Integrated Profiling of Extracellular Vesicle microRNA Impact on Trabecular Meshwork mRNA Expression: Insights From Microarray Analysis

Abstract

Purpose

Extracellular vesicles (EVs) secreted by non-pigmented ciliary epithelial (NPCE) cells under oxidative stress may contribute to primary open-angle glaucoma (POAG) pathogenesis by altering gene expression in human trabecular meshwork (HTM) cells. This study investigated the impact of microRNAs (miRNAs) carried by NPCE-derived EVs on HTM cell gene expression under oxidative stress conditions.

Methods

NPCE cells were exposed to oxidative stress, and EVs were isolated from control and stressed cells. HTM cells were treated with these EVs, followed by microarray analysis to identify differentially expressed miRNAs in EVs and messenger RNAs (mRNAs) in HTM cells. Bioinformatics analysis was used to explore miRNA–mRNA interactions, enriched Gene Ontology (GO) terms, and miRNA–mRNA–GO networks.

Results

The study identified 54 differentially expressed miRNAs in stressed NPCE EVs. In HTM cells treated with stressed NPCE EVs, 88 genes were upregulated and 58 downregulated. GO analysis of upregulated genes showed enrichment in processes such as extracellular matrix organization, cell proliferation, and adhesion. Downregulated genes were associated with oxidative phosphorylation and adenosine triphosphate (ATP) biosynthesis. Notably, 59 out of 88 upregulated genes are known targets of downregulated miRNAs. Network analysis identified interactions between downregulated miRNAs and upregulated genes involved in key biological processes relevant to POAG pathogenesis.

Conclusions

This study provides new insights into the potential role of NPCE-derived EVs and their miRNA cargo in POAG, suggesting novel mechanisms for disease progression and potential therapeutic targets for further investigation.

Primary open-angle glaucoma (POAG) is an eye disease that progressively damages retinal cells and optic nerves, potentially leading to blindness if left untreated.¹ The aqueous humor (AH) is a clear fluid that fills the front portion of the eye, between the lens and the cornea. An imbalance in the production and/or drainage of AH can lead to elevated intraocular pressure (IOP), which is harmful to the optic nerve.² The trabecular meshwork (TM) is responsible for draining the AH. It plays a crucial role in maintaining IOP balance by remodeling its extracellular matrix (ECM).^3,4 In contrast,a dysfunctional TM can lead to increased IOP, subsequently causing degeneration of the optic nerves and POAG.^4–7

Oxidative stress (OS) exhibits a higher positive correlation with POAG. Studies have shown raised levels of AH antioxidant enzymes such as superoxide dismutase, glutathione peroxidase, and catalase in the POAG group compared to the control group.^8–11 In addition, long-term accumulation of oxidative damage due to mitochondrial failure and endothelial dysfunction causes molecular changes in the ocular anterior chamber during the early stages of glaucoma, leading to decreased antioxidant defenses in AH and the activation of apoptosis in TM cells, altering tissue function and integrity.¹²

Extracellular vesicles (EVs), including exosomes, are membrane-enclosed structures composed of a lipid bilayer that facilitates intercellular communication, even across distant cells. In line with the International Society for Extracellular Vesicles nomenclature guidelines, we consistently refer to these vesicles as “EVs” throughout the study.¹³ Studies have confirmed that EVs originating from cells exposed to OS possess the ability to protect target cells from the adverse effects of OS-induced cellular dysfunction and death.^14–16 The effectiveness of this protective mechanism can be ascribed to the precise adjustment of microRNA (miRNA) profiles within these EVs, enabling them to selectively influence crucial pathways such as Nrf2/Keap1, which is recognized as a pivotal regulatory pathway for cellular adaptation during OS in the target cells.^14,17,18

In theory, EVs originating from non-pigmented ciliary epithelium (NPCE) cells, the site of AH production, should readily reach TM cells. Building upon our prior research, we have successfully demonstrated the regulatory influence of NPCE cells on TM cells in an in vitro setting. This impact involves downregulation of the Wnt signaling pathway within TM cells, which is mediated bymammalian target of rapamycin (mTOR) and leads to decreased cadherin expression, thereby mitigating the buildup of collagen and ECM by TM cells.^19,20

To decipher the message codes among cells in remote sites, the connection between miRNAs in donor EVs and messenger RNAs (mRNAs) in recipient cells should be elucidated. miRNAs exert negative regulation on their target mRNAs, making it essential to explore miRNA and mRNA groups where this control is reversed (i.e., downregulated miRNAs with upregulated mRNAs, and vice versa) to fully understand their role in these interactions.

Microarray assay is a high-throughput, genome-wide miRNA/mRNA expression profiling technique widely used in the research of biological mechanisms and various disease diagnoses.^21–23 The binding of miRNA/mRNA to the probes immobilized on a silicon chip generates solid signals, which are detected and recorded by a machine that transforms the data for downstream analysis. This includes assessing differential expression of miRNAs/mRNAs and miRNA–mRNA interactions.

