12/15-Lipoxygenase-dependent ROS production is required for diet-induced endothelial barrier dysfunction.

To understand the mechanisms of 15(S)-HETE-induced endothelial cell (EC) barrier dysfunction, we examined the role of xanthine oxidase (XO). 15(S)-HETE induced junction adhesion molecule A (JamA) phosphorylation on Y164, Y218, and Y280 involving XO-mediated reactive oxygen species production and Src and Pyk2 activation, resulting in its dissociation from occludin, thereby causing tight junction (TJ) disruption, increased vascular permeability, and enhanced leukocyte and monocyte transmigration in vitro using EC monolayer and ex vivo using arteries as models. The phosphorylation of JamA on Y164, Y218, and Y280 appears to be critical for its role in 15(S)-HETE-induced EC barrier dysfunction, as mutation of any one of these amino acid residues prevented its dissociation from occludin and restored TJ integrity and barrier function. In response to high-fat diet (HFD) feeding, WT, but not 12/15-lipoxygenase (LO)−/−, mice showed enhanced XO expression and its activity in the artery, which was correlated with increased aortic TJ disruption and barrier permeability with enhanced leukocyte adhesion and these responses were inhibited by allopurinol. These observations provide novel insights on the role of XO in 12/15-LO-induced JamA tyrosine phosphorylation and TJ disruption leading to increased vascular permeability in response to HFD.

fetal bovine serum, 100 IU/ml penicillin, and 100 g/ml streptomycin and used in transmigration assays. Cultures were maintained at 37°C in a humidifi ed 95% air and 5% CO 2 atmosphere. HUVECs between 6 and 10 passages were growth-arrested by incubating in Medium 200 for 12 h and used to perform the experiments unless otherwise indicated.

Adenoviral vectors
Construction of Ad-GFP, Ad-dnSrc, Ad-dnPyk2, and Ad-LacZ are described previously ( 20 ). Wherever adenoviral vectors were used, cells are transduced with the indicated adenovirus at 40 multiplicity of infection (moi) overnight in complete medium . After transductions, cells were allowed to reach to 70% confl uence in complete medium and growth-arrested for 24 h before being used.

Animals
WT (C57BL/6) and 12/15-LO Ϫ / Ϫ mice (B6129S2-Alox15 tm1fun /J) were purchased from the Jackson Laboratory (Bar Harbor, ME). Mice were bred and maintained according to the Institutional Animal Facility Guidelines. All the experiments involving animals were approved by the Animal Care and Use Committee of the University of Tennessee Health Science Center, Memphis, TN. To study the diet-induced effect, mice were kept on chow diet (CD) or HFD for 3 months and used as needed. When mice were administered with allopurinol, it was added (6.8 g/l) to drinking water ad libitum.

Western blot analysis
An equal amount of protein from cell or tissue extract was analyzed by Western blotting for the indicated antigen using its specifi c antibody as described previously ( 20 ).

Immunoprecipitation
An equal amount of protein from cell or tissue extract was immunoprecipitated with the indicated antibody as described previously ( 20 ).

Immunofl uorescence
HUVECs were grown to a confl uent monolayer on cell culture grade glass cover slips, quiesced, and treated with or without inhibitors, followed by vehicle or 15(S)-HETE (0.1 M) addition for the required time periods. In studies with adenoviral vectors, HUVECs were fi rst transduced with the desired adenoviral vectors and then were grown to a monolayer, quiesced, and treated with 15(S)-HETE. After treatment, cells were washed with cold PBS, fi xed with 95% ethanol for 30 min at 4°C, permeabilized in TBS [10 mM Tris-HCl (pH 8.0), 150 mM NaCl] containing 0.1% Triton X-100 for 10 min at room temperature and blocked in 2% BSA in TBS with 1.33 g/l CaCl 2 , 1 g/l MgCl 2 , and 0.1% saponin overnight at 4°C. Then, the cells were incubated with the desired primary antibodies followed by Alexa Fluor conjugated secondary antibodies, counterstained with Hoechst 33342 (1:3,000 dilution in PBS) for 1 min at room temperature, and mounted onto glass slides with Prolong Gold Antifade mounting medium. Fluorescence images of cells were captured using an inverted Zeiss fl uorescence microscope (AxioObserver Z1) via a 40× NA 0.6 objective and AxioCam MRm camera without any enhancements using the microscope operating and image analysis software AxioVision version 4.7.2 (Carl Zeiss Imaging Solutions GmbH). The TJ localized JamA levels were quantifi ed with Nikon's NIS Elements AR 3.1 imaging software and expressed as relative fl uorescence units (RFUs).

Cell culture
Human umbilical vein ECs (HUVECs) were obtained from Invitrogen and cultured in Medium 200 containing low serum growth supplements, 10 g/ml gentamycin, and 0.25 g/ml amphotericin B. THP1 cells which were supplied by the ATCC were grown in RPMI 1640 medium containing 50 M 2-mercaptoethanol, 10% fetal bovine serum, 100 IU/ml penicillin, and 100 g/ml streptomycin. Human peripheral blood mononuclear cells (PBMCs) were purchased from BioreclamationIVT (Baltimore, MD), suspended in RPMI 1640 medium containing 50 M 2-mercaptoethanol, 10% Devices). Wherever adenoviral vectors or plasmids were used, cells were transduced or transfected with these viral or plasmid vectors, respectively, prior to seeding onto the transwell. In the case of inhibitors, they were added to cells on the apical side just 30 min prior to treatment with or without 15(S)-HETE. The fl ux was expressed as percent dextran diffused per hour per square centimeter.

