Transcriptomic Characterization of Human Choroidal Neovascular Membranes Identifies Calprotectin as a Novel Biomarker for Patients with Age-Related Macular Degeneration

Recent studies deciphering the transcriptional profile of choroidal neovascularization (CNV) in body donor eyes with neovascular age-related macular degeneration are limited by the time span from death to preservation and the associated 5′-RNA degradation. This study therefore used CNV and control specimens that were formalin-fixed and paraffin-embedded immediately after surgical extraction and analyzed them by a 3′-RNA sequencing approach. Transcriptome profiles were analyzed to estimate content of immune and stromal cells and to define disease-associated gene signatures by using statistical and bioinformatics methods. This study identified 158 differentially expressed genes (DEGs) that were significantly increased in CNV compared with control tissue. Cell type enrichment analysis revealed a diverse cellular landscape with an enrichment of endothelial cells, macrophages, T cells, and natural killer T cells in the CNV. Gene ontology enrichment analysis found that DEGs contributed to blood vessel development, extracellular structure organization, response to wounding, and several immune-related terms. The S100 calcium-binding proteins A8 (S100A8) and A9 (S100A9) emerged among the top DEGs, as confirmed by immunohistochemistry on CNV tissue and protein analysis of vitreous samples. This study provides a high-resolution RNA-sequencing–based transcriptional signature of human CNV, characterizes its compositional pattern of immune and stromal cells, and reveals S100A8/A9 to be a novel biomarker and promising target for therapeutics and diagnostics directed at age-related macular degeneration. Recent studies deciphering the transcriptional profile of choroidal neovascularization (CNV) in body donor eyes with neovascular age-related macular degeneration are limited by the time span from death to preservation and the associated 5′-RNA degradation. This study therefore used CNV and control specimens that were formalin-fixed and paraffin-embedded immediately after surgical extraction and analyzed them by a 3′-RNA sequencing approach. Transcriptome profiles were analyzed to estimate content of immune and stromal cells and to define disease-associated gene signatures by using statistical and bioinformatics methods. This study identified 158 differentially expressed genes (DEGs) that were significantly increased in CNV compared with control tissue. Cell type enrichment analysis revealed a diverse cellular landscape with an enrichment of endothelial cells, macrophages, T cells, and natural killer T cells in the CNV. Gene ontology enrichment analysis found that DEGs contributed to blood vessel development, extracellular structure organization, response to wounding, and several immune-related terms. The S100 calcium-binding proteins A8 (S100A8) and A9 (S100A9) emerged among the top DEGs, as confirmed by immunohistochemistry on CNV tissue and protein analysis of vitreous samples. This study provides a high-resolution RNA-sequencing–based transcriptional signature of human CNV, characterizes its compositional pattern of immune and stromal cells, and reveals S100A8/A9 to be a novel biomarker and promising target for therapeutics and diagnostics directed at age-related macular degeneration. Choroidal neovascularization (CNV) is a common cause of irreversible vision loss in patients with age-related macular degeneration (AMD), which is the leading cause of blindness in the elderly, affecting approximately 200 million individuals worldwide.1Wong W.L. Su X. Li X. Cheung C.M.G. Klein R. Cheng C.Y. Wong T.Y. Global prevalence of age-related macular degeneration and disease burden projection for 2020 and 2040: a systematic review and meta-analysis.Lancet Glob Health. 2014; 2: e106-e116Abstract Full Text Full Text PDF PubMed Scopus (2072) Google Scholar The introduction of antiangiogenic therapy has enabled significant advances in the treatment of neovascular AMD (nAMD). However, the treatment's success is hampered by the relatively short half-life of antibodies in the eye and depends on frequently repeated intraocular injections that pose a significant cumulative risk of local complications. Approximately one-third of patients with nAMD lose vision despite continuous anti–vascular endothelial growth factor therapy,2Wecker T. Ehlken C. Bühler A. Lange C. Agostini H. Böhringer D. Stahl A. Five-year visual acuity outcomes and injection patterns in patients with pro-re-nata treatments for AMD, DME, RVO and myopic CNV.Br J Ophthalmol. 