Supplemental material to this article can be found at: http://molpharm.aspetjournals.org/content/suppl/2016/11/09/mol.116.105031.DC1

1521-0111/91/1/1–13$25.00 http://dx.doi.org/10.1124/mol.116.105031 MOLECULAR PHARMACOLOGY Mol Pharmacol 91:1–13, January 2017 Copyright ª 2016 by The American Society for Pharmacology and Experimental Therapeutics

Novel Small Molecule JP-153 Targets the Src-FAK-Paxillin Signaling Complex to Inhibit VEGF-Induced Retinal Angiogenesis s

Jordan J. Toutounchian, Jayaprakash Pagadala, Duane D. Miller, Jerome Baudry, Frank Park, Edward Chaum, and Charles R. Yates Department of Pharmaceutical Sciences (J.J.T., J.P., D.D.M., F.P., C.R.Y.) and Department of Ophthalmology (E.C., C.R.Y.), University of Tennessee Health Science Center, Memphis, Tennessee; Department of Biochemistry and Cellular and Molecular Biology at The University of Tennessee, Knoxville, Tennessee; and UT/ORNL Center for Molecular Biophysics, Oak Ridge National Laboratory, Oak Ridge, Tennessee (J.B.) Downloaded from Received May 9, 2016; accepted October 28, 2016

ABSTRACT Targeting vascular endothelial (VEGF) is a common downstream activation of Akt (S473), resulting in reduced treatment strategy for neovascular eye disease, a major cause migration and proliferation of retinal endothelial cells stimu- of vision loss in diabetic retinopathy and age-related macular lated with VEGF. However, this effect did not prevent the initial molpharm.aspetjournals.org degeneration. However, the decline in clinical efficacy over time activation of either Src or FAK. Furthermore, topical application of in many patients suggests that monotherapy of anti-VEGF a JP-153-loaded microemulsion affected the hallmark features of therapeutics may benefit from adjunctive treatments. Our pathologic retinal angiogenesis, reducing neovascular tuft previous work has shown that through decreased activation of formation and increased avascular area, in a dose-dependent the cytoskeletal protein paxillin, growth factor–induced ischemic manner. In conclusion, our results suggest that using small retinopathy in the murine oxygen-induced retinopathy model molecules to modulate the protein paxillin is an could be inhibited. In this study, we demonstrated that VEGF- effective strategy for treating pathologic retinal neovasculari- dependent activation of the Src/FAK/paxillin signalsome is zation. To our knowledge, this is the first paradigm validating required for human retinal endothelial cell migration and pro- modulation of paxillin to inhibit angiogenesis. As such, we have liferation. Specifically, the disruption of focal adhesion kinase identified and developed a novel class of small molecules at ASPET Journals on September 27, 2021 (FAK) and paxillin interactions using the small molecule JP-153 aimed at targeting focal adhesion protein interactions that are inhibited Src-dependent of paxillin (Y118) and essential for pathologic neovascularization in the eye.

Introduction advanced stages of both diseases owing to aberrant ocular angiogenesis and neovascularization (Aiello et al., 1994; Diabetic retinopathy and age-related macular degeneration Ferris et al., 1984). Vascular endothelial growth factor are among the most common causes of blindness in adults (VEGF) plays a key role in this pathophysiology and is the (Pascolini and Mariotti, 2012). Vision loss occurs in the target of current FDA-approved antiangiogenic protein ther-

This work was funded by the University of Tennessee College of Pharmacy apeutics (Ozaki et al., 1999; Osborne et al., 2004; Nowak, 2006; (Pharmaceutical Sciences) Research Enhancement Seed Grant (2014) and the Wilkinson-Berka et al., 2013; http://www.fda.gov/NewsEvents/ ’ University of Tennessee Research Foundation s Technology Maturation Fund Newsroom/PressAnnouncements/ucm433392.htm). However, Program (2015). Conflict of interest statement: Jordan J. Toutounchian, Jayaprakash Pagadala, Duane D. Miller, Frank Park and Charles R. Yates prospective studies show a decline in long-term efficacy, which are listed on the patent application entitled “Inhibitors of paxillin binding and is believed to result from the emergence of VEGF-independent related compositions and methods” US Patent Application number 61/935,616. JP-153 is a patent-pending technology owned by the University of Tennessee mechanisms and expression of other growth factors and Research Foundation. No competing financial interests exist for authors cytokines involved in maintaining the abnormal angiogenic Jerome Baudry or Edward Chaum. Portions of this work were previously presented at the annual meeting of the milieu (Bergers and Hanahan, 2008; van Beijnum et al., Association for Research in Vision and Ophthalmology (ARVO) in Denver, CO, 2015). In addition, the further decline in visual function with June 2015, and published as Toutounchian JJ, Pagadala J, Miller DD, Steinle JJ, and Yates R (2015) The role of a Src/FAK-paxillin signalsome in VEGF- long-term anti-VEGF therapy has been linked to the loss of induced retinal neovascularization. Invest Ophthalmol Vis Sci 56:208–208. the choroidal blood supply, which is in part VEGF-dependent dx.doi.org/10.1124/mol.116.105031. s This article has supplemental material available at molpharm. and which supports the integrity and health of the overlying aspetjournals.org. retinal pigment epithelium and neural retina (Marneros et al.,

ABBREVIATIONS: AV, avascular area; DAPI, 49, 6-diamidino-29-phenylindole; DMSO, dimethyl sulfoxide; ERK, extracellular signal-regulated kinase; FA, focal adhesion; FAC, focal adhesion complex; FAK, focal adhesion kinase; GIT-1, ADP ribosylation factor GTPase-activating protein; LY294002, 2- morpholin-4-yl-8-phenylchromen-4-one; MAPK, mitogen-activated protein kinase; NV, neovascularization; OIR, oxygen-induced retinopathy; PARP, poly(ADP ribose) polymerase; PBS, phosphate-buffered saline; PI, propidium iodide; PI3K, phosphatidylinositol-4,5-bisphosphate 3-kinase; REC, retinal endothelial cell; RNV, retinal neovascularization; 6-B345TTQ, 6-Bromo-3,4-dihydro-4-(3,4,5-trimethoxyphenyl)-benzo[h]quinolin-2(1H)-one; SU6656, (3Z)- N,N-dimethyl-2-oxo-3-(4,5,6,7-tetrahydro-1H-indol-2-ylmethylidene)-2,3-dihydro-1H-indole-5-sulfonamide; VEGF, vascular endothelial growth factor.

1 2 Toutounchian et al.