Previous studies from our laboratory have described some physiological effects of NPCE EVs on TM cells under OS conditions.^24–26 However, the specific miRNA–mRNA interactions mediating these effects remain largely unknown, representing a significant gap in our understanding of the molecular mechanisms at play. To address this knowledge gap, our study employed a comprehensive approach using miRNA/mRNA microarray assays and advanced bioinformatic tools. Our objectives were threefold:

• 1.Elucidate the miRNA profiles and identify differentially expressed miRNAs (DEmiRNAs) in NPCE EVs under OS conditions, providing insight into potential regulatory molecules.
• 2.Identify and characterize differentially expressed genes (DEGs) in TM cells following exposure to OS-induced NPCE EVs, thereby uncovering potential gene targets of DEmiRNAs.
• 3.Investigate the intricate relationships between DEmiRNAs in NPCE EVs and DEGs in TM cells.

By pursuing these objectives, we aimed to bridge the current knowledge gap in miRNA–mRNA interactions and provide a more comprehensive understanding of the molecular mechanisms underlying NPCE EV–mediated changes in TM under OS conditions. This research has the potential to unveil novel regulatory pathways and contribute to the development of more targeted treatments for glaucoma and other ocular disorders associated with TM dysfunction.

Materials and Methods

Cell Culture

The immortalized NPCE cell line²⁷ was kindly donated by M. Coca-Prados (Yale University, New Haven, CT, USA). Cells were used for experiments when they reached approximately 80% confluence, as this ensures optimal cell growth and function, minimizing variability. Cells were passaged no more than 20 times to preserve their physiological properties and avoid genetic drift.

Cell line authentication tests were performed at the Genomics Center of the Biomedical Core Facility (Technion, Israel) using the GenePrint 24 System (Promega Corporation, Madison, WI, USA) (Supplementary Materials A). Human donor eye tissue was obtained with authorization for research use, as specified by the CorneaGen Tissue Request Form. The donor was a 68-year-old individual who provided two corneas (identification numbers W4192 22 014171 V0003000 and W4192 22 014171 V0004000). The donor's gender and race were not provided. All experiments were conducted in accordance with the tenets of the Declaration of Helsinki for the use of human tissue. The isolated TM cells were characterized according to the consensus recommendations to ensure their identity and purity (Supplementary Material B).^28,29

Oxidative Stress

To induce OS in NPCE cells, we utilized 2,2′-azobis (2-amidinopropane) dihydrochloride (AAPH; 440914; Sigma-Aldrich, St. Louis, MO, USA), a free-radical–generating azo compound. AAPH initiates oxidation reactions by continuous production of peroxyl radical followed by alkoxyl radical via nucleophilic and free radical mechanisms.³⁰ Based on a previous OS study conducted in our lab, NPCE cells were treated with 1.5-mM AAPH for 24 hours.^24,25

EV Isolation

EV-depleted media were created by mixing Dulbecco's modified Eagle's medium with regular fetal bovine serum, followed by ultracentrifugation at 110,000g for 16 hours at 4°C and filtration through a 0.22-µm filter. The media were utilized for subsequent experiments. Post-experiment, EVs from the conditioned media were isolated using a series of ultracentrifugation steps. These included centrifugation at 300g for 10 minutes to remove cells, 2000g for 10 minutes to eliminate dead cells, and 10,000g for 30 minutes to clear cell debris. The final steps involved centrifugation at 100,000g for 70 minutes, repeated twice, as previously described.³⁰

EV Size and Concentration Analysis

Isolated EVs were assessed using the NanoSight NS500 instrument (Malvern Panalytical, Malvern, UK) with a 405-nm blue laser. Nanoparticle Brownian motion was recorded for 60 seconds, and videos underwent nanoparticle tracking analysis (NTA) through NanoSight software (NTA 2.0). Camera sensitivity was set to levels 16 to 17 for optimal particle detection without saturation. Size distributions were averaged across replicates and normalized to total concentrations or cell counts. At a 1:1000 dilution, EV concentrations were approximately 6 × 10⁷/mL, and data are presented as mean EV size and mode.

EV miRNA Extraction and Characterization

Total RNA extraction from EVs utilized the Total Exosome RNA Purification Kit (17200; Norgen Biotek Corp., Thorold, ON, Canada) following the manufacturer's guidelines. Briefly, 200 µL of EVs in PBS underwent 1-minute lysis in buffer through vortexing. The miRNA separation employed spin columns at 3500g for 1 minute, with four subsequent washes at 14,000g for 1 minute and elution in 50 µL at 600g for 2 minutes. The miRNA quality was assessed with the Bioanalyzer 2100 (Agilent Technologies, Santa Clara, CA, USA). Samples exceeding 130-ng/µL mRNA concentrations were chosen for further analysis. Exosomal miRNA profiling utilized the Applied Biosystems GeneChip miRNA 4.0 Assay and Flashtag Bundle (902445; Thermo Fisher Scientific, Waltham, MA, USA).

TM Cell Treatment

Primary TM cells were seeded at 0.75 × 10⁶ cells/well in six-well plates. After 48 hours, when the cells had doubled (∼1.5 × 10⁶ cells/well), they were treated with 2 µg EV protein per mL of media (2 mL/well), totaling 4 µg EV protein (∼6 × 10⁹ EVs) per well. Total mRNA extraction was conducted following the manufacturer's guidelines (17200; Norgen Biotek Corp.). The concentration and quality of mRNA were evaluated using a BioDrop (Harvard Bioscience, Hollistan, MA, USA), ensuring that the mRNA concentration exceeded the minimum requirement for chip analysis (130–1000 ng/mL).

