Role of ephrin B2 in human retinal endothelial cell proliferation and migration

Cellular Signalling 15 (2003) 1011 – 1017 www.elsevier.com/locate/cellsig Role of ephrin B2 in human retinal endothelial cell proliferation and migration Jena J. Steinle a,*, Cynthia J. Meininger a, Usha Chowdhury a, Guoyao Wu a,b, Harris J. Granger a a Cardiovascular Research Institute and Department of Medical Physiology, College of Medicine, The Texas A&M University System Health Science Center, 702 SW HK Dodgen Loop, Medical Research Building, Room 202A, Temple, TX 76504, USA b Department of Animal Science, Texas A&M University, College Station, TX 77843, USA Received 8 January 2003; accepted 2 April 2003 Abstract This study was designed to determine the presence of Eph B4 or ephrin B2 in human retinal endothelial cells (REC) and their signal transduction. Human retinal endothelial cells were stimulated with an Eph B4/Fc chimera and probed for phosphorylation of phosphatidylinositol-3-kinase (PI3K), Src, and mitogen-activated protein kinase (MAPK) pathways. Proliferation and migration were investigated after Eph B4/Fc stimulation in the presence of various pathway inhibitors. Human retinal endothelial cells express ephrin B2, with little expression of Eph B4. Treatment with EphB4/Fc chimera resulted in activation of PI3K, Src, and MAPK pathways. Eph B4stimulated endothelial cell proliferation was mediated via PI3K, nitric oxide synthase, and extracellular signal-regulated kinase 1/2 (ERK1/2). Blockade of Src-PI3K pathways produced significant attenuation of Eph B4/Fc-stimulated migration. These results demonstrate for the first time that ephrin B2 is present in human retinal endothelial cells. Additionally, it appears that vascular growth may be modulated in the retina through activation of the PI3K pathway and its downstream components. D 2003 Elsevier Science Inc. All rights reserved. Keywords: Retina; Endothelial cell; Angiogenesis; Matrix metalloproteinase; Eye; Blood vessel 1. Introduction Erythropoietin-producing hepatoma amplified sequence (Eph) receptors and their ligands (ephrins) have recently been shown to regulate vasculogenesis in mice [1,2]. Knockout mice for the Eph B4 receptor or ephrin B2 ligand exhibit significant defects in capillary remodelling [3,4]. However, it is unclear what role this receptor– ligand pair plays in adult neovascularization. Studies of genetically engineered mice did reveal that ephrin B2 demarcates arteries in the adult [5]. Additionally, stimulation of other members of the Eph receptor family induced tube formation in renal endothelial cells, but not in umbilical vein endo- Abbreviations: PI3K, phosphatidylinositol-3-kinase; MAPK, mitogenactivated protein kinase; ERK1/2, extracellular signal-regulated kinase 1/2; REC, retinal endothelial cells; MM1, mesenteric microvascular endothelial cell line; MMP, matrix metalloproteinase. * Corresponding author. Tel.: +1-254-742-7144; fax: +1-254-7427145. E-mail address: jsteinle@siumed.edu (J.J. Steinle). 0898-6568/$ - see front matter D 2003 Elsevier Science Inc. All rights reserved. doi:10.1016/S0898-6568(03)00072-X thelial cells in vitro [6]. However, expression and signal transduction of ephrin B2 remains unclear in adult tissues. We have recently reported that human mesenteric endothelial cells express only Eph B4, and activation of Eph B4 by an ephrin B2/Fc chimera produces significant increases in mesenteric endothelial cell proliferation and migration [7]. However, since the cells used in our previous study were a cloned cell line, the interactions between more traditional endothelial cells of both arterial and venous origin are unclear. Additionally, we