A, Geographical map of the 3 different tissue samples. B, Detailed image of the human choroid-RPE tissue punch biopsy specimens collected for this study. C, Fundus image of neovascular age-related macular degeneration (arrowhead indicates subretinal hemorrhage from choroidal neovascularization; small circle, fovea; large circle, macula area). D, Fundus image of macular geographic atrophy sparing the fovea (large circle indicates the macula area). E, Fundus image of pigment degeneration in the periphery.
Unbiased clustering of proteins differentially expressed (P < .05) in the 3 different regions of the human choroid-RPE complex. Proteins represented in this cluster analysis were compared using analysis of variance (P < .05). The heatmap is divided into regions. A, Proteins with highest expression in the macula. B, Proteins with highest expression in the periphery. C, Proteins with expression exclusive to the periphery. D, Proteins with expression exclusive to the fovea. E, Proteins with highest expression in the fovea. F, Proteins absent in the macula but present in the fovea and periphery. Numbers indicate the number of proteins in each category. The bar at the bottom depicts a logarithmic color scale for relative expression where orange represents high and black represents low levels.
Analysis reveals unique groups of proteins. A, Differentially expressed proteins were determined by analysis of variance. Proteins with P < .05 were chosen from each geographical region. Proteins were grouped into subcategories of biological processes, molecular functions, and cellular component for each region of choroid-RPE. B, Top ten differentially expressed protein pathways represented in the choroid-RPE. ACM indicates muscarinic acetylcholine receptor M3; EGF, endothelial growth factor; GPCR, G-protein coupled receptor; LRRK2, leucine-rich repeat serine/threonine-protein kinase 2; LTD4, leukotriene D4; and PDGF, platelet-derived growth factor.
Nineteen proteins had statistically significant differences in expression among the 3 different geographical regions, with the highest expression in the fovea. Proteins are divided into categories. A, Matrix proteins. B, Rab proteins. C, Oxidative stress proteins. D, Protein expression regulators. E, Immune modulators. The statistically significant changes are available in eTable 4 in the Supplement. Abundance of protein is displayed as mean number of spectra. ACAA1 indicates acetyl coenzyme A acyltransferase 1; ERG, transcription factor ERG; HM13, minor histocompatibility antigen H13; LAM, laminin; MYH, myosin; NID1, nidogen 1; PRDX1, peroxiredoxin 1; PSMD11, proteasome 26S subunit 11; and TUFM, Tu translations elongation factor. Error bars indicate SEM.
Each pie chart shows the relative protein representation in the fovea (pink), macula (blue), and periphery (yellow). The size of the pie chart represents a relative abundance of each complement protein in the data set. Activators (green arrows) and inhibitors (red lines) are also listed in eTable 6 in the Supplement. The 3 different pathways that initiate complement activation are shown. All pathways converge at the formation of C3 and C5 convertases and result in the formation of anaphylatoxins C3a and C5a (blue) and the membrane attack complex in the terminal pathway. The (n) indicates that multiple C9 molecules can be linked to this protein complex. CD55 indicates decay accelerating factor for complement; CD59, complement regulatory protein; CFD, complement factor D; CFH, complement factor H; CFI, complement factor I; CLU, clusterin; MASP, mannan-binding serine peptidase; MBL, mannose-binding lectin; and SERPING1, serpin peptidase inhibitor, clade G (c1 inhibitor), member 1.
eFigure 1. Choroid–retinal pigment epithelial (RPE) proteomic analysis pipeline
eFigure 2. Gene ontology (GO) distributions and pathway analysis of human choroid–retinal pigment epithelial (RPE) protein show tissue similarity
eMethods. Tissue collection, mass spectrometry, and analysis
eTable 1. Complete table of proteins identified in the human choroid-RPE using LC-MS/MS.
eTable 2. Top 50 signaling pathways in the human choroid-RPE using MetaCore.
eTable 3. Top 50 differentially expressed signaling pathways in the human choroid-RPE using MetaCore.
eTable 4. Differentially expressed proteins in the fovea, macula, and periphery of the human choroid-RPE.
eTable 5. Differentially expressed transcription proteins were identified in foveal, macular, and peripheral choroid-RPE (p<0.05).
eTable 6. Complement proteins identified in the human choroid-RPE.
eTable 7. Proteins associated with age related macular degeneration (AMD).
