Advertisement

Angiogenic and Immunologic Proteins Identified by Deep Proteomic Profiling of Human Retinal and Choroidal Vascular Endothelial Cells: Potential Targets for New Biologic Drugs

Published:March 17, 2018DOI:https://doi.org/10.1016/j.ajo.2018.03.020

      Purpose

      Diseases that involve retinal or choroidal vascular endothelial cells are leading causes of vision loss: age-related macular degeneration, retinal ischemic vasculopathies, and noninfectious posterior uveitis. Proteins differentially expressed by these endothelial cell populations are potential drug targets. We used deep proteomic profiling to define the molecular phenotype of human retinal and choroidal endothelial cells at the protein level.

      Methods

      Retinal and choroidal vascular endothelial cells were separately isolated from 5 human eye pairs by selection on CD31. Total protein was extracted and digested, and peptide fractions were analyzed by reverse-phase liquid chromatography tandem mass spectrometry. Peptide sequences were assigned to fragment ion spectra, and proteins were inferred from openly accessible protein databases. Protein abundance was determined by spectral counting. Publicly available software packages were used to identify proteins that were differentially expressed between human retinal and choroidal endothelial cells, and to classify proteins that were highly abundant in each endothelial cell population.

      Results

      Human retinal and/or choroidal vascular endothelial cells expressed 5042 nonredundant proteins. Setting the differential expression false discovery rate at 0.05, 498 proteins of 3454 quantifiable proteins (14.4%) with minimum mean spectral counts of 2.5 were differentially abundant in the 2 cell populations. Retinal and choroidal endothelial cells were enriched in angiogenic proteins, and retinal endothelial cells were also enriched in immunologic proteins.

      Conclusions

      This work describes the different protein expression profiles of human retinal and choroidal vascular endothelial cells, and provides multiple candidates for further study as novel treatments or drug targets for posterior eye diseases. NOTE: Publication of this article is sponsored by the American Ophthalmological Society.
      In the 21st century, the leading causes of irreversible blindness in the United States and other industrialized countries include diseases that frequently involve the vascular beds of the posterior segment of the eye: age-related macular degeneration (AMD); retinal ischemic vasculopathies related to diabetes mellitus or premature birth; and noninfectious posterior uveitis. These diseases affect approximately 2% of adults in the United States aged 18 years or older, and they may account for more than 60% of blindness in the population, depending on race and ethnicity.
      • Klein R.
      • Klein B.E.
      The prevalence of age-related eye diseases and visual impairment in aging: current estimates.
      • Gritz D.C.
      • Wong I.G.
      Incidence and prevalence of uveitis in Northern California: the Northern California Epidemiology of Uveitis Study.
      • Nussenblatt R.B.
      The natural history of uveitis.
      Treatment outcomes for these vision-threatening conditions have improved over the past decade, primarily related to the introduction of drugs designed to target proteins that mediate key stages in disease pathogenesis. Although successful therapeutically, however, the first generation of biologically targeted drugs has been associated with serious complications in some patients. Consequently, considerable effort is being directed toward the identification of new therapeutic targets for these posterior eye diseases.

       Burden of Posterior Segment Eye Disease in Industrialized Countries

      AMD is a progressive retinal disease that involves the macula and is based at the level of the retinal pigment epithelium and Bruch membrane.
      • Lim L.S.
      • Mitchell P.
      • Seddon J.M.
      • Holz F.G.
      • Wong T.Y.
      Age-related macular degeneration.
      Early AMD is characterized by the presence of retinal drusen and/or pigmentary abnormalities, and late AMD is manifest as choroidal neovascularization and/or geographic atrophy.
      • Lim L.S.
      • Mitchell P.
      • Seddon J.M.
      • Holz F.G.
      • Wong T.Y.
      Age-related macular degeneration.
      A new prospective survey of 3811 persons aged 46-86 years, enrolled in the Multi-Ethnic Study of Atherosclerosis (MESA) and representing 4 racial/ethnic groups in the United States, estimates 8-year incidence of early AMD at 4.1% and late AMD at 2.3%.
      • Fisher D.E.
      • Klein B.E.
      • Wong T.Y.
      • et al.
      Incidence of age-related macular degeneration in a multi-ethnic United States population: the Multi-Ethnic Study of Atherosclerosis.
      Age-related macular degeneration is particularly common in middle-aged and elderly persons, with cumulative incidence climbing rapidly from the seventh decade.
      • Klein R.
      • Klein B.E.
      • Knudtson M.D.
      • Meuer S.M.
      • Swift M.
      • Gangnon R.E.
      Fifteen-year cumulative incidence of age-related macular degeneration: the Beaver Dam Eye Study.
      Accounting for approximately one half of late AMD cases,
      • Klein R.
      • Klein B.E.
      • Knudtson M.D.
      • Meuer S.M.
      • Swift M.
      • Gangnon R.E.
      Fifteen-year cumulative incidence of age-related macular degeneration: the Beaver Dam Eye Study.
      choroidal neovascularization carries a high risk of vision loss; an early report from the Beaver Dam Eye Study, conducted before effective treatment was developed, recorded the frequency of legal blindness in persons with choroidal neovascularization at 47%.
      • Klein R.
      • Wang Q.
      • Klein B.E.
      • Moss S.E.
      • Meuer S.M.
      The relationship of age-related maculopathy, cataract, and glaucoma to visual acuity.
      Biologic drugs that target the master angiogenic regulator, vascular endothelial growth factor (VEGF), have substantially reduced the risk of vision loss from choroidal neovascularization in late AMD.
      • Bressler N.M.
      • Doan Q.V.
      • Varma R.
      • et al.
      Estimated cases of legal blindness and visual impairment avoided using ranibizumab for choroidal neovascularization: non-Hispanic white population in the United States with age-related macular degeneration.
      • Campbell J.P.
      • Bressler S.B.
      • Bressler N.M.
      Impact of availability of anti-vascular endothelial growth factor therapy on visual impairment and blindness due to neovascular age-related macular degeneration.
      However, while resolving neovascularization, therapeutic VEGF blockade has been associated with retinal atrophy, which itself leads to reduction of the visual acuity.
      • Channa R.
      • Sophie R.
      • Bagheri S.
      • et al.
      Regression of choroidal neovascularization results in macular atrophy in anti-vascular endothelial growth factor-treated eyes.
      • Grunwald J.E.
      • Pistilli M.
      • Ying G.S.
      • et al.
      Growth of geographic atrophy in the comparison of age-related macular degeneration treatments trials.
      Diabetic retinopathy is a retinal ischemic vasculopathy that evolves from an early “nonproliferative” phase, characterized by microaneurysms and other structural vascular changes, and microvascular occlusions and leakage, to a late “proliferative” phase with neovascularization; macular edema can occur at any stage in this course.
      • Cheung N.
      • Mitchell P.
      • Wong T.Y.
      Diabetic retinopathy.
      Examination of a cross-sectional, nationally representative sample of the 2005-2008 National Health and Nutrition Examination Survey (NHANES) reveals that 28.5% of persons aged 40 years and older who have diabetes mellitus—equating to 3.8% of the United States population—suffer from this condition.
      • Zhang X.
      • Saaddine J.B.
      • Chou C.F.
      • et al.
      Prevalence of diabetic retinopathy in the United States, 2005-2008.
      Vision loss in diabetic retinopathy is generally related to macular edema and/or complications related to retinal neovascularization.
      • Cheung N.
      • Mitchell P.
      • Wong T.Y.
      Diabetic retinopathy.
      In the Wisconsin Epidemiologic Study of Diabetic Retinopathy, the 25-year incidence of vision impairment was 13% and of legal blindness was 3% for a group of 955 persons with diabetes diagnosed before age 30 years and requiring treatment with insulin.
      • Klein R.
      • Lee K.E.
      • Gangnon R.E.
      • Klein B.E.
      The 25-year incidence of visual impairment in type 1 diabetes mellitus the Wisconsin Epidemiologic Study of Diabetic Retinopathy.
      Retinal laser photocoagulation has been the mainstay treatment of diabetic retinopathy for several decades. However, VEGF blockade provides superior visual acuity outcomes for diabetic macular edema
      • Diabetic Retinopathy Clinical Research Network
      • Elman M.J.
      • Aiello L.P.
      • et al.
      Randomized trial evaluating ranibizumab plus prompt or deferred laser or triamcinolone plus prompt laser for diabetic macular edema.
      and has become preferred practice for this disease manifestation.
      • American Academy of Ophthalmology Retina/Vitreous Panel
      Preferred Practice Pattern Guidelines. Diabetic Retinopathy.
      At this time, laser photocoagulation continues to be recommended for preproliferative and proliferative diabetic retinopathy,
      • American Academy of Ophthalmology Retina/Vitreous Panel
      Preferred Practice Pattern Guidelines. Diabetic Retinopathy.
      and though it is often effective, reduced visual field, color vision, and contrast sensitivity are well documented in treated patients.
      • Fong D.S.
      • Girach A.
      • Boney A.
      Visual side effects of successful scatter laser photocoagulation surgery for proliferative diabetic retinopathy: a literature review.
      Retinopathy of prematurity (ROP) is a serious retinal ischemic vasculopathy presenting in infants born preterm, when normal retinal angiogenesis is perturbed outside the intrauterine environment.
      • Hartnett M.E.
      Pathophysiology and mechanisms of severe retinopathy of prematurity.
      The disease manifests at the junction of vascular and avascular retina, where neovascularization may develop, often heralded by vascular engorgement and haziness of the vitreous, and with the potential to progress to retinal detachment. In studies conducted across a 27-year period ending in 2013,
      • Quinn G.E.
      • Barr C.
      • Bremer D.
      • et al.
      Changes in course of retinopathy of prematurity from 1986 to 2013: comparison of three studies in the United States.
      overall incidence of ROP in the United States has remained relatively stable at mean of 61% for infants born with a weight of 1250 g or less; the National Vital Statistics Reports for 2013 indicate that approximately 1.0% of babies are born in this weight category.
      • Martin J.A.
      • Hamilton B.E.
      • Osterman M.J.K.
      • Curtin S.C.
      • Mathews T.J.
      Births: final data for 2013.
      Although clearly less common than AMD and diabetic retinopathy, ROP has a substantial impact on health-related quality of life from childhood.
      • Quinn G.E.
      • Dobson V.
      • Saigal S.
      • et al.
      Health-related quality of life at age 10 years in very low-birth-weight children with and without threshold retinopathy of prematurity.
      Though treatment with retinal photocoagulation reduces the risk of progression of severe disease to retinal detachment, structural complications may occur.
      • McLoone E.
      • O’Keefe M.
      • McLoone S.
      • Lanigan B.
      Long term functional and structural outcomes of laser therapy for retinopathy of prematurity.
      Biologic drugs directed against VEGF have been used in a limited number of infants with severe ROP, but a recent Cochrane Neonatal analysis
      • Sankar M.J.
      • Sankar J.
      • Mehta M.
      • Bhat V.
      • Srinivasan R.
      Anti-vascular endothelial growth factor (VEGF) drugs for treatment of retinopathy of prematurity.
      concludes that “insufficient data precludes strong conclusions favoring routine use of intravitreal anti-VEGF agents in preterm infants” and, importantly, raises concern about the potential for systemic adverse effects, including myocardial dysfunction and neurodevelopmental abnormalities, in these patients.
      Uveitis—defined as intraocular inflammation—has an incidence of approximately 25/100 000 person-years in the United States; most cases of the disease begin during the working years.
      • Suhler E.B.
      • Lloyd M.J.
      • Choi D.
      • Rosenbaum J.T.
      • Austin D.F.
      Incidence and prevalence of uveitis in Veterans Affairs Medical Centers of the Pacific Northwest.
      • Acharya N.R.
      • Tham V.M.
      • Esterberg E.
      • et al.
      Incidence and prevalence of uveitis: results from the Pacific Ocular Inflammation Study.
      Although it is a relatively uncommon disease, data from United Kingdom specialist clinics indicate uveitis causes vision loss in 70% of individuals, with 55% of these persons experiencing legal blindness.
      • Durrani O.M.
      • Tehrani N.N.
      • Marr J.E.
      • Moradi P.
      • Stavrou P.
      • Murray P.I.
      Degree, duration, and causes of visual loss in uveitis.
      Quality of life for patients with uveitis is considerably lower than would be expected from level of vision alone, probably reflecting the common association with systemic diseases and frequent need for oral medications.
      • Multicenter Uveitis Steroid Treatment Trial Research Group
      • Kempen J.H.
      • Altaweel M.M.
      • Holbrook J.T.
      • Jabs D.A.
      • Sugar E.A.
      The multicenter uveitis steroid treatment trial: rationale, design, and baseline characteristics.
      The economic cost for persons who are blind or vision impaired from uveitis is equivalent to the cost for persons with diabetes mellitus, both estimated at USD 240 million per year in 2011.
      • de Smet M.D.
      • Taylor S.R.
      • Bodaghi B.
      • et al.
      Understanding uveitis: the impact of research on visual outcomes.
      When inflammation involves the posterior eye, impacts on vision and quality of life are particularly high.
      • Rothova A.
      • Suttorp-van Schulten M.S.
      • Frits Treffers W.
      • Kijlstra A.
      Causes and frequency of blindness in patients with intraocular inflammatory disease.
      Most patients who develop noninfectious posterior uveitis are treated with conventional immunosuppressive drugs, including antimetabolites and calcineurin inhibitors.
      • Jabs D.A.
      • Rosenbaum J.T.
      • Foster C.S.
      • et al.
      Guidelines for the use of immunosuppressive drugs in patients with ocular inflammatory disorders: recommendations of an expert panel.
      These drugs control the disease in just 30%-51% of patients, and to maintain control, many of those individuals require additional treatment with corticosteroid.
      • Gangaputra S.
      • Newcomb C.W.
      • Liesegang T.L.
      • et al.
      Methotrexate for ocular inflammatory diseases.
      • Pasadhika S.
      • Kempen J.H.
      • Newcomb C.W.
      • et al.
      Azathioprine for ocular inflammatory diseases.
      • Daniel E.
      • Thorne J.E.
      • Newcomb C.W.
      • et al.
      Mycophenolate mofetil for ocular inflammation.
      • Kacmaz R.O.
      • Kempen J.H.
      • Newcomb C.
      • et al.
      Cyclosporine for ocular inflammatory diseases.
      Also of concern in using conventional immunosuppressive drugs is the wide range of side effects related to nonspecific mechanisms of action.
      • Jabs D.A.
      • Rosenbaum J.T.
      • Foster C.S.
      • et al.
      Guidelines for the use of immunosuppressive drugs in patients with ocular inflammatory disorders: recommendations of an expert panel.
      Over the past 10 years, biologic drugs that specifically inhibit the activity of the inflammatory cytokine, tumor necrosis factor (TNF)-α, have been used to treat uveitis. The most widely cited prospective clinical trial to evaluate TNF-α blockade by infliximab in patients with recalcitrant uveitis
      • Suhler E.B.
      • Smith J.R.
      • Wertheim M.S.
      • et al.
      A prospective trial of infliximab therapy for refractory uveitis: preliminary safety and efficacy outcomes.
      • Suhler E.B.
      • Smith J.R.
      • Giles T.R.
      • et al.
      Infliximab therapy for refractory uveitis: 2-year results of a prospective trial.
      suggests that, although infliximab is effective in approximately 75% of patients, inhibition of homeostatic functions of TNF-α may cause serious toxicity, leading one quarter of patients to cease the drug.

       Pathogenesis of Disease as a Target for Innovative Therapies

      To identify relevant biologic drug targets for any disease, consideration must be given to the key processes that mediate the pathology. A diverse array of molecules and cell populations participate in the different basic pathogenic mechanisms that characterize late AMD, advanced retinal ischemic vasculopathy, and/or noninfectious posterior uveitis: neovascularization, increase in vascular permeability, and/or leukocyte–endothelial cell interactions.
      • Campochiaro P.A.
      Molecular pathogenesis of retinal and choroidal vascular diseases.
      • Perez V.L.
      • Caspi R.R.
      Immune mechanisms in inflammatory and degenerative eye disease.
      However, one cell type is a major player in all 3 processes; each of these posterior eye diseases involves an ocular vasculature, and the endothelial cells that line the vasculature have critical roles in the initiation of the pathology.
      Neovascularization in late AMD, or proliferative diabetic retinopathy and advanced ROP, are characterized by the abnormal development of new blood vessels from the choroidal or retinal circulations, respectively. Some forms of noninfectious posterior uveitis are also complicated by ocular neovascularization, albeit infrequently. Neovascularization is initiated by an endothelial “sprout.”
      • Gerhardt H.
      VEGF and endothelial guidance in angiogenic sprouting.
      Within this sprout, endothelial cells exist as tip cells and stalk cells. Tip cells migrate to direct the sprout, and stalk cells proliferate to build the sprout. As the sprout advances and expands, formation of a lumen and addition of other cells results in a complete blood vessel. Activated endothelial cells extend ring-shaped “podosomes” to remodel the basement membrane as angiogenic sprouts form.
      • Linder S.
      The matrix corroded: podosomes and invadopodia in extracellular matrix degradation.
      As the cells that direct blood vessel growth, endothelial cells play a key role in neovascularization. In choroidal neovascularization in late AMD, accumulating extracellular debris triggers a molecular signaling cascade that ultimately initiates new blood vessel growth within the choroid; these vessels grow into the retina.
      • Ding X.
      • Patel M.
      • Chan C.C.
      Molecular pathology of age-related macular degeneration.
      In proliferative diabetic retinopathy, toxic metabolites and other molecules cause vascular dysfunction in the retina, limiting oxygen supply to the tissue; hypoxia stimulates the growth of retinal blood vessels.
      • Antonetti D.A.
      • Klein R.
      • Gardner T.W.
      Diabetic retinopathy.
      High oxygen tension in the retina of premature infants who are treated with supplemental oxygen leads to degeneration of retinal vessels; the resulting hypoxia triggers retinal neovascularization in ROP.
      • Sapieha P.
      • Joyal J.S.
      • Rivera J.C.
      • et al.
      Retinopathy of prematurity: understanding ischemic retinal vasculopathies at an extreme of life.
      Neovascularization in noninfectious posterior uveitis is explained by the angiogenic properties of cytokines and growth factors that are released in the course of inflammation.
      In contrast to the choroidal vascular endothelium, which is fenestrated and therefore inherently “leaky,” the retinal vascular endothelium has no fenestrations, and presents adherens junctions and high-resistance tight junctions. These features of the retinal endothelium are maintained through interactions with other retinal cells, including pericytes, neurons, and glial cells, and they contribute to the substantial blood-retinal barrier that exists in the healthy eye.
      • Cunha-Vaz J.G.
      The blood-retinal barriers system. Basic concepts and clinical evaluation.
      • Runkle E.A.
      • Antonetti D.A.
      The blood-retinal barrier: structure and functional significance.
      An increase in retinal vascular permeability is a defining feature of the retinal ischemic vasculopathies that may appear early in the course of disease. This has been best studied in experimental models of diabetic retinopathy, in which VEGF and inflammatory cytokines induce alterations in the molecular components of retinal endothelial junctional complexes.
      • Titchenell P.M.
      • Antonetti D.A.
      Using the past to inform the future: anti-VEGF therapy as a road map to develop novel therapies for diabetic retinopathy.
      • Klaassen I.
      • Van Noorden C.J.
      • Schlingemann R.O.
      Molecular basis of the inner blood-retinal barrier and its breakdown in diabetic macular edema and other pathological conditions.
      Retinal vascular leakage is also a feature of noninfectious posterior uveitis, and, although less well studied, is associated with the release of an overlapping spectrum of molecules and may reflect overlapping mechanisms.
      Noninfectious posterior uveitis is initiated when lymphoid and myeloid leukocyte subsets infiltrate the retina and/or choroid.
      • Van Ginderdeuren R.
      • Van Calster J.
      • Stalmans P.
      • Van den Oord J.
      A new and standardized method to sample and analyse vitreous samples by the Cellient automated cell block system.
      Experimental work in rodents has convincingly showed that leukocytes migrate into the posterior segment of the eye across the retinal vasculature,
      • Parnaby-Price A.
      • Stanford M.R.
      • Biggerstaff J.
      • et al.
      Leukocyte trafficking in experimental autoimmune uveitis in vivo.
      • Chen X.
      • Kezic J.M.
      • Forrester J.V.
      • et al.
      In vivo multi-modal imaging of experimental autoimmune uveoretinitis in transgenic reporter mice reveals the dynamic nature of inflammatory changes during disease progression.
      supporting the observation that many patients with noninfectious posterior uveitis have retinal vasculitis as a component of their disease.
      • Levy-Clarke G.A.
      • Nussenblatt R.
      Retinal vasculitis.
      Migration of leukocytes from the circulation into a tissue across the blood vessel wall is controlled by the vascular endothelium, through its surface expression of adhesion molecules and chemokines that interact with ligands and receptors on leukocytes.
      • Nourshargh S.
      • Alon R.
      Leukocyte migration into inflamed tissues.
      The constitutive and induced expression of these endothelial proteins directs stages of the migration that include rolling, firm adhesion, spreading and crawling, and transmigration. Although the process is less well characterized, leukocytes also interact with the retinal vascular endothelium in retinal ischemic vasculopathies, inducing endothelial cell injury and death,
      • Joussen A.M.
      • Murata T.
      • Tsujikawa A.
      • Kirchhof B.
      • Bursell S.E.
      • Adamis A.P.
      Leukocyte-mediated endothelial cell injury and death in the diabetic retina.
      • Joussen A.M.
      • Poulaki V.
      • Le M.L.
      • et al.
      A central role for inflammation in the pathogenesis of diabetic retinopathy.
      • Tian P.
      • Ge H.
      • Liu H.
      • et al.
      Leukocytes from diabetic patients kill retinal endothelial cells: effects of berberine.
      and potentially with the choroidal vascular endothelium in AMD, when neovascular lesions may be infiltrated with monocytes.
      • Tsutsumi C.
      • Sonoda K.H.
      • Egashira K.
      • et al.
      The critical role of ocular-infiltrating macrophages in the development of choroidal neovascularization.
      • Luhmann U.F.
      • Robbie S.
      • Munro P.M.
      • et al.
      The drusenlike phenotype in aging Ccl2-knockout mice is caused by an accelerated accumulation of swollen autofluorescent subretinal macrophages.
      • Krause T.A.
      • Alex A.F.
      • Engel D.R.
      • Kurts C.
      • Eter N.
      VEGF-production by CCR2-dependent macrophages contributes to laser-induced choroidal neovascularization.
      Products of the choroidal or retinal vascular endothelial cells that mediate neovascularization, vascular permeability changes, and/or leukocyte–endothelial cell interactions may represent novel therapeutic targets for AMD, retinal ischemic vasculopathies, and/or noninfectious posterior uveitis. Vascular endothelial cells in different tissues exist in different microenvironments and perform different functions.
      • Atkins G.B.
      • Jain M.K.
      • Hamik A.
      Endothelial differentiation: molecular mechanisms of specification and heterogeneity.
      Accordingly, the molecular composition of each endothelial population is different and specific to function, and may predispose to tissue-specific diseases. The implication is that local endothelial cell populations may present potential targets for novel biologic therapies. The principle of targeting a “vascular zipcode” has already been applied in humans for disease outside the eye, including cancer.
      • Pasqualini R.
      • Millikan R.E.
      • Christianson D.R.
      • et al.
      Targeting the interleukin-11 receptor alpha in metastatic prostate cancer: a first-in-man study.
      Recent research from our group has provided proof-of-concept for application to eye pathologies. Our microarray profiling study identified high levels of intercellular adhesion molecule (ICAM)-1 on human retinal endothelial cells, in comparison to choroidal endothelial cells. We subsequently showed that transmigration of human lymphoid cells—including Th1 and Th17 helper T cells, as well as B cells
      • Bharadwaj A.S.
      • Schewitz-Bowers L.P.
      • Wei L.
      • Lee R.W.
      • Smith J.R.
      Intercellular adhesion molecule 1 mediates migration of Th1 and Th17 cells across human retinal vascular endothelium.
      • Bharadwaj A.S.
      • Stempel A.J.
      • Olivas A.
      • et al.
      Molecular signals involved in human B cell migration into the retina: in vitro investigation of ICAM-1, VCAM-1, and CXCL13.
      —across the retinal vascular endothelium, as occurs in noninfectious posterior uveitis, may be inhibited by antibody-mediated blockade of ICAM-1.