felt it was important to investigate a cell line that may be significantly altered in a disease state, such as diabetic retinopathy. Retinal endothelial cells (REC) may play a major role in vascular disease of the eye. Angiogenesis, or new blood vessel growth, is a common symptom and potential cause of vision loss in both diabetic retinopathy [8] and age-related macular degeneration [9]. Although vascular endothelial growth factor (VEGF) [10], basic fibroblast growth factor (bFGF) [11], and many other growth factors are increased in both of these diseases, the exact mechanism of ocular angiogenesis is not known. Because the ephrins were first 1012 J.J. Steinle et al. / Cellular Signalling 15 (2003) 1011–1017 discovered as factors regulating neurite outgrowth in the developing nervous system, it is highly likely that they may also regulate vascular growth in the neural retina. To date, this remains to be demonstrated. It was the goal of the present study to demonstrate Eph B4 and ephrin B2 protein expression in human retinal endothelial cells. Once this was established, we sought to determine whether activation of ephrin B2 could induce proliferation and migration of the retinal endothelial cells, two markers of angiogenesis. Finally, we investigated the signalling pathways involved in both proliferation and migration induced by Eph B4/Fc. 2. Methods 2.1. Cell culture Human retinal endothelial cells were purchased from Cell Systems (Kirkland, WA) and used at passages 3 – 7. Cells are grown in attachment factor-coated dishes and maintained in serum-free medium (Cell Systems) supplemented with 10% fetal bovine serum, 100 IU/ml penicillin, 100 Ag/ml streptomycin, and 0.25 Ag/ml amphotericin B. After the cells reach 80– 90% confluence, they are passaged with the use of passage reagent group. Starvation medium contains all of the above ingredients except that 0.1% bovine serum albumin is substituted for fetal bovine serum. 2.2. Determination of Eph receptor expression in human retinal endothelial cells Western blotting was conducted to determine if retinal endothelial cells expressed either Eph B4 receptor or ephrin B2 ligand or both. Cells in 60-mm dishes were lysed (50 mM Tris – HCl, pH 7.4; 1% NP-40; 0.25% Na-deoxycholate; 150 mM NaCl; 1 mM EDTA; 1 mM PMSF; 1 Ag/ml each of aprotinin, leupeptin, pepstatin; 1 mM Na3VO4; 1 mM NaF; 0.1% SDS) and 50 Ag of protein was loaded into each well and separated on a 4 –12% pre-cast polyacrylamide gel (Invitrogen, Carlsbad, CA), blotted onto a nitrocellulose membrane, and blocked with Super Block (Pierce, Rockford, IL) for 1 h at room temperature. Primary antibodies to either Eph B4 (H-200) or ephrin B2 (P-20; 5 Ag/ ml; Santa Cruz, Santa Cruz, CA) were then applied overnight at 4 jC. Membranes were probed with horseradish peroxidase-conjugated anti-rabbit secondary antibodies applied at a 1:10,000 dilution for 2 h at room temperature. Immunoreactive bands were detected by enhanced chemiluminescence (LumniGlo, Cell Signalling, Beverly, MA) using Kodak BioMax ML film and scanned into the computer using reflectance scanning. Intensity of the bands was quantified using NIH Image. Western blots to evaluate phosphorylation states of Akt (phospho Akt-473, 1:1000; Biosource, Camarillo, CA), MAP kinase (phospho p42/44, 1:1000, extracellular sig- nal-regulated kinase 1/2 (ERK1/2); Cell Signalling), Src (phospho Src-416, 1:1000; Cell Signalling) following stimulation with Eph B4/Fc chimera (50 nM; R&D Systems, Minneapolis, MN) for 0, 5, 10, 15, or 30 min were done as described above. Western blotting was also used to determine the optimal concentration of Eph B4/Fc for phosphorylation of Akt as described above. 