Data Processing and Analysis

The miRNA (six controls and five OS, NPCE) and mRNA (two controls and three OS, HTM) raw CEL files were normalized, filtered, and analyzed using Transcriptome Analysis Console software (TAC 4.0.3). DEGs and DEmiRNAs between samples (treatment vs. control) were identified using the eBayes method in TAC. The normalized signals, along with fold change and P values were exported from TAC for further analysis. Data files were deposited to the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) under accession number GSE275895. Altogether, 19,959 gene and 2578 miRNA entries were examined (Supplementary Material C). The distances between samples were calculated using the R package Rtsne 0.17 (R Foundation for Statistical Computing, Vienna, Austria) and visualized using ggplot2 3.4.4.

DEGs and DEmiRNAs were defined by fold change ≥ 2 (up) or fold change ≤ –2 (down) and P < 0.05. Volcano plots depicting gene and miRNA fold changes and P values were generated with ggplot2. ClusterProfiler 4.10.1³¹ was employed for Gene Ontology (GO) enrichment analysis of up/down DEGs (parameters: pAdjustMethod = BH, pvalueCutoff = 0.05, qvalueCutoff = 0.1). Then, the identified GO terms were subjected to clustering analysis and heatmap visualization with simplifyEnrichment 1.12.0³² to cluster enriched GO terms based on semantic similarity (method = binary cut; parameters: measure = Rel, remove_orphan_terms = FALSE).

Target genes of DEmiRNAs were collected from TarBase 9.0,³³ a comprehensive database of experimentally validated miRNA–target interactions obtained through various experimental approaches (filters: experimental type = direct, gene location = 3′UTR) and processed crosslinking, ligation, and sequencing of hybrids (CLASH)-like datasets.³⁴ The latter source provides bonafide (with direct evidence) miRNA–target interactions within the 3′UTR region, derived from four datasets generated from various human cell types. Our analysis specifically focused on the 3′UTR region, as it is widely considered to contain the primary functional sites for miRNA regulation. The network of DEGs, DEmiRNAs, and GO terms was visualized using networkD3 0.4.

In this study, we investigated the influence of acute OS (15 mM AAPH for 1 hour) on NPCE cells in parallel. Significant changes were observed in the miRNA content of NPCE EVs, indicating a notable impact.³⁵ However, the effects on HTM mRNA were relatively minor, so they are not detailed in the main paper; and the raw data are available at GEO (GSE275895).

Results

t-Distributed Stochastic Neighbor Embedding Analysis of HTM mRNA and NPCE EVs miRNA Profiles Under OS and Control Treatments

NPCE-derived EVs were subjected to OS to induce potential changes in miRNA expression, with untreated NPCE-derived EVs used as controls for comparison. HTM cells were subsequently treated with these NPCE-derived EVs, and their mRNA expression profiles were analyzed to assess the impact of the EV treatments under OS conditions. Two separate t-distributed stochastic neighbor embedding (t-SNE) analyses were performed on the microarray results, covering both OS and control treatments: one to compare the mRNA expression profiles of HTM genes (Fig. 1A), and another to compare the miRNA expression profiles from NPCE EVs (Fig. 1B). The HTM mRNA analysis revealed two distinct clusters: control samples were grouped while OS-treated samples formed a separate cluster. Similarly, the NPCE EV miRNA analysis showed clear segregation between the control samples and the OS samples. The t-SNE analysis yielded a clear visual representation, offering valuable insights into the relationships within our dataset.

Figure 1.

Results of t-SNE microarray analysis of HTM mRNA and NPCE EV miRNA samples, showing profiles of HTM mRNA (A) and NPCE EVs miRNA (B) under both OS and control (CTL) treatments.

Differential Gene and miRNA Expression Analysis

We identified DEGs in HTM cells post-treatment with OS-NPCE–derived EVs compared to NPCE EVs. In HTM-treated samples versus controls (Fig. 2Ai), we identified 88 upregulated and 58 downregulated mRNAs. In OS-NPCE EVs versus controls (Figs. 2Aii, 2B), 27 upregulated and 27 downregulated miRNAs were identified.

Figure 2.

Differential expression analysis of miRNAs in OS-NPCE EVs and mRNAs in HTM cells treated with OS-NPCE EVs. (A) The volcano plot illustrates changes in mRNA expression in HTM cells following treatment with OS-NPCE EVs compared to NPCE EVs (i), and changes in miRNA abundance in OS-NPCE EVs versus control treatments (ii). Blue dots denote significantly upregulated expression (fold change ≥ 2, P < 0.05), red dots signify significantly downregulated expression (fold change ≤ –2, P ≤ 0.05), and gray dots indicate no significant difference. The top 10 genes/miRNAs with the highest fold changes among up- and down regulated sets are highlighted. The total numbers of examined miRNAs and genes were 2578 and 19,959, respectively. (B) Information about the fold change, P value, and number of experimentally validated targets according to TarBase 9.0 and processed CLASH datasets (see Methods) for downregulated DEmiRNAs (i) and upregulated DEmiRNAs (ii).

