Enhancement of Glucose Transport by Vascular Endothelial Growth Factor in Retinal Endothelial Cells

Abstract

purpose. To investigate effects of vascular endothelial growth factor (VEGF) on glucose transport and GLUT1 glucose transporter expression in primary bovine retinal endothelial cell (BREC) cultures.

methods. Glucose transport in control and VEGF-treated BREC cultures was determined by measurement of[ ¹⁴C]-3-O-methylglucose (3MG) uptake. GLUT1 protein and mRNA was determined by Western and Northern blot analyses, respectively. Protein kinase C (PKC) activity was measured in control and VEGF-treated cultures, and glucose transport was determined with and without prior PKC depletion and PKC inhibition.

results. Dose-dependent increases in 3MG uptake were seen in the VEGF-treated cultures, with an increase of 69% after a 24-hour exposure to 50 ng/ml VEGF (P < 0.001). Total cellular GLUT1 mRNA or protein, however, was unchanged. Western blot analysis of plasma membrane fractions revealed a 75% increase in plasma membrane GLUT1 in VEGF-treated cultures (P = 0.02), suggesting that the VEGF-stimulated increase in glucose transport was due to a translocation of GLUT1 to the cell membrane. VEGF stimulated a 90% increase in PKC activity in membrane fractions from cultures treated with VEGF, and VEGF-stimulated enhancement of glucose transport was abolished by cellular PKC depletion and by general and PKC β inhibition.

conclusions. The present study demonstrates VEGF-mediated enhancement of retinal endothelial cell glucose transport and suggests that this increase is due to PKC β–mediated translocation of cytosolic GLUT1 to the plasma membrane surface. Upregulation of retinal endothelial cell glucose transport by various factors associated with the development of retinopathy may be responsible for the metabolic derangements observed in the diabetic inner blood–retinal barrier in vivo.

The underlying molecular processes that give rise to the pathologic microvascular changes of diabetic retinopathy (DR) have yet to be fully elucidated. The association between duration of diabetes and elevated glycosylated hemoglobin values with increased risk of DR ¹ on the one hand, and the substantial decrease in risk associated with intensive diabetic management ² on the other, indicate the importance of chronic exposure of the retinal microvasculature to elevated blood glucose concentrations. Several pathways have been proposed to play a role in linking the hyperglycemia of diabetes with the characteristic microvascular lesions of DR. ^{3 4} These pathways include elevated protein kinase C (PKC) activity, ⁵ oxidative stress, ⁶ nonenzymatic glycation, ⁷ and direct toxic effects of glucose on endothelial cell replication and viability. ⁸ A common denominator among these pathways, however, is the exposure of the intracellular milieu of the endothelial cell to elevated glucose concentrations: indeed, increased glucose flux into the endothelial cell has been assumed in the conceptualization of several of these pathways. ^{3 8}  

Glucose entry into and through the retinal endothelial cells of the inner blood–retinal barrier (BRB) occurs exclusively through the GLUT1 glucose transporter, ^{9 10 11 12} a member of the sodium-independent glucose transporter family. ^{13 14} GLUT1, which is characteristically expressed in cells that serve barrier functions, ^{9 10 15 16} such as the brain capillary endothelia of the blood–brain barrier (BBB) and the endothelia and retinal pigment epithelium of the inner and outer BRB, is a high-affinity transporter (1–5 mM), and transport through GLUT1 is therefore at near-saturation levels at normal physiological glucose concentrations. ^{17 18 19 20} In the absence of changes in the density of GLUT1 at the endothelial membrane, increasing extracellular glucose concentrations may do little to increase intracellular glucose, a critical molecular event in the development of DR. In contrast, an increase in inner BRB GLUT1 expression may have a profound impact in providing substrate for the molecular processes leading to retinal microvascular disease. In a previous study, ²¹ a dramatic focal increase of GLUT1 abundance was demonstrated on the inner BRB of individuals with long-standing diabetes without clinical evidence of retinopathy. These results suggest that the earliest stages in the development of DR are associated with a localized increase in inner BRB GLUT1 density. 

