Skip to main content
NCBI home page
Journal of Extracellular Vesicles logoLink to Journal of Extracellular Vesicles
. 2023 Oct 19;12(10):12373. doi: 10.1002/jev2.12373

Extracellular vesicles from retinal pigment epithelial cells expressing R345W‐Fibulin‐3 induce epithelial‐mesenchymal transition in recipient cells

Mi Zhou 1, Yuanjun Zhao 1, Sarah R Weber 1, Christopher Gates 2, Nicholas J Carruthers 2, Han Chen 3, Xiaoming Liu 4, Hong‐Gang Wang 4, Michael Ford 5, Matthew T Swulius 6, Alistair J Barber 1, Stephanie L Grillo 1, Jeffrey M Sundstrom 1,✉
PMCID: PMC10585439  PMID: 37855063

Abstract

We have shown previously that expression of R345W‐Fibulin‐3 induces epithelial‐mesenchymal transition (EMT) in retinal pigment epithelial (RPE) cells. The purpose of the current study was to determine if extracellular vesicles (EVs) derived from RPE cells expressing R345W‐Fibulin‐3 mutation are sufficient to induce EMT in recipient cells. ARPE‐19 cells were infected with luciferase‐tagged wild‐type (WT)‐ Fibulin‐3 or luciferase‐tagged R345W‐Fibulin‐3 (R345W) using lentiviruses. EVs were isolated from the media by ultracentrifugation or density gradient ultracentrifugation. Transmission electron microscopy and cryogenic electron microscopy were performed to study the morphology of the EVs. The size distribution of EVs were determined by nanoparticle tracking analysis (NTA). EV cargo was analysed using LC‐MS/MS based proteomics. EV‐associated transforming growth factor beta 1 (TGFβ1) protein was measured by enzyme‐linked immunosorbent assay. The capacity of EVs to stimulate RPE migration was evaluated by treating recipient cells with WT‐ or R345W‐EVs. The role of EV‐bound TGFβ was determined by pre‐incubation of EVs with a pan‐TGFβ blocking antibody or IgG control. EM imaging revealed spherical vesicles with two subpopulations of EVs: a group with diameters around 30 nm and a group with diameters over 100 nm, confirmed by NTA analysis. Pathway analysis revealed that members of the sonic hedgehog pathway were less abundant in R345W‐ EVs, while EMT drivers were enriched. Additionally, R345W‐EVs had higher concentrations of TGFβ1 compared to control. Critically, treatment with R345W‐EVs was sufficient to increase EMT marker expression, as well as cell migration in recipient cells. This EV‐increased cell migration was significantly inhibited by pre‐incubation of EVs with pan‐TGFβ‐neutralising antibody. In conclusion, the expression of R345W‐Fibulin‐3 alters the size and cargo of EVs, which are sufficient to enhance the rate of cell migration in a TGFβ dependent manner. These results suggest that EV‐bound TGFβ plays a critical role in the induction of EMT in RPE cells.

Keywords: EMT, extracellular vesicles, Fibulin‐3, RPE, TGFβ

1. INTRODUCTION

RPE cells form a monolayer of highly polarised cells that lie posterior to the neuroretina and are essential for maintaining the health and function of the adjacent photoreceptors. Loss of terminal differentiation and acquisition of a mesenchymal cell phenotype in RPE cells has been shown to occur in degenerative retinal diseases (Ghosh et al., 2018; Zhou et al., 2020). Early in RPE dysfunction, RPE cells create drusen and sub‐RPE deposits. The hallmark clinical lesions of sub‐RPE lipoprotein deposits and pigmentary changes appear within the macula in diseases such as age‐related macular degeneration (AMD) and Doyne honeycomb macular dystrophy (Stone et al., 1999; Vincent et al., 2012).

An inherited macular degeneration in which RPE dysfunction plays a significant role is Doyne honeycomb macular dystrophy, which results from a single arginine‐to‐tryptophan point mutation, R345W, in Fibulin‐3 (Marmorstein, 2004; Narendran et al., 2005). Fibulin‐3 is an extracellular matrix protein that contains six epidermal growth factor (EGF)‐like domains followed by a fibulin domain (Zhang & Marmorstein, 2010). The R345W mutation causes Fibulin‐3 misfolding, poor Fibulin‐3 secretion, and activation of the unfolded protein response (UPR) (Hulleman & Kelly, 2015; Marmorstein et al., 2002).

Previous studies have shown that RPE cells are capable of undergoing EMT (Ghosh et al., 2018; Zhou et al., 2020; Sripathi, Hu, Liu, et al., 2021; Tamiya & Kaplan, 2016). Recently, we have shown that overexpression of R345W‐Fibulin‐3 in RPE cells attenuates RPE differentiation, and induces EMT in RPE cells (Zhou et al., 2020). Extracellular vesicles (EVs) play a critical role in cell‐cell communication and modulate cellular differentiation (Yuyama et al., 2008; Weber et al., 2020). In numerous tissues, EVs contribute to the regulation of epithelial‐mesenchymal transition (EMT), including in the lungs, breasts, liver, and brain (Vella, 2014; van de Vlekkert et al., 2019). For example, myofibroblast‐derived EVs are sufficient to induce normal fibroblasts to become myofibroblasts that possess mesenchymal features by upregulating transforming growth factor beta (TGFβ) pathways and EMT drivers (van de Vlekkert et al., 2019). Recent studies have shown that alterations in EV size and cargo are dependent upon the secretory mechanisms and phenotypic status of their parental cells (Zhang et al., 2018). While RPE cells have been shown to secrete EVs (Klingeborn et al., 2017; Shah et al., 2018), the role of EVs in RPE dysfunction remains to be determined.

In this study, we investigated whether EVs derived from RPE cells expressing R345W‐Fibulin‐3 mutation are sufficient to induce EMT in recipient cells. EVs were isolated from ARPE‐19 cells expressing either WT (WT‐ARPE) or R345W (R345W‐ARPE) Fibulin‐3; labelled as WT‐ARPE‐EVs or R345W‐ARPE‐EVs, respectively. EV size, cargo and function were examined in ARPE cells. We found that RPE‐derived EVs are sufficient to induce EMT in RPE cells in a TGFΒ‐dependent manner.

2. METHODS

2.1. ARPE‐19 cell culture

Naïve ARPE‐19 cells (CRL‐2302, p19) were purchased from American Type Culture Collection (ATCC), which originated from a 19‐year‐old male human. ARPE‐19 cell lines were maintained in DMEM (Dulbecco's Modified Eagle's Medium)/Hams F‐12 (10‐092‐CV, Corning) supplemented with 10% fetal bovine serum (100106, Gemini) and 100 I.U. Penicillin; 100 μg/mL Streptomycin (30‐002‐CI, Corning) at 37°C and 5% CO2. Viability and phenotype were assessed by a microplate fluorometer and microscopy routinely. Trypan blue (25‐900‐CI, 0.4%, Corning) was used to stain cells with 1:1 dilution. Cell viability was confirmed to be at least 95% before cells were used for analysis. ARPE‐19 Tet‐On cells with Lentiviral GLuc‐tagged WT‐ or R345W‐Fibulin‐3 were described previously, named as WT‐ARPE or R345W‐ARPE (Hulleman et al., 2013; Zhou et al., 2020). Inserted genes have inducible expression in the presence of doxycycline (1 μg/mL, Dox, D9891, MilliporeSigma, Burlington, MA, USA).

2.2. Extracellular vesicle isolation (differential centrifugation)

EVs used for characterisation and Western blot studies were isolated by differential centrifugation. Briefly, ARPE‐19 Tet‐On WT or R345W Fibulin‐3 eGLuc2 cells were seeded (5 × 106 cells per 100 mm dish [Corning]) initially in DMEM/F12 media supplemented with 10% FBS and 100 I.U. Penicillin; 100 μg/mL Streptomycin. When cells reached 60%–70% confluency, they were rinsed 3 times with PBS containing Ca2+ and Mg2+, and serum‐free DMEM/F12 with doxycycline (1 μg/mL) was added. Cells were in a monolayer and were positive for tight junction proteins (ZO‐1) and polygonal morphology one day following doxycycline treatment (Figure S1). Two days later, EVs were collected from conditioned media after cells were grown in FBS‐depleted media (serum‐free). Cell culture media (100 mL) was filtered using a 0.22‐μm filter and then concentrated to 1/10th of its original volume (Amicon Centrifugal Filter Ultra‐15 mL 3K, UFC900324, Millipore Sigma). Concentrated cell culture media was ultracentrifuged at 100,000 g for 17 h and washed using HBS at 100,000 g for 5 h. EV pellets were re‐suspended in HEPES based saline (HBS, 150 mM NaCl, 20 mM HEPES, pH 7.4, 0.22 μm filtered) buffer. A schematic of the protocol is shown (Figure 1a). Fresh EVs were stored at 4°C for use within 1 week. Unused EVs were frozen at −20°C and kept for less than 6 months with ≤ 3 freeze/thaw cycles.

FIGURE 1.