Skeie JM, Mahajan VB. Proteomic Landscape of the Human Choroid–Retinal Pigment Epithelial Complex. JAMA Ophthalmol. Published online July 24, 2014. doi:10.1001/jamaophthalmol.2014.2065.
Skeie JM, Mahajan VB. Proteomic Landscape of the Human Choroid–Retinal Pigment Epithelial Complex. JAMA Ophthalmol. 2014;132(11):1271-1281. doi:10.1001/jamaophthalmol.2014.2065
Differences in geographical protein expression in the human choroid–retinal pigment epithelial (RPE) complex may explain molecular predisposition of regions to ophthalmic diseases such as age-related macular degeneration.
To characterize the proteome of the human choroid-RPE complex and to identify differentially expressed proteins in specific anatomic regions.
Design, Setting, and Participants
Experimental study of choroid-RPE tissue from 3 nondiseased eyes. The choroid-RPE complex underwent biopsy from beneath the foveal, macular, and peripheral retina. Protein fractions were isolated and subjected to multidimensional liquid chromatography and tandem mass spectrometry. A bioinformatic pipeline matched peptide spectra to the human proteome, assigned gene ontology classification, and identified protein signaling pathways unique to each of the choroid-RPE regions.
Main Outcomes and Measures
Mean number of mass spectra, statistically significant differentially expressed proteins, gene ontology classification, and pathway representation.
We identified a mean of 4403 unique proteins in each of the foveal, macular, and peripheral choroid-RPE tissues. Six hundred seventy-one differentially expressed proteins included previously known risk factors for retinal diseases related to oxidative stress, inflammation, and the complement cascade. Gene ontology analysis showed that unique categories in the foveal and macular regions included immune process proteins as well as protein complexes and plasma membrane proteins. The peripheral region contained unique antioxidant activity proteins. Many proteins had the highest expression in the foveal or macular regions, including inflammation-related proteins HLA-A, HLA-B, and HLA-C antigens; intercellular adhesion molecule 1 (ICAM-1); S100; transcription factor ERG; antioxidant superoxide dismutase 1 (SOD1); chloride intracellular channel 6 ion (CLIC6); activators of the complement cascade C1q, C6, and C8; and complement factor H. Proteins with higher expression in the periphery included bestrophin 1 (BEST1), transcription factor RNA binding motif protein 39 (RBM39), inflammatory mediator macrophage migration inhibitory factor, antioxidant SOD3, ion channel voltage-dependent anion-selective channel protein 3 (VDAC3), and complement inhibitor CD55. The complement activation was among the highest represented pathways (P < 7.5e−13).
Conclusions and Relevance
This proteomic data set identifies novel molecular signatures in anatomically sensitive regions of the choroid-RPE complex. The findings give mechanistic insight into choroid-RPE function, reveal important choroid-RPE processes, and prioritize new pathways for therapeutic targeting.
The choroid is a high-flow vascular network with fenestrated capillaries that surround and nourish the neurosensory retina. Disruption of the choroid–retinal pigment epithelial (RPE) complex is a frequent cause of blinding retinal diseases. These diseases include age-related macular degeneration (AMD), central serous retinopathy, infectious and noninfectious chorioretinitis, and retinal degeneration.1 Many of these conditions localize to specific regions of the fundus. Neovascular AMD is frequently localized to the central foveal macula. Dry AMD, on the other hand, often spares the fovea. Infectious and noninfectious choroiditis occur throughout the peripheral fundus (Figure 1). Why choroid-RPE diseases show regional susceptibility is not known. Anatomic variation of the choroid-RPE complex is one commonly proposed explanation for disease localization.2- 5 Higher metabolic demands imposed by the overlying retina compared with the periphery is another. We considered an alternative hypothesis that geographic differences in protein expression patterns could make some regions prone to disease.