       Defining the Human Retinal and Choroidal Vascular Endothelial Cell Phenotypes

      In planning to target the ocular vascular endothelium therapeutically, it would be essential to focus on the vascular bed that is primarily involved in the pathology: the choroidal vasculature in AMD, and the retinal vasculature in ischemic retinopathies and noninfectious posterior uveitis. Specifically directing drug at the pathogenic endothelial cell population should effectively inhibit disease, without toxicity to the noninvolved vasculature. To this end, our research team has developed methods for isolating retinal and choroidal vascular endothelial cells from human cadaveric eyes,
      • Bharadwaj A.S.
      • Appukuttan B.
      • Wilmarth P.A.
      • et al.
      Role of the retinal vascular endothelial cell in ocular disease.
      and conducted multiple published studies aimed at defining the molecular profile of each cell type.
      • Bharadwaj A.S.
      • Appukuttan B.
      • Wilmarth P.A.
      • et al.
      Role of the retinal vascular endothelial cell in ocular disease.
      • Smith J.R.
      • Choi D.
      • Chipps T.J.
      • et al.
      Unique gene expression profiles of donor-matched human retinal and choroidal vascular endothelial cells.
      • Choi D.
      • Appukuttan B.
      • Binek S.J.
      • et al.
      Prediction of cis-regulatory elements controlling genes differentially expressed by retinal and choroidal vascular endothelial cells.
      • Zamora D.O.
      • Riviere M.
      • Choi D.
      • et al.
      Proteomic profiling of human retinal and choroidal endothelial cells reveals molecular heterogeneity related to tissue of origin.
      • Chipps T.J.
      • Appukuttan B.
      • Pan Y.
      • Smith J.R.
      CD44 isoforms in human retinal and choroidal endothelial cells.
      • Smith J.R.
      • Chipps T.J.
      • Ilias H.
      • Pan Y.
      • Appukuttan B.
      Expression and regulation of activated leukocyte cell adhesion molecule in human retinal vascular endothelial cells.
      Because molecular phenotype may vary considerably between different individuals, retinal endothelial cells have been profiled against choroidal endothelial cells from the same human donor.
      Gene expression microarrays provided our first opportunity to define the ocular endothelial cell phenotype at a molecular level, generating transcriptomes of a size that was primarily limited by the number of oligonucleotide probes on the array. We used the Affymetrix Human Genome Focus Array to compare expression of 8743 transcripts in human retinal and choroidal vascular endothelial cells isolated from 6 human cadaver eyes.
      • Smith J.R.
      • Choi D.
      • Chipps T.J.
      • et al.
      Unique gene expression profiles of donor-matched human retinal and choroidal vascular endothelial cells.
      This work showed that despite a high degree of correlation—Pearson coefficient of 0.95-0.99—the gene expression profiles of retinal and choroidal endothelial cells were distinct, and differences between the cell populations were more striking than inter-individual differences. Gene ontology classification revealed that 779 (8.9%) differentially expressed transcripts in human retinal and choroidal endothelial cells included a high representation of molecules involved in cell proliferation, which could be expected to participate in neovascularization. In addition, human retinal endothelial cells had high representation of molecules involved in the immune response and inflammation.
      Because proteins ultimately determine the function of a cell, the logical next step for our investigation was comparison of the proteomes of human retinal and choroidal vascular endothelial cells. Our first attempt to study the protein complement of human retinal and choroidal endothelial cells was performed by 2-dimensional difference gel electrophoresis.
      • Zamora D.O.
      • Riviere M.
      • Choi D.
      • et al.
      Proteomic profiling of human retinal and choroidal endothelial cells reveals molecular heterogeneity related to tissue of origin.
      This labor-intensive technique involves separation of labeled protein mixtures in gels in 2 dimensions; spots of interest are extracted and studied by mass spectrometry. The method is limited to detection of abundant proteins within a gel-determined range of protein isoelectric points and molecular weights. This approach is also biased toward water-soluble proteins and generally does not detect membrane-bound proteins. The latter is an important issue for studies of endothelial diversity because much specialization occurs at the cell surface. Co-migration to the same gel spot and differential processing, producing multiple gel spots, are other disadvantages that limit protein identification. These issues were reflected in our results, which took multiple years to obtain. In the gels, we identified just 31 protein spots that were significantly differentially expressed between the ocular endothelial cell populations, and it was possible to confidently match only 17 spots to single proteins. As a result, this work did not allow us to draw any conclusions about the global proteomes of human retinal and choroidal endothelial cells.
      Advances in mass spectrometry instrumentation, separation techniques, and informatics have enabled deep proteomic profiling, often termed “shotgun proteomics.”
      • Ahrens C.H.
      • Brunner E.
      • Qeli E.
      • Basler K.
      • Aebersold R.
      Generating and navigating proteome maps using mass spectrometry.
      • Beck M.
      • Claassen M.
      • Aebersold R.
      Comprehensive proteomics.
      Multiple dimensions of separation are performed in liquid phases on enzymatic digests of complex protein mixtures. The separated peptides are ionized using electrospray ion sources and introduced into mass spectrometers, where ion currents are measured for abundance estimates and peptides are fragmented for amino acid sequencing. The basic technique is largely automated, is reproducible, and has broad coverage, with relatively little dependence on protein abundance, size, charge, or water solubility. All proteins are detected above a very low minimum detection limit. In addition to differentially expressed proteins, proteins that are present at similar levels between samples are also identified. We established the feasibility of applying deep proteomic profiling to ocular endothelial cells in a study of one retinal endothelial cell isolate from a single individual.
      • Bharadwaj A.S.
      • Appukuttan B.
      • Wilmarth P.A.
      • et al.
      Role of the retinal vascular endothelial cell in ocular disease.
      In the present study, we greatly extend this investigation to characterize and compare protein expression in retinal and choroidal vascular endothelial cells from 5 human donors. Our central hypothesis is that a key to understanding the pathogenesis of blinding posterior eye diseases and developing effective therapeutic interventions is defining the unique set of molecular signals expressed by human retinal and choroidal vascular endothelial cells.

      Methods

       Human Study Statement

      Human ocular tissue was purchased from Lions VisionGift (Portland, Oregon, USA) as paired posterior globes. The US Office for Human Research Protections does not consider the deceased to be human subjects (Code of Federal Regulations, DHHS 45CFR46.102). Therefore, the Oregon Health & Science University Institutional Review Board waived the need for their approval of this work.

       Overview of Experimental Design

      Retinal and choroidal vascular endothelial cells were isolated from paired eyes of 5 human cadaveric donors, by selection on CD31 expression, and cultured under standardized conditions. Endothelial cell protein extracts were digested, and peptides were fractionated on a polysulfoethyl A cation exchange column. Fractions were analyzed by reverse-phase liquid chromatography and tandem mass spectrometry. Peptide sequences were assigned to fragment ion spectra by searching Swiss-Prot and human reference proteome, UP000005640 protein databases. Publicly available software tools were used to infer and annotate proteins. Proteins were quantified by spectral counting, to permit calculation of differential protein expression and enrichment analyses. Our experimental design is illustrated in Figure 1.
      Figure thumbnail gr1
      Figure 1Flow chart illustrating experimental design of deep proteomic profiling of human retinal and choroidal vascular endothelial cells.

       Isolation of Human Retinal and Choroidal Vascular Endothelial Cells

      Age at death and sex of the anonymous donors of human ocular tissue were as follows: 39-year-old female, 46-year-old male, 48-year-old male, 18-year-old female, and 36-year-old male. Death to endothelial cell isolation time varied from 11 to 22.5 hours. Our method for isolation of endothelial cells from human retina and choroid has been published in detail.
      • Bharadwaj A.S.
      • Appukuttan B.
      • Wilmarth P.A.
      • et al.
      Role of the retinal vascular endothelial cell in ocular disease.
      In brief, the retina and choroid were dissected from both posterior globes and separately digested with 0.3 mg/mL Dispase (Thermo Fisher Scientific-GIBCO, Grand Island, New York, USA) and 0.25-1 mg/mL type II collagenase (Sigma-Aldrich, St Louis, Missouri, USA). After 7-10 days of culture in MCDB-131 medium (Sigma-Aldrich) supplemented with 2% fetal bovine serum (FBS; GE Healthcare Life Sciences-HyClone, Logan, Utah, USA) and endothelial growth factors (EGM-2 SingleQuots supplement, omitting FBS, hydrocortisone, and gentamicin; Lonza-Clonetics, Walkersville, Maryland, USA) at 37 C, endothelial cells were purified using magnetic Dynabeads (Thermo Fisher Scientific-Invitrogen Dynal, Oslo, Norway) coated with mouse anti-human CD31 antibody (BD Pharmingen, San Diego, California, USA) and were grown in modified MCDB-131 medium with 10% FBS. Subculturing of retinal endothelial cells was performed with 0.05% trypsin (Thermo Fisher Scientific-GIBCO). The cell isolates were used at passage 2 or 3.

       Preparation of Protein Samples

      Human retinal and choroidal endothelial cells were grown to confluence in modified MCDB-131 medium with 10% FBS in separate 10-cm-diameter dishes (2 dishes per endothelial cell population). The medium was replaced with fresh MCDB-131 medium supplemented with 5% FBS and endothelial growth factors, and the cells were cultured for a further 4 hours. Subsequently the dishes were gently washed 4 times with phosphate-buffered saline (Thermo Fisher Scientific-GIBCO) at room temperature to remove serum proteins and snap frozen at -80 C ahead of protein isolation.
      On thawing, 500 μL of 100 mM ammonium bicarbonate buffer was added to the first of each set of 2 dishes. Adherent endothelial cells were dislodged using a disposable plastic cell scraper; the cell suspension was transferred to the second of each set of 2 dishes; and the process was repeated. Cells collected from each set of 2 dishes were transferred to a single centrifuge tube, and an additional 500 μL of ammonium bicarbonate buffer was used to collect any remaining cells left in the plates. Samples were dried by vacuum centrifugation, subsequently suspended in 200 μL of 8 M deionized urea containing 1 M Tris (pH 8.5) and 8 mM calcium chloride, and finally sonicated using a Fisher Scientific Model 60 Sonic Dismembrator (Thermo Fisher Scientific, Waltham, Massachusetts, USA) at a setting of 2, using 3 treatments of 15 seconds each, with an intervening 30 seconds of cooling on ice. Protein concentrations were determined using the Pierce Bicinchoninic Acid Protein Assay Kit (Thermo Fisher Scientific - Thermo Scientific, Rockford, Illinois, USA), with bovine serum albumin as the standard.
      Portions of each sample (1 mg, approximately 125 μL) were combined with 12.5 μL of 2 M methylamine and reduced by addition of 12.5 μL of 0.9 M dithiothreitol and incubation at 50 C for 15 minutes. Samples were alkylated by addition of 25 μL of 1 M iodoacetamide and incubation in the dark at room temperature for 15 minutes, followed by addition of a second 12.5 μl of 0.9 M dithiothreitol to remove unreacted iodoacetamide. Water was added at a volume of 272 μL, followed by 40 μL of 1 μg/uL Trypsin Gold (Promega Corporation, Madison, Wisconsin, USA) dissolved in 1 mM hydrochloric acid. Following an overnight digestion at 37 C, formic acid was added to a final concentration of 5%, and the peptides were extracted in solid phase using Sep-Pak Light cartridges (Millipore, Billerica, Massachusetts, USA).

       Two-dimensional Liquid Chromatography and Tandem Mass Spectrometry

      Sep-Pak-cleaned protein digests were injected onto a 100 × 2.1 mm polysulfoethyl A cation exchange column (The Nest Group, Southborough, Massachusetts, USA) at a flow rate of 200 μL/min. Mobile phase A contained 10 mM sodium phosphate (pH 3.0) and 25% acetonitrile, and mobile phase B contained the same solutions plus 350 mM potassium chloride. Following 5 minutes of loading and washing in mobile phase A, peptides were eluted using a linear gradient of 0-50% B over 45 minutes, followed by a linear gradient of 50%-100% B over 20 minutes. One-minute fractions were collected, dried by vacuum centrifugation, and redissolved by shaking in 100 μL of 5% formic acid. Fractions at the beginning or end of the salt gradient were combined, based on ultraviolet absorbance, to reduce the number of fractions to approximately 40 per sample. Volumes of 10 μL of each fraction were analyzed by liquid chromatography/mass spectrometry on an Agilent 1100 series capillary liquid chromatography system and an LTQ linear ion trap mass spectrometer (Thermo Fisher Scientific - Thermo Scientific, San Jose, California, USA), using a standard electrospray source fitted with a 34 gauge metal needle (catalogue number 97144-20040). Samples were applied at 20 μL/min to a C18 micro trap cartridge (Optimize Technologies, Oregon City, Oregon, USA), and subsequently switched onto a 0.5 × 250 mm Zorbax SB-C18 column (Agilent Technologies, Palo Alto, California, USA) using a mobile phase A containing 0.1% formic acid. The gradient was 7%-35% acetonitrile over 90 minutes at 10 μL/min flow rate. Data-dependent collection of spectra used the dynamic exclusion feature of the instrument's control software to obtain tandem mass spectra of the 3 most abundant parent ions following each survey scan. Parameters were as follows: repeat count of 1; exclusion duration of 30 seconds; list size of 50; exclusion mass width low of 1.0 Da; and exclusion mass width high of 4.0 Da. The tune file was configured with no averaging of microscans, a maximum inject time of 0.2 second, and automatic gain control targets of 3 × 104 in MS mode and 1 × 104 in MSn mode.

       Identification of Peptides and Protein Inference

      Thermo RAW files were converted to readable text files using Proteowizard
      • Chambers M.C.
      • Maclean B.
      • Burke R.
      • et al.
      A cross-platform toolkit for mass spectrometry and proteomics.
      (release 3.0.9490), and all tandem mass spectra scans in the compressed text files with a minimum ion count of 15 and a minimum absolute intensity of 100 were converted to spectral format files with Python.
      • McDonald W.H.
      • Tabb D.L.
      • Sadygov R.G.
      • et al.
      MS1, MS2, and SQT - three unified, compact, and easily parsed file formats for the storage of shotgun proteomic spectra and identifications.
      For fragment ions with a mass-to-charge ratio greater than parent ion value, intensity fraction was used to distinguish peptides with 1+ charge. Duplicate 2+ and 3+ charges were created for peptides with higher charge.
      Data were processed using the Proteomic Analysis Workbench (or PAW) pipeline,
      • Wilmarth P.A.
      • Riviere M.A.
      • David L.L.
      Techniques for accurate protein identification in shotgun proteomic studies of human, mouse, bovine, and chicken lenses.
      with steps that included control of peptide-sequencing errors, inference of protein identities from observed peptides, and quantitative estimates of protein abundance. Two protein databases were accessed from UniProt (http://www.uniprot.org) on January 19, 2016: (1) Swiss-Prot, providing reviewed canonical human protein sequences (20 146 sequences); and (2) human reference proteome, UP000005640, including isoforms and nonreviewed sequences (91 738 sequences). A total of 179 common contaminant sequences and sequence-reversed “decoy” sequences were added to permit estimation and control of the protein false discovery rate (FDR). Databases were searched using Comet
      • Eng J.K.
      • Jahan T.A.
      • Hoopmann M.R.
      Comet: an open-source MS/MS sequence database search tool.
      (version 2016.01 rev. 2). Parameters included mass parent ion tolerance of 2.5 Da and monoisotopic fragment ion tolerance of 1.0005 Da; tryptic enzyme specificity; static modifications of +57 Da on cysteine residues; variable modifications of +16 Da on methionine residues with a maximum of 2 modifications; and scoring with y- and b-ions. Comet search scores were transformed to discriminant scores using functions modeled on PeptideProphet.
      • Keller A.
      • Nesvizhskii A.I.
      • Kolker E.
      • Aebersold R.
      Empirical statistical model to estimate the accuracy of peptide identifications made by MS/MS and database search.
      Target and decoy discriminant scores were displayed as histograms and overlaid, in order to set thresholds for peptide-spectral matches and control the FDR. Peptides were separated into subclasses based on charge and modification (ie, unmodified or containing oxidized methionine), and thresholds were set independently in each subclass. Protein inference used basic parsimony logic.
      • Nesvizhskii A.I.
      • Aebersold R.
      Interpretation of shotgun proteomic data: the protein inference problem.
      Identical peptide sets were taken to represent 1 protein, and proteins inferred from subsets of peptide sets were excluded. Inference was experiment-wide, with the additional requirement of at least 2 distinct peptides per protein per biological sample. A final stage of processing was implemented to deal with highly homologous proteins. Nearly identical peptide sets were grouped together, and proteins inferred from peptide sets that were nearly subsets were grouped together. For protein families generated by these groupings, spectral count fractions were assigned to indistinguishable peptides based on the spectral counts assigned to each protein. Fractional peptide spectral counts provided more accurate protein spectral counts, referred to below as “corrected spectral counts.”
      Protein FDR was estimated using the target/decoy method.
      • Elias J.E.
      • Gygi S.P.
      Target-decoy search strategy for increased confidence in large-scale protein identifications by mass spectrometry.
      Protein inference was performed for peptides matched to decoy proteins exactly as for peptides matched to target proteins, and total number of those decoy proteins was used to calculate the protein FDR. Although the protein FDR was below 0.01 in each sample, the experiment-wide error rate exceeded 0.07. An experiment-wide protein ranking value was computed as the sum of all discriminant function scores for all peptide-spectral matches made for each protein.
      • Zhang Y.
      • Xu T.
      • Shan B.
      • et al.
      ProteinInferencer: confident protein identification and multiple experiment comparison for large scale proteomics projects.
      Target and decoy proteins were ranked from highest to lowest scores, and target and decoy scores were used to compute protein q-scores.
      • Kall L.
      • Storey J.D.
      • MacCoss M.J.
      • Noble W.S.
      Posterior error probabilities and false discovery rates: two sides of the same coin.
      Final protein identifications were reported at protein FDR of 0.01.

       Differential Expression and Enrichment Analyses

      Analysis of differential protein expression between human retinal and choroidal endothelial cell isolates was based on quantitative estimates made using the Swiss-Prot protein database, owing to the minimal peptide redundancy of that database. Corrected spectral counts for each sample were normalized to mean corrected spectral counts across the 10 samples; normalization factors that equalized total protein ranged from 0.8 to 1.2 across samples. Proteins with mean spectral counts below 2.5 were excluded. Analysis of differential protein expression was performed using R and the edgeR Bioconductor package
      • Robinson M.D.
      • McCarthy D.J.
      • Smyth G.K.
      edgeR: a Bioconductor package for differential expression analysis of digital gene expression data.
      (release 3.3). The built-in trimmed mean of M-values (or TMM) normalization method
      • Robinson M.D.
      • Oshlack A.
      A scaling normalization method for differential expression analysis of RNA-seq data.
      was applied, and a paired study design was specified and processed using the global linear modeling edgeR extensions,
      • McCarthy D.J.
      • Chen Y.
      • Smyth G.K.
      Differential expression analysis of multifactor RNA-Seq experiments with respect to biological variation.
      following example 4.1 in the edgeR User's Guide (revised June 30, 2016). Differentially expressed proteins were defined in 3 groups as high, medium, and low, according to FDR less than 0.01, from 0.01 to less than 0.05, and from 0.05 to less than 0.1, respectively.
      Richly annotated, flat-format text files for human Swiss-Prot entries available in UniProt—including protein and gene names, gene ontology groupings, pathways, and keywords—were matched to endothelial proteins. Single-protein enrichment analysis of differentially abundant retinal and choroidal endothelial cell proteins, as defined by an FDR less than 0.05, was performed with DAVID
      • Huang da W.
      • Sherman B.T.
      • Lempicki R.A.
      Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources.
      (version 6.8), with focus on keyword correlations, gene ontology, and pathway mapping. Annotation categories selected for the enrichment analyses included UP_KEYWORDS (based on UniProt designations) for keyword analysis, (2) GOTERM_BP_DIRECT for gene ontology analysis, and REACTOME_PATHWAY (using the Reactome database,
      • Fabregat A.
      • Sidiropoulos K.
      • Garapati P.
      • et al.
      The Reactome pathway knowledgebase.
      version 59) for pathway analysis. Each annotation category was selected individually, and the Functional Annotation Chart tool was run using the human genome as the background gene set. Enriched categories were defined as those achieving a DAVID-defined EASE score below 0.05 (equivalent to uncorrected P value of .05) and surviving the Benjamin-Hochberg multiple test correction.

