Microarray Analysis of Small Extracellular Vesicle-Derived miRNAs Involved in Oxidative Stress of RPE Cells

The aim of this study was to investigate the miRNA profiles of nanosized small extracellular vesicles (sEVs) from human retinal pigment epithelial (RPE) cells under oxidative damage. ARPE-19 cells were cultured with ox-LDL (100 mg/L) or serum-free medium for 48 hours, sEVs were then extracted, and miRNA sequencing was conducted to identify the differentially expressed genes (DEGs) between the 2 groups. RNA sequence results were validated using quantitative real-time PCR. The Gene Ontology (GO) enrichment, Kyoto Encyclopedia of Genes and Genomes pathway, and ingenuity pathway analyses (IPA) were performed for the DEGs. Results revealed that oxidative stress inhibited RPE cell viability and promoted sEV secretion. A total of 877 DEGs from sEVs were identified, of which 272 were downregulated and 605 were upregulated. In total, 66 enriched GO terms showed that the 3 most significant enrichment terms were cellular processes (biological processes), cell (cellular component), and catalytic activity (molecular function). IPA were used to explore DEGs associated with oxidation damage and further construct a miRNA-target regulatory network. This study identified several DEGs from oxidation-stimulated RPE cells, which may act as potential RNA targets for prognosis and diagnosis of RPE degeneration.


Introduction
Oxidative damage is one of the major contributors to retinal degenerative diseases such as age-related macular degeneration (AMD) [1]. AMD is a multifactorial disease in which oxidative stress serves as a key component. The retinal pigment epithelium (RPE) is a highly specialized, polarized epithelium, which is in intimate contact with the outer segments of the photoreceptor and Bruch's membrane [2]. PRE cells are particularly metabolically active, highly oxygenated, and vulnerable to oxidative stress under exposure to photosensitizers such as lipofuscin [3]. Oxidative stress induces cell apoptosis through reactive oxygen species, thereby leading to RPE dysfunction [4].
Exosome refers to one form of extracellular nanometersized vesicle, which mediates multiple extra-and intercellular activities, including cell-cell communication, immune modulation, extracellular matrix turnover, stem cell division/differen-tiation, neovascularization, and cellular waste removal [5]. RPE cells secrete extracellular vesicles (EVs) in response to oxidative stress, resulting in angiogenesis in endothelial cells [6]. Exosomal biological markers CD63 and LAMP2 have been found in the drusen of AMD patients and stressed RPE cells, which suggests that the drusen is initiated by intracellular proteins from RPE cells and becomes extracellular via the exosome [7]. Small extracellular vesicles (sEVs) contain multiple functional molecules such as mRNA, microRNA (miRNA), and proteases. miRNAs are small noncoding RNA molecules, which inhibit several targeting mRNA expressions. Genetic mutations of miRNAs induce pathophysiological and immunological dysfunction in RPE cells. A related study suggested exosomal miRNA variations as predictive biomarkers in AMD disease [8]. Here, we aimed to screen for differentially expressed miRNA profiles in sEVs derived from oxidative stress-stimulated RPE cells and identify potential functional miRNAs, which may be associated with RPE oxidation.
2.2. CCK8 Assay for RPE Cell Viability. ARPE-19 cells were seeded at a density of 1 × 10 4 cells/100 μL/well in 96-well plates. After the treatment mentioned above, 10 μL of XTT (BBI Life Sciences, China) solution was added into each well for 1 hour at 37°C. Cell viability was determined by measuring absorbance at 450 nm using a microplate spectrophotometer (Molecular Devices, Sunnyvale, CA, USA).  hours. Cell viability was tested by CCK8 assay. Data are expressed as mean ± SD (n = 3). Experiments were repeated 3 times. * * * p < 0:001 vs. the control group. 2 Oxidative Medicine and Cellular Longevity

