Exosomal miR-184 in the aqueous humor of patients with central serous chorioretinopathy: a potential diagnostic and prognostic biomarker | Journal of Nanobiotechnology


This prospective study included 42 eyes from 42 patients diagnosed with CSC in the Department of Ophthalmology at Yeungnam University College of Medicine, Daegu, Korea. Twenty eyes from 20 patients undergoing cataract surgery, without diabetes or diagnosed retinal disease, were selected as controls. For the exosome size analysis and miR-184 measurement of individual eyes, 15 patients with CSC and 22 with cataracts as controls were enrolled at Asan Medical Center. The study protocol was approved by the Institutional Review Board (IRB No. 2019-10-056-002 in Yeungnam University, IRB No. 2020-1945-0002 in Asan Medical Center). The study was performed in accordance with the tenets of the Declaration of Helsinki. The diagnosis of CSC was established by the presence of a typical fluorescein leakage pattern on fluorescein angiography (FA) and subretinal fluid accumulation evident on spectral domain optical coherence tomography (SD-OCT) [4]. We enrolled subacute CSC patients with a symptom duration ranging between 6 weeks and 4 months under the same criteria as a previous study [12]. Patients were excluded based on the presence of choroidal neovascularization, prior treatment for CSC, concurrent ophthalmologic disease or history of diabetes, any intraocular surgery, using systemic or topical carbonic anhydrase inhibitor within 1 month, and any history of intravitreal steroid injection to study eye.

Multimodal imaging analysis of patients

All patients received bilateral ophthalmic examination, including biomicroscopic examination, fundus examination, and imaged with ultra-wide-field fundus photography (UWF, Optos California; Optos plc, UK), ultra-wide-field autofluorescence (UWF-AF) images, and ultra-wide-field FA (UWF-FA), or imaged with scanning laser ophthalmoscope infrared images (SLO-IR), blue autofluorescence (BAF), and infrared autofluorescence (IRAF), and the corresponding SD-OCT (Spectralis; Heidelberg Engineering, Heidelberg, Germany) [51]. Choroidal images were also obtained using the enhanced depth imaging (EDI) technique in SD-OCT. Ultra-wide-field indocyanine green angiography (UWF-ICGA) or optical coherence tomography angiography (OCTA) were performed as needed to rule out choroidal neovascularization. Central retinal thickness (CRT) was measured automatically on the central 1 mm zone from the fovea using SD-OCT segmentation analysis. Subfoveal choroidal thickness (SFChT) was manually measured by determining the vertical distance between the interface of the Bruch membrane and the sclerochoroidal junction on the B-scan of the OCT [52]. The height of the PED was assessed by measuring the vertical distance between the Bruch membrane and the apex of the retinal pigment epithelium (RPE) using the OCT scan in areas with the most prominent lesion [53]. SRF height was defined as the maximum distance between the RPE and the border of the detached neurosensory retina within a 3 mm nasal and 3 mm temporal area from the center of the fovea as previously described [54]. Measurement of SFChT, SRF height, and PED height were manually performed by utilizing a virtual caliper within the software (Heidelberg Eye Explorer ver. [55] (Fig. 1A).

Patient subgrouping

All patients with CSC were treated with intravitreal bevacizumab (IVB) and classified into responder and non-responder groups according to the result of the follow-up examination. Among the various anti-VEGF agents, some have not received approval for use in CSC, while IVB has been permitted for off-label use with IRB approval in our institute. Every patient was treatment naïve. CSC responders (CSC-Rs) were defined as patients who achieved complete absorption of SRF on SD-OCT at 1 month after IVB injection. CSC non-responders (CSC-NRs) were patients who remained SRF on SD-OCT at 1 month after IVB injection. Complete absorption was defined as the absence of SRF in an OCT 6 X 6 mm macular cube scan (Heidelberg Spectralis, Heidelberg, Germany); otherwise, it is defined as remaining SRF.