      Results

      Raw and processed proteomic data files have been deposited to the ProteomeXchange Consortium via the PRoteomics IDEntifications (PRIDE)
      • Vizcaino J.A.
      • Csordas A.
      • del-Toro N.
      • et al.
      2016 update of the PRIDE database and its related tools.
      partner repository with the dataset identifier PXD005972. The results files, which are cited below, are contained in ZIP archive files that are lodged within the PRIDE repository deposit.
      Chromatographic separations of the 10 human ocular endothelial cell samples produced a dataset of 4 574 538 tandem mass spectra. Processing with the Proteomic Analysis Workbench pipeline, and using the UP000005640 human reference proteome protein database (holding approximately 90 000 protein sequences), resulted in peptide assignments to 1 410 959 spectra, which equated to a 30.8% identification rate. There were 15 530 spectra assigned to decoy peptide sequences for an overall peptide-spectral match FDR of 0.01. Peptides were mapped to 33 965 proteins, but after basic parsimony principles were applied and only proteins detected by 2 or more distinct peptides per biological sample were retained, 6367 noncontaminant proteins (or groups of proteins with indistinguishable sets of identified proteins) were inferred, including 458 matches to decoy proteins, for an overall protein FDR of 0.07. An experiment-wide protein score heuristic was employed to rank target and decoy protein matches and apply a protein-level false discovery control. This identified 5042 proteins at a protein FDR of 0.01 [PRIDE file path: ∼/OTHER/human_reference_proteome/results_files/; file name: HCEC_HREC_protein_summary_reference_2.xlsx].
      Approximately 90% of the proteins identified using the UP000005640 human reference proteome protein database were also present in the Swiss-Prot protein database (holding approximately 20 000 protein sequences). The highly curated Swiss-Prot database includes superior annotations and has lower peptide redundancy. Therefore, processing was repeated using this database, for a quantitative comparison of proteins expressed by human retinal vs choroidal endothelial cell populations with relative protein quantity based on spectral counts [PRIDE file path: ∼/OTHER/human_Swiss-Prot_canonical/results_files/; file name: HCEC_HREC_protein_summary_sprot.xlsx]. Homologous proteins were grouped into families before the comparative analysis was performed [PRIDE file path: ∼/OTHER/human_Swiss-Prot_canonical/results_files/; file name: HCEC_HREC_quant_protein_summary_sprot.xlsx]. Setting a mean spectral count cutoff of 2.5, to address the complication of missing data points, 3454 proteins were identified. Among these 3454 proteins, 3369 had 2 or fewer missing data points (97.5%), and 2926 (84.7%) were identified in all 10 samples. The 3454 quantifiable proteins accounted for 98.3% of the total corrected spectral counts from 4343 proteins that were confidently identified from the Swiss-Prot database, and the 2926 quantifiable proteins present in all 10 samples accounted for 96.6% of the total corrected spectral counts. The proteomics data and identification statistics are summarized in Figure 2.
      Figure thumbnail gr2
      Figure 2Flow chart summarizing proteomics data generated in profiling of human retinal and choroidal vascular endothelial cells.
      Quantification of different proteins by spectral counting separated the ocular endothelial cell samples by donor, and by endothelial cell subpopulation, as illustrated in the multidimensional scaling plot shown in Figure 3. To identify proteins that were highly abundant in human retinal vs choroidal endothelial cells, a differential expression analysis was performed with edgeR using a paired study design [PRIDE file path: ∼/OTHER/human_Swiss-Prot_canonical/results_files/; file name: HCEC_HREC_quant_protein_summary_sprot.xlsx]. With the FDR set at less than 0.10, there were 626 differentially expressed proteins, including 342 with significantly higher expression in retinal endothelial cells and 284 with significantly higher expression in choroidal endothelial cells. These 626 proteins included 338 proteins with high differential expression (FDR below 0.01), 160 proteins with medium differential expression (FDR from 0.01 to less than 0.05), and 128 proteins with low differential expression (FDR from 0.05 to less than 0.1). Differentially expressed proteins are highlighted as a correlation plot of mean corrected spectral count for each endothelial cell subpopulation in Figure 4. The proteins are listed by cell subtype and according to level of differential expression in Tables 1 and 2.
      Figure thumbnail gr3
      Figure 3Multidimensional scaling plot presenting comprehensive protein expression in human retinal and choroidal endothelial cell isolates from 5 human donors. Distances on plot correspond to leading-logarithmic fold-changes between samples. Numbers indicate retinal or choroidal endothelial cell samples from different donors. Red indicates retinal endothelial cell samples, and blue indicates choroidal endothelial cell samples.
      Figure thumbnail gr4
      Figure 4Correlation plot showing comprehensive protein expression for human retinal and choroidal endothelial cells. The 3454 quantifiable proteins are shown. The x- and y-axes represent mean corrected spectral count per protein in the 5 retinal endothelial samples and the 5 choroidal endothelial samples, respectively. Axes have logarithmic scales. Proteins having similar expression levels in both cell types are shown in light blue. Differential expression candidates are highlighted by false discovery rate: red, .05 ≤ P < .1; dark blue, .01 ≤ P < .05; and green, P < .01.
      Table 1Proteins Abundant in Human Retinal Endothelial Cell Isolates
      Protein IdentityAccession NumberFold DifferenceFalse Discovery Rate
      High differential expression (P < .01)
       Reticulon-1Q16799322.351.9287E-42
       Vascular cell adhesion protein 1 (V-CAM 1; VCAM-1)P19320140.692.5651E-19
       UMP-CMP kinase 2, mitochondrialQ5EBM0121.539.5016E-19
       Interferon-induced GTP-binding protein Mx1, N-terminally processedP2059186.281.2541E-122
       Cadherin-6P55285104.811.6841E-15
       Bone marrow stromal antigen 2 (BST-2)Q1058963.771.7632E-16
       Interferon-induced protein with tetratricopeptide repeats 1 (IFIT-1)P0991448.801.4805E-23
       Thrombospondin type-1 domain-containing protein 4Q6ZMP059.332.9679E-14
       Collagen alpha-1(XII) chainQ9971517.891.2598E-03
       Interferon-induced protein with tetratricopeptide repeats 3 (IFIT-3)O1487938.125.4275E-18
       Band 4.1-like protein 3, N-terminally processedQ9Y2J222.612.6735E-18
       Latent-transforming growth factor beta-binding protein 1 (LTBP-1)Q1476622.243.5679E-32
       Selenoprotein M (SelM)Q8WWX933.857.6337E-08
       Plasminogen activator inhibitor 2 (PAI-2)P0512020.271.1526E-10
       Lymphocyte cytosolic protein 2Q1309428.883.5115E-07
       MAP1 light chain LC2P7855915.781.7474E-12
       Poly [ADP-ribose] polymerase 9 (PARP-9)Q8IXQ621.482.1642E-08
       KynureninaseQ1671916.951.8017E-08
       Sulfotransferase family cytosolic 1B member 1 (ST1B1; Sulfotransferase 1B1)O4370417.434.7670E-09
       Thrombomodulin (TM)P0720416.885.3417E-08
       Thymidine phosphorylase (TP)P1997114.911.5711E-21
       Receptor-interacting serine/threonine-protein kinase 2O4335315.242.1159E-07
       PalladinQ8WX9310.133.8817E-18
       Ephrin-B2P5279912.391.4938E-08
       Interferon-stimulated gene 20 kDa proteinQ96AZ610.002.2520E-05
       Aspartyl/asparaginyl beta-hydroxylaseQ127977.813.8296E-36
       Helicase with zinc finger domain 2Q9BYK88.226.7407E-05
       von Willebrand factor A domain-containing protein 1Q6PCB08.063.3121E-08
       Cathepsin SP257746.226.9262E-07
       Ubiquitin-like protein ISG15P051615.931.8084E-18
       Schlafen family member 5Q08AF35.694.5551E-15
       2'-5'-oligoadenylate synthase 2 ((2-5')oligo(A) synthase 2; 2-5A synthase 2)P297284.882.6432E-07
       Fibronectin type III domain-containing protein 3BQ53EP05.751.2542E-16
       N-acetylgalactosaminyltransferase 7Q86SF25.738.7115E-09
       Stimulator of interferon genes protein (hSTING)Q86WV65.254.7853E-12
       BAG family molecular chaperone regulator 2 (BAG-2)O958165.071.4585E-08
       Filamin A-interacting protein 1-likeQ4L1805.402.2306E-09
       Contactin-associated protein 1 (Caspr; Caspr1)P783574.741.1156E-03
       Endothelial cell-specific molecule 1 (ESM-1)Q9NQ304.991.3800E-03
       Agrin C-terminal 22 kDa fragment (C90; C22)O004684.524.0361E-27
       Antigen peptide transporter 2 (APT2)Q035194.541.6080E-04
       Poly [ADP-ribose] polymerase 10 (PARP-10)Q53GL75.393.3746E-03
       Stromal membrane-associated protein 2Q8WU795.162.4286E-03
       Sterile alpha motif domain-containing protein 9 (SAM domain-containing protein 9)Q5K6514.562.0138E-06
       Sialate O-acetylesteraseQ9HAT25.304.5882E-03
       Tapasin (TPN; TPSN)O155334.222.4745E-04
       CD44 antigenP160703.956.7320E-10
       Transmembrane protein 2Q9UHN63.954.4693E-09
       Alpha-mannosidase 2Q167063.901.4740E-05
       Fermitin family homolog 3Q86UX73.872.1159E-07
       Cordon-bleu protein-like 1Q53SF73.945.8264E-09
       Pre-B-cell leukemia transcription factor-interacting protein 1Q96AQ63.918.9420E-04
       A disintegrin and metalloproteinase with thrombospondin motifs 1 (ADAM-TS 1; ADAM-TS1; ADAMTS-1)Q9UHI83.809.2275E-06
       Tripartite motif-containing protein 16O953613.744.2054E-09
       E3 ubiquitin-protein ligase DTX3LQ8TDB63.621.0363E-09
       LeupaxinO607113.623.6207E-04
       2'-5'-oligoadenylate synthase 3 ((2-5')oligo(A) synthase 3; 2-5A synthase 3)Q9Y6K53.101.7134E-08
       Nuclear factor NF-kappa-B p52 subunitQ006533.683.9752E-08
       Integrin beta-3P051063.633.0644E-13
       Secernin-3Q0VDG43.843.5917E-09
       Sorting nexin-18Q96RF03.386.3083E-03
       Prolyl 3-hydroxylase 3Q8IVL63.365.8450E-03
       Peroxidasin homologQ926263.313.5679E-32
       Urokinase plasminogen activator surface receptor (U-PAR; uPAR)Q034053.649.3995E-03
       TumstatinQ019553.575.5203E-03
       E3 ubiquitin-protein ligase RNF213Q63HN83.332.0305E-23
       Diphosphoinositol polyphosphate phosphohydrolase 2 (DIPP-2)Q9NZJ93.127.1523E-10
       Matrix metalloproteinase-14 (MMP-14)P502813.121.6330E-06
       Processed cysteine-rich motor neuron 1 proteinQ9NZV13.232.2781E-07
       N-myc-interactor (Nmi)Q132873.001.6482E-04
       RAS protein activator like-3Q86YV03.221.5416E-04
       Interferon-induced 35 kDa protein (IFP 35; Ifi-35)P802173.085.9184E-05
       Angiopoietin-2 (ANG-2)O151232.879.9565E-05
       Dihydropyrimidinase-related protein 3 (DRP-3)Q141952.785.1127E-20
       Deoxynucleoside triphosphate triphosphohydrolase SAMHD1 (dNTPase)Q9Y3Z32.949.6475E-23
       Probable ATP-dependent RNA helicase DDX58O957862.861.2912E-10
       Galectin-9 (Gal-9)O001822.821.2270E-06
       Integrin alpha-V light chainP067562.754.5718E-15
       Antigen peptide transporter 1 (APT1)Q035182.779.7709E-09
       Arylsulfatase A component CP152892.847.3993E-05
       Plasminogen activator inhibitor 1 (PAI; PAI-1)P051212.912.8521E-03
       Rho-related GTP-binding protein RhoBP627452.714.1612E-05
       Signal transducer and activator of transcription 1-alpha/betaP422242.644.8680E-16
       HLA class I histocompatibility antigen, B-49 alpha chainP304872.613.1315E-08
       Solute carrier family 2, facilitated glucose transporter member 1P111662.574.5685E-04
       Poly [ADP-ribose] polymerase 14 (PARP-14)Q460N52.592.2781E-07
       ADP-dependent glucokinase (ADP-GK; ADPGK)Q9BRR62.571.6071E-03
       Kinesin-like protein KIF13BQ9NQT82.744.8397E-04
       EGF-containing fibulin-like extracellular matrix protein 1Q128052.601.1544E-11
       Cathepsin ZQ9UBR22.586.1315E-12
       Cytosol aminopeptidaseP288382.518.9036E-21
       SupervillinO954252.551.5763E-07
       Serine/threonine-protein kinase receptor R3 (SKR3)P370232.562.8286E-04
       HLA class I histocompatibility antigen, B-13 alpha chainP304612.431.6845E-05
       Connective tissue growth factorP292792.478.9350E-11
       Netrin-4Q9HB632.501.2771E-04
       Intercellular adhesion molecule 1 (ICAM-1)P053622.434.5122E-08
       Superoxide dismutase [Mn], mitochondrialP041792.531.2777E-04
       Armadillo repeat-containing X-linked protein 2Q7L3112.493.1421E-03
       Prolyl 4-hydroxylase subunit alpha-2 (4-PH alpha-2)O154602.324.5882E-03
       Signal transducer and activator of transcription 2P526302.411.0566E-03
       Guanylate-binding protein 4Q96PP92.327.5330E-06
       Plexin-A2O750512.403.9718E-03
       Guanylate-binding protein 1P324552.373.8970E-10
       Aldo-keto reductase family 1 member C3P423302.405.8115E-08
       Four and a half LIM domains protein 3 (FHL-3)Q136432.368.8799E-05
       Ubiquitin/ISG15-conjugating enzyme E2 L6O149332.371.7013E-03
       Procollagen galactosyltransferase 1Q8NBJ52.213.5917E-09
       HLA class I histocompatibility antigen, B-44 alpha chainP304812.213.4407E-06
       Fibulin-1 (FIBL-1)P231422.152.9192E-04
       Zinc finger CCCH domain-containing protein 7BQ9UGR22.306.8369E-03
       A-kinase anchor protein 13 (AKAP-13)Q128022.346.2379E-03
       HLA class I histocompatibility antigen, Cw-4 alpha chainP305042.221.3130E-04
       SynaptopodinQ8N3V72.233.7206E-06
       Alpha-synucleinP378402.211.7265E-05
       Procollagen-lysine,2-oxoglutarate 5-dioxygenase 2O004692.122.1050E-08
       Protein CYR61O006222.135.4506E-13
       Procollagen-lysine,2-oxoglutarate 5-dioxygenase 1Q028092.057.4378E-09
       Interferon-induced, double-stranded RNA-activated protein kinaseP195252.072.5114E-05
       Leukocyte elastase inhibitor (LEI)P307402.105.0849E-07
       Beta-2-microglobulin form pI 5.3P617692.031.3317E-08
       Mannosyl-oligosaccharide glucosidaseQ137241.963.5595E-03
       Laminin subunit alpha-4Q163631.951.4306E-10
       GDP-fucose protein O-fucosyltransferase 1Q9H4881.836.2886E-03
       Receptor-type tyrosine-protein phosphatase beta (Protein-tyrosine phosphatase beta; R-PTP-beta)P234671.871.6689E-03
       Ugl-Y3P027512.073.9674E-04
       Protein-methionine sulfoxide oxidase MICAL1Q8TDZ21.924.7574E-03
       Apolipoprotein L2Q9BQE51.898.6603E-04
       OCIA domain-containing protein 2Q56VL31.921.8040E-03
       Insulin-like growth factor-binding protein 7 (IBP-7; IGF-binding protein 7; IGFBP-7)Q162701.894.8961E-05
       E3 ubiquitin-protein ligase TRIM21P194741.891.2526E-03
       Neuropathy target esteraseQ8IY171.918.4178E-03
       Adenylyl-sulfate kinaseO953401.881.1542E-06
       Tyrosine-protein kinase YesP079471.861.1557E-03
       DnaJ homolog subfamily C member 3Q132171.882.1030E-03
       Glucose 1,6-bisphosphate synthaseQ6PCE31.835.6776E-03
       Unconventional myosin-VIQ9UM541.857.1128E-07
       DnaJ homolog subfamily C member 10Q8IXB11.863.2000E-03
       PDZ and LIM domain protein 1O001511.791.2355E-09
       Ubiquitin-like modifier-activating enzyme 7 (Ubiquitin-activating enzyme 7)P412261.844.5280E-03
       Procollagen-lysine,2-oxoglutarate 5-dioxygenase 3O605681.786.9420E-06
       CD2-associated proteinQ9Y5K61.802.0799E-04
       HLA class I histocompatibility antigen, A-34 alpha chainP304531.806.4261E-04
       Delta(3,5)-Delta(2,4)-dienoyl-CoA isomerase, mitochondrialQ130111.873.4569E-03
       Golgin subfamily A member 3Q083781.777.7162E-03
       Cathepsin B heavy chainP078581.781.2355E-09
       BTB/POZ domain-containing protein KCTD12Q96CX21.812.8592E-07
       Endothelin-converting enzyme 1 (ECE-1)P428921.747.2495E-03
       Neutral cholesterol ester hydrolase 1 (NCEH)Q6PIU21.722.3990E-03
       CanstatinP085721.757.6394E-06
       Protein NOXP20Q8IWE21.712.1380E-03
       A-kinase anchor protein 2 (AKAP-2)Q9Y2D51.723.9024E-07
       Laminin subunit gamma-1P110471.719.7334E-08
       HLA class I histocompatibility antigen, A-2 alpha chainP018921.675.0540E-04
       Microtubule-associated protein RP/EB family member 2Q155551.707.5616E-03
       ArrestenP024621.771.3402E-03
       Thrombospondin-1P079961.721.6213E-10
       Laminin subunit beta-1P079421.672.4053E-07
       Cysteine and glycine-rich protein 1P212911.651.5529E-06
       Sulfotransferase 1A4 (ST1A4)P0DMN01.675.3898E-03
       Reticulocalbin-1Q152931.791.8568E-04
       PodocalyxinO005921.674.4765E-03
       LG3 peptideP981601.649.9117E-10
       Complement component C1q receptorQ9NPY31.603.0497E-04
       PDZ and LIM domain protein 4P504791.632.3121E-03
       Protein transport protein Sec24DO948551.595.6776E-03
       ERO1-like protein alpha (ERO1-L; ERO1-L-alpha)Q96HE71.607.7942E-03
       Cytoskeleton-associated protein 4Q070651.594.6796E-07
       Protein PMLP295901.562.0099E-04
       Peptidyl-prolyl cis-trans isomerase FKBP10 (PPIase FKBP10)Q96AY31.591.3187E-03
       Aldose reductase (AR)P151211.573.6207E-04
       Enoyl-CoA delta isomerase 2, mitochondrialO755211.571.3748E-04
       Poly [ADP-ribose] polymerase 4 (PARP-4)Q9UKK31.577.9152E-04
       Hypoxia up-regulated protein 1Q9Y4L11.528.6914E-07
       Filamin-binding LIM protein 1 (FBLP-1)Q8WUP21.521.9970E-03
       3-ketoacyl-CoA thiolase, mitochondrialP427651.533.5073E-04
       Catenin beta-1P352221.527.4091E-05
       TestinQ9UGI81.491.4913E-03
       Metallothionein-2 (MT-2)P027951.445.8450E-03
       Cytosolic acyl coenzyme A thioester hydrolaseO001541.484.2576E-03
       SPARCP094861.466.2462E-03
       Tropomyosin alpha-1 chainP094931.441.3187E-03
       Adipocyte plasma membrane-associated proteinQ9HDC91.447.7889E-03
       Transgelin-2P378021.471.0113E-04
       Stabilin-1Q9NY151.439.3248E-03
       Caldesmon (CDM)Q056821.441.2672E-04
       Cytosolic non-specific dipeptidaseQ96KP41.437.4091E-05
       Thioredoxin domain-containing protein 5Q8NBS91.412.2927E-04
       Major vault protein (MVP)Q147641.426.7194E-05
       Early endosome antigen 1Q150751.442.0165E-03
       AP-2 complex subunit alpha-2O949731.427.4421E-03
       Inorganic pyrophosphataseQ151811.422.3105E-03
       Calponin-2Q994391.401.4892E-03
       VinculinP182061.431.4413E-04
       Protein transport protein Sec31AO949791.366.5656E-03
       Protein disulfide-isomerase (PDI)P072371.356.9291E-04
       Protein disulfide-isomerase A6Q150841.351.3187E-03
       5'-nucleotidase (5'-NT)P215891.323.7419E-03
       Platelet endothelial cell adhesion molecule (PECAM-1)P162841.285.6776E-03
       78 kDa glucose-regulated protein (GRP-78)P110211.266.5825E-03
      Medium differential expression (.01 ≤ P < .05)
       Fibrillin-1P355554.162.9556E-02
       Colorectal mutant cancer protein (Protein MCC)P235083.611.1231E-02
       Interferon regulatory factor 9 (IRF-9)Q009783.501.0662E-02
       Gamma-glutamylaminecyclotransferase (GGACT)Q9BVM43.441.0282E-02
       Cytosolic 5'-nucleotidase 3AQ9H0P03.191.6227E-02
       Caspase-8 subunit p10Q147903.034.2523E-02
       Cysteine-rich protein 1 (CRP-1)P502382.943.0301E-02
       Bisphosphoglycerate mutase (BPGM)P077382.854.9454E-02
       TRAF family member-associated NF-kappa-B activatorQ928442.742.4167E-02
       FERM domain-containing protein 8Q9BZ672.721.1044E-02
       Protein sel-1 homolog 1Q9UBV22.713.8157E-02
       Receptor-type tyrosine-protein phosphatase epsilon (Protein-tyrosine phosphatase epsilon; R-PTP-epsilon)P234692.681.6818E-02
       Ras-related protein Rab-23Q9ULC32.672.8626E-02
       Protein YIF1BQ5BJH72.613.5580E-02
       GTPase HRas, N-terminally processedP011122.491.7867E-02
       Follistatin-related protein 1Q128412.374.3127E-02
       Transferrin receptor protein 1, serum formP027862.353.9331E-02
       Rho GTPase-activating protein 24Q8N2642.303.4022E-02
       Uncharacterized protein DKFZp434B061Q9UF832.223.0152E-02
       Laminin subunit beta-2P552682.222.6043E-02
       Armadillo repeat-containing X-linked protein 3Q9UH622.212.4141E-02
       Death-associated protein kinase 3 (DAP kinase 3)O432932.183.6324E-02
       Alpha-2-macroglobulin receptor-associated protein (Alpha-2-MRAP)P305332.182.3939E-02
       Alpha-2-macroglobulin (Alpha-2-M)P010232.181.7867E-02
       Calcium-transporting ATPase type 2C member 1 (ATPase 2C1)P981942.103.8855E-02
       Protein FAM107BQ9H0981.972.8626E-02