Insights From Clustering Analysis of Enriched GO Terms for Upregulated and Downregulated DEGs

We identified enriched biological process (BP), molecular function (MF), and cellular component (CC) GO terms for both up- and downregulated DEGs using ClusterProfiler.³¹ These terms were then subjected to clustering analysis and heatmap visualization with simplifyEnrichment³² based on semantic similarity. As a result, we identified 17, 13, and five clusters for upregulated BP, MF, and CC terms and five, two, and three clusters for downregulated BP, MF, and CC terms, respectively (Fig. 3). To enhance clarity, the term with the highest count of DEGs within each cluster was selected to represent the cluster, as indicated on the heatmap y-axis. This selection process aimed to identify pathways with the most pronounced changes in gene expression, providing key insights into the prominent molecular events associated with each cluster. These terms, designated as dominant, were used in further analysis described below.

Figure 3.

Clustering analysis of enriched GO terms for up- or downregulated DEGs. (A) Upregulated DEG BP (17 clusters). (B) Upregulated DEG MF (13 clusters). (C) Upregulated DEGs CC (five clusters). (D) Downregulated DEG BP (five clusters). (E) Downregulated DEGs MF (two clusters). (F) Downregulated DEGs CC (three clusters). Heatmap visualization of GO terms clustering, with each matrix panel representing visually similarities between pathways. Darker colors indicate stronger similarity. For each cluster, a representative term with the largest number of genes is displayed on the right, provided it contains at least five genes. These terms are designated as dominant in what follows. The number of DEGs in each term is indicated in brackets. Clustering results, where each cluster is annotated with word clouds highlighting the most common keywords for each cluster, can be found in Supplementary Material D. Full clustering information is available in Supplementary Material E.

Figure 3.

Continued.

The clustering analysis of enriched GO terms for upregulated DEGs revealed several notable BP, MF, and CC terms of particular interest in contexts involving OS or the glaucoma drainage system. Noteworthy among the upregulated BP terms are epithelial cell proliferation, ECM organization crucial for tissue integrity under stress, wound healing critical for repair mechanisms, and epithelial cell migration important in cellular response. Additionally, regulatory processes such as positive regulation of cell adhesion, transferase activity, cytokine production, and various metabolic processes including lipid metabolism are indicated (Fig. 3A). In terms of MF, the upregulated DEGs exhibit enrichments in ECM structural constituents, growth factors, integrin binding crucial in cellular adhesion, and DNA-binding transcription factors pivotal in stress response regulation (Fig. 3B). The enrichment in collagen-containing ECM and endoplasmic reticulum lumen among CC terms underscores their role in maintaining structural integrity and cellular homeostasis under stress conditions (Fig. 3C).

GO analysis of downregulated genes identified significant impacts on BP, MF, and CC terms within the context of OS and EVs miRNA detection. Downregulated BP terms such as oxidative phosphorylation, adenosine triphosphate (ATP) biosynthetic process, and proton transmembrane transporter indicate reduced cellular energy production and transport, potentially reflecting cellular adaptation to OS (Fig. 3D). In MF, reductions in the structural composition of ribosomes and proton transmembrane transporter activity suggest compromised protein synthesis and ion transport, which may influence miRNA processing in extracellular vesicles (Fig. 3E). CC analysis showed decreases in the mitochondrial inner membrane, ribosomes, and mitochondrial matrix, highlighting diminished mitochondrial function, critical in OS responses and miRNA trafficking (Fig. 3F). Our GO analysis revealed significant changes in gene expression, including both upregulation and downregulation of genes involved in ECM organization, cell adhesion, ECM structural components, integrin binding, collagen-containing ECM, and focal adhesion. These alterations, along with changes in other GO categories, point to a specific cellular response to OS and EV-mediated miRNA regulation. The targeted downregulation observed under certain experimental conditions highlights the nuanced nature of these responses, providing deeper insights into how cells adapt to OS and extracellular vesicle–mediated miRNA signaling

DEmiRNAs Target Analysis

To explore the relationship between DEmiRNAs and DEGs, we examined the experimentally validated target genes of upregulated and downregulated DEmiRNAs. The miRNA–target interactions were collected from TarBase³³ and additional datasets (see Methods). We identified 8923 target genes associated with 27 downregulated DEmiRNAs. Notably, as shown in Figure 2B, despite an equal number of miRNAs, downregulated DEmiRNAs have a significantly higher number of experimentally identified targets compared to upregulated DEmiRNAs. This disparity facilitates a deeper exploration of the axis between downregulated DEmiRNAs and upregulated DEGs, but not vice versa. Among the 88 upregulated genes in HTM cells, 59 were found to be targeted by downregulated DEmiRNAs, suggesting potential regulatory interactions. In contrast, our analysis of 27 upregulated DEmiRNAs revealed 2158 target genes, with only five of them overlapping with downregulated DEGs (Table).

Table.