Aqueous humor sampling

Aqueous humor (AH) samples were collected before IVB injection in the subacute CSC and during cataract surgery in the controls. To mitigate the occurrence of IOP spikes following injections [56], we performed a paracentesis procedure while sampling aqueous humor for study purpose prior to intravitreal injection after the permission from the patients. All the paracentesis procedure that we performed was very safe and prevented complications such as IOP surge. Patients were informed and signed a written consent form for the collection and scientific use of the specimen before the analysis. Approximately 100 to 200 µl of aqueous humor were collected from each patient, based on the amount of a previous study [16, 22]. The collected aqueous humor samples from the patients with CSC were sorted by the response to the bevacizumab as CSC responders or non-responders. Each group was pooled into Protein LoBind Tube. Pooling volumes of each group were control (N = 20) 3.5 mL, CSC responder (N = 17) 2.0 mL, and CSC non-responder (N = 25) 3.0 mL, and the mean volume for each group is 175 µl (3500/20), 117 µl (2000/17), and 120 µl (3000/25) for control, CSC responder, and CSC non-responder, respectively.

Exosomal miRNA-sequencing

Exosomal RNA was isolated using ExoLutE® Exosome Isolation Kits (Rosetta, Seoul, Korea), following the manufacturer’s instructions. Quality control and quantity measurement of the RNA was performed by Agilent 2100 Bioanalyzer using the RNA 6000 Pico Chip (Agilent Technologies, Amstelveen, The Netherlands). RNA quantification was performed using a NanoDrop 2000 Spectrophotometer system (Thermo Fisher Scientific, Waltham, MA, USA).

For control and test RNAs, library construction was performed using the NEBNext Multiplex Small RNA Library Prep kit (New England BioLabs, Inc., USA) according to the manufacturer’s instructions. Briefly, for library construction, 1 µg of each total RNA sample was used to ligate the adaptors, and then cDNA was synthesized using reverse-transcriptase with adaptor-specific primers. PCR was performed for library amplification, and libraries conducted clean-up using QIAquick PCR Purification Kit (Qiagen, Inc, German) and AMPure XP beads (Beckman Coulter, Inc., USA). The yield and size distribution of the small RNA libraries were assessed by the Agilent 2100 Bioanalyzer instrument for the high-sensitivity DNA assay (Agilent Technologies, Inc., USA). High-throughput sequences were produced by the NextSeq500 system as a way of single-end 75 sequencings (Illumina, SanDiego, CA, USA).

Raw and processed data have been deposited at GEO ( (accession no. GSE227142).

Data analysis

Sequence reads were mapped by the bowtie2 software tool to obtain the bam file (alignment file). A mature miRNA sequence is used as a reference for mapping. Read counts mapped on mature miRNA sequences were extracted from the alignment file using bedtools (v2.25.0) and Bioconductor, which uses R (version 3.2.2) statistical programming language (R Development Core Team, 2011). Read counts were used to determine the expression level of miRNAs. The quantile normalization method was used for comparison between samples. Differentially expressed genes (DEGs) expression analysis was performed using DESeq2 [26]. Significant DEGs were defined as P < 0.05 and an absolute log2 fold change of > 2. For miRNA target and functional study, miRNet and miRWalk 2.0 were utilized [57, 58]. Graphical representations were generated using ggplot2.

Single patient-derived aqueous humor preparation

Aqueous humor samples were delivered in individually sealed sterile syringes on ice, fresh or fresh-frozen. All samples were delivered to the lab and prepared within 4 h of acquisition. Fresh-frozen aqueous humor samples were thawed at 4 °C for 2 h. Fresh or fresh-frozen aqueous humor samples were filtered using a 4 mm RC membrane syringe filter (Corning, #431212, US) to minimize sample loss and collected at Protein LoBind Tube (Eppendorf, #022431081, Germany). The filtered aqueous humor was centrifuged at 15000 × g, 4 °C, for 30 min to remove cell debris and proteins. The supernatant was collected and proceeded for exosome purification and analysis.