       Serine/threonine-protein kinase Nek7Q8TDX71.931.5320E-02
       HLA class I histocompatibility antigen, A-30 alpha chainP161881.894.4553E-02
       Laminin subunit alpha-5O152301.851.9076E-02
       Spectrin beta chain, non-erythrocytic 2O150201.814.4814E-02
       Erlin-2O949051.801.3225E-02
       Protein transport protein Sec23BQ154371.794.7127E-02
       Amyloid-like protein 2 (APLP-2)Q064811.761.3225E-02
       Torsin-1A-interacting protein 2Q8NFQ81.751.1044E-02
       Coiled-coil domain-containing protein 50Q8IVM01.703.1064E-02
       Proteasome subunit beta type-9P280651.671.5417E-02
       Lysyl oxidase homolog 2Q9Y4K01.643.8233E-02
       Poliovirus receptorP151511.644.2771E-02
       Protein disulfide-isomerase TMX3Q96JJ71.602.6557E-02
       Insulin-like growth factor 2 mRNA-binding protein 3 (IGF2 mRNA-binding protein 3; IMP-3)O004251.584.6171E-02
       Zinc finger CCCH-type antiviral protein 1Q7Z2W41.571.6349E-02
       Heme oxygenase 1 (HO-1)P096011.551.8745E-02
       Phosphoglucomutase-1 (PGM 1)P368711.553.9040E-02
       Protein ERGIC-53P492571.544.7127E-02
       GDP-L-fucose synthaseQ136301.542.3939E-02
       Protein ABHD14BQ96IU41.534.1440E-02
       GTPase IMAP family member 1Q8WWP71.484.1086E-02
       ATPase family AAA domain-containing protein 3AQ9NVI71.453.1318E-02
       Omega-amidase NIT2Q9NQR41.444.2464E-02
       Prolyl 3-hydroxylase 1Q32P281.422.7857E-02
       Proteasome subunit beta type-8P280621.411.5320E-02
       Ras-related protein Ral-AP112331.402.7822E-02
       Sequestosome-1Q135011.393.1987E-02
       Proteasome subunit beta type-4P280701.392.4847E-02
       Brain acid soluble protein 1P807231.393.0947E-02
       Dipeptidyl peptidase 3Q9NY331.381.2704E-02
       Phosphoacetylglucosamine mutase (PAGM)O953941.374.9282E-02
       Junction plakoglobinP149231.372.0335E-02
       Glutamine–fructose-6-phosphate aminotransferase [isomerizing] 1Q062101.374.6755E-02
       PDZ and LIM domain protein 7Q9NR121.363.3534E-02
       LIM and SH3 domain protein 1 (LASP-1)Q148471.363.6781E-02
       Electron transfer flavoprotein subunit beta (Beta-ETF)P381171.353.4198E-02
       Very long-chain specific acyl-CoA dehydrogenase, mitochondrial (VLCAD)P497481.351.3806E-02
       Serpin B9P504531.343.3959E-02
       Cell surface glycoprotein MUC18P431211.331.0760E-02
       Tubulointerstitial nephritis antigen-likeQ9GZM71.324.6375E-02
       Integrin alpha-5 light chainP086481.322.8626E-02
       Transmembrane emp24 domain–containing protein 10P497551.324.2321E-02
       Ras-interacting protein 1 (Rain)Q5U6511.302.3846E-02
       Catenin delta-1O607161.271.6955E-02
       Coatomer subunit deltaP484441.263.3959E-02
       PDZ and LIM domain protein 5Q96HC41.253.0301E-02
       Glycine–tRNA ligaseP412501.244.6259E-02
       Catenin alpha-1P352211.241.4979E-02
       Coronin-1CQ9ULV41.244.0132E-02
       Myosin-9P355791.221.6729E-02
       EndoplasminP146251.222.0361E-02
       Protein disulfide-isomerase A4P136671.222.3846E-02
       Ribosome-binding protein 1Q9P2E91.204.1378E-02
       Fructose-bisphosphate aldolase AP040751.194.7127E-02
      Low differential expression (.05 ≤ P < .1)
       ADAMTS-like protein 1 (ADAMTSL-1)Q8N6G63.045.4283E-02
       Pyruvate dehydrogenase protein X component, mitochondrialO003302.905.0233E-02
       N-sulphoglucosamine sulphohydrolaseP516882.885.8851E-02
       Cancer-related nucleoside-triphosphatase (NTPase)Q9BSD72.816.5206E-02
       Probable ATP-dependent RNA helicase DDX60Q8IY212.586.0765E-02
       MICOS complex subunit MIC13Q5XKP02.548.6633E-02
       Glutathione S-transferase Mu 3P212662.535.5532E-02
       HLA class I histocompatibility antigen, B-8 alpha chainP304602.536.1005E-02
       Carnitine O-palmitoyltransferase 2, mitochondrialP237862.458.1245E-02
       Golgin subfamily A member 2Q083792.095.6388E-02
       Torsin-1AO146562.088.2911E-02
       HLA class I histocompatibility antigen, A-36 alpha chainP304552.058.2512E-02
       Protein TSSC1Q53HC92.049.1018E-02
       Plastin-2P137962.039.1198E-02
       Zinc finger protein-like 1O951591.979.4984E-02
       Armadillo repeat protein deleted in velo-cardio-facial syndromeO001921.805.3859E-02
       Latency-associated peptide (TGF-beta-1; LAP)P011371.785.2776E-02
       Nitrilase homolog 1Q86X761.777.7289E-02
       Signal-induced proliferation-associated protein 1 (Sipa-1)Q96FS41.697.9404E-02
       Pyridoxal kinaseO007641.676.2313E-02
       Clathrin light chain B (Lcb)P094971.668.8203E-02
       Hematopoietic progenitor cell antigen CD34P289061.667.5125E-02
       HLA class I histocompatibility antigen, Cw-2 alpha chainP305011.645.5029E-02
       Caspase-7 subunit p11P552101.615.7422E-02
       Oxidoreductase HTATIP2Q9BUP31.609.8853E-02
       Erythrocyte band 7 integral membrane proteinP271051.606.0691E-02
       Uncharacterized protein FLJ45252Q6ZSR91.578.2911E-02
       Syntaxin-binding protein 2Q158331.557.5917E-02
       Methylcrotonoyl-CoA carboxylase beta chain, mitochondrial (MCCase subunit beta)Q9HCC01.526.3742E-02
       Unconventional myosin-IeQ129651.507.0378E-02
       Interleukin-6 receptor subunit beta (IL-6 receptor subunit beta; IL-6R subunit beta; IL-6R-beta; IL-6RB)P401891.495.5532E-02
       Peptidyl-prolyl cis-trans isomerase FKBP9 (PPIase FKBP9)O953021.485.9813E-02
       Isovaleryl-CoA dehydrogenase, mitochondrial (IVD)P264401.479.9996E-02
       Carnitine O-palmitoyltransferase 1, liver isoform (CPT1-L)P504161.475.1511E-02
       Inosine-5'-monophosphate dehydrogenase 1 (IMP dehydrogenase 1; IMPD 1; IMPDH 1)P208391.466.8953E-02
       GMP reductase 2 (GMPR 2)Q9P2T11.447.7820E-02
       CalumeninO438521.449.8142E-02
       Charged multivesicular body protein 4bQ9H4441.416.8844E-02
       Tubulin-folding cofactor BQ994261.415.8383E-02
       Xaa-Pro dipeptidase (X-Pro dipeptidase)P129551.419.5579E-02
       Sec1 family domain-containing protein 1Q8WVM81.415.7422E-02
       PalmdelphinQ9NP741.407.3210E-02
       EF-hand domain-containing protein D2Q96C191.388.0797E-02
       Myeloid-derived growth factor (MYDGF)Q969H81.385.8334E-02
       Proteasome subunit alpha type-3P257881.347.0251E-02
       3-ketoacyl-CoA thiolase, peroxisomalP091101.338.7591E-02
       CTP synthase 1P178121.336.8844E-02
       N-acetylglucosamine-6-sulfataseP155861.316.5805E-02
       Cadherin-5P331511.306.0765E-02
       Switch-associated protein 70 (SWAP-70)Q9UH651.309.1018E-02
       3-hydroxyacyl-CoA dehydrogenase type-2Q997141.305.0032E-02
       Disabled homolog 2P980821.298.5011E-02
       Aldehyde dehydrogenase, mitochondrialP050911.289.2258E-02
       Guanine nucleotide-binding protein G(I)/G(S)/G(T) subunit beta-2P628791.289.3198E-02
       Eukaryotic translation initiation factor 2 subunit 3P410911.258.3067E-02
       UDP-glucose:glycoprotein glucosyltransferase 1 (UGT1; hUGT1)Q9NYU21.258.2245E-02
       Macrophage migration inhibitory factor (MIF)P141741.238.6633E-02
       VimentinP086701.226.0139E-02
       Serpin H1P504541.215.6312E-02
       ZyxinQ159421.215.6752E-02
       Glyceraldehyde-3-phosphate dehydrogenase (GAPDH)P044061.177.3674E-02
       Protein disulfide-isomerase A3P301011.179.2770E-02
      Table 2Proteins Abundant in Human Choroidal Endothelial Cell Isolates
      Protein IdentityAccession NumberFold DifferenceFalse Discovery Rate
      High differential expression (P < .01)
       Retinal dehydrogenase 1 (RALDH 1; RalDH1)P00352249.453.4484E-94
       Estradiol 17-beta-dehydrogenase 2P3705975.581.1605E-10
       Ras-related protein Rab-3BP2033752.971.5811E-07
       Actin-binding protein anillinQ9NQW651.562.0713E-07
       Chondroitin sulfate proteoglycan 4Q6UVK141.012.5316E-08
       E3 ubiquitin-protein ligase UHRF1Q96T8829.437.6337E-08
       Lysosomal Pro-X carboxypeptidaseP4278523.374.5944E-11
       Lambda-crystallin homologQ9Y2S222.884.1068E-05
       WD repeat and HMG-box DNA-binding protein 1O7571718.609.3592E-07
       DNA topoisomerase 2-alphaP1138813.598.6215E-07
       Nidogen-2 (NID-2)Q1411211.857.2284E-05
       Kinetochore-associated protein 1P507489.881.2777E-04
       Ig alpha-1 chain C regionP018767.492.5799E-03
       Antigen KI-67P460137.332.0684E-07
       Probable ATP-dependent RNA helicase DDX47Q9H0S47.042.7926E-03
       Nesprin-3Q6ZMZ36.921.0620E-04
       Thymidine kinase, cytosolicP041836.831.4874E-03
       cGMP-dependent 3',5'-cyclic phosphodiesteraseO004086.692.4832E-11
       Clusterin alpha chainP109096.641.0942E-07
       Squalene synthase (SQS; SS)P372686.387.6337E-08
       Retinoblastoma-associated proteinP064006.262.0918E-03
       Fatty acid desaturase 2O958645.421.3130E-04
       Ribonucleoside-diphosphate reductase subunit M2P313505.236.3821E-03
       RelA-associated inhibitorQ8WUF55.041.6689E-03
       SH3 domain-containing kinase-binding protein 1Q96B975.031.8318E-06
       RNA-binding protein 42Q9BTD84.968.1362E-03
       Neuropilin-1O147864.901.2476E-11
       Filamin-interacting protein FAM101BQ8N5W94.784.4805E-03
       Sodium/potassium-transporting ATPase subunit beta-1P050264.733.5826E-03
       DNA (cytosine-5)-methyltransferase 1 (Dnmt1)P263584.737.3441E-11
       Apoptosis-associated speck-like protein containing a CARD (hASC)Q9ULZ34.561.6482E-04
       Kinetochore protein Spc24 (hSpc24)Q8NBT24.495.3607E-03
       Solute carrier family 12 member 2P550114.487.9919E-10
       Dynein light chain 2, cytoplasmicQ96FJ24.368.7140E-03
       Histone deacetylase 2 (HD2)Q927694.342.3764E-05
       Microtubule-associated protein 2 (MAP-2)P111374.247.0789E-06
       DNA mismatch repair protein Msh6 (hMSH6)P527014.248.6255E-07
       Lymphokine-activated killer T-cell-originated protein kinaseQ96KB54.092.3746E-03
       Friend leukemia integration 1 transcription factorQ015434.079.1066E-05
       Cohesin subunit SA-2Q8N3U44.042.6038E-03
       Structural maintenance of chromosomes protein 4 (SMC protein 4; SMC-4)Q9NTJ33.924.8729E-08
       Histone-lysine N-methyltransferase EHMT1Q9H9B13.758.4456E-03
       Condensin complex subunit 2Q150033.627.4091E-05
       Cyclin-dependent kinase 1 (CDK1)P064933.589.7911E-07
       Ephrin type-A receptor 2P293173.562.0412E-09
       PaladinQ9ULE63.541.3317E-08
       Condensin complex subunit 1Q150213.509.5377E-07
       D-3-phosphoglycerate dehydrogenase (3-PGDH)O431753.489.5283E-06
       Ephrin type-B receptor 4P547603.471.6857E-03
       DNA replication licensing factor MCM2P497363.462.5502E-12
       DNA replication licensing factor MCM7P339933.451.0288E-13
       Thymidylate synthase (TS; TSase)P048183.446.3148E-05
       Serine protease HTRA1Q927433.353.4762E-10
       ABI gene family member 3Q9P2A43.341.9598E-03
       DNA replication licensing factor MCM3P252053.245.8742E-16
       Flotillin-2Q142543.242.7424E-08
       Neuropilin-2O604623.231.1182E-04
       Lanosterol 14-alpha demethylase (LDM)Q168503.194.4765E-03
       Condensin complex subunit 3Q9BPX33.185.0390E-03
       DNA replication licensing factor MCM4P339913.171.5293E-14
       Non-syndromic hearing impairment protein 5O604433.113.2677E-05
       Serine/threonine-protein kinase VRK1Q999863.088.7826E-03
       Prostaglandin reductase 1 (PRG-1)Q149143.045.0412E-14
       Serine hydroxymethyltransferase, cytosolic (SHMT)P348963.035.7608E-03
       WolframinO760243.002.5006E-05
       DNA replication licensing factor MCM6Q145662.981.2355E-09
       Redox-regulatory protein FAM213AQ9BRX82.941.5213E-10
       ThymopentinP421662.903.5115E-07
       E3 ubiquitin-protein ligase NEDD4P469342.895.9330E-03
       28S ribosomal protein S27, mitochondrial (MRP-S27; S27mt)Q925522.899.5381E-03
       DNA ligase 1P188582.892.3063E-03
       DNA-directed RNA polymerase I subunit RPA1 (RNA polymerase I subunit A1)O956022.856.4900E-03
       Ras GTPase-activating protein 3Q146442.787.4596E-03
       Adenylyl cyclase-associated protein 2 (CAP 2)P401232.762.2429E-03
       Pentatricopeptide repeat domain-containing protein 3, mitochondrialQ96EY72.692.3585E-03
       DNA replication licensing factor MCM5P339922.686.2350E-08
       Mevalonate kinase (MK)Q034262.643.7045E-03
       Protein diaphanous homolog 2O608792.601.1367E-04
       Insulin-like growth factor 2 mRNA-binding protein 2 (IGF2 mRNA-binding protein 2; IMP-2)Q9Y6M12.577.2284E-05
       Homer protein homolog 3 (Homer-3)Q9NSC52.523.2998E-04
       Flotillin-1O759552.485.0711E-06
       Breast cancer anti-estrogen resistance protein 1P569452.471.8311E-07
       Isopentenyl-diphosphate Delta-isomerase 1Q139072.391.0400E-03
       Band 4.1-like protein 2O434912.382.2771E-05
       Biotin carboxylaseQ130852.365.6515E-05
       Phosphoribosyl pyrophosphate synthase-associated protein 1 (PRPP synthase-associated protein 1)Q145582.356.0725E-03
       Annexin A6P081332.341.0288E-13
       DNA-directed RNA polymerase I subunit RPA34O154462.292.3990E-03
       Structural maintenance of chromosomes protein 2 (SMC protein 2; SMC-2)O953472.253.1907E-04
       Absent in melanoma 1 proteinQ9Y4K12.253.6700E-04
       DysferlinO759232.211.9599E-14
       Mini-chromosome maintenance complex-binding protein (MCM-BP; MCM-binding protein)Q9BTE32.216.4917E-03
       Hydroxymethylglutaryl-CoA synthase, cytoplasmic (HMG-CoA synthase)Q015812.201.4892E-03
       High mobility group protein B3O153472.201.5416E-04
       Clustered mitochondria protein homologO751532.178.4027E-03
       Deoxyuridine 5'-triphosphate nucleotidohydrolase, mitochondrial (dUTPase)P333162.131.4572E-05
       Myosin-10P355802.102.1159E-07
       Acetyl-CoA acetyltransferase, cytosolicQ9BWD12.041.0339E-07
       Glia maturation factor gamma (GMF-gamma)O602342.021.3187E-03
       Adenosine kinase (AK)P552632.011.0472E-03
       Flap endonuclease 1 (FEN-1)P397481.991.2853E-03
       Diphosphomevalonate decarboxylaseP536021.958.2121E-03
       Structural maintenance of chromosomes protein 3 (SMC protein 3; SMC-3)Q9UQE71.953.3429E-03
       RadixinP352411.959.7911E-07
       Nuclear pore complex protein Nup205Q926211.936.4606E-03
       Structural maintenance of chromosomes protein 1A (SMC protein 1A; SMC-1-alpha; SMC-1A)Q146831.931.4892E-03
       Cytoskeleton-associated protein 5Q140081.932.1030E-03
       DnaJ homolog subfamily B member 4Q9UDY41.922.8286E-04
       Importin subunit alpha-1P522921.909.1461E-04
       Glycogen phosphorylase, liver formP067371.891.5530E-03
       Exportin-2 (Exp2)P550601.882.5010E-06
       N(G),N(G)-dimethylarginine dimethylaminohydrolase 1 (DDAH-1; Dimethylarginine dimethylaminohydrolase 1)O947601.882.7970E-03
       Ribonucleoside-diphosphate reductase large subunitP239211.856.5806E-04
       Histone H2A type 2-BQ8IUE61.856.2379E-03
       Beta-arrestin-1P494071.807.5528E-03
       High mobility group protein B2P265831.793.8021E-04
       Septin-10Q9P0V91.785.6776E-03
       Four and a half LIM domains protein 2 (FHL-2)Q141921.766.0753E-03
       Replication protein A 70 kDa DNA-binding subunit, N-terminally processedP276941.715.8017E-03
       Phosphoglucomutase-2 (PGM 2)Q96G031.719.1461E-04
       Processed integrin alpha-6P232291.701.2896E-03
       Myristoylated alanine-rich C-kinase substrate (MARCKS)P299661.694.2576E-03
       Histone H2A.Z (H2A/z)P0C0S51.682.7771E-03
       Oleoyl-[acyl-carrier-protein] hydrolaseP493271.671.1542E-06
       Protein-glutamine gamma-glutamyltransferase 2P219801.651.0316E-06
       Protein RCC2Q9P2581.652.2429E-03
       MyoferlinQ9NZM11.646.2969E-05
       LIM domain and actin-binding protein 1Q9UHB61.639.1810E-03
       Calpain-2 catalytic subunitP176551.605.5490E-04
       Serum deprivation-response proteinO958101.591.8600E-03
       Importin-5 (Imp5)O004101.591.0319E-03
       Nuclear autoantigenic sperm protein (NASP)P493211.561.6367E-03
       Lamin-B1P207001.563.1734E-04
       Calponin-3Q154171.561.4892E-03
       Nuclear mitotic apparatus protein 1 (NuMA protein)Q149801.541.7576E-03
       Proliferating cell nuclear antigen (PCNA)P120041.544.5280E-03
       Poly [ADP-ribose] polymerase 1 (PARP-1)P098741.542.1304E-03
       DNA-dependent protein kinase catalytic subunit (DNA-PK catalytic subunit; DNA-PKcs)P785271.444.2576E-03
      Medium differential expression (.01 ≤ P < .05)
       Formin-binding protein 1-likeQ5T0N54.782.2239E-02
       Ankyrin repeat and KH domain-containing protein 1Q8IWZ34.042.0828E-02
       Calcitonin gene-related peptide type 1 receptor (CGRP type 1 receptor)Q166024.011.5644E-02
       Replication factor C subunit 5P409373.932.4520E-02
       Protein farnesyltransferase/geranylgeranyltransferase type-1 subunit alphaP493543.901.4215E-02
       Serine/threonine-protein phosphatase 2A 56 kDa regulatory subunit gamma isoformQ133623.884.0401E-02
       5'(3')-deoxyribonucleotidase, cytosolic typeQ8TCD53.793.8518E-02
       Protein CIP2AQ8TCG13.774.4714E-02
       Metastasis-associated protein MTA1Q133303.634.8201E-02
       Actin filament-associated protein 1Q8N5563.471.8745E-02
       Serine/threonine-protein kinase RIO2Q9BVS43.394.8201E-02
       Ras-related protein Rab-27A (Rab-27)P511593.374.9374E-02
       Paired amphipathic helix protein Sin3aQ96ST33.334.2523E-02
       THO complex subunit 1 (Tho1)Q96FV93.293.6481E-02
       Sperm-associated antigen 7O753913.204.4542E-02
       Nuclear pore complex protein Nup160Q127693.132.0321E-02
       Disks large homolog 1Q129592.904.5429E-02
       Oxysterol-binding protein-related protein 3 (ORP-3; OSBP-related protein 3)Q9H4L52.872.9544E-02
       Tensin-1Q9HBL02.841.5008E-02
       Putative E3 ubiquitin-protein ligase UBR7Q8N8062.834.4696E-02
       Zinc finger CCHC domain-containing protein 8Q6NZY42.794.1955E-02
       Ras GTPase-activating-like protein IQGAP3Q86VI32.773.2269E-02
       Survival of motor neuron-related-splicing factor 30O759402.762.6732E-02
       Prolactin-inducible proteinP122732.683.7672E-02
       Ras-specific guanine nucleotide-releasing factor RalGPS2Q86X272.671.5603E-02
       Mediator of DNA damage checkpoint protein 1Q146762.591.9333E-02
       Putative ATP-dependent RNA helicase DHX30Q7L2E32.561.1716E-02
       Sister chromatid cohesion protein PDS5 homolog BQ9NTI52.531.7867E-02
       Aspartate–tRNA ligase, mitochondrialQ6PI482.473.8611E-02
       SH2 domain-containing protein 3CQ8N5H72.451.7371E-02
       Histone H2AX (H2a/x)P161042.311.2691E-02
       Nesprin-2Q8WXH02.252.2988E-02
       Dihydrofolate reductaseP003742.201.5417E-02
       Interferon regulatory factor 2-binding protein 2 (IRF-2-binding protein 2; IRF-2BP2)Q7Z5L92.161.6858E-02
       Baculoviral IAP repeat-containing protein 6Q9NR092.143.6481E-02
       WW domain-binding protein 11 (WBP-11)Q9Y2W22.071.4979E-02
       Centrin-2P412082.034.2464E-02
       39S ribosomal protein L1, mitochondrial (L1mt; MRP-L1)Q9BYD62.034.5276E-02
       Hippocalcin-like protein 1P372352.021.1336E-02
       Neural Wiskott-Aldrich syndrome protein (N-WASP)O004012.014.6888E-02
       RNA cytidine acetyltransferaseQ9H0A01.974.7127E-02
       Negative elongation factor E (NELF-E)P186151.962.7265E-02
       E3 ubiquitin-protein ligase HECTD1Q9ULT81.911.5714E-02
       Phosphatidylinositol transfer protein alpha isoform (PI-TP-alpha; PtdIns transfer protein alpha; PtdInsTP alpha)Q001691.883.7106E-02
       Nuclear pore complex protein Nup133Q8WUM01.871.9405E-02
       Histone acetyltransferase type B catalytic subunitO149291.861.9597E-02
       PininQ9H3071.852.6623E-02
       Chromobox protein homolog 5P459731.801.2588E-02
       SWI/SNF-related matrix-associated actin-dependent regulator of chromatin subfamily A member 5 (SWI/SNF-related matrix-associated actin-dependent regulator of chromatin A5)O602641.801.0662E-02
       Uveal autoantigen with coiled-coil domains and ankyrin repeatsQ9BZF91.801.7867E-02
       U4/U6.U5 tri-snRNP-associated protein 1O432901.792.7822E-02
       RUN and FYVE domain-containing protein 1Q96T511.783.2937E-02
       Cullin-4B (CUL-4B)Q136201.773.8457E-02
       Unconventional myosin-VaQ9Y4I11.763.7672E-02
       FACT complex subunit SPT16Q9Y5B91.741.5289E-02
       General transcription factor II-I (GTFII-I; TFII-I)P783471.711.1053E-02
       Replication protein A 14 kDa subunit (RP-A p14)P352441.702.8870E-02
       Cell division cycle 5-like protein (Cdc5-like protein)Q994591.653.2810E-02
       Nidogen-1 (NID-1)P145431.641.8293E-02