Only a few of the principal factors that modulate inner BRB GLUT1 expression have been identified. Takagi et al. ²² have reported recently that in vitro, hypoxia increases GLUT1 protein and mRNA and glucose transport in bovine retinal endothelial cell (BREC) cultures, a process that is regulated by adenosine and the cyclic adenosine monophosphate (cAMP)–PKA pathway. To investigate regulation of retinal endothelial cell glucose transport and GLUT1 expression by growth factors that are known to be associated with the development of DR, we studied the effects of vascular endothelial growth factor (VEGF) on BREC glucose transport and GLUT1 expression. VEGF, a cytokine that has endothelial cell–specific mitogenic and angiogenic properties, ²³ has been demonstrated to play a major role in the proliferative stages of DR. ^{24 25} Although there is some controversy, ²⁶ numerous studies have documented increased retinal VEGF expression in nonproliferative DR as well. In both human diabetic retina ^{27 28 29} and the retina of experimental animal models of DR, ^{30 31} VEGF is increased in the earliest stages of retinopathy, in which it may act to increase microvascular permeability ³⁰ and changes in microvascular blood flow. ³² In the present study we report a PKC-mediated increase in BREC glucose transport in response to VEGF. Unlike the situation in hypoxia, however, VEGF-mediated increases in glucose transport were not the result of increased total cellular GLUT1 abundance but appeared to follow VEGF-mediated translocation of GLUT1 to the cytoplasmic membrane. 

Methods

Cell Cultures

Primary bovine retinal endothelial cell (BREC) cultures were established from fresh calf eyes. Under sterile conditions, the retinas were isolated and extensively rinsed in Dulbecco’s modified Eagle’s medium (DMEM, Gibco, Grand Island, NY) and suspended in an enzyme solution containing collagenase (500 μg/ml) and pronase (100 μg/ml) in 10 mM phosphate-buffered saline (PBS; all from Sigma, St. Louis, MO). Pieces of adherent retinal pigment epithelium were dissected, and the retinas were minced and incubated with shaking at 37°C for approximately 25 minutes. Sequential sieving was performed over 210- and 52-μm nylon mesh, and the remaining retinal capillary fragments were plated and grown in DMEM with 15% fetal calf serum (FCS), endothelial growth supplement (EGS, 100 μg/ml; Sigma), heparin (88μ g/ml), and antibiotic–antimycotic solution (Sigma) on fibronectin-coated dishes in 5% CO₂ at 37°C. Cultures were passaged every 8 to 10 days, and nearly confluent (80% to 90%) cultures from passages two to six were used for all experiments. Purity of cultures was confirmed by either more than a 90% uptake of acetylated low-density lipoprotein (LDL) or a more than 90% immunopositivity for factor VIII. 

On the day before each experiment, the medium was changed to DMEM with 0.5% FCS and antibiotics without EGS. The next day, these media were exchanged with either fresh medium alone or medium containing the required concentrations of human recombinant VEGF (R&D Systems, Minneapolis, MN). 

Glucose Transport Studies

Uptake of the [¹⁴C]-labeled glucose analogue, 3-O-methylglucose ([¹⁴C]-3MG, kindly provided by C. Carter-Su (University of Michigan), was performed according to the protocol of Tai et al., ³³ with modifications. Briefly, BREC cultures were grown to near-confluence in 35-cm² fibronectin-coated dishes. After treatment with VEGF, the media from control and VEGF-treated cultures were aspirated and replaced with low-calcium Krebs–Ringer phosphate buffer containing 1% bovine serum albumin (KRP-BSA) and incubated at 37°C for 30 minutes to deplete intracellular glucose concentrations. The cultures were washed twice with KRP-BSA at room temperature for 5 minutes. The assay was performed by incubation of each cell culture with 1 ml KRP-BSA containing 0.25μ Ci/ml [¹⁴C]-3MG (specific activity, 55.2 mCi/mmol) for 7 seconds at room temperature. To assess nonspecific binding, parallel cultures were preincubated with KRP-BSA containing 40μ M cytochalasin-B (Sigma) for 5 minutes at room temperature, followed by incubation with the [¹⁴C]-3MG solution containing cytochalasin-B. Maximum uptake was assessed by incubation of parallel cultures with the radioisotope for 90 seconds. Uptake was terminated by rapid aspiration of the isotope with repeated washes with ice-cold phloretin (0.2 mM phloretin in KRP without BSA, Sigma), an inhibitor of glucose transport. Cells were solubilized in 0.1% SDS and 0.1 M NaOH at 60°C for 30 minutes. Protein was measured by a BCA assay (Pierce, Rockford, IL). ¹⁴C was measured by liquid scintillation counting (Tri-Carb Scintillation Counter; Packard Instruments, Downers Grove, IL). 