2005; Saint-Geniez et al., 2008; Tokunaga et al., 2014). Thus, this report is the first paradigm validating small-molecule targeting downstream signaling linked to pathologic modulation of paxillin within FAs to prevent pathologic angio- neovascularization represents an alternative or adjunctive genesis in neovascular disease. With this study, we have approach to approved anti-VEGF treatments that may reduce exploited paxillin as our molecular target and have identified the damaging effects of antiangiogenic therapy. a novel class of small-molecule modulators of the FA protein VEGF activates endothelial cells, in part, by stimulating interactions essential for retinal neovascularization. pathways that regulate the enzymatic turnover of adhesion complexes, or “signalsomes” consisting of adaptor proteins and kinases, Materials and Methods e.g., Src-family kinases, focal adhesion kinase (FAK), and Reagents/Antibodies. Recombinant human VEGF-165A protein paxillin (Waltenberger et al., 1994; Abedi and Zachary, 1997; was purchased from R&D Systems (Minneapolis, MN). Total VEGFR- Provenzano and Keely, 2011). Targeting focal adhesion (FA) 2, Akt, and p44/42 MAPK [extracellular signal-regulated kinase kinases downstream of growth factor receptor activation (ERK1/2)] as well as phosphorylated VEGFR-2 (Tyr1175), FAK has recently emerged as an effective strategy for inhibiting (Y397, Y576/577, Y925), Akt (Ser473), cleaved and total poly(ADP retinal angiogenesis (Wary et al., 2012). In ischemic models ribose) polymerase (PARP), GAPDH, and ERK 1/2 (Thr202/Tyr204) of retinopathy, the local silencing of Src or FAK expression were acquired from Technologies (Danvers, MA). Phosphorylated paxillin (Y118) and FAK (Y861) were purchased from Downloaded from causes a significant reduction in pathologic neovascular Abcam (Cambridge, MA). Mouse antibodies against human paxillin disease (Kornberg et al., 2004; Werdich and Penn, 2006). (clone 349) and FAK (clone 77) were purchased from BD Biosciences However, evidence of resistance is also accumulating, as re- (San Jose, CA). Mouse a- primary antibody and secondary cently demonstrated when cells deficient in FAK protein showed antibodies IRDye 800CW goat anti-rabbit and IRDye 680LT goat anti- enhanced expression of its homolog, -rich kinase mouse were purchased from LI-COR Biotechnology (Lincoln, NE). 2 (PYK2), which is also known to regulate expression and Calcein-AM was obtained from BD Biosciences. DAPI nuclear stain endothelial budding or sprouting via VEGF-dependent mecha- was purchased from ThermoFisher Scientific (Pierce; Sunnyvale, CA). molpharm.aspetjournals.org nisms (Bergers and Hanahan, 2008; Weis et al., 2008; Shen 6-B345TTQ and the Src kinase inhibitor SU6656 were purchased et al., 2011; Eke and Cordes, 2015). Thus, there is a critical need from Sigma-Aldrich (St. Louis, MO). LY294002 (PI3K inhibitor) was to identify alternative drug targets that serve as common acquired from Cell Signaling Technologies. Primary antibody names, catalog numbers, species of origin, and dilutions are included in “interface points” shared by proteins within the focal adhesion Supplemental Table 1. complex (FAC). JP-153 was synthesized in accordance with the methods devised for Paxillin is a multidomain adaptor protein that binds to both ortho-functionalization of aniline derivatives (Houlden et al., 2010). FAK and PYK2, as well as numerous other FA proteins (e.g., Briefly, naphthylisocyanate 1 (5.9 mmol, 1.0 g) was added to a solution GIT-1, , and actopaxin) (Turner, 2000). Studies char- of t-butylisopropylamine (5.9 mmol, 0.9 ml) in diethyl ether (10 ml) acterizing these protein-protein interactions at the structural under stirring at room temperature. The colorless solution was stirred at ASPET Journals on September 27, 2021 level have identified highly conserved four-helix bundled for 3 hours and subsequently cooled to 0°C. Tetramethylethylenedi- regions, or so called paxillin-binding subdomains, which spe- amine (12.98 mmol, 2.0 ml) was added followed by n-butyllithium cifically engage the paxillin N-terminal -rich domains (11.8 mmol, 2.43 M in hexanes, 3.0 ml). The clear yellow solution was (Brown et al., 1998; Arold et al., 2002; Vanarotti et al., 2014). then stirred for 3 hours, during which time a white precipitate formed. The reaction mixture was cooled to –78°C and aldehyde 2 (8.85 mmol, Paxillin, together with Src and FAK, recruit other proteins to ’ 1.7 g) in tetrahydrofuran (5 ml) was added dropwise over 4 minutes. the cell s leading edge where actin filaments coalesce around Following the addition, ethanol (5 ml) was added rapidly and the “ ” (cellular anchors ) to provide mechanical forces mixture was allowed to warm to room temperature and stirred for needed to pull the cell forward. Since these complexes help 1 hour. The reaction mixture was then concentrated in vacuo, diluted assemble and support the connections between the actin with dichloromethane, and washed with saturated ammonium chlo- cytoskeleton and the , targeting these ride, NH4Cl (aqueous). The organic layer was evaporated onto silica proteins with small molecules would dismantle the FA com- and purified by column chromatography. JP-153 purity was charac- plexes and obstruct proliferative and migratory signal trans- terized with high-resolution mass and nuclear magnetic resonance duction during angiogenesis (Fig. 9). spectroscopy. JP-153 and 6-B345TTQ structures and calculated LogP We have identified a proliferative response phenotype of hu- values are presented in Supplemental Fig. 1. Primary Retinal Endothelial Cell Culture. Primary human man primary retinal endothelial cells (REC) exposed to high-dose retinal endothelial cells (Lot 181) were purchased from Cell Systems ionizing radiation (Toutounchian et al., 2014). Irradiation- Corporation (Kirkland, Washington). Cells were grown on attachment- enhanced paxillin Y118 phosphorylation, which was reduced by factor surfaces in M131 medium containing microvascular growth mitogen-activated protein kinase (MAPK) inhibition. Under these supplements (Invitrogen, Carlsbad, CA) gentamicin (10 mg/ml) and same mechanisms, inhibiting MAPK and, thus, paxillin phos- amphotericin B (0.25 mg/ml). Only primary cells up to passage six were phorylation caused a reduction in in vivo retinal angiogenesis. used. For immunoassays, RECs were plated into six-well plates, Our data suggested a direct role for activated paxillin in radiation- cultured for 2 days, and serum-deprived using 0.1% bovine serum induced retinopathy, an ischemic inflammatory disease with a albumin (Sigma-Aldrich) overnight prior to experiments. RECs were neovascular component (Boozalis et al., 1987; Finger, 2008). pretreated with inhibitors, SU6656 (1 mM), LY294002 (10 mM), or m However, the mechanisms by which paxillin coordinates JP-153 (1 M), for 1 hour prior to VEGF (100 ng/ml) stimulation, unless mentioned otherwise. All chemical compounds were solubi- VEGF-signaling through the FAC is not well understood, as lized in dimethyl sulfoxide (DMSO) and further diluted into serum- most focus has been on targeting kinase activity of either Src free cell culture medium, reaching a final vehicle concentration or FAK. It was shown, however, that paxillin deletion caused of ,0.01% (v/v) DMSO. dysfunctional cell spreading and stunted migration, similar to REC Proliferation Assays. To evaluate paxillin-dependent mod- the phenotypes of cells without FAK (Eliceiri et al., 1999; ulation of retinal endothelial cell proliferation, 50,000 cells were Brown and Turner, 2004; Brown et al., 2005). To our knowledge, seeded into each well of a 96-well dish and allowed to adhere Targeting Src/FAK/Paxillin Signalsome in Neovascular Disease 3 overnight. RECs were serum-deprived for 1 hour in 0.1% bovine Immunoblot (Western) Analysis. Cellular proteins were ana- serum albumin, stimulated with VEGF (100 ng/ml), treated with lyzed by Western blotting after SDS-PAGE using human specific vehicle, kinase inhibitors, or test compounds and incubated for primary antibodies, as previously described (Toutounchian et al., 24 hours. Cellular proliferation was determined using the tetra- 2014). Whole REC lysates were collected in radioimmunoprecipitation zolium salt WST-1 according to the assay manufacturer’sinstruc- assay lysis buffer with protease/phosphatase inhibitor (1Â)cocktail tions (Quick Cell Proliferation Assay Kit II; Abcam, Cambridge, (Roche, Indianapolis, IN). Total protein was measured by BCA assay MA). Optical density as a measure of cellular proliferation was (Pierce/ThermoFisher Scientific) then processed in 4Â LDS loading buffer measured using a microplate reader at an absorbance of 450 nm. containing 2.5% 2-mercaptoethanol (Sigma-Aldrich), heated to 70°C for Data represent mean optical density (OD) 6 S.D., n 5 8 per group. 10 minutes, and loaded into NuPAGE 4–12% Bis-Tris Gels (Invitrogen/ In parallel to the 24-hour viability experiments, RECs were incu- ThermoFisher Scientific). Immunoblotting was performed with nitrocel- bated with calcein-AM for 30 minutes and imaged using the EVOS lulose membranes (Bio-Rad, Hercules, CA), blocked using Odyssey Blocking FL Cell Imaging System (ThermoFisher Scientific) to observe cell Buffer (LI-COR), and then incubated with specific primary antibodies numbers. overnight at 4°C. Analysis of phosphorylation is presented as a ratio Annexin V/Fluorescein Isothiocyanate Staining and Flow of phosphorylated protein to total protein (e.g., P-Y397 FAK/total FAK); Cytometry Analysis for Apoptosis. REC apoptosis was measured cellular lysates analyzed for both phosphorylated and nonphosphorylated by detection of phosphatidylserine translocation to the external protein were normalized to total cellular/housekeeping proteins, i.e., surface of the cell membrane (Fadok et al., 1992). Annexin V/ GAPDH or a-tubulin. Secondary antibodies (IRDye 800CW goat anti-rabbit