FIGURE 1

Open in a new tab

Methods for EV purification. (a) Conventional filtration and ultracentrifugation was used to isolate EVs for Western blot and other studies. (b) Sucrose cushion and iodixanol buoyant gradient ultracentrifugation was used to isolate EVs for proteomics studies. EVs, extracellular vesicles; HBS, HEPES‐buffered saline; PBS, phosphate buffered saline.

2.3. Extracellular vesicle isolation (iodixanol buoyant density gradient ultracentrifugation)

EVs used in proteomics studies were isolated using iodixanol buoyant density gradient ultracentrifugation described previously in order to isolate a more homogeneous fraction (Choi & Gho, 2016; Zhou et al., 2020). ARPE‐19 Tet‐On WT or R345W Fibulin‐3 eGLuc2 cells were seeded in 100 mm dishes (353003, Corning) initially in DMEM/F12 media supplemented with 10% FBS and 100 I.U. Penicillin; 100 μg/mL Streptomycin. Cells were split into 145 mm dishes (639160, Greiner) by trypsinization when they reached 80%–90% confluency. After another three rounds of expansion, there were 84 and 112 dishes (15 mL media per dish) of WT and R345W cells respectively. When cells reached 60%–70% confluency, media was removed, cells were rinsed 3 times with PBS containing Ca2+ and Mg2+ for 10 mL each time, and serum‐free DMEM/F12 with Doxycycline (1 μg/mL) was added for 15 mL in each dish. Two days later, media was collected (∼2 × 107 cells/145 mm dish) and filtered using vacuum filtration system with 0.22 μm polyethersulfone (PES) membrane (10040‐448, VWR), and concentrated using Centricon Plus‐70 Centrifugal Filter Device (100K MWCO, UFC710008, Millipore Sigma) by spinning at 3000 g for 6–8 min at 4°C in Eppendorf 5804R centrifuge and A‐4‐62 bucket rotor. Fifty millilitres of media were loaded each time and the concentrate (2.5–5 mL) was recovered after spinning every 150 mL media by spinning the device invertedly with collection cup at 1000 g for 2 min at 4°C. After all the media was concentrated, the filter device was rinsed 3 times with HEPES based saline (HBS, 150 mM NaCl, 20 mM HEPES, pH7.4, 0.22 μm filtered). Each time, 5 mL HBS was added and spun for 5 min at 3000 g at 4°C. The rinsed HBS was collected, combined with the concentrated media and adjusted to 75 mL using HBS. Cell viability assay was performed by trypsinization and counting using hemocytometer (3100, Hausser Scientific) after Trypan Blue (25‐900‐CI, Corning) staining. There were 1.5–2.0 × 107 viable cells from each 145 mm dish and the percentage of dead cells was 1%–3%. Concentrated culture media was loaded over 1 mL of 0.8 M sucrose buffer (0.8 M sucrose, 150 mM NaCl, 20 mM HEPES, pH7.4) and 0.5 mL of 2.0 M sucrose buffer (2.0 M sucrose, 150 mM NaCl, 20 mM HEPES, pH7.4) and spun at 100,000 g for 2 h at 4°C. Two mL of interface between 0.8 M and 2.0 M sucrose cushion diluted 1:5 in HBS, loaded over 0.35 mL of 0.8 M sucrose buffer and 0.15 mL of 2.0 M sucrose and spun at 100,000 g for 2 h at 4°C. The interface (0.5 mL) between 0.8 M and 2.0 M sucrose cushion buffers was collected and stored at 4°C. The sample was mixed with 1.42 mL HBS and 2.88 mL 50% iodixanol prepared from 5 volumes of OptiPrep Density Gradient Medium (D1556, Millipore Sigma) and 1 volume of the buffer containing 0.25 M Sucrose, 900 mM NaCl and 120 mM HEPES (pH = 7.4), resulting in 30% iodixanol solution and overlaid with 3 mL of 20% and 2.5 mL 5% iodixanol. Finally, the sample was spun at 200,000 g for 2 h at 4°C and the pellet collected from the 3rd fraction (Pgrad).

Ten fractions were collected, and each 1 mL fraction was diluted to 10 mL with HBS and spun at 150,000 g for 3 h at 4°C. The supernatant was removed and the pellet was re‐suspended in 50 μL HBS. Fifteen μL of each sample were stored at 4°C for protein assay, electrophoresis, and TEM. A schematic of the protocol adapted from (Choi & Gho, 2015) is shown (Figure 1b). EV samples for proteomics were stored at −80°C for 2 weeks until shipment to MS Bioworks for mass spectrometry analysis.

2.4. Sample preparation and proteomic analysis

EVs for proteomics analysis were isolated using iodixanol buoyant density gradient ultracentrifugation as described above. A total of 20 μg of protein was processed using sodium dodecyl sulfate‐polyacrylamide gel electrophoresis (SDS‐PAGE) with a 4%–12% Bis‐Tris Mini‐gel (Thermo Fisher) and the 3‐(N‐morpholino) propanesulfonic acid (MOPS) buffer system. Following electrophoresis, the gel lanes were excised into 40 equal‐sized segments and processed by in‐gel digestion with trypsin (DigiLab). Briefly, fragments were washed with 25 mM ammonium bicarbonate followed by acetonitrile. Samples were reduced with 10 mM dithiothreitol at 60°C followed by alkylation with 50 mM iodoacetamide at room temperature. Digestion was performed with Sequencing Grade Trypsin (Promega, Madison, WI, USA) at 37°C for 4 h. Samples were quenched with formic acid, and the supernatant was analysed directly without further processing.

Half of each gel digest was analysed by nano LC‐MS/MS with a Waters NanoAcquity high performance liquid chromatography system interfaced to a ThermoFisher Q Exactive mass spectrometer. Peptides were loaded on a trapping column and eluted over a 75 μm analytical column at 350 nL/min. Both columns were packed with Luna C18 resin (Phenomenex). The mass spectrometer was operated in data‐dependent mode, with the Orbitrap operating at 60,000 and 17,500 full widths at half maximum (FWHM) for MS and MS/MS, respectively. The 15 most abundant ions were selected for MS/MS. The mass spectra were processed using Mascot set to search the SwissProt Human database (forward and reverse appended with common contaminant proteins) set with Trypsin/P as the enzyme. The resultant Mascot DAT files were parsed into Scaffold (Proteome software) for validation, filtering, and creation of non‐redundant protein lists for samples. Data were filtered using a 1% protein and peptide false discovery rate (FDR) and requiring at least two unique peptides per protein.

2.5. Bioinformatics

Proteomic data analysis was conducted in R (v4.1.1). Bos Taurus proteins (from bovine serum) and uncharacterised proteins were removed from the analysis. Spectral counts were used for protein quantification. EV quality was assessed by comparing the subcellular origin profile of our EV proteins with other high quality EV preparations (data not shown) (Burton et al., 2023; Burton et al., 2022). Known EV and non‐EV marker proteins (Askeland et al., 2020) were also used for quality assurance. Marker proteins were highlighted on a plot of protein abundance versus abundance rank for the pooled data from both EV samples and enrichment was assessed visually.

This analysis didn't use replicate analyses but relied on a single quantitative value in R345W and in WT EV for protein quantification and differential expression analysis. The log differential expression (Log2DE) was calculated by determining the differential expression of R345W and WT proteins, then taking the log base 2 of that value. Log2DE was submitted to iPathway Guide for functional enrichment analysis. In addition, EdgeR (Robinson et al., 2010) was used to generate differential expression statistics using a nominal dispersion value (variance coefficient of 0.4) in lieu of biological variance. The resulting t‐statistics were submitted to gene set analysis using PIANO (Väremo et al., 2013) against EMT gene sets downloaded from the EMTome (http://www.emtome.org) (Vasaikar et al., 2021). PIANO was set to use the mean t‐statistic as the gene set statistic and gene set sampling to test significance. Results from directional testing were used for this analysis.

They were generated using EdgeR (Robinson et al., 2010) with a variance coefficient of 0.4. GSA was carried out using PIANO (Väremo et al., 2013) against EMT gene sets downloaded from the EMTome (http://www.emtome.org) (Vasaikar et al., 2021).

2.6. Nanoparticle tracking analysis

EV suspensions (n = 6 per group) were diluted to 1 mL (1:50 to 1:1000) with particle‐free water. Each sample was loaded by syringe pump into the NanoSight NS300 (Malvern Instruments Ltd, Malvern, Worcestershire, UK) set in scatter mode (blue 488 nm laser) and five 60‐s movies were generated. Movies were recorded at camera level 11 (Shutter: 890; Gain 146) and 13 (Shutter: 1232; Gain: 219). The size distribution and concentration of particles were analysed and images were acquired using NanoSight software version 3.2 (Malvern Instruments Ltd).