2017; 101: 353-359PubMed Google Scholar highlighting the complexity of this disease and indicating a relevance of other disease-associated molecular mediators. Although genome-wide association studies, animal models, and cell culture systems have provided important insights into nAMD pathogenesis, the underlying cellular and molecular mechanisms in the cascade of events leading to CNV have not been identified in detail. Risk factors for nAMD include increasing age, environmental factors such as smoking, and dietary habits, as well as mutations in susceptibility genes, which predispose an individual to developing AMD. Most AMD gene-association studies published thus far are genome-wide association studies investigating single nucleotide polymorphisms or noncoding single nucleotide polymorphisms associated with CNV formation. These studies have identified several risk factors, such as polymorphisms in the CFH, ARMS2, or C3 genes,3Maguire M.G. Ying G.S. Jaffe G.J. Toth C.A. Daniel E. Grunwald J. Martin D.F. Hagstrom S.A. CATT Research GroupSingle-nucleotide polymorphisms associated with age-related macular degeneration and lesion phenotypes in the comparison of age-related macular degeneration treatments trials.JAMA Ophthalmol. 2016; 134: 674-681Crossref PubMed Scopus (11) Google Scholar but they do not provide a detailed view on the biological pathways underlying CNV formation. Gene expression studies on human CNV membranes are therefore becoming increasingly important to move beyond genetic association and comprehensively uncover CNV-associated pathways.4Morgan D.J. DeAngelis M.M. Differential gene expression in age-related macular degeneration.Cold Spring Harb Perspect. 2014; 5: a017210PubMed Google Scholar To that end, numerous studies have used microarray technology to explore the expression profiles of normal human eye tissues5Strunnikova N.V. Maminishkis A. Barb J.J. Wang F. Zhi C. Sergeev Y. Chen W. Edwards A.O. Stambolian D. Abecasis G. Swaroop A. Munson P.J. Miller S.S. Transcriptome analysis and molecular signature of human retinal pigment epithelium.Hum Mol Genet. 2010; 19: 2468-2486Crossref PubMed Scopus (177) Google Scholar,6Yoshida S. Yashar B.M. Hiriyanna S. Swaroop A. Microarray analysis of gene expression in the aging human retina.Invest Ophthalmol Vis Sci. 2002; 43: 2554-2560PubMed Google Scholar and to determine genes differentially expressed in eyes of patients with nAMD.7Chowers I. Wong R. Dentchev T. Farkas R.H. Iacovelli J. Gunatilaka T.L. Medeiros N.E. Presley J.B. Campochiaro P.A. Curcio C.A. Dunaief J.L. Zack D.J. The iron carrier transferrin is upregulated in retinas from patients with age-related macular degeneration.Invest Ophthalmol Vis Sci. 2006; 47: 2135-2140Crossref PubMed Scopus (86) Google Scholar Those studies, however, were limited by their dependence on the standard genome annotations required for probe design and possible cross-hybridization problems.8Liu F. Jenssen T.K. Trimarchi J. Punzo C. Cepko C.L. Ohno-Machado L. Hovig E. Patrick Kuo W. Comparison of hybridization-based and sequencing-based gene expression technologies on biological replicates.BMC Genomics. 2007; 8: 153Crossref PubMed Scopus (56) Google Scholar The more recent availability of RNA-sequencing (RNA-Seq) techniques improved examination of the transcriptome of healthy and diseased tissues, enabling a more accurate and unbiased measurement of gene expression abundance. Moreover, RNA-Seq can assess novel and particularly rare transcripts that have been overlooked by conventional microarray technology so far, and it is also more reproducible due to less technical variation and fewer false-positive findings.9Ozsolak F. Milos P.M. RNA sequencing: advances, challenges and opportunities.Nat Rev Genet. 2011; 12: 87-98Crossref PubMed Scopus (1356) Google Scholar,10Yendrek C.R. Ainsworth E.A. Thimmapuram J. The bench scientist's guide to statistical analysis of RNA-Seq data.BMC Res Notes. 