propidium iodide (PI) staining was performed according to manu- and IRDye 680LT goat anti-mouse; 1:12,500; LI-COR) were incubated in Downloaded from facturer’s instructions (BioLegend, San Diego, CA). Briefly, RECs the dark at room temperature for 45 minutes. Dual-channel infrared scan treated with either JP-153 or vehicle for 24 hours were trypsinized and quantitation of immunoblots were conducted using the Odyssey Sa and washed twice with ice-cold phosphate-buffered saline (PBS) infrared imaging system with Image Studio (Ver. 3.1.4; LI-COR). containing two-percent fetal bovine serum. Pelleted RECs were In Vitro Scratch-Wound Assay. REC migration was performed resuspended in Annexin V Binding Buffer at 5.0 Â 106 cells/ml and in accordance with methods previously described (Ghosh et al., 2013). incubated with fluorescein isothiocyanate–annexin V and PI stain- RECs (106 cells/well) were seeded to 12-well plates and cultured to ing solution (BioLegend) at room temperature for 15 minutes in the confluence. RECs were washed twice with 1Â PBS and prewarmed molpharm.aspetjournals.org dark. Cells were then resuspended in binding buffer and analyzed by serum-free Medium 131 (Invitrogen) was introduced to wells for 1 hour to fluorescence flow cytometry using the BD LSRII Flow Cytometry remove any residual effects of supplemented growth factors. Using a Analyzer (BD Biosciences). Data were statistically assessed using sterile 200-ml pipette tip, a straight scratch down the center of the well FlowJo analysis software (V10.0.6; Tree Star Inc., Ashland, OR). provided the baseline for the analysis and quantification of REC migration Apoptotic cells were defined as annexin V-positive and PI-negative, and proliferation over 24 hours. Wells were then washed one time with and necrotic cells are defined as annexin V-positive and PI-positive. PBS to remove any detached cells. Growth factor–supplemented medium Viable cells were considered annexin V and PI-negative. with or without JP-153 (0.10–10 mM) was added to each well, and plates at ASPET Journals on September 27, 2021

Fig. 1. VEGF-induced FA signaling in RECs. (A) Retinal endothelial cells were stimulated with VEGF (100 ng/ml) and cellular lysates were collected over four hours and focal adhesion protein activation was measured using Western blotting as described in Materials and Methods. Initially, VEGFR-2 is activated at Y1175 upon VEGF ligation which triggers FAK Y397 autophosphorylation (representative Western blots on the left, analysis of FAK pY397 levels on the right) (*P , 0.05, ***P , 0.001). Subsequently, Src-kinase binds to FAK and further activates the kinase-domain loop FAK Y576/577 and the focal adhesion targeting domain FAK Y925. (B) Src- dependent activation of FAK coincides with paxillin Y118 phosphorylation over 4 hours (**P , 0.01, ***P , 0.001). Data represent mean 6 S.D., n =4–8. 4 Toutounchian et al. were immediately imaged using a CoolSNAP charge-coupled device EVOS FL Cell Imaging System (ThermoFisher Scientific). A camera (Roper Technologies, Inc., Sarasota, FL) mounted on an Eclipse representative image from each group was used to depict extent of TE300 Inverted Microscope (Nikon, Melville, NY). Using 4Â magnifica- wound closure. tion and a computer-controlled stage, images at three specific coordinates Transwell Cellular Migration Assays. Cell migration was per well at the time of the initial wounding were obtained in Metamorph performed using Transwell polycarbonate membranes (Corning, software (Universal Imaging, West Chester, PA). Plates were returned to Corning, NY), as previously described (Cheranov et al., 2008). Briefly, incubator for 24 hours. The next day, previous coordinates were recalled cell-culture inserts containing membranes 6.5 mm in diameter and and images were again collected in Metamorph and then transferred to 8.0-mm pore size (Corning) were placed in a 24-well tissue culture Adobe Photoshop (CS5 Extended, Ver. 12.1; Adobe Systems, Inc., San plate (Corning). The upper surface of the porous membrane was coated Diego, CA). Using the magnetic lasso tool in Photoshop, the outline of with attachment factor at 37°C for 1 hour. Human RECs were serum- protruding/migrating cells from the periphery of the scratch toward the starved overnight in medium 131 containing 0.1% bovine serum center was measured. The area devoid of migrating cells was recorded albumin, trypsinized, pelleted, and resuspended in medium 131 with and quantified as a percentage change from the previous day’sarea vehicle (0.1% DMSO) or JP-153 at respective concentrations. RECs quantification: were then seeded into the upper chamber at 1 Â 105 cells/well. Medium 1 2   131 containing either vehicle or VEGF (100 ng/ml) / JP-153 was added to the lower chamber. After 24 hours of incubation at 37°C, % 5 2 A24 hours Area 1 (1) nonmigrated cells were removed from the upper side of the membrane A0 hours

with cotton swabs and the cells on the lower surface of the membrane Downloaded from Data represent mean percent wound closure 6 S.D. RECs from each were fixed in 4% paraformaldehyde for 15 minutes and washed twice group were fixed at 24 hours, stained with DAPI, and imaged using the with 1Â PBS. Nuclei were then stained with DAPI in PBS for five molpharm.aspetjournals.org at ASPET Journals on September 27, 2021

Fig. 2. Src-dependent activation of FAK and paxillin in RECs. (A) Src-inhibition with SU6656 (1 mM) inhibited VEGF’sactivation of FAK Y576/577, Y861, and Y925 and paxillin Y118 (* P , 0.05,††P , 0.01) but did not prevent autophosphorylation of FAK Y397 (P . 0.05). Data (n = 3) represent mean 6 S.D. (B) VEGF-mediated prolifera- tion of RECs was performed as described in Materials and Methods. VEGF-induced pro- liferation in RECs was reduced in the pres- ence of SU6656 (1 mM), which correlated with FA activation in panel A (***,†††P , 0.001). Data represent mean 6 S.D., n =8. Targeting Src/FAK/Paxillin Signalsome in Neovascular Disease 5 minutes and images were collected using the EVOS FL Cell Imaging were quantified from six random fields (n 5 3). Data represent mean System (ThermoFisher Scientific). Images were imported into Adobe number of migrating cells/field 6 S.D. Photoshop (Adobe Systems, Inc.) and cells were counted using batch Retinal Angiogenesis: Murine Oxygen-Induced Retinopathy image processing with automation. Briefly, the batches of images Model. C57BL/6N (Charles River Laboratories, Wilmington, MA) from all experimental groups were processed using color correction to mice were used in all experiments. All animal studies were performed enhance DAPI signal against background. Nuclei were outlined under the guidelines of the Association for Research in Vision and using the color-selection tool. The automation protocol was estab- Ophthalmology for the humane use of animals in vision research, lished on the basis of the first image processed in Photoshop to ensure and under the guidance and approval of the Institutional Animal that the processing of each subsequent image was done without any Care and Use Committee at the University of Tennessee Health biasing or manipulation of quality and/or integrity. Migrating RECs Science Center. Downloaded from molpharm.aspetjournals.org at ASPET Journals on September 27, 2021

Fig. 3. Discovery of JP-153 as a potent inhibitor of VEGF-induced proliferation. (A) REC proliferation was used to investigate compound 6-B345TTQ, a known paxillin disruptor, which was found to inhibit REC proliferation at concentrations greater than 10 mM(†P , 0.05, †††P , 0.001). Owing to potency issues, we redesigned a derivative, JP-153, that inhibits REC proliferation substantially in concentrations as low as 0.25 mM(†††P , 0.001). Data represent mean 6 S.D., n = 3. (B) We observed cell numbers using calcein-AM as described in Materials and Methods. (C) We investigated apoptosis using cleaved-PARP signaling in Western blots and showed that JP-153 (1 mM) did not significantly enhance apoptotic signaling (panel a, P = 0.239 versus 10% fetal bovine serum controls; data are presented as the mean 6 S.D.; n = 3). Flow cytometry quantified apoptotic cells within the population treated with JP-153 (1 mM, 24 hours to confirm that cell death was not induced with treatment, compared with controls; panel b, n = 50,000 cells). 6 Toutounchian et al.