2.7. Cell migration assay

Cell migration assays (n = 6 wells, assay repeated twice) were performed as described previously (Zhou et al., 2020). Briefly, ARPE‐19 cells were cultured in 96‐well ImageLock Microplates (Essen Bioscience Inc., Ann Arbor, MI, USA) to confluence. A 96‐pin WoundMaker (Essen Bioscience Inc., Ann Arbor, MI, USA) was used to make scratches. The wells were then washed with PBS to remove cell debris. EVs were added to the cell culture media. The EV concentration (1 × 1010 vesicles/mL) was chosen according to a previously published study (van de Vlekkert et al., 2019), and EV uptake in ARPE‐19 cells was confirmed as described in the previous section. To determine whether EVs derived from RPE cells induce EMT via TGFβ signaling, ARPE‐19 cells were treated with WT‐ARPE‐EVs or R345W‐ARPE‐EVs after the wounds were created. EVs were pre‐incubated with pan‐TGFβ‐neutralising antibody (10 μg/mL; Cat# MAB1835‐100, Quantikine; R&D Systems, Minneapolis, MN, USA) or the isotype IgG (10 μg/mL; Cat # IX2721082, Quantikine; R&D Systems, Minneapolis, MN, USA) as a negative control prior to scratch assays. EVs without pre‐incubation with antibody or IgG served as an additional control. Relative wound recovery time were monitored. Wound images were acquired automatically and data were analysed by the IncuCyte software system (Essen Bioscience Inc., Ann Arbor, MI, USA).

2.8. Real‐Time PCR

Real‐time (RT)‐PCR was performed as described previously (Zhou et al., 2020). In brief, total RNA was extracted from samples using the AllPrep DNA/RNA/Protein Mini Kit (80004, Qiagen Sciences, Inc.), and 0.25 μg of total RNA was reverse transcribed by SuperScript IV First‐Strand cDNA Synthesis System (18091050, Thermo Fisher Scientific) with a mixture of oligo dT and random hexamer. Amplification of cDNA was performed with optimised PCR primers designed by Primer3 software (https://bioinfo.ut.ee/primer3/) and synthesised by Integrated DNA Technologies (Coralville, Iowa) in SYBR Green master (Rox) system (04913850001, Roche, Basel, Switzerland) on QuantStudio 3 RT‐PCR Systems (ThermoFisher). The volume of PCR reaction was 10 μL, and the procedure included hold stage at 95°C for 10 min; PCR stage at 95°C for 15 s and at 60°C for 1 min for 40 cycles; and melt curve stage at 95°C for 15 s, 60°C for 1 min, and 95°C for 15 s. PCR primers were validated by amplicon size analysis on agarose gel, melting curve and standard curve of RT‐PCR products. Detailed information on PCR primers used in this study are provided in Table S1. The relative quantity of mRNA was normalised to glyceraldehyde 3‐phosphate dehydrogenase (GAPDH) using the comparative 2−∆∆Ct method. GAPDH mRNA expression was found to be consistent between different EV groups (Figure S2).

2.9. TGFβ1 enzyme‐linked immunosorbent assays (ELISAs)

Cell culture media (serum‐free) was collected from ARPE‐19 cells expressing either WT‐Fibulin‐3 or R345W‐Fibulin‐3. EVs from media were isolated by using conventional ultracentrifugation, and media left over after EVs were centrifuged out were collected as EV‐depleted media. TGFβ1 ELISAs (Quantikine; R&D Systems, Minneapolis, MN, USA) were performed on media, EV‐depleted media, and EVs from both groups according to kit instructions. Fifty microliters of media, EV, or EV‐depleted media was used for loading on each test. Before the analysis, dilution factor was first determined to ensure all the values fell within the detection range of the assay. To activate latent TGFβ1 to immunoreactive TGFβ1 detectable by the Quantikine TGFβ1 immunoassay, all samples were incubated with 1 N HCL for 10 min and neutralised by 1.2 N NaOH/0.5 M HEPES. Activated samples were assayed immediately after activation. Optical densities were determined within 30 min. with a SpectraMax 190 microplate reader (Molecular Devices, San Jose, CA, USA) at 450 nm with wavelength correction at 570 nm. All TGFβ1 ELISA experiments were run in technical duplicates and biological triplicates.

2.10. Western blot analysis

Cells were lysed in RIPA buffer (R0278, Millipore Sigma) and sonicated for 8 pulses. Protein concentrations in cell lysate and EVs were quantified using the DC Protein Assay (5000116, Bio‐Rad). Total protein (2.5 μg) for each sample was mixed with LDS sample buffer (NP0007, Thermo Fisher Scientific) and 0.1 M dithiothreitol (except when probing with CD63 or CD81 antibodies) and heated at 70°C for 15 min, then loaded on NuPAGE 4%–12% gradient (NP0323BOX, Thermo Fisher Scientific) or 12% (NP0343BOX, Thermo Fisher Scientific) Tris‐Bis Gel and ran for 1.5 h at 150 V in MOPS SDS Running buffer (NP0001, Thermo Fisher Scientific). A Coomassie stain of the gel was used as the reference to show the protein content between WT and R345W‐ARPE‐EVs (Figure S3). For Coomassie staining, gels were incubated in staining buffer (0.25% Coomassie brilliant blue R250 (97063‐820, VWR), 45% methanol, 10% acetic acid) overnight at room temperature, and then de‐staining buffer (45% methanol, 10% acetic acid) and scanned using FluorChem M imaging System (ProteinSimple, San Jose, CA). For Western blot, samples were transferred from gel to 0.22 μm nitrocellulose membrane (EP2HY00010, Fisher) at 100 V for 1.5 h in Transfer buffer (NP0006‐1) with 20% methanol and 0.05% SDS. The membrane was blocked at room temperature for 1 h in blocking buffer (5% non‐fat milk in TBST (10 mM Tris, 150 mM NaCl, 0.05% Tween‐20, pH 7.2)). The membrane was then incubated with primary antibodies (Table S2) diluted in blocking buffer at 4°C overnight, washed in TBST 4 times for 10 min each time, incubated with secondary antibody diluted (1: 5000) in blocking buffer for 1.5 h at room temperature, and washed another 4 times in TBST for 10 min. each time. In this study, we used either alkaline phosphatase conjugated goat anti‐rabbit IgG (H+L) (111‐055‐144, Jackson ImmunoResearch Laboratories, Inc.) or anti‐mouse IgG+IgM (H+L) (115‐055‐068, Jackson ImmunoResearch Laboratories, Inc.) secondary antibodies and Vistra ECF Substrate (RPN5785, GE Healthcare) to detect signals and then Typhoon 9400 (GE Healthcare) to scan the image; or IRDye 800CW goat anti‐rabbit IgG (H+L) (925‐32211, LI‐COR) or anti‐mouse IgG (H+L) (925‐32210, LI‐COR) secondary antibodies to detect signals and Odyssey CLx Imaging system (LI‐COR) to scan the image using software Image Studio Ver 5.2 (LI‐COR).

2.11. Transmission electron microscopy (TEM)

Negative staining was conducted as follows. Briefly, the resuspended pellet was fixed in 2% paraformaldehyde. The fixed vesicles (10 μL) were deposited on Formvar carbon‐coated TEM grids and incubated for 2 min. The WT‐ARPE‐EVs were then stained with uranyl acetate solution and air dried. Vesicles were observed using the JEM1400 TEM (JEOL Ltd., Tokyo, Japan). TEM images were from gradient ultracentrifugation isolation. WT‐ARPE‐EV had 11 images and R345W‐ARPE‐EV had 10 images. Magnification was 10,000x.

2.12. Cryogenic electron microscopy (Cryo‐EM)

Four microliters of WT‐ARPE‐EVs (1 × 1010 EVs) was pipetted onto a freshly glow‐discharged Quantifoil R2/2 holey carbon grid. Grids were hand blotted from behind using Whatman paper #1 and plunged into liquid ethane using a Mark IV Vitrobot (Thermo Fisher Scientific) for vitrification. Samples were then transferred under liquid nitrogen into a cryo side‐entry holder (Gatan) and loaded into a JEM 2100 cryo TEM (JEOL) operating at 200 k. Holes containing vesicles were targeted for data collection. A nominal magnification of 40,000x was used, which corresponded to a pixel size of 0.29 nm/pixel on our 4k ultrascan CCD (Gatan).

2.13. Statistics

Power analyses were conducted using Graphpad StatMate software 2.0 (Graphpad, San Diego, CA, USA), Type I error (alpha) was set to 0.05 and type II error (beta) was set to 80%. Based on the power analysis, a group size of n = 6 will provide more than 99% power to confirm the results. Data are presented as mean ± standard error of the mean (SEM). Statistical analysis was performed by using Prism 8 (GraphPad, Inc., La Jolla, CA, USA). One‐way analysis of variance (ANOVA) was used to determine differences among multiple groups, where individual differences were tested using the Tukey post‐test. Where appropriate, an unpaired Student's t‐test was used to compare differences between groups. A p‐value of < 0.05 was considered statistically significant.

3. RESULTS

3.1. Expression of R345W‐Fibulin‐3 alters EV size in RPE cells

ARPE‐19 cells were transfected with luciferase‐tagged WT‐Fib3 or R345W‐Fib3 to overexpress these proteins. The cell lysate and cell culture media were collected to examine the extent of Fibulin3 secretion using the western blots and the luciferase assay. Results showed that transfected ARPE‐19 cells secrete significantly lower levels of Fibulin3 compared to WT cells (Figure S4a,b), suggesting that the Fibulin‐3 mutation led to reduced Fibulin‐3 secretion in RPE cells.