Comparison of Target Genes of DEmiRNAs With DEGs in HTM Cells After OS-NPCE EV Treatment

DEmiRNAs		Target Genes of miRNAs		DEGs		miRNA-Targeted DEGs
Upregulated	Downregulated	Upregulated	Downregulated	Upregulated	Downregulated	Upregulated	Downregulated
27	27	2158	8923	88	58	59	5

Analysis of DEmiRNA–Target Genes in Dominant GO Categories

We explored the interplay among DEmiRNAs, DEGs, and GO terms. To that end, we retrieved dominant GO terms identified within clusters, as shown in Figure 3. We identified DEGs targeted by miRNAs with opposing regulation for each dominant GO term. In GO terms enriched with upregulated DEGs, many genes were targeted by downregulated miRNAs (Fig. 4A). Consistent with the low number of target genes of upregulated miRNAs reported earlier (Fig. 2B; Table), a low proportion of target genes was observed for downregulated DEGs within GO terms (Fig. 4B).

Figure 4.

Target genes of DEmiRNAs in dominant GO categories. (A) Proportion of genes targeted by downregulated DEmiRNAs among upregulated DEGs in dominant GO terms. (B) Proportion of genes targeted by upregulated DEmiRNAs among downregulated DEGs in dominant GO terms. The most dominant upregulated (A) and downregulated (B) terms were identified utilizing a selection criterion focused on the highest counts of DEGs within specific clusters identified in Figure 3. The black portion of each bar represents DEGs targeted by differentially regulated miRNAs (in the opposite direction), and those in gray indicate non-target DEGs. Terms are ordered as in the heatmaps in Figure 3.

Characterization of miRNA–Target Gene Interactions and Their Impact on Glaucoma-Related BP, MF, and CC GO Analyses

Our analysis concentrated on miRNA and mRNA molecules exhibiting significant expression changes to explore their underlying relationships. We constructed networks using Sankey diagrams to visualize the interconnections between these molecules. Additionally, we associated relevant GO terms with the identified mRNA molecules to provide functional context. We focused on downregulated miRNAs and upregulated genes while excluding upregulated miRNAs and downregulated mRNAs due to limited interaction data for the latter. Sankey graphs illustrating these connections in HTM cells are shown in Figure 5 and are organized by BP- (Fig. 5A), MF- (Fig. 5B), and CC-related (Fig. 5C) GO terms. Within these networks, some miRNAs target a few specific genes, whereas others act as master regulators, influencing a broader range of targets. Similarly, although some mRNAs are regulated by a single miRNA, many are influenced by multiple miRNAs, demonstrating a complex network of one-to-many and many-to-one connections.

Figure 5.

Exploring networks with OS-NPCE EV DEmiRNAs, HTM DEGs, and GO terms. (A) Downregulated DEmiRNAs, upregulated DEGs, and BP terms. (B) Downregulated DEmiRNAs, upregulated DEGs, and MF terms. (C) Downregulated DEmiRNAs, upregulated DEGs, and CC terms. The Sankey diagram visualizes the miRNA–mRNA–GO term network in HTM cells. Left rectangles depict miRNAs, and middle rectangles represent genes. The right rectangles illustrate the corresponding GO terms: BP (A), MF (B), and CC (C). Rectangle sizes signify the number of connections.

Figure 5A reveals complex regulatory interactions related to biological processes relevant to POAG and OS. A key finding is the miRNA–gene interactions involving the downregulation of hsa-miR-19b-3p, which correlates with the upregulation of 29 target genes, including hyaluronan synthase 2 (HAS2) and bone morphogenetic protein 2 (BMP2). These genes play crucial roles in ECM dynamics within TM cells. HAS2 influences outflow resistance through hyaluronic acid production, a major ECM component.³⁶ The diagram in Figure 5A shows the connection of HAS2 to wound healing processes, suggesting its role in tissue repair and remodeling within the TM. BMP2, a transforming growth factor beta (TGFβ) family member, is linked to ossification and positive regulation of transferase activity in the Sankey diagram. This connection highlights the potential role of BMP2 in promoting ECM calcification, potentially disrupting AH outflow and contributing to TM dysfunction.^37,38 The Sankey diagram illustrates heparin-binding EGF-like growth factor (HBEGF) involvement in epithelial cell proliferation and migration pathways, as well as positive regulation of cell adhesion. These connections suggest that HBEGF promotes ECM gene expression, indicating a potential mechanism for ECM alterations in glaucomatous conditions.³⁹