Retinal angiogenesis was induced using a mouse model of with symbols *, **, ***, whereas significances ,0.05, 0.01, 0.001 oxygen-induced retinopathy (OIR), as previously described (Smith among treatment arms are represented with †, ††, †††,respectively. et al., 1994; Toutounchian et al., 2014). Five independent litters on three separate occasions were used for OIR experiments. Mouse pups exposed to the oxygen chamber were shuffled into three Results groups prior to dosing (P12) to provide intralitter controls. Exper- imental groups were as follows: 1) mice reared in normal atmo- Src/FAK-Paxillin Signaling Pathway in REC spheric conditions (negative-control; normoxia); 2) mice exposed to Proliferation. FAK and paxillin are coordinators of FA turn- OIR/hyperoxic chamber and treated with vehicle microemulsion over during VEGF-induced proliferation and migration—two (1 ml/g; positive-control); 3) OIR-mice treated with JP-153-loaded seminal events of angiogenesis (Brown et al., 2005). To microemulsion at 0.5 mg/kg; and 4) JP-153 at 5.0 mg/kg. Mouse confirm the relevance of these two players in VEGF-induced pups were exposed to 75% oxygen at postnatal day seven (P7) for proliferation of RECs, we stimulated RECs with VEGF and 5 days and then returned to normal oxygen (P12). Ocular micro- analyzed cell lysates for FAK and paxillin phosphorylation emulsion used for drug delivery comprised Capryol 90 (10.5% v/v), over time. Fig. 1A shows that rhVEGF (100 ng/ml) activates Triacetin (10.5% v/v), Tween-20 (24.5% v/v), and Transcutol P VEGF receptor-2 (VEGFR-2) with maximal phosphoryla- (24.5% v/v) (Gattefossé Pharmaceuticals, Saint-Priest, France) generated via homogenization and water titration methods, as tion occurring within 15 minutes at a major phosphorylation site, Tyr-1175. Activation of VEGFR-2 triggers autophos- previously described (Toutounchian et al., 2014). JP-153 was first Downloaded from loaded into the oil-phase and then incorporated into the final phorylation of FAK Y397 (as seen in Western blot images, microemulsion formulation and stored at room temperature away with analysis to the right; *P , 0.05, ***P , 0.001), which from light until dosing. OIR mice were weighed prior to receiving promotes association of Src with FAK (Schaller et al., 1994) each daily dose to both eyes using either JP-153 or vehicle-loaded and subsequently leads to Src-dependent FAK phosphoryla- microemulsion from P12 to P17 (vehicle control, N 5 8; JP-153 tion of its kinase domain loop, Y576/577 and focal adhesion 0.5 mg/kg, N 5 14; JP-153 5.0 mg/kg, N 5 14). On P17, retinas were targeting domain Y925 (Fig. 1A). Src-dependent activation removed, dissected, mounted, and stained for endothelial cells to and binding of FAK forms the Src/FAK focal adhesion molpharm.aspetjournals.org investigate retinal angiogenesis. At the conclusion of the study, complex (FAC), which phosphorylates paxillin Y118 (Fig. anesthetized animals were humanely euthanized according IACUC , , guidelines. 1B, **P 0.01, ***P 0.001). Retinal Whole-Mount Imaging and Analysis. Enucleated To determine if the Src/FAK complex is necessary for whole eyes from P17 mouse pupsunderwentimmediateweak paxillin activation in RECs and thus proliferation, we exam- fixation in 4% paraformaldehyde in PBS for 1 hour and washed ined FAK and paxillin phosphorylation in VEGF-stimulated three times in ice-cold PBS. Retinas were carefully isolated under a RECs treated with Src-kinase inhibitor SU6656 (1 mM) (Blake Leica S6E dissecting stereomicroscope (Leica Microsystems, Buf- et al., 2000). In Fig. 2A, we show that inhibiting Src kinase falo Grove, IL) and mounted onto microscope slides. Whole retinas reduces the phosphorylation of FAK Y576/577, Y925, and were incubated overnight at 4°C with isolectin B4-594 (Alexa Fluor Y861 (††P , 0.01) but does not affect autophosphorylation of at ASPET Journals on September 27, 2021 594; Molecular Probes, Eugene, OR), as previously described (Connor Y397. An inactive Src/FAK complex fails to phosphorylate et al., 2009; Toutounchian et al., 2014). Isolectin-stained retinas were †† paxillin Y118 (Fig. 2A, P , 0.01). We again treated RECs then washed three times in 1 PBS, sealed under coverslips using with SU6656 (1 mM) for 24 hours and showed that inhibition Vectashield mounting medium (Vector Laboratories, Inc.), and stored at 4°C until imaging. of Src-mediated phosphorylation of FA proteins leads to a Images were acquired using a Zeiss LSM 710 system attached to a Zeiss Axio Observer inverted microscope with Zen 2010 v.6.0 software (Carl Zeiss Microscopy, Peabody, MA). Multidimensional acquisition was carried out using Z-stacks with ,4-mmslicing intervalsandtile-scanautomationwithan8%tileoverlapata resolution of at least 512  512 pixels per tile and digitally stitched together. Quantification of avascular area (AV) and neovasculariza- tion (NV) in retinal whole mounts was performed in Adobe Photo- shop (Adobe Systems, Inc.), as previously described (Toutounchian et al., 2014). The area devoid of vascularization around the optic disc was characterized as percentage of total retinal area (%AV). Photo- shop color-range analysis tool were used to outline NV formations after intensity thresholds were set to exclude normal vasculature. Data were recorded as a percentage of total retinal area (%NV). Representative whole-mounted retinas were displayed using the exact quantified outlined areas and layered back into place onto the original whole-retina image. Using the linear light-blending method in Photoshop, both avascular and neovascular areas were trans- posed in white. Statistical Analyses. All data represented herein were per- formed in replicates of three or more and presented as the mean 6 S.D., unless otherwise indicated. Differences among groups were analyzed using one-way analysis of variance. When overall analysis revealed significance among groups, means were compared and Fig. 4. JP-153 inhibits VEGF-induced activation of paxillin Y118. (A) tested using Tukey’s posthoc analysis. Statistical significance was REC lysates were collected at 4 hours post-VEGF activation, and , phosphorylation of paxillin Y118 was measured using Western blotting. set at P 0.05. All statistical analyses were performed in SigmaPlot (B) JP-153 significantly reduced phosphorylation in cells stimulated with 12.0 software (Systat Software, Inc., San Jose, CA). P values VEGF (**,††P , 0.01) but did not affect constitutive/unstimulated levels representing significances of ,0.05, 0.01, and 0.001 are denoted (P = 0.749 versus vehicle control). Data represent mean 6 S.D.; n =3. Targeting Src/FAK/Paxillin Signalsome in Neovascular Disease 7 significant decrease in VEGF-induced proliferation (Fig. 2B, panel b, †††P , 0.001). JP-153 and 6-B345TTQ structures, ††† P , 0.001). IC50,andcalculatedLogP values depict JP-153 as more Discovery of JP-153 as a Potent Inhibitor of VEGF- pharmaceutically favorable (Supplemental Fig. 1) (Lipinski Induced Proliferation. Src-dependent FAK and paxillin et al., 2001). We used calcein-AM staining (Fig. 3B) to show phosphorylation correlated with VEGF-induced prolifera- that live cell number is reduced with JP-153 treatments tion in RECs (Fig. 2B). We used this phenotypic response in addition to reduced proliferative activity, as measured to derive compounds related to a known paxillin protein by WST-1 in Fig. 3A. Yet, JP-153 does not promote apopto- disruptor, 6-B345TTQ (Kummer et al., 2010). Our initial sis in cells, as characterized by PARP cleavage (Fig. 3C, lead identification efforts yielded the analog JP-153, which panel a, *P , 0.05 versus serum-free controls) and annexin was ∼50 times more potent than 6-B345TTQ in REC pro- V/PI staining at 1-mM concentration over 24 hours (Fig. 3C, liferation assays (Fig. 3A, panel a, †P , 0.05, †††P , 0.001; panel b). Downloaded from molpharm.aspetjournals.org at ASPET Journals on September 27, 2021

Fig. 5. JP-153 acts by reducing effector signaling through Src/FAK/paxillin FA complex to inhibit VEGF-induced proliferation. A) Western blot images (left) and respective analyses (right, panels a-f) of RECs activated by VEGF (100 ng/mL for 15 minutes) show FA and effector signaling after one hour pre- treatments with JP-153 (1mM), Src-inhibitor SU6656 (1 mM) or PI3K inhibitor LY294002 (10 mM). JP-153 and SU6656 significantly reduce levels of VEGF-induced paxillin Y118 phosphorylation (panel a; **, ††P , 0.01), but only SU6656 inhibits FAK phosphorylation at Y576/577 (panel d; *, †P , 0.05), Y861 (panel e; *, †P , 0.05), and Y925 (panel f; *P , 0.05, ††P , 0.01), in agreement with earlier experiments shown in Figure 3. VEGF-induced pAKT (S473) phosphorylation was inhibited by JP-153, SU6656 and LY294002 (panel b; **, ††P , 0.01,†††P , 0.001). Neither SU6656 nor JP-153 causes any significant change to VEGF-induced pERK 1/2 activation (panel c; P . 0.05), while LY294002 caused an increase in activation of ERK (†P , 0.01 vs. VEGF controls). The dividing lines in the Western blot panel convey where samples from the same blot were shifted over to the left by one lane for data presentation consistency. B) We confirmed Akt-dependent REC-proliferation by treating cells with LY294002 which resulted in the potent inhibitionof proliferation in a more pronounced manner than JP-153 or SU6656 (***, †††P , 0.001, n = 8). Data represent mean 6 SD. 8 Toutounchian et al.