Transmigration
Endothelial transmigration was measured as described previously ( 21 ). Whenever adenoviral vectors were used, HUVECs were transduced with the control or the indicated adenovirus prior to seeding onto the transwell. In the case of inhibitors, they were added to cells just 30 min prior to the treatment. HUVECs were treated with or without 15(S)-HETE for 30 min, at which time the BCECF-AM-labeled quiescent THP1 cells (1 × 10 5 cells/well) or PBMC (1 × 10 5 cells/well) were added and incubation continued overnight at 37°C. The transmigration of THP1 cells or PBMCs through the HUVEC monolayer was measured by capturing the images by a fl uorescence microscope.

ROS production
Intracellular ROS production was determined using membrane permeable CM-H 2 DCFDA, as described previously ( 22 ). After the treatments, HUVECs were incubated with 10 mM CM-H 2 DCFDA for 30 min, washed with PBS, and resuspended in serum-free medium. The fl uorescence intensities of the resuspended cells were measured in a SpectraMax Gemini XS spectrofl uorometer (Molecular Devices) with excitation at 485 nm and emission at 535 nm. ROS production was expressed as RFUs.

XO activity
XO activity was measured using a kit according to the manufacturer's instructions (Cayman Chemicals) and expressed as RFUs.

Leukocyte adhesion
Leukocytes were isolated using Lymphoprep solution (catalog number 0460539, Fisher Scientifi c) from mouse blood. Mice were anesthetized by isofl urane inhalation and peripheral blood was collected by cardiac puncture into an Eppendorf tube containing EDTA. One volume of EDTA-treated blood was mixed with four volumes of 0.84% ammonium chloride and incubated at 4°C for 5 min. To ensure the elimination of erythrocytes, this process was repeated several times. Following the centrifugation at 1,000 g for 10 min, the white cell pellet was resuspended in PBS. Then, an equal volume of Lymphoprep solution was laid over the cell suspension and centrifuged at 800 g for 23 min. The leukocyteenriched layer was collected and resuspended in PBS followed by labeling with BCECF-AM. Aortas from WT and 12/15-LO Ϫ / Ϫ mice were dissected out, cleaned from connective and fat tissue, and treated with and without 15-(S)-HETE (0.5 M) in the presence and absence of the indicated inhibitors for 30 min. The aortas were then opened longitudinally and the BCECF-AM-labeled leukocytes (5.6 × 10 5 cells /ml) were added in serum-free medium and incubated for 1 h at room temperature. The aortas were fi xed with 3% paraformaldehyde for 30 min followed by 0.2% picric acid for 1 h, permeabilized in TBS [10 mM Tris-HCl and 150 mM NaCl (pH 8.0)] containing 3% BSA and 0.2% Triton X-100 for 10 min, blocked in 3% BSA for 1 h at room temperature, and probed with rabbit anti-JamA antibodies followed by Alexa Fluor 568-conjugated goat anti-rabbit secondary antibodies. In another set of experiments, aortas were isolated from both WT and 12/15-LO Ϫ / Ϫ mice that were on CD or HFD for 3 months, fi xed and stained for JamA and leukocytes using rabbit anti-JamA and mouse anti-CD45 immunocomplexes were separated by SDS-PAGE and visualized by Coomassie Brilliant Blue R250 staining. The JamA band was excised and subjected to in-gel digestion with trypsin. The resulting peptides were analyzed by LC-CID-MS/MS to identify the phosphorylated amino acid residues at the Bioanalytical and Mass Spectrometry facility at University of Alabama at Birmingham School of Medicine as described previously ( 23 ).

Construction of JamA expression vector
The human JamA (NM_016946) coding sequence was cloned from enriched cDNA by PCR amplifi cation. RNA was isolated from HUVECs and cDNA synthesized by reverse transcription reaction with Superscript III First-Strand synthesis system for RT-PCR (catalog number 1808051, Invitrogen) following the manufacturer's instructions. To enrich JamA mRNA, PVDF membrane ( ‫ف‬ 3 mm 2 pieces) coated with primers 5 ′ -CTTGTATGGTCTCT-GAGGAAGG-3 ′ and 5 ′ -ATACCCATTCCGTGCCTCACAG-3 ′ were incubated with cDNA at ambient temperature for 30 min and eluted in 10 mM Tris-HCl (pH 8.0) for 5 min at 60°C. The eluted cDNA was used as a template to amplify JamA coding region using the forward primer 5 ′ -ATGGGGACAAAGGCGCAAGT-3 ′ and the reverse primer 5 ′ -TCACACCAGGAATGACGAGGT-3 ′ with the high fi delity TaKaRa Ex Taq DNA polymerase. Restriction sites Bgl II and Kpn I were incorporated to the amplifi ed JamA coding region by PCR using the forward primer 5 ′ -TCGACC-G AGATCT CTACCATGGGGACAAAGGCGCA-3 ′ and the reverse primer 5 ′ -CTCGCC GGTACC TCGCACCAGGAATGACGAG-GTC-3 ′ (the restriction enzyme sites are underlined). The resulting 930 bp PCR product was then digested with Bgl II and Kpn I and then inserted into the Bgl II/ Kpn I sites of pCMV-Myc-C vector (catalog number 635689, Clonetech) to generate recombinant JamA (rJamA). The construct was confi rmed by restriction analysis and DNA sequencing.