       Nuclear pore complex protein Nup93Q8N1F71.593.1987E-02
       YLP motif-containing protein 1P497501.564.3127E-02
       SH3 and multiple ankyrin repeat domains protein 3 (Shank3)Q9BYB01.564.6263E-02
       MARCKS-related proteinP490061.531.4468E-02
       Cytoplasmic FMR1-interacting protein 1Q7L5761.521.7558E-02
       Rho GTPase-activating protein 18Q8N3921.494.3127E-02
       U5 small nuclear ribonucleoprotein 200 kDa helicaseO756431.462.4682E-02
       Lamin-B2Q032521.452.1183E-02
       Splicing factor 3B subunit 3Q153931.452.3846E-02
       Aminopeptidase N (AP-N; hAPN)P151441.433.2886E-02
       Leucine-rich PPR motif-containing protein, mitochondrialP427041.433.2886E-02
       Transportin-1Q929731.424.3127E-02
       Leucine–tRNA ligase, cytoplasmicQ9P2J51.422.8741E-02
       Splicing factor 3B subunit 1O755331.423.5661E-02
       Polymerase I and transcript release factorQ6NZI21.411.4215E-02
       Heat shock protein beta-1 (HspB1)P047921.401.2059E-02
       Bifunctional purine biosynthesis protein PURH, N-terminally processedP319391.392.5437E-02
       Alanine--tRNA ligase, cytoplasmicP495881.372.2662E-02
       C-1-tetrahydrofolate synthase, cytoplasmic, N-terminally processed (C1-THF synthase)P115861.372.8741E-02
       Ubiquitin carboxyl-terminal hydrolase isozyme L1 (UCH-L1)P099361.354.8789E-02
       Histone H4P628051.343.7270E-02
      Low differential expression (.05 ≤ P < .1)
       Ubiquitin-conjugating enzyme E2 SQ167633.478.9505E-02
       DNA polymerase epsilon subunit 3Q9NRF93.356.6835E-02
       Protein LAP2Q96RT13.145.8970E-02
       NHS-like protein 1Q5SYE73.068.9532E-02
       SWI/SNF-related matrix-associated actin-dependent regulator of chromatin subfamily D member 1Q96GM52.948.9858E-02
       Putative ATP-dependent RNA helicase DHX57Q6P1582.927.0378E-02
       Cysteine and glycine-rich protein 2Q165272.899.6375E-02
       Peptidyl-prolyl cis-trans isomerase G (PPIase G; Peptidyl-prolyl isomerase G)Q134272.708.6633E-02
       Lysine-specific histone demethylase 1AO603412.696.6003E-02
       Protein AAR2 homologQ9Y3122.677.0378E-02
       Thioredoxin-related transmembrane protein 2Q9Y3202.648.7477E-02
       Ribosomal protein S6 kinase alpha-3 (S6K-alpha-3)P518122.455.2230E-02
       Zinc finger protein 598Q86UK72.406.9294E-02
       DNA ligase 3P499162.395.3219E-02
       Probable ATP-dependent RNA helicase DDX49Q9Y6V72.396.5206E-02
       Nuclear factor of activated T-cells, cytoplasmic 2 (NF-ATc2; NFATc2)Q134692.378.3792E-02
       CUGBP Elav-like family member 2 (CELF-2)O953192.316.5455E-02
       Myeloid differentiation primary response protein MyD88Q998362.308.8203E-02
       KIF1-binding proteinQ96EK52.305.7952E-02
       RNA polymerase II-associated factor 1 homolog (hPAF1)Q8N7H52.256.8632E-02
       BRCA1-associated ATM activator 1Q6PJG62.208.0430E-02
       Zinc finger CCCH domain-containing protein 18Q86VM92.166.9760E-02
       28S ribosomal protein S31, mitochondrial (MRP-S31; S31mt)Q926652.159.1137E-02
       Exocyst complex component 3O606452.148.7993E-02
       Rab GTPase-activating protein 1Q9Y3P92.095.8851E-02
       55 kDa erythrocyte membrane protein (p55)Q000132.026.9675E-02
       Phosphatidylinositol 4-phosphate 3-kinase C2 domain-containing subunit alpha (PI3K-C2-alpha; PtdIns-3-kinase C2 subunit alpha)O004432.009.6375E-02
       Tyrosine-protein kinase BAZ1BQ9UIG01.996.1715E-02
       Microsomal glutathione S-transferase 3 (Microsomal GST-3)O148801.985.7982E-02
       Eukaryotic translation initiation factor 4E-binding protein 1 (4E-BP1; eIF4E-binding protein 1)Q135411.938.9511E-02
       Lamin-B receptorQ147391.926.9675E-02
       Dipeptidyl peptidase 9 (DP9)Q86TI21.889.1018E-02
       Serine/threonine-protein phosphatase 2A 56 kDa regulatory subunit delta isoformQ147381.875.2776E-02
       Probable rRNA-processing protein EBP2Q998481.875.6022E-02
       Tetratricopeptide repeat protein 37 (TPR repeat protein 37)Q6PGP71.867.7289E-02
       SWI/SNF complex subunit SMARCC1Q929221.858.8203E-02
       5'-nucleotidase domain-containing protein 1Q5TFE41.849.3464E-02
       G-rich sequence factor 1 (GRSF-1)Q128491.828.7164E-02
       Cadherin-13P552901.816.0926E-02
       Sister chromatid cohesion protein PDS5 homolog AQ29RF71.647.2757E-02
       Protein NDRG1Q925971.635.1336E-02
       FACT complex subunit SSRP1Q089451.599.4540E-02
       Protein SONP185831.585.5298E-02
       Copine-3O751311.575.7168E-02
       Peptidyl-prolyl cis-trans isomerase FKBP5 (PPIase FKBP5)Q134511.557.0911E-02
       Thymidylate kinaseP239191.558.7591E-02
       Thymosin alpha-1P064541.546.8383E-02
       Splicing factor U2AF 35 kDa subunitQ010811.548.2342E-02
       Protein AHNAK2Q8IVF21.548.6300E-02
       Epididymal secretory protein E1P619161.537.1398E-02
       Probable ATP-dependent RNA helicase DDX46Q7L0141.525.6022E-02
       Endophilin-A2Q999611.528.4677E-02
       Glutaredoxin-3O760031.525.6259E-02
       Hypoxanthine-guanine phosphoribosyltransferase (HGPRT; HGPRTase)P004921.437.7289E-02
       Acidic leucine-rich nuclear phosphoprotein 32 family member BQ926881.438.0689E-02
       Calcyclin-binding protein (CacyBP; hCacyBP)Q9HB711.437.3928E-02
       Transcription factor BTF3P202901.405.2165E-02
       Pre-mRNA-processing-splicing factor 8Q6P2Q91.386.6374E-02
       Peroxiredoxin-5, mitochondrialP300441.359.1446E-02
       Nucleoprotein TPRP122701.358.2970E-02
       Heterogeneous nuclear ribonucleoprotein H, N-terminally processed (hnRNP H)P319431.355.6072E-02
       Annexin A1P040831.338.4335E-02
       Histone H2B type 1-NQ998771.335.6463E-02
       14-3-3 protein thetaP273481.327.5796E-02
       ATP-citrate synthaseP533961.299.1201E-02
       Neuroblast differentiation-associated protein AHNAKQ096661.259.8574E-02
      To delineate the molecular phenotype of human retinal and choroidal endothelial cells, enrichment analyses were performed with DAVID, classifying the 498 proteins that were differentially expressed at the standard FDR of 0.05 [PRIDE file path: ∼/OTHER/human_Swiss-Prot_canonical/results_files/; file names: choroid_DAVID_All.xlsx and retina_DAVID_All.xlsx]. Analyses for both cell populations yielded multiple annotations related to processes involved in blood vessel growth and/or vascular permeability. For retinal endothelial cells, keywords included “extracellular matrix,” “angiogenesis,” and “cell junction”; gene ontology groupings included “extracellular matrix organization,” “extracellular matrix disassembly,” “angiogenesis,” “negative regulation of angiogenesis,” “positive regulation of cell migration,” and “negative regulation of cell migration”; and pathways included “non-integrin membrane-extracellular matrix interactions” and “degradation of extracellular matrix.” For choroidal endothelial cells, keywords included “cell cycle,” “DNA replication,” “mitosis,” and “cell division”; gene ontology groupings included “cell division,” “cell-cell adhesion,” and “cell proliferation”; and pathways included a range of intracellular processes that characterize cell division, such as “activation of the pre-replicative complex,” “unwinding of DNA,” and “cohesin loading onto chromatin.” The enrichment analyses of retinal endothelial cells also yielded multiple annotations associated with the immune response: keywords included “immunity” and “innate immunity”; gene ontology groupings included “type I interferon signaling pathway,” “interferon-γ signaling pathway,” “leukocyte migration,” and “leukocyte cell-cell adhesion”; and pathways included “interferon α/β signaling,” “interferon-γ signaling,” and “integrin cell surface interactions.” Table 3, Table 4, and 5 list the annotations enriched in proteins that were significantly more abundant in human retinal or choroidal endothelial cells, by keyword, gene ontology, and pathway, respectively.
      Table 3Keywords Enriched in Proteins That Are Differentially Abundant in Human Ocular Endothelial Cells
      KeywordFold EnrichmentEASE Score
      EASE score is generated by enrichment analysis in DAVID (version 6.8) and refers to uncorrected P value.
      Abundant in human retinal endothelial cells
       Cell adhesion5.451.93E-15
       Phosphoprotein1.571.24E-14
       Endoplasmic reticulum3.492.65E-14
       Basement membrane24.231.02E-13
       Antiviral defense11.572.63E-13
       Extracellular matrix6.947.87E-13
       Host-virus interaction5.422.34E-12
       Cytoplasm1.807.63E-12
       Acetylation1.942.24E-10
       Polymorphism1.307.12E-10
       Signal1.797.28E-10
       LIM domain12.432.92E-09
       Laminin EGF-like domain21.655.55E-09
       Glycoprotein1.691.00E-08
       Disease mutation2.021.28E-08
       Angiogenesis8.291.45E-08
       Immunity3.881.74E-08
       Disulfide bond1.761.79E-07
       Isomerase7.337.51E-07
       Alternative splicing1.288.52E-07
       Vitamin C22.375.40E-06
       Osteogenesis imperfecta21.317.01E-06
       Innate immunity4.291.20E-05
       Actin-binding4.082.07E-05
       Redox-active center10.654.65E-05
       Cell junction2.541.15E-04
       Secreted1.781.22E-04
       Dioxygenase7.191.23E-04
       Heparin-binding6.781.78E-04
       Coiled coil1.571.94E-04
       Host cell receptor for virus entry8.162.11E-04
       MHC I29.832.64E-04
       Metal-binding1.474.41E-04
       Hydrolase1.794.52E-04
       Oxidoreductase2.438.92E-04
       Cytoskeleton1.901.40E-03
       EGF-like domain3.451.42E-03
       Nucleotide-binding1.632.91E-03
       Cardiomyopathy5.524.61E-03
       Serine protease inhibitor5.205.93E-03
       TPR repeat3.577.30E-03
      Abundant in human choroidal endothelial cells
       Acetylation3.507.51E-44
       Phosphoprotein2.042.10E-36
       Cell cycle5.372.37E-16
       Chromosome6.842.37E-15
       DNA replication17.072.95E-15
       Cytoplasm2.021.81E-14
       Nucleus1.939.01E-14
       Isopeptide bond3.597.39E-13
       Mitosis7.575.79E-12
       Ubl conjugation2.822.12E-11
       Cell division5.842.53E-11
       ATP-binding2.655.59E-08
       Nucleotide-binding2.388.68E-08
       DNA repair5.489.53E-08
       Actin-binding5.512.35E-07
       DNA damage4.561.11E-06
       Cholesterol biosynthesis29.811.28E-06
       Steroid biosynthesis18.881.61E-06
       Sterol biosynthesis22.665.57E-06
       DNA condensation33.721.11E-05
       Centromere6.891.52E-05
       Helicase6.602.14E-05
       Lipid biosynthesis6.054.28E-05
       Coiled coil1.715.65E-05
       Methylation2.361.54E-04
       Alternative splicing1.233.16E-04
       Cholesterol metabolism9.943.34E-04
       Spliceosome5.954.03E-04
       Cytoskeleton2.164.20E-04
       mRNA splicing3.994.58E-04
       Sterol metabolism8.457.07E-04
       Polymorphism1.187.74E-04
       mRNA transport6.239.11E-04
       DNA recombination6.991.66E-03
       Steroid metabolism6.592.17E-03
       mRNA processing3.132.86E-03
       Calmodulin-binding4.355.58E-03
      Table presents UniProt keywords that are significantly enriched by EASE score, and survive Benjamin-Hochberg multiple correction testing.
      a EASE score is generated by enrichment analysis in DAVID (version 6.8) and refers to uncorrected P value.
      Table 4Gene Ontology Processes Enriched in Proteins That Are Differentially Abundant in Human Ocular Endothelial Cells
      Gene Ontology ProcessTermFold EnrichmentEASE Score
      EASE score is generated by enrichment analysis in DAVID (version 6.8) and refers to uncorrected P value.
      Abundant in human retinal endothelial cells
       Extracellular matrix organizationGO:00301989.124.17E-18
       Type I interferon signaling pathwayGO:006033716.961.98E-15
       Interferon-gamma-mediated signaling pathwayGO:006033314.392.22E-13
       AngiogenesisGO:00015256.878.09E-13
       Cell adhesionGO:00071554.591.28E-12
       Defense response to virusGO:00516076.978.71E-10
       Antigen processing and presentation of peptide antigen via MHC class IGO:000247419.151.37E-08
       Proteolysis involved in cellular protein catabolic processGO:005160313.304.71E-08
       Negative regulation of viral genome replicationGO:004507114.371.58E-07
       Leukocyte migrationGO:00509006.804.59E-07
       Antigen processing and presentation of endogenous peptide antigen via MHC class IGO:001988545.611.95E-06
       Response to endoplasmic reticulum stressGO:00349768.512.44E-06
       Endodermal cell differentiationGO:003598716.553.06E-06
       Platelet degranulationGO:00025766.824.75E-06
       Response to hypoxiaGO:00016664.831.68E-05
       Regulation of cell migrationGO:00303347.771.95E-05
       Cell-cell adhesionGO:00986093.772.43E-05
       Substrate adhesion-dependent cell spreadingGO:003444611.762.47E-05
       IRE1-mediated unfolded protein responseGO:00364988.663.55E-05
       Leukocyte cell-cell adhesionGO:000715915.323.66E-05
       Negative regulation of angiogenesisGO:00165258.244.91E-05
       Antigen processing and presentation of exogenous peptide antigen via MHC class I, TAP-dependentGO:00024798.115.46E-05
       Response to virusGO:00096155.805.62E-05
       Cell adhesion mediated by integrinGO:003362721.286.90E-05
       Protein foldingGO:00064574.261.25E-04
       Positive regulation of cell migrationGO:00303354.161.52E-04
       Integrin-mediated signaling pathwayGO:00072295.801.59E-04
       Extracellular matrix disassemblyGO:00226176.721.82E-04
       Single organismal cell-cell adhesionGO:00163375.691.83E-04
       Viral entry into host cellGO:00467186.382.50E-04
       Antigen processing and presentation of exogenous peptide antigen via MHC class I, TAP-independentGO:000248028.382.94E-04
       Collagen catabolic processGO:00305746.984.79E-04
       Negative regulation of endopeptidase activityGO:00109514.756.22E-04
       Negative regulation of cell migrationGO:00303365.387.14E-04
       Antigen processing and presentation of endogenous peptide antigen via MHC class I via ER pathway, TAP-independentGO:000248663.857.20E-04
       ER to Golgi vesicle-mediated transportGO:00068883.999.30E-04
       Protein folding in endoplasmic reticulumGO:003497519.659.56E-04
       Cytoskeleton organizationGO:00070103.979.72E-04
      Abundant in human choroidal endothelial cells
       DNA replicationGO:00062609.768.92E-13
       Cell divisionGO:00513015.681.36E-11
       G1/S transition of mitotic cell cycleGO:000008211.703.35E-11
       Protein sumoylationGO:00169259.522.49E-09
       Sister chromatid cohesionGO:00070628.506.48E-07
       Mitotic nuclear divisionGO:00070674.813.00E-06
       DNA unwinding involved in DNA replicationGO:000626839.794.71E-06
       Regulation of transcription involved in G1/S transition of mitotic cell cycleGO:000008320.768.17E-06
       DNA repairGO:00062814.748.44E-06
       Isoprenoid biosynthetic processGO:000829928.422.16E-05
       Mitotic chromosome condensationGO:000707626.532.91E-05
       Telomere maintenance via recombinationGO:000072214.924.46E-05
       DNA replication initiationGO:000627014.924.46E-05
       DNA damage response, detection of DNA damageGO:004276913.268.02E-05
       Nucleotide-excision repair, DNA incision, 5'-to lesionGO:000629612.919.18E-05
       Cholesterol biosynthetic processGO:000669512.571.05E-04
       Nucleotide-excision repair, DNA incisionGO:003368312.571.05E-04
       Cell-cell adhesionGO:00986093.821.61E-04
       Double-strand break repairGO:00063028.441.73E-04
       Nucleotide-excision repair, DNA gap fillingGO:000629716.582.08E-04
       Mitotic nuclear envelope disassemblyGO:000707710.852.13E-04
       Nucleotide-excision repair, preincision complex assemblyGO:000629413.724.42E-04
       Cell proliferationGO:00082833.047.22E-04
      Table presents gene ontology terms that are significantly enriched by EASE score, and survive Benjamin-Hochberg multiple correction testing.
      a EASE score is generated by enrichment analysis in DAVID (version 6.8) and refers to uncorrected P value.
      Table 5Pathways Enriched in Proteins That Are Differentially Abundant in Human Ocular Endothelial Cells
      PathwayTermFold EnrichmentEASE Score
      EASE score is generated by enrichment analysis in DAVID (version 6.8) and refers to uncorrected P value.
      Abundant in human retinal endothelial cells
       Interferon alpha/beta signalingR-HSA-90973312.581.63E-13
       Interferon gamma signalingR-HSA-8773009.021.69E-10
       Non-integrin membrane-ECM interactionsR-HSA-300017114.882.18E-10
       Antigen presentation: folding, assembly, and peptide loading of class I MHCR-HSA-98317019.071.00E-09
       Integrin cell surface interactionsR-HSA-2160838.651.33E-09
       ECM proteoglycansR-HSA-30001789.262.37E-09
       Laminin interactionsR-HSA-300015716.534.29E-09
       Collagen biosynthesis and modifying enzymesR-HSA-16508148.887.56E-08
       Platelet degranulationR-HSA-1146085.341.87E-06
       Degradation of the extracellular matrixR-HSA-14742286.993.18E-06
       ER-phagosome pathwayR-HSA-12369745.904.35E-05
       Endosomal/vacuolar pathwayR-HSA-123697720.666.83E-05
       Molecules associated with elastic fibersR-HSA-21293799.149.70E-05
       Elastic fiber formationR-HSA-156694813.783.84E-04
       ISG15 antiviral mechanismR-HSA-11694085.436.21E-04
       XBP1(S) activates chaperone genesR-HSA-3810386.208.46E-04
       PECAM1 interactionsR-HSA-21099016.531.53E-03
      Abundant in human choroidal endothelial cells
       Switching of origins to a postreplicative stateR-HSA-6905257.808.33E-09
       Activation of ATR in response to replication stressR-HSA-17618714.061.61E-07
       Condensation of prometaphase chromosomesR-HSA-251485331.535.98E-07
       SUMOylation of DNA damage response and repair proteinsR-HSA-31082148.266.89E-07
       Cholesterol biosynthesisR-HSA-19127320.237.44E-07
       Activation of the prereplicative complexR-HSA-6896214.459.09E-07
       Unwinding of DNAR-HSA-17697428.901.01E-06
       Assembly of the prereplicative complexR-HSA-6886723.123.68E-06
       Activation of gene expression by SREBF (SREBP)R-HSA-242616811.016.31E-06
       Resolution of sister chromatid cohesionR-HSA-25002575.736.84E-06
       Removal of licensing factors from originsR-HSA-6930020.407.37E-06
       PCNA-dependent long patch base excision repairR-HSA-565180116.522.29E-05
       Mismatch repair (MMR) directed by MSH2:MSH6 (MutSalpha)R-HSA-535856519.279.89E-05
       E2F mediated regulation of DNA replicationR-HSA-11351017.001.68E-04
       G1/S-specific transcriptionR-HSA-6920517.001.68E-04
       Meiotic synapsisR-HSA-12216325.853.99E-04
       NoRC negatively regulates rRNA expressionR-HSA-4274134.914.62E-04
       Cohesin loading onto chromatinR-HSA-247094623.125.47E-04
       Removal of the flap intermediate from the C-strandR-HSA-17443723.125.47E-04
       Dual incision in GG-NERR-HSA-56964008.466.40E-04
       Establishment of sister chromatid cohesionR-HSA-246805221.027.43E-04
       Gap-filling DNA repair synthesis and ligation in GG-NERR-HSA-569639711.568.02E-04
       SUMOylation of DNA replication proteinsR-HSA-46158857.719.88E-04
       Separation of sister chromatidsR-HSA-24678133.441.30E-03
       Orc1 removal from chromatinR-HSA-689495.701.34E-03
       Removal of the flap intermediateR-HSA-6916616.521.58E-03
       Recognition of DNA damage by PCNA-containing replication complexR-HSA-1103149.631.63E-03
       Condensation of prophase chromosomesR-HSA-22997185.471.66E-03
       Mismatch repair (MMR) directed by MSH2:MSH3 (MutSbeta)R-HSA-535860615.411.95E-03
       Translesion synthesis by REV1R-HSA-11031214.452.37E-03
       Translesion synthesis by POLIR-HSA-565612113.602.84E-03
       Translesion synthesis by POLKR-HSA-565586213.602.84E-03
       Nuclear pore complex (NPC) disassemblyR-HSA-33018548.262.91E-03
       NS1 mediated effects on host pathwaysR-HSA-1682767.813.57E-03
       Translesion synthesis by POLHR-HSA-11032012.173.94E-03
       B-WICH complex positively regulates rRNA expressionR-HSA-52509244.454.70E-03
       Gap-filling DNA repair synthesis and ligation in TC-NERR-HSA-67822105.345.07E-03
      Table presents Reactome pathways that are significantly enriched by EASE score, and survive Benjamin-Hochberg multiple correction testing.
      a EASE score is generated by enrichment analysis in DAVID (version 6.8) and refers to uncorrected P value.