Conventional filtration and ultracentrifugation (P100K) and density gradient centrifugation (Pgrad) were used to isolate EVs. TEM and Cryo‐EM were performed to study the morphology of the WT‐ARPE‐EV (Pgrad). TEM imaging showed concave‐appearing vesicles (Figure 2a) and cryo‐EM showed spherical vesicles (Figure 2b), both with two subpopulations of EVs: a group with diameters around 30 nm and a group with diameters over 100 nm. NTA showed that, in the R345W group, the particle size distributions were smaller than those of the WT‐ARPE‐EV (Figure 2c). There were no significant differences in the total amount of particles in these two groups or in the amount of protein per vesicle between WT and R345W groups (Figure 2d,e). Western immunoblots confirmed the presence of EV markers (P100K), including Alix, CD81, and CD63, in both WT and R345W‐ARPE‐EVs, as well as the absence of negative controls including mitochondrial import inner membrane translocase subunit Tim23 (TIM23), Golgi matrix protein 130 (GM130), calnexin, and histone H3 (Figure 2f). GAPDH was not detected in EVs derived from ARPE‐19 cells. A loading gel was used as a reference (Figure S3).

FIGURE 2.

FIGURE 2

Open in a new tab

Expression of R345W‐Fibulin‐3 in RPE cells alters EV size. Representative (a) TEM and (b) cryo‐EM images of WT‐ARPE‐EVs reveals two subpopulations: vesicles ∼30 nm in diameter (red arrowhead) and vesicles over 100 nm in diameter (white arrowhead). Cryo‐EM images show the EVs possess a lipid bilayer, with peripheral proteins visible on the surface of the bilayer and protein densities are seen within the lumen of the vesicles (asterisks represent surface contamination). Scale bar: 100 nm. (c) Representative nanoparticle tracking analysis trace confirms a median diameter of 143.5 and 94.5 nm for WT‐ARPE‐EVs and R345W‐ARPE‐EVs, respectively. (d,e) No significant differences in the total amount of vesicles or in the amount of protein per vesicle between WT and R345W groups are observed (n = 7). (f) Western immunoblots confirmed the presence of EV markers (P100K), including Alix, CD81, and CD63, in both WT‐ARPE‐EV and R345W‐ARPE‐EVs and the absence of negative controls including TIM23, GM130, GAPDH, calnexin, and histone H3 (n = 3). A student's unpaired t‐test was used to compare differences between groups. Data are presented as mean ± SEM. EVs, extracellular vesicles. Alix, apoptosis‐linked gene‐2 interacting protein X; CL, cell lysates; EVs, extracellular vesicles; GAPDH, glyceraldehyde 3‐phosphate dehydrogenase; GM130, Golgi matrix protein 130 kD; TIM23, translocase of inner mitochondrial membrane 23.

3.2. Expression of R345W‐Fibulin‐3 in RPE cells alters EV protein cargo

WT‐ARPE‐EV and R345W‐ARPE‐EV were purified using sucrose cushion and iodixanol buoyant density gradient ultracentrifugation (Pgrad) (Figure 1b). Final ultracentrifugation yields 10 remaining fractions, with EVs in the third fraction from the top (density = 1.096) (Choi et al., 2021). Ten fractions were collected, and each 1 mL fraction was diluted to 10 mL with HBS and spun at 150,000 g for 3 h at 4°C. The supernatant was removed and the pellet was re‐suspended in 50 μL HBS. SDS‐PAGE gels showed distinct differences in protein distribution pattern between these groups in both fraction and pellet samples (Figure S5a,b). Western immunoblots confirmed the presence of EV markers (Pgrad), including Alix, HSP70, Flotillin‐1, CD81, and CD63 in the WT‐ARPE‐EV pellets (Figure S5c).

WT‐ARPE‐EV and R345W‐ARPE‐EV (Pgrad) were subjected to proteomic analysis (Figure 3). A total of 2852 proteins were identified in WT‐ARPE‐EV, and 2868 proteins were identified in R345W‐ARPE‐EV (Figure 3a). Known EV proteins were found in high abundance in our EV preparations, while non‐EV proteins were dispersed evenly across the protein abundance curve (Figure S6) indicating that the EV samples were of high purity. The proteomics screen was validated by comparing expression levels (normalised to total protein) of select target proteins on western blotting, to expression levels (normalised to spectral counts) of the same target proteins from the proteomics screen. This confirmed that WT‐ARPE‐EV exhibit greater ALIX, ZO‐1, CD63 and CD81, and lesser Flotilin‐1, Annexin A5 and Fibulin‐3 content compared to R345W‐ARPE‐EV (Figure S7). Differential protein expression data were submitted to iPathway Guide for functional enrichment analysis (Figure S8). Pathway analysis revealed that sonic hedgehog (SHH) pathways were found to be less abundant in EVs derived from R345W‐ARPE‐19 cells, while ribosome and lysosome pathways were more abundant (Figure 3b). Proteomic results were compared to gene sets within the EMTome (Vasaikar et al., 2021) using gene set analysis. The EMTome lists 3597 genes associated with EMT which are organised as gene sets from 87 publications; 963 EMTome genes were identified within the 2951 proteins in our dataset (Figure 3c). The set of EMT genes identified by Gotzmann et al. (Gotzmann et al., 2006) were significantly increased in R345W relative to WT (mean t‐statistic = 0.73, adjusted p = 0.075; Figure 3d). 15 of 22 genes in the set from Gotzmann et al. (Gotzmann et al., 2006) were increased in R345W relative to WT. Of particular interest to us, Gotzmann et al. (Gotzmann et al., 2006) identified TGFβ as a critical mediator of EMT in hepatocytes. Therefore, we investigated the capacity of EV‐associated TGFβ to induce EMT in RPE cells.

FIGURE 3.

FIGURE 3

Open in a new tab

Proteomic analysis of RPE EVs reveals markers of EMT. (a) A total of 2852 proteins were identified in WT‐Fibulin‐3 EVs, and 2868 proteins were identified in R345W‐Fibulin‐3 EVs. (b) Pathway analysis revealed that sonic hedgehog (SHH) pathways were found to be more abundant in EVs derived from WT ARPE‐19 cells, while the ribosome and lysosome pathway were decreased. (c) A total of 2951 proteins were identified in our EVs where 963 were identified in a EMTome gene set (3597 proteins). (d) Test statistics for protein differential expression between R345W‐ and WT‐ARPE‐EVs are summarised as boxplots for proteins that were EMT‐related according to Gotzmann et al. (2006) or not EMT‐related (n = 1 EV isolation per group).

3.3. R345W‐RPE‐derived EVs have increased TGFβ1 protein

Studies in post‐mortem AMD eyes found upregulation of well‐known drivers of EMT, namely, TGFβ and vimentin (VIM) (Ghosh et al., 2018; Hirasawa et al., 2011; Touhami et al., 2018). TGFβ is a well‐known regulator of EMT (Xu et al., 2009). Several studies have shown that TGFβ is located on the surface of EVs (Shelke et al., 2019; Yin et al., 2020) and can initiate TGFβ‐induced EMT in recipient cells (Kim et al., 2016; Shelke et al., 2019; van de Vlekkert et al., 2019; Webber et al., 2010). To test whether induction of R345W Fibulin‐3 alters TGFβ protein content, we quantified TGFβ1 protein in cell culture media, EV‐depleted media, and EVs in both WT and R345W groups using a commercial ELISA. Cell culture media were collected from WT‐ARPE and R345W‐ARPE after culturing for 48 h. EVs (P100K) were isolated from WT‐ARPE and R345W‐ARPE and EV‐depleted media (supernatant) was also collected. Compared to the WT media and EV‐depleted media, the amount of TGFβ1 protein was significantly greater in the R345W media and EV‐depleted media (n = 3, p < 0.01, each experiment was performed in duplicate, Figure 4a). Compared to the WT‐ARPE‐EV, TGFβ1 protein was significantly more abundant in the R345W‐ARPE‐EVs (each experiment was performed in duplicate, n = 3, p < 0.05) (Figure 4b). These results suggest that expression of R345W Fibulin‐3 induces a higher abundance of TGFβ1 protein in RPE cells and EVs derived from RPE cells. We have also quantified the percentage of EV‐bound‐TGFβ1 to free‐TGFβ1. The results show that approximately 74% and 67% of total TGFβ1 is EV‐bound in media of WT‐ARPE and R345W‐ARPE, respectively, suggesting that EV‐bound TGFβ1 plays an important role in mediating EMT in RPE cells (Figure 4c). Critically, these ELISAs were performed without lysing the EVs, suggesting that TGFβ is on the surface of the EVs. This topology is consistent with previous studies (Shelke et al., 2019; Yin et al., 2020).

FIGURE 4.