Transmission electron microscopy (TEM) for exosome

100 μl of freshly filtered-collected aqueous humor samples were ultracentrifuged for 120000 × g, 4 °C, 2 h. The supernatant was discarded so as not to disturb the pellet. The pellet was then reconstituted at 10 μL sterile PBS. The reconstituted liquid was dropped on a nickel-carbon grid (Electron Microscopy Sciences, #CF200-Ni-50, US), and 2.5% glutaraldehyde was dropped on the grid for 1:1 volume for 10 min for fixation. 1% uranyl acetate solution was dropped for gentle flow-thru and left in RT for 2 min for exosome staining. The grid was washed with sterile distilled water 3 times with gentle dropping-flow-thru and let dry. The images were immediately obtained using a transmission electron microscope (H-7000, Hitachi, Japan).

Exosome purification

100 μL of freshly filtered-collected aqueous humor samples were collected in the new tube, and the exosome was isolated using ExoQuick-TC Ultra for Tissue Culture Media (System Bioscience, #EQULTRA-20TC-1, US), modifying the manufacturer’s protocol to accommodate the smaller sample volume. Briefly, the buffers were used for 1/20 volume than the recommended protocol. The isolated exosomes were ultracentrifuged for 120000 × g, 4 °C for 2 h, and the supernatant was discarded without disturbing the pellet. Pellets were reconstituted or lysed in buffers for the next analyses.

Exosome size analysis

For the size analysis of purified exosomes, the pellet was reconstituted in sterile DEPC water-based 1 × PBS to create a final volume of 100 μL. The fresh aqueous humor or the purified-reconstituted exosome samples were diluted to 1:3 with sterile PBS. For DLS (Direct Light Scattering) analysis, particle size was measured using Zetasizer (NANO-ZS ZEN3600, Malvern Pananlytical, UK). The analytical report was obtained after performing 10 measurements per each sample for three different samples in the disease group. The concentration was compensated using a dilution factor of 3.00e + 0. Nanoparticle Tracking Analysis (NTA) was performed for individual samples using NanoSight (NS300, Malvern Pananlytical, UK). The traces were recorded for 60 s, 5 times per sample at 25 °C. The concentration was compensated using a dilution factor of 3.00e + 0.

Exosomal RNA isolation and analysis

For the exosome-purified pellet, 35 μL of QIAzol was added and proceeded for total RNA isolation including both mRNA and miRNA using miRNeasy Micro Kit (Qiagen, #1071023, Germany) following the provided protocol. For mRNA analysis, weverse transcription was performed to synthesize cDNA using PrimeScript 1st strand cDNA synthesis kit (TAKARA, #6110A, JAPAN). Real-time qPCR for STC2 mRNA was performed with BioRad CFX Connect using SsoAdvance Universal Supermix (BioRad, #1725270, USA). For miRNA analysis, reverse transcription was performed with 10.0 ng RNA using miR-184-specific reverse transcription primer and quantified with miR-184 probe-based real-time qPCR using TaqMan MicroRNA Reverse Transcription Kit (Applied Bioscience, #4366596, US), miR-184-specific TaqMan MicroRNA Assay (Applied Bioscience, #4427975, US), and TaqMan Universal Master Mix II (Applied Bioscience, #4440040, US) following the manufacturer’s protocol. Real-time qPCR was performed with BioRad CFX Connect and analyzed with CFX Maestro software.

Aqueous humor miRNA isolation and analysis

Fresh aqueous humor sample or exosome purified liquid was filtered using a 4 mm RA membrane syringe filter, and filtered aqueous humor was collected. TRIzol LS (Invitrogen, #10296028, US) was added to the filtered aqueous humor following the manufacturer’s protocol. miRNA isolation and miR-184 detection were performed as an exosomal miRNA analysis procedure.