      Discussion

      We have used shotgun proteomics to profile the proteins expressed by human retinal and choroidal vascular endothelial cells, which were separately isolated from 5 pairs of eyes. In contrast to previous work by ourselves
      • Bharadwaj A.S.
      • Appukuttan B.
      • Wilmarth P.A.
      • et al.
      Role of the retinal vascular endothelial cell in ocular disease.
      • Smith J.R.
      • Choi D.
      • Chipps T.J.
      • et al.
      Unique gene expression profiles of donor-matched human retinal and choroidal vascular endothelial cells.
      • Choi D.
      • Appukuttan B.
      • Binek S.J.
      • et al.
      Prediction of cis-regulatory elements controlling genes differentially expressed by retinal and choroidal vascular endothelial cells.
      • Zamora D.O.
      • Riviere M.
      • Choi D.
      • et al.
      Proteomic profiling of human retinal and choroidal endothelial cells reveals molecular heterogeneity related to tissue of origin.
      and independent groups using gene expression microarray, PCR array, and/or protein array,
      • Browning A.C.
      • Halligan E.P.
      • Stewart E.A.
      • et al.
      Comparative gene expression profiling of human umbilical vein endothelial cells and ocular vascular endothelial cells.
      • Mammadzada P.
      • Gudmundsson J.
      • Kvanta A.
      • Andre H.
      Differential hypoxic response of human choroidal and retinal endothelial cells proposes tissue heterogeneity of ocular angiogenesis.
      • Saker S.
      • Stewart E.A.
      • Browning A.C.
      • Allen C.L.
      • Amoaku W.M.
      The effect of hyperglycaemia on permeability and the expression of junctional complex molecules in human retinal and choroidal endothelial cells.
      this study is the first investigation to take a comprehensive or “deep” discovery approach
      • Richards A.L.
      • Merrill A.E.
      • Coon J.J.
      Proteome sequencing goes deep.
      in seeking to define the molecular phenotype of human ocular endothelial cells. We identified 5042 nonredundant proteins expressed by one or both endothelial cell subpopulations. Although no other team has reported a shotgun proteomics analysis of human ocular endothelial cells, the proteomes of human extraocular endothelial cell subtypes have been described, including umbilical vein,
      • Gautier V.
      • Mouton-Barbosa E.
      • Bouyssie D.
      • et al.
      Label-free quantification and shotgun analysis of complex proteomes by one-dimensional SDS-PAGE/NanoLC-MS: evaluation for the large scale analysis of inflammatory human endothelial cells.
      • Jorge I.
      • Navarro P.
      • Martinez-Acedo P.
      • et al.
      Statistical model to analyze quantitative proteomics data obtained by 18O/16O labeling and linear ion trap mass spectrometry: application to the study of vascular endothelial growth factor-induced angiogenesis in endothelial cells.
      yielding a similar number of protein identifications. Of the total of 5042 proteins, 3454 proteins had sufficiently high mean spectral counts to be included in a differential expression analysis. The majority of these 3454 proteins were expressed at similar levels by human retinal and choroidal human endothelial cells; however, 498 proteins (14.4%) were differentially expressed between subpopulations, applying the standard FDR of 0.05. Enrichment analyses showed that the list of proteins enriched in human retinal endothelial cells included groups of molecules involved in the regulation of angiogenesis, and in innate and adaptive immune responses, which are processes directly relevant to the development of retinal ischemic vasculopathies and posterior uveitis. Proteins that were enriched in human choroidal endothelial cells also included molecules that regulate angiogenesis and thus may participate in processes that control the onset and/or progression of neovascular AMD.

       Molecular Profiling by Shotgun Proteomics

      Methodologies and bioinformatics tools that have been implemented in proteomics over the past 10 years provide an unprecedented capacity to realize the field's ultimate goal of “characterizing the entire protein content present in a cell, tissue, or bodily fluid at a given point in time.”
      • Gillet L.C.
      • Leitner A.
      • Aebersold R.
      Mass spectrometry applied to bottom-up proteomics: entering the high-throughput era for hypothesis testing.
      We employed liquid chromatography–tandem mass spectrometry and took a shotgun approach for the purpose of characterizing human ocular endothelial cell proteomes. The shotgun—also known as “bottom-up”—approach to protein discovery involves the specific identification of peptides present in digested biological samples, followed by protein inference by extrapolation from peptide sequences to protein identities.
      • Zhang Y.
      • Fonslow B.R.
      • Shan B.
      • Baek M.C.
      • Yates 3rd, J.R.
      Protein analysis by shotgun/bottom-up proteomics.
      An alternative “top-down” strategy refers to direct identification of intact proteins. Although assumption is not involved in the latter technique, various technical issues related to working with longer amino acid chains presently limit its scope for discovery. For this reason, deep proteomics is almost always shotgun in nature. Other considerations in undertaking a proteomic profiling analysis are the methods to accurately define the proteome and to compare the abundance of individual proteins.
      In shotgun proteomics, identification of proteins is limited primarily by the protein database that one selects. To identify the maximum number of proteins, we used the UniProt human reference proteome, UP0000005640 (combination of Swiss-Prot—manually curated—and TrEMBL—computer-annotated—proteins), which held over 90 000 sequences at the time of our analysis. In comparison, the Swiss-Prot database, which contains reviewed canonical sequences only, held approximately 20 000 sequences. The stringent Proteomic Analysis Workflow pipeline
      • Wilmarth P.A.
      • Riviere M.A.
      • David L.L.
      Techniques for accurate protein identification in shotgun proteomic studies of human, mouse, bovine, and chicken lenses.
      was employed to control errors in peptide spectral matching, with approximately one third of spectra being matched to peptides. For high accuracy of protein identification, positive identification required the presence of at least 2 unique peptides per protein in each biological sample, and parsimony processing assigned overlapping peptide sets to single proteins. By comparison of matches for actual protein sequences vs sequence-reversed decoy sequences, and application of an experiment-wide protein score heuristic, the FDR for protein identification was set to just 0.01.
      To identify proteins that were differentially abundant in human retinal vs choroidal endothelial cells, it was first necessary to measure the level of expression of all proteins. In quantification, redundancy poses a challenge, and for that reason we used the Swiss-Prot database for this aspect of the work. We used spectral counting, which is a simple but robust method; within a complex sample, higher-abundance proteins produce more peptides and consequently a larger number of mass spectra, and the number of mass spectra assigned to a protein is directly related to abundance in the sample.
      • Liu H.
      • Sadygov R.G.
      • Yates 3rd, J.R.
      A model for random sampling and estimation of relative protein abundance in shotgun proteomics.
      A potential complication in this type of comparative analysis is missing data points. Many protein identifications in large-scale experiments have small spectral counts and large fractions of missing data points. Consistent identification becomes likely once abundance rises above the mass spectrometry detection threshold, which is typically a spectral count of 2.
      • Gupta N.
      • Pevzner P.A.
      False discovery rates of protein identifications: a strike against the two-peptide rule.
      Instead of requiring a missing data threshold (eg, protein detected in at least 4 of 5 samples in each cell type), we required a minimum average spectral count, with the average calculated across all 10 samples. This was more tolerant of a protein present in 1 cell type, but absent in the other cell type. We used a mean spectral count minimum of 2.5, just above the detection threshold of 2. Of the 3454 proteins exceeding this minimum, 2926 proteins were detected in all 10 samples, and 97.5% of the proteins had 2 or fewer missing data points.

       Molecular Heterogeneity of Human Ocular Vascular Endothelial Cells

      Our observations demonstrate that human ocular endothelial diversity is manifest at a protein level, which has immediate relevance for physiology and pathology of the human eye. We first described the molecular heterogeneity of human retinal and choroidal endothelial cells in a study that used gene expression microarray to define molecular phenotypes of multiple cell isolates at the transcript level.
      • Smith J.R.
      • Choi D.
      • Chipps T.J.
      • et al.
      Unique gene expression profiles of donor-matched human retinal and choroidal vascular endothelial cells.
      • Choi D.
      • Appukuttan B.
      • Binek S.J.
      • et al.
      Prediction of cis-regulatory elements controlling genes differentially expressed by retinal and choroidal vascular endothelial cells.
      Our finding of human retinal vs choroidal endothelial transcriptomic diversity across humans was subsequently replicated by an independent group led by Amoaku,
      • Browning A.C.
      • Halligan E.P.
      • Stewart E.A.
      • et al.
      Comparative gene expression profiling of human umbilical vein endothelial cells and ocular vascular endothelial cells.
      who additionally differentiated retinal and choroidal endothelial cell transcripts from those expressed by iris and umbilical vein endothelial cells. We have reported differences in the transcriptomic responses of human retinal vs choroidal endothelial cells to inflammatory stimuli, including lipopolysaccharide,
      • Smith J.R.
      • Choi D.
      • Chipps T.J.
      • et al.
      Unique gene expression profiles of donor-matched human retinal and choroidal vascular endothelial cells.
      and different responses following exposure to high glucose conditions are described by Amoaku's team.
      • Saker S.
      • Stewart E.A.
      • Browning A.C.
      • Allen C.L.
      • Amoaku W.M.
      The effect of hyperglycaemia on permeability and the expression of junctional complex molecules in human retinal and choroidal endothelial cells.
      Mammadzada and associates
      • Mammadzada P.
      • Gudmundsson J.
      • Kvanta A.
      • Andre H.
      Differential hypoxic response of human choroidal and retinal endothelial cells proposes tissue heterogeneity of ocular angiogenesis.
      performed directed, small-scale profiling using “human endothelial cell biology” and “human angiogenesis” PCR arrays to study 133 transcripts, and a “human angiogenesis” antibody array to study 55 proteins from retinal and choroidal endothelial cells of a single human donor under normoxia and hypoxia, also showing differential responses.
      An interesting comparison is the molecular diversity of human retinal and choroidal endothelial cells according to our previous transcriptomic study
      • Smith J.R.
      • Choi D.
      • Chipps T.J.
      • et al.
      Unique gene expression profiles of donor-matched human retinal and choroidal vascular endothelial cells.
      vs this new proteomic study. Overall, approximately one quarter of the molecules that were differentially expressed at a protein level were identified as differentially expressed at a transcript level also: 140 of 626 proteins (22.4%). These proteins were split evenly between retinal and choroidal endothelial groups: 74 of 342 highly expressed retinal endothelial proteins (21.6%) and 66 of 285 highly expressed choroidal endothelial proteins (23.1%). Differences in the identifications may be explained on methodological or biological grounds. Cell isolation and culture techniques did not vary between studies, but sample processing was necessarily different. Arguably of greater importance, the transcriptomic study was performed by gene expression microarray, which is a directed form of profiling; in comparison, the shotgun proteomics study used a discovery approach. Data analysis also differed, in part directed by the need to identify molecules in the discovery-driven study. It is tempting to speculate that concordance would have been higher had the transcriptome been generated by RNA sequencing, given that this technique takes a discovery approach.
      • Wang Z.
      • Gerstein M.
      • Snyder M.
      RNA-Seq: a revolutionary tool for transcriptomics.
      However, some discordance between transcriptomic and proteomic data sets is likely to reflect cell biology, including rates of generation and degradation of mRNA or protein and/or regulation during and after translation.
      • Cox B.
      • Kislinger T.
      • Emili A.
      Integrating gene and protein expression data: pattern analysis and profile mining.
      While concordance for specific molecules was relatively low, results of enrichment analyses were similar: human retinal and choroidal endothelial cells were enriched in transcripts and proteins involved in processes of angiogenesis, and retinal endothelial cells were enriched in transcripts and proteins involved in immune responses.
      There are strengths and limitations to the cultured ocular endothelial cell model that we and other groups use. Differences in the morphology and molecular composition of endothelial cells between species and across vascular beds
      • Aird W.C.
      Endothelial cell heterogeneity.
      underpins the importance of studies of human ocular endothelial cells, prior to or as an essential complement to in vivo animal studies of normal physiology and pathology of the posterior eye. However, although it is unknown to what degree ocular endothelial cells change as a result of culture in isolation from their microenvironment, this concern is real and relates to any work using cultured cells. The phenomenon has been well illustrated in multiple rodent studies that would be difficult to conduct on healthy humans, in which gene expression of cultured endothelial cells from a range of extraocular tissues was found to vary from that of in vivo or ex vivo endothelial cell preparations.
      • Calabria A.R.
      • Shusta E.V.
      A genomic comparison of in vivo and in vitro brain microvascular endothelial cells.
      • Durr E.
      • Yu J.
      • Krasinska K.M.
      • et al.
      Direct proteomic mapping of the lung microvascular endothelial cell surface in vivo and in cell culture.
      • Geraud C.
      • Schledzewski K.
      • Demory A.
      • et al.
      Liver sinusoidal endothelium: a microenvironment-dependent differentiation program in rat including the novel junctional protein liver endothelial differentiation-associated protein-1.
      These studies also showed that culture resulted in drift toward a common phenotype, which is reassuring in the sense that we are likely to encounter false-negative protein identifications, rather than false-positive identifications. It is clear that primary endothelial cells retain a variety of endothelial characteristics in culture, including cobblestone morphology; constitutive expression of endothelial markers, including von Willebrand factor and CD31; induced expression of cell adhesion molecules; and formation of capillary tubes on Matrigel.
      • Unger R.E.
      • Krump-Konvalinkova V.
      • Peters K.
      • Kirkpatrick C.J.
      In vitro expression of the endothelial phenotype: comparative study of primary isolated cells and cell lines, including the novel cell line HPMEC-ST1.6R.
      The endothelial cells we isolate exhibit all these features,
      • Bharadwaj A.S.
      • Appukuttan B.
      • Wilmarth P.A.
      • et al.
      Role of the retinal vascular endothelial cell in ocular disease.
      and to reduce the possibility of phenotypic drift, we use cells in early passage. In addition, we study multiple endothelial cell isolates. Most research on endothelial cells is conducted using isolates from a single donor or pooled from several donors. Yet, we have observed distinct expression profiles across retinal and choroidal endothelial cell isolates from different donors.
      • Smith J.R.
      • Choi D.
      • Chipps T.J.
      • et al.
      Unique gene expression profiles of donor-matched human retinal and choroidal vascular endothelial cells.
      The paired design—directly comparing retinal and choroidal endothelial cells isolated from the same human eye pairs—addresses the concern of inter-individual variation.