Effector Signaling through an Activated Src/FAK- VEGF-induced REC migration was significantly inhibited in Paxillin Signaling Complex during VEGF-Induced Pro- JP-153 treatments over a range of concentrations (0.10–10 liferation Is Akt-Dependent. We postulated that JP-153 mM)(Fig.6;*,†P , 0.05, ††† P , 0.001). Next, we validated inhibits REC proliferation through disruptions in FA pro- our scratch-wound results with the Transwell migration/ tein interactions, as shown by Kummer et al. (2010) with invasion assay with VEGF as the chemotactic inducer (Yoshida 6-B345TTQ. Disrupting Src/FAK binding to paxillin results et al., 1996). Our results show that JP-153 inhibits REC inva- in decreased activation of paxillin Y118 (Richardson et al., sion at submicromolar concentrations (0.10–0.50 mM) (Fig. 7, 1997). Thus, we treated RECs with JP-153 (1 mM) for 1 hour ***,†††P , 0.001). and then stimulated them with VEGF for 4 hours. In cells Signal Disruption of Src/FAK/Paxillin Complex by JP-153 significantly reduces Y118 phosphorylation (Fig. 4, JP-153 In Vivo Inhibits Retinal Neovascularization in **,††P , 0.01) but did not inhibit constitutive levels of the Murine Oxygen-Induced Retinopathy Model. Our unstimulated RECs treated with JP-153 (P 5 0.749). in vitro mechanism of action studies in RECs suggested Next, we examined downstream FA effector signaling that JP-153 inhibited proliferation and migration by dis- during early VEGF activation at 15 minutes. We pretreated rupting Src/FAK/paxillin signaling pathway. Therefore, we RECs with JP-153 for 1 hour prior to VEGF-activation and hypothesized that JP-153 could inhibit retinal angiogenesis measured phosphorylation of FAK phosphorylation sites, in vivo by reducing Src/FAK/paxillin activity. We used the as well as downstream angiogenic markers AKT and ERK. murine OIR model of retinal neovascularization (RNV) to Downloaded from Our results again confirmed that JP-153 reduces activation test JP-153 at low and high topical doses applied daily to of paxillin Y118 compared with VEGF controls (Fig. 5A, panel a; each eye during the hypoxic period (P17 retinal whole- *, †P , 0.05) but does not change autophosphorylation of FAK mountsinFig.8A,andsubsequentanalysisinFig.8B).Our Y397; these results mimic the activity of SU6656 (††P , 0.01). data shows that JP-153 inhibits neovascularization by However, when we probed for FAK Y576/577, Y861, and Y925 40 and 45% in a dose-dependent manner (0.5 and 5 mg/kg, in cells, JP-153 did not affect levels of Src-dependent FAK respectively), compared with vehicle-treated eyes (panels molpharm.aspetjournals.org phosphorylation sites (Fig. 5A, panels d–f; P . 0.05), whereas a–c, ***P , 0.001). However, only JP-153 at the higher dose SU6656 inhibited these levels strongly (†P , 0.05, ††P , 0.01). To rule out kinase inhibition, we show that JP-153 was not a direct kinase inhibitor of FA signaling effectors per se, as measured by the Z9-LYTE SelectScreen Single Point bio- chemical assay (ThermoFisher Scientific) (Supplemental Table 2). Src-mediated activation of paxillin Y118 primes the com- plex for recruitment to focal contacts, where interactions with at ASPET Journals on September 27, 2021 PI3K and MEK activate their respective downstream sub- strates, AKT and ERK, to promote cytoskeletal rearrange- ments during proliferation and migration (Fujikawa et al., 1999; Akagi et al., 2002; Du et al., 2011). Thus, we compared RECs treated with JP-153 and SU6656 with those treated with PI3K inhibitor LY294002 (10 mM). Both p-Akt (Ser473) and p-ERK 1/2 levels rose under VEGF, but only Akt was effectively blocked by SU6656 and JP-153 (Fig. 5A, panels b and c; *,†P , 0.05, ††P , 0.01), since neither show significant inhibition of p-ERK 1/2 at concentrations tested (P . 0.05). However, complete inhibition of Akt phosphorylation by LY294002 caused no reductions in FAK or paxillin activa- tion, suggesting the Src/FAK/paxillin activation cascade precedes PI3K-induced Akt phosphorylation. However, unlike JP-153 or SU6656, LY294002 significantly induced ERK activation (†P , 0.05; LY294002 versus VEGF). To validate an Akt-dependent proliferation pathway, cells treated with LY294002 potently inhibited proliferation, with levels far exceeding serum starvation, Src-inhibition, and JP-153 treatments (Fig. 5B, ***,†††P , 0.001). Together, these data suggest JP-153 acts to inhibit REC proliferation through an Akt-dependent but ERK- independent mechanism. Paxillin Modulation with JP-153 Inhibits VEGF-Induced Migration of Retinal Endothelial Cells. We have shown Fig. 6. JP-153 inhibited VEGF-induced REC migration in the scratch- that JP-153 inhibited REC proliferation through disruptions wound assay. The scratch-wound migration assay was performed in RECs in Src/FAK activation of paxillin Y118 and pAkt (Fig. 5). Since exposed to VEGF for 24 hours, as described in Materials and Methods.(A) angiogenesis requires two distinct but cooperative mecha- Data analysis show JP-153 inhibits VEGF-induced migration in a concentration-dependent manner. (B) Representative DAPI-stained im- ’ † nisms, proliferation and migration, we examined JP-153 s ages after 24 hours. Data are presented as the mean 6 S.D. (n =6;*,P , effect on migration using the standard scratch wound assay. 0.05, †††P , 0.001). Targeting Src/FAK/Paxillin Signalsome in Neovascular Disease 9