Site-directed mutagenesis
The tyrosine (Y) residues at 52, 83, 107, 164, 218, and 280 in JamA were mutated to phenylalanine (F) with QuickChange sitedirected mutagenesis kit (Agilent Technologies) using the Myctagged JamA expression vector and the following primers: Y52F mutant: forward, 5 ′ -GAAGTTGTCCTGTGCCT T CTCGGGCTTTT- The mutated nucleotide is shown as a bold letter and the mutations were verifi ed by DNA sequencing using the vector-specifi c primers.

Flux assay
HUVECs were grown to a monolayer on the apical side of a polycarbonate membrane of the transwell (0.4 m pore size) and growth-arrested for 6 h. The monolayer was treated with and without 15(S)-HETE for 30 min, at which time, FITC-conjugated dextran (molecular mass ‫ف‬ 70,000 Da) was added to the basal chamber at 100 g/ml concentration, and after 2 h the fl uorescence intensity of the medium from each chamber was measured using SpectraMax Gemini XS spectrofl uorometer (Molecular An equal amount of protein from control and each treatment was either analyzed by Western blotting for the steady state levels of JamA and occludin using their specifi c antibodies (A) or immunoprecipitated with pTyr (PY20) antibodies followed by immunoblotting with the indicated antibodies (B). C: Equal amounts of proteins from control and 30 min of the indicated HETE-treated HUVECs were analyzed for JamA tyrosine phosphorylation as described in (B). D: All the conditions were the same as in (B) except that an equal amount of protein from control and each treatment was immunoprecipitated with anti-JamA antibodies and the immunocomplexes were analyzed by Western blotting for the indicated proteins using their specifi c antibodies. E: Quiescent HUVEC monolayer was treated with vehicle or 15(S)-HETE (0.1 M) for the indicated time periods, fi xed, and double immunofl uorescently stained for JamA and occludin using rabbit anti-JamA and mouse antioccludin antibodies followed by developing with Alexa Fluor 568-conjugated goat anti-rabbit and Alexa Fluor 488-conjugated goat antimouse secondary antibodies, respectively. Images were captured using an inverted Zeiss fl uorescence microscope (AxioObserver Z1) via a 40× NA 0.6 objective and AxioCam MRm camera without any enhancements. The TJ-localized JamA levels were quantifi ed using NIS Elements AR 3.1 imaging software (Nikon, Japan). The bar graphs in B, D, and E represent mean ± SD values of three experiments. * P < 0.05 versus vehicle control . Because 15(S)-HETE induced JamA tyrosine phosphorylation more robustly than occludin, we next focused the rest of the study on its role in 15(S)-HETE-induced TJ disruption and barrier dysfunction. Coimmunoprecipitation assays showed that JamA dissociates from occludin in response to 15(S)-HETE in a time-dependent manner with a maximum effect at 30 min and starting reassociation at 120 min ( Fig.  1D ). JamA dissociation/reassociation with occludin was correlated with its state of tyrosine phosphorylation. Colocalization for JamA and occludin by double immunofl uorescence staining revealed that both JamA and occludin were present in the TJs, and upon treatment with 15(S)-HETE, they both were found dislocated from the TJs ( Fig. 1E ).

Src and Pyk2 mediate 15(S)-HETE-induced JamA tyrosine phosphorylation
Previous work from our laboratory showed that Src and Pyk2 mediate 15(S)-HETE-induced ZO2 tyrosine phosphorylation leading to TJ disruption ( 20 ). Therefore, to understand the mechanism(s) involved in 15(S)-HETEinduced JamA tyrosine phosphorylation, we tested the role of Src and Pyk2. In line with our previous observations, 15(S)-HETE enhanced tyrosine phosphorylation of both Src and Pyk2 in a time-dependent manner with maximum increases between 5 min and 30 min and declining thereafter ( Fig. 2A ). Adenovirus-mediated expression of dnSrc or dnPyk2 attenuated JamA tyrosine phosphorylation and prevented its dissociation from occludin ( Fig. 2B, C ). Blockade of Src and Pyk2 by their dominant negative mutants also reduced 15(S)-HETE-induced JamA dislocation from TJs ( Fig. 2D ). In addition, inhibition of Src and Pyk2 by their dominant negative mutants blocked 15(S)-HETE-induced increase in EC barrier permeability and attenuated the transmigration of monocytes and leukocytes ( Fig. 2E-G ). Because dnSrc was able to block the tyrosine phosphorylation of both Src and Pyk2, and dnPyk2 inhibited only Pyk2 phosphorylation ( Fig. 2B ), it is likely that Src acts upstream to Pyk2 in the mediation of JamA tyrosine phosphorylation. Based on these fi ndings, it appears that 15(S)-HETE induces JamA tyrosine phosphorylation via Src-Pyk2 signaling and inhibition of JamA tyrosine phosphorylation by dnSrc and dnPyk2 is able to protect EC TJs and its barrier permeability from 15(S)-HETEinduced disruption .