2.3. Reverse-transcription polymerase chain reaction (RTPCR) for Eph B4 RNA was isolated from four dishes of retinal endothelial cells, passage 5, using a TotallyRNA kit (Ambion, Austin, TX). The quality of RNA was evaluated using an Agilent 2100 Bioanalyzer (Palo Alto, CA). Once the quality of RNA was found to be suitable, 1 Ag of RNA was used in the SuperScript One-Step RT-PCR with Platium Taq kit (Invitrogen). Thirty-seven cycles of PCR amplification were used for these experiments. The Eph B4 sense primer was CCCCAGGGAAGAAGGAGAGCTG, and the antisense primer was GCCCACGAGCTGGATGACTGTG [12]. Once the RT-PCR product was complete, the samples were run on a 2% agarose gel to determine whether transcript was amplified. A picture of the gel was taken under ultraviolet light and the picture scanned into Adobe Photoshop. 2.4. Migration assay All migration assays were done using BD BioCoat Angiogenesis System-Endothelial Cell Invasion plates according to supplied protocols with little variation. Briefly, 100,000 retinal endothelial cells were added to top chambers in 250 Al of medium. Starvation medium, Eph B4/Fc chimera (50 nM), Eph B4/Fc chimera (50 nM) with an inhibitor, or inhibitor alone was added to the lower chambers. The inhibitors were LY294002 (Calbiochem; 2 AM, phosphatidylinositol-3-kinase (PI3K) inhibitor), ML-9 (Biomol, Plymouth Meeting, PA; 100 AM, Akt inhibitor), KT5823 (Calbiochem; 1 AM, protein kinase G inhibitor), PD98059 (Calbiochem; 10 AM, ERK1/2 inhibitor), matrix metalloproteinase (MMP)-2/MMP-9 inhibitor III (MMP-I; Calbiochem; 1 AM, matrix metalloproteinase 2 and 9 inhibitor) or PP2 (Calbiochem; 1 AM, Src inhibitor). All inhibitors were added for 30 min before Eph B4/Fc was added to allow for complete pathway blockade. Cells were not added to the bottom chamber, so that background fluorescence of Calcein AM could be calculated from these wells. Plates were allowed to incubate for 26– 29 h at 37 jC to allow for migration through the Matrigel-coated membrane. Chambers were then transferred to wells containing Calcein AM (Molecular Probes, Eugene, OR) in Hanks Balanced Salt Solution for 1.5 h and read on a fluorescence plate reader (Bio-Tek, Winooski, VT, Model FL600, gain of 100) at 485/530 nm. Calcein AM is a fluorophore that will cause the cells that invaded the membrane to become fluorescent at 480 nm, which is then read on the fluorescent J.J. Steinle et al. / Cellular Signalling 15 (2003) 1011–1017 plate reader [13,14]. Migration of Eph B4- and inhibitortreated cells was expressed as a percentage of controls (those receiving only starvation medium), after background fluorescence was subtracted. Data were analysed using Prism software (GraphPad, San Diego, CA) and differences were considered significant at P < 0.05. 2.5. Cell proliferation assay Endothelial cell proliferation was assessed using an assay based on the cleavage of the tetrazolium salt WST-1 to formazan by cellular mitochondrial dehydrogenases. Expansion in the number of viable cells results in an increase in the overall activity of the mitochondrial dehydrogenases in the sample. The augmentation in enzyme activity leads to an increase in the formazan dye formed. The formazan dye produced by viable cells can be quantified by a multiwell spectrophotometer by measuring the absorbance of the dye solution at 440 nm. To perform the experiments, an aliquot of 50,000 retinal endothelial cells was added to each well of a 96-well tray in medium with 10% foetal bovine serum. After cell attachment, the cells were washed and high serum medium was replaced with starvation medium overnight. Negative control wells