Human primary retina and RPE-choroid tissue culture

Human primary retina and RPE/choroid tissues were isolated from fresh donor eyes; a 60-year-old female donor without metabolic or ocular disease history. The retina and choroid tissue were separated from the donor eye immediately after cornea buttoning and was radially cut into four leaves. The tissues were cultured in a 60pi culture dish each in a mixture of culture media for eye tissue culture; Exo-free FBS (System Biosciences, #EXO-FBS-250A-1, US) was mixed with serum-free DMEM F/12, Neurobasal Media supplemented with B27, Pericyte Growth Media and Endothelial Growth Media was mixed in adequate proportions for each retina and RPE-choroid and cultured for 36 h.

Human primary tissue-derived exosomal miRNA analysis

The media soup was collected 36 h after tissue culture. Exosomes from the collected soup and naïve media were purified using ExoQuick-TC Ultra (System Biosciences, #EQULTRA-20TC-1, US) for Tissue Culture Media for the aqueous humor exosome purification, following the provided protocol, and isolated as a pellet by ultracentrifugation.

Human primary choroidal endothelial cell isolation and culture

Human primary choroidal endothelial cells (hCEC) were isolated from fresh donor eyes aged approximately 30 without metabolic or ocular disease history; 33-year-old female donor and a 31-year-old male donor. The choroid tissue was separated from the donor eye immediately after cornea buttoning and was dissociated into single cells using 1U/mL of Collagenase/Dispase (Roche, #10269638001, Germany) in a shaking incubator (37 °C, 200 rpm, 2 h) and filtered through a 40 μm cell strainer. The CD31-positive endothelial cells were isolated using Dynabeads magnet-based cell isolation system with anti-CD31 Dynabeads (ThermoFisher, #11155D, US) following the manufacturer’s protocol. CD31-positive hCECs were cultured in a 1% gelatin-coated dish with microvascular endothelial cell growth medium-2 (EGM2-MV; LONZA, #CC-3202, Switzerland) in 37 °C, 5% CO2 condition up to passage 2.

miRNA mimic/inhibitor transfection

miR-184 mimic, inhibitor, and non-targeting scrambled negative control RNAs (Bioneer Co. Ltd, Daejeon, Korea) were used. Cells were transfected with miR-184 mimics, inhibitors, or negative control RNAs using Lipofectamin™ RNAiMAX Transfection reagent (Invitrogen, #13778150), according to the manufacturer’s instructions. Synthetic miRNA oligomers were complexed with the transfection reagent in Opti-MEM, reduced serum medium, and added to cells. Cells were then used for the following assays 24–48 h after transfection.

STC2 siRNA transfection

STC2 siRNA and scrambled siRNA for negative control (Bioneer Co. Ltd, Daejeon, Korea) were used to induce STC2 knock-down. Cells were transfected with siRNAs using Lipofectamin™ 3000 Transfection reagent (Invitrogen, # L3000008), according to the manufacturer’s instructions. Synthetic siRNA oligomers were complexed with the transfection reagent in Opti-MEM, reduced serum medium, and added to cells. Cells were then used for the following assays 24 h-36 h after transfection.

Total RNA real-time quantitative PCR (qPCR)

Total RNA was extracted from the cultured cells by TRIzol Reagent (Invitrogen, Carlsbad, CA, USA) per the manufacturer’s instructions. Reverse transcription was performed to synthesize cDNA using PrimeScript 1st strand cDNA synthesis kit (TAKARA, #6110A, JAPAN). Real-time qPCR was performed with BioRad CFX Connect using iQ SYBR Green Supermix (BioRad, #1708880, USA) and analyzed with CFX Maestro software.

The qPCR primers and their sequences are listed in Additional file 3: Table S1.

3D tube formation assay

hCEC transfected with miRNA oligo (miR-184 mimic, inhibitor or negative controls) and siRNA oligo (STC2 siRNA or negative control) were seeded in a Matrigel® Matrix (Corning, NY, USA) coated 96-well plate at 2 × 104 cells/well density. Cells were incubated at 37 °C, 5% CO2 for 4 h, and imaged at 4 h after seeding. Tube formation parameters were analyzed using WimTube online software (Onimagin, Córdoba, Spain).