Figure 5B, which focuses on MF, shows an overlap in the distribution of miRNAs and DEGs across both MF and BP categories. Distinct differences in the specific GO terms associated with these categories, particularly highlighting pathways related to collagen-containing ECM, endoplasmic reticulum lumen, and focal adhesion were observed. Notably, hsa-miR-19b-3p emerged as a central regulator, showing upregulation of all 26 TM genes presented in the analysis, suggesting a potentially crucial role for this miRNA in TM function and glaucoma pathogenesis. Key upregulated genes include insulin-like growth factor 2 (IGF2), previously reported to have functional receptors in cultured human TM cells.⁴⁰ The involvement of IGF2 in the endoplasmic reticulum lumen pathway suggests its role in protein synthesis and processing within TM cells. ADAM22 (a disintegrin and metalloproteinase domain 22) suggests potential involvement in ECM remodeling, particularly in the collagen-containing ECM, indicating its significance in modulating the structural components of the outflow pathway. Also, integrin subunit alpha 6 (ITGA6), expressed in putative Schlemm's canal endothelial cells,⁴¹ is crucial in focal adhesion pathways. The upregulation of genes involved in integrin binding, including ITGA6, underscores the significance of integrin-mediated processes in TM function. The upregulation of genes involved in integrin binding, including ITGA6, underscores the significance of integrin-mediated processes in TM function, aligning with recent research highlighting the role of integrin crosstalk in modulating the biological functions of the TM complex.^42,43 These findings collectively suggest a complex regulatory network involving miRNAs, particularly hsa-miR-19b-3p, and various genes crucial for TM function. The upregulation of genes related to growth factor signaling (IGF2), ECM interactions (ADAM22), and cell adhesion (ITGA6) indicates potential mechanisms by which miRNA dysregulation could contribute to altered AH dynamics and, consequently, glaucoma pathogenesis, providing new insights into molecular functions affected in glaucoma and identifying potential targets for therapeutic interventions.

Finally, in Figure 5C, which focuses on CC, the Sankey data reveal a complex network of interactions among miRNAs, genes, and cellular components in glaucoma-related processes, centered around three primary GO terms: collagen-containing ECM, endoplasmic reticulum lumen, and focal adhesion. Collagen-containing ECM, the most represented category, features genes such as laminin alpha 2 chain (LAMC2), Dermatopontin (DPT), Tenascin C (TNC), Lumican (LUM), chondroitin sulfate proteoglycan 4 (CSPG4), and cadherin 13 (CDH13), regulated by miRNAs such as hsa-miR-19b-3p. The endoplasmic reticulum lumen category includes TNC, prostaglandin-endoperoxide synthase 2 (PTGS2), insulin-like growth factor binding protein 5 (IGFBP5), also regulated by hsa-miR-19b-3p. Focal adhesion involves genes such as CSPG4, CDH13, ITGA6, Fms-like tyrosine kinase 1 (FLT1), and ITGB3, with diverse miRNA regulation, including hsa-miR-30d-5p and hsa-miR-19b-3p. Several genes, such as TNC, CSPG4, and CDH13, are involved in multiple biological processes, highlighting the interconnected nature of these pathways. Notably, hsa-miR-19b-3p emerges as a key regulator targeting genes across all three GO categories, emphasizing its significance in glaucoma pathogenesis.

Discussion

POAG is a severe eye condition leading to vision loss due to retinal and optic nerve damage. It arises from imbalanced IOP influenced by AH dynamics, particularly involving the TM and its ECM. OS exacerbates POAG, contributing to tissue damage and TM dysfunction. EVs, carrying miRNAs, play a role in cellular communication and offer the potential to mitigate OS-induced damage, particularly from NPCE cells near the TM.

Previous research has highlighted differential gene expression in HTM cells under various conditions, including POAG,⁴⁴ steroid treatment,^45,46 TNFα and IL-1 exposure,^47,48 mechanical stretching,^49,50 and OS.⁵¹ In particular, HTM-derived EVs have been studied as carriers of miRNAs.^52,53 Additionally, investigations have focused on analyzing miRNA content in AH EVs from patients with cataracts^54,55 and myopia.⁵⁶ The AH encompasses EVs originating from various tissues within the ocular drainage system, possibly originating from stromal cells, limbal epithelial/stem cells, ciliary epithelium, and corneal endothelium.⁵⁷ Given the significant role of the ciliary epithelium in continuous AH production and release, our focus has centered on NPCE EVs. However, it is essential to acknowledge that the contribution of EVs from other sources within the AH should not be overlooked and merits further research.

Our lab's prior work has shown the potential of NPCE EVs to modulate the TM ECM structure and composition.⁵⁸ Recently, our study contributed to this body of research by conducting microarray analysis to examine miRNA content in NPCE-derived EVs, comparing the effects of OS on miRNA profiles in NPCE cells and their corresponding EVs.³⁵ Our present study addressed a critical knowledge gap by elucidating the miRNA profiles within NPCE-derived EVs under OS conditions and their impact on TM cells. This investigation is crucial, as the role of these specific miRNAs in glaucoma pathogenesis remains poorly understood. We opted for the miRNA/mRNA microarray assay due to its ease of standardization, which allows for more reliable comparisons across different diseases and studies. Although this method meets our current needs, future studies could incorporate next-generation sequencing for more comprehensive insights into miRNA/mRNA expression and function.

Data Visualization and Insights From t-SNE Analysis

Overall, the t-SNE analysis of both miRNA and mRNA samples provided a comprehensive and visually insightful representation of the dataset relationships. It allowed for the identification of clear patterns and separations between samples based on their treatment conditions, enhancing our understanding of the molecular responses of HTM cells and NPCE EVs to control and OS treatments.