received starvation medium and positive control wells received 50 nM Eph B4/Fc only. Inhibitors, including PP2 (Src inhibitor, 1 AM), LY294002 (PI3K inhibitor, 2 AM), KT5823 (PKG inhibitor, 1 AM), or PD98059 (ERK1/2 inhibitor, 10 AM), were added for 30 min before Eph B4/Fc was added to allow for complete blockade. Controls treated with inhibitor alone were also included to determine their effect on proliferation. Cells were allowed to incubate for 48 h. After this time, the WST-1 reagent dissolved in Electro Coupling Solution (Chemicon) was applied for 4 h to measure cell proliferation. Data are presented as a percentage of negative control proliferation with P < 0.05 being significant. 1013 The conversion of nitrate to nitrite is 98% complete as determined with known amounts of both standards [15]. One hundred microliters of sample [diluted with doubledistilled water (DD-H2O)], diluted blank medium OR sodium nitrite standard (0 –2 AM) was mixed with 100 Al of DD-H2O and 20 Al of 316 AM DAN (in 0.62 M HCl). These reaction mixtures were incubated at room temperature for 10 min, followed by addition of 10 Al of 2.8 M NaOH. After mixing, 15 Al of the derivatized nitrite-DAN solution was injected into a 5-Am C8 column guarded by a 40-Am C18 column for chromatographic separation of NAT. The mobile phase (1.3 ml/min) was 15 mM sodium phosphate buffer (pH 7.5) containing 50% methanol (1 l of 30 mM Na2HPO4 and 125 ml of 30 mM NaH2PO4 mixed with 1.125 l of 100% methanol) (0.0 –3.0 min), followed sequentially by 100% HPLC-grade water (3.1 –5.0 min), 100% methanol (5.1 –8.0 min), 100% HPLC-grade water (8.1 – 10.0 min), and the initial 15 mM sodium phosphate buffer (pH 7.5) –50% methanol solution (10.1 –15.0 min). The use of 100% HPLC-grade water before and after 100% methanol is necessary to prevent abrupt marked increases in column pressure, and is sufficient to regenerate the columns for automatic analysis of multiple samples. All chromatographic procedures were carried out at room temperature. Fluorescence was monitored with excitation at 375 nm and emission at 415 nm. The retention time for NAT is 4.4 min. 3. Results 3.1. Ephrin B2 and Eph B4 expression Western blot analysis of Eph B4 and ephrin B2 in human retinal endothelial cell lysates shows limited 2.6. Measurement of nitrite accumulation Nitrite, a stable end product of nitric oxide metabolism, was assessed in the medium by reaction with 2,3-diaminonaphthalene (DAN) under acidic conditions to yield 2,3 naphthotriazole (NAT), a highly fluorescent product [15]. Reversed-phase HPLC separates NAT from DAN [and other fluorescent compounds present in biological samples] before fluorescence detection of NAT. Two days prior to the experiments, the medium was replaced with retinal serum-free medium containing 0.4 mM glutamine. Eph B4/Fc (50 nM) was added to the cells to stimulate ephrin B2 and downstream pathways. Medium was collected after 24 h. Nitrate was converted to nitrite using nitrate reductase as follows: 200 Al of diluted sample or nitrate standard (0 – 2 AM), 10 Al of 1 U/ml nitrate reductase (Roche, Indianapolis, IN), and 10 Al of 120 AM NADPH were mixed and incubated at room temperature for 1 h. This solution was then used directly for nitrite analysis. Fig. 1. Retinal endothelial cells (REC) lysates were probed with either antiEph B4 or anti-ephrin B2 primary antibodies and anti-rabbit secondary antibodies. Three independent lysates of retinal endothelial cells are compared to microvascular mesenteric endothelial cells (MM1) for Eph B4 (A). Results of ephrin B2 expression in REC are from four independent experiments (B). RNA from four separate experiments was amplified using RT-PCR to determine if Eph B4 transcripts were present (C). Eph B4 (120 kD); ephrin B2 (42 kD); Eph B4 mRNA (250 bp). 