2012; 5: 506Crossref PubMed Scopus (24) Google Scholar All previously published microarray and RNA-Seq studies on CNV membranes, however, are limited by the fact that they have been performed on human donor eyes.7Chowers I. Wong R. Dentchev T. Farkas R.H. Iacovelli J. Gunatilaka T.L. Medeiros N.E. Presley J.B. Campochiaro P.A. Curcio C.A. Dunaief J.L. Zack D.J. The iron carrier transferrin is upregulated in retinas from patients with age-related macular degeneration.Invest Ophthalmol Vis Sci. 2006; 47: 2135-2140Crossref PubMed Scopus (86) Google Scholar,11Voigt A.P. Mulfaul K. Mullin N.K. Flamme-Wiese M.J. Giacalone J.C. Stone E.M. Tucker B.A. Scheetz T.E. Mullins R.F. Single-cell transcriptomics of the human retinal pigment epithelium and choroid in health and macular degeneration.Proc Natl Acad Sci U S A. 2019; 116: 24100-24107Crossref PubMed Scopus (77) Google Scholar, 12Newman A.M. Gallo N.B. Hancox L.S. Miller N.J. Radeke C.M. Maloney M.A. Cooper J.B. Hageman G.S. Anderson D.H. Johnson L.V. Radeke M.J. Systems-level analysis of age-related macular degeneration reveals global biomarkers and phenotype-specific functional networks.Genome Med. 2012; 4: 16Crossref PubMed Scopus (159) Google Scholar, 13Hunter A. Spechler P.A. Cwanger A. Song Y. Zhang Z. Ying G. Hunter A.K. deZoeten E. Dunaief J.L. DNA methylation is associated with altered gene expression in AMD.Invest Ophthalmol Vis Sci. 2012; 53: 2089-2105Crossref PubMed Scopus (97) Google Scholar, 14Kim E.J. Grant G.R. Bowman A.S. Haider N. Gudiseva H.V. Chavali V.R.M. Complete transcriptome profiling of normal and age-related macular degeneration eye tissues reveals dysregulation of anti-sense transcription.Sci Rep. 2018; 8: 3040Crossref PubMed Scopus (26) Google Scholar Because the postmortem interval and the time from death to conservation may range from at best 1 hour to well over 24 hours,15Blair J.A. Wang C. Hernandez D. Siedlak S.L. Rodgers M.S. Achar R.K. Fahmy L.M. Torres S.L. Petersen R.B. Zhu X. Casadesus G. Lee H. Individual case analysis of postmortem interval time on brain tissue preservation.PLoS One. 2016; 11 ([Erratum appeared in PLoS One 2016;11:e0157209]): e0151615Crossref PubMed Scopus (49) Google Scholar it can be assumed that the RNA is already partially degraded. This scenario is particularly critical given the short half-life of mRNA, which ranges between 10 and 25 minutes.16Wada T. Becskei A. Impact of methods on the measurement of mRNA turnover.Int J Mol Sci. 2017; 18: 2723Crossref Scopus (27) Google Scholar Although retinal tissue can provide stable RNA, if processed within 5 hours after death, the retinal pigment epithelium (RPE), which is one of the tissues primarily affected in AMD, must be processed immediately to prevent its decay.17Malik K.J. Chen C.D. Olsen T.W. Stability of RNA from the retina and retinal pigment epithelium in a porcine model simulating human eye bank conditions.Invest Ophthalmol Vis Sci. 2003; 44: 2730-2735Crossref PubMed Scopus (48) Google Scholar To overcome this limitation, the current study analyzed CNV membranes that had been formalin-fixed and paraffin-embedded (FFPE) immediately after surgical extraction from patients with nAMD. Because RNA degradation in FFPE samples is subjected to chemical degradation, which occurs from 5′ to 3′ ends,18Abdueva D. Wing M. Schaub B. Triche T. Davicioni E. Quantitative expression profiling in formalin-fixed paraffin-embedded samples by affymetrix microarrays.J Mol Diagn. 2010; 12: 409-417Abstract Full Text Full Text PDF PubMed Scopus (78) Google Scholar a 3′-end RNA sequencing approach called Massive Analysis of cDNA Ends (MACE) was used. This approach allows deciphering of expression signatures, assessment of differential expression,19Lange C.A.K. Lehnert P. Boneva S.K. Zhang P. Ludwig F. Boeker M. Hoffmeier K. Horres R. Schlunck G. Reinhard T. Böhringer D. Auw-Haedrich C. Increased expression of hypoxia-inducible factor-1 alpha and its impact on transcriptional changes and prognosis in malignant tumours of the ocular adnexa.Eye (Lond). 