In vitro wound healing assay

hCEC transfected with miRNA oligo (miR-184 mimic, inhibitor or negative controls) and siRNA oligo (STC2 siRNA or negative control) were seeded at a concentration of 2 × 105 cells/well at a 24-well plate to reach the confluent monolayer in 24 h. Cell monolayers were scraped with a 200 μL pipette tip and incubated at 37 °C, 5% CO2. Since the doubling time of the endothelial cells are more than 17 h, images of the cells were taken immediately after scraping and at 12 h to analyze the wound closure by cell migration [59]. The closed wound spaces were quantified and normalized to the originally generated wound spaces.

In vitro 3D microfluidic angiogenic assay

Microfluidic plastic chips and chip holders to maintain chip humidity were purchased from AIM Biotech company (AIM Biotech, Singapore). Collagen type I solution (Corning, NY, USA) at 2 mg/mL was gently pipetted into the gel-filling inlet of the devices and polymerized for 30 min at 37 °C and 5% CO2 assembled in the humidified chip holder. 1/10 diluted human plasma fibronectin (Sigma-Aldrich, #F0895, US) was injected into the microchannels, and the device was incubated for 1 h at 37 °C and 5% CO2 in the humidified chip holder. hCEC transfected with miR-184 mimic/inhibitor or negative controls were seeded in one of the fluidic channels at 6 × 104 cells/channel. Twenty-four hours after seeding, recombinant Human VEGF165 (Peprotech, #100–20, Rocky Hill, NJ, USA) at a final concentration of 40 and 20 ng/mL according to the manufacturer’s instructions were added to the growth media to generate a VEGF gradient, as confirmed in our previous study [60]. Cells were incubated at 37 °C and 5% CO2, and the medium with VEGF was changed every 24 h. Cell sprouting was monitored and imaged every 24 h. The number of tip cells/mm was counted from the acquired image for sprouting angiogenic property parameters.

STC2 Western blot analysis

Cells transfected with miR-184 mimic/inhibitor or negative controls for 48 h were washed with three changes of PBS. After removing PBS, cells were immediately frozen with liquid N2. Cold 1 × Pierce RIPA Buffer (Thermo Scientific, #89,900, US) containing 1 × HALT Phosphatase Inhibitor Cocktail (Thermo Scientific, #78420, US) and 1 × HALT Phosphatase Inhibitor Cocktail (Thermo Scientific, #78430, US) were added immediately for cell lysis. Cells were scraped from the plastic surface and collected in Protein LoBind Tubes. After 10 min incubation on ice, cell lysates were centrifuged (13000 g, 4 °C, 30 min), and supernatants containing proteins were acquired. Protein concentration was measured using the Pierce BCA Assay Kit (Thermo Scientific, #23227, US) following the manufacturer’s protocol. Samples were blotted to the NC membrane (GE Healthcare Life Science, #10600114, Germany) after electrophoresis. The blotted membrane was then blocked using EveryBlot Blocking Buffer (BioRad, #12010020, US). The antibodies used for detection were rabbit anti-STC-2 (Abcam, #ab63057, US), and mouse anti-α-tubulin (Santa Cruz, #SC-5286, US), goat-anti-rabbit-HRP (Genetex, #GTX213110-01, US), and goat-anti-mouse-HRP(Genetex, #GTX213111-01, US) diluted to the manufacturer’s recommended concentration.

Statistical analysis

The representative values are presented as mean ± standard deviation (SD). Statistical significance was analyzed with Welch’s t-test and Student’s t-test and defined as * P < 0.05, **P < 0.01, and *** P < 0.001. Statistical analyses were performed using R × 64 v4.1.1.

Image analysis

The microscopic images for cell functional assay analysis were imaged using an inverted microscope (Olympus IX70, Japan) using DP controller software. The images were quantified and analyzed using Java-based imaging software (ImageJ, v.1.52p, in the public domain at; National Institutes of Health (NIH), Bethesda, MD, USA) [26].