1014 J.J. Steinle et al. / Cellular Signalling 15 (2003) 1011–1017 Fig. 2. Eph B4/Fc chimera induces phosphorylation of Akt in a (A/B) dose-and (C/D) time-dependent manner. Note that optimal phosphorylation of Akt occurred between 25 and 50 nM Eph B4/Fc chimera and between 10 and 15 min. *P < 0.05 vs. control; N = 3 for each set of experiments. pAkt (60 kD). expression of Eph B4 in comparison with a microvascular mesenteric cell line (MM1) (Fig. 1A). Ephrin B2 expression was apparent in retinal endothelial cells (Fig. 1B) in contrast to previous findings of no ephrin B2 expression in mesenteric cells. The four lanes on the ephrin B2 blot and the three lanes on the Eph B4 blot represent four and three independent retinal endothelial cell lysates, respectively. Since the signal for Eph B4 was weak and could have been background, RT-PCR was used as an independent method of confirmation. All four samples analysed had mRNA for Eph B4 (Fig. 1C); therefore, it is clear that limited amounts of Eph B4 protein are present in retinal endothelial cells. 3.2. Phosphorylation of Akt, ERK1/2, and Src 416 following Eph B4/Fc stimulation Western blot analysis of Eph B4/Fc-stimulated retinal endothelial cells indicated that 25 nM to 50 nM was the optimum dose for ephrin B2-mediated phosphorylation of Fig. 3. Eph B4/Fc chimera produces phosphorylation of ERK1/2 (B, D) and Src (A, C). Cells treated with 50 nM Eph B4/Fc chimera were stimulated for 0, 5, 10, 15, or 30 min. N = 3 for each signalling pathway time course. pERK1/2 (44 kD); pSrc (60 kD). J.J. Steinle et al. / Cellular Signalling 15 (2003) 1011–1017 1015 Fig. 4. Ephrin B2 activation results in significant retinal endothelial cell migration. All inhibitors were administered 30 min prior to Eph B4/Fc chimera. *P < 0.05 vs. control; #P < 0.05 vs. Eph B4 alone; N = 5 for all treatments. Akt ( P < 0.05, Fig. 2A and B). Phosphorylation of Akt occurs most efficiently between 10 and 15 min of stimulation ( P < 0.05, Fig. 2C and D). Administration of Eph B4/Fc produced phosphorylation of Src on tyrosine 416 at 10 min (Fig. 3A and C) while ERK 1/2 appears to be phosphorylated following Eph B4/Fc stimulation at 5 min (Fig. 3B and D). PP2 did not alter Eph B4-induced increases in retinal endothelial cell migration and was significantly different from control migration ( P < 0.05 vs. control). These blockers do not affect migration in the absence of Eph B4/Fc (data not shown). 3.4. Ephrin B2 mediates retinal endothelial cell proliferation 3.3. Ephrin B2 mediates retinal endothelial cell migration Stimulation of ephrin B2 with Eph B4/Fc chimera resulted in a 23% increase in migration over starvation medium controls ( P < 0.001, Fig. 4). This response could be blocked by prior administration of LY294002 ( P < 0.001 vs. B4, P < 0.001 vs. control), ML-9 ( P < 0.01 vs. B4, not significant vs. control), PD98059 ( P < 0.01 vs. B4, not significant vs. control), KT5823 ( P < 0.001 vs. B4, P < 0.01 vs. control), and MMP-2/MMP-9 inhibitor (MMP-I, P < 0.001 vs. B4, not significant vs. control) in the presence of Eph B4 (Fig. 4). Stimulation of ephrin B2 with Eph B4/Fc chimera produced a 42% increase in cell proliferation ( P < 0.001, Fig. 5). The increased proliferation was inhibited by PP2 ( P < 0.001 vs. B4, P < 0.05 vs. control), LY294002 ( P < 0.001 vs. B4, not significant vs. control), KT5823 ( P < 0.001 vs. B4, not significant vs. control) and PD98059 ( P < 0.01 vs. B4, not significant vs. control) in the presence of Eph B4 (Fig. 5). ML-9 decreased cell numbers (even in the absence of Eph B4), probably due to an effect on cell attachment; therefore, the role of Akt inhibition could not be evaluated directly. All Fig. 5. Ephrin B2 activation produces significant increases in retinal endothelial cell proliferation. All inhibitors were administered 30 min prior to Eph B4/Fc chimera. *P < 0.05 vs. control; #P < 0.05 vs. B4; N = 5 for all treatments. 