Exosome Isolation and Transmission Electron Microscopy
Imaging. Exosomes were isolated from ARPE-19 cells using multistep differential centrifugation [9]. ARPE-19 cells were centrifuged at 300 × g for 10 minutes at 4°C. Subsequently, the supernatant was subjected to the following centrifugation steps: 2000 × g for 10 minutes, 10,000 × g for 30 minutes, and 100,000 × g for 70 minutes. The resulting sEVs were finally resuspended in PBS and centrifuged at 100,000 × g for 70 minutes again. The morphology of sEVs was visualized using the Hitachi transmission electron microscope operated at 80 kV (Hitachi, Japan).
2.4. Western Blot Analysis. As described previously [10], after incubating for 5 min at 90°C with loading buffer (Life Technologies, Australia), 10 μg of exosomes in each group was electrophoresed on NuPAGE Novex 4-12% Bis-Tris Gels (Life Technologies, USA). Gels were transferred onto PVDF membranes using the Trans-Blot Turbo system. Membranes   3 Oxidative Medicine and Cellular Longevity were blocked in 2% BSA solution for 3 hours and then probed overnight with primary exosomal marker protein antibodies: anti-Hsp70 (ab134045, Abcam, Cambridge, UK), anti-CD63 (ab181606, Abcam), and anti-TSG101 (ab125011, Abcam) at 4°C, followed by incubation with a secondary antibody for 3 hours. The ChemiDoc XRS gel documentation system (Bio-Rad Laboratories, USA) was used to quantify the immune-reactive proteins, and β-actin was used as a loading control for each lane. Each indicated band was quantified and normalized to β-actin through ImageJ software.
2.5. miRNA Extraction and miRNA Sequencing. As reported previously [11], RNA extraction was performed using the Total Exosome RNA and Protein Isolation Kit (catalog # 4478545; Invitrogen, USA) according to the provided instructions. 200 ng-1 μg RNA in final volume of 30 μL solution was collected for each sample. Total RNA quantity and quality (260/280 absorbance ratio) were assessed using NanoDrop 2000 (Thermo Fisher Scientific, Waltham, MA, USA) and Agilent 2100 Bioanalyzer to test concentration and inorganic ions or polycarbonate contamination. miRNA sequence was isolated by BGI Company (China) based on previous instructions [12]. cDNA libraries were constructed using the Ion Total RNA-Seqv2 kit (Life Technologies, USA) (n = 3 for each group) and purified using AMPure beads (Beckman Coulter). Emulsion PCR and enrichment of cDNA-conjugated particles were performed with an Ion OneTouch 200 Template Kit v2 DL (Life Technologies). The final cDNA samples were sequenced single end on the HiSeq 2000 System with a 50 bp read length.
2.6. Bioinformatics Analysis of the Data. Raw data was filtered to eliminate low-quality reads, primers, adaptors, and other contaminants. Following this, we summarized the length distribution and common and specific sequences between samples. After filtering, the remaining tags were called clean tags and stored in FASTQ format. Bowtie2 was used to map clean reads to the reference genome and other sRNA databases. To identify differentially expressed genes (DEGs), differentially expressed miRNAs (DEMIs) were screened out using the limma package through the thresholds of fold change > 2 or <0.5 and adjusted p value of <0.05 [13].
To perform Gene Ontology (GO) enrichment analysis, we mapped all genes to GO terms in the database, which calculated the gene numbers for every term. The hypergeometric test was then used to find significantly enriched GO terms in the input gene list.
The Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway was used to perform pathway enrichment analysis. This analysis identified significantly enriched metabolic or signal transduction pathways from target genes of DEGs when compared with the whole genome background. The p value was corrected using the Bonferroni method; a corrected p value < 0.05 was considered significant.
In this study, experimentally verified miRNA-mRNA regulatory pairs were obtained using TargetScan and miRanda and a miRNA-target regulatory network was constructed by comparing the DEGs with obtained miRNA-gene regulatory pairs.

Validation of miRNA Expression Using Quantitative
Real-Time PCR (RT-PCR). In order to validate initial miRNA sequence results, the 10 most significant up-or downregulated miRNAs were selected for further RT-PCR tests as reported previously [14]. Total RNA was isolated using TRIzol reagent and the quality and quantity of RNA was measured using a NanoDrop 2000 spectrophotometer. Each reverse transcription reaction mixture contained 10 mL of SYBR Green Master Mix, 0.5 mL of miR-RT primers F (10 mM), 0.5 mL of miR-RT primers R (10 mM), and RNase-free H 2 O. The RT-PCR reactions for the selected 10 miRNAs were performed using the ViiA 7 Real-Time PCR System (ABI, USA) under the following conditions: 95°C for 1 min, followed by 40 PCR cycles (95°C for 10 s and then 60°C for 20 s). miRNA expression was normalized to the endogenous reference gene GAPDH. Each sample was  Figure 4: Validation of the top 10 selected DEGs screened from miRNA sequence by qRT-PCR tests. ARPE-19 cells were treated as before, and miRNAs were extracted from sEVs. Data was expressed as mean ± SD (n = 6). Experiments were repeated three times. * p < 0:05 vs. the control group. ns: no significance. 4 Oxidative Medicine and Cellular Longevity analyzed in triplicate. Specific primers were produced by BIOTNT Company (Shanghai, China). Relative quantification was achieved by the comparative 2 −ΔΔct method.
2.8. Statistics. The data were analyzed with a one-way analysis of variance (ANOVA) using the statistical program SPSS 17.0. All data were presented as mean ± SD. p value < 0.05 was considered statistically significant.

ox-LDL Decreases ARPE-19
Cell Viability. We first measured the cytotoxicity of ox-LDL to ARPE-19 cells after 48 hours. Figure 1 shows that cell viability in the ox-LDL group was significantly lower than that in the control group (p < 0:05), which indicated its cytotoxicity and aligned with previous conclusions [15,16].