Several molecules are associated with choroid-RPE diseases. Complement and inflammatory molecules are strongly associated with AMD.6 Oxidative stress may also predispose the macula to AMD, because it can activate the innate immune response via the complement system.7- 10 Glucocorticoids and ion channels are implicated in central serous retinopathy.11,12 Innate and adaptive immune responses are involved in chorioretinitis.13 Also, a number of molecules in the outer retina rely on a functional interaction with the RPE and choroid, including those involved in energy metabolism, ion/water transport, absorption of light, and waste removal.14 A comprehensive molecular map of the choroid-RPE complex could provide critical insight into disease mechanisms, but the normal distribution of these molecules across the choroid-RPE complex has not been explored on a large scale. We used multidimensional liquid chromatography and tandem mass spectrometry in an unbiased systems biology approach to identify large numbers of proteins in regions of the choroid-RPE complex.15
This study was approved by the institutional review board of the University of Iowa and adheres to the tenets of the Declaration of Helsinki. Human donor tissue was obtained from the Iowa Lions Eye Bank, Iowa City. Tissue was obtained within 5 hours after death from 1 man and 2 women, all in their eighth and ninth decades of life. None of the eyes showed signs of retinal disease. Eyes were flowered into 4 quadrants as previously described.16 The choroid-RPE complex was collected using a 4-mm (fovea and periphery) or an 8-mm (macula around foveal punch) knife biopsy punch and stored in our biorepository until processed for mass spectrometry (Figure 1 and eFigure 1A in the Supplement).17
Mass spectrometry was performed as previously described using a nanoflow system (Agilent 1100 series; Agilent Technologies) and a dual-pressure linear ion trap device (LTQ Velos, Thermo Fisher).18- 20 The Mascot generic format (MGF) files were searched with X!Hunter21 against the latest library available in 2010 on the global proteome machine22 and X!!Tandem23,24 using the native and k score25 algorithms and the open mass spectrometry search algorithm (OMSSA) (eFigure 1B-D in the Supplement).26
Proteins were considered identified if they had an expectation value of less than 0.01 (ie, <1% chance of being a random assignment). A systems biology approach was implemented using several bioinformatic analyses to determine significant protein expression (Partek Genomics Suite 6.6; Partek, Inc), gene ontology (GO terminology, Protein Analysis Through Evolutionary Relationships software, version 7.2 [PANTHER]; http://www.pantherdb.org/), and pathway representation (MetaCore; GeneGo) (eFigure 1E in the Supplement).27 Statistically significant proteins (analysis of variance, P < .05) were visualized using an undiscriminated clustered heatmap with a normalized clustering function. Proteins that showed a trend but did not meet statistical significance were analyzed further if they were an important component of a specific pathway or classification (eMethods in the Supplement).
Choroid-RPE samples underwent trypsinization and multidimensional liquid chromatography before analysis by tandem mass spectrometry. In the periphery, we identified 193 460 spectra with 21 829 unique peptides, corresponding to 4204 unique proteins. The macula had 189 447 spectra with 24 915 unique peptides, corresponding to 4595 unique proteins, and the fovea had 196 028 spectra with 23 379 unique peptides, corresponding to 4409 unique proteins. The mean total spectra for the 3 samples showed excellent correlation with an SD of only 1.7% of the mean total. This correlation indicated the protein load for each sample was highly consistent. The most abundant proteins identified were albumin, 6 classes of tubulin, 5 classes of hemoglobin, serpin peptidase inhibitor, vimentin, transferrin, actin, cathepsin D, and complement component C3 (eTable 1 in the Supplement). Infectious agents, which have been suggested as a cause of chronic chorioretinal degeneration, were not detected when the peptide data set was searched against nonhuman protein libraries.
When we compared the protein profiles for all 3 choroid-RPE regions, the gene ontology summaries were similar. The categories with the highest representation were metabolic process, binding, catalytic activity, and intracellular region (eFigure 2A in the Supplement).