Differential Expression Analysis of miRNAs in OS-NPCE EVs and mRNAs in HTM Cells Treated With OS-NPCE EVs

The differential expression analysis revealed significant changes in mRNA expression in HTM cells and miRNA abundance in OS-NPCE EVs, underscoring the complex molecular interactions in OS responses. We identified 88 upregulated and 58 downregulated mRNAs in HTM cells treated with OS-NPCE EVs, along with 27 upregulated and 27 downregulated miRNAs in the EVs, indicating a specific transcriptional and post-transcriptional response to OS. This aligns with prior research by Wang et al.⁵⁹ showing that OS alters gene expression profiles in TM cells, particularly affecting ECM remodeling and cellular senescence. The balanced changes in miRNA profiles within OS-NPCE EVs are noteworthy, especially because previous studies have reported a bias toward miRNA upregulation under OS conditions.⁶⁰ This discrepancy may stem from tissue-specific responses or differences in OS models. Our volcano plots and miRNA target data offer valuable insights into the most affected molecular players, building on our earlier work on exosome-mediated signaling and OS conditions in the TM.^20,58

Although previous studies indicated that OS-induced EVs from retinal cells could modulate gene expression in recipient cells,⁶¹ our findings extend this concept to the TM, a critical tissue in glaucoma pathogenesis. Collectively, these results suggest a complex interplay between OS-NPCE EVs and HTM cells, driven by direct mRNA changes and altered miRNA content, supporting the idea that OS-induced changes in the TM contribute to glaucoma. However, our study differs from reports that found more pronounced effects on pro-inflammatory gene expression in HTM cells under OS,⁶² which may be attributed to our focus on EV-mediated effects rather than direct oxidative insults.

Insights From Clustering Analysis of Enriched GO Terms for Upregulated and Downregulated DEGs

Our analysis of DEGs in response to OS and the glaucoma drainage system revealed key insights through enriched GO terms. Using ClusterProfiler and simplifyEnrichment, we identified distinct patterns in BP, MF, and CC associated with upregulated and downregulated DEGs, highlighting complex molecular interactions. The clustering of upregulated BP terms—such as epithelial cell proliferation, ECM organization,⁶³ and wound healing—aligns with prior research emphasizing their roles in ocular health and tissue repair.⁶⁴ Similarly, the enrichment of genes related to ECM constituents and integrin binding reflects adaptive mechanisms documented in other studies that activate stress-responsive pathways.³ Conversely, downregulated DEGs showed a marked impact on energy metabolism and protein synthesis, with reductions in oxidative phosphorylation and ATP biosynthesis suggesting a conserved strategy during OS. This finding is consistent with existing literature that indicates a shift in cellular resource allocation under stress.⁶⁵ Overall, our findings illustrate the intricate interplay between DEGs in OS, providing a foundation for further investigation into the molecular events influencing glaucoma and related conditions.

Analysis of DEmiRNAs Target Genes in Dominant GO Categories: A Novel Regulatory Perspective

Our integrated analysis of DEmiRNAs, DEGs, and GO terms revealed intriguing patterns in the regulatory landscape of OS-induced changes. Interestingly, we found a limited number of downregulated DEGs targeted by upregulated miRNAs within GO terms, consistent with our earlier observation of reduced target gene numbers for upregulated miRNAs (Fig. 2B; Table). In the future, as miRNA–target interaction datasets expand, it may be possible to investigate this axis, as well.

Key miRNAs in Glaucoma Pathogenesis and Trabecular Meshwork Regulation

In the present research, several miRNAs have been identified as key regulators in TM cells, playing crucial roles in glaucoma pathogenesis. For example, hsa-miR-27b-3p regulates TGFβ signaling in TM cells, a pathway known to be dysregulated in glaucoma.⁵¹ hsa-miR-143-3p is involved in the regulation of ECM production in TM cells, essential for maintaining proper intraocular pressure.⁶⁶ Similarly, hsa-miR-199a-3p and hsa-miR-199b-3p have been associated with the regulation of ECM genes in TM cells.⁶⁷

Positive and Negative Effects of miRNAs on TM Cell Function and POAG

miRNAs play diverse roles in regulating HTM cell function and POAG pathogenesis, with both positive and negative effects. On the positive side, miRNAs such as miR-29b help maintain TM homeostasis by regulating ECM proteins such as collagen and laminin, potentially preserving outflow facility.⁵¹ miR-200c has been linked to lower IOP, suggesting a beneficial role in POAG management.⁶⁸ Conversely, downregulation of miR-29b can lead to excessive ECM deposition, increasing outflow resistance.⁶⁶ miR-183 promotes HTM cell apoptosis by targeting integrin β1,⁶⁹ and miR-153 exacerbates OS by targeting Nrf2.⁷⁰ miR-155 upregulation promotes inflammation in TM cells.⁷¹ These findings highlight the complex role of miRNAs in POAG, influencing IOP regulation, TM dysfunction, and retinal ganglion cell loss. Further research is needed to fully understand their roles and explore their potential as therapeutic targets.

Other miRNAs are known to be involved in POAG but may not be part of EV-mediated signaling under OS according to current research. For example, hsa-miR-24-3p is frequently discussed alongside hsa-miR-27b and hsa-miR-23b in the context of glaucoma.⁷² Similarly, hsa-miR-29b is notable for its role in ECM regulation, a key factor in POAG pathogenesis.^51,66 These miRNAs, although significant in POAG research, might not be directly involved in the EV-mediated signaling pathways influenced by OS being investigated in our current study.