FIGURE 4

Open in a new tab

R345W‐ARPE‐EVs demonstrate increased TGFβ1 content. TGFβ1 content in (a) media, EV‐depleted media, and (b) EVs from WT and R345W groups were quantified by ELISA. (a) R345W media and EV‐depleted media contain significantly more TGFβ1 protein (n = 3, each experiment was performed in duplicate). (b) R345W‐ARPE‐EVs show significantly greater TGFβ1 protein compared to WT‐ARPE‐EVs (n = 3, each experiment was performed in duplicate). (c) We have quantified the percentage of EV‐bound TGFβ1 to free TGFβ1. The results show that approximately 74% and 67% of total TGFβ1 is EV‐bound in media of WT‐ARPE and R345W‐ARPE, respectively, suggesting that EV‐derived TGFβ1 plays an important role in mediating EMT in RPE cells (Figure 4c). Respectively, suggesting that EV‐derived TGFβ1 play an important role in mediating EMT in RPE cells. A student's unpaired t‐test was used to compare differences between groups. Data are presented as mean ± SEM, *p < 0.05, **p < 0.01. EVs, extracellular vesicles.

3.4. R345W‐ARPE‐EVs are sufficient to enhance migration abilities and upregulate EMT markers in recipient RPE cells

Confocal images show RPE‐derived EVs are taken up by recipient RPE cells (Figure S9a). Green puncta located within the cell body of ARPE‐19 cells (labelled for phalloidin staining for actin [greyscale]) that were incubated with PKH67‐labelled WT‐ or R345W‐EVs at 1 × 1010 vesicles/mL for 8 h. Green puncta were absent in cells that were not incubated with EVs. Number of EVs was quantified using ImageJ. There are no significant differences regarding EV uptake per area (1mm2) between WT and R345W groups (n = 3, ns, p = 0.3391) (Figure S9b).

EV treatment experiments were conducted to determine if R345W‐ARPE‐EVs are sufficient to promote EMT in naive cells. EVs (P100K) were isolated from WT‐ARPE or R345W‐ARPE cell lines and were taken up in naïve cells (Figure S9), suggesting the EVs can signal in a paracrine manner. Naive ARPE‐19 cells were seeded onto a 96‐well plate (5 × 104/well). On Day 1, the wound was created, each well was rinsed with PBS, and media containing WT‐ARPE‐EVs or R345W‐ARPE‐EVs were added to the appropriate wells. ARPE‐19 cells grown on FBS depleted media without adding EVs were used as controls. A total of n = 8 was used for each group, and the experiment was repeated three times with new EV preparations. EV concentrations were chosen according to a previously published study (van de Vlekkert et al., 2019). For these studies, 1 × 1010/mL EVs were used. A schematic for the scratch assay timeline is shown (Figure 5a). To compare across groups, the maximum recovery at 72 h was used to determine the time required to close half of the wound area. The mean half‐max wound recovery time across three experiments revealed that R345W‐ARPE‐EVs had the fastest wound recovery time (n = 3) (Figure 5b,c). No significant differences were found between the no EV and WT‐ARPE‐EV groups (p = 0.116).

FIGURE 5.

FIGURE 5

Open in a new tab

RPE‐derived EVs are sufficient to enhance migration abilities in RPE cells. (a) Schematic for the EV transplant and Cell Migration Assay. On Day 0, naive ARPE‐19 cells were seeded onto a 96‐well plate. On Day 1, the wound was created, each well was rinsed with PBS, and media containing WT or R345W‐ARPE‐EVs were added to the appropriate wells. On Days 2–6, recovery rate was automatically analyzed by the software. (b) Cell migration assays (n = 6 wells, assay repeated thrice) show that WT‐ARPE cells treated with WT‐ARPE‐EVs (dashed line) or R345W‐ARPE‐EVs (solid line) resulted in a faster recovery rate, where R345W‐ARPE‐EVs were significantly faster than WT. (c) The relative quantity of mRNA in naive ARPE‐19 cells was normalised to GAPDH and then normalised to the control group (No EVs). RT‐PCR analysis show that, compared to the control (No EVs) and WT‐ARPE‐EVs, the mRNA expression levels of EMT markers TGFβ1 and VIM were significantly greater in cells treated with R345W‐ARPE‐EVs (n = 6). A two‐way ANOVA with multiple comparison post‐test (b) or student's unpaired t‐test (c) was used to compare differences between groups. Data are presented as mean ± SEM, *p < 0.05, **p < 0.01, ***p < 0.01. CDH2, Cadherin 2; EVs, extracellular vesicles; VIM, vimentin.

Using RT‐PCR, we compared the expression levels of EMT‐promoting factors (TGFβ1, VIM, cadherin 2 (CDH2)) in each group 24 h after EV treatment. ARPE‐19 cells grown on FBS depleted media without adding EVs were used as controls. The relative quantity of mRNA was normalised to GAPDH and then normalised to the control group (no EVs). The mRNA expression levels of GAPDH did not change among conditions (Figure S1). Compared to the control group (No EVs) and WT‐ARPE‐EVs, increased mRNA expression levels of EMT markers were found in cells treated with R345W‐ARPE‐EVs (n = 6) (Figure 5d). Based on the consensus criteria for defining EMT (Yang et al., 2020), these results suggest that EVs are sufficient to induce EMT in RPE cells.

3.5. Pre‐incubation with TGFβ‐antibody reduced EV‐induced migration in RPE cells

Previous studies have shown that EV‐bound TGFβ1 induces EMT in various cell types (Kim et al., 2016; Shelke et al., 2019; Webber et al., 2010). To determine if TGFβ bound to EVs is required for EMT induction, EVs were incubated with neutralising antibodies prior to the cell migration assay. WT‐ARPE‐EVs induced a significant (p < 0.05) but modest increase in cell migration rate. Pre‐incubation WT‐ARPE‐EVs with pan‐TGFβ‐neutralising antibody did not alter the migration rate (Figure 6a). R345W‐ARPE‐EVs induced a highly significant (p < 0.001) increase in cell migration rate. Pre‐incubation R345W‐ARPE‐EVs with pan‐TGFβ‐neutralising antibody reduced the migration rate to the untreated (No EV) group (Figure 6b). EVs pre‐incubated with non‐specific IGG antibodies did not alter the migration rate. Taken together, these results suggest that EV‐bound TGFβ is required to induce EMT in the ARPE‐19 cell line. As discussed below, future studies are required to determine if TGFβ is required to induce EMT in primary RPE cells, iPSC‐derived RPE cells and in vivo.

FIGURE 6.

FIGURE 6

Open in a new tab

Pre‐incubation with TGFβ‐antibody reduced EV‐induced migration in RPE cells. EVs pre‐incubated with pan‐TGFβ‐neutralising antibody induced a significantly slower recovery rate compared to the control group and isotype IgG group in both (a) WT‐ARPE‐EVs and (b) R345W‐EVs, however, there was a greater statistical difference in the R345W‐EVs. Two‐way ANOVA with multiple comparison post‐test was used to compare differences between EV groups (EVs alone [*], or EVs preincubated with either IGG [#] or TGFβ AB [ns]) and the no EV control group. Data presented as mean ± SEM (n = 6 wells, assay repeated twice) #,*p < 0.05, ##,**p < 0.01, ###,***p < 0.001.

4. DISCUSSION

Under normal conditions, RPE cells are terminally differentiated (Kozlowski, 2012). Recent studies in post‐mortem age‐related macular degeneration (AMD) eyes have found upregulation of critical EMT drivers such as TGFβ, VIM, and Snail (Ghosh et al., 2018; Hirasawa et al., 2011; Ishikawa et al., 2016). Our group and previous studies have shown that RPE cells are capable of undergoing EMT (Ghosh et al., 2018; Zhou et al., 2020; Sripathi, Hu, Turaga, et al., 2021; Tamiya & Kaplan, 2016). In other tissues, EVs have been shown to regulate EMT (Vella, 2014; Kim et al., 2016). The current study suggests that EVs promote EMT in RPE cells in a TGFβ‐dependent manner.

The effect of R345W‐Fibulin‐3 expression on EV secretion was examined. The size of R345W‐ARPE‐EVs are smaller on average than the WT‐ARPE‐EVs. The change in size distribution may suggest an alternate secretion mechanism of R345W‐ARPE‐EVs, however, there is no definitive mechanistic links that connect EV size to EV mechanism. This is an evolving area in the field of EVs and has been extensively reviewed in Buzas et al. 2023 (Buzas, 2023). The secretion rate and the amount of protein per vesicle were not significantly different. A recent study demonstrated, however, that the complexities of heterogeneous EV subpopulations may imply distinct biological functions (Zhang et al., 2018), further emphasising the importance of separating and studying subpopulations of EVs.