Initial hexose uptake rates were calculated according to the methods of Carter-Su and Okamoto, ³⁴ using the formula devised by Foley et al. ³⁵  

\[dU/dt(\mathrm{at}\ t{=}0){=}(U_{\mathrm{max}}){\times}{[}\mathrm{ln}\ U_{\mathrm{max}}-\mathrm{ln}\ (U_{\mathrm{max}}-U){]}/t\]

where U equals the uptake at time t and U _max equals maximum uptake. Initial uptake rates were first calculated using this formula in triplicate control BREC cultures for a variety of time points (2, 5, 10, 30, 60, 90, and 300 seconds). These initial studies demonstrated linear uptake between 5 and 10 seconds, and therefore a 7-second incubation was used for all experiments. Specific transport of[ ¹⁴C]-3MG was calculated by subtracting the average radioactivity in cytochalasin-B cultures from the total radioactivity in each sample. For each experiment, n = 6 to 7 cultures in each group were used. 

Crude Membrane Preparations

Total cellular membranes were isolated from control and VEGF-treated BREC cultures according to the methods of Kaiser et al. ³⁶ Briefly, the media of control and VEGF-treated cultures were aspirated, and the cultures were washed with cold PBS. Four milliliters of homogenization buffer (0.25 M sucrose, 10 mM NaHCO₃, 5 mM NaN₃, and 0.1 mM phenylmethylsulfonyl fluoride) was added to each culture, and the cells were scraped and transferred to a 10-ml glass tissue grinder. The cells were homogenized by 60 strokes of a Teflon-coated pestle, followed by centrifugation at 1200g for 10 minutes at 4°C. The supernatant was subsequently collected and centrifuged at 9000g for 10 minutes to pellet out cellular organelles and nuclei. The total cellular membranes in the resultant supernatant were collected by centrifugation at 100,000g for 60 minutes at 4°C and resuspended in Laemmli loading buffer withoutβ -mercaptoethanol (β-ME). 

Plasma Membrane Preparations

Plasma membranes were isolated according to the methods of Jaffe et al. ³⁷ Briefly, the media from control and treated cultures grown in 150-cm² flasks were aspirated, and the cultures were thoroughly washed with ice-cold PBS. Ten milliliters of ice-cold PBS containing protease inhibitors was added to each flask, and the cells scraped from two flasks were pooled in a chilled 50-ml tube. The cells were pelleted, resuspended in PBS with protease inhibitors and homogenized with 20 strokes in a homogenizer (Dounce; Kontes Glass, Vineland, NJ). The homogenate was centrifuged at 1800g at 4°C for 10 minutes, and the pellet was resuspended, homogenized again and centrifuged as before. The two supernatants were pooled and centrifuged at 30,000g, 4°C, for 30 minutes. The pellet, which contained nuclei, cellular membranes, and organelles, was resuspended in 250 mM sucrose in 10 mM Tris, layered on a discontinuous sucrose gradient, and centrifuged at 55,000 rpm at 4°C for 120 minutes. The interfaces containing the plasma membranes (i.e., between 20% and 27% and between 32% and 40%) were carefully aspirated, resuspended in PBS with protease inhibitors, and collected by another centrifugation at 55,000 rpm at 4°C for 60 minutes. The final pellet was resuspended in Laemmli loading buffer without β-ME. 

Protein Measurements

Protein concentrations were measured by BCA assay (Pierce) with BSA as a standard. For Western blot samples, β-ME was added after protein measurement to a final concentration of 5%. 