Identifi cation of the 15(S)-HETE-induced tyrosine phosphorylation sites in JamA
To identify the tyrosine residues phosphorylated by 15(S)-HETE, we performed mass spectrometry. We observed six tyrosine residues, namely Y52, Y83, Y107, Y164, Y218, and Y280, that are phosphorylated in 15(S)-HETEtreated ECs ( Fig. 3A, B ). To identify which of these tyrosine residues showed increased phosphorylation in antibodies, followed by Alexa Fluor 568-conjugated goat anti-rabbit and Alexa Fluor 488-conjugated goat anti-mouse secondary antibodies. Fluorescence images of the luminal side of aortas were captured and quantifi ed as described above in immunofl uorescence staining of HUVECs.

Miles assay
The level of vascular permeability was determined by quantitative measurement of the Evans Blue dye diffusion into the aorta according to Miles test ( 24 ). WT and 12/15-LO Ϫ / Ϫ mice that were on either CD or HFD for 3 months were anesthetized, and 0.1 ml of 1% Evans Blue dye was injected into the inferior vena cava. After 30 min, blood vessels were perfused with PBS through the left ventriculum and aortas were isolated. Evans Blue dye was extracted from the arteries by incubating in formaldehyde at 55°C for 24 h, cleared by centrifugation and the optical density was measured at 610 nm. Vascular permeability was expressed as the amount of Evans Blue extravasated per milligram of artery.

Statistics
All the experiments were performed three times and the data are presented as mean ± SD. The treatment effects were analyzed by one-way ANOVA followed by Student's t -test and P values <0.05 were considered to be statistically signifi cant.

15(S)-HETE induces tyrosine phosphorylation of JamA in the disruption of EC TJs
Loss of barrier function is a hallmark of endothelial dysfunction ( 1, 2, 6 ). TJs play a crucial role in the maintenance of EC barrier function ( 3,4 ). Previously, we have reported that 15(S)-HETE increases EC barrier permeability by promoting threonine and tyrosine phosphorylation of ZO1 and ZO2, respectively, and their dissociation from TJ complexes ( 20,21 ). In the present study, we examined the role of transmembrane proteins, JamA and occludin, that are present in the TJ complexes in 15(S)-HETE-induced EC barrier disruption. First, we tested the effect of 15(S)-HETE on the steady state levels of JamA and occludin. 15(S)-HETE did not affect the steady state levels of either JamA or occludin ( Fig. 1A ). We then tested the effect of 15(S)-HETE on their tyrosine phosphorylation. 15(S)-HETE stimulated tyrosine phosphorylation of both JamA and occludin in a time-dependent manner with maximum increase between 5 min and 30 min and declining thereafter ( Fig. 1B ). The specifi city experiment showed that 15(S)-HETE was found to be more potent than 5(S)-HETE or its enantiomer 15(R)-HETE in the stimulation of JamA tyrosine phosphorylation ( Fig. 1C ). However, 12(S)-HETE, that is preferentially produced by the murine ortholog of human 15-LO1, namely, 12/15-LO ( 25 ), also stimulated JamA tyrosine phosphorylation almost to the same level as that of 15(S)-HETE. Cell extracts were immunoprecipitated with anti-JamA antibodies and resolved on SDS-PAGE, stained with Coomassie Brilliant Blue, and the molecular mass band corresponding to JamA was subjected to high resolution mass spectrometric analysis (Orbitrap Velos) with fragmentation carried out in both Collision-induced dissociation (CID) and Electron transfer dissociation (ETD) modes . The resulting MS/MS scans are shown for the confi rmed phosphorylated tyrosine residues indicated in (A). The phosphorylated amino acid residues for each spectra indicated were manually confi rmed and included a minimal probability score of 100%, localization score of >98%, A-score of >20 (with the exception of Y107 at 13.4), charge state of у 2, delta ppm of р 3, and Sequest "XCorr" score of >4.0 (with the exception of Y280 at 3.8). Of note, all cysteine residues were carbamidomethylated with iodoacetamide resulting in an increased mass of 57.0293 Da. C, D: HUVECs were transiently transfected with empty vector (EV) or Myctagged rJamA expression vector with or without Y52F, Y83F, Y107F, Y164F, Y218F, or Y280F mutations, grown to confl uence, quiesced, treated with vehicle or 15(S)-HETE (0.1 M) for 30 min and an equal, amount of protein from control and each treatment was immunoprecipitated with anti-pTyr or anti-Myc antibodies and the immunocomplexes were analyzed by Western blotting using the indicated antibodies. In (C), an equal amount of protein from control and each treatment was also analyzed by Western blotting for cMyc using its specifi c antibodies to show the expression of rJamA. E: All the conditions were the same as in (C) except that after quiescence and JamA ( Fig. 4E ). Thus, these data confi rm that 15(S)-HETE stimulates ROS production, specifi cally H 2 O 2 , leading to tyrosine phosphorylation of Src, Pyk2, and JamA by enhancing XO activity. Based on these observations, we examined the role of XO in 15(S)-HETE-induced Src, Pyk2, and JamA tyrosine phosphorylation as well as EC barrier dysfunction. Allopurinol abrogated 15(S)-HETE-induced tyrosine phosphorylation of Src, Pyk2, and JamA and prevented JamA dissociation from occludin ( Fig. 4F-H ). In accordance with these fi ndings, double immunofl uorescence staining for JamA and occludin pointed out that inhibition of XO activity by allopurinol prevents 15(S)-HETEinduced dislocation of JamA and occludin from TJs ( Fig. 4I ). 15(S)-HETE-induced EC permeability as well as monocyte transmigration through the EC monolayer were also attenuated by allopurinol ( Fig. 4J, K ). These results indicate that XO-dependent ROS production is required for 15(S)-HETE-induced Src-Pyk2-mediated JamA tyrosine phosphorylation, JamA dissociation from occludin, TJ disruption, and EC barrier dysfunction. These results infer that 15(S)-HETE-induced XO activation appears to be upstream to Src-Pyk2 signaling in mediating JamA tyrosine phosphorylation and TJ disruption .