       Molecular Phenotype of Human Retinal and Choroidal Vascular Endothelial Cells

      Our in silico analysis indicates that human retinal and choroidal vascular endothelial cell proteomes are enriched in proteins with angiogenic regulatory properties. Certain proteins, including the potent ocular angiogenic promoter, VEGF,
      • Miller J.W.
      • Le Couter J.
      • Strauss E.C.
      • Ferrara N.
      Vascular endothelial growth factor a in intraocular vascular disease.
      and its receptors, are present at comparable levels in both retinal and choroidal endothelial cells. The finding that some other proteins are differentially expressed between these cell populations supports the hypothesis that there are differences in the molecular regulation of angiogenesis in the retinal and choroidal vascular beds. The implication of the observation is that differentially expressed proangiogenic proteins may be targets for new biologic drugs, while antiangiogenic proteins have potential for therapeutic use, in retinal vs choroidal neovascularization and/or vascular leakage. Of particular interest are those proteins that have not been identified in previous ocular endothelial profiling studies, conducted in a targeted manner. Although it is clearly outside the scope of this thesis to discuss each novel protein, some examples selected from the list of high-differential-expression proteins illustrate the potential implications of our work.
      Proteins with potential to regulate angiogenesis that are identified for the first time in relatively high abundance in human retinal endothelial cells are thrombospondin type-I domain-containing protein 4 (THSD4, approximately 60-fold difference), netrin-4 (NET4, approximately 2.5-fold difference), and testin (TES, approximately 1.5-fold difference). As a member of the ADAMTS (“a disintegrin-like and metalloprotease with thrombospondin type I motif”) superfamily, THSD4—also termed ADAMSL6—is a secreted protein involved in extracellular matrix homeostasis, including the interaction between the matrix and cells.
      • Dubail J.
      • Apte S.S.
      Insights on ADAMTS proteases and ADAMTS-like proteins from mammalian genetics.
      Turnover of the basement membrane occurs as a blood vessel grows. First described in 2010,
      • Tsutsui K.
      • Manabe R.
      • Yamada T.
      • et al.
      ADAMTSL-6 is a novel extracellular matrix protein that binds to fibrillin-1 and promotes fibrillin-1 fibril formation.
      THSD4 has not yet been investigated in relation to angiogenesis, but because it is a molecule that promotes microfibril assembly, it is highly likely that the protein promotes this process. Netrins are secreted proteins that promote the formation of neuronal networks and the vasculature during development.
      • Larrieu-Lahargue F.
      • Thomas K.R.
      • Li D.Y.
      Netrin ligands and receptors: lessons from neurons to the endothelium.
      Netrin-4 has been localized to the retina in the mouse, and NET4 gene–deficient mice have been employed to evaluate the role of NET4 in experimental retinal and choroidal neovascularization, that is, oxygen-induced retinopathy and laser-induced choroidal neovascularization. A NET4 deficiency results in faster revascularization of the retina after hypoxia in oxygen-induced retinopathy, but has no effect on laser-induced choroidal neovascularization; this observation has been interpreted as indicating a role for NET4 in protecting the eye from hypoxic, as opposed to inflammatory, insult.
      • Kociok N.
      • Crespo-Garcia S.
      • Liang Y.
      • et al.
      Lack of netrin-4 modulates pathologic neovascularization in the eye.
      Our data provide support for an alternate explanation: NET4 may participate angiogenesis that involves the retinal endothelial cell, but not the choroidal endothelial cell. While not extensively studied to date, TES is a cytoskeleton protein that participates in cell-cell adhesion.
      • Li Y.N.
      • Pinzon-Duarte G.
      • Dattilo M.
      • Claudepierre T.
      • Koch M.
      • Brunken W.J.
      The expression and function of netrin-4 in murine ocular tissues.
      TES has been identified as a tumor suppressor gene in mice
      • Drusco A.
      • Zanesi N.
      • Roldo C.
      • et al.
      Knockout mice reveal a tumor suppressor function for Testin.
      and a prognostic marker in human carcinomas.
      • Ma H.
      • Weng D.
      • Chen Y.
      • et al.
      Extensive analysis of D7S486 in primary gastric cancer supports TESTIN as a candidate tumor suppressor gene.
      • Zhu J.
      • Li X.
      • Kong X.
      • et al.
      Testin is a tumor suppressor and prognostic marker in breast cancer.
      In an in vitro human breast cancer model, TES inhibits angiogenesis,
      • Zhu J.
      • Li X.
      • Kong X.
      • et al.
      Testin is a tumor suppressor and prognostic marker in breast cancer.
      implying the potential to function as an angiogenesis blocker in the human retina.
      Focusing on the regulation of angiogenesis in the choroid, human choroidal endothelial cells express high levels of actin-binding protein anillin (ANLN, approximately 50-fold difference); nesprin-3 (SYNE3, approximately 7-fold difference); and neuronal precursor cell–expressed developmentally downregulated NEDD4 (NEDD4, approximately 3-fold difference). The intracellular scaffold protein, anillin, plays a key role in cytokinesis, which is the final stage in cell division.
      • Piekny A.J.
      • Maddox A.S.
      The myriad roles of Anillin during cytokinesis.
      Since endothelial cell proliferation is a necessary component of angiogenesis, an obvious hypothesis is that anillin promotes choroidal angiogenesis. The nesprin family includes 4 large proteins that link nucleus and cytoskeleton, and participate in fundamental processes such as organelle positioning, cell division, and cell polarity and migration.
      • Cartwright S.
      • Karakesisoglou I.
      Nesprins in health and disease.
      Though SYNE3 has not been studied in relation to angiogenesis specifically, silencing expression in human aortic endothelial cells with small interfering RNA (siRNA) slows migration of these cells.
      • Morgan J.T.
      • Pfeiffer E.R.
      • Thirkill T.L.
      • et al.
      Nesprin-3 regulates endothelial cell morphology, perinuclear cytoskeletal architecture, and flow-induced polarization.
      Consistently, siRNA-mediated blockade of nesprin-1 or nesprin-2 decreases vascular loop formation in an in vitro assay of human umbilical vein endothelial cells.
      • King S.J.
      • Nowak K.
      • Suryavanshi N.
      • Holt I.
      • Shanahan C.M.
      • Ridley A.J.
      Nesprin-1 and nesprin-2 regulate endothelial cell shape and migration.
      Together, these observations suggest SYNE3 may act to promote blood vessel growth in the choroid. The NEDD4 protein is an E3 ubiquitin-protein ligase, and is thus involved in the ubiquitin-proteasome pathway that controls turnover of cellular proteins.
      • Ingham R.J.
      • Gish G.
      • Pawson T.
      The Nedd4 family of E3 ubiquitin ligases: functional diversity within a common modular architecture.
      Ubiquitination is a multi-step enzymatically controlled process that ultimately targets a protein for degradation in the proteome; E3 ubiquitin-protein ligases participate in the final stage of transfer of ubiquitin to a protein.
      • Schmidt M.
      • Finley D.
      Regulation of proteasome activity in health and disease.
      Involvement of NEDD4 in the ubiquitin-proteasome pathway suggests a potential role in choroidal angiogenesis, since human choroidal endothelial sprouting is potently inhibited by proteasome inhibitor epoxomicin.
      • Bargagna-Mohan P.
      • Ravindranath P.P.
      • Mohan R.
      Small molecule anti-angiogenic probes of the ubiquitin proteasome pathway: potential application to choroidal neovascularization.
      However, since the ubiquitin-proteasome pathway degrades many proteins, including those that promote angiogenesis, the impact of NEDD4 blockade is likely to be complex. Indeed, NEDD4 is implicated in the degradation of VEGF receptor 2, which suggests anti-angiogenic property.
      • Murdaca J.
      • Treins C.
      • Monthouel-Kartmann M.N.
      • et al.
      Grb10 prevents Nedd4-mediated vascular endothelial growth factor receptor-2 degradation.
      Yet separately, NEDD4 has been implicated in the p38 mitogen-activated protein kinase signaling pathway, which increases permeability of the vascular endothelium
      • Grimsey N.J.
      • Aguilar B.
      • Smith T.H.
      • et al.
      Ubiquitin plays an atypical role in GPCR-induced p38 MAP kinase activation on endosomes.
      ; this activity would be expected to promote vascular leakage in neovascular AMD.
      Further investigation of other proteins that are highly expressed in human retinal endothelial cells may prove informative of blood-retinal barrier function, including vascular endothelial permeability and leukocyte transendothelial migration. These proteins include selenoprotein M (SELM, approximately 35-fold increase); plasminogen activator inhibitor 2 (SERPINB2, approximately 20-fold difference); stimulator of interferon genes protein (hSTING, approximately 5-fold difference); and stabilin-1 (STAB1, approximately 1.5-fold difference). SELM and SERPINB2 are proteins that would be expected to strengthen the blood-retinal barrier, whereas hSTING and STAB1 are proteins that are likely to promote breakdown of this barrier. Selenium is a trace metal, incorporated into a family of selenoproteins that are involved in redox homeostasis.
      • Rayman M.P.
      Selenium and human health.
      • Gladyshev V.N.
      • Arner E.S.
      • Berry M.J.
      • et al.
      Selenoprotein gene nomenclature.
      Oxidative stress contributes to increased vascular permeability in ischemic retinal vasculopathy and posterior uveitis.
      • Kaur C.
      • Foulds W.S.
      • Ling E.A.
      Blood-retinal barrier in hypoxic ischaemic conditions: basic concepts, clinical features and management.
      • Yadav U.C.
      • Kalariya N.M.
      • Ramana K.V.
      Emerging role of antioxidants in the protection of uveitis complications.
      Given that SELM promotes the activity of antioxidant enzymes,
      • Hwang D.Y.
      • Sin J.S.
      • Kim M.S.
      • et al.
      Overexpression of human selenoprotein M differentially regulates the concentrations of antioxidants and H2O2, the activity of antioxidant enzymes, and the composition of white blood cells in a transgenic rat.
      this protein may participate in maintenance of the blood-retinal barrier. Although SERPINB2 functions primarily to inhibit plasminogen activation, the protein has potent anti-inflammatory effects including suppression of Th1 inflammatory cytokine expression by different leukocyte subsets,
      • Schroder W.A.
      • Le T.T.
      • Major L.
      • et al.
      A physiological function of inflammation-associated SerpinB2 is regulation of adaptive immunity.
      which also would limit breakdown of the blood-retinal barrier. As an adaptor molecule to the cytosolic DNA sensor known as cGAS, hSTING activates immune responses against DNA-containing pathogens, as well as self DNA.
      • Chen Q.
      • Sun L.
      • Chen Z.J.
      Regulation and function of the cGAS-STING pathway of cytosolic DNA sensing.
      The inflammatory cytokine cascade that characterizes these immune responses increases endothelial permeability and promotes transendothelial migration. We have previously identified several well-characterized leukocyte adhesion molecules at high levels on human retinal endothelium,
      • Bharadwaj A.S.
      • Appukuttan B.
      • Wilmarth P.A.
      • et al.
      Role of the retinal vascular endothelial cell in ocular disease.
      • Smith J.R.
      • Choi D.
      • Chipps T.J.
      • et al.
      Unique gene expression profiles of donor-matched human retinal and choroidal vascular endothelial cells.
      • Smith J.R.
      • Chipps T.J.
      • Ilias H.
      • Pan Y.
      • Appukuttan B.
      Expression and regulation of activated leukocyte cell adhesion molecule in human retinal vascular endothelial cells.
      and demonstrated the major role of ICAM-1 in particular, in the transendothelial migration of T cells and B cells into human retina that is essential for the development of noninfectious posterior uveitis.
      • Bharadwaj A.S.
      • Schewitz-Bowers L.P.
      • Wei L.
      • Lee R.W.
      • Smith J.R.
      Intercellular adhesion molecule 1 mediates migration of Th1 and Th17 cells across human retinal vascular endothelium.
      • Bharadwaj A.S.
      • Stempel A.J.
      • Olivas A.
      • et al.
      Molecular signals involved in human B cell migration into the retina: in vitro investigation of ICAM-1, VCAM-1, and CXCL13.
      This thesis work identifies STAB1 as another adhesion molecule that may be involved in retinal transendothelial migration of one or more leukocyte subsets. In nonocular tissues, transendothelial migration blockade of several leukocyte populations, including CD4+ and CD8+ helper T cells, B cells, monocytes, and neutrophils in mice, and regulatory T cells in humans, has been reported.
      • Karikoski M.
      • Irjala H.
      • Maksimow M.
      • et al.
      Clever-1/Stabilin-1 regulates lymphocyte migration within lymphatics and leukocyte entrance to sites of inflammation.
      • Shetty S.
      • Weston C.J.
      • Oo Y.H.
      • et al.
      Common lymphatic endothelial and vascular endothelial receptor-1 mediates the transmigration of regulatory T cells across human hepatic sinusoidal endothelium.
      There are multiple in vitro and ex vivo experimental methods that might be employed to evaluate the functional role of proteins in human retinal vs choroidal angiogenesis. Growth of blood vessels may be simulated in various in vitro assays. As one example, in the capillary-like tube assay, endothelial cells form tubules in 2 or 3 dimensions on manufactured extracellular matrix.
      • Aguilar E.
      • Dorrell M.I.
      • Friedlander D.
      • et al.
      Chapter 6. Ocular models of angiogenesis.
      As another example, in the bead assay, sprouts grow from microbeads coated with endothelial cells and embedded in a matrix.
      • Nakatsu M.N.
      • Hughes C.C.
      An optimized three-dimensional in vitro model for the analysis of angiogenesis.
      Growth is measured as variables such as length of tubes and number of branch points. An ex vivo assay uses intact tissue, dissected from the eye, which is also embedded in matrix and following for budding of microvessels.
      • Knott R.M.
      • Robertson M.
      • Muckersie E.
      • et al.
      A model system for the study of human retinal angiogenesis: activation of monocytes and endothelial cells and the association with the expression of the monocarboxylate transporter type 1 (MCT-1).
      Different aspects of blood vessel formation—endothelial cell proliferation, endothelial cell migration, and basement membrane degradation—may be interrogated separately. Proliferation of endothelial cells is readily quantified by counting, with automated methods usually requiring cells to be tagged for detection.
      • Pan Y.
      • Appukuttan B.
      • Mohs K.
      • Ashander L.M.
      • Smith J.R.
      Ubiquitin carboxyl-terminal esterase L1 promotes proliferation of human choroidal and retinal endothelial cells.
      Migration of endothelial cells across a plane surface or through a perforated membrane may be counted by microscopy.
      • Pan Y.
      • Appukuttan B.
      • Mohs K.
      • Ashander L.M.
      • Smith J.R.
      Ubiquitin carboxyl-terminal esterase L1 promotes proliferation of human choroidal and retinal endothelial cells.
      The most simple technique for studying the ability of endothelial cells to degrade extracellular matrix involves seeding cells on fluorescently labeled commercial matrix, and subsequently measuring areas of matrix degradation under fluorescence microscopy.
      • Bowden E.T.
      • Coopman P.J.
      • Mueller S.C.
      Invadopodia: unique methods for measurement of extracellular matrix degradation in vitro.
      Permeability of human retinal endothelium, or leukocyte transmigration across this endothelium, may be studied in vitro with the Boyden chamber assay. The assay uses a transwell, in which 2 chambers that are separated by a perforated membrane are coated with commercial extracellular matrix and seeded with endothelial cells. Diffusion of a high-molecular-weight tracker molecule across a confluent monolayer of cells indicates permeability of the endothelium.
      • Harhaj N.S.
      • Felinski E.A.
      • Wolpert E.B.
      • Sundstrom J.M.
      • Gardner T.W.
      • Antonetti D.A.
      VEGF activation of protein kinase C stimulates occludin phosphorylation and contributes to endothelial permeability.
      Transcellular electrical resistance provides an alternative measure of permeability, and today this also may be performed in real time over multiple days in an incubation chamber.
      • Srinivasan B.
      • Kolli A.R.
      • Esch M.B.
      • Abaci H.E.
      • Shuler M.L.
      • Hickman J.J.
      TEER measurement techniques for in vitro barrier model systems.
      Interactions between leukocytes and endothelial cells may be studied by different methodologies. The transwell Boyden chamber assay is the most commonly employed system. Leukocytes move from upper to lower chambers across the simulated endothelium.
      • Muller W.A.
      • Luscinskas F.W.
      Assays of transendothelial migration in vitro.
      At the end of the assay, leukocytes are collected from the lower well for counting and/or immunophenotyping. Leukocyte adhesion molecule expression on the endothelium may be influenced by wall shear stress, exerted by the flow of blood.
      • Morigi M.
      • Zoja C.
      • Figliuzzi M.
      • et al.
      Fluid shear stress modulates surface expression of adhesion molecules by endothelial cells.
      Flow chamber assays are an evolving technology that evaluates leukocyte binding to endothelial cells activated by wall shear stress. In the flow chamber, an endothelial monolayer is perfused with fluid alone, followed by leukocyte suspension, and leukocyte binding is commonly imaged in real time by phase-contrast microscopy, for quantification of leukocyte–endothelial cell interactions.
      • Muller W.A.
      • Luscinskas F.W.
      Assays of transendothelial migration in vitro.
      The Woodruff-Stamper assay addresses the concern that cell phenotype may change in culture. In this assay, leukocytes bind to endothelium of blood vessels in intact fresh frozen tissue section under flow conditions, and at the end of the assay leukocyte binding to intact endothelium is quantified under microscopy.
      • Hill T.A.
      • Stanford M.R.
      • Graham E.M.
      • Dumonde D.C.
      • Brown K.A.
      A new method for studying the selective adherence of blood lymphocytes to the microvasculature of human retina.
      Manipulations of the relevant experimental system may be implemented to test the specific involvement of candidate proteins in a disease process. Significant reduction in the outcome measure when a protein is blocked—and significant increase when it is augmented—in one endothelial cell population, but not the other population, would support the differential involvement of that protein in retinal vs choroidal vascular disease. Blockade may be achieved by antibody, small molecular inhibitor, and siRNA. Supplementation may be effected by recombinant protein or expression plasmid. As one example, to test the hypothesis that THSD4 specifically promotes human retinal angiogenesis, one would expect blockade to reduce the length of capillary-like tubules formed by retinal endothelial cells grown on extracellular matrix, and the number of microvessel buds from retinal explants. Blockade of THSD4 in retinal endothelial cells would also be expected to decrease the number of proliferated cells, the number of migrated cells, and the area of basement membrane defect per cell. Moreover, one would expect supplementing THSD4 to increase these same parameters of blood vessel growth. Conversely, in hypothesizing that the effect of THSD4 is retinal endothelial cell–specific, one would anticipate no differences between test and control conditions for choroidal endothelial cells in the same assays.

      Conclusions

      We have described comprehensive proteomes of the human retinal vascular endothelial cell and the human choroidal vascular endothelial cell. This work provides strong evidence that the protein phenotypes of these cells are unique, confirming a hypothesis of ocular endothelial cell molecular diversity that to date has been based on large data sets generated at the RNA level only. Both retinal and choroidal endothelial cell populations produce an abundance of proteins that participate in the regulation of angiogenesis, but differences in enriched proteins between cell populations suggest differences in the molecular regulation of proliferative retinal ischemic vasculopathies and neovascular AMD, respectively. Human retinal endothelial cells are also enriched in immunologic proteins, implying that this cell population participates in ocular immune privilege, and in uveitis when privilege is breached. Application of RNA sequencing and deeper proteomic technologies that allow differentiation of protein polymorphisms and/or post-translationally modified proteins may expand understanding of the molecular diversity of ocular endothelial cells in the future. At this time, however, our demonstration of enriched human retinal endothelial cell and human choroidal endothelial cell proteins provides a substantial list of candidates for further study as novel disease-directed biologic treatments or drug targets.
      Funding/Support: This work was supported by grant R01 EY019875 (Dr Smith) and grant P30 EY010572 (Dr David) from the National Institutes of Health, Bethesda, Maryland; and grant FT130101648 (Dr Smith) from the Australian Research Council, Canberra, Australia. Financial Disclosures: The following authors have no financial disclosures: Justine R. Smith, FRANZCO, PhD, Larry L. David, PhD, Binoy Appukuttan, PhD, and Phillip A. Wilmarth, PhD. All authors attest that they meet the current ICMJE criteria for authorship.
      Other Acknowledgments: The authors thank Mr Timothy Chipps (Oregon Health & Science University, Portland, Oregon) and Mrs Yuzhen Pan (Oregon Health & Science University, Portland, Oregon) for their support with preparation of endothelial cell samples, and Ms Kyra Patton (Oregon Health & Science University, Portland, Oregon) for her assistance with the rich protein annotation programming.