After identifying proteins that were differentially expressed in specific choroid-RPE regions (Figure 2), differences in gene ontology categorization emerged. For biological processes, proteins with the highest expression in the periphery fell into the category of metabolites/energy. Proteins with the highest expression in the macula and fovea fit into the response to stimulus, reproduction, homeostasis, or immune system process categories, whereas these categories were absent in the periphery (Figure 3A). Catalytic proteins were represented slightly more in the macula than the other 2 regions, whereas proteins associated with antioxidant activity were represented more in the periphery. For the cellular component category, the fovea and macula had more proteins that localized to the plasma membrane and formed complexes compared with the periphery. In the periphery, far more ribonucleoprotein complex proteins were found (Figure 3A).
A molecular pathway analysis identified groups of functionally related proteins. The global top ten pathways in the choroid-RPE complex were cytoskeleton remodeling, chemokine cell adhesion, integrin-mediated cell adhesion and migration, transforming growth factor and Wnt cytoskeleton remodeling, leucine-rich repeat serine/threonine-protein kinase 2 (LRRK2) in neurons, lectin-induced and classic complement pathway immune responses, clathrin-coated vesicle cycle transport, histamine H1 receptor signaling cell adhesion, and blood coagulation (eFigure 2B and eTable 2 in the Supplement). The top ten pathways determined from differentially expressed proteins were LRRK2 in neurons, airway smooth muscle contraction, muscle contraction, chemokine and integrin-mediated cell adhesion, development and process of inhibitory action of lipoxin A4, netrin 1 in regulation of axon guidance, muscarinic acetylcholine receptor M3 (ACM3) transport, and ACM regulation of nerve impulse (Figure 3B and eTable 3 in the Supplement).
Comparison of the 3 anatomic regions of the choroid-RPE complex revealed groups of proteins that were differentially expressed (eTable 4 in the Supplement).
We found high expression of 66 proteins in the foveal choroid-RPE complex only, and 251 more proteins in the fovea and macula choroid-RPE complex compared with the periphery (Figure 2). These proteins included HLA-A, HLA-B, and HLA-C antigens, which play an important role in the immune system antigen presentation. Also included in this area were Rab proteins (RAB7A, RAB8A, RAB8B, and RAB35), nidogen 1 (NID1), laminins (LAMA5 and LAMB2), Tu translations elongation factor (TUFM), acetyl coenzyme A acyltransferase 1 (ACAA1), minor histocompatibility antigen H13 (HM13), myosins (MYH9, MYH10, and MYH14), peroxiredoxin 1, proteasome 26S subunit 11 (PSMD11), and transcription factor ERG (Figure 4). Mutations in TUFM have been associated with oxidative phosphorylation deficiency and fatal encephalopathy.28 Pseudo-Zellweger syndrome is associated with ACAA1.29 Because HM13 is responsible for peanut allergies and possibly graft rejection, it may have a role in inflammatory responses. Transcription factor ERG was found only in the fovea region and is implicated in embryonic development, cell proliferation, cell differentiation, angiogenesis, inflammation, and apoptosis. Peroxiredoxin 1 is an antioxidant enzyme, which reduces hydrogen peroxide expression, linked to angiogenesis and cancer and tumor progression.30 Nidogen 1 binds LAMC1, and mutations in these genes are associated with Dandy-Walker malformation and occipital cephaloceles.31 Different myosins are associated with deafness. As a component of the immunoproteasome, PSMD11 processes class I major histocompatibility complex peptides.
Several proteins had higher levels of expression in the peripheral choroid-RPE complex compared with the macula and fovea. These proteins included bestrophin 1 (BEST1), rhodopsin (RHO), ras homolog family member A (RHOA), and ras homolog family member C (RHOC). Bestrophin 1 forms a chloride channel expressed by RPE cells.32,33 Rhodopsin is the retinal outer segment rod opsin responsible for light detection, and RHOA and RHOC are involved in cell cycle cytokinesis.34
Transcription proteins profoundly influence tissue functions by controlling gene expression. We found several transcription proteins with statistically significant expression differences in regions of the choroid-RPE complex. Expression of transcription factor ERG (V-ets erythroblastosis virus E26 oncogene homolog) was highest in the foveal choroid-RPE region. In the macular choroid-RPE region, expression of cell division cycle and apoptosis regulator 1 (CCAR1) and histone deacetylase 6 (HDAC6) were higher. In the peripheral choroid-RPE region, transcription proteins with elevated levels included RNA binding motif protein 39 (RBM39), cyclin-dependent kinase 9 (CDK9), and nascent polypeptide-associated complex subunit alpha (NACA) (eTable 5 in the Supplement). These transcription proteins control processes such as apoptosis, cell differentiation, angiogenesis, and inflammation.