15(S)-HETE disrupts aortic endothelial TJs via XO-, Src-, and Pyk2-dependent JamA tyrosine phosphorylation ex vivo
To gain additional support for the role of JamA tyrosine phosphorylation in 15(S)-HETE-induced EC barrier disruption, we examined its effects on aortic endothelial TJs ex vivo. Aortas from WT mice were exposed to 15(S)-HETE in the presence and absence of pharmacological inhibitors of Src, Pyk2, or XO, and were then examined for JamA tyrosine phosphorylation, JamA dissociation from occludin, JamA dislocation from TJs, and leukocyte adhesion. 15(S)-HETE induced JamA tyrosine phosphorylation and its dissociation from occludin, Src, and Pyk2 tyrosine phosphorylation, JamA disappearance from TJs and leukocyte adhesion to the artery and inhibition of Src or Pyk2 by their pharmacological inhibitors, PP1 and PF431396, respectively ( 26,27 ), prevented these effects ( Fig. 5A-C, E, F ). PP3, a structural analog of PP1 that does not inhibit Src ( 26 ), had no apparent effect on 15(S)-HETE-induced Src, Pyk2, or JamA tyrosine phosphorylation, which confi rms the specifi city of PP1 ( Fig. 5D ). Allopurinol also efficiently blocked JamA tyrosine phosphorylation and its dissociation from occludin, Src, and Pyk2 phosphorylation, JamA disappearance from TJs, and the adhesion of leukocytes to the artery ( Fig. 6A-E ). These observations reveal that 15(S)-HETE can stimulate JamA tyrosine phosphorylation in a XO-, Src-, and Pyk2-dependent manner, even in intact artery, and these events trigger JamA dissociation from occludin and thereby their dislocation from TJs facilitating response to 15(S)-HETE, we constructed a Myc-tagged rJamA expression vector and mutated the tyrosine (Y) residues to phenylalanine (F) by site-directed mutagenesis . Then, ECs were transfected with WT or mutant rJamA expression vectors, treated with 15(S)-HETE, and the rJamA tyrosine phosphorylation and its association/ dissociation with occludin was measured . In response to 15(S)-HETE, WT and Y52F, Y83F, and Y107F mutant rJa-mAs showed ‫ف‬ 5-to ‫ف‬ 7-fold increase in their tyrosine phosphorylation as compared with control ( Fig. 3C ). In contrast, Y164F, Y218F, and Y280F mutant rJamAs failed to be phosphorylated by 15(S)-HETE ( Fig. 3C ), suggesting that 15(S)-HETE stimulates JamA phosphorylation on these tyrosine residues. Consistent with their phosphorylation states, while the WT and Y52F, Y83F, and Y107F rJamA mutants showed dissociation from occludin, the Y164F, Y218F, and Y280F rJamA mutants remained complexed with endogenous occludin and TJs ( Fig. 3D, E ). Similarly, while WT and Y52F, Y83F, and Y107F rJamA mutants failed to protect ECs from 15(S)-HETE-induced increase in their permeability, the Y164F, Y218F, and Y280F rJamA mutants exhibited resistance to this effect of 15(S)-HETE ( Fig. 3F ). These observations demonstrate that Y164, Y218, and Y280 residues of JamA are critical for its capacity to form a TJ complex with occludin and maintain barrier function.

