A novel role for activating transcription factor-2 in 15(S)-hydroxyeicosatetraenoic acid-induced angiogenesis.

To investigate the mechanisms underlying 15(S)-HETE-induced angiogenesis, we have studied the role of the small GTPase, Rac1. We find that 15(S)-HETE activated Rac1 in human retinal microvascular endothelial cells (HRMVEC) in a time-dependent manner. Blockade of Rac1 by adenovirus-mediated expression of its dominant negative mutant suppressed HRMVEC migration as well as tube formation and Matrigel plug angiogenesis. 15(S)-HETE stimulated Src in HRMVEC in a time-dependent manner and blockade of its activation inhibited 15(S)-HETE-induced Rac1 stimulation in HRMVEC and the migration and tube formation of these cells as well as Matrigel plug angiogenesis. 15(S)-HETE stimulated JNK1 in Src-Rac1-dependent manner in HRMVEC and adenovirus-mediated expression of its dominant negative mutant suppressed the migration and tube formation of these cells and Matrigel plug angiogenesis. 15(S)-HETE activated ATF-2 in HRMVEC in Src-Rac1-JNK1-dependent manner and interference with its activation via adenovirus-mediated expression of its dominant negative mutant abrogated migration and tube formation of HRMVEC and Matrigel plug angiogenesis. In addition, 15(S)-HETE-induced MEK1 stimulation was found to be dependent on Src-Rac1 activation. Blockade of MEK1 activation inhibited 15(S)-HETE-induced JNK1 activity and ATF-2 phosphorylation. Together, these findings show that 15(S)-HETE activates ATF-2 via the Src-Rac1-MEK1-JNK1 signaling axis in HRMVEC leading to their angiogenic differentiation.

In contrast to the above-mentioned observations, some reports provided clues that 15(S)-HETE induces retinal angiogenesis (23,24). Angiogenesis plays a critical role in the progression of various diseases including atherosclero-sis, tumors, and diabetic retinopathy (25)(26)(27). To learn more about the capacity of 15(S)-HETE in the stimulation of angiogenesis, we reported that it induces the migration and tube formation of microvascular endothelial cells from different vascular beds (8,28,29). Substantial evidence shows that Rac1, a small GTPase protein, plays a major role in the regulation of cell motility (30)(31)(32). We have previously shown that 15(S)-HETE-induced angiogenesis requires activation of ERK1/2 and JNK1 (8). Rac1 plays a role in the activation of JNK1 (33). Toward understanding the mechanisms of 15(S)-HETE-induced angiogenesis, here we report that 15(S)-HETE-induced JNK1 activation requires Src-dependent Rac1-mediated MEK1 stimulation and that Src-Rac1-MEK1-JNK1 signaling targets ATF-2 in influencing human retinal microvascular endothelial cell (HRMVEC) angiogenic differentiation.

Adenoviral vectors
The construction of pAd-GFP and pAd-dnMEK1 were described previously (34). To clone dnRac1 into adenoviral vector, Rac1N17 was released from pCEV-Rac1N17 by digestion with BamHI and EcoRI (35) and subcloned into the same sites of pENTR3C to yield pENTR3C-Rac1N17. The pENTR3C-Rac1N17 was then subjected to recombination with pAdCMVV5DEST to obtain pAd-Rac1N17. To construct adenoviral vector for dnJNK1, Flag-JNK1apf was released from pCDNA1FlagJNK1apf by digestion with Hind III and XbaI (36) and subcloned into the same sites of pBluescript II SK (1). The Flag-JNK1apf was excised by digestion with KpnI and NotI and cloned into the same sites of pENTR3C to yield pENTR3C-flag-JNK1apf. The pENTR3C-flag-JNK1apf was recombinated with pAdCMVV5DEST to yield pAdflag-JNK1apf. To make adenoviral vector for dnATF-2, dnATF-2 fragment was released from pCDNA3-dnATF-2 by digestion with KpnI and NotI (37) and cloned into the same sites of pENTR3C to yield pENTR3C-dnATF-2, which was then recombinated with pAdCMVV5DEST to yield pAd-dnATF-2. The plasmids, pAd-GFP, pAd-dnMEK1, pAd-Rac1N17 pAd-dnATF-2, and pAd-flag-JNK1apf were linearized by digestion with PacI and transfected into HEK293A cells. The resultant adenovirus was further amplified by infection of HEK293A cells and was purified by cesium chloride gradient ultracentrifugation (38). The construction of Ad-dnSrc was described previously (39) and provided by Dr. Richard C. Vanema of Georgia Medical College, Augusta, GA.