ALIX, an endosomal sorting complexes required for transport (ESCRT)‐related protein, is required for EV secretion (Colombo et al., 2013). Tetraspanins, such as CD63 and CD81, are highly enriched in EVs relative to their content in the respective producing cells (Colombo et al., 2013). Both our proteomic analyses and western blot analysis have shown that WT‐ARPE‐EV exhibit greater ALIX, CD63, and CD81 content compared to R345W‐ARPE‐EV. This EV compositional heterogeneity results from EV biogenesis operating across a spectrum of endosomal and plasma membranes (Kowal et al., 2014; Théry et al., 2018). These results suggest that expression of R345W‐Fibulin‐3 may alter the mechanism of EV secretion. However, future studies are required to identify these possible alternate secretion mechanisms.

We next examined whether expression of R345W‐Fibulin‐3 alters EV cargo. We found that R345‐ARPE‐EVs have reduced levels of proteins within the SHH pathway. Interestingly, SHH coordinates photoreceptor differentiation during development (Stenkamp et al., 2000). Moreover, nearly one third of our EV proteins, are found within the EMTome (Vasaikar et al., 2021). Critically, when compared to gene sets within the EMTome, EMT proteins were significantly upregulated in R345W‐ARPE‐EVs specifically within the gene set from Gotzmann et al. (2006) which describes a TGFβ‐driven EMT pathway.

Prior work from our group has shown that expression of R345W‐Fibulin‐3 induced EMT in RPE cells (Zhou et al., 2020). In the current study, we chose to measure EMT drivers at a timepoint coincident with the early change in migration. We show that treatment of RPE cells with EVs derived from cells expressing R345W‐Fibulin‐3 are sufficient to enhance cell migration and induce expression of EMT‐promoting factors, such as CDH2, SNAIL and TGFβ itself. These data are consistent from those published showing the change in EMT drivers over time in RPE cells treated with TGFβ or stimulated to undergo EMT through trypsin digestion (Sripathi, Hu, Liu, et al., 2021; Sripathi, Hu, Turaga, et al., 2021). WT‐ or R345W‐ARPE‐EVs were shown to be taken up by cells they were not secreted from, suggesting that they may act in a paracrine manner. TGFβ ELISAs revealed greater TGFβ1 protein abundance in R345W‐ARPE‐EVs. Moreover, pre‐incubation of EVs with pan‐TGFβ antibody slowed the migration rate to that of controls. However, data also reveal that conditioned media from R345W cells contain more TGFβ, both free and EV‐bound, thus it is possible that some soluble protein (if a portion is pelleted at 100K x g) may be exerting an effect on migration rate. Taken together, these results suggest that EV‐bound TGFβ may play a critical role in the induction of EMT in RPE cells, but further investigations using alternate EV isolation protocols need to be performed.

Limitations of the current study are as follows. ARPE‐19 cells lack some of the essential characteristics of RPE cells, such as high barrier function (as assessed by transepithelial resistance), a deficiency which could reflect some flaws in polarity. Though ARPE‐19 cells were cultured to confluence so that a monolayer and tight junction had formed, the RPE characteristics of these cells could be improved if they had been left to differentiate for 2 weeks with nicotinamide (Hazim et al., 2019). Ultimately, primary RPE or iPSC‐derived RPE models are the preferable model for RPE research and will be used for future studies. The current study used density gradient preparations for the proteomic analysis. Though, for functional studies, we elected to culture the cells on large dishes to obtain large media volumes and use filtration and centrifugation methods, which can result in lower EV purity (Théry et al., 2018) but would effect both WT and R345W in the same manner. These choices were made due to the practicality of scale of cell culture and size of transwell plates available at the time of the current study. Future experiments using larger transwell plates will be required to achieve full differentiation and polarisation of RPE cells prior to experimentation. This approach will also allow for the characterisation and EMT‐inducing capacity from both apical and basolateral EVs. These studies are critical as EV proteins have been found in deposits (Garland et al., 2014; Crabb, 2014) and it has become clear that differences exist in apical and basolateral EVs cargo and function (Klingeborn et al., 2017; Klingeborn et al., 2018; Matsui et al., 2021).

5. CONCLUSION

In summary, our experimental results indicate that expression of a misfolded protein in RPE cells alters EV protein cargo, resulting in EVs that can induce EMT in recipient cells. The current study and prior literature suggest a paracrine signalling model, whereby EVs from dysfunctional RPE cells may propagate EMT in RPE monolayers (Figure 7). This model will guide future experiments to further elucidate the mechanisms underlying RPE dysfunction in macular degenerations.

FIGURE 7.

FIGURE 7

Open in a new tab

RPE EVs act in a paracrine manner to impact adjacent RPE cell differentiation. Simplified schematic illustration showing the proposed theory that EVs, secreted both apical and basolateral (only apical shown), act in a paracrine manner to impact adjacent RPE cell differentiation. The green circles represent EVs derived from RPE cells expressing R345W‐Fibulin‐3.

AUTHOR CONTRIBUTIONS

Mi Zhou: Formal analysis; investigation; methodology; visualisation; writing – original draft and writing – review & editing. Yuanjun Zhao: Formal analysis; investigation; methodology and supervision. Sarah R. Weber: Writing – original draft and writing – review & editing. Christopher Gates: Data curation; formal analysis; methodology and writing – review & editing. Nicholas J. Carruthers: Data curation; formal analysis; methodology and writing – review & editing. Han Chen: Formal analysis; investigation and methodology. Xiaoming Liu: Methodology and resources. Hong‐Gang Wang: Methodology; resources and supervision. Michael Ford: Methodology and resources. Matthew T. Swulius: Investigation; methodology; resources and supervision. Alistair J. Barber: Conceptualization; methodology; project administration; supervision; writing – original draft and writing – review & editing. Stephanie L. Grillo: Conceptualization; formal analysis; methodology; project administration; supervision; visualization; writing – original draft and writing – review & editing. Jeffrey M. Sundstrom: Conceptualization; funding acquisition; project administration; resources; supervision; visualization; writing – original draft and writing – review & editing.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

Supporting information

Supporting Information

Click here for additional data file. (5.9MB, docx)

ACKNOWLEDGEMENTS

This work was funded by the Bennett and Inez Chotiner Early Career Professorship in Ophthalmology endowment (JMS).

Zhou, M. , Zhao, Y. , Weber, S. R. , Gates, C. , Carruthers, N. J. , Chen, H. , Liu, X. , Wang, H.‐G. , Ford, M. , Swulius, M. T. , Barber, A. J. , Grillo, S. L. , & Sundstrom, J. M. (2023). Extracellular vesicles from retinal pigment epithelial cells expressing R345W‐Fibulin‐3 induce epithelial‐mesenchymal transition in recipient cells. Journal of Extracellular Vesicles, 12, e12373. 10.1002/jev2.12373