1016 J.J. Steinle et al. / Cellular Signalling 15 (2003) 1011–1017 4.2. Ephrin B2 regulation of retinal endothelial cell proliferation of the other inhibitors had no effect when administered in the absence of Eph B4/Fc (data not shown). Endothelial cell proliferation is required if enough cells are to be present to form the nascent tube in neovascularization. In the present study, proliferation of retinal endothelial cells in response to Eph B4/Fc chimera uses a common signalling pathway of phosphorylation of PI3K and Akt. Phosphorylation of Akt stimulates nitric oxide production through phosphorylation of eNOS on serine 1177, and the expected rise in cyclic GMP is known to elicit mitogenactivated protein kinase (MAPK) activation through PKG phosphorylation of c-raf-1 [16]. Since the MAPK pathway is a known proliferative pathway in response to other growth factors, such as VEGF [2], it is not surprising that this mechanism is employed by retinal endothelial cells for proliferation. Similar activation of this pathway was noted in mesenteric endothelial cells following stimulation of Eph B4 receptor-expressing cells with ephrin B2 [7]. 3.5. Eph B4/Fc stimulation increases nitrite production 4.3. Retinal endothelial cell migration Treatment of retinal endothelial cells with Eph B4/Fc for 24 h produced a significant induction of nitrite in the medium, a measure of nitric oxide production, relative to untreated cells (control, P < 0.05, Fig. 6). Retinal endothelial cell invasion requires the activation of a signalling cascade leading to destruction of the extracellular matrix, thereby facilitating cell motility. Retinal cells use a different signalling pathway for migration than did the mesenteric endothelial cells. Our results indicate that stimulation with Eph B4/Fc chimera resulted in significant migration that was inhibited by either LY294002 or ML-9. Since LY294002 decreased migration to a level below ‘‘basal’’ in the presence of Eph B4/Fc, it may indicate that PI3K and Akt have separate, but equally important, functions, in retinal endothelial cell migration. PI3K is capable of phosphorylating ras [17], which would lead to activation of ERK1/2. Akt could potentially be phosphorylated through a mechanism that is insensitive to LY294002. This would allow phosphorylation of Akt without PI3K activation, leaving regulation of each compound independent. The phosphorylation of Akt has been shown by others to induce both MMP-2 [18] and MMP-9 [19] activation. Cell migration is mediated by phosphorylation of Src, leading to PI3K and Akt activation. It remains clear that retinal endothelial cells employ both PI3K and Akt to launch cellular migration of retinal endothelial cells. Interestingly, blockade of Src phosphorylation affected only proliferation and not migration in retinal endothelial cells. This is the direct opposite of the effects of Src phosphorylation blockade in mesenteric endothelial cells, where it appears that Src is only involved in migration and not proliferation. This difference may be related to interactions of Eph B4 and ephrin B2 with their downstream partners. This remains to be determined. It is clear, however, that stimulation of ephrin B2 in retinal endothelial cells produces significant migration that is regulated through induction of both the MAPK pathway and MMPs following Akt phosphorylation. Fig. 6. Administration of Eph B4/Fc chimera produced significant increases in nitric oxide production, as measured by analysis of nitrites in the medium. *P < 0.05, N = 3. 