Western Blot Studies

For studies involving total cell lysates, electrophoresis was performed on 25-μg aliquots of each sample on a 10% SDS-polyacrylamide gel and transferred to polyvinylidene fluoride (PVDF) membranes. For total cell membranes and plasma membranes, aliquots of 7.5-μg aliquots were used. An aliquot of 7.5 ng of human erythrocyte glucose transporter, purified from whole blood as described previously, ^{38 39} was loaded on each gel for comparison of molecular mass. Western blot analysis was performed as previously described ⁴⁰ with a rabbit polyclonal anti-human erythrocyte glucose transporter antiserum (a kind gift of C. Carter-Su, University of Michigan) at a concentration of 1:10,000. This antibody has been characterized previously. ³³ For the plasma membrane preparations, Western blot analysis for NaK-ATPase was performed using a polyclonal rabbit antiserum (UBI, Lake Placid, NY) at a dilution of 1:4000. After washing, the blots were reacted with an anti-rabbit secondary antibody coupled to horseradish peroxidase at a concentration of 1:7500, followed by development with chemiluminescence reagents (ECL or ECL Plus; Amersham Pharmacia, Piscataway, NJ), according to the manufacturer’s instructions. Blots were exposed to radiographic film (X-OMAT; Eastman Kodak, Rochester, NY), and the autoradiographs were scanned and quantified by computer (NIH Image software; National Institutes of Health, Bethesda, MD). After development of the autoradiograms, each Western blot was stained with Coomassie blue to confirm equal transfer of total proteins to the membrane. 

Northern Blot Studies

Twenty-five-microgram aliquots of total RNA isolated from confluent 75-cm² flasks of control and VEGF-treated (50 ng/ml for 24 hours) BREC cultures were isolated (RNeasy; Qiagen, Valencia, CA), loaded on a 1.1-M formaldehyde 1% agarose gel, and run overnight at 20 V. RNA was transferred to a membrane (Genescreen Plus; Dupont–NEN, Boston, MA), and the membrane was dried in a vacuum at 80°C for 2 hours. Northern blot analysis for GLUT1 and mouse actin were performed as previously described, ³⁹ using a 512-kb PstI fragment of the bovine blood–brain barrier glucose transporter cDNA ⁴¹ linearized with HindIII and with a mouse actin clone, pAM-91 (generously provided by Michael J. Getz, Mayo Clinic/Foundation, Rochester, MI), linearized with EcoRI. Both cDNAs were labeled with[ ³²P]-dCTP using a random primer method, as described previously. ³⁹  

PKC Studies

Partially purified PKC was prepared from cytosolic and total membrane preparations of BREC cultures according to the method of Xia et al. ⁴² Total PKC activity was determined in the cytosolic and membrane preparations by measurement of the transfer of ³²P from [γ-³²P]ATP (100 μCi/mmol) to a PKC-specific peptide (neurogranin_(28–43)) ⁴³ in the presence of phosphotidylserine, Ca²⁺, and diacylglycerol. ⁴² Involvement of PKC with VEGF-mediated effects on BREC glucose transport was studied in BREC cultures that had previously undergone PKC depletion or preincubation with LY379196, a selective inhibitor of PKC-β (kindly provided by K. Ways, Eli Lilly, Indianapolis, IN). Cellular PKC depletion ⁴⁴ was achieved by overnight incubation of the cultures with 1 μM phorbol 12-myristate 13-acetate (TPA, Sigma) before addition of the growth factor. The inhibitory profile of LY379196 for various PKC isoforms is similar to that of LY333531, which has been used in various in vitro studies of PKC-mediated pathogenic mechanisms in diabetic vascular and renal complications ^{45 46 47 48} (Table 1) . LY379196 shows general PKC inhibition at a concentration of 600 nM; at 30 nM, LY379196 demonstrates PKC-β1 and β2-selective inhibition, with median effective doses (ED₅₀) of 0.05 and 0.03 μM, respectively (J. R. Gillig, Eli Lilly). For PKC inhibition, BREC cultures were preincubated with the inhibitor for 60 minutes at 37°C before the addition of recombinant VEGF. 

Statistical Analysis

All the results are expressed as means ± SEM. Comparisons between control and VEGF-treated cultures for the 3MG transport assays, Western blot analyses, and PKC activity assays were performed by Student’s t-test. Comparisons between multiple groups in 3MG transport experiments involving PKC depletion or inhibition and in Northern blot analyses were performed by analysis of variance (ANOVA). For all 3MG transport assays, n = 6 to 7 for each group tested. Each transport experiment was performed at least three times. For Western blot analysis, n = 3 to 4 for each group. Western and Northern blot analyses and PKC activity assays were repeated a minimum of three times on separate cell cultures. P < 0.05 was considered significant for all experiments. 