REFERENCES

  1. Alvarez‐Erviti, L. , Seow, Y. , Schapira, A. H. , Gardiner, C. , Sargent, I. L. , Wood, M. J. , & Cooper, J. M. (2011). Lysosomal dysfunction increases exosome‐mediated alpha‐synuclein release and transmission. Neurobiology of Disease, 42(3), 360–367. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Askeland, A. , Borup, A. , Østergaard, O. , Olsen, J. V. , Lund, S. M. , Christiansen, G. , Kristensen, S. R. , Heegaard, N. H. H. , & Pedersen, S. (2020). Mass‐spectrometry based proteome comparison of extracellular vesicle isolation methods: Comparison of ME‐kit, size‐exclusion chromatography, and high‐speed centrifugation. Biomedicines, 8(8), 246. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Atienzar‐Aroca, S. , Serrano‐Heras, G. , Freire Valls, A. , Ruiz de Almodovar, C. , Muriach, M. , Barcia, J. M. , Garcia‐Verdugo, J. M. , Romero, F. J. , & Sancho‐Pelluz, J. (2018). Role of retinal pigment epithelium‐derived exosomes and autophagy in new blood vessel formation. Journal of Cellular and Molecular Medicine, 22(11), 5244–5256. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Biasutto, L. , Chiechi, A. , Couch, R. , Liotta, L. A. , & Espina, V. (2013). Retinal pigment epithelium (RPE) exosomes contain signaling phosphoproteins affected by oxidative stress. Experimental Cell Research, 319(13), 2113–2123. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Burton, J. B. , Carruthers, N. J. , Hou, Z. , Matherly, L. H. , & Stemmer, P. M. (2022). Pattern analysis of organellar maps for interpretation of proteomic data. Proteomes, 10(2), 18. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Burton, J. B. , Carruthers, N. J. , & Stemmer, P. M. (2023). Enriching extracellular vesicles for mass spectrometry. Mass Spectrometry Reviews, 42(2), 779–795. [DOI] [PubMed] [Google Scholar]
  7. Buzas, E. I. (2023). The roles of extracellular vesicles in the immune system. Nature Reviews. Immunology, 23(4), 236–250. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cao, D. , Leong, B. , Messinger, J. D. , Kar, D. , Ach, T. , Yannuzzi, L. A. , Freund, K. B. , & Curcio, C. A. (2021). Hyperreflective foci, optical coherence tomography progression indicators in age‐related macular degeneration, include transdifferentiated retinal pigment epithelium. Investigative Ophthalmology & Visual Science, 62(10), 34. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Chen, Y. , Zeng, C. , Zhan, Y. , Wang, H. , Jiang, X. , & Li, W. (2017). Aberrant low expression of p85α in stromal fibroblasts promotes breast cancer cell metastasis through exosome‐mediated paracrine Wnt10b. Oncogene, 36(33), 4692–4705. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Choi, D. , & Gho, Y. S. (2015). Isolation of extracellular vesicles for proteomic profiling. Methods in Molecular Biology (Clifton, N.J.), 1295, 167–177. [DOI] [PubMed] [Google Scholar]
  11. Choi, D. , & Gho, Y. S. (2016). Isolation of extracellular vesicles for proteomic profiling. Methods in Molecular Biology (Clifton, N.J.), 2261, 193–206. [DOI] [PubMed] [Google Scholar]
  12. Colombo, M. , Moita, C. , van Niel, G. , Kowal, J. , Vigneron, J. , Benaroch, P. , Manel, N. , Moita, L. F. , Théry, C. , & Raposo, G. (2013). Analysis of ESCRT functions in exosome biogenesis, composition and secretion highlights the heterogeneity of extracellular vesicles. Journal of Cell Science, 126(Pt 24), 5553–5565. [DOI] [PubMed] [Google Scholar]
  13. Crabb, J. W. (2014). The proteomics of drusen. Cold Spring Harbor Perspectives in Medicine, 4(7), a017194. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Garland, D. L. , Fernandez‐Godino, R. , Kaur, I. , Speicher, K. D. , Harnly, J. M. , Lambris, J. D. , Speicher, D. W. , & Pierce, E. A. (2014). Mouse genetics and proteomic analyses demonstrate a critical role for complement in a model of DHRD/ML, an inherited macular degeneration. Human Molecular Genetics, 23(1), 52–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Ghosh, S. , Shang, P. , Terasaki, H. , Stepicheva, N. , Hose, S. , Yazdankhah, M. , Weiss, J. , Sakamoto, T. , Bhutto, I. A. , Xia, S. , Zigler, J. S., Jr , Kannan, R. , Qian, J. , Handa, J. T. , & Sinha, D. (2018). A role for βA3/A1‐crystallin in type 2 EMT of RPE cells occurring in dry age‐related macular degeneration. Investigative Ophthalmology & Visual Science, 59(4), AMD104–AMD113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Goldberg, M. F. , McLeod, S. , Tso, M. , Packo, K. , Edwards, M. , Bhutto, I. A. , Baldeosingh, R. , Eberhart, C. , Weber, B. H. F. , & Lutty, G. A. (2018). Ocular histopathology and immunohistochemical analysis in the oldest known individual with autosomal dominant vitreoretinochoroidopathy. Ophthalmology. Retina, 2(4), 360–378. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Gotzmann, J. , Fischer, A. N. , Zojer, M. , Mikula, M. , Proell, V. , Huber, H. , Jechlinger, M. , Waerner, T. , Weith, A. , Beug, H. , & Mikulits, W. (2006). A crucial function of PDGF in TGF‐beta‐mediated cancer progression of hepatocytes. Oncogene, 25(22), 3170–3185. [DOI] [PubMed] [Google Scholar]
  18. Grillo, S. L. , Etzel, J. D. , Weber, S. R. , Ondeck, C. , Wang, W. , Zhao, Y. , Barber, A. J. , & Sundstrom, J. M. (2021). Descriptive analysis of Fibulin‐3 and the extracellular vesicle marker, Alix, in drusen from a small cohort of postmortem human eyes. Experimental Eye Research, 203, 108422. [DOI] [PubMed] [Google Scholar]
  19. Guidry, C. , Medeiros, N. E. , & Curcio, C. A. (2002). Phenotypic variation of retinal pigment epithelium in age‐related macular degeneration. Investigative Ophthalmology & Visual Science, 43(1), 267–273. [PubMed] [Google Scholar]
  20. Hazim, R. A. , Volland, S. , Yen, A. , Burgess, B. L. , & Williams, D. S. (2019). Rapid differentiation of the human RPE cell line, ARPE‐19, induced by nicotinamide. Experimental Eye Research, 179, 18–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Hirasawa, M. , Noda, K. , Noda, S. , Suzuki, M. , Ozawa, Y. , Shinoda, K. , Inoue, M. , Ogawa, Y. , Tsubota, K. , & Ishida, S. (2011). Transcriptional factors associated with epithelial‐mesenchymal transition in choroidal neovascularization. Molecular Vision, 17, 1222–1230. [PMC free article] [PubMed] [Google Scholar]
  22. Hulleman, J. D. , Brown, S. J. , Rosen, H. , & Kelly, J. W. (2013). A high‐throughput cell‐based Gaussia luciferase reporter assay for identifying modulators of fibulin‐3 secretion. Journal of Biomolecular Screening, 18(6), 647–658. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Hulleman, J. D. , & Kelly, J. W. (2015). Genetic ablation of N‐linked glycosylation reveals two key folding pathways for R345W fibulin‐3, a secreted protein associated with retinal degeneration. FASEB Journal: Official Publication of the Federation of American Societies for Experimental Biology, 29(2), 565–575. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Ishikawa, K. , Kannan, R. , & Hinton, D. R. (2016). Molecular mechanisms of subretinal fibrosis in age‐related macular degeneration. Experimental Eye Research, 142, 19–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kim, J. , Kim, T. Y. , Lee, M. S. , Mun, J. Y. , Ihm, C. , & Kim, S. A. (2016). Exosome cargo reflects TGF‐β1‐mediated epithelial‐to‐mesenchymal transition (EMT) status in A549 human lung adenocarcinoma cells. Biochemical and Biophysical Research Communications, 478(2), 643–648. [DOI] [PubMed] [Google Scholar]
  26. Klingeborn, M. , Dismuke, W. M. , Skiba, N. P. , Kelly, U. , Stamer, W. D. , & Bowes Rickman, C. (2017). Directional exosome proteomes reflect polarity‐specific functions in retinal pigmented epithelium monolayers. Scientific Reports, 7(1), 4901. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Klingeborn, M. , Stamer, W. D. , & Bowes Rickman, C. (2018). Polarized exosome release from the retinal pigmented epithelium. Advances in Experimental Medicine and Biology, 1074, 539–544. [DOI] [PubMed] [Google Scholar]
  28. Kowal, J. , Tkach, M. , & Théry, C. (2014). Biogenesis and secretion of exosomes. Current Opinion in Cell Biology, 29, 116–125. [DOI] [PubMed] [Google Scholar]
  29. Kozlowski, M. R. (2012). RPE cell senescence: A key contributor to age‐related macular degeneration. Medical Hypotheses, 78(4), 505–510. [DOI] [PubMed] [Google Scholar]
  30. Marmorstein, L. (2004). Association of EFEMP1 with malattia leventinese and age‐related macular degeneration: A mini‐review. Ophthalmic Genetics, 25(3), 219–226. [DOI] [PubMed] [Google Scholar]
  31. Marmorstein, L. Y. , Munier, F. L. , Arsenijevic, Y. , Schorderet, D. F. , McLaughlin, P. J. , Chung, D. , Traboulsi, E. , & Marmorstein, A. D. (2002). Aberrant accumulation of EFEMP1 underlies drusen formation in Malattia Leventinese and age‐related macular degeneration. Proceedings of the National Academy of Sciences of the United States of America, 99(20), 13067–13072. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Matsui, T. , Osaki, F. , Hiragi, S. , Sakamaki, Y. , & Fukuda, M. (2021). ALIX and ceramide differentially control polarized small extracellular vesicle release from epithelial cells. EMBO Reports, 22(5), e51475. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Narendran, N. , Guymer, R. H. , Cain, M. , & Baird, P. N. (2005). Analysis of the EFEMP1 gene in individuals and families with early onset drusen. Eye (London, England), 19(1), 11–15. [DOI] [PubMed] [Google Scholar]