Inflammatory proteins are associated with ophthalmic diseases including AMD, retinal degeneration, and retinochoroiditis. Intercellular adhesion molecule 1 (ICAM-1), S100, and advanced glycosylation end product–specific receptor (AGER) were more highly expressed in the fovea and macula. In contrast, expression of macrophage migration inhibitory factor (MIF) and proteasome subunit beta type 8 (PSMB8) were higher in the periphery. Intercellular adhesion molecule 1 is an endothelial cell surface protein that aids in the tethering of leukocytes to the endothelial cell surface and extravasation of leukocytes from the circulation into nearby tissues.35 The S100 proteins are a diverse group of proteins that play a role in the regulation of proliferation, differentiation, metabolism, and inflammation.36 Advanced glycosylation end product–specific receptor is an immunoglobulin cell surface receptor that binds advanced glycosylation end products and proteins involved in inflammation and disease. Macrophage migration inhibitory factor is a proinflammatory cytokine that helps to regulate macrophage defenses against bacterial infections. Proteasome subunit beta type 8 is a proteasome with catalytic activity involved in regulating inflammation. The G201V mutation in this protein is directly associated with inflammatory Nakajo-Nishimura syndrome.37
Several antioxidant proteins were found in the choroid-RPE complex. These proteins included superoxide dismutases (SOD1, SOD2, and SOD3), glutathione peroxidases (GPX1, GPX3, and GPX4), glutathione reductase (GSR), vitronectin (VTN), and peroxiredoxins (PRDX1, PRDX2, PRDX3, PRDX4, PRDX5, and PRDX6). Some of these proteins had higher expression in the periphery (SOD3 and GPX3), and others had higher expression in the fovea (SOD1, GPX1, GPX4, PRDX1, PRDX2, PRDX3, and VTN). Some oxidative stress proteins associated with AMD, such as nuclear factor erythroid 2–related factor 2 (NRF2), kelchlike ECH-associated protein 1 (KEAP1), and NLR pyrin–containing (NLRP) inflammasome proteins were not found in the data set.8,38
The high level and number of complement component proteins found in the choroid-RPE complex was striking. We found a total of 30 complement cascade proteins, including all of the proteins that constitute the membrane attack complex (MAC; C5b-9) (eTable 6 in the Supplement). Most of these proteins were evenly distributed. However, some regulatory proteins of the complement cascade were differentially expressed. Expression of C8G, C6, complement factor H (CFH), and clusterin (CLU) was significantly higher in the fovea than in the macula and the periphery, whereas expression of CD55 was highest in the periphery and almost completely absent in the fovea.
Macular diseases such as central serous retinopathy are associated with subretinal fluid that may be owing to malfunctioning ion channels in the RPE. We identified 27 ion channels, including several voltage-sensitive chloride, calcium, and potassium channels. Voltage-dependent anion channels 1 and 3 (VDAC1 and VDAC3) displayed the highest expression in the periphery, whereas expression of chloride intracellular channel 6 (CLIC6) was highest in the fovea choroid-RPE region. The VDACs are small mitochondrial membrane proteins that regulate the diffusion of small molecules.39 The CLICs play a role in chloride ion transport, which may help regulate osmotic concentration.40
We interrogated the expression of 84 proteins associated with AMD that were identified in genetic screens.41 We detected apolipoprotein E (APOE), endothelial growth factor–containing fibulinlike extracellular matrix protein 1 (EFEMP1), HtrA serine peptidase 1 (HTRA1), ICAM-1, serpin peptidase inhibitor, clade G (c1 inhibitor), member 1 (SERPING1), and von Willebrand factor (vWF) (eTable 5 in the Supplement). Several AMD biomarkers were not detected in the data set, including adenosine triphosphate–binding cassette subfamily A member 1 (ABCA1), age-related maculopathy susceptibility 2 (ARMS2), complement factor B (CFB), interleukin 6 (IL-6), interleukin 8 (IL-8), and vascular endothelial growth factor A (VEGFA) (eTable 7 in the Supplement). These proteins may represent good biomarkers for the study of AMD progression.