Results

Exposure of the BREC cultures to recombinant VEGF resulted in a dose-dependent increase in initial 3MG transport rates (Fig. 1A ). Increased 3MG uptake in response to VEGF was seen at 5 hours and was maximal after 24 hours (data not shown). Exposure of BREC to 50 ng/ml for 24 hours resulted in a 69.2% ± 14.4% increase in initial transport of 3MG (P < 0.001, Fig. 1B ). 

To determine whether the observed VEGF-mediated increases in BREC glucose transport were due to increased abundance of endothelial cell GLUT1, quantitative Western blot analysis was performed on total cell lysates and crude total cellular membrane preparations from control and VEGF-treated BREC cultures. Treatment with 50 ng/ml VEGF for 24 hours did not result in a significant increase in total cellular GLUT1 (Fig. 2A ) or in cellular total membrane GLUT1 (Fig. 2B) . Similarly, there was no significant increase in GLUT1 mRNA abundance in BREC cultures exposed to 50 ng/ml VEGF at any time point up to 24 hours (Fig. 2C) . Quantification of arbitrary densitometric units from GLUT1 and actin Northern blot analysis from three separate experiments showed no significant difference in GLUT1–actin ratios between control and VEGF-treated cultures for any time point (data not shown). 

A possible explanation for the discrepancy between a VEGF-mediated increase in glucose transport rate in the absence of increased total cellular GLUT1 could be VEGF-mediated translocation of preexisting glucose transporters from cytoplasmic stores to the plasma membrane. As has been demonstrated, ²¹ approximately 50% of GLUT1 in the human inner BRB is localized to the cytoplasm; therefore, translocation of even a fraction of these transporters to the plasma membrane would result in a significant increase in glucose transport without a change in the total abundance of GLUT1 protein or mRNA. To test this hypothesis, plasma membranes were isolated from control BREC cultures and cultures treated with VEGF (50 ng/ml for 24 hours). Western blot analysis for the α isoform of NaK-ATPase, which is localized to the plasma membrane, ⁴⁹ demonstrated a greater than 20-fold enrichment of the plasma membrane fraction in control preparations (Fig. 3A ). Quantitative Western blot analysis from control and VEGF-treated cultures revealed a 75.4% ± 32.0% increase in plasma membrane GLUT1 in the VEGF-treated cultures (P = 0.02, Figs. 3B 3C ). 

Because the actions of VEGF are thought to be, in part, due to activation of PKC, and in particular, the β isoform, ^{46 47} the role of PKC in VEGF-mediated increases in retinal endothelial cell glucose transport was investigated. PKC isoforms α, βI, βII, and ε have been demonstrated in the rat retina, ⁵⁰ and α and βII isoforms have been documented in primary BREC cultures. ⁵⁰ Stimulation of BREC cultures by VEGF resulted in an increase in total PKC activity in the membrane fraction of VEGF-treated cultures by 90% (P < 0.01, Fig. 4 ). Total PKC activity in the cytosolic fraction of VEGF-treated cultures did not change significantly (Fig. 4) . 

Depletion of cellular PKC by chronic exposure to phorbol ester abolished the VEGF-mediated increase in BREC glucose transport (Fig. 5A ). Treatment of BREC cultures with 1 μM TPA alone resulted in a modest, statistically insignificant, increase in[ ¹⁴C]-3MG uptake. Prior experiments involving measurement of PKC activity in BREC cultures after 8, 16, and 24 hours of stimulation with TPA demonstrated a decrease in cytosolic PKC activity of more than 95% at 16 hours (data not shown), thus verifying near-total PKC depletion by overnight treatment with TPA in this cell type. No significant increase in [¹⁴C]-3MG uptake was seen in VEGF-treated cultures after preincubation with the LY379196 inhibitor at a concentration that resulted in generalized PKC inhibition (600 nM, Fig. 5B ). Similarly, inhibition of VEGF-mediated increases in BREC glucose transport was seen by preincubation with LY379196 at a β isoform-selective concentration of 30 nM (Fig. 5B) . Exposure of the BREC cultures to the LY379196 inhibitor alone at 30 or 600 nM had no significant effect on BREC glucose transport (Figs. 5A 5B) . 