  34. Robinson, M. D. , McCarthy, D. J. , & Smyth, G. K. (2010). edgeR: A Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics (Oxford, England), 26(1), 139–140. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Shah, N. , Ishii, M. , Brandon, C. , Ablonczy, Z. , Cai, J. , Liu, Y. , Chou, C. J. , & Rohrer, B. (2018). Extracellular vesicle‐mediated long‐range communication in stressed retinal pigment epithelial cell monolayers. Biochimica et biophysica acta. Molecular Basis of Disease, 1864(8), 2610–2622. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Shelke, G. V. , Yin, Y. , Jang, S. C. , Lässer, C. , Wennmalm, S. , Hoffmann, H. J. , Li, L. , Gho, Y. S. , Nilsson, J. A. , & Lötvall, J. (2019). Endosomal signalling via exosome surface TGFβ‐1. Journal of Extracellular Vesicles, 8(1), 1650458. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Sripathi, S. R. , Hu, M. W. , Liu, M. M. , Wan, J. , Cheng, J. , Duan, Y. , Mertz, J. L. , Wahlin, K. J. , Maruotti, J. , Berlinicke, C. A. , Qian, J. , & Zack, D. J. (2021). Transcriptome landscape of epithelial to mesenchymal transition of human stem cell‐derived RPE. Investigative Ophthalmology & Visual Science, 62(4), 1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Sripathi, S. R. , Hu, M. W. , Turaga, R. C. , Mertz, J. , Liu, M. M. , Wan, J. , Maruotti, J. , Wahlin, K. J. , Berlinicke, C. A. , Qian, J. , & Zack, D. J. (2021). Proteome landscape of epithelial‐to‐mesenchymal transition (EMT) of retinal pigment epithelium shares commonalities with malignancy‐associated EMT. Molecular & Cellular Proteomics: MCP, 20, 100131. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Stenkamp, D. L. , Frey, R. A. , Prabhudesai, S. N. , & Raymond, P. A. (2000). Function for Hedgehog genes in zebrafish retinal development. Developmental Biology, 220(2), 238–252. [DOI] [PubMed] [Google Scholar]
  40. Stone, E. M. , Lotery, A. J. , Munier, F. L. , Héon, E. , Piguet, B. , Guymer, R. H. , Vandenburgh, K. , Cousin, P. , Nishimura, D. , Swiderski, R. E. , Silvestri, G. , Mackey, D. A. , Hageman, G. S. , Bird, A. C. , Sheffield, V. C. , & Schorderet, D. F. (1999). A single EFEMP1 mutation associated with both Malattia Leventinese and Doyne honeycomb retinal dystrophy. Nature Genetics, 22(2), 199–202. [DOI] [PubMed] [Google Scholar]
  41. Tamiya, S. , & Kaplan, H. J. (2016). Role of epithelial‐mesenchymal transition in proliferative vitreoretinopathy. Experimental Eye Research, 142, 26–31. [DOI] [PubMed] [Google Scholar]
  42. Théry, C. , Witwer, K. W. , Aikawa, E. , Alcaraz, M. J. , Anderson, J. D. , Andriantsitohaina, R. , Antoniou, A. , Arab, T. , Archer, F. , Atkin‐Smith, G. K. , Ayre, D. C. , Bach, J. M. , Bachurski, D. , Baharvand, H. , Balaj, L. , Baldacchino, S. , Bauer, N. N. , Baxter, A. A. , Bebawy, M. , … & Zuba‐Surma, E. K. (2018). Minimal information for studies of extracellular vesicles 2018 (MISEV2018): A position statement of the International Society for Extracellular Vesicles and update of the MISEV2014 guidelines. Journal of Extracellular Vesicles, 7(1), 1535750. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Touhami, S. , Beguier, F. , Augustin, S. , Charles‐Messance, H. , Vignaud, L. , Nandrot, E. F. , Reichman, S. , Forster, V. , Mathis, T. , Sahel, J. A. , Bodaghi, B. , Guillonneau, X. , & Sennlaub, F. (2018). Chronic exposure to tumor necrosis factor alpha induces retinal pigment epithelium cell dedifferentiation. Journal of Neuroinflammation, 15(1), 85. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. van de Vlekkert, D. , Demmers, J. , Nguyen, X. X. , Campos, Y. , Machado, E. , Annunziata, I. , Hu, H. , Gomero, E. , Qiu, X. , Bongiovanni, A. , Feghali‐Bostwick, C. A. , & d'Azzo, A. (2019). Excessive exosome release is the pathogenic pathway linking a lysosomal deficiency to generalized fibrosis. Science Advances, 5(7), eaav3270. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Väremo, L. , Nielsen, J. , & Nookaew, I. (2013). Enriching the gene set analysis of genome‐wide data by incorporating directionality of gene expression and combining statistical hypotheses and methods. Nucleic Acids Research, 41(8), 4378–4391. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Vasaikar, S. V. , Deshmukh, A. P. , den Hollander, P. , Addanki, S. , Kuburich, N. A. , Kudaravalli, S. , Joseph, R. , Chang, J. T. , Soundararajan, R. , & Mani, S. A. (2021). EMTome: A resource for pan‐cancer analysis of epithelial‐mesenchymal transition genes and signatures. British Journal of Cancer, 124(1), 259–269. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Vella, L. J. (2014). The emerging role of exosomes in epithelial‐mesenchymal‐transition in cancer. Frontiers in Oncology, 4, 361. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Vincent, A. , Munier, F. L. , Vandenhoven, C. C. , Wright, T. , Westall, C. A. , & Héon, E. (2012). The characterization of retinal phenotype in a family with C1QTNF5‐related late‐onset retinal degeneration. Retina (Philadelphia, Pa.), 32(8), 1643–1651. [DOI] [PubMed] [Google Scholar]
  49. Webber, J. , Steadman, R. , Mason, M. D. , Tabi, Z. , & Clayton, A. (2010). Cancer exosomes trigger fibroblast to myofibroblast differentiation. Cancer Research, 70(23), 9621–9630. [DOI] [PubMed] [Google Scholar]
  50. Weber, B. H. , Vogt, G. , Pruett, R. C. , Stöhr, H. , & Felbor, U. (1994). Mutations in the tissue inhibitor of metalloproteinases‐3 (TIMP3) in patients with Sorsby's fundus dystrophy. Nature Genetics, 8(4), 352–356. [DOI] [PubMed] [Google Scholar]
  51. Weber, S. R. , Zhou, M. , Zhao, Y. , & Sundstrom, J. M. (2020). Exosomes in retinal diseases. In Exosomes (pp. 415–431). Academic Press. [Google Scholar]
  52. Wu, D. , Kanda, A. , Liu, Y. , Kase, S. , Noda, K. , & Ishida, S. (2019). Galectin‐1 promotes choroidal neovascularization and subretinal fibrosis mediated via epithelial‐mesenchymal transition. FASEB Journal: Official Publication of the Federation of American Societies for Experimental Biology, 33(2), 2498–2513. [DOI] [PubMed] [Google Scholar]
  53. Xu, J. , Lamouille, S. , & Derynck, R. (2009). TGF‐beta‐induced epithelial to mesenchymal transition. Cell Research, 19(2), 156–172. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Yang, J. , Antin, P. , Berx, G. , Blanpain, C. , Brabletz, T. , Bronner, M. , Campbell, K. , Cano, A. , Casanova, J. , Christofori, G. , Dedhar, S. , Derynck, R. , Ford, H. L. , Fuxe, J. , García de Herreros, A. , Goodall, G. J. , Hadjantonakis, A. K. , Huang, R. Y. J. , Kalcheim, C. , … EMT International Association (TEMTIA) . (2020). Guidelines and definitions for research on epithelial‐mesenchymal transition. Nature Reviews. Molecular Cell Biology, 21(6), 341–352. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Yin, Y. , Shelke, G. V. , Lässer, C. , Brismar, H. , & Lötvall, J. (2020). Extracellular vesicles from mast cells induce mesenchymal transition in airway epithelial cells. Respiratory Research, 21(1), 101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Yuyama, K. , Yamamoto, N. , & Yanagisawa, K. (2008). Accelerated release of exosome‐associated GM1 ganglioside (GM1) by endocytic pathway abnormality: Another putative pathway for GM1‐induced amyloid fibril formation. Journal of Neurochemistry, 105(1), 217–224. [DOI] [PubMed] [Google Scholar]
  57. Zhang, H. , Freitas, D. , Kim, H. S. , Fabijanic, K. , Li, Z. , Chen, H. , Mark, M. T. , Molina, H. , Martin, A. B. , Bojmar, L. , Fang, J. , Rampersaud, S. , Hoshino, A. , Matei, I. , Kenific, C. M. , Nakajima, M. , Mutvei, A. P. , Sansone, P. , Buehring, W. , … Lyden, D. (2018). Identification of distinct nanoparticles and subsets of extracellular vesicles by asymmetric flow field‐flow fractionation. Nature Cell Biology, 20(3), 332–343. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Zhang, Y. , & Marmorstein, L. Y. (2010). Focus on molecules: Fibulin‐3 (EFEMP1). Experimental Eye Research, 90(3), 374–375. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Zhou, M. , Geathers, J. S. , Grillo, S. L. , Weber, S. R. , Wang, W. , Zhao, Y. , & Sundstrom, J. M. (2020). Role of epithelial‐mesenchymal transition in retinal pigment epithelium dysfunction. Frontiers in Cell and Developmental Biology, 8, 501. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Zhou, M. , Weber, S. R. , Zhao, Y. , Chen, H. , Barber, A. J. , Grillo, S. L. , Wills, C. A. , Wang, H. G. , Hulleman, J. D. , & Sundstrom, J. M. (2020). Expression of R345W‐Fibulin‐3 induces epithelial‐mesenchymal transition in retinal pigment epithelial cells. Frontiers in Cell and Developmental Biology, 8, 469. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Zhou, M. , Weber, S. R. , Zhao, Y. , Chen, H. , & Sundstrom, J. M. (2020). Chapter 2 ‐ Methods for exosome isolation and characterization. Exosomes, 23–38. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting Information

Click here for additional data file. (5.9MB, docx)

Articles from Journal of Extracellular Vesicles are provided here courtesy of Wiley

RESOURCES