2018; 32: 1772-1782Crossref PubMed Scopus (16) Google Scholar and estimation of the compositional patterns of immune and stromal cells in CNV using the computational method xCell.20Aran D. Hu Z. Butte A.J. xCell: digitally portraying the tissue cellular heterogeneity landscape.Genome Biol. 2017; 18: 220Crossref PubMed Scopus (948) Google Scholar This study included 11 patients with nAMD and 18 control patients. For RNA-Seq analysis of human CNV membranes, we retrospectively analyzed four consecutive patients with nAMD who underwent subretinal CNV extraction during vitreoretinal surgery between 1992 and 1999. All CNV membranes were treatment-naive classic CNV and were processed for formalin fixation and paraffin embedding immediately following extraction. Four age-matched FFPE RPE/choroidal specimens obtained in the same period from the macular region of enucleated eyes with ciliary body melanoma served as control samples (Figure 1). For immunohistochemical analysis, the aforementioned FFPE CNV membranes and three randomly chosen and macroscopically normal eyes of three body donors admitted to the Institute of Anatomy were analyzed. Eyes were enucleated in consent with the body donors and secured by contract; no data other than age, sex, weight, and cause of death were disclosed. For protein quantification, seven consecutive patients with nAMD and 11 patients without AMD who underwent vitrectomy for subretinal bleeding or macular pucker between 2017 and 2019 were enrolled. Patients with vitreous hemorrhage, history of previous vitreoretinal surgeries, intraocular inflammations, or concomitant vitreoretinal diseases were excluded. Ethics approval was granted from the Ethics committee of the University of Freiburg, and written informed consent was obtained from each patient. CNV membranes for RNA-Seq were obtained from four patients by using the currently obsolete technique of subretinal CNV extraction.21Bressler N.M. Bressler S.B. Hawkins B.S. Marsh M.J. Sternberg P. Thomas M.A. Submacular Surgery Trials Pilot Study InvestigatorsSubmacular surgery trials randomized pilot trial of laser photocoagulation versus surgery for recurrent choroidal neovascularization secondary to age-related macular degeneration: I. Ophthalmic outcomes submacular surgery trials pilot study report number 1.Am J Ophthalmol. 2000; 130: 387-407Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar Following a 20-gauge vitrectomy, retinotomy was performed temporal to the macula, and the CNV membrane was surgically extracted by using a flexed forceps. The diagnosis was made after a thorough funduscopic examination of both eyes. Only patients exhibiting a CNV associated with AMD changes, such as drusen and RPE alterations, were included in the study. CNV membranes from patients with other diseases (eg, myopia, history of central serous chorioretinopathy) were not included in this study. Control RPE/choroid samples for RNA-Seq were obtained from the macula region of four enucleated eyes with ciliary body melanoma. For protein analysis, undiluted vitreous samples (0.5 to 0.8 mL) were obtained from the mid-vitreous of 7 patients with nAMD and 11 patients with epiretinal membrane (control) undergoing vitrectomy using a vitreous cutter at the start of a vitrectomy, before intraocular infusion of fluid. Plasma samples (2 mL) were collected by venous puncture from the same patients after surgery. The vitreous and plasma samples were transferred directly into sterile plastic tubes on ice. The samples were aliquoted and stored at −80°C until analysis. Formalin fixation and paraffin embedding of CNV membranes and control RPE/choroidal tissue was performed immediately after surgery according to routine protocols, as previously described.19Lange C.A.K. Lehnert P. Boneva S.K. Zhang P. Ludwig F. Boeker M. Hoffmeier K. Horres R. Schlunck G. Reinhard T. Böhringer D. Auw-Haedrich C. Increased expression of hypoxia-inducible factor-1 alpha and its impact on transcriptional changes and prognosis in malignant tumours of the ocular adnexa.Eye (Lond). 