The choroid-RPE is a complex tissue where proteins arise from multiple origins. Gene expression studies might not account for proteins originating from distant organs.15,18 Proteomic analysis using mass spectrometry is a powerful method that accounts for global protein expression without bias to specific cells. Using this platform, our data set demonstrates significant molecular differences between regions of the human choroid-RPE complex that may account for the regional susceptibility observed in diseases such as AMD, central serous retinopathy, and retinitis pigmentosa and those without geographic preferences such as inflammatory choroiditis. The identification of transcription factors with differential expression confirms that similarly appearing cells carry unique molecular properties, because transcription regulators control the identity of cells through regulation of expression of hundreds of downstream genes.
Our study shows that molecular control of infection and inflammation regulation is a primary activity within the choroid-RPE complex. Among the most abundant proteins were transferrin and complement component C3 of the innate immune system. Differential expression of inflammatory proteins could be correlated directly with region-specific diseases. These proteins are important in understanding differential susceptibility of the choroid-RPE complex to inflammation. Proteins with the highest expression in the fovea and macula included ICAM-1, S100, and AGER. In contrast, MIF and PSMB8 had highest expression in the periphery. A previous investigation42 has shown ICAM-1 to be associated with elevated levels of anaphylatoxin C5a, indicating that increased complement activation, as shown in AMD, can mediate an increase in leukocyte tissue infiltration. Macrophage migration inhibitory factor is a proinflammatory cytokine that is involved with innate immunity, specifically in response to bacterial cells. We found that expression of MIF was highest in the peripheral choroid-RPE region and lowest in the fovea. High expression of HLA-A, HLA-B, and HLA-C antigens was found in the fovea and macula compared with the periphery. These proteins play an important role in presenting antigens to the immune system. The HLA antigen polymorphisms are associated with AMD and uveitis.43,44
One of the most highly represented pathways was the complement cascade (Figure 5),45 which provides an important context for interpreting the prior genetic studies, specifically how genome-wide association study (GWAS) risk alleles relate to potential therapeutics. Gene mutations in complement and complement regulators such as C3, CFH, C2, and complement factor I were associated with AMD.41,46- 57 In diseased eyes, these molecules localize to sites of degeneration. For example, MAC was found in drusen and on the surface of choroid-RPE complex blood vessels in AMD.58,59 Complement factor H inhibits the formation of MAC, so the Y402H variant potentially causes a loss of function. We found relatively more CFH expression in the fovea than the periphery, which may represent a protective mechanism that can be disrupted by loss of function mutations in patients with AMD who have high-risk alleles. Expression of CD55 was higher in the peripheral choroid-RPE region, which is an inhibitor of MAC formation.60- 62 No CD55 gene variants associated with AMD have been reported, but the lower CD55 expression in the posterior pole may make the fovea and macula more susceptible to complement-activated damage.