Fig. 7. JP-153 inhibited VEGF-induced REC invasion using the Transwell migration assay. RECs were seeded onto porous membranes and chemotactic factor VEGF was used to stimulate REC migration, as described in Materials and Methods. (A) Results show that JP-153 inhibited REC invasion in a concentration-dependent manner (data are mean 6 S.D.; ***,†††P , 0.001; n =6). Downloaded from (B) Cells traversing the membrane were fixed and stained with DAPI, and representative images of each group are shown (image labels A–E: serum-free, VEGF, V + 0.10, V + 0.25, V + 0.50 mM, respectively). molpharm.aspetjournals.org at ASPET Journals on September 27, 2021 enhanced the AV compared with vehicle (panels d and A major finding in this study was that in human primary e, ***P , 0.001). Mouse pups kept outside the OIR chamber RECs, Src/FAK activation of paxillin directs VEGF-induced for the entire study were also dosed with JP-153 (5 mg/kg) signaling during REC proliferation and migration, a signaling under identical age-based regimens to evaluate any impact pathway well characterized in cancer cells and other trans- on retinal vascular development. There were no obvious formed cell lines but previously undescribed in primary differences between vehicle and JP-153-treated retinas in human RECs (Abedi and Zachary, 1997; Birukova et al., mice not exposed to the OIR chamber (Supplemental Fig. 2). 2009; Yang et al., 2015). We hypothesized that targeting These findings suggest that JP-153 can act to regulate REC Src/FAK or paxillin would limit the activation of pathologic RNV without affecting normal retinal blood downstream effector proteins important for retinal angio- vessel growth or vasculogenesis. genesis. First, we showed VEGF induces activation of Src kinase leading to the phosphorylation of FAK and paxillin, which could be prevented by pharmacological inhibition of Discussion Src. We then used a small-molecule probe of paxillin binding interactions, 6-B345TTQ (Kummer et al., 2010), to investi- In previous work, paxillin Y118 activation in high-dose radi- gate paxillin’s role during VEGF-induced REC proliferation. ation injury was an important signaling component driving REC Blocking interactions that involve paxillin effectively re- proliferation in ischemic retinopathy (Toutounchian et al., 2014). duced REC proliferation in vitro, but owing to inherently low We demonstrated in this study that VEGF-dependent activation potency and solubility, we derived a more effective deriva- of the Src/FAK/paxillin signaling complex, or signalsome, drives tive, JP-153. REC migration and proliferation (Fig. 9). Moreover, we showed An unexpected and novel finding during in vitro mech- that modulation of the Src/FAK/paxillin signaling complex with anistic studies was that JP-153 reduced phosphorylation small molecule JP-153 reduced paxillin Y118 activation and of paxillin Y118, a critical tyrosine activation site, but did inhibited migration and proliferation of RECs; and that this not affect FAK phosphorylation, distinguishing JP-153’s effect did not interfere with VEGF-dependent activation of either activity from Src inhibitor SU6656. Thus, we have shown Src or FAK. Furthermore, topical application of a JP-153-loaded that paxillin Y118 is an important downstream biomarker ocular microemulsion inhibited hallmark features of pathologic for VEGF-induced REC proliferation. Additionally, JP-153 retinal angiogenesis in mice; both neovascular tuft formation did not inhibit the kinase activities of Src or FAK (Supple- and vascular regrowth in the murine OIR model were reduced in mental Table 2); strongly suggesting that JP-153’santipro- a dose-dependent manner. liferative phenotype in RECs is through paxillin-dependent 10 Toutounchian et al.

Fig. 8. JP-153 inhibited retinal angiogenesis Downloaded from in the murine oxygen-induced retinopathy model. P17 retinal whole-mounts were stained for endothelial cells using isolectin B4-594 as described in Materials and Methods. Mice were dosed daily from P12-17 using either topical microemulsion-loaded vehicle, 0.5 mg/kg, or 5.0 mg/kg JP-153. (A) Representative im-

ages of retinal whole-mounts depicting: neo- molpharm.aspetjournals.org vascular area (a–c) and AV (d–f). (B) Data analysis of retinal vasculature revealed that JP-153 inhibited NV and increased AV in a dose-dependent manner. Data represent mean 6 S.D.; ***P , 0.001; N =8–14/group. at ASPET Journals on September 27, 2021

signaling, independent of kinases that may regulate its phos- other studies that established the important stepwise role phorylation. In fact, mutagenesis of FAK- or paxillin-binding of the Src/FAK complex as a crucial activator of the PI3K- domains are known to inhibit their interaction and prevent Akt pathway (Thakker et al., 1999; Bullard et al., 2003; activation of paxillin and other downstream proteins (Subauste Thamilselvan et al., 2007). Therefore, our results show et al., 2004; Kadaré et al., 2015). that paxillin is an important signaling intermediary that Activation of Src/FAK drives proliferation and migration connects the activated Src/FAK complex and Akt in through intermediates ERK and Akt (Yan et al., 2008; angiogenesis. Pylayeva et al., 2009). Our data show that PI3K-inhibitor The uncoupling of an active Src/FAK complex from paxillin LY294002 remained unchanged, although effective at pre- suggested it is a key regulator of pathologic FA signal trans- venting both Akt phosphorylation and REC proliferation, duction and potentially represents a novel in vivo target levels of FAK, or paxillin phosphorylation. SU6656 and distinct from anti-VEGF therapies aimed at silencing JP-153 both caused reductions in Akt phosphorylation, receptor-mediated kinase signaling. Studies using tar- suggesting that activation of FAK and paxillin precedes geted deletions of FA proteins FAK and Src in the mouse VEGF-induced activation of Akt in RECs. However, since retina disrupt the progression of RNV (Kornberg et al., JP-153 did not disrupt FAK phosphorylation levels and still 2004; Werdich and Penn, 2006); these findings correspond reduced p-Akt, we concluded that paxillin Y118 plays a with our in vitro results using the Src inhibitor SU6656, crucial role in coordinating events that drive Akt-dependent which affects all downstream binding and activation angiogenesis in RECs. These results are in agreement with partners. We show similar in vitro effects with JP-153 on Targeting Src/FAK/Paxillin Signalsome in Neovascular Disease 11

can potently inhibit vascular growth and retinal function. These deficits were a result of significant structural changes to the retinal layers, despite their prevention of classic neo- vascular pathology (Tokunaga et al., 2014). These findings raise concerns as to whether enhancing the AV, or preventing revascularization with anti-VEGF treatment, may exacerbate ischemic injury in neuroretinal tissues (Bautch and James, 2009). We used the same dosing regimen of JP-153 in mice reared in atmospheric conditions (room air) and found that even high-dose treatments did not affect normal vasculo- genesis, as there were no obvious defects in “normal” vessel growth patterns (Supplemental Fig. 2). Our findings point to an important difference between anti-VEGF therapies and JP-153 with respect to dose effect on vasculogenesis, findings that suggest that JP-153 might help to avoid adverse effects associated with anti-VEGF monotherapy in patients long- term by sparing normal physiologic homeostasis and neuro- Downloaded from retinal function. In conclusion, our results detail an effective strategy to treat pathologic RNV using the small molecule JP-153. Aberrations in FA protein signaling underlie many aggres- sive hyperproliferative diseases, including cancer metastasis and polycystic kidney disease, making the Src/FAK/paxillin molpharm.aspetjournals.org signalsome an attractive therapeutic target (Ischenko et al., 2007; Sweeney et al., 2008; Lee et al., 2015). Recently, small- Fig. 9. Summary diagram of JP-153’s proposed target of action. Data molecule kinase inhibitors of paxillin binding partners, suggests that JP-153 targets the interaction between an active Src/FAK Src and FAK have advanced to late-stage clinical trials in signaling complex and paxillin. Inhibiting this interaction resulted in humans, which suggests FA signal transduction can be decreased paxillin activation (Y118), preventing activation of down- effectively and safely modulated in humans (Sulzmaier stream effector protein Akt. This effect translated into potent inhibition of REC proliferation and migration,invitro,andinhibitionofRNV, et al., 2014; Taylor et al., 2015). Paxillin, however, has in vivo. never been successfully targeted by pharmacologic inter-

vention for the treatment of any proliferative disease, even at ASPET Journals on September 27, 2021 though its expression has been correlated with highly proliferation as with SU6656, specifically with decreased invasive cancers (Jagadeeswaran et al., 2008). Moreover, paxillin Y118 phosphorylation and inhibition of p-Akt the ability of paxillin to function as a scaffold that binds downstream, resulting in potent inhibition of movement multiple FA proteins makes it an interesting target for and growth. From these studies, we can assert that the development of novel inhibitors of pathologic neovasculari- activation of the FAC may be a crucial component in the zation. Since adaptive resistance is a major obstacle plagu- regulation of pathologic retinal angiogenesis, in vivo. We ing the efficacy of current antiangiogenic treatments (Bergers tested this hypothesis by administering JP-153 topically in and Hanahan, 2008), the novelty of this current study can the OIR model, which resulted in significantly reduced be characterized by two major findings: 1) paxillin is an im- retinal angiogenesis, as measured by both neovasculariza- portant and viable target in pathologic retinal angiogenesis; tion and the AV. Intriguingly, we found that only the higher and 2) JP-153 effectively modulates paxillin-dependent sig- doses of JP-153 were able to significantly enhance AV, naling in vitro and in vivo to treat RNV. Thus, the target suggesting perhaps that our small molecule affects patho- and mechanism of JP-153 has extensive applicability across a logic neovascularization more than vasculogenesis. How- wide range of proliferative indications and warrants further ever, since genetic knockdown of paxillin in mice leads to pharmaceutical development and refinement as a novel early embryonic lethality (Hagel et al., 2002), paxillin has therapeutic. been conditionally silenced in the developing mouse retina. These studies actually showed that paxillin knockdown Acknowledgments induced migration and endothelial cell sprouting during The authors thank Drs. Bilal Aleiwi and Shivaputra Patil for help development (German et al., 2014). Thus, knocking down with the synthetic chemistry of JP-153 and the University of paxillin may not be a strategy as clear as one would expect, Tennessee College of Pharmacy and the University of Tennessee since the coordination of FAs, and thus angiogenesis, may Research Foundation for financial support. rely on differential or contextual interactions and/or phos- phorylation patterns (Birukova et al., 2009). We are cur- Authorship Contributions rently investigating the effects of JP-153 on paxillin with Participated in research design: Toutounchian, Miller, Yates. respect to its critical binding partners and how these Conducted experiments: Toutounchian, Pagadala. interactions trigger differential phosphorylation that pro- Contributed new reagents or analytic tools: Yates, Miller. mote FA signaling during angiogenesis. Performed data analysis: Toutounchian, Park, Chaum, Yates. VEGF participates in both pathologic and physiologic Wrote or contributed to the writing of the manuscript: Toutounchian, growth. Thus, it is not surprising that anti-VEGF therapeutics Park, Baudry, Chaum, Yates. 12 Toutounchian et al.