Cell culture
Primary HRMVEC (Cat. No. ACBRI 181) were bought from Applied Cell Biology Research Institute (Kirkland, WA). Cells were grown in medium 131 containing microvascular growth supplements, 10 mg/ml gentamycin, and 0.25 mg/ml amphotericin B. Cultures were maintained at 37°C in a humidified 95% air and 5% CO 2 atmosphere. Cells were quiesced by incubating in medium 131 for 24 h and used between 4 to 10 passages to perform the experiments unless otherwise indicated. 15(S)-HETE was added to cells in ethanol. Thus, controls received an equal volume of ethanol and it does not exceed 0.03%.

Pull-down assay
To measure Rac1 activation, an equal amount of protein from control and each treatment was incubated with GST-PAK1 beads overnight at 4°C. After collection by centrifugation, the beads were heated in 50 ml of Laemmeli sample buffer for 10 min, and the released proteins were resolved by 0.1% SDS-12% PAGE followed by Western blot analysis for Rac1 using its specific antibodies.

JNK1 assay
JNK1 activity was measured by immunocomplex kinase assay as described previously (40). Protein extracts from cells or tissues were prepared using lysis buffer (20 mM HEPES, pH 7.4, 150 mM NaCl, 1% NP-40, 10 mg/ml aprotinin, 10 mg/ml leupeptin, 50 mM glycerophosphate, 10 mM NaF, and 1 mM sodium orthovanadate). Equal amount of protein (200 mg) from each sample was immunoprecipitated with anti-JNK1/2 antibodies (2 mg) overnight at 4°C followed by incubation with protein-A sepharose beads for 1 h at room temperature. The immunocomplexes were washed four times with lysis buffer, one time with kinase buffer, and incubated in kinase reaction mix containing 25 mM HEPES (pH 7.4), 10 mM MgCl 2 , 1 mM EGTA, 200 mg/ml GST-c-Jun fusion protein, 1 mCi [g 32 P]ATP, and 50 mM ATP for 30 min at 30°C. The kinase reactions were stopped by addition of SDS-PAGE sample buffer and subsequent boiling for 5 min. The reaction products were separated by electrophoresis on 0.1% SDS and 10% polyacrylamide gels. The [ 32 P]labeled GST-c-Jun protein was visualized by autoradiography and the band intensities were quantified using NIH Image J.

Cell migration assay
Cell migration was performed using a modified Boyden Chamber method as described by Nagata, Mogi, and Walsh (41). The cell culture inserts containing membranes with 10 mm in diameter and 8.0 mm pore size (Nalge Nunc International, Rochester, NY) were placed in a 24-well tissue culture plate (Costar, Corning Incorporated, Corning, NY). The lower surface of the porous membrane was coated with 70% Matrigel at 4°C overnight and then blocked with 0.1% heat-inactivated BSA at 37°C for 1 h. HRMVEC were quiesced for 24 h in medium 131, trypsinized, and neutralized with trypsin neutralizing solution. Cells were seeded into the upper chamber at 1 3 10 5 cells/well. Vehicle or 15(S)-HETE were added to the lower chamber at the indicated concentration. The upper and lower chambers contained medium 131. When the effect of dominant negative Src, Rac1, JNK1, and ATF-2 mutants was tested on 15(S)-HETE-induced HRMVEC migration, cells were infected with Ad-GFP, Ad-dnSrc, Ad-dnRac1, Ad-dnJNK1, or Ad-dnATF-2 at a moi of 80 and quiesced before they were subjected to the migration assay. After 6 h of incubation at 37°C, nonmigrated cells were removed from the upper side of the membrane with cotton swabs and the cells on the lower surface of the membrane were fixed in methanol for 15 min. The membrane was then stained with Hematoxylin and observed under a light microscope (Eclipse 50i, Nikon, Japan). Images were captured using a Nikon Digital Sight DS-L1 system. Cells were counted in five randomly selected squares per well and presented as the number of migrated cells/field.