XO-dependent ROS production mediates 15(S)-HETEinduced JamA tyrosine phosphorylation and EC barrier dysfunction
Oxidative stress plays a critical role in EC dysfunction ( 7 ). Many studies have shown that XO-mediated ROS generation is linked to EC dysfunction and atherosclerosis (16)(17)(18)(19). Hence, to fi nd whether XO-mediated ROS production has any role in 15(S)-HETE-stimulated JamA tyrosine phosphorylation and TJ disruption, we fi rst tested its effect on ROS production in ECs. 15(S)-HETE stimulated ROS production in a time-dependent manner with maximum increase at 10 min ( Fig. 4A ). We then tested the effect of 15(S)-HETE on XO activity. 15(S)-HETE induced XO activity in a time-dependent manner with a maximum ‫ف‬ 5-fold increase at 10 min as compared with control ( Fig.  4B ). Furthermore, allopurinol, a highly specifi c inhibitor of XO ( 16 ), completely blocked 15(S)-HETE-induced ROS production ( Fig. 4C ). To further confi rm the ROS produced by 15(S)-HETE, we studied the effect of catalase. Incubation of cells with PEG-catalase (25 U/ml) completely abolished 15(S)-HETE-induced ROS production ( Fig. 4D ). To explore the role of redox-sensitive mechanisms in 15(S)-HETE-induced Src, Pyk2, and JamA tyrosine phosphorylation, we tested the effect of catalase on these events. PEG-catalase (25 U/ml) prevented 15(S)-HETE-induced tyrosine phosphorylation of Src, Pyk2, and treatments, the HUVEC monolayer was stained for endogenous JamA and rJamA using goat anti-JamA and rabbit anti-cMyc antibodies followed by Alexa Fluor 488-conjugated donkey anti-goat and Alexa Fluor 568-conjugated goat anti-rabbit secondary antibodies, respectively. Fluorescence images were captured and quantifi ed as described in Fig. 1E . F: All the conditions were the same as in (C), except that after quiescence the HUVEC monolayer was subjected to 15(S)-HETE (0.1 M)-induced dextran fl ux assays. The bar graphs represent mean ± SD values of three experiments. * P < 0.05 versus vehicle control; ** P < 0.05 versus WT + 15(S)-HETE.   1 M) in the presence and absence of either XO inhibitor, allopurinol (50 M) (C) or PEG-catalase (25 U/ml) (D) for 30 min and ROS production was measured. E: All the conditions were the same as in (D) except that after the indicated treatments cell extracts were prepared and an equal amount of protein from control and each treatment was analyzed for phospho and total Src, Pyk2, and JamA levels as described in Fig. 2B . F-H: All the conditions were same as in (C) except that after the indicated treatments cell extracts were prepared and an equal amount of protein from control and each treatment was analyzed by Western blotting for pSrc and pPyk2 levels using their phospho-specifi c antibodies and the blots were reprobed for total Src and Pyk2 levels for normalization (F) or immunoprecipitated with anti-pTyr (G) or anti-JamA (H) antibodies and the immunocomplexes were analyzed by Western blotting for JamA and occludin using their specifi c antibodies. In (G), an equal amount of protein from no adverse effects on TJ organization ( 35 ). Endothelial TJs are comprised of transmembrane (occludin, claudins, and junction adhesion molecules) and intracellular (zona occludens 1-3, protein incorporated later into TJs, and junction enriched and associated protein) proteins ( 33,36,37 ). A large body of evidence shows that interference in the interactions between TJ proteins perturbs the TJs affecting the barrier permeability of the endothelium ( 38 ). It was further observed that the TJ formation or disintegration depends on the phosphorylation state of TJ proteins (39)(40)(41). However, most of these studies were focused on epithelial TJs, and therefore, the role of phosphorylation of TJ proteins in the preservation or perturbation of endothelial TJ integrity and its barrier function has not been well-studied. In this context, a few studies, including ours, showed that phosphorylation of TJ proteins leads to endothelial TJ disruption and loss of its barrier function ( 20,21,42,43 ). In the present study, we have extended our previous observations on the role of tyrosine phosphorylation of TJ proteins in the disruption of TJ integrity and barrier function to JamA. We found that 15(S)-HETE, the 15-LO conversion product of AA, induces tyrosine phosphorylation of JamA leading to its dissociation from occludin and disrupting TJs and barrier function both in vitro in ECs and ex vivo in intact arteries. Because interference with Src and Pyk2 activation attenuated 15(S)-HETE-induced JamA tyrosine phosphorylation and its dissociation from occludin and prevented TJs and barrier function from disruption by 15(S)-HETE in ECs and arteries, it is likely that 15(S)-HETE mediates its adverse effects on TJs via Src-and Pyk2-dependent JamA tyrosine phosphorylation. As 15(S)-HETE induced JamA phosphorylation on tyrosine residues 164, 218, and 280 and mutation of any one of these residues to phenylalanine prevented 15(S)-HETE-induced JamA phosphorylation, and its dissociation from occludin and TJs restoring EC barrier function, these tyrosine residues appear to be crucial for the capacity of JamA in forming a TJ complex with occludin . Furthermore, because these three tyrosine residues are present in the carboxyl terminus, it may be suggested that JamA interacts with other TJ proteins in establishing functional TJs via its carboxyl region. Previously, we have also demonstrated that Src-and Pyk2dependent phosphorylation of ZO2 leads to its dissociation from claudin 1/5 affecting endothelial TJs and barrier function. In view of these observations, it may be concluded that 15(S)-HETE disrupts endothelial TJs and their barrier function by tyrosine phosphorylation of several TJ proteins, including ZO2 and JamA, and affecting their interactions with other TJ proteins such as claudins 1/5 and occludin , respectively. These observations may also infer that 15(S)-HETE triggers simultaneous phosphorylation of several TJ proteins leading to disruption of EC TJ complexes and barrier function in a coordinate manner, as blockade of enhanced trapping of leukocytes onto the dysfunctional endothelium .

Deletion of 12/15-LO gene blocks XO activity and protects aortic endothelial TJs from HFD-induced disruption
In humans the sources of 15(S)-HETE production are 15-LO1 and 15-LO2 ( 25,28,29 ); however, in rodents neither of these LOs is expressed. Rather, 12/15-LO, the mouse homolog of 15-LO1 is expressed and possesses the capacity to convert AA to 12(S)-HETE and 15(S)-HETE, and both appear to exhibit at least some similar cellular responses ( 30,31 ). Therefore, to validate the role of 15(S)-HETE in EC TJ disruption and barrier dysfunction, here we used a 12/15-LO knockout mouse model. WT, but not 12/15-LO Ϫ / Ϫ , mice fed with HFD for three months showed an increase in the expression of XO and its activity as compared with control diet ( Fig. 7A-C ). Consistent with these results, feeding with HFD also disrupted aortic EC TJs, as demonstrated by decreased JamA staining, and this effect was correlated with an increase in aortic endothelial permeability and leukocyte adhesion only in WT mice, but not in 12/15-LO Ϫ / Ϫ mice ( Fig. 7D, E ). To confi rm the role of XO in HFD-induced vascular permeability, WT mice were fed with HFD, alone or in combination with administration of allopurinol for 3 months, and XO activity, Src, Pyk2, and JamA tyrosine phosphorylation, JamA dissociation from occludin, endothelial TJ integrity, and vascular permeability were measured. Arteries from WT mice fed with HFD showed increased XO activity, Src, Pyk2, and JamA tyrosine phosphorylation, JamA dissociation from occludin, endothelial TJ disruption, and vascular permeability as compared with those from CD-fed mice, and all these effects were prevented by simultaneous administration of allopurinol ( Fig. 7F-J ).