2018; 32: 1772-1782Crossref PubMed Scopus (16) Google Scholar Briefly, specimens were fixed immediately after surgery in 4% formalin for 12 hours, dehydrated in alcohol, and finally processed for paraffin embedding. Sections were cut (4-μm thick), mounted on silanized slides, and deparaffinized in xylol-alcohol. After routine histologic staining, each specimen's histologic diagnosis was made by two experienced ophthalmic pathologists. Fifteen FFPE sections (4-μm thick) from each CNV membrane were collected and stored in tubes before RNA extraction. For the initial histopathologic evaluation, FFPE eyes of patients with ciliary body melanoma had been dissected by removing two scleral shells, leaving a central sclerocorneal ring segment. For the current study, the FFPE blocks were melted, and using a dissection microscope, the retina was carefully lifted off the RPE layer. Next, the RPE and choroid were cut laterally with a scalpel, and the central RPE/choroid complex, which easily detaches from the sclera in FFPE eyes, was retrieved and stored in low binding tubes until RNA extraction was performed. RNA isolation from FFPE specimens was performed as previously described.19Lange C.A.K. Lehnert P. Boneva S.K. Zhang P. Ludwig F. Boeker M. Hoffmeier K. Horres R. Schlunck G. Reinhard T. Böhringer D. Auw-Haedrich C. Increased expression of hypoxia-inducible factor-1 alpha and its impact on transcriptional changes and prognosis in malignant tumours of the ocular adnexa.Eye (Lond). 2018; 32: 1772-1782Crossref PubMed Scopus (16) Google Scholar Briefly, total RNA was extracted from FFPE samples by using the Quick-RNA FFPE Kit (Zymo Research, Irvine, CA). Following DNase I digestion using the Baseline-ZERO Kit (Epicentre Technologies, Madison, WI), the RNA concentration was quantified by using the Qubit RNA HS Assay Kit on a Qubit Fluorometer (Thermo Fisher Scientific, Waltham, MA). RNA quality was determined via the RNA Pico Sensitivity Assay on a LabChip GXII Touch (PerkinElmer, Waltham, MA). The fragment size of all RNA samples ranged between 120 and 150 bp. The preparation of MACE libraries was performed by GenXPro GmbH (Frankfurt, Germany) using 1 μg of total RNA as previously described.19Lange C.A.K. Lehnert P. Boneva S.K. Zhang P. Ludwig F. Boeker M. Hoffmeier K. Horres R. Schlunck G. Reinhard T. Böhringer D. Auw-Haedrich C. Increased expression of hypoxia-inducible factor-1 alpha and its impact on transcriptional changes and prognosis in malignant tumours of the ocular adnexa.Eye (Lond). 2018; 32: 1772-1782Crossref PubMed Scopus (16) Google Scholar Briefly, eight barcoded libraries (four CNV membranes and four RPE/choroidal control tissues) were sequenced simultaneously on the NextSeq 500 (Illumina, San Diego, CA) with 1 × 75 bp. The sequencing data reported in the current article have been deposited in the Gene Expression Omnibus database (https://www.ncbi.nlm.nih.gov/geo; accession number GSE146887). Vitreous and plasma calprotectin levels were measured on vitreous and plasma samples from patients with nAMD undergoing vitrectomy for subretinal macular hemorrhage. Patients with macular pucker undergoing vitrectomy for epiretinal peeling served as control subjects. Frozen vitreous and plasma samples were thawed, and calprotectin levels were measured by using the Human S100A8/S100A9 Heterodimer Quantikine ELISA Kit according to the manufacturer's protocol (R&D Systems, Minneapolis, MN). Human choroidal CNV membranes and control FFPE samples were cut into 5-μm sections and deparaffinized according to standard protocol. Sections were blocked with 5% normal goat serum in phosphate-buffered saline (PBS)/Triton-X 0.03% for 60 minutes at room temperature. Primary anti-S100 calcium-binding protein A9 (S100A9) antibody (sc-376772; Santa Cruz Biotechnology, Dallas, TX) was incubated in a 1:200 dilution in PBS containing 0.5% bovine serum albumin and 0.03% Triton-X overnight at 4°C. No primary antibody served as negative control. After extensive washing with PBS/Triton-X 0.03%, sections were incubated with an Alexa Fluor 647–labeled goat anti-mouse secondary antibody 1:500 in PBS/Triton-X 0.03% (Thermo Fisher Scientific) at room temperature for 1 hour in the dark. After washing at least three times with PBS/Triton-X 0.03% and Aqua Dest, autofluorescence was quenched with Sudan Black B (0.3% in 70% ethanol overnight, stirring at 50°C, filtrated twice, and incubated for 5 minutes at 37°C on the slides). Nuclei were counterstained with DAPI 1:10,000 for 5 minutes, washed three times with Aqua Dest, and mounted in Fluorescence Mounting Medium (Agilent Dako, Santa Clara, CA). Slides were imaged by using an Olympus BX-40 with a color camera (XM10, Olympus, Tokyo, Japan). Sequencing data were uploaded to and analyzed on the Galaxy web platform (usegalaxy.eu) as previously described.22Afgan E. Baker D. Batut B. van den Beek M. Bouvier D. Čech M. Chilton J. Clements D. Coraor N. Grüning B.A. Guerler A. Hillman-Jackson J. Hiltemann S. Jalili V. Rasche H. Soranzo N. Goecks J. Taylor J. Nekrutenko A. Blankenberg D. The Galaxy platform for accessible, reproducible and collaborative biomedical analyses: 2018 update.Nucleic Acids Res. 2018; 46: W537-W544Crossref PubMed Scopus (1414) Google Scholar,23Boeck M. Thien A. Wolf J. Hagemeyer N. Laich Y. Yusuf D. Backofen R. Zhang P. Boneva S. Stahl A. Hilgendorf I. Agostini H. Prinz M. Wieghofer P. Schlunck G. Schlecht A. Lange C. Temporospatial distribution and transcriptional profile of retinal microglia in the oxygen-induced retinopathy mouse model.Glia. 2020; 68: 1859-1873Crossref Scopus (24) Google Scholar Quality control was performed with FastQC Galaxy version 0.72 (Barbraham Bioinformatics, Babraham, UK; http://www.bioinformatics.babraham.ac.uk/projects/fastqc, last accessed August 5, 2019). Reads were mapped to the human reference genome (hg38, Galaxy built-in reference genome) with RNA STAR Galaxy Version 2.6.0b-224Dobin A. Davis C.A. Schlesinger F. Drenkow J. Zaleski C. Jha S. Batut P. Chaisson M. Gingeras T.R. STAR: ultrafast universal RNA-seq aligner.Bioinformatics. 2013; 29: 15-21Crossref PubMed Scopus (15795) Google Scholar (default parameters) using the GENCODE annotation file (GENCODE 31, release June 2019, downloaded on August 5, 2019; https://www.gencodegenes.org/human/releases.html).25Frankish A. Diekhans M. Ferreira A.-M. Johnson R. Jungreis I. Loveland J. GENCODE reference annotation for the human and mouse genomes.Nucleic Acids Res. 2019; 47: D766-D773Crossref PubMed Scopus (988) Google Scholar Reads mapped to the human reference genome were counted by using featureCounts Galaxy Version 1.6.426Liao Y. Smyth G.K. Shi W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features.Bioinformatics. 2014; 30: 923-930Crossref PubMed Scopus (7013) Google Scholar (default parameters) using the aforementioned annotation file. The output of featureCounts was imported to RStudio version 1.2.1335) (R version 3.5.3). Gene symbols and gene types were determined based on Ensembl release 97 (July 2019) (Human genes, GRCh38.p12, downloaded on August 31, 2019).27Zerbino D.R. Achuthan P. Akanni W. Amode M.R. Barrell D. Bhai J. et al.Ensembl 2018.Nucleic Acids Res. 2018; 46: D754-D761Crossref PubMed Scopus (1614) Google Scholar Genes with 0 reads were removed from the analysis. Based on Ensembl gene types, genes were filtered for protein coding genes and small RNAs (long noncoding RNA, miRNA, miscellaneous RNA, small conditional RNA [scRNA], small nucleolar RNA, snRNA, and small RNA). Genes coding for ribosomal or mitochondrial proteins were removed from the analysis. After principal component analysis, differential gene expression with Benjamini-Hochberg adjusted (adj.) P values was analyzed by using the R package DESeq2 version 1.22.2 (default parameters).28Love M.I. Huber W. Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2.Genome Biol. 2014; 15: 550Crossref PubMed Scopus (25756) Google Scholar Transcripts with log2fold change (log2FC) >2 or <−2 and Benjamini-Hochberg adj. P < 0.05 were considered as differentially expressed genes (DEGs). Heatmaps were created with the R package ComplexHeatmap 1.20.0.29Gu Z. Eils R. Schlesner M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data.Bioinformatics. 2016; 32: 2847-2849Crossref PubMed Scopus (2034) Google Scholar Data visualization with volcano plots was performed by using the ggplot2 package.30Wickham H. ggplot2: Elegant Graphics for Data Analysis, ed 2. Springer International Publishing, Basel, Switzerland2016Crossref Google Scholar Gene enrichment analysis and its visualization with dotplots and cnetplots were performed by using the R package clusterProfiler 3.10.1.31Yu G. Wang L.G. Han Y. He Q.Y. clusterProfiler: an R package for comparing biological themes among gene clusters.OMICS. 2012; 16: 284-287Crossref PubMed Scopus (8447) Google Scholar The 10 most significant Gene Ontology (GO) terms were visualized by using dotplots. A cnetplot was used to illustrate the four most disease-associated GO terms and their associated DEGs. To determine the cellular profile within the CNV microenvironment, the computational method xCell, which estimates the abundance scores of 64 immune and stromal cell types, including hematopoietic progenitors, epithelial cells, extracellular matrix (ECM) cells, and adaptive and innate immune cells, was applied.20Aran D. Hu Z. Butte A.J. xCell: digitally portraying the tissue cellular heterogeneity landscape.Genome Biol. 2017; 18: 220Crossref PubMed Scopus (948) Google Scholar Although the final xCell abundance scores cannot be directly interpreted as cell fractions, they showed high correlation with the true cell proportions.32Finotello F. Trajanoski Z. Quantifying tumor-infiltrating immune cells from transcriptomics data.Cancer Immunol Immunother. 2018; 67: 1031-1040Crossref PubMed Scopus (154) Google Scholar For xCell analysis, transcripts per million were calculated based on the output of featureCounts (assigned reads and feature length), as previously described.33Wagner G.P. Kin K. Lynch V.J. Measurement of mRNA abundance using RNA-seq data: RPKM measure is inconsistent among samples.Theory Biosci. 2012; 131: 281-285Crossref PubMed Scopus (938) Google Scholar Enrichment scores were compared between different groups by using U-tests. Cell types with P < 0.05 were considered to be significantly differentially enriched. For MACE RNA-Seq analysis, four eyes of four consecutive patients with nAMD were included; four eyes of four patients with ciliary body melanoma requiring enucleation served as control samples. Three macroscopically normal eyes of three body donors served as control samples for immunohistochemical analysis. Patients' mean age was 79.5 years (range, 72 to 83 years) in the nAMD group, 73.0 years (range, 66 to 75 years) in the control group, and 82.7 years (range, 78 to 89 years) in the body donor group. CNV membranes were stored in paraffin for a mean of 7.264 days (range, 4346 to 7848 days) before RNA-Seq, and control samples were stored for 6.197 days (range, 3410 to 8057 days). Routine hematoxylin and eosin histology confirmed the diagnosis of CNV and revealed a microscopically intact retinal, RPE, and choroidal histology in the control and body donor samples (Figure 1). For protein quantification, seven consecutive patients with nAMD and 11 patients without AMD who underwent vitrectomy for subretinal bleeding or macular pucker were enrolled. The mean age of patients with nAMD was 77 years (range, 72 to 90 years); the mean age of the control patients was 70 years (range, 64 to 76 years). None of the patients exhibited concomitant macular disease, and only two patients with nAMD had received intravitreal bevacizumab injections 4 months before vitrectomy. Using MACE RNA-Seq, a mean number of 907.233 reads (range, 297,004 to 1.744.933) in CNV samples and 4.889.361 reads (range, 348.446 to 11.898.565) in control samples were obtained. Because the RNA-Seq profile is a mixture of RNA from multiple cell types that include resident and infiltrating cells, bulk RNA cell type enrichment analysis was first used to recover the identity of the cell types found in human CNV, using the gene signature expression–based cell type enrichment tool xCell.20Aran D. Hu Z. Butte A.J. xCell: digitally portraying the tissue cellular heterogeneity landscape.Genome Biol. 2017; 18: 220Crossref PubMed Scopus (948) Google Scholar Cell type enrichment scores across 64 immune and stromal cell types were obtained for both CNV and control specimens (Figure 2A and Supplemental Figure S1). These data show that immune scores were overall higher in CNV samples (0.025 ± 0.029) compared with the scores of healthy RPE/choroid samples (0.014 ± 0.013; P = 0.27). Among the different immune cell subpopulations, CD4 T-effector memory phenotype cells, CD4+ naive T cells, M1 macrophages, and natural killer T cells were significantly increased in CNV samples compared with levels observed in healthy control samples (P < 0.05) (Figure 2A). In contrast, T helper cells type 2 and basophils were significantly decreased in CNV samples compared with control samples. Of note, subpopulations of M2 macrophages were similar in CNV samples compared with control samples (Supplemental Figur