References Kornberg LJ, Shaw LC, Spoerri PE, Caballero S, and Grant MB (2004) Focal adhe- sion kinase overexpression induces enhanced pathological retinal angiogenesis. Abedi H and Zachary I (1997) Vascular endothelial growth factor stimulates tyrosine Invest Ophthalmol Vis Sci 45:4463–4469. phosphorylation and recruitment to new focal adhesions of focal adhesion kinase and paxillin in endothelial cells. J Biol Chem 272:15442–15451. Kummer C, Petrich BG, Rose DM, and Ginsberg MH (2010) A small molecule that inhibits the interaction of paxillin and alpha 4 inhibits accu- Aiello LP, Avery RL, Arrigg PG, Keyt BA, Jampel HD, Shah ST, Pasquale LR, mulation of mononuclear leukocytes at a site of inflammation. JBiolChem Thieme H, Iwamoto MA, Park JE, et al. (1994) Vascular endothelial growth factor 285:9462–9469. in ocular fluid of patients with diabetic retinopathy and other retinal disorders. N Engl J Med 331:1480–1487. Lee BY, Timpson P, Horvath LG, and Daly RJ (2015) FAK signaling in human cancer as a target for therapeutics. Pharmacol Ther 146:132–149. Akagi T, Murata K, Shishido T, and Hanafusa H (2002) v-Crk activates the phos- phoinositide 3-kinase/AKT pathway by utilizing focal adhesion kinase and H-Ras. Lipinski CA, Lombardo F, Dominy BW, and Feeney PJ (2001) Experimental and Mol Cell Biol 22:7015–7023. computational approaches to estimate solubility and permeability in drug discov- ery and development settings. Adv Drug Deliv Rev 46:3–26. Arold ST, Hoellerer MK, and Noble ME (2002) The structural basis of localization and signaling by the focal adhesion targeting domain. Structure 10:319–327. Marneros AG, Fan J, Yokoyama Y, Gerber HP, Ferrara N, Crouch RK, and Olsen BR (2005) Vascular endothelial growth factor expression in the retinal pigment epi- Bautch VL and James JM (2009) Neurovascular development: The beginning of a thelium is essential for choriocapillaris development and visual function. Am J beautiful friendship. Cell Adhes Migr 3:199–204. Pathol 167:1451–1459. Bergers G and Hanahan D (2008) Modes of resistance to anti-angiogenic therapy. Nat Nowak JZ (2006) Age-related macular degeneration (AMD): pathogenesis and ther- Rev Cancer 8:592–603. apy. Pharmacol Rep 58: 353–363. Birukova AA, Cokic I, Moldobaeva N, and Birukov KG (2009) Paxillin is involved in Osborne NN, Casson RJ, Wood JP, Chidlow G, Graham M, and Melena J (2004) the differential regulation of endothelial barrier by HGF and VEGF. Am J Respir Retinal ischemia: mechanisms of damage and potential therapeutic strategies. Cell Mol Biol 40:99–107. Prog Retin Eye Res 23:91–147. Blake RA, Broome MA, Liu X, Wu J, Gishizky M, Sun L, and Courtneidge SA (2000) SU6656, a selective src family kinase inhibitor, used to probe growth factor sig- Ozaki H, Yu AY, Della N, Ozaki K, Luna JD, Yamada H, Hackett SF, Okamoto N, – Zack DJ, Semenza GL, et al. (1999) Hypoxia inducible factor-1alpha is increased in naling. Mol Cell Biol 20:9018 9027. Downloaded from ischemic retina: temporal and spatial correlation with VEGF expression. Invest Boozalis GT, Schachat AP, and Green WR (1987) Subretinal neovascularization from – the retina in radiation retinopathy. Retina 7:156–161. Ophthalmol Vis Sci 40:182 189. Pascolini D and Mariotti SP (2012) Global estimates of visual impairment: 2010. Br J Brown MC, Cary LA, Jamieson JS, Cooper JA, and Turner CE (2005) Src and FAK – kinases cooperate to phosphorylate paxillin kinase linker, stimulate its focal ad- Ophthalmol 96:614 618. hesion localization, and regulate cell spreading and protrusiveness. Mol Biol Cell Provenzano PP and Keely PJ (2011) Mechanical signaling through the cytoskeleton 16:4316–4328. regulates cell proliferation by coordinated focal adhesion and Rho GTPase sig- – Brown MC, Curtis MS, and Turner CE (1998) Paxillin LD motifs may define a new naling. J Cell Sci 124:1195 1205. family of protein recognition domains. Nat Struct Biol 5:677–678. Pylayeva Y, Gillen KM, Gerald W, Beggs HE, Reichardt LF, and Giancotti FG (2009) Ras- and PI3K-dependent breast tumorigenesis in mice and humans requires focal

Brown MC and Turner CE (2004) Paxillin: adapting to change. Physiol Rev 84: molpharm.aspetjournals.org – 1315–1339. adhesion kinase signaling. J Clin Invest 119:252 266. Bullard LE, Qi X, and Penn JS (2003) Role for extracellular signal-responsive Richardson A, Malik RK, Hildebrand JD, and Parsons JT (1997) Inhibition of kinase-1 and -2 in retinal angiogenesis. Invest Ophthalmol Vis Sci 44: cell spreading by expression of the C-terminal domain of focal adhesion 1722–1731. kinase (FAK) is rescued by coexpression of Src or catalytically inactive – Cheranov SY, Karpurapu M, Wang D, Zhang B, Venema RC, and Rao GN (2008) An FAK: a role for paxillin tyrosine phosphorylation. MolCellBiol17:6906 essential role for SRC-activated STAT-3 in 14,15-EET-induced VEGF expression 6914. and angiogenesis. Blood 111:5581–5591. Saint-Geniez M, Maharaj AS, Walshe TE, Tucker BA, Sekiyama E, Kurihara T, Connor KM, Krah NM, Dennison RJ, Aderman CM, Chen J, Guerin KI, Sapieha P, Darland DC, Young MJ, and D’Amore PA (2008) Endogenous VEGF is required for Stahl A, Willett KL, and Smith LE (2009) Quantification of oxygen-induced reti- visual function: evidence for a survival role on müller cells and photoreceptors. nopathy in the mouse: a model of vessel loss, vessel regrowth and pathological PLoS One 3:e3554. angiogenesis. Nat Protoc 4:1565–1573. Schaller MD, Hildebrand JD, Shannon JD, Fox JW, Vines RR, and Parsons JT (1994) Du J, Xu R, Hu Z, Tian Y, Zhu Y, Gu L, and Zhou L (2011) PI3K and ERK-induced Autophosphorylation of the focal adhesion kinase, pp125FAK, directs SH2-dependent Rac1 activation mediates hypoxia-induced HIF-1a expression in MCF-7 breast binding of pp60src. Mol Cell Biol 14:1680–1688.