Tube formation assay
Tube formation assay was performed as described by Nagata, Mogi, and Walsh (41). Twenty-four well culture plates (Costar, Corning Incorporated) were coated with growth factor-reduced Ad-GFP or Ad-dnRac1 at a moi of 80, quiesced, and subjected to 15(S)-HETE-induced migration (B) or tube formation (C). D: C57BL/6 mice were injected subcutaneously with 0.5 ml of Matrigel premixed with vehicle or 50 mM 15(S)-HETE with and without Ad-GFP or Ad-dnRac1 (5 3 10 9 pfu/ml). One week later, the animals were sacrificed, and the Matrigel plugs were harvested from underneath the skin and either immunostained for CD31 expression using anti-CD31 antibodies or analyzed for hemoglobin content using Drabkinʼs reagent. The values in the bar graphs in panels A, B, C, and D are the means 6 SD of three independent experiments or four plugs from four animals. * P , 0.01 vs. control or Ad-GFP; ** P , 0.01 vs. Ad-GFP 1 15(S)-HETE. The white bars represent controls to their respective treatments.
Matrigel (BD Biosciences) in a total volume of 280 ml/well and allowed to solidify for 30 min at 37°C. HRMVEC were trypsinized, neutralized with trypsin neutralizing solution, and resuspended in medium 131 at 5 3 10 5 /ml and 200 ml of this cell suspension was added into each well. Vehicle or 15(S)-HETE, at the indicated concentration, were added to the appropriate well and the cells were incubated at 37°C for 6 h. When the effect of dominant negative Src, Rac1, JNK1, and ATF-2 mutants was tested on 15(S)-HETE-induced HRMVEC tube formation, cells were infected with Ad-GFP, Ad-dnSrc, Ad-dnRac1, Ad-dnJNK1 or Ad-dnATF-2 at a moi of 80 and quiesced before they were subjected to tube formation. Tube formation was observed under an inverted microscope (Model, Eclipse TS100, Nikon, Japan). Images were captured with a CCD color camera (Model, KP-D20AU, Hitachi, Japan) attached to the microscope and tube length was measured using the NIH Image J.

Western blot analysis
After appropriate treatments and rinsing with cold PBS, HRMVEC were lysed in 500 ml of lysis buffer (PBS, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 100 mg/ml PMSF, 100 mg/ml aprotinin, 1 mg/ml leupeptin and 1 mM sodium orthovanadate) and scraped into 1.5 ml Eppendorf tubes. After standing on ice for 20 min, the cell lysates were cleared by centrifugation at 12,000 rpm for 20 min at 4°C. Samples of cell lysates containing an equal amount of protein were resolved by electrophoresis on 0.1% SDS and 10% polyacrylamide gels. The proteins were transferred electrophoretically to a nitrocellulose membrane (Hybond, Amersham Pharmacia Biotech, Piscataway, NJ). After blocking in 10 mM Tris-HCl buffer, pH 8.0, containing 150 mM sodium chloride, 0.1% Tween 20 and 5% (w/v) nonfat dry milk, the membrane was treated with appropriate primary antibodies followed by incubation with Horseradish Peroxidase (HRP)-conjugated secondary antibodies. The antigen-antibody complexes were detected using chemiluminescence reageant kit (Amersham Pharmacia Biotech, Piscataway, NJ).

Matrigel plug angiogenesis assay
Matrigel plug assay was performed essentially as described by Medhora et al. (42). C57BL/6 mice (8 wks old) were lightly anesthetized with sodium pentobarbital (50 mg/kg, i.p.) and were injected subcutaneously with 0.5 ml of Matrigel that was premixed with vehicle or 50 mM of 15(S)-HETE along the dorsal midline. Although this high concentration of 15(S)-HETE was used in the Matrigel plug angiogenesis experiments during the course of the present study, ongoing experiments showed that 15(S)-HETE even at much lower concentrations (5 mM) also induces Matrigel plug angiogenesis to a level similar to that of 50 mM. The injection was made rapidly with a B-D 26G1/2 needle to ensure the entire content was delivered as a single plug. Whenever the effects of Ad-GFP, Ad-dnSrc, Ad-dnRac1, Ad-dnJNK1, or Ad-dnATF-2 (5 3 10 9 pfu/ml) were tested on 15(S)-HETEinduced angiogenesis, these adenoviruses were added to the Matrigel prior to the injection. The mice were allowed to recover and 7 days later, unless otherwise stated, the animals were sacrificed by inhalation of CO 2 and the Matrigel plugs were harvested from underneath the skin. The plugs were homogenized in 1 ml of deionized H 2 O on ice and cleared by centrifugation at 10,000 rpm for 6 min at 4°C. The supernatant was collected and used to measurements of hemoglobin content with Drabkinʼs reagent along with hemoglobin standard essentially according to the manufac-turerʼs protocol (Sigma Chemical Co.). The absorbance was read at 540 nm in an ELISA plate reader (Spectra Max 190, Molecular Devices, Sunnyvale, CA). These experiments were repeated at least three times with four mice for each group, and the values are expressed as g/dl of hemoglobin/mg plug.