Discussion

In the experiments in the present study, we examined the effects of VEGF on retinal endothelial cell glucose transport and GLUT1 abundance in an in vitro model of the inner BRB. Our results demonstrate that in addition to its effects on retinal endothelial cell proliferation and microvascular permeability, VEGF had the ability to upregulate retinal microvascular glucose transport and that it did so through activation of PKC and in particular, the β isoform. This enhancement of glucose transport is observed at concentrations similar to those reported in the vitreous of patients with proliferative diabetic retinopathy ²⁵ ; however, because intraretinal production of VEGF occurs in cells that are contiguous to the retinal capillary endothelia, such as the Müller cells, ²⁹ and perhaps even occurs in an autocrine manner by endothelial cells themselves, ⁵¹ intravitreal VEGF concentrations may actually represent an underestimation of the levels of VEGF to which the retinal endothelial cells are exposed within the living retina. Therefore, it is possible that the concentrations of VEGF used in these studies may be present within the retina before the onset of proliferative retinopathy. 

The effect of VEGF on retinal endothelial cell glucose transport is not that of a nonspecific response of cellular metabolism to a mitogen. Although mitogenic factors such as basic fibroblast growth factor (bFGF), tumor necrosis factor (TNF)-α, ^{52 53} phorbol esters, ⁴⁴ and transformation ⁵⁴ are known to cause increased glucose transport and/or GLUT1 abundance in a variety of cell types, glucose transport in endothelial cells is not responsive to insulin, ^{53 55} which acts as a major growth factor in the central nervous system during development. ⁵⁶ Indeed, GLUT1, which is the predominant glucose transporter in the inner BRB in vivo, ^{9 10 11 12} is not insulin sensitive. ¹⁴ Given evidence that VEGF may act as a survival factor in the retina during development, ⁵⁷ the ability of VEGF to regulate endothelial cell glucose transport in conjunction with proliferation may serve to ensure adequate substrate delivery as well as blood flow during development. Maintenance of adequate nutrient transport to the retina during development and after maturation is of critical importance, because neuroretinal metabolism is completely dependent on glucose. 

VEGF increases retinal endothelial cell glucose transport, not through an increase in total cellular GLUT1 transcript and protein, but by an apparent translocation of preexisting cytoplasmic transporters to the plasma membrane (Fig. 3) . In this sense, the actions of VEGF on GLUT1 are similar to those of insulin on GLUT4 in insulin-sensitive tissues. ⁵⁸ The VEGF-stimulated increase in glucose transport in the absence of an increase in total cellular abundance of GLUT1 in retinal endothelial cells is in apparent contrast to its effects on primary cultures of bovine aortic endothelial cells (BAECs), in which exposure to comparable concentrations of VEGF results in an approximate threefold increase in 2-deoxyglucose uptake and a fivefold increase in GLUT1 transcript. ⁵³ In the present studies, exposure of BREC to VEGF at comparable concentrations for up to 24 hours did not result in a statistically significant difference in GLUT1 mRNA compared with control cultures (Fig. 2C) . Differential effects of VEGF on aortic and retinal endothelial cells were not directly compared in these studies, which concentrated on the effects of this cytokine on glucose transport in a microvascular endothelial cell type associated with diabetic complications. Nonetheless, one may speculate that the discrepancy in the results of the present study with those reported by Pekala et al. ⁵³ may be due to inherent differences in endothelia isolated from microvascular versus macrovascular sources. In this regard, Thieme et al. ⁵⁹ have demonstrated that although BRECs and BAECs possess the same types of high-affinity receptors for VEGF, BRECs possess a threefold higher density of these receptors than do BAECs. The differences in VEGF receptor abundance, or perhaps the relative levels of expression of the different receptors for VEGF, in retinal and aortic endothelia may account for the different VEGF-mediated responses in glucose transport and GLUT1 expression in these two cell types. 