DISCUSSION
AJs and TJs are essential structural components of the endothelial barrier ( 3,4 ). While AJs form cell-to-cell and cell-to-substratum contacts, the TJs, by forming strand-like structures, close the clefts between the neighboring ECs and regulate the paracellular permeability of the endothelium to ions and macromolecules ( 3 ); although, some reports demonstrated that disruption of AJs also results in endothelial hyperpermeability ( 32 ). Because dynamic interactions between AJs and TJs are required to modulate the EC barrier function ( 33,34 ), the EC hyperpermeability caused by AJ disruption could still be attributed to perturbation in TJs as well. In fact, it was proposed that AJs are linked to the development of TJs ( 34 ); however, some studies disputed this claim, as interfering with the formation of AJs has the same cell extracts was also analyzed by Western blotting for JamA levels for normalization. I: All the conditions were the same as in (C) except that quiescent the HUVEC monolayer, after treatment with vehicle or 15(S)-HETE (0.1 M), was stained for JamA and occludin and the fl uorescence images were captured and the TJ-localized JamA levels were measured as described in Fig. 1E   The same tissue extracts were analyzed by Western blotting for total JamA levels shown in (A) or for pSrc and pPyk2 levels and normalized to their total levels in (C). D: Aortas from WT mice were incubated with and without 15(S)-HETE (0.5 M) in the presence and absence of PP3 (10 M) for 30 min, tissue extracts were prepared and analyzed for Src, Pyk2, and JamA tyrosine phosphorylation as described in (A) and (C). E, F: All the conditions were same as in (A) except that after the treatments, the aortas were opened longitudinally, either left alone (E) or further incubated with BCECF-labeled leukocytes for 60 min (F), fi xed, permeabilized, blocked and stained for JamA using rabbit anti-JamA antibodies, followed by developing with Alexa Fluor 568-conjugated goat anti-rabbit secondary antibodies. The immunofl uorescence images were captured and the TJ-localized JamA levels were quantifi ed as described in Fig. 1E . The bar graphs represent mean ± SD values of three experiments or six animals. * P < 0.05 versus vehicle control; ** P < 0.05 versus 15(S)-HETE. could allow the transendothelialization of leukocytes and monocytes, the hallmarks in the initiation of infl ammation. In the present study, we observed that when the EC monolayer or aortas were exposed to 15(S)-HETE, their TJs were disrupted and caused enhanced leukocyte and monocyte adhesion and transmigration. Furthermore, feeding WT, but not 12/15-LO Ϫ / Ϫ , mice with HFD also disrupted aortic endothelial TJs and exhibited increased leukocyte adhesion to the endothelium. In addition, hypercholesterolemia and hyperlipidemia were reported to induce the expression of 12/15-LO and generate 15(S)-HETE (46)(47)(48). Furthermore, a number of studies have demonstrated 12/15-LO plays a role in atherogenesis (49)(50)(51). These fi ndings, along with the observations that the effects of HFD feeding on TJ integrity and barrier function were substantially downregulated in 12/15-LO Ϫ / Ϫ mice, may suggest that 12/15-LO plays a crucial role in endothelial TJ disruption and leukocyte or monocyte phosphorylation of any of these proteins is suffi cient in restoring TJs and barrier function. In both ZO2 and JamA tyrosine phosphorylation, Src was found to act upstream to Pyk2. An association between Pyk2 and Src was reported in epithelial cells in response to EGF-induced wound healing ( 44 ). Similarly, Pyk2 was found to bind to Src and mediate its activation in PC12 cells in response to lysophosphatidic acid ( 45 ). In both of these reports, Pyk2 was observed to act upstream to Src. Based on these observations and the present fi ndings, it can be viewed that the spatiotemporal interactions between Src and Pyk2 differ in response to various agents in mediating their cellular effects.
In addition to their role in the regulation of barrier permeability, endothelial TJs play a role in the maintenance of cell polarity ( 39 ), signaling ( 38 ), and gene expression ( 4,5 ). Most importantly they provide fencing to the endothelium in intact blood vessels. Therefore, the disruption of TJs may result in leaky EC barrier permeability, which and tissue extracts were prepared. An equal amount of protein from control and each treatment was analyzed for JamA tyrosine phosphorylation and its dissociation from occludin as described in Fig. 5A, B , respectively. C: The same tissue extracts were analyzed for pSrc and pPyk2 levels and normalized to their total levels as described in Fig. 5C . D, E: Aortas from WT mice after being subjected to the treatments described in (A) were opened longitudinally, left alone, or further incubated with BCECF-labeled leukocytes as described in Fig. 5F and stained for JamA using rabbit anti-JamA antibodies followed by developing with Alexa Fluor 568-conjugated goat anti-rabbit secondary antibodies. The bar graphs represent mean ± SD values of three experiments or six animals. * P < 0.05 versus vehicle control; ** P < 0.05 versus 15(S)-HETE. Ϫ / Ϫ mice that were fed with CD or HFD for 3 months were isolated and analyzed for XO levels by Western blotting using XO specifi c antibodies (A) or its activity using a kit (B ). The blot in (A) was normalized for ␣ -tubulin. C: All the conditions were the same as in (A) except that after isolation, the arteries were fi xed in OCT compound, cross-sections were made and stained for XO and CD31 using rabbit anti-XO and rat anti-CD31 antibodies followed by Alexa Fluor 568-conjugated goat anti-rabbit and Alexa Fluor 488-conjugated goat anti-rat secondary antibodies. D: All the conditions were the same as in (A) except that after dissection the arteries were opened longitudinally and stained for JamA and leukocytes using rabbit anti-JamA and mouse anti-CD45 antibodies followed by Alexa Fluor 568-conjugated goat anti-rabbit and Alexa Fluor 488-conjugated goat anti-mouse secondary antibodies. E: After feeding with CD or HFD, the mice were anesthetized and 0.1 ml of 1% Evans Blue (EB) dye was injected into the inferior vena cava. After 30 min, the blood vessels were perfused with PBS through the left ventriculum and the aortas were isolated and photographed. After taking the pictures, the aortas were minced, incubated in formamide at 55°C for 24 h and the optical density was measured at 610 nm. The aortic endothelial permeability was expressed as nanograms of EB dye extravasated per milligram of tissue. F-J: WT mice were fed with CD or HFD, alone or in combination with allopurinol (6.8 g/l) for 3 months and either arteries were isolated and analyzed for XO activity (F), phospho and total Src, Pyk2, and JamA levels (G), JamA dissociation from occludin (H), endothelial TJ integrity (I), or vascular leakage (J) as described in (B), (A), (D), and (E), respectively. The bar graphs represent mean ± SD values of six animals. * P < 0.05 versus CD; ** P < 0.05 versus WT (HFD). recruitment to the dysfunctional endothelium, which are considered to be early events in atherogenesis ( 1,2 ). However, how the disruption of endothelial TJs enhances the leukocyte and monocyte recruitment is not known. One proposed mechanism is that dysfunctional endothelium expresses cell adhesion molecules, such as ICAM and VCAM, which may aid in the adhesion of circulating leukocytes and monocytes to the perturbed endothelium, and the chemotactic molecules such as MCP1 produced by the dysfunctional endothelium may enhance their migration through the endothelium. Support toward this mechanism can be drawn by the fi ndings that 12/15-LO, via stimulating protein kinase C ␣ -dependent activation of NF B, enhances ICAM expression in ECs and mediates endothelium-monocyte interaction ( 52 ). In addition, it was shown that TJ disruption infl uences NF B activation leading to expression of infl ammatory molecules and promotes leukocyte extravasation ( 53,54 ). Based on all these observations, we believe that HFD-induced expression of 12/15-LO enhances the tyrosine phosphorylation of TJ proteins such as JamA and ZO2, disrupts TJs, and damages the fencing function of the TJs, which may lead to the expression of cell adhesion and/or infl ammatory molecules, thereby enhancing the recruitment of monocytes to the vessel wall, the events that exacerbate the progression of atherosclerosis. However, it is also possible that monocytes adhere to the endothelium, disrupt the TJs, and transmigrate through the disrupted TJs into the subendothelial space, and both of these events may be required 12/15-LO expression/activation by HFD . XO, by its capacity to produce ROS, is considered to be a predominant mediator of EC dysfunction ( 12,18,19 ). But the mechanism of XO activation in EC dysfunction is not well-understood. In this context, our present study demonstrates that 15(S)-HETE induces XO activity producing ROS. Similarly, feeding mice with HFD induces XO expression only in WT mice, but not in 12/15-LO Ϫ / Ϫ mice, suggesting that 12/15-LO and its conversion product of AA, 15(S)-HETE, play a role in the induction of XO expression and its activity. Furthermore, inhibition of XO was suffi cient in blocking Src-Pyk2-mediated JamA tyrosine phosphorylation and its dissociation from occludin and preventing TJ disruption and barrier dysfunction in vitro, ex vivo, and in vivo. These observations reveal that XO is not only induced by, but also is required for, 12/15-LO-15(S)-HETE-induced TJ disruption and barrier hyperpermeability. Many other reports show that hypercholesterolemia and HFD induce XO expression, and inhibition of XO attenuates atherosclerosis ( 16,17 ). Similarly a large body of data indicates a role for 12/15-LO in atherogenesis (49)(50)(51). However, whether there is any interaction between these two enzymes in atherogenesis has never been addressed. In this aspect, our fi ndings show for the fi rst time that 12/15-LO, via activation of XO, generates ROS, which in turn, enhances non-receptor tyrosine kinases, such as Src and Pyk2, and thereby mediating JamA tyrosine phosphorylation affecting endothelial TJs and barrier function. 12/15-LO has been reported to mediate LDL oxidation ( 55 ). Because 12/15-LO triggers ROS production via activation of XO, it