Immunohistochemistry
After retrieving the Matrigel plugs from mice, they were snapfrozen in OCT compound. Cryo-sections (5 mm) were made using Leica Cryostat machine (Model: CM3050S) and stained with anti-CD31 antibody (1:500 dilution, BD Pharmingen, Palo Alto, CA), followed by sequential incubation with biotinylated anti-rat IgG (1:300 dilution), avidin-biotin peroxidase, and diaminobenzidine substrate (Vector Laboratories, Burlingame, CA). The sections were counterstained with Hematoxylin and examined under a light microscope (Model: Eclipse 50i, Nikon, Japan) in six random fields (4003 magnification) from each group. Images were captured using Nikon Digital Sight DS-L1 system.

Statistics
All the experiments were repeated three times and data are presented as Means 6 SD. The treatment effects were analyzed by Studentʼs t -test and the P values , 0.05 were considered to be statistically significant. In the case of Western blotting, one representative set of data is shown.

15(S)-HETE activates Rac1 in HRMVEC
Previously we have demonstrated that 15(S)-HETE induces HRMVEC migration more potently than 5(S)-HETE or 12(S)-HETE. In addition, we reported that 15(S)-HETE induces angiogenic differentiation of HRMVEC involving MEK1-dependent activation of ERK1/2 and JNK1 (8). To investigate further the mechanisms by which 15(S)-HETE induces retinal angiogenesis, here we have studied the role of Rac1. Quiescent HRMVEC were treated with and without 15(S)-HETE (0.1 mM) for the indicated times and an equal amount of protein from control each treat- HETE with and without Ad-GFP or Ad-dnSrc (5 3 10 9 pfu/ml). One week later, the animals were sacrificed, and the Matrigel plugs were harvested from underneath the skin and either immunostained for CD31 expression using anti-CD31 antibodies or analyzed for hemoglobin content using Drabkinʼs reagent. The values in the bar graphs in panels A, B, and C are the means 6 SD of three independent experiments or four plugs from four animals. * P , 0.01 vs. Ad-GFP; ** P , 0.01 vs. Ad-GFP 1 15(S)-HETE. The white bars represent controls to their respective treatments. ment was analyzed by pull-down assay for PAK-bound Rac1 as previously described. 15(S)-HETE stimulated Rac1 activation in a time-dependent manner with a near maximum 3-fold increase at 10 min that was sustained until 60 min and declined thereafter (Fig. 1A). Next we tested the role of Rac1 in 15(S)-HETE-induced HRMVEC migration and tube formation. 15(S)-HETE stimulated HRMVEC migration by approximately 2-fold as measured by a modified Boyden chamber method, and adenovirus-mediated expression of dnRac1 attenuated this effect (Fig. 1B). Similarly, 15(S)-HETE induced HRMVEC tube formation by approximately 2-fold, and this effect was blocked by adenovirusmediated expression of the dominant negative Rac1 mutant (Fig. 1C). In order to obtain additional evidence for the role of Rac1 in 15(S)-HETE-induced angiogenesis, we used the Matrigel plug angiogenesis model. As shown in Fig. 1D, 15(S)-HETE (50 mM) induced Matrigel plug angiogenesis and dnRac1 significantly inhibited this effect.