      References

        • Klein R.
        • Klein B.E.
        The prevalence of age-related eye diseases and visual impairment in aging: current estimates.
        Invest Ophthalmol Vis Sci. 2013; 54: ORSF5-ORSF13
        • Gritz D.C.
        • Wong I.G.
        Incidence and prevalence of uveitis in Northern California: the Northern California Epidemiology of Uveitis Study.
        Ophthalmology. 2004; 111: 491-500
        • Nussenblatt R.B.
        The natural history of uveitis.
        Int Ophthalmol. 1990; 14: 303-308
        • Lim L.S.
        • Mitchell P.
        • Seddon J.M.
        • Holz F.G.
        • Wong T.Y.
        Age-related macular degeneration.
        Lancet. 2012; 379: 1728-1738
        • Fisher D.E.
        • Klein B.E.
        • Wong T.Y.
        • et al.
        Incidence of age-related macular degeneration in a multi-ethnic United States population: the Multi-Ethnic Study of Atherosclerosis.
        Ophthalmology. 2016; 123: 1297-1308
        • Klein R.
        • Klein B.E.
        • Knudtson M.D.
        • Meuer S.M.
        • Swift M.
        • Gangnon R.E.
        Fifteen-year cumulative incidence of age-related macular degeneration: the Beaver Dam Eye Study.
        Ophthalmology. 2007; 114: 253-262
        • Klein R.
        • Wang Q.
        • Klein B.E.
        • Moss S.E.
        • Meuer S.M.
        The relationship of age-related maculopathy, cataract, and glaucoma to visual acuity.
        Invest Ophthalmol Vis Sci. 1995; 36: 182-191
        • Bressler N.M.
        • Doan Q.V.
        • Varma R.
        • et al.
        Estimated cases of legal blindness and visual impairment avoided using ranibizumab for choroidal neovascularization: non-Hispanic white population in the United States with age-related macular degeneration.
        Arch Ophthalmol. 2011; 129: 709-717
        • Campbell J.P.
        • Bressler S.B.
        • Bressler N.M.
        Impact of availability of anti-vascular endothelial growth factor therapy on visual impairment and blindness due to neovascular age-related macular degeneration.
        Arch Ophthalmol. 2012; 130: 794-795
        • Channa R.
        • Sophie R.
        • Bagheri S.
        • et al.
        Regression of choroidal neovascularization results in macular atrophy in anti-vascular endothelial growth factor-treated eyes.
        Am J Ophthalmol. 2015; 159 (e11–e12): 9-19
        • Grunwald J.E.
        • Pistilli M.
        • Ying G.S.
        • et al.
        Growth of geographic atrophy in the comparison of age-related macular degeneration treatments trials.
        Ophthalmology. 2015; 122: 809-816
        • Cheung N.
        • Mitchell P.
        • Wong T.Y.
        Diabetic retinopathy.
        Lancet. 2010; 376: 124-136
        • Zhang X.
        • Saaddine J.B.
        • Chou C.F.
        • et al.
        Prevalence of diabetic retinopathy in the United States, 2005-2008.
        JAMA. 2010; 304: 649-656
        • Klein R.
        • Lee K.E.
        • Gangnon R.E.
        • Klein B.E.
        The 25-year incidence of visual impairment in type 1 diabetes mellitus the Wisconsin Epidemiologic Study of Diabetic Retinopathy.
        Ophthalmology. 2010; 117: 63-70
        • Diabetic Retinopathy Clinical Research Network
        • Elman M.J.
        • Aiello L.P.
        • et al.
        Randomized trial evaluating ranibizumab plus prompt or deferred laser or triamcinolone plus prompt laser for diabetic macular edema.
        Ophthalmology. 2010; 117: 1064-1077
        • American Academy of Ophthalmology Retina/Vitreous Panel
        Preferred Practice Pattern Guidelines. Diabetic Retinopathy.
        American Academy of Ophthalmology, San Francisco, CA2016
        • Fong D.S.
        • Girach A.
        • Boney A.
        Visual side effects of successful scatter laser photocoagulation surgery for proliferative diabetic retinopathy: a literature review.
        Retina. 2007; 27: 816-824
        • Hartnett M.E.
        Pathophysiology and mechanisms of severe retinopathy of prematurity.
        Ophthalmology. 2015; 122: 200-210
        • Quinn G.E.
        • Barr C.
        • Bremer D.
        • et al.
        Changes in course of retinopathy of prematurity from 1986 to 2013: comparison of three studies in the United States.
        Ophthalmology. 2016; 123: 1595-1600
        • Martin J.A.
        • Hamilton B.E.
        • Osterman M.J.K.
        • Curtin S.C.
        • Mathews T.J.
        Births: final data for 2013.
        Natl Vital Stat Rep. 2015; 64: 1-65
        • Quinn G.E.
        • Dobson V.
        • Saigal S.
        • et al.
        Health-related quality of life at age 10 years in very low-birth-weight children with and without threshold retinopathy of prematurity.
        Arch Ophthalmol. 2004; 122: 1659-1666
        • McLoone E.
        • O’Keefe M.
        • McLoone S.
        • Lanigan B.
        Long term functional and structural outcomes of laser therapy for retinopathy of prematurity.
        Br J Ophthalmol. 2006; 90: 754-759
        • Sankar M.J.
        • Sankar J.
        • Mehta M.
        • Bhat V.
        • Srinivasan R.
        Anti-vascular endothelial growth factor (VEGF) drugs for treatment of retinopathy of prematurity.
        Cochrane Database Syst Rev. 2016; 2: CD009734
        • Suhler E.B.
        • Lloyd M.J.
        • Choi D.
        • Rosenbaum J.T.
        • Austin D.F.
        Incidence and prevalence of uveitis in Veterans Affairs Medical Centers of the Pacific Northwest.
        Am J Ophthalmol. 2008; 146: 890-896
        • Acharya N.R.
        • Tham V.M.
        • Esterberg E.
        • et al.
        Incidence and prevalence of uveitis: results from the Pacific Ocular Inflammation Study.
        JAMA Ophthalmol. 2013; 131: 1405-1412
        • Durrani O.M.
        • Tehrani N.N.
        • Marr J.E.
        • Moradi P.
        • Stavrou P.
        • Murray P.I.
        Degree, duration, and causes of visual loss in uveitis.
        Br J Ophthalmol. 2004; 88: 1159-1162
        • Multicenter Uveitis Steroid Treatment Trial Research Group
        • Kempen J.H.
        • Altaweel M.M.
        • Holbrook J.T.
        • Jabs D.A.
        • Sugar E.A.
        The multicenter uveitis steroid treatment trial: rationale, design, and baseline characteristics.
        Am J Ophthalmol. 2010; 149: 550-561
        • de Smet M.D.
        • Taylor S.R.
        • Bodaghi B.
        • et al.
        Understanding uveitis: the impact of research on visual outcomes.
        Prog Retin Eye Res. 2011; 30: 452-470
        • Rothova A.
        • Suttorp-van Schulten M.S.
        • Frits Treffers W.
        • Kijlstra A.
        Causes and frequency of blindness in patients with intraocular inflammatory disease.
        Br J Ophthalmol. 1996; 80: 332-336
        • Jabs D.A.
        • Rosenbaum J.T.
        • Foster C.S.
        • et al.
        Guidelines for the use of immunosuppressive drugs in patients with ocular inflammatory disorders: recommendations of an expert panel.
        Am J Ophthalmol. 2000; 130: 492-513
        • Gangaputra S.
        • Newcomb C.W.
        • Liesegang T.L.
        • et al.
        Methotrexate for ocular inflammatory diseases.
        Ophthalmology. 2009; 116: 2188-2198
        • Pasadhika S.
        • Kempen J.H.
        • Newcomb C.W.
        • et al.
        Azathioprine for ocular inflammatory diseases.
        Am J Ophthalmol. 2009; 148: 500-509
        • Daniel E.
        • Thorne J.E.
        • Newcomb C.W.
        • et al.
        Mycophenolate mofetil for ocular inflammation.
        Am J Ophthalmol. 2010; 149: 423-432
        • Kacmaz R.O.
        • Kempen J.H.
        • Newcomb C.
        • et al.
        Cyclosporine for ocular inflammatory diseases.
        Ophthalmology. 2010; 117: 576-584
        • Suhler E.B.
        • Smith J.R.
        • Wertheim M.S.
        • et al.
        A prospective trial of infliximab therapy for refractory uveitis: preliminary safety and efficacy outcomes.
        Arch Ophthalmol. 2005; 123: 903-912
        • Suhler E.B.
        • Smith J.R.
        • Giles T.R.
        • et al.
        Infliximab therapy for refractory uveitis: 2-year results of a prospective trial.
        Arch Ophthalmol. 2009; 127: 819-822
        • Campochiaro P.A.
        Molecular pathogenesis of retinal and choroidal vascular diseases.
        Prog Retin Eye Res. 2015; 49: 67-81
        • Perez V.L.
        • Caspi R.R.
        Immune mechanisms in inflammatory and degenerative eye disease.
        Trends Immunol. 2015; 36: 354-363
        • Gerhardt H.
        VEGF and endothelial guidance in angiogenic sprouting.
        Organogenesis. 2008; 4: 241-246
        • Linder S.
        The matrix corroded: podosomes and invadopodia in extracellular matrix degradation.
        Trends Cell Biol. 2007; 17: 107-117
        • Ding X.
        • Patel M.
        • Chan C.C.
        Molecular pathology of age-related macular degeneration.
        Prog Retin Eye Res. 2009; 28: 1-18
        • Antonetti D.A.
        • Klein R.
        • Gardner T.W.
        Diabetic retinopathy.
        New Engl J Med. 2012; 366: 1227-1239
        • Sapieha P.
        • Joyal J.S.
        • Rivera J.C.
        • et al.
        Retinopathy of prematurity: understanding ischemic retinal vasculopathies at an extreme of life.
        J Clin Invest. 2010; 120: 3022-3032
        • Cunha-Vaz J.G.
        The blood-retinal barriers system. Basic concepts and clinical evaluation.
        Exp Eye Res. 2004; 78: 715-721
        • Runkle E.A.
        • Antonetti D.A.
        The blood-retinal barrier: structure and functional significance.
        Methods Mol Biol. 2011; 686: 133-148
        • Titchenell P.M.
        • Antonetti D.A.
        Using the past to inform the future: anti-VEGF therapy as a road map to develop novel therapies for diabetic retinopathy.
        Diabetes. 2013; 62: 1808-1815
        • Klaassen I.
        • Van Noorden C.J.
        • Schlingemann R.O.
        Molecular basis of the inner blood-retinal barrier and its breakdown in diabetic macular edema and other pathological conditions.
        Prog Retin Eye Res. 2013; 34: 19-48
        • Van Ginderdeuren R.
        • Van Calster J.
        • Stalmans P.
        • Van den Oord J.
        A new and standardized method to sample and analyse vitreous samples by the Cellient automated cell block system.
        Acta Ophthalmol. 2014; 92: e388-e392
        • Parnaby-Price A.
        • Stanford M.R.
        • Biggerstaff J.
        • et al.
        Leukocyte trafficking in experimental autoimmune uveitis in vivo.
        J Leukoc Biol. 1998; 64: 434-440
        • Chen X.
        • Kezic J.M.
        • Forrester J.V.
        • et al.
        In vivo multi-modal imaging of experimental autoimmune uveoretinitis in transgenic reporter mice reveals the dynamic nature of inflammatory changes during disease progression.
        J Neuroinflammation. 2015; 12: 17
        • Levy-Clarke G.A.
        • Nussenblatt R.
        Retinal vasculitis.
        Int Ophthalmol Clin. 2005; 45: 99-113
        • Nourshargh S.
        • Alon R.
        Leukocyte migration into inflamed tissues.
        Immunity. 2014; 41: 694-707
        • Joussen A.M.
        • Murata T.
        • Tsujikawa A.
        • Kirchhof B.
        • Bursell S.E.
        • Adamis A.P.
        Leukocyte-mediated endothelial cell injury and death in the diabetic retina.
        Am J Pathol. 2001; 158: 147-152
        • Joussen A.M.
        • Poulaki V.
        • Le M.L.
        • et al.
        A central role for inflammation in the pathogenesis of diabetic retinopathy.
        FASEB J. 2004; 18: 1450-1452
        • Tian P.
        • Ge H.
        • Liu H.
        • et al.
        Leukocytes from diabetic patients kill retinal endothelial cells: effects of berberine.
        Mol Vis. 2013; 19: 2092-2105
        • Tsutsumi C.
        • Sonoda K.H.
        • Egashira K.
        • et al.
        The critical role of ocular-infiltrating macrophages in the development of choroidal neovascularization.
        J Leukoc Biol. 2003; 74: 25-32
        • Luhmann U.F.
        • Robbie S.
        • Munro P.M.
        • et al.
        The drusenlike phenotype in aging Ccl2-knockout mice is caused by an accelerated accumulation of swollen autofluorescent subretinal macrophages.
        Invest Ophthalmol Vis Sci. 2009; 50: 5934-5943
        • Krause T.A.
        • Alex A.F.
        • Engel D.R.
        • Kurts C.
        • Eter N.
        VEGF-production by CCR2-dependent macrophages contributes to laser-induced choroidal neovascularization.
        PLoS One. 2014; 9: e94313
        • Atkins G.B.
        • Jain M.K.
        • Hamik A.
        Endothelial differentiation: molecular mechanisms of specification and heterogeneity.
        Arterioscler Thromb Vasc Biol. 2011; 31: 1476-1484
        • Pasqualini R.
        • Millikan R.E.
        • Christianson D.R.
        • et al.
        Targeting the interleukin-11 receptor alpha in metastatic prostate cancer: a first-in-man study.
        Cancer. 2015; 121: 2411-2421
        • Bharadwaj A.S.
        • Schewitz-Bowers L.P.
        • Wei L.
        • Lee R.W.
        • Smith J.R.
        Intercellular adhesion molecule 1 mediates migration of Th1 and Th17 cells across human retinal vascular endothelium.
        Invest Ophthalmol Vis Sci. 2013; 54: 6917-6925
        • Bharadwaj A.S.
        • Stempel A.J.
        • Olivas A.
        • et al.
        Molecular signals involved in human B cell migration into the retina: in vitro investigation of ICAM-1, VCAM-1, and CXCL13.
        Ocul Immunol Inflamm. 2017; 25: 811-819
        • Bharadwaj A.S.
        • Appukuttan B.
        • Wilmarth P.A.
        • et al.
        Role of the retinal vascular endothelial cell in ocular disease.
        Prog Retin Eye Res. 2013; 32: 102-180
        • Smith J.R.
        • Choi D.
        • Chipps T.J.
        • et al.
        Unique gene expression profiles of donor-matched human retinal and choroidal vascular endothelial cells.
        Invest Ophthalmol Vis Sci. 2007; 48: 2676-2684
        • Choi D.
        • Appukuttan B.
        • Binek S.J.
        • et al.
        Prediction of cis-regulatory elements controlling genes differentially expressed by retinal and choroidal vascular endothelial cells.
        J Ocul Biol Dis Inform. 2008; 1: 37-45
        • Zamora D.O.
        • Riviere M.
        • Choi D.
        • et al.
        Proteomic profiling of human retinal and choroidal endothelial cells reveals molecular heterogeneity related to tissue of origin.
        Mol Vis. 2007; 13: 2058-2065
        • Chipps T.J.
        • Appukuttan B.
        • Pan Y.
        • Smith J.R.
        CD44 isoforms in human retinal and choroidal endothelial cells.
        Graefes Arch Clin Exp Ophthalmol. 2013; 251: 1245-1246
        • Smith J.R.
        • Chipps T.J.
        • Ilias H.
        • Pan Y.
        • Appukuttan B.
        Expression and regulation of activated leukocyte cell adhesion molecule in human retinal vascular endothelial cells.
        Exp Eye Res. 2012; 104: 89-93
        • Ahrens C.H.
        • Brunner E.
        • Qeli E.
        • Basler K.
        • Aebersold R.
        Generating and navigating proteome maps using mass spectrometry.
        Nat Rev Mol Cell Biol. 2010; 11: 789-801
        • Beck M.
        • Claassen M.
        • Aebersold R.
        Comprehensive proteomics.
        Curr Opin Biotechnol. 2011; 22: 3-8
        • Chambers M.C.
        • Maclean B.
        • Burke R.
        • et al.
        A cross-platform toolkit for mass spectrometry and proteomics.
        Nat Biotechnol. 2012; 30: 918-920
        • McDonald W.H.
        • Tabb D.L.
        • Sadygov R.G.
        • et al.
        MS1, MS2, and SQT - three unified, compact, and easily parsed file formats for the storage of shotgun proteomic spectra and identifications.
        Rapid Commun Mass Spectrom. 2004; 18: 2162-2168
        • Wilmarth P.A.
        • Riviere M.A.
        • David L.L.
        Techniques for accurate protein identification in shotgun proteomic studies of human, mouse, bovine, and chicken lenses.
        J Ocul Biol Dis Inform. 2009; 2: 223-234
        • Eng J.K.
        • Jahan T.A.
        • Hoopmann M.R.
        Comet: an open-source MS/MS sequence database search tool.
        Proteomics. 2013; 13: 22-24
        • Keller A.
        • Nesvizhskii A.I.
        • Kolker E.
        • Aebersold R.
        Empirical statistical model to estimate the accuracy of peptide identifications made by MS/MS and database search.
        Anal Chem. 2002; 74: 5383-5392
        • Nesvizhskii A.I.
        • Aebersold R.
        Interpretation of shotgun proteomic data: the protein inference problem.
        Mol Cell Proteomics. 2005; 4: 1419-1440
        • Elias J.E.
        • Gygi S.P.
        Target-decoy search strategy for increased confidence in large-scale protein identifications by mass spectrometry.
        Nat Methods. 2007; 4: 207-214
        • Zhang Y.
        • Xu T.
        • Shan B.
        • et al.
        ProteinInferencer: confident protein identification and multiple experiment comparison for large scale proteomics projects.
        J Proteomics. 2015; 129: 25-32
        • Kall L.
        • Storey J.D.
        • MacCoss M.J.
        • Noble W.S.
        Posterior error probabilities and false discovery rates: two sides of the same coin.
        J Proteome Res. 2008; 7: 40-44
        • Robinson M.D.
        • McCarthy D.J.
        • Smyth G.K.
        edgeR: a Bioconductor package for differential expression analysis of digital gene expression data.
        Bioinformatics. 2010; 26: 139-140
        • Robinson M.D.
        • Oshlack A.
        A scaling normalization method for differential expression analysis of RNA-seq data.
        Genome Biol. 2010; 11: R25
        • McCarthy D.J.
        • Chen Y.
        • Smyth G.K.
        Differential expression analysis of multifactor RNA-Seq experiments with respect to biological variation.
        Nucleic Acids Res. 2012; 40: 4288-4297
        • Huang da W.
        • Sherman B.T.
        • Lempicki R.A.
        Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources.
        Nat Protoc. 2009; 4: 44-57
        • Fabregat A.
        • Sidiropoulos K.
        • Garapati P.
        • et al.
        The Reactome pathway knowledgebase.
        Nucleic Acids Res. 2016; 44: D481-D487
        • Vizcaino J.A.
        • Csordas A.
        • del-Toro N.
        • et al.
        2016 update of the PRIDE database and its related tools.
        Nucleic Acids Res. 2016; 44: D447-D456
        • Browning A.C.
        • Halligan E.P.
        • Stewart E.A.
        • et al.
        Comparative gene expression profiling of human umbilical vein endothelial cells and ocular vascular endothelial cells.
        Br J Ophthalmol. 2012; 96: 128-132
        • Mammadzada P.
        • Gudmundsson J.
        • Kvanta A.
        • Andre H.
        Differential hypoxic response of human choroidal and retinal endothelial cells proposes tissue heterogeneity of ocular angiogenesis.
        Acta Ophthalmol. 2016; 94: 805-814
        • Saker S.
        • Stewart E.A.
        • Browning A.C.
        • Allen C.L.
        • Amoaku W.M.
        The effect of hyperglycaemia on permeability and the expression of junctional complex molecules in human retinal and choroidal endothelial cells.
        Exp Eye Res. 2014; 121: 161-167
        • Richards A.L.
        • Merrill A.E.
        • Coon J.J.
        Proteome sequencing goes deep.
        Curr Opin Chem Biol. 2015; 24: 11-17
        • Gautier V.
        • Mouton-Barbosa E.
        • Bouyssie D.
        • et al.
        Label-free quantification and shotgun analysis of complex proteomes by one-dimensional SDS-PAGE/NanoLC-MS: evaluation for the large scale analysis of inflammatory human endothelial cells.
        Mol Cell Proteomics. 2012; 11: 527-539
        • Jorge I.
        • Navarro P.
        • Martinez-Acedo P.
        • et al.
        Statistical model to analyze quantitative proteomics data obtained by 18O/16O labeling and linear ion trap mass spectrometry: application to the study of vascular endothelial growth factor-induced angiogenesis in endothelial cells.
        Mol Cell Proteomics. 2009; 8: 1130-1149
        • Gillet L.C.
        • Leitner A.
        • Aebersold R.
        Mass spectrometry applied to bottom-up proteomics: entering the high-throughput era for hypothesis testing.
        Annu Rev Anal Chem (Palo Alto Calif). 2016; 9: 449-472
        • Zhang Y.
        • Fonslow B.R.
        • Shan B.
        • Baek M.C.
        • Yates 3rd, J.R.
        Protein analysis by shotgun/bottom-up proteomics.
        Chem Rev. 2013; 113: 2343-2394
        • Liu H.
        • Sadygov R.G.
        • Yates 3rd, J.R.
        A model for random sampling and estimation of relative protein abundance in shotgun proteomics.
        Anal Chem. 2004; 76: 4193-4201
        • Gupta N.
        • Pevzner P.A.
        False discovery rates of protein identifications: a strike against the two-peptide rule.
        J Proteome Res. 2009; 8: 4173-4181
        • Wang Z.
        • Gerstein M.
        • Snyder M.
        RNA-Seq: a revolutionary tool for transcriptomics.
        Nat Rev Genet. 2009; 10: 57-63
        • Cox B.
        • Kislinger T.
        • Emili A.
        Integrating gene and protein expression data: pattern analysis and profile mining.
        Methods. 2005; 35: 303-314
        • Aird W.C.
        Endothelial cell heterogeneity.
        Cold Spring Harb Perspect Med. 2012; 2: a006429
        • Calabria A.R.
        • Shusta E.V.
        A genomic comparison of in vivo and in vitro brain microvascular endothelial cells.
        J Cereb Blood Flow Metab. 2008; 28: 135-148
        • Durr E.
        • Yu J.
        • Krasinska K.M.
        • et al.
        Direct proteomic mapping of the lung microvascular endothelial cell surface in vivo and in cell culture.
        Nat Biotechnol. 2004; 22: 985-992
        • Geraud C.
        • Schledzewski K.
        • Demory A.
        • et al.
        Liver sinusoidal endothelium: a microenvironment-dependent differentiation program in rat including the novel junctional protein liver endothelial differentiation-associated protein-1.
        Hepatology. 2010; 52: 313-326
        • Unger R.E.
        • Krump-Konvalinkova V.
        • Peters K.
        • Kirkpatrick C.J.
        In vitro expression of the endothelial phenotype: comparative study of primary isolated cells and cell lines, including the novel cell line HPMEC-ST1.6R.
        Microvasc Res. 2002; 64: 384-397
        • Miller J.W.
        • Le Couter J.
        • Strauss E.C.
        • Ferrara N.
        Vascular endothelial growth factor a in intraocular vascular disease.
        Ophthalmology. 2013; 120: 106-114
        • Dubail J.
        • Apte S.S.
        Insights on ADAMTS proteases and ADAMTS-like proteins from mammalian genetics.
        Matrix Biol. 2015; 44-46: 24-37
        • Tsutsui K.
        • Manabe R.
        • Yamada T.
        • et al.
        ADAMTSL-6 is a novel extracellular matrix protein that binds to fibrillin-1 and promotes fibrillin-1 fibril formation.
        J Biol Chem. 2010; 285: 4870-4882
        • Larrieu-Lahargue F.
        • Thomas K.R.
        • Li D.Y.
        Netrin ligands and receptors: lessons from neurons to the endothelium.
        Trends Cardiovasc Med. 2012; 22: 44-47
        • Kociok N.
        • Crespo-Garcia S.
        • Liang Y.
        • et al.
        Lack of netrin-4 modulates pathologic neovascularization in the eye.
        Sci Rep. 2016; 6: 18828
        • Li Y.N.
        • Pinzon-Duarte G.
        • Dattilo M.
        • Claudepierre T.
        • Koch M.
        • Brunken W.J.
        The expression and function of netrin-4 in murine ocular tissues.
        Exp Eye Res. 2012; 96: 24-35
        • Drusco A.
        • Zanesi N.
        • Roldo C.
        • et al.
        Knockout mice reveal a tumor suppressor function for Testin.
        Proc Natl Acad Sci U S A. 2005; 102: 10947-10951
        • Ma H.
        • Weng D.
        • Chen Y.
        • et al.
        Extensive analysis of D7S486 in primary gastric cancer supports TESTIN as a candidate tumor suppressor gene.
        Mol Cancer. 2010; 9: 190
        • Zhu J.
        • Li X.
        • Kong X.
        • et al.
        Testin is a tumor suppressor and prognostic marker in breast cancer.
        Cancer Sci. 2012; 103: 2092-2101
        • Piekny A.J.
        • Maddox A.S.
        The myriad roles of Anillin during cytokinesis.
        Semin Cell Dev Biol. 2010; 21: 881-891
        • Cartwright S.
        • Karakesisoglou I.
        Nesprins in health and disease.
        Semin Cell Dev Biol. 2014; 29: 169-179
        • Morgan J.T.
        • Pfeiffer E.R.
        • Thirkill T.L.
        • et al.
        Nesprin-3 regulates endothelial cell morphology, perinuclear cytoskeletal architecture, and flow-induced polarization.
        Mol Biol Cell. 2011; 22: 4324-4334
        • King S.J.
        • Nowak K.
        • Suryavanshi N.
        • Holt I.
        • Shanahan C.M.
        • Ridley A.J.
        Nesprin-1 and nesprin-2 regulate endothelial cell shape and migration.
        Cytoskeleton (Hoboken). 2014; 71: 423-434
        • Ingham R.J.
        • Gish G.
        • Pawson T.
        The Nedd4 family of E3 ubiquitin ligases: functional diversity within a common modular architecture.
        Oncogene. 2004; 23: 1972-1984
        • Schmidt M.
        • Finley D.
        Regulation of proteasome activity in health and disease.
        Biochim Biophys Acta. 2014; 1843: 13-25
        • Bargagna-Mohan P.
        • Ravindranath P.P.
        • Mohan R.
        Small molecule anti-angiogenic probes of the ubiquitin proteasome pathway: potential application to choroidal neovascularization.
        Invest Ophthalmol Vis Sci. 2006; 47: 4138-4145
        • Murdaca J.
        • Treins C.
        • Monthouel-Kartmann M.N.
        • et al.
        Grb10 prevents Nedd4-mediated vascular endothelial growth factor receptor-2 degradation.
        J Biol Chem. 2004; 279: 26754-26761
        • Grimsey N.J.
        • Aguilar B.
        • Smith T.H.
        • et al.
        Ubiquitin plays an atypical role in GPCR-induced p38 MAP kinase activation on endosomes.
        J Cell Biol. 2015; 210: 1117-1131
        • Rayman M.P.
        Selenium and human health.
        Lancet. 2012; 379: 1256-1268
        • Gladyshev V.N.
        • Arner E.S.
        • Berry M.J.
        • et al.
        Selenoprotein gene nomenclature.
        J Biol Chem. 2016; 291: 24036-24040
        • Kaur C.
        • Foulds W.S.
        • Ling E.A.
        Blood-retinal barrier in hypoxic ischaemic conditions: basic concepts, clinical features and management.
        Prog Retin Eye Res. 2008; 27: 622-647
        • Yadav U.C.
        • Kalariya N.M.
        • Ramana K.V.
        Emerging role of antioxidants in the protection of uveitis complications.
        Curr Med Chem. 2011; 18: 931-942
        • Hwang D.Y.
        • Sin J.S.
        • Kim M.S.
        • et al.
        Overexpression of human selenoprotein M differentially regulates the concentrations of antioxidants and H2O2, the activity of antioxidant enzymes, and the composition of white blood cells in a transgenic rat.
        Int J Mol Med. 2008; 21: 169-179
        • Schroder W.A.
        • Le T.T.
        • Major L.
        • et al.
        A physiological function of inflammation-associated SerpinB2 is regulation of adaptive immunity.
        J Immunol. 2010; 184: 2663-2670
        • Chen Q.
        • Sun L.
        • Chen Z.J.
        Regulation and function of the cGAS-STING pathway of cytosolic DNA sensing.
        Nat Immunol. 2016; 17: 1142-1149
        • Karikoski M.
        • Irjala H.
        • Maksimow M.
        • et al.
        Clever-1/Stabilin-1 regulates lymphocyte migration within lymphatics and leukocyte entrance to sites of inflammation.
        Eur J Immunol. 2009; 39: 3477-3487
        • Shetty S.
        • Weston C.J.
        • Oo Y.H.
        • et al.
        Common lymphatic endothelial and vascular endothelial receptor-1 mediates the transmigration of regulatory T cells across human hepatic sinusoidal endothelium.
        J Immunol. 2011; 186: 4147-4155
        • Aguilar E.
        • Dorrell M.I.
        • Friedlander D.
        • et al.
        Chapter 6. Ocular models of angiogenesis.
        Methods Enzymol. 2008; 444: 115-158
        • Nakatsu M.N.
        • Hughes C.C.
        An optimized three-dimensional in vitro model for the analysis of angiogenesis.
        Methods Enzymol. 2008; 443: 65-82
        • Knott R.M.
        • Robertson M.
        • Muckersie E.
        • et al.
        A model system for the study of human retinal angiogenesis: activation of monocytes and endothelial cells and the association with the expression of the monocarboxylate transporter type 1 (MCT-1).
        Diabetologia. 1999; 42: 870-877
        • Pan Y.
        • Appukuttan B.
        • Mohs K.
        • Ashander L.M.
        • Smith J.R.
        Ubiquitin carboxyl-terminal esterase L1 promotes proliferation of human choroidal and retinal endothelial cells.
        Asia Pac J Ophthalmol (Phila). 2015; 4: 51-55
        • Bowden E.T.
        • Coopman P.J.
        • Mueller S.C.
        Invadopodia: unique methods for measurement of extracellular matrix degradation in vitro.
        Methods Cell Biol. 2001; 63: 613-627
        • Harhaj N.S.
        • Felinski E.A.
        • Wolpert E.B.
        • Sundstrom J.M.
        • Gardner T.W.
        • Antonetti D.A.
        VEGF activation of protein kinase C stimulates occludin phosphorylation and contributes to endothelial permeability.
        Invest Ophthalmol Vis Sci. 2006; 47: 5106-5115
        • Srinivasan B.
        • Kolli A.R.
        • Esch M.B.
        • Abaci H.E.
        • Shuler M