Additionally, several miRNAs are relatively new to this field or have limited evidence for direct involvement in POAG, TM function, or OS, including hsa-miR-1246, hsa-miR-132-3p, hsa-miR-195-5p, hsa-miR-19b-3p, hsa-miR-224-5p, hsa-miR-30d-5p, hsa-miR-324-5p, hsa-miR-335-5p, hsa-miR-4496, and hsa-miR-99b-5p. These miRNAs represent potential new avenues for research in glaucoma pathogenesis and TM function, as their roles in these contexts are not yet well defined or extensively studied.

DEGs and Their Role in ECM Regulation

The identified DEGs in HTM cells, including potassium inwardly rectifying channel subfamily J member 2 (KCNJ2), protocadherin 9 (PCDH9), and secreted frizzled-related protein 1 (SFRP1), exhibit significant downregulation and potential influence on cellular functions intersecting with ECM-related processes. It is speculated that KCNJ2 downregulation may interfere with ion channel activity, potentially affecting the regulation of TM cell volume, which could influence AH outflow. Similarly, a decrease in PCDH9 expression might impair cell–cell adhesion within the TM, possibly compromising tissue integrity. Although these are plausible hypotheses based on the known functions of these genes, further experimental evidence is needed to confirm their roles in TM physiology.⁷³ Although direct evidence linking SFRP1 downregulation to TM dysfunction remains speculative, its established role, as Wnt signalinginhibitor supports the idea that altered SFRP1 signaling could influence intraocular pressure and potentially increase outflow resistance in the TM. These changes could collectively lead to alterations in ECM production and turnover, TM stiffness, and outflow facility, all of which are critical factors in glaucoma pathogenesis.

In contrast, upregulated HTM genes, such as BMP2, BMP6, PLAT, SULF2,^50,74 DPT,^74,75 IGF2,⁷⁶ TNC,^77,78 and LUM,⁷⁹ are prominently associated with ECM regulation and CD34,⁸⁰ ITGA6, and ITGB3⁸¹ to cell adhesion. Furthermore, the intricacies of EV packaging and releasing processes add complexity to comparing experiments, with current miRNA findings partially overlapping with reported results. Our analysis, employing heatmap visualization and clustering, underscores crucial pathways influenced by our experimental conditions, particularly emphasizing ECM organization, collagen cell adhesion, cellular response to OS, mitogen-activated protein kinase (MAPK) signaling, and glycosaminoglycan (GAG) binding. This enrichment of upregulated genes in transmembrane receptor protein tyrosine kinase signaling and ECM components aligns with our study's focus on the central roles of ECM and adhesion processes. The intricate interplay between differentially expressed miRNAs and their target genes underscores complex regulatory mechanisms, emphasizing the crucial role of miRNA-mediated gene expression in maintaining cellular homeostasis.

Our findings regarding the impact of OS-NPCE EVs on HTM mRNA expression and associated pathway influences, as depicted in Figure 5, underscore the significance of EV-mediated miRNA signaling in modulating the ECM. The ECM, a crucial component of IOP homeostasis and stress response, is intricately linked to the regulation of various biological processes. The representation in Figure 5 highlights the pivotal role of EVs in orchestrating miRNA-mediated signaling, emphasizing their contribution to ECM modulation and, consequently, their influence on maintaining IOP equilibrium and responding to stress. In line with these observations, we have previously examined key signaling pathways, such as glycogen synthase kinase-3 beta (GSK3β) and β-catenin, which are regulated by the Wnt signaling pathway in NPCE EVs under OS.²⁰ This work, published by Lerner et al. in 2020,²⁰ further reinforces the importance of miRNA-driven EV signaling in ECM regulation, a process essential for both IOP homeostasis and the cellular stress response.

Potential for Therapeutic Applications of NPCE EVs in Glaucoma Treatment

Although the current research focuses on understanding the cargo of NPCE EVs, which could offer therapeutic insights, this work is primarily basic rather than translational research. However, NPCE EVs hold promise in uncovering novel mechanisms that could be harnessed for future therapeutics. By elucidating their beneficial cargo—such as miRNAs, proteins, or lipids—targeting cellular processes relevant to glaucoma, we can gain a better understanding of their potential therapeutic effects. For example, NPCE EVs may carry molecules that modulate IOP through interactions with TM cells or by influencing ECM remodeling. In the future, if these cargo components are better understood, they could be selectively modified or enhanced to help restore IOP homeostasis. However, it is worth considering the practical challenges that come with using NPCE-derived EVs as a therapeutic approach. The production, isolation, and potential scaling for therapeutic use present difficulties. Mesenchymal stem cell–derived EVs could offer an alternative platform that is more viable for therapeutic development. MSCs are easier to culture and scale, and their EVs can be engineered to contain specific therapeutic cargos or surface molecules that enhance their efficacy.

In summary, our study delved into the impact of OS on TM cells and the potential therapeutic role of EVs derived from NPCE cells. Through microarray analysis, we explored miRNA content in NPCE-derived EVs and its influence on TM cells. The results highlight differential gene expression in TM cells, emphasizing the regulatory role of microRNAs in maintaining cellular homeostasis. These findings prioritize validated gene expression changes and underscore the significance of EV-mediated miRNA signaling in modulating the ECM, essential for IOP homeostasis and stress.