Src mediates 15(S)-HETE-induced Rac1 activation in HRMVEC
Many studies have shown that Src mediates Rac1 in response to receptor tyrosine kinase agonists and cytokines (43,44). To determine whether Src mediates 15(S)-HETEinduced activation of Rac1, we first studied the time course of the effect of this eicosanoid on tyrosine phosphorylation  1 mM) for the indicated times, and cell extracts were prepared and analyzed by Western blotting for pJNK1 using its phosphospecific antibodies. The blot was reprobed with anti-JNK1 antibodies for normalization. B, C: HRMVEC were transduced with Ad-GFP, Ad-dnSrc, or Ad-dnRac1 at a moi of 80, quiesced, treated with and without 15(S)-HETE (0.1 mM) for 10 min, and cell extracts were prepared and analyzed for JNK1 activation either by Western blotting using its phosphospecific antibodies or by immunocomplex kinase assay using GST-c-Jun protein and [g-32 P]ATP as substrates. The blots in panels A and B were reprobed either with anti-JNK1 antibodies for normalization or anti-Src antibodies to show overexpression of Src. The values in the bar graphs in panels A and B are the means 6 SD of three independent experiments. * P , 0.01 vs. control or Ad-GFP; ** P , 0.01 vs. Ad-GFP 1 15(S)-HETE. The white bars represent controls to their respective treatments. of Src. 15(S)-HETE induced the tyrosine (Tyr416) phosphorylation of Src in a time-dependent manner with a maximum 3-fold increase at 10 min, that was sustained for 60 min and declined thereafter ( Fig. 2A). Next, we tested the effect of dnSrc on 15(S)-HETE-induced activation of Rac1. Adenovirus-mediated expression of dnSrc without affecting the Rac1 levels significantly inhibited its PAK-bound levels (Fig. 2B). To test the role of Src in 15(S)-HETE-induced angiogenic events, we further tested the effect of dnSrc on HRMVEC migration and tube formation and Matrigel plug angiogenesis. Adenovirus-mediated expression of dnSrc completely inhibited 15(S)-HETE-induced HRMVEC migration and tube formation and Matrigel plug angiogenesis (Fig. 3A-C).

MEK1 mediates 15(S)-HETE-induced ATF-2 via JNK1-dependent manner
Previously we have reported that MEK1 acts as an upstream kinase in 15(S)-HETE-induced JNK1 activation (8). To understand the upstream signaling events of MEK1 activation by 15(S)-HETE in HRMVEC, we studied the role of Src and Rac1. Interference with Src-Rac1 signaling via adenovirus-mediated expression of their dominant negative  1 mM) for the indicated times, and cell extracts were prepared and analyzed by Western blotting for pATF-2 using its phosphospecific antibodies. B, C: HRMVEC were transduced with Ad-GFP, Ad-dnSrc, Ad-dnRac1, or Ad-dnJNK1 at a moi of 80, quiesced, treated with and without 15(S)-HETE (0.1 mM) for 10 min, and cell extracts were prepared and analyzed for ATF-2 phosphorylation as described in panel A. The blots in panels A, B, and C were reprobed either with anti-ATF-2 antibodies for normalization or anti-Src antibodies to show overexpression of Src. The values in the bar graphs are the means 6 SD of three independent experiments. * P , 0.01 vs. control or Ad-GFP; ** P , 0.01 vs. Ad-GFP 1 15(S)-HETE. The white bars represent controls to their respective treatments. mutants inhibited 15(S)-HETE-induced MEK1 activation (Fig. 8A, B). Next we tested the role of MEK1 in 15(S)-HETE-induced JNK1 and ATF-2 activation. Blockade of MEK1 activation by adenovirus-mediated expression of its dominant negative mutant suppressed 15(S)-HETEinduced phosphorylation and activity of JNK1 and phosphorylation of ATF-2 (Fig. 8C, D). Recently, v-6 PUFAs such as AA have been reported to stimulate proangiogenic processes, including enhancing the simultaneous effect of vascular endothelial growth factor (VEGF) and bFGF on angiopoietin-2 expression and tube formation in human endothelial cells (48). Because AA is the substrate for 15(S)-HETE production via the 15-LOX1/2 or 12-LOX pathways, we asked whether it has any effect on activation of ATF-2 in HRMVEC. As expected, AA (5 mM) stimulated the phosphorylation of ATF-2 in a time-dependent manner in HRMVEC (Fig. 9A). Similarly, AA (50 mM) induced Matrigel plug angiogenesis (Fig. 9B, C).