cancer cells. PLoS One 6:e25213. Shen CJ, Raghavan S, Xu Z, Baranski JD, Yu X, Wozniak MA, Miller JS, Gupta M, at ASPET Journals on September 27, 2021 Eke I and Cordes N (2015) Focal adhesion signaling and therapy resistance in cancer. Buckbinder L, and Chen CS (2011) Decreased cell adhesion promotes angiogenesis Semin Cancer Biol 31:65–75. in a Pyk2-dependent manner. Exp Cell Res 317:1860–1871. Eliceiri BP, Paul R, Schwartzberg PL, Hood JD, Leng J, and Cheresh DA (1999) Smith LE, Wesolowski E, McLellan A, Kostyk SK, D’Amato R, Sullivan R, Selective requirement for Src kinases during VEGF-induced angiogenesis and and D’Amore PA (1994) Oxygen-induced retinopathy in the mouse. Invest Oph- vascular permeability. Mol Cell 4:915–924. thalmol Vis Sci 35:101–111. Fadok VA, Voelker DR, Campbell PA, Cohen JJ, Bratton DL, and Henson PM Subauste MC, Pertz O, Adamson ED, Turner CE, Junger S, and Hahn KM (2004) (1992) Exposure of phosphatidylserine on the surface of apoptotic lymphocytes Vinculin modulation of paxillin-FAK interactions regulates ERK to control survival triggers specific recognition and removal by macrophages. J Immunol 148: and motility. J Cell Biol 165:371–381. 2207–2216. Sulzmaier FJ, Jean C, and Schlaepfer DD (2014) FAK in cancer: mechanistic findings Ferris FL 3rd, Fine SL and Hyman L (1984) Age-related macular degeneration and and clinical applications. Nat Rev Cancer 14:598–610. blindness due to neovascular maculopathy. Arch Ophthal 102:1640–1642. Sweeney WE, Jr, von Vigier RO, Frost P, and Avner ED (2008) Src inhibition ame- Finger PT (2008) Radiation retinopathy is treatable with anti–vascular endo- liorates polycystic kidney disease. J Am Soc Nephrol 19:1331–1341. thelial growth factor bevacizumab (Avastin). Intl J Rad Oncol Biol Physics 70: Taylor JW, Dietrich J, Gerstner ER, Norden AD, Rinne ML, Cahill DP, Stemmer- 974–977. Rachamimov A, Wen PY, Betensky RA, Giorgio DH, et al. (2015) Phase 2 study of Fujikawa K, de Aos Scherpenseel I, Jain SK, Presman E, Christensen RA, bosutinib, a Src inhibitor, in adults with recurrent glioblastoma. J Neurooncol 121: and Varticovski L (1999) Role of PI 3-kinase in angiopoietin-1-mediated mi- 557–563. gration and attachment-dependent survival of endothelial cells. Exp Cell Res Thakker GD, Hajjar DP, Muller WA, and Rosengart TK (1999) The role of phos- 253:663–672. phatidylinositol 3-kinase in vascular endothelial growth factor signaling. J Biol German AE, Mammoto T, Jiang E, Ingber DE, and Mammoto A (2014) Paxillin Chem 274:10002–10007. controls endothelial cell migration and tumor angiogenesis by altering neuropilin Thamilselvan V, Craig DH, and Basson MD (2007) FAK association with multiple 2 expression. J Cell Sci 127:1672–1683. signal proteins mediates pressure-induced colon cancer cell adhesion via a Src- Ghosh MC, Gorantla V, Makena PS, Luellen C, Sinclair SE, Schwingshackl A, dependent PI3K/Akt pathway. FASEB J 21:1730–1741. and Waters CM (2013) Insulin-like growth factor-I stimulates differentiation of Tokunaga CC, Mitton KP, Dailey W, Massoll C, Roumayah K, Guzman E, Tarabishy ATII cells to ATI-like cells through activation of Wnt5a. Am J Physiol Lung Cell N, Cheng M, and Drenser KA (2014) Effects of anti-VEGF treatment on the re- Mol Physiol 305:L222–L228. covery of the developing retina following oxygen-induced retinopathy. Invest Hagel M, George EL, Kim A, Tamimi R, Opitz SL, Turner CE, Imamoto A, Ophthalmol Vis Sci 55:1884–1892. and Thomas SM (2002) The adaptor protein paxillin is essential for normal de- Toutounchian JJ, Steinle JJ, Makena PS, Waters CM, Wilson MW, Haik BG, Miller velopment in the mouse and is a critical transducer of fibronectin signaling. Mol DD, and Yates CR (2014) Modulation of radiation injury response in retinal en- Cell Biol 22:901–915. dothelial cells by quinic acid derivative KZ-41 involves p38 MAPK. PLoS One 9: Houlden CE, Lloyd-Jones GC, and Booker-Milburn KI (2010) Facile double-lithiation e100210. of a transient urea: vicarious ortho-metalation of aniline derivatives. Org Lett 12: Turner CE (2000) Paxillin interactions. J Cell Sci 113:4139–4140. 3090–3092. van Beijnum JR, Nowak-Sliwinska P, Huijbers EJ, Thijssen VL, and Griffioen AW Ischenko I, Guba M, Yezhelyev M, Papyan A, Schmid G, Green T, Fennell M, (2015) The great escape; the hallmarks of resistance to antiangiogenic therapy. Jauch KW, and Bruns CJ (2007) Effect of Src kinase inhibition on metasta- Pharmacol Rev 67:441–461. sis and tumor angiogenesis in human pancreatic cancer. Angiogenesis 10: Vanarotti MS, Miller DJ, Guibao CD, Nourse A, and Zheng JJ (2014) Structural and 167–182. mechanistic insights into the interaction between Pyk2 and paxillin LD motifs. J Jagadeeswaran R, Surawska H, Krishnaswamy S, Janamanchi V, Mackinnon AC, Mol Biol 426:3985–4001. Seiwert TY, Loganathan S, Kanteti R, Reichman T, Nallasura V, et al. (2008) Waltenberger J, Claesson-Welsh L, Siegbahn A, Shibuya M, and Heldin CH (1994) Paxillin is a target for somatic mutations in lung cancer: implications for cell Different signal transduction properties of KDR and Flt1, two receptors for vas- growth and invasion. Cancer Res 68:132–142. cular endothelial growth factor. J Biol Chem 269:26988–26995. Kadaré G, Gervasi N, Brami-Cherrier K, Blockus H, El Messari S, Arold ST, Wary KK, Kohler EE, and Chatterjee I (2012) Focal adhesion kinase regulation of and Girault JA (2015) Conformational dynamics of the focal adhesion targeting neovascularization. Microvasc Res 83:64–70. domain control specific functions of focal adhesion kinase in cells. J Biol Chem 290: WeisSM,LimST,Lutu-FugaKM,BarnesLA,ChenXL,GöthertJR,ShenTL, 478–491. Guan JL, Schlaepfer DD, and Cheresh DA (2008) Compensatory role for Pyk2 Targeting Src/FAK/Paxillin Signalsome in Neovascular Disease 13

during angiogenesis in adult mice lacking endothelial cell FAK. JCellBiol Yang WJ, Yang YN, Cao J, Man ZH, Li Y, and Xing YQ (2015) Paxillin regulates 181:43–50. vascular endothelial growth factor A-induced in vitro angiogenesis of human um- Werdich XQ and Penn JS (2006) Specific involvement of SRC family kinase activation bilical vein endothelial cells. Mol Med Rep 11:1784–1792. in the pathogenesis of retinal neovascularization. Invest Ophthalmol Vis Sci 47: Yoshida A, Anand-Apte B, and Zetter BR (1996) Differential endothelial migration 5047–5056. and proliferation to basic fibroblast growth factor and vascular endothelial growth Wilkinson-Berka JL, Rana I, Armani R, and Agrotis A (2013) Reactive oxygen spe- factor. Growth Factors 13:57–64. cies, Nox and angiotensin II in angiogenesis: implications for retinopathy. Clin Sci (Lond) 124:597–615. Yan W, Bentley B, and Shao R (2008) Distinct angiogenic mediators are required for Address correspondence to: Dr. Charles R. Yates, The University of Tennessee, basic fibroblast growth factor- and vascular endothelial growth factor-induced Memphis, Department of Pharmaceutical Sciences, 881 Madison Avenue, Phar- angiogenesis: the role of cytoplasmic c-Abl in tumor angiogenesis. macy Building Room 446, Memphis, TN 38163. E-mail: [email protected] Mol Biol Cell 19:2278–2288. Downloaded from molpharm.aspetjournals.org at ASPET Journals on September 27, 2021