4. Discussion 4.1. Ephrin B2 expression in human retinal endothelial cells Based upon the work of others, the expression patterns and signal transduction of Eph B4 and ephrin B2 in particular subtypes of endothelial cells remained unclear. It was noted that renal endothelial cells formed tubes in the presence of ephrin B1 but not ephrin A1. Human umbilical vein endothelial cells show opposite preferences, with tube formation induced by ephrin A1 and not ephrin B1 [6]. Therefore, it appears that endothelial cells of different vascular beds may possess a unique complement of Eph receptors and ephrin ligands, which produce different responses upon activation. We had found that cloned human microvascular mesenteric endothelial cells contain predominantly Eph B4, with little to no ephrin B2 protein expression [7]. However, given that ephrin expression is critical in the nervous system and these molecules may regulate vascular changes, we sought to determine the expression pattern of ephrin B2 and Eph B4 in human retinal endothelial cells. Ephrin B2 appears to be the predominant protein expressed in retinal endothelial cells relative to Eph B4. These findings are dissimilar to those found in mesenteric endothelial cells. Therefore, it was important to determine if activation of ephrin B2 signalling induced changes in proliferation and migration that could modulate ocular (retinal) angiogenesis. J.J. Steinle et al. / Cellular Signalling 15 (2003) 1011–1017 In conclusion, ephrin B2 stimulation results in significant migration and proliferation of retinal endothelial cells. Since these are two important steps in neovascularization, alterations in ephrin B2 signalling may lead to retinal angiogenesis and retinal disease. Acknowledgements This work was supported by NIH grant NHLBI 446221 (H.J.G.) and JDRF grants 2000-437 and 2002-228 (C.J.M. and G.W.). References [1] Gerety SS, Wang HU, Chen ZF, Anderson DJ. Mol Cell 1999;4: 403 – 14. [2] Gale NW, Yancopoulos GD. Genes Dev 1999;13:1055 – 66. [3] Adams RH, Wilkinson GA, Weiss C, Diella F, Gale NW, Deutsch U, et al. Genes Dev 1999;13:295 – 306. [4] Wang HU, Chen ZF, Anderson DJ. Cell 1998;93:741 – 53. [5] Gale NW, Baluk P, Pan L, Kwan M, Holash J, DeChiara TM, et al. Dev Biol 2001;230:151 – 60. 1017 [6] Daniel TO, Stein E, Cerretti DP, St John PL, Robert B, Abrahamson DR. Kidney Inter, Suppl 1996;57:S73 – 81. [7] Steinle JJ, Meininger CJ, Wu G, Wu MH, Granger HJ. J Biol Chem 2002;277:43830 – 5. [8] Garner A. Int Rev Exp Pathol 1986;28:249 – 306. [9] Amin R, Puklin JE, Frank RN. Invest Ophthalmol Vis Sci 1994;35: 3178 – 88. [10] Aiello LP, Avery RL, Arrigg PG, Keyt BA, Jampel HD, Shah ST, et al. N Engl J Med 1994;331:1480 – 7. [11] Kumar R, Yoneda J, Bucana CD, Fidler IJ. Int J Oncol 1998;12: 749 – 57. [12] Raynor M, Stephenson SA, Walsh DC, Pittman KB, Dobrovic A. BMC Cancer 2002;2:14. [13] Braut-Boucher F, Pichon J, Rat P, Adolphe M, Aubery M, Font J. J Immunol Methods 1995;178:41 – 51. [14] Yang A, Cardona DL, Barile FA. Cell Biol Toxicol 2002;18:97 – 108. [15] Li H, Meininger CJ, Wu G. J Chromatogr, B, Biomed Sci Appl 2000; 746:199 – 207. [16] Hood J, Granger HJ. J Biol Chem 1998;273:23504 – 8. [17] Blancher C, Moore JW, Robertson N, Harris AL. Cancer Res 2001; 61:7349 – 55. [18] Park BK, Zeng X, Glazer RI. Cancer Res 2001;61:7647 – 53. [19] Kim D, Kim S, Koh H, Yoon SO, Chung AS, Cho KS, et al. FASEB J 2001;15:1953 – 62.