DISCUSSION
The important findings of the present study are that 1) 15(S)-HETE activated Src, Rac1 and ATF-2 in a timedependent manner in HRMVEC; 2) 15(S)-HETE-induced ATF-2 activation was dependent on Src-Rac1-MEK1-JNK1 Ad-GFP or Ad-dnATF-2 at a moi of 80, quiesced, and subjected to 15(S)-HETE-induced migration (A) or tube formation (B). C: C57BL/6 mice were injected subcutaneously with 0.5 ml of Matrigel premixed with vehicle or 50 mM 15(S)-HETE with and without Ad-GFP or Ad-dnATF-2 (5 3 10 9 pfu/ml). One week later, the animals were sacrificed, and the Matrigel plugs were harvested from underneath the skin and either immunostained for CD31 expression using anti-CD31 antibodies or analyzed for hemoglobin content using Drabkinʼs reagent. The values in the bar graphs in panels A, B, and C are the means 6 SD of three independent experiments or four plugs from four animals. * P , 0.01 vs. Ad-GFP; ** P , 0.01 vs. Ad-GFP 1 15(S)-HETE. The white bars represent controls to their respective treatments. signaling; and 3) 15(S)-HETE-induced migration and tube formation of HRMVEC and Matrigel plug angiogenesis in vivo exhibited a requirement for activation of Src-Rac1-MEK1-JNK1-ATF-2 signaling. A large body of evidence indicates that Rac1 via its targeting of a wide variety of signaling molecules plays a role in the regulation of cell growth and in motility (30)(31)(32). One of the mechanisms of Rac1 activation by growth factors and cytokines is the involvement of tyrosine kinases such as the Src family of nonreceptor tyrosine kinases. A role for Src and Fyn in the activation of Rac1 by EGF and SCF in the mediation of intestinal epithelial cell and mast cell migration has been reported (43,49). Similarly, the involvement of Src in the activation of Rac1 in mechanical strain-induced epithelial cell mitogenesis has been reported (50). Our results reveal that 15(S)-HETE, an eicosanoid, also possesses the capacity to activate Rac1 via recruiting Src as an upstream signaling molecule in the stimulation of HRMVEC migration and tube formation. It was reported that oxidants induce Src activation via the involvement of PKC (51). Many studies also showed that HETEs, including 15(S)-HETE, possess the capacity to activate PKC (52)(53)(54). Based on these findings, it is conceivable that 15(S)-HETE activates Src via a mechanism involving PKC. One study has reported that 12(S)-HETE binds to high affinity binding sites in the cytoplasm of lung carcinoma cells and thereby interacts with steroid receptor coactivator-1 (55). Although no further characterization was provided for these receptors, it is possible that 15(S)-HETE may also possess similar receptors in HRMVEC, which upon binding to 15(S)-HETE, may lead to activation of Src. While further studies are required for elucidation of the mechanisms by which 15(S)-HETE activates Src in HRMVEC, our results provide convincing evidence for the ability of 15(S)-HETE in the activation of this important nonreceptor tyrosine kinase leading to Rac1 activation in the mediation of migration and tube formation of these cells.
Rac1 plays a role in mediating receptor tyrosine kinase receptor and cytokine receptor-mediated signaling to activate JNK1 via recruiting MAPK kinase kinases, such as MEKK1-4, and MAPK kinases, such as MKK4/7 (56). Although, a role for MEKK1-4 and MKK4/7 cannot be ruled out in 15(S)-HETE-induced JNK1 activation, our previous observations revealed that PD98059, a potent inhibitor of MEK1/2 and a dominant negative mutant of MEK1 blocked 15(S)-HETE-induced JNK1 activation (8). In the present study, we showed that dnMEK1 also inhibits 15(S)-HETEinduced JNK1 activity. In addition, we found that suppression of Src-Rac1 signaling blocks 15(S)-HETE-induced MEK1 activation, suggesting that 15(S)-HETE activates JNK1 via Src-Rac1-MEK1-dependent mechanism in HRMVEC, leading to their migration and tube formation. A large body of data indicates that MEK1/2 phosphorylates and activates only ERK1/2 (56). However, our findings revealed a novel role for MEK1 in JNK1 activation as well, at least in human retinal microvascular endothelial