Characterization of sEVs and Biological Marker Protein
Detection. Transmission electron microscopy images of sEVs derived from both groups revealed the presence of distinct vesicles with an average diameter of 106 ± 7:62 nm (Figure 2(a)). The vesicles were also positive for exosomal markers. CD63 is the general tetraspanin protein used as the exosomal "star marker" [17]. TSG101 and Hsp70 are also commonly used for exosome detection [18,19]. As shown in Figure 2(b), we found that the expression levels of TSG101 and Hsp70 were statistically higher in the ox-LDL group than the control, but it did not reach a significant difference (p > 0:05).     Oxidative Medicine and Cellular Longevity 3.3. Differential Expression of miRNA Profiles in sEVs Isolated from the Oxidative and Control RPE Cells. In order to identify the influence of oxidation on miRNA profiles from sEVs of RPE cells, miRNA sequence was sequenced after treatment for 48 hours. Figure 3 shows that 877 significantly differentially expressed miRNAs had been screened between the ox-LDL and control groups, among which 272 were downregulated and 605 were upregulated. The top 10 differentially expressed genes are listed in Table 1.

DEG Validation
Using RT-PCR. The top 10 selected DEGs were further validated using RT-PCR (Figure 4). PCR tests revealed similar results of RNA sequence, except that miR-138-5p showed insignificant differences between the two groups (p > 0:05).

Gene Ontology Enrichment
Analysis. The GO analysis contains three ontologies: biological processes, molecular function, and cellular components. We identified 66 enriched GO terms, among which 26 belong to biological processes, 21 belong to molecular function, and 19 belong to cellular components. The three most enriched biological process terms were cellular processes, single-organism processes, and metabolic processes. Meanwhile, cell, cell part, organelle, catalytic activity, transporter activity, and transporter activity were the most enriched GO terms of cellular components and molecular function, respectively ( Figure 5).
3.8. miRNA-Target Regulatory Network Analysis. We used TargetScan and miRanda software to predict possible targeted mRNAs for DEGs and associated their intersections with the IPA results; a total of 10 miRNAs and 43 targeted mRNAs formed a miRNA-target regulatory network (Figure 7).

Discussion
Oxidative stress has been recognized as a major influence in AMD pathophysiology, and RPE appears to be the main site of damage [2]. Oxidative damage of the RPE layer originates from the digestion of photoreceptor outer segments and other reactive oxygen species. RPE damage occurs in multiple locations within the central part of the eye and finally forms a region of atrophy by the bystander effect, which is mediated via EVs [20]. In this research, ox-LDL decreased ARPE-19 cell viability and promoted sEV secretion. RNA sequences and RT-PCR tests confirmed a downregulation of miR-1910-5p in sEVs of the ox-LDL group, which is contrary to a similar study that found that H 2 O 2 increases miR-1910-5p concentrations in ARPE-19 cells [21]. miR-324-5p was reported to be expressed in plasma of wet AMD patients, and we found a decreased expression in the ox-LDL group [22]. Desjarlais et al. [23] demonstrated an upregulation of let-7g-5p (>570%) in oxygen-induced retinopathy models during the neovascularization phase, which is consistent with our result. Other authors reported significant changes of miR-192, let-7c, miR-183, miR-27a, miR-27b, miR-361-5p, miR-335, and miR-30c in experimental AMD models, which were also observed in our study (see Table 3) [24][25][26]. KEGG analysis suggested cytokine-cytokine receptor interactions and phagosome and protein processing in the endoplasmic reticulum to be the most significant enrichment items. In agreement with other reports, cancer-related pathways are also involved in DEGs of sEVs in our study [15]. Further research is needed to explore the specific roles of these pathways. IPA screened out 6 key pathways related to AMD and 4 GO terms related to oxidative damage. miR-138-5p, miR-  Figure 7: microRNA-target regulatory networks of differentially expressed genes (DEGs). Green triangles represent miRNAs; red circles represent targeting genes. Oxidative Medicine and Cellular Longevity 345-5p, miR-210-5p, miR-34a-5p, miR-1908-5p, miR-1343-3p, miR-485-5p, miR-423-5p, and miR-4488 are associated with oxidative stress and AMD.
Compared with previous attempts at this type of analysis, a lower number of identified DEGs coincided with this study, which is probably due to the use of different oxidative injury models in RPE cells. Our research investigated acute responses of RPE cells to oxidative stress, which could not represent pathogenesis of AMD since it is a long-term effect.

Conclusion
n conclusion, exploring oxidative stress-induced miRNA profiles has led us to potential prospects in evaluating RNA variation in sEVs, which may be useful as prognostic and diagnostic tools in the future.

Data Availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.

Conflicts of Interest
The authors report no conflicts of interest.

Authors' Contributions
Professor Xingwei Wu conceived the study, participated in its design and coordination, and helped draft the manuscript. Dr. Ke Mao participated in the design and operation of this experiment, performed the statistical analysis, and drafted the manuscript. All authors have read and approved the final manuscript.