Activation of PKC by hyperglycemia, presumably through de novo synthesis by diacyl glycerol, has been proposed as one of the principal biochemical pathways responsible for the development of diabetic microvascular complications. ³ The actions of VEGF in binding to its receptors on endothelial cell membranes are in part mediated by activation of PKC. ⁴⁶ These actions include changes in retinal blood flow, ³² microvascular permeability, ⁴⁷ and endothelial cell mitogenesis. ⁴⁶ The present study demonstrates that VEGF-mediated increases in retinal endothelial cell glucose transport occur through activation of PKC. This conclusion is supported by increased localization of PKC to the plasma membrane in VEGF-stimulated BREC cultures (Fig. 4) and abrogation of VEGF-stimulated increases in glucose transport by depletion of PKC intracellular stores (Fig. 5A) and by generalized inhibition of PKC (Fig. 5B) . The observation of the present study that VEGF increases PKC activity in BREC cultures is in close agreement with that of Xia et al., ⁴⁶ who have documented similar effects in bovine aortic endothelial cells. Furthermore, the demonstration of the ability of the β isoform–selective inhibitor LY379196 to abolish VEGF-stimulated increases in BREC glucose transport (Fig. 5B) suggests that VEGF’s actions in modulating retinal endothelial glucose transport are mediated by PKC-β, the PKC isoform that is thought to be responsible for characteristic changes in retinal blood flow ⁴⁵ and microvascular permeability ⁴⁷ observed in experimental models of diabetes. 

The principal factors modulating retinal endothelial cell GLUT1 expression have yet to be fully elucidated. In a recent publication, Takagi et al. ²² have demonstrated that hypoxia causes an eightfold increase in GLUT1 mRNA and two- and threefold increases in 2-deoxyglucose transport and immunoreactive GLUT1, respectively, in BREC cultures after a 12-hour exposure to hypoxic conditions. With regard to the direct effect of glucose on retinal endothelial glucose transport and GLUT1 expression, Mandarino et al. ⁶⁰ have reported no change in 3MG transport in BREC cultures exposed to elevated glucose concentrations for 5 days. In Mandarino et al., however, changes in the abundance of BREC GLUT1 mRNA and protein were not reported. Nonetheless, focal upregulated immunoreactive GLUT1 expression has been documented in the human diabetic inner BRB. ²¹  

Because hyperglycemia per se does not appear to cause an increase in glucose transport nor in GLUT1 expression, it is unlikely that hyperglycemia-mediated changes in glucose transport and/or GLUT1 expression represent the initiating event in the molecular processes underlying the development of DR. Although speculative at this point, it is possible that in the setting of long-standing diabetes, interactions of growth factors or advanced glycation end products with their respective endothelial cell receptors ^{61 62} or the interaction of these receptors with cell surface integrins ^{63 64} may initiate processes that upregulate glucose transport and GLUT1 expression on the endothelial cell surface. This increase in glucose flux into the endothelia of the inner BRB may have toxic effects on the endothelial cells by exposure of the intracellular environment to elevated glucose concentrations. We propose that hypoxia, elevated VEGF production, and other as yet unidentified factors associated with the development of diabetic retinopathy contribute to causing an upregulation of glucose transport in the endothelial cells of the diabetic inner BRB and that this enhancement exacerbates the deleterious effects of hyperglycemia on the retinal microvasculature. ¹⁶  

 Supported by National Eye Institute Grant EY000369 (AKK); Juvenile Diabetes Foundation, Atlanta, Georgia (AKK); the Midwest Eye Banks and Transplantation Center, Ann Arbor, Michigan (AKK); Japan Society for the Promotion of Science, Tokyo; the Mochida Foundation for Medical and Pharmaceutical Research, Tokyo (HS); a generous gift from Thelma Prior (AKK); and National Institutes of Health Grant 5PO60DK-20572 to the Michigan Diabetes Research and Training Center. HS is a recipient of postdoctoral fellowship awards from the Japan Society for the Promotion of Science and the Mochida Foundation for Medical and Pharmaceutical Research.

 Submitted for publication May 27, 1999; revised December 29, 1999; accepted January 26, 2000.

 Commercial relationships policy: N.

 

Table 1.

 

Selectivity of LY379196 for PKC Isoenzymes and Other Kinases

Enzyme	ED50 (μM)
PKC-α	0.6
PKC-β1	0.05
PKC-β2	0.03
PKC-γ	0.6
PKC-δ	0.7
PKC-ε	5
PKC-ζ	48
PKC-μ	0.3
Cyclic AMP kinase	>100
Ca+2-calmodulin kinase	5
Casein kinase	>100
Src tyrosine kinase	4.4

 Each of the human PKC isoenzymes was purified from Sf9 cells in a baculovirus system. Activity was assessed by measurement of the incorporation of [³²P]-ATP into either histone or myelin basic protein. Data kindly provided by J. R. Gillig, Eli Lilly Research Laboratories, (Indianapolis, IN). ED₅₀, median effective dose.