cells. A role for MAPKs, particularly JNKs and p38MAPK, has been shown in the activation of ATF-2 (45)(46)(47). Similarly, a requirement for Rac1 activation in syndecan-4-induced ATF-2 transcriptional activity has been reported (57). Along with these findings, our observations show that 15(S)-HETE-induced ATF-2 activation depends on the Src-Rac1-MEK1-JNK1 signaling axis. In addition, because suppression of Src, Rac1, MEK1, and JNK1 inhibited ATF-2 activation and adenovirus-mediated expression of dominant negative ATF-2 mutant abrogated 15(S)-HETE-induced HRMVEC migration and tube formation, it appears that this transcriptional factor is an effector of Src-Rac1-MEK1-JNK1 activation signaling in the mediation of HRMVEC angiogenic differentiation by this eicosanoid. Fig. 9. AA stimulates phosphorylation of ATF-2 in HRMVEC in vitro and induces Matrigel plug angiogenesis in vivo. A: Quiescent HRMVEC were treated with and without AA (5 mM) for the indicated times, and cell extracts were prepared and analyzed by Western blotting for pATF-2 using its phosphospecific antibodies. B, C: C57BL/6 mice were injected subcutaneously with 0.5 ml of Matrigel premixed with vehicle or 50 mM AA. One week later, the animals were sacrificed, and the Matrigel plugs were harvested from underneath the skin and either immunostained for CD31 expression using anti-CD31 antibodies or analyzed for hemoglobin content using Drabkinʼs reagent. The values in the bar graph are the means 6 SD of four plugs from four animals. * P , 0.01 vs. control. Furthermore, because adenovirus-mediated expression of dominant negative mutants of Src, Rac1, JNK1, MEK1, or ATF-2 negated 15(S)-HETE-induced Matrigel plug angiogenesis, it is likely that these signaling molecules may also influence angiogenesis in vivo in a manner similar to that seen in HRMVEC in vitro. Because AA, a substrate for the LOX enzymes, stimulated the phosphorylation of ATF-2 in HRMVEC in vitro and Matrigel plug angiogenesis in vivo, it is possible that similar signaling may also be effective in mediating angiogenesis by other hydroxy fatty acids such as 12(S)-HETE. Based on the efficacy of AA to induce Matrigel plug angiogenesis, it is possible to assume that besides its conversion via the cyclooxygenase, LOX, or cytochrome P450 monooxygenase (CYP) pathways, free fatty acid itself may induce angiogenesis.
Several studies have shown that hypoxia-induced angiogenesis requires FGF-2 and VEGF (58,59). Previously, others and we have demonstrated that hypoxia induces the expression of 15-LOX1 and production of 15(S)-HETE in arteries and microvascular endothelial cells (8,24,60). In addition, we reported that 15(S)-HETE induces the expression of VEGF in human dermal microvascular endothelial cells (29). Based on these observations, it is possible that 15(S)-HETE may influence the expression of angiogenic factors such as FGF-2 and VEGF via activation of signaling pathways such as Src-Rac1-MEK1-JNK1-ATF-2 during hypoxia-induced angiogenesis. It was reported that lipoxin A 4 , a metabolite of 15(S)-HETE, inhibits VEGFinduced angiogenesis (61). Based on this result, it may be conceivable that 15(S)-HETE-induced angiogenic differentiation of HRMVEC may not require its conversion to its metabolites.
In summary, the findings of the present study suggest that 15(S)-HETE-induced human retinal microvascular endothelial cell angiogenic differentiation requires Src-Rac1-MEK1-JNK1-dependent activation of ATF-2, as depicted in Fig. 10.
T.Z. performed cell migration and tube formation, immunocomplex kinase assays, pull-down assays, and Western blot analysis; D.W. performed Matrigel plug injections and immunohistochemistry; S.C. measured hemoglobin levels in Matrigel plugs; M.K. performed immunocomplex kinase assays and Western blot analysis; K.R.C. performed Western blot analysis; V.K.S. contributed in data analysis; D.A.J. provided collaborative discussions and participated in writing the manuscript; J.S.P. provided critical reagents and participated in collaborative discussions; G.N.R. designed the experiments, interpreted the data, and wrote the manuscript.