1:200 dilution of free antibody. B: probed with a 1:200 dilution of the VEGF antibody that had been preadsorbed with VEGF peptide antigen.

Effect of HG culture on levels of VEGF secreted into medium by HSMC. To evaluate the effect of HG culture on the secretion of VEGF, we used a specific ELISA to evaluate the levels of VEGF peptide secreted into the culture medium. Conditioned medium from HSMC cultured in HG or NG medium containing 0.2% BSA and 0.4% FCS was analyzed for VEGF levels by ELISA. Results seen in Fig. 8 show that the cells cultured in HG secreted significantly greater amounts of VEGF into the medium. We have also observed similar results with PSMC (NG 306 ± 17 vs. HG 436 ± 13 pg/mg protein, P < 0.02).

View larger version (8K): [in this window] [in a new window]   Fig. 8.   Effect of HG culture on levels of VEGF secreted into culture medium. Serum-depleted HSMC were incubated for 4 h and conditioned medium saved for quantitation of levels of secreted VEGF by a specific enzyme-linked immunosorbent assay as described in MATERIALS AND METHODS. Results shown are means ± SE. * P < 0.01 vs. NG by analysis of variance (n = 9).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

In the present study, we have observed for the first time that chronic HG exposure significantly increased VEGF mRNA as well as protein expression in HSMC and PSMC. Furthermore, HSMC exposed to HG secreted greater amounts of VEGF into the culture medium, indicating the potential biological significance of this effect. The observed effect was not due to hyperosmolality, since culture of the cells in mannitol did not upregulate VEGF expression. In addition, we also noted for the first time that the LO metabolite of arachidonic acid, 12-HETE, could directly induce VEGF mRNA and protein expression in VSMC cultured in both NG as well as HG.

ANG II is a potent vasoactive agent as well as a growth-promoting agent for VSMC and has therefore been implicated in the pathogenesis of hypertension and atherosclerosis (7, 11, 23, 28). The vasoactive effects of ANG II have been shown to be enhanced in the diabetic state (41), and furthermore, the growth-promoting effects of ANG II were enhanced in VSMC cultured under HG conditions (25). In the present study, ANG II led to an increase in VEGF mRNA and protein expression only in VSMC cultured under HG conditions. Endothelial dysfunction accompanied by an increase in endothelial permeability is a key early event in the pathogenesis of atherosclerosis (30). Earlier studies have indicated that ANG II could induce the expression of the potent vascular permeability factor VEGF in VSMC, thus indicating a mechanism whereby ANG II could affect endothelial function by influencing vascular permeability (42). Our new observations that ANG II-induced VEGF expression is clearly greater in the VSMC cultured in HG raise the possibility for a novel mechanism for the accelerated cardiovascular disease observed in diabetes. Whereas the concentration of ANG II used in these studies is higher than those seen in vivo, these doses are typically employed for ANG II in in vitro studies mainly due to the lability of ANG II in tissue culture medium. It is also likely that local concentrations of ANG II generated in the vasculature may be significantly higher than the circulating levels.

Recent evidence has indicated that VEGF expression can also be induced by acute and rapid hypoglycemia as well as hypoxia in glioma cells (33, 37) and in retinal pigmented epithelial cells (34). Other ischemia-induced genes were also induced in the glioma cells, and it was suggested that these proteins satisfy the tissue needs by stimulating local blood supply (33, 37). Furthermore, in the retinal epithelial cells, Sone et al. (34) observed that chronic hyperglycemia also upregulated VEGF. This is similar to our present results in VSMC. In the context of diabetes, Sone et al. suggested that upregulation of VEGF by chronic HG can lead to the worsening of proliferative diabetic retinopathy. On the other hand, the induction by acute and rapid glucose deficiency represents a stressful situation, which may aggravate diabetic retinopathy. This is supported by observations that although strict glycemic control is clearly associated with reduced rates of proliferative diabetic retinopathy, achieving this goal too rapidly through very intensive treatment may induce progression (albeit transiently) of proliferative diabetic retinopathy. Tilton et al. (40), who recently showed that vascular dysfunction induced by HG in rats is mediated by VEGF, suggested that HG induces VEGF in tissues as a result of hypoxia-like redox imbalance resulting from increased flux of glucose via the sorbitol pathway.

Treatment of vascular cells with HG has been shown to upregulate the production of growth factors such as PDGF-BB and transforming growth factor- (21, 22). Chronic HG culture can also lead to the formation of advanced glycation products (3), which in turn could induce production of growth factors and cytokines. Recent studies indicate that key VSMC growth factors such as PDGF-BB and fibroblast growth factor could upregulate VEGF expression in VSMC and also displayed a synergistic effect with hypoxia (35, 36). It is possible that the observed HG-induced VEGF may arise from the action of these cytokines or growth factors. However, further studies are needed to confirm this hypothesis.

LO products of arachidonate metabolism such as 12- and 15-HETE have been shown to have angiogenic properties (17, 32). The 12-LO pathway has been shown to play a key role in the growth-promoting and vasoactive effects of ANG II (25, 38), and LO products also have direct growth-promoting effects in VSMC (25). We have also shown that HG culture of VSMC can upregulate the activity and expression of the 12-LO enzyme (26). 12-LO products have potent chemotactic effects and can also upregulate key growth-related signaling kinases (24, 29). The present studies showing that 12-HETE can induce VEGF expression indicate another mechanism of action linking these lipid metabolites to vascular disease. Because 12-HETE increased VEGF expression in NG as well as in HG, it is likely that there are distinct mechanisms for 12-HETE and HG such as activation of specific isoforms of protein kinase C. Activation of the LO pathway can also lead to the formation of reactive oxygen species (31), and a very recent report has demonstrated that reactive oxygen species can increase VEGF gene expression (15). Therefore, the formation of reactive oxygen species may be a common pathway linking activation of VEGF by various lipids, growth factors, and glucose, since it is now clear that ANG II, PDGF, and HG can lead to the formation of reactive oxygen species such as hydrogen peroxide and superoxide (12, 39, 44).

Thus a combination of physicochemical changes in the arterial wall, along with the growth factors and cytokines released on endothelial injury, can lead to the induction of VEGF. In turn, VEGF can promote early atherogenesis by increasing endothelial permeability and monocyte migration and also promote the late stages of atherosclerosis by acting as an angiogenic factor leading to plaque neovascularization (16).

In summary, our data showing the induction of VEGF by hyperglycemia, potentiation of VEGF expression by ANG II in HG, and induction of VEGF by the lipid 12-HETE indicate novel mechanisms for the accelerated vascular complications observed in diabetes. An improved understanding of the exact function of VEGF in VSMC and the mechanisms of its action may lead to new therapeutic modalities to combat diabetic complications.