Proteins associated with inherited retinal degeneration were also detected in our data set and showed greater expression in the peripheral choroid-RPE complex. The protein with the greatest peripheral expression trend was C1qTNF5. Mutations in the C1qTNF5 gene cause late-onset retinal degeneration (LORD [OMIM 605670]), characterized by drusenlike deposits, peripheral and central atrophy of the choroid and retina, and neovascularization.63 Rhodopsin mutations (RHO [OMIM 180380]) cause retinitis pigmentosa, and expression was also highest in the periphery. Retinal pigment epithelial cells phagocytose retinal outer segments, and the topographical ratio of rods to cones is higher in the periphery compared with the macula and the fovea.64 Expression of BEST1 was highest in the periphery, and this finding is supported by a previous immunohistochemistry study.32 Mutations in BEST1 cause Best vitelliform macular dystrophy (VMD [OMIM 153700]) and autosomal dominant vitreoretinochoroidopathy (ADVIRC [OMIM 193220]), which display degenerative phenotypes in the macula and periphery. The adenosine triphosphate–binding cassette subfamily A, member 4 (ABCA4 [OMIM 601691]) was found in all 3 regions, but had higher expression in the periphery. Mutations of ABCA4 cause Stargardt disease, which can have degenerative patterns in the peripheral retina and macula.65
Central serous retinopathy typically occurs in the posterior pole, may be owing to failure of an RPE ion pump, and is made worse by glucocorticoids.11,12 We identified differential expression of specific ion channels.66 Expression of VDAC1 and VDAC3 was highest in the periphery and expression of CLIC6 was highest in the fovea. The VDACs were associated with Alzheimer disease and Down syndrome.67 A CLIC6 mutation was associated with goiter. None of the differentially expressed ion channels have previous links to chorioretinal diseases.68
Oxidative stress is implicated in acute and chronic diseases owing to the destruction of proteins caused by unmanaged free radicals. Oxidative stress proteins were found in human drusen, including modified forms of metalloproteinase inhibitor 3 (TIMP3) and VTN, which may lead to the buildup of waste products in the Bruch membrane.69 We found a consistent presence of TIMP3 across the entire choroid-RPE complex but found VTN expression was higher in the foveal region. We also found other oxidative stress proteins with differential expression in the periphery (SOD3 and GPX3) and fovea (SOD1, GPX1, GPX4, PRDX1, PRDX2, and PRDX3). Interestingly, Sod1−/− mice display a neovascular phenotype similar to AMD.70 The differential expression of oxidative stress enzymes in our proteomics data set may indicate regionally specific protective mechanisms.
Several previously reported proteins linked to chorioretinal disease were not found in this data set of normal human eyes. For example, VEGF is highly associated with choroid-RPE–related diseases, but we found no evidence of VEGF proteins or receptors. These proteins may only be present in disease states and could represent good biomarkers for the investigation of diseased eyes.
Proteomic analysis is a powerful method for investigating the function of complex tissues. Even without prior genetic studies, this data set would point to the complement cascade as a potential therapeutic target in AMD. Antibodies directed at activators of the complement cascade are currently on the horizon and offer promise for geographic atrophy. Lampalizumab, an antibody directed against complement factor D, is currently in phase 2 testing and was reported to show a 44% reduction in the rate of geographic atrophy.71 Instead of suppressing complement activators, our study suggests that targeted overexpression of complement inhibitors, such as CD55, could also be considered as a potential therapy for AMD. Complement inhibitors might also be effective in other retinal diseases with choroid-RPE inflammation such as uveitis, in which corticosteroid-sparing agents are needed. Further interrogation of the choroid-RPE proteomic data set should generate additional hypotheses for future validation studies in the laboratory and clinic.
Submitted for Publication: February 8, 2014; final revision received April 13, 2014; accepted April 14, 2014
Corresponding Author: Vinit B. Mahajan MD, PhD, Omics Laboratory, Department of Ophthalmology and Visual Sciences, University of Iowa Carver College of Medicine, 200 Hawkins Dr, Iowa City, IA 52242 (firstname.lastname@example.org).
Published Online: July 24, 2014. doi:10.1001/jamaophthalmol.2014.2065.
Author Contributions: Dr Mahajan had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: All authors.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: All authors.
Critical revision of the manuscript for important intellectual content: Mahajan.
Statistical analysis: Skeie.
Obtained funding: All authors.
Administrative, technical, or material support: Mahajan.
Study supervision: Mahajan.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was supported by the Bright Focus Foundation and by grants 1F32EY022280-01A1 (Dr Skeie) and K08EY020530 (Dr Mahajan) from the National Institutes of Health.
Role of the Sponsor: The funding organizations had no role in design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.