Figure 1.

 

Initial 3MG uptake rates in VEGF-stimulated BREC cultures. (A) Dose–response curve of BREC 3MG uptake after exposure for 24 hours to increasing concentrations of recombinant human VEGF (hVEGF). Data are representative of one of three separate experiments, each of which had six to seven cultures per group. Uptake at 50 ng/ml (∗) was significantly different from control values (P < 0.04). (B) VEGF-stimulated 3MG uptake after exposure to 50 ng/ml recombinant VEGF for 24 hours. Data are the mean ± SE of 3MG uptakes, normalized for control values, from five separate experiments, each of which had six to seven samples in each experimental group (P < 0.001).

Figure 2.

 

GLUT1 Western and Northern blot analysis of control and VEGF-treated BREC cultures. BREC cultures were treated with 50 ng/ml recombinant VEGF for 24 hours. (A) Left: Representative GLUT1 Western blot of triplicate 25-μg aliquots of solubilized whole-cell lysates from control (Cont.) and VEGF-treated (VEGF) cultures. Approximate molecular mass is to the right of the blot. A 7.5-ng aliquot of purified human erythrocyte glucose transporter (hGT) was loaded for comparison of molecular mass. (A) Right: Quantification of arbitrary densitometric units from GLUT1 Western blot depicted at left. Western blot analysis and results are representative of three separate experiments, each experiment with n = 3 to 4 in each group. (B) Representative GLUT1 Western blot of whole-cell membranes in control (C) and VEGF-stimulated (V) BREC cultures under conditions identical with those in (A). (C) Representative Northern blot analysis of total RNA from BREC cultures treated without (C) or with (VEGF) 50 ng/ml recombinant VEGF for the times indicated. Top: GLUT1; bottom:β -actin.

Figure 3.

 

Western blot of plasma membrane fractions from BREC cultures. (A) Representative Western blot for the α subunit of NaK-ATPase of 25-μg aliquots from whole-cell lysates (WCL) or plasma membrane (PM) fractions. (B) Representative GLUT1 Western blot analysis of 7.5-μg aliquots of plasma membrane preparations in control (Cont.) and VEGF-treated (VEGF) BREC cultures. (C) Immunoreactive GLUT1 in control (Cont.) and VEGF-stimulated (VEGF) BREC cultures. Data are mean ± SEM of ratios, expressed as a percentage of control arbitrary densitometric units of Western blot analysis of control and VEGF-treated cultures from five separate experiments. Each experiment consisted of two to four pooled 150-cm² flasks of cells for n = 1 experiment (P = 0.02).

Figure 4.

 

VEGF effects on BREC PKC translocation. Total PKC activity (expressed as picomoles per milligram per minute of ³²P transferred to a PKC-specific substrate) measured in cytosolic and membrane fractions from BREC cultures treated with (VEGF) or without (Cont.) 50 ng/ml VEGF for 24 hours. Results are means ± SEM derived from three separate experiments.

Figure 5.

 

3MG uptake in control and VEGF-treated BREC cultures. Effects of PKC. (A) Representative 3MG assay from control (Cont.) and VEGF-treated (VEGF, 50 ng/ml for 24 hours) cultures as well as VEGF-treated cultures exposed overnight to 1 μM TPA for cellular PKC depletion. (B) Representative 3MG assay from control (Cont.) and VEGF-treated (VEGF) cultures with and without LY379196, a selective inhibitor of PKC-β. BREC cultures were treated with the inhibitor at both non–β-selective (600 nM) and β-selective (30 nM) concentrations. Data in both (A) and (B) are normalized against control cultures and expressed as the mean ± SE of control 3MG uptake for six to seven samples in each experimental group. Data are representative of three separate experiments, each with n = 6 to 7 per group.

The authors thank Christin Carter-Su for her gift of anti-GLUT1 antisera; [¹⁴C]-3MG, Kirk Ways and Eli Lilly & Co. for the LY379196 inhibitor; Rubén J. Boado for the bovine BBB GLUT1 cDNA; and Michael J. Getz for the mouse actin cDNA. The authors are indebted to Frank C. Brosius, III, Christin Carter-Su, Douglas A. Greene, Rubén J. Boado and Dennis Larkin for invaluable discussions and